Report No. 39303-PK Pakistan Promoting Rural Growth and Poverty Reduction March 30, 2007 Sustainable and Development Unit South Asia Region Document of the World Bank Table of Contents ABBREVIATIONSANDACRONYMS................................................................................................. ..I1 ACKNOWLEDGEMENTS..................................................................................................................... IX EXECUTIVE SUMMARY........................................ .................................................................................. X BACKGROUND ............................................................................................................................................ x TOWARDSSTRATEGYFORRURALGROWTH POVERTYREDUCTION............................................. A AND XIV PROMOTINGEFFICIENTAND SUSTAINABLEAGRICULTURALGROWTH ................................................... X N Agricultural Production and Technology ................................ xiv Increasing Efficiency of Water Use..................................................................................................... xv Agricultural Markets and Trade Policies .............. xvi CREATINGAN ENABLINGENVIRONMENT SERVICE DELIVERY................................................................................................................................ FORTHERU XVII Enhancing Access to Credit......................................... ................. xvii IMPROVINGTHE EFFECTIVENESSAND GOVERNANCE RURALINSTITUTIONS OF ....................................... XIX EMPOWERING .................................................... CONCLUSIONS....................................................................................................................................... THEPOORAND PROTECTINGTHE MOST VULNERABLE XXII XXIV INTRODUCTION ............................................................................................................................... 1 PLAN OF THEREPORT .................................................................................................................................. 3 ANNEXTO CHAPTER 1:POPULATIONGROWTH PAKISTAN...................................................................... IN 4 Urban and Rural Population Growth.................................................................................................... 5 RURAL POVERTY INPAKISTAN....... ................................................................................... ..,.....8 POVERTYTRENDS....................................................................................................................................... 8 CHANGESINREAL INCOMES ..................................................................... ACROSS HOUSEHOLD GROUPS 10 LONG-RUN DETERMINANTSPOVERTYTRANSITIONS OF ........................................................................... 17 NON-MONETARY INDICATORS OF WELFARE AND POVERTY ..................................................................... 20 SUMMARY ................................................................................................................................................. 24 PROMOTINGAGRICULTURAL GROWTH .............................................................................. 38 AGRICULTURALPRODUCTIONAND TECHNOLOGY .................................................................................... 38 The Crop Sub-Sector............................................................................................................................ 41 Livestock and 45 FACTORS PRO OF .......................................................................... .49 Land Distribution and Productivity .................................................................................................... .49 Water Availability and Efficiency of Use. 52 Other Agricultural Inputs ................................................................................................................... .58 MARKETS TRADEPOLICIES............................................................................................................... 59 AND WheatPolicy ........................................................................................................................................ 63 AGRICULTURALGROWTH LINKAGES RURALPOVERTYREDUCTION................................................. AND 66 POLICY IMPLICATIONS .............................................................................................................................. 70 GROWTHINTHE RURAL NON-FARMECONOMY .............................................................. .71 STRUCTUREOF RURALNON-FARM ENTERPRISES Non-farm enterprises' contribution to income .............................. ,I6 .................... Technology and Business Services .... Security and Rule oflaw ..................................................................................................................... 87 ... 111 Rural Enterprise Ownership: Start-up and Performance ................................................................... 88 POLICY IMPLICATIONS .............................................................................................................................. 90 ANNEXTO CHAPTER4: THEPAKISTANRURALINVESTMENTCLIMATE SURVEY, 2005............................. 91 5 PUBLICEXPENDITURES AND RURAL SERVICE DELIVERY . ................................................ 103 BACKGROUND: DEVOLUTIONPAKISTAN............................................................................................. IN 103 FISCALOUTCOMES AND ISSUES .............................................................................................................. 104 ADMINISTRATIVE ISSUES CAPACITY............................................................................................................................................... ........................................................................................................................ 117 118 ACCOUNTABILITY................................................................................................................................... 119 SUMMARY OF ISSUES .............................................................................................................................. 120 POLICY IMPLICATIONS ............................................................................................................................ 122 6. SOCIAL MOBILIZATION AND ENHANCINGLIVELIHOODS ........................................... 124 MAJOR RURALDEVELOPMENT PROGRAMS PAKISTAN ........................................................................ IN 124 SOCIAL MOBILIZATION AND RURALLIVELIHOODS DEVELOPMENT ........................................................ 127 MICROFINANCE INTERVENTIONS ............................................................................................................. 134 SOCIAL PROTECTION ............................................................................................................................... 137 Heavily-indebted Laborers ................................................................................................................ 141 7. SUMMARY AND CONCLUSIONS .............................................................................................. 142 RURALPOVERTY: STRUCTUREAND TRENDS.......................................................................................... 142 TOWARDSSTRATEGY FORRURALGROWTH A AND POVERTY REDUCTION ............................................. 143 PROMOTINGEFFICIENTAGRICULTURAL GROWTH .................................................................................. 144 THERURALNON-FARM SECTORAND THERURALINVESTMENT CLIMATE .............................................. 152 IMPROVING PUBLIC SERVICE DELIVERY................................................................................................. 153 IMPROVING THE EFFECTIVENESS AND GOVERNANCE RURALINSTITUTIONS OF ...................................... 155 EMPOWERING POORAND PROTECTING THEMOST THE VULNERABLE ..................................................... 156 CONCLUSIONS ......................................................................................................................................... 157 REFERENCES .......................................................................................................................................... 159 PAKISTANAT A GLANCE .................................................................................................................... 162 iv List of Tables TABLE2.1: POVERTYESTIMATESPAKISTAN, 1998-99, 2001-02 AND 2004-05 ............................................ IN 9 TABLE2.2. RURALPOVERTY ACROSSHOUSEHOLD GROUPS PAKISTAN, 2004-05 ............... IN TABLE2.3. SOURCESOFINCOMEBY RURALHOUSEHOLD GROUPINPAKISTAN, 2004-05 ....... TABLE2.4. REALPERCAPITA EXPENDITURESRURALHOUSEHOLDS ..................................................................................... OF INPAKISTAN, 1998-99,2001-02 TABLE2.5. GROW RATESOFAGRICULTURALOUTPUTANDINCOMEINPAKISTAN, ....... AND 2004-05 .................................. 15 TABLE2.6. REALWAGERATES OFUNSKILLEDLABORPAKISTAN, 1998-99 TO 20 IN TABLE2.7. REALINCOMES INFOURRURAL DISTRICTS PAKISTANBY SOURCE, 1986-87 TO 1991-92 IN ............................................................................................................................ 18 TABLE2.8. POVERTYTRANSITIONS RURALPAKISTAN: 1987-91 TO 2002 .......................... AND 2001-02 IN TABLE2.9: SOCIALINDICATORS OFRURALHOUSEHOLDS INPAKISTAN, 1988-99,2001-02 AND 2004-05 ....21 TABLE2.10: EDUCATIONAND LITERACY RATES INPAKISTAN, 1998-99,2004-05 ................. TABLE2.11: SOCIALINDICATORS INSOUTH ASIA .................................................................. TABLE3.1. AGRICULTURAL GDPGROWTH PAKISTAN, 1999-2000 TO 2005-06* ...................................... IN 39 TABLE3.2. ESTIMATES TOTALFACTOR OF PRODUCTIVITY GROWTH AGRICULTURE PUNJAB, OF IN PAKISTAN ................................................................................................................................. 40 TABLE3.3. PAKISTAN:AREA,YIELDANDPRODUCTIONOFMAJOR CROPS, 2004-06 (AVERAGE). .................42 TABLE3.4. AVERAGE CROPYIELDS INPAKISTANAND OTHER COUNTRIES, 1997-98 TO 2001-02 ...... TABLE3.5. COSTSANDRETURNSFORCOOPERATIVEDAIRYPRODUCTION, PUNJAB2005 ............................ 49 TABLE3.6. DISTRIBUTIONOF LAND OWNERSHIP INPAKISTAN .......................................................... TABLE3.7. SEASONAL WATER FLOWS THEINDUSRIVERBASIN,1975-76 TO 2003-04.................. IN TABLE3.8. SOURCESOF CREDITINPAKISTAN, 2005-06 ................................................................................ TABLE3.9. NOMINAL RATES OF ASSISTANCEFORSELECTEDAGRICULTURALPRODUCTSINPAKISTAN .......58 60 TABLE3.10. FOODGRAINPOLICY INPAKISTAN, INDIAAND BANGLADESH, 2001-05 ................................... 64 TABLE3.11.RURALAGRICULTURAL INCOMES INPAKISTAN ......................................................................... 67 TABLE4.1. PERCENTAGEOFRURALHOUSEHOLDS INPAKISTANTHAT OWNSHOPSAND/OR OTHER BUSINESSES 71 TABLE4.2. PROPORTIONOF THERURALWORKINGPOPULATION INPAKISTAN....................... .............................................................................................................................. .72 TABLE4.3. ANNUALHOUSEHOLD NON-FARM PAKISTAN (IN2001-02 Rs) ....................................................................................................... INCOME SHARES BY EXPENDITUREQUINTILES 73 TABLE4.4. PROFILEOFNON-FARM ENTERPRISES .................................................................................... TABLE4.5. ENTERPRISE SALES AND ASSETS TABLE4.6: PAKISTAN:RURALAND SMALLTOWNFIRMS'ACCESS FINANCE,2005 . ('000 RS) ........................................................................... TO TABLE4.7. QUALITY OF, AND ACCESS INFRASTRUCTURE........................................ TO, TABLE5.1: REALFEDERAL PROVINCIALREVENUESAND EXPENDITURESPAKISTAN, 2001-02 TO AND 2005-06, (INBILLIONRS, 2004-2005 = 100)......................................... IN .......................... 105 TABLE5.2: DEVOLUTION PAKISTAN: REALEXPENDITURES IN AND TRANSFERS, 200 TO 2005-06 .......IO8 TABLE5.3: PUBLICEXPENDITURESRURALDEVELOPMENT SOCIAL SERVICESINSOUTH ASIA ON AND (US$/PERSON) ........................................................................................................................ 109 TABLE5.4: REALREVENUESAND EXPENDITURES INPUNJAB,PAKISTAN,2001-02 TO 2005-06 .................111 TABLE6.3: DISTRIBUTION HOUSEHOLDSINTHEMAJORSAFETYNETPROGRAMSINPAKISTAN, 20 OF POVERTYREDUCTION ................................................................ ......................... 146 V ListofAnnex Tables TABLEAI-1 ESTIMATES AGRICULTURAL OF GROWTH . TABLEAI-2 POPULATION AND FERTILITYRATES PAKISTAN ............................................... PERCAPITA INPAKISTAN, 1960-2004 IN TABLEA i - 3 BIRTHRATESINRURALAND URBANAREASINPAKISTAN,2001AND2002.............................. 6 TABLEA1-4 URBANAND RURALPOPULATIONGROWTH SCENARIOS ..................... ............................. 7 TABLEA2- 1.RURALHOUSEHOLDS'PERCAPITA EXPENDITURESPUNJAB, REAL IN 2001-02AND 2004-05 ..... TAN, 1998-99, ........................... 25 TABLEA2-2.RURALHOUSEHOLDS'PERCAPITA EXPENDITURESINSINDH, PAKISTAN, 1998-99,2001- REAL 02 AND 2004-05 ............. ............................................................................ ......26 ~ TABLEA2-3. RURALHOUSEHOLDS' REA A EXPENDITURESNWFP, PAKISTAN, IN 1998-99, 2001-02AND 2004-05 ........................................................................................................... 27 TABLEA2-4. RURALHOUSEHOLDS'PERCAPITA EXPENDITURES BALUCHISTAN, 1998-99, 2001-02AND 2004-05 ..................................................................................... REAL IN PAKISTAN, TMLE A2-5. WHEAT AND COTTONPRODUCTIONANDINCOMES INPAKISTAN, 1998-99,2001-02 AND TABLEA2-6. STIMATES OFAGRICULTURALOUTPUT ANDINCOMESINPAKISTA, 1998-99 TO 2005-06 .........30 2004-05 ................................................................................................................................ 29 TABLEA2-7. REALINCOMESINFOURDISTRICTS RURALPAKISTAN BY SOURCE, 1986-87T0 1991-92 IN AND 2002 ................................................................................................... TABLEA2-8. REALINCOMESOFPOORHOUSEHOLDS DISTRICTSINRURALPAKIS 1986-87 TO 1991-92AND 2002............................................................................................. INFOUR 32 TABLEA2-9.DETERMINANTSREALINCOMES ANDEXPENDITURES RURALPAKISTAN,REGRESSION OF IN RESULTS ............................................................................................................................... 33 TMLE A4.1 DISTRIBUTION FARMANDNON-FARM OF HOUSEHOLDS TABLEA4.2 BUSINESSREGISTRATION...................................... ....IN RURALPAKISTAN. ....................... 92 .................... ......,...............................92 TABLEA4.3 SALESREVENUES....................................................................................................................... 92 TABLEA4.4 PRECENTAGEOF FIRMS IDENTIFYINGVARIOUSINVESTMENTCLIMATEFACTORSAS CONSTRAMT ..........,...... .......................................... ..............................93 TABLEA4.5 MAJOR& SEVEREINVESTMENTCLIMATECONSTRAINTS IDENTIFIEDBY ENTERPRISES...............93 TABLEA4.6 ACCESS TO MARKETS ................................................................ ...,,,..................................93 TABLEA4.7 MARKETLINKAGESGEOGRAPHIC LOCATIONOF SALESAND TRA PARTNERS ....................... 95 TABLEA4.8 TECHNOLOGICAL IMPROVEMENTS/INNOVATIONS ............................................................ TABLEA4.9 PARTICIPATIONINBUSINESSASSOCIATIONS............................................................................... 96 TABLEA4.10 AVAILABILITY AND USEOFBUSINESSSERVICES....................................................................... 97 TABLEA4.11LAWS AND REGULATIONS ........................................................................................................ 98 TABLEA4.12 SOCIAL COHESION ................................................................................................................... 99 TABLEA4.13 ENTERPRISEOWNERSHIPAND START-UPPROBITREGRESSION .............................................. 100 TABLEA4.14 LABORPRODUCTIVITY REGRESSIONS ................................................................................... TABLEA5.1 REALEXPENDITURES REVENUESPUNJABPROVMCE, 2001-02 TO 2005-06.................102 AND IN 123 vi List of Figures FIGURE2.1. DISTRIBUTIONOFRURALHOUSEHOLDEXPENDITURES ................................................... PAKISTAN, 9 FIGURE2.2. REALPERCAPITA RURALHOUSEHOLD EXPENDITURES PAKISTAN, 1998-99 TO 2004-05......11 IN FIGURE2.3. REALPERCAPITA RURALPOORHOUSEHOLD EXPENDITURES: FIGURE2.4. PERCAPITA AGRICIJLTWRAL OUTPUTANDREALINCOMETO2005-06..,.,,..,.. 1998-99 TO 2004-05 .................11 INPAKISTAN,1998-99 FIGURE.2.5. CGEMODEL SIMULATIONSOFSHOCKSTOKEYDETERMINANTS OFRURALINCOMEINPAKISTAN.. . FIGURE2.6. RURALPOVERTYTRANSITIONSINPAKISTAN, 1987-88 TO 2001-02........................................... 19 FIGURE2.7.PERCAPITA GDPAND SOCIALINDICATORS INSOUTHASIA.................................. FIGURE3.1. REALAGRICULTURALGDP INPAKISTAN, 1989-90TO 2004-05 ................................................ 40 FIGURE3.2. ANNUALINDUSFLOW,DIVERSION AND ESCAPAGE SEASON, 1975-76 TO 2003-04 BY FIGURE3.3. GROWTH INSTITUTIONALAGRICULTURALCREDITINPAKISTAN(RS2004-05 BN) OF FIGURE3.4. NOMINAL .........60 FIGURE.3.5. PAKISTAN AND INTERNATIONALWHEAT PRICES, 2000-2006 .................................................... RATES OF ASSISTANCE FORMAJOR CROPS INPAKISTAN, 1961-62 TO 2004-05 61 FIGURE3.6. PAKISTAN: REAL PRICESOF SEEDCOTTON, 1980-81 TO 2004-05 .............................................. 62 FIGURE3.7. SIMULATION RESULTS OFA 10-PERCENTOUTPUTSHOCKTO WHEAT, ALLMAJOR CROPS AND LIVESTOCKPAKISTAN .................................................................................. IN FIGURE3.8. SIMULATION RESULTS OFA 10-PERCENTOUTPUT SHOCKTO WHEAT, ALLMAJOR CROPS AND LIVESTOCK INPUNJAB,PAKISTAN .......................................................................................... 69 FIGURE4.1. CLIMATECONSTRAINTSFORSMALL-TOWNAND RURALENTERPRISES, ............................ RURALAND SMALLTOWNINVESTMENTCLIMATE CONSTRAINTS ...............78 2005 FIGURE4.2. PAKISTAN:MAJOR FIGURE4.3. INVESTMENTCLIMATECONSTRAINTS IDENTIFIEDBYURBANFIRMSINPAKISTAN, 2003 ..........79 FIGURE4.4. ESTIMATEDNETWORKTRAVEL ANDRURALINVESTMENTCLIMATE SURVEY SITES ........79 TIME 85 FIGURE4.5. CONFIDENCEINTHERULEOFLAW THREEPROVINCES.......................................................... IN 88 FIGURE5.1. GOVERNMENT TOTALEXPENDITURE AND FINANCING*,2001-02 TO 2005-06 FIGURE5.2. FEDERAL AND PROVINCIAL GOVERNMENT EXPENDITURES*, 2001-02 TO 2005-06 ..................106 ......................... 107 FIGURE5.3. FEDERAL 2004-05 ................................................................................................................................ AND PROVINCIALAGRICULTURE AND IWGATION EXPENDITURES,1987-88 TO 107 FIGURE5.4. EXPENDITURES AGRICULTURE RURALDEVELOPMENT ....................... 110 FIGURE5.5. EXPENDITURES SOCIALSERVICESINSOUTHASIA ........................ ON AND INSOUTHASIA ON FIGURE5.6. REALREVENUESAND EXPENDITURESINPUNJAB,PAKISTAN, 2001-0 FIGURE5.7. TRANSFERSLOCAL TO GOVERNMENTSPROVINCESINPAKISTAN, BY FIGURE6.1. MAJOR GOVERNMENT NON-GOVERNMENT AND RURALPROGRAMSINPAKISTAN, 2004-05 ...128 List of Annex Figures FIGUREA 1-1. 1998POPULATIONDISTRIBUTION INPAKISTANBY CITY SI FIGUREA 2-1. REALPERCAPITA HOUSEHOLD EXPENDITURESOF RURAL PAKISTAN, 1998-99, 2001-02 AND 2004-05 ..................... FIGUREA 2-2. REALPERCAPITA HOUSEHOLD EXPENDITURESRURAL OF PAKISTAN: 1998-99, 2001-02 AND 2004-05 ....................................................................... 34 FIGUREA 2-3. REALPERCAPITA HOUSEHOLD EXPENDITURESOF RURALHOUSEHOLDS INSINDH, PAKISTAN, 1998-99, 2001-02 AND 2004-05 ........................................................................ 35 FIGUREA 2-4. SINDH: REALPERCAPITA HOUSEHOLD EXPENDITURES RURALPOORHOUSEHOLDS: OF 1998-99, 2001-02 AND 2004-05 .......................................................................................... 35 FIGUREA 2-5. REALPERCAPITAHOUSEHOLD EXPENDITURESOFRURALHOUSEHOLDS INNWFP, PAKISTAN, 1998-99, 2001-02 AND 2004-05 ........................................................................ 36 FIGUREA 2-6. NWFP:REALPERCAPITA HOUSEHOLD EXPENDITURESOFRURALPOORHOUSEHOLDS: 1998-99,2001-02 AND 2004-05 ....... ..................................................... FIGUREA 2-7. BALUCHISTAN:REALPERCAPITA Ho OF 1998-99,2001-02 AND 2004-05 ....... D EXPENDITURES RURALHOUSEHOLDS: ....... FIGUREA 2-8. REALPERCAPITA HOUSEHOLD EXPENDITURESOFRURALPOOR HOUSEHOLDS IN BALUCHISTAN, PAKISTAN, 1998-99,2001-02 AND 2004-05 .. ............................... vii Listof Boxes BOX 2.1. HUMAN CAPITAL AND HOUSEHOLD INCOMES INRURALPAKlSTAN ................................................ 20 Box2.2: REDUCINGINFANT MORTALI Box3.1: INDIA'S NATIONAL AGRICUL BOX 3.2: FOOD Box3.3: IDARA-E-KISSAN(HALA ................... QUALITY AND SAFETY DAIRY) BOX 3.4. LAND TENANCYAND PRODUCTIYITY............................. BOX 3.5. IMPROVINGEFFICIENCYOF WATER USEINPAKISTAN BOX 3.6. WHEAT POLICY REFORMS: TOWARDSMORE EFFICIENTWHEATMARKETPAKISTAN.............56 .................................................................... A IN 65 BOX.3.7. AGRICULTURALGROWTH AND RURALPOVERTY ................................................................ Box5.1: DECENTRALIZATIONINDIA IN .............................................................................................. Box 5.2: CITIZEN COMMUNITYBOARDS (CCBs) ....................................... ............................................ 121 Box 6.1: RURALDEVELOPMENT PROGRAMSINPAKISTAN ......................... ............................................ 125 BOX 6.2: KHUSHAL PAKISTAN PROGRAMS1AND 11 ..................................................................................... 127 BOX 6.3: SOCIALMOBILIZATION THE RSPS AND ........... ................................................... 129 Box6.4: AGAKHANRURALSUPPORTPROGRAMME (AKRSP) ................................................................... 130 Box6.5: SOCIAL MOBILIZATIONANDHRAPRADESH,INDIA IN ................................................................... 132 Box6.6: THECHILD SUPPORTPROGRAM(CSP)-A CONDITIONAL CASHTRANSFER .............................. 139 BOX 6.7: PROBLEMSOF HEAVILY-INDEBTED LABORERS INPAKISTAN ........................................................ 141 viii ACKNOWLEDGEMENTS This report was prepared by a core team led by Paul Dorosh (SASSD) and including Mona Sur (rural investment climate) and Raja Rehan Arshad (public service delivery). Parvez Hasan provided overall guidance for the work and participatedin the main mission o f the study. Major ' contributions and background papers were provided by Sohail J. Malik and Hina Nazli (rural poverty), Abdul Salam (agriculture), Muhammad Khan Niazi, Nuzhat Ahmad, and Mahmood Khalid (rural service delivery), Jian Zhang (rural' investment climate), Ridley Nelson (CDD and micro-credit) and Rasmus Heltberg (safety nets). Marika Krausova and Kelly Jones contributed to the statistical analysis. Olufemi Terry edited the final manuscript; Lilac Thomas and Yoshiko Masuyama formatted this document. The Rural Investment Climate Survey was managed by Mona Sur and conducted by Innovative Development Strategies, Ltd. (Islamabad) under the direction o f Sohail J. Malik. The survey was funded in part by the Food and Agriculture Organization (FAO) o f the United Nations and the U.K.Department for InternationalDevelopment. Helpful comments and other inputs were received from Zafar Altaf, Shweta Banerjee, Adolfo Brizzi, Manuel Contijoch, Caesar Cororaton, Csaba Csaki, Carlo del Ninno, Tekola Dejene, Gershon Feder, Manuela Ferro, Qazi Azmat Isa, Hanan Jacoby, Willem Janssen, Ijaz Nabi, Ambreen Malik, Ghazala Mansuri, David Orden, Sarfraz Qureshi, Stephen Rasmussen, Khalid Riaz, Geeta Sethi, Andrea Silverman, Dina Umali-Deininger, Tara Vishwanath, John Wall and Nobuo Yoshida. The report has also benefited from comments by participants at numerous workshops and seminars held at the Ministry o f Food, Agriculture and Livestock (MINFAL), the Pakistan Agricultural Research Council, the Punjab Planning Department in Lahore and the World Bank in Washington, D.C., as well as a video-conference with NGO groups and others in February 2007. ix EXECUTIVESUMMARY Over the last century, Pakistan has successfully harnessed the Indus River to develop a major irrigated-agriculture sector that remains the backbone o f its economy. However, the country i s divided by income inequality and geographic disparities that are an increasing source o f concern, and which are likely to be potentially destabilizing. These divisions are particularly pronounced in the agriculture sector, as most o f the rural poor lack access to land and irrigation water. Ensuringefficient use o f water and buildingproactive partnerships that involve the private sector can contribute to fulfillment o f agriculture's potential for diversification and growth. However, .for the rural non farm poor, escaping poverty will also likely require a combination o f migration and a revitalization o f the rural economy, which generates substantial rural employment. Expansiono f the ruralnon-agriculture sector and improvements inthe welfare o f the rural poor will not occur however, without the undertaking o f significant investments in infrastructure and improvements in social services, including safety nets. Two critical elements must underpin the necessary transformation o f the rural sector. First is the efficiency o f public institutions and the need to make these more accountable and flexible. Second i s the capacity to organize the "people sector" so that farmers, communities and villages gain a voice and achieve the scale needed to attract the private sector and financial services as well as to strengthen the demand side o f development by making government more accountable. BACKGROUND After a decade of moderate growth but little or no long-term change in rural poverty in Pakistan, agricultural output, rural incomes, rural poverty and social welfare indicators all showed marked improvement between 2001-02 and 2004-05: Real agricultural GDP per capita rose by 7.4 percent; average per capita rural expenditures rose by 5.9 percent; and per capita expenditures o f the poorest two quintiles rose by 3.1 percent. Further, rural poverty declined from 39.1 percent to 34.0 percent according to World Bank estimates. Other rural welfare indicators improved even more sharply. Child immunization rates rose by 26 percentage points, from 46 percent to 72 percent; net primary school enrollment rose by 10 percentage points, from 38 percent to 48 percent; and access to tap water in the home rose by 13 percentage points, from 10percent to 23 percent. These impressive achievements notwithstanding, there is not yet cause for complacency. Approximately 35 million rural people remain poor (accounting for about 80 percent o f Pakistan's poor), and rural poverty rates in 2004-05 were still at levels approximating those o f the 1990s. And in spite o f improvements, non-income measures o f welfare related to health and education are low in comparison with those o f other countries in South Asia. Infant mortality per 1000 live births i s 82 in Pakistan (88 in rural areas), compared to only 62 in India, 56 in Bangladesh and 12 in Sri Lanka. The national primary school enrollment rate for girls in Pakistan i s only 48 percent (42 percent inrural areas), compared to 86 percent inIndia. X Figure 1: Social Indicators of RuralHouseholds in Pakistan, 1988-99,2001-02 and 2004-05 780 * O I ~ 70 775 60 770 g 50 765 B 760 -Net Enrollment Rate E E 40 -- 755 Access to Tap Water , 30 --. - 750 Real Per Capita 20 745 10 740 735 1998-99 2001-02 2004-05 Source: PIHS 1998-99,2001-02; PSLM 2004-05. Notes: * Real per capita expenditures o f the bottom 40 percent of rural households, Rs (2004-05)icapita. Moreover, not all improvements in incomes are the result of government policies or sustainable increases in private-sector productivity. Impressive gains in agricultural output and real incomes o f the rural poor relative to 2001-02 levels i s in part a reflection o f low output and incomes in that year due to drought and other adverse shocks. The longer term agricultural GDP per capita growth rate (1999-2000 to 2004-05) was only 0.3 percent annually. In addition, much o f the improvement in total incomes can be attributed to a steep rise in net private unrequited transfers from abroad (including workers' remittances). By 2005-06 these transfers averaged more than Rs 3,000 per person for Pakistan's entire population, equivalent to more than two-thirds the real output o f crop agriculture or livestock production, (Figure 2). Yet, these transfer incomes, some o f which accrue to rural households or are spent on rural and small-town products and services, may not continue to grow at the rates o f recent years. Workers' remittances, a major component o fthese transfers, have stagnated over the last three years. Pakistan has already taken important steps to spur agricultural growth and farmer incomes through liberalization o f trade and pricing policies. Implicit taxation o f tradable agricultural products through trade policy and real exchange distortions was largely eliminated in the late 1980s and the 1990s. Likewise, direct government intervention in agricultural markets i s now minimal, with the exception o f wheat. The country i s also taking steps to promote its agricultural exports and has joined the Cairns group in the WTO negotiations. Pakistan's agriculture, particularly its irrigated crop and livestock sub sectors, are poised for gains in competitiveness if the country can develop and diversify higher value supply chains and if the recent exchange-rate appreciation does not become a limitingfactor to its export potential. x i Figure2: Trends inAgriculturalGDP and Net Private Transfers inPakistan 1 5,000 1 I 0.40 4,500 0.35 e 4,000 60 3,500 0.30 x Q 3,000 0.25 8B 2,500 0.20 0 2,000 2. 0.15 2 1,500 1,000 0.10 500 0.05 0 I , , , 1 I I , 0.00 1999-00 2000-01 2001-02 2002-03 2003-04 2004-05 2005-06 +Crops --a- Livestock - Private Transfers +Remittances -Agric Share of GDP Source: PakistanEconomic Survey, (various years) and authors' calculations. Notes: Per capita figures are calculated using the total populationof Pakistan. Who are the RuralPoor? Although agriculture is at the heart of the rural economy, the majority of Pakistan's rural poor are neither tenant farmers nor farm owners. Farmers (including both owners and tenants) comprised only 43 percent o f households in the bottom 40 percent o f the rural per capita expenditure distribution in 2004-05. Non-farm households (excluding agricultural laborer households) accounted for slightly more than half (52 percent) o f the poor. Overall, agriculture (including both crop and livestock production) accounts for only about 40 percent o f rural household incomes; the poorest 40 percent o f rural households derive only about 30 percent o f their total income from agriculture. Growing Disparities A major reasonfor the large proportion of rural non-farm poor in Pakistan, as well as poverty levels among small farmers, is the prevailing highly unequal distribution of land and access to water. According to the 2000 Agricultural Census, only 37 percent o f rural households owned land, and 61 percent o f these land-owning households owned fewer than five acres, or 15 percent o f total land. Access to usable water i s also quite unequal, a major cause o f lower productivity in the dry lands (buruni) relative to irrigated land, land at the tail end o f watercourses relative to land at the head end, and areas with saline groundwater as compared with areas that have sweet groundwater. Because o f this skewed distribution o f ownership and access to productive assets, much o f the direct gains in income from crop production, particularly irrigated agriculture, accrue to higher-income farmers.' 1 Inprinciple, landreformcouldhelpreducethe inequalityo f landdistribution. Past landreforms in Pakistaninvolving land ownership ceilings faced substantial opposition, however, and had only limited effects. One alternative i s sales or grants o f government land to the poor. Another option i s market-based xii Figure3: RuralPoor inPakistanby HouseholdGroup, 2004-05 Agricultural Farmers Laborers Non-farm Self- Baluchistan ^^I -I5% i ,/Employed 17% Farmers Farmers Sindh 11% Non-farm Other Punjab 35% 24% Source: Calculated usingPSLM 2004-05 Notes: The poor are definedas households inthe bottom40 percent of the rural adult equivalent per capitaexpendituredistribution.Livestockherderswho do not receive any crop income, (four percentof the ruralpoor), are includedinthe Non-farmother category. Additional measures to accelerate rural income growth and poverty reduction are essential if Pakistan i s to extend the promising trends in growth, poverty reduction and social welfare of the past several years. In addition to land and water, unequal distribution o f household income and access to public services, both within and across provinces, i s predominant. These disparities threaten not only future growth and poverty reduction, but social cohesion and stability, as well. Thus, inclusive growth that enhances the welfare o f the rural poor i s essential for Pakistan's future. The alternative i s more extreme inequality and polarization, with large pockets o f poverty in rural areas and accelerated migration o f unskilled people to urban areas that will find it increasingly difficult to provide needed employment and social services. Agricultural growth is not enouph Agricultural growth is a necessary, but not sufficient, conditionfor rapid reduction of rural poverty. Agricultural growth remains important to raise the incomes o f small farmers and to generate growth linkages by increasing demand for rural non-farm goods and services. However, inmost o f rural Pakistan, the impact o f agricultural growth linkages on rural poverty i s limited for two reasons. First, much o f the gains in rural incomes are spent on urban goods and services. Second, gains to non-agricultural rural incomes and employment as a result o f growth linkages are shared among a large number o f rural poor. Pakistan's rural and small-town non-farm sector, however, faces numerous constraints, particularly concerning access to credit and inadequate infrastructure. As in other South Asian countries, the non-farm sector in Pakistan's rural villages and small towns primarily consists o f family based micro-enterprises, averaging only about two workers per enterprise. According to a 2005 survey, more than 30 percent o f these enterprises rated access to landreforminvolving voluntary sales of land(which couldinvolve large fiscal costs). See World Bank (2004b). xiii finance as the most important overall constraint to the operation and growth of their business. Poor road infrastructure raises transport costs and reduces profitability. Lack o f access to reliable electricity (half o f village enterprises reported power outages o f 20 days or more in a typical month) limitsproduction or necessitates private investment in generators. Moreover, high rates of rural-urban migration and the increasing integration of the rural and urban economies point to the importance of investment in human capital, in addition to policies addressing growth of the agricultural and rural non-farm sectors. As in most o f Asia, Pakistan's economy i s in the midst o f a major transformation involving growing urbanization and increased linkages among rural, urban and international markets. About two- thirds o f Pakistan's population lives in rural areas, where average per capita expenditures are 31 percent lower than those in urban areas (Rs 1259/month and Rs 1818/month in 2004-05, respectively). Many members o f today's rural households are likely to migrate to more urbanized areas. Others that remain in rural areas will derive an increasingly large share o f their incomes from employment in small towns. Investment inthe human capital o f these workers will not only enhance their productivity in the rural economy, but also enable them to earn higher incomes if they choose to migrate. TOWARDSSTRATEGYFORRURAL A GROWTH AND POVERTY REDUCTION Given the context, a comprehensive rural growth and poverty-reduction strategy is needed,with four main pillars: Promote efficient and sustainable agricultural growth to raise the incomes o f small farmers and to generate growth linkages inthe rural non-farm economy Create an enabling environment for the rural non-farm sector to enhance employment and incomes, and improve rural public service delivery in infrastructure, health and education to serve as a foundation for growth and to increase household welfare 0 Enhance the effectiveness and governance of rural institutions through decentralization and strengthen local demand for enhanced accountability through more proactive use o f public-private partnership Empower the poor and protect the most vulnerable through social mobilization, safety nets and by facilitating access to productive assets for income-generating activities. PROMOTINGEFFICIENT AND SUSTAINABLEAGRICULTURAL GROWTH Agricultural growth has historically played a major role in Pakistan's development and continues to be crucial for overall growth and poverty reduction. Investments in large- scale surface irrigation in the 19`h century, the Indus Waters Treaty agreed in 1960, and subsequent major investments in the Tarbela dam (the world's largest earth-fill dam) and link canals provided the basis for a vibrant agncultural sector and the development o f the country's modem economy. Even though agriculture's share o f GDP has fallen from about 40 percent inthe 1960s to about 20 percent today, strong overall growth i s infeasible without strong agricultural growth also. Agriculture remains the largest source o f household income for 38 million Pakistanis, including 13 million o f the poorest 40 percent o f rural households. Substantial scope exists for increasingproductivity and overall economic efficiency inthe sector. Aaricultural Production and Technoloav Because of constraints on arable land with access to water, crop sub-sector growth cannot take place without increases in productivity. Past investments in crop technology xiv (especially cereals) have had very highpayoffs in South Asia interms o f poverty reduction, both by raising farmer incomes and by reducing the real prices o f major staples (wheat and rice). Furtherinvestment in agricultural technology, especially for the major cereals and cotton, (crops for which Palustan has a comparative advantage), i s needed both to offset declining soil fertility inmany regions andto augment productivity. Increasedrecourse to public-privatepartnerships in research and extension, an approach that has been successful in other developing countries, i s one promising way to boost the effectiveness o fpublic spending. Achieving substantial future growth in agricultural crop incomes will also require additional diversification into higher-value crops appropriate for Pakistan's various agro- ecological regions. Substantially greater efforts are needed in agricultural research on higher value crops such as oilseeds, vegetables and h i t s as well as livestock, and improvements in seed certification and quality control. Increasing the price o f water rights so that they more closely reflect the opportunity costs o f water also would encourage cultivation o f high-value h i t s and vegetables. Perhaps most importantly, formation o f farmers' organizations and new forms o f marketing arrangements (such as public-private partnerships and contract farming) can significantly reduce transactions costs and increase sales volumes and prices paid to farmers. Programs to increase livestock production, especially production of dairy cattle and milk in Punjab and Sindh, and sheep and goats in NWFP and Baluchistan, could have significant direct impact on the incomes of the rural poor. Productivity gains for small-holder dairy farmers are feasible through the use o f improved feeds, better veterinary services, and more efficient marketing channels to help ensure more remunerative milk prices for farmers. Multi- national firms have enjoyed some success inimproving the efficiency o f marketing channels, but pasteurized and UHT milk still account for only about 10 percent o f total milk consumption in Pakistan; the remainder i s fresh milk. Vertically integrated cooperatives such as the Idara-e- Kissan (Hala Dairy) that deliver both the requisite technology and veterinary and marketing services to small farmers are promising alternatives. Designing effective programs requires attention to the cultural and economic constraints facing rural women, who play a major role in the care o f livestock, including the collection o f fodder, cleaning o f sheds and the processing o f animal products. Zncreasinp Efficiency of Water Use Increasing the efficiency of water use is critical for the sustainability of agriculture in Pakistan. Water remains the key resource constraint on Pakistan apculture. Irrigation water i s critical for agriculture in much o f the country, and massive public investments in dams and canals in the Indus river basin, as well as mainly private investment in wells for groundwater irrigation have been essential for agricultural production. However, current patterns o f water use in Pakistan are causing significant environmental degradation. Losses of soil fertility due to agricultural soil degradation (soil salinity and erosion) are estimated at Rs 70 billion per year or 1.5 percent o f total GDP and 6.8 percent o f agricultural GDP, according to 2004-05 GDP figures. Inaddition to environmental loss, inefficiency inthe allocation of water within canal systems causes lower crop productivity at both the head ends o f canals (due to over-use o f water and water-logging) and the tail ends (due to water shortages). Improvements in water delivery (including investments in drainage, control structures and conveyance mechanisms), better water management (through strengthening o f water course organizations, farmers organizations, area water boards or professional canal management agencies) and greater efficiency of water use at the farm level (crop choice and adoption o f water saving technologies) are the most promising approaches in the short to medium-term. xv Irrigation Departments need to establish asset management plans that indicate measures for divestment o f assets that can be turned over to the private sector or farmer organizations. Irrigation Departments also need operation and maintenance plans that include medium-term expenditure frameworks, assignment o f financial responsibilities to the various actors, pricing policies and cost-recovery mechanisms. Service delivery and allocation of water can be improved through promotion of water user associations and farmer organizations (FOs), as well as through improvements in the existing infrastructure. FOs could be given authority to collect and hold fees for canal water (abiana), as well as for operation and maintenance o f distributaries and branch canals. Financial incentives could also rationalize water use. Given that water i s increasingly scarce inPakistan, a shift from water-intensive crops like sugar cane to less water-intensive crops like oilseeds (sunflower and canola), maize, and vegetables irrigated with drip irrigation techniques, could result in large water savings, increasing water availability downstream and reducing environmental degradation. For example, shifting 60 percent o f land under sugar cane cultivation to a maize-wheat rotation or converting about one-quarter o f ordinary rice land to maize would save enough water to supply the current water needs o f Karachi. Charges for canal water (abiana) should be revised to better reflect opportunity costs of water use. Abiana can be used to incentivize farmers to diversify to higher value crops. However, to minimize problems o f nonpayment, fee increases must be accompanied by improvements in service delivery. Raising water charges to levels approximating the opportunity cost o f water i s likely to be politically infeasible, however. An alternative would be the development o f markets for tradable water rights such as exists in Chile and Mexico. This would require laws establishing tradable water rates and the strengthening o f local institutions to enable these to serve as brokers in water markets. Projects and programs involving agricultural extension, technical assistance, market development and credit may be needed to ease the transition o f farmers andprocessors to alternative crops and products. Apricultural Markets and TradePolicies Direct government involvement in markets through domestic purchases and sales has declined since its peak in the 1970s and 1980s. Interventions in sugar cane and rice markets have ceased, but in recent years provincial and federal agencies have intervened in cotton and wheat markets. These interventions, which typically lead to large financial losses and discourage private sector investment in storage and trade, should be, if possible, phased out. Promotion o f farmer associations to increase farmer's market power, particularly for production o f highly perishable fruits, vegetables and dairy products, can improve producer prices and volumes sold. Several steps can be taken to improve market efficiency and reduce the fiscal burden associated with government intervention in wheat markets. International experience suggests that the main objectives o f Pakistan's wheat policy-to secure the provision o f adequate supplies o f wheat flour for consumers at affordable market prices and to maintain prices paid to producers at sufficiently high levels to ensure domestic production and real incomes for wheat farmers+an be achieved more efficiently through greater reliance on market mechanisms. The federal government has actively promoted liberalization o f wheat markets in recent years by advocating avoidance o f domestic marketing and storage restrictions and promoting a liberalized import policy. Significantly lower volumes o f public procurement by provincial food departments and PASSCO would immediately reduce subsidies and increase the private sector's market share. Increased transparency and consistency in the implementation o f these policies could increase price stability and enhance food security inPakistan. xvi Faster growth in Pakistan's exports of non-traditional agriculturalproducts (fruits and vegetables, meat, fish) will require improved capacity to meet Sanitary and Phyto- sanitary (SPS) standards imposedby importingcountries.Pakistan presently lacks a coherent strategy (or fkamework o f strategies) for quality and SPS management in relation to its trade. Organizationally, there i s a need to consolidate and enhance the fragmented, isolated, and non- coordinated capacity o f SPS management institutions and regulations. This can be achieved by better defining the roles and responsibilities o f the various agencies involved in SPS management, strengthening technical capacity for risk assessment, and re-institutionalizing early warning and surveillance systems for pests, diseases and contaminants. Trade policies and governance issues that reduce efficiency of supply chains for processing, transport and commercialization of high-value agriculturalproducts also need to be addressed. These issues are particularly important for products destined for highly competitive international markets. For example, import tariffs on cold and sanitary packaging materials inflate processing costs o f milk and milk powder. Similarly, governance problems related to clearance o f goods at customs and ports add to costs for export goods. And unofficial payments for security protection add to marketing costs for both domestic and internationally traded goods. CREATINGAN ENABLINGENVIRONMENT FOR THE RURALNON-FARM SECTORAND IMPROVING PUBLIC SERVICE DELIVERY Although agricultural growth remains an important pathway out o f poverty in rural Pakistan, particularly for the 40 percent o f Pakistan's poor who are farmers, 45 percent o f the rural poor rely on non-farm activities as their main source o f income. Poor farmers also derive about half o f their total incomes from non-farm activities. Moreover, nearly 40 percent o f the labor force o f rural Pakistan i s employed in non-farm sectors, either in self-employment (26 percent) or wage workers (13 percent). For these reasons, a well developed and functioning rural non-farm sector i s essential to generate employment, ensure income diversification and reduce poverty. Achieving this will require substantial improvements in rural service delivery, particularly for investments in rural and small-town infrastructure. Greater investments in health and education are also required to both improve living standards today and to build human capital for the future. EnhancinpAccess to Credit A number of successful financial-sector reforms have been implemented in recent years, but rural and small town non-farm enterprises, which primarily tend to be micro enterprises involved in trade and services (about 60 and 30 percent of enterprises, respectively),still lack access to formal credit. Survey results indicate that access to credit is a major constraint and that many firms seeking credit are unable to get it. Potential suppliers o f credit are often reluctant to lend to rural and small-town businesses because o f high transaction costs and perceived high risks. Various measures are needed to increase the flow o f credit and spur investment and growth. Among these: Develop alternative forms o f collateral acceptable to banks Make procedures for loan applications less cumbersome Provide practical training for entrepreneurs in accounting and preparation o f business plans to improve their ability to signal credit-worthiness to financial institutions xvii Figure 4: Percentage of the RuralWorking Population* Employedin Non-farm Sectors 50% 45% 40% 35% 30% 25% 20% 15% 10% 5yo 0.Y" ._ Pakistan Punjab Sindh NWFP Source:PSLM 2004-05. Notes: *Active labor force, aged 10 years and above. Organize potential borrowers in groups or professional associations to reduce risks and transaction costs for lenders Strengthen the institutional and regulatory framework for the enforcement o f contracts, arbitration and conflict resolution FacilitatingLinkages toInput and Sales Markets Market linkages could be facilitated by supporting the creation and development of more effective local business associations or organizations. These organizations could arrange product fairs, help develop business directories, and provide information services and details on how to undertake market research. They could also share information on prices, quality standards and how to access technical, financial and organizational services for greater value addition. Facilitating group marketing and business clusters could also help rural enterprises take advantage of scale economies, thereby allowing them to purchase inputs at lower prices, access larger markets and share the use o f equipment and infrastructure. ImprovingPublic-service Delivery Because spending on rural public service delivery is highly fragmented, with major componentsallocated across the various levels of government, a concerted effort is neededto coordinate government's rural development efforts. Addressing the major infrastructure constraints (e.g. roads and electricity) that hinder growth o f the rural non-farm sector will require greater flow o f resources to local governments and increased efficiency o f public current and development spending. Improvements inpublic-service delivery inhealth, education, water supply and sanitation also are needed to enhance welfare o f the poor and build human capital for future income growth. Given that there i s no longer a formal administrative difference between rural and urban areas in Pakistan, there i s no mechanism to track spending in rural areas apart from information on agriculture spending. Accurate monitoring o f total spending inrural areas i s needed to ensure that rural development needs are addressed and impact measured. Provision of infrastructure, particularly roads and reliable electricity, can reduce operating and marketing costs, making investmentsin enterprisesin rural areas and small towns more profitable. Access to electricity remains a major challenge in many villages and xviii small towns, and even among enterprises with access, reliability o f supply i s uncertain. The median number o f days with power outages ina typical month was reportedas 20 days invillages and 15 days in small towns. Access to and use o f telecommunications among enterprises i s also surprisingly limited. Only 31 percent o f entrepreneurs in small towns (9 percent in villages) owned fixed line phones or cellular phones. Weeded measures include: 0 Better road maintenance and extension o fbasic motorable access; 0 Improved institutional arrangements for ownership, management and financing o f the rural transport system inaccordance with the realities o f devolution; Promoting greater community involvement in planning and managing transport infrastructure improvements to ensure that infrastructure meets local needs. 0 Increased investments in power distribution and transmission, together with accelerating key reforms inthe energy sector. Increased public expenditures are particularly important in health, education and population. In spite o f improvements since 2001-02, Pakistan's expenditure and outcome in health and education remain low by comparison with other south Asian countries. These sectors also appear to be under-funded relative to other public expenditures that are less pro-poor. Substantial investments in education, including vocational training programs and programs targeted to women, are needed to increase human capital and income- earning potential, so as to break the inter-generational poverty trap. Measures to promote slower population growth are also a highpriority because o f the direct impact smaller family size exerts on per capita household incomes. IMPROVINGTHE EFFECTIVENESSAND GOVERNANCERURAL OF INSTITUTIONS Pakistan needs to review the role of its public institutions to change a control- oriented, supply-driven system to one that i s more decentralized, efficient and demand- driven. The devolution that occurred in2002 was an important step, butmuchremains to be done to reap the benefits o f devolution, including improving administrative capacity o f local government and further strengthening o f Citizen Community Boards (CCBs). Greater interaction with private-sector groups o f farmers and enterprises through public-private partnerships could facilitate government responsiveness and improve efficiency o f investments and programs. In addition, measures that improve governance and accountability such as stricter enforcement o f laws, legal reforms and wider dissemination o f information on spending and effectiveness o f programs can help spur both greater efficiency o f government, as well as growth and investment inthe private sector. Reaping the Benejits of Devolution Fiscal resources for rural development should be increased, particularly at local government (districts, tehsils and unions) leveland efficiencyof spending enhanced. These flows should be monitored closely and renewed efforts should be made to ensure that resourcesreach local governments inthese provinces. Inorder to strengthen the local government planningprocess and the transparency of resource allocation, discretionary transfers' share o f total transfers should be reduced and provincial non-discretionary transfers to local governments should increase more quickly than local government wage bills. Special attention should be given to monitoring public service delivery inrural areas. xix Five years after devolution, there is still confusion regarding the roles and responsibilities of the various levels of government, as well as apparent jurisdictional overlaps. Immediate steps should be taken to begin to alleviate these administrative constraints and improve the efficiency o f spending. New staff should be hired and practical training provided, particularly at the Tehsil Municipal Authority level. 0 Jurisdictional overlaps should be reviewed and corrected following a systematic assessment o f administrative responsibilities at the various levels o f local government 0 The transfer o f resources through vertical programs and other higher-level government programs should be discouraged and reduced, with particular attention to the water and sanitation sector, which currently receives substantial transfers through Members o f the National Assembly (MNAs), Members o f Provincial Assemblies (MPAs) and directives o f the ChiefMinister. The delegation of power to local governments through devolution has brought decision making closer to the communities, enabling them to access government officials and this has led to some improvements in service delivery. Survey evidence suggests that citizens' degree o f access to their representatives has increased considerably and problems are solved relatively quickly under the new system. Priorities o f the local communities are also increasingly reflected in development schemes. The effective participation o f communities could be increased further, however, through: 0 Widely disseminating user-friendlyguidelines for CCB processes 0 Developingtraining programs for councilors to improve their awareness and ensuring implementation o f the Local Government Ordinance (LGO) provisions for CCB funding Reforming Institutionsfor Agricultural Development Irrigation, agricultural research and agricultural extension institutions also need to be made more accountable to farmers. Refoms o f irrigation departments are neededto increase accountability, efficiency, transparency and competition in surface water supply. Key steps include: 0 Unbundling the business into bulk (operation of dams and barrages), transmission (canal management) and distribution (management o f tertiary systems through water users' associations) components, with relations among the parts governed by contracts 0 Encouraging private-sector involvement in management o f canals and promoting competition between the irrigation departments and private service .providers at the distributary level Effectiveness o f agriculturalresearchat the provincial level could be enhanced by: 0 Creating autonomous research organizations with the ability to attract qualified scientists and with adequate administrative and managerial flexibility 0 Tailoring research agendas to the needs o f individual provinces, particularly regarding soil fertility, water-logging and development o f new varieties and products that meet market demand 0 Forging partnerships with the private sector and knowledge institutions to develop new technologies, as well as to disseminate andtransfer these on a large scale. xx Decentralizationo f agriculture to the district level has been implementedwithout the necessary measures to ensure adequate capacity and funding.Programs to promote agricultural production can be made more effective by: 0 Increasing the capacity o f district governments to set priorities and develop implementation strategies 0 Strengthening the link between district and provincial agriculture departments to ensure consistency o f design and implementation o f agncultural programs and po1icies 0 Increasing the amount o f funds allocated for agnculture at the district level. Public-Private Partnerships Public-private partnerships (PPPs) should be pursued aggressively with respect to agricultural marketing, research and extension, health, education, infrastructure and other sectors to enhance productivity of public-sector programs. Itwill be necessary to: 0 Put in place an incentive structure for the private sector to enter into contractual arrangements with farmer organizations and link delivery o f public services to business development 0 Create public-private partnerships in agricultural research and extension to help public research systems become more responsive to farmers' actual needs 0 Strengthen recent government education-sector reforms that enable engagement with private and NGO schools to improve access and quality o f education, including government-supported per child subsidies 0 Increase government financial support for NGOs, including Rural Support Programmes (RSPs) that deliver micro-credit, skills training and other private services. Governance and Accountability Improving governance is crucial for reducing poverty and promoting growth in both the farm and non-farm sector. Governance issues, including political instability, corruption, insecurity and lingering conflict, are a major bar to investment, economic growth and efforts to reduce poverty. Bureaucracies with low levels o f accountability to stakeholders tend toward inefficiency and inability to deliver rural services, including agricultural and livestock research, extension, irrigation, natural-resource management, and market support services. To address these issues, the Government o f Pakistan should strengthen recently begun efforts to: i)reformgovernment institutionsthat oversee economic andfinancialmanagement, thepolice, the judicial system, and the civil service; (ii)improve public financial management, accountability, and increased transparency and information on government activities to facilitate public oversight; (iii)strengthen local government institutions; (iv) privatize and deregulate public enterprises to reduce opportunities for rent-seeking behavior. Improving the mechanisms for contract enforcement, promoting transparency in the legal system and allowing non-farm enterprises a greater say in policy making would help spur rural and small-town enterprise growth. Further, improved mechanisms for contract enforcement would also facilitate market linkages by enabling smaller enterprises to enter into sub-contracting arrangements with larger firms and thereby gain access to larger markets. xxi All rural developmentefforts couldbe improvedthrough enhanced monitoring and evaluation. Adopting a national, results-based management framework is a crucial step to ensure effective use o fresources. 0 Both projects and larger programs should establish indicators at the outset, carry out simple baseline assessments and monitor indicators against targets. Improvementsin collection o f basic data are also needed. Inparticular, data on wage rates in small towns and rural areas should be collected to help monitor labor- earnings trends inthe agricultural and rural non-farm sectors. Publishinglocal, provincial and federal government plans and financial reports would improve access to information and make the local government system more transparent and accountable. EMPOWERING POORAND PROTECTINGTHE MOST THE VULNERABLE Increasing the flow o f resources to local governments and improving administrative procedures address mainly the supply of public services. However, a major reason for the limited impact o f rural development efforts in Pakistan and many other countries is a lack o f participation and influence by rural poor households, which limits effective demand for public services and hampers efficiency in development programs. Social mobilization can empower the poor, enabling them to play a greater role in the development process, not only to improve delivery o f public services, but also to increase their market power by building so-called voice and scale in the farm and non-farm sectors (farmer organizations, water users, self help groups). Finally, although inclusive economic growth should be the main mechanismfor reducing poverty, increasedefforts toward social protection are neededto protect the most vulnerable. Social Mobilization Too often, a top-down approach is implemented - one that sees the rural poor simply as beneficiaries of public programs supplied by the government. Instead, a development paradigm i s needed that puts the household and its community at the origin of development initiatives. Empowering the rural poor to take on this role, however, requires social mobilization, a central pillar o f the government's Mid-Term Development Framework (MTDF), which covers 2005-2010 and a key feature o f the Rural Support Programmes (RSPs). Community Boards (CCBs) are also the product o f social mobilization, although these have a more limited focus on development projects implementedthrough local governments. Social mobilization, along with economic empowerment and eventual graduation o f groups, should be at the core o f the rural livelihood development strategy. The benefits o f broad economic growth trickle down very slowly when the poor have little access to key physical, social and financial endowments, To overcome highly unequal distribution o f these and to achieve rapid pro-poor growth, poor people neednew opportunities to organize, generate business and link with mainstream development activities. 0 Social mobilization, as a way to develop institutions OF the poor as opposed to institutions FOR the poor, i s a central component o f effective rural livelihood development. Mobilizing poor people provides them with a voice and the scale required to more effectively engaging with the range o f institutions and individuals providing public and private services. xxii 0 Economic empowerment, whether through micro-credit, grants or skills training, facilitates access to assets, increases incomes, and demonstrates the creditworthiness o f individuals and groups. 0 Ultimately these groups may graduate to form other federative and associative movements involved in income-earning activities including public-private partnerships, various types o f franchising and contract farming arrangements. These movements can provide the organization and critical mass needed to reduce transaction costs and leverage private sector and financial services. To strengthen and expand these and other efforts at social mobilizatfon, the government could: 0 Initiate a social mobilization program aimed at buildingsocial capital and promoting organizations and institutions of the poor 0 Facilitate access to productive assets by poor people through a matching grant window or the establishment o f revolving funds at the community level 0 Include social mobilization components in the design and implementation o f rural development programs Micro-Finan ce In addition to social mobilization efforts, microfinance programs can empower the poor by alleviating severe credit constraints that limit their incomes and increase their vulnerability to adverse shocks. Major microfinance programs exist that are run by NGOs in rural areas, such as the Pakistan Poverty Alleviation Fund(PPAF), already exist, though coverage inPakistan is smaller than inother south Asian countries. Pakistan's microfinanceprograms may improve their effectiveness through adoption o f the following international best practices: Focus on savings and the efficiency o f savings services, as savings may be even more important for the poor than credit and micro-credit without savings institutions may not be sustainable Organize the demand side and help poor communities demonstrate their creditworthiness Continually assess product demand to design saving and lending products to meet that demand (informal lenders, and even the staff o f credit institutions themselves, have been able to develop suitable products); Increase lendingto women's groups (the majority o f sustainable successes worldwide have been with women) through socially acceptable mechanisms suited to specific localities; Find efficient ways o f linking local informal rural finance institutions to formal institutions. Social Protection Safety nets for both the rural and urban poor should be made more efficient and scaled up to cover all ultra-poor and some poor households (4-11 million households). 0 Cash transfer programs such as the Food Support Program (FSP) can be scaled up and the scope o f these programs expanded to encourage long-term human-capital investments among the poor; one example i s the Child Support Programpilot. Nutritional interventions aimed at improving health and nutrition o f women and infants, such as The Tawana Pakistan school feeding program should be, as needed, redesigned or piloted and scaled up. xxiii 0 The government should also consider piloting a major rural workfare program to provide temporary employment for vulnerable households. Programs are also needed to address the problems o f heavily-indebtedworkers subject to forced labor o f various types, such as landless sharecroppers in rural southern Punjab and northern Sindh, and brick kilnworkers. Possible interventions include: 0 Increase public awareness o f unfair labor practice issues through seminars and the media 0 Strengthen enforcement o f existing laws 0 Launch a pilot project to provide access to credit and alternative income sources for these groups CONCLUSIONS Rural growth i s crucial to Pakistan's future. Two-thirds o f the country's population and 80 percent o f the poor live inrural areas; unless there i s sustained progress in these areas, rapid overall economic growth and poverty reduction are impossible. Stagnation o f the rural economy could also threaten the social cohesion o fthe country, drive massive migration to urban areas, and result in massive urban unemployment. Achieving rapid rural growth and poverty reduction, however, requires overcoming major constraints related to unequal distribution o f land and access to water, low productivity o f crop agriculture, inadequate infrastructure, ineffective public-service delivery, and insufficient participation by rural people in most public-sector development programs. Nonetheless, Pakistan has made important strides in the last several years to promote rural growth and poverty reduction. There has been substantial progress in liberalizing agricultural markets, promoting diversification and exports, and increasing expenditure on infrastructure and public services in rural areas. Progress has also been made in empoweringthe poor through social mobilization (RSPs and CCBs) and micro-credit (the PPAF). These efforts should be continued and strengthened. Inaddition, rapid and sustained reduction o f rural poverty will require an even greater focus on building human capital, through improved delivery o f health services and sanitation, basic education and appropriate technical training. Going forward, it is crucial to maintain a multi-faceted and holistic approach to rural development and poverty reduction and to ensure that sufficient resources are invested in this task. Efficient and sustainable agricultural growth is a necessary condition for robust rural economic growth, but in itself cannot be expected to drive substantial reductions in rural poverty given the unequal distribution o f access to land and water resources in Pakistan. Thus, policies to promote agricultural growth must be complemented by investments and policies that create an enabling environment for the rural non-farm sector, including increased public investment in rural and small-town infrastructure. Finally, because the ultimate objective o f a poverty-reduction strategy i s to increase the welfare o f the poor, health and education programs that increase the human capital o f the rural poor, social mobilization that empowers the poor in their interactions with government and markets and social protection programs that protect the most vulnerable should all receive increased resources and attention. xxiv 1. INTRODUCTION After a decade of moderate growth but little or no long-term change in rural poverty in Pakistan, agriculturaloutput, rural incomes,ruralpoverty and social welfare indicatorsall showed marked improvements between 2001-02 and 2004-05. Real agricultural GDP per capita rose by 7.4 percent, average per capita rural expenditures rose by 5.9 percent, and the per capita expenditures o f the poorest two quintiles o f the rural population rose by 3.1 percent. Rural poverty declined from 39.1 percent to 34.0 percent according to World Bank estimates. Other rural welfare indicators improved even more dramatically. Child immunization rates rose by 26 percentage points, from 46 percent to 72 percent; net primary school enrollment rose by 10 percentage points, from 38 percent to 48 percent; and access to tap water in the home rose by 13 percentage points, from 10 percent to 23 percent. These impressiveachievements notwithstanding,there is little reason for complacency.Not all improvements are the result o f government policies or sustainable increases in private-sector productivity. Impressive gains in agricultural output and real incomes o f the rural poor relative to 2001-02 reflect in part low output and incomes in that year due to drought and other adverse shocks. Longer term, the agricultural GDP per capita growth rate (1999-2000 to 2004-05) was only 0.3 percent per year. Rural poverty rates in 2004-05 remain at a similar level to those o f the 1990s. And social welfare indicators in Pakistan are still significantly below those o f other south Asian countries. Moreover, problems related to timing and availability o f water for irrigation, inadequate rural infrastructure, a skewed distribution o f assets, and low levels o f health and education continue to hamper the progress o f economic growth and poverty reduction. Pakistan's Poverty Reduction Strategy Paper (PRSP), which was written in 2003, recognizesthe importance of reducingruralpoverty as part of the country's overallpoverty reduction efforts. The PRSP gives a key role to the rural sector in accelerating growth and reducing rural poverty, placing major emphasis on employment-generating growth in agriculture and other sectors. However, it lacks an explicit integrated rural-development strategy that covers all major aspects o f the rural economy, including agricultural production and markets, input markets (seeds, fertilizer, extension), factor markets (land, water, labor and credit), the rural non- farm sector, targeted interventions, and perhaps most important, how local governments and communities can more effectively deliver basic rural services and strengthen the rural investment climate. This report is designedto help fill these gaps through analysis of the major constraints to rural income growth and poverty reduction in Pakistan and by offering specific policy recommendationsfor achieving these objectives. Although much o f the report's focus is on agriculture, which i s at the core o f Pakistan's rural economy, the rural non-farm economy i s also examined at length, as are government and non-government policies and programs related to rural service delivery, social mobilization and safety nets. Background:RuralPoverty andEconomicGrowthinPakistan About two-thirds of Pakistan's population and almost 80 percent of the country's poor live in rural areas. According to the 1998 census, 89.3 million people lived inrural areas o f Pakistan 1 in that year.' Household incomes are lower and poverty rates are higher in rural areas than in urban areas. Average per capita expenditures o f rural households in 2004-2005 were 31 percent lower than those o f urban households (Rs 1259/month and Rs 1818/month, respectively). The poverty rate inrural areas i s estimated at 34.0 percent, about 15 percentage points higher than the 19.1 rate inurban areas (World Bank 2006b). Rural, as well as total population growth rates are declining, which bodes well for future per capita income growth. Average fertility rates (the average number o f birthsper woman over her lifespan) have declined sharply since the mid-1980s from about 6.8 children per woman to 4.1 in 2001, although this is still high compared to that o f Bangladesh (3.3) and India (3.2). Rural population growth rates have also fallen from 3.5 to 2.6 percent. Nonetheless, the rural population i s likely to continue to grow, reaching 122 million in 2015 (64 percent o f the total population) at historic rates o f migration of 1.2 percent per year. The urban population would reach 70 million inthis scenario, and 82 million (43 percent ofthe population) ifnet migrationrates doubledto 2.4 percent per year (See the Annex to Chapter one). During the 1970s and 1980s, agricultural growth was accompanied by substantial reductions in rural poverty, but rural poverty rates in Pakistan did not decline in the 1990s despite substantialgrowth in agriculturalGDP.Eventhough real agricultural GDP rose by 4.6 percent per annum, the percentage o f the rural population living below the poverty line remained essentially unchanged between 1990-91 (36.9 percent) and 1998-99 (33.8 per~ent).~ Several factors explain the non-correlation between relatively rapid agricultural growth and rural poverty reduction, including a possible overestimate of agricultural GDP growth and an increase inreal consumer prices o f major staples since the mid-1990s. Inaddition, because o f the skewed structure o f ownership and access to factors o f production in rural Pakistan, 46 percent o f the rural poor in non-farm households do not share directly in incomes derived from agricultural crop production. Moreover, even accounting for growth linkage, effects from increases in traditional crop agriculture are relatively small in comparison with the sector's large size. Positive developments in agricultural output, rural poverty reduction and social welfare indicators in 2004-05 represent a sharp break with 1990s trends. These improvements have been achieved through a combination o f increased overall development expenditures and improved service delivery at the local level (in some localities at least), supported by a sound macro-economic environment including a liberalized trade and exchange-rate policy regime with relatively low inflation. The latter in turn spurred high GDP growth and increased demand for construction and other labor-intensive services. Factors that are likely to be more transitory also played a role, including increased increases in workers' remittances and bumper crops related to good eat her.^ The challenge now will be to extend the success of recent years to the medium term so as to further reduce still-high rates of rural poverty in Pakistan. The strategies adopted to achieve 2 "Rural" here is defined according to administrative definitions at the time o f the 1998 census. Since the 2002 devolution, there has beenno formal administrative distinction between urban and rural areas. 3 Official government estimates show poverty at 35.9 percent in 1998-99 (Pakistan Economic Survey, 2005-06). Poverty estimates vary, however, because o f changes indefinitions o f poverty lines over time and methodological issues related to price deflators. See Chapter 2. 4 Though levels o f remittances and crop yields may be maintained, growth rates are likely to fall. 2 these goals will need to take into account gradual, but increasingly important long-term changes inPakistan's economy, particularlyrural-urban migration, a declining share of agriculture intotal economic output and increasingly severe environmental constraints (particularly related to growing water demand inurban areas and environmental degradation due to drainage problems). Most importantly, a rural poverty-reductionstrategy should focus not on rural sectors, but on rural people. Many o f today's rural poor may migrate to small towns and large cities or be employed outside o f rural areas for all or part o f the year. Because o f this, reducing rural poverty rates will not be merely a function o f agricultural and rural non-farm growth, but also o f development in urban areas (including small towns) and overall economic growth. Investment in rural and small-town infrastructure can facilitate these economic linkages. Likewise, investments in human capital (education and health) can increase the productivity and welfare of the rural poor irrespective o f whether they remain inrural areas or not. PLANOFTHEREPORT The analysis begins with a detailed examination o f recent trends inpoverty and rural incomes and determinants o f rural incomes, as well as an overview o f changes to non-monetary measures o f welfare. The focus o f this analysis i s not on estimates o f poverty and rural incomes, per se, but on their determinants, so as to better understand the effects o f agricultural growth, growth in other sources o f income, and public investments inthe welfare o f the poor. Chapters three and four discuss programs and policies for spurring growth in the agricultural and rural non-farm sectors, respectively. Chapter three examines the structure and constraints to growth o f Pakistan's agricultural sector that remain at the heart o f the rural economy in spite o f that sector's declining share in the overall economy. This chapter includes an assessment o f water, the key resource constraint in agriculture, measures to increase agricultural productivity and crop diversification, the importance o f the livestock sector, agricultural markets and implications for Pakistan's agricultural trade and pricing policies. Chapter four highlights the importance o f spurringthe rural non-farm sector through more than agricultural growth linkages in order to accelerate reduction o f rural poverty. The results o f a survey o f rural non-farm enterprises undertaken in2005 follow, along with an analysis o f constraints to growth o f the rural non-farm sector. The chapter concludes with suggestions on policy measures to overcome these constraints. Chapter five examines rural service delivery in Pakistan, beginning with a study o f fiscal flows among federal, provincial and local levels o f government since the 2002 devolution. Subsequently, the results are presented o f a 2006 survey on rural public-service delivery in 7 districts and 14 tehsils, along with recommendations for enhancing community participation and program effectiveness. Chapter six covers government and NGO initiatives aimed at social mobilization and enhancing rural livelihoods, including an overview o f the major programs in Pakistan and lessons learned from the experiences o f other developing countries in social mobilization and microfinance. Chapter six also highlightsthe role o f other direct interventions to improve the welfare o f the poor, including safety nets and programs targeted to disadvantaged groups, particularly heavily indebted laborers. Chapter seven offers a summary o f the main findings and policy recommendations o f the study, including a rural development strategy comprised o f four main pillars: i)promoting efficient and sustainable agricultural growth; ii)creatinganenabling environment fortheruralnon-farmsector andimprovingruralpublic- service delivery; iii)improving the effectiveness and governance o f rural institutions; and iv) empoweringthe poor and protecting the most vulnerable. 3 ANNEXTO CHAPTER 1:POPULATION GROWTH PAKISTAN IN Available evidence suggests that average fertility rates (average number of births per woman over her lifespan) in Pakistan were stable but remained high until the mid-l980s, but have since declined sharply. Feeney and Alam (2003) show that the average total fertility rate as derived from various population surveys in Pakistan was essentially constant at 6.8 children per woman from 1960 through the mid-1980s. Data from the annual Pakistan Demographic Surveys (PDS), which were begun in 1984, show a sharp decline in fertility (at a rate o f 1.8 children per woman per decade) from 6.9 children per woman in 1987 to 4.3 in 2000, and 4.1 in 2001 (PDS 2001). Nonetheless, this figure i s high compared to those o f Bangladesh (3.3), India (3.2), Iran (2.6) and Sri Lanka (2.1).5 Population growth rates have also declined since the early 1990s, which bodes well for future per capita income growth. Official estimates o f population from the four national population censuses (1961, 1972, 1981 and 1998) suggest a decline in population growth rates from 3.62 to 3.01 to 2.65 over the inter-census periods. Feeney and Alam (2003), however, use projections o f population growth based on fertility and life expectancy data to argue that the censuses o f 1972 and 1981 overstated the population relative to the 1961 and 1998 censuses. Their population projections suggest that population growth rates actually increased from 2.6 percent in 1961-65 to 3.5 percent in 1986-90 before falling to 3.0 percent in 1991-95 and then 2.6 percent in 1996-2001,6 The implication o f these alternative estimates which adjust population growth downward prior to the mid-1980s and upward thereafter is that per capita agricultural growth in the 1980s and 1990swas substantiallyless than indicated by the official data. Real agricultural GDP growth per capita inthe 1980-90 period is thus only 0.6 percent, instead o f 1.3 percent; for the 1990s, the adjusted growth rate i s 1.6 percent, as compared to the official figure o f 1.9 percent (Table Al.1). Slower than reported agricultural growth rates per capita are another contributing factor to the rural poverty puzzle o f the 1980s and 1990s in which there i s a non-correlation between agricultural growth and rural poverty reduction. Table A 1.1. Estimatesof AgriculturalGrowth Per Capita in Pakistan, 1960-2004 1960-70 1970-80 1980-90 1990-00 1990-2004 Real GDP growth 7.19% 4.71% 6.32% 3.75% 3.62% Real agricultural GDP growth 4.89% 2.33% 4.04% 4.42% 3.54% Population growth (official) 2.79% 3.18% 2.70% 2.49% 2.47% Rural population growth (official) 2.42% 2.73% 2.34% 2.11% 2.06% ' Population growth (Feeney and Alam, 2003) 2.68% 3.12% 3.48% 2.74% 2.52% Real agricultural GDP growth per capita (official) 2.04% -0.82% 1.31% 1.88% 1.04% Realagricultural GDP growth per capita (adjusted) 2.16% -0.76% 0.55% 1.63% 0.99% Difference in growth rates (adjusted less official) 0.11% 0.06% -0.76% -0.25% -0.05% Source: PakistanEconomic Survey (various years), authors' calculations. Notes: Adjusted per capita agricultural GDP growth uses population figures from Feeney and Alam (2003). 5 2001World Populationdata sheet, Population ReferenceBureau, Washington DC. Quoted in PakistanDemographic Survey 2001). 6 Note that analysis based on fertility and mortality rates provides a consistency check on population growthrates, but not levels. Data from a future population census may resolve the issue (Feeney and Alam 2003, p. 83). 4 Urban and Rural Population Growth Pakistan's population remains predominantly rural, with half of the urban population concentratedinthe six largest cities. Interpretationo fcensus data on urban andruralpopulation levels and growth are complicated by changes in the definition o f urban in the various censuses. In the 1981 and 1998 censuses, urban areas were defined according to an administrative definition (municipal corporations, town committees and cantonments), as opposed to the population-based definition used in earlier censuses.' Using the former definition, Pakistan's population was 32.5 percent urban in 1998. About one-sixth o f the population (16.3 percent) lived in cities of more than one million people: Karachi (9.339 million), Lahore (5.443 million), Faisalabad (2.009 million), Rawalpindi (1.970 million), Multan (1.410 million) and Hyderabad (1.167). Another 7.3 percent o f the population lived in the 43 cities with populations o f 100,000 people or more. Only 4.3 and 7.5 percent o f the population lived in cities o f from 50 to 99.9 thousand people and urban areas with less than 50 thousand people, respectively (Figure Al.l). Figure Al-1. 1998PopulationDistributioninPakistanby Cit Size (mns) I .3% 3.3% .7% 67 HI + mn 100K 999K - 50-99.9K <50K Rural Source: Government o f Pakistan 2000; The 1998 Census Report; Arif 2003. Notes: The figure is based on a total population of 132.4million. Pakistan's urban population has grown significantly faster than the rural population over time, but problems with the various censuses make exact calculations for recent years difficult. Using a consistent administrative definition o f urban as adopted in the 1981 and 1998 censuses, Arif (2003) estimated that the proportiono f Pakistan's population living inurban areas rose from 17.4 percent in 1951 to 22.4 percent in 1961 and again to 32.5 percent in 1998. Annual growth rates o f urban and rural populations over the 1961 to 1998 period were 4.13 and 2.71 percent, respectively. Unadjusted population data imply growth rates o f 3.47 percent (urban) and 2.30 percent (rural) from 1981 to 1998, but because the total population in 1981 appears to be overstated relative to 1998, one or both o f these growth rates i s likely understated. Applying the same percentage adjustment (-9.8 percent) to both urban and rural populations for 1981 as used ~~ ~~ 7The 1951, 1961 and 1972 censuses defined urban as areas with a minimumpopulationbase o f 5,000 people, though exceptions were made for some localities with less than 5,000 people that had urban characteristics. 5 by Feeney and Alam (2003) to adjust total population, estimated urban and rural population growth rates in the 1981 to 1998 period were 4.1 percent and 2.9 percent, respectively (Table Al.2). Table A 1.2. Population and Fertility RatesinPakistan Natural Growth due to Growth Rate Growth Rate Net Migrationa 1981 1998 1981-98 1981-98 1981-98 Population Census Data Total 84.3 132.4 2.7% 2.7% 0.0% Urban 23.8 43.0 3.5% 2.6% 0.9% Rural 60.4 89.3 2.3% 3.0% -0.7% Adjusted Population Datab Total 76.4 132.1 3.3% 3.3% 0.0% Urban 21.6 42.9 4.1% 2.9% 1.2% Rural 54.8 89.1 2.9% 3.4% -0.5% Fertility Rate 6.8' Source:PakistanDemographic Surveys 1984-1997,2001; Karim andNasar 2003; Feeney and Alam 2003 Notes a Natural growth rates for figures based on census data derived from Pakistan Demographic Surveys 1984-97 (Karim and Nasar, 2003), p. 172. Natural growth rates for adjustedpopulation figures are calculated on relative natural growth rates of urban andrural populations from Karimand Nasar (2003). Official census data for 1981 (and 1972) appear to be inconsistent with census data from 1961 and 1998. Adjusted population series total is projectionby Feeney and Alam (2003) based on fertility rate and life expectancyestimates and 1961 and 1998 census data. Adjusted total population estimate for 1981 i s 9.8 percent below the 1981 census figure; urban and rural population estimates for 1981 use the same percentage adjustment (-9.8 percent) to the 1981 census figures. Adjusted average from various surveys as calculatedby Feeney and Alam (2003). 2001 estimate from Populationand Demographic Survey, 2001. Birthrates remain substantially higher in rural areas than in urban areas, however. Crude birthrates (the number o fbirthsper 1000 o fthe population) for 2001 were 29.4 inrural areas, 18 percent higher than inurban areas (25.0). General fertility rates (the number o f birthsper 1000 of the population o f women aged 15-49 years) were 131.6 in rural areas, 28 percent higher than the 103 for urban areas (Table A1.3). Table A 1.3. BirthRatesinRuraland Urban Areas inPakistan, 2001 and 2002 AII Urban Rural RurallUrban Crude Birth Rate 2001 27.8 25.0 29.4 1.I8 2000 29.1 25.8 31.8 1.23 General Fertility Rate 2001 120.8 103.0 131.6 1.28 2000 127.6 108.2 144.9 1.34 Source: PakistanDemographic Survey 2001 Notes: The crude birth rate i s the number of births per 1000 total population. The general fertility rate i s the number of births per 1000 women aged 15-49years. Migration to urban areas accounts for an estimated one-quarter of total urban growth. Estimates by Karim and Nasar (2003) using unadjusted population census and estimated natural 6 population growth rates in urban and rural areas suggest that net migration added about 0.9 percent per year to urban populations, out of a total of 3.5 percent annual growth. Estimates using figures adjusted for apparent over-enumeration in the 1981 census suggest that urban net migration increased urban populations by 1.2 percent per year out of a total of 4.1 percent per year growth. In spite of lower fertility rates and natural population growth rates in urban areas, their share of the population is likely to increase, but the absolute rural population will also increase.Assuming a decline innatural population growth rates to 2.0 and 2.3 percent for urban and rural areas, respectively over the 1998-2015 period, if urban populations continue to grow at an additional 1.2 percent per year because of net migration, total urban population growth will average 2.9 percent (Table Al.4). Inthis scenario, the urban population would rise by 62 percent from 42.9 million, or 33 percent of the total population, to 69.6 million (36 percent of the population). The rural population would increase by 37 percent to 121.8 million. Doubling the rate of migration to an annual 2.4 percent would raise the projected2015 urbanpopulation to 81.5 million (43 percent of the population). Table A1.4. Urban and RuralPopulation Growth Scenarios Scenario 1 Scenario 2 1998 2015 2025 2015 2025 Migration rate to urban ' 1.2% 1.2% 1.2% 2.4% 2.4% Natural growth rates Total 3.3% 2.2% 2.2% 2.2% 2.2% Urban 2.9% 2.0% 2.0% 2.0% 2.0% Rural 3.4% 2.3% 2.3% 2.3% 2.3% Population (mns) Total 132.1 191.3 238.0 191.3 238.0 Urban 42.9 69.6 89.2 81.5 111.8 Rural 89.1 121.8 148.8 109.9 126.2 Population share Urban 33% 36% 37% 43% 47% Source: World Bank staff calculations based on Karim and Nasar 2003. Notes: Scenario 1: Migration rate to urban: 1.2 percent (based on estimates for 1980-1998 from Karim and Nasar, 2003); Scenario 2: Migration rate to urban: 2.4 percent per year. 7 2. RURALPOVERTY IN PAKISTAN POVERTYTRENDS Ruralpoverty in Pakistan, which declined sharply in the 1980s, remained stubbornly high in the 1990s. Inthe 1980s, agricultural GDP growth averaged 3.9 percent per year, contributing to a steady decline in rural poverty fiom 49.3 percent in 1984-85 to 36.9 percent in 1990-91. In spite o f even faster growth in agricultural real GDP in the 1990s (4.6 percent), however, rural poverty did not decline further. Rather, the percentage o f people living in poverty remained essentially unchanged between 1990-91 (36.9 percent) and 1998-99 (35.9 percent).* Several factors help explain rural poverty's stasis in the 1990s, among these overestimates o f livestock income growth, a rise in the real consumer price o f major staples, unequal distribution o f returns to land and the fact that the crop sector contributed a declining share o f overall GDP.' Since 1998-99, real household incomes, income-based poverty indicators and agricultural output in Pakistan have fluctuated sharply, with only gradual improvement over the medium term. Recent household survey results indicate sharp reductions in rural poverty in Pakistan over the 2001-02 to 2004-05 period. Longer-term trends are less encouraging though, as these suggest no major changes in real expenditures o f the poorest 40 percent o f households between 1998-99 and 2004-05. Changes in agricultural output, due in large part to weather, mirror the changes inrural real incomes over the periods inquestion, but like real expenditures o f the poor, agricultural output and incomes have increased only modestly over the entire six-year period (1998-99 to 2004-05). Non-agricultural factors, especially increases in workers' remittances have also contributed to increased rural (and national) incomes since 2001-02. Inthe medium term, however, econometric evidence suggests that investments in human capital and physical infrastructure have been among the most important determinants o f increased real incomes inrural Pakistan. Preliminary analysis of the 2004-05 Pakistan Social and Living Standards Measurement Survey (PSLM) data indicates that both rural and urbanpoverty have declined since 2001- 02. Planning Commission estimates based on a poverty line o f Rs 723.4 in 2001-02 suggest that national poverty fell by 10.6 percent, from 34.5 to 23.9 percent between 2001-02 and 2004-05. Their estimates o f rural poverty show a decline o f 11.2 percent in the same period, from 39.3 percent to 28.1 percent. World Bank estimates for the same period show smaller declines in poverty: from 34.4 to 29.2 percent (5.2 percentage points) at the national level and from 39.1 to 34.0 (5.1 percentage points) for rural households (Table 2.1). Disparities in these estimates o f changes in poverty levels are mainly due to the differing inflation factors used to determine poverty lines.'' 8 See World Bank 2002,20; Government o f Pakistan2003; Palustan Economic Survey 2002,2003. 9 See Mal& 2005; Dorosh, Niazi, and Nazli 2003. 10 The Planning Commission estimates use the Consumer Price Index to express the 2001-02 poverty line in 2004-05 prices, following the same methodology used inpoverty estimates for earlier years. World Bank estimates use inflation rates calculated from price information collected as part o f the PSLM survey. Differences in definitions o f poverty lines in the base year and data cleaning protocols also contribute to differences inthe estimates. See World Bank 2006c for details. 8 Estimates of poverty in Pakistan vary Table 2.1: Poverty Estimates in Pakistan, 1998-99, considerably depending on the methodology 2001-02 and 2004-05 used, because a high percentage of rural 1998-99 2001-02 2004-05 households haveper capita expendituresclose poverty National 30.0 34.4 29.2 to the Official poverty line (Figure 2.1). 10.9 Headcount Urban 21.0 22.8 19.1 percent o f rural households in 2001-02 had per Rate Rural 33.8 39.1 34.0 capita expenditures within +/- 5 percent o f the National 6.3 7.0 6.1 official poverty line; in 2004-05, 8.95 percent of Poverty Gap Urban 4.3 4.6 3.9 rural households were within +/- 5 percent o f the Rural 7.1 8.0 1.2 Planning Commission official poverty line (Rs Squared Nation 2.0 2.1 2.0 878.6). Thus, small changes in calculated real PovertyGap Urban 1.3 1.4 1.2 incomes (expenditures), whether due to actual Rural 2.2 2.4 2.3 changes in expenditure, price deflators or other Source: World Bank staff estimates based on PIHS 1998-99, 00- methodological factors related to updating a 01; and PSLM 2004-05, (World Bank, 2006~). 6 - 9 I- --..- 1998/1999 - 2001/2002 2004/2005 I Source: HIES 1998-99,2000-01,2001-02; PSLM 2004-05. Notes: P1 i s the 20th percentile of 2004-05 expenditure distribution (809 Rs (2004-05)/adult equivalentlmonth); P2 i s the 40th percentile of 2004-05 expenditure distribution (995 Rs (2004- 05)ladult equivalentlmonth). The Planning Commission Poverty Line i s 937.5 Rs (2004-05)/adult equivalent'month, 5.7 percent less than the cutoff for the 40th percentile of the 2004-05 rural expenditure distribution. I1 Note, however, that focusing o n the bottom 40 percent of the per capita household expenditure distributiondoes not eliminate the issue o f the appropriate price index used incomparisons o f household expenditures over time (see below). 9 CHANGESINREAL INCOMES ACROSS HOUSEHOLD GROUPS The majority of Pakistan's rural poor are neither tenant farmers nor farm owners. Non- farm households (excluding agricultural laborer households) accounted for slightly over half (57 percent) o f the rural poor in 2004-05 (Table 2.2). Farmers comprised only 35 percent o f households inthe bottom 40 percent o frural per capita expenditure distribution. The remainder (8 percent were agricultural laborer households. This distribution o f rural poverty closely reflects land distribution, which i s highly unequal in Pakistan. According to the 2000 Agricultural Census, only 37 percent o f rural households owned land, and 61 percent o f land-owning households owned less than 5 acres, or 15 percent o f total land. Two percent o f households owned 50 acres, or 30 percent o f total land.12 Moreover, returns to land are estimated to be about half o f incomes (value added) from crop agriculture, with only about five percent o f value added paidto hired agricultural 1ab0r.l~ Table 2.2. RuralPoverty across Household Groups in Pakistan, 2004-05 Households Expenditures Poor (L40) Poor (L40) Poor (L40) (millions) (Rslperson) (percent) (milIions) % of Rural Poor Farm 5.65 1,346 27.1 1.53 34.9 Agricultural Laborers 0.72 1,028 50.3 0.36 8.2 Rural Non-Farm 6.68 1,209 37.3 2.49 56.9 Rural Self-Employed 2.22 1,244 31.2 0.69 15.8 Rural Non-Farm Other 4.46 1,190 40.3 1.80 41.I Total Rural 13.05 1,259 33.6 4.38 100.0 Source: HIES 1998-99,2000-01, 2001-02; PSLM 2004-05; andWorld Bank staff calculations. Moreover, non-farm income i s a major source of revenue, even for farmers with land. According to 2004-05 PSLM data, crop, livestock and agricultural wage labor incomes account for only 25, 8 and 4 percent respectively, o f total rural incomes; non-farm incomes (40 percent), remittances (9 percent), and "other income" (15 percent) comprise the remainder. Even for farm households, crop incomes account for only about half (49 percent) o f total income (Table 2.3).14 Estimates from a 2001-02 Social Accounting Matrix for Pakistan are consistent with national accounts figures for value added by sector," which indicates high proportions o f non-farm income in total rural incomes: 36 percent for farm households and 57 percent for all rural households (Dorosh, Niazi and Nazli 2003). Real per capita expenditures of the bottom two quintiles of rural households rose by 3.1 percent between 2001-02 and 2004-05, considerably less than the increase in per capita expenditures of rural households overall (5.9 percent) in the same period (Table 2.4).16 The 12 The overall Gini coefficient o flandownership in2000 inPakistanwas 0.66; including rural landless households, the Gini coefficient was 0.86. By comparison, the Ginicoefficient for land ownership inIndia is 0.71, inBangladesh, 0.42, andinBrazil, 0.85. See World Bank2004b. 13 This estimate is from a Social Accounting Matrix for Pakistan2001-02. See Dorosh, Niazi and Nazli 2003. 14 Unfortunately, data from the P S L M survey on land assets were not available for this report, so no analysis o f income or expenditure data by farm size is possible. 15 Note that P S L M 2004-05 income data show substantially lower earnings from livestock relative to crop agriculture than do the national accounts. 16 Itshouldbe emphasizedthat the bottomtwo quintiles o fthe per capita expenditure distributiondo not comprise the same households inthe two surveys, because many households likely moved into or out o f the bottomtwo quintiles during the period inquestion. 10 largest gains in real per capita expenditures (8.7 percent) were achieved by the richest quintile (the top 20 percent o f the household distribution); the poorest quintile saw the smallest gains (1.5 percent). Among the rural poor (the bottom two quintiles o f the rural per capita expenditure distribution), the real expenditure o f non-farm households rose by nearly twice as much as those o f rural farm households (3.6 percent versus 2.3 percent). However, comparing 2004-05 per capita expenditures with those of 1998-99, there was almost no change (only -0.2 percent) in real expenditures per capita of households of the poorest 40 percent of households. Inother words, the gain inreal per capita expenditures for the poorest 40 percent in the second period simply offset the loss in real per capita expenditures incurred in the first period. The same overall pattern holds for rural farmers and non-farmers (Figures 2.2 and 2.3). In contrast, the top 10 percent o f farmers (in terms o f gross crop income) gained 6.9 percent from 1998-99 to 2004-05; rural agricultural labor households suffered a 4.6 percent decline inthe same period. Figure 2.2. Real Per Capita RuralHousehold Expenditures inPakistan, 1998-99 to 2004-05 1,350 3E$!p:1,300 1,250 a 38 1,200 0 1,150 &! 1,100 v 1,050 Farmers Non-farmers Total Rural Figure 2.3. RealPer Capita RuralPoor Household Expenditures: 1998-99 to 2004-05 780 %t? E 770 760 $Q) 3 750 0 740 N 3* 730 720 Poor Farmers Poor Non-Farmers Total Rural Poor Source:PIHS 1998-99, 2001-02; PSLM2004-05; and World Bank staff calculations. 11 $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ + -sm9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 z z z g s z 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ~ ~ ~ 0 0 0 0 0 0 0 0 0 0 0 7 7 7 7 7 7 7 7 Y - 7 7 8 8 8 8 8 8 Y Y c q ? ? h ? ? ? ? ? ? 7 7 0 7 7 7 7 Similar patterns-declines in real expenditures per capita between 1998-99 and 2001-02 and subsequent increases from 2001-02 to 2004-05-are observable in each province. In Punjab, the total change in real expenditures per capita over the 1998-99 to 2004-05 period was small, only 0.3 percent (-1.1 percent for the poorest 40 percent" o f households). Likewise, in Sindh, the total change for all rural households and for households inthe poorest 40 percent o f the national rural distribution were small (0.3 and 2.2 percent, respectively), though farmers' gains were substantially larger at 7.4 percent. Farmers in the NWFP and the poorest 40 percent o f households also enjoyed substantial overall gains over the six-year period (11.O percent and 5.9 percent, respectively), although the total rural population gained only 1.5 percent. InBaluchistan, the large decline inper capita expenditures between 1998-99 and 2001-02 (-17.4 percent for the total rural population and -11.9 percent for the poorest 40 percent) outweighed the rather small improvements experienced by most household groups in the 2001-02 to 2004-05 period. As a result, average real per capita expenditures actually fell by 14.0 percent from 1999-2000 to 2004- 05 inrural Baluchistan(Tables A2.1-A2.4 and Annex Figures A 2.1-A 2.8). Changes in agricultural output and prices, driven to a large extent by weather, mirror the stagnation of real per capita expenditures, as well as their variation over the survey periods (Table 2.5 and Figure 2.4).'* There was substantial growth (4.2 percent per year) in real agricultural GDP between 2001-02 and 2004-05, but this growth rate does not reflect longer-term trends because real incomes and expenditures in 2001-02 were negatively affected by a drought that lowered crop production inmany parts o f Pakistan. Over the 1998-99 to 2004-05 period, the trend growth rate o f real agricultural value-added was only 2.3 percent per year, (and only 0.2 percent per year inper capita terms). Moreover, during these periods, agricultural prices declined relative to overall rural prices.Ig Deflating agricultural incomes (as measured by nominal GDP) by the overall rural price index, real agricultural income growth was only 0.2 percent (1.1 percent ifthe CPIused) andper capita agricultural incomes fell by 1.8 percentperyear. Fluctuations in wheat and cotton-the two most important crops in terms of gross value of production-were even larger than those of overall agricultural incomes, and help explain regional variations. Wheat productionrose by only 2 percent between 1998-99 and 2001-02, but then increasedby 16 percent between 2001-02 and 2004-05 (Annex Table A 2.5). Real wholesale prices also increased by 16 percent in this latter period, so that the value o f production increased by 34 percent, to the benefit o f net wheat sellers (84.1 percent o f producers and 32.0 percent of rural households, overall). Similarly, the real value o f cotton production, which had fallen by 9 percent between 1998-99 and 2001-02, increased by 29 percent in2004-05, as productionrose by 38 percent relative to 2001-02.20 This gain in gross cotton incomes helped boost real per capita expenditure o f farm households in cotton-wheat cropping system districts innorthern Sindh" by 22 percent. Given the large number o f farm households with expenditures just below the official 17 Data are for the poorest 40 percent o f the rural per capita expenditure distribution. 18 Note that expenditure fluctuations are expected to be smaller than income fluctuations because o f consumption smoothing (i.e. borrowing to offset declines inincome). 19 The rural price index used here i s basedon the national CPI for the first triennium and the survey- based price deflator for the second triennium. See Annex Table 2.1 for details. 20 Agricultural Census for 2000 data indicate little fluctuation in cropping pattern by size o f fann, suggesting that fluctuations in yields and prices may have been similar across a range o f farm sizes (World Bank, 2004b). However, larger farms' better access to water may have limited the decline in their yields during the 2001-02 drought. 21 The cotton-wheat cropping pattern region includes Khairpur, Sukkur, Nawabshah, Noshero Feroz, Ghotki, Hyderabad, Tharparkar and Sanghar. The 2004-05 PSLM survey sample included 532 farm households from these districts. 14 poverty line in 2001-02, gains in cotton incomes were likely a major factor behind the sharp reduction inrural poverty in Sindh.22 Table2.5. GrowthRatesofAgriculturalOutputandIncomeinPakistan, 199899to200405 1998-99 2001-02 1998-99 to 2001-02 to 2004-05 TO2004-05 Real Agricultural GDP 0.9% 4.2% 2.3% Agricultural Income I* -2.0% 4.8% 1.1% Agricultural Income II** -2.0% 2.9% 0.2% Per Capita Real Agricultural GDP -1.2% 2.2% 0.2% Per Capita Agricultural Income I* -4.1% 2.7% -1.O% Per Capita Agricultural Income II* -4.1% 0.9% -1.8% Source: PakistanEconomic Survey, various years; World Bank staff calculations. Notes: Nominalagricultural GDP divided by private consumption price index Eromnational accounts. ** Nominalagricultural * GDP divided by the Survey Based Price Index that uses PRHS 2001-02, 2005-06 price data. See Annex Table A 2.6 for details of the calculations. Other factors have also contributed to good overall economic performance inrecent years, as well as to gains in the real incomes (expenditures) o f the rural Workers' remittances (some o f which accrue directly to rural households), increased sharply from $1.1 to $4.2 billion between 2001-02 and 2002-03. However, as remittances have changed little between 2002-03 and 2004- 05, further large increases seem unlikely. Growing demand for labor inthe overall economy and in the construction sector in particular have also spurred urban and rural employment. Though increases in demand may have contributed to increased employment and incomes, they did not lead to increasedreal wage rates. Inthe 2001-02 to 2004-05 triennium, trend growth inreal wages was 0.4 percent using the CPI as a deflator, although real wage rates fell by 1.1percent per year using the national accounts price index o f personal consumption as a deflator (Table 2.6). Over the six-year period from 1998-99 to 2004-05, real wage rates declined when assessed by either measure, by -0.9 percent or by -2.4 percent per year, respectively. 22 UsingPIHS 2001-02 data, Orden and others (2005) estimate that a 10 percent increase in cotton prices leads to 11 percent and 6 percent increases in real incomes o f cotton farmers in Sindh and Punjab, respectively. Cotton farmers account for 13 percent o f the rural population nationally; poor cotton farmers are 3.5 percent o f the rural population. Given that many cotton farmers in Sin& have incomes just below the poverty line, the poverty rate in Sindh drops the poverty rate among cottonfarmers in Sindh fiom 50 percent to 39 percent (and from 36 percent to 31percent among cottonfarmers inPunjab). Nationally, rural poverty rates would fall by 0.5 percentage points. Unfortunately, no data on crop production was collected inthe 2004-05 survey, so it is not possible to compare real incomes or poverty rates among cotton farmers in2004-05 with data fromearlier years. 23 Smallholder farmers that derive much o f their income fiom cotton are likely to be an exception to this generalpattern. 15 Figure 2.4. Per CapitaAgriculturalOutputandRealIncomeinPakistan,199899to 2005-06 100 g i r k 95 90 .F Agric. GDP --e-Agric Income I* -Agric Income II* Source: Pakistan Economic Survey, various years; World Bank staff calculations. Notes: P indicates provisional data. * Nominal agricultural GDP divided by the price index for private consumption from the national accounts. ** Nominal agricultural GDP divided by the Survey Based Price Index that uses PRHS 2001-02,2005-06 price data. See Annex Table A 2.6 for details o f the calculations. Table 2.6. RealWage Rates of Unskilled Labor inPakistan, 1998-99 to 2004-05 1998-99 2001-02 1998-99 to 2001-02 to 2004-05 to 2004-05 1998-99 2001-02 2004-05 (% change) (% change) (% change) Wage Rate (Rslday) 116.5 126.3 152.2 -2.3% 5.9% 3.5% Real Wage (Rs 2004-05/day)* 153.7 148.8 152.2 -1.5% 0.4% -0.9% Real Wage (Rs 2004-05/day)** 166.9 155.8 152.2 -3.0% -1.I% -2.4% Source: Pakistan Economic Survey 2006; World Bank staff calculations. Notes: Wage rates are the average daily wage o f unskilledconstruction labor inNovember infive major cities (provincial capitals plus Islamabad). ***Consumer Price Index used as a deflator. National accounts personal consumption price index used as a deflator. Simulations using a computable general equilibrium(CGE) model for Pakistan also suggest that increases inworkers' remittances and other capital inflows were the dominant factor in explaining the gain in rural incomes between 2001-02 and 2004-05. Model simulations indicate that the increases in wheat, paddy and cotton productivity (and area harvested o f paddy) that occurred between 2001-02 and 2004-05 result in a 1 percent gain in both real incomes of rural poor households and o f rural poor farm h o ~ s e h o l d sAn~increase in workers' remittances . ~ has a much larger effect, however, raising incomes o f rural households by four percent, though ~ 24 Simulations o f these agricultural productivity shocks with a fixed-price semi-input-output (SIO) multiplier model result inlarger real income gains, about 5 percent for poor farmers and 3 percent for rural non-farm poor. This is so largely because prices do not decline as production increases, as occurs in the CGE simulations (particularly for cotton and rice). Note that these simulations model the productivity shocks as gains in output per capita with constant technology; the CGE simulations implicitly assume that land cultivated per capita does not change. 16 real incomes o f poor farmers rise by only three percent because o f a six-percent appreciation of the real exchange rate, which depresses real prices o f tradable agricultural products. Including the effects o f increased capital inflows and higher petroleum import prices, the total combined effects suggest that these four shocks raised rural incomes by 3.5 percent and real incomes o f poor farmers by about 2.5 percent (Figure 2.5). LONG-RUN DETERMINANTS POVERTYTRANSITIONS OF Analysis of a panel data set of rural households from four districts of Pakistan (Attock, Badin, Dir and Faisalabad) coveringthe periods 1986-87 to 1990-91 and 2001-0225indicates a decline in real incomes over time, though with considerablevariation across households and across districts (Annex Tables A 2.7 and A 2.8). Real incomes o f many households declined between the early 1990s and 2002, despite modest gains in provincial and national agricultural output. Net crop income increased by 38 percent for the total sample, and by 81 percent for poor farmers, whose total incomes rose by 23 percent. Nevertheless, rural non- agricultural incomes fell by 30 percent overall and by 16 percent for poor households. Figure2.5. CGEModelSimulationsof ShockstoKeyDeterminantsofRuralIncomeinPakistan I 5.0 , 1 4.0 Crop Productiity Shock 3 3.0 Increased Remittances m 5 2.0 r 1.0 Increased Capital Inflows U C 8 0.0 Petroleum Price 4 -1.0 b Rural Poor Rural Poor/Fahers Increase -2.0 - Combined Effects Source: Cororaton and Orden (2007). The panel data also indicate considerable movementby householdsinto and out of poverty, known as transitory poverty episodes (Figure 2.6). Although poverty rates tended to rise slightly over this period, (as reflected in the combined area of "poor" and "became poor" in Figure 2.6), about one quarter o f households moved into or out o f poverty each year. Between 1990-91 and 2001-02, however, the percentage o f "chronic poor" (households who were poor in both the current and preceding periods) was essentially the same (about 35 percent). Third, poverty in the sample rose substantially in 2001-02, though 15 percent o f households made a 25 The household data set used in this analysis is made up o f 14 rounds o f the International Food Policy Research Institute (IFPRI) sample from 1986/87 to 1990191, together with a sub-sample o f panel data households included in PRHS 2001102. The 571 household sample used in the analysis includes only the "base" households. Households that have split off from the base household are not included in this analysis. Note also that 103 households that had data for all five years o f the IFPRI survey could not be traced after 11 years. On average, these households were poorer than the average household that could be traced: Their real incomes per adult equivalent were 24 percent lower, at Rs (2002) 11,756 compared with 13,842 and their value o f household assets were 39 percent lower (Rs (2002) 160, 314 compared with 264,144. 17 transition out o f poverty in this 11-year period, essentially the same percentage as in the earlier period. Poverty transitions using the 1987 to 1991 average income and 2001-02 income also indicate an increase inpoverty, with 40 percent o f households falling below the poverty line and only 9 percent o f households escaping poverty (Figure 2.6).26 Regressionanalyses indicate that primary and secondary education, land ownership, village electrification and paved roads are all significant factors determining changes in household welfare over time. Using a two-period panel data set consisting o f the average o f 1989-90 to 1990-91 as the first period and 2001-02 as the second period, and controlling for random effects across hou~eholds,~~the presence o f an additional male with a secondary education in a household increases real expenditures (a measure o f long-term incomes) by 10.2 percent; having a household head with primary education increases real expenditures by 21.5 percent (Annex Table A 2.9, equation 4).28Owning five acres o f either irrigated or non-irrigated land (a small farm by Palustan standards) raises real expenditures by about 13 percent, or about half as much as having a household head with primary education. The regression coefficients also imply that village electrification raises real expenditures by about 75 percent and paved roads approximately double real expenditures. Table 2.7. RealIncomes in Four RuralDistricts inPakistan by source, 1986-87 to 1991-92 and 2001-02 Crops Livestock Non-Farm Remittances Total Full sample 1986-87 to 91-92 (Rdhh) 4,592 1,882 5,111 1,780 13,522 1986-87 to 91-92 (share) 0.340 0.139 0.382 0.132 1.ooo % change to 2002 22.2% -18.1% -30.1% -49.4% -13.1% Poor households 1986-87 to 91-92 (Rdhh) 1,920 1,476 3,028 648 7,239 1986-87 to 91-92 (share) 0.265 0.204 0.421 0.090 1.ooo YOchange to 2002 70.5% -6.9% -11.9% 2.5% 13.9% Poor farm households 1986-87 to 91-92 (Rdhh) 2,558 1,733 2,436 593 7,498 1986-87 to 91-92 (share) 0.341 0.231 0.328 0.079 1.ooo % change to 2002 72.5% 3.0% -11.1% 6.3% 23.1% Poor non-farm households 1986-87 to 91-92 (Rdhh) 237 797 4,588 793 6,556 1986-87 to 91-92 (share) 0.036 0.122 0.700 0.121 1.ooo % change to 2002 - 15.9% -63.4% -13.1% -5.0% -14.0% Source: Dorosh and Malik 2006. Notes: Calculated from IFPRIpanel data set andPRHS 2001-02 for Attock, Badin, Dir and Faisalabaddistricts. Agricultural wage incomes, which account for less than two percentof total incomes, are not shown. 26 InFigure 2.6, the percentage ofhouseholds escaping poverty is indicatedby"Became Non-Poor" 27 Village level dummy variables are also includedto control for unobserved fixed effects. 28 Comparing separate regressions from the two periods (Annex Table 2.9, columns 2 and 3), the effect o f education on the headof household appears to have increased over time. 18 i 100% 90% f! UI 80% c 0 70% a 60% a, I50% 2 Y- O CI 40% 1987188 21301iQ2 : 4 Pt 8to thc 19 Box 2.1. Human Capital and HouseholdIncomes inRural Pakistan Usingdata from an International Food Policy ResearchInstitute (IFPRI) survey o f nearly 1000 rural households in four districts o f Pakistan from 1986 to 1989, Fafchamps and Quisumbing(1999) found that education has significant effects on rural incomes. Controlling for the effects o f background characteristics (land owned by father, inherited land, father's and mother's schooling) and innate ability (measured using the results o f an administered test (Raven's Colored Progressive Matrices Test), one additional year o f schooling for all adult males in a household (the mean number o f adult males in a household in the sample was 2.0) raises household incomes by 8.9 percent. One-fifth o f this additional income ensues from the re- allocation o f labor away from agricultural activities and toward non-farm work. The remainder i s due to higher non-farm productivity. However, this study found little evidence o f the effect o f male education on agricultural productivity, nor o f female education on agricultural or non-farm productivity, though. It should be noted, however, that this analysis o f household data derives from the context o f a "rural labor market with a very low supply o f educated people and a mediocre nutritional status in general." (Fafchamps and Quisumbing 1999, 401) Given this environment, the marginal economic returns to education and nutrition for individuals are high. However, these economic returns would almost certainly be lower if a large number o f individuals (relative to the size o f the labor market) improved their skills and health status, as demand for labor might be insufficient to match the increase in effective labor supply without a decline in real wages. Thus, as the proportion o f individuals in the labor force with improved skills and health status in Pakistan increases over time, the marginal returns to education and nutrition may diminish. Changes in the structure o f economic output and increases in the share o f slulled jobs in overall employment would tend to increase returns to education over time, however. Source: Fafchamps and Quisumbing1999. NON-MONETARY INDICATORSOFWELFARE AND POVERTY Despite essentially no change in the real per capita expendituresof the poorest 40 percent of the rural population between 1998-99 and 2004-05, there have been major improvements in other welfare indicators (Table 2.9 and Figure 2.7). The gross primary school enrollment rate rose 16 percentage points, from 63 to 79 percent over this period; net enrollment rates rose by 11 percentage points, from 37 to 48 percent.29The proportion o f fully immunized children increased even more sharply, from 55 to 72 percent. Access to tap water nearly doubled, from 12 to 23 percent. For all non-monetary measures o f welfare shown, however, urban residents fare much better than do rural residents, particularly with respect to access to tap water, adult literacy rates, and gross primary school enrollment rates. 29 Gross primary school enrollment rates, defined as the number o f children attending primary level (class 1-5) dividedby the number o f children aged 5-9 years, are an indicator o f total attendance rates, and include children above the age cutoff. Net primaryschool enrollment rates are ratios and cover only the number o f children aged 5-9 who attend primary school, and thus exclude children above 9 years or below 5 years. 20 Table 2.9. SocialIndicators of RuralHouseholds inPakistan, 1988-99,2001-02 and 2004-05 Rural Rural Rural Urban Rural-Urban 1998-99 2001-02 2004-05 2004-05 2004-05 Gross EnroIIment Rate 63 66 79 104 -25 Net Enrollment Rate 37 38 48 64 -16 AduIt Literacy Rate 33 34 40 69 -29 Fully Immunized Children 45 46 72 87 -15 Electricityfor Lighting --- --- 74 96 -22 Access to Tap Water 12 I O 23 60 -37 Real Per Capita Expenditures 1268 1189 1259 1818 -31% Bottom 40 Percent 775 750 773 824 -6% Source: PIHS 1998-99,2001-02; PSLM 2004-05. Notes: * Realper capitaexpenditures of the bottom40 percent inRs (2004-05)/capita. GrossEnrollmentrate is the number of children attendingprimarylevel (classes 1-5)dividedby the number of children aged 5-9 years. Net EnrollmentRate is the number of children aged 5-9 years attending primary level (classes 1-5)dividedby the total number of children aged 5-9 years. Fully Immunized Children: The percentage of children aged 12-23 months that have been fully immunizedbased on recordor recall. Large disparitiespersist betweenmale and female education and literacy rates (Table 2.10). For girls, gross primary school enrollment rates rose by 18 percentage points, from 50 percent in 1998-99 to 68 percent in2004-05, but they remain 21 percentage points below those for boys (89 percent). Similarly, net primary school enrollment rates for girls rose by 12 percentage points, to 42 percent in 2004-05, but this i s still 11percentage points below rates for boys. Similar patterns are observed for middle school enrollment rates in rural areas: 29 percent for girls as compared with 46 percent for boys. Gender gaps are declining only slowly: the disparity in gross primary and middle school enrollment rates declined by a mere five percentage points between 1998-99 and 2004-05. Gender gaps innet enrollment rates fell by only one or two percentage points. Adult literacy o f rural women increased by 8 percentage points, to 24 percent in 2004-05, as compared to adult literacy o f rural men, which rose by 5 percentage points to 56 percent. Despite progress, Pakistan's social indicators still lag those of several other South Asian countries(See Figure 1and Table 2.11). According to PSLM 2005-06 data, the primary school net enrollment rate for girls in Pakistan in rural areas i s only 42 percent (48 percent nationally), compared to 86 percent in India (national), 84 percent in Bangladesh (national) and 98 percent in Sri Lanka (national). National infant mortality per 1000 live births i s 80 in Pakistan (88 in rural areas) whereas it i s only 62 and 12 in India and Sri Lanka, respectively. The infant mortality rate in Bangladesh is 56 per 1000 live births, although Bangladesh's per capita income of $187O/person i s 84 percent that o f Pakistan ($2225/person). These indicators suggest that there i s considerable scope for continued improvement inthese welfare indicators inPakistan. 21 Table 2.10. Education and Literacy Rates inPakistan, 1998-99,2004-05 1998-99 PIHS 2004-05 PSLM MALE FEMALE TOTAL MALE FEMALE TOTAL Gross Primary Level Enrollment Rates (age 5-9) All Pakistan 80 61 71 94 77 86 Rural 75 50 63 89 68 79 Urban 95 92 94 107 100 104 Net Primary Level Enrollment Rates (age 5-9) All Pakistan 47 37 42 56 48 52 Rural 43 30 37 53 42 48 Urban 58 56 57 66 63 64 Gross Middle Level Enrollment Rates (age 10-12) All Pakistan 48 32 40 51 40 46 Rural 43 21 32 46 29 38 Urban 62 60 61 64 63 64 Net Middle Level Enrollment Rates (age 10-12) All Pakistan 19 13 16 20 16 18 Rural 16 9 13 17 11 14 Urban 27 25 27 27 27 Adult Literacy -2615 Years and Older All Pakistan 58 28 43 63 36 50 Rural 51 16 33 56 24 40 Urban 73 53 63 78 59 69 Source: PIHS 1998-99,2001-02; PSLM2004-05. Figure .2.7. Per Capita GDP and SocialIndicators in South Asia 120 n 80 61 Bangladesl 60 India Pakistan 40 oSri Lanka 20 0 Per Capita GDP Net Primary Infant Mortality Index School Enrollment Rate Rate (Girls) Source: World Bank World Development Indicators 2006; UNESCO 2006; PSLM 2004-05; Deolalikar 2005. Notes: Per Capita GDP Index figures for Bangladeshare for 2003, and for India, Pakistan and Sri Lanka are ' for 2004. Net Primary School Enrollment rates for girls figures for Bangladeshare for 2002, for India and Pakistan are for 2003, and for Sri Lanka, 2004. Infant mortality rates figures for Bangladeshare for 2003, for India, Pakistanand Sri Lanka are for 2004. 22 Box 2.2: Reducing Infant Mortality in Pakistan Econometric analysis o f household data from the Pakistan Socio-Economic Survey (PSES), 2000- 01 indicates that household income i s not a major determinant o f infant mortality in Pakistan and that it i s only one o f several determinants o f child malnutrition o f children. The survey data indicate that the average infant mortality rate inrural areas in 2000-01 (covering children born over the 1991-99 period) was 85 per 1000 live births, (compared to 61 in urban areas). The major determinants o f infant mortality (defined as the probability o f an infant dying duringthe first 12months o f his or her life) include: household sanitation (having a flushtoilet in the household correlates with a decline in infant mortality o f 15 per 1000 live births); immunizations (a 1-percent increase in the proportion o f children immunized in a district results in a decrease of 9.6 deaths per 1000 live births); and level of mother's education (post-middle school and above). Controlling for other factors, infant mortality rates actually increase as per capita expenditures rise, although this i s likely due to spurious correlation. Nonetheless, results strongly suggest that public health measures (immunizations), household sanitation and mother's education all contribute to reducing infant mortality, even inthe absence o f income growth. As expected, the probability o f a child aged 0-59 months being underweight falls with rising per capita household consumption: a 1 percent increase in per capita household consumption i s associated with a 7.6 percentage point decrease in the probability o f a child being underweight, from an average o f 47.5 percent (PSES 2001-02) to 39.9 percent. But other factors such as mother's formal schooling, electrification o f the home and the installation o f a flush toilet in the household each contributes to a five percent decline in the proportion o f underweight children). Clearly, public investments and public-service delivery related to formal schooling, electricity and sanitation can have a major impact on malnutrition, even inthe absence o f household income growth. Source: Deolalikar, A., May 2005. Attaining the Millennium Development Goals in Pakistan. South Asia HumanDevelopment Sector, Discussion Paper Series Report No. 8. Washington, D.C.: World Bank. SUMMARY Survey evidence points to an encouraging 5.1 percentage point decline inrural poverty, from 39.1 percent to 34.0 percent between 2001-02 and 2004-05, and a 3.1 percent increase in the incomes o f the poorest 40 percent of the rural income distribution (World Bank estimates). However, longer-term trends (1998-99 to 2004-05) show only very small changes in real per capita expenditures o f the rural poor or in the level o f rural poverty. Analysis o f the structure of household incomes and sectoral output suggest that much improvement in the three-year period reflects a dip in 2001-02 incomes due to drought and lower cotton production. Nonetheless, in spite o f the relatively small change in real per capita expenditure over the six-year period, there have been marked improvements in non-monetary welfare indicators, including child immunization, access to drinking water, and electrification o f the home. Moreover, educational levels, which are a major factor in raising incomes, have also risen, suggesting improved prospects for future income growth. 24 7 7 7 7 0 d .-J C PL. 9 8 O m cy 0 0 cy Q) Q) Q) 7 0 9 E I I I I cor-m h l m o hl. . . 7- 7 7 7 i z n v ) D !?% O U 0 ? Z Y v ) 3 0 3 0 Y e l U 0 n m POD ? !!EO r U 0 z v) 0 0 (Y el ? 6 0 (Y Q) 2 Q) Q) 7 r r II 8 "m"o"o 0 0 Q,Q)cnQ) Q, 0 m 0 0 0 0 c C c c 'FE c a 6 : ita C U : a C ! i i I t c d c d! c 3 m I +C i! I i I C i C I c de r c i :C E 030s a 0 t0 2 - 1 * * * * * * * * * * * * * * * * * * * * * * + I -4 G P I C eIE G * ** * * * * * P * * * * I c d c dt e I r 3! ** ** ** * * * * 7 si m m 7I -; C dc * * * * * * * * * * * * * * * * * * * C c ';i I L 2 C c -2 2c c Figure A 2-1. RealPer Capita HouseholdExpendituresofRuralHouseholdsinPunjab, Pakistan, 1998-99,2001-02 and 2004-05 1,400 1,350 . 1,300 1,250 1,200 1,150 1,100 1,050 Farmers Non-Farmers Total Rural I El!1998-99 W 2001-02 0 2004-05 1 FigureA 2-2. RealPer Capita HouseholdExpendituresofRuralPoor Householdsin Punjab, Pakistan: 1998-99,2001-02 and 2004-05 790 780 770 760 750 740 730 720 Poor Farmers Poor Non-Farmers Total Rural Poor 1E4 1998-99 W 2001-02 0 2004-05 1 34 Figure A 2-3. RealPer Capita Household Expenditures of Rural Householdsin Sindh, Pakistan, 1998-99,2001-02 and 2004-05 1,400 n 5S 1,350 I 1,300 s 0 1,250 Q) P 3j 4 1,200 0 1,150 N $ 1,100 v 1,050 Farmers Non-Farmers Total Rural 1998-99 2001-02 0 2004-05 I Figure A 2-4. Sindh: Real Per Capita Household Expenditures of Rural Poor Households: 1998-99,2001-02 and 2004-05 790 780 770 760 750 740 730 720 Poor Farmers Poor Non-Farmers Total Rural Poor 35 Figure A 2-5. RealPer Capita HouseholdExpendituresof Rural HouseholdsinNWFP, Pakistan, 1998-99,2001-02 and 2004-05 1,250 I I 1,200 1,150 1,100 1,050 1,000 Farmers Non-Farmers Total Rural 1 1998-99 2001-02 0 2004-05 I Figure A 2-6. NWFP: RealPer Capita HouseholdExpendituresof RuralPoor Households: 1998-99,2001-02 and 2004-05 Poor Farmers Poor Non-Farmers Total Rural Poor 1 1998-99 2001-02 2004-05 1 36 FigureA 2-7. Baluchistan: RealPer CapitaHouseholdExpendituresofRuralHouseholds: 1998-99,2001-02 and 2004-05 1,500 1,400 1,300 1,200 1,100 1,000 Farmers Non-Farmers Total Rural 1998-99 2001-02 2004-05 I FigureA 2-8. RealPer CapitaHouseholdExpenditures of RuralPoor Householdsin Baluchistan,Pakistan, 1998-99,2001-02 and 2004-05 950 900 850 800 750 700 Poor Farmers Poor Non-Farmers Total Rural Poor 3 1 3. PROMOTINGAGRICULTURAL GROWTH Pakistan's agricultural sector accounts for about 70 percent o f rural household income and nearly one-quarter o f national GDP. Raising agricultural productivity i s thus crucial for rapid rural growth. This chapter begins with analyses o f total factor productivity in the sector; production constraints; potential for growth in output o f major crops and livestock products; and pertinent agricultural technology issues. The next section examines the availability of water and the efficiency o f its use for irrigation, which i s the biggest constraint to production for much o f Pakistan's agriculture. The third section covers existing agricultural markets and trade policies, which are crucial if productivity gains are to be translated into real income gains. The prospects for raising real incomes o f the rural poor through agricultural growth are analyzed in section four. The chapter concludes with a brief summary o fthe main implications for policy. AGRICULTURAL PRODUCTIONAND TECHNOLOGY Annual agricultural growth in Pakistan averaged 3.7 percent over the roughly four decades from 1959-60 to 2001-2002, although there were wide year-to-year variations. Apart from a period o f slow growth in the first half o f the 1970s, average agricultural growth exceeded 3.2 percent annually in each quinquenniumfrom 1960 to 2000, due inlarge part to highgrowth inthe crop sector in the 1970s and 1980s as a result o f Green Revolution technology (improved seeds, increased fertilizer use, and irrigation). However, the performance o f agriculture (particularly the crop sub-sector), has suffered inrecent years because o f severe droughts inthe country, as well as environmental factors (increased soil salinity and deteriorating groundwater quality). Pakistan's agricultural sector grew at a modest rate of 2.6 percent per year from 1999-2000 to 2005-06, (0.3 percent per year on a per capita basis, Table 3.1). Real value added o f major crops (wheat, basmati and other rice, cotton, sugar cane and maize), which accounted for about two-thirds o f agricultural crop GDP, grew by 2.6 percent per year over the period. There have been substantial fluctuations in real crop GDP in recent years due largely to variations in water availability. For example, real crop income fell by 3.6 percent in 2005-06, after it rose by 17.8 percent the previous year due to a record cotton crop (production increased by 42 percent). Total crop GDP grew by 2.3 percent annually, almost entirely due to a 2.1 percent yearly increase in crop GDP/hectare, while the cropped area increased by only 0.2 percent per year. Livestock (which accounts for half o f agricultural GDP) grew by an average o f 3.5 percent per year (Figure 3.1). The rapid growth in output during the Green Revolution of the 1970s and 1980s was due both to a large expansion in inputs (seed, fertilizer, irrigation), as well as productivity. Major investments in land, and most importantly in the Pakistan context, in water supply, particularly inthe form o f tube wells facilitated increases inthe net area Non-factor input 30 Increases in cropping intensity, (net gross cultivated area divided by net area sown), through multiple-cropping were also made possible, in part, by increases in water availability. In Ali (2005), land use i s defined in terms of net area sown, so increases in cropping intensity are captured mainly through increases in land and total factor productivity. In Ali and Byerlee 2002a, land use i s measured as a flow variable, (the rental value of the land for a period). By this measure, land use rose rapidly over the past three decades, whereas the measured increase inland and total productivity i s smaller. 38 use (fertilizer, pesticide, tractor services) also expanded rapidly. Nonetheless, total factor productivity in crop agriculture, as well as total agriculture, have increased substantially since about 1970, with estimates ranging from about 1.5 to 2.3 percent per year, depending on the definitions of variables, methodologies and time periods o f analysis.31 Table3.1. AgriculturalGDP GrowthinPakistan, 1999-2000to 2005-06* 2005-06 2005-06 2005-06 2000-06 Value Added Share Share Growth (Rs billion) Agric. GDP Total GDP Rate Agriculture 1055.2 100.0% 21.6% 2.6% Major Crops 371.1 35.2% 7.6% 2.8% Minor Crops 129.9 12.3% 2.7% 0.9% Livestock 523.5 49.6% 10.7% 3.5% Fishing 14.2 1.3% 0.3% -0.9% Forestry 16.5 1.6% 0.3% -5.7% Industry 1270.1 ___ 26.0% 7.9% Services 2554.2 --- 52.3% 5.9% GDP (factor cost) 4879.5 --- 100.0% 5.6% Population (000,000) 155.4 --- --_ 2.1% Agric GDPlcapita (Rs) 6790 _ _ _ __- 0.5% Cropped Area (000,000 has) 22.5 --- --- 0.2% Crop GDPlHa (000 Rs) 22.3 __- --- 2.1% Source: PakistanEconomic Survey 2005-06. Notes: * 2005-06 data are provisional. Econometric analysis suggests, however, that since the early 1990s Punjab's total factor productivity (TFP) in the crop sector has at best remained constant, and may even have declined. From 1993-2003 decade, total input use increasedby 1.80 percent per year, but output increased by only 1.69 percent per year. Thus, total factor productivity inPunjab crop agriculture actually fell by 0.11 percentage points per year in this period (Table 3.2; Ali and Byerlee 2002, 2004). Severe droughts in several years are part o f the explanation. There i s also evidence of long-term deterioration in water and soil quality. Reduced effectiveness o f agricultural research and extension services also played a role. Labor productivity (measured as a partial factor productivity) also declined, althoughthis could be due to a combination o f a deteriorating natural- resource base, drought, and increased underemployment in agriculture due to a slowdown inrural non-farm employment and out-migration. 31 For details, see Annexes 2.8, 2.9 and2.10 o fWorldBank 2006. Pakistan: Growth and Export Competitiveness. 39 Figure 3.1. Real Agricultural GDP inPakistan, 1989-90 to 2004-05 I 1200 1000 m h u) 0 800 L 3 s 600 Q) - 400 Pc 200 0 +Agriculture -dr Major Crops Minor Crops +Liwstock Source: Pakistan Economic Survey 2005-06. Notes: P denotes provisional data Table 3.2. Estimates of Total Factor ProductivityGrowth of Agriculture inPunjab, Pakistan Crops Livestock Agriculture output TFP TFPlOutput TFP TFP Green Revolution, 1966-74 3.23 -1.38 -42.7% -0.38ns -1.04 Input-Intensification, 1975-84 2.84 1.40 49.3% 0.71 1.37 Post-Green Revolution, 1985-94 3.27 2.86 87.5% 3.29 3.51 Total Period, 1966-94 3.23 1.26 39.0% 1.25 1.51 1971-81 1982-92 1993-2003 1971-2003 3.01 1.54 51.2% _ _ _ _ _--- Source: Ali and Byerlee 2002,2004. Notes: ns denotes not significantly different from zero. Data on output, input and TFP changes across agro-ecological zones inPunjab suggeststhat fluctuation in cotton yields i s a major cause of the stagnation in overall TFP productivity in the crop sector. Output and TFP growth were highest in the 1982-92 period, (3.91 and 1.54 percent, respectively), a period in which output and TFP growth in the wheat-cotton cropping systemwere 5.10 and 3.96 percent, respectively.Yield growth of cotton inPunjab, overall, inthis periodwas 9.29 percent per year (1981-90). By contrast, inthe 1991-2002 period, cotton yields fell by 2.59 percent per year due to virus problems and poor weather. Improving cotton yields through better pest control and improved seeds would likely substantially add to overall TFP growth inPunjab (as well as innorthern Sindh). 40 There are no detailed analyses of productivity growth in other provincesbut available data suggest that water availability i s the main constraint in Sindh. Wheat output in Sindh has fluctuated in tandem with water availability, particularly during the 1999-2000 to 2001-02 drought years. From 1972 to 1996, canal water releases in Sindh were negatively correlated with rainfall, thus stabilizing total water availability in the province. From 1995 to 2001, however, releases were positively correlated with rainfall, and remained essentially constant at about 50 MAF/year fkom 1995-98, which were years o f normal rainfall, but fell to about 35 MAF in 1999- 2001, years o f protracted drought that reduced availability o f canal water as well. Analysis o f average water productivity in Sindh indicates that cereal yields are higher in areas o f above average rainfall; however cereal yields are negatively correlated with the availability o f canal water. In contrast, data for area under cultivation are positively correlated with both rainfall and availability o f canal water.32 As a result, total production o f cereals i s not highly correlated with rainfall or canal water availability, except during the drought years o f 1999-2000 to 2001-02 when wheat production fell by 23 percent relative to the 1996-97 to 1998-99 average. Nonetheless, there is evidence in Sindh of substantial inefficiency inwater use. Net returns to irrigation water use vary widely across crops, which suggests that substantial savings in water would be possible by shifting away from water-intensive ordinary (non-aromatic) rice, which has a net income o f 66 Rs/acre-inch and water use o f 64 acre-inches, to cotton (with a net income of 361 Rs/acre-inch and water use o f only 20 acre-inches), inareas with adequate drainage. The Crop Sub-Sector Cropping patterns vary across Pakistan dependingon soil type and water availability. Inthe northern irrigated areas o f Punjab, basmati rice i s cultivated in the monsoon season (kharzj), followed by wheat in the winter season (rabi).Insouthern Punjab and northern Sindh, cotton i s the major kharif crop, typically followed by wheat. Further south, in southern Sindh, where drainage problems inhibit cotton cultivation and warm nighttime temperatures constrain wheat yields, ordinary (non-aromatic) rice i s the major crop. Sugar cane, a long-duration, water- intensive crop, i s grown almost exclusively on irrigated land. Non-irrigated areas, including hilly and mountainous regions o f NWFP and Baluchistan, typically have more diverse cropping patterns that include wheat, maize and pulses. From 1990-91 to 2005-06, average grain yields continued to increase in Pakistan, albeit at rates below those of the 1980s. The fastest yield gains were those for maize, 4.7 percent per year and basmati rice, 3.6 percent annually (Table 3.3). Yields o f wheat (the major crop in Pakistan in terms o f both area cultivated-37 percent o f total area cultivated in 2003-04 to 2005-06-and in terms o f share o f agricultural GDP, 14 percent in 2004-05) also increased, by an average o f 2.2 percent per year. Yield growth for other crops was generally much slower. Yields o f other food grains (mainly millet and sorghum) increased by only 0.6 percent per year; cotton yields . increased by only 0.8 percent per year, and yields o f sugar cane increasedby only 0.9 percent per year. 32 See Damania, R., "Water andAgriculturalProductivityin Sindh," DRAFT, April 2005. 41 Table 3.3. Pakistan: Area, Yield and Production of Major Crops, 2004-06 (average) Area Area Yield Yield Production Production (000,000 ha.) (YOgrowth) (tonslha) (YOgrowth) (000,000 tons) (% growth) Wheat 8,292 0.19 2.52 2.19 20,937 2.39 Rice 2,534 1.33 2.03 1.92 5,140 3.28 Basmati 1,539 2.51 1.65 3.60 2,538 6.20 Other rice 951 -0.52 2.52 1.45 2,398 0.93 Maize 984 1.01 2.80 4.74 2,751 5.80 Other Food 850 -1.37 0.60 0.55 514 -0.68 grains Gram 1,047 -0.38 0.64 2.08 669 1.70 Sugar cane 982 0.71 49.19 0.92 48,325 1.64 Oilseeds 315 -0.84 0.74 0.80 234 0.22 Cotton 3,093 0.77 0.67 0.79 2,086 1.56 Tobacco 51 -0.09 2.01 1.05 102 0.96 Other 4,506 -0.66 _ _ _ Total 22,653 0.15 _ _ _ _ _ _ --- _ _ _ Source: PakistanEconomic Survey 2005-06. Notes: Growthrates are averages, 1990-91to 2005-06; 2005-06 data are provisional Basmati and other rice figures are for data through 2004-05. Other Food grains consists of bajra,jowar and barley. Oilseeds includesrapeseed, mustard and sesamum. Other includes fruits, vegetables and all other crops. Further increases in wheat yields are possible, but water availability and quality remain major constraints. Between 1998 and 2002, average national wheat yields on irrigated land in Pakistan (2.5 tonsha) were significantly below average yields in Punjab, India (4.4 tonsha) and Haryana, also in India (4.0 tonsha), (Table 3.4). More disaggregated analysis, however, shows that wheat yields in Haryana (4.48 tonsha) and in the Jhelum canal system inPunjab, Pakistan33 (4.1 1 tons/ha) are similar, with high variation in yield across watercourses within a given distributary, especially in Pakistan (Hussain and others 2003). Major inequities in canal-water distribution, as well as significant variability in groundwater quality, are major factors. Insystems with adequate canal water, reallocation to areas where groundwater quality i s low may substantially increase average yields. There is considerablepotential to increase average cotton yields, which fluctuated between 0.49 and 0.62 tonsthectarefrom 1992-93 to 2003-04. Good weather and low incidence o f pests contributed to a rise in cotton yields in 2004-2005 o f 0.20 tonsha (33 percent) to 0.76 tonshectare, almost equal to the record yield of 1991-92 (0.77 tonslhectare). Sustained yield increases in yields, however, will require continued successful pest control through the use of improved varieties, pesticides andor integrated pest-management techniques. Improvements in groundwater quality are also needed to help prevent yield declines during years o f low rainfall. Nonetheless, average cotton yields in Pakistan (0.58 tons/ha) are nearly double those of Punjab and Haryana inIndia (0.3 1and 0.32 tonsha, respectively). 33 The data are from a sample o f 216 farms in the Bhakra canal system in Haryana, India and 218 farms inthe lower Jhelum canal systemo f Punjab, Pakistan inthe October 2000 to May 2001 rabi season. 42 Table 3.4. Average Crop Yields inPakistan and other Countries, 1997-98 to 2001-02 Wheat Rice Cotton Sugarcane Maize Haryana, India 3.99 2.53 0.32 55.74 2.23 Punjab, India 4.39 3.40 0.31 60.96 2.49 Gujarat, India 2.32 1.41 0.26 70.63 1.48 Rajasthan, India 2.54 1.17 0.22 45.36 1.18 Maharashtra, India 1.24 1.60 0.13 84.71 1.47 India (average, 5 states) 2.90 2.02 0.25 63.48 1.77 India (natl. average) 2.67 1.96 0.21 69.95 2.23 Bangladesh 2.19 2.11 --- 40.51 2.98 Punjab, Pakistan 2.42 1.50 0.56 44.36 1.91 High yielda 2.57 1.86 --_ Low yieldb 1.14 1.39 _-_ Sindh, Pakistan 2.51 2.82 0.67 56.8 0.55 High yielda 2.56 2.82 --- Low yieldb 1.26 _ _ _ _ _ _ All Pakistan 2.30 1.94 0.58 47.47 1.72 High yielda 2.52 2.50 Low yieldb 1.oo 1.42 __ __ __ PakistaniIndia 86.1% 98.9% 278.0% 67.9% 77.0% Puniab PaWPuniab Ind 55.0% 44.1% 180.6% 72.8% 76.7% Source: India datafrom CMIE data files; Bangladesh(BBS);PakistanMINFAL; authors' calculations. Notes aHigh yield wheat figuresare for irrigatedwheat; high yield rice figures are for ordinary rice. Low yield wheat figures are for non-irrigatedwheat; low yield rice figures are for basmatirice. The mostpromisingstrategyfor raising agricultural crop incomesinthe future, however,is diversification into higher value crops that can maximize returns to scarce factors of production (land and water). Domestic demand for these products will continue to rise rapidly as incomes increase and diets become more diversified, provided marketing channels function efficiently. Capturing these opportunities will require greater investment in agricultural research for higher value crops such as oilseeds,34 vegetables, fruits and livestock; better coordination between research, extension and the private sector through public-private partnerships; and improvements in seed certification and quality control (Box 3.1). Pakistan is also well-positioned to compete in expanding export markets for citrus, dates and other fruits, particularly in the Middle East. Although production has stagnated inrecent years, exports o f higher-quality kinnos (a type o f orange) have increased sharply. This export development has been private sector-led: all marketing o f kinnos for export i s done through sales to brokers or directly to factories, rather than through the domestic wholesale market. To ensure 34 As mentioned below, oilseeds production already benefits fiom trade protection of processed vegetable oil. Increases inproductivity and improved marketing could potentially make vegetable oil fiom domestic oilseeds competitive with imported vegetable oil even without trade protection, however. 43 Box3.1:India'sNationalAgriculturalInnovationProject The National Agricultural InnovationProject (NAP) is a U S $250 million Government o f India project (with $50 million in financing from the World Bank) designed to help transform India's agricultural research and development efforts from their traditional focus on food self-sufficiency to a more market-oriented approach aimed at poverty alleviation and income generation. Inthe last ten years the Indian agricultural sector has undergone considerable change. National food self-sufficiency has been achieved and market demand for fruits, vegetables, meat, eggs, fish, milk, sugar, and edible oils has increased and is encouraging diversification o f production. Given that there i s limited scope for expansion o f area cultivated, enhanced productivity, profitability, and competitiveness will be the main sources o f future agricultural growth. While agricultural development has long been supply oriented, its future will be more market driven. The challenge for Indian agriculture is to increase profitability, competitiveness and income earning capacity in a market-driven system for the 200 million poor people dependant on agriculture for their livelihoods. Activities. NAP will support the development and implementation o f agricultural innovations through collaboration among farmers, the private sector, civil society, and public-sector organizations. The project will be implemented from 2006 through 2012 and i s designed around the four components. Indian Council for Agricultural Research (ICAR) as the catalyzing agent for management of change in the Indian National Agricultural Research System (NARS): ICAR has to provide leadership and empower the NARS and the expanding group o f research providers to adapt to emerging challenges. Through this component the ICAR system will also increase its capacity for dialogue and interaction. Research on production to consumption systems: The specific objective of this component i s to establish market-oriented collaborative research for sustainable improvement o f selected agricultural production to consumption systems, in recognition o f the growing importance o f the market. Research on sustainable rural livelihood security: The core objective of this component i s the sustained improvement o f incomes and well-being o f farm families that have been left behind indevelopment. Basic and strategic research in the frontier areas of agricultural science: This component i s intended to strengthen the capacity o f Indian NARS to participate at the forefront o f science, in recognition o f the need to develop capacity for solving the problems o f tomorrow and the need to be a global leader in the field o f agncultural research. NAP is open to all Indian scientific organizations and will actively engage the private sector, farmer organizations and NGOs in a series o f large consortia (defined as "a formal group o f likeminded partners that will carry on a given task by sharing a research agenda and research resources as per an agreed work plan") principally selected through competitive processes. NAP has put in place a series o f support mechanisms to ensure that every organization interested to participate will be able to compete. Source:WorldBank205e. IndiaNationalAgriculturalInnovationProject ProjectInformationDocument(PID). 44 quality control, harvesting o f kinnos for export i s done by teams hired directly by brokers or factories. Currently, almost all kinnos grown for export are produced on large farms. Increasing the participation o f small farmers in kinno production for export will likely require creation o f some form o f farmer organization to help achieve economies o f scale in quality control and processing. Public support for research on varietal improvement, which would benefit all kinno farmers, may also be warranted. Faster growth in non-traditional agricultural exports (fruits and vegetables, meat, fish) will require improved capacity to meet Sanitary and Phytosanitary (SPS) standards imposed by importing countries. Pakistan has already faced restrictions on exports o f its shrimp because o f an inability to meet European Union SPS standards (Box 3.2). Quality and SPS standards issues are likely to be a major constraint on expansion o f Pakistan's high-value agricultural exports unless Pakistan develops and implements a coherent strategy (or set o f strategies) for quality and SPS management. Organizationally, there i s a need to remedy the fragmented, isolated, and non- coordinated capacity o f SPS management institutions and regulations by better defining the roles and responsibilities o f various agencies involved in SPS management, strengthening technical capacity for risk assessment, and re-institutionalizing early warning and surveillance systems for pests, diseases and contaminants. Livestock and Poultry The livestock sub-sector (cattle, dairy, sheep, goats, and poultry), which is dominated by small holders, accounts for half of agricultural GDP, (11 percent of total GDP), and is the fastest growing component of the agricultural sector. From 1989-90 to 2005-06, the average growth rate o f the livestock sub-sector was 3.5 percent, outpacing official population growth o f 2.0 percent. Because livestock ownership i s more evenly spread across rural households than i s land ownership or even access to land, productivity gains in livestock are generally more pro- poor than productivity gains o f major crops.35 According to the 1996 Livestock Census, the majority o f livestock holders are small farmers: 83 percent o f households that owned cattle andor buffaloes owned six or fewer animals. Moreover, there i s an important gender dimension to the livestock sector, as rural women play a major role inthe care o f livestock.36 Production technology and constraints vary according to location and herd size. Households with only two or three animals typically use very few inputs and produce mainlyto satisfy family needs. Households with four to six animals typically stall-feed their animals, and often add concentrates to the basic feed o f green fodder and straw. This group accounts for the bulk o f marketed milk supply. Rural commercial farms, including specialized dairy farms and mixed crop-livestock farms, typically have more than 40 animals, mostly buffaloes. Economies o f scale and availability o f adequate financial resources make it feasible to maintain breeding bulls on such farms. Peri-urban commercial dairy farms (gawala colonies) typically have herds of 15-50 animals, mostly buffaloes. These farms face relatively highfodder costs, but benefit fiom lower marketing costs and in some cases market directly to final consumers. 35 Simulationresults suggest that, ifmultiplier effects on non-farm incomes are included, gains from a 10- percent increase inlarge livestock production (mainly cattle, goats and dairy products) raise total incomes o f small farmers and landless agricultural householdsby about 10percent (Dorosh, Niazi and Nazli, 2003). 36 See Ahmad, Asghar and Khan 1993. 45 Box 3.2: Food Quality and Safety Standards inPakistan Pakistan presently lacks a coherent strategy for quality SPS (sanitary and phytosanitary) management inrelation to its trade. Inthe absence o f a coherent national strategy, Pahstani firms currently pursue their own strategies, in many cases reacting to events and adopting defensive actions designed to limit potential damage that might ensue from non-compliance with standards. For example, Pakistan banned seafood exports to the EU in March 2005, a step intended to pre- empt a possible EU ban, and this resulted in losses estimated at between $10 and $40 million duringthe first months o fthe ban. Improving food quality and safety in fisheries, horticulture, and meat/livestock i s a highpriority for Pakistan's export competitiveness. In place o f the existing fragmented, isolated, and non- coordinated state o f SPS management, a regulatory framework i s needed to support Pakistan's international trade objectives and obligations. This framework should: 4. Clearly define and delineate the roles and responsibilities o f the different federal and provincial ministries and agencies with respect to SPS management; 5. Strengthen existing technical capacity to administer science-based SPS measures including risk assessment; and 6. (Re-)institutionalize early warning and/or surveillance systems for pests, diseases and contaminants. Inaddition, the Consultative Group (CG) onWTO matters and its SPS Committee should be fully activated and a Task Force formed comprising a balanced cross-sectiono f stakeholders. Institutional changes and capacity-buildingactivities are also needed. Recommended highpriority actions for the short-term include: 7. Stratem and prioritv setting. HighlightSPS management constraints and issues, prioritize them and develop an action plan (short term, very highpriority); 8. Raising awareness. Engage the private sector in dialogue on SPS management capacity issues and develop an SPS information system inthe public domain (short-term); 9. Training on food-safetv controls. Awareness-raising and training on Hazard Analysis and Critical Control Point (HACCP), Good Agricultural Practice (GAP), and Good Manufacturing Practice (GMP) in fish products, horticulture, meat and livestock sectors (short-term); 10. Implementation of food-safetv controls. Implement and enhance food safety controls in slaughterhouses, fish processing plants, and pack-houses via awareness-raising, certification, surveillance, and auditing (short to medium-term); 11. Investment in hygiene facilities. Invest and upgrade hygiene facilities at Karachi Harbor (short to medium-term); 12. Support for implementation ofstandards. Provide support for implementation o f HACCP, GAP, GMP and other standards throughout the supply chain through loans and matching grants (medium-term); and 13. Phvtosanitary control measures. Raise awareness and provide training in practices for animal and plant health control including GAP, Integrated Crop Management (ICM) and IntegratedPest Management (IPM) (medium-term). Source: World Bank GrowthandCompetitivenessReport 2006 46 Milk production is expanding especially rapidly, although most milk continues to be marketed through traditional channels. The supply o f fresh milk increased at an average annual growth rate of 11.4percent between 1971-72 and 2004-05, and by 6.6 percent between the two livestock censuses of 1986 and 1996. Increased tractor use for land preparation i s one major factor that has driven increased milk production since the 1970~.~'Because fewer male buffaloes and bullocks are needed for draught purposes, herd composition has changed over time and available feed and fodder are increasingly used for female animals. The milk marketing system in Pakistan continues to be dominated by traditional channels that market unprocessed milk, which constitutes about 90 percent o f total milk production. Non-traditional channels, which specialize primarily in marketing processed milk from commercial milk processing firms account for the remaining 10 percent o f supply. Of total processed milk, approximately half, or five percent o f total supply, i s UHT (Ulb-a Heat Treatment) milk.The remainder i s mainly open pasteurized milk sold at milk shops (SDI2005). Significant productivity gains for small-holders are feasible through use of improved feeds, better veterinary services, and more efficient marketing channels to help ensure remunerative prices of milk for farmers. Multi-national firms have enjoyed some success in improving the efficiency o f marketing channels, but this accounts for only about 10 percent o f total milk consumption in Pakistan. One promising alternative mechanism for delivering the requisite technology as well as veterinary and marketing services to small farmers i s through cooperatives such as the Idaru-e-Kissan (IK), also known as Hala Dairy, a vertically integrated dairy cooperative (Box 3.3). The cooperative's output i s small relative to the total milk market, however, accounting for only about 0.2 percent o f total milk production in Palustan ( or about 2 percent o f total pasteurized and UHTmilk). The poultry sector, which accounts for about three percent of the livestock sector, value- added) i s expanding rapidly and medium-term prospects are good, due to increased urbanization and rising per capita incomes. Total number o f birds and egg production both more than doubled between 1990-01 and 2002-03, increasing by 135 and 145 percent, respectively. Production o f maize, used mainly for poultry feed, grew by an annual average o f 4.69 percent from 1990-91 to 2004-05; yields grew by an annual average 3.95 percent due to adoption o f improved seeds.38 However, Pakistan needs to be prepared to contain outbreaks o f poultry disease (including the Asian avian flu), as well as to provide improved general veterinary services. The timely flow of information regarding outbreaks of disease is crucial to facilitate government response, which may include mandatory culling of flocks in some areas to reduce spread of the disease. To encourage cooperation by farmers and to prevent poultry- owning households and firms from suffering major loss o f income, some form o f monetary compensation or credit may be necessary. However, as poultry and eggs supply a small share o f total animal protein consumed in Pakistan-only 3.6 percent and 2.7 percent, respectively-the nutritional consequences o f any outbreak would be limited. ~ 37 A comparison of results from the 1986 and 1996 Livestock Censuses reveals that in the inter census period the number o f bullocks reportedly used `for work' declined 33.5%. Government o f Pakistan, (1998). Livestock Census 1996. Punjab. Agriculture Census Organization Statistics Division. Gulberg, Lahore. 38 Production of maize i s still only about one-tenth that of wheat, or 2.1 million tons as compared to 19.9 million tons (2003-05 averages). 47 Box 3.3: Idara-e-Kissan (Hala Dairy) Idara-e-Kissan i s a vertically integrated dairy cooperative with 20,164 members in519 villages in Punjab (2004 figures). The cooperative i s open to any livestock farmer in a target village that owns one buffalo or cow and i s able to supply 300 liters o f milk duringa 6-month period. Unlike traditional cooperatives that own dairy farms, IK collects milk from thousands o f geographically dispersed members. Milk i s processed in one o f the cooperative's milk processing plants and marketed to urban consumers through retail outlets. The cooperative uses profits from its commercial operations to subsidize a package o f veterinary and livestock extension services delivered to members through private contractors. In2004, IK farms produced about 47 million liters o f milk, about 0.2 percent o f total milk production in Pakistan and about 2 percent o f total pasteurized and UHT milk. According to a 2005 survey o f 36 small dairy farmers (3-5 animals per farm) in the Kasur (IK farmers) and Okara (non-IK farmers) districts o f Punjab, milk production by IK farmers was 25 percent higher than for non-IK member farms (Table 3.16). Two major factors account for the difference in productivity: higher-quality feed and better veterinary services. IK members fed their animals three times as much cotton seed cake per milkbuffalo as didnon-IK members. They also benefited from access to veterinary services provided through private contractors; IK farmers also utilized veterinary services at about twice the rate o f non-IK farmers. Moreover, cooperative members' satisfaction with these services was higher than non-members' satisfaction with services provided to them by government agencies and the private sector. However, member farmers have not widely adopted other useful innovations promoted by the cooperative, such as balanced feed rations, improved fodder seed, non-traditional fodders, artificial insemination services, and hygienic practices to control disease. Improved market access i s also key to the success o f cooperative members. IK farmers received 13 percent higher prices for milk (which was priced according to fat content) than did non-IK members, who sold to local milk collectors (dhodis). Total revenue per milk buffalo for IK members was 38 percent higher than for non-members. Thus, despite their costs being 44 percent higher, IK cooperative members enjoyed 34 percent higher net returns per milk buffalo than did non-members (Rs 18,158 versus Rs 13,582), implying a marginal returnon cash costs invested of about 112 percent. Sensitivity analysis o f these results suggests that increasing market access for non-cooperative members, thereby raising both their milk prices and fodder costs by 20 percent would reduce the difference in net returns between the two groups to only 21 percent and reduce the marginal return on cash costs invested to about 93 percent. However, it must be noted the cooperative does not provide credit; members bear the higher feed costs themselves. However, for farmers without access to credit, higher input costs may hinder adoption o f improved feeding practices. Source: Riaz 2006. 48 Table 3.5. Costs and Returns for Cooperative Dairy Production, Punjab 2005 IKvs. Non-IK IKfarms Non-IK farms YOdifference Number of animalunits 6.83 6.84 -0.1% Milk-buffaloedfarm 1.70 1.56 9.0% Per milkbuffalo: Milkproduction(liters/year) 2,003 1,599 25.3% Price milk (Rskter) 14.58 12.85 13.5% Value of milk production (Rs) 29,209 20,548 42.2% Value of calves (Rs) 2,383 2,383 0.0% Total value of production (Rs) 31,592 22,93 1 37.8% costs Fodder 8,116 6,669 21.7% Other feed 5,233 2,604 100.9% Health treatments 85 75 13-9% Total costs 13,434 9,348 43.7% Net returns/milk buffalo 18,158 13,582 33.7% Source: 2005 field survey of 36 farmers inKasur (IK farmers) and Okara(non-IK farmers) districts, Punjab, as reportedinRiaz 2006. Notes: IK denotesIduru-e-Kissun, a dairy cooperativethat markets productsunderthe product name "Hula Dairy." FACTORSPRODUCTIONANDAGRICULTURAL INPUTS OF Land Distribution and Productivity Unequal land distribution is a major cause of income inequality in rural Pakistan. Returns to land are estimated to be about half o f incomes (value added) from crop agriculture. However, according to the 2000 Agricultural Census, only 37 percent o f rural households owned land, and 61 percent o f these owned fewer than 5 acres (15 percent o f total land). Two percent o f households owned 50 acres or more, accounting for 30 percent o f total land. The overall Gini coefficient o f land ownership in 2000 in Pakistan was 0.66; if rural landless households are included, the Gini coefficient was 0.86. By comparison, Gini coefficients for land ownership are 0.71 inIndia, 0.42 inBangladesh, and 0.85 inBrazil. As Pakistan's rural population is rising, the number of small farms is increasing over time despite the total area under cultivation being essentially fixed, constrained by access to water). According.to Agricultural Census data for 1990, 54.4 percent o f farm households owned less than 5 acres (accounting for only 11.4 percent o f total area), while 2.8 percent o f households owned 50 acres or more (accounting for 34.0 percent o f total land). By 2000, 61.2 percent o f farm households owned less than 5 acres o f land, (accounting for 14.8 percent o f total farm area), and 2.0 percent farmers owned 50 or more acres, but these farms accounted for 29.7 percent o f total area. In spite o f the increase in the number o f small farms, the overall Gini coefficient for land ownership inPakistan remained approximately constant from 1972to 2000 (Table 3.6). 49 Table3.6. DistributionofLandOwnership inPakistan 1972 1980 1990 2000 Gini Coefficient 0.66 0.65 0.66 0.66 % of landless households 62.0 63.3 % Share of Holdings e5 acres a. Households 47.3 n.a. 54.4 61.2 b. Land 5.4 n.a. 11.4 14.8 % Share of Holdings 50+ acres a. Households 3.3 n.a. 2.8 2.0 .b. Land 22.4 n.a. 34.0 29.7 Source: Agriculture Census (various issues). Most empiricalevidenceindicatesthat landproductivityon largefarms inPakistanis lower than that of small farms, holding other factors constant. For example, plot-level regressions o f productivity, correcting for plot characteristics and some household characteristics (such as tractor ownership and number o f household workers) suggests a doubling o f operated area leads to 10 percent lower wheat yields and 13 percent lower rice yields (World Bank 2002). Farm household income data also show that small farmers have higher net returns per hectare than do large farms. These findings suggest that increases inthe share o f land cultivated by small-holders would tend to increase overall farm productivity in Paki~tan.~' Although, land sales markets are thin in Pakistan:' land rentals in various forms are more common. According to 2001-02 HIES data, 33 percent o f land cultivated in 2000 was under some form o f tenancy arrangement; 18 percent was under share-cropping. Econometric evidence also suggests that tenure arrangements have a significant impact on productivity. Sharecropper productivity i s about 20 percent lower than landowner productivity, holding other factors constant, because there i s less incentive for sharecroppers' own-labor inputs (Box 3.6). Though share-cropping i s declining as a form o f tenancy, the benefits it offers to share-croppers (shared financial risks o f crop failure and credit for purchase o f inputs) and landowners (greater owner inputinto management decisions) alike explainits persistence as a tenancy arrangement. 39 Econometric analysis o f district-level data inPunjab indicates that districts with higher average farm size tend to have higher productivity than districts with smaller average farm size, however. See Ahmad (2002). 40 Land is rarely bought and sold, because hightransactions costs, potential disputes about accuracy o f land records, land prices inexcess of the discounted value o f potential agricultural earnings from the land, and lack o f access to credit by those without land. The low frequency o f landtransactions helps perpetuate the highlyunequal distributiono f land, hampers labor mobility, and reduces returns to family labor. 50 Box3.4. LandTenancy andProductivity Econometric analysis o f the effects o f tenancy suggest that the productivity o f share-croppers that are relatively unsupervisedai s 18 percent less than productivity o f supervised share-croppers and farmers who cultivate their own landab Since only 19 percent o f land in Pakistan i s share- cropped, and an estimated 35 percent o f tenants are relatively unsupervised, "unsupervised" tenants account for only about 7 percent o f total area cultivated and the effect on total production o f major crops i s small, only 1.4percent (Jacoby and Mansuri 2007). The lack o f secure tenancy arrangements also has negative implications for productivity because it disincentivizes tenants to make long-term investments to improve land quality. Econometric analysis o f PRHS 2001-02 data, (Jacoby and Mansuri 2006), shows that farmers used less manure on leased plots than on their own plots, reducing yields by about one percent in the first year and a cumulative total o f 2-2.5 percent in subsequent years.' These effects on yields are relatively small, but they may be larger if there i s a need for investment in such larger inputs as canal lining and tube wells. Productivity increases arising out o f a shift from share-cropping to fixed rents could lead to substantial benefits for share-croppers; however this i s o f relatively small benefit for most o f the rural poor. An 18 percent increase in productivity o f small farm share-croppers resulting from a shift to fixed-rent tenancy would only increase average incomes o fsmall-farmrenters (11percent o f the rural population) by 3 percent. Since the size o f this productivity shock i s small relative to the rural economy, there are only minimal effects on the average incomes o f other household groups, even accounting for linkage effects to the rural non-farm economy. Source: World Bank 2004b. Jacoby and Mansuri (2006,2007). a Inthis study, tenants were consideredto be unsupervised, ifduringthe cropping season there were fewer than ten meetings between the tenant and landlord to discuss management o f the plot. Share-cropping arrangements benefit share-croppers by enabling them to share the risk o f output fluctuations with the landlord, and also to share the costs o f purchased inputs. However, they provide fewer incentives for labor input by the tenant, as output is shared, as well. Application o f manure increases yields, by providing additional soil nutrients and by improving soil structure, aeration, water retention, and ability to retainnutrients. Three major attempts at redistributivelandreforminPakistanhavefailed (the most recent was in 1977), and land reform has neither political support nor the backing of Islamic religious authorities. Because o f these institutional constraints to traditional redistributive land reform, any attempt to increase access to land by poor households must involve a win-win approach. Land purchase schemes that include grant components for the poorest landless households are one option, though fiscal costs could limit the scale o f such programs. Policy measures to increase access to land must also include: increased access to credit to enable poor households to purchase land, land taxation to minimize the holding o f land for speculative purposes, and measures to improve the efficiency o f land sales and rental markets. Administrative reform of enforcement of existing norms and regulations, including contracts and the rights of tenants, and the liftingof restrictionson rentalswould also help improve the functioning of land markets. Simplification o f complicated legal procedures that result inlong processing delays and a backlog o f cases would reduce land transactions costs. Poor households are at a particular disadvantage in cases brought before civil courts, as they often 51 cannot afford to pursue these cases. Possible steps include: imposing limits on the number o f appeals; introducing procedural reforms and alternative conflict resolution mechanisms; and removal o f land revenue matters from the jurisdiction o f civil courts. Improved information dissemination on landprices and transactions would also enhance efficiency o f land markets. Improvements in land records4' could also enhance the efficiency of land markets and increase security of tenure. Without secure rights to their land, farmers not only have few incentives to invest, but also will devote fewer resources to defending their rights.Lack o f secure title means landowners are less willing to risk renting out land, which in turn reduces access by landless households to land. In addition, banks are reluctant to lend money if land i s used as collateral because they do not trust the current recording system. Uncertainty over the validity and enforceability o f titles also makes it harder to buy and sell land. Computerization o f land records, as tested in a recent pilot project in Punjab, i s one option for establishing clearer land records and reducing fraudulent dealings by land administration officials. WaterAvailability and Efficiency of Use Irrigation water i s critical for agriculture in most o f Pakistan. Massive public investment in dams and canals inthe Indusriver basin, as well as mainly private investment inwells for groundwater irrigation has been essential for agricultural production. Nearly 80 percent o f the cropped area i s irrigated, and agriculture i s by far the largest user o f available water inthe country, consuming on average about 95% o f available water resources. Increases in water resource availability for irrigation and the expansion of irrigated areas have played a central role in agricultural growth since the 1960s. The total irrigated area increased by 80 percent between 1960 and 2005, from 10.4 to 18.8 million hectares, mainly due to an expansion in tube well irrigation. In 2004-05, 37 percent o f irrigated land was irrigated solely with canal water, 41 percent was irrigated with canal and tube well water, and 18 percent solely with tube well water. Other forms o f irrigation, such as wells, canals with wells and tanks accounted for the remaining four percent. About 60 percent o f irrigated water available at farm head i s providedby canal water; the remaining 40 percent i s supplied by gr~undwater.~~ Nearly 97% o f irrigated areas are in the Indus basin, the largest contiguous irrigation system in the world. On average, approximately two-thirds o f total net water in the Indus basin (102.0 in 147.5 million acre feet) i s diverted to canals. Of this total, about 45 percent goes for consumptive use; the remainder i s comprised o f surplus flows to the estuary (29.4 MAF); system losses, including groundwater recharge (9.9 MAF) and Indus Treaty Allocations to India (8.4 MAF), (Table 3.7). In some recent years (2001-02 through 2003-04), however, total water flows in the Indus basin have been sharply curtailed by drought, reducing availability of water at the tail end of the system and limiting crop production, as well as necessary drainage (Figure 3.10). Kharifseason water canal diversions were 3.8 MAF (6 percent) lower than the long-term average. Rabi season shortfalls fell much more sharply, by 11.8 MAF (33 percent). Surplus flows were 41 InPunjab, land disputesare primarily due to fraudulent dealingbecause the records are not clear andpatwars (land administration officials) often accept informal payments. 42 Government o f Pakistan Water Vision 2025 Country Report, cited in World Bank 2003a; Pakistan Public Expenditure Management, Accelerated Development o f Water Resources and Irrigated Agriculture, Vol. 11. 52 greatly affected in both seasons, falling by 95 percent in the kharifseason and by 99 percent in rabi season. Table 3.7. SeasonalWater Flowsinthe IndusRiver Basin, 1975-76 to 2003-04 Drought Years Drought Years Drought Years 1975-76 2000-01 Less - Average Flows to 2003-04 to 2002-03 Average Flows (as % of Average) INFLOW Western Rivers 140.65 106.02 -34.62 -25% Eastern Rivers* 6.85 1.11 -5.73 -84% Total Inflow** 147.49 107.14 -40.36 -27% OUTFLOW lndus Treaty Allocation 8.40 8.40 0.00 0% Canal Diversions 102.93 87.40 -15.53 -15% Kharif 66.81 63.05 -3.76 -6% Rabi 36.12 24.35 -11.77 -33% System Losses*** 1.24 9.65 8.41 678% Surplus flow 34.93 1.69 -33.23 -95% Kharif 32.77 1.67 -31.10 -95% Rabi 2.15 0.02 -2.13 -99% Total Outflow 147.49 107.14 -40.36 -27% Source: Pakistan Statistical Yearbook 2004-2005; PakistanWater Sector Strategy October 2002; Ministry of Water and Power; WAPDA. Notes: Water flows are expressed inmillion acre feet (MAF). *Eastem Rivers Flows data available for 1991-2004, means calculated 1991-2001and 1991-2004. **Total Inflow calculated as sum of Mean Western Rivers Inflow andMean EastemRivers Inflow. ***System Losses calculated as Total Mean Outflow - Mean Indus Treaty Allocation - Mean Canal Diversions - Mean SurplusFlow. Current patterns of water use in Pakistan are causing significant environmental degradation: costs associated with loss o f soil fertility due to agricultural soil degradation (soil salinity and erosion) are estimated at Rs 70 billion per year (1.5 percent o f total GDP and 6.8 percent of agricultural GDP, based on 2004-05 GDP estimates). In addition to environmental losses, inefficient water use also leads to lower agricultural productivity and incomes. Problems related to seasonal water shortages are particularly acute at the tail end o f distributaries throughout the system and at the tail end o f the system as a whole (in southern parts o f Sindh).43 Optionsto increaseavailability of irrigationwater: There is little storage capacity downstream o f the Mangla and Tarbela dams innorthern Punjab or water control in Pakistan's run o f the river irrigation system. Adding storage capacity in the form o f dams will in principle help increase availability o f water in the dry season by spreading the flow o f water throughout the year.44 Nonetheless, even if investments are made immediately to increase storage, capacity will not increase in the next ten years, and i s likely to actually decrease somewhat in this period due to silting o f existing faci~ities.~' 43 See World Bank 2005a Sindh Economic Report. 44 Inrecent years o f low rainfall and snow melt, however, Pakistan's existing dams have not been filled to capacity and little Indus water has reached the coast. Additional dams can alleviate but probably not eliminate, the problems o f reduced water supply infuture drought years. 45 However, the major economic benefits o f increased storage capacity do not accrue from increased agricultural productivity, but from power generation. About 90 percent o f the anticipated benefits from the 53 54 downstream and reducing environmental degradation. For example, shifting 60 percent o f land under sugar cane cultivation to a maize-wheat rotation or about one-quarter o f ordinary rice land to maize would save enoughwater to supply Karachi's cui-rent needs. One way to encourage the transition to less-intensive water use is to increase water charges (abialza) for cultivation of water-intensive crops so that they more closely reflect the economic cost of water. Currently, Sindh levies canal water charges according to crop grown; in Punjab by contrast abiana does not vary by crop cultivated.) To minimize problems o f non- payment, however, any water-charge increases would need to be accompanied by improvements inservice delivery. Nonetheless, raising water charges to levels approximating the opportunity cost of water i s likely to be politically infeasible. An alternative would be the development o f markets for tradable water rights such as existing in Chile and Mexico. This would require laws establishing tradable water rates and the strengthening o f local institutions to enable them to serve as brokers inwater markets. Trade ofwater rightsabove the water course level would be limited because of the minimal degree o f water control above the water course level (the tail end o f the water distribution system), though some transactions conceivably could be brokered by water course associations or water users associations at the irrigation minor level (i.e. higher in the water distribution system). Projects and programs involving agricultural extension, technical assistance, market development and credit may be needed to ease the transition o f farmers and processors to alternative crops and products. For example, if increased water costs reduced the profitability o f sugar cane, programs could be developed to reduce the burden o f converting processing.facilities such as sugar mills for other uses (such as oilseed processing facilities). 55 Box 3.5. Improving Efficiency of Water Use inPakistan Major improvements in the efficiency and sustainability o f water use in Pakistan can be achieved by addressing problems related to asset management, water allocation and service delivery. Irrigation Departments need to establish and implement asset management plans that set out measures to be taken to ensure: - Divesture o f assets for activities that can be taken over by the private sector or - farmer organizations Operation and maintenance plans including medium-term expenditure frameworks, assignment o f financial responsibility to the different actors, measures to reduce establishment costs, pricing policies and cost recovery mechanisms; 0 Service delivery and allocation of water can be improved by devolving water management and canal command operation to Area Water Boards (AWBs) and Farmers Organizations (FOs). FOs should also exercise oversight o f - -- Collection and retention o f abiana Operationand maintenance at the distributary and branch canal level Improving existing infrastructure To improve transparency, Irrigation Departments should publicize information on water flows, expected discharge and allocation o f water for each canal, as stipulated inthe 1991 Water Accord between the provinces and the federal government. 0 Increase availability o f irrigation water through - Investment in improved drainage, control structures and measures to reduce conveyance and distribution losses, particularly in areas where groundwater - quality i s poor. Investment in large-scale storage only after careful evaluation o f environment, social and economic consequences 0 Rationalize water usage and review water charges so that these better reflect opportunity costs o f water usage and can be used as an incentive to improve water productivity - To minimize problems o f non-payment, increases in water charges should be accompaniedby improvements inservice delivery 0 Develop effective markets for entitlements to canal water to make the system more equitable, particularly for farmers whose land i s located at the tail end o f water courses - Enforce the present system o f water allocation to facilitate legal tradable water - rightsat the watercourse, FO and AWB levels Strengthen local institutions (such as Water Course Associations, FOs, AWBs and professional canal management agencies) to increase water productivity and - enable them to serve as brokers inwater markets Any system o f tradable water rights must account for the minimal degree o f water control above the water course level in the current Indus basin system caused by the absence o f sufficient storage capacity and control structures (gates) to regulate water flows. 56 AgriculturalCredit Total formal sector rural credit has expandedrapidly since the Zarai Taraqiati Bank Ltd. (ZTBL), (then called the Agriculture Development Bank o f Pakistan) reformed its policies in the late 1990s to simplify lending procedures and make credit more accessible to small landowners. Territorial restrictions on commercial bank lending operations were also removed. As a result, total rural credit grew by an average annual rate o f 12.0 percent inreal terms between 1999-2000 and 2005-06 ( Figure 3.3). Commercial credit more than quadrupled inreal terms over the same period, increasing its share o f total formal credit from an average o f only 23 percent in the 1990sto over halfo f all credit in2005-06.46 Figure3.3. Growthof InstitutionalAgriculturalCredit inPakistan(Rs2004-05 bn) 120 I I / \ 100 II 80 60 40 20 0 -+-ADBPETBL -a- CooperativeslFBC t-CommercialBanks -Total Source: Malik 1999; PakistanEconomic Survey (various issues). According to PRHS 2001-02 data, nearly 80 percent of cultivatorhouseholdsparticipatein the credit market, with two-thirds of total rural credit coming from the informal sector (Table 3.8). Access to institutional credit i s highly restricted. Few farmers in the sample fully cover their financial needs through formal credit, and those without land have almost no access to formal-sector credit. Overall, only 11 percent o f farmers had obtained formal sector loans, often usingthese loans to purchase tractors. Informal sector credit i s muchmore widespread. 75 percent o f farmers had obtained informal sector loans, including 72 percent o f land owners and 84 percent o f non-landowners. Total average borrowing by land-owners (including both borrowers and non- borrowers) was Rs 29,500, 65 percent higher than total average borrowing by landless farmers (Rs 17,900). Agricultural credit (particularly production loans through commercial banks) was highly subsidized from the late 1970s through the mid-1980s. Subsidies on these commercial loans were 46 Nonetheless, deposits made to rural and small town banks far outpace advances to the rural sector, indicating a net financial outflow from rural areas. Commercialbank interest rates o n deposits that are less than overall rates o f inflation represent a tax on savings o f rural (and urban) households and firms. 57 discontinued in the late 1980s, however. Nonetheless, low repayment rates on loans through ADBP/ZTBL continue to constitute an implicit credit subsidy. As discussed above, because uncertainties about land titles in the current system make banks reluctant to lend money on the basis of these titles, improvements in land administration and land titling could improve access to credit, as well as facilitate more efficient use of land, as well as increasesecurity of tenure. Amount (bn Rs) Share formal Share rural Urban (formal) 2005.4 95.7% n.a. Rural (formal) 91.2 4.3% 28.0% ZTBL 29.0 1.4% 8.9% Domestic Private Banks 11.0 0.5% 3.4% Cooperatives 4.2 0.2% 1.3% Commercial Banks 47.0 2.2% 14.4% Rural (informal)* 185.1 n.a. 67.0% Total (formal) 2096.6 100.0% n.a. Total (rural) 276.2 n.a. 100.0% Other Agricultural Inputs There have been no direct subsidies on pesticides, seeds and most fertilizers since 1996, but subsidies on surface irrigation water and electricity (in Baluchistan only) remain. The pesticide and seed subsidies were eliminated in 1981-82 and 1982-83, respectively. The subsidy on locally produced fertilizer was eliminated in 1993; a small subsidy on imported fertilizer was eliminated in 1996.47 Likewise, a subsidy on tube wells o f Rs 16,000 to Rs 20,000 per unit was abolished in 1994-95. Subsidies on canal irrigation (calculated as the difference between revenues collected from farmers as water charges and the cost incurred by the government for operation and maintenance, excluding capital costs) averaged Rs 4.6 billion from 1996-97 to 1999-00, equivalent to 0.8 percent o f agricultural GDP. The price o f nitrogenous fertilizer remains substantially below border prices, however, due to subsidies on domestic p r o d ~ c t i o n .Domestic prices o f both urea and Di-ammonium phosphate ~ ~ (DAP) were on average 25 to 40 percent below import parity in the late 1970s and the 1980s. Since the early 1990s, there has been little or no subsidy on DAP and other major fertilizers. The exception has been urea, for which the domestic price was on average 38 percent below border 47 All fertilizers now face a 15 percent sales duty, and imported fertilizers also face a 5 percent customs duty. 48 Surface irrigation water i s also implicitly subsidized as water charges (abiana) are insufficient to cover the cost o f maintenance o f dam, canals and other water channels. Measurement o f the economic value o f these subsidies involves assessment o f overall investment and maintenance costs, as well as problems attributing these costs to various crops, and i s not attempted inthis paper. 58 prices from 1990-91 to 2005-06. Given that the cost o f urea i s about eight percent o f the value o f wheat production during this period according to APCOM cost-of-production estimates, the implicit subsidy on fertilizer was equivalent to only about three percent (0.08 * 0.38) o f the value o f wheat production from 1990-91 through 2005-06. The value o f the urea fertilizer subsidy as a share o f output values for paddy, cotton and sugar cane respectively i s similarly small. MARKETS TRADEPOLICIES AND Until the mid-l980s, most major agricultural commodities were implicitly taxed through negative indirect effects of overall trade and exchange rate policy that exacerbated the effects of agricultural domestic price interventions.Inthe 1983-87 period, the direct effects o f domestic price and trade policies reduced wheat prices to an average o f 25 percent below border prices. The indirect effects o f trade and exchange rate policies that led to an appreciation o f the real exchange rate further reduced these prices to an average o f 42 percent below border prices.49 These price disincentives reduced wheat production by an estimated 24 percent (relative to what it would have been during this period). Taking into account price distortions on all crops, total farm incomes were reducedby an estimated 29 percent." Real exchange rate distortions caused by trade policies have decreased sharply, however, due to major trade liberalization in the 1980s and 1990s. Import tariffs and other restrictions on imports reduce demand and increase supply o f importables, and thus lead to an appreciation o f the real exchange rate. In the 1980s, the net effect o f implicit import (and export) taxes was to cause the real exchange rate to appreciate by an estimated 17 percent compared to its value inthe absence o f these trade distortions. As a consequence o f overall trade policy reforms, however, estimated implicit tariffs (adjusted for export taxes) were reduced to 30 percent in the 1990s and to only 15 percent from 2000 to 2005. Thus, the appreciation in the real exchange rate due to these trade distortions was also reduced, from 17 to 11to 6 percent over the same period, and the implicit indirect taxation o f agricultural tradables was reduced, as well (Dorosh and Salam, 2007)." Other factors, however, led to a sharp appreciation o f the Pakistan rupee in 2005, resulting in lower incomes for producers o f tradable goods. The real exchange rate appreciated by 14 percent between December 2004 and November 2005, due largely to substantial public and private capital inflows accommodated by increases in domestic money supply. This appreciation reduced domestic prices o f major tradable agricultural products (such as wheat, cotton, rice, sugar cane, vegetable oils, and citrus fruit for export) and reduced farmers' real incomes. Moreover, an appreciating real exchange rate reduces the competitiveness o f Pakistan's high-value fruits, vegetables and livestock products in international markets, thus diminishing growth in a sub- sector crucial for raising agricultural incomes. In addition to the effects of overall trade and exchange rate policy, commodity-specific trade and pricing policy reforms undertaken for the most part from the late 1980s until the mid-1990s greatly reduced price distortions as measured by divergences between domestic output prices and import or export parity border prices. Significant trade taxes and 49 For cotton, this real exchange rate appreciation reduced the protectionafforded to domestic cotton production from 177 percent down to 27 percent. 50 Dorosh and Valdes 1990; andHamid, Nabi and Nasim 1990 give similar figures for nominal rates o f protection, as well as a detailed account o f the policy interventions from 1960 through the mid-1980s. 51 These calculations use an econometric estimate o f the elasticity o f the real exchange rate with respect to a change in(one plus) the implicit tariff rate o f -0.41 (Dorosh and Valdes 1990). 59 restrictions on major agricultural products remain only for vegetable oil and powdered milk imports and basmati rice exports. Thus, direct trade and pricing policy distortions, as measuredby nominal rates of protection using official exchange rates, have generally declined in absolute magnitude since the early 1980s. Including the effects of exchange-rate distortions, the pattern i s clearer: there i s a sharp reduction in policy distortions over the past two decades (Table 3.9 and Figure 3.4). Direct Effects Total Effects 1981-1990 1991-2000 2001-2005 1981-1990 I1 1991-2000 2001-2005 II 11Wheata -0.12 -0.05 -0.09 -0.28 -0.19 -0.15 Wheata'" II -0.03 II -0.17 II 0.03 II -0.03 I1 -0.13 I1 0.04 1I Cotton 0.00 -0.06 0.09 -0.14 -0.09 0.06 Cotton' n.a. -0.14 -0.12 n.a. -0.15 -0.17 Basmati Paddy -0.56 -0.30 -0.35 -0.65 -0.39 -0.39 IRRl Paddy -0.29 -0.29 -0.29 -0.38 -0.29 -0.08 Sugar 1.06 0.70 0.83 0.83 0.65 0.88 Sugar Cane 0.66 0.06 0.37 0.33 0.02 0.44 Vegetable Oil 0.00 0.28 0.27 -0.08 0.25 0.24 Implicit Tariff _ _ _ _ _ _ 0.48 0.32 0.16 RER Effect __ __ __ --- --- 0.17 0.12 0.06 Figure 3.4. NominalRates of Assistancefor Major Crops in Pakistan, 1961-62 to 2004-05 0.60 I 0.40 0.20 0.00 -0.20 -0.40 -0.60 -0.80 Source: Doroshand Salam 2007. Notes: * Calculatedusing estimated autarky prices for years in which estimated trade i s zero under free trade. 60 Domestic price and trade policies, especially large-scale government imports, have kept farm gate prices of wheat at an average of 15 percent below import parity since the early 1980s. Net government injections in wheat markets (domestic sales less domestic procurement), inmost years supplied through government imports, averaged 2.10 and 0.93 million tons in the periods from 1990-91 to 1999-2000 and 2002-03 to 2004-05, respectively. This i s equal, respectively, to 12.4 and 5.2 percent o f net availability. Only following the record harvest in MarcWApril 2000 and in 2005-06 has the net effect o f government interventions been to reduce total net availability and thus raise domestic prices.52 Under a free-trade regime, domestic wheat markets would likely have cleared at a price below the import parity level and imports would have been zero. Using estimated autarky (no trade) prices in these years, actual domestic wheat prices were on average 10 percent below no-intervention prices in the 1990-91 to 1999-2000 period, and 3 percent above no-interventionprices inthe 2000-01 to 2005-06 period. Between 2004-05 and during the last six months of 2006, however, an increase in world prices combinedwith stagnant domestic prices (heavily influenced by government trade and stock as well as stock management policies) to bring down domestic prices from near import parity to near export parity. Average wholesale wheat prices inLahorewere unchanged betweenthese two periods, while the world price o f wheat (including estimated cost and freight to Karachi) rose by 25 percent. Combined with an exchange-rate depreciation of 3 percent, import parity in Karachi rose by 28 percent. With this shift in world prices, private sector exports o f wheat may be profitable, helping to stabilize domestic prices at export parity levels (Figure 3.5). Figure 3.5. Pakistan and International Wheat Prices, 2000-2006 16 14 12 6 Paklstan(Lahore) - ExportPanty Lahore ~ -c Import Panty Lahore(ex US) Source: Authors' calculations. 52 Without these interventions, domestic prices would have been an estimated 30 percent lower in 2000101 and 8 percent lower in2001102 (Dorosh and Salarq 2007). 61 Trade-policy distortions in domestic markets for cotton lint are minimal, although seed cotton is protected through import tariffs on vegetable oils that compete with cotton seed oil. Using domestic cotton seed prices in the calculations, estimated average nominal rates o f protection for seed cotton were only five percent in 2001-05. The protection provided to cotton seed oil, however, raises the nominal rate o f protection for seed cotton to 29 percent in the same period. Moreover, annual price movements suggest that domestic seed cotton prices are essentially determined by world prices o f cotton lint and the domestic price o f cotton seed ( Figure 3.6 and Orden et al. 2005). Simulations using various models o f world cotton markets generally suggest that the elimination o f cotton subsidies in developed countries would raise world cotton prices by an estimated 11-28 percents3 If Pakistan's own cotton policies remain unchanged, domestic cotton lint prices would rise by a similar amount. Figure 3.6. Pakistan:RealPricesof SeedCotton, 1980-81to 2004-05 45 40 35 30 25 20 15 10 5 0 ! ' l 8 ' I 8 8 I I 8 8 8 I 8 8 8 8 S I S I 8 , 1980/81 1983/84 1986/87 1989/90 1992/93 1995/96 1998/99 2001/02 2004/- I 11I* .,SeedCottonPrice(Farmgate)+-Border Price(Farmgate)-m-SupportPriceSeedCotton Source: Orden et al. 2005 and authors' calculations. The most significant policy distortions on output prices of agricultural products are the result of import restrictions and tariffs on sugar and vegetable oils. Domestic sugar prices (wholesale in Punjab) have been on average nearly double world prices (import parity) over the last two-and-a-half decades, due mainly to quantitative import restrictions. Estimated protection for sugar cane producers i s lower although still significant (it averaged 19 and 58 percent in the 1991-2000 and 2001-05 periods, respectively, a reflection o f the high costs o f converting sugar cane to refined sugar.54 Trade policy also encourages import substitution in vegetable oils with fixed rupee import taxes on vegetable oils averaging an equivalent o f 32 percent for soybean oil and 40-48 percent for palm oil. Separate tax rates apply to refined, bleached and de-odorized (RBD) and to crudehefined palm oil. The average ad valorem rates quoted here are calculated using the respective fixed rates divide by the average price o f palm oil imports o f all kinds. 53 The range o f results o f model simulations i s based on those reported inOrden et al. 2005, Table 3, with the highest and lowest estimate excluded. The simulations are based on 2000101 or 2001/02 data. Sumner (2003) estimates that removal o f US.subsidies alone would raise world prices by 12.6 percent. 54 Protection rates for sugar and sugar cane are highly sensitive to assumptions regarding processing costs and marketing margins. 62 Ad valorem tariffs on whole milk in2005-06 were 25 percent; the rate on milkpowder in2001-02 was 20 percent. Moreover, direct government intervention (public sector procurement and sales) in domestic markets has been sharply curtailed for most agricultural products since the mid- 1990s. Procurement and sales o f rice (both basmati and IRRI) and sugar cane have essentially stopped. Direct intervention in the cotton market has been minimal since 1994 following removal o f the export duty on cotton as Pakistan became a net importer; in 2004-05 however, the Trading Corporation o f Pakistan bought 270 thousand tons o f cotton (11 percent o f domestic production) inaneffort to boost domestic prices after abumper WheatPolicy The government continues to be substantially involved in wheat markets, however. Provincial governments (mainly Punjab and to a lesser extent Sindh) and PASSCO (Pakistan Agriculture Storage and Supplies Corporation) procure about 20 percent o ftotal wheat production each year. Total procurement o f wheat as a share o f production (20 percent) inPakistan i s similar to that in India for wheat (25 percent) and rice (23 percent), and significantly higher than that of Bangladesh (10 percent for wheat and 3 percent for rice). The latter, a country with higher levels o f poverty appears to have stabilized domestic grain prices and to maintain food security stocks with considerably less recourse to government procurement (Table 3.10). Large scale government wheat procurement offers little direct benefit to consumers of wheat flour, however, because almost all procured wheat i s bought and then sold again to flour millers (at a financial loss to the government) in the same wheat marketing year. Most recently, millers have enjoyed quotas for purchase o f government subsidized wheat at below-market prices and have thus been able subsequently to sell wheat flour at market-clearing prices. This has contributed to a large expansion in wheat mills, over-capacity in the milling industry and a large number o f mills that operate only when subsidized wheat i s available for purchase. Reducing the volume of domestic procurement (and subsequent intra-annual sales) would directly reduce the wheat'subsidy that arises from the disparity between the sale price o f government wheat (whether procured locally or domestically) to wheat mills and the full cost o f wheat procurement and handling.56Reducing the volume o f procured wheat and subsequent intra- annual sales would have little adverse effect on consumers, though, as wheat flour produced from government wheat i s typically sold at open market prices. The potential savings fkom eliminating the wheat subsidy through lower intra-annual domestic procurement and increases in release (sales) prices to flour mills would have increased funds for targeted programs inPunjab in 2002- 03 by t w o - t h i r d ~ . ~ ~ 55 The Trading Corporation o f Pakistan (TCP) also bought 35 thousand tons (2 percent o f roduction) in2001-02 to support prices. p6 Raising the sales price relative to the procurement price would also lower total subsidies, but setting the sales (release) price too highrelative to open-market prices would induce flour mills to purchase less, as was the case in2005-06. 57 Officially, wheat subsidies inPunjab fell between 2002103 and 2003104. Although the procurement price was unchanged, the per kilogram subsidy fell by Rs 1.43 as the cost o f incidentals fell by Rs 1.06kg (from Rs 1.92 /kg to Rs 0.86kg). Details o f how these subsidies are calculated are not published. 63 Table 3.10. Food Grain Policy inPakistan, India and Bangladesh,2001-05 Pakistan India India Bangladesh Bangladesh Wheat Wheat Rice Wheat Rice Production (mn tons) 19.49 70.54 85.72 1.40 25.22 Procurement (mn tons) 3.78 17.73 19.40 0.15 0.82 Share of production 19.5% 25.2% 22.6% 9.6% 3.2% Stocks (mn tons) 1.30 21.56 18.04 0.26 0.56 Stocks (kgkapita) 3.0 20.6 17.2 1.9 4.0 Distribution channels Subsidized Subsidized Subsidized Targeted Targeted sales to sales sales Distrib Distrib; flour mills PDS PDS (FFW, etc.) Sales Other wheat-policy reforms should accompany changes in procurement to ensure consistency across policies and flexibility to tailor policy to changes in the market (Box 3.6). Consistent and transparent trade policy with low tariffs to encourage international trade, level playing fields and avoiding marketing restrictions aimed at promoting competitive domestic market^,^' and enhanced analytical capacity and information flow are key complements to a policy of reduced government intervention in markets. Because Pakistan's wheat markets are generally well-integrated, using cash instead o f food for safety nets would .allow welfare objectives to be met while eliminating the need for government procurement. Current average year-end stocks are not excessively large relative to the size o f the domestic market, and could play a useful role in emergency preparedness. 58 Pakistan took significant steps towards market liberalization and improved efficiency o f wheat markets in 2004 and 2005, including liberalization o f private sector wheat imports and the removal o f restrictions on grain movement inPunjab. This policy represents a dramatic improvement over the policies o f 2004, when transport o f wheat across districts inPunjab was banned as part o f an initiative by the Punjab provincial government to meet its procurement targets. 64 Box 3.6. Wheat Policy Reforms: Towards a More EfficientWheat Market inPakistan Pakistan's current wheat system involves implicit procurement quantity targets, purchases at a fixed procurement price, and large scale same-year subsidized sales to wheat millers. In years when internal procurement quantity targets, (equivalent to about 20 percent o f domestic production), are not met, the government has resorted to raising procurement prices or imposing restrictions on transport o f grain. Shifting to a system o f minimal market intervention designed to stabilize prices in years o f extreme supply shocks, would achieve similar price stabilization results, at the same time significantly reducing costs. A market- oriented system o f this sort should include the following components: 1. Lower levels of government procurement. The government should set a floor price at which procurement in years o f normal or below-normal harvest i s equal to minimal levels required to rotate and maintain stocks. (Current average levels o f procurement are about four million tons or 20 percent o f production.) 2. Modest government carry-over of stocks as a security measure. The government should maintain annual carryover stocks equal to current levels (0.5 to 1.0 million tons). Average year end per capita carryover stocks or "strategic reserves" in Pakistan are already relatively small in comparison with other South Asian countries, and do not involve large financial costs. 3, Liberalization of private sector international trade. The government should provide clear signals to the private sector regarding tariffs and import restrictions, as well as promoting a level playing field with no special advantages for government agencies (PASSCO, State Trading Corporation o f Palustan, etc.). Experience o f other Asian countries, including Afghanistan and Bangladesh, has shown that private sector import trade can supplement domestic supplies and prevent price spikes inyears o f crop shortfalls. 4. Cash-based safety nets. The government should use cash instead o f transfers-in-kind or subsidized sales o f commodities, to save on administrative and distribution costs. 5. Promotion of a competitive domestic market. The government should provide fair access to credit, avoid limits on private storage, and maintain clear and consistent policy signals. Competition among traders and a level playing field can maximize the potential efficiency and social welfare gains from competitive markets, even as individual traders seek to maximize their own profits. 6. Enhanced analytical capacity and information sharing. The government should strengthen its capacity to monitor and analyze market developments, and establish systems to more effectively share market and policy change information. 65 AGRICULTURAL GROWTH LINKAGES RURALPOVERTYREDUCTION AND Historically, demand linkages ensuing from increased agricultural output and incomes have been the most important mechanism for spurring growth in the rural non-agricultural economy o f Pakistan and other developing countries. Increases in agricultural production generally involve increased demand for agricultural inputs, processing and marketing services. Also, as household incomes rise, consumer demand for both urban and rural products and services increases. To the extent that the supply o f goods and services i s elastic, increases in demand can spur increases in production which inturn further boosts demand. (See Box 3.7.) Box 3.7. Agricultural Growth and RuralPoverty Tamil Nadu, India Despite concerns that Green Revolution technology for rice (improved seeds, increased fertilizer use and irrigation) would lead to increased income inequality, most rural household groups benefitedfrom increased rice production in the 1970s and early 1980s (Hazel1 and Ramasamy, eds., 1991). Household surveys o f farmers indicatedthat, although the first adopters o f the rice technology tendedto be larger farmers, by the early 1980s small farmers invillages with access to irrigation had adopted the new technologies at the same rate as large farmers. Land distribution remained essentially unchanged, and agricultural wage earnings for agncultural laborers increased although real agricultural wage rates didnot increase. Moreover, increased rice production ledto substantial rural growth linkages, adding an additional 0.87 rupees o f value added for every 1 rupee o f value added from additional rice production. Half o f these multiplier effects were due to increased demand for agricultural inputs, marketing and processing services; the remainder derivedfrom increased consumer demands as householdincomes rose. Survey results showed that real incomes o f small paddy farmers and landless laborers rose by 90 and 125 percent, respectively, between 1973-74 and 1983-84. Incomes o f non-agricultural households also rose by 55 percent, due in part to the growth linkage effects ensuing from agricultural growth. Large paddy farmers' real incomes rose only 18 percent, due to increased fertilizer and labor costs. Notably, non-paddy farmers with no access to irrigated landexperienced real income gains o f only 17 percent. Similarly, analysis o f state-level data on growth and poverty inIndia (Datt and Ravallion 1998) shows that agricultural productivity growth from 1958-1994 played a major role inreducing poverty. During this period, agricultural output per acre ofnet sownarea grewby 2.91percent per year, contributingto higher real wages for agricultural laborers (which grew by 2.84 percent annually) and a decline inthe real price o f food. The price o f food relative to agricultural laborers' consumer price index fell by 0.15 percent per year from 1976-94, after having increased by 0.62 percent per year from 1958-75. Higher farm yields for small producers also contributed directly to their incomes, through increased value o f output and indirectly through increased employment. In Pakistan because much of the value added in agriculture accrues to land (in the form of implicit or explicit rents) and becauseland distribution is highly skewed, much of the rural populationhas relatively little agricultural income. Medium and large land owners (those with 12.5 acres or more) account for 10 percent o f agncultural households and receive an estimated 32 percent o f agricultural incomes. However, agricultural incomes (including livestock) generally account for only 56 percent o f total income for agncultural households, which constitute 41 percent of Pakistan's total population. Including rural non-agricultural households (3 1 percent o f the national population), the .share o f agricultural incomes in total rural incomes i s only 37 percent (Table 3.11). 66 Table 3.11. RuralAgricultural Incomes inPakistan PRHS PRHS SAM* SAM* Agric Inc Agric Inc Agric Inc Agric Inc Per Capita Share Per Capita Share ('000 Rs) (percent) ('000 Rs) (percent) Medium and Large Farms 15.7 83.5 29.9 57.2 Small Farms 6.1 67.9 8.6 54.8 Landless Farmers 7.2 87.7 5.3 59.7 Rural Agric Workers 2.2 53.1 5.5 53.1 Rural Non-Farm Non-Poor 0.3 1.9 0.1 0.4 Rural Non-Farm Poor 0.2 6.3 0.2 4.5 Total Rural 6.1 69.7 5.6 37.2 Rural Agric Households 7.1 74.8 9.6 55.9 Source: Pakistan SAM 2001-02; PakistanRural Household Survey 2001-02. Notes: * denotes Social Accounting Matrix The extent to which non-agricultural households gain from agricultural growth is determined by the magnitude of growth linkage effects. These linkage effects can be estimated using a semi-input-output model, based on a social accounting matrix ( S A M ) for Pakistan that includes disaggregated information on production activities, returns to factors o f production (e.g. land, labor and capital), supply and demand of goods and services, and incomes and expenditures o f households and other institutions. The model assumes that output o f tradable goods (major crops, cattle products, most industrial products) i s inelastic, and does not expand due to increases indemand. For these products, increaseddemandresults inincreasednet imports.For elastically supplied products, (minor crops, poultry, and services), increased demand i s assumed to induce increases in Simulation results o f a 10 percent increase in the output o f all major crops (wheat, basmati and IRRIrice, cotton and sugar cane) suggest that the largest gains of increasedagricultural crop production accrue to large and medium land owners, whose incomes rise by 7.2 percent. Incomes of small farm owners and pure tenants also rise by about 4.6 percent. Due to multiplier effects, incomes o f non-farm rural household groups also rise by 3.4 percent on average. The poorest rural household groups (agricultural laborers and rural non-farm poor (29 percent o f the rural population) reap only 6.7 percent o f the total income gains, and their incomes rise by only 2.6-4.1 percent. Much o f the income gains accrue to the owners o f capital inboth rural and urban areas (Figure.3.7) .60 59 Correct classification o f a sector as elastic insupply i s crucial to the model results. Note also that the model uses average input-output, value added and factor payment coefficients derived directly from the SAM, though consumption parameters are basedon econometrically estimated income elasticities o f demand. Thus, the simulation analysis assumes there i s no change inthe structure o f Pakistan's land distribution and factor payments. For details, see Dorosh, Niazi and Nazli (2003). 60 The simulations show the average income gains for the various household groups. If non- agricultural labor incomes are concentrated among only a segment o f rural non-farm poor households, the income gains o f this segment o f households will be larger, (and the gains to other rural non-farm poor households correspondingly smaller). 67 Figure.3.7. Simulation Resultsof a 10-percent Output Shock to Wheat, All Major Crops and Livestock inPakistan II - 8.0 al m E -s 6.0 Major Crops 0 Livsstock 4.0 2.0 0.0 Lrg/Med Sm Sm Agr Non-farm Non-farm Farm Owners Renters Workers Non-poor Poor Source: Dorosh, NiaziandNazli2003. Simulations using a disaggregatedS A M that separates activities and households in Punjab produce similar results. A 10 percent increase in the output o f all major crops raises overall incomes by 2.5 percent and large and medium land owners' incomes by 6.2 and 4.3 percent, respectively. Incomes o f small farm owners and pure tenants in Punjab rise by 3.3 and 3.7 percent, respectively. Multiplier effects raise incomes o f non-farm rural non-poor household groups in Punjab by 2.7 percent, on average, and the rural non-farm poor households by only 2.3 percent (Figure.3.8). Gains from a 10 percent increase in large livestock production (mainly cattle, goats and dairy products) are more evenly spread out given the distribution of livestock ownership. A 10 percent increase in national livestock production results in an overall income gain o f 4.5 percent; incomes o f small farmers and landless agncultural households (both o f which also have significant livestock income) rise by 9.7 and 10.9 percent, respectively. A 10 percent increase in livestock production in Punjab only raises overall incomes in the province by 3.4 percent. Incomes o f Punjab small farm owners, small farm renters, and landless agricultural households (the latter o f which also have significant livestock income) rise by 6.4, 10.0 and 7.1 percent, respectively. Thus, the simulationsshow that agricultural growth has substantial benefitsfor low income farmers and tenants (approximately 37 percent of the rural poor), but another source of demand (besides agricultural growth linkages) is needed to rapidly raise the earnings of the rural non-farm sector. The limited impact o f agricultural growth on incomes o f the rural non- farm poor i s the result o f two major factors: segmentation o f Pakistan's agricultural labor market; and agriculture's declining contribution to both total GDP and rural household incomes. Because much crop farm labor (about 80 percent) is own-family labor, most of the direct gains in labor earnings from increased agricultural output accrue to farm households, not 68 to hired agriculturalworkers. As a result, incomes of landless agricultural workers (10 percent o fthe rural poor) rise only moderately inthese scenarios. Figure.3.8.SimulationResultsof a 10-percent Output Shockto Wheat,All Major Crops andLivestockinPunjab,Pakistan 12.0 10.0 - 3 8.0- P Wheat 6.0 Major Crops E0 4.0 2.0 0.0 Source: Dorosh, NiaziandNazli, (2003). Moreover, because over time, agriculture's share in boththe rural and national economies has shrunk, multipliereffects originatingfromthe agriculturalsector havea smaller impact on the non-agriculturaleconomy.61 Closer integration with the urban economy also reduces consumption linkages within the rural economy, as rural households spend more o f their incomes outside o f rural areas or on goods produced outside rural areas. Both these changes are typical patterns ineconomic development. The implicationis that growth linkage effects have a smaller impact on GDP growth and rural poverty reduction today than they did in the past, and that an additional source of demandis neededfor rapidgrowthof the rural non-farmsector. Agricultural growth still has a positive impact on rural poor incomes, but this i s smaller, relative to the size o f the economy, than it was three decades ago. Even with agricultural growth o f 3.9 percent per year (the historical average from 1975 to 1990, a period o f rapid real agricultural GDP growth) and the higher growth multiplier o f 2.5, non-agricultural income growth induced through multiplier effects would be only about 2.9 percent, implying a per capita growth rate o f less than 1.0 percent per year. 61 The share of major crops intotal GDP inPakistan has fallen dramatically, from 0.234 in 1970 to 0.091 in2000. Therefore, the direct effect o f a 10 percent gain inmajor crop production is smaller: it was equivalent to 2.3% o f GDP in 1970, but only 0.9% o f GDP in2000. Moreover, assuming a constant value- added multiplier of 1.5,the total effects (including multiplier effects) on the rest o f the economy are also smaller (4.6% o f GDP in 1970 compared with 1.5% o f GDP in2000), (Dorosh, Niazi andNazli 2003). 69 POLICY IMPLICATIONS Agricultural growth inPakistan throughout most o f the last three decades has depended to a large extent on the major crops (wheat, rice, cotton and sugarcane). Although there i s still scope for production increases o f these crops, future increases inagricultural productivity and rural incomes are likely to be driven by diversification into high-value crops and livestock. Research and extension services need to be restructured to meet the needs o f a more diverse agriculture sector, including provision o f region-specific information packages. Facilitating new contracting arrangements and investment by supermarkets will also facilitate marketing o f high value fruits, vegetables and animal products. Investment in irrigation and drainage, in tandem with reform o f irrigation management are also crucial to arrest environmental degradation, especially in some major irrigated areas o f Punjab and Sindh. Expanded use o f water-conservation technologies such as drip irrigation, can increase the efficiency o f use o f scarce available water. Finally, more resources are needed for the provision o f research, extension and veterinary services for livestock, particularly for larger animals. The poultry sector i s already expanding rapidly, spurring substantial growth in maize production (a major feed) and in poultry and egg production. Productivity gains in the livestock sector are especially important for pro-poor rural income growth, as livestock ownership in rural Pakistan i s distributed far more equitably than i s land ownership. 70 4 GROWTHINTHE RURALNON-FARM ECONOMY As discussed in the previous chapter, substantial agricultural growth alone is insufficient to achieve rapid reduction inrural poverty. This i s so because agricultural earnings accrue mainly to those with access to the key factors o f production (land and water) and because growth linkage effects on incomes in the rural non-agricultural sector are small relative to the number of households deriving incomes from this sector. Thus, increasing rural non-farm incomes, in addition to those arising from growth linkages associated with increases in agricultural incomes, i s critical for rapid rural poverty reduction. This chapter examines the structure o f the rural non-farm economy and the constraints it faces. W e utilize two sources o f data for the analysis. Inpresenting an overview o f the rural non-farm economy and the characteristics o f households that derive their incomes from it, we use data from national household surveys and focus on households residing in rural areas (using the Federal Bureau o f Statistics definition o f rural areas based on administrative boundaries and population). The analysis o f the economic characteristics and constraints o f the non-farm sector, however, i s based on data from a separate rural investment climate survey o f enterprises (including household-based enterprises). Because many o f the enterprises that reap the benefits o f agricultural growth linkages and that employ members o f rural households are located in small towns, the survey covers both villages (rural areas as defined by the Federal Bureau o f Statistics) and small towns with populations o f less than 100,000 people. STRUCTUREOF RURAL NON-FARM ENTERPRISES Non-farm enterprises' contribution to income and employment There are no precise figures on the number of rural non-farm enterprises in Pakistan, but extrapolation from recent surveys suggests there are about 3.8 million rural non-farm enterprises.62About 70 percent o f these are located in Punjab, 15 percent in NWFP and 10 percent in Sindh. Data from the 2004-05 PSLM survey indicates that about 29 percent o f rural households in Pakistan operate a non-farm enterprise (Table 4.1). The rate o f enterprise ownership varies from about 19 percent inrural Sindh to 34 percent inrural Punjab. Table 4.1. PercentageofRuralHouseholdsinPakistanthat Own Shops and/or other Businesses Expenditure Quintile Pakistan Punjab Sindh NWFP Poorest Quintile 24% 28% 14% 20% Quintile 2 27% 33% 18% 24% Quintile 3 29% 35% 16% 30% Quintile 4 30% 35% 20% 33% Quintile 5 37% 41% 24% 45% Total 29% 34% 19% 30% Source: Authors' calculations basedon PSLM 2004-2005. __ ~ 62 This estimate is based on PSLM 2004/2005 data and utilizes the Federal Bureau o f Statistics definition o f rural areas that does not include small towns. Note that these PSLM-based estimates are considerably higher than estimates for rural households those based on HIES 2001102, which collected information on households with enterprises o f fewer than nine workers. HIES data indicate about 2.5 million rural non-farm enterprises. 71 While wealthier householdsare morelikely to own non-farmenterprises,poorer households are reliant on employment and thus for a significant share of household income in these enterprises. In all three provinces, enterprise ownership tends to increase monotonically with wealth. For instance, in rural NWFP 45 percent o f the wealthiest households own and operate a non-farm enterprise, compared to 20 percent o f the poorest households (Table 4.1). About 27 percent of the income o f the poorest households in rural Pakistan comes from non-agricultural wages (Table 4.3). In absolute terms, the poorest rural households earn three times as much income fkom non-agricultural wage employment as from agricultural wage employment. Almost 40 percent of the rural working population in Pakistan is either self-employed(13 percent) or employedas wage or salariedworkers innon-farmactivities (25 percent) (PSLM 2004-05).63Nearly half of all self-employed persons are engaged inwholesale or retail trade (The construction and social/personal services sectors are major sources of non-agricultural wage employment and account for roughly 25 percent and 40 percent o f paid non-agricultural work, respectively. Note that this non-farm employment need not take place within rural areas (as definedby the Federal Bureau o f Statistics), but may take place innear-by small towns (or even larger cities). Table4.2. Proportionof the RuralWorkingPopulationinPakistan (Over 10Years of age) that is Self-employedor a PaidNon-AgriculturalEmployee Pakistan Punjab Sindh NWFP Self- Non- Self- Non- Self- Non- Self- Non- employed Agricultural employed Agricultural employed Agricultural employed Agricultural persons workers persons workers persons workers persons workers as%of as%of as%of as%of as%of as%of as%of as%of Total Total Total Total Total Total Total Total Working Working Working Working Working Working Working Working Industry Population Population Population Population Population Population Population Population Mining & Quarrying 0.0% 0.2% 0.0% 0.1% 0.0% 0.2% 0.0% 0.4% Manufacturing 1.4% 3.2% 2.0% 4.4% 0.4% 1.5% 0.7% 1.8% Electricity 0.1% 0.3% 0.1% 0.3% 0.0% 0.5% 0.1% 0.5% Construction 0.4% 6.1% 0.3% 5.9% 0.4% 3.9% 0.5% 10.7% Wholesale & Retail Trade 6.2% 1.7% 6.3% 1.5% 4.6% 1.4% 8.5% 2.6% Transport & Storage 1.2% 2.1% 1.2% 1.5% 0.4% 1.8% 2.0% 3.9% Real Estate, Insurance 0.1% 0.0% 0.1% 0.0% 0.0% 0.1% 0.1% 0.1% Social & Personal Services 3.3% 9.6% 4.4% 9.2% 1.4% 9.2% 3.1% 11.7% Other 0.6% 2.1% 0.7% 1.9% 0.4% 2.1% 0.4% 2.9% Total 13.2% 25.5% 15.3% 24.9% 7.6% 20.7% 15.4% 34.5% Source:Authors' calculations based on P S L M 2004-2005.. Incomefrom self-employmentandwage employment in non-agriculturalactivities accounts for 30 percent of income for the average household (Table 4.3). The relative importance of non-farm incomes varies across provinces, with total non-farm income accounting for 67 percent o f household income in rural NWFP, 43 percent in rural Punjab and 32 percent in rural Sindh. Both poor and wealthy rural households are highly dependent on non-farm income. The poorest households are relatively more dependent on non-farm self-employment and wage incomes than richer households. 63 These estimates correspond closely to others from the Pakistan Labor Force Survey 2001-02. In 2001-02 nearly 40 percent o f the rural labor force was employed in the manufacturing, construction, wholesale and retail trade, transport and communicationand social-services sectors (Malik 2005). 72 Table4.3. AnnualHouseholdNon-farmIncomeSharesby ExpenditureQuintiles inRural Pakistan(in2001-02 Rs) Average Poorest Income Expenditure Quintile Quintile 2 3 4 5 AII (Rs) Punjab AgriculturalWages 6Yo 4% 2Yo 1% 0% 2% 1,489 Total Farm (excl. agric. wages) 38% 47% 53% 55% 64% 55% 40,175 Net Business Income 15% 12% 11Yo 11% 9% 11% 7,864 Non-agriculturalWages 26% 21% 20% 17% 12% 17% 12,738 Total Non-farm 57% 48% 45% 44% 36% 43% 31,246 Sindh Agricultural Wages 24% 17% 8% 4% 2Yo 7yo 5,791 Total Farm (excl. agric. wages) 42% 52% 61Yo 63% 67% 61% 52,865 Net Business Income 3% 4% 4% 5% 5% 5% 4,139 Non-agriculturalWages 28% 24% 26% 25% 19% 23% 19,809 Total Non-farm 34% 31% 32% 34% 31% 32% 27,587 NWFP AgriculturalWages 2% 2% 1% 1% 0% 1% 481 Total Farm (excl. agric. wages) 24% 30% 36% 36% 30% 32% 21,882 Net Business Income 12% 10% 12% 16% 16% 14% 9,761 Non-agriculturalWages 28% 25% 23% 19% 20% 21Yo 14,444 Total Non-farm 75% 68% 63% 63% 70% 67% 45,676 Pakistan Agricultural Wages 9% 7% 4% 2Yo 1% 3% 2,285 Total Farm (excl. agric. wages) 37% 45% 51Yo 53% 58% 52% 33,576 Net Business Income 12% 10% 9Yo 10% 9% 10% 7,451 Non-agriculturalWages 27% 24% 23% 21% 16% 20% 15,247 Total Non-farm 54% 48% 45% 45% 41% . . .- 45% 34,005 Source: Authors' calculations based on HIES 2001-2002; PIHS 2001-2002. A Profile of Pakistan's Rural Non-farm Enterprises As in other South Asian countries, the non-farm sector in both rural villages and small towns in Pakistan primarily consists of family based micro-enterprises. According to data from the Pakistan Rural Investment Climate (RIC) Survey, the average small town enterprise employed roughly two workers, and village enterprises employed about 1.7 workers), including paid and un-paid family workers and hired workers. Only three percent o f non-farm enterprises in small towns and one percent invillages had more than five workers. The vast majority o f labor in small town and village enterprises i s family workers; only 23 percent o f enterprises in small towns and 11percent invillages report hiringnon-family workers. Trading enterprises dominate in Pakistan's villages and small towns. About 57 percent o f small town and village enterprises are engaged inwholesale or retail trade, followed by around 31 percent in services; the remainder are in productiodmanufacturing (Table 4.4).64The very thin 64 These estimates of the sectoral distribution of enterprises from the RIC survey are very similar to those from PSLM Round 1 2004/05. According to PSLM, about 47 percent o f people self-employed in non-agricultural activities were engaged in wholesale or retail trade and about 34 percent were in the 73 manufacturing base in small towns and villages i s noteworthy, particularly by comparison to the rural non-farm sector in other South Asian countries. In Sri Lanka, production-oriented firms account for 40 percent o f rural enterprises, while in Bangladesh manufacturing and construction account for about 27 percent o f non-farm enterprises (ADB and World Bank 2004; World Bank 2005).65Ago-processing enterprises, primarily flour mills, bakeries and sweet shops account for about 32 percent o f manufacturing enterprises in small towns and 36 percent in rural villages.66 Typical service-sector businesses are repair shops for cycles, vehicles and small electronics; hotels and restaurants; tailors; and hairdressers and barbers. While the rural sectoral composition o f enterprises i s quite similar in all three provinces, small towns in Punjab have significantly more manufacturingand service enterprises than do small towns in Sindh and NWFP. The average enterprise is fairly young, and 50 percent havebeeninoperationfor less than 5 years, reflecting high startup and closure rates. Manufacturing/productionenterprises tend to be older than service and trading enterprises. The average age o f a production enterprise in the sample i s around 16 years for small-town firms and 11years for firms invillages; the average age o f both village and small town trading and services enterprises i s about eight years. About 43 percent o f production enterprises have been in operation for more than 10 years, compared to 29 percent o f trading enterprises and 25 percent o f service enterprises. The vast majority of owners had establishedtheir own businesses; about nine percent had inherited businesses and another five percent reported purchasing their business. On average enterprise owners had about three years o f industry experience prior to establishing or buying an enterprise. Owners acquired relevant management and technical skills in a number of different ways. Just under half o f business owners (42 percent) acquired skills from relatives and friends; 20 percent were self-taught; and about the same proportion had gained experience from prior employment. Most enterprises were established/purchased using funds borrowed from family and friends, or income from non-farm activities (Table 4.6). The predominant formof ownershipis sole proprietorship, and 80 percentofbusinessesare owner-managed. Most of these enterprises can be considered informal businesses as they have fewer than 10 workers, and very few are registered or pay taxes. Although rates o fregistration are generally low, non-farm enterprises inPunjab are more likely to be registered than those in Sindh or NWFP. Only 9 percent o f small-town enterprises and 10 percent o f village enterprises in Sindh are registered, as compared to 32 percent o f small town enterprises and 21 percent of village enterprises inPunjab (Table 4.4). On average it takes about nine days (ifnecessary) for a village enterprise to get a permit to operate and about 12 days in small towns (Annex Table A 4.2). Between one and two government agencies are involved inthe issuingof permits and costs range from around Rs 500 in villages to Rs 1,000 in small-towns. In comparisons with corresponding data from Sri Lanka, it i s clear that inPakistan the registrationprocess for non-farm enterprises i s longer and costs are higher. While few non-farm enterprises in villages and small towns are registered inPakistan, more than half o f comparable businesses in Sri Lanka are registered. service sector (defined as including social andpersonal services, real estate and insurance and transport and storage). 65 The sectoral distribution o f enterprises in Pakistan i s similar to that o f Nicaragua, where manufacturing accounted for about 6 percent o f rural non-farm enterprises, 42 percent were in the service sector and the remainders were intrade. 66 The main agricultural commodities grown inthe surveyed villages were wheat, rice and sugarcane inPunjab; wheat, sugarcane and cotton inSindh; and wheat, maize andjawar (sorghum) inNWFP. 74 Table 4.4. Profile of Non-farm Enterprises Punjab Sindh NWFP AII Urban Rural Urban Rural Urban Rural Urban Rural Employment Average number of workers 1.81 1.65 2.25 2.15 1.87 1.55 1.99 1.71 Size distribution 2 or fewer workers 84% 92% 76% 72% 87% 89% 81% 88% 2-5 workers 14% 6% 21Yo 28% 10% 11% 16% 11% More than 5 workers 2% 1% 4% 1% 3% 0% 3% 1% Firms hiringworkers 19% 9% 25% 30% 31% 3% 23% 11% Age Average age of firm 9.89 8.29 7.41 8.76 6.81 6.01 8.42 7.91 Age Distribution 2 years or less 14% 26% 15% 8% 15% 27% 15% 23% 2-5 years 26% 29% 39% 34% 29% 30% 31% 30% 5-10 years 24% 19% 19% 20% 30% 17% 23% 19% More than 10years 36% 26% 27% 38% 27% 26% 31% 28% Manager profile Manageris male 98% 99% 100% 100% 93% 100% 98% 99% Years of experience 3.09 1.87 2.92 3.35 5.11 4.13 3.35 2.56 Level of Education None 13% 35% 8% 10% 24% 23% 13% 29% CompletedSecondary 27% 13% 30% 29% 33% 26% 29% 18% F.A./FSc 11% 7% 18% 18% 13% 18% 14% 11% Professional Degree 1% 1% 1% 3% 0% 1% 1% 1% UniversityDegree 5% 3yo 13% 3% 6% 9% 8% 4% Sectoral Distribution Production 12% 11% 4% 19% 3% 8% 8% 12% Services 41% 32% 27% 28% 25% 32% 33% 31% Trade 46% 57% 68% 53% 73% 60% 59% 57% Registration& Taxes Registeredfirms 32% 21% 9% 10% 12% 9% 19% 17% Pay income taxes * 17% 5% 12% 16% 4% 0% 13% 6% Sole-proprietorships 91% 94% 92% 96% 93% ' 91% 92% 94% Source: Authors' calculationsbasedon the PakistanRural InvestmentClimate Survey, 2005. . * There were a large number o f missing observations for the question on income taxes. The numbers reported inthis table assume that firms did not pay income taxes if this question was unanswered. Owners and managers of enterprises tend to be relatively inexperienced, with an average of three years experience. Managers o f small-town enterprises are slightly better educated than village entrepreneurs. Small town entrepreneurs were less likely to report having no education than rural entrepreneurs (13 percent versus 29 percent) and slightly more likely to have a university degree (8 percent versus 4 percent) (Table 4.5). Virtually all non-farm enterprises are owned and operated by men. This i s not surprising as female participation in Pakistan's labor force, for activities outside the home was 11percent in 1999-00, the lowest in South Asia among countries with employment data by gender (World Bank 2004). Median annual sales revenue for small town enterprises, (Rs 180 thousand) are about 1.8 times larger than median annual sales revenues for village enterprises (Rs 102,000) (Table 75 4.5).67 Revenue from both village and small-town enterprises in Sindh are almost twice as large as revenue for Punjab and NWFP enterprises. Even the smallest enterprises in Sindh have median revenues 2.3 times those o f the smallest NWFP enterprises and 1.7 times those o f the smallest Punjab enterprises (Annex Table A 4.3). Across provinces and regions, median revenues are consistently highest for traders and lowest for service-sector enterprises. Table 4.5. Enterprise Sales and Assets (`000 Rs) Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Sales revenues (mean) 328 206 453 265 192 138 355 202 Sales revenues (median) 151 98 295 180 160 96 180 102 Value of fixed assets (mean) 102 46 121 51 39 31 99 44 Value of fixed assets (median) 10 14 25 14 10 7 15 12 Fixed assets per worker (mean) 72 35 69 27 31 21 64 30 Fixed assets per workers (Median) 7 11 13 7 6 6 8 10 Value added (mean) 59 30 87 53 (38) 24 55 32 Value added (median) 48 25 46 31 14 11 43 24 Value added per worker (mean) 41 24 53 29 (3) 21 39 24 Value added per worker (median) 28 20 28 18 7 7 27 18 Source: Authors' calculations based on the PakistanRural Investment Climate Survey, 2005. The median value added per worker was Rs 18,000 (about $300) for village enterprisesand around Rs 27,000 ($450) for small-town enterprises. These estimates are very close to those of rural non-farm enterprises in Bangladesh ($480) and Vietnam ($420). The median value added per worker in NWFP i s considerably lower than that in Sindh and Punjab. In both o f these provinces the median value added per worker in small-town enterprises i s about 50 percent higher than it i s for village enterprises; however in NWFP the median value added per worker i s virtually the same inboth small towns and villages. The medianvalue of fixed assets was around $200 for village enterprisesand $250 for small- town enterprises; the mean however was considerably higher, about $2,830 for small-town enterprises and $682 for village enterprises. Once again, these figures are quite similar to those observed for similar types o f enterprises in other countries. Recent survey data from Vietnam indicate median fixed assets in rural enterprises o f around $168. Although Pakistan's enterprise sector does not appear to be particularly dynamic, data indicate some employment growth in this sector. The average annual compound employment growth has been about one percent for village enterprises and three percent for small-town enterprises.68 THERURALINVESTMENT CLIMATE It is widely recognizedthat a good investment climate provides opportunities and incentives for firms to invest productively, create jobs and expand (World Bank 2004). To gauge the effects o f the investment climate on the performance o f non-farm enterprises in Pakistan's small " This is based on an exchange rate ofUS$1=60 Pakistanrupees. Employment growth rates are calculated for the subset o f firms that reported data for both 2001 and 2004. 76 towns and villages, the RIC Survey asked entrepreneurs to subjectively appraise the various political, administrative, economic and infrastructural conditions affecting their enterprise's operation and growth. In addition to these subjective indicators, data were also collected on various objective measures o f the investment climate. The most significant obstacles identified in the survey by small-town and village entrepreneursin the provinces of Sindh, Punjab and NWFP were access to formalfinance, the cost of finance and cumbersomeloanprocedures (Figure 4.1). Withthe exception of small towns in Sindh, more than a third o f entrepreneurs cited access to finance as a serious obstacle to business operation (Annex Table A4.5). Access to finance was rated the single most important overall constraint facing firms in Punjab and NWFP. In Sindh cumbersome loan procedures topped the list o f the most important overall constraint. Poor infrastructurealso ranks as a serious constraint for businessesin villages and small towns. Access to electricity, the quality o f electricity, road quality, and availability o f transport are among the main infrastructure constraints identified. The disparity in perceptions between village entrepreneurs and small-town entrepreneurs with respect to these infrastructure constraints i s quite strikmg. A larger proportion o f village entrepreneurs reportedinfrastructure constraints as a major obstacle to business than did small-town entrepreneurs (11 percent compared to 5 percent). While infrastructure bottlenecks pose more o f a challenge to village enterprises, enterprises in small towns perceive economic policy uncertainty, crime and corruption as relatively more serious than their village counterparts. L o w market demand also ranks fairly high among the various constraints faced by firms inboth rural and small-town Punjab, indicating that entrepreneurs there perceive that they do not have access to a large enough consumer base. The Pakistan Rural Investment Climate data indicate considerable differences in the way rural and smalltown entrepreneurs perceive investmentclimate constraintsas comparedto entrepreneurs in larger cities, but there are also some similarities. The most significant constraints identified by urban (large city) firms in the Pakistan Investment Climate Assessment 2003 related to tax administration and tax rates; more than 45 percent o f surveyed urban firms reportedtax-related issues as major or severe obstacles (World Bank, 2005g). By contrast, a mere three percent o f rural firms considered taxation a major obstacle. This i s hardly surprising given that many small-town and village enterprises are informal firms and few pay taxes. On the other hand, issues related to the cost o f finance and access to finance appear to pose a challenge for rural, small town, and urban firms alike. Electricity and corruption are also common constraints, although the proportion o f rural and small town firms that rate these problems as major constraints i s considerably smaller. This may reflect differences in the types o f businesses found incities as opposed to small towns and villages. 77 vl 0 0 e4 7 a, L E %J ._ 'C t + t -E + -m m 0 +0 v) v) m a, 8 2 U Figure4.2: Pakistan:Major Ruraland SmallTownInvestment Climate Constraints Top IC Constraints Percent of firms reporting majorkevere constraints E E Punjab Punjab Sindh Small Town Rural Small Town I Access to finance Loan procedures Cost of finance Cost of finance Cost of finance Low market demand Loan procedures Access to finance Loan procedures Low marketdemand Crimeltheft Econ. policy uncertainty NaturalGas Cornrpton 0 10203040 0 1020304050 0 10 20 30 40 % of firms % of firms % of firms E IE I? Sindh NWFP NWFP Rural Small Town Rural Loan procedures Access to finance Cost of finance Access to finance Cost of finance Loan procedures Roadquality Loan procedures Access to finance Cost of finance Access to electricity Availabilityof transport Availabilityof transport Quaiityof electricity Access to electricity 0 10203040 % of firms 0 1020304050 QcB01c60 % of f i n s 01% of f i n s Source:PakistanRuralInvestmentClimate Survey,2005. Figure4.3: InvestmentClimate ConstraintsIdentifiedbyUrban FirmsinPakistan, 2003 % of firms reporting majorhevere constraints 0 10 20 30 40 50 Tax Administration Tax Rates Cost of Financing Corruption Regulatory Policy Uncertainty Electricity Access to Financing Source: World Bankand SMEDA 2003 Though the business challenges faced by non-farm enterprises in villages and small towns are appreciably differentto those for businessesoperatinginlarge cities, the former are not dissimilar to those reported by rural non-farm enterprises in other countries. When the subjective assessments of Pakistan's small town and village entrepreneurs are compared to those made by their counterparts in Sri Lanka, Nicaragua and Tanzania, countries for which 79 comparable data are available, numerous similarities are evident (World Bank 2006~).In Nicaragua and Tanzania access to finance ranks as a major problem for 27 percent and 47 percent of rural entrepreneurs, respectively. While access to finance i s o f less concern in Sri Lanka, (where only nine percent o f businesses identified it as a major or severe constraint), 29 percent of Sri Lanka's rural entrepreneurs reported that the cost o f finance was a major challenge. Infrastructure problems rank high among the constraints identified by non-farm enterprises in Pakistan, but fewer firms there complain about infrastructure than in other countries. This does not necessarily mean that infrastructure conditions inrural Pakistan are superior to those o f other countries; perceptions o f what benefits and what hampers investments depend on business size, activities, capabilities, and the degree o f formality. Data on objective measures would need to be carefully compared in order to accurately assess differences in the availability and quality o f infrastructure across comparator countries. Relatively few rural entrepreneurs in Pakistan identify governance issues and policy uncertainty as a major problem.InNicaragua and Tanzania about 12percent and 17 percent of firms, respectively, identified crime as a major constraint compared to 3 percent o f village enterprises and 7 percent o f small-town enterprises in Pakistan. Thirty-seven percent and 17 percent o f Nicaraguan and Tanzanian enterprises cited economic policy uncertainty as a major concern. The corresponding figures for Pakistan are two percent for village enterprises and seven percent for small-town enterprises. However, economic policy uncertainty i s a more important concern in small towns in Punjab where 15 percent o f firms reported this as a major or severe problem. The proportion o f entrepreneurs that identify governance as a major problem in Sri Lanka and Pakistan are comparable; however slightly more firms in Pakistan complained about corruption and crime than in Sri Lanka. Access to Finance Access to formal credit rates as the top business constraint among village and small-town non-farm enterprises inPakistan. Roughly similar proportions of enterprises also rated the cost of finance and cumbersome loan procedures as major obstacles. O f surveyed business owners, 28 percent inrural areas (34 percent in small towns) wanted to apply,for a formal loan in the five years preceding the survey. However, only about 14 percent inrural areas (20 percent in small towns) o f these enterprises did actually go on to apply for a loan. Of those wanting a loan, but that did not apply for one, 40 percent stated that loan procedures were too complicated, 27 percent felt that the interest rate would be too high, 16 percent felt that they had insufficient collateral and 8 percent stated that the duration o f the loan would be too short. Overall, few firms (four percent in villages and seven percent in small towns) applied for a formal loan to finance investment or working capital inthe five years preceding the survey. Enterprises in Sindh seem to have relatively better access to finance. There, the proportion o f enterprises applying for loans was highest (8 percent o f village enterprises and 14 percent o f small-town enterprises). By contrast, only one percent o f small-town enterprises and three percent o f village enterprises in Punjab and NWFP applied for loans. Comparison o f these results with those from other countries reveals that the proportion o f small-town and village enterprises that receive loans from formal lenders inPakistan i s relatively low. InSri Lanka, 25 percent o f enterprises receivedformal loans in 2002-03. Nicaraguan businesses also received formal loans at a similar rate in 2004 (28 percent). Inthe same year in Tanzania, however, only six percent o f firms received formal loans, which i s comparable to Pakistan (World Bank 2006d). 80 , Table4.6: Pakistan: Ruraland SmallTownFirms'Access to Finance,2005 Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Firm wanted to apply for a loan in past 5 years 41% 36% 38% 24% 7% 7% 34% 28% Firms that applied for loans (of those wanting a loan) 3% 8% 38% 34% 17% 39% 20% 14% Firms that applied for a loan (amongall firms) 1% 3% 14% 8Yo 1% 3% 7% 4% Firm has a PLS account 10% 12% 12% i8% 5% 9% 10% 12% Firm has current account 3% 2% 13% 17% 4% 6% 8% 6% Firm has overdraftfacility 1% 2% 10% 4% 4% 2% 6% 2% Firm has access to supplier credit 53% 43% 59% 48% 45% 51% 54% 46% Repayment periodsupplier credit (median days) 10 15 15 15 10 15 15 15 Share of goodslinputs purchasedon credit 50% 50% 50% 50% 50% 50% 50% 50% Preparesa financial statement 2% 1Yo 22% 13% 6% 10% 11Yo 5% Made new fixed investment 39% 31% 80% 87% 57% 44% 58% 43% Mean value of new fixed investment (Rs.) 26,570 22,554 21,479 13,973 14,561 25,868 21,889 20,452 Medianvalue of new fixed investment (Rs.) 10,000 5,000 10,000 8,000 4,000 10,000 8,500 8,000 Share of investmentfrom own savings 95% 94% 88% 90% 89% 67% 90% 87% Share of investmentsfrom formal loans 0% 0% 2% 0% 0% 0% 1% 0% Share of investmentfrom relatives/friends 4% 5% 5% 8% 10% 33% 5% 12% Share of investmentsfrom private lenders 0% 1% 5Yo 1% 1% 0% 3% 1% Source of start-up capital Familyor friends 53% 38% 43% 28% 32% 50% 46% 39% lncome from agriculture 6% 17% 3% 20% 7% 14% 5% 17% lncome from non-farm activities 36% 39% 19% 35% 36% 25% 29% 36% Remittances 0% 3% 2% 0% 8% 3% 2% 2% Sale of assets 7% 6% 19% 13% 9% 0% 12% 6% Bank loan 0% 1% 2% 1% 0% 0% 1% 0% Private money lenders 2% 1% 1% 1% 0% 0% 1Yo 1% Other 5% 8% 20% 9% 11% 14% 11% 9% Share of start-up capital from: Family or friends 49% 33% 41% 25% 29% 48% 42% 35% Income from agriculture 5% 16% 2% 18% 7% 13% .4% 16% Income from non-farm activities 31% 37% 15% 34% 35% 24% 25% 34% Remittance 0% 2% 2% 0% 9% 3% 2% 2% Sale of assets 6% 5% 19% 13% 9% 0% 12% 5% Bank loan 0% 0% 2% 1% 0% 0% 1% 0% Private money lenders 2% 1% 1% 1% 0% 0% 1% 1% Other 5% 6% 18% 8% 11% 11% 11% 7% Source: Authors' calculations based on the Pakistan RuralInvestment Climate Survey, 2005. The major source of formal enterprise finance is commercial banks. O f 1651 enterprises surveyed, 70 enterprises reported applying for at least one formal loan.69 Firms that applied for loans are evenly split between small towns and villages and around 40 percent o f such firms were in Punjab, another 40 percent were in Sindh and the remainder inNWFP. Sixty-two percent o f f i r m s applied to commercial banks for loans, 17 percent applied to NGOS,8 percent applied to institutions that finance Small and Medium Enterprises (SMEs) and the remainder to other 69 Only five enterprises reportedapplying for more than one loan 81 source^.'^ Two-thirds of loan applications were approved by lenders. Among rejected applications, 36 percent were refused because o f insufficient collateral, 8 percent o f applicants had no co-signer and the remaining were rejected for a variety o f other reasons. Almost all formal enterprise loans require collateral and are fairly short-term in nature. Median loan size was Rs 25,000 (the average was about Rs 72,000) and about 76 percent o f approved applicants received the full amount they had applied for.'l Loan terms are short, ranging from about two months to five years; the median loan term was for one year and the average was about 15 months. The median annual interest rate on loans was 12 percent. For almost 80 percent o f approved loans, collateral was required by the lender. This i s unsurprisingas previous studies on the availability o f farm credit also found that collateral to be a requirement for all formal loans (World Bank 2004). The most common forms o f collateral were propertybuildings (used in 28 percent o f loan applications) and land (used in 26 percent o f loan applications). On average the value o f collateral required was equal to 132 percent o f the loan amount (the median value was 100 percent). The fact that only a small fraction of enterprises that seek loans are actually able to apply for one is consistent with observed trends in Pakistan's formal sector, which is largely urban based and which advances to the rural sector loans worth only a small fraction (3.5 percent) of those advanced to the urban sector. Commercial banks appear to face higher costs and risks in lending to small businesses, while the latter often do not know what i s needed to qualify for bank financing and lack critical market information. Small businesses are geographically dispersed, and lenders usually face greater costs in identifying potential borrowers, conducting due diligence, and maintaining contact with the borrower after a loan has been made. Small entrepreneurs also often lack formal financial records and have typically more limited track records. Only about 5 percent o f village enterprises and 11 percent o f small town enterprises inPakistan prepare financial statements (Table 4.6). Overall very few firms appear to deal with the formal banking sector. At most 10-12 percent of enterprises keep savings accounts and less than 10 percent maintain checking accounts. A very tiny fraction o f village (two percent) and small-town enterprises (six percent) have access to an overdraft facility to finance investment or working capital needs. There i s a strong correlation between lack o f access to banking services and the degree to which firms cite finance as a major constraint. In small towns in Sindh only 14 percent o f enterprises mentioned access to finance as a major or severe constraint. Thirty-eight percent of firms in small towns in Sindh that wanted a loan actually applied for one (Table 4.6); 22 percent o f Sindh's small-town enterprises prepared financial statements. The primary sources of start-up capital for village and small town non-farm enterprises are funds borrowed from family and friends and income from nonfarm activities (Table 4.6). The share o f start-up capital from bank loans i s negligible (one percent for small town enterprises and less than one percent for village enterprises). These findings are not dissimilar to results from other countries. New business owners have unproven track records and are therefore less likely to secure formal finance. The major source of investment finance was savings and funds from family and friends. About 43 percent o f village enterprises and 58 percent o f small-town enterprises made new investments inthe year preceding the survey. Among firms making new investments, the median investment was Rs 8,000 ($133) for village enterprises and Rs 8,500 ($142) for small-town enterprises. Close to 90 percent o f all new investments were made using savings. The average ' O NGO finance for enterprises appears to be used only inSindh. 71 The average size of the loan received was equal to 92 percent of the amount applied for. 82 investment which ranged between Rs 20,400 and Rs 21,900 ($340 and $365) was less than half the size o f the average investment made by similar types o f enterprises in Sri Lanka. While access to long-term finance and formal credit for working capital and investment is limited, Pakistan's non-farm enterprises seem to have better access to supplier credit than do firms in other countries.. Approximately 46 percent o f village enterprises and about 54 percent o f small-town ones purchase inputs/goods on supplier credit. Small-town enterprises in Sindh appear to have the best access to supplier credit o f the three provinces. This type o f credit i s generally only extended on a very short term basis with most enterprises required to make repayment within two weeks. Across the various sectors, traders are significantly more likely to have access to supplier credit than manufacturing or service enterprises. Firms purchase about halftheir inputsusing supplier credit (Table 4.6). Infrastructure Village enterprises are more constrained by poor infrastructure than those located in small towns. Road quality, availability o f transport, and access to electricity ranked among the top constraints reported by village entrepreneurs. Quality and access to electricity i s also a concern for businesses in small towns. Entrepreneurs in villages in Sindh were more likely to identify road- and transport-related problems as a major constraint those in other provinces. Data on road quality (the type o f internal roads in the community), availability o f public transport and connectivity reveal that villages in Sindh are indeed disadvantaged compared to enterprises located elsewhere. For instance, 93 percent o f enterprises in rural Sindh are located in communities where dirt roads are the most common type o f internal road surface (Table 4.7). By contrast, only about 54 percent o f village business owners in NWFP and 26 percent in Punjab report being located in a community where dirt roads are commonest. Twenty-eight percent o f village enterprises in Sindh are located in communities with public transportation to the nearest major city. For Punjab and NWFP the corresponding figures are 85 percent and 47 percent, respectively. Although the distance to the nearest main city from villages in Sindh i s comparable to those for villages in Punjab and NWFP, travel times are considerable longer (Table 4.7). This may reflect both the poor quality o f roads and the lack o f available public transport. Fewer small- town entrepreneurs complain about road quality and availability o f public transport than do village entrepreneurs; it i s not surprising that small-town enterprises have better roads and access to public transport. Access to electricity remains a major challenge in many rural villages and small towns, and even for enterprises that have access, supply is unreliable. The last urban investment climate assessment in Pakistan found that the typical business loses 5.6 percent o f annual output due to power outages, much higher than in China (1.99 percent) and Bangladesh (2.35 percent) (World Bank and SMEDA 2003). Although electricity connectivity in small towns and villages in Pakistan appears to be quite high(83 percent o f village enterprises and 96 percent o f small town enterprises reported having access to electricity), almost all enterprises with access to the grid experience power outages.'* In villages and small towns o f Sindh as well as villages in NWFP, entrepreneurs report almost daily outages. The median number o f days with power outages in a typical month was reported as being 20 days in villages and 15 days in small towns. Frequent outages increase production costs and generators are usually prohibitively expensive for these small business. Approximately five percent o f village enterprises and around seven percent o f small-town enterprises reportedowning or sharing a generator. 72 Connectivity to the grid appears relatively high compared to rural Sri Lanka where slightly less than 70 percent of enterprises reported using electricity. 83 Table4.7. Qualityof, andAccess to, Infrastructure Punjab Sindh NWFP AII Urban Rural Urban Rural Urban Rural Urban Rural Roads 8 Transport Proportionof enterpriseslocated in areas where: Internalroads are dirt roads 1Yo 26% 37% 93% 0% 54% 15% 42% Publictransport is availableto the nearest main city 98% 85% 85% 28% 92% 47% 92% 68% A main road connectsthe communityto nearestcity 91% 97% 93% 82% 100% 98% 93% 95% Distance and Travel Time Median distance to nearestmain city (Kin) 40 15 32 14 12 15 32 15 Median travel time to nearest main city by mainmeans of transport (minutes) 60 30 40 60 25 45 45 30 Median travel speed to nearest main city (kmlhr) 40 30 48 14 29 19 43 29 Electricity Enterpriseswith accessto electricity 97% 79% 94% 81Yo 98% 96% 96% 83% Enterprisesconnectedto the grid reporting power outages 98% 99% 100% 100% 100% 100% 99% 100% Median days/monthwhen outages occur 10 10 30 30 20 30 15 20 Own or share a generator 4% 7% 12% 4% 5% 0% 7% 5yo Median days to get newelectricity connection 40 45 20 30 30 20 30 30 Unofficialfees neededfor a new connection 81Yo 90% 100% 100% 97% 100% 91% 94% Telecommunications Owns fixed-line phone 28% 9% 24% 4% 39% 6% 28% 7% Owns cellular phone 10% 3% 3yo 3% 19% 6% 9Yo 4% Uses fax 0Yo 0% 1Yo 0% 3% 0% 1Yo 0% Uses email 1% 0% 0% 0% 2% 0% 1% 0% Source: Pakistan Rural Investment Climate Survey, 2005 Access to and use of telecommunications among enterprises surveyed is surprisingly limited, particularly consideringthat close to 60 percent of the enterprises are engaged in trade. About 28 percent o f respondent small-town entrepreneurs owned fixed line phones and 9 percent owned cellular phones (Table 4.7). The comparable figures for village entrepreneurs are seven percent for fixed-line phones and four percent for cellular phones. On the whole, small- town entrepreneurs inM P appear to have slightlybetter connectivity than their counterparts in the other two provinces. The most recent urban investment climate assessment found that only 30 percent o f urban businesses inPakistan regularly communicate with customers or suppliers using the Internet-far less than the 71 percent o f firms that do so in China or the 45 percent that do so inIndia. Access to Internet connectivity for rural and small-town firms is extremely limited and only about one percent o f the latter reportedusingemail. 84 k Travel. Tittie nrrd Krrrrtl Itwestrnent CiirnateSur Marketing About 18 percent of rural enterprises and 26 percent of small-town enterprises reported that they face obstacles with respect to marketing. Somewhat surprisingly small-town entrepreneurs were more likely to report marketing-related problems despite being located in larger communities (the average small town's population was around 43,000 compared to a population o f 7,300 in villages) and having access to more customers and markets (Annex Table A 4.6). Low market demandrankedamong the top five constraints reported bybusinesses inboth small towns and villages. There are however, significant differences across provinces. Entrepreneursin Punjab were more likely to report marketing obstacles (42 percent in small towns and 25 percent in villages) than their counterparts in Sindh and NWFP (13 percent in small towns and 6-7 percent invillages) (Annex Table A 4.4). This is despite the fact that Punjab appears to have the best access to markets o f the three provinces. The small towns and villages surveyed in Punjab are considerably larger than those surveyed in Sindh and NWFP (Annex Table A 4.6). Other measures o f market access, including travel times to larger cities and a market accessibility index, also suggest that Punjab has relatively better access to markets. In small towns in Punjab, production-oriented enterprises are almost 30 times as likely to report problems with marketing as service-sector firms, but there i s no significant difference between the ratings of production- related enterprises and traders respectively with respect to the constraints imposed by low demand. On the other hand, traders in villages were significantly more likely than production- related enterprises to cite low demand as a problem. With the exceptionof production-oriented firms in NWFP, the markets and customer base of village and small-town enterprises are limitedto local markets and most sales are made to customerswithinthe same tehsil (administrative division).Seventy percent o f sales by both village and small-town production enterprises in NWFP are to buyers in other provinces (Annex Table A 4.7). The exact reasons for these differences are not readily apparent. In all three provinces most services-sector firms and traders sell locally either within the same village, or to a different village within the same union council. Service-sector firms and traders in small towns make significantly larger sales to wider markets, although most sales still take place within the same tehsil. The vast majority o f sales are made directly to households or to other small firms. Technologyand BusinessServices Lack of innovation and poor product quality also limits the access of rural firms to wider markets. Few rural firms in Pakistan reported making improvements and/or technological innovationsover time. Only six percent o f small-town and village enterprises reportedthat they had made new innovations and/or technological improvements (Table A 4.8). The propensity of undertaking technological innovation i s significantly higher among larger firms, with better educated managers. One possible explanation could be that larger firms have the required resources to develop innovative products, whereas smaller firms do not. Lack o f financial resources, technological or management know-how, and the lack o f access to networks (such as business associations) are all factors that contribute to low levels o f product innovation among rural firms. Few rural and small-town entrepreneurs tap into formal business networks. Evidence from other countries suggests that participation in business organizations and local chambers o f commerce could potentially strengthen marketing channels. This allows businesses to share information on prices, quality standards and obtain technical, financial and organization services 86 for greater value addition. A very small proportion o f non-farm enterprises in small towns and villages in Pakistan are able to take advantage o f these networking opportunities. Even though there are no significant differences in participation across industry types, participation varies by location; village enterprises are significantly less likely to participate than small-town enterprises inall provinces. Two percent of village enterprises are members of a Chamber of Commerce or other business association compared to 11 percent o f small-town enterprises (Table A 4.9). Participation in business associations i s highest in small towns in Punjab, followed by small towns in Sindh and NWFP. Controlling for industry type and location, enterprises with more educated managers are significantly more likely to belongto a business association. Very few firms reported that they need or use formal business services, which is unsurprising given the informal nature of many rural and small-town businesses. Annex Table A 4.10 provides a summary o f the various business services that are accessible within an hour's travel time fiom the town or village in which the enterprises are located. While about a third o f small-town businesses have access to various types o f services such as legal services, accounting services and insurance only 5-15 percent o frural enterprises have similar access. Even though various services are available, virtually none o f the businesses reported needing or using available services. Security and Rule of Law Crime appears to be much less of a concern for enterprises located in small towns and villages than for those operating in large cities. A `politically and economically stable environment i s necessary to unleash productive investments (World Development Report 2005). The 2003 Investment Climate Assessment (ICA) found a high incidence o f urban crime and widespread concern among urban business managers over theft, arson, extortion, intimidation o f employees and destruction o f business property. NWFP fared particularly poorly; 32 percent o f businesses in NWFP reported having been victims o f crime, compared to 22 percent in Punjab and 11 percent in Sindh. On average, urban firms in NWFP reported spending 4.5 percent o f revenue on security; the comparable figures for firms in Sindh and Punjab were one and two percent. Of all small-town and village enterprises surveyed, crime ranks as one o f the top five constraints in only small towns in Sindh, where about 12 percent o f entrepreneurs identified crime as a major or severe problem. Overall about five percent of enterprises in villages and seven percent of those in small towns reported having actually been a victim of robbery or theft (Table A 4.11). The proportions o f village and small-town enterprises that report having experienced crime are similar across the three provinces. About a third (32 percent) o f small-town enterprises reported paying for security, with another 28 percent claiming they pay for protection. The corresponding figures for village enterprises are 16 percent and 11percent. Large numbers o f firms in Sindh reported making protectionpayments, worth an average o f about one percent o f sales revenue. In addition to having experienced more crime, businesses in Sindh are also more likely to claim to be locatedin neighborhoodsaffected by tension, conflict and violence. More than a third of small-town and village enterprises in Sindh were located in areas with some degree o f conflict, compared to 5 and 10 percent in Punjab respectively, and 14 percent o f small-town enterprises inNWFP, albeit none inrural NWFP (Table A4.12). The data also indicate that about 86 percent o f village enterprises in Sindh are located in communities divided along social, religious, economic or other lines. A relatively large proportion o f small-town enterprises in Sindh (67 percent) and rural businesses in NWFP (66 percent) are also located in areas that are 87 divided along social/religious/economic lines. In small towns and rural Sindh, the primary factor contributing to perceptions o f difference pertained to land ownership. This was a leading factor in perceptions about difference in rural areas o f NWFP, although differences in political affiliation was most frequently identified as the source o f problems. Insmall towns inNWFP, differences in ethnic/caste/tribal background as well as political affiliations contribute to perceived and real divisions inthe community. Very few Punjab enterprises whether invillages (seven percent) or in small towns (five percent) reported being located in communities that experienced conflict or other divisions. A majority of entrepreneurs (69 percent of those in small towns and 62 percent of those in rural areas) described laws and regulations affecting their businesses as predictable, although a large proportion of business owners in Sindh perceive laws and regulations in their communities that affect their businesses as highly unpredictable. Assurances that contracts will be honored, disputes handled fairly and efficiently by the legal system, and that legalrulings will be enforced are important for attracting investment as well as for doing business with geographically dispersed firms. The majority o f enterprises (67 percent o f small-town enterprises and 64 percent o f village enterprises) asserted that contracts offer them protection from being cheated (Figure 4.5). Confidence in contracts and contract enforcement i s highest in NWFP where 90 percent o f small-town entrepreneurs and 77 percent o f village entrepreneurs felt protected by a contract. Confidence incontracts and the legal system's ability to enforce them and property rights i s lowest insmall towns inSindh. Figure 4.5: Confidenceinthe Ruleof Law inThree Provinces Percent o f enterprises 0% 10% 20% 30% 40% 50% Punjab SmallTown Punjab Rural SindhSmallTown SindhRural NWFP Small Town NWFP Rural All Small Town All Rural Disagreethat contracts offer protection 0 Disagreethatthe legalsystemw illupholdcontracts in disputes Source: Authors' calculations based on the PakistanRIC Survey, 2005. Rural Enterprise Ownership: Start-up and Performance While infrastructuresuch as electricity and fixed phone lines seem to have a positive effect on enterprise ownership in small towns, in rural areas other factors such as the number of government programs in a community appear to have greater importance. As part of the Pakistan RIC survey, households were asked whether any member was a proprietor of, or partner in, a non-agricultural enterprise within the last 12 months. About 40 percent o f the 947 participating households had members that operated a non-farm enterprise. The determinants o f 88 owning a non-farming business were analyzed usingregression analysis (Table A 4.13). Inboth rural villages and small towns the effects o f parents' occupation are powerful. Whether as a result o f parental influence or because o f the likelihood o f actual business inheritance, ownership o f a non-farm enterprise increased by fifteen to twenty percentage points the likelihood o f one's offspring also owning one. Another important household characteristic i s size; larger households are more likely to diversify income by owning a non-farm enterprise. While household size i s a factor inthe rural sector, it i s more influential in small towns. More important inthe rural areas i s the amount o f land cultivated. Factors affecting start-up of non-farm enterprises include proximity to large cities for small-town enterprises and the number of government programs and access to communications facilities for those in villages. In addition to examining the determinants o f enterprise ownership the survey also examined the factors that influence households' decisions to start up a non-farm enterprise. For this regression, households were excluded that own older businesses (those more than three years old), in order to compare two categories: those households that recently decided to start a business and those that decided not to. One finding i s that households in small towns farther away from large cities are more likely to start businesses. This is not surprising as proximity to large cities means more competition. Small-town households also gain positive effects from having a male head o f household and more valuable assets. Inrural areas, government programs continue to play an important role and, as market access and communications are more significant constraints in villages, the effects o f cell phone usage are significant. Inaddition, in rural areas older heads o f household are more likely to start up a non-farm enterprise than younger ones. The regression results indicate that there is no significant effect o f parents' occupation in start-up decisions in either rural areas or in small towns. In addition to the overall investment climate, human capital, as measured by a manager's education, is an important determinant of village and small-town enterprise performance. The impact o f various investment-climate constraints on enterprise performance i s assessed by studyingthe impact o fthese factors on firmproductivity, as measuredby value added per worker (Annex Table A 4.14). For enterprises in small towns, availability o f public transport and the presence o f a chamber of commerce in the tehsil have a significant positive impact on performance. On the other hand, more power outages negatively affect firms located in small towns. For village enterprises, access to finance in the form o f a commercial bank in the community has a positive and significant impact on productivity; businesses located in communities with dirt roads seem to be negatively affected (although this variable i s only significant at the 10%level). Unsurprisingly the impact of the investment climate on a firm's performance varies by industry. Power outages have a significant negative effect on production-oriented firms although this is far less with service or trade businesses. Access to commercial banks has a positive and significant correlation with firm productivity for service-sector enterprises and production enterprise (although only significant at the 10% level for production enterprises), but i s not significant in the trade regression. Access to commercial banks i s probably more important for service- and production-related firms, as traders have relatively greater access to supplier credit. It is noteworthy that in virtually all circumstances, human capital as measured by the manager's education i s an importance determinant o f enterprise performance. Firms with more educated managers consistently perform better and, as mentioned earlier, are also more likely to innovate and access business networks. 89 POLICY IMPLICATIONS Although Pakistan's rural non-farm sector is an important source of employment and contributes significantly to improving household welfare, particularly of the rural poor, improvingthe sector's productivityand dynamismwill requiretargeted policies and public investments. While a number o f successful financial-sector reforms have been implemented in recent years, rural non-farm enterprises, which are primarily micro enterprises, seem to have benefited little from these. Access to formal credit for investment purposes or working capital is extremely limited, and evidence o f large unmet demand for credit suggests a need for alternative measures to ensure enterprises with bankable business plans access to external sources o f finance. Easing some o f the highcollateral requirements for formal loans and developing alternative forms o f collateral acceptable to banks would improve access to credit. There i s also evidence to suggest that loan application procedures need to be less onerous. At the same time, practical training for entrepreneurs in accounting and the preparation o f business plans would greatly improve the ability o fmany businesses to signal their credit-worthiness to financial institutions. Provision and improvement of infrastructure,particularly roads and reliable electricity, can reduce operating and marketingcosts, thereby makinginvestmentsin rural and small- town enterprises more profitable. There is a particular need in Sindh to investment in rural roads, as distances and transport costs there are higher than in Punjab and NWFP. Among vital measures to improve the quality o f roads and transport services are: better road maintenance and extension o f basic motorable access; improved institutional arrangements for ownership, management and financing o f rural transport systems in accordance with the realities o f devolution; and promoting greater community involvement in planning and managing transport infi-astructure improvements to ensure that infrastructure meets local needs (World Bank 2005). An unreliable electricity supply is also a severe obstacle to the development and growth o f non- farm enterprises, particularly in the manufacturing sector. Unlike their urban counterparts, few rural non-farm enterprises are able to afford generators, and unreliable electricity has likely limitedthe emergence of small manufacturing units, which also tend to generate relatively larger numbers o fjobs. Increased investment inpower distribution and transmission i s needed, but there has been little progress in implementing key reforms in the energy sector that would make these investments forthcoming . Other important measures to promote the growth of rural non-farm enterprises include: facilitating market linkages; improving the mechanisms that underpin contract enforcement; promoting transparency in the legal system; and allowing non-farm enterprises a greater say inthe policy makingthat pertainsto them. Stronger market linkages could be facilitated through the creation and development o f more effective local business associations and organizations. These organizations could arrange product fairs, help develop business directories, provide information services and educate entrepreneurs on how to undertake market research. Organizations could also share information on prices, quality standards and how to access technical, financial and organizational services for greater value addition. Facilitating group marketing and formation o f business clusters could also help rural enterprises take advantage o f economies o f scale by allowing the purchase o f inputs at lower prices, offering access to larger markets and facilitating shared use o f equipment and infkastructure. Improving mechanisms for contract enforcement would also facilitate market linkages because this would enable smaller enterprises to enter into sub-contracting arrangements with larger firms, thereby gaining access to larger markets. 90 ANNEXTO CHAPTER4: THEPAKISTANRURALINVESTMENTCLIMATE SURVEY, 2005 Despite the the non-farm sector's large contribution to rural incomes in Pakistan, there is limited information on the factors that affect the performance and growth of these enterprises. This knowledge gap i s in large part due to inadequate data in general on rural non- farm enterprises. Much existing information on enterprise performance in Pakistan i s based on surveys o f formal urban manufacturing firms or on household surveys that do not elicit data on the investment climate or measures o f enterprise performance. Given the nature and scale o f rural enterprises, it seems likely that they face considerably different challenges from their urban counterparts. To better understand the rural non-farm enterprise landscape and identify policies that could potentially spur growth in this sector, the World Bank with support from Department for International Development and the FA0 recently conducted a survey o f the investment climate for rural non-farm enterprises inPakistan (Box 4.1). This annex presents additional tables and findings from the survey not included inthe main section o f chapter 4. 91 Annex Table A 4.1: Distributionof Farm andNon-farmHouseholdsinRural Pakistan Punjab Sindh NWFP Baluchistan All Rural HHs Rural HHs Rural HHs Rural HHs Rural HHs FarmTotal 48.14% 45.19% 58.60% 44.81% 48.21% Farmer 43.21% 39.99% 52.09% 42.28% 44.83% Agricultural Labor 4.93% 5.19% 6.50% 2.54% 3.37% Non-FarmTotal 51.86% 54.81% 41.40% 55.19% 51.79% Self-employed 17.99% 20.85% 12.00% 18.06% 11.12% Other 33.88% 33.96% 29.40% 37.13% 40.68% Source:PSLM2004-05. Annex Table A 4.2: BusinessRegistration Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Registering an enterprise Business requiresa permit or license 6% 2% 100% 91% 100% 46% 54% 19% Days to complete registration process 21 14 33 11 8 7 24 9 Days to get a permit to operate 13 14 11 9 7 12 9 Numberof governmentagencies that need to approve a permitto operate 1 2 1 2 1 2 1 Cost of a permit to operate (Rs) 500 5,000 2,000 500 1,000 500 1,000 500 Enterprisespunished if not registeredin communitieswhere registration is required 100% 100% 74% 100% 84% 8% 79% 53% Source: Authors' calculations based on the Pakistan Rural Investment Climate Survey, 2005 Annex Table A 4.3: Sales Revenues Punjab Sindh NWFP AII Quintiles Small Small Small Small of Sales Town Rural Town Rural Town Rural Town Rural 1 Mean 28,929 25,211 41,373 42,466 35,863 18,932 26,432 25,543 Median 27,000 27,000 36,000 45,000 36,000 19,800 25,000 26,250 2 Mean 85,729 54,009 126,187 92,967 81,814 55,397 72,482 65,689 Median 90,000 52,540 132,000 100,000 80,000 56,000 74,000 62,000 3 Mean 157,794 100,190 254,090 171,486 121,587 89,462 140,020 130,633 Median 151,000 98,400 246,000 180,000 100,000 98,000 144,000 124,000 4 Mean 352,840 214,389 439,334 306,808 241,406 151,604 289,021 276,717 Median 320,000 200,000 428,000 280,000 240,000 144,000 295,000 280,000 5 Mean 1,170,954 632,310 1,307,414 973,406 588,096 441,581 984,940 818,557 Median 840,000 480,000 1,260,000 714,000 432,000 310,000 726,000 660,000 Total Mean 327,551 206,147 453,129 265,012 192,287 138,073 354,822 202,245 Median 151,000 98,400 295,000 180,000 160,000 96,000 180,000 102,000 Source: Authors' calculations based on the Pakistan Rural Investment Climate Survey, 2005 92 Annex Table A 4.4: PercentageofFirmsIdentifyingVariousInvestmentClimateFactorsas a Constraint Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Public Utilities 45% 43% 29% 34% 21% 21% 35% 37% Transport 20% 33% 26% 46% 12% 25% 21% 33% Financing 68% 63% 59% 63% 80% 81% 66% 67% Marketing 42% 25% 13% 7% 13% 6% 26% 18% Registration, Licenseand Permits 1% 0% 3% 2% 0% 1Yo 1% Taxation 5% 4% 15% 6% 3% 2% 9% 4% Labor 2% 1Yo 0Yo 0% 1% 0Yo 1% 1% Land 0% 1% 1% 8% 6% 1% 2Yo Agricultural Policy 1Yo 2% 0% 2Yo 0% 5yo 0% 2% Trade Policy 0Yo 0% 1% 0% 0Yo Environmental Policy 1Yo 0% 0% 1Yo 0Yo Governance 35% 23% 20% 14% 3yo 0% 24% 17% Source: Authors' calculations based on the PakistanRural Investment Climate Survey, 2005 Annex TableA 4.5: Major & Severe Investment ClimateConstraintsIdentifiedby Enterprises Punjab Sindh NWFP All Small Small Small Small Town Rural Town Rural Town Rural Town Rural Access to finance 43% 49% 14% 32% 46% 36% 32% 37% Cost of finance 29% 26% 26% 25% 46% 38% 31% 28% Loan procedures 17% 19% 40% 37% 40% 37% 30% 33% Low marketdemand 19% 14% 5% 6% 1% 1% 10% 6Yo Quality of electricity 10% 10% 6% 8% 4% 2Yo 7yo 7yo Access to electricity 6% 8% 8% 14% 8Yo 9Yo 7yo 12% Economicpolicyuncertainty 15% 8% 1% 0% 0Yo 0% 7yo 2Yo Crime/ theft 5% 5% 12% 4% 0% 0% 7yo 3% Corruption 1% 1% 12% 3% 0Yo 0% 6Yo 2Yo Access to markets 9% 6% 3% 1% 0% 1% 5% 2% Roadquality 5% 9% 6% 24% 0% 3% 5% 16% Market information 5% 1% 4% 6% 0% 0% 4% 3% Naturalgas 8% 12% 0% 3% 1% 1% 4% 5% High tax rates 1% 0% 7% 5% 0% 0% 3% 3% Availability of transport 2% 6% 5% 21% 0% 6% 3% 14% RoadAccess 5% 7yo 1% 8% 0% 2Yo 3% 6% Illegaltax levies 3yo 1Yo 3% 0% 2Yo 0% 3% 0% Complicatedtax rules 0% 0% 5% 3% 0% 0% 2% 2% Road Blocks 2% 1Yo 2% 6% 0% 1Yo 2% 4% Water Supply 1% 1Yo 2% 2% 0% 9Yo 1% 4% Telecommunications 1% 2% 0% 3% 2% 1% 1Yo 2% Postal Service 1% 1% 0% 3% 0% 0% 0% 2% Source: Authors' calculations based on the PakistanRural Investment Climate Survey, 2005 93 Annex Table A 4.6: Access to Markets Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Populationof village/smalltown 50.841 9,080 38,999 4,379 31,393 4,078 43,026 7,337 Main Markets HouseholdsUse to Buy/Sell Goods Groceryin the townhillage 81% 27% 45% 0% 6% 22% 55% 22% Groceryinnearby townhillage 0% 0% 13% 2% 0% 1% 5% 0% Commercialcenter 0% 11% 0% 17% 76% 3% 12% 10% Nearest city 19% 62% 38% 81% 18% 63% 26% 65% Median distance to the nearest commercialcenter (Km) 1 15 19 16 1 14.5 1 15 Median time to the nearest commercialcenter (minutes) 5 45 30 60 5 25 10 45 Public transport availableto commercial center 98% 85% 91% 100% 100% 100% 96% 90% Median cost of public transport to commercial center (Rs) 10 14 15 20 3 10 15 14 Transport hire cost from the nearest commercial center Bus (Rs) 1,500 2,400 2,000 3,000 300 500 2,000 2,000 Three wheeler (Rs) 100 100 200 150 50 100 100 100 Tractor (Rs) 1000 800 600 740 300 500 600 700 Bus fare from commercialcenter (Rs) 20 15 20 15 3 10 20 15 Bus travel time from commercial center (minutes) 75 45 45 60 5 25 45 45 Three wheeler fare from commercial center (Rs) 16 12 20 15 3 5 15 14 Three wheeler travel time from commercialcenter (minutes) 60 30 20 35 5 5 17 30 Source: Authors' calculations based on the Pakistan Rural Investment Climate Survey, 2005 94 Annex Table A 4.7: Market Linkages; Geographic Location of Sales and Trading Partners Production Services Trade Village Enterprises Punjab Sindh NWFP Punjab Sindh NWFP Punjab Sindh NWFP Sales to Buyers in: Same village 29 94 25 74 71 56 81 80 87 Same UC, differentvillage 34 2 2 24 25 38 17 16 9 Same Tbesil, different Union Council (UC) 12 0 0 1 2 0 1 2 1 Same district,different Tbesil 16 0 0 1 0 5 0 2 1 Same Province, different district 8 4 2 0 2 0 0 0 0 Other Provinces 0 0 71 0 0 0 0 0 1 Other countries 0 0 0 0 0 0 0 0 0 Small town Enterprises Sales to Buyers in: Same town 41 74 21 55 74 60 44 76 39 Same UC, differenttown 12 20 11 25 17 33 28 19 16 Same Tbesil, different UC 42 0 0 16 6 4 26 3 38 Same district,different Tbesil 3 1 0 4 4 1 1 2 7 Same Province,different district 2 4 0 1 0 0 0 0 0 Other Provinces 0 0 68 0 0 1 0 0 0 Other countries 0 0 0 0 0 0 0 0 0 Village Enterprises Types of Buyers: Government 3 0 0 0 0 5 0 0 0 Traders 15 1 0 Multinationalslocated in your country 0 0 0 2 0 0 0 0 0 Your parent company or affiliated subsidiaries 2 0 0 1 10 0 0 0 0 Large domestic firms 3 0 0 0 0 0 0 1 0 Agricultural producers and cooperatives 4 17 71 2 0 0 3 0 9 Households 60 75 27 85 79 92 92 77 88 Other (salesto small firms, etc) 13 7 2 10 12 3 5 22 3 Small-town Enterprises Government 0 0 0 0 0 1 0 0 0 Traders 6 0 0 Multinationalslocatedin your country 0 0 0 0 0 2 0 2 0 Your parent company or affiliated subsidiaries 0 0 0 0 0 0 0 0 0 Large domesticfirms 0 0 0 0 1 0 1 0 0 Agricultural producers and cooperatives 2 1 0 4 0 0 1 1 1 Households 76 76 29 77 75 84 84 83 82 Other (sales to small firms, etc) 16 23 71 19 24 14 14 14 17 Source: Authors' calculations'based on the Pakistan Rural Investment Climate Survey, 2005 95 Annex Table A 4.8: TechnologicalImprovements/Innovations Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Proportionof enterprisesthat made improvements/technology innovations 10% 6% 2% 10% 5% 1% 6% 6% Type of improvementsmade Innovation in product 40% 48% 0% 1% 4% 50% 29% 32% IrnprovemenVinnovationin equipmenthachinery 37% 9% 38% 96% 47% 24% 39% 38% Improvement in workers through hiring 0% 0% 62% 0% 14% 0% 11% 0% Improvementin management 18% 19% 0% 3% 6% 27% 14% 14% Other 5% 23% 0% 0% 30% 0% 7% 15% Source: Authors' calculationsbasedonthe PakistanRuralInvestmentClimate Survey, 2005 Annex TableA 4.9: Participation inBusinessAssociations Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Proportionof enterpriseslocatedin a tehsilwith Chamber of Commerce 0% 0% 32% 10% 25% 2% 17% 2% Businessassociationfor a//types of businesses 43% 2% 79% 51% 100% 100% 66% 29% lndustry specific businessassociation 10% 0% 76% 47% 86% 36% 48% 14% Proportionof enterprisesbelonging to Any businessassociation 15% 2% 9% 3% 5% 0% 11% 2% Chamber of Commerce 1% 1% 2% 3% 4% 0% 2% 1% Businessassociationfor a//types of businesses 14% 2% 7% 3% 5% 0% 10% 2% lndustry specific business assocation 1% 0% 0% 0% 0% 0% 0% 0% Source: Authors' calculationsbasedonthePakistanRuralInvestmentClimate Survey, 2005 96 Annex Table A 4.10: Availability and Use of Business Services Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Proportionof enterpriseslocatedin Tehsils where the followingservicesare available: Engineeringservices 12% 0% 46% 14% 38% 60% 26% 15% Managementservices 0% 0% 56% 16% 33% 28% 21% 9% Marketingservices 14% 1% 56% 14% 38% 35% 30% 11% Accountingservices 8% 0% 56% 12% 33% 21% 25% 7% Legal services 22% 7% 59% 16% 39% 61% 35% 19% Insuranceservices 17% 1% 58% 14% 100% 100% 32% 15% Informationtechnology services 7% 2% 61% 17% 39% 61% 27% 16% Proportionof enterpriseswith businessservicesavailablewithin a hour's travel time from town/village: Engineeringservices 59% 4% 13% 6% 9% 7% 33% 5% Management services 42% 4% 32% 16% 23% 17% 35% 8% Marketingservices 23% 7% 32% 19% 9% 17% 24% 11% Accountingservices 37% 7% 19% 13% 23% 17% 28% 10% Legal services 41% 16% 39% 20% 9% 7% 35% 15% Insuranceservices 65% 9% 32% 13% 8% 7% 43% 9Yo Informationtechnology services 55% 8% 37% 19% 8% 7% 40% 9% Proportionof enterprisesreportingthat they needed: Engineeringservices 0% 0% 1% 0% 1% 0% 1% 1% Management services 0% 1% 0% 0Yo 0% 0% 0Yo 0% Marketingservices 0% 0% 0% 0% 0% 0% 0% 0% Accountingservices 0% 0% 0% 0% 0% 0Yo 0% 0% Legal services 1% 0% 1% 0Yo 0% 0% 1Yo 0% Insuranceservices 0% 0% 0% 0% 0% 0% 0% 0% Informationtechnology services 1% 0% 0% 0Yo 0% 0% 0Yo 0% TechnicalSUDDO~from ~ buyers/suppl'iers 0% 0% 5% 6% 0% 0% 2% 1% Source: Authors' calculations based o n the Pakistan Rural Investment Climate Survey, 2005 Note: There are quite a few missing observations for the sections on business services in both the community and enterprise questionnaires. The numbers in the table should be interpreted in the following way: 12 percent o f enterprises inPunjab are located intehsils where engineering services are available. The other 88 percent are located in tehsils where theses services are either not available or respondents in the survey where unaware as to whether these services were available. Similarly 1 percent o f enterprises .in small towns reported that they needed engineering services, the remaining either reported that they did not need these services or that they didnot know ifthey needed these services. 97 Annex Table A 4.11: Laws and Regulations Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Proportionof enterprisesreportingthat laws andregulationaffectingtheir businessesare: Highly predictable 4% 7yo 4% 7yo 44% 46% 11Yo 15% Somewhat predictable 62% 50% 67% 67% 26% 24% 58% 47% Unpredictable 25% 25% 8% 8% 23% 19% 18% 21% Highly unpredictable 3% 3% 16% 18% 5% 6% 9% 6Yo Proportionof enterprisesreportingthat laws, rules andregulationbeingimplementedin their communityare: Highly predictable 3yo 7% 6% 6Yo 36% 39% 10% 13% Somewhat predictable 61% 48% 52% 55% 28% 24% 52% 44% Unpredictable 26% 26% 22% 24% 28% 25% 24% 26% Highly unpredictable 3% 3% 16% 13% 5% 5% 8Yo 5% A contractoffersprotection from being cheated-proportion of enterprisesthat: Strongly agree 4% 4% 5% 6% 52% 21% 12% 8% Agree 63% 53% 52% 67% 38% 56% 55% 56% Disagree 26% 26% 40% 25% 8% 18% 28% 24% Strongly disagree 0% 0% 0% 1% 0% 0% Legalsystem will upholdcontractsandproperty rightsin disputes-proportion of enterprisesthat: Strongly agree 7% 5yo 1Yo 0% 37% 19% 9% 7% Agree 63% 45% 46% 49% 47% 56% 54% 48% Disagree 23% 33% 45% 39% 15% 19% 30% 31% Strongly disagree 0% 0% 3% 9% 0Yo 0% 2% 1% Proportionof enterprisesthat incurred expenses for: Providing security 30% 15% 22% 23% 57% 15% 32% 16% Providing protection payments 15% 7% 50% 34% 9% 5% 28% 11% Due to theft, robbery, vandalism 0% 0% 1% 1% 0% 0Yo 0% 0% Enterprises reportingbeing a victim of theft or robbery 8Yo 5yo 8Yo 5% 9% 4% 7% 5yo Proportionof enterprisesthat resolved dispute using Neighbors and friends 4% 2Yo 1% 10% 1% 0% 2% 3% Community leaders 0% 0% 3% 7% 0% 0% 1% 1Yo Police station 2% 2% 1% 9% 1% 1% 2% 3% Magistrate court 0% 0% 0% 0% 0% 0% 0% 0Yo Source: Authors' calculations basedon the PakistanRural Investment Climate Survey, 2005 98 Annex TableA 4.12: Social Cohesion Punjab Sindh NWFP AII Small Small Small Small Town Rural Town Rural Town Rural Town Rural Proportionof enterpriseslocated in communitiedneighborhoods that are Very united 17% 54% 12% 21% 5% 44% 18% 45% Somewhat united 76% 41% 29% 50% 81% 56% 58% 46% Affected by some quarrelsand conflicts 6% 0% 36% 7% 14% 0% 19% 1% Affected by great tension and conflict 2% 5% 9% 29% 0% 0% 4% 8% Affected by conflict and violence 0% 0% 0% 2% 0% 0% 0% 0% Homogenous 44% 42% 0% ,30% 66% 49% 30% 41% Have problems due to religious/ethnic/socialand other differences 5% 7% 67% 85% 24% 66% 33% 31% Differencesthat cause problems Differencein education 43% 0Yo 0% 0% 0% 0Yo 3yo 0% Differencein landholding 18% 20% 42% 37% 7% 37% 36% 35% Differencein wealthhaterial possessions 7% 27% 0% 23% 1% 0% 1% 14% Differencein social status 16% 0% 12% 16% 0% 0Yo 11Yo 7% Differencebetween men and women 0Yo 0% 6% 0% 0% 0Yo 5yo 0% Differencebetween young and old 0Yo 0% 0% 6% 0% 0% 0% 3yo Differencein politicalaffiliation 9% 0% 21% 15% 36% 60% 22% 32% Differencein religiousbeliefs 0% 0% 19% 0% 0% 0% 15% 0Yo Differencein ethnic backgroundlracelcastehribe 7% 53% 0% 3% 56% 3% 7% 10% Source: Authors' calculations based onthe PakistanRural Investment Climate Survey, 2005 99 100 I I I I I I I I Source: Authors' calculations based on the Pakistan Rural Investment Climate Survey, 2005 101 Annex Table A 4.14: Labor ProductivityRe-gressions (OLS) (1) Small town (2) Rural (3) Production (4) Services (5) Trade Ln(FixedAssets) 0.029 0.042 0.059 0.126 0.000 (1.25) (1.44) (155) (3.70)"' (0.02) Age of firm 0.018 0.011 0.016 0.013 0.017 (1.79) (1.27) (1.66) (1.47) (1.78) Manager's experience (years) 0.030 -0.015 0.051 -0.005 -0.002 (1.98) (0.91) (2.83)"* (0.32) (0.06) Manager's gender 1.413 0.121 1.I82 2.254 0.518 (1.83) (0.25) (1.90) (1.78) (1.38) Manager's education (none) -0.686 -0.650 -0.919 -0.451 -0.960 (2.36)* (1.65) (2.47)" (1.47) (2.66)** Manager's education (up to secondary) -0.423 -0.414 -0.382 -0.420 -0.479 (2.63)" (2.41)" (1.00) (2.40)" (2.36)" Production -0.527 -0.231 (3.56)"* (0.95) Services -0.363 -0.372 (2.04)" (1.88) Commercial Bank 0.342 0.391 0.674 0.945 -0.154 (0.97) (3.34)"* (1.72) (4.16)"* (0.85) Internalroad: dirt road 0.392 -0.285 0.136 0.003 -0.109 (1.63) (1.92) (0.41) (0.01) (0.56) Publictransport avaiIable 0.658 0.009 0.483 0.272 0.085 (3.70)"* (0.08) (1.42) (0.94) (0.58) Electricityoutages (dayslmonth) -0.016 0.002 -0.034 -0.001 -0.007 (2.07)" (0.44) (2.97)"* (0.13) (1.28) % households with fixed-linephones 0.006 -0.002 -0.006 0.005 -0.002 (1.63) (0.64) (0.68) (1.17) (0.62) YOhouseholds with cell phones -0.006 0.002 -0.003 -0.001 -0.006 (1.79) (0.59) (0.52) (0.24) (2.03)" Community conflict -0.093 0.294 -0.930 -0.286 0.267 (0.26) (1.56) (1.33) (1.25) (1.34) Community Population 0.000 0.000 0.000 0.000 0.000 (0.64) (4.24)"* (154) (153) (1.71) Distanceto city 100K 0.001 -0.001 -0.001 0.001 0.000 (1.12) (1.17) (0.71) (0.54) (0.48) Chamber of Commerce 0.548 0.446 -0.100 -0.186 0.295 (2.60)" (1.60) (0.16) (0.53) (1.61) Attock -0.429 0.545 0.178 0.022 0.529 (2.36)" (2.70)** (0.44) (0.07) (2.30)" Faislabad 0.138 -0.191 -0.486 0.460 0.346 (1.19) (1.48) (1.54) (3.40)"* (1.75) Other Punjab 0.532 -0.001 0.376 0.082 0.311 (3.30)"* (0.01) (0.74) (0.31) (1.70) Sindh 0.072 0.166 -0.638 0.460 0.124 (0.36) (0.57) (1.76) (1.86) (0.60) % households with electricity 0.001 (0.26) Rural 0.296 0.375 -0.276 (0.61) (1.37) (1.56) Constant 7.734 10.473 8.610 5.575 10.908 (7.31)"* (11.64)** (4.60)** (3.46)** (17.64)** Observations 623 483 140 401 565 R-squared 0.22 0.19 0.48 0.30 0.17 *Robustt-statisticsinparentheses significantat 5% level; ** significantat 1% level Source: Authors' calculations based on the PakistanRural Investment Climate Survey, 2005 102 5. PUBLIC EXPENDITURESAND RURALSERVICE DELIVERY As mentioned in the previous chapter, inadequate roads, electricity and telecommunications are major constraints on productivity and growth of Pakistan's rural and small-town enterprises. Similarly, low levels o f public service related to health and education contribute to a disparity between Pakistan and other south Asian countries on many social welfare indicators. Improving rural service delivery i s thus crucial bothfor economic growth and improved welfare for the poor. This chapter analyzes this issue by examining: the changes in total fiscal resources available at lower levels o f government over time; the composition o f expenditures; and constraints to more efficient service delivery. Following a brief examination o f devolution in Pakistan, the chapter presents data on total revenues and expenditures at various levels o f government; these show that total resources increased at the federal and provincial levels from 2001-02 to 2005-06, but also that the transfer o f resources to local governments remains uneven. Based on the results o f a 2006 survey o f selected district and Tehsil Municipal Authority (TMA) governments, the chapter also identifies administrative issues and provides suggestions for reform. BACKGROUND: DEVOLUTION PAKISTAN IN International experience demonstrates that decentralization o f administrative and financial authority to lower levels o f government offers great scope for enhancing efficiency o f public- sector service delivery. Decentralization can increase efficiency in administration and allocation, improve equity in service provision, enhance participation and inclusion, empower local stakeholders and accelerate poverty reduction. Local governments are more likely to engage in participatory procedures that can identify local priorities and provide feedback on implementation and progress. Reforming the fiscal framework entails implementing incentives and fiscal arrangements that promote efficient use o f resources according to local needs, budgetary certainty, autonomy in preparing the budget and incentives to increase local revenues. Local revenues may also contain intrinsic incentives for efficiency, as local taxpayers are likely to be more vigilant than distant ones. Accountability relationships also can be improved through the political process o f citizens influencing policymakers and policymakers influencing service providers. Pakistan, like many other countries has devolved service-delivery responsibilities to local governments through an elaborate program initiated in 2001. The main features o f this program are: restructuring the bureaucratic set up and devolution o f administrative authority to the district level and below; refocusing administrative systems to allow public participation in decision- making; facilitating oversight o f government functionaries by the monitoring committees o f local councils and rationalizing administrative structures to improve efficiency. The program was implemented through creation o f three tiers o f local governments at the district, tehsil and union levels. The most significant change under the new system i s the transfer o f responsibilities for budgeting, planning and development o f service provision, previously decided at the provincial level, to the district governments. Health, education, assessment and collection o f land taxes, cesses and agricultural income tax, agricultural extension and farm water management, planning and design o f district roads and buildings are all now devolved from the provincial governments to the district governments. Tehsils and towns have assumed responsibility for delivering water 103 and sanitation, drainage, roads under their authority, street lighting and fire fighting.73No service provision responsibilities have been assigned to the union admini~tration.~~ The local governments have also been assigned powers to raise additionalrevenues under the new system and the distribution o f resources through the Provincial Finance Commission (PFC) formula75ensures direct distribution o f resources from the provinces to local governments as well as distribution among local governments. Under the new system o f government, citizen representation i s ensured through direct election o f the Union Councilors, and the Nazims (elected heads o f local government) have been made more accountable to the people. In addition Local Government Ordinance (LGO) 2001 also provides for the existence o f numerous new bodies through which citizens are able to access and potentially influence government decisions at the local level. These oversight bodies fall into three broad categories. There are those that provide external checks, like the Monitoring Committees at the district, tehsil and union levels. Through these committees citizens are able to influence the performance o f service delivery providers through their elected representatives, who are members on these committees. Village Development Councils (VDCs) and Neighborhood Councils are appointed under the provisions o f the LGO, which are designed to mobilize community resources. The VDCs may, with funds raised through voluntary contributions or on self help basis, develop and maintain municipal and community welfare facilities and can also assume authority to deliver services in their areas. Last are a group o f bodies responsible for community development and management o f facilities, The Citizen Community Boards (CCBs), formed voluntarily by citizens to mobilize resources aimed at improving service delivery and conveying community concerns to local governments. FISCALOUTCOMESANDISSUES Overall fiscal resources in Pakistan have increased at both the federal and provincial levels since 2001-02. Spending on rural development at the federal and provincial levels has increased as well. Total transfers to local government (districts, tehsils and unions) varied widely by province, as well as by district within provinces. Increases in government revenues since 2001-02 have provided the fiscal space for an expansion of expenditures at both the federal and district levels. From 2001-02 to 2005-06, total consolidated expenditures (fiscal and provincial), government revenue and the budget deficit all increased by 23 to 26 percent in real terms. Expenditures grew by an average o f 5.9 percent per year and 23 percent overall in the period. The increase in overall spending was largely financed by stronger revenue: 62 percent o f the revenue increase came from tax increases; the remainder was derived from non-tax sources. The overall government budget deficit increased by 23 percent in real terms, but remained equal to about one-quarter o f total expenditures over the period (Table 5.1 and Figure 5.1) 73 Under the old system these services were simultaneously providedby boththe rural and urban local councils and by the respective provincial line departments. 74 The federal government has sole responsibility for providing road transport, electricity, energy and telecommunication services. It shares policing andjudicial responsibilities with the provincial government, and health, education, water and sanitation provision with lower levels o f government. There are therefore multiple levels of government providing the same services. 75 The ProvincialFinance Commission formula defines the allocation o fthe transfers o frevenues from provincial governments to local governments. See World Bank (2004; Pak Public Exp M g m t , V o l I). 104 Table 5.1: Real Federaland ProvincialRevenuesand Expenditures inPakistan, 2001-02 to 2005-06, (in billion Rs,2004-2005 = 100) 2001-02 2002-03 2003-04 2004-05 2005-06 FYO2-06 Percent Change 2002-06 Consolidated Total Expenditures 1019.6 1029.4 1019.3 1195.5 1257.6 5.9% 23.3% Current Expenditures 864.1 907.4 827.1 943.1 970.3 2.7% 12.3% Development Expenditures 155.8 148.1 174.5 227.7 288.7 18.1% 85.3% Net Lending to PSEs -0.2 -26.0 17.7 24.8 -1.4 _-- --- Statistical Discrepancy -14.4 3.7 0.0 78.5 0.0 -_- --_ Total Revenue 770.2 826.1 873.4 900.0 968.3 5.6% 25.7% Tax Revenue 590.0 637.0 669.8 659.4 712.0 4.2% 20.7% Non-Tax Revenue 180.2 189.1 203.7 240.7 256.3 9.9% 42.2% Surplus/Deficit -235.0 -207.0 -145.8 -217.0 -289.3 __- 23.1% Federal Total Expenditures 768.6 824.6 764.2 823.9 893.1 3.0% 16.2% Current Expenditures 647.4 687.4 631.3 688.6 691.2 1.3% 6.8% Development Expenditures 121.2 137.2 133.0 135.3 201.9 10.6% 66.6% Transfers to Provinces -216.1 -221.8 -229.1 -245.3 -251.3 12.8% 16.3% Total Revenues 720.7 772.0 808.5 843.0 892.6 5.3% 23.9% Tax Revenue 566.8 612.0 632.8 624.8 674.1 3.7% 18.9% Non-Tax Revenue 153.9 160.0 175.7 218.3 218.5 10.6% 42.0% Surplus/Deficit -47.9 -52.6 44.2 19.1 -0.5 -__ -98.9% Provincial Total Expenditures 251.3 230.8 237.4 346.9 365.9 12.3% 45.6% Current Expenditures 216.7 219.9 195.9 254.5 279.1 6.7% 28.8% Development Expenditures 34.6 10.9 41.5 92.4 86.8 48.8% 150.7% Total Revenues 265.6 275.9 290.0 302.3 326.9 5.2% 23.1% Tax Revenue 23.2 25.0 36.9 34.6 37.8 13.9% 63.0% Non-Tax Revenue 26.3 29.1 24.0 22.4 37.8 4.8% 43.9% Transfers to Provinces 216.1 221.8 229.1 245.3 251.3 4.1% 16.3% SurpMDeficit 14.3 45.0 56.7 -44.6 -39.0 --- Source: Pakistan Economic Survey 2002-2003,2004-2005,2005-2006. Notes: Nominalvalues deflated with National Accounts Index of the Price of Private Consumption; Consolidated, Federal and Provincial Budget Numbers are from PakistanEconomic Survey 2005-2006, Table 4.2. Federal andProvincial Development Expendituresbasedon calculations from: PakistanEconomic Survey 2005-2006, Table 4.1;Pakistan Economic Survey 2004-2005, Table 5.9; andPakistanEconomic Survey 2002-2003, Table 5.6. Transfersto provinces basedon PakistanEconomic Survey 2005-2006, Table 5.13; PakistanEconomic Survey 2004- 2005, Table 5.10; andPakistan Economic Survey 2002-2003, Table 5.7. Transfersto provinces includeboth federal grants and federal tax receipts. 105 Figure 5.1: Government Total Expenditure and Financing*, 2001-02 to 2005-06 1 1.4 i 1.2 1.o 0.8 -.-% 8 0.6 0.4 0.2 0.0 2001-02 2002-03 2003-04 2004-05 2005-06 I I Tax Revenue Non-Tax Reenue IDeficit Financing I I Source: Pakistan Economic Survey 2005-06. Notes: * Consolidated budget figures. Deficit financing defined as total expenditures less tax and non-tax revenues. Development expenditures increased by 85 percent (18.1 percent per year) from 2001-02 to 2005-06, reflecting the government's increased commitment to development investments. Growth in development expenditures was particularly rapid at the provincial level, (48.8 percent per year) as compared to the federal level (10.6 percent per year). Combined, current federal and provincial government expenditure increased by only 2.7 percent per year in the same period. As a result, the share o f development in total consolidated expenditures rose from 15.3 percent in 2001-02 to 22.9 percent in2005-06 (Figure 5.2). Total expenditures by the provinces increased much more rapidly (45.6 percent) than they did at the federal level (16.2 percent), excluding transfers to provinces and net lending to Public Sector Enterprises (PSEs). The rapid growth in provincial expenditures was not due to increased transfers from the federal government, however; these increased by 16.3 percent, (essentially the same rate as other federal spending), (Table 5.2). Rather, Rs 53.2 (2004-05) bn of the Rs 114.6 (2004-05) bn increase was financed by a change in provincial budget surpluses, which fell from 5.7 percent to -10.7 percent o f total expenditures from 2001-02 to 2005-06. Moreover, federal and provincial development expenditures on agriculture and irrigation, which declined dramatically during the 1990s, have increased sharply since 2001-02 (Figure 5.3). Federal and provincial development expenditures on agriculture and irrigation fell from an average o f more than Rs (2004-05) 15 billion between 1987-88 and 1995-96, to less than Rs (2004-05) 5 billion in 1999-00 and 200-01, mainly due to a steep decline in development expenditures on irrigation. Large increases in irrigation expenditures at the federal level, from less than Rs (2004-05) 1 billion in 2000-01 to more than Rs (2004-05) 15 billion in 2004-05 largely account for the more than fivefold increase in total real federal and provincial expenditures on agriculture and irrigation inthis period. 106 Fi ire 5.2: FederalandProvincial GovernmentExpenditures*, 2001-02 to 2005-06 1400 I i 1200 1000 800 -.-5 ' 600 400 200 0 2001-02 2002-03 2003-04 2004-05 2005-06 Source: PakistanEconomic Survey 2005-06. Notes: * Consolidatedbudget figures Figure5.3: FederalandProvincial Agriculture and IrrigationExpenditures, 1987-88to 2004-05 20 15 10 5 0 Source: Pakistan federal and provincialbudgets. 107 z \o * 0 0 z F4 0 0 c, 0 N F4 ? 3 v) 0 4 0 hl B N0 z d 00 N z(3 0 M 0N 4: N 4 r 0 0N .-ca 0 n v) Y to W 3 0 n Q k 0 Ucn 2a .. .-'0 C E ?i - Q) U c3 a! E v) Pakistan's per capita total expenditure on agriculture and irrigation remain far below those o f other South Asian countries, however (Table 5.3 and Figure 5.4). Per capita total expenditures on agriculture and irrigation in India (including both current and development expenditures) are more than four times those of Pakistan; in Sri Lanka total expenditures are more than double the level o f Pakistan. Much o f India's current expenditures on agriculture and irrigation are on price subsidies that encourage excessive use o f fertilizer and water, and have contributed to depletion o f groundwater and other environmental damage (World Bank 2005). Spending on subsidies rather than on investments has contributed to poor apcultural performance in Latin America (de Ferranti et al., 2005). Pakistan lags its neighbors in development expenditures as well, however. Per capita development expenditures on agriculture and irrigation inIndia are double those of Pakistan; those in Sri Lanka are three times as large. Table 5.3: Public Expenditureson Rural Development and Social Services in South Asia (US$/person) Pakistan Pakistan Banglades India Sri Lanka h 2000-01 2004-05 2004-05 2004-05 2004-05 Agriculture Current 1.84 1.49 1.20 16.08 7.40 Of which: Agricultural Extension n.a. n.a. 0.62 n.a. n.a. Agriculture Development 0.19 0.51 2.88 0.56 10.50 Of which: Agricultural Extension n.a. n.a: 0.46 n.a. n.a. Irrigation Current 1.26 1.84 --- 3.09 --_ Irrigation Development 0.53 2.86 --- 6.12 -__ Rural Development Current I.47 7.47 0.00 Rural Development Capital __ _ _ _ _ __ __ __ 5.55 0.93 0.00 Total 3.82 6.70 12.19 34.26 17.90 Current 3.10 3.33 2.68 26.64 7.40 Development 0.72 3.38 8.43 7.61 10.50 Education Current 8.14 10.15 9.87 17.97 24.88 Of which: Primary Education 3.89 4.50 Education Development 0.31 3.08 3.33 0.27 7.10 Of which: Primary Education 0.10 0.70 Health Current 2.19 2.71 1.95 4.92 13.69 Health Development 0.43 0.79 2.26 0.28 6.32 Water Supply Current 0.21 0.18 _ _ _ 1.89 --- Water Supply Development 0.15 0.12 --- 1.06 -__ Total 15.42 22.23 17.42 26.39 51.99 Current 10.54 13.04 11.83 24.77 38.57 Development 0.89 3.99 5.59 1.61 13.42 Rural Population as % of Total * (in 65.9 65.9 73.0 71.7 76.2 2003) Rural PoDulation in Millions 92.8 100.2 101.6 788.7 14.8 Source: Sri Lanka CentralBank Annual Report2005; World Bank countryoffice, Bangladesh; World Bank country office, India; and World Bank countryoffice, Pakistan. Notes: Sri Lankawater supplyfiguresnot included. Sri Lankaand BangladeshAgriculture figures include irrigation. Pakistan2004-05 Figuresare preliminary 109 Figure5..4: ExpendituresonAgricultureandRuralDevelopmentin SouthAsia 40 =35 E 30 ; 25 m 20 3n 15 0 3 10 tft v) 3 5 0 2000-01 2004-05 2004-05 2004-05 2004-05 Pakistan Pakistan Bangladesh India Sri Lanka ~ ~ ~ ~ ~ Agriculture Current Agriculture Development Irrigation Current Irrigation Development w Rural Development Current Rural Development Capital Source: Sri LankaCentralBankAnnual Report 2005; World Bank countryoffice, Bangladesh;World Bank countryoffice, India; and World Bank countryoffice, Pakistan. Notes: Sri LankaandBangladeshAgriculture figures includeirrigation. Pakistan2004-05 Figures are preliminary. Likewise, in spite of improvements made since 2001-02, Pakistan's total (both urban and rural) per capita expenditureson health, educationand water supply remainbelow those of India and Sri Lanka (Figure 5.5). In 2004-05, Palustan spent $17 per person on these key public services, essentially the same as Bangladesh. Inthe same year, India spent $26 per person and Sri Lanka spent $32 per person, 1.6 times and 3.1 times as much as Pakistan, respectively. InPunjab,the only province for which detailedtime series budgetdatawere available, total expenditures increased by 66 percent between 2001-02 and 2005-06, or 13.1 percent per year. Current expenditures increased by only 19 percent inreal terms, however (4.0 percent per year), permitting a more than four-fold increase in development expenditures from Rs (2004-05) 14.3 to 79.2 billion over the period. An increase in federal transfers by Rs 2004-05) 52.6 billion funded most of the 88.0 (2004-05) billion increase in total Punjab expenditures. A further one quarter o f the increase in expenditures, however, was financed by a change in the provincial budget deficit, which increased from 0.2 billion to 22.1 billion over the period (Table 5.4 and Figure 5.6). 110 Figure 5.5: Expenditureson Social Services in South Asia 60 I 2000-01 2004-05 2004-05 2004-05 2004-05 ~ ~ Pakistan Pakistan Bangladesh India Sri Lanka ~ ~ Education Current Education Devdopment Health Current Health Development Water Supply Current Water Supply Development Source: Sri LankaCentral Bank Annual Report 2005; World Bank country office, Bangladesh; World Bank country office, India; and World Bank country office, Pakistan. Notes: Sri Lankawater supply figures not included. Pakistan2004-05 Figures are preliminary Table 5.4: Real Revenues and ExpendituresinPunjab, Pakistan, 2001-02 to 2005-06 2001-02 2002-03 2003-04 2004-05 2005-06 Growth Percent Rate Change 2002 - 2002- 2006 2006 Total Expenditures 133.2 160.2 185.5 199.2 221.1 13.1% 66% Current Expenditures 118.8 131.8 141.9 136.4 141.9 4.0% 19% Transfers to LOC.Gov 60.7 69.5 67.5 63.8 66.7 1.0% 10% Development Expenditures 14.3 28.4 43.6 62.8 79.2 52.4% 453% Total Revenues* 132.9 150.4 173.9 186.5 199.0 10.8% 50% Tax Revenue 13.0 14.4 19.7 19.1 23.6 15.9% 81% Non-Tax Revenue 12.4 10.9 18.8 15.3 15.3 7.8% 23% Transfers to Punjab 107.5 122.9 130.1 146.9 160.1 10.2% 49% Development Revenue ___ 2.2 5.3 5.2 0.0 54.5% 139% Surplus/Deficit -0.2 -9.9 -11.6 -12.8 -22.1 --- --- Source: Calculated from Punjab Budget Reports2003-06; Punjab Economic Report March 31,2005 Notes: Figures for the growth rate and percentagechange of Development Revenuesare for 2003 to 2005. 2005-06 data are preliminary. 2001-02 data for Development Revenue are unavailable. Real values are calculated using aprice index of private consumption from the National Accounts.. 111 Figure 5.6: RealRevenuesand Expenditures inPunjab, Pakistan, 2001-02 to 2005-06 250 225 200 cc 175 8c!150 125 -5 .- 100 a 75 50 25 0 2001-02 2002-03 2003-04 2004-05 2005-06 Source: Calculated from Punjab Budget Reports2003-06; Punjab Economic Report March 31,2005 Note: 2001-02 developmentexpenditure data includerural development expenditures. Punjab's expenditures on rural development as a share of total development expenditures varied between 12 and 15 percent from 2002-0376to 2005-06. Real expenditures on the Rural Development Programme, the largest budget item, almost doubled over the period, from Rs (2001-02) 2.3 to 4.5 billion. Real expenditures on rural water supplies increased even more rapidly, fi-om only Rs (2001-02) 0.05 billion in 2002-03 to Rs (2001-02) 3.25 billion in 2005-06. Livestock spending increased fourfold to Rs (2001-02) 0.72 billionby 2005-06, bringingit almost to parity with the level o f spending for crop agriculture and agricultural credit, which was Rs (2001-02) 0.94 billion in 2005-06, having itself increased by 25 percent per year in real terms between 2001-02 and 2005-06. Total transfers to local government (districts, tehsils and unions) varied by province, however. Transfers by the Punjab provincial government changed little inreal terms from 2001- 02 to 2005-06. The real rupee value o f transfers to district and tehsil governments increased by only 11 and 9 percent, respectively in the period, and transfers to union level governments actually fell by 12 percent in real terms. Similarly, in NWFP transfers to local governments remained essentially unchanged in real terms between 2002-03 and 2005-06. By contrast, however, in Sindh transfers to local governments inboth 2004-05 and 2005-06 were 50 percent or more higher in real terms than the province's average transfers in 2002-03 and 2003-04 (Figure 5.7). India experienced similar problems with the flow o f resources to local governments in the early years o f fiscal and administrative decentralization (Box 5.1). 76 No data were available for 2001-02. 112 Box 5.1: DecentralizationinIndia India has made bold efforts to strengthen the voice of the rural poor through decentralization to local governments. A 1993 Constitutional Amendment created a decentralized government structure. Bringing government closer to the rural people improves the match between diverse local preferences and public services. Decentralization strengthens the voice o f local people and accountability to them o f public sector decision-making. These benefits flow from the democratic process, as well as because o f greater transparency, as local actions are more visible to local people. India's size and diversity in terms o f ethnicity, language, geography, resource endowment, climate and levels o f economic development, makes designing effective central and state level programs extremely difficult. Thus, decentralization has become the critical core o f India's strategy to improve service delivery. The 73rd Constitutional Amendment o f 1993 created three tiers o f democratically elected sub-state rural governments: at district level (district panchayat or DP); at block level (blockpanchayat or BP); and at the village level (gram panchayat or GP)..The Amendment specifies a list o f 29 subjects that states can choose to devolve to rural governments at these three levels. Since decentralization was defined as a state subject, different states have pursued varying strategies to empower rural governments. Rural decentralization in India i s constrained by state finances. State fiscal deficits in India hover around 5 percent o f GDP. With decentralization defined as a state subject, this raises a critical issue of financing decentralization. As a result, even states that have a relatively well designed and mature decentralization program become captive to the state fiscal situation, and the result i s again under-funding o f decentralization initiatives relative to entitlement. For example, two states, Karnataka and Kerala, both showed large surpluses at the Gram Panchayat level (the level o f local government with discretion). The natural conclusion was that weak capacity at that level led to the accumulation o f huge fiscal surpluses. However, further investigation revealed that because o f the fiscal situation o f the state, actual disbursements to local governments were smaller than the corresponding figures inthe statebudgets. To correct this problem, financial management systems need to be updated so as to portray actual fiscal positions o f local governments, instead o f only budget figures. In addition, accelerating the pace o f rural decentralization will require either a debt-restructuring program for the states or improving the financial positions o f large cities so as to free up resources at the state level. Source: Fiscal Decentralization to Rural Governments inIndia (World Bank 2004). 113 Figure 5.7: Transfers to Local Governments by Provinces inPakistan, 2001-02 to 2005-06 I 8 0 / I s ; : I I i 40 E 30 P, 20 10 0 2001-02 2002-03 2003-04 2004-05 2005-06 +Punjab -INWFP --IQ---Sindh Source:Provincial Finance Departments; Niaziet al. 2006. A 2006 survey o f seven districts in Punjab, Sindh and NWFP77highlights the considerable variation in levels o f resources available at the district level. Per capita availability o f funds ranged from 981 Rs/person in Sargodha (Punjab) to 2000 Rs/person in Chakwal (P~njab).~'It i s noteworthy that large expenditure increases for health and education at the district level are not apparent in these budgets. On average, 88 percent o f the budgetary resources across the seven districts came from provincial government grants.79 In two Punjab districts, Chakwal and Bawalpur, however, drawdown o f cash balances accounted for 16 percent o f resources, reflecting unused funds from previous budgets. On average, current expenditures account for 80 percent of total expenditures; salaries alone account for 63 percent o f total expenditures. The share o f resources devoted to development expenditures nevertheless varies widely: from 3 percent in Lower Dir (NWFP) and 9 percent in Sargodha to 35 percent in Chakwal (Annex Table A 5.1). Likewise, the performance o f CCBs in terms o f their financial contribution as a percentage o f the government financing o f development expenditures varies from only 2 to 6 percent in the three sample districts in Sindh up to between 22 and 34 percent inthe three Punjab districts." 77 Sample districts were Sargodah, Chakwal and Bahawalpur in Punjab; Khairpur, Mirpurkhas and Badin in Sindh and Lower Dir in NWFP. Two tehsils from each district were selected for case studies: Sargodah and Silanwali in Sargodah District; Chakwal and Talagang in Chakwal District; Bahawalpur and Ahmedpur East inBahawalpur District; Khairpur and Kotdigi inKhairpur District; Mirpurkhas and Digri in Mirpurkhas District; Badin and Matli in Badin District; and Temargarah and Samarbagh in Lower Dir District. Field work was conducted from February 9 to March29,2006 78 This variation is inpart by design, as Provincial Finance Commissions (PFCs) awards to districts in Punjab and Sindh take into account district rankings of indicators of health, education, housing, employment and residential services, and those in NWFP take into account a backwardness index. Other factors taken into consideration in PFC awards include tax effort in Sindh, infrastructural deficiencies in NWFP and size o f the district interms o f geographical area inBaluchistan. 79 It should be noted, though, that special development grants are not consistently recorded across districts. ' O As inthe case o f special development grants, it is likely that CCB resources are not consistently recorded inthe government budgets. 114 An even wider variation in fiscal resources is found at the tehsillTMA level. Expenditureper capita inDigri(381 Rdperson) and Khairpur (361 Rdperson) are more than eight times as highas those in Timergarah and Samarbagh in NWFP (both 44 Rdperson). In part, this difference in resources reflects a greater ability by some tehsils to raise their own revenues through local taxes and fees such as urban immovable property tax, cattle mandi (market) and adda (transport station) fees and rents. Own revenues per capita are highest in the three Punjab district capitals sampled (131 Rdperson) where these account for 48 percent o f revenues; by contrast, own revenues are lowest in NWFP at only 28 Rdcapita in the district capital o f Lower Dir, or 63 percent o f (very low district) revenues. District awards to tehsils also vary, from only 11 Rdcapita in the two NWFP tehsils sampled to 210 Rdcapita in Mirpurkas and 240 Rdcapita in Khairpur. Table 5.5: Budgets in Selected Tehsils and TMAs, 2005-06 Punjab Punjab Sindh Sindh Total Total Capitals Remote Capitals Remote Capitals Remote Population(mns) 1.01 0.54 0.41 0.34 0.70 0.41 Budgetlcapita (Rs) 275 174 268 236 239 182 ExpenditureShares Salary 38% 35% 43% 36% 41% 33% Non-Salary 26% 18% 28% 19% 28% 20% Development 36% 47% 29% 45% 32% 48% Total 100% 100% 100% 100% 100% 100% RevenueShares Grants 33% 47% 62% 73% 44% 55% Own Revenue 48% 40% 30% 19% 43% 33% Cash Balance 19% 13% 8% 7% 13% 11% Total Revenues 100% 100% 100% 100% 100% 100% Source: TMA budgets; Niazi et a]. 2006. Notes: The total sample includesthree remote and three district headquartersTMAs in both Punjab and Sindh, and one remote and one district headquartersTMA inNWFP . TMAs are unable to tap additional sources of revenues because the tax base is small and taxpayers cannot pay moreaE1For example, the tehsils o f Sargodah, Silanwali, Bahawalpur, Khairpur, Kotdigi and Mirpurkhas face difficulty in even collecting the existing taxes. Raising revenue through new taxes or higher rates i s politically unpopular. Attempts to raise revenues through additional taxes were made in Talagang and Mirpurkhas tehsils but could not be sustained due to political pressures to withdraw them. In Mirpurkhas, the cattle mandi has been abolished as a result o f devolution, thereby eroding the revenue base. By contrast, the resource base inBadin has improved slightly due to inclusion o f a cattle Piri initsjurisdiction. In some tehsils, the devolution process has resulted in an expanded jurisdictions and responsibilities without a corresponding increase in resources, threatening service delivery to remote areas. In Silanwali and Talagang tehsils in Punjab, and Khairpur, Kotdigi and Mirpurkhas tehsils in Sindh, government officials reported that providing services to the remote areas now under their jurisdiction has increased costs, and that because o f the lack o f a 81 This is the generalview fromthe field andneeds to be exploredmore. 115 corresponding increase in resources they have been unable to provide adequate services to these areas. Since there i s now no formal administrative distinction between rural and urban areas and no mechanism to track spending, it i s difficult to earmark resources for these under-serviced areas. In general, establishment costs (expenditures on salaries and overheads) have increased faster than provincial government transfers to local governments through PFC awards. As a result, development expenditures have tended to rise more slowly than the overall increase in revenues. Nonetheless, the situation for lower tiers o f local government i s muchbetter than it was previously when they received virtually no transfers from the province except in the form o f discretionary, purpose-specific grants. Under the current system, district governments are reimbursed for their actual wage bills and the funding i s earmarked for salaries," while transfers for non-salary part contingencies are made on an incremental basisaX3.The district councils have little say in the prioritization o f establishment-related allocations, which are incremental and input-focused. Thus the DGs do not exercise control over a large part o f their budgets and allocationo fresources does not necessarily reflect local priorities. Although relatively more resources are available for development following devolution, one major reason still cited for inadequate service delivery is that development resources available to the local government remain limited. The development budget is allocated to the DGs and TMAs by the respective PGs through the PFC on a residual basis. Development funds constitute a small proportion, 10-12 percent o f DG budgets. Utilization rates are very low, and some districts have accumulated large cash balances because o f cany-over from previous years o f development funds, including funds for CCB projects. The releases have generally been smooth inPunjab andNWFPbut Sindhhas faced some problems, with the release o f funds occurringlate and this has contributed to even lower utilization rates there. Moreover in Sindh these funds are often utilized to finance the DGs' fiscal deficit. TMAs have more development funds at their disposal but local priorities are not always reflectedinthe allocation o f funds. The transfer of resources through the vertical programs and other higher-level government programs adversely affect LGs' autonomy in preparing development budgets and undermines accountability and operational efficiency. Substantial resources are allocated, especially in the water and sanitation sector, through the Members o f the National Assembly (MNAs), Members of Provincial Assemblies (MPAs) and directives o f the Chief Minister. A number o f donor programs, Punjab Devolved Social Service Program, NWFP-Rural Water Supply and Sanitation Project, Punjab-Community Water Supply & Sanitation Program, Sindh Devolved Social Service Program, Sindh-Presidents' Revamping/Rehabilitationo f Irrigation and Drainage Systems, Khusal Pakistan Programme, and the Drought Emergency Relief Assistance project also operate in some districts. All these allocations normally bypass LGs budgets and are spent on programs which have formally been assigned to LGs. It i s also not clear how these vertical programs have been treated in the TMA budgets; sometimes they appear with some allocation and in other years they are not mentionedat all. Inter-LG funding comparisons become very difficult because o f this varying treatment o f funds. Some LGs have welcomed these transfers because they fill the huge gap in development funding but these transfers nevertheless disturbthe localplanningprocess and create unwanted liabilities for LGs. 116 Problems of insufficient finances are particularly acute for TMAs that now formally have responsibility for maintaining water-supply schemes, many of which are old and have very high operation and maintenance costs. Officials in many o f the TMAs covered in the study districts stated that they lacked the financial ability to maintain these schemes. There are severe fundingconstraints in Sargodah, Silanwali, Bahawalpur Ahmedpur and Khairpur, and a number o f schemes have been abandoned. Funds are just sufficient to maintain schemes in Talagang, Kotdigi, Badin and Matli, but this leaves no resources for development. Moreover, most tehsils lack the financial resources to implement new schemes. As a result, previously planned schemes are either not undertaken or are built with funds from provincial governments, vertical programs, M P A allocations or directives o f the Chief Minister. To address this problem, the former Public Health Engineering Department has been revived to undertake water and sanitation projects in some provinces. Similar problems exist at the village levelwhere most water supply, drainage and sanitation schemes are either non-operational or were very poorly maintained because of ineffective Village Development Committees (VDCs). Most o f the water supply and sanitation schemes were constructed by the PHED and later handed over to these committees for maintenan~e.~~ VDCs are not able to operate and maintain water supply systems in Sargodah, Silanwali, Talagang, Bahawalpur, Ahmedpur, Khairpur, Kotdigi, and Digri because o f a lack o f funds. VDCs are also unable to recover user charges from the communities to operate and maintain the expensive facilities built by the PHED. All TMAs were concerned about this and feared that they would be obliged to maintain these schemes in future without additional financial support.85 ADMINISTRATIVE ISSUES Even five years after devolution, there still exists confusion regarding roles and responsibilities of the various levels of government, as well as apparent jurisdictional overlaps. Members o f local government staff are still not absolutely clear about their roles and responsibilities and the line o f reporting under the devolved system and often look to the District Government for direction. There was also some confusion regarding functions assigned to different levels o f government when multiple levels are involved in delivering the same service and the public i s unable to attach credit or blame for a particular service to a particular politician. The mismatch between expectations and jurisdictional responsibility creates problems o f accountability without responsibility. Many union nazims and naib (deputy) nazims report that voters are demanding services that they have neither responsibility nor resources to provide. Yet, despite their ex officio presence on tehsil and district councils, they are not credited with any o f the successes o f higher tiers o f government. Still, the delegation of power to the localgovernments has brought decision making closer to the communities and enabled them to access government officials, and there has been some improvement in service delivery. As reportedby all tehsils (except Bahawalpur) citizens' access to their representatives has increased considerably and problems are solved relatively quickly under the new system. The priorities o f local communities are also increasingly reflected in development schemes. There has also been substantial improvement inawareness and knowledge 84 InPunjab no funds are being allocated to the VDCs for operation andmaintenance while inSindh they were received funding for maintenance for four years. 85 Operation and maintenance are less important in the NWFP tehsils as most o f the water supply schemes there are spring water gravity supply schemes. There are therefore no user charges for them. 117 and flow o f information across all areas.86 In terms o f perceptions o f improvement in service delivery, this i s mixed, with a number o f tehsils showing improvement while in others the majority o frespondents claimed to not have experienced any improvement. Political priorities are reflected in decision making in most tehsils. Surveys reveal that councilors are most effective in getting their priorities reflected in development schemes but have no role to play in revenue raising and monitoring. NGO feedback on whether service delivery has improved and decision making has been placed in the hands of the communities following devolution i s mixed. Little improvement i s reported in water supply and sanitation in Sargodah, Khairpur Mirpurkhas Badin and Dir. Inthe other two district, Chakwal and Bahawalpur slight improvements in both decision making and improvement in service delivery i s reported. Access roads have improved most sharply in the opinion o f NGOs in all districts. This i s also reflected in district Annual Development Plans as most o f the development funds are allocated to this sector. The tehsils still look toward the District administration for guidance and are dependent on the District in a number of ways. Data indicate that the most common interaction occurs when projects are financed through the DG sources or when it provides technical help in the design o f projects as happened in Bahawalpur, Ahmedpur, Mirpurkhas, and Digri. This way the DG can influence where the money i s spent and affects the independence o f the tehsil administration. There are bothnegative and positive aspects o f this linkage. Some tehsils are better o f f as they are able to draw on the larger resources o f the DGand undertake costly water and sanitation schemes, which they would not have been able to from their own resources. In others such as Sargodah, Silanwali, Talagang, Digri, and Badin the role o f the district i s considered an encroachment on the TMA's authority and against the spirit o f devolution. CAPACITY Staff shortages and lack of appropriate skills are serious constraints to efficient service delivery. The capacity issues at the TMA level are more serious and almost all the TMAs surveyed lacked the capacity to perform their planning functions. Some staff serve in positions that require greater seniority and experience. The key post o f Tehsil Municipal Officer (TMO) was vacant in two tehsils while officers at a lower grade were serving as TMOs in three tehsils. Staff capacity was especially weak in the finance department^.^^ The planning capacity was consistently a deficiency o f TMA staff strength, and the position o f Tehsil Officer (TO) incharge o f planning was vacant in seven o f the fourteen tehsils in the study. The T O (Infrastructure) was usually a qualified engineer but all tehsils lackedthe capacity to deal with the technical aspects o f service delivery. There seems to be little attention paid to the post o f TO (community development) in most tehsils: often the officer was not physically present inthe tehsil or ajunior level staff was assigned these responsibilities. Weak financial capacity especially at the T M A level adversely affects their performance in many ways. First, it hampers the exploitation o f new revenue resources, which requires staff and with the skills to assess the new revenue base. Other key tasks such as preparing a case for 86 This was inmany ways inevitable withhundreds o funion councilors across the entire district but this is also a visible difference fromthe first three years o fdevolutionexperience when improvements were not so widely felt by the stakeholders, as reported inthe ADB, DFID, World Bank study (2004). 87 Inmany districts, EDO/TOFinance didnothave the requiredbackground infinance with no accounts and audits training. 118 approval by counciVPG, arranging for the publicity, conducting public hearings and dealing with public comments have not been adequately done. None o f the TMAs have ever made a case for their PFC allocations or transfers. There are also frequently inconsistencies in LG budgetary practices. Financial planning has been ineffective and there has been no planning beyond the yearly budgets. The latter also do not reflect policy priorities. Provincial governments still hold the Appointment, Promotion and Transfer power over local government and thus exercise a considerable influence over them. Many o f the senior staff at TMAs are appointed by the provincial government. The lack o f transparency in the APT process, in the form o f opaque recruitment and promotion criteria, means that the threat o f a transfer can exert pressure on senior staff to accede to the wishes o f the provincial government. Transfers o f staff are frequent and few officials are able to complete their full term o f office. ACCOUNTABILITY Beneficiary and stakeholder responses on the council process indicate that it is functioning well and that regular council sessions are held in Silanwali, Chakwal and Matli tehsils to debate the development budget, although attendance i s thin. The Council process i s reported to be effective in Sargodah, Silanwali, Chakwal, Mirpurkhas, Matli and Samarbagh. Overall, 82 percent o f the respondents believed that LG decisions are based on political priorities rather than other considerations. Political priorities are reflected in the distribution o f funds in the tehsils o f Sargodah, Silanwali, Chakwal, Digri, Badin and Samarbagh, which have also reported the process to be effective. Accounts and works committees have been formed in two tehsils each in Punjab and Sindh and one inNWFP but they have been functional and effective only inTalagang tehsil. The ineffectiveness of the LG councils and their elected committees in overseeing the budgets and ensuring that LG plans reflect peoples' preferences is one of the main reasons cited by the NGOs for service delivery deficits. Overall 70 percent o f NGO representatives considered the council process to be weak. The council process i s reported to be ineffective for a number o f reasons. First councilors are not powerful enough to influence the Nuzim who controls a large proportion o f funds and uses these at his own discretion. Second, councilors are also reluctant to raise their voices against political hand outs because o f fear o f isolation. Third, council resolutions are not always implemented to the letter. Fourth, councilors lack a strong interest in the process, as the allocations to them, out o f the development funds, are small. Finally, access to information i s poor and citizens are not aware o f their rights. Focus group discussions with citizens, beneficiaries and councilors show that the CCBs process is gaining ground and a number of projects have been successfully undertaken (Box 5.2). Overall 65 percent o f respondents were familiar with the process. However, CCBs have been only partially effective, especially inremoter areas, for a number o freasons. Citizens are not aware o f the procedures for CCB formation; in some cases they had no knowledge o f the process at all. The registration o f the CCBs i s to be undertaken by the Executive District Officer ED0 o f Community Development to whom communities do not have ready access. Registration guidelines are also not always clear to citizens. In remote areas there is the additional problem that people do not have the capacity to develop a PC 1 for projects due to lack o f education and were thus unable to utilize funds. In five o f the fourteen tehsils visited the funds allocated to the CCBs under the development program were not fully utilized. In Sargodah, Talagang, Bahawalpur and Matli tehsils CCBs were not effective because o f a lack o f awareness, while financial constraints prevented their effective functioning in Silanwali, Ahemedpur and Badin, 119 where poor communities were unable to make the 20% contribution. It was also observed that in Sargodah, Silanwali and Talagang community participation was low as the communities were not united in their efforts. CCBs are less effective at the TMA level because they are registered at the district level, and no one guides their formation at TMAs. NGOs have generally not contributed to CCB formation while the community development department (CD) at the district lacks motivation, capacity and mobility. CCBs have not been a priority for local government leadership and sometimes there is general political reluctance to release government funds to private entities. In some cases hostility has been observed toward CCBs on the part o f elected representatives. CCBs have also been the victims in power struggles between district and local governments. However, there are some examples (Khairpur) that suggest innovative leadership would have a positive impact on community mobilization and on service delivery. Mushawarat Committees, with members from the district Nuzimeen and the DCO, were envisaged as forums for evolving strategic plans and coordinating the development o f entire districts. This i s the only mechanism to coordinate the district and the tehsil under the new system. The overall conclusion from field visits is that they do not meet regularly and are not effective. Only inKhairpur do Mushawarat Committees meet regularly. They are somewhat effective inKhairpur and Mirpurkhas tehsils. The Local Government Commission (LGC), which is a reconciliation mechanism for inter- government dispute resolution, is not effective. In almost all districts and tehsils, except the tehsils o f Bahawalpur inPunjab and Digri in Sindh, it was reportedto be totally ineffective. Even inthe tehsils o fBahawalpur andDigrithe LGC was effective only inthat a number o f caseswere referred but there were numerous delays in resolving matters due to a cumbersome lengthy process. All the officials interviewed at the district and TMA level showed no faith in the institution's ability to resolve their inter-government disputes. SUMMARY OF ISSUES Budget certainty and predictable transfers are the key to successful decentralization. In Pakistan local planning and budgeting have not been a strong substitute for central direction although local administrations have been given the powers to manage their budgets and raise their own revenues. Lack o f adequate resources remains a serious constraint to efficient service delivery. A too-large share o f resources are taken up by establishment costs and little i s left for operations and maintenance; there are often no funds available for development o f infrastructure. Also, many TMAs have a weak tax base and some local taxes and user charges are difficult to impose inpractice. The process i s highly political and unpopular. The transfers o f funds through the PFC award are negotiated on the basis o f establishment size and provide a disincentive to reduce costs and improve efficiency. The allocation o f resources through vertical programs or through MNA, MPA and CM funds also creates a disincentive for local politicians to exert their authority to improve service delivery in their areas. The budget process under the new system allows the councils to allocate resources according to their priorities and toward service delivery objectives. Political relationships between the provinces and the local government also has a significant impact on the flow of funds. Jurisdictional overlaps and unclear responsibilities under the new system make it difficult for the public to assign credit or blame for delivery o f specific services. 120 This overlap weakens incentives to perform well on service delivery and may encourage politicians to target services aimed only or largely at their core supporters. Box 5.2: CitizenCommunityBoards(CCBs) Background Citizen Community Boards (CCBs) were created as part o f the devolution process begun in Pakistan in 2001 to address problems o f public service delivery, the lack o f need-based planning and the overall lack o f accountability o f government to people. These citizens' groups, comprising at least 25 members, register with government and receive public funds to undertake development projects or monitor public service providers. At least 25 percent o f the total development budget o f each tier o f local government (district, tehsil, and union) must be earmarked for CCB projects. Because local governments are required to devote such a large share o f their annual development budgets to CCB projects, they have an incentive to include CCBs in the decision making, planning, implementation, and monitoring o f development process. CCB development funds cannot be re-appropriated for other activities ifunused by the end o f the fiscal year and must be carried forward to subsequent financial years. Current Status As o f 2006, over 25,000 CCBs have been registered inPakistan. Thousands o f these have formed partnerships with government for community driven development and projects have been launched and completed. In these projects, CCBs are required to contribute at least 20 percent o f the total capital costs o f the project. Participation of the CCBs has also led to increased use o f appropriate technologies, introducedthe concept o f service standards, and strengthened operation and maintenance undertaken by project users through the CCBs. A recent study supported by the World Bank at the request o f the Government o f the Punjab suggests a needto simply the systems and procedures related to CCBs (especially registration and project execution). A technical or financial cap also i s needed on CCB projects to reflect communities' ability to implement projects. Weak field outreach o f government and lack o f adequate evaluation systems are other major drawbacks. Good CCB practices occur mainly in districts where there i s a strong political will and commitment. CCBs themselves could be strengthened, however, by malung them more broad-based; at present, they are driven mostly by a few community activists. LookingAhead The CCB idea i s innovative and ambitious and appears already to be taking o f f . However, true community participation as an integral part o f the work o f local governments will require a continuous review and revision o f rules, procedures and partnership arrangements between government and CCBs, as well as a robust monitoring and evaluation system. In many cases, it would also be useful to have an interlocutor between CCBs and the government to facilitate interactionbetween them. Finally, project standards and written agreements that take into account community capacities and clearly delineate responsibilities for projects between CCBs and government could improve transparency and accountability for individual projects. 121 Staff capacity, especially at the TMA level i s not strong and many local government officials lack the managerial skills to undertake their increased responsibilities under the devolved system. Since ATP powers remain with the provincial government after devolution there i s pressure on the senior staff to adhere to the demands o f the provinces. Many water supply and sanitation schemes are outdated and expensive to maintain. POLICY IMPLICATIONS To improve the resource position of the local governments and to ensure their financial autonomy, more and more predictable resources should be made available to them. Allocations through pre-determined and pre-announced capital budgets will ensure that LGs know in advance what financial resources are available to them and plan accordingly. Autonomy and flexibility can also be ensured by allowing LGs greater local discretion over the use o f funds and by ensuring a more meaningful planning process. Incentives for more efficient use of resources may be provided in the form of performance indicators into the grant system, e.g., development expenditure allocations for good performance; and by linking LG performance to a range o f criteria, such as quality o f planning, fiscal effort, compliance with procurement procedures, financial management, transparency, and operations and maintenance arrangements. A block grant funding approach will also encourage LGs and communities to focus efforts on mobilizing their own resources to complement this amount and shift the focus o f LG politicians away from lobbying for funds from higher level politicians. As a substantial amount o f funds i s channeled through the federal government, providing information on vertical programs and converting existing vertical programs into grants, or alternatively making LG managers responsible for them will give LGs more control over their resources and thus greater autonomy. Making more resources available to the LG and ensuring the allocation o f operational funds to VDCs will help maintain and manage public facilities. Strengthening LG capacity for finance and planningon a priority basis will ensure better budget preparationand implementation; planning can go beyond assessment and look at such issues as appraisal, cost-benefit analysis, and O&M arrangements. Demand-driven capacity- building can play an important role inoptional or location-specific needs. Giving APT powers to local governments and improving the transparency o f the APT process will ensure that there i s no interference from higher levels o f government interms o f appointments. Monitoring and evaluation capabilities should be enhanced to improve quality of development spending. Monitoring committees should have a formal role in developing sector budgets and receive training in sector specific plans and budgets. Requiring Executive District Officers (EDOs) to report on sector performance to Monitoring Committees, providing councils with resources to follow up reports and complaints, and giving councils the authority to take action against EDOs will all also improve accountability and increase citizens' power over service providers. Accountability can be increased through increased dissemination of information and greater community involvement. Publishing district and TMA plans, budgets, transfers to service providers, sector expenditures and performance reports, and encouraging the local media to discuss issues will improve access to information and make the local government system more transparent. Likewise, CCBs can be made more effective by publishingand disseminating widely user-friendly guidelines for CCB processes, developing training programs for councilors to improve their awareness and ensuring implementation o f the LGOprovisions for CCB funding. 122 Annex TableA5.1: RealExpendituresandRevenuesinPunjabProvince, 2001-02 to 2005-06 2001-02 2002-03 2003-04 2004-05 2005- Growth Change 06 Rate FYO2-06 FYO2-06 Total Expenditures 133.15 160.24 185.48 199.23 221.I3 13.1Yo 66% Current Expenditures 118.83 131.83 141.90 136.38 141.89 4.0% 19% Salaries Other Transfers to Local Government 60.74 69.55 67.53 63.81 66.69 1.O% 10% Total* - District Governments 53.12 58.29 58.31 54.79 59.00 1.5% 11% - Tehsil/TMAs -- 4.65 8.16 5.06 4.88 5.06 -3.4% 9% Union Administrations 2.98 3.10 2.93 2.71 2.63 -3.7% -12% Other 0.00 1.23 1.43 0.00 Development Expenditures 14.33 28.42 43.58 62.85 79.24 52.4% 453% AgriculturaVRural Total* - Agriculture 3.58 6.02 9.48 10.11 42.9% 183% and Agricultural 0.52 0.74 0.89 2.12 0.94 24.9% 81% Credit - Livestock Development --- 0.17 0.16 0.45 0.72 71.1% 326% Forestry/Wildlife/Fisheries 0.08 0.05 0.14 0.44 81.9% 423% Rural Dev. Programme 2.28 3.92 4.63 4.52 24.8% 98% -- Rural Water Supplies: 0.05 0.81 1.81 3.25 278.7% 6380% Urban Water Supplies: 0.23 0.01 0.63 0.73 128.9% 218% Irrigation 0.26 0.25 0.19 0.33 0.24 5.5% 0% Other 24.84 37.55 53.37 69.13 40.8% 178% Total Revenues 132.94 150.39 173.88 186.46 199.01 9.5% 32% Current Revenues 132.94 148.23 168.61 181.30 198.97 10.6% 50% Federal Tax Receipts 102.86 110.20 114.63 124.57 129.58 6.0% 26% Federal Grants 4.62 12.75 15.44 22.28 30.50 54.2% 560% Provincial Taxes 13.03 14.40 19.71 19.15 23.61 15.9% 81% Provincial Non-Tax Receipts 12.42 10.88 18.84 15.30 15.28 7.8% 23% Development Revenues 0.00 2.16 5.26 5.16 0.04 -69.9% -98% Current Surplus/Deficit 14.11 16.41 26.71 44.93 57.08 46.3% 304% Total Surdus/Deficit -0.21 -9.85 -11.60 -12.76 -22.12 125% Source: Calculated using Punjab Budget Reports 2003-06; Punjab Economic Report from March 31,2005. 123 6. SOCIAL MOBILIZATIONAND ENHANCING LIVELIHOODS Numerous government and non-government programs and approaches have been used to promote rural development and poverty alleviation. Among these have been initiatives designed to build rural infrastructure, mobilize communities, enhance access to rural finance, and provide a social safety net. This chapter provides an overview o f the major programs inPakistan and underscores the importance o f social mobilization in creating effective demand for public services at the local level. Other direct interventions aimed at improving welfare o f the poor are also examined, including micro-credit programs, safety nets and programs targeted at disadvantaged groups, particularly heavily indebted laborers. MAJOR RURALDEVELOPMENT PROGRAMSINPAKISTAN Pakistan has a long history of major rural development programs involving various approaches, but most have had only limited success. Early programs were characterized by a top-down planning process with little input from the rural communities. Over time, these programs have come to involve increased participation by rural communities and NGOs in planning and implementation. The devolution process begun in 2001 institutionalized efforts toward greater involvement by local governments and communities in development programs and public service delivery (Box 6.1). Spending on agricultural and rural development, including both government current and development expenditures on agriculture and irrigation, and major non-government programs, totaled Rs 59 billion in 2004-05, or about 0.9 percent of Pakistan's GDP. Government expenditure on agriculture and irrigation, both current and development, accounted for 71 percent o f this total: government expenditures on irrigation alone accounted for 40 percent o f total spending (both government and non-government) on rural programs (Table 6.1 and Figure 6.1). Nonetheless, several ad hoc development programs are ongoing, although these function outside the main Annual Development Programme framework and established government department budgets at the provincial or local government levels. These programs, such as Khushal Pakistan- I,whilepotentiallymoreresponsiveto localneeds(to the extentthat localgroupshaveaccessto high level elected government officials), may result in projects that are inconsistent or incompatiblewith overall development strategies or priorities (Box 6.2). 124 Box 6.1: RuralDevelopmentProgramsinPakistan Numerous rural development programs involving various approaches, donors and government administrations have been implementedinPakistan, but generally with only limited success. The first major program, the Village Agricultural and Industrial Development Programme (V- AID), from 1953-1962, was designed to increase agricultural production, establish rural small- scale industries and provide social and physical infrastructure through village councils, youth clubs, cooperatives and social centers. Subsequently, the Rural Works Programme was launched inWest Pakistan in 1963, after initial success inEast Pakistan (current-day Bangladesh) the previous year. Although this program made an impressive contribution to the development o f local infrastructure, recourse to local resources for repairs and maintenance reduced the long-term impact o f the investments. In 1972, the initiative was renamed the Peoples Works Programme and placed under the Federal Ministry o f Finance and Planning, where it remained until 1983. Projects under the Peoples Works Programme, which was concurrent with the Integrated Rural Development Programme (1972-80), included road construction, school buildings, small irrigation dams, drinking water facilities and other physical infrastructure, although in many cases, other necessary inputs and services were not provided.* Numerous broadly similar short-lived initiatives were launched in the mid-1980s. The Five- Point Special Development Programme (1985-88) was later integrated into other existing development programs: the Peoples Programme (1989-90) and subsequently the Tameer-e- Watan ProgramlPeople'sProgramme (1991-99). These programs generally included a broad range o f rural development activities, among these education, health, roads, electrification, water supply and sanitation. The Social Action Programmes (SAP) I(1992-93 to 1995-96) and I1 (1998-99 to 2003-04) were designed to expand access to basic services for the poor, particularly education and health for women and girls, and improve service quality; these were intended to include involvement o f NGOs, the private sector and community participation to build political consensus and ensure bureaucratic support. Delays in funding and other implementation problems hampered SAP-I, particularly federal programs for population welfare and health.' SAP-I1had a greater focus on technical and vocational education and on improvements in the quality o f service delivery. Implementation o f SAP was decentralized to the provincial and district levels for the majority o f relevant expenditures, consistent with the government's devolution policies formally adopted in 2001. The KhushalPakistanProgram(Pakistan Welfare Program, January 2000 to June 2002) provided Rs 36 billion duringthis period for small-scale public works schemes (Rs. 0.05 million to Rs. 5 million per scheme inrural areas and Rs.0.05 million to Rs. 8 million per scheme in urban areas). The Tameer-e-Pakistanprogramme(TPP), a federal programbegun inFebruary 2003 which was later renamed the KhushalPakistanProgramme-I(KPP-l), initially provided Rs 5 million in2003-04 to eachMember o fthe National Assembly (MNA) for local development schemes. The amount per MNA has increased over time, and the total budget equaled Rs 4.42 billion in2005-2006. * An assessment byNadeem(2003) points out aproblemencounteredinmany top-down development efforts: "the programme was riddled with irregularities inthe choice o f projects, determination o f priorities and locations by politically influential people with little regard for the needs o f the community, overwhelming reliance on contractors rather than o n project committees, and preference for large projects as well as widespread corruptionand misuse o fpublic funds." 125 Table 6.1:Major Government and Non-GovernmentRuralProgramsinPakistan,2004-05 Amount Share (Rs million) (percent) Federal Government 10,307 18.8 Agriculture 1,843 3.4 Current 1,233 2.3 Development 610 1.I Irrigation 8,464 15.5 Current 101 0.2 Development 8,363 15.3 Federal Special Programmes 7,840 14.3 Khushal Pakistan Programme I 5,340 9.8 Khushal Pakistan Programme II 2,500 4.6 Provinces 31,521 57.6 Agriculture 16,414 30.0 Current 11,137 20.3 Development 5,277 9.6 Irrigation 15,107 27.6 Current 7,780 14.2 Development 7,327 13.4 Rural Support Programmes 1,000 1.8 Cost of Public Infrastructure Schemesa 1,000 1.8 PPAF 4,063 7.4 Micro-credit 3,106 5.7 Community Physical Infrastructure 575 1.I Capacity Building Assistance 382 0.7 Total Rural Expenditures 54,731 100.0 Current (Government) 20,251 37.0 Development (including RSPs, PPAF) 23,534 43.0 PPAF Micro-Credit 3,106 5.7 Government Special Programmes 7,840 14.3 Source: Government of Pakistan, Ministry of Finance and Provincial Departments o f Finance data; RSP website; PakistanPovertyAlleviation Fund(PPAF) website . Notes: aEstimated based on cumulative figures and information from annual reports.Note that RSP figures do not includecreditdisbursed. In addition, several large NGO programs supported by international donors have undertaken major rural development efforts in Pakistan in parallel with government programs. The most notable o f these are the Aga Khan Rural Support Program (AKRSP), the National Rural Support Programme (NRSP) and various regional Rural Support Programmes (RSPs). These programs have achieved substantial success in organizing rural communities, building local community infrastructure and delivering basic health services. For the most part RSPs, like the special programs outside the ADP framework, operate largely inparallel with local government. 126 Box 6.2: KhushalPakistanProgramsIand II The federal government finances two parallel programs for the development o f municipal services in the country, the Khushal Palustan Programs, KPP-I and KPP-11. The Khushal Palustan Program was formerly called the Tameer-e-Pakistan program under the Nawaz Sharif government. KPP-I i s managed by the federal department o f Local Government and Rural Development (LGRD). Under KPP-I, each o f the 342 members o f the National Assembly (the lower house) and each o f the 100 members o f the Senate (the upper house) receive Rs. 5 million per year to carry out minor development projects intheir constituencies. Each MNA and Senator identifies two to three development schemes in his or her constituency each year and submits these schemes to the federal LGRD department. Each MNA and Senator must also identify the agency through which he or she wishes to execute a project, whether at the federal, provincial, district or tehsil level. The relevant district or tehsil governments do not have any control over the identification and execution o fthese projects. KPP-I covers multiple sectors like health, education, sanitation, roads, electrification o f villages, gas, and telephones, although most projects involve improvements to small roads and drains. KPP-11, also called the Roshan Palustan program, i s managed by the Prime Minister- Secretariat in Islamabad and deals primarily with the development work carried out under the Prime Minister's directive. Most o f these directives are issued as political instruments during public meetings held by the Prime Minister or the President duringtheir visits to various parts o f the country. Source: Stone et al. 2006., Assessment Report: Pakistan Devolution Support Project, prepared for USAID. Washington, D.C.:The UrbanInstitute(unpublishedreport). SOCIAL MOBILIZATION RURALLIVELIHOODS AND DEVELOPMENT A major reason for the limited success of rural development efforts in Pakistan and many other countries is a lack of effective participationand influenceby ruralpoor householdsin development programs and rural society in general. Too often, a top-down approach is implemented-one that treats the rural poor simply as beneficiaries o f public programs supplied by the government. Instead, the development paradigm should be one that putsthe household and its community at the center o f development initiatives. Empowering the rural poor to take on this role, however, requires collective action and social mobilization, central features o f Pakistan's Rural Support Programmes (RSPs), (Box 6.3). Not only do strong institutions help to create effective demand for public services and hold governments more accountable to their constituencies, these institutions can increase market power and enhance business opportunities for the poor. 127 llrc (3.1 ;Major Gowr nt RuralProgramsin P 4.1, 7% 12.6, 23% Box 6.3: SocialMobilization and the RSPs Social mobilization, which can be defined as the enhancement o f a community's capacity for undertaking collective action for its own betterment, i s a central pillar o f the Government's Mid-Term Development Framework (2005-10). The MDTF explicitly recognizes the need to reach each poor household through a concentrated process o f social mobilization: "Pro-poor policies cannot be effective without creating strong outreach mechanisms in each village through which people can interact with government." Social mobilization i s the primary means by which a community's capacity for collective action i s harnessed and strengthened-creating the basis for sustainable welfare impacts as well as broader social and political empowerment. Effective mobilization has been able to achieve this by reducing information problems (eliciting development priorities directly from communities and allowing communities to identify projects and eligible recipients o f private benefits), expanding the resources available to the poor through credit, capacity building, and occupational training, and strengthening the civic capacities o f communities by nurturingthe organizations that represent them. Mobilization o f economically, socially or politically marginalized groups has been an integral part o f Pakistan's Rural Support Programs (RSPs), non-governmental organizations registered as "not for profit" joint stock companies with the securities and exchange commission o f Pakistan. All RSPs have independent and autonomous boards o f directors that work in a voluntary capacity and a team o f highly qualifiedprofessionals. All RSPs share a common objective to reduce poverty and improve the quality o f life o f the rural poor. However, given the diverse socio-cultural milieu in Pakistan, their strategies are context specific and this i s one reason why they replicate rather than expand. Since creation o fthe AKRSP in 1982, the number o f RSPs inPakistan has now reached 10. The RSPs see their main strategic focus as the "provision o f social guidance for community mobilization and organization leading to community empowerment". A second strategic goal i s to increase communities' income-earning potential through micro-credit, grant funding o f community physical infrastructure, capacity building and training for employment. A third strategic goal i s to enhance community access to available public services. All interventions are identified inpartnership with organized communities. Support extended by RSPs to communities i s typically in the form o f technical and financial support for buildingphysical village infrastructure, provision o f micro-credit, support in the natural-resource sectors and small-enterprise development and support to poor communities in the social sectors. RSPs also act as facilitators among government, other development organizations and organized communities to improve service delivery to the grassroots. An evaluation o f the RSP program by the World Bank in conjunction with the Pakistan Poverty Alleviation Fund i s currently underway. The evaluation covers four main areas: (a) the returns to social mobilization-focusing inparticular on the quality ofthe institutionscreatedby the programas demonstrated by their internal structure and dynamics, their ability to engage in collective action, and their ability to link with local government and ensure more effective and equitable delivery o f public services and program benefits; (b) performance o f community infrastructure schemes carried out by the RSPs relative to similar schemes delivered by the more traditional `top-down' delivery mechanisms o f the government; (c) mechanisms that are likely to enhance the sustainability o f the program and its impact on poverty; (d) how the quality o f facilitation carried out by field staff can be optimized ina context o frapid expansion and scale up. 129 Box 6.4: Aga KhanRural Support Programme(AKRSP) The Aga Khan Rural Support Programme (AKRSP) was launched in 1982 for the purpose of implementing a rural development program in the Northern Areas of Pakistan. The essential elements of the AKRSP are the establishment of effective Village Organizations (VOs). Each V O starts its operation with one grant-assisted productive physical infrastructure project. Women's VOs focus on various aspects o f raising women's productivity and welfare. Examples o f the activities in which these organizations are engaged are: (i) introduction and dissemination o f improved varieties and cultivation techniques for crops; (ii)improved livestock husbandry practices; and (iii)new productivity-raisingtechnology for apricot-drying tents, nut-cracking machines, and butter-chumers. AKRSP assists village organizations to obtain access to institutional credit by acting as a loan guarantor through the promotion o f savings by members, which serve both as a form o f loan collateral for credit and to provide supportive marketing services for differentproducts. The savings and credit scheme of the AKRSP is based on a system o f collective savings and borrowings. Each member deposits his or her savings with the manager o f the village organization. The VO's collective savings are deposited in a single account held at a commercial bank. Credit facilities are extended in a similar manner. The VO will then take out a single collective loan and disburse sub-loans to members according to individual's credit need. This system minimizes the transaction costs for borrowers in both savings and borrowings. Due to peer-group pressure, the transparent nature o f the credit operations, and the group guarantees for loans, loan repayment rates are almost 100percent. Based on substantial positive results from the social mobilization, group-based savings and credit programs and other activities o f the AKRSP, a National Rural Support Programme (NRSP) was begun in 1991 and expanded in 1992 with a grant from the Government of Pakistan. NRSP is buildinga network of grassroots organizations that will plan and undertake their own development. It uses a development strategy that mobilizes communities and maximizes the use of local resources, similar to that o f AKRSP and the Orangi Pilot Project (OPP). Evaluations o f the AKRSP and OPP show that the projects are financially sustainable and that they have yielded significant benefits for the poor. Source: World Bank 2004b; Malik, Sohail J. 2003. Given the AKRSP's success, other Rural Support Programs now operate across Pakistan, with a total cumulative rural investment (including the AKRSP) o f Rs 15.0 billion. Much o f this total (Rs 11.8 bn) derives from earlier investment by the AKRSP. Estimatedtotal public and public- private partnership investments by the network in 2004-05 were about Rs 1.0 billion, (four percent o f total public investments for rural development by federal and provincial governments, RSPs and PPAFg8). The largest of these, the National Rural Support Programme (NRSP), begun in 1992, was set up with federal government fundingto work in all provinces and Azad Jammu and Kashmir. The NRSP i s also the largest RSP interms o f funding, with cumulative projects o f Rs 1.3 billion. NRSP activities are similar to those o f the other RSPs and include programmes in human-resource development, rural credit and enterprise development, natural-resource management, physical infrastructure, and social-sector services. All the RSPs are linked through the Rural Support Programme Network (RSPN), fundedbyDfID,which provides a forum for the RSPs to share lessons o f experience and to undertake strategic planning. 88 The total includes Rs 957 million invested by the Pakistan Poverty Alleviation Fund in its physical infrastructuredevelopment program in2004-05. 130 Another example of successful social mobilization and livelihoods development is the Andhra Pradesh Rural Poverty Reduction Project in India. Operating in a state with a rapidly growing economy and generally good infrastructure, the project focuses on creating women's self-help groups (SHGs). Group formation relies entirely on members' self-initiative and on their willingness to mobilize savings and inter-loan mechanisms conducive to establishing strong trust and discipline among the group members. While focusing on capacity building and social mobilization, the project also provides a financing facility for accessing productive assets, income generating activities and for starting micro-businesses. The investment facility i s provided as a grant at the village level, which represents the first-tier federation o f the SHGs, and i s recycled by the community itself into a revolving fund extending loans to its own members according to agreed procedures and terms. Loans are revolved so that over a period o f about five years all member households get access to a loan. The SHGs and VOs also use their savings to leverage credit from commercial banks. As the SHGs movement continues to expand, the federative movement has been able to add two more tiers at the "block" and district level. The scale reached and the level of organization i s such that the poor have over time become an extremely credit-worthy and attractive market for banks, micro-finance institutions, agro-business, traders and insurance companies. As a result, total credit and investments from the financial institutions and private sector have increased substantially. Households have also invested in education, health care, housing and further multiplied their assets as a result o f additional income. Finally, because o f their `voice' and `scale' they have become more effective in influencing the responsiveness and quality o f service delivery by the public sector (Box 6.5). ProjectDesignand Sustainability Components of social mobilizationhave been incorporatedinto various rural-development projectsand programs,often in conjunctionwith micro-creditand public investmentunder the broad rubric of community-driven development (CDD). Inprinciple, governments may be best suited for delivery o f public goods and services (and the investments required to support these) and the private sector and NGOs best suited for delivery o f private goods and services (including micro-credit, and skills training). However, the best institutional pathway to promote the active participation o f communities and their linkages with both the public and private sectors i s debatable, and depends on the particular characteristics o f communities, government institutions and development constraints. In areas where local government institutions remain unable to effectively deliver public services, direct interventions through community organizations and NGOs are needed. These programs work best where strong community groups exist; rural micro-level infrastructure and public health are especially important in isolated areas. To ensure long-term sustainability, however, it i s important that these efforts do not undermine devolution initiatives intended to strengthen local-government institutions. 131 Box 6.5: SocialMobilization in Andhra Pradesh, India Social mobilization through the creation o f self-help groups (SHGs) has been a centerpiece o f efforts by the government and NGOs in Andhra Pradesh, India. Following a successful pilot project covering 66,000 households begun in 1996, the Government o f Andhra Pradesh incorporated' the social-mobilization approach into its major poverty-reduction project, Velugu, which was subsequently renamed Indira Kranti Patham (IKP). The first phase, the Andhra Pradesh District Poverty Initiatives Project (AF'DPIP), was implementedbetween 2000 and 2006, covered 930,000 rural poor families in the state's six poorest districts. The second phase, the Andhra Pradesh Rural Poverty Reduction Project (APRPRP), is designed to cover two million rural poor families at a cost o f 15 billion rupees (nearly $300 million) over five years (2003- 2008). The projects were built on the existing women's self-help thrift movement that had begun 15 years earlier, and added capital, federation o f institutions and technical support. Under the projects, groups o f 12-15 poor and "poorest o f the poor" women form an SHG; 150-200 SHGs make up a Village Organization (VO), 4,000-6,000 VOs federate to form a Mahila Mandal Samakhya (MMS) (Women's Group Confederation). This process is initiated by community resource persons ( 0 s ) or community coordinators, which are selected, remunerated, and evaluated by the community and based in the village for a period o f at least three months. Typically, a CRP team consists o f two women members, one Senior Bookkeeper and one Activist. This team meets with village officials, provides assistance in establishing group management and financial norms to existing SHGs, and helps form new SHGs. The larger VOs and M M S ' are designed to have the critical mass to address social, economic and political issues that the smaller affinity groups cannot through the power o f a combined voice and by pooling their supplies and demands o f marketed goods and services. Inaddition to group mobilization, the programs focus on expanding the assets of the poor and creating economic opportunities connected to household livelihoods. To reduce, mitigate and manage risk IKP also supports the Community Investment Fund, which promotes investments in sub-projects for the poor, as well as the Comprehensive Insurance Package, which seeks to develop community-based delivery o f life and health insurance services. Source: ReducingPovertyon a GlobalScale (World Bank 2005~)."Case study on AndhraPradesh"; Project documents. Financing rural development for poor households and local groups at the local level: credit versus grants. International best practice suggests that, generally, loans should be the principal financing instrument for `private goods.' For the poorest households, for which there i s almost no institutional credit available however, grants or matching grants may be a more appropriate tool to facilitate access to productive assets. Yet providing matching grants through NGOs or government agents involves inherent risks related to supply-driven bias. Thus, the establishment o f revolving funds at the community level for administering internal lending mechanisms has proven to be more reliable and effective. Ideally, it is preferable to avoid mixing credit and grant financing in the same vehicle. Rather, separate programs should be developed. A program for livelihoods development of poor communities should include social mobilization, matching grants (or revolving funds) for asset creation and the development o f market linkages. A separate effort to develop micro-finance 132 services should be run as a business with appropriate linkages to financial-sector requirements and private sector-oriented client-selection methods. INCORPORATING SOCIAL MOBILIZATION INDEVELOPMENT PROJECTS A social mobilization approach involving greater community involvement can enhance the effectiveness o f development efforts across a wide range o f rural sub-sectors, including forest management, watershed development, irrigation or agriculture. By buildinginstitutions o f natural resource users, communities are better placed to pursue collective action and exercise management authority over the sustainable use o f the resources. . In forestry, international experience suggests that a purely conservationist approach to forest management is infeasible given that this ignoresthe livelihoods needs of poor forest- dependentpeople. Rather, devolution ofresponsibilities to local communities for the sustainable use o f natural resources must be a priority, as this will promote gains inproductivity and increase communities' income, without compromising forest sustainability. As with all natural `common pool' resources (CPR), multiple benefit streams flow to multiple stakeholders, and multiple service deliverers and decision makers are engaged in overseeing access to and management o f these benefits. Projectsinwatershed development can also serve as a point of entry for the mobilizationof poor communitiesand establishment of an institutionalbasefor planningand implementing multi-sector activities to improvelivelihoods. Watershed development projects typically have the multiple objectives o f improving agricultural activity levels, sustaining and improving management o f the natural-resource base, and increasing income levels among beneficiaries within the watershed areas. In addition to introduction o f watershed protection and treatment technologies, projects also support community-based institutional development for collective decision-making and management. This i s intended to promote participation o f the local population in the sustainable management o f watershed areas (common productive assets) and in activities that promote savings and income (individual productive assets). As with forests, management o f watershed areas involves critical externalities; gains to one individual may entail costs to another (upstreaddownstream trade-offs). For this reason, watershed management should not rest entirely in the hands o f one set o f stakeholders. Instead, incentives and institutional frameworks that promote cooperation and participation o f watershed community groups, line agencies and local governments must be supported. Community participation through Water Users Associations (WUA) can improve maintenance and efficiency of surface irrigation. Confronted with the failure o f traditional top-down irrigation bureaucracies in delivering services and ensuring proper function o f the systems, W A Scan provide farmers with greater control over operation and maintenance o f the watercourse system in their locality. W A S can also serve as a mechanism for confronting governance and transparency issues in the delivery o f water, thereby creating an environment o f greater accountability in terms o f public service delivery. WUA formation and capacity building are typically addressed through irrigation projects rather than livelihoods projects although the same principles apply. In addition, studying organized water users as the core element o f irrigation projects, i s instructive in articulating critical linkages with agriculture services and markets, which have typically been a weakness o f irrigation projects. 133 Farmers' organizations @`Os) can also serve as a mechanism for achieving the required scale and scope neededto increasefarmers' influenceof and access to markets, traders and public services. These associations are typically organized along commodity lines and are typically driven by market opportunities rather than management o f a public asset. As these associations gain confidence and capacity, they are likely to evolve into more sophisticated organizations such as cooperatives and be in a much better position to intermediate with markets forces and provide services to their members. The demise o f the cooperative movement in many countries around the world i s mostly the result o f excessive government influence in the decision malung process o f these institutions as well as the lack o f a clearly established mutual approach driven by genuine ground up principles based on member needs and social cohesion. Where successful, for instance in Western Europe, the cooperative structure has evolved outside direct government intervention and was also able to cater to farmers' financial needs through a savings and loan approach. MICROFINANCE INTERVENTIONS In Pakistan, microfinance has been slow to scale up, and outreach to women has been limited. It is estimated that only about 8 percent o f poor households receive credit from formal sources. Microfinance lending i s concentrated. The top five microfinance programs, one bank and four NGOs, had between them 84 percent o f the total market. Inthe past, much microfinance has been delivered through integrated development projects with credit components. There have been historically interest rate caps and subsidies but this has changed in recent years. However, most microfinance institutions still receive below market cost funds in various forms. The PPAF currently funds 29 NGOs providing microfinance but half have fewer than a thousand clients. Savings services are encouraged by RSPs but they are not permitted to take deposits directly so these are held and investedby community organizations and deposited with banks. Followingrecent legislation, Pakistan's microfinance policy environment now appears to be quite conducive to efficient delivery o f microfinance to poor households. The challenge now i s for the various actors to seek out efficiency and scale. Moreover, there continues to be significant subsidies in various forms in Pakistan microfinance, which warrants examination for efficiency. The subsidies have largely been related to low interest rates on loans, but microfinance projects generally have failed to leadto longer term institutional development. Experience with microfinance in Indonesia, Egypt, India, Vietnam and Pakistan (the Kashf Foundation project) suggests the existence o f a few enabling public actions that can enhance programeffectiveness (Table 6.2). The Pursuit of Efficiency and Public Policy Options. The overarching aim o f most microfinance institutions and country systems i s the pursuit o f efficiency in the delivery o f rural financial services (above and beyond credit) to poor households in a sustainable .manner. The quite broad country experience and the wider literature suggest the following priority public policy areas that may be o f relevance to Pakistan: (i) supporting institutional innovation through policy research aimed at better understanding the efficiency o f comparative microfinance experiences; (ii) supporting capacity buildingin institutions and communities; (iii) supporting the provision and sharing o f information; (iv) avoiding subsidized lending whenever possible but analyzing the comparative poverty impact o f current subsidies; (v) supporting savings; (vi) avoiding add-on credit components; (vii) meeting the product needs o f the poor; and (viii) improving communications (Nelson 2006). 134 SupportingInstitutional Innovation. Although there are no global best practice cases, arguably the most important area o f public policy support for microfinance in Pakistan i s in the area o f knowledge. While individual microfinance experiences globally are well documented, there are still no country studies that compare the cost effectiveness and impact o f alternative approaches to service delivery o f rural finance. Available evidence from multi-country studies suggest some several contributing factors for the profitability o f microfinance organizations. Government could strengthen microfinance development by supporting research on comparative efficiency and impact across the available range o f alternative approaches, perhaps through some competitive grant funding associated with the Microfinance Network, an international associating o f leading microfinance institutions. It i s critical to not overlook the impact and cost effectiveness o f the many informal savings and credit networks and relationships. These provide competition to formal microfinance institutions and offer more-tailoredproducts at quite competitive costs. Supporting Capacity Building. The use o f public funds could justifiably be spent to improve national microfinance skills at both institutional and community level. Support for Information. The cost o f information about individuals or groups in scattered rural communities i s high.This i s one o f the main reasons for the highcost o f rural lending. This can be reduced by a widely adopted group approach that passes on a share o f these costs and responsibilities to the group itself. While the government's role keeping records on groups has privacy risks, in India, at least one project supporting groups now provides lending institutions with ratings on group performance, which greatly aids lending decisions. This i s a local variation o f the publicly available credit ratings often available to microfinance institutions. Risks of Subsidized Lending and the Analysis of Subsidy Impact. The existing case studies and global experience do not support an argument for broad subsidized lending, although they also do not provide strong enough evidence against utilization o f a judicious initial subsidy to enable microfinance institutions to reach a sufficient scale for profitability. The history o f global rural finance i s dotted with failures due to large-scale subsidized lending, among these examples i s the pre-reform period o f Indonesia's now profitable Bank Rakyat Indonesia (BRI). However, the Vietnam Bank for Agriculture and Rural Development (VBARD) and its new offshoot the Vietnam Bank for Social Policies (VBSP), have, in fact, reached large numbers o f poor people through directed subsidized lending. Still, this i s likely a unique case inan unusual country with a strong pre-existing network o f communes. Moreover, it i s too early to predict what the outcome o f the VBSP initiative will be. Nevertheless, the global experience suggests that with any new microfinance operation, there will likely be a need for initial subsidy in the early years. The evidence shows few if any such cases that have truly reached the poor on a large scale and been profitable from the outset. In Pakistan there is still a substantial array o f below-market rate funds going towards microfinance through both direct and indirect routes. The cost and impact warrant close study. Much appears to have gone to keeping interest rates down instead o f building sustainable institutions. However, there has been little if any analysis o f just what the cost and impact has been. It i s important to know for the future what course o f action would ensure the best impact and what pattern o f support would offer the fewest distortions o f the market. Encouraging Savings. There may be some role for governments in encouraging savings services, including post office savings. Literature and some country case studies suggest that savings may be more important for the very poor for consumption smoothing and small 135 investment than credit, as well as offering lower risk8'.Savings services have often been limited partly due to regulations that control operation o f institutions offering savings accounts. Managing very small individual savings accounts i s costly, as the new Pakistan First Microfinance Bank i s discovering. But demand may grow rapidly. In Indonesia, BRI's savers (averaging about $100) are about 10 times the number o f active borrowers (averaging about $500) (Table 6.2). Avoiding the Error of the Add-on Credit Component. The evidence shows that microfinance needs a financially sound and sustainable institutional home to enable effective management. It should not simply be added as a component to a CDD project. There are many examples o f failed add-on, often subsidized, credit components. For example, the World Bank's Independent Evaluation Group (IEG) concluded recently that in the Indonesia Kecamatan Development Program (KDP) committed an error offering credit in addition to its grant program outside the framework o f a sound microfinance institution. The poor repayment performance o f that component could have been anticipated. For increasedpoverty impact, the same funds could have been used for a larger volume o f targeted infrastructure. Add-on credit components have been a feature o f some Pakistan CDDs. There are some partial successes such as AKRSP's credit component but this didnot reach profitability despite the institution having substantial slulls. Meetingthe Product Needs of the Poor. Country experience suggests that both lending- and saving-product flexibility and diversity are critical for the poor. While this i s predominantly a private-sector role, governments may be able to help by creating an enabling environment for innovation and supporting research, again perhaps through some research-grant arrangements. Market research i s important. The poor often need to transact loans over very short time frames. In Vietnam, for example, one NGO has a daily, weekly, and monthly loan product. VBSP is unable to compete with this. Institutions that have done well, such as Grameen Bank in Bangladesh, SEWA and others, offer products geared closely to demand. They are agile institutions. Diversity o f institutions may help; however there i s some trade-off between the number o f institutions and efficiency as, with a given demand, volume, which aids efficiency, is reduced by a multitude o f small institutions. Ultimately, inPakistan, the efficient can be expected to thrive as the less efficient will be forced to close. ImprovingCommunications.New lending technologies linked to products are being vigorously pursued in case countries. There is increasing use o f ATM-card technology (one example is India's Kisan (farmer) bank cards, which enable flexible borrowing). Computerized systems are lowering costs and enabling greater focus on efficiency. But Pakistan lags the other case countries in terms o f communications and technology. This would be a key area for public investment in support o f microfinance. Policies and Practices of MicrofinanceInstitutions. Global experience suggests that, other things beingequal, poverty and sustainable scale probably have a better chance o f success if an institution: (i)lends to women initially; (ii)starts off in more densely populated urban areas to keep initial costs down; (iii) charges full costs; (iv) builds to scale rapidly; (v) uses groups and supports internal group lending; (vi) listens and develops highlytailored saving and loan products for the poor; (vii) benefits (in some cases) from some cross-subsidy flowing from higher profit clients to lower profit clients; (viii) gives priority to capacity building for both itself and clients; 89 However, this calls for some qualification. It has been suggested that demand for savings in Pakistan may be somewhat less than it i s in some other countries due to a preference inPakistan rural areas to keep livestock or fixed assets. But, even iftrue, this i s likely to change. 136 (ix) i s generally fairly unyielding about repayment, but (in some cases) attempts to offer insurance cover; (x) has tight management with an administration pyramid very narrow at the waist with strong focus on keeping down localhranch level costs including use o f simple procedures and records; (xi) adopts new technologies when and where available; (xii) piggy- backs on already available client howledge; (xiii) initially spreads costs across other program activities beyond credit; (xiv) benefits from some start-up subsidy but spends that wisely to enhance future efficiency and reach; and (xv) i s flexible and agile. The institution also must operate within a reasonably conducive enabling environment, which it may or may not be able to influence. SOCUL PROTECTION Social protection programs to supplement or stabilize incomes of the poorest households remain very limited in Pakistan, particularly in rural areas. Social security payments, made almost exclusively to former government and formal sector workers in urban areas, account for about two-thirds of the estimated Rs 52 billion in expenditures on social protection (Table 6.3). The wheat subsidy, which costs Rs 8 billion or about 15 percent of total expenditures on social protection, i s untargeted and i s likely to provide little benefit to poor rural households, except to the extent that net wheat sales from public sector imports and drawdown o f government stocks reduce market prices. Safety net programs targeted to the poor account for less than 18 percent of total social protection expenditures. The largest targeted program, Zakat, is financed through a voluntary levy o f 2.5 percent o f the value o f financial assets deducted once a year and covers 1.7 million beneficiaries at a cost o f Rs 5.9 billion. The Pakistan Bait-ul-Mal i s a Federal government agency delivering vertical social welfare programs financed from general revenues, including the Food Support Program, which in 2003/04 delivered Rs 2.5 billion cash transfers to 1.25 million beneficiaries. Including the Tawana Palustan program that provides midday meals to 530, 000 girls in government rural primary schools at a cost o f Rs 0.70 billion annually, these safety nets total less than Rs 10billion, or 0.17 percent o f GDP. Zakat, the largest targeted safety net program, has only a limited impact on poverty and inequality. A major problem with Zakat is its targeting, which features significant leakage to the non-poor, especially in rural areas (Table 6.3). Moreover, benefits are small and delivered inii-equently. For example, Zakat 's guzara transfers and rehabilitation grants (its main components) represents 14 and 23 percent o f average recipient household income, respectively These figures are higher among ultra poor households (25 and 42 percent respectively) and lower among poor and non-poor households (around 6 to 8 percent for each program modality). By comparison conditional cash transfer programs in Mexico and Nicaragua provide approximately 21 percent o f average household expenditures. Given the relatively small size o f Zakat 's program benefits, it i s not surprising that poverty and inequality decline only slightly as a consequence o f these transfers. Moreover, irregularities inbenefit payments also limit Zakat 's impact on poverty. It is common for beneficiary households to receive fewer than the stipulated number o f transfers; survey results indicate that the total amount receivedby the average household from guzara was equal to only about half o f the full entitlement o f Rs. 6,000 per year. Bait-ul-Mal, by contrast, delivers better targeted (albeit even smaller) benefits. 137 E c V I \c t 9 Y Ultra-poor Poor Non-poor Yo of Yo of Yo of O h of Yo of % of Total beneficiary program beneficiary program beneficiary program households' funds households' funds households' funds Zakat guzara 39 33 33 35 27 32 100 Zakat Rehabilitation 44 23 19 32 37 45 100 Bait-ul-Mal 46 45 30 32 23 23 100 All sample households 38 35 27 100 Safety netsfor boththe rural andurbanpoor shouldbe mademore efficient and scaledup so as to cover a major share of the chronic poor and some of the transitory poor as well (largely through the expanded use of conditional cash transfers). The wheat subsidy program could be replaced by a targeted program based on either food or cash. Cash transfer programs such as the Food Support Program could be scaled up and the scope o f these programs expanded to encourage long- term human capital investments among the poor, for example based on the model o f the pilot Child Support Program, a conditional cash transfer (Box 6.6). Likewise the Tawana Pakistan school feeding program could be redesigned and scaled up. Nutritional interventions aimed at improving health and nutrition o f women and infants ought to be piloted and scaled up. The government may also consider piloting a major rural workfare program in order to provide temporary employment to vulnerable households. Box 6.6: The ChildSupport Program(CSP) -A ConditionalCash Transfer Pakistan will soon begin to implement a new Child Support Program (CSP). CSP i s a conditional cash transfer for education targeted to existing beneficiaries o f the Food Support Program (FSP). CSP will provide an additional benefit to FSP beneficiaries provided their children o f school- going age regularly attend school. Under CSP, a family with one child aged 5-12 years will receive an additional Rs.200 per month, while a family with two or more children aged 5-12 years will receive an additional Rs. 350 for as long as they keep all their children 5-12 enrolled in school and attending at least 85 percent o f the time. Families with no children o f school-going age or whose children do not attend school will continue to receive the regular FSP benefits and will therefore not lose out as a result o f this new program. CSP was launched as a pilot in three districts in late 2006. A rigorous scientific impact evaluation, based on comparisons o f outcomes with a control group o f FSP beneficiaries not receiving the CSP transfers, i s planned and scheduled to be completed by October 2007. Depending on the outcome o f this evaluation, the FSP pilot could later be scaled up to cover the entire country. 139 Table 6.4: Safety NetsinPakistan Name of program Type of Target group Source of Annual Number of benefit funds Expenditures Beneficiaries (Rs bn) Zakat (cash transfers and Cash The destitute Special levy 5.86 1,733,000" other) on bank balances Bait-ul-Mal (Food Support Cash (and The destitute Federal budget 4.5 1,250,000" and other) some schools) Tawana Pakistan Midday meals Girls in Federal budget 0.7 530,000" government rural primary schools Wheat subsidy All wheat Federal and 8 n.a. consumers provincial (untargeted) budgets Total,social assistance 19.06 n.a. Social security Employees Old Age Old-age Formal-sector Employers 1.74 850,000b Benefits (EOBI) pension (cash) workers Workers Welfare Fund Housing, Formal-sector Employers 2.6 n.a. (WWF) schools, workers earning health less than Rs 5,000 facilities per month Employees' Social Security Insurance Formal-sector Employers 2.08 464,000b Institutions (ESSI) against death, workers earning disability, less than Rs 5,000 maternity per month (cash); medical care inown hospitals (in- kindbenefits) Total, social security 6.42 n.a. Total, social assistance and social security 25.48 n.a. Total, as share o f GNI 0.47% n.a. 'ource:World Bank Pakistan Safety Nets report (forthcoming). Notes'a Number of beneficiaries ofrecurrent benefits; Number of workers coveredby insurancescheme. 140 Heavily-indebted Laborers Social protection programs are also needed to address the problems of heavily-indebted laborers subject to forced labor of various types. This practice is particularly persistent among landless sharecroppers in rural southern Punjab and northern Sindh. Measures to assist households trapped in unfair labor and credit arrangements would have a major impact on some o f the poorest rural households inPakistan. One option, implementedby the government of Nepal in 2000, was to issue a decree freeing such laborers from ancestral debts and imposing harsh penalties on employers. An alternative to manumission by decree would be some sort o f debt buyout implemented through local governments or NGOs. Whatever means i s used to eliminate or ease the debt burden o f these laborers, these efforts should be complemented by other assistance such as skills development, access to credit, and perhaps even land grants, in order to .enable laborers to develop alternative sources o f income not linked to erstwhile employers-cum-creditors. Given the political sensitivities involved and the need to test the modalities for reductiodelimination o f debts and income-generation programs, pilot projects may be neededbefore full scale r e f o m i s attempted. Box 6.7: Problemsof Heavily-IndebtedLaborersin Pakistan Laborers subject to unfair labor and credit arrangements work in various sectors and industries in Pakistan, e.g. agriculture, carpet weaving, brick-kilns, stone quarrying, hotels and restaurants and construction. Debt bondage i s the worst form o f these unfair practices, in which workers are bound to employers by debts owed to the latter as well as through formal or informal labor contracts. Because o f their lack o f assets, extremely low incomes and lack o f funds, poor laborers may become dependent on the landlord or employer for credit with high implicit interest rates that must be repaid in the form o f free labor to their creditor. This type o f labor bondage often continues over generations as when adults grow old or are disabled, their children become liable for the debt. Similarly, when husbands die, wives must assist in the repayment o f these debts. In its worst form, laborers are not allowed to work at other places without the permission o f the landlordemployer. In some cases, women and children must remain resident on the farm to vouchsafe good behavior. Religion, ethnicity and caste play major roles in determining the burden o f these labor practices. According to a report carried out by Australian consultants for the Government o f Sindh and the Asian Development Bank under the Sindh Rural Development Project, "There are some 1.7 million landless agricultural workers (haris) and sharecroppers in five districts o f Sindh Province (Thatta, Dadu, Badin, Mirpurkhas and Umerkot). Most o f these people are in debt bondage. While bonded labor exists throughout Sindh, the majority o f those bonded in northern Sindh belong to the Muslimmajority, while most o f the bonded agricultural laborers in southern Sindh belong to dalit (untouchable) and to tribal communities who have migrated from the drought- prone area o f Tharparkar desert. Poverty and starvation have forced these communities to accept cash advances from landlords, and to submit to work for them from dawn until dusk. Bonded laborers may be detained or guarded to stop them escaping and in these situations o f total ownership rape o f women i s not uncommon" (Agrodev 2000). Though the number o f workers subject to these extremes may be only a small share o f haris and sharecroppers, the plight o f these people led the Pakistan government to pass the Bonded Labor System (Abolition) Act o f 1992. Despite the serious penalties prescribed by the legislation, however, cases o f bonded labor persist because o f the political and financial strength o f some landlords insouthern Punjab and Sindh. Sources: Nazli 2003 and Agrodev 2000. 141 7. SUMMARY AND CONCLUSIONS After a decade of moderate growth but little or no long term change in rural poverty in Pakistan, agricultural output, rural incomes, rural poverty and social welfare indicators all showed marked improvements between 2001-02 and 2004-05. However, longer term trends suggest there i s little reason for complacency. Some o f the improvement in rural expenditures between 2001-02 and 2004-05 appear to reflect low levels o f income related to drought and low levels o f workers' remittances in 2001-02. The agricultural GDP per capita growth rate (1999- 2000 to 2004-05) was only 0.3 percent per year; rural poverty rates in 2004-05 are still at levels that approximate those o f the 1990s; and social welfare indicators inPakistan remain significantly below those o f other countries in south Asia. Moreover, problems related to timing and availability o f water for irrigation, inadequate rural infrastructure, a skewed distribution o f assets, and low levels o f health and education continue to slow the progress o f economic growth and poverty reduction. RURAL POVERTY: STRUCTUREAND TRENDS About two-thirds of Pakistan's population lives in rural areas (1998 census data), where average per capita expenditures are 31 percent lower than those in urban areas (Rs 1259/month and Rs 1818/month in 2004-05, respectively). Rural, as well as total population growth rates are declining, which bodes well for future per capita income growth. Nonetheless, the rural population i s likely to continue to grow, and may reach 122 million in2015 (64 percent o f the population) at historic rates o f migration o f 1.2 percent per year. Urban population would reach 70 million inthe same time frame according to this scenario, and 82 million (43 percent o f the population) ifthe net migration rate doubledto 2.4 percent per year. The majority of the rural poor in Pakistan are not farmers (neither tenants nor owners). Farmers comprised only 35 percent o f the households inthe bottom 40 percent o frural per capita expenditure distribution. Non-farm households (not including agricultural laborer households) accounted for slightly over half (57 percent) o f the poor in 2004-05. This distribution o f rural poverty in large measure reflects the distribution o f land, which i s highly skewed in Pakistan. According to the 2000 Agricultural Census, only 37 percent o f rural households owned land, and 61 percent o f these land-owning households owned fewer than 5 acres, (15 percent o f total land) Because of the skewed distribution of land and agricultural income, as well as non- agricultural income's large share of total income for the rural poor, agricultural growth is not highly correlated with reductions in rural poverty. During the 1970s and 1980s, agricultural growth was accompanied by substantial reductions inrural poverty, but rural poverty rates inPakistan did not decline inthe 1990s in spite o f substantial growth in agricultural GDP. In ~ ~ Pakistanhas attempted to implement several land reforms (the most recent inthe early 1970s) that faced substantial opposition and ultimately had only limited success. Sale or grants,ofgovernment land to the poor are potential alternatives. Market-based land reform involving voluntary land sales i s another alternative, but could involve large fiscal costs (See World Bank 2004b). 142 addition to skewed land distribution, possible overestimates o f agricultural GDP growth and an increase inreal consumer prices o f major staples in the mid-1990salso contributed to stubbornly highrural povertyrates. Data from national household surveys indicate that real expenditures of the poorest 40 percent of householdsincreased and rural poverty decreased between 2001-02 and 2004-05. However, comparing 2004-05 per capita expenditures with those in 1998-99, there was almost no change (only -0.2 percent) inreal expenditures per capita o f the poorest 40 percent o f households. A decline in crop incomes due to drought in 2001-02 and then an increase in2004-05 (inlarge part because o f a record cotton crop following good weather) are in part responsible for the fall and subsequent rise in rural household incomes, particularly for cotton farmers. In addition, workers' remittances (some o f which accrue directly to rural households), increased sharply from $2.4 billion to $4.2 billion over the three years, although as these have changed little between 2002-03 and 2004-05, further large increases seem unlikely. Finally, rural households also benefited from overall growth in the economy that spurred demand for construction and other labor-intensive services. Analysis of rural household panel data indicates that primary and secondary education, land ownership, village electrification and paved roads are all significant factors that determinechanges in householdwelfare over time. For example, the presence of an additional male with a secondary education in a household increases real expenditures (a measure o f long- term incomes) by 10.2 percent, a magnitude approximately equivalent to owning five acres o f land (a small farm by Pakistan standards). Econometric evidence also suggests substantial positive returns o f education on income levels in rural Pakistan persist, even after taking into account innate cognitive ability, mainly due to higher productivity innon-farm activities. Social welfare indicators also showed improvement in 2004-05, but unlike income and expenditure-based measures, these gains appear to reflect permanent (structural), rather than transitory factors. Urban residents fare much better than do rural residents, particularly with respect to access to tap water, adult literacy rates, and gross enrollment rates in primary school. All o f these measures show consistent improvement for rural households in surveys from 1998-99, 2001-02 and 2004-05, though there remain large differences between males and females interms ofeducation and literacy. For example, gross enrollmentrates for girls at primary school rose by 18 percentage points, from 50 percent in 1998-99 to 68 percent in 2004-05, but they remain 21 percentage points below those for boys (89 percent). Moreover, these gender gaps are declining only slowly: the gender gaps in gross enrollment rates at both the primary and middle school levels declinedby only five percent between 1998-99 and 2004-05. TOWARDSSTRATEGY FORRURAL A GROWTH AND POVERTYREDUCTION Overcoming the constraints to rural income growth and poverty reduction in Pakistanwill require a strategy that goes beyond agriculture, however, and for several reasons. First,the rural economy involves much more than just agriculture and much o f the gains made in agricultural production accrue to the non-poor. Agriculture accounts for only about 40 percent o f rural household incomes (30 percent for the poorest 40 percent o f rural households) and the majority o f the rural poor inPakistan are not farmers. Secondly, household welfare involves more thanjust income and Pakistan's rural poor are particularly disadvantaged innon-income measures o f welfare. Third, poverty reduction i s ultimately about people, not location or economic sectors. As inmost of Asia, Pakistan's economy is inthe midsto f a major transformationcharacterized by growing urbanizationand increased linkages among the rural, urban and internationalmarkets. 143 Given this context, a comprehensive rural growth and poverty reduction strategy is needed, predicatedon four mainpillars: Promoting efficient and sustainable agricultural growth to raise incomes o f small farmers and to generate growth linkages inthe rural non-farm economy. Creating an enabling environment for the rural non-farm sector to enhance employment and incomes, and improving rural public-service delivery in infrastructure, health, education and population to serve as a foundation for growth and to increase household welfare Improving the effectiveness and governance of rural institutions through the decentralization and strengtheningo f local demand for enhanced accountability as well as through more proactive use o f public-privatepartnerships Empowering the poor and protecting the most vulnerable through social mobilization, safety nets and facilitating access to productive assets for income generating activities. PROMOTINGEFFICIENTAGRICULTURALGROWTH Annual agricultural growth in Pakistan averaged 3.7 percent over the four decades from 1959-60 to 2001-2002, due in large part to high growth in the crops sub sector in the 1970s and 1980s as a result of Green-Revolution technology (improved seeds, increased fertilizer use, and irrigation). However, the performance o f the agriculture sector (particularly the crops sub-sector), has suffered in recent years because o f severe droughts, as well as environmental factors (increased soil salinity and deteriorating groundwater quality). From 1999-2000 to 2004- 05, agricultural GDP grew at a modest 2.3 percent per year (only 0.3 percent per year on a per capita basis), in spite o f a 7.5 percent growth from 2003-04 to 2004-05 due partly to a record cotton harvest. Total crop GDP grew by only 1.8 percent per year, mainly because o f a 1.4 percent annual increase in crop GDPhectare, while the total cropped area increased by only 0.4 percent per year. Livestock, which accounts for nearly half o f agricultural GDP, grew by 3.0 percent per year. Four crops (wheat, cotton, basmati and ordinary rice, and sugar cane) accounted for roughly two-thirds (68 percent in 2002) of agricultural crop GDP and slightly more than one-third (34 percent) of total agricultural GDP in 2004-05. Much o f the focus o f agricultural policy, research and extension has been on these crops, and adoption o f Green-Revolution technology o f improved seeds, increased fertilizer use and irrigation (especially private tube wells, which provide better water control) contributed to substantial increases in yield and production o f wheat and rice. Comparatively little attention has been givento increasing livestock productivity, despite the sub sector's highshare o f total agricultural GDP. Although grain yields have continued to increase since the onset of the Green Revolution, further productivity increases are possible. From 1990-91 to 2004-05, average growth in wheat, basmati rice and maize yields were 2.1, 3.73, and 3.95 percent, respectively. Cotton yields, by contrast, stagnated over this period, increasing by only 0.59 percent per year (although 2004- 05 produced a bumper crop), and yields o f sugar cane increased by only 0.91 percent per year. Past investments in agricultural crop technology (especially for cereals) have had very high payoffs in south Asia in terms o f poverty reduction, both by raising farmers' incomes and by reducing real prices o f major staples (wheat and rice). Further investments in agricultural technology, especially for the major cereals and cotton, (for which Pakistan has a comparative 144 advantage are needed to offset declining soil fertility in many regions as well as to increase productivity. 145 L 0 0 . 0 0 0 . 0 . a U C 2E -0>m Q) 0, -00 53 it 2 c 0 .-C 0 c E w .-.-Cw U Q) C 9 Q) U C m .-E:0 e .-m C P 0 8 Q) -E g> a" cn D! u- 0 c I- -> v) 120 .-m B C m U v) 0 m 2 v) Q) 0 C m L wC 0 . 0 a a W C m Ee ca, a a a e L v) E E z2 m 3 0 v) . . . . e . . . e e 8Cm F S .-E g Lc 0 Achieving substantial growth in agricultural incomesin the future will also require greater diversification into higher-value crops that can maximize returns to scarce factors of production (land and water). This will require greater investment in agricultural research on crops such as oilseeds, vegetables and fruits as well as livestock; better coordination with the private sector in the form o f public-private partnerships; and improvements in seed certification and quality control. Sound macro-management will also be required to avoid an appreciation o f the real exchange rate, which would not only reduce the incomes o f wheat, cotton and rice farmers, but also reduce price incentives for producers o f fruits, vegetables and other items for export. Programs to increaselivestockproduction, especiallyof dairy cattle and milk inPunjab and Sindh, and sheep and goats in NWFP and Baluchistan, can have direct and significant impact on the incomes of the rural poor. As livestock ownership is more evenly spread across rural households than i s access to land, productivity gains in livestock are generally more pro- poor than productivity gains o f major crops. Smallholder's gains are feasible `through use o f improved feeds, better veterinary services, and more efficient marketing channels to help ensure remunerative prices o f milk for farmers. Designing an effective program will require special attention to cultural and economic constraints faced by rural women, who have a major role inthe care o f livestock. Water: The Kev Resource Constraint Increasing the efficiency of water use is critical for the sustainability of agriculture in Pakistan. Water remains the key resource constraint in Pakistan agriculture. Irrigation water is critical for agriculture in most areas o f the country and massive public investments in dams and canals in the Indus river basin, together with mainly private investment in wells for groundwater irrigation have been essential for agricultural production. However, current patterns o f water use in Pakistan are causing significant environmental degradation: losses o f soil fertility due to agricultural soil degradation (soil salinity and erosion) are estimated at Rs 70 billion per year (1.5 percent o f total GDP and 6.8 percent o f agricultural GDP, 2004-05 GDP estimates). Inaddition to environmental losses, inefficiencies in water allocation within canal systems result in lower crop productivity at both the head end (due to over-use o f water and water-logging) and the tail end (due to water shortages) o f canals. Increasing availability of surface irrigation water can only be a partial solution. There is little storage capacity (downstream o f the Mangla and Tarbela dams) innorthern Punjab or water control in Pakistan's run o f the river irrigation system. Adding storage capacity in the form o f dams could in principle help improve water availability in the dry season by spreading out the flow o f water over the year. Nonetheless, even if investments to increase storage are undertaken immediately, storage capacity will not increase within the next ten years, and will likely actually decrease somewhat in this period due to silting of existing facilities. In addition, there i s also a need to improve drainage systems to reduce the salinity and water logging that occur inthe Indus basin irrigation system. Improvements in water delivery (including investments in drainage, control structures and conveyance mechanisms), better water management (through strengthening o f water course organizations, farmer organizations, area water boards or professional canal management agencies) and greater efficiency of water use at the farm level (crop choice and adoption of water-saving technologies) are the most promising approaches in the short to medium term. Irrigation Departments need to establish asset management plans that indicate measures for divestment o f assets to the private sector or farmer organizations, as well as operation and 150 maintenance plans that include medium-term expenditure frameworks, assignment o f financial responsibilities to the various actors, pricing policies and cost-recovery mechanisms. Service delivery and water allocation can be improved through promotion of water user associations (WUAs) and farmer organizations @Os), as well as through improvements inthe existing infrastructure. FOs could be given the authority to collect abiana, as well as operation and maintenance at the distributary and branch canal level. Financial incentives for efficient, sustainable and equitable water use can also help rationalize water use. Given that water i s increasingly scarce in Pakistan, a shift away from water-intensive crops like sugar cane to less water-intensive crops such as oilseeds (sunflower and canola), maize, and vegetables irrigated with drip irrigation techniques could result in large water savings, increasing water availability downstream and reducing environmental degradation. For example, shifting 60 percent o f land cultivated with sugar cane to a maize-wheat rotation or about one-quarter o f ordinary rice land to maize would save enough water to supply the current needs o f Karachi. Charges for canal water (abiana) should be revised so that they better reflect opportunity costs of water usage and thus used as an incentive for farmers to diversify to higher value crops. However, to minimize problems o f nonpayment, any increase in water charges would need to be accompanied by improvements in service delivery. Raising water charges to levels commensurate with the opportunity cost o f water i s likely to be politically infeasible, however. One alternative i s the development o f markets for tradable water rights such as exist in Chile and Mexico. This would require laws establishing tradable water rates and the strengthening o f local institutions to enable them to serve as brokers in water markets. Projects and programs involving agricultural extension, technical assistance, market development and credit may be needed to ease the transition o f farmers and processors to alternative crops and products. Apricultural Markets and Trade Policies Agricultural trade and pricing policies can also significantly affect producer incentives, but distortions to prices of most agricultural products in Pakistan are small. Direct trade and pricing policy distortions, as measured by nominal rates o f protection using official exchange rates, have generally declined in absolute magnitude since the early 1980s. Including the effects o f exchange rate distortions, the pattern i s clearer, with a sharp reduction in policy distortions over the past two decades. Significant trade taxes and restrictions on major agricultural products remain only for vegetable oil, powdered milk imports and basmati rice exports. However, the government continues to intervene heavily in wheat markets at high fiscal costs and generally little benefit to most producers or consumers. Provincial governments (mainly Punjab and to a lesser extent Sindh) and PASSCO (Pakistan Agriculture Storage and Supplies Corporation) procure about 20 percent o f total wheat production each year. On average, a similar amount i s resold before the next harvest, ostensibly to prevent market manipulation by traders and millers and thus stabilize wheat market prices and supplies. One major effect o f these interventions, however, i s to crowd out private trade and storage. Reducing the volume of domestic procurement (and subsequent intra-annual sales) would directly reduce the wheat subsidy that arises from the difference between the sales price o f government wheat (whether procured locally or domestically) to wheat mills and the full costs o f procurement o f wheat and handling. Maintaining incentives for private-sector trade through avoidance o f restrictions on transport o f grain (as were imposed by the government o f Punjab in 2004), continued liberalization o f private-sector imports, and transparent and consistent government policies can help stabilize prices and supplies, while lowering fiscal costs. 151 Faster growth in exports of non-traditional agricultural products (fruits and vegetables, meat, fish) will require improved capacity to meet Sanitary and Phytosanitary (SPS) standards imposedby destination countries. Pakistan presently lacks a coherent strategy (or set o f strategies) for quality and SPS management inrelation to its trade. Organizationally, there i s a need to consolidate and augment the fragmented, isolated, and non-coordinated capacity o f SPS management institutions and regulations by better defining roles and responsibilities o f the various agencies involved in SPS management, strengthen technical capacity for risk assessment, and re-institutionalize early warning and surveillance systems for pests, diseases and contaminants. THERURALNON-FARM SECTORAND THE RURAL INVESTMENTCLIMATE Agricultural growth is an important pathway out of poverty for the 40 percent of Pakistan's rural poor that are farmers; for the 45 percent of the rural poor who do not earn income from the crop sector, however, escapingpoverty requires growth in rural non-farm incomes or out-migration. Simulation analyses based on a 2001-02 Social Accounting Matrix ( S A M ) for Pakistan indicates that growth linkage effects ensuing from increases in traditional crop agriculture are likely to result in only relatively small gains in total rural non-farm incomes relative to the large size o f this sector. As a result the average percentage increase in incomes o f rural non-farm poor may likewise be small. For these households, a well-developed and well- functioning rural non-farm sector i s essential for employment generation, ensuring income diversification and reducing poverty (World Bank 2004b). Results from a 2005 survey of 1651 entrepreneurs in rural villages and small towns" in Punjab, Sindh and NWFP indicate that finance is a major constraint on the growth of the rural non-farm enterprise sector. More than 30 percent o f enterprises surveyed rated access to finance as the most important overall constraint to the operation and growth o f their business. There appears to be considerable unmet demand for formal credit; less than 20 percent o f entrepreneurs that seek formal credit for start-up capital or investment purposes received credit. The major source o f formal financing i s from commercial banks, but access to these institutions i s limited inrural villages; most small-town enterprises however are located in close proximity to a commercialbank. Only about one percent o f start-up and working capital for rural and small town enterprises comes from formal banking sources. Even when entrepreneurs have physical access to commercial banks, they are constrained by an inability to provide sufficient collateral. Eighty percent o f approved loans required collateral. On average the value o f collateral required for a loan was about 132 percent o f the loan amount, making it difficult for households without substantial land or property to gain access to formal credit. While improving financial access for non-farm enterprises is important, survey findings also reveal that most rural enterprises are not able to put together a financial statement, and face problems in marketing their products and hence are probably unable to meet the expectations of financial institutions' due-diligence processes. Therefore in addition to improving supply-side conditions, technical assistance in bookkeeping, documentation, training 91 Small towns are defined as town councils and municipal councils with populations less than 100,000. 152 in business management and entrepreneurial skills for rural enterprises would significantly improve their access to credit and overall business development. Although access to formal medium- to long-term financing is limited, Pakistan's non-farm enterprises seem to have comparativelybetter access to supplier credit. Traders are the mainbeneficiaries o f supplier credit which i s extremely short-term innature (with a median loan term o f 2 weeks). According to the survey, village enterpriseswere more affected by poor infrastructure than those located in small towns. Road quality, availability o f transport, and access to electricity also rank among the top constraints reported by village entrepreneurs. Quality and access to electricity were the major infrastructure problem cited by small-town businesses. Of the three provinces surveyed, villages in Sindh fared the worst in terms o f road quality and availability of transport facilities for goods. Access to electricity remains a major challenge in many rural villages and small towns, and even among enterprises with access, reliability o f supply is uncertain. The median number o f power outage days in a typical month were reported as 20 in villages and 15 in small towns. Frequent outages increase production costs and generators are usually unaffordable for small businesses. Only five percent o f village enterprises and around seven percent o f small town enterprises reported owning or sharing a generator. Power outages are particularly costly for manufacturing enterprises. While small-town and village entrepreneurs surveyed did not complain much about telecommunications, access to and use o f telecommunications among enterprises i s surprisinglylow, particularly considering that close to 60 percent o fthe enterprises were traders. Only 31percent o f entrepreneurs insmall towns owned fixed line phones or cellular phones; the comparable figure invillages i s nine percent. Relatively few firms cited as a constraint governanceissues.This differs significantly from the findings o f the 2003 urbanICA, inwhich 40 percent o fbusinesses rated corruption as a major or severe constraint. There are however regional differences. Firmslocatedin small towns o f Punjab were more likely to complain o f policy uncertainty and those in small towns in Sindh were more likely to cite problems with crime and corruption. Overall however, less than three percent o f approximately 900 firms reported the need to pay an illegal fee. Although few firms mention crime or corruption, contract enforcement and confidence in the legal system remain a concern for many enterprises, particularly in Sindh. As close to half of all rural households in Pakistan are primarily non-farm households and thus heavily dependent on non-agricultural incomes, attention to the constraints facing small-town and village entrepreneurs will be necessary to ensure that this sector is able to grow and generate more employment. Doing so will require a combination o f investments, improvements in service delivery and policy reforms. Technical assistance in the forms o f business development services and strengthening local business organizations would also be beneficial. Overall improvements in human-resource development and greater levels o f literacy are also needed as the survey data clearly indicate that firms with more educated managers are more likely to innovate and tap into business networks and consistently seem to performbetter. IMPROVINGDELIVERY PUBLICSERVICE Provision of basic rural infrastructure and delivery of basic public services such as water supply, sanitation and roads, is crucial for rural economic growth and improved welfare of the rural poor. Pakistan's devolution program, begun in2002, was designed to improve public- service delivery by increasing the decision-making and implementation authority o f local governments and augmenting their fiscal resources. Evidence from case studies in eight districts 153 in Punjab, Sindh and NWFP suggests that devolution has been broadly successful in achieving these goals, though more remains to be done even inthe most successful cases. Fiscal resources for rural development have increased at the federal and provincial levels since 2001-02; changes in the level of resources at the district levels vary by province. From 2001-02 to 2005-06, total consolidated expenditures (fiscal and provincial), government revenues and the budget deficit all increased by 23 to 26 percent in real terms. A large share o f these additional resources was devoted to development expenditures, which increased by 85 percent (18.1 percent per year), reflecting the government's commitment to development investments. Growth in development expenditure was particularly rapid at the provincial level (48.8 percent per year) as compared to the federal level (10.6 percent per year). Total transfers to local government (districts, tehsils and unions) varied by province, however. Average transfers to local governments in Sindh in 2004-05 and 2005-06 were 50 percent higher in real terms than the average o f 2002-03 and 2003-04; in Punjab and NWFP, by contrast, there was essentially no change over time. It i s noteworthy that large increases in expenditure for health and education at the district level are not shown inthese budgets, however. Resources vary enormously across districts, tehsils and villages. Ina survey o f seven districts inPunjab, Sindh and NWFP, availability of resources per capita ranged from 981 Rs/person in Sargodha (Punjab) to 2000 Rs/person in Chakwal, also in Punjab. On average, current expenditure accounted for 80 percent o f total expenditure (salaries alone accounted for 63 percent) and the share of resources devoted to development expenditures nevertheless ranged from 3-35 percent. Even wider variation in fiscal resources is found at the TMA level, with expenditure per capita more than eight times as high in Digri (381 Rs/person) and Khairpur (361 Rdperson) as in Timergarah and Samarbagh o f NWFP (both 44 Rdperson). In part, this difference in resources reflects a greater ability o f some tehsils to raise their own revenues through local taxes and fees; district awards to tehsils also vary, from only 11Rdcapita inthe two NWFPtehsils sampled to 240 Rdcapita inKhairpur. Even five years after devolution, confusion remains over the roles and responsibilities of the various levels of government, as well as apparent jurisdictional overlaps. Staff shortages and lack o f appropriate skills are also serious constraints to efficient service delivery, particularly at the TMA level. Moreover, the transfer o f resources through the vertical programs and other higher-level government programs adversely affects local governments' autonomy in preparing development budgets as well as undermining accountability and operational efficiency. Substantial resources are allocated, especially in the water and sanitation sector, through Members o f the National Assembly (MNAs), Members o f Provincial Assemblies (MPAs) and directives o f the Chief Minister. The delegation of power to local governments through devolution has brought decision making closer to the communities and allowed them to access government officials and there has been some improvement in service delivery. As reported by all 14 tehsils surveyed (except Bahawalpur) citizens' degree o f access to their representatives has increased considerably and problems are solved relatively quickly under the new system. Priorities o f the local communities are also increasingly reflected in development schemes. However, the performance o f Citizen Community Boards (CCBs) in terms o f their financial contribution as a percentage of government financing o f development expenditures at the district level varies substantially, from only 2-6 percent in the three districts o f Sindh sampled to 22-34 percent in the three Punjab districts. 154 IMPROVING THE EFFECTIVENESS GOVERNANCERURAL AND OF INSTITUTIONS Devolution in 2002 was an important step toward a more decentralized, efficient and demand-driven government in Pakistan. In addition to increasing resources for local government and improving its administrative capacity, other measures are needed to improve the effectiveness and governance o f rural institutions. Greater interaction between private-sector groups and farmers in the form o f public-private partnerships could facilitate government responsiveness and improve efficiency o f investments and programs. Measures that improve governance and accountability such as stricter enforcement o f laws, reforms in the legal system and wider dissemination o f information on spending and effectiveness o f programs would also help spur both greater efficiency o f government, as well as growth and investment in the private sector. Public-private partnerships (PPPs) should be pursed aggressively in agricultural marketing, *researchand extension, health, education, infrastructure and other sectors to enhance productivity of public-sector programs. An incentive structure should be put inplace to encourage the private sector to enter into contractual arrangements with farmer organizations and link delivery o f public services to business development. Public-private partnerships in agricultural research and extension would encourage public research systems to become more responsive to farmers' actual needs. Recent education sector reforms undertaken by the government should be strengthened to enable engagement with private and NGO schools to improve access and quality o f education, including government-supported per child subsidies. Finally, government financial support could be increased for NGOs, including Rural Support Programmes for delivery o f micro-credit, skills training and other private services. Improving governance is crucial for reducing poverty and promoting growth in both the farm and non-farm sector. Governance issues, including political instability, corruption, insecurity and lingering conflicts, are a major drag on investment, economic growth and efforts to reduce poverty. Bureaucracies with low accountability to stakeholders give rise to inefficiency and shortfalls in delivering rural services. To overcome these challenges, the government should strengthen efforts begun inrecent years to: i)reform government institutions related to economic and financial management, the police, the judicial system, and the civil service; (ii)improve public financial management, accountability, and increased transparency and information on government activities to facilitate public oversight; (iii) strengthen local government institutions; and (iv) privatize and deregulate public enterprises to reduce opportunities for rent-seeking behavior. Monitoring and evaluation within a national and results-based management framework is increasingly important with burgeoning CDD and microfinance programs as there is greater uncertainty about what will work than intraditional projects. There is a need for up- front poverty indicators and simple baseline assessments and ongoing monitoring o f indicators against targets. Although "better M&E" i s so often mentioned as to be almost a clichCd recommendation, the increase in CDD projects and microfinance, which by their nature are uncertain and require rapid management reactions, effective M&E with a poverty impact focus becomes muchmore important. 155 EMPOWERING POORAND PROTECTINGTHEMOST THE VULNERABLE Increasing the flow o f resources to local governments and improving administrative procedures addresses mainly the supply o f public services. Improving the effectiveness o f development programs requires more effective demand for public services, as well. In addition, although inclusive economic growth should be the main mechanism for reducing poverty, increased social protection efforts are needed to protect the most vulnerable. A major reason for the limited impact o f rural development efforts in Pakistan and many other countries is a lack o f effective participation and influence o f r u r a l poor households in development programs and rural society. Too often, a top-down approach has been implemented, one inwhich the rural poor are simply beneficiaries o f public programs suppliedby the government. Rather, the development paradigm shouldput the household and the community at the origin o f development initiatives. Empowering the rural poor to take on this role, however, requires social mobilization, a central pillar o f the government's Mid-Term Development Framework (2005-10). Social mobilization can be defined as the enhancement o f the community's capacity for collective action for its own betterment. Together with economic empowerment and graduation, it should be at the heart o f the rural livelihood development strategy. Mobilizing the poor, a major component o f Pakistan's Rural Support Programmes (RSPs), provides them with the voice and the scale required to more effectively engage with the range o f institutions and individuals that provide public and private services. Economic empowerment, e.g. through micro- credit or skills training, facilitates access to assets, increases incomes, and demonstrates the creditworthiness o f individuals and groups. Ultimately these self-help groups may graduate to form other federative and associative movements involved in income-earning activities such as: public-private partnerships, various types o f franchising and contract farming arrangements. Social-mobilization components have been incorporated into various rural development projects and programs, often together with micro-credit and public investment under the broad rubric o f community-driven development (CDD). In principle, governments may be best suited for delivery o f public goods and services (and the investments required to support these) and the private sector and NGOs best suited for delivery o f private goods and services (including micro-credit and skills training). However, the best institutional pathway to promote active community participation and their linkages with both public and private sectors i s debatable, and depends on the particular characteristics o f communities, government institutions and development constraints. In areas where local government institutions are still unable to effectively deliver public services, direct interventions through community organizations and NGOs are still necessary. These programs work best where there are existing strong community groups; rural micro-level infrastructure and public health are especially important in isolated areas. To ensure long-term sustainability, however, it i s important that these efforts do not undermine efforts intended to strengthen local government institutions. Microfinance programs, typically run through NGOs, including the Pakistan Poverty Alleviation Fund (PPAF), also channel money to r u r a l areas, although this i s mainly to individual households and enterprises. Global experience with microfinance suggests a number o f relevant lessons for Pakistan, such as: (i) the need to focus on savings and the efficiency o f savings services as these may be even more important for the poor than credit, and micro-credit 156 without savings institutions i s likely unsustainable; (ii) need to continually assess product the demand and to design savings and lending products to meet that demand (informal lenders, and even the staff o f credit institutions themselves, have been able to develop suitable products); (iii) exceptionalqualifiedsuccesswithlendingtowomen(the majorityofsustainable loan the initiatives worldwide have involved women) suggests that the comparative advantage within the household may be large, although socially acceptable mechanisms to lend to women may differ across localities; and, (iv) the importance o f finding efficient ways o f linkinglocal informal rural finance institutions to formal institutions. Greater efforts need to be made to protect the most vulnerable households and individuals, as well. Safety net programs to supplement the incomes o f the poorest households remain very limited, particularly in rural areas. The largest program, Zakat, i s financed through a voluntary levy o f 2.5 percent o f the value o f financial assets deducted once a year and covers 1.733 million beneficiaries at a total cost o f Rs 5.86 billion. The Pakistan Bait-ul-Mal, a Federal government agency that delivers vertical social welfare programs financed from general revenues, includes the Food Support Program, which delivers Rs 4.5 billion cash transfers to 1.25 million beneficiaries. Together with the Tawana Pakistan program that provides mid-day meals to 530,000 girls in government rural primary schools at an annual cost of Rs 0.70 billion, these safety net schemes total only Rs 11.1 billion, 0.23 percent o f GDP. The other major social assistance program, the wheat subsidy, costs Rs 8 billion, i s untargeted and i s unlikely to provide muchbenefit to poor rural households, except to the extent that net wheat sales from public sector imports and drawdown o f government stocks lower market prices. Safety nets for both the rural and urban poor should be made more efficient and scaled up to cover all ultra-poor households and some poor households (4-1 1 million households), largely through increased use of conditional cash transfers. The wheat subsidy program could be replaced by a food- or cash-based targeted program. Cash transfer programs such as the Food Support Program can be scaled up and the scope o f these programs widened to encourage long- term human capital investments among the poor, for example based on the pilot o f the Child Support Program. The TawanaPakistan school feeding program should, similarly, be redesigned and scaled up. Nutritional interventions aimed at improving health and nutrition o f women and infants ought to be piloted and scaled up. The government could also consider piloting a major rural workfare program in order to provide temporary employment for vulnerable households. Programs are also needed to address the problem o f heavily-indebted laborers, which persists particularly among landless sharecroppers inrural areas o f southern Punjab and northern Sindh. CONCLUSIONS Many of the constraints on the efficient use of resources, economic growth and poverty reduction discussed in this report have long been recognized: inefficient water usage, unequal distribution of land and incomes, low levels of infrastructure and delivery of public services. Several factors have changed, however. Environmental degradation has become increasingly serious, as has competition for rural and urban water uses. Agriculture's share o f the overall economy has declined and the rural economy has become more closely integrated with urban and international economies. The latter two factors, along with the highly skewed distribution o f land, offset but do not eliminate the impact o f agricultural growth on rural poverty. Nonetheless, Pakistan has made important strides in the last several years in promoting rural growth and poverty reduction. There has been substantial progress in liberalizing agricultural markets, promoting diversification and exports, and increasing expenditures on 157 infrastructure and public services in rural areas. Progress has also been made in empowering the poor through social mobilization (Rural Support Programmes and the creation o f CCBs) and microfinance (the Pakistan Poverty Alleviation Fund). These efforts should be continued and strengthened. In addition, rapid and sustained reduction o f rural poverty will require an even greater focus on building human capital (through improved delivery o f health services and sanitation, basic education and appropriate technical training). Moving forward, it will be crucial to maintain a multi-faceted approach to rural development and poverty reduction and to ensure that sufficient resources are invested in the task. Efficient agricultural growth is a necessary condition for robust rural economic growth, but inand o f itself cannot be expected to drive substantial reductions inrural poverty in light of the highly unequal distribution o f access to land and water resources inPakistan. Thus, policies to promote agricultural growth must be complemented by investment and policies that create an enabling environment for the rural non-farm sector, including increased public investment inrural and small-town infrastructure. This will require shoring up the effectiveness and governance o f rural institutions through decentralization, strengthening local demand for public services and public-private partnerships. Finally, because the ultimate objective o f any poverty reduction strategy i s to increase the welfare o f the poor, health and education programs that promote the human capital o f the rural poor, social mobilization that empowers and social protection programs that protect the most vulnerable should all receive increasedresources and attention. 158 REFERENCES Agrodev. 2000. Sindh Rural DevelopmentProject-FinalReport prepared for Planning and Development. The Government of Sindhand the Asian DevelopmentBank, October 2000, Agrodev Canada Inc. Ahmad, Munir. 2002. "Productivity Growth Differential in Punjab, Pakistan: A District-Level Analysis," The Pakistan Development Review, Vol. 40, No. 1: 1-25. Ahmad, Munir, Chand, Asghar and Nasir, Alam Khan. 1993. "Participation of rural Women in Agricultural and Household Activities: A Micro-Level Analysis," Haider, A.S., Hussain, Z., McConnen, R. and Malik, S.J. (eds.) Agriculture Strategies in the 1990s: Issues and Policies. Islamabad: Pakistan Association of Agricultural Social Sciences. Ali, Mubarak and Derek Byerlee. 2002. "Productivity Growth and Resource Degradation in Pakistan's Punjab: A Decomposition Analysis," Economic Development and Cultural Change Vol. 50: 839-863. , 2004. Productivity Growth and Resource Degradation in Pakistan's Punjab: A Decomposition Analysis, World Bank Policy ResearchWorking Paper No 2480. Arif, G. M.2003. UrbanizationinPakistan: Trends, Growth and Evaluation of the 1998 Census, inUNFPA (2003) Population of Pakistan: An Analysis of 1998Population and Housing Census Kemal, A.R., Irfan, M.and Mahmood, N.(eds.) Pakistan Institute of DevelopmentEconomics. Asian Development Bank and World Bank. 2005. Sri Lanka: Improving the Rural and Urban Investment Climate. Washington DC: World Bank. Cororaton, Caesar B. and David Orden. 2007. "Inter-sectoral and Poverty Implications of Changes in Cotton and Textile Policies in Pakistan: A CGE Analysis". Draft Research Report. Washington, D.C. InternationalFoodPolicy ResearchInstitute. Datt, Gaurav and Martin Ravallion. 1998. "Farm productivity and rural poverty in India," Journal of Development Studies. No. 34 (April). ' Dorosh, Paul A., Muhammad Khan Niazi and Hina Nazli. 2003. "Distributional Impacts of Agricultural Growth in Pakistan: A Multiplier Analysis," The Pakistan Development Review. 42(3): 249-275. Dorosh, Paul. and Sohail J. Malik. 2006. Transitions Out of Poverty: Drivers of Real Income Growth for the Poor inRural Pakistan. BackgroundPaper. Washington, D.C. Dorosh, Paul and Abdul Salam. 2006. Wheat Markets and Pricing in Palustan: Political Economy and Policy Options. Background Paper.Washington, D.C. . 2007. Trade and Price Policy Distortions in Pakistan Agriculture. Background Paper. Washington, D.C. 159 Dorosh, Paul. 2006. Distributional Impacts o f Agricultural Growth. Background Paper. Washington, D.C. de Ferranti, David, Guillenno E. Perry, William Foster, David Lederman, and Albert0 ValdCs. 2005. "Beyond the City: The Rural Contribution to Development." Latin American and Caribbean Studies, World Bank. Washington, D.C. Deolalikar, A. 2005. Attaining the Millennium Development Goals in Pakistan. South Asia Human Development Sector, Discussion Paper Series Report No. 8. World Bank. Washington, D.C.: Fafchamps, Marcel and Agnes R.Quisumbing.1999. Human Capital, Productivity, and Labor Allocation in Rural Pakistan. FCND Discussion Paper No. 48. Food Consumption and Nutrition Division. InternationalFood Policy ResearchInstitute.Washington, D.C. Food and Agriculture Organization (FAO). 2006. Food Balance Sheets. www.fao.org. Feeney, Griffth and Iqbal Alam. 2003. Fertility, Population Growth and Accuracy of Census EnumerationinPakistan 1968-1998. Populationand DevelopmentReview, October 15 Government of Pakistan. various issues. Household Income and Expenditure Survey. Federal Bureau of Statistics, Islamabad. . various issues. Pakistan Rural Household Survey. Federal Bureau of Statistics, Islamabad. . various issues. Pakistan Economic Survey. Economic Advisor`s Wing, Ministry of Finance, Islamabad. . various issues. Monthly Statistical Bulletin. Federal Bureau o f Statistics, Islamabad. . 1985. Rural Credit Survey o f Pakistan. Agricultural Census Organization, Lahore. . 1999. Agricultural Statistics ofPakistan. Ministryo fFood, Agriculture and Livestock, Islamabad .2002. Water Vision 2025 Country Report, cited inWorld Bank 2003. PalustanPublic Expenditure Management, Accelerated Development of Water Resources and Irrigated Agriculture, Vol. 11. . 2003. Cost of Production series data for various crops. Agricultural Prices Commissions (APCOM), Ministry o f Food, Agriculture and Livestock, Islamabad. . 2005. Pakistan Social and Living Standards Measurement Survey. Federal Bureau of Statistics, Islamabad. Government of Punjab (various issues), Annual Budget Reports. Punjab Department of Finance, Islamabad. 160 Jacoby, Hanan G. and Mansuri, Ghazala. 2007. "Incomplete Contracts and Investment: A Study o f LandTenancy inPakistan. World Bank Policy ResearchWorking Paper No. 3826. . 2007. "Incentives, Supervision, and Sharecropper Productivity". World Bank Policy ResearchWorking PaperNo. 4125. Karim, M. S. and A. Nasar. 2003. MigrationPatterns and Differentials inPakistan inKamalA.R., et. al. Population of Pakistan. Pakistan Institute of Development Economics, Islamabad. Malik, Sohail J. 1999. Poverty and Rural Credit: The Case of Pakistan. Pakistan Institute of DevelopmentEconomics, Islamabad. . 2003. Rural Credit Markets Pakistan: Institutions and Constraints. Background paper in for World Bank Pakistan Rural Factor Markets study. Nazli, Hina. 2003. Rural Labor Markets in Pakistan: Institution and Constraints. Background paper for World Bank Pakistan Rural Factor Markets study. Niazi, Muhammad Khan, Nuzhat Ahmad and M. Khalid. 2006. Rural Service Delivery in Pakistan. Case Study. Background paper for Pakistan Spurring Rural Growth and Poverty Reductionreport. Nuzhat, Ahmad 2006. Improving Rural Service Delivery: Rural Institutions at the Community and Local Government Levels. Background paper for Pakistan Spurring Rural Growth and Poverty Reductionreport. Orden, David, Abdul Salam, Reno Dewina, Hina Nazli and Nicholas Minot. 2005. TheImpact of Global Cotton Markets on Rural Poverty in Pakistan. Pakistan Poverty Assessment Update, BackgroundPaper Series. Asian DevelopmentBank, Islamabad. Pakistan Poverty Alleviation Fund(PPAF). 2003. Annual Report. World Bank. Population Reference Bureau. 2001. World Population Data Sheet. Quoted in Pakistan Demographic Survey 2001. Washington, D.C. Riaz, Khalid. 2006. Analyzing Success Stories in Agricultural Development in Pakistan: Case Studies of Milk Processing Sector, Tunnel Farming for Vegetables, Citrus Marketing and Exports. Background Paper. Islamabad. Stone et al. 2006. Assessment Report: Pakistan Devolution Support Project, preparedfor USAID. The Urban Institute (unpublishedreport). Washington, D.C. Sur, Mona and Zhang Jian. 2006. The Pakistan Rural Investment Climate Survey. Background Paper. World Bank. Washington, D.C. WAPDA 2002. Pakistan Water Sector Strategy. Ministry of Water and Powers Report. World Bank. 2003. Poverty Reduction Strategy Paper on Pakistan, Report No. 25463-PAK. Washington, D.C. 161 . 2004a. AgricultureInvestmentSourcebook. Agriculture and Rural DevelopmentUnit. Washington, D.C. . 2004b. PakistanRural Factor Markets Policy Reforms for Growth and Equity. Report No. 30381-PK.Washington, D.C. . 2004c. Trade Policies in South Asia: An Overview. vol. 11: An Overview. Washington, D.C. . 2004d. Devolution Pakistan - Main Report: An in Assessment & Recommendations for Action. ReportNo. 29912-PK. Washington, D.C. .2005a. SindhEconomicReport.Washington, D.C. . 2005b. Pakistan Country Water Resources Assistance Strategy. Water Economy: RunningDry,ReportNo. 24081-PK. Washington, D.C. . 2005c. ReducingPoverty on a Global Scale. "Case study on Andhra Pradesh"; project documents.Washington, D.C. . 2005d. India: Re-energizing the Agricultural Sector to Sustain Growth and Reduce Poverty.New Delhi.Oxford University Press/WorldBank. . 2005e. IndiaNational Agricultural Innovation Project. Project Information Document (PID). http://www.worldbank.orq.in/external/defaultimain?parrePK=64027221&uiPK=64027220&theSit ePK=295584&menuPK=295 617&Proi ectid=P092735 .2006a. Pakistan:Growth and Export Competitiveness;ReportNo. 35499-PK. Washington, D.C. ,2006b. Summary of Key Findings andRecommendations(Note on poverty estimates by Tara Vishwanath andNobuo Yoshida). http://siteresources.worldbanl<.orp/ INTSAREGTOPPOVRED/1337567-1152551765388/20987772/PovertyHCR2000-2005.pdf .2006c. TheRural Investment Climate: It Differs and I t Matters. Agriculture andRural DevelopmentDepartmentReportNo. 36543-GLB. Washington, D.C.: World Bank. .forthcoming. Pakistan: SocialProtection.Washington, D.C. Websites consulted www.pak.gov.pk www.fa0.org www.finance.gov.pk w w .us.iIo.org/archive/iIofocus/2001/winter www.worldbank.org wuw.adb.org www.unesco.org www.iae.a.org 162 Pakistan at a glance 4/6/07 POVERNand SOCIAL 'South Low. Pakistan Asia income Developmentdiamond. 2006 Population. mid-year (millions) 159.0 1,470 2,353 GNI per capita (Atlas method, US$) 770 684 580 Life GNI (Atlas method, US$ billions) 122.1 1,005 1.364 - Average annual growth, 2000-06 Population (%) 2.4 1.7 1.9 Labor force 1%) 3.8 2.1 2.3 GNI Gross per pnmary Most recent estimate (latest year available, 2000-06) capita enrollment Povew (% of population below nationalpovertyline) Urban population (% of totalpopulation) 35 29 31 Life expectancy at birth (years) 65 63 59 I Infant mortality (per 1,000live births) 80 66 80 Child malnutrition (% of children under 5) 38 45 39 Access to improved water source Access to an improvedwater source (% ofpopulation) 91 84 75 Literacy (% ofpopulation age 75+) 47 60 62 - Gross primarqenrollment (% of school-agepopulation) 82 110 104 Pakistan Male 95 116 110 Low-income group ~ Female 69 105 99 KEY ECONOMIC RATIOS and LONG-TERM TRENDS 1986 1996 2005 2006 Economicratios* GDP (US$ billions) I 31.9 63.3 110.9 128.9 Gross capital formationlGDP 18.8 19.0 18.1 20.0 Exports of goods and services/GDP 11.9 16.9 15.5 15.5 Trade Gross domestic savings/GDP 8.0 14.5 14.3 11.1 Gross national savingslGDP 19.7 18.2 20.6 17.3 Current account balance/GDP -2.4 -7.3 -0.8 -3.7 Interest paymentdGDP 1.8 1.9 0.7 Domestic Capital Total debVGDP 46.9 47.1 30.4 savings formation Total debt service/exports 25.0 28.3 10.9 Present value of debffGDP 25.4 Present value of debffexports 125.9 Indebtedness 1986-96 1996-06 2005 2006 2006-10 (average annual growth) GDP 5.0 4.2 7.3 6.2 6.7 Pakistan GDP per capita 2.4 1.7 4.7 4.1 4.5 Exports of goods and services 8.0 8.3 9.6 12.9 17.0 STRUCTUREof the ECONOMY 1986 1996 2005 2008 I 1%of GDP) Growth of capltal and GDP (Oh) I Agriculture 27.6 25.5 22.2 20.5 IsT Industry 23.4 24.2 26.5 26.7 10 Manufacturing 16.3 16.0 17.8 18.3 5 Services 49.0 50.4 51.3 52.9 0 1 5 Householdfinal consumption expenditure 79.2 72.9 78.0 81.4 -10 1 General gov't final consumption expenditure 12.8 12.6 7.7 7.6 - Imports of goods and services 22.7 21.4 19.3 24.4 GCF *GDP 1986-96 1996-06 2005 2006 (averageannual growth) Agriculture 4.3 2.5 6.7 2.5 industry 5.8 5.5 11.4 5.9 Manufacturing 5.2 6.8 12.6 8.6 Services 5.1 4.8 8.0 8.8 Householdfinal consumption expenditure 4.7 4.4 13.1 8.1 General gov't final consumption expenditure 3.0 0.5 1.7 4.8 - 2 0 1 Gross capital formation 4.3 1.6 9.3 9.9 - Exports -0-lrnpolts Imports of goods and services 4.3 3.3 40.5 23.9 Note: 2006 data are preliminaryestimates. Group data are to 2005 The diamonds show four key indicators in the country (in bold) compared with its income-groupaverage. If data are missing, the diamond will be incomplete. 163 NCE 4986 1996 2005 2008 I 108 6 3 7 9 3 3 8 4 R ? 30 3 1 7 5 180 14 1 14 4 -0 Y -2 8 0 4 -8 i -7 8 -4 2 1986 1996 20Q5 2006 and Import 1 2,945 8 313 1 16 185 923 51,P 124 342 504 933 1 a1z 2,055 4,989 8 268 9,882 6,002 72075 18996 23 967 1,519 706 783 1039 2010 4,534 l l Y 324 128 113 131 142 10s 84 YO 1486 2846 2005 2006 Current accountbalancoto GDP (VO) I 3798 s977 z 20 254 7,230 18227 2 33 098 4 4 3 4 -5,2513 -12844 -640 -f 9% "2 386 6 3,302 26111 6 3.14 7 -712 -4593 .849 .4,8'13 1,200 4163 239 5 6.13 428 431 610 -1 137 1,838 2,839 1' 516 12468 ? a t 335 59 4 59 9 1886 1986 2005 2006 i4BB.j 23829 33 E75 603 3007 1580 3480 1626 3287 2 443 38 436 404 26 BJ 167 315 zoo P I1 398 I115 590 58 556 933 138 922 2 3 285 758 558 1550 f 249 2351 a45 I F 50 2% 427 i 193 386 41s 63 234 i95 138 152 274 I64