Report No. 37858-SAS South Asia Growth and Regional Integration December 2006 Poverty Reduction and Economic Management Sector Unit South Asia Region Document of the World Bank TABLE OF CONTENTS 1 SouthAsia's Growthand RegionalIntegration: An overview . .............................................................. 1 Introduction ............................................................................................ ............................ 1 I s Growth Sustainable ................ ......................................................................................................... 4 Cost ofDoingBusiness ......................................................................................................................... 7 Sources ofGrowth................................................................................................................................. 5 Institutions............................................................................................................................................. 8 Knowledge........................................................................................................................................... 13 Infrastructure ....................................................................................................................................... 19 Is Regional Integration Desirable ............................................................................................................ 22 AgglomerationBenefits....................................................................................................................... 22 Trade.................................................................................................................................................... 24 Loweringthe InfrastructureConstraint throughRegionalCooperation.............................................. 28 Conclusion................................................................................................................................................ 31 2.EconomicGrowthin SouthAsia: A GrowthAccountingPerspective ................................................ 33 Introduction ........ .................................................................................... 33 Growth in South Asia: An Overview................................................................... ........33 Output Growthversus Living Standards.............................................................................................. 34 Growth Accounts ............................................................................................................. InvestmentRa EducationalAttainment:A LaborQuality Index................................................................................. s CapitalAccumulation........................................................................ 43 SomeImplications Looking Forward ....................................................................................................... 44 Annex: Growth Accounts ............................................................................................... ..........46 3.Improvingthe Climatefor Investment and BusinessinSouthAsia .................................................... 47 Introduction........... .......................................... 47 DisaggregatingInvestment Climate Measures......................................................................................... 48 48 Doing Business ExaminesRegulatoryHurdlesFacingEntrepreneurs................................................ SubjectiveRankings ............................................................................................................................ 49 UsingInvestment ClimateSurveysto Identify Key Challengesfor Improvingthe Investment Climate inSouthAsia ................................................................................... ............................................... 50 Linking Investment Climate Measures to Firm Pe~ormance InvestmentClimateandProductivity................................................................................................... 58 ................................................................... 56 InvestmentClimateandExporting...................................................................................................... 59 Conclusions .......................................................................................................................... 4.PropertyRightsInstitutions. ContractingInstitutions. and Growthin SouthAsia: Macro and Micro Evidence ........................................................................................................................................ 61 Introduction .............................................................................................................................................. 61 Cross-country Evidence on Institutions and Growth............................................................................... 64 MeasuringPropertyRightsInstitutionsandContractingInstitutions.................................................. 66 DeepHistoricalDeterminants ofInstitutionalQuality ........................................................................ 66 Do InstitutionsMatterfor Development: IdentifyingCausation........................................................ 69 Firm-Level Evidence on Institutions and Growth..................................................................................... 71 Dothe Micro andMacroDataAgree .................................................................................................. 71 DoProperty RightsInstitutionsand ContractingInstitutionsMatterfor Firms................................... 75 HowDoFirmsCopeWith Weak ContractingInstitutions.................................................................. 78 Conclusions ................................................................................................................................ 80 Annex 1.First-Stage Regressionsfor Property Rights and Contracting Institutions. Small Sample .......81 Annex 2.Instrumental Variables Regressions of In (GDP Per Capita) on Property Rights and Contracting Institutions. Small Sample........................................................................ Annex 3. VariableDefinitions and Country Coveragefor the WBESand PICS Datas 5.Improving Technology. Skills. and Innovation in South Asia ............................................................. 85 Introduction .............................................................................................................................................. 85 TheNew Competitive Context .................................................................................................................. 86 South Asian Countries in the Context ofthe Knowledge Economy .......................................................... RelativePositionof SouthAsianCountries......................................................................................... 89 World Bank KnowledgeAssessment Methodology............................................................................ 88 88 Global Trends in Education and the Situation of South Asian Countries................................................. 91 EducationandSkills inSouthAsianCountries................................................................................... GlobalTrends inEducationandTraining............................................................................................ 91 93 Global Trends in Innovation and Situation of South Asian Countries ..................................................... 95 GlobalInnovationTrends.................................................................................................................... 95 InnovationinSouthAsianCountries................................................................................................... 98 Overall Assessment................................................................................................................................... 99 Key Actions ............................................................................................................................................. 100 Improvingthe Economic IncentiveandInstitutional Regime ........................................................... 100 100 Tapping into GlobalKnowledge........................................................................................................ StrengtheningEducationand Skills................................................................................................... 101 CreatingKnowledge.......................................................................................................................... 102 DisseminatingExistingKnowledge, Especiallyto the Very LargeTraditional Sectors inEach Country.............................................................................................................................................. 103 Networking and Collaboration Among South Asian Countries ........ ................................................. 104 6.Power Sector Reform. Private Investment. And Regional Cooperation ........................................... 107 ......................................................................... 113 Obstacles to Private Investment ............................................................ Alternative Sources and Uses of Investment Finance .................................. 119 ............................................................................. 120 Conclusions ........................................................................................ Annex: Data and Analysis of Energy Supply and Demand...... The RelationshpbetweenElectricity ConsumptionandGDP.......................................................... 124 7. Trading Choices of South Asia ............................................................................................................. 129 Introduction ............................................................................................................................................ 129 Triumph of TradeLiberalization ........................................................ ..................................... 130 Preferential Trading in South Asia: A Brief Historical Background..................................................... .I32 SAFTA: A Critical Examination ............................................................................................... Sri Lanka-India FTA: Sectoral Exceptions and Rules of Origin ......... ............................................. 138 The WayForward: An Asian Bloc via India-China FTA ....................................................................... 139 Concluding Remarks ................................................................................................. 141 8. The Role of Trade Facilitation inExport Growth and Inter-regional Trade ................................... 143 Introduction ........................................................................ ........................ 143 Characteristics of Trade in South Asia. .......................................................... 144 Logistics and CompetitiveAdvantage ................................................................................ .I47 ................................................................................. ........................ 150 Road Transport....... ............................................................ 155 Rail Services.......... ...................................................................... 156 ...................................................................... ......................... 158 LandBorders ..................................................................................................................................... ..................................................... 159 161 Ports................................................................................................................................................... 161 Regulatory Impediments.... ............................................................ 164 Duties and Other Taxes Regional Tradeand TransitAgreements....................................... .................167 Supply Chain Management....................... .......................................... I70 Conclusions and Recomme 9. CuttingTrade Costs and Improved BusinessFacilitation In South Asia: Estimating The Benefits OfReform ............................................................................................................................................... 179 Introduction,......... ............................................................... 180 TradeFacilitation.. ........................................................................................................ Measuring the Impact of TradeFacilitation Summaly Overview of Conditions in South Asia.................................................................... Port Infrastructure and Eficiency .................................................. Land Transportation................................. Border Crossings and Customs........................................ InformationTechnology and Services Sector Infrastructure ............................................................. Standards and Technical Regulations ................................................................................................ 192 193 Estimating the Gains From Capacity Building in South Asia .................................................. Expanding Intraregional Trade..................................................................................... Global Trade andthe SouthAsia Region ................................................................. Conclusions .......... ................. ............................................................ 200 Annex ....................................... ................................................................................... References...................................................................................................................................................... 203 TABLES Table 1.1:South Asia: Selected Indicators...................................................................................................... 5 Table 1.2: Sources of Growth. SouthAsia. 1960-2003 .................................................................................. 6 Table 1.3: Infrastructure Stocks inSouth Asia, East Asia, and China, 2003................................................. 20 Table 1.4: South Asia is a Recent Global Integrator: Exports andImports o f Goods and Services as Proportion o fthe GDP............................................................................................................... 24 Table 1.5: Long- TermPotential for Energy Trade inSouthAsia............................................................... 29 Table 2.1: South Asia: Selected Indicators.................................................................................................... 34 Table 2.2: Components o f Gross NationalIncome per Capita ...................................................................... Table 2.3: Sources o f Growth, SouthAsia, 1960-2003 ................................................................................ 36 40 Table 2.4: Sources o fGrowthby Region ...................................................................................................... 41 Table 2.5: Average Years o f Schooling (Population Aged 15+) ................................................................... Table 3.1: Governance Indicators ................................................................................................................. 43 Table 3.2: New Data Sources from the World Bank Allow for Comparisons Across Countries ..................49 49 Table 3.3: A Better Investment Climate is Associated With Higher Productivity ........................................ Table 3.4: A Better InvestmentClimate i s Associated With Higher Wages and Investment Rates..............58 59 Table 3.5: InvestmentClimate Conditions Affect the Probability a FirmWill Export................................. 60 Table 4.1:First-Stage Regressions for Property Rights and Contracting Institutions................................... 67 Table 4.3: Regressions o fMicro Measures from WBES on Macro Measures: Contracting Institutions ....73 Table 4.2: Instrumental Variables Regressions o fIn(GDP Per Capita) ....................................................... 70 Table 4.4: Regressions o fMicro Measures FromWBES on Macro Measures: Property Rights Institutions Table 4.5: Regressions o fMicro Measures FromPICS on Macro Measures: Contracting Institutions ......73 ................................................................................................................................................... 74 Table 4.6: Regressions o f MicroMeasures FromPICS onMacro Measures: Property Rights Institutions.75 Table 4.7: FirmPerformance, Contracting, and Property Rights .................................................................. 77 Table 4.8: Getting Around Weak Contracting Institutions: Regressions Using PICS Data.......................... 78 Table 4.9: Institutional Dependence, Contracting andProperty Rights Institutions Across Industries: Regressions Using PICSDataset............................................................................................... 79 Table 4.10: Institutional Dependence and Ways to Get Around Weak contracting Institutions Across Industries: RegressionsUsing PICS Data ................................................................................. 79 Table 5.1: Population, GDP, and Exports 2003 ............................................................................................ 85 85 Table 5.3: Percentage Share inWorld Totals................................................................................................ Table 5.2: Growth o f GDP ............................................................................................................................ 86 Table 5.4: Various Indicators of Trade Structure .......................................................................................... 86 Table 5.5: Knowledge Economy Indicator and Components: Changes 1995 to Most Recent...................... 91 Table 5.6: Various Proxy Indicators ofTraditiona1 Sector............................................................................ 99 Table 5.7: Knowledge Failures inthe Use o fKnowledge and Corrective Actions ..................................... 104 Table 6.1: Rates o f Growth o f Energy Intensity 1986-88 to 1996-98 (percent per annum)....................... 108 Table 7.2: Exports and Imports o f Goods and Services as Proportion ofthe GDP..................................... Table 7.1: Annual Growth Rates o f GDP.................................................................................................... 130 131 Table 7.3: Direction o fExports by Major Destinations (Percent) ............................................................... Table 7.4: Direction o fImports by Major Origins (Percent) ....................................................................... 135 135 Table 7.5: Population, GDP, andTrade inSAARC-5 (2001) ..................................................................... 136 Table 7.6: Informal Trade Between India and its SAARC Partners............................................................ 137 Table 8.1: Percentage o f Export Goods Trade According to Commodity Groups Based on Value (2-Digit HSCode)................................................................................................................................. 145 Table 8.3: ExportTrade by Volume, inPercent (2003) .............................................................................. Table 8.2: Major NonBulk Export ............................................................................................................. 145 147 Table 8.4: Relationship of Logistics andProducts for the Apparel Industry............................................... Table 8.5: Principle Corridors..................................................................................................................... 150 152 Table 8.6: General Performance for Main Corridors .................................................................................. 152 154 Table 8.8: Estimated Truck Operating Costs .............................................................................................. Table 8.7: Extent o fDivestiture Transport Services ................................................................................... 155 Table 8.9: Railroad Performance (FY03) .................................................................................................... 156 Table 8.10: Container Statistics 1990-2004 ................................................................................................ 158 Table 8.11: Transshipment Share for Indian Container Traffic FY04 ........................................................ 159 Table 8.13: Tariffs to Europe US$/TEU. 2005 ........................................................................................... Table 8.12: Typical Shipping Times ........................................................................................................... 159 159 Table 8.15: Regulatory Functions ............................................................................................................... Table 8.14: Major Gateways ....................................................................................................................... 160 160 Table 8.16: Performance Statistics for Petrapole/Benapole (Hours) ........................................................... 161 Table 8.17: Drafts inRegional Ports ........................................................................................................... 162 Table 8.19: Additional Taxes at Border ...................................................................................................... Table 8.18: Standard Tariff Bands .............................................................................................................. 165 165 Table 8.20: Customs Reforms..................................................................................................................... 166 Table 8.21: DocumentsRequired for Clearance o f Goods Relatedto the Garment Trade.......................... 166 Table 8.22: Pakistan Supporting Documents .............................................................................................. 167 Table 8.23: Trade andTransit Agreements ................................................................................................. 168 Table 8.25: Priority Initiatives for Improving Regional Logistics .............................................................. Table 8.24: Signatories to U.N.Conventions Relatedto Transit Cargo...................................................... 169 175 Table 8.26: Specific Benefits andRisks Associated with the Initiatives..................................................... 175 Table 9.1:Trade Gains (US$ Million) FromCapacity Buildingby Each o f South Asian Countries and Entire South Asia Region inTrade Facilitation ...................................................................... 198 Table 9.2: Trade Gains (US$ Million) FromUnilateral and Collective Capacity Buildings between South Asia and the Rest o f the World ............................................................................................... 199 FIGURES Figure 1.1:SouthAsia is the Least IntegratedRegion inthe World............................................................... 2 Figure 1.2: Share o f South AsianFirms Reportingthe Issues as a "Major" or "Severe" Constraint on the Figure 1.3: Cost o fDoingBusiness is HighinSouthAsia ............................................................................. Operation o fTheir Business........................................................................................................ 8 9 Figure 1.4: Rule o f Law and Per Capita Incomes.......................................................................................... 10 Figure 1.5: Contracting and Property Rights Institutions Across Cities: Bangladesh, India, Pakistan, and Sri Figure 1.6: Overall Knowledge Economy Indexfor South Asia: 1995 Versus the Most Recent..................14 Figure 1.7: Comparison o fKEIComponent Parts for World Regions with South Asian Countries (Most 15 Figure 1.8: SouthAsia on Education............................................................................................................. Recent inTop Line. Comparedto 1995 BottomLine for Each Group) .................................... 16 Figure 1.9: SouthAsia on Innovation........................................................................................................... 17 21 Figure 1.11:SouthAsia Has Done Well inService Exports......................................................................... Figure 1.10: Electricity Losses inSouthAsia and China .............................................................................. 23 Figure 1.12: Remittances Inflow and Skill Composition o f Labor Migration............................................... 24 Figure 1.13: SouthAsia Has Reduced Import Tariffs: .................................................................................. 25 Figure 1.14: Cost o f Trading Across Borders for South Asia: An International Comparison....................... Figure 1.15: Cost o f Trading Across Borders for South Asian Countries ..................................................... 26 26 Figure 1.16: Who Gains From SAFTA? ....................................................................................................... 28 Figure 1.17: SouthAsia is Weak on Trade Facilitation................................................................................. 30 Figure 2.1: Output Per Worker And Its Components SouthAsia, 1960-2003 .............................................. 38 Figure 2.2: Comparison o f Investment Share and Change inthe Capital Stock, 1980-2003 (all countries Figure 2.3: Average Schooling in 107 countries ........................................................................................... 43 except IND,LKA, PAK, BGDthrough 2000) .......................................................................... 42 Figure 3.1:Share o fFirms Reporting the Issues as a "Major" or "Severe" Constraint on the Operation o f Figure 3.2: Average Time to Clear Customs................................................................................................. Their Business ........................................................................................................................... 51 52 Figure 3.3: Extent o fVariation to Clear Customs ......................................................................................... 53 Figure 3.4: Evidence from India That Poor Transportation Raises Inventory Costs ..................................... 53 Figure 3.5: Share ofFirms Exporting: Responseto a One Standard DeviationImprovement in Transportation. Access to Electricity. and Finance................................................................... 54 Figure 3.6: Highest Firing Costs ................................................................................................................... 54 Figure 3.7: Time to Enforce a Contract......................................................................................................... 55 Figure 3.8: A Weak Investment Climate Disproportionately Hurts Small Firm........................................... 56 65 Figure 4.2: Property Rights and Contracting Institutions inSouth Asia ....................................................... Figure 4.1 :Rule of Law and Per Capita Incomes.......................................................................................... 68 Figure 4.3: Contracting and Property Rights Institutions Across Countries: Macro Versus WBES Measures ................................................................................................................................................... 72 Figure 4.4: Contracting and Property Rights Institutions Across Countries: Macro Versus PICS Measures ................................................................................................................................................... 72 Figure 4.5: Contracting and Property Rights Institutions Across Cities: ...................................................... 76 Figure 5.1: Overall KEI 1995 Versus Most Recent....................................................................................... 89 Figure 5.2: ComparisonofKEIComponent Parts for World Regions With SouthAsian Countries (Most 90 Figure 5.3: Education Scorecards for South Asian Countries ....................................................................... Recent inTop Line, Compared to 1995 Bottom Line for Each Group) .................................... 94 Figure 5.4: Innovation Scorecards for South Asian Countries ...................................................................... 96 Figure 6.1:Comparisons o f Electricity Intensity o f Southeast Asia, the UnitedSates. China, and the European Union-15 ................................................................................................................. 108 Figure 6.2: Electricity Consumption inSouth Asia Comparedto PredictedperUS$ Thousand PPP .........109 110 Figure 6.4: Load Shedding inMaharashtra ................................................................................................. Figure 6.3: Electricity Losses inSouthAsia and China .............................................................................. 111 Figure 7.1: Measuring GDP at Preliberalization Domestic Prices Understates Growth ............................. 132 134 Figure 8.1: Export Share FY2005 .............................................................................................................. Figure 7.2: South Asia's Intraregional Trade as a Share o f Total Trade, 1948-99 ..................................... 144 Figure 8.2: India Export Trade byValue 2003............................................................................................ 146 146 Figure 8.4: Maldives Export Trade by Value 2003 ..................................................................................... Figure 8.3: Sri Lanka Export Trade byValue 2004 .................................................................................... 146 Figure 8.5: Interregional Trade as a Percentageo fTotal Trade by Value................................................... 147 Figure 8.7: Map o fMajor Corridors............................................................................................................ Figure 8.6: Exports o f Fabric and Garments, 2003 ..................................................................................... 148 151 Figure 8.8: Relationship o f Transport Time and Cost ................................................................................. Figure 8.9: N'hava Sheva Container Traffic ................................................................................................ 162 153 Figure 8.10: Traffic inKarachi and Qasim ................................................................................................. 162 Figure 8.11: Chittagong Container TraMic .................................................................................................. 163 163 Figure 8.13: Clearance Times for All Imports ............................................................................................ Figure 8.12: GrowthinColombo Container Traffic.................................................................................... 167 Figure 8.15: Shrimp Export Supply Chain.................................................................................................. 171 Figure 8.14: Countries Adopting the TIR Convention ................................................................................ 170 Figure 8.16: Timeline for Garment Productions ......................................................................................... 172 Figure 9.1: Exports from Developing Countries, 1980-2002 ..................................................................... 181 Figure 9.2: Intraregional Trade as a Share inRegion's Total Trade ........................................................... 181 Figure 9.4: Port Efficiency Indicators (Maritime andAir) .......................................................................... Figure 9.3: Actual to PredictedRatios for Intraregional and InterregionalTrade....................................... 182 188 Figure 9.5: Average Days Required for Customs Clearance by Sea........................................................... 191 Figure 9.6: Technical Regulations and Standards: Percent o fFirmsRanking Regulations Important to 193 194 Figure 9.8: Internet Users per 10,000 Inhabitants in2003 .......................................................................... Figure 9.7: Internet Hostsper 10,000Inhabitants in2003 .......................................................................... Export Expansion.................................................................................................................... 194 Figure 9.9: Trade Facilitation Indicators..................................................................................................... 196 Figure 9.10: Trade Gains (Percentage) FromCollective Capacity.............................................................. 198 Figure 9.11:Trade Gains to South Asia (Percent) FromIts Own (Unilateral) Capacity Buildingand Capacity Buildingbythe Rest ofthe World ........................................................................... 200 PREFACE South Asia has experienced remarkable growth rate since the 1980s. This growth was triggered by reforms aimed at increasing global integration, improving macroeconomic management, and unleashing the private sector as the engine o f growth. Highgrowth rate, inturn, has created the interest and allowed political space for greater regional integration. This raises two issues: Is South Asia's highgrowth sustainable? I s regional integrationdesirable? This report looks at several aspects o f South Asia's growth and how regional integration can contribute to growth. The two themes are not tightly linked ina causality sense. Rather, the themes emerged as an outcome o f a knowledge partnership between the World Bank and the South Asian Association for Regional Cooperation (SAARC) Chamber of Commerce and Industry (SCCI), which i s the apex business organization o f SAARC. The SCCI i s composed o f the national chambers o f commerce and industry in Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, and Sri Lanka. This partnership resulted inthe first SAARC Business Conclave, held in November 2005 inNew Delhi. The conference was organized by SCCI and was implementedby the Federation o f Indian Chambers o f Commerce and Industry and brought together the private sector, academics, civil society, and policy makers from the South Asian countries to discuss growth and regional integration. Several authors were invited to prepare background papers for this conference on issues o f growth and regional integration. The report is essentially a compilation o fthese backgroundpapers. Past growth was helpedby the implementation o f first-generation policy reforms aimed at global integration, macroeconomic stabilization, and reducing the scope o f the state while strengthening the role o f the private sector. These reforms have made South Asia more competitive, stable, and adaptable. South Asia now faces increasing challenges from second- generation reforms. These include the high cost of doing business, weak institutions, weak knowledge economy, and poor infrastructure. Increasing investment rates will require reducing costs o f doing business, improving institutions, and addressingthe infrastructure constraint. South Asia also needs to expand its knowledge economy to raise the productivity o f investment. These second-generation policy reforms have started; continued strong implementation will support the sustainability o f growth. Is regonal integration desirable? From the very narrow perspective o f trade flows, the economic characteristics o f the South Asia region, such as the small regional market relative to the world both interms o f gross domestic product and trade flows and the highlevel o fprotection, would suggest that focusing on regional integration alone will not generate the beneficial productivity and growth effects o f integration. Instead, further multilateral trade liberalization i s the way to go. Nevertheless, regional integration is desirable from other perspectives. Regional cooperation can be an effective tool in addressing energy shortage, ensuring that no regiodcountry i s left behind, land-lockedregions/countries have full access to markets, and peace and stability are promoted. There i s a need for greater people-to-people contact through improved connectivity, phasing out o f visa restrictions, and liberalizing the restrictions on the trade of services (for example, tourism, education, and health) where the risk o f trade diversion i s low. These initiatives would help increase investment and growth by reducing the infrastructure constraint and by lowering transaction costs. Better regional cooperation and integration can also increase welfare by improving the regional political environment, thereby reducing conflicts and associated social andeconomic costs. REPORTTEAMAND ACKNOWLEDGEMENTS This report was prepared by a team consisting o f Sadiq Ahmed, Susan Collins, Carl J. Dahlman, Ana Margarida Fernandes, Ejaz Ghani, Mary C. Hallward-Driemeier, Aart C. Kraay, DavidNewberry, Tsunehiro Otsuki, Arvlnd Panagariya,John S. Wilson, and John H.Arnold. We would like to acknowledge the contribution o f the conference participants and thank several colleagues for their cornmentshggestions, including M. K. Saharia, W. Ahmed, R. Kohli, A. Sarkar, G. Ghosh, A. McKechnie, S. Devarajan, H.Kharas, A. Winters, F. Sobhan, A. R. Kemal, C. Harris, S. A. Mahmood, R. Agarwala, N.Kumar, T. Baysan, A. Subramanian, and A. Mattoo. Finally, we would like to thank B. Kalimili for research support, and to R. Soni, J. Teodosio, A. Reyes and M.Schwartz for providingvery competent team support. 1. SOUTHASIA'S GROWTHAND REGIONALINTEGRATION:AN OVERVIEW INTRODUCTION 1.1 South Asia has experienced unprecedented growth, averaging close to 6 percent per annum since the 1990s. This growth is impressive because many developing countries grew more slowly during this period. As India accounts for more than three quarters of the region's gross domestic product (GDP), its growth has had a decisive impact on the overall regional growth. India grew at 3.2 percent during 1965-81 ,acceleratedto 5.1 percent during 1981-7, and then to 6 percent during 1987-2004. While India has led the way, the other South Asian countries including Bangladesh and Pakistan have also shown remarkable improvements in economic growth (Ahmed 2006). It i s this steady rise in growth that has attracted the world's attention to the SouthAsia region. 1.2 Growth in South Asia was triggered by first-generation policy reforms, includinggreater global integration, macroeconomic stabilization, and economic deregulation (Ahmed 2006). Trade restrictions including import tariffs were reduced. The scope of the state was reduced through economic deregulation to enhance the role of the private sector as the engine of growth. These reforms made the South Asian countries more stable, competitive, and adaptable. It helped to overcome three key constraints to growth (Spence 2005): Exports to the rest of the world relaxed the constraint from capacity to consume domestically. Inflow o fremittancesrelaxedthe constraint from capacity to save domestically. In-bound technology and knowledge transfers through increased trade moved the productionpossibility frontier outward rapidly. 1.3 While South Asia made significant progress in integrating with the global economy, integration within the region remained limited. South Asian countries have maintained a higher level o fprotection within the region than with the rest of the world. Restrictive policies within the region have neutralizedthe beneficial effects o f common cultural affinity, common geography,1 and the "gravitational" pullo fproximity on movement o f goods and people withinthe region. 1.4 South Asia is the least integrated region in the world, where integration is measured by intraregional trade in goods, capital, and ideas. Intraregional trade as a share of total trade is the lowest for South Asia. There i s little cross-border investment within South Asia. The flow o f ideas, crudely measured by the cross-border movement of people, or the number of telephone calls, or the purchase of technology and royalty payments, are all low for South Asia. InSouth Asia, only 7 percent of internationaltelephone calls are regional, compared to 71 percent for East Asia. Poor connectivity, cross-border conflicts, and concerns about security, have all contributed to South Asia beingthe least integrated region inthe world (Figure 1.1) 'Indiahas a common(land) border with most SouthAsian counties inthe region. Figure 1.1: South Asia is the Least Integrated Region in the World Intraregional Trade (Percent of World Trade) 60.0 - 55.0 - -)--.I 50.0 - 45.0 - 40.0 - / 35.0 - 30.0 - 25.0 - -4.9 LAC 15.0 - 10.0 - I ' Sources: U.N.COMTRADE, SITC 1 classification and GDP from World Development Indicators. S A R - South Asia Region,E M - East Asia & Pacific, SSA - Sub-SaharanAfrica, LAC - Latin American & Caribbean, MENA - Middle East &NorthAfrica. 1.5 The rapid growth that South Asia has experienced, however, has generated interest in, and political support for, increased regional integration. On January 6, 2004, the South Asian countries signed a South Asia Free Trade Agreement (SAFTA). The prospect that rapid growth will facilitate regional integrationraises two fundamental questions. I s rapid growth sustainable inSouthAsia? Ifso, what role canregional integrationplay? 1.6 I s growth sustainable? South Asia has made significant progress in implementing the first-generation policy reforms. Increasingly, South Asia faces the challenge o f second-generation policy reforms, which have become the key downside risks to growth. These include: 0 Highcost o f doing business 0 Weak institutions Weak knowledge economy 0 Weak infrastructure 1.7 Despite past progress, the cost o f doing business in South Asia remains high. Corruption and energy supply have been identified as the two biggest problems faced by the firms in South Asia. The problem o f infrastructure deficit was ignored in the past because o f political reasons (opposition to increasing costs when services were almost once free) and economic reasons (many consumers doubt that increasing charges will improve services, gwen decades of corrupt public administration). The competitiveness o f South Asia i s held back by inadequate education and skills. Although South Asia, and in particular India, has made a name for itself in high-skill services and highs h l l manufacturing (Kochar et. a1 2005) South Asia lags on most indicators of knowledge economy (for example, gross enrollment rates for secondary education was 49 percent for South Asia compared to 69 percent for East Asia in2004). 1.8 The policy makers in South Asia have begun to address the second-generation policy reforms (Ahmed 2006). They have deepened attention to microeconomic policies aimed at 2 reducing the cost of doing business and improving competitiveness. There is increased attention to improving institutions and governance by expanding market-based allocation o f resources, improving transparency through better disclosure and market discipline, reforming regulation to minimize corruption and opportunities for rent seeking, and seeking opportunities to introducing better checks and balances to improve governance. Policy reforms are being implemented to encourage private and public partnerships in the provision o f infrastructure services. These reforms have, however, started recently and they need to be deepened and sustained. Political opposition is a challenge that needs to be met with determination, especially in regards to establishing proper institutional arrangements required for sustaining these second generation reforms. With the full implementation of the second-generation policy reforms, South Asia should be able to sustain highgrowth rates inthe future. 1.9 I s regional integration desirable? Regional integration creates both challenges and opportunities. The challenge i s to minimize the possible harmful effects of regional integration. The economic characteristics of the South Asia region, such as the small regional market relative to the world both in terms of GDP and trade flows, and the high level of external protection, would suggest that focusing on regional integration alone may not generate the beneficial productivity and growth effects in South Asia. South Asia i s a relative new comer to global integration. Despite liberalization, it lags other regions (for example, East Asia) on openness. When external protection i s high, trade diversion i s likely to dominate trade creation, and so the risks that regional integration will become a drag on growth in South Asia i s high.There are three reasons why South Asia will need to further lower external trade bamers: to generate classical gains from trade, to lessen the chances that trade diversion will occur, and to reduce income transfers between member countries resulting from regional integration and the tensions that can arise from such transfers (Hoekman and Schiff 2002). History shows that a successful regonal integrationi s often preceded by global integration. 1.10 But regional integration also provides opportunities to make progress in areas that otherwise would not take place in the absence o f regional cooperation. Some of these opportunities include addressing the problems o f energy shortage, relaxing the mobility constraints for lagging and land-locked regions, overcoming high transaction cost due to poor trade facilitation across regions, and reaping the positive benefits emerging from reputation effects/political risk premiudpeace dividendthrough regional cooperation. The gains from these opportunities can contribute to higher sustainedgrowth. Importantly, better economic cooperation can lead to better political relations thereby reducingconflicts and associatedsocial and economic costs. 1.11 Regional cooperation can play an important role in addressing the problem of energy needs inthe region. Energyendowments differ among the South Asian countries, but energy trade in the region is low. Only India, Bhutan, and Nepal currently trade electricity. Bangladesh is endowed with natural gas reserves, but gas trade i s constrained by the region's inadequate infrastructure and political misconceptions. Pakistan and Afghanistan can play an important role as transit states for the rest of South Asia, as they provide the best route for access to Central Asia's energy. 1.12 Regional cooperation, along with national initiatives, could play a useful role inensuring that no regiodcountry in South Asia i s left behind. Rising inequality across regions and within countries i s becoming a concern to the policy makers as rising inequality i s a threat to the region's growth and stability. Several lagging regions in South Asia are border economies. They suffer fiom the disabilities typically associated with land-locked countries or geographical isolation. Examples include northeast India, northwest Pakistan, northern Bangladesh, and parts 3 of Nepal and Afghanistan. Typically, these sub-regions have poor connectivity with the markets within the country and with the neighboring countries. Regional cooperation on transport and trade facilitation can transformthese land-lockedregions into land-linked regions. There are other areas where the region can benefit through cooperation. These include tourism, education, health, and professional services, where the risk of trade diversion is low. 1.13 In conclusion, provided external protection is further reduced globally, regional integration could play a beneficial role in allowing the South Asian countries to gain from geographical proximity, improved transport and trade facilitation, improved management of cross-border resources (for example, energy and water), and reduced conflict and friction. 1.14 Against the backdrop of the above, this book looks at several aspects of South Asia's growth and how regional integration can contribute to growth. The two themes are not tightly linkedina causality sense. Rather, the themes emerged as an outcome of aknowledgepartnership between the South Asia Chamber o f Commerce; national chambers of commerce in Bhutan, Bangladesh, India, Maldives, Nepal, Pakistan, and Sri Lanka; and the World Bank. This partnership resulted in the first South Asian Association for Regional Cooperation (SAARC) Business Conclave that was held in November 2005 in New Delhi. This conference was organized by the Federation o f Indian Chamber o f Commerce and Industryand SAARC Chamber of Commerce and brought together the private sector, academics, civil society, and policy makers from all South Asian countries to discuss growth and regional integration. Several authors were invited to prepare background papers for this conference on issues of growth and regional integration. The book i s essentially an edited compilation ofthese back groundpapers. 1.15 Thisbookhasthree parts: 0 Part 1,Growth and RegionalIntegration, comprises this overview. 0 Part 2, I s Growth Sustainable, includes papers on growth accounting by Susan Collins, investment climate by Mary Hallward-Driemeier, institutions by Ana Margarida Fernandes and Aart Kraay, knowledge economy by Carl J. Dahlman, and energy infrastructure by DavidNewbery,. 0 Part 3, I s Regional Integration Desirable, includes papers on trading choices by Arvind Panagariya, trade facilitation and exports by John Arnold, and trade facilitation and regional integrationby John S. Wilson and Tsunehiro Ostuki. 1.16 This overview chapter summarizes some o f the key issues on growth and integration drawing on the analysis of the papers inthis volume. While there i s a broad consensus on the key challenges that South Asia faces to sustain growth, there are differences inviews on the role that regional integration should play in South Asia. The differences in views nevertheless enrich the evolving thinking on the subject as well as the political dynamics by bringing different perspectives from researchers, private sector, policy makers, and the civil society on the desirability o fregional integrationinSouth Asia. I s GROWTH SUSTAINABLE 1.17 What are the sources o f growth in South Asia? What will it take for South Asia to increase growth rates from 6 to 10 percent per annum? What role will business climate, institutions, knowledge economy, and infrastructureplay insustaining growth? 4 Sources of Growth 1.18 South Asian economies have achieved impressive rates of economic growth since the 1980s. Chapter 2, on economic growth in South Asia by Susan Collins, explains that output for India, Pakistan, Bangladesh, and Sri Lanka has grown more rapidly since 1980than for any other region except East Asia. Table 1.1 shows the key features o f South Asia's growth. During the period 1980-2000, India and Bangladesh increased their GDP growth rates relative to the rates they had sustained in the two decades prior to 1980. Sri Lanka maintained a steady pace throughout from the decade o f the 1960s. Pakistan maintained rapid growth until the 1980s but growth faltered in the 1990s. South Asia, as a regon, has increased the growth rate by nearly 2 percentage points since the 1980s. Growthrates o f these magnitudes are impressive achievements that have helped South Asia to reduce poverty rates and raise living standards. Table 1.1: SouthAsia: SelectedIndicators GNIPer Capita Population Annual Ratesof Change Investment (PPP)a (Mil1ions)a GDP Labor Force Sharepercent India 1965-80 3.2 2.2 17.5 1980-04 US$3,116 1,080 5.8 1.8 22.6 Bangladesh 1973-80 3.7 2.1 10.1 1980-04 US$1,969 139 4.3 2.3 19.0 Pakistan 1965-80 5.8 2.7 16.6 1980-04 US$2,174 152 5.2 2.9 18.4 Sri Lanka 1965-80 4.7 2.2 19.0 1980-04 US$4,208 19 4.6 1.5 25.1 SouthAsia 1965-80 3.6 16.6 1980-04 US$2,854 1,447 5.5 1.9 21.9 East Asia Pacific 1965-80 6.7 26.4 1980-04 US$5,332 1870 8.0 2.0 33.5 Source: World Bank,2006. a. These data are for 2004. Note:PPP, purchasingpower parity, GNI, gross national income. 1.19 Table 1.2 reports the growthdecompositionfor the South Asian countries. At 3.3 percent per annum since 1980, growth in output per worker in South Asia has been well above the world average, rivaling East Asia's experience (3.9 percent). Total factor productivity (TFP)-a measure o f efficiency with which resources are used-surged for South Asia in the 1980s, after two decades o f little growth. TFP contributed twice as much to growth in South Asia as in East Asia, while increases in capital per worker contributed just half as much. However, increases in education contributed far less to growth in South Asia compared to East Asia. Increases in education among the South Asian economies have not been impressive; that is, average years o f schooling remains quite low in South Asia. 5 Table 1.2: Sourcesof Growth, SouthAsia, 1960-2003 Contribution of Output Labor Outputper Physical Factor Regionlperiod force worker Capital Education productivity India 1960-70 4.06 1.94 2.08 1.05 0.30 0.72 1970-80 3.OO 2.26 0.72 0.63 0.36 -0.27 1980-90 5.79 1.87 3.85 1.02 0.33 2.46 1990-2003 5.59 2.11 3.41 1.33 0.47 1.57 1960-2003 4.67 2.05 2.57 1.03 0.37 1.15 Bangladesh 1970-80 1.53 2.46 -0.91 -0.23 0.30 -0.97 1980-90 4.75 2.55 2.15 0.47 0.18 1.49 1990-2003 4.84 2.17 2.62 1.10 0.34 1.16 1960-2003 3.84 2.25 1.55 0.64 0.24 0.66 Pakistan 1960-70 7.22 2.44 4.67 3.75 0.35 0.54 1970-80 4.68 3.04 1.59 0.68 0.23 0.68 1980-90 6.28 2.63 3.55 0.98 0.92 1.61 1990-2003 3.82 2.71 1.08 0.46 0.04 0.57 1960-2003 5.37 2.71 2.60 1.39 0.36 0.83 Sri Lanka 1960-70 4.57 2.36 2.16 0.02 0.33 1.80 1970-80 4.40 2.10 2.25 1.51 0.39 0.33 1980-90 4.19 2.04 2.11 2.04 0.22 -0.15 1990-2003 4.64 2.05 2.53 1.10 0.34 1.08 1960-2003 4.46 2.13 2.28 1.16 0.32 0.78 SouthAsia 1960-80 3.8 - 1.5 1.o 0.3 0.3 1980-2003 5.5 - 3.3 1.1 0.4 1.8 1960-2003 4.7 - 2.5 1.o 0.4 1.1 EastAsia less China 1960-80 7.0 - 4.0 2.2 0.5 1.2 1980-2000 6.4 - 3.9 2.4 0.5 0.9 1960-2000 6.7 - 3.9 2.3 0.5 1.o Source: Susan Collins (see chapter 2). 1.20 Where will South Asia need to focus to sustain growth in the future? The sources o f growth analysis inTable 1.2 show that both capital accumulation and productivity growth played important roles. Investment rates grew substantially from a l o w average rate o f about 12 percent per year in the 1970s to 23 percent in the decade o f 2000s (Ahmed 2006). Similarly the contribution o f TFP growth increased significantly inthe post 1980periodrelative to the 1960-80 period. As explained in detail in Ahmed (2006), these developments resulted from the markedly improved overall policy environment in South Asia. Since both capital accumulation and productivity matter, a prudent policy i s to foster both. For the future, much of the potential productivity gains for South Asia will come from the reallocation of labor from agnculture, where productivity is relatively low, to the rest o f the economy (manufacturing and services), where it i s considerably higher. However, for South Asia to successfully shift underemployed workers out of agriculture into higher productivity activities in the manufacturing and services 6 sectors, it will need to make investments to increase both physical and human capital stocks. Without the appropriate physical infrastructure, and human skills, the structural transformation of South Asia will be hampered. 1.21 What will it take for South Asia to boost its growth rate from 6 to 10percent? Inaddition to improving productivity, investment rates will need to rise substantially. Output growth rates o f about 6 percent per annum in South Asia are consistent with maintaining investment rates o f 23- 24 percent of GDP, the average inrecent years. However, increasing the region's growth rate to 10 percent will require increasing the investment rate to more than 35 percent. As explained below, the infrastructure gap in South Asia i s large, and meeting this gap will require substantial additional investment. Higher investment could be achieved by reducing fiscal deficits, by lowering the cost o f doing business, and by improving institutions. Increased physical capital stock will need to be accompanied by human capital stock. The average year o f schooling for South Asia is low at 5 years compared to nearly 15 years for Korea. Increasing human capital stock will require increased emphasis on the knowledge economy. Cost of Doing Business 1.22 Increasing investment rates will require improving the investment climate and reducing the cost o f doing business. H o w does South Asia compare with other regions on the cost o f doing business and investment climate? Chapter 3, on investment climate by Mary Hallward-Dnemeier, uses cross-country data on the cost o f doingbusiness and firm level surveys that were carried out inBangladesh, India, Pakistan, and Sri Lankato assessthe investment climate. Firmlevel surveys on the investment climate draw data directly from f m s and cover both objective and subjective indicators. They cover eight indicators that influence investment decisions, from policy uncertainty and corruption to reliability of electricity and labor regulations. 1.23 Figure 1.2 shows a comparison o f the investment climate in South Asia with East Asia and Central Europe. Access to infrastructure has been cited as the number one problem faced by the firms in South Asia. The key concern for most firms is the access to reliable electricity. Corruption figures as the second biggest problem facing the firms in South Asia. When officials have discretion in how particular regulations will be implemented; there i s an opportunity for unofficial payments to determine the outcomes. The costs o f these bribes and the uncertainty of the standards that will be imposedcan reduce the incentive to invest or to expand. Across South Asia the size o f bribes was reported to be between 2.2 and 2.5 percent of firm sales (See Figure 1.2). 7 Figure 1.2: Shareof South Asian FirmsReportingthe Issues as a "Major" or "Severe" Constraint on the Operation of Their Business Electricity -/??=- Telecom/transport< /I \ Corruption Informality - -Eastern Europe& CentralAsia ~ -SouthAsia Asia East and Pacific Source: Investment Climate Surveys. (See Chapter 3) 1.24 Figure 1.3 compares the cost of doing business in South Asia with East Asia and Organisation for Economic Co-operation and Development (OECD) countries. Despite substantial deregulation, regulatory burden, particularly tax, customs, and labor regulation, remain high in South Asia. Access to finance has improved but credit information and secured lendingsystems lag other regions. Time to enforce a contract, or register a property, i s high in South Asia relative to East Asia. Policies combined with institutions influence the investment climate and cost of doing business. Institutions 1.25 It is increasingly recognized that weak institutions are a dragon growth and development (World Bank 2001). The cost o f poor institution and governance i s largely borne by the poor (World Bank 2006a). Inthe literature, five types of institutions have been identifiedas important for growth and development (Rodrick 2002; Knack 2006): 0 Market-creatinginstitutions (property rights, rule of law) 0 Market regulating institutions (financial services, telecom, transport) 0 Market stabilizing institutions (central bank, budget rules) 0 Social insurance or market-substituting institutions (provision of education, health) 0 Institutions for conflict management or market legitimizing institutions (democracy, equity, justice) 8 Figure 1.3: Cost of DoingBusinessis Highin SouthAsia Time and Cost to Register Property Time and Costto Enforce Contracts High Income Highlnwme OECD OECD East Asia EastAsia South Asia 6.1 SoumAsia 0 20 40 60 80 0 100 200 300 400 500 I Days 0PercentageofPropertyValue 1 ~~~~ ~ Difficulty of Hiring and Firing Workers Price and Cost to Close a Business HighIncome CECD 40 EastAsa 20 0 SouthAsa SouthAsia HighIncome East Asia OECD 0 20 40 60 80 100 HHiringIndex OFiringindex .Years for Insolvency OCents on $ Not Recovered Source: World Bank (2005). 1.26 What i s the relationship between institutions and growth? How should institutions be measured?Can institutions be unbundledto identify those that are more critical for growth? How does South Asia compare on institutions with other parts o f the world? Chapter 4, on growth and institutions in South Asia by Ana Margarida Fernandes and Aart Kraay, is an attempt to investigate the links between institutions and growth in the context of South Asia. It i s worth emphasizing that institutions are difficult to measure, and the link between institutions and growth i s complex. Although cross-country data are available to assess how South Asia i s doing relative to other regions, time series data are not available to judge whether institutions in South Asia have improved or deteriorated over time. 1.27 Several researches have produced evidence to suggest that there i s a strong positive correlation between measures of institutional quality and log-levels of per capita income. Figure 1.4 shows one such typical relationship. On the horizontal axis i s a widely used measure o f rule of law produced by KaufmaM, Kraay, and Mastruzzi (2005) who combine information from a large number of cross-country sources measuring perceptions of governance and construct composite indicators summarizing these perceptions. The rule of law measure in particular captures the perceptions of individuals, firms, commercial risk rating agencies, nongovernmental organizations, think tanks, and multilateral development banks on issues relating to the protection of property. For example, it captures perceptions o f the likelihood that property will be expropriated by the state, the likelihood that contracts will be enforced, the likelihood that property is secure fi-om crime, and so on. 9 Figure 1.4: RuleofLaw and Per CapitaIncomes 10.5 - - -2.5 -1.5 -0.5 0.5 1.5 2.5 Rule of Law Note: Real GDP per capita refers to 1996 and rule o f law to 2004. Higher values o f the rule o f law measure correspond to worse outcomes. (See Chapter 4). 1.28 Figure 1.4 shows a strong negative correlation between this broad measure of institutional quality and levels o f development: Countries with worse institutional quality are on average poorer than countries with good institutional quality. A further striking feature is the relative position o f countries inthe South Asia region. All countries inthe South Asia region fall below the regression line. Bangladesh, Pakistan, and Nepal have fairly poor scores on this measure, which place them in the bottom quartile o f all countries, while India, Sri Lanka, and Bhutan fare much better, around the median o f all countries worldwide. 1.29 However, interpreting the relationship between growth and rule of law i s not easy. The evidence provided by the above research has been controversial. Jn particular, the research has not been able to reconcile the very low rating on the rule of law for South Asia with the regions sustained rapid growth, among the highest inthe world. Ingeneral, the low rating on governance as a whole for South Asia emerging from the Kaufmann, Kraay, and Matsurazzi research in the face o f the region's rapid sustained growth has generated considerable debate. The debate centers on two issues: First, are we measuring the right things when ranking governance performance? And second, is there necessarily a causal relationshipbetween institutions and growth and which way does this causality run? 1.30 The measurement issue has led to research seeking to unbundled governance and institutions. Acemoglu and Johnson (2005) unbundled institutions into two distinct dimensions, and identified them as important determinants o f growth. n e first o fthese, which they refer to as "property rights institutions," capture the extent to which private property i s secure from predation by the state (for example, through outright expropriation or, less dramatically, from corrupt officials demanding bribes in exchange for favors to the firm or individual). The second, which they refer to as "contracting institutions," captures how the effectiveness o f institutionsthat are usedto mediate disputes between private parties, such as the courts andthejudicial system. 10 1.31 Using the above framework, Fernandez and Kraay measure the property rights and contracting institutions for the South Asian countries usingmacro and micro data. They measure property rights institutions using the Kaufmann, Kraay, and Mastruzzi (2005) measure o f corruption. While perceptions o f corruption are clearly not institutions themselves, the prevalence o f corruption i s considered a good proxy for the absence o f well-functioning institutions that prevent the arbitrary or abusive exercise o f authority. They measure contracting institutions using an estimate o f the lengtho f time required to resolve a dispute over an unpaid commercial debt, constructing by using the World Bank's annual Doing Business report. 1.32 They find great diversity on institutional performance within South Asia. Bangladesh, Sri Lanka, and Palustan have better contracting institutions compared to India, but India has better property rights institutions. They find that property right institutions are more critical, relative to contracting institutions, to growth in South Asia. 1.33 The poor performance o f countries such as Bangladesh on corruption based on the views of firms in these countries suggests that this is an area where reforms, although difficult, are likely to have substantial impact. The evidence for India points to deficiencies in contracting institutions, which can be interpreted more broadly as failures in the overall regulatory environment. While cross-country evidence suggests that institutional weaknesses in this dimension have smaller development impacts than do property rights institutions, this does not mean that there are no retums to improvements inthis area. Infact, firm-level evidence suggests that firms needto develop alternative strategies to circumvent weak contracting institutions, and it i s likely that these alternative strategies are inefficient compared with the benchmark o f good contracting institutions. 1.34 Do institutions vary within countries? Fernandez and Kraay investigate this using firm level data. Figure 1.5 shows on the horizontal axis the judiciary aspect of rights in business disputes as the proxy for contractinginstitutions, and it shows on the vertical axis the importance o f corruption as an obstacle to business. As indicatedby the labels, the data points in the figure represent each o f the cities in Bangladesh, India, Pakistan, and Sri Lanka. Each data point corresponds to an average at the city level o f the firm-level measures o f contracting and property rightsinstitutions. 1.35 The strihng feature o f Figure 1.5 i s that there is tremendous variability o f institutional quality within countries. For India, variability in institutional quality within the country exceeds variability across countries. India has some o f the best and worst performers on institutional performance in South Asia. The city o fKanpur inthe state o fUttar Pradesh (a laggmg regon) in 1.36 India has the worst performance on property rights in the region, while Calicut (in Kerala) and Gurgaon (in Haryana), both leading states, do well on institutional performance. The measure on institutional quality is worse according to the firm-level measures than according to the cross-country measures (Figure 1.5). 11 Figure1.5: ContractingandProperty RightsInstitutionsAcross Cities: Bangladesh,India, Pakistan, and Sri Lanka Contracting and Prop. Rights Institutionsacross Cities Bangladesh(B), India (I), Pakistan (P), and Sri Lanka (S) 0123 0118 B 1 Dhaka I16 Surat I33 Hosur P I 1 Quetta B 2 Chittagong I17 Vadodara I34 Madurai PI2 Peshawar I 1 Ahmedabad I18 Gurgaon I35 Ghaziabad S I Colombo I2 Bangalore I19 Faridabad I36 Noida S2 Gampana Shahjahanpur- I3 Calcutta I20 Panipat I37 Lakimpur S3 Kalutara Hubli- I 4 Chandigarh I21 Dhanvad I38 Howrah S4 Kandy I5 Chennai I22 Calicut I39 Mangalore S5 Matale I 6 Cochin I23 Palakkad P I Karachi S6 NuwaraEliya I7 Delhi I24 Bhopal P2 Lahore S7 Galle I 8 Hyderabad I25 Gwalior P3 Sheikhupura S8 Matara I 9 Kanpur I26 Indore P4 Sialkot S9 Kurunegala I10 Mumbai I27 Nagpur P5 Faisalabad SI0 Puttalam I11 Pune I28 Nashik P6 Gujranwala S11 Anuradhapura I12 Mysore I29 Thane P7 Wazirabad SI2 Badulla I13 Vijayawada I30 Jalandhar P8 Islamabad/Rawalpindi SI3 Monaragala I14 Lucknow I31 Ludhlana P9 sukkur SI4 Ratnapura I15 Guntur I32 Coimbatore P10 Hyderabad SI5 Kegalle Note: City averagesof the variablesjudiciary respects rights inbusinessdisputes and corruption as an obstacleto businessfrom the PICSdataare shown. 12 1.37 However, there remains the major question of how South Asia sustained high growth rates despite weak institutions. A large part o f the answer i s that South Asia has performed very well inimplementingmajor policyreforms that are good for growth ( b e d 2006). So, making a sharp distinction between good policies and good institutions i s not very helpfulbecause policies and institutions are interrelated through a time dimension. Many good policies can be implemented immediately while institutions are built over time. Sustained good policies help build institutions. Indeed, it can be argued that not all institutions are weak in South Asia (Subramanian 2006). Many of them have improved over time with the implementation of the first-generation policy reforms. First, aspects o f market creating institutions improved starting in the 1980s, as the policy makers substituted planned allocation with market-based allocation of resources. Global integration also helped to strengthen market creating institutions. Second, market stabilizing institutions improved as the policy makers strengthened economic management. The good aspects of the market stabilizinginstitutions have resultedin South Asia averaging one o f the lowest inflation rates in the world. Output variability in South Asia over the period 1960- 2000 has also been low compared to other regions. Third, institutions on conflict management have helped to avoid extreme outcomes in South Asia, such as famines (Sen 198l), or the disintegration o f states (for example, Soviet Russia, Yugoslavia, and Sudan). Some researchers have credited South Asia with achieving large growth responses with small policy changes. It i s the underlying institutions that helped magnify the effect of policy changes on growth. Governance institutions have improved as a result o f the increasedrole of the civil society, access to information, and education. 1.38 In conclusion, the first-generation policy reforms aimed at maintaining a stable macroeconomy, strengthening the role of the private sector, reducing the scope of the state, and global integrationhave contributed to growth. However, while the role o f statehas been curtailed, the effectiveness o f the state has not improved. Cost of doing business and corruption are still highinSouth Asia. The delivery ofbasic services has not improved as muchas neededin South Asia (as an example of the Indian case, see World Bank 2006b). Inequality is on the rise. The second-generation policy reforms will need to focus on improving the effectiveness o f the state. Knowledge 1.39 The generation and application o f new technology, knowledge, or ideas i s widely acknowledged to be a crucial driver of growth and competitiveness (Romer 1993). Given the rapid rate o f development and dissemination o f new knowledge globally, low capital and labor costs can no longer be viewed as the only drivers of competitiveness and growth. Although they continue to be important, new drivers of competitiveness have emerged, including the ability to rapidly redeploy resources in order to capture new opportunities; ensure the quality, skills, and flexibility o f labor force (and management); keep up with rapidly changing technological and organizational advances; move to higher value parts of value chain (researchldesign and marketing, branding, and managing of customer information); make effective use o f information technologes to reduce transactions costs; and improve capacity to respond quickly to changing opportunities and threats. 1.40 Chapter 5, on technology, slulls, and innovation by Carl J. Dahlman, compares South Asia with the rest of the world on different indicators of knowledge economy. The key indicators o f knowledge economy include: 13 Education Innovation 0 Information infrastructure 0 Institutionalregime governingknowledge 1.41 South Asia does poorly compared to other developing regions, except Afr-ica, on knowledge economy (Figure 1.6.) Within South Asia, India does the best, although it does not show improvement over time. Its higher knowledge economy index is largely due to its high index on innovation, given the large absolute size o f scientists and engineers in R&D as well as the absolutevolume of scientific andtechnical publications.2 Figure 1.6: Overall Knowledge Economy Index for South Asia: 1995 Versus the Most Recent 10 Q 8 - 7 6 c Argentina u Q) 2c. 5 0 v) I 4 T y y indonesia i a 3 2 Nepal 1 0b 1 I 2I 3I I 4 I 5I 8I 7I 8I Q 1.42 Sri Lanka i s the second highest among the South Asian countries, and it shows some improvement over the period. The biggest improvement i s in the economic incentive and institutional regime, where it gets the highest ranking among the South Asian group. It made significant improvements inthe information and communicationtechnology (ICT) indexwhere it moves from second to first. While it makes a small improvement in the innovation index, it * All the indicators in the methodologywere scaledby population. However,becauseknowledge is not consumedin its use, for the innovation variables the indicator was also computed based on absolute values, which is how it i s reportedhere. Inthe full KAMdatabase it is possibleto use the innovationvariablesnormalizedbypopulation. 14 actually loses ground in the education variable even though it still remains the highest in education among the group. 1.43 Pakistan, Bangladesh, and Nepal all lose ground in the aggregate knowledge economy index (KEI). Most notable i s the sharp fall in the economic incentive regime in Nepal which considerably pulls down its overall, average. Nepal also loses in the ICT indicator. Pakistan also loses inthe economic incentive, and inthe ICT indicator as well as inthe education indicator, and ends up with the lowest score among the group in the later. Bangladesh slips most in the innovation index and also slips in the economic incentive regime, but makes some gains in the ICT and a smaller gain inthe education index (Figure 1.7). Figure 1.7: Comparisonof KEIComponent Partsfor World Regionswith South Asian Countries (Most RecentinTop Line, Comparedto 1995 BottomLine for EachGroup) I I G7 Western Europe EastAsia Europe and Central Asia Latin America Middle East and North Africa India Sri Lanka South Asia Pakistan Africa Bangladesh Nepal Source: www.worldbank.orgkam. Note: The top bar represents the most recent aggregate KEI score for a selected region or country, split into the four KE pillars. Each color band represents the relative weight of a particular pillar to the overall country's or region's knowledge readiness, measured by the KEI. The first line for each country is its position in the most recent year for which data are available (generally 2002-3). The second line is for 1995. This figure has been computed using the un- weighted variables for the innovation index. 1.44 Education i s becoming more important because o f the increase in speed o f the creation and dissemination o f new knowledge. Education is the fundamental enabler o f the knowledge economy and a key to long-term competitiveness and growth. What is critical is no longer just basic or even secondary education, buthigher education and the constant upgradingo f skills. This i s a challenge for all countries o f the world but especially for South Asia. 1.45 The development o f a knowledge economy demands a flexible education system. It begins with basic education that provides the foundation for learning; continues with secondary and tertiary education that develops core, including technical skills; and encourages creative and critical thinkingthat i s key to problem solving and innovation, extending into a system o f lifelong learning. Such a system is one that encompasses learning from early childhood to retirement and includes formal training (schools, training institutions, and universities) and non-formal learning (on-the-job training, and skills learned from family members or people in the community). The basic elements o f such a system are comprehensiveness, new basic skdls (acting autonomously, 15 using tools interactively, and functioning in socially heterogeneous groups), multiple pathways, and multiple providers (Figure 1.8). I 1995 I 1.46 South Asian countries are in a relatively weak position in terms o f education and skills. As a group they have high illiteracy rates, low enrollment ratios at the secondary and tertiary levels, very low average educational attainment among the adult population, extremely low percentage o f professional and technical workers among the labor force, low quality of math and science education, little staff training even among firms in the modern sector, and a serious problem o f emigration of the highly slalled workers. 1.47 There is considerable variation with the South Asia region. Nepal and Bangladesh are much weaker on all the variables. Paktstan ranks somewhat better. Sri Lanka and India score much higher. Sri Lanka has the highest literacy, enrollment rates, and average educational attainment. 1.48 However, India is ranked higher in terms o f the quality o f science and math education, extent of staff training, and availability o f management education. India has the world-renowned Indian Institutes o f Technology and Indian Institutes o f Management, which produce world-class graduates. These institutes, along with many other lesser known regional colleges, have given India a critical mass o f highly skilled people. These high quality English-speaking human resources are a large part of the reason why India has been able to develop the information technology export services that have moved up from simpler back office functions and call centers to software design and innovation services. Many o f the highly slalled Indians have immigrated to the United States and Europe in search o f higher payingjobs. However, some o f this brain drain has been turned into a brain gain as they have started to outsource highly skilled services from India. 16 1.49 This strong high-slulled ICT service sector has not developed in the other South Asian countries because o f their smaller scale and less prevalence of Englishintheir education systems. However, this is a very small sector inIndia relative to its total population, and the average levels o f educational attainment in India are very low. Therefore, improving education and skills i s a challenge for all the South Asian countries including India. 1.50 Innovationis another important element o f competitiveness and growth as there is greater mobility o f factors, products, services, and knowledge. The innovation system plays an important role inacquiring, creating, adapting, and disseminating knowledge, which is crucial for success in the knowledge economy. It consists of the network o f institutions, rules, and procedures that affect how the country acquires, creates, disseminates, and uses knowledge. Innovation in a developing country does not just concern domestic development of knowledge on the global frontier. It also concerns the application and use o f existing knowledge to the local context. For the countries o f South Asia, which are still far behindthe global frontier inmany sectors, tapping into and making effective use o f existing global knowledge will likely have a greater economic impact than directing most of its resources to develop frontier knowledge, no matter how prestigious the latter may be (Figure 1.9) Figure 1.9: South Asia on Innovation I $0 0 - 8 - 7 - e- E Q) Kenya d u 5- 2vi c 4- 3- 2 - 1- 0 I I 0 1 2I I 3 4 I 5I eI 7I 8I 0I 1995 1.51 As a region, the South Asian countries do better on the innovation pillar than on any o f the others, and that is largely because o f the capabilities o fthe large countries, India inparticular, but also Palustan and to a lesser extent Bangladesh. The main strength comes from the large absolute number o f scientists and engineers in R&D as well as the number o f scientific and technical journal articles. Bangladesh also has strength in the high science and engineering enrollment ratios inhigher education, although this advantage i s diluted by the very low tertiary enrollment rates. 1.52 Another area o f relative strength i s a strong state o f cluster development, although this is mostly concentrated in India (where it includes not just IT services but pharmaceuticals, textiles, and metal engineering industries), Pakistan (medical instruments, sporting goods, textiles, and garments), and Sri Lanka (textiles and garments). 17 1.53 The overall formal R&D effort o f the South Asian countries is very small. R&D expenditures as a share o f GDP average 0.48 percent, with a higho f only 0.78 percent inIndia. In general, the vast majority o f the research done in the South Asian economies i s done in public R&D labs. An area where India shows some strength though is in patenting. India has a large public research network and recently there have been some reforms that are strengthening the incentive regime to produce more commercially relevant output. 1.54 The private sector, with the exception o f some o f the larger Indian groups, does very little research. Inaddition, the relatively little research done by the public sector i s not commercially relevant, and there are poor mechanisms to get it out to the productive sectors. It i s also generally quite burdensome to start up new businesses particularly technology-based business that face the additional challenge o f raising funds for risky new technology projects. The overall business environment is somewhat more supportive in Sri Lanka. 1.55 Another area o f weakness o f the innovation system in South Asian countries i s the poor links between university and company researchers. This is a little stronger in India than in the other countries, but i s still quite weak by the standards of developed countries. 1.56 However, not all innovations are done through formal research. Inall countries there are informal innovationefforts. InIndia some o f this effort i s being collected through an organization called the Honey Bee network, which has documented more than 12,000 small indigenous innovations, mostly in the agncultural sector. Furthermore, the government i s beginning to pay attention to supporting and scaling up this indigenous effort. 1.57 All five countries with a partial exception of Sri Lanka, however, do not draw very much on global knowledge. This is revealed by the very low share o f foreign direct investment to GDP, which is just a fraction o f 1percent for all countries except Sri L a n k (where it is 1.4 percent), and by very low formal purchase o f foreign technology as shown by very low royalty or licensing fee payments (US$0.33 per person in India where it is the highest followed by US$0.12 in Paktstan, and virtually nil inthe others). This contrasts with the situation o f East Asian countries where the average share o f gross direct foreign investment as a share o f GDP is 8.26 percent and the average royalty and licensing fees per population are US$30.82. In addition, with the exception o f Sri Lanka, the share o f manufactured trade (imports and exports as proxy for access to embodied knowledge and pressures to keep up with global technology) in GDP is less than 25 percent (and in India it is only 13 percent) compared to an average o f 99 percent for East Asian countries. 1.58 Inconclusion, the South Asian countries are significantly behind the global frontier in education and innovation and ICT. There is considerable diversity among the five South Asian countries. India i s clearly ahead in its skills, technology, and innovation capability because o f its much larger size and the critical mass in the absolute number o f highly skilled population, number or researchers in R&D, resources allocated to R&D, and the vast network o f public research laboratories, universities, and large private companies that are already undertaking research. Nepal is at the other extreme because o f its very small population, much lower per capita income, and much less developed technology infrastructure. 1.59 There are many generic actions that are similar across the South Asian countries, and all the countries share the need to find more effective ways to extend education and technology to the large part oftheir population that is outside the modem economy. Sharing the experiences that each country has in dealing with the issues identified would be very beneficial to others. In addition, countries that are less advanced in a particular area could learn from those that have 18 more experience o f successful programs inthat area. There i s also scope for collaboration across countries in tackling similar issues and even in doingjoint research on common problems. An excellent example o f such knowledge sharing across a region as well as a formal framework for joint research i s given by the European Community's program in education and in research. Besides regional collaboration, the key policy actions needed to strengthen knowledge economy and competitiveness include improving the economic incentive and institutional regme, strengthening education and skills, tapping global knowledge, and networking and collaboration. Infrastructure 1.60 Modern infrastructure, particularly electricity, telecoms, and roads, i s critical to economic development. As noted earlier, a part o f the reason for highcost o f doing business in South Asia i s the inadequacy o f infrastructure. Electricity provides light, the ability to use modern equipment, computers, and access to ICT. Telecoms facilitate information exchange and access to the rest of the world, while transport infrastructure i s critical for trade and by lowering transport costs extends the market and increases competition. Studies of the productivity o f infrastructure suggest that infrastructure has strong complementarities with other human and physical capital. If there i s a surplus o f infrastructure, more investment adds little to total output, but if there i s a deficit, then shortages constrain total output, magnifying the impact, so that the return to reducing that deficit can be very high indeed. In chapter 6, David Newbeny deals with South Asia's challenges inthe provision o f one major infrastructure service, electricity. 1.61 Firm level surveys o f investment climate, as mentioned earlier, have identified infrastructure, particularly power, as a major constraint to growth in South Asia. The concern on lack o f electricity i s striking in South Asia compared to other regions in Asia. In India, for instance, investment climate surveys have found that, on average, manufacturers face almost 17 significant power outages per month versus 1in Malaysia and less than 5 in China. InPakistan, the typical business estimates that it loses 5.6 percent in annual sales revenue owing to power outages against a reported loss o f 2 percent by its Chinese counterparts. InBangladesh, the most frequent common complaint i s the constraint imposedby the poor electricity system. 1.62 Cross-country data reinforce the findings o f the firm level surveys. Despite some recent improvements, infrastructure coverage and quality inthe region do not compare well with the rest o f the world. South Asia ranks the last among all world regions interms o froad density, rail lines, and mobile tele-density per capita. It is slightly ahead o f the Sub-Saharan AfXca region in terms o f mainlines coverage, electricity, improved water sources, and sanitation. South Asia is the only region in the world that has no city that can provide 24/7 piped water. Poor transport and communications still hinder the integration o f many rural areas into the wider economy. For example, in Palustan urban tele-density is 28 per 1,000 households versus 0.9 for rural areas. Rural access to all-season roads i s as l o w as 39 percent inBangladesh and 30 percent inNepal. 1.63 When compared to its neighbor East Asia and the Pacific ( E M ) and one of its main economic competitors China, South Asia region and India have significantly stayed behind in terms o f infrastructure expansion and improvement. The gap in infrastructure coverage between South Asia and East Asia has enlarged more dramatically inthe past decade. While in 1980 India was characterized by higher levels of infrastructure coverage (that is, electricity, pavedroads, and mainlines), duringthe 1990s China poured substantial investments in infrastructure and overtook India and South Asia infrastructure coverage by 2002. Were the region to try to reach China's present level o f infrastructure stocks per capita by 2015, it would have to invest more than 12 percent o f GDP each year for the next 10years (Table 1.3). 19 Table 13: InfrastructureStocks inSouthAsia, East Asia, and China, 2003 EAP LatinAmerica Sector India China South Asia (2000)p and the Caribbean Electricity generationcapacity (million kilowatts) 126.3 356.1 151.9 66.5 Paved roads (kmper 1,000 km') 267.2 171.6 227.5 Rail routes (kmper 1,000 km2) 19.2 6.3 14.6 IMadines (lines per 1,000 hab) 46.3 209.0 38.9 (35) 49 (76 in 2o03) 192 Mobile (lines per 1,000 hab) 24.7 215.0 22.6 (37) 48 (96 in 2o03) 249 Access to improved water (percent o fpopulation)b - - 88 (86) 77 86 (72) 75 90 Access to improved sanitation (percent ofpopulation)b 34(30) 44 39 (48) 60 77 Sources: E M Infrastructure Flagship, WorldDevelopmentIndicators, and SASE1databases. a. Infrastructure stocks for EAP countries do not include stocks of China. b.Water and sanitationfigures correspondto 2002. 1.64 The challenges involved in addressing the infrastructure deficit, and in particular the shortcomings in policies and service delivery that has lead to this infrastructure deficit, i s enormous. The politicization o f tariff setting for services has lead to prices well below costs, and those services are not well targeted to the poor. 1.65 Regarding electricity supply, Newberry notes that while not all South Asia countries suffer from the same problems, as a generalization the regon has inherited the legacy o f state- owned vertically integrated electricity supply industries, often with the characteristic politicization o f tariff setting that leads to excessively cheap electricity to domestic consumers, highlevels o fnon-technical losses (that is, theft or failure to collect bills), highlevels o f debt or arrears, hgh levels o f manning, and poor commercial performance (as measured by the ability o f revenues to cover costs). As a result, it i s difficult for the sector to finance its investment needs on commercial terms. The shortage o f revenue leads to poor maintenance with frequent equipment failures (for example, as measured by transformer failures and low generation availability), resulting in power shortages and load shedding. Figure 1.10 gives time series o f losses, although for India these are considerably below those reportedby various states. 1.66 Prompted by the apparent success of reforms in Latin America, many South Asian countries in the region have considered or embarked upon reform programs to allow private investment in the sector. The main obstacle to private investment is the fear that once the investment is sunk, it will not be allowed to earn a remunerative return. The electricity sector i s particularly problematic as private investors supply an essential service directly to a large fraction o f the voting population in competition with underpriced supply from the state-owned sector. As prices will have to rise to ensure that the investments are remunerative, the price rise will be associated with the reforms that brought in private investors, and will be doubly resisted on that account. Many o f the current beneficiaries o f opaque accounting, cross-subsidies, patronage in the appointment o f regulators and senior management, and so on will have an interest in preserving the status quo, including the low prices that deter efficient commercial competition. 20 Figure 1.10: ElectricityLossesin SouthAsia and China (Losses as percent of generation) 20 15 10 5 0 I 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 I Source: World Bank 2002. (See Chapter 6). 1.67 There remainreasons for being optimistic. Governments are increasingly confronting the challenges of poor sector performance. In the power sector, South Asia provides examples of innovative approaches to expanding rural access and to improving performance, including cooperative rural schemes in Bangladesh, and a thrivingprivate sector playing an extensive role insmall scalerenewable-basedpower inSri Lanka. 1.68 Perhaps the main leadership role that governments in the region could contribute would be to agree and enforce a regional energy charter to underwrite increased energy trade. Such steps have been effective in integrating the transition countries of Central Europe into the European Union, and stimulating foreign direct investment (FDI) into the power sector, and might have similarly stimulative effects inSouth Asia, quite apart from creating profitable trade opportunities and increasing regional security of supply and greater resilience against external oil shocks. Opening access to industrial customers would help assure the financial viability o f investments in cross-border infrastructure. 1.69 Inconclusion, South Asia can sustainhighgrowthprovidedit can aggressively manage the downside risks arising from a poor investment climate, weak institutions, weak knowledge economy, and a poor infrastructure. At the same time, it needs to complete the any remaining agenda on global integrationand macroeconomic stabilization. 1.70 We now turn to the second issue, which deals with the role of regional integration in supporting highgrowth inSouth Asia. 21 I s REGIONAL INTEGRATION DESIRABLE? 1.71 Regional integration can promote growth through several channels: agglomeration benefits, increased investmentsas a result of enlarged markets and economies of scale, flow o f information and technology and knowledge spillovers, increased foreign direct investments, and deeper integration through regulatory cooperation and harmonization. These benefits of regional integration typically materialize when external protection is low or when integration is between north-south rather than south-south (Hoekman and Schiff 2002). When extemal protection is high, trade diversion dominates trade creation. Countries are more likely to benefit from north-south integration because endowment differences are larger between north-south, and this will encourage countries to exploit their comparative advantages better. Knowledge and technology spillovers are also likely to be higherbetween north-south. Since institutions that protect property rights and promote competition are likely to be superior in industrialized countries, north-south integrationi s likely to generatemore benefits from deep integration. 1.72 But regional integration can also provide opportunities that are unlikely to be addressed between north-south, and where a case can be made for a regional approach (for example, energy trade, transport andtrade facilitation, andreduced conflict and fiction). 1.73 This section examines whether regional integration i s desirable for South Asia. I s there evidence on agglomeration benefits arising &om integration in South Asia? Is trade preference in South Asia likely to lead to trade diversion rather than trade creation? What role can regional cooperation play in addressing energy shortage in South Asia? Can integration play a role in ensuringthat laggingregionsand land-locked countries have full access to markets? AgglomerationBenefits 1.74 There is not much evidence on agglomeration benefits from regional integration in South Asia, at least not inthe manufacturing sector, although a potential role for it inthe services sector can not be ruled out. Unlike East Asia and the European Community, South Asia has not benefited from the neighborhood effects o f regional integration. The benefits o f integration are clearly evident in the manufacturing sector in East Asia, where production networks have been formed within the region to take advantage of economies o f scale, specialization, and agglomeration. The literature on new trade theory, new growth theory, and new economic geography all highlightthe role that integration can play inenhancing productivity and growth by taking advantage o f economies o f scale and specialization, which, in turn, could give rise to agglomeration at the regional level (Feenstra 1999; Bhagwati and Dehejia 1994; Krugman 1996). 1.75 UnlikeEast Asia, South Asia appears to have done well inthe services sector, benefiting &om outsourcing and specialization. South Asia i s the fastest growing region in the export of services. Exports of services from South Asia grew at 14 percent per annum over the period 1995-2003 compared to less than 8 percent for East Asia (Figure 1.11). It is not only India that did well, but Pakistan and Sri Lanka, too, which have grown faster than East Asia in service exports. Bangladesh services exports have also grown fairly rapidly, averaging about the rate in East Asian economies. India and Bangladeshhave performed well inthe exports o f computer and information communications and other commercial services, while Pakistan has done well inthe export of transport services and Sri Lanka intravel services. 22 Figure 1.11: South Asia Has Done Well in Service Exports Sri Lanka East Asia & Pacific Bangladesh United States 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 Average Annual Growth Rate (1995-2003) in current US$ 1.76 There are two different channels through which services can be exported. It can be outsourced to another country or labor can migrate to produce the services inthat country. We do not have data on labor migration by sectors, but we have data on the skill composition of labor migration and remittances inflow. Remittances inflow and the skill composition of labor migration are other crude indicators on South Asia's performance on exports. 1.77 First, as a region, South Asia has the largest ratio of remittances to GDP. This ratio significantly exceeds FDI inflows. Inflow o f remittances into South Asia has in many respects performed the same functions for South Asia as FDI inflows have done for East Asia. It i s an important source o f capital and ideas for the region. 1.78 Second, global demand for South Asian skills i s high, as reflected by the skill composition of labor migration to OECD countries. Nearly 64 percent of the labor force that migrates out of South Asia to OECD belongs to the high s b l l category. This ratio i s significantly higher for South Asia compared to other regions. The proportion of workers that migrate out of South Asia to OECD countries with low slulls is small compared to the medium and high skill categories. 1.79 Why did East Asia benefit from outsourcing in the manufacturing sector while South Asia seems to be doing the same in the services sector? This question cannot be fully addressed here. In part, the growth of the manufacturing sector in South Asia was constrained by poor physical infrastructure. But infrastructure was not a binding constraint for the services sector. English language and IT slulls strengthened the competitiveness of South Asia in the services sector. The services sector in South Asia has also benefited from a relatively liberal regulatory regime, which attracted FDIandthe flow of ideasandtechnology from abroad (Figure 1.12). 1.80 However, service exports from South Asia are largely global and not regional. Like the manufacturing sector, services sector also benefit from economies of scale and specialization. But transportation costs for the services sector are significantly low compared to the transportation costs inthe export o f manufactured goods. Internet and modern telecommunication have resulted in the transportation cost of exporting IT services to the United States, for example, beingno more expensive than transporting it to Sri Lanka. Nevertheless, there is a potential for South Asian countries to gain from liberalizing regional trade in education, health, and tourism at the regional level. 23 High- and re in 70 0 4.5 *4 0 +3 5 3 0 2 5 2 0 + 1.s I10 -00 5 0 Trade Table 1.4: rts 24 Figure 1.13: SouthAsia Iias Reduced rt into SouthAsia c 25 import rt 26 for South Asian countries to push multilateral trade liberalization instead o f focusing on regional trade agreements (Baysan, Panagariya, and Pitigala 2006; Newfarmer and Pierola 2006) 1.87 Panagariya notes that SAFTA faces many challenges. First, a critical factor in determining whether SAFTA would raise or lower the real incomes of the South Asian countries depends on whether it will be predominantly trade creating or trade diverting. Thus, for example, when Bangladesh allows Indian cement to be imported duty free and this leads the more efficient Indian cement industry to out-compete the less efficient Bangladesh cement industry, there is trade creation: Increased imports into Bangladesh represent a shift from high-cost Bangladeshi producers to low-cost Indian producers. On the other hand, if duty-free access to Indian computers into Bangladesh allows the less efficient Indian computer manufacturers to displace more efficient Korean suppliers who remain subject to the duty, there is trade diversion: Increased imports from India inthis case represent a switch from low-cost outside sources to the high-cost within-union sources o f supply. Some characteristics o f the South Asia region (for example, small regonal market relative to the world bothin terms o f GDP and trade flows, high level o f protection among SAARC countries) increase the probability that SAFTA i s likely to be largely trade diverting. 1.88 The second challenge i s the political economy o f the selection of excluded sectors and rules o f origin. When countries are allowed to choose sectors that can be excluded from tariff preference o f free trade, domestic lobbies make sure that the sectors in which they may not withstand competition from the union partner are the ones that get excluded. On the other hand, lobbies go along with free trade in the sectors in which they are competitive and the preference will threaten the imports from outside countries. Inthe same vein, lobbies tend to go for tight rules o f origin or outright quantitative restrictions inprecisely those sectors inwhich they fear the competition from the partner most. On the other hand, when the threat i s mainly to the imports from outside countries, they are willing to accept greater liberalization. The rules o f origin can also be subject to abuse by the bureaucrat administering them. In cases where imports from the partner may be threatening an inefficient domestic competitor, bureaucratic discretion may be employed to block entry o fthe imports. 1.89 The reservations to the SAFTA based on sectoral exceptions and rules o f origin were applied extensively under the Sri Lanka-India Free Trade Agreement. For example, the top 20 exports o f Sri Lanka (to the world and not just India) at the 6-digit H S level accounted for 46 percent o f Sri Lanka's total exports in 1999. India subjects 15 out o f these 20 products to either a tariff rate quota (meaning the tariff preference applies only up to a pre-specified quantity o f imports) or negative-list exception. Thus, the exclusionary policies applied with potency to products in which Sri Lanka showed the greatest comparative advantage. The rules o f origin and rules o f destination requirements further restrict exports. For example, apparel exports from Sri Lanka are not only subject to the tariff rate quota o f 8 millionpieces but at least 6 million o f these pieces should be manufactured from fabrics o f Indian origin exported to Sri Lanka from India. Likewise, exports o f tea from Sri Lanka at the preferential tariff are not to exceed 12.5 million kilograms within a calendar year. Both products are also subject to a uniquely South Asian restriction called the rule o f destination: The preference applies only ifthe products enter through specific Indianports. 1.90 The third challenge is that SAFTA overlooks the role of the services sector. The welfare effects o f trade preferences for services are likely to be more positive compared to trade preference for goods, as preferential liberalization in services leads to trade creation with little or no trade diversion. Loss o f tariff revenue from services is less of an issue and it allows countries to take advantage o f increasingreturns to scale. Regulatory cooperation, o f particular importance 27 Y&.O 1Fi.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 UPf3c aint throu a1 Cooper 28 T af 29 30 1.100 Better trade facilitation would reduce substantially the transactions costs o f intraregional trade. But streamlining transport and trade systems i s also needed to facilitate interregional trade. As formal tariff barriers fall, transport-related costs are becoming one of the prime factors determining competitiveness. Many of South Asia's competitors have already dramatically reduced customs and port clearance times, South Asia risks being left behind. Pnvately managed port terminals, with international levels o f productivity, are increasingly the norm, though Bangladesh remains an exception. But such terminals have to be complemented by full customs computerization, port community IT systems, reliance on self declaration and risk management systems, and streamlining o f other non-customs trade-related procedures if SouthAsia is to make the transitionfromthirdworldto first class external transport systems. 1.101 Finally, one other area not covered in this volume where regional cooperation will have large payoff concerns cross-border management o f water resources. Watershed management and storage inNepal would generate hydro and irrigation benefits inNepal and flood control benefits inBihar. Water storage innortheast India could provide hydropower and flood control in India and flood control and dry season water augmentation in Bangladesh. Similarly, there are significant cross-border water management issues between Afghanistan and Pakistan. High priorities include agreements between India and Bangladesh and between Afghanistan and Pakistan on regionalwater issues. CONCLUSION 1.102 South Asia has experienced remarkable growth rate since the 1980s. This growth was triggered by reforms aimed at increasing global integration, improving macroeconomic management, and unleashing the private sector as the engine o f growth. Highgrowth rate, inturn, has created the interest in, and political space for, greater regional integration. Ths raises two issues of importance to the region: I s highgrowth sustainable?I s regional integration desirable? 1.103 Past growth was helpedbythe implementationo f first generation policy reforms aimed at global integration, macroeconomic stabilization, and reducing the scope o f the state while strengthening the role of the private sector. These reforms have made South Asia more competitive, stable, and adaptable. 1.lo4 South Asia now faces increasing challenges from second-generation reforms. These include the high cost of doing business, weak institutions, weak knowledge economy, and poor infrastructure. Refoms in these areas are inter-related and hold prospects for large payoffs. Increasing investment rates will require reducing costs of doing business, improving institutions, and addressing the infrastructure constraint. South Asia also needs to expand its knowledge economy. Increasesinbothphysical and human capital investments are needed to bringabout the structural transformation of the economy, shifting unemployed and low-skill workers from agriculture and infomal services (low productivity) to other sectors such as manufacturing and formal services (highproductivity) which require high skills and better infrastructure. South Asia has begun to address the second-generation policy reforms and continued implementation will ensure sustainability of growth. 1.lo5 I s regional integration desirable? From the very narrow perspective o f trade flows, the economic characteristics of the South Asia regon, such as the small regional market relative to the world both interms o f GDP and trade flows, and the highlevel of protection, would suggest that focusing on regional integration alone will not generate the beneficial productivity and growth effects o f integration. South Asia i s a relative newcomer to global integration. Despite recent liberalization, it lags other regons such as East Asia interms o f openness. When external 31 protection is high, trade diversion i s likely to dominate trade creation, and so the risks that regonal integration will be a drag on growth in South Asia i s high. History shows that a successfulregional integration i s often precededby global integration. 1.106 Nevertheless, regional integration is desirable from other perspectives. Regonal cooperation can be an effective tool in addressing energy shortage, ensuring that no regiodcountry i s left behind, land-lockedregions/countries have full access to markets, and peace and stability are promoted. Above all, there is a need for greater people-to-people contact through improved connectivity, phasing out of visa restrictions, and liberalizing the restrictions on the trade of services (for example, tourism, education, and health) where the risk o f trade diversion i s low. These initiatives would help increase investment and growth by reducing the infrastructure constraint and by lowering transaction costs. Better regional cooperation and integration can also increase welfare by improving the regonal political environment, thereby reducing conflicts and associatedsocial and economic costs. 32 2. ECONOMICGROWTH INSOUTHASIA: A GROWTHACCOUNTING PERSPECTIVE4 INTRODUCTION 2.1 South Asian economies have achieved impressive rates o f economic growth since the 1980s. Output for India, Pakistan, Bangladesh, and Sri Lanka (SA4) has grown more rapidly since 1980 than for any other region except East Asia. However, unlike East Asia, these countries have not been characterized by particularly highrates of investment. Indeed, as a share o f output, investment has averaged just one-half to two-thirds o f the levels typical in East Asia during its sustainedperiods of highgrowth. This fact-which some see as a puzzle-could be taken to imply that capital accumulation has been a relatively unimportant factor in the region's growth experience. If capital accumulation mattered little in the past, perhaps it need not be a central focus for South Asia to sustain, and increase, its rates o f growthin the future. 2.2 The main objective here i s to take issue with that perspective.' It is argued instead that capital accumulation i s indeed important for growth and that this is evident in South Asia's experience. Furthermore, countries inSouthAsia will needto increasetheir rates o f investment so as to accumulate (both physical and human) capital more rapidly ifthey are to achieve the desired rise in growth rates going forward. It i s also important to emphasize that capital accumulation is certainly not all that matters. Strong, sustained growth also requires countries to increase the efficiency with which they use domestic factors to produce goods and services, which economists label total factor productivity (TFP). The two should both be seen as central to the growth process, and indeedmany policies undertaken to foster one will also tendto promote the other. 2.3 Thus, the perspective that capital accumulation matters goes hand-in-hand with the extensive and convincing new findings linkingpositive growth experience with strong domestic institutions (such as those protecting property rights). Growth accounting provides a useful framework for exploring the roles o f capital accumulation and TFP and i s the basis for the empirical analysis presented. The approach, including its strengths and limitations, i s discussed further below. Some o f the implications will be discussed from work to date from ongoing research on economic growth in South Asia and elsewhere. GROWTH SOUTHASIA: ANOVERVIEW IN 2.4 Table 2.1 highlights some key features o f South Asia's growth. A s shown, during the period 1980-2000, two o f the economies (India and Bangladesh) increased their gross domestic product (GDP) growth rates by roughly 2 percentage points per annum relative to the rates they had sustained in the two decades prior to 1980. Sri Lanka's growth increased only marginally, but from the initially strong rate o f 4.5 percent per year. While average output growth declined after 1980 inPakistan, itremained about 5 percent per year. Growthrates o f these magnitudes are impressive achievements that have helped these countries to reduce poverty rates and raise living standards. Indeed, South Asia grew more rapidly than any other region except East Asia. This chapter was preparedby SusanM. Collins,Professorof Economics, GeorgetownUniversity and a Senior Fellow in Economic Studies, the BrookingsInstitution, USA. Key argumentsare drawn from work that the author has done with BanyBosworth. The author would also like to thank GabrielChodorow-Reichfor expert research assistance. The views in this paper are solely the authors and do not necessarily reflect those of the World Bank or its Executive Directors. See BosworthandCollins (2003). Table 2.1: South Asia: SelectedIndicators Annual ratesof change GWcapita Population GDP Labor Investment share (PPP)" (millions)" force (percent) India 1960-80 3.5 2.1 21.9 1980- 2003 $2,880 1,064 5.7 2.0 21.3 Bangladesh 1960-80 2.4 2.2 22.7 1980- 2003 $1,870 138 4.4 2.3 18.2 Pakistan 1960-80 5.9 2.7 22.1 1980- 2003 $2,040 148 4.9 2.7 19.5 Sri Lanka 1960-80 4.5 2.2 23.2 1980- 2003 $3,740 19 4.4 2.0 22.6 Source: World Bank 2005. Note:PPP, purchasingpower parity. a. These dataare for 2003. 2.5 However, as also shown in Table 2.1, these growth achievements do not seem closely relatedto what has happenedto underlyinginputs.Inparticular, investment as a share o f GDP did not rise after 1980 inthe countries where growth surged. Instead, it declined slightly inIndia and more sharply in Bangladesh. Moreover, in none of the four countries has investment rates approached the 3040 percent range typical for East Asian economies during their rapid growth periods. The table also shows that, while the labor force continued to grow rapidly in all four countries, this growth didnot accelerateineither India or Bangladesh. 2.6 Thus, the Figures inTable 2.1 raise questionsrelatedto South Asia's growth experience. In particular, should they be taken to imply that capital accumulation played little role in explaining strong growth in the region? And what are the implications for the importance of investment, if South Asia i s to sustain, and hopefully increase, its output growth over the next decade? While the investment rate relates to the role of physical capital, it is also o f interest to explore the role of increasesinhuman capital, which i s done brieflybelow. OutputGrowthversus LivingStandards 2.7 Most of this examines the growth in South Asia's GDP. The growth accounting methodology focuses attention on the proximate role of underlyingfactor inputs:physical capital, labor augmented for changes inlabor quality usingeducational attainments, and the residual role of increases in the efficiency with which those factors are used. Thus, much o f the analysis emphasizes growth inGDP scaledby labor inputs(GDPLF), an indicator o fproductivity. 2.8 However, it is important to keep in mind that a primary objective is to raise living standards for residents in each o f these countries. GDP does not measure how much o f the income earned fi-om production actually goes to locals. Given the wide disparities in age 34 distribution and labor force participation across countries, labor force i s not an accurate indicator o fpopulation. Gross national incomeper resident (GWcapita) i s a much betterindicator o f living standardsthan GDP/LF. 2.9 There i s a close linkage between the two indicators. As shown in equation (2.1) below, each country's per capita income can be decomposed into productivity, the portion of domestic income that accrues to residents and the labor force as a share of the total population. Data permitting, this last term can be further decomposed into a demographic effect-the population share o f labor force age-and the labor force participation rate (LFPR). The split between labor force participation and a demographic effect i s not presented because the results for some countries are quite sensitive to the selected start and end dates, raising questions about data reliability. However, the demographic effect does appear to have been positive in all four countries as dependencyrates have declined.6 (GDP /LF)' GNI/GDP)' ( (LF/pup) = GN$up where GD%F =productionper memberofthe labor force GN%DP =the proportionofincome fromproductionthat accruestoresidents =theproportionofthe populationthat is economicallyactive GN%*p =nationalincomeper capita 2.10 Table 2.2 reports this decomposition for each country. Inaddition to measures for 1980 and 2003, it shows the percentage change over the period. The top panel shows that for India, productivity and living standards both more than doubled, increasing by roughly the same percentage. Nearly all income accrued to residents in2003, as in 1980. Similarly, there was little change in the labor force as a share of the total population. In contrast, living standards have grown more slowly than productivity in Pakistan, reflecting a shift from net inflow to a net outflow of factor payment^.^ However, in Bangladesh and Sri Lanka, growth in per capita incomes have significantly exceeded growth in productivity. For Bangladesh and especially Sri Lanka, this reflects a rise in the share of the population that i s economically active. Bangladesh has also experienced a net inflow of factor payments. 2.11 Table 2.2 also highlightsthat a relatively small share o f the people in these four South Asian countries are economically active. As discussed further below, this i s attributable inpart to the relatively low labor force participation for women in India, Sri Lanka, and especially Pakistan. With employment data, productivitycan be measuredas GDP per employed worker, and the last term can be split betweenthe demographiceffect, the LFPR, andthe employment rate. However, net factor payment to Pakistan have been quite variable, in part due to remittances, and do not exhibit a clear trend. 35 Table 2.2: Components of Gross NationalIncome per Capita GDPhabor force participant (2000 Labor GNUcapita (2000 internationalUS$) GNI/GDP force/population international US$) India 1980 2,705 1.oo 0.44 1,185 2003 6,144 0.99 0.44 2,721 Percent change 127.2 -0.7 2.0 129.5 Pakistan 1980 2,916 1.08 0.37 1,148 2003 5,277 0.97 0.38 1,927 Percent change 81.0 -9.9 2.3 67.8 Bangladesh 1980 2,257 0.97 0.49 1,074 2003 3,259 1.06 0.51 1,767 Percent change 44.4 8.5 5.2 64.5 Sri Lanka 1980 4,550 0.99 0.38 1,704 2003 7,791 0.99 0.46 3,533 Percent change 71.2 -0.4 21.0 107.4 South Korea 1980 1,180 0.98 0.41 4,482 2003 32,594 1.oo 0.52 17,004 Percent change 191.5 1.9 27.8 279.4 Thailand 1980 4,705 0.99 0.53 2,463 2003 12,041 0.98 0.60 7,038 Percent change 155.9 -1.1 12.7 185.7 Source: WorldDevelopmentIndicators and author's calculations. 2.12 Increases in the percentages of labor force-aged women who become economically active, combined with continued declines in dependency rates, are important channels for raising the growth of income per capita above the growthrate o fproductivity. GROWTHACCOUNTS 2.13 Growth accounting provides a means for decomposing increases in output per worker into the contributions from accumulation o f physical and human capital (per worker) and a residual measure of the change intotal factor productivity.' Inrecent work with Barry Bosworth, we have constructed growth accounts for 84 industrial and developing economies over the period 1960-2003. This sample o f countries, which encompasses 95 percent of world GDP and 85 percent of world population, gives a useful means for comparing growth experiences across countries, regions and other time periods. Methodology and data sources are explained in the annex. *As discussed above, "worker" refers to a member o fthe labor force except for the industrial countries. 36 2.14 Growth accounting is a useful framework for organizing data. It is simple and internally consistent, and has been used ina wide variety o f contexts. For example, growth accounts (often based on industry-level decompositions) are central to the ongoing debate over recent productivity growth in the United States and Europe, and the role o f computers and other information technology goods. Given their extensive usage within industrial countries, it i s somewhat surprising that some in the development literature view them with skepticism. Bosworth and Collins (2003) argue that this partially reflects issues of measurement and interpretation, and that when appropriately implementedand interpreted, they provide extremely informativebenchmarks for studylnggrowth experiences. 2.15 It is also important to keep in mind their limitations. These are noted briefly here and discussed further in Bosworth and Collins (2003). First, growth accounting shows only the proximate sources of growth and i s not intended to determine the underlying causes o f growth. Consider a country with rapid increases in both accumulation of capital per worker and factor productivity. The decomposition provides no information about whether the productivity growth caused the capital accumulation (for example, by increasing the expected retums to investment) or whether the capital accumulationmade additional innovationspossible, or some combination. 2.16 Second, growth accounts measure TFP as a residual. Inaddition to changes in economic efficiency, this residual will reflect a range of other determinants o f growth, not accounted for by the measured increases in factor inputs. Changes in TFP should not be taken as a proxy for technological inno~ation.~ 2.17 Third, some express concern that the decomposition is sensitive to measurementofinputs and outputs, and to the underlying assumptions about the production process. Thus, we pay considerable attention to measurement issues inconstructing the dataset. 2.18 Fourth, growth accounts are an appropriate tool for examining growth experiences over longer run periods of a decade or more. The supply side approach i s not designed to capture cyclical relationships between variables or effects of short-term shocks such as droughts. By construction, cyclical movements in output simply will be reflected in the residual measure o f TFP. 2.19 Table 2.3 reports the resulting growth decompositions for each of the SA4 by decade and for the entire period. Figure 2.1 provides graphical representations of the decomposition, with the contributions from physical and human capital combined for ease o f presentation. Regional averages are shown inTable 2.4 for comparison purposes. China i s reported separately from the rest of East Asia owing to concerns about data reliability. (See Heston, 2001.) The discussion below refers to these tables and figures inhighlightingthe main findings. ~ See Hulten(2001) for adetaileddiscussionof growth accountingandthe TFP residual. 37 2,t: Per !1: I t s 3 5 3 2.5 2 1 5 2 0.5 0 38 Pakistan 3 5 3 2 5 2 1 5 1 ts.5 0 Sri Lanka 3 2 s 2 I 5 1 0 5 a 39 Table 2.3: Sourcesof Growth, SouthAsia, 1960-2003 Contributionof Output Outputper Physical Factor Regiodperiod worker capital Education productivity India 1960-70 3.74 1.81 1.34 0.17 0.74 1970-80 3.08 0.70 0.73 0.33 -0.21 1980-90 5.50 3.48 1.06 0.36 2.05 1990-2003 5.82 3.99 1.48 0.46 1.95 1960-2003 4.62 2.59 1.17 0.34 1.19 Bangladesh 1960-70 3.99 2.09 0.10 0.13 1.85 1970-80 0.81 -1.61 -0.73 0.30 -1.18 1980-90 3.72 1.07 0.84 0.18 0.05 1990-2003 4.84 2.67 1.61 0.34 0.70 1960-2003 3.43 1.15 0.53 0.24 0.37 Pakistan 1960-70 7.22 4.67 3.75 0.35 0.54 1970-80 4.68 1.59 0.68 0.23 0.68 1980-90 6.28 3.55 0.98 0.92 1.61 1990-2003 3.82 1.08 0.46 -0.08 0.69 1960-2003 5.37 2.60 1.39 0.33 0.87 Sri Lanka 1960-70 4.57 2.16 0.08 0.33 1.73 1970-80 4.40 2.25 1.55 0.39 0.29 1980-90 4.19 2.11 2.05 0.22 -0.16 1990-2003 4.64 2.53 1.10 0.34 1.07 1960-2003 4.46 2.28 1.19 0.32 0.75 Source: Authors' calculations as explained intext. 2.21 Itis also interesting to note that strong growth associated withrapid increases inTFP was a standard pattern observed during the 1960s (not shown separately). During this decade, efficiency increases were associated with roughly 60 percent o f the rise in output per worker among industrial economies, still recovering fkom the World War 11, as well as in Latin America and the Middle East. 2.22 However, a more careful look at the results to distinguish across countries and decades suggests a much more nuanced view o f the contribution o f TFP to growth in South Asia (Table 2.3). First, the regional average (GDP weighted) i s of course dominated by India's experience. Only two o f the other three (Bangladesh and Pakistan) exhibited strong TFP growth since 1980, and both well below that for India. 40 Table 2.4: Sourcesof Growthby Region Contribution of Output Outputper Physical Factor Regiodperiod worker capital Education productivity South Asia (4) 1960-80 3.61 1.41 0.85 0.32 0.24 1980-2003 5.52 3.35 1.17 -0.36 1.76 1960-2003 4.63 2.44 1.02 0.35 1.05 World (84) 1960-80 4.48 2.77 1.19 0.39 1.17 1980-2003 3.25 1.82 0.85 0.29 0.68 1960-2003 3.82 2.26 1.oo 0.33 0.91 Industrialcountries(22) 1960-80 4.25 2.95 1.19 0.42 1.32 1980-2003 2.59 1.62 0.75 0.22 0.63 1960-2003 3.36 2.23 0.95 0.3 1 0.95 China (1) 1960-80 4.04 1.83 0.80 0.38 0.64 1980-2003 9.48 7.77 2.78 0.35 4.49 1960-2003 6.91 4.97 1.86 0.36 2.68 EastAsia less China (7) 1960-80 7.02 3.98 2.23 0.53 1.18 1980-2003 6.08 3.67 2.22 0.53 0.88 1960-2003 6.52 3.81 2.23 0.53 1.02 Latin America (23) 1960-80 5.74 2.75 1.04 0.32 1.37 1980-2003 1.96 -0.58 0.10 0.41 -1.08 1960-2003 3.70 0.95 0.54 0.36 0.05 Africa (19) 1960-80 4.40 1.94 1.01 0.12 0.80 1980-2003 2.16 -0.55 -0.08 0.41 -0.88 1960-2003 3.20 0.60 0.43 0.28 -0.1 1 Middle East (9) 1960-80 5.43 3.20 1.81 0.38 0.99 1980-2003 3.81 0.82 0.35 0.52 -0.05 1960-2003 4.56 1.92 1.02 0.45 . . 0.43 Source:Authors' calculations as explained intext. 2.23 Second, in all three o f these economies, TFP surged during the 1980s after two decades o f little growth. For India and Bangladesh, this surge partially reflected a recovery from the poor performance during the 1970s. Especially for India and Pakistan, TFP growth then slowed substantially during 1990-2003. 2.24 Third, the most recent period i s certainly not one inwhich physical capital accumulation was relatively unimportant for growth. For Bangladesh, Palustan, and Sri Lanka, its contribution accounts for 43 percent o f total growth inoutput per worker, a share identical to the overall global average. For India this share i s lower, but not substantially, at 39 percent. 41 InvestmentRatesVersusCapitalAccumulation 2.25 The growth accounting methodology highlights the fact that the change in a country's capital stock is the appropriate way to measure the contribution o f capital accumulation to growth-not the country's investment rate. This isparticularlyrelevant here, because the average ratios o f investment to GDP understate capital's contribution for all o fthe SA4. The change inthe capital stock is gwen by where d represents depreciation. Dividing through by K and assuming a steady-state constant value (y) for the inverse o f the capital-output ratio allows the rate of change of the capital stock (k) to be measuredby the investment rate (i =I/Q: k = i y - d (2.3) However, if the capital output ratio i s not constant, I can be a very poor proxy for k. The assumption o f a constant capital output ratio and steady-state seems particularly unreasonable for developing countries struggling to catch up. Indeed capital-output ratios not only differ across countries, but also evolve over time for many of the developing countries inour sample." 2.26 Infact, the correlation between the two measures is surprisinglylow, as illustrated in Figure 2.2. Each point represents the average investment rate and the average change inphysical capital stock for one o f the countries inthe sample. The SA4 (as well as Korea and Thailand) are each denoted with the first letter o f their country name. Countries that devote modest shares o f their output to investment can exhibit relatively large increases in their capital stocks if their output i s growing rapidly. Incontrast, countries with very slow output growth will tend to have small changes intheir capital stocks despite highinvestment shares. Figure2.2: Comparisonof InvestmentShareand Changeinthe CapitalStock, 1980-2003 (all countriesexcept IND,LKA,PAK, BGD through 2000) y = 0.2321~-1.1869 R2= 0.3495 K - -I -2 loNote that with investment equal to 20 percent of GDP and a 5 percent depreciation rate, the growth in the capital stock would equal 5 percent with a capital-output ratio equal to 2, but just 2 percent with a capital-output ratio just under 3. 42 EducationalAttainment: A Labor QualityIndex 2.27 The growth accounts also highlight increases in labor quality, which we measure using average years of schooling and an assumed rate of return to each additional year o f schooling." Overall, for our sample, increases in education contribute 0.3 percentage points per annum to growth. Duringmuch of its rapid growth period, East Asia stands out for very rapid rise inyears of schooling, contributingroughly 0.6 percent to annual growth, or twice the global average. 2.28 Increases in education among the South Asian economies have not been as impressive (Figure 2.3) and the average level of schooling remains quite low (Table 2.5). Data constructed byRobertBarro andJong-Wha Lee for 2000show that the percentagesofthe populationover age 15 who had completed at least the first level of schooling (six years) were just 17 percent in Bangladesh, 28 percent in India and Pakistan,12and 52 percent in Sri Lanka. However, in India the population share with post-secondary school education has risen relatively rapidly. Our growth accounts show the contribution fkom increased schooling ranging fkom under 0.3 to just over 0.4 percent. b 4 - 0 E o 3 5 N O . 3 a - a 0 3%2.5 2 -. - . - .E 5 1 5 ....P . cuo - . .-I-. -.-..-= -.- s - - . * - c I l s I c 1 , * 2c u 0.5 -:. . % : - - 1 0 n 1 . I I I I I 1 I n 2 4 6 8 i n 12 14 Average years ofschooling in 1980 - Table 2.5: Average Years of Schooling (Population Aged 15C) Average years of school 1960 1980 2000 India 1.7 3.3 5.1 Bangladesh 0.6 1.9 2.6 Pakistan 0.7 2.1 3.9 Sri Lanka 3.9 5.6 6.9 Source:Providedby JongWha Lee (Barro and Lee 2000 dataset). '' Our work to date assumes a constant rate of return to each additionalyear of schooling,. The results reportedhere use 7 percent, which is in the lower range of estimates from microeconomicstudies. We have also consideredhigher rates of return, and explored approaches to controlling for the quality as well as number of years, of schooling. However, some estimates suggest that the returns to educationmay not be constant and it would be usehl to explore this further. '*The educationaldatareportedfor Pakistanappearsparticularlynoisyandmaybeunreliable. 43 SOME IMPLICATIONS LOOKINGFORWARD 2.29 Insum, growthaccounts show thatbothcapitalaccumulationandincreased efficiency of factor usage have been important for South Asia's growth. Modest investment rates to the contrary, capital accumulation has not played an unusually small role. At the same time, there i s considerable room for increasing investments in both physical and human capital. This final section provides a much more speculative discussion, loolung to the future. What are the prospects for achievingdouble digit growth, and what wouldit require? 2.30 Could South Asia achieve this goal through concentrating on approaches intended to accelerate TFP growth? There are at least two reasons one might reach this conclusion. First, it has become popular in the literature on economic growth to emphasize the role of TFP. Thus, researchers now frequently model capital accumulationas endogenous, such that increases inTFP automatically induce the investment required to maintain the capital-output ratio. (See Easterly and Levine, 2001, and Klenow and Rodriguez-Clare, 1997) However, there is little evidence for this in our data. Capital accumulation and TFP growth exhibit surprisingly little correlation, consistent with the view that investment decisions are influencedby a great many factors (such as availability o f finance and tax consideration) inaddition to changes inTFP. It i s also worthnoting in this context that initial conditions and available policy indicators explain investment and capital accumulation at least as well as they explain changes in TFP (for example, see Bosworth and Collins 2003.) There are no magic policy bullets for generating TFP growth. Since both capital and TFP matter, a prudent policy stance should seek to foster both. 2.31 Second, India has achieved strong TFP growth with modest investment. Over this period, it has also achieved very impressive growth inits information technology (IT) sector, which the Ministry o f Finance forecasts will account for fully 25 percent o f value added by 2020, up from roughly 7 percent today. However, in a recent paper, Srinivasan (2006) identifies a variety of potential constraints to the sectors' growth. The domestic ones include shortages o f appropriately shlled labor, poor infrastructure, and difficult labor and bankruptcy laws. External ones include concerns about increasedprotectionismifU.S. fears about offshoringintensify. 2.32 Perhaps more importantly, one can agree that there are strong prospects for continued growth of India's IT sector without concluding that IT is a likely engine o f more rapid growth that could be sustained with modest investment. As discussed above, our data show somewhat more impressive Indian TFP duringthe 1980sthan inthe past decade. Furthermore, the IT sector i s concentrated in a few urban areas. A s in the other South Asian economies, a substantial share o f India's labor force i s relatively low skilled, living inrural areas, and involved with agricultural activities. Though difficult to estimate, underemployment is extremely high.Substantial increases inlivingstandards will necessitate reallocatingmuch o f this labor to activities inthe industrialor service sector, where their productivity will be significantly higher. And for them to be productive in these activities will require investments in their human capital, as well as increases ininfrastructure and other types ofphysical capital. 2.33 However, one should not be surprised if this reallocation generates relatively little TFP increase in the sectors to which these workers move. For example, there i s considerable scope for expanding production o f manufactured goods, using technology that i s standard in the global marketplace. Assuming the required capital i s purchased at world prices, and the goods are sold in world markets, this is unlikely to generate increases in TFP in the manufacturing sector. However, to the extent that labor pulled out of agriculture was at least partially redundant, one would expect to observe measured productivity in agnculture increase. Much o f the aggregate 44 productivity gains would be associated with the reallocation o f labor from agriculture where productivity i s relatively low, to the rest o fthe economy, where it i s considerably higher. 2.34 T o study the implications o f this reallocation for aggregate growth requires consistent data series so as to disaggregate the growth accounts by sector. We hope to extend our growth accounts to the broad categories o f agriculture, industry, and services for at least some o f the South Asian countries. However, this type of analysis has been done recently for Thailand, and the results for the period 1977-96, during which agnculture's share o f employment fell from roughly 65 to 45 percent, are quite interesting. Inparticular Bosworth (2005) finds that hlly two- thirds o f the rise in Thailand's TFP during this period can be attributed to reallocation o f labor (1.1 percentage points out o fa total TFP contribution to growth o f 1.6percent per annum). 2.35 Finally, it is interesting to construct some benchmark scenarios that project the implications for required investment o f achieving alternative rates o f growth over the coming decade. While based on relatively simple assumptions, and focusing on the results for India, the results are provocative. We begin by assuming that India maintains balanced growth, with investment sufficient for capital to grow at the rate o f output. Sustainable output growth would be given by the growth o f the labor force (adjusted for labor quality), and the rate o f TFP increase, scaled by labor's share o f income. Consider growth in the number o f workers ranging from 2 to 2.8 percent per year. (ADB [2005] projects India's labor force will grow at 1.8 percent through 2015. These figures are augmented to reflect prospects for increased labor force participation o f women, and reduced underemployment.) Next suppose the feasible range for India to increase average years o f schooling is from one to one and a halfyears over the next decade. (Note that the changes in Figure 2.3 reflect a twenty year period.) The implied increases in labor quality would add an additional 0.7-1.2 percent per year to effective labor force growth. Suppose that India can achieve TFP growth o f at least 1.5 percent per year, but perhaps as much as 3 percent per year. Finally, assume labor share equal to 65 percent, as for the growth accounts. The result i s GDP growth ranging from 4.9 to 8.7 percent per annum. Using the capital output ratio for India from our data, and assuming a 5 percent rate o f capital depreciation, the required investment ranges from 22 to 30 percent o f GDP. And with the other variables all at the top o f their ranges, India could achieve output growth o f 10 percent per annum with an investment share of 38 percent o f GDP. 2.36 These scenarios suggest that output growth rates o f about 5 percent per annum in South Asia are quite consistent with maintaining investment rates o f 22 percent o f GDP, the average since 1980. However, these scenarios support the view that sustained increases in the region's growth will require significant increases inthe investment rate, as well as efforts to increase labor force participation and increase worker skills through schooling. 45 ANNEX:GROWTHACCOUNTS 2.37 Earlier, growth accounting is used to examine South Asia's growth experience over the period 1960-2003. As discussedinBosworth and Collins (2003), we have usedthis methodology to construct consistent accounts for an additional 80 countries during 1960-2000.'3 Readers are referred to Bosworth and Collins (2003) for additional details about the data and methodology and for further discussion of its strengths and shortcomings. 2.38 A country's output inany givenyear dependson its factor inputs-labor and (human and physical) capital-as well as on the efficiency with which factors are used inproduction. Thus, the key pieces to the procedure are a growth accounting equation for constructing the decomposition, parameter assumptions, and data on output and factor inputs. 2.39 Define Y as GDP, K as the physical capital stock, and A as the level of technology. L is labor inputs (measured as "bodies of economically active persons") which we assume is "augmented" by H, an index o f the average "level o f labor quality," which we measure by average years of schooling. We also assume that a country's output can be expressed as a function o fthese inputs,usingthe specific functional form shownin (A.l).14 Y = A K ~ ( H L ) ( ~ - ~ ) (A-1) 2.40 We report our results in a form that decomposes growth in output per worker into the contributions from the growth o f physical capita per worker, the growth of education per worker, and the growth in total factor productivity, as shown below. (Lower case letters denote a variable's average annual growth rate.) yll= CY(kA)+ (1 - CY )h + a (A.2) 2.41 Given an estimate for CY and measures of Y, L,K, and H, it is straightforward to solve for A (or a) and construct the decomposition. As for our 84-country study, this decomposition assumes a capital share: CY = 0.35. An analysis that used the actual income shares in each period would allow for the consideration of a much wider range of underlying production functions. However, few countries are able to allocate the incomes of the self-employed between capital and 1ab0r.I~ 2.42 As noted above, Y is real GDP. L is employment for industrial countries and labor force for all nonindustrial countries. The capital stock measure i s constructed from investment data using the perpetual inventory method, with a depreciation rate o f 0.05. Finally, to construct H, we assume that human capital is directly relatedto average years o f schooling (S) ,and that there i s a 7 percent return to eachadditional year of schooling: H=(1.07)' (A.3) 2.43 Information about data sources for the 1960-2000 sample i s provided in Bosworth and Collins (2003). Data from World Development Indicators (2005) were used to update GDP, L, and K to 2003 for India, Bangladesh, Palustan, and Sri Lanka. Average years of schooling was interpolated to 2003 for these countries, basedon trends during 1990-2000. l3 Country coverage was determined by data availability. The main exclusions are the transition economies and countries with population less than one million. l4 Equation (A.1) assumes constant returns to scale. l5 See Bosworth and Collins (2003) for further discussion andreferences inthe cross-country context. 46 3. IMPROVING THE CLIMATE FORINVESTMENTAND BUSINESSIN SOUTH ASIAI6 INTRODUCTION 3.1 Increasing growth is a central goal o f policy makers interested in improving the opportunities and living standards o f their constituents. The sustained highgrowth rates of China continue to grab headlines. Recent growth in the South Asia region has also been impressive, although less consistent across the region. To understand why growth is higher in some locations than others and to highlight ways to foster higher growth, we need to understand the investment climate faced by entrepreneurs. The institutional, policy, and regulatory environment affects the ease o f doing business and the scope o f opportunities facing entrepreneurs. 3.2 Here we will focuses on the microeconomic underpinnings o f growth in the South Asia region. Decisions to invest, to hire new workers, or to expand production are made by firms every day. These decisions are then aggregated across the whole economy to give the country performance measures. Understanding the factors influencing these decisions from the firms' perspective can gwe insights into the aggregated dynamics and performance in the aggregate measures o f growth and productivity. 3.3 Growth can be achieved by increasing inputs, the amount of machinery, or human capital. It can also be achievedbyimproving productivity or producingmore with the same inputs. Most see the challenge as how to increase productivity. There are two components: The first i s related to technology. The second i s the environment, which enables entrepreneurs to work more efficiently, such as fewer difficulties accessing raw materials, fewer distractions dealing with bureaucratic harassment and demands for bribes, or less exposure to theft. 3.4 Total factor productivity (TFP) captures these last two sources. Much o f the focus has traditionally been on the first source, on differences in technology. More recently, attention has shifted to differences in the broader institutional, social infrastructure, or investment climate setting.17 The World Bank's World Development Report (2005) focused on the latest evidence this perspective brings to growth and poverty reduction (see World Bank 2004). We will compare countries across South Asia and with China on key dimensions o f the investment climate: access to infrastructure and financial services, the security o f property rights, the burden o f regulatory requirements, and the quality o f governance, providing evidence on how they can . improve productivity and the prospects o f doing business inSouth Asia. 3.5 In this view, the investment climate can be understood as a crucial link between the efforts o f sowing and reaping. If the local government is highly bureaucratic and corrupt, if government's own provision or regulation o f infrastructure and financial services i s inefficient so that firms cannot get reliable services, then returns on potential investments will be low and uncertain, and there would not be much accumulation and growth inthese environments. On the other hand, in developing locations that create a good governance and business environment, returns and accumulation should be high. Thus, the expected rate o f return on a project i s determined not only by the degree o f efficiency on the factory floor but also by the broader l6 This Chapterwas preparedby MaryC. Hallward-Driemeier, Senior Economist, World Bank.The views inthis paper are solely the author's anddo not necessarilyreflect those ofthe World Bankor its ExecutiveDirectors. 17 See Acemoglu andJohnson (2005), Hall andJones (1999), ParenteandPrescott(2000), and Klenow and Rodriguez- Clare (1997),. environment, including the ability to appropriate the returns. Infact, the costs associated with a poor investment climate can more than offset technical efficiency gains. DISAGGREGATING INVESTMENT CLIMATEMEASURES 3.6 New measures are available to benchmark the quality of a location's investment climate. Polls o f expert opinions including the Global Competitiveness Report, Transparency International, and International Country Risk Guide have provided subjective measures for more than a decade. Two recent initiatives provide more objective measures. The World Bank annual Doing Businessreport providestime and cost measures associated withfully complying with a set ofregulations according to statutory requirements. The World Bank Investment Climate Surveys also provide time and cost measurementsbased on the actual experience o f a wide range o f firms. The topics covered include key regulatory areas, access to finance and infrastructure services, as well as firmperformance measures. 3.7 Using these measures o f the investment climate in South Asia identifies areas where entrepreneurs could benefit from reforms. Providing comparisons across countries allow policy makers to benchmark conditions intheir countries. This can highlight areas for reform, providing demonstration effects o f what can be achieved. Subjective Rankings 3.8 The Global Competitiveness Index publishes a Business Competitiveness Index (see World Economic Forum 2004). In2004, 4 countries from South Asia were included in its list of 101 countries. India ranked thirty-seventh, Sri Lanka fifty-seventh, Palustan seventy-fifth, and Bangladesh ninety-first according to the perceptions o f expert respondents. The report then details specific areas o f relative strengths and weaknesses for each country. India's strong performance is due to the strength o f its pool o f scientists and engineers, quality o f management schools, and the availability o f locally based competitors. The factors identified as "most problematic" for doing business in the region were strikingly similar across the countries. Corruption, inefficient bureaucracy, inadequate infrastructure, and policy instability are the top concerns inBangladesh, Sri Lanka, Pakistan, and India. Restrictive labor regulations joins the list inIndia. 3.9 Loolung at measures o f institutional quality and governance the ranking shifts somewhat. Table 3.1 gives indicators based on the compilation o f all the major indices published from a variety o f sources, normalized around an average o f 0 and a standard deviation of 1. On these indicators Sri Lanka performs relatively better, except for the measure of political stability. Bhutan and the Maldives have reasonably favorable results, although they are based on a smaller number o f underlying ratings. Still, the concern regarding corruption inmuch of the region does come through, as does the dampening effect o fpolitical instability. 3.10 These subjective measures do provide a picture o f the relative strengths and weaknesses of different countries. But the subjective nature o f the rankings makes it difficult to make comparisons or to measure improvements over time. The World Bank has launched two extensive new efforts to fillthe gap and to provide more objective measures o f key dimensions of climate for investing and doing business (see Table 3.2). 48 Table 3.1: GovernanceIndicators I Government IPolitical IRegulatory IRuleof I Source: Kauhann, Kraay, andMastruzzi,2005. Note: Scores are centeredon 0 with astandarddeviationof 1. Table 3.2: New Data Sourcesfrom the World Bank Allow for ComparisonsAcross Countries Investment Climate Surveys Doing Business Country Launchedin2001; the reportdraws onmore than 50,000 Launchedin2003; 145 countriesare coverage firms in70 countries. Eachyear an additional 15-20 now covered each year. surveys are fielded. Secondroundsurveysare now underway, including inIndia, Pakistan, and China. Investment The standardquestionnaire of 82questions covers Ten areas of regulation: business climate regulations, governance,accessto finance, and registration, propertyregistration, dimensions infrastructure services. It also collectsdataon firm insolvency, contract enforcement, covered productivity, investment, and employmentdecisions. hiring and firing workers, accessing credit, customs, licenses, investor protection, andtax payments. Types of Coversboth objectiveand perceptiondata. The Objectivemeasuresof the number of variables objectivedataincludethe time to completeprocessesor procedures, the time to completethem, receiveservices andthe monetary costs of various andthe fees andcosts associated with disruptionsandregulations. Inaddition, respondents full compliance. giveperceptions ofpotentialconstraintsand assessments of risks and competition. Whose Surveys cover adiverserangeof sizes andactivities, Use a single, defined, hypothetical perspective with randomsamples of severalhundredfirms. Dataare firmandtransaction. Judgmentsbased gatheredthroughface-to-face interviewsconductedwith on assessmentof upto five local senior managers and accountants. experts(lawyers, accountants). Differences Samples cover multiple locationswithin each country. A single indicatoris given for the within a largest city inthe country. For some country largecountries, including India and Pakistan, additionalcities are available. Basis of Indicatorsare based on the experiencereportedby firms, Indicatorsmeasure formal regulatory assessment providingranges ofhow policiesare implemented in requirements. practice. ource: Adar d from World Bank 2004. Doing Business ExaminesRegulatoryHurdlesFacingEntrepreneurs 3.11 Doing Business provides detailed information on the procedures, their time, and costs of complying with a set of regulations fiom opening a business, accessing credit, hiring and firing workers, and closing a business (World Bank 2005b). Requirements for specific, tightly defined transactions can then be compared across countries and over time. The 10 regulatory areas are reported separately, with an aggregate index also reported. See World Bank (2005a). 49 3.12 Taking a simple average o f all the Doing Business indicators, India ranks in the bottom 25 percent o f the 155 countries, Sri Lanka at the midpoint, and Bangladesh, Nepal, and Pakistan in the second quintile. On the other hand, China is in the third quintile, while Malaysia and Thailand are inthe top 15 percent o f countries according to these measures. The differences with the Global Competitiveness Report is in part because these rankings only include the regulatory environment in 10 specific areas, and does not address other issues such as infrastructure services, availability o f skdled labor, or corruption. Certainly the regulatory environment is an extremely important component, but the ranlungs would likely shift ifother dimensions were included. 3.13 This simple ranlung also masks a number of features. First, looking across the 10 regulatory areas, the same country can do extremely well on some dimensions and rather poorly on others. It is not the case that the efficient registration o f property or businesses implies that contracts can be enforced easily or that investors will necessarily be protected. Thus, the Maldives ranks the best on the ease o f paying taxes, but is 134 out o f 155 on the ease o f getting credit. Bhutani s 42 on the ease o fregistering propertybut 143 on the ease o f closing a business. 3.14 Second, there i s considerable variation within the region as to where countries' relative strengths are. Thus, while the region as a whole generally makes it easy to register property, Bangladesh stands out with some particularly onerous requirements, being 151 compared to India's being 29. 3.15 Third, Doing Business' comparisons are based on the formal requirements and may or may not be a good proxy for how the laws and regulations are actually applied in practice. In fact, where there are gaps between the investment climate survey results and the Doing Business indicatorspoint to areas where reforms would be desirable and are often associated with opening upopportunities for bribes. UsingInvestmentClimate Surveysto IdentifyKey Challengesfor Improvingthe InvestmentClimateinSouthAsia 3.16 Results from the Investment Climate Surveys are consistent with the themes highlighted inthe perceptiondata of international experts as reflected by the Global Competitiveness Index. The investment climate surveys are based on large, random samples o f firms in specific sectors in four South Asian countries: Bangladesh, India, Pakistan, and Sri Lanka. For additional comparison, results are also compared to those of China. All the surveys used a common instrument and sample design. Key manufacturing sectors were included in each country based on their contribution to output and exports, including garments, textiles, leather products, food products, electronic and electrical equipment, metal products, and chemicals. Within each country, major industrial centers were included. The sample sizes are 1,001 inBangladesh; 1,900 in India; 879 in Pakistan; 452 in Sri Lanka; and 1,500 in China. These firms report the significant constraints are a unreliable access to electricity, corruption, a lack of government efficiency, and burdensome labor regulations (see Figure 3.1). 50 Figure3.1: Share of FirmsReportingthe Issuesas a "Major" or "Severe" Constraint on the Operation of Their Business Electricity Telecomltranspo Informality Skills available uncertainty Tax -Eastern Europe & Central Asia -South -EastAsia Asia and Pacific Source: Investment Climate Surveys. 3.17 Access to Reliable Electricity. Access to a reliable source of electricity topped the list o f concerns for the region as a whole. The level of the constraint i s particular striking compared to other regions in Asia. Firms in Bangladesh report that power interruptions are virtually a daily occurrence. Compared to all other regions, firms inSouth Asia were likely to lose a higher share of their output due to power losses. InBangladesh the losses amount to 3.3 percent of sales, which in turn i s much better than either Pakistan (5.4 percent) or India (5.5 percent) but significantly higher than in China (2.2 percent o f sales). For all o f these measures there i s also muchvariation within countries. InIndia, for example, Bangalore is similar to China at 2 percent o f sales lost, while Calcutta i s very poor for a large city (6 percent of sales). This i s due not only to time lost during an outage, but also the production spoiled owing to the interruption in the process andthe time neededto reset machinery. 3.18 Given the losses stemming from power failures, many firms respond by running their own generator. While it i s not uncommon for large firms in any location to have their own power generators, for small and medium enterprises (SMEs) the cost o f maintaining a power generator i s quite high and burdensome. Thus, another gauge o f the reliability o f the power supply is the proportion of firms that have their own generators. The patterns are similar; the share of firms with their own generator i s only 17 percent in the China sample but half of the SMEs sampledinSri LankaandBangladeshdo and 60percent inIndia andNepal. 3.19 Government regulations. The issue of corruption was the second greatest source of concern reportedinthe South Asia region as a whole. This correlates significantly with concerns that the interpretationo f regulations can be unpredictable. When officials have discretion inhow particular regulations will be implemented, there i s an opening for unofficial payments to determine the outcomes. The costs of these bribes and the uncertainty of the standards that will be imposedcan reduce the incentive to invest or to expand. The magnitude o f the issue is seen in the share of firms reporting to make such payments and the delays faced in the provision of public services. 51 3.20 The survey includes a number o f questions that get at the efficiency o f the government in providing services that are essential for firms. It includes a general measure o f the quality of government efficiency as experienced by firms. The share o f firms that consider the provision o f government services to be inefficient varies from less than 30 percent in China to 60 percent in Pakistan, Sri Lanka, and Nepal. The time management has to spend with officials can be more than 10 percent in India and Pakistan, while closer to 4 percent in Sri Lanka. Requests for additionalpayments to "get things done" are reported as extremely common. Across the countries the size o f bribes was between 2.2 and 2.5 percent o f sales. The efficiency of providing services is also illustratedby the time it takes to get hooked up to utilities. For a new phone line, firms have to wait on average 35 days inIndia and 130 days inBangladesh. 3.21 Particular attention is paid to two areas o f regulation given their importance to growth and the prospect of increased regional integration. They are customs administration and labor regulations. 3.22 Customs administration. Customs administration i s o f particular interest in assessing the constraints to better regional integration. The data show there i s variation in the time it takes to get goods through customs. This i s both at the level o f securing the necessary paper work as well as the actual inspection process. The Doing Business database reports on the official time to obtain the paperwork to clear goods, with the InvestmentClimate Surveys reporting on the actual time it takes to clear customs once goods have left the production facility. This variation not only introduces greater uncertainty in the delivery process, but it can mean the difference between fulfilling an export order or not. With buyers facing many potential suppliers, reliable delivery schedules can be key. Figure 3.2 shows the average time to clear customs, with India reporting just over nine days, while firms inPalustancan face double that. However, Figure 3.3 shows the extent o f variation aroundthese averages. Figure 3.2: Average Time to Clear Customs I 52 Figure 3.3: Extent of Variation to Clear Customs (Around the Averages) .-.-r 2 1 U 5 I.5 L m> 1 L0 0.5 6 0 3.23 A related issue i s the cost o f transportation. This is both across borders and withm a country. Firmsreport significant delays in transportation owing to poor infrastructure conditions as well as the need to pass often repeated inspections. Costs are even higher taking into account the potential o ftheft andbreakage that occurs duringthe transportation o f goods to market. 3.24 The impact o f such customs and transportation delays can be seen in different dimensions. One impact i s that firms respond by increasing their holdings o f inventories (see Figure 3.4). Such increasedholdings are needed to ensure against supplier delays and can tie up significant worlung capitalthat might have been directed to more efficient purposes. Figure3.4: Evidence from India That Poor Transportation RaisesInventory Costs Karnataka Tamil Nadu Kerala West bengal Uttar Pradesh Maharashtra Punjab Delhi 60 40 20 0 20 40 60 ElDaysofinventoryofmajorinputs 6 Percentidentifyingtransportas amajor-to-severebottlenecktogrowth Source: 2002 InvestmentClimate Survey inIndia. 3.25 Another impact can be seen on the ability o f firms to export. Here the experience with customs administrationis linked with access to infrastructure, finance, and public utilities. Figure 3.5 illustrates the results if the access to these services were improved by one standard deviation and what the implication would be on the probability that a firmwould be an exporter. The share could rise by over 10 percent in Chenai and double in Karachi (see the next section for more detailed results). 53 5 0 Potentia I 1 C a l c u t t a Chenai K a r a c h i Figure 3-6: Highest FirlsgCosts The highestfiring casts ar AS33 54 Nepal gladesh Pakistan IndiaI Sri Canka 0 1QQ 200 300 400 500 aays 55 rnportion 20 : G o ntract Enforcement Regu la tion 1 5 * D B r i b e% I 10 C rim e I s - infra stru ctu re I informal, Srnali, Medrum, Large, Very c 5 4 2 0 20-4 4 50-24 4 large, 2 on up LlNM Q FXRRX control for human capital and the potential endogeneity of input selections.20 TFP calculated from the data is thenrelated to differences inthe investment climate acrossthe firm locations. 3.33 Incontrolling for investment climate conditionsanumber ofthe variables were discussed in the preceding section. Production losses from power outages captures concerns about the access to electricity. Access to an overdraft facility is the measure of access to finance. A number of proxies of government efficiency are used. The days to get goods through customs andthe time to get a telephone connection are reported here. The share of firms reporting bribes, an index of overall government efficiency, and the confidence in the courts are also used with similar findings. 3.34 There are a number of potential concerns that need to be addressedinthis analysis. One i s that the firm's own investment climate indicators could be endogenous to their performance. This i s most likely the case for subjective opinions. However, it i s not necessarily clear which way the bias would go. Firms that are performing poorly may be more likely to complain and blame external factors inthe investment climate for their lack o f success or successful firms may be more constrained by a weak investment climate in expanding their business. A significant benefit of using objective data avoids issues of how much a given entrepreneur is prone to complain. However, even objective measures can still be endogenous. For example, a firm's performance could affect the probability that it faces more harassment (and demands for bribes) by officials; that is, growing or successful firms may be targets as they are more likely to attract attentionand becausethey are more likely to have the resourcesto make additional payments. To deal with this potential endogeneity, investment climate indicators are averaged for a particular sector ineach city. The interpretationi s then not the specific investment climate conditions o f a firmbutofthe broader locationinwhich itis operating. 3.35 Another potential source of endogeneity regards the location choice o f firms; that is, that more productive firms chose to locate in the better environments. This can be controlled for by restricting the analysis to exclude firms that are mobile (that is, excluding multinationals and larger domestic firms). Migration within these countries is remarkably low, and other surveys confirm that the location choice of domestic firms i s overwhelmingly determined by where the founder was born or brought up. So we define the latter group as the subsample of domestically owned small firms employing no more than 150workers.21 3.36 There could also be a competing hypothesis to explain differences inperformance across locations. The literature on economic geography points to the importance of the size of the local market. To explore this, the regressions include the population o f the local city, the distance to the nearest port, and the distance to the nearest major overseas market. The investment climate results are indeed robust to their inclusion. This holds out the potential for sound investment climate policies to be able to offset certain disadvantages of geography. 3.37 Finally, the results are robust to including country dummies, so that identification comes only from the variation across locations within a country. Even where policies may be set nationally, they are often implementedlocally. The results confirm that local governance matters. 2o The technique outlinedby LevinsohnandPetrin(2000) was used. Resultsare available uponrequest. 21 We repeatedthe analysis using a cutoff of 100 workers and even 50 workers and the results were robust to these cutoffs, too. 57 InvestmentClimateand Productivity 3.38 Table 3.3 shows the results o f regressing firm productivity on the objective investment climate indictors. The indicators are averaged at the city level to control for endogeneity, and errors are clustered to correct for potential correlation within cities. The signs o f the coefficients are all as expected. Increasing the time it takes through customs, the longer it takes to receive a telephone connection, and the greater the losses from electricity are all associated with lower productivity. Greater access to finance, on the other hand, is positively correlated with productivity. The coefficients are significant. Adding in controls for the geographic characteristics (the size o f the local market and distance to other major markets) do not remove the significance o f the investment climate indicators. The results are robust even after adding country dummies so all the variation is coming from differences within a given country. * significantat 10 percent;** significantat 5 percent;***Significant at 1 percent. 3.39 The third column looks only at the subsample o f small domestic firms to assuage concerns that the results are driven by large, mobile firms choosing to locate inbetter investment climates. If better performing firms have a tendency to locate in more hospitable investment climates, the effect o f deficiencies in investment climate that estimated for this group o f firms should be smaller than the estimates for the full sample. On the other hand, if inherently more productive firms tolerate more deficiency in investment climate, estimates o f the effect o f investment climate on performance would be larger for this group than the estimate for the full sample. That the coefficients are marginally larger is consistent with the latter view; that is, more mobile firms are better equipped to deal with or have private means o f circumventing weak investment climate conditions. 3.40 The results are also economically significant. If investment climate conditions were to improve by one standard deviation, productivity could go up by 2 percentage points. If the conditions in India were the same as those enjoyed by their Chinese competitors, productivity could be one percentage point higher, much o f the gap because of the differences in access to reliable electricity. 58 3.41 Table 3.4 extends the approach to look at the impact o f investment climate variables on wages and on investment rates. The results in the first column show that firms in good investment climate locations pay higher wages. This is true even controlling for a number of additional firm characteristics that can be associated with higher wages, such as the average schooling and age (experience) of the employee and whether the firm i s a multinational. This i s still true including country dummies. However, including the economic geography variables in the second column does soften the results somewhat. The economic impact i s also significant. A one standard deviation in the investment climate variables is associated with wages that are 25 percent higher. IfPalustan and Bangladesh were to raise their investment climate variables to the level o f China, wages could be 23 and 31percent higher. Table3.4: A Better InvestmentClimateis AssociatedWith Higher Wages andInvestmentRates Note:Variables are in logarithms. Robust t statistics inparentheses. *significant at 10percent;** significantat 5 percent; *** significantat 1percent. 3.42 The results are also robust looking at the impact on the rate o f investment by firms. The impact of a one standard deviation improvement in the average investment climate conditions could raise investment by a third. InvestmentClimateandExporting 3.43 Table 3.5 looks at the effect the investment climate conditions have on the probability that a firm will export at least some o f its output. These have implications for the types o f constraints needed to be addressed to further the regional integration in South Asia. The geographic controls are included, as are country dummies. The results are very robust. Additional robustness checks are reported using alternative measures of the efficiency o f 59 government services. Concerns that better firms are more likely to both export and select to locate inbetter investment climate cities, the fifth column reports the results for the subsample o f small domestically owned firms that are not likely to be mobile. Again, the results are robust. Improvementsin the investment climate o f one standard deviation are associated with raising the probability that a firmexports by 30 percent. Table 3.5: InvestmentClimateConditionsAffect the ProbabilityaFirmWill Export Countrydummies Yes Yes Yes Yes Yes Sectordummies Yes Yes Yes Yes Yes Observations 4386 4461 5370 5288 3953 X" 2 186 292 269 465 118 Prob>X"2 0.0 0.0 0.0 0.0 0.0 PseudoR 2 0.13 0.17 0.18 0.24 .12 included.'Errors are clusteredby'cities. Robustz statistics&e reported. ** percent;*** *significantat 10percent; significant at 5 significantat 1percent. CONCLUSIONS 3.44 A range o f indicators were provided that measure the investment climate in South Asia, allowing for comparisons of countries in the region with each other and with China. Access to electricity, access to finance, and efficient government regulation in areas such as customs administration, corruption, and property rights are raised as priority areas by entrepreneurs themselves. 3.45 Linking these indicators to firm performance shows that plants that exist in better investment climate locations are more productive, producing more value with a given level o f capital and labor. The benefits are enjoyed as higher profits, but are also shared with workers, with firms in better investment climate locations paying higher wages. These superior returns also serve to encourage greater capital accumulation. Given the large differences in investment climate that we find in our surveys, it i s not surprising that growth rates vary so much across these locations 60 4. PROPERTYRIGHTSINSTITUTIONS,CONTRACTINGINSTITUTIONS, AND GROWTHINSOUTHASIA: MACROAND MICRO EVIDENCE" INTRODUCTION 4.1 There i s by now widespread consensus among academics and development practitioners that institutionsare important for economic development. Here we will examine how countries in the South Asia region fare interms o f their institutional quality and the consequent development impact. A first glance at the data suggests a number o f puzzles that make the performance of South Asian counties of considerable interest. Bangladesh, and to a lesser extent India, rank poorly inmost cross-country ranlungs o f corruption and regulation, yet have grown quite quickly in recent years. Both countries, and Sri Lanka as well, have had sharp growth accelerations around 1980 without obvious changes in various measures o f governance or institutional quality.23Turning to the long-run relationship between institutions and levels o f per capita income, it is striking that all countries in the region fall below the simple regression line of per capita incomes on measures of institutional quality. Inother words, despite recent rapid growth, these countries have not yet reached the per capita incomes typical of countries with similar levels of institutional quality. 4.2 A naive interpretationof these facts might suggest that institutional performance has not been important for economic performance for countries inthe region. Inthis paper we dig deeper into both cross-country and within-country fm-level evidence in order to come up with a more nuanced view. We begin by following Acemoglu and Johnson (2005) in distinguishingbetween "property rights institutions," which measure the extent to which private property i s secure from predationby the state, and "contracting institutions," which measurehow well institutions such as the courts allow private parties to contract with each other. The former may be thought of as institutions that mediate the vertical relationshipbetween individuals or firms and the state, while the latter mediate the horizontal relationshipbetween f m s and individuals. 4.3 While conceptually this distinction i s a useful one, coming up with empirical proxies i s difficult. Inthis paper we measure the (absence of) good property rights institutions using the Kaufmann, Kraay, and Mastruzzi (2005) measure of corruption. While perceptions of corruption are clearly not institutions themselves, we argue that the prevalence of corruption i s a good proxy for the absence of well-functioning institutions that limit the arbitrary exercise of state power. In particular, since corruption i s defined as the use of public office for private gain, the tahng of bribesbypublic officials canbe thought ofas the expropriationofprivatepropertybythe ~tate.2~ 4.4 We measure contracting institutions using an estimate o f the length of time required to resolve a dispute over an unpaid commercial debt, constructing by the World Bank's annual Doing Business report. This measure is the same as the one used by Acemoglu and Johnson 2' This Chapter was prepared by Ana Margarida Fernandes, Economist, and Aart C. Kraay, Lead Economist, the Development EconomicResearch Group Unit, World Bank. The views in this paper are solely the authors and do not necessarilyreflect those ofthe World Bank or its ExecutiveDirectors. '3 Consistent with common usage we will use the terms "governance" and "institutional quality" interchangeably throughout the paper. Below we spell out in more detail precisely which dimensions of governance or institutional quality are of interest for this paper. 24 We prefer this measure to the one used by Acemoglu and Johnson (2005). They use a measure of "expropriation risk" produced by ICRG, a political risk consultancy. ICRG produced this index between the mid-1980s and mid- 199Os, but since has discontinued it. We prefer our corruption measure because it covers more countries, is more timely, and is based on a largenumber of individual measures of corruptionand hence is likely to be more informative thananyindividual source. The ICRGmeasureishoweverhighly correlatedwith our corruptionmeasure. (2005). It measures the number of legal procedures, and the estimated number of days to complete them, for a plaintiff who successfully collects on an unpaid debt. This measure emphasizes statutory or de jure procedures that must be followed, and does not attempt to measure the de facto procedures that are actually followed. These may differ significantly fiom the de jure ones, particularly in countries where courts are Clearly, there other possibilities for empirical proxies for property rights institutions and contracting institutions. For example, measures o f expropriation risk provide direct evidence on perceptions of the risk of overt predation by the state. Similarly, excessively burdensome regulation makes contracting between private parties more cumbersome (for example, strict labor market regulations impede efficient contracting betweenf m s and workers). 4.5 We next ask how countries in South Asia fare on these two key dimensions of institutional quality. In order to do so it is useful to have some hnd of benchmark for institutional performance. 4.6 We obtainbenchmarks fiom the growing literature on the deep historical determinants of institutional quality. Inthe case ofproperty rights institutions, Acemoglu, Johnson, and Robinson (2001) have argued that the strength of modem institutions can be traced back to the interaction between countries' geographical features and their colonial experience. Inparticular they argue that in countries where the disease environment was inimical to European settlement in the eighteenth and nineteenth centuries, colonial powers were more likely to set up "extractive" states whose main purpose was to exploit local resources. In contrast, in more hospitable climes colonial powers were more likely to set up the institutions o f limited government that served the interests of European settlers who established themselves in the colonies. The very different experiences of Britishcolonial power inthe Gold Cost in West A h c a versus inCanada or New Zealand provide examples of these two extremes. They show that these early differences in institutions cast long shadows that are still seen in cross-country differences in modem institutions. 4.7 We use their measure o f settler mortality in the eighteenth and nineteenth centuries to measure this deep determinant of property rights institutions. Turning to contracting institutions, Djankov et al. (2003) have shown that countries that inherited common-law legal traditions tend to have more effective courts, inthe sense of simpler and faster procedures and fairer outcomes. We use a dummy variable indicating British legal origins to capture this deep historical determinant o f contracting institutions. 4.8 After controlling for the deep historical determinants of these two types of institutions, we are better able to assess how well or how poorly countries in the South Asia region fare on these distinct dimensions of institutional quality. For example, we find that while India has fairly average performance on contracting institutions when loohng at all countries, it stands out as being one of the poorest performers among countries with British legal origins, a deep determinant o f this type of institutions. By contrast, it fares reasonably well on measures of property rights institutions. The reverse i s true for Bangladesh, which stands out for having quite poor property rights institutionsbutreasonablecontractinginstitutions. 4.9 We next document the Acemoglu and Johnson (2005) core finding that of the two types of institutions, property rights institutions matter muchmore for development than do contracting 25 The possibility of corruption in the courts raises an interesting question: Does this signal the failure of property rights institutions or contracting institutions? The answer is surely a bit of both. On the one hand, bribes paid by privatepartiestojudges who are public officials representstate expropriationas discussed.At the same time corruption likely makesthe conflict resolutionservicesprovidedby courts less efficient. 62 institutions. The rationale that they offer for this finding i s that private parties can respond to weak formal contracting institutions by developing alternative informal institutions. Incontrast, private parties have little recourse to alternatives when they face a predatory state. For this reason contracting institutions appear to matter less for long-run growth than secure property rights institutions. 4.10 We also ask where countries in the region fall in the causal relationship running from these two types o f institutions to economic performance. While, as noted, all countries in the region seem to fall below the simple regression line, once we control for the direction o f causation a rather different situation emerges. W e find that India and Bangladesh again stand out inthis estimateo fthe causal impact ofinstitutions on development. Indiastands out for having a large negative residual, suggesting that even India's good performance on property rights institutions has not yet had the development impact that cross-country analysis would suggest. By contrast, Bangladesh stands out injust the opposite direction, with a predicted level of per capita income based on its institutional quality that is much lower than its actual income levels. One interpretation o f this findingis that India has a great deal o f scope for future growth given its relatively good institutional performance, while Bangladesh's current relatively high income levels (gwen its weak institutional quality) are more likely to be unsustainable. 4.1 1 We next turn to large cross-country firm-level datasets from two main sources, the World Business Environment Survey (WBES) and the World Bank Investment Climate Surveys (PICS) to dig deeper into the relationship between different types o f institutions and economic performance at the firm level. We first identify specific questions from these two firm-level datasets that can be used to measure these distinct types o f institutions. Firms'perceptions about the quality of the courts capture contracting institutions and their views on corruption capture property rights institutions. W e then show that there is a broad correspondence betweenthe firm- level data on institutional quality and the cross-country data described above. Countries with poor contracting and property rights institutions according to the cross-country measures also have poor functioning o f the judiciary and corruption as a major obstacle to business according to the WBES and PICS firm-level measures. Interestingly, we generally find that for South Asian countries institutional quality i s worse according to the firm-level measures than according to the cross-country measures. 4.12 We then use the firm-level data for India, Palustan, Bangladesh, and Sri Lanka to investigate the consequences o f successes and failures in the provision o f these two types o f institutions. We find that there i s a great deal o f dispersion inthe two types o f institutions across firms and across cities within countries. Interestingly, perceptions o f the two types o f institutional quality are not very highly correlated: Cities where firms find corruption to be a problem are not necessarily cities where firms findpoorly functioning courts to be a problem, and vice versa. 4.13 This dispersion inthe two types o f institutional outcomes i s useful because it provides us with an opportunity to examine the consequences o f successes and failures in both dimensions. W e find that for firms in Bangladesh, India, Pakistan, and Sri Lanka, property rights institutions have a more detrimental effect on performance-measured by salesper worker-than contracting institutions. A possible interpretation for this finding is that firms can find ways to contract around poorly performing contracting institutions while that i s not possible for weak property rights institutions. Our evidence supports this interpretation: Incities and industries with weaker institutional quality, firms resort more frequently to informal channels-for example, belonging to a business association and usingit to resolve business disputes-to get around those. W e also show that in industries that rely more heavily on contracting institutions, firm-level concerns 63 about institutional quality are higher, but at the same time firms get around the weaknesses in contracting institutionsbyusinginformal channels more frequently. 4.14 Overall, our evidence suggests that governance or institutional quality does matter for economic development, around the world and for countries inthe South Asia region. The poor performance of countries such as Bangladesh and Pakistan on cormption-based not just on outside assessments but particularly so based on the views of firms in these countries-suggests that this is an area where reforms, although difficult, are likely to have substantial impact. Our evidence for India points to deficiencies in contracting institutions, which we primarily measure as weaknesses in the ability o f courts to settle business disputes, but which can be interpreted more broadly as failures in the overall regulatory environment. While cross-country evidence suggests that institutional weaknesses in this dimension have smaller development impacts than do property rights institutions, this does not mean that there are no retums to improvements in this area. In fact, our firm-level evidence suggests that firms need to develop alternative strategies to circumvent weak contracting institutions, and while we do not provide direct evidence, it i s very plausible that these alternative strategies are inefficient compared with the benchmark of good contracting institutions. CROSS-COUNTRY EVIDENCE ONINSTITUTIONSAND GROWTH 4.15 As noted, there is a growing consensus among academics and policy makers that good institutional performance is o f fundamental importance for economic development. This consensus goes back at least to the seminal work o f Douglass North. Intheir classic study of the economic effects of the Glorious Revolutioninseventeenthcentury England, North and Weingast (1989) document how the emergence of a limited monarchy subject to checks and balances by Parliament made the commitments of the state more credible. In particular the Glorious Revolution was followed by a flourishing of public debt markets as private parties came to trust the state's promises to service public debt, which in turn catalyzed the development of private financial markets inEngland. Taking a similar historical perspective, Delongand Shleifer (1993) empirically documented that cities governed by absolute rulers subject to no checks on their authority grew more slowly than those with more limitedgovernment duringthe 800 years before the industrial revolution. 4.16 A large body of cross-country empirical work has identified the importance of institutional quality for either growth or income levels. Mauro (1995) and Knack and Keefer (1995) documented the relationship between corruption and economic growth, while Hall and Jones (1999), Rodrik, Subramanian, and Trebbi (2004), Rigobon and Rodrik (2004), and Alcala and Ciccone (2004) have shown that institutions exert an important causal effect on per capita incomes across countries. 4.17 A growing body of research also has looked for more disaggregated evidence on the importance of institutions. For example, Banejee and Iyer (2005) show that Indian districts in which property rights were historically given to landlords had lower levels of investment and productivity than districts where property rights were assigned cultivators, arguing that this reflects the greater security o f tenure innonlandlord districts. 4.18 The starting point for many discussions of the role o f institutional quality, or good governance, in development i s the strong positive correlation observed between measures of institutional quality and log-levels o f per capita income. Figure 4.1 shows one such typical relationship. On the horizontal axis we have a widely used measure o f rule of law producedby Kaufmann, Kraay, and Mastruzzi (2005). These authors combine information from a large 64 number o f cross-country sources measuring perceptions of governance and construct composite indicators summarizing these perceptions. The rule o f law measure in particular captures the perceptions o f individuals, firms, commercial risk rating agencies, nongovernmental organizations, think tanks, and multilateral development banks on issues relating to the protection o f property. For example, it captures perceptions of the likelihood that property will be expropriated by the state, the likelihood that contracts will be enforced, the likelihood that property i s secure fkom crime, and so on. Throughout the paper, we will orient our data such that higher values for measures o f institutional quality mean worse outcomes. Figure4.1: Rule of Law and Per CapitaIncomes 10.5 - 4.19 The strilung feature o f such a graph, and many others like it, i s that there i s a very strong negative correlation between this broad measure o f institutional quality and levels of development: Countries with worse institutional quality are on average poorer than countries with good institutional quality. A further striking feature o fparticular interest for this paper i s the relative position o f countries in the South Asia region. All six countries in the region for which we have data fall below the regression line, as shown in the Figure.26Bangladesh, Pakistan, and Nepal have similar fairly poor scores on this measure, which place them inthe bottom quartile of all countries, while India, Sri Lanka, and Bhutan fare much better, around the median o f all countries worldwide. 4.20 Interpreting these striking regularities i s more difficult. Consider, for example, the position o f the six countries in the region relative to the regression line. Should we interpret this graph horizontally and conclude that these countries have much better institutional quality than one might expect given their relatively low income levels (that is, emphasize the fact that the countries fall to the left o f the regression line)? Or should we interpret the graph vertically and observe that these countries have substantially lower income levels than we would expect given their governance performance (that is, emphasize the fact that the countries fall below the regression line)? Clearly sorting this out requires an understanding o f the causal relationship between institutions and per capita incomes, not merely the correlation shown inFigure 4.1. The second problem o f interpretation has to do with the measure o f institutional quality itself. What 26We use the World Banksregional groupingof countries. The South Asia regionalso includes Afghanistanand the Maldives, for which we do nothave data inFigure4.1. 65 exactly does a measure such as rule o f law mean? Do countries that fare poorly on this measure have a highrisk of expropriationby the state? Or do they fare poorly because crime is an issue, or because private contract enforcement i s weak? And do all of these distinct ingredients of rule o f law matter equally for development outcomes? 4.21 Inthispaper we use a framework developedby Acemoglu and Johnson (AJ)(2005) to make progress on these fundamental questions of interpretation inthe context of countries inthe South Asia region. AJ suggest that it is important to unbundle measures of rule o f law into two distinct dimensions. The first o f these, which they refer to as "property rights institutions," capture the extent to which private property i s secure from predation by the state (for example, through outright expropriation) or, less dramatically, from corrupt officials demanding bribes in exchange for favors to the firm or individual. The second, which they refer to as "contracting institutions," captures how the effectiveness o f institutions that are used to mediate disputes between private parties, such as the courts and the judicial system. Fromthe definition o f rule o f law given above, it is clear that this aggregate measure combines-and conflates-these distinct conceptions of institutional quality. MeasuringPropertyRightsInstitutionsandContractingInstitutions 4.22 The first step i s to find empirical proxies for these two distinct types o f institutions. In our main results, we use a measure of corruption also produced by Kaufmann, Kraay, and Mastruzzi (2005) to measure property rights institutions. As noted inthe introduction, we reason that corruption reflects the illicit capture o f private property by corrupt public officials, and as such constitutes expropriation. We orient this variable such that higher values correspond to greater corruption. We also follow AJ inusing a novel measure o f the functioning o f the courts to capture contracting institutions. We, and they, use a variable that measures the number of days, and the number of formal procedures, that are required to resolve a hypothetical business disputebetweentwo privateparties over anunpaiddebt. This measurewas originally constructed byDjankov et al. (2003) and is updatedand maintainedinthe DoingBusiness reports. 4.23 Higher values, both in terms o f the length o f time, and the number o f procedures, requiredto settle a business dispute correspond to worse institutional outcomes. Inthe annexes we also report results obtained using two measures of institutional quality preferred by AJ, but which are available only for a smaller sample o fcountries. These are a measureo fperceptions of the likelihood o f expropriation, reported by Political Risk Services, a commercial risk rating agency, and "executive constraints" reportedinthe Polity IV database, which measures the extent to which the power ofthe executive i s limitedby other branches of government. The rationale for the latter i s that expropriationis less likelyincountries where executive power i s limited. DeepHistoricalDeterminantsofInstitutionalQuality 4.24 The secondkey step ininterpreting this evidence i s to find deep historical determinants of these two types of institutional quality that can be used to statistically identify the effects o f institutions on development. The basic problem i s straightforward. The simple correlation between institutions and per capita incomes in Figure 4.1 does not tell us anything about causation. It could be that better institutional quality causes higher incomes, or it could also be that richer countries can afford better institutions. Sorting out the direction o f causality i s typically done using instrumental variables. These are variables that can explain cross-country differences in institutional quality but have relatively little direct effect on per capita incomes today. By exploiting only the cross-country variation in institutional quality that is driven by 66 these deep historical determinants it is possibly to statistically identify the causal effects of institutionson long-run economic performance. 4.25 Ina highly influential paper, Acemoglu, Johnson, and Robinson (2001) propose using mortality rates o f European settlers around the world inthe eighteenth and nineteenth centuries as an instrument for property rights institutions. As discussed in the introduction, the argument is that in countries where settler mortality was low, it was inthe interests o f the colonial powers to create institutions that protected property rights, at least those o f the occupying settlers, and these institutions persisted to the present. In contrast, in countries where settler mortality was high, colonial powers were interested only in extracting wealth and had no interest in developing any lundof formal institutions. We follow this approach here, using an expanded dataset on settler mortality.*' We also follow AJ inusingdata on the legal traditions o f countries as an instrument for contracting institutions. Here the insight i s that countries with British legal origins tend to have much more streamlined procedures for dispute resolution, as opposed to the highly formalized and codified procedures in countries particularly with French legal origins. This effect i s reflectedinthe complexity or current dispute resolutionmechanisms. Table 4.1: First-Stage Regressionsfor Property Rights and Contracting Institutions Ruleof law Corruption Number In (number of procedures days) Britishlegal origins -0.268 -0.268 -9.886 -0.32 1 (0.173) (0.177) (3.436)*** (0.143)** In (settler mortality) 0.336 0.327 1.154 0.084 (0.078)*** (0.078)*** (1.158) (0.054) India 0.057 0.384 13.604 (0.1lo)*** 0.433 (0.173) (0.188)** (2.465)*** Bangladesh 0.696 1.029 2.161 0.248 (0.160)* ** (0.171)*** (2.649) (0.107)* * Sri Lanka -0.125 0.111 -9.813 0.437 (0.161) (0.172) (2.636)*** (0.107) * * * Pakistan 0.846 1.029 19.920 0.383 (0.184)* ** (0.202)*** (2.376)*** (0.115)*** Constant -1.006 -1.073 (5.8lo)* 31.800 5.614 (0.391)** (0.400)*** ** (0.293)*** Observations 82 82 82 82 R squared 0.34 0.32 0.22 0.14 Note:Robust standarderrors inparentheses.All measuresof institutionsare orientedso that highervalues correspond to worse institutions. * significantat10percent; ** significant 5 at percent;***significantat 1percent. 4.26 In Table 4.1 we summarize the relationship between these deep determinants of institutional quality and institutionaloutcomes. We regress our four measures o f institutions (rule o f law, corruption, number of procedures, and the logarithm of the number of days) on the two instrumental variables (a dummy variable identifying countries with British legal origins and settler mortality). We also include dummy variables for the four major countries in the South Asia region: Bangladesh, India, Pakistan, and Sri Lanka. These dummy variables capture the extent to which these countries differ significantly from the average relationship estimated using a large cross-country sample o f 82 countries. The main regressions o f interest are in the second *' This i s based on ongoing work by Kaufmann, Kraay, and Mastruzzi, who use Acemoglu, Johnson, and Robinson (2001) estimates of settler mortality but extendthem to geographicallyproximate countrieswith similar climates that were not included in the original Acemoglu, Johnson, and Robinson (2001) analysis owing to the limited country coverageoftheir preferredmeasure ofinstitutionalquality. 67 and third columns. For corruption, we find that settler mortality enters very significantly and positively; that is, countries with higher settler mortality 200 years ago have worse corruption today, which we interpret as worse property rights institutions. We also find that British legal origm is very significantly associated with the complexity o f dispute resolution procedures, as measuredby number o f procedures or number o f days. Countries with Britishlegal origins have substantially lower values for these two measures, indicating that they tend to have better contractinginstitutions.28 These findings mirror those o fAJ inthe substantially smaller sample of 51countries that they study. 4.27 Of particular interest for this paper is how countries in South Asia fare in these relationships. All four countries have Britishlegal origins, and so inassessingthe quality of their contracting institutions we should focus on how they fare relative to other countries with British legal origins. The dummy variables for India and Pakistan are positive and highly significant, which tells us that these countries have substantially worse contracting institutions than other countries with British legal origins. For number o f procedures, for example, the estimated coefficients imply that inIndia and inPalustan it takes 13.6 and 19.2 days longer (respectively) to resolve a business dispute than it does in a typical British legal origin country. T h s i s summarized graphically in Figure 4.2. Inboth panels o f this graph we plot corruption on the horizontal axis and number o fprocedures on the vertical. Inthe top panel, the green dots identify countries with Britishlegal origins. Although both India and Palustan are around the middle of the pack for all countries taken together on contracting institutions as measured by number o f days, they are in the worst five performers among the group o f countries with British legal origins. In short, India and Pakistan both perform quite poorly on this measure of institutional quality. Incontrast, the dummy variable for Sri Lanka enters negatively, and in the graph it i s clear that Sri Lanka ranks among the best o f all countries interms o f speedy dispute resolution. Figure4.2: PropertyRights and ContractingInstitutionsin SouthAsia q . . 80 - 70 - . U .. y .B.llOX + 31.44 e 60 - e - . R'.0.19 U 5 0 - .. '.*Pakistan u y = 6 . 8 0 ~+ 31.44 ! 9Pakistan I , , , , I -3 -2.5 -2 -1.5 .t 4.5 0 0.5 1 1.5 2 -3 2 5 .2 -15 -1 -05 0 0 5 1 1 5 2 Corruption Corruptlon Vote: All variables refer to 2004. Corruption i s oriented so that higher values correspond to worse outcomes. In the left panel the large green dots identify counties with British legal origins. In the right panel the large green dots identify countrieswith lower-than-averagesettler mortality. 4.28 Turningto property rights institutions, we see that the dummy variables for Pakistan and Bangladesh inparticular are very large, positive, and significant. The interpretationo f this is that these two countries have particularly poor performance on the corruption variable, even after taking into account the deep historical determinants o f this type o f institution. This i s also *' As a narrow, but important, technical point, note that settler mortality is a good predictor of property rights institutionsbut not of contractinginstitutions, while the opposite is true of British legal origins. This is crucial for the instrumentalvariables strategyto work. Intuitively, both instrumentspredictedbothtypes of institutions,andthen they would not behelpful for isolatingthe separatecausaleffects ofthe two types of institutions. 68 visually striking from the right panel o f Figure 4.2. Inthis panel, the green dots correspond to countries with low settler mortality. These are countries where we should expect that their historical situation was favorable to developing strong property rights institutions. However, Pakistan and Bangladesh stand out as being among the worst performers o f all countries in this group, suggesting fundamental problems with corruption. 4.29 In summary so far, the cross-country data suggest that Pakistan stands out for its poor performance inboth types of institutions. Bangladesh stands out for having poor performance on property rights institutions, but reasonable performance on contracting institutions, while the converse is true for India. Finally, Sri Lanka has very good contracting institutions but only average property rights institutions. We note also that we find broadly similar results when we look at other measures of property rights institutions ina smaller sample of countries, as reported inAnnex 1. Do InstitutionsMatter for Development: IdentifyingCausation 4.30 The next step i s to assess how these differences in institutional quality have translated into economic development, as summarized by real GDP per capita. We report the results of the instrumental variables estimates o f the causal impacts o f these two types o f institutions on real per capita GDP in Table 4.2. Consistent with the findings of AJ, we find a clear pattern that property rights institutions, as proxied by corruption, have a very large and significant causal impact on development, while contracting institutions, as proxiedby either number o f procedures or number o f days, matter much less. The magnitude of the effect of institutions on development i s very large. For example, a one-standard-deviation improvement in corruption, corresponding roughly to the difference between Sri Lanka and Palustan or Bangladesh, would result inincomes that are higher by a factor o f 6 in the long run. The result that property rights institutions matter much more than contracting institutions i s quite intuitive. As AJ argue, a key feature of contracting institutions such as the courts i s that firms and individuals have a variety of opportunities for circumventing them if the services they provide are weak. For example, firms might rely on business associations or informal networks to enforce contracts if the courts are ineffective or slow. Incontrast, it is much more difficult for firms to contract around a predatory state when propertyrightsinstitutions are weak. 4.3 1 It is again ofparticular interest to see how the four South Asian countries differ from this average relationship. Consider for example the regression which uses the number of days to proxy for contracting institutions. Here we see a very strikingpattern inthe dummy variables for the four countries. Bangladesh and Pakistan stand out for having very significant positive residuals, while India stands out with a modestly significant negative residual. The interpretation o f this pattern i s that Bangladesh and Palustan have per capita incomes that are substantially higherthan expected given their institutionalquality inthesekey dimensions. The point estimates suggest that Bangladesh and Pakistan have per capita incomes that are between two and three times higher than their very weak institutional performance would suggest based on average cross-country relationships. Incontrast, India's per capita income i s only about half of what i s expected given its per capita income. Finally, Sri Lanka falls more or less on the average cross- country relationship, witha dummy variable that i s not significantly different from zero. 69 Table4.2: InstrumentalVariablesRegressionsof In(GDP Per Capita) on PropertyRightsand ContractingInstitutions Corruption -1.843 -2.23 1 (0.419)*** (0.632)*** Numberofprocedures 0.070 (0.038) * In(number ofdays) 2.467 (1SOO) India -0.661 -0.633 (0.269)* * (0.354)* Bangladesh 1.185 1.121 (0.494)** (0.420)*** Sri Lanka 1.068 -0.651 (0.607)* (0.490) Pakistan -0.137 0.704 (0.341) (0.219)* ** constant 6.114 -5.939 (1.200)*** (8.572) Observations 82 82 Note: Robuststandarderrors inparentheses. Corruptionis orientedso that higher values correspondto worse outcomes. * significantat10percent; ** significantat 5 percent;*** significantat 1percent. 4.32 Inannex 2we showhowtheseresults differ ina smaller sample ofcountries that iscloser to that studied by AJ, in which we can use their preferred measures o f institutional quality. The finding that India has relatively low incomes given its institutional quality comes through very consistently and significantly in this sample as well while Bangladesh stands out for its higher- than-expected income levels, although not consistently so. The evidence for Sri Lanka and Palustan is somewhat more mixed. 4.33 How do we interpret these findings? By far the clearest findings we have are that India i s much poorer than expected based on the cross-country relationship between institutions and economic performance while Bangladesh, and to a lesser extent Pakistan, i s far richer than expected. A somewhat naive interpretation o f this evidence focusing on these countries alone might be that institutional quality simply does not matter for these countries. After all, our evidence suggests that Bangladesh and Palustan have managed to attain quite high per capita incomes despite poor institutional quality while India has lagged despite its relatively good performance onthe property rights institutions that matter most for growth. 4.34 W e think, however, that this interpretation is too narrow, because it ignores the quite strong cross-country evidence that property rights institutions do matter on average for economic performance. A more nuanced view mightbe that institutions do infact matter for growth, but in the case o f India they have not yet delivered their full development impact. The good news in this interpretation is that India has considerable room to grow based on its reasonably good property rights institutions. The flip side o fthis is a rather more gloomy view for Bangladesh and Pakistan, where it could be argued that their past income gains, whatever their source, are fragile because they are not supported by commensurately strong property rightsinstitutions. 4.35 Based on our reading o f the evidence on institutions and growth, we prefer the latter interpretation. However, we recognize that cross-country analysis based on these rather crude proxies for property rights institutions and contracting institutions can only bring us so far in understanding what the causes and consequences o f institutional weaknesses are. They also provide relatively little guidance for policy advice. 70 4.36 Inthe rest ofthis paper we therefore turnto more detailedfirm-level data from a large number of countries, and from major South Asian countries in particular, to probe further the links between different types of institutions and economic performance at the firm level. In particular we ask whether firms' perceptions o f institutional quality mirror those in the cross- country macro data that we have used so far. We also examine the evidence on how these types o f institutions matter for firm performance, and show how firms make use o f alternative, and possibly inefficient, strategiesfor getting around failures of contracting institutions. FIRM-LEVELEVIDENCE ONINSTITUTIONSAND GROWTH 4.37 Inthis section we dig deeper into the linksbetween institutions and firmperformance, usingfirm-level data. Throughout t h s section we will refer to this as "micro data," as opposed to the "macro cross-country data" o f the previous section. We draw on two large cross-country datasets o f firms, WBES and PICS, to analyze at the micro level the relationship between different types o f institutions and firm performance, and we also study the use o f informal channels by firms to get around weak contracting institutions. In both firm-level datasets, we identify specific questions that proxy for contracting institutions and property rights institutions. Inbroadterms, weusefms' perceptions aboutthe quality ofthe courts andthejudiciary system to capture contracting institutions and firms' views on corruption to capture property rights institutions. Detailed definitions and country coverage o f the measures used are provided in Annex 3. Do the Micro and Macro Data Agree 4.38 We first conduct a validation exercise where we relate the micro data and the macro data on institutional quality. For this purpose, we average the firm-level data ineach country to obtain micro-founded macro measures o f institutional quality that we relate to the macro measures of institutional quality employed inthe first part o f the paper: rule of law and control o f corruption indicators, number o f days, and number ofprocedures to solve a payment dispute. 4.39 The micro and the macro data on institutional quality are generally consistent with one another, as can be seen in Figure 4.3 for the WBES measures and Figure 4.4 for the PICS measures. Inthe left panel of both Figures, we show the measures of contracting institutions: on the horizontal axis the macro measure (the number o f procedures) and on the vertical axis the micro measure (quality o f courts inFigure 4.3 for WBES orjudiciary respect of rightsinbusiness disputes in Figure 4.4 for PICS). In the right panel o f both Figures, we show the measures of property rights institutions: on the horizontal axis the macro measure (the control of corruption indicator and on the vertical axis the micro measure (frequency o fpayments to get things done in Figure 4.3 for WBES or percentage of firms that pay bribes during government inspections in Figure 4.4 for PICS). For all measures, higher values indicatepoorer quality of the corresponding type o f institution. The South Asian counties are highlighted in both Figures. The regression lines suggest a positive cross-country relationship between the micro and macro measures of institutional quality, whether it i s contracting institutions or property rightsinstitutions. 71 Figure43: Contractingand PropertyRightsInstitutionsAcross Countries: Macro Versus WBES Measures Note: Highervalues o f the quality of courts and of the control of corruptionmeasurescorrespondto worse outcomes. Figure4.4: Contractingand PropertyRightsInstitutionsAcross Countries: Macro Versus PICSMeasures correspondto worse outcomes. 4.40 W e summarize the consistency between the macro and micro data more systematically in Tables 4.3 and 4.4 for the WBES measures and in Tables 4.5 and 4.6 for the PICs measures. In these tables we report the results of regressing each of the micro measures o f property rights institutions and contracting institutions on their corresponding macro proxies, treating each country as an observation. We also include dummy variables for the South Asian countries included in each sample, so that we can see the extent to which these countries do or do not deviate from the average cross-country relationshipbetween the micro and the macro measures. 4.41 The results in Tables 4.3 and 4.5 indicate that, on average, countries with poor contracting institutions according to the number o f procedures also have poor contracting institutions according to the micro measures on the quality o f courts and on the functioning o f the judiciary inbusiness disputes. Similar findings are obtained when the macro proxy for contracting institutions is the number o f days. Tables 4.4 and 4.6 suggest that countries with weak property rights institutions measured by the control o f corruption indicator have weak property rights 72 institutionsproxied at the micro level by the importance o f corruption as an obstacle to business and the percentageof firms payingbribes to government officials duringin~pections.'~ Table 4.3: Regressionsof Micro Measures from WBES on MacroMeasures: Contracting Institutions Functioningof Functioningof Judiciarywith Quality of Judiciarywith DependentVariable is: courtsOf respectto respectto Business courts Business Disputes Disputes Number of Procedures 0.027" 0.020* (0.015) (0.015) Number of Days(log) 0.669*** 0667"' (0.118) (0.142) Bangladesh -0.173 -0.053 -0.257*** -0.129 (0.107) (0.116) (0.098) (0.099) India -0.806 -0.623*** -0.699*** -0.578*** (0.198) (0.193) (0.101) (0.099) Pakistan -0.222 -0.436 0.094 -0.221** (0.276) (0.268) (0.099) (0.098) Observations 46 45 46 45 R-squared 0.15 0.08 0.34 0.33 1Note: Robust standard errors in parentheses. The regressions include a constant. The dependent variable is the average of the firm-level responses on quality of courts or on functioning ofjudiciary with respectto businessdisputesineach country. * significantat 10percent; ** significantat 5 percent; *** significantat 1 percent. Table 4.4: Regressionsof Micro Measures FromWBES on Macro Measures: Property Rights Institutions Frequencyof Frequencyof Corruptionas Additional Corruptionas Additional DependentVariable is: an Obstacleto Paymentsto an Obstacleto Paymentsto Business GetThings Business GetThings Done Done Ruleof Law 0.569*** 0.755'** (0.058) (0.080) Controlof Corruption 0.557"' 0.748*** (0.047) (0.077) Bangladesh 0.473'*' 1.&lo+** 0.297*** 1.599'*' (0.069) (0.107) (0.073) (0.116) India 0.208*'* 0.961'** 0.023 0.716*'* (0.058) (0.093) (0.057) (0.091) Pakistan 0.304*** 0.947*** 0.207"' 0.813*** (0.067) (0.105) (0.067) (0.107) Observations 45 46 45 46 R-squared 0.66 0.62 0.68 0.65 Note: Robust standard errors in parentheses. The regressions include a constant. The dependen variable is the average of the firm-level responses on Corruptionas an Obstacle to Business or on Frequency of Additional Paymentsto Get ThingsDone ineach country. * significantat 10percent; ** significant at 5 percent;*** significantat 1percent. 29 The exercisesshown inthese tables differ from those in AJ that also make use of WBES data. AJ regress the proxies for contractingand propertyrights institutionsin WBES at the firm-level (instead of taking country averagesas we do here) on macro measures (executiveconstraints shown in annex 1) and the deep determinants discussedin the second section. 73 4.42 The less significant results obtained using the PICS measures in Tables 4.5 and 4.6 though still positive are likely due to the smaller number o f countries included in the corresponding regressions relative to the WBES sample. Annex 3 shows that while the WBES dataset covers a large number of developing and middle-income countries plus some developed countries (United States, United Kingdom, and Canada), the PICS dataset covers a smaller numberofdeveloping andmiddle-incomecountries only. 4.43 The average relationship between macro and micro measures of institutional quality estimated using the cross-sections of countries in the WBES or the PICS datasets indicate that macro and micro measuresare consistent with one another. We now focus on the dummies for the major countries in South Asia region, shown inTables 4.34.6. These dummies show the extent to which the relationshipbetween macro and micro measures of institutional quality differs fiom the averagerelationship. The regressionsbased on the WBES dataset (Tables 4.3 and 4.4) suggest that Bangladesh and India have a better ranlung interms o f their contracting institutions' quality according to the fm-level perceptions than they do according to the macro measures. However, WBES firm-level perceptions in Bangladesh, India, and Pakistan reveal significantly worse quality o fproperty rights institutions than i s suggestedby the macro measures. 4.44 The regressions based on the PICS dataset (Tables 4.5 and 4.6) suggest that Bangladesh has clearly worse institutional quality--both contracting and property rights-according to the firm-level perceptions than according to the macro measures. Given that the overall fit o f the regressions based on the PICS dataset is weaker owing to the aforementioned smaller sample size, it i s not surprisingthat the dummies for Bangladesh are less significant inthose regressions. We find these results for Bangladesh to be o f particular interest, given that aggregate cross- country measures of corruption based on perceptions data are often viewed with considerable skepticism, particularly incountries that fare poorly on these indicators. Our results here suggest that not only do the cross-country aggregate indicators agree quite well on average with the view of firms within countries, but moreover firms inBangladesh seem to view corruption to be even more o f a problem than do the outside "expertf' assessments that feature prominently inthe macro data on corruption. Table 4.5: Regressions of Micro Measures From PICS on Macro Measures: Contracting Institutions Functioning of Functioning of DependentVariable is: Judiciary with Judiciary with Respect to Respectto Business DisputesBusiness Disputes Numberof Procedures ,0.026 Numberof Days (log) (0 n4m -.- . -, 0.343 1.192* * (0.291) Bangladesh 1.192*** India -&0g?l* -h0;l;?2* (0.230) (0.132) Pakistan 0.23 0.647' * (0.311) (0.133) Observations 20 20 R-squared 0.42 0.39 I Note: Robust standard errors in parentheses. The regressions include a constant. The dependent variable is the average of the firm-level responses on functioning of judiciary with respect to businessdisputes ineach country. * significantat 10percent; ** significant 5 at percent;*** significant at 1percent. 74 Table 4.6: Regressions of Micro Measures F r o m P I C S on Macro Measures: Property Rights Institutions Percentageof Percentageof Corruptionas FirmsAsked to Corruptionas FirmAsked to DependentVariableis: an Obstadeto Pay Bribesin an Obstadeto Pay Bribesin Business Gov. Business Gov. Inspections Inspections Ruleof Law 0.335 0.136' (0.351) (0.071) Controlof Cwruption 0.208 0.139" (0.311) (0.065) Bangladesh 0.29 0.622"' 0.294 0.584- (0.181) (0.063) (0.219) (0.075) India -0.067 0.483"' -0.16 0.448"' (0.203) (0.037) (0,154) (0.039) Pakistan -0.292' -0.267 (0.166) (0.172) Observations 22 19 22 19 R-squared 0.14 0.62 0.08 0.64 Note: Robust standard errors in parentheses. The regressions include a constant. The dependent variable is the average of the firm-level responses on corruption as an obstacle to business or on whether firm are asked to pay bribes during government inspections in each country. * significant at 10percent;**significant at 5 percent; ***significant at 1percent. Do PropertyRightsInstitutionsand ContractingInstitutionsMatter for Firms 4.45 We next turn to the question o f whether the cross-country evidence on the importance of these two types of institutions holds up at the firm level. There is a great deal o f variability in contracting and property rights institutions across firms, and across cities within countries, as can be seen in Figure 4.5, which reports our two preferred measures of institutions from the PICS dataset. On the horizontal axis, we show thejudiciary respect of rightsinbusiness disputes as the proxy for contracting institutions, and on the vertical axis we show the importance o f corruption as an obstacle to business. As indicatedby the labels, the data points inthe Figure represent each o f the cities inBangladesh, India, Palustan, and Sri Lanka where the PICS was conducted. Each data point corresponds to an average at the city level o f the firm-level measures of contracting and property rightsinstitutions. 4.46 Interestingly, this Figure also shows that the perceptions about the two types of institutional quality are not very highly correlated: The cities where firms find corruption to be a major problem arenot necessarily cities where firms find a poorly functioningjudiciary system to be a major problem, and vice versa. The variability inthe two types of institutional outcomes i s usefulbecause it provides an opportunity to examine the consequencesof successes and failures inbothcontracting andpropertyrightsinstitutions for firms inSouthAsia. 4.47 We are interested in relating firm performance to the measures of institutional quality using the PICS dataset. Conceptually, contracting and property rights institutions should be exogenous to a given firm. However, our analysis is based on firm-level perceptions o f institutional quality, and thus there i s a potential endogeneity problem in relating firm-level performance to those perceptions. On the one hand, firms with better performance may be more aware of the difficulties that weak institutions cause and therefore may complain more about those. On the other hand, firms with better performance may be able to undertake actions to avoid dealing with the weak institutions. This endogeneity would lead to biased estimates of the relationshipbetween institutional quality and firmperformance. 75 Figure4.5: Contractingand PropertyRightsInstitutionsAcross Cities: B2 Chittagong I17 Vadodara I34 Madurai P12 Peshawar I 1 Ahmedabad I18 Gurgaon I35 Ghaziabad S1 Colombo 12 Bangalore I19 Faridabad I36 Noida S2 Gampana I3 Calcutta I20 Panipat I37 Shahjahanpur-Lakimpur S3 Kalutara Hubli- I4 Chandigarh I21 Dhanvad I38 Howrah S4 Kandy I5 Chennai I22 Calicut I39 Mangalore S5 Matale I6 Cochin I23 Palakkad P1 Karachi S6 NuwaraEliya I7 Delhi I24 Bhopal P2 Lahore S7 Galle I 8 Hyderabad I25 Gwalior P3 Sheikhupura S8 Matara I9 Kanpur I26 Indore P4 Sialkot S9 Kurunegala I10 Mumbai I27 Nagpur P5 Faisalabad S10 Puttalam I11 Pune I28 Nashik P6 Gujranwala S11 Anuradhapura I12 Mysore I29 Thane P7 Wazirabad S12 Badulla I13 Vijayawada I30 Jalandhar P8 IslamabadlRawalpindi S13 Monaragala I14 Lucknow I31 Ludhiana P9 Sukkur S14 Ratnapura I15 Guntur I32 Coimbatore P10 Hyderabad S15 Kegalle "Vote: City averagesof the variablesjudiciary respectsrights inbusiness disputes andcorruptionas an obstacleto businessfrom the PICSdataare shown. 4.48 T o circumvent this problem, we consider the measures o f institutional quality averaged at the city-industry level. There are several reasons why institutional quality will vary at the city or industrylevel, although institutional quality has an important national dimension. For example, in large countries such as India or Pakistan, the decentralization of economic policy making likely leads to important differences in the enforcement o f contracting institutions and even property rights institutions across states and cities depending on the political power and the quality and the degree o f corruption o f local government officials. The significant variability in institutional quality across cities (seen inFigure 4.5), which i s even higher for city-industry cells, allows us to identify the effects of institutions on firmperformance. 4.49 T o assess how institutional weaknesses affect firms in Bangladesh, India, Palustan, and Sri Lanka, we regress firm-level sales per worker (in logs) on city-industry average institutional quality. Sales per worker is a measure o f labor productivity, and varies substantially across 76 industries within each country owing to differences in capital intensity and in production processes. Our regressions therefore include industry-fixed effects to control for such industry differences. We include in our regressions only those firms that have been in operation for more than 15 years. The purpose o f this choice i s to make our measure o f sales per worker close to a long-run measure o f firm growth, in a similar vein to what i s done inthe second section, where a country's real GDPper capita is usedas a proxy for that country's long-run growth.30 4.50 Our main finding in Table 4.7 is that for firms in Bangladesh, India, Pakistan, and Sri Lanka, property rights institutions have a more detrimental effect on performance than contracting institutions. Since we use institutional measures as averages at the city-industry level, we are relatively confident in interpreting the regression coefficients as a causal effect of institutionson firmperformance." 4.5 1 The coefficients inthe third column o f Table 4.7 indicate that f m s in a city-industry cell with worse property rights institutionsby 1standard deviation (that is, a city-industry cell where corruption is more important as an obstacle to business) would have on average 6.6 percent lower sales per worker than firms in a city-industry cell with (sample) average property rights institution^.^^ We note however, that weak contracting institutions also have a negative effect on firmperformance, albeit six times smaller than that o fpropertyrights institutions. Table 4.7: FirmPerformance, Contracting, and Property Rights Institutions: Regressions Using PICS Dataset DependentVariable is: Sales per Worker Sales perWorker Sales perWorker (1) (2) (3) Functioningof Judiciarywith Respectto Bus. Disputes -0.012 -0.002 (0.063) (0.063) Corruption as an Obstacleto Business -0.066 -0.066 (0.062) (0.062) Country Dummies Yes Yes Yes Industry Dummies Yes Yes Yes Observations 1820 1820 1820 R-squared 0.91 0.91 0.91 Note: Robust standard errors inparentheses. The regressions include a constant. The dependent variable is the logarithm o f sales per worker with sales converted to US$. The variables functioning o f judiciary with respect to business disputes and corruption as an obstacle to business are normalized to have a mean of 0 and variance o f 1. The independent variables are averages o f these normalized variables for city-industry cells. The regressions include only firms that are older than 15 years o f age. Outliers in the sales per worker distribution for each industry and country (top andbottom 2 percent) are eliminated from the regressions. * significant at 10percent; ** significant at 5 percent;*** significant at 1percent. 4.52 While this firm-level evidence o f a link between performance and institutional quality is less clear cut than the earlier cross-country evidence, that likely reflects a variety o f weaknesses inthe firm-level data. Inparticular, the important variability o f institutional quality across cities and industries may still be lower than the variability o f institutional quality across countries used inthe second sectionto identify the impactofinstitutionson cross-country incomes. 30 Inthe cross-country data, the argument is that per capita income levels across countries were not very different inthe distant past, and so cross-country differences in levels o f per capita income reflect differences in countries' very long- rungrowth performance. 31We note, though, that we would need to use instrumental variables for the city-industry institutional quality measures to be fully confident on the causal interpretation o f the regression coefficients. Unfortunately, those instrumental variables are not available. 32 The measures o f contracting and property rights institutions included in the regressions are normalized to have a mean o f 0 and a standard deviation o f 1. 77 How Do FirmsCopeWith Weak ContractingInstitutions 4.53 The firm-level data allow us to investigate in more detail one hypothesis that emerges from the cross-country evidence: The reason why poor contracting institutions do not have a strong impact on growth is that firms can find ways to contract around poorly performing institutions in this dimension while that i s not possible in the case of weak property rights institutions that allow governments to prey on the private sector. We investigate the extent to which firms are able to use informal channels inBangladesh, India, Pakistan, and Sri Lanka to get around weak contracting institutions. The informal channels considered are: belongng to a business association, using the business association to resolve business disputes with customers or suppliers, using informal sources of finance for working capital and for investment needs, and solving overdue payments problemsby alternative mechanisms instead o f courts. 4.54 Table 4.8 provides strong evidence that in cities and industries where the quality o f contracting institutions i s weaker, firms resort to different informal ways to get around those. For example, firms operating in city-industry cells with worse contracting institutions by 1 standard deviation (that is, where the functioning o f the judiciary in business disputes is worse) are 14.6 percent more likely to belong to a business association than firms in a city-industry cell with (sample) average contracting institution^.^^ Also, in cities and industries where the weak quality of the judicial system makes it difficult to enforce debt contracts, firms resort to the use of informal credit channels. Table 4.8: GettingAround Weak ContractingInstitutions:RegressionsUsingPICS Data Importanceof Percentageof Percentageof Percentageof Membershipin Dispute Working Capital Investment Overdue DependentVariable is: Business Resolution by Financedby Financed by PaymentsNot Association Business Informal Informal Resolvedby Association Sources Sources courts (1) (2) (3) (4) (5) Functioningof Judiciarywith Respectto BusinessDisputes 0.146" 0.273"' 0.01 l** 0.018' 0.133 (0.071) (0.074) (0.005) (0.010) (0.111) CountryDummies Yes Yes Yes Yes Yes IndustryDummies Yes Yes Yes Yes Yes Observations 4038 2052 1908 1164 328 R-squared 0.21 0.02 0.04 0.11 Note: Robust standard errors inparentheses. The regressions include a constant. The variable functioning ofjudiciary with respect to businessdisputes is normalizedto have a mean of 0 and variance of 1. The independent variable is the average of these normalizedvariables for city-industry cells. The first column shows the marginal effects from a probit regression. * significant 10percent; ** significant 5 percent; *** at at significant at 1 percent. 4.55 Finally, we use the PICS dataset to examine whether firm-level concerns about weak institutional quality are stronger in industries that are more institutionally dependent. We follow Levchenko (2004) ranking o f industries according to their institutional dependence as measured by industry-level Gini coefficient o f intermediate input use and number o f supplyingindustries. An industry is more institutionally dependent if its contracting intensity i s high; that is, if producers that industryneed to enter into complex production relationships with many different suppliers. Industries with more concentrated intermediate input use are less institutionally dependent than industries whose intermediate inputs used originate in many different industries. Industries where producers enter in interactions with a large number o f suppliers have a high 33 The measures o f contracting institutions included in the regressions are normalized to have a mean of 0 and a standarddeviation of 1. 78 contract intensity and thus are more dependent on good contracting institutions. Applying the ranlung to the industries covered in the PICS dataset, it indicates that auto parts, transport equipment, and metals and machinery industries are more institutionally dependent than leather and food industries. 4.56 Table 4.9 shows that indeed firm-level concerns about the quality o f contracting institutions are higher in industries that are more institutionally dependent. In contrast, we find contracting institutions use more frequently informal channels to get around institutional weaknesses, as shown inTable 4.10. For example, firms in auto parts, transport equipment, and metals and machinery find it more important to use their business association to solve payment disputes with customers and suppliers than firms inindustries such as leather and food. Table 4.9: InstitutionalDependence,Contracting and PropertyRightsInstitutionsAcross Industries: RegressionsUsingPICSDataset Functioning of Judiciary with Functioning of Judiciary with Corruption as an Corruptionas an DependentVariable is: Obstacle to Respectto Bus. Disputes Respect to Bus. Disputes o~~~~~~ Business I (1) (2) (3) (4) Inst. Depend. Ranking N. Interactions - 0.069' 0.003 Inst. Depend. Ranking Gini Coefficient - (0.041) (0.005) 0.076' 0.004 (0.041) (0.005) Country Dummies Yes Yes Yes Yes observations 25 25 25 25 R-squared 0.89 0.89 0.66 0.66 Note:Robust standard errors inparentheses. The regressions include a constant. The dependent variables hnctioningo f judiciary with respect to business disputes or corruption as an obstacle to business are normalized to have a mean of 0 and variance o f 1. Higher values o f the independent variables N. Interactions or Gini coefficient correspond to a higher institutional dependence o fthe industry. significant at 10 percent;** significant at 5 percent; *** significantat 1percent. Table 4.10: InstitutionalDependenceandWays to Get Around Weak contractingInstitutionsAcross Industries: RegressionsUsingPICS Data Importanceof Perc.of Importanceof Perc. of Membership Dispute Overdue Membership Dispute Overdue DependentVariable is: in Business Resolutionby Payments in Business Resolutionby Payments Association Business Not Resolved Association Business Not Resolved Association by Courts Association by Courts (1) (2) (3) (4) (5) (6) Inst.Depend. Ranking N. Interactions - 0.029 0.006 -0.002 (0.033) (0.046) (0.004) Inst. Depend. Ranking Gini Coefficient - 0.011 -0.047 -0.004 (0.035) (0.048) (0.004) Country Dummies Yes Yes Yes Yes Yes Yes Observations 25 25 11 25 25 11 R-squared 0.46 0.87 0.02 0.43 0.88 0.07 Note: Robust standard errors in parentheses. The regressions include a constant. Higher values of the independent variables N. Interactions or Gini coefficient correspondto a higher institutional dependence ofthe industry. significant at 10percent;** significant at 5 percent; *** significant 1percent. at 79 CONCLUSIONS 4.57 We have examined the evidence on institutions and growth in South Asian countries at the macro and micro levels, distinguishing between two key dimensions o f institutional performance: propertyrights institutions and contracting institutions. 4.58 First, we document how countries inthe South Asia region fare along these two types of institutional quality, using cross-country macro data and controlling for deep historical determinants of these two types of institutions. We find that while India has poor performance on contracting institutions, it fares reasonably well on measures of property rights institutions. The reverse is true for Bangladesh, which stands out for having quite poor property rights institutions butreasonablecontracting institutions. 4.59 Second, we show that property rights institutions matter muchmore for economic growth across countries than contracting institutions. Relative to this strong cross-country empirical regularity, India stands out for having a large negative residual, suggestingthat even India's good performance on property rights institutions has not yet had the development impact that cross- country analysis would suggest. Incontrast, Bangladesh stands out injust the opposite direction, with a predicted income levelbased on its institutional quality that is muchlower than its actual income level. A possible interpretation o f these findings is that India has a great deal of scope for future growth given its good institutionalperformance, while Bangladesh's current relatively high income levels (given its weak institutionalquality) are more likely to be unsustainable. 4.60 Third, using micro data, we confirm the importance of property rights institutions for firm performance in Bangladesh, India, Pakistan, and Sri Lanka. Finally, we show how firms in the South Asian countries are able to circumvent failures in formal contracting institutions by resorting frequently to informal channels such as belongingto a businessassociation. 80 ANNEX 1.FIRST-STAGE REGRESSIONSFORPROPERTYRIGHTSAND CONTRACTINGINSTITUTIONS, SMALL SAMPLE Rule of Control of Executive Expropriation Numberof In(number law corruption constraints risk procedures of days) Britishlegal -0.394 -0.418 0.417 -0.675 -10.014 -0.385 origins (0.210)* (0.227)* (0.590) (0.363)* (4.567)** (0.200)* In(settler 0.472 0.489 1.244 0.742 1.251 0.153 mortality) (0.099)*** (0.099)*** (0.148)*** (0.119)*** (1.381) (0.079)* India 0.404 0.785 -2.121 -0.705 14.507 0.586 (0.203)* (0.235)*** (0.479)*** (0.330)** (3.637)* ** (0.178)*** Bangladesh 0.991 1.368 2.401 2.110 3.026 0.375 (0.192)*** (0.219)*** (0.483)*** (0.322)*** (3.792) (0.167)** Sri Lanka 0.173 0.454 -0.571 1.236 -8.946 0.566 (0.193) (0.220)** (0.483) (0.322)*** (3.781)** (0.167)*** Pakistan 1.231 1.474 -1.781 1.716 20.849 0.555 (0.215)*** (0.248)*** (0.479)*** (0.340)*** (3.571)*** (0.188)*** Constant -1.757 -1.951 -10.130 -9.783 30.646 5.257 (0.479)*** (0.488)*** (0.848)*** (0.587)*** (7.038)*** (0.430)*** Observations 57 57 57 57 57 57 R squared 0.5 1 0.52 0.49 0.49 0.22 0.18 Note: Robuststandarderrors inparentheses. * significantat10percent;** significantat5 percent; *** significantat 1percent. 81 ANNEX 2. INSTRUMENTALVARIABLESREGRESSIONSOFLN(GDPPERCAPITA) ON PROPERTY RIGHTSAND CONTRACTINGINSTITUTIONS, SMALL SAMPLE Rule o flaw 1.740 2.459 (0.402)* ** (0.924)** Control 1.686 2.403 corruption *(0.371)** (0.991)** Expropriati 1.120 1.638 onrisk (0.293)*** (0.932)* Executive 0.568 0.550 constraints (0.086)** * (0.121)** * Number 0.083 0.085 0.090 -0.009 procedures (0.052) (0.054) (0.057) (0.031) Number 2.902 2.993 3.254 -0.213 days (2.096) (2.357) (3.222) (0.703) India -1.428 -0.833 -3.020 -1.997 -1.631 -0.790 -3.985 -1.965 (0.442)* ** (0.329)** (0.845)*** (0.409)** * (0.804)** (0.532) (2.364)* (0.481)** * Bangladesh 0.485 1.061 1.102 0.402 0.360 1.177 1.247 0.414 (0.471) (0.596)* (0.679) (0.299) (0.3 13) (0.670)* (1.072) (0.266) Sri Lanka 0.796 1.276 1.941 -0.654 -1.467 -0.853 -0.066 -0.444 (0.774) (0.923) (1.093)* (0.521) (0.927) (0.749) (0.566) (0.280) Pakistan -0.796 -0.490 -1.159 -2.026 0.214 0.682 -0.195 -2.062 (0.520) (0.487) (0.623)* (0.556)** * (0.213) (0.330)* * (0.360) (0.440)** * Constant 5.905 5.618 -2.202 6.038 -8.063 -8.907 -21.614 7.058 (1.639)* ** (1.788)** * (3.729) (1.238)** * (11.853 (13.490) (24.570) (4.532) Observations 57 57 57 57 57 57 57 57 Note: Robust standard errors inparentheses. * significantat 10percent;** significant 5 percent; *** significant at at 1percent. 82 ANNEX 3. VARIABLEDEFINITIONS AND COUNTRYCOVERAGEFORTHE WBES AND PIcsDATASETS Description Countrycoverage An assessmento f the quality o f courts based on the rating Bangladesh, India, (WBES) o f "the overall quality and efficiency o f services delivered Pakistan, and by thejudiciary and courts." Answers rangedfrom 1=very othersa pood to 6 = very bad An assessmento fthe functioning o fthejudiciary with Bangladesh, India, withrespect to business respectto business disputesbased on the question, "In Pakistan, and resolvingbusiness disputes do you believe your country's others court system is (a) fadimpartial, (b) honesthncorrupt, (c) quick, (d) affordable, (e) consistent, (f) decisions enforced?" Answers ranged from 1= always to 6 =never. We use the average o fthe firms' answers to the six items Corruption as an obstacle A n assessmento f corruption based on "how problematic is" Bangladesh, India, to business (WBES) corruption "for the operation and growth" o f the firm. Pakistan, and Answers ranged from 1= no obstacle to 4 = major obstacle others Frequency o f additional The frequency o fbribes based on the rating o fthe Bangladesh, India, payments to get things statement, "It is common for firms inmy line o fbusiness to Pakistan, and done (WBES) have to pay some irregular additional payments to get others things done." Answers ranged from 1= always to 5 = seldom Functioning o fjudiciary An assessmento fthe functioning o fthe judiciary with Bangladesh, India, with respect to business respect to business disputes based on the rating o f the Pakistan, and Sri disputes (PICS) statement, "I confident that thejudicial system will am Lanka enforce my contractual and property rights inbusiness disputes." Answers ranged from 1=fully disagree to 6 = fully agree Corruption as an obstacle An assessmento fcorruptionbasedon "how problematic is" Bangladesh, India, to business (PICS) corruption "for the operation and growth" o f the firm. Pakistan, and Sri + Answers rangedfrom 0 =no obstacle to 4 =major obstacle Lanka Percentage o f firms The percentage o f firms that were asked to pay bribes Bangladesh, India, asked to pay bribesin during inspections by government officials from several Pakistan, and Sri government inspections agencies Lanka IMembership inbusiness ~~Bangladesh, India, association Pakistan, and Sri Lanka Importance o f dispute An assessmento fthe value o fthe services provided by the Bangladesh, India, resolutionbybusiness business association regarding the resolution o f disputes Pakistan, and Sri association with officials, workers or other firms. Answers rangedfrom Lanka 0 =no value to 4 = criticalvalue to the fm Percentage o f worlung Informal sources o f finance are listed as, for example, Bangladesh, capital financed by moneylender Pakistan, and Sri informal sources Lanka Percentage o f investment Informal sources o f finance are listed as, for example, Bangladesh, financed by informal moneylender Pakistan, and Sri sources Lanka Percentage o f overdue Percentageof fmdisputesabout overdue payments inthe Bangladesh and Sri Ipayments not resolvedby last two years that were not resolvedby court action Lanka courts Kenya, Madagascar, Malawi, Malaysia, Mexico, Namibia, Nicaragua, Nigeria, Panama, Peru, Philippines, Senegal, Singapore, SouthAfrica, Tanzania,Tunisia, Uganda,United States, Uruguay, Venezuela, Zambia, andZimbabwe. 83 84 5. IMPROVINGTECHNOLOGY, SKILLS, AND INNOVATIONINSOUTH ASIA34 INTRODUCTION 5.1 Since the 1980sthe South Asian countries35have been growing faster than the rest of the world. However, the international environment i s becoming more competitive and demanding. In addition, higher education and innovation are becoming more critical for countries to be able to benefit from the increasingly globalized international environment. Therefore, South Asian countries have to improve their skills and innovation capabilities. Here we will assesses the position of South Asian countries and proposes some o f the key actions that they need to take to strengthen their technology and innovation capabilities to improve their economic performance and welfare. 5.2 This paper covers the five largest South Asian countries. They range from Sri Lanka and Nepal, which have around 20 million inhabitants each, to India, the second most populous country inthe world with slightly more than 1billion people. Interms of gross domestic product (GDP) per capita, four are inthe World Bank's low income category (per capita income less than US$765 in2003) with Nepal at the very low income level, to Sri-Lanka, which just makes it into the lower middle income category (US$766-3,035) (Table 5.1). Population GDP GDPkapita Merchandise exports (millions) (US$ billion) (US%millions) Bangladesh 138 55 400 6942 India 1064 571 540 55982 Nepal 25 6 240 662 Pakistan 148 78 520 11930 Sri Lanka 19 18 930 5125 LWorld 6273 I1 34577 I1 5510 I1 7578698 5.3 Each o f the five has had rates o f growth above the world average for the period 1980- 1990 and 1990-2003, with Pakistan having the highest rate o f growth in the decade of the 1990s and India the highest in the 1990-2003 period (Table 5.2). In fact in the last years India has achieved a spectacularrate of growth of 8 percent. 1980-90 1990-2003 Bangladesh 3.7 4.9 India 5.7 5.9 Nepal 4.6 4.6 Pakistan 6.3 3.6 Sri Lanka 4.0 4.7 World 3.3 2.8 34This Chapter was prepared by Carl J. Dahlman, Professor o f International Relations and Information Technology, EdmundA. Walsh School o f Foreign Services, Georgetown University, USA. The views in this paper are solely the author's and do not necessarily reflect those o f the World Bankor its Executive Directors. 35 We examine the situation o f five o fthe South Asian countries: Bangladesh, India, Nepal, Pakistan, and Sri Lanka. Unfortunately, it was not possible to include Bhutan or the Maldives because of data limitations. 5.4 To put them in the global context it i s instructive to compare their shares in the global population with those in global GDP and trade. Because they are low income countries, their share in global GDP is much lower that that inglobal population. At the higher end, Sri Lanka's share in global GDP i s 17 percent o f its share inpopulation, while that o f Nepal isjust 5 percent. With respect to exports, Sri Lanka's share o f global exports actually exceeds it share of global GDP by 40 percent, indicating it is an export-driven economy. The shares of the four other countries range from just 44 percent (India, the least export oriented) to 70 percent (Palustan). While the share of exports in global exports between 1990 and 2003 increased by 80 percent for Bangladesh and 40 percent Sri Lanka and 45 percent for India, Pakistan and Nepal just maintainedtheir relative shares (Table 5.3). Population GDP2003 Merchandise Merchandise 2003 exports 2003 exports 1990 02.20 00.16 00.09 00.05 India 16.96 Ki 00.40 "?kistan 02.36 00.23 00.16 L a n h Source:Computedfrom World Bank 2005 -I 00.30 5.5 Interms of the share of exports of goods and services inGDP, only Sri Lanka, with 36 percent, is above the world average of 24 percent (Table 5.4). India is particularly low at just 14 percent, in fact one of the lowest in the world in spite of its considerable service exports particularly inthe information technology services for which it is so well known worldwide. With respect to merchandise exports, the majority inall countries are manufactured products. With the exception o f India and Sri Lanka, the share of manufacture in merchandise exports exceeds the world average. However, only a very small percentage o f them are hightechnology products. Table 5.4: Various Indicators of Trade Structure Source:World Bank 2005. THENEWCOMPETITIVE CONTEXT 5.6 The world i s in what could be called a knowledge revolution. There has been a speed up in the production and dissemination of knowledge based, in part, on advances in information processing and communications technologies, as well more general advances in the science base and inability to codify knowledge. 5.7 The rapid reduction of transportation and communications costs made possible by technologicalprogress inall means of transportation and information technologies (IT), combined with liberalizationof trade ingoods and services, are leadingto a rapid increase inthe volume of 86 goods and services that are traded. Between 1990 and 2003. the share o f imports and exports in GDP increased from 38 to 48 percent o f global GDP. In addition, 27 percent of global value added i s being producedby multinational corporations (MNCs), meaning that nearly one-third of world GDP is produced by corporations spanning multiple markets and national jurisdiction^.^^ This is actually an underestimate of the degree to which global production is interlinked because it does not include all the indirect effects through the integration of supply and distribution chains. 5.8 MNCs are also the key agent in the creation o f knowledge since the R&D done by multinationals accounts for the large majority o f R&D done by the private sector, and private R&Dhas become larger than public R&D as a result of the decrease in global defense budgets that occurred after the endo f the cold war. 5.9 Combinedwith the trend toward liberalizationo f product, service, and financial markets, greater globalization means increasing global competition. Capital markets move capital to where they expect the highestrisk adjusted returns, while MNCs redirect their resources inline with the global dynamics of markets, driven by size and growth potential, andthe creationof endowments, among which knowledge stands out. Thus, countries like China and India have become magnets inthe creation ofmajornewplatforms inthe exports ofproductionandservices. 5.10 The nature of competitiveness i s also changing. Traditionally it was based on lower capital or labor costs, or of other local inputs, including infrastructure services, while also depending on the economic and business environment. Although these fundamentals continue to play a key role, given the very rapid rate of development and dissemination of new knowledge globally and the pressureto restructure, there are important new elements, includingthe ability to: 0 Rapidly redeployresourcesinorder to capture new opportunities 0 Ensure the quality, skdls, and flexibility of labor force (and management) 0 Keep up withrapidly changing technologcal and organizational advances 0 Move to highervalue parts of value chain (researcwdesign and marketing, branding, and managing of customer information) 0 Make effective use of ITto reduce transactions costs and improve capacity to respond quickly to changing opportunities and threats 5.11 As a result there is increased attention across countries on improving their overall business environment and the flexibility and speed o f their economies to respond to rapidly changing circumstances, improving education and skills systems, and improving their innovation systems and their information infrastructure. With the exception o f India, the South Asian countries are not so well placed to take advantage o f these changes in the competitive environment because they lack the critical mass of high skilled human resources and technological capability necessaryto do so. 36 Basedon calculationsfrom UNCTAD (various years). 87 SOUTHASIAN COUNTRIESINTHE CONTEXTOFTHE KNOWLEDGEECONOMY World Bank KnowledgeAssessment Methodology 5.12 The World Bank Institute has developed a useful benchmarlung tool that helps to rank countries interms of their readinessto use howledge for their de~elopment.~~The methodology consists o f examining a country's global rank ordering infour pillars. The four pillars are: An economic and institutional regime that provides incentives for the efficient use of existing and new knowledge and the flourishing o f entrepreneurship; this is critical because it affects not only the incentive to improve performance, but also the ease with which better knowledge can be put intopractice An educatedand skilledpopulation that can create, share, and use knowledge well; this is critical because, as argued above, this i s the key enabler to make effective use of knowledge as well as to produce it A dynamic information infiastructure that can facilitate the effective communication, dissemination, and processing of information; this has become critical for access to knowledge and to reduce transactions costs An efficient innovation system o f firms, research centers, universities, consultants, and other organizations that can tap into the growing stock o f global knowledge, assimilate, and adapt it to localneeds, and createnew technology 5.13 The rationale for including the economic incentive and institutional regime as one of the pillars of the knowledge economy i s that it sets the broader framework for the more structural pillars. In addition, an economy needs to have the flexibility to adjust to the rapidly changing opportunities created by the rapid development and dissemination o f new knowledge. An effective economic and institutional regime that permits such rapid adjustment i s critically important. 5.14 Knowledge Assessment Methodology (KAM) consists of a set o f 80 structural and qualitative variables that serve as proxies for the four pillars (roughly 20 per pillar) that are critical to the development of a knowledge economy. The comparison is undertaken for a group of 128 countries that includes most of the developed Organisation for Economic Co-operation and Development (OECD) economies and more than 90 developing countries. A reducedindex of the KAMcalled the knowledge economy index (KEI) based on three indicative variables in each o f the four pillars has been developed to give a quick summary index of a country's overall position.38 37 See www.wor1dbank.orgikam. KnowledgeAssessment Methodologyis designed to help countries understandtheir strengths and weaknesses in making the transition to the knowledge economy. It provides a preliminary knowledge economy assessment of acountry,which can form the basisfor moredetailedsector-specific work. 38 The actual indicatorsused for eachofthe pillar are as follows. Economicandinstitutionalregime: tariff andnontariff barriers, regulatory quality, and rule of law. Education and human resources: adult literacy rate (percent age 15 and above), secondary enrollment, and tertiary enrollment. Innovation system: researchers in R&D, patent applications grantedby the U.S. Patent andTrademark Office, and scientific and technicaljournal articles (all weightedper million people). Information infrastructure:telephonesper 1,000 persons, computersper 1,000 persons, and Internet users per 10,000 persons. 88 Relative Position of South Asian Countries 5.15 Figure 5.1 presents the KEI for the six developing regions, the Group o f Seven and Western Europe, as well as some key developing countries and the five South Asian countries for which we have data. Several interesting aspects stand out. First, as a whole the developed countries (withthe notable exception o f Finland) have lost some of their relative ranking as some middle income countries have made faster relative progress. Second, some developing countries such as Brazil, China and Jordan have made significant relative improvements. Other middle income countries such as Argentina and SouthAfrica have lost ground3' Figure 5.1: Overall KEI 1995 Versus Most Recent 10 0 - 8 - 7 - e- 6 - 4- 3 - 2- 1- 0 I I 1 2I 3I 4 5I 6I 7I 8I 0I 1995 ~~ Source: www.worldbank.org/kam. Note: The horizontal axis represents the relative position o f the country or region in 1995. The vertical axis represents the positionin the most recent year (generally 2000-3). The graph is split by a 45 degree line. The most advanced countries are on the northeastern section o f the diagonal. But the position relevant to the diagonal is also critical. Those countries or regions that are plotted below the line indicate a regression in their performance between the two time periods. Countries or regions that are marked above the line signify improvement between the two time periods, while those countries that are plotted on the line indicate stagnation. The KAM methodology allows the user to check performance in the aggregate KEI or knowledge index (KI), as well as the individual pillars that define them: economic incentive regime, education, and information communications technologies, and the innovation index. This figure has been computed using the unweighted variables for the innovation index 39Countries can lose ground in one o f two ways. They may have an actual decline in an indicator or they may not improve an indicator as fast as other countries so they fall behind relative to the rest o f the world. InSouth Africa, for example, both elements are at work. In secondary education, the actual enrollment rates declined. In the information communications technologies indicators, even though South Africa made significant improvements in the penetration ratios increasing some by a factor o f six times, the rest o f the world moved much faster so South Africa fell behind relative to the rest o f the word. The KAM score cards also show that Africa as a whole i s particularly weak in the innovationpillar variables 89 5.16 South Asia does worse than the other developing regions except Africa. Within South Asia, India does the best, although it does not show any improvement over time. Its higher knowledge economy index i s largely due to its high index on innovation (see Figure 5.2 and Table 5.5, which give the breakdown in the composition o f the indices and their change over time) given the large absolute size o f scientists and engineers in R&D as well as the absolute volume o f scientific and technical publication^.^^ Figure5.2: ComparisonofKEIComponentPartsfor WorldRegionsWith SouthAsian Countries (MostRecentinTop Line,Comparedto 1995 BottomLinefor EachGroup) G7 I I I Ip*"'-.e I#.h+.... >.--. Western Europe EastAsia Europe and Central Asia LatinAmerica Middle East and NorthAfrica India Sri Lanka South Asia Pakistan Africa Bangladesh Nepal 0 0 2:o 4:O 6.0 8.0 1 burce: www.wor1dbank.orglkam. Note: The top bar representsthe most recent aggregate KEI score for a selectedregion or country, split into the four KE pillars. Each color band represents the relative weight of a particular pillar to the overall country's or region's knowledgereadiness, measuredby the KEI. The first line for each country is its position in the most recent year for which data are available (generally 2002-3). The second line is for 1995. This figure has been computed using the unweightedvariables for the innovationindex. 5.17 Sri Lanka is the second highest among the South Asian countries, and it shows some improvement over the period. The biggest improvement i s in the economic incentive and institutional regime, where it gets the highest ranking among the South Asian group. It made significant improvements inthe ICT index where it moves from second to first. While it makes a small improvement in the innovation index, it actually loses ground in the education variable even though it still remainsthe highest in education among the group. 5.18 Pakistan, Bangladesh, and Nepal all lose ground in the aggregate KEI. Most notable is the sharp fall inthe economic incentive regime inNepal which considerably pulls down it overall average. Nepal also loses in the information communications technologies (ICT) indicator. Pakistan also loses inthe economic incentive, and inthe ICT indicator as well as inthe education indicator, and ends up with the lowest score among the group inthe latter. Bangladesh slips most inthe innovationindex and also slips inthe economic incentiveregime, butmakes some gains in the ICT and a smaller gain inthe education index. 40 All the indicatorsinthe methodologywere scaled by population.However, becauseknowledgeis not consumedin its use, for the innovationvariables the indicatorwas also computedbasedon absolutevalues, which is how it is reportedhere. Inthe full KAMdatabase it is possibleto use the innovationvariablesnormalizedby population. 90 Table 5.5: Knowledge Economy Indicator and Components: Changes 1995 to Most Recent Most recent 1995 Source: wv.worldbank.org/kam. Note:These numbersare the normalizedrankings. Forthe raw scores see the database. 5.19 See the annex for individual basic scorecards with changes between the two periods for each of the countries. The annex also includes more detail data on the economic incentive and institutional regime as well at the information and communications indicators because they are not the main focus o fthis paper. GLOBAL TRENDS EDUCATION THE SITUATIONOFSOUTHASIAN COUNTRIES IN AND Global Trends inEducation and Training 5.20 Education i s the fundamental enabler o f the knowledge economy. Well-educated and skilled people are key to creating, sharing, disseminating, and using knowledge effectively. Critical is no longerjust basic or even secondary education, but higher education and the constant upgrading o f skills. This is a challenge for all countries. There i s also increasing competition for people with high level skills, which makes their education and effective employment a central aspect of development strategy. 5.21 The development of a knowledge economy demands a flexible education system. It begins with basic education that provides the foundation for learning; continues with secondary and tertiary education that develops core, includingtechnical, skills; and encourages creative and critical thinking that i s key to problem solving and innovation, extending into a system of lifelong learning. Such a system is one that encompasses learning from early childhood to retirement and includes formal training (schools, training institutions, and universities) and nonformal learning (on-the-job training, and skills learned from family members or people in the community). The basic elements o f such a system are comprehensiveness, new basic shlls (acting autonomously, using tools interactively, and functioning in socially heterogeneous groups), multiple pathways, and multiple providers. 91 5.22 Thus, the knowledgerevolution meansthat higher levelsofeducation are neededto keep upwith andmake effective use o frapidly changing knowledge. Italso means that highlevel scientific and technical manpower is needed to create new knowledge. Butbecause the halflife o f knowledge i s getting shorter and there are new slulls to learn, it also means that there is a need for a system o f continuous training inorder to constantly upslull or re-skill people who have already passedthrough the formal educational system. 5.23 Countries are therefore paying more attention to education as part o f their development and competitiveness strategy, putting great efforts into increasing the levels o f educational attainment. Between 1990/1 and 2002/3, for instance, enrollment rates at the secondary level increased from 55 to 71 percent and from 16 to 26 percent at the tertiary level. For high income countries, inparticular the increases, were from 94 to 107 percent and from 47 to 66 percent for the secondary and tertiary levels, re~pectively.~'For Korea, an economy that has traditionally placed a very high value on education and is now well known as a knowledge economy, the increase in tertiary enrollment rates was from 39 to 85 percent, which put it second only to Finland (with86 percent), the other well-known knowledge-based economy (World Bank 2005). 5.24 Related to the need to keep up with new slulls and functions-in computer literacy, communication slulls, and the ability to work in groups-is the need for constant up-skillingand re-skilling o f persons who have already left the formal educational system. This i s reflected inthe very high percentage of adults who are talung additional courses at work, in specialized institutions, or even going back to tertiary institutions for formal education (inFinland, this i s true for more than 50 percent o f adults). This is also reflected inthe number o f students at universities who are older than the typical university age cohort o f 18-24 years old. Inthe United States, more than 40 percent o f undergraduates are over 25 years o f age. InAustralia, N e w Zealand, Denmark, Norway, and Sweden, more than 20 percent o f first-time entering students were over the age o f 27 in2000.42 5.25 Inadditionthereisatrendfor anincreasingnumberofprivateeducationalinstitutionsall the way from nursery schools to the university level that have arisen to fill in the needs not adequately addressed by public education. Firms are undertaking increasing amounts of in-house training to give their workers the skills they need to compete. Inaddition, some o f the larger firms are even setting up their own in-house universities to provide the most advanced specialized slulls needed to be competitive. Firms are also more proactive in approaching universities and specialized training centers to get them to develop specific trainingprograms to meet their needs 5.26 Equally noteworthy i s the increasing use o f information-based technologies, which has been gathering speed as the technology has improved and more experience acquired on its use. In the United States, 16 percent o f tertiary-level students are taking at least one course online, and 40 percent o f those are full time on line. E-learning i s expanding very rapidly and much is being provided by nontraditional universities (new entrants, including publishers and mass media). E- education, by crossing boundaries, has also facilitated the internationalization of education. In 2005 there were 2.1 million students in tertiary education outside their home countries, which combined with a growing tendency for foreign universities to set up facilities abroad, i s increasing competition inthe education sector worldwide. 41In most countries, in spite of rapid expansion of higher education, the relative wage differencesbetween college graduates and high school grades is not narrowing. This is surprising given the very large increase in supply. The explanationfor this is that rapid technological change and higher education are complementary. Higher education is becomingincreasinglyimportantto take advantageofthe rapidadvancesinknowledge. 42 OECD (2005). 92 5.27 Finally, there is also increasing competition for high level human capital across countries because there i s clear understanding that to be globally competitive and to be able to innovate, countries need high level human resources. It is telling, for example, that w h l e there is no free trade in labor, the exception is for highly trained persons, where even the U.S. has had more liberal immigration policies and developed a program for the temporary immigration o f specialized manpower inthe ICT sector. EducationandSkills in SouthAsian Countries 5.28 South Asian countries43are in a weak position in terms of education and slulls. As a group they have high illiteracy rates, low enrollment ratios at the secondary and tertiary levels, very low average educational attainment among the adult population, extremely l o w percentage o f professional and technical workers among the labor force, low quality o f math and science education, little staff training even among firms inthe modem sector, and a very serious problem o f emigration o fthe highly slulled workers. 5.29 As can be seen from the spider charts (Figure 5.3 next page) for individual countries there i s considerable variations among the five countries. Nepal and Bangladesh are much weaker on all the variables. Palustan ranks somewhat better. Sri Lanka and India score much higher. Sri Lanka has the highest literacy, enrollment rates, and average educational attainment. 5.30 However, India is ranked higher in terms of the quality o f science and math education, extent o f staff training, and availability o f management education. India has the world-renowned Indian Institutes o f Technology and Indian Institutes o f Management, which produce world-class graduates. These institutes, along with many other lesser known regional colleges, have given India a critical mass o f highly skilled people. These high quality English-speaking human resources are a large part of the reason why India has been able to develop the information technology export services that have moved up from simpler back office functions and call centers to software design and innovation services. Many o f the highly skilled Indians have immigrated to the United States and Europe in search o f higher paying jobs. However, some o f this brain drain has been turned into a brain gain as they have started to outsource highly skilled services from India.44 5.31 This strong high-slulled I C T service sector has not developed in the other South Asian countries because o f their smaller scale and less prevalence o f English intheir education systems. However, this i s a very small sector in India relative to its total population, and that the average levels o f educational attainment in India are very low. Therefore, improving education and skills i s challenge for all the SouthAsian countries. ~ 43Figure5.3 shows the positionof the regionor countryrelativeto all the other countries inthe world. Being on the outsideperimeter indicatesrankingamongthe top 10percent of countries inthe world andbeingat the center ofthe circle indicatesranking inthe bottom 10 percent. For more on education and skills in Indiaincludingthe rise ofthe high technology service exports, see Dahlmanand Utz(2005). 93 Figure 5.3: Education Scorecards for South Asian Countries SouthAsia, NOrmaliKatiOnGroup: ail Adit Lteracy Rate (% age 15 and above) (57.06) (2.40) Wen EducatedPeopleDo NotEmigrste Abr rage Years of Schocimg(4.1 6) (4.22) Availability of Managemet4 Educatio ondary Enrollment(49.16) (3.15) Extentof Staff Training Tertiary Enrollment(6.18) (3.78) Qualityof Science and MathEducatio fe Expectancyat Birth,years (64.56) (6.00) Prof.and Tech.Workers as % of the Labor For rnet Access in Schools (2.88) (2.59) Fubk Spendingon E&catiin as %of GDP India,NormalizationGroup: A11 AcWt LiteracyRate(% age 15 andabove) (61.03) (2.83) W EducatedPeopleDa NdEmigrateAtxo uage Years of Schoeihg(5.06) (5.70) Avakbiky of ManagementEducation e m * Enrohent (48.47) (3.80) Extent of Staff TraWg Tettiary Enrohent (10.58) (5.50) h R y of Science and MathEducation LifeExpectancyat Bkth,years(6320) (4.10) Public SpencWg MEducationas %of G rnet Access inSchools(3.60) Bangladesh,NormalizationGroup: AH Aduk Lteracy Rate(%age 15 md hove) (41.J9) (2.01) WeH EducatedPeople Do Not hnigrateA b rageYears of Schooling (2.58) (3.00) AvaWebPy of ManagemeniEducat econdaryEnrohent (46.88) (2.30)Extent of Staff Train Tertiary Enrollment(6.07) (2.80) Quatily of Sciencemd MathEducat ife Expectancy at Birth,years(62.10) (3.08) Prof.and Tech.Workers as %of the Lrkm F Access hSchools(1.BO) (2.32) PublicSpendingon Edwation as %of GDP Pakistan,NormalizationGroup: AU AduR LiteracyRate(%age 15 and above) (47.10) (2.48) Well EducatedPeople Do Not Emigrate A ge Years of S c M (3.68) (4.10) AvaiiabMyof ManagementEducation econdary EnroHmat(25.81) (3.00) Extentof Staff Training Tertiary Enrobnt (3.55) (2.60) Quality of Science andMsthEducation LifeExpectancyat Bkth,years (63.80) (5.32) Prof. and Tech.Workersas %of the Labor Access inSchds (3.00) (1.80)PublicSpendingon Educationas % of GDP Nepal,NormalizationGroup: Ail Aduk Lteracy Rate (% age 15 and above) (44.01) (3.40) Public Spending on Educstiinas % of GDP Average Years of Schooling(2.43) (59.90) Life Expectancy at Birth,years Secondary Enrollment (43.87) (5.37) Tertiary Enrollment SriLanka,NormalizationGroup: A1 Aduk LiteracyRate (% age 15 and above) (92.08) (2.26)Web EducatedPeople DoNot Emigrate Years of Schooling (6.87) (4.10) AvalWIlXy of Manage& Educatb condary Enrolhent(80.77) (3.50) Extentof Staff Tra- Tutiary EnroHment (5.33 (4.20) GualRy of Science and M hE W m fe Expectancyat Birth,years (n.sO) (9.60) Prof.andTech.Workers as %of the LaborForc emet Access in%hods (2.90) (1.31) PubicSpending on Education as %of GW GLOBAL TRENDS ININNOVATIONAND SITUATIONOF SOUTHASIAN COUNTRIES GlobalInnovation Trends 5.32 Innovation i s becoming a critical element o f competitiveness and growth as there i s greater mobility of factors, products, services, and knowledge. A larger percentage o f a country's economic growth can be attributed to more effective use of knowledge, even in developed countries. Countries behind the global frontier can dramatically increase their performance by improvingtheir ability to innovate. 5.33 Expenditures on R&D globally have been increasing, particularly the share contributed by the productive sector. In OECD countries the share of R&D in GDP averages around 2.2 percent and 70-80 percent is undertaken by the private sector. In developing countries the share o f R&D in GDP tends to average below 1percent, and 70-90 percent tends to be undertaken by the public sector. The R&D effort undertaken by private firms is oriented toward commercially relevant applications, while the R&D effort undertaken by the government i s usually more oriented toward basic knowledge and military research. As noted earlier, MNCs are the main producers o f commercially oriented knowledge and they are also the main disseminators of such knowledge throughout the world. That i s why it is important for developing countries to attract foreign investment that i s relevant to their development strategies. 5.34 The innovation system plays an important role in acquiring, creating, adapting, and disseminating knowledge, which i s crucial for success in the knowledge economy. It consists o f the network o f institutions, rules, and procedures that affect how the country acquires, creates, disseminates, and uses knowledge. Innovation in a developing country does not just concern domestic development o f knowledge on the global frontier. It also concerns the application and use of existing knowledge to the local context. For the countries o f South Asia, which are still far behind the global frontier in many sectors, tapping into and makmg effective use of existing global knowledge will have a greater economic impact than directing most of its resources to . develop frontier knowledge, no matter how prestigious the latter may be. 5.35 The concept of innovation encompasses not only technological innovation, that is, diffusion o f new products and services of a technological nature into the economy, but equally includes nontechnological forms o f innovation, such as organizational innovations. The latter includes the introductionofnew management or marketing techniques, adoption o fnew supply or logistic arrangements, and improved approaches to internal and external communications and positioning. 5.36 The concept of a national innovation system rests on the premise that understanding the linkages among the various actors involved in innovation are key to improving a country's technology performance. These actors include private enterprises, universities, research institutes, think tanks, and consulting firms. The innovative performance of a country depends to a large extent onhow these actors relate to each other as elements of a broader system. Linkages can take the form of joint research, personnel exchanges, cross patenting, licensing o f technology, purchase of equipment, and a variety of other channels. 95 Figure 5.4: Innovation Scorecardsfor SouthAsian Countries SoulhAsia (mostrecent) Gross ForetgnDiect krvestmerdas %of GDP (0.66) FeesPaymerds(t~ T(123.673 J censeFeesPelmentsId.~.(0.16) (3.83) State Of ChstU Dew and LicsnseFees Retapto($ nd)(4.67) (2.90) AWkMIy of Verhue C@d Researdrers hR8D (28737.50) (4816)CosttoEnforceaConbact(%of GMpercrpita) L4wasity-CMnpsny Research CoHaboratiim(2.42) (7628) Costto Register a @usmess(% of GM per capta) lormalization Group All1 India(mostrecent) GrossForagnDvedInvestment8s % of GDP (080) (3 50) Privatesectwsp UcenscFeesPaymrls(f ml)(350 W) (5 W) Hgh-TechExportsas % of Manuf kyardLLrenseFeesPaFentsIrrd p4, (033) (4 10) Stateof cluster andLcenseFeesRecapts (rg nrl)(I2 W) (033) PdCrt AppwcaDnsGrwtcdby andLicemeFeesRecepts I ml p4, (355 00) Pdnrt AWaUons Grantedbythe ebe4 8 Ugneerw EmdmerdRelio(% of $wmlts)(25 00) (3 59)Avdablbty of Venturecaplal ResearchsrsnR8D(95428 W) mw (4 m)Acfiun Burdenfa Start-Ups ResearchersnR8DImKm (98 85) (9 23) Swntific ardTedrdcd JwnalArtcles I ml TdalExpadhxsfa R8Das %of GDP(0 78) (9217 W) SckntIfbadT e c W JourndMlclct M w f Trade as %of GW (t 3 02) (95M1)Costto Enforcea C d r a d(49 of GN per capla) (% UFnvs&y-canpary ResearchCoWatim(3 ZD) Sa) Costto Regster a Busmess(%of GN per capta) lormalization Group: All Bangladesh(mostrecent) GrossFwagnDidhvestMnt as % of GDP (0.20) (2.30) Pride Sector S p e w on R aty and k m s e FeesPaymcrts(S wd) (3.00) (0.20) W T e c hE m s8s %of hianuf. y a y and LicenseFeesPaywefts /mt POD.(0.02) (320)State Of tkrstef DSV~pme yaty and License FCCSReceipts (I mu) (0.m) (0.01)PalentAppHc~sGrwtedbyth andLicenseFeffReceiptsI~M.pop. (I.W)PatentAppHcMiMsGrmtedby Engmeeriq EnrolmentR& (% of e4 StUderlS) (47.W) (2.00)Availabililyof Venture C esearchers nR8D (6097.W) (3.30) AMn. Burdenfw archers nRID ImUim(50.77) INormalization Group: All 96 PakMan(mortrecent) GrossF o r mDiredinvestmentas % of GDP (1.M) (2.30)Privde Sedw FeesPayments(S ms) (18.00) (1 .W)High-TechE m s as % of M m . LicenseF ~ a Fs' a y d s /nil.pop.(0.12) (4.20) Sde of uustuOsvelo andLicenseFees Recelpts(f ma) (2.00: 0.03)PatentApplic&ono Grantedby erd UcsnseFeesRccnptt I nil.pq t* S a m elevelstWlentS) EmmdmentRatio(% 8Enpneerlng(32.00) (2.40) Avrilabilty of VentureC@al ResearchersnR80(9977.00) lorrnalization Group All Nepal(most recent) Gross ForelgnW e d investmentas % of GDP (0 IO) (0.03)High-TechExpotsas %01 Manuf. Export nce 8 EngineeringEnrohneniRatio(% of tertiary students)(13.00) (1.73) Scientificand TechrucalJournslArticles Imilpop. Manuf.Trade as %of GDP (23.82) (39.00) Sckrdfic and TechnicalJwrnal M i c to Registera Business(%of GM per cap4ta) (44.20) Costto Enforce a Contract (% of GNI per capita) Normalization Group: All sriLanka(mostrecent) GrossFmeignNed hvestmentas %of GDP (1.40) ( 3 3 ) PrivateSectwS~~R&-&I mg Enrohrd Ratio(% of totiary kvel (I.m)W-Tech ~xportsas %of Manuf.E ear-s inR8D(3448.W) (3.80) Stateof aster Dev esmrchers inR8D/milon (197.03) (0.21)PatentA$+4cationt Grantedbythe Total Expendduefor R8Das % of GDP (0.18) (4.00)Patent ApplicationsGrantedbythe Manuf.Trade as %of GOP (46.25) (3.40)A v d a b qof V&e Capt! ivusity-CompanyResearchCc#aborstion (4.30)Admin.Wden for stat-up egster a Business(% of GNIper cap,ita) (4.61) Scientlfband T e W d JournalArticles Contract(% of GNI per capta) (7.60) (84.00) Scientific and TechnrdJowMlMiles I Normalization Group: All 97 Innovation in South Asian 5.37 As a region the South Asian countries do better on the innovation pillar than on any of the others, and that i s largely because o f the capabilities o f the large countries, India inparticular, but also Pakistan and to a lesser extent Bangladesh. The main strength comes from the large absolute number o f scientist and engineers in R&D as well as the number o f scientific and technicaljournal articles. Bangladesh also has a strength inthe very highscience and engineering enrollment ratios inhigher education, although that is diluted by the very low tertiary enrollment rates. See Figure 5.4. 5.38 Another area o f relative strength is a strong state of cluster development, although this is mostly concentrated in India (where it includes notjust IT services but pharmaceuticals, textiles, and metal engineering industries), Pakistan (medical instruments, sporting goods, textiles, and garments), and Sri Lanka (textiles and garments). 5.39 The overall formal R&D effort o f the South Asian countries i s very small. R&D expenditures as a share o f GDP average 0.48 percent, with a highof only 0.78 percent inIndia. In general, the vast majority o f the research done in the South Asian economies is done in public R&D labs. An area where India shows some strength though is in patenting. India has a large public research network and recently there have been some reforms which are strengthening the incentive regime to produce more commercially relevant output. 5.40 The private sector, with the exception o f some o f the larger Indian groups, does very little research. In addition, the relatively little research done by the public sector i s not commercially relevant, and there are poor mechanisms to get it out to the productive sectors. It i s also generally quite burdensome to start up new businesses, particularly technology-based business that also face the additional challenge or raising funds for risky new technology projects. The overall business environment is somewhat more supportive inSri Lanka. 5.41 Another area o f weakness o f the innovation system in South Asian countries i s the poor links between university and company researchers. This is a little stronger in India than in the other countries, but is still quite weak by the standards o f developed countries. 5.42 However, it should be noted not all innovations are done through formal research. Inall countries there are informal innovation efforts. In India some o f this effort is being collected through an organization called the HoneyBee network, which has documented more than 12,000 small indigenous innovations, mostly in the agricultural sector. Furthermore, the government i s beginning to pay attention to supporting and scaling up this indigenous effort. 5.43 All five countries with a partial exception of Sri Lanka, however, do not draw very much on global knowledge. This i s revealed by the very low share o f foreign direct investmentto GDP, which isjust a fraction o f 1percent for all countries except Sri Lanka (where it i s 1.4 percent) and by very low formal purchase o f foreign technology as shownby very low royalty or licensing fee payments (US$0.33 per person in India where it is the highest followed by US$0.12 in Pakistan, and virtually nil inthe others). This contrasts with the situation o f East Asian countries where the 45The innovationscorecardshave asmallglitch inthat the variables on the cost of registeringacontractandon the cost to register a business were inversely scaled. In their case, the higher the cost, the higher the ranking shown. To be consistent with the scaling of all the other variables, they should have been scaled inversely: the higher the cost, the lower the ranking.This inversionshould be taken into account ininterpretingthe high rankingson the two variables in this section. 98 average share of gross direct foreign investment as a share o f GDP i s 8.26 percent, and the averageroyalty and licensing fees per population are US$30.82. 5.44 In addition, again with the exception of Sri Lanka, the share of manufactured trade (imports and exports as proxy for access to embodied knowledge and pressures to keep up with global technology) in GDP i s less than 25 percent (and in India it is only 13 percent) compared to an average of 99 percent for East Asian countries. 5.45 Inshort, although the SouthAsian countries do relatively betteronthe innovationpillar than on the other, they still have relatively small R&D effort relative to their needs. Also, the little that they do is not well integrated into the production system. In addition, as will be stressed below, they have a majority of the population in traditional subsistence sectors of the economy, and very little of this effort gets out to them except for some work on agricultural research and extension. Finally, these countries are not drawing very much or very effectively on the rapidly growing stock of global knowledge. OVERALL ASSESSMENT 5.46 The South Asian countries are significantly behind the global frontier in education and innovation and ICT. While their GDP growth has been higher than the world average, they also have high population growth rates. Competing on the international stage i s becoming more demanding. 5.47 There i s considerable diversity among the five South Asian countries analyzed in this paper. India is clearly ahead in its skdls, technology, and innovation capability because of its much larger size and the critical mass in the absolute number of highly skilled population; number or researchers in R&D, resources allocated to R&D, and the vast network of public research laboratories, universities, and large private companies that are already undertalung research. Nepal is at the other extreme because o f its very small population, much lower per capita income, and much less-developed technology infrastructure. 5.48 Keeping in mind these differences in scale and in the degree of development of their human and technological infrastructure, there are still some generic actions that all the countries can take. These will be treated under five headings, although the details under each will have to be adjusted to the specificities o f each country. However, there i s an ovemding reality that cuts across all these recommendations that needs to be addressed. It i s that inall South Asian countries the modem sector i s just a small part of their economies. Two-thirds or more of their population arerural, and agnculture i s still a largepart of economic activity. The share o f agriculture inGDP ranges from 19 percent in Sri Lanka to a high o f 41 percent in Nepal, compared to the world average o fjust 4 percent. Inaddition, with the exception of Sri Lanka (where it i s only 8 percent) the share of the populationbelow US$1 a day i s inthe double digits, ranging from 13 percent in Pakistan to 38 percent inNepal. Furthermore, the South Asian countries have very highilliteracy ratesranging from 29 to 50 percent for menand from 51to 74 percent for women. See Table 5.6. Table 5.6: Various Proxv Indicators of Traditional Sector 99 5.49 This means that there needs to be a massive effort to increase the educational and technological level of the large population that i s still not integrated into the modem sector. This population tends to be largely the rural population, although it also includes a large portion of the urban population elung out a marginal subsistence in urban slums in the informal urban sector. This will be a major focus ofpolicy action and will bebuilt into the recommendations for actions that follow. An additional recommendation will be the value o f networlung and cooperation among the South Asian countries. KEYACTIONS Improvingthe EconomicIncentiveandInstitutionalRegime 5.50 Although Sri Lanka does better than the others, South Asian countries need to improve their economic and institutional regime. This includes increasing the depth and flexibility of capital and labor markets, strengtheningsocial safety nets and competition policy, and improving governance and the rule of law, all of which are part of the basics o f development. From the knowledge economy perspective, an additional area that needs to be improved is participation in international trade. Imports of capital goods, components, and products embody much knowledge. Exports also force companies to reduce costs and improve quality, which stimulates greater technological effort. With the exception o f Sri Lanka, where the ratio of exports and import o f goods to GDP i s 57 percent, the share o f trade in goods to GDP i s below the weighted world average o f 42 percent in 2003. The share in India is one of the lowest inthe world at only 21 percent (World Bank 2005). India is missingout inthe benefit of specialization and exchange and access to global knowledge. China, where the trade in goods i s 60 percent of GDP, has benefited greatly from this greater global integration. Clearly there i s room to lower tariff and nontariff barriers in the South Asian economies to create more competitive pressure to improve performance as well as to get greater access to embodied global knowledge. StrengtheningEducationandSkills 5.51 As notedearlier, education and skills are critical for countries to be able to make effective use of new knowledge, let alone to develop new knowledge. Here the challenges for South Asian countries are enormous. They have five challenges: Expand literacy. All five countries have female illiteracy rates over 50 percent and male illiteracy rates from 29 to 50 percent. Clearly they all have to undertake major literacy drives. Expand basic formal education to the large number of children not in school and who drop out. Primary enrollmentrates are still less than 100 percent, and primary completion rates are even lower. The share of first grade students reaching the fifth grade i s only 61 percent for males and 65 percent for females (World Bank 2005). At the same time they have to expand secondary education. Secondary enrollment rates averagejust 49 percent. Expand higher education where enrollment rates average just 6 percent compared to a world weighted average of 26 percent. Improve the quality of education at all levels. Most countries have outdated curriculums and do not teach the basic shlls needed by students to continue to learn throughout their lives no matter at what level they leave formal education. Develop ways of effectively retraining adults who have left formal education, but need new skills demandedbyrapidly changing technology. 100 5.52 Hard choices of priorities on budget allocations will be needed to be made as public sector budgets are clearly insufficient. Moreover, it is necessary to move on the various fronts as they are all critical and none can be completely unattended. Thus, ways have to be foundto make more resources available for the challenge. 5.53 Three types o f actions can be helpful. First, increase efficiency in the use o f existing budgets. In all countries there are many inefficiencies and waste in educational expenditure. Better monitoring and incentive mechanismneedto beput inplace. 5.54 Second, make effective use o f the private sector. This includes charging tuition for private education, particularly at the higher educational level. China, for example, has increased enrollments at the entering class at the tertiary level by 50 percent per year since 1997 and has been able to increase the enrollment rates from 6.5 percent in 1997 to 21percent today. Itdid this inpartbyraisingstudenttuitionsto cover from25 to 40 percent o fpublic tertiary education costs. It also includes allowing entry o f private providers o f education, again especially at the higher education level since the government has the obligation to provide basic education as a public good. InChina there are now more than 4 millionstudents innongovernment institutions. 5.55 Third, use the potential o f new technologes to extend access and improve quality of education. In India, for example, Tata Consulting has developed a computer-based functional literacy program based on a multimedia program with animated characters and voiceover that allows students to acquire a 300-500-word vocabulary in their own language or dialect in30-45 hours.46 It is being piloted successfully in Andra Pradesh and has the potential to help make a large proportion o f India functionally literate injust a few years. There is also great potential to add new Internet- and satellite-based educational delivery mechanism to more traditional radio- or television-based distance education program^.^' TappingintoGlobalKnowledge 5.56 All five countries could do better at tapping into global knowledge inaddition to greater trade with the rest o f the world. All could benefit by trying to attract more direct foreign investment. India inparticular has great potential. Surveys o f investors such as those done be A. T. Kearny find that executives rank India as a top potential destination. 5.57 However, little o f this investment materializes for various reasons. One i s that many sectors are still reserved for the state and other are only partially open. The second is the poor power and road infrastructure. The latter, in particular, inhibits investments that have to move goods in and out o f the country. By contrast, investment inthe services area that need only good Internet bandwidth such as back-office functions, call centers, engineering services, and software are not as inhibited because all they really need are good telecommunications andbackup power. Liberalization o f foreign investment restrictions and more proactive policies to attract foreign investment need to be put in place. Recall that the MNCs are the main producers and disseminators o f appliedknowledge. 46 See DahlmanandUtz (2005). 47Cisco for example has developed a full day-by-day Internet-based cumculum to teach math in Arabic from kindergardento twelfth grade that is beingpiloted inJordanin400 discoveryschoolsthrough ajoint programwith the Ministry of Education. It has also developed online training programs for skills ranging from installing local area networksof its own systemsto basicplumbing andelectricity for workers inthe UnitedKingdom. 101 5.58 South Asian countries also are not very active in purchasing foreign technology. Licensing payments per million population are extremely low, compared to East Asian countries (average o f US$0.16 in South Asia versus US$30.82 average in East Asia). Overly restrictive technology transfer policies that inhibitpurchase o f foreign technology need to be relaxed. 5.59 Another means o f tapping into foreign knowledge that needs to be strengthened i s attracting back nationals who have gone abroad for studies and have acquired not only academic knowledge but practical experience fi-om working in foreign countries. All the South Asian countries have a significant Diaspora that can be tapped. Economies such as Korea and Taiwan started doing this decades ago with great success. China has also launched major programs to attract its Diaspora by offering attractive incentives and special h-tech parks dedicated exclusively for returning nationals. While India has begun to do some o f this, it is still far behind China, and other countries in South Asia are even further behind. 5.60 Finally, much more can be done by South Asian countries to take advantage o f the extensive amount o f technical information and knowledge that can be tapped through technical publications and databases, especially now with the advent of the Internet. Some o f the issues here are how to reap economies o f scale in subscriptions to and dissemination o f such information. Creating Knowledge 5.61 As noted, the five South Asian countries, with the partial exemption o f India, are relatively weak in creating knowledge. Part o f the reason for that is the still relatively weak pressure for improving performance for the economy as they generally are less open to international trade (although Sri Lanka is an exceptionhere). It could be argued that they all need to increase expenditures on research and development. However, before pushing for an increase in public resources for R&D (as the bulk of spending is currently done by the government) it would be appropriate to improve the use o f the resources that are currently being allocated. The productivity o f the public resources allocated to R&D i s very low. It i s necessary to improve the efficiency inthe allocationofpublic resources to R&D. 5.62 Some o f the mechanisms that should be strengthened inthe allocation o f public funds for research are competitive bidding, peer review, and more monitoring and accountability inthe use o f the funds. In addition, it i s necessary to improve the micro incentive regimes in the public R&D labs to that they use their resources more efficiently to focus on the needs o f the respective economies. 5.63 It is also necessary to get the private sector to do more research o f its own. Increasing competitive pressure inthe economies should help. Inaddition, the government should stimulate more private research by strengthening various incentive mechanisms including matching R&D grants, tax subsidies for increases inR&D, and special programs to foster collaboration between public R&D labs, private firms, and universities (for example, by earmarking some o f the matching grants for programs that actually involve such collaborations). 5.64 The latter should also include a review to remove disincentives for university researchers to do collaborative work with industry. Instead there should be strong incentives for pubic laboratory and university researchers to work with industry by allowing them to share in the benefits o f the technologies developed through consultancy contracts and profit sharing in royalties. There should also be a special emphasis on publicly funded research to address the needs to the very large part o f the population in all these countries that still live at close to the 102 subsistence level. These should include preventive health programs as well as programs to help them make more effective use o f their limited resources, as well as how to extend to them the benefits o f informationtechnology by developing low-cost ICT solutions for them. 5.65 Inall countries more can be done to set up special mechanisms and institutionsto foster greater technology spin o f f from public research labs and universities. These mechanisms should include technology transfer offices in the public labs and universities, science-based industrial parks, and business incubator programs. These will make more sense in the larger economies such as India, Pakistan, and Bangladesh than in Nepal and Sri Lanka, although Sri Lanka i s making efforts in this direction even in spite o f its small size. Inaddition for the start up o f new technology-based firms, it i s necessary to help these f m s to develop realistic business plans and provide them assistance in getting capital as it i s very difficult for start ups with no assets to get loans from the banking sector. 5.66 It is also necessary to integrate these policies to allocate more resources to research and to strengthen the researchinfrastructure and collaboration among the main innovation agents with policies to develop the hgh level scientific, engmeering, and technical persons who are to create new knowledge. This should also include strengthening business and entrepreneurship programs for scientists and engineers so that they are more attuned to identifyingbusiness opportunities. 5.67 The governments also have to strengthen programs to support grass root innovations and to support the scale up and broader dissemination o f these innovations, as is being done by the Indian government. Disseminating Existing Knowledge, Especially to the Very Large Traditional Sectors inEach Country 5.68 This i s perhaps the most important area given the very large percentage o f the total population in these economies that has not yet benefited from the modem economy. Inaddition, this does not have the risk involved in the creation o f new knowledge. It is the application and adaptation o f already existing knowledge. In some cases there i s the issue o f purchasing existing proprietary knowledge. However, there is a tremendous amount o f knowledge in the public domain that is not being used. This may be due to lack o f information on its existence, lack of understanding o f its relevance, lack o f education or skills to make use o f it, or lack o f access to complementary inputs or supporting infrastructures and institutions to use it efficiently. A combination of these factors explains why it i s not so easy for developing countries to catch up with developed countries, or for poor people to use knowledge that may be very relevant for them. The most relevant actions to address the lack o f use o f knowledge differ according to the nature o f the problem (Table 5.7). They range from the more straightforwardaction o f increasing access to information to the need to build up domestic institutions and domestic capabilities in terms o fpeople's skills and ability to make effective use o f knowledge. 103 Natureof the problem Actionsto solve the problem Lack o f information ofexistence o fknowledge Provide more technical informationthrough print, radio, television, andInternet Lack o f understanding of its relevance 0 Educate people on the value and relevance of different types o fknowledge and how to look for an access to it Lack o f education or skills to be able to use it Beyondgeneral educationtrain people with the specific skills necessary to use the knowledge Lack o f access to complementary inputsor Provide easier access to finance supportinginfrastructures and institutions to be 0 Provide extension services and other able to use it technical assistanceto demonstrate the proper use o fthe new knowledge and to support its application; remove regulations that make it difficult to access the knowledge or to start upnewbusinessor activities with it Provide a supportive environment interms o f good rule o f law and contract and enforceability o f contracts 5.69 Butwhile it is easy to say this needs to be done, it is quite difficult inpractice. There are various mechanisms that need to be strengthened to disseminate knowledge. They include technical information services; extension services in agriculture, industry, and services; productivity organizations; and strengthening the metrology, standards, and quality control infrastructure. These should be given stronger priority than they are usually given by most governments that tend to focus too much on developing new knowledge rather than supporting the broader dissemination o f knowledge to the large populationintheir economies who are still in the traditional sectors. 5.70 The kinds o f knowledge that are needed for development are not just hard technical knowledge such as how to manufacture engines or produce steel or fertilizers or petrochemicals. Also important are organizational and managerial knowledge: how to set up transportation or distribution systems, and not just for goods but also for services. Some examples are how to set up effective tax collection andrevenue administrationsystems for government or how to provide effective health and education services or business services. NETWORKINGCOLLABORATIONAMONGSOUTHASIAN COUNTRIES AND 5.71 A sixth action recommendation is to set up a system for networking and collaboration among the South Asian countries inthe areas o f education and training as well as in technology and innovation. As noted, in spite o f their differences, there are many generic actions similar across the South Asian countries, and all the countries share the need to find more effective ways to extend education and technology to the large part of their population outside the modern economy. Sharing the experiences that each country has in dealing with the issues identified would be very beneficial to others. 104 5.72 Inaddition countries that are less advanced ina particular area could learn from those that have more experience o f successful programs in that area. There i s also scope for collaboration across countries in tackling similar issues, and even in doing joint research on common problems. An excellent example o f such knowledge sharing across a region as well as a formal framework for joint research i s given by the European Union's program in education and inresearch. 5.73 In 2000 the European Union set itself the ambitious target of becoming the most competitive and dynamic knowledge-based economy o f the world by 2010. To achieve this the European Union developed numbers o f program to transform education and training. These include sharing o f experiences and working toward common goals and learning from what works best elsewhere. It also includes programs to encourage mobility o f students teachers and researcherswithin the European Union and with outside countries, as well as programs for mutual recognitions of degrees. 5.74 The European Union has even more ambitious collaboration programs in the area of research. The Sixth Framework Program for research was launched in 2002 with a budget of 17.5 billion (about 5.5 percent o f the total R&Deffort), and was subsequently increasedto 19.2 billion, to foster joint research across countries and between firms, universities, and enterprises and includes the creation of 25 technology platforms as well as an openmethod of coordinationto contribute to policy learning and policy integration by encouraging and facilitating mutual exchange of knowledge and best practice.48 5.75 The European Union's programs are extremely ambitious and may be to more than what i s feasible at the current time for South Asian countries, because South Asian countries do not yet have the formal integrationof the European community. However, these programs do indicate the importance o f cross-country collaboration and learning within a region, as well as the rationale for funding of research across national boundaries. Clearly it would make sense to start some of this cross-country sharing of experience and learningamongthe South Asian countries, and even consider some more formal exchange andjoint-funding programs. 5.76 Inconclusion, there is a greatneedto upgrade technology, skills, and innovationinSouth Asian economies to help improve their productivity and growth potential and to increase the welfare of their people. Fortunately, there i s useful experience of the kinds o f policies, mechanisms, and institutions that can be used to do this. The first thing that is needed is to raise awarenessamong policy makers, the businesscommunity, and the populationat large o f what can be done. The second step is to launch some concrete programs that begin to make a difference in order to demonstrate what can be done. The third i s to publicize the scale up of the successful project. 48 For more information see the European Union's Web site and look under the variouseducation and research programs: http://europa.eu/index-en.htm. 105 106 6. POWERSECTORREFORM,PRIVATEINVESTMENT,AND REGIONALCOOPERATION49 INTRODUCTION 6.1 Modeminfrastructure, particularly electricity, telecoms, and roads, i s critical to economic development. Electricity provides light, the ability to use modem equipment and computers, and access to information and communication technology (ICT). Telecoms facilitate information exchange and access to the rest o f the world, while transport infrastructure i s critical for trade, and, by lowering transport costs, extends the market and increases competition. Studies o f the productivity o f infrastructure (Canning 1999; Canning and Bennathan 2000) suggest that infrastructure has strong complimentarities with other human and physical capital. If there i s a surplus of infrastructure, more investment adds little to total output, but ifthere is a deficit, then shortages constrain total output, magnifying the impact, so that the return to reducing that deficit can be very highindeed. 6.2 This can be seenmost clearly for electricity. Once there is an adequate reserve margin o f generation and adequate transmission and distribution to deliver the power to customers, more capacity has almost no extra value, and the efficient (and competitive) price o f power falls to its short-run avoidable cost, essentially the cost o f the fuel used in the least efficient plant dispatched. Ifthere is a shortage, the value o f lost load can be tens or even a 100times as high.In Britainduring the period 1990-2001, the wholesale market (the electricity pool) set a capacity payment based on the value o f lost load, initially taken as ;E2,000/MWh (megawatts per hour) when the average wholesale price net o f the capacity payment was less than E20/MWh.50The value o f lost load reflects the considerable inconvenience o f unexpected disconnection, while the value o f unserved energy ina country familiar with power shortages may be lower, as users take precautions such as installingbackup or stand-alone systems. Nevertheless, these are often many times as costly as reliable centrally generated power, showing the potentially high returns to investing to deliver that power, particularly to customers with a high willingness to pay (commercial and industrial customers inparticular). 6.3 The demand for infrastructure, and particularly electricity, is growing rapidly in the region, and at low levels o f income per head, can be expected to grow more rapidly than GDP as the economies modernize and shift resources from agriculture to industry. Figure 6.1 shows the electricity intensity o f a selection of South Asian and other countries, measured by production o f electricity per thousand US$ gross domestic product (GDP) (at 1995 constant prices). Indian electricity production i s rising considerably faster than GDP, as i s that inPaktstan and even more so inBangladesh (although the arithmetic scale may not show this clearly). 49This Chapter was preparedby DavidNewberry,Professorof Economics, Cambridge University, England. The views inthis paperare solelythe author's anddonotnecessarilyreflect those ofthe WorldBankor its ExecutiveDirectors. The pricepaidto generatorswas the sum of the system marginalprice, the price of the most expensivebid accepted (iftheyweredispatched)andthecapacitypaymentfor plantdeclaredavailable. Figure 6.1: Comparisonsof ElectricityIntensity ofSoutheastAsia, the UnitedSates, China, and the EuropeanUnion-15 Electricity Intensity SE Asia and comparators1971-99 1800 1600 1400 -China 1200 wiz -India e 1000 +Pakistan 3 zf -Malaysia 800 3, +I+SriLanka 600 -Bangladesh +Singapore 400 200 0 Table 6.1: Ratesof Growthof Energy Intensity 198688 to 199698 (percentper annum) Source: World Bank 2002. 6.4 Table 6.1 gives the rate of growth of electricity production and of GDP (at constant US$ and also constant purchasing power parity (PPP) dollars) and of electricity intensity (which is also the rate of growth of electricity production less the rate of growth of GDP)." India (and even more so China) looks surprisingly electricity intensive at market exchange rates (nearly three ~~~ ~ PPP is a measure of the real standard of living, taken from the Penn World Tables version 6. The relationship betweenthe growth rates is not exact as the ratesof growth are found by averagingthe initial and terminalvalues over three years. 108 times as much as the United States andmore than four times as much as the European Union-15). This continues to be true for India and Pakistanwhen GDP i s measured at PPP. Figure 6.2 shows consumption (rather than production) per thousand $PPP, 52 together with the predicted consumption for South Asian countries as a whole (excluding Bhutan) using the regression estimates presented in the annex. India and Palustan track each other closely, and are considerably more electricity intensive than (more than three times as much as) the smaller countries, Sri Lanka, Bangladesh, andNepal, and almost twice the predicted level. Figure6.2: Electricity ConsumptioninSouth Asia Comparedto Predictedper US$ Thousand PPP Electricityconsumptionper $'OOO (1996)PPP 325 i l 300 275 250 225 xE 200 0 f 175 .-S 0 150 g8 8B125 100 75 50 25 0 6.5 Thus, whether electricity intensity is compared at market exchange rates or at PPP rates, India and Pakistan appear more electric intensive than might be expected. One obvious explanation o f this high intensity is that electricity i s underpriced in many countries (both directly, and effectively through the failure to collect bills and prevent theft). Countries that have a lower than expected electric intensity usually also have a low penetration o f electricity, particularly in rural areas. In that respect India appears to do quite well, given its per capita income. 52 The difference between consumption and production is mainly losses, except for exporting countries like Bhutan (not shown) or India (which imports a very small share o f consumption from Bhutan). PPP measures come from the Penn World tables and attempt to correct for different relative prices in poorer countries, primarilythe lower prices of nontradables, as well as distortions that cause differences between domestic andworld prices for tradables. 109 THEPROBLEM 6.6 Not all South Asia countries suffer from the same problems, but as a generalization the region still has the legacy o f state-owned vertically integrated electricity supply industries, often with the characteristic politicization o f tariff setting that leads to excessively cheap electricity to domestic consumers, high levels o f nontechnical losses (that is, theft or failure to collect bills), high levels o f debt or arrears, high levels o f manning, and poor commercial performance (as measured by the ability o f revenues to cover costs). A s a result, it is difficult for the sector to finance its investment needs on commercial terms. The shortage o f revenue leads to poor maintenance with frequent equipment failures (for example, as measured by transformer failures and low generation availability), resulting inpower shortages and load shedding. Figure 6.3 gives time series of losses as reported by the World Bank, although for India these are considerably below those reportedby various states. Figure6.3: Electricity Losses inSouthAsia and China Losses as percent of generation 401 I 35 30 +Pakistan 25 +Nepal -India 15 +Srilanka I Source: World Bank 2002. I 6.7 The India National Electricity Plan notes that "the country faced energy shortage o f 7.1 percent and peaking shortage of 11.2 percent in 2003-04" (CEA 2004). ICRA (2004) gives statewise score card data for India for 2004 stating that "the power sector inthe country i s grossly over staffed leadingto low productivity.. ." and "the proportion o f billing on metered basis at less than 50 percent o f energy input into the system" (ICR4 2004, p. IO). "Despite progress, the coverage ... i s very low and typically less than 35 percent" (ICRA 2004, p. 11). Taking a sample coverage o f costs through revenues is still low for most states." "... in the North East, the o f states, we find that for Delhi "the generation plants are aged and have a low PLF (48.5 percent) improvements inthe low level o f metered billing ... reduction inthe highlevels o f AT&C losses and low availability (62.2 percent). The commercial viability o f DISCOMs is contingent upon o f 52.8 percent and improvement indistribution infiastructure. The power sector as a whole has 110 6.8 For Mlxh and mly c tb 5 IIf 6.4.'4 been advocating the multi-year tariff policy as per G o 0 directions to embark upon long-term business plans by power utilities." Lengthy legal litigation has delayed the introduction of SERC's tariff orders, and generation availability i s 10 percent below norms. "AT&C losses were quite high at 54 percent ..." and ... long-term viability o f the power sector reforms heavily " depends on the state govemment's support and far reaching operational improvements in distribution segment" (ICRA 2004, p. 61). 6.10 The former Minister of Power, Yoginder Alagh, who introduced the Electricity Regulatory Bill to the Lok Sabha in 1997, noted that "by early 2001, SEBs [State Electricity Boards] as a whole faced an average 50 percent level o f technical plus non-technical losses, and they collectively owe around [uS]$5 billion to the Government o f India undertakings" (Ruet 2005a, p.13). The situation does not seem to have improved since then, with losses o f the SEBs overall reported as Rs21,OOO crores (about US$4.2 billion) by the prime minister in May 2005. H e expressed concern over continuing electricity shortages, and argued that the power sector neededurgentreforms, including ~nbundling.'~ 6.11 These problems are not peculiar to India, although their sheer scale there dwarfs those elsewhere. Thus, Bangladesh suffered energy shortages for much of the 1990s. Inthe fiscal year to June 1998, the Bangladesh Power Development Board, the main electricity producer, provided unintermptedsupply on only 49 days. Mucho f the time, 25 percent o f peak power was unserved (World Bank, 1998). Unreliable power is estimated to have led to a loss of 10 percent of industrial output. As industry accounts for 15 percent o f GDP, compared to electricity at only 1 percent, the social cost of electricity shortfalls are substantially larger than just the value of the unproduced power. Only 2,400 megawatts (MW) or 77 percent of nameplate capacity of 3,100 MW was available in 1998. The plant load factor was only 55 percent, despite excess demand. The poor availability and load factors result from poor maintenance and plant derating. This situation appears to have continued, with the Energy Information Agency (EM) reporting in August 200556that "The World Bank has estimated that Bangladesh loses around US$1 billion per year in economic output due to power outages and unreliable energy supplies. [Olnly two- thirds of Bangladesh's total electric generating capacity is considered to be `available.' Problems inthe Bangladeshi electric power sector includehighsystem losses (up to 40 percent), delays in completion o f new plants, low plant efficiencies, natural gas availability, erratic power supply, electricity theft, blackouts, shortages of funds for power plant maintenance, and unwillingness of customers to pay bills. Overall, the country's generation plants have been chronically unable to meet systemdemand over the past decade." 6.12 The same EM sources7notes that in Pakistan "Rotating blackouts ("load shedding") are, however, still necessary in some areas. Losses are about 30 percent, due to poor quality infrastructure and a significant amount o f power theft. Periodic droughts affect the availability of hydropower." The World Bank notes that in the KESC "System losses have increased from 17 percent in 1985-6 to 40 percent in 2001-2. The experiments with public sector management through non-traditional methods including the induction o f army personnel in uniform as top managers since 1999 have not shown any signs of significant improvement" (Alexander, Raza, and Wright 2003). The government of Pakistan still heavily subsidizes the power sector.'* Losses 55The Hindu (2005). 56Available at http://www.eia.doe.gov/emeu/cabs/bangla.htrnl. "Availableathttp://www.eia.doe.gov/emeu/cabs/pakistan.pdf. 58The Water and PowerDevelopmentAuthority was reportedinMay 2005 as requesting Rs26 billion (nearly $US500 million) for 2004-5 (http://www.dawn.corn/2OO5/05/04/top5.htm), while inNovember 2004 the Daily Times reported that the federal governmenthaddecidedto pay Rs15 billion (US$250million) "to subsidizeelectricity for domestic and 112 inKESChave fallen a small amount since then, as havetotal losses for the public utilities (down to 27.6 percent for the year up to June 30, 2003, according to the Pakistan Energy Yearbook 2003). REFORMS 6.13 The high rates o f growth of electricity production shown in Table 6.1 (typically 7-8 percent per anum [p.a.]) and the high levels o f unserved demand in some parts o f the region (certainly in Bangladesh, India, and Pakistan) appear to require high rates of investment in generation if supply i s not to become an increasing constraint on growth. State and central budgets are under stress, and the electricity companies are often effectively bankrupt, so the apparent solution has beento bringinprivate capital. 6.14 Under pressure from the International Financial Institutions (IFIs) and prompted by the apparent success o f reforms in Latin America, many countries in the region have considered or embarked upon reformprograms to allow private investment inthe sector. The first step involves passing an electricity law to allow private investment, then establishing regulatory agencies to set tariffs, unbundlingthe natural monopoly transmission and distributionbusinesses, and in some cases privatizing distribution companies and some generation assets. The typical form of private participation has been by independent power producers (IPPs) signing long-term Power Purchase Agreements (PPAs) with the single buyer (normally the incumbent power company or SEB, but the standard model recommended i s with a separate transmission company buying in a nondiscriminatoryway from existing and new generation companies). 6.15 The results of these reforms have often been disappointing (Ranganathan 2003). 59 Elsewhere inAsia, currency crises underminedthe ability o f the single buyer to honor the PPAs, which were often largely denominated in foreign currency (Newbery 2002). More generally, the tariffs needed to finance foreign direct investment (given the perceived level o f risk and the short tenor o f most debt finance) has led to high initial charges for electricity purchased from these IPPs. The mismatchbetween the cost o f these new PPAs, the average cost o f existing generation (withtariffs based on written-down asset values and often underpricedfuel),60 the lower average tariff o fretail electricity, and the even lower average revenue per unit generated, placed the SEBs or their counterparts under increasing financial stress. 6.16 Ruet's (2005a, b) analysis o f the problems of the SEBs i s that they currently act as administrative bodies that are unresponsive to incentives, and for whom the concept o f cost do not apply. Their concept o f rationality involves following procedures (particularly based on paperwork) rather than taking cost-benefit-based decisions. The lack o f proper management accounting, the absence o f a proper internal allocation o f discretionary power (to take sensible day-to-day operational decisions), and the pre-eminence o f administrative rules, makes attempts to reform the structure o f incentives and property rights (through privatization) ineffective without first changing the nature o f the SEBs. Of course, there are major beneficiaries from this opacity o f accounting that may account for the extreme resistance to meaningful reform. Reflecting on the reforms o f the 1 9 9 0 ~Ruet (2005, p. 206) notes "the structural reforms as ~ agriculturalconsumersthroughoutPakistan" as the distributioncompaniescouldnot cover the cost ofthe subsidies and requestedgovernment support (Daily Times 2004). 59There i s growing recognitionthat the earlyenthusiasmfor privatization,particularlyby the World Bank, was "oversimplified, oversold, andultimately somewhat disappointing" (Kessides2005), but see Kikeri andNellis (2004) andKessides(2004). 6oInIndia,therequiredrateofreturnonassetswas set at 3 percentin 1947, which eveninrealterms iswell belowa sensible economicrate. 113 envisagedby the Government o f India and enforced by some States do not leadto any substantial changeneither of the organization nor of the administrative nature o f SEBs." 6.17 The solution Ruet proposes i s "enterprisation," to be contrasted with corporatisation, whichjust changesthe legal status of the SEBs, and which inany case would be a necessaryfirst step toward the kind of restructuringenvisaged by the reform program. The institutional changes required involve fundamental changes in management accounting, creating cash flow rights for the enterprise and allocating rights to and control over these cash flows to the relevant decentralized units, limiting the executive instructions from the state, while providing the information neededto expose corruption and clientalism.61 6.18 One central problemi s that much electricity i s not sold at remunerative prices, and a large fraction of customers are either not charged at all (agricultural consumers in some areas) or bills are not collected. The obvious solution is to install meters where these are laclung, to set remunerative tariffs by regulators charged to ensure that tariffs are cost-reflective and capable of financing both operations and investment of efficient companies, and then to privatize distribution companies to provide incentives to collect bills due. This strategy worked well in Chile (Gala1 et al. 1994), but appears not to have been successful in Orissa, where it was first tried. Apart from Ruet's diagnosis o fthe needto createproper enterprisesbeforeprivatization (as was done in Chile in a lengthy preparation to eventual privatization), there are several serious difficulties facing distribution companies in India. Dealing with the high levels of nontechnical losses requires installing (and reading) meters, ensuring that the meters are not tampered with, ensuring that those collecting the money are not corrupt, and protecting them when reading the meters and collecting bills, and, most important, having the legal authority and actual will to cut off those not paying. 6.19 One practical problem is forecasting a realistic set o f targets over time at which the nonpayment rate will be reduced (too low and the distribution company will make a windfall gain, while too stringent may cause financial distress, and an inability to make investments to reduce losses) (Ranganathan 2005). This requires sophisticated multiyear regulation (probably with profit-sharing arrangements) insulated from political pressures that keep tariffs low and provide free electricity. Here the record is disappointing, with frequent political reversals of cost- justified tariff increases. Of course, when the whole industry i s fraught with over-manning, poor maintenance, poor bill collection, and other obvious inefficiencies, it i s easy for politicians to argue that removing such inefficiencies would deal with losses without tariff increases, and there i s some force inthese arguments, as discussedbelow. OBSTACLESTO PRIVATEINVESTMENT 6.20 The main obstacle to private investment is the fear that once the investment is sunk, it will not be allowed to earn a remunerative return. The electricity sector is particularly problematic as private investors supply an essential service directly to a large fraction o f the voting population incompetition with underpriced supply from the state-owned sector. As prices will have to rise to ensure that the investments are remunerative, the price rise will be associated withthe reforms that brought inprivate investors, andwill be doubly resistedon that account. 6.21 Many o f the current beneficiaries of opaque accounting, cross-subsidies, patronage inthe appointment of regulators and senior management, and so on will have an interest inpreserving the status quo, including the low prices that deter efficient commercial competition. The fact that 6' See also IrwinandYamamoto (2004). 114 external bodies such as the World Bank are pressingfor such reforms provides additional reasons for populist resistance, for the price rises that are needed to ensure investment adequacy yield current pain while the benefit o f improved quality of service may be some way in the future, and beyondthe politician's invariably short time horizon. 6.22 Private foreign investors are wary o f investing in hydro capacity, which is both capital intensive, with long construction periods, and often subject to water management regimes that may conflict with power generation. Coal-fired power stations can be similarly problematic where they are dependent on domestic coal, as coal mining i s often fiaught in terms o f labor relations.62Coal-fired stations using imported coal could be economically attractive but may be discouraged if there are inefficient domestic mines whose employees may object. The logical choice for IPPs i s therefore gas-fired combined cycle plant usingeither indigenous piped gas (as in Bangladesh, Pakistan, or India), or where local gas is not available, liquefied natural gas (LNG) imports. India is increasingly turning to that source to supplement inadequate domestic gas. Petronet LNG has a 5 million tonne LNG terminal at Dahej that i s being expanded to 12.5 million tonne^.^^ Petronet is also setting up a new 5 million tonne facility at Kochi, and i s also talung over completion o fthe 5 million tonne terminal at Dabhol (as described below). 6.23 The Dabhol power plant in Maharashtra illustrates some o f the problems facing private investors. In 1992 India opened up the electricity sector to foreign investment, and officials visitedthe UnitedStates to encourage investors, an invitation that Enronrapidly followed up with a proposal to build a large LNG terminal to supply a combined cycle gas turbine generating station of about 2,000 MW at Dabhol, on the coast some 180 kilometers south o f Mumbai. Negotiations ensued with respect to the project contracts and led to the signing o f a PPA in 1993.64The first phase of 740 MW was commissioned in 1997 before the LNG terminal was completed and ran on liquid fuel (initially distillate but then naphtha). The price charged by MSEB to consumers for power was less than it cost to generate power at the Dabhol plant, given the high cost of fuel and the capital costs associatedwith the project. As the amount of power purchased increased, the financial ability of MSEB to pay came under increasing stress, in large part because of the government o f Maharashtra's failure to effect necessary (albeit politically unpopular) reforms in the power sector, such as charging market rates for the power produced. "Payment problems with the ... MSEB, however, prompted Enron-backed Dabhol Power Corporation (DPC) to serve notice o fbreach of contract on MSEB inMay 2001. Construction on phase IIwas halted in June 2001." 65 The resulting acrimonious dispute lasted from 2001 until July 2005, when settlements were reached between MSEB and GE and Bechtel (the surviving equity holders after the bankruptcy o fEnron) ofvarious arbitrationclaims. That does not stop some apparently uneconomic coal-fired stations being built. In June 2005, a consortium of the ChinaNational MachineryImport and Export Corporationand the Xuzhou CoalMining GroupCompany Ltd. signeda contract to run the management and production of the Barapukuriamine in NorthwestBangladesh. The only use for such coal would be for electricity generation in a country well endowed with cheap indigenous gas. See http://www.eia.doe.gov/emeu/cabs/bangla.html. 63 Petronet is ajoint venture between Oil and Natural Gas Corporation, the Indian Oil Corporation, the Gas Authority of India Ltd., the NationalThermalPowerCorporation,and Gazde France. 6.1There i s extensive materialon Dabhol; http://www.atimes.com/reports/CAl3AiOl .htmland for more recentmaterial and an archive http://www.rediff.com/money/enron.htm. The author should declare an interest as an expert witness acting for the some ofthe investorsin DPC, which was, however, settled before going to arbitration. The text is based solely on material publicly available from the web and whose accuracy is not guaranteed, and makes no use of any confidentialmaterialthat may havebeen seenby the author. 6s Fromhttp://www.eia.doe.gov/emeu/cabs/india.html. 115 6.24 The successor company Ratnagiri Gas and Power Pvt. Ltd. (RGPPL) with the National Thermal Power Corporation and the Gas Authority o f India Ltd (GAIL) subsequently took over the almost completed 2,148 MW plant.66On September22,2005, the Bombay HighCourt issued an order on a consent term jointly filed by the DPC, RGPPL, and the IDBI Bank-led lenders.67 Petronet LNG was then asked by the Ministry o f Petroleum and Gas to complete the LNG terminal that will supply the power plant. Effectively, what was originally the largest IPP inIndia has now been taken back into public ownership after a lengthy and costly dispute during which the 2,000 MW that was potentially available to deal with shortages that were typically o f the order of at least 2,000 MW (see Figure6.4 above) were not available to MSEBand the 6.25 Three problems compound the difficulties facing such plants in India. The first i s that while GAIL sells its domestic gas at a price below import parity (that is, subsidizes it), imported liquid fuels such as naphtha and distillate are taxed. There are sound public finance principles arguing that inputs into production should not be taxed (except to correct externalities such as pollution or C02emissions), and any taxes should fall on final consumers (Diamond and Mirrlees 1971; Newbery 2005).69As the public power is invariably subsidized to most final consumers, it i s particularly perverse to tax fuel inputs into ele~tricity.~' Not surprisingly, RGPPL appealed to the Maharashtra government in October 2005 for a "waiver in sales tax and excise in a bid to maintain the per unit tariff of Dabhol phase-I (740 MW) at around Rs3.60 as against Rs2.50 by use of LNG." 71 6.26 The second is that LNGprices are both volatile and typically linkedto oil prices, which have considerably increased inrecent years, underminingthe apparent attractiveness o f gas-fired power stations. The third i s that the energy cost o f gas-fired generation can rise above that o f indigenous fuels, encouraging dispatch centers to dispatch gas at lower load factors, further increasing the average cost o f electricity, and straining the contractual relationship with the IPP when it is the average andnot the marginalcost that isreported.72 6.27 Pakistan's experience with IPPs also dates back to the mid-l990s, under encouragement from the World Bank's Power Sector Development Project. The Bank's project goals included restructuringand privatization, investment, and technical assistanceto improve the operations and managerial efficiency o f the power system (World Bank 1994). Before that date, there was little investor interest inPakistan's power sector after the government first allowed private investment in 1992, because of the highduties on imported equipment and the time taken to deal with the bureaucracy. In 1994 the Government publishedits Policy Framework and Package of Incentives for Private SectorPower Generation Projects in Pakistan. This providedan attractive formula for setting the PPA terms (according to the World Bank a bulk tariff o f 6.5 UScentskWh indexedto fuel prices, United States and Pakistani inflation, exchange rate fluctuations, operations and 66 See http:/~www.tribuneindia.com/2005/2005072 Unation.hW2. 67 See http://www.financialexpress.com/fe_full_story.php?content_id=103402. See The Times oflndia, Oct. 10,2005. 69 The case for taxing distillate, giventhe prevailing emissions standards, while underpricingdomestic coal i s perverse on environmentalgrounds 70 Oil taxation is justified on a variety of grounds, of which the most relevant here are to prevent substitution of legitimatelytaxed roadfuels, andas part of an optimalimport tariff to cover the costsof maintainingsecurityof supply. Nevertheless, there shouldbe less distortingways ofmeetingthese objectives for largepower stationfuel supplies. 7' http://www.financialexpress.com/fe_full_sto~.php?content-id= 105176 72 The situationi s further complicatedby the terms underwhich the gas is purchased.LNG contracts are typically long- termtake-or-paycontractsandthe relevantopportunitycost maybe nearzero, arguingfor base-loaddispatch, while the apparent energy cost may suggest that it should run only at the peak. Such confusions undermine rational discussion between IPPs, regulatory agencies, politicians, and consumers who may appeal tariff decisions and may delay tariff adjustments and reduceinvestorconfidence. 116 maintenance costs, and so on), tax holidays, and a standardized security package that included model agreements. These and other incentives resulted in considerable foreign interest, and ledto the development of the Hub Power Company (Hubco). Hubco built a 1,300 MW oil-fired power station located on the Hub River estuary owned by a consortium o f International Power (United Kingdom), Xenal (Saudi Arabia), and Mitsui Corporation. It i s described by Water and Power Development Authority (WAPDA) as "the first and largest power station to be financed by the private sector in Southern Asia and one of the largest private power projects in the newly industrialized 6.28 The Hubco project, completed ahead of schedule and on budget (US$1.6 billion) in 1997, was structured through four detailed agreements following the standardized model: the PPA, the Fuel Supply Agreement, the Implementation Agreement, and the Operations and Maintenance Agreement. These four key agreements formed the security package against which project fundingwas secured. In1998atariffdispute causedthe suspensionofall dividendpayments for a three-year period until it was resolved in December 2001. Shares in the company are locally traded, allowing the foreign investors to withdraw equity and investinother power projects. Thus InternationalPower had cut its holdings inHubco from 26 percent to 16 percent by May 2004 but recently acquired 40 percent equity stake in the 586 MW Uch Power project, a 400 MW dual- firedproject to supply the textile industryinFaisalabad. 6.29 Reforms o f the state-owned companies continue, and WAPDA has recently been unbundled "in an attempt to create a more competitive, market-oriented environment." 74 However, "Due to weak investor interest, KESC was not privatized as planned during 2002 and the process ofunbundlingWAPDA, althoughformally completed byDecember 2003, hasnot yet created the needed financial and managerial autonomy" (World Bank 2004). WAPDA continues to control all financial flows in the sector, including practically all decisions on allocation of funds, while nontechnical losses and subsidies remain. KESC was reported to be finally privatized by the Daily Dawn on February 22, 2005, although the deal fell through, and the cabinet then approved the sale o f 73 percent to the second highest bidder, Hassan associate^.^^ The sale was finally signed on November 19,2005. 6.30 Investors appear to have responded positively to the Private Power Investment Board's announcement of three large power projects, "including a proposal from AES to develop a US$1 billion coal project inThar, and the announcement of increased investment to the tune o f US$1 billion from CDC Group's Globeleq, which already owns 50 percent of the Lahore-based Orient Power."76Inaddition, ifthe government finds enough unintemptible gas, Hubco may be allowed to build two gas-fired power generation plants (of 300 and 600 MW) at Karachi. Whether these positive responses will translate into investments will in part depend on resolving the sector's chronic financial problems. 6.3 1 Bangladesh has also been successful inattracting foreign private investment inelectricity generation. According to EIA, "Given Bangladesh's electricity supply shortage, in 1996 the government issued the Private Sector Power Generation Policy o f Bangladeshand began to solicit proposals fiom international companies for IPPs. Among the first IPPs were a 360-MW gas-fired combined-cycle plant at Haripur, which began operation in October 2001, and a 450-MW gas- 73WAPDA was created in 1958 and is one of two vertically integratedstate-owned electricity companies, the other being KESC, which serves only Karachiand surroundingareas. Together, WAPDA and KESC transmit and distribute all power inPakistan,morethanhalfto householdconsumers. See http:llw.hubpower.com/n/about.html. ''New 74 York Times (2005). ReportedinKhaleej Times,Nov. 3,2005. 76New York Times (2005). 117 fired combined-cycle plant at Meghnaghat, which began operation in November 2002. Both plants were sold to the British firm CDC Globeleq in December 2003. India's Bharat.Heavy ElectricalsLtd. completed a 124-MW gas-fired Baghabari generating unit inNovember 2001."77 6.32 It is worth aslung why Bangladesh (and to some extent Pakistan) appear to have been more successful than India in continuing to attract foreign private investment. The case of Bangladesh i s particularly interesting as it was, according to Transparency International, the most corrupt country in the world in 2005.78The most obvious reason is that indigenous gas is cheap, and Bangladesh was lucky inattractingAES to invest. AES is a company noted for its enthusiasm to buildpower stations inrisky parts of the world supported by PPAs with very reasonable terms. The combination of cheap gas, moderately cheap capacity charges and excess demand for power made the project attractive both to the investor and also to the government. Unfortunately, AES's share price fell from a peak o f $70 in 2000 to less than $2 in 2002, forcing asset sales, and making expensive foreign ventures both unattractive and infeasible. It i s perhaps noteworthy that AES sold to CDC Gl~baleq,~'which has a mission to help developing countries improve their power sectors, and may not be the best test of commercial willingness to invest in the depressed post-2000 power sector investor climate. 6.33 IfBangladeshhasthe advantageofcheapgas, thenPalustan,which appearsto have some 33 years ofreserves at current rates o f consumption, also might expect to be attractive to gas-fired private generation investment, but reliable supplies o f the required volumes appear to be problematic. It is hardly surprising that an oil-fired plant like Hubco (using expensive imported fuel) experienced difficulties over the tariff, although it i s not clear to what extent the domestic cost of fuel oil i s insulated fiom world oil price movements. Given that domestic consumers are still heavily subsidized, and nontechnical losses remain high (losses had only fallen from a high of 41 percent in 2002 to 38 percent in 2004, when KESC was to be privatized), one must be cautious injudging whether the apparent recent interest inprivate investment reflects confidence inthe reformprogramor reassurancethat the model agreementswill adequately protectinvestors. The fact that the army was called in to manage KESC in 1999 may indicate that there is more evidence of political commitment to reformthan inIndia, although as noted above their arrival in 1999 didnot noticeably improve KESC's performance.80 6.34 To summarize, private investors need confidence that the necessary contractual underpinnings (PPA, fuel purchase agreements, sovereign guarantees, and so on) will be honored, that any legal disputes will be settled rapidly and fairly, with appeal to expeditious international arbitration, and that the underlying causes of disputes (inability of the counter party, either the single buyer, SEB, or the distribution companies to pay because of inadequate revenue) will be sustainably addressed. Opening access to the national transmission grid i s one obvious step to reduce reliance upon populist-swayed state governments and bankrupt SEBs, providing there is enough transmission capacity and a sensible way o f pricing access and use, and providing large customers can secure reliable power as a result (whichmay require direct connection to the gnd rather than to the local distribution network). The Indian Electricity Act 2003 requires 77 http://www.eia.doe.gov/emeu/cabs/bangla.html. 78 Indiawas ranked 88 out o f 159andPakistanis ranked 144. 79 The CommonwealthDevelopmentCorporation (CDC) was set up in 1948 as the U.K. government's instrumentfor investing in the private sector in developing economies. In 2002, CDC launched Globeleq, an operating power company solely focusedonthe emergingmarketsofAfrica, the Americas, andAsia. "Improvement of the availability andquality ofpower supplyinKarachiis apriority for the government ofPakistan. To facilitate this, management control of the company (KESC) was handed to the army in 1999. The army managementhas made progress in a number of areas, such as reductions in commerciallosses, decreases in accounts receivable, and restructuring of the organisation aimed at increasing the quality of customer services and the profitability o fthe company." TheDawn (2005). 118 nondiscriminatory open access in transmission and the adoption o f multiyear tariff principles, while the creation o f the Power Trading Corporation in 1999 is gradually increasing power exchanges between state utilities and private generators, reaching 4.2 TWh in 2002-3 (Singh 2006). The five regional grids are interconnected with high voltage DC lines with a capacity of 5,000 MW (4 percent o f installed generation capacity) and interregional transfer capacity i s 9,450 MW carrying 12 TWh. POWERGRID has plans to increase this to 37,150 MW by 2012.*' If these principles are effectively implemented, they will go some way to improvinga commercial approach to the sector, although reforming the distribution companies remains critical. 6.35 If anything, private involvement is far more important inthe distribution sector than in generation, for without commercial distribution charging cost-reflective tariffs, the counter parties to any power contracts will be financially weak and the PPAs will lack the credibility needed to attract private investment into generation. Reforming the distribution companies therefore has highpriority. There is general agreement that sustained improvements will require privatization, although preparingthe companies or boards for privatization requires considerable care, not least inensuring adequate information (from metering, management budgetary systems, and so on) is available to regulators and investors before final privatization, to avoid costly mistakes and painful policy reversals (Ranganathan 2005; Ruet 2005a). ALTERNATIVE SOURCESAND USESOFINVESTMENTFINANCE 6.36 Foreign private investment in power has a major advantage but one obvious disadvantage. The advantage i s that it brings best practice in terms o f contracting and efficiency (both inconstruction and operation) that putspressure on the country's electricity supply industry to shape up, make necessary reforms, and establish sensible regulatory bodies and tariff-setting practice. The model agreements proposed by Palustan," and insurance against opportunistic tax and legal changes, may provide a good model, but will have to be supported by evidence o f enforcement. There is ample evidence from elsewhere that private ownership delivers more efficient construction and operation than state ownership. The disadvantage is that the cost o f finance i s likely to be high, as sovereign and regulatory risk are perceived as high. If a larger fraction of the finance could come fkom low cost debt finance (supplied either by the government through its state banks or IFIs, with suitable exchange rate guarantees), then the overall cost of finance will fall (although the equity component i s still likely to be costly, its share may be low enough to reduce the overall cost). 6.37 Some South Asian governments have now accumulated considerable foreign exchange reserves that allow both increased domestic lending and foreign exchange rate guarantees, although it still ought to be preferable to agree conditions under which IF1 finance becomes available, as this i s likely to give better signals to the private investment community. Clearly, the lending by India to Bhutanto finance dams (discussed below) suggests that public funds can be used effectively; and for high capital cost low running cost projects such as dams and transmission, cheap finance can be critical to economic success. 6.38 The central problem in making use o f this cheaper finance i s that unless the distribution companies are reformed to become commercial and regulated to set (and collect) cost-reflective tariffs the revenue flows even to service cheaper debt will be laclung. If capital (including the revalued modem equivalent asset value of existing plant) earns a sensible retum (8-10 percent *' ~~ http://www.cea.nic.in/power-systems/National-Elec~ci~-Pl~index-National-Elec~ci~-Plan.h~published in July 2005. 82This requires PPAs that reward capacity availability and passthrough energy costs at a specified energy efficiency to encourage efficiency operation, to give sensible average and marginal cost signals to the dispatch center. 119 real) then the electricity supply industry would become largely self-financing at current demand growth rates (from Table 6.1 of 7-10 percent). Ifpower were sensibly priced, then excessdemand might rapidly disappear, providing the time and resources to improve maintenance and availability, further reducing costly load shedding. The paradox i s that without reforming distribution (which will eventually require privatization to sustain the reforms) private investment ingeneration may fail, andwith effective reformindistribution and a more intelligent approach to losses, private finance in generation may not be necessary (although good practice management and operations still argue for private ownership or at least management).83 6.39 Ruet(2005b) has demonstratedmost effectively that finance shouldnot be a constraint by comparing the profitability o f investing in new generation capacity (and the associated transmission) inIndia with alternatives. H e estimates that eliminating nontechnicallosses at 2002 tariff levels would give an internal rate o f return (IRR) of 339 percent, compared with new generation yielding 8.6 percent, although increasing the plant load factor (PLF) from 67 to 70 percent would reduce the amount o f new capacity required and would deliver an IRRon the total investment needed of 13.4 percent. Just investing in refurbishment and maintenance to raise the PLF alone yields 116 percent return, while investing in better transmission and distribution to reduce technical losses yields 27 percent retum. The implication is that much can be done to bridge the supply-demand gap with less finance thanjust building more generation capacity. If existing resources can be reallocated to reduce various losses, then considerable extra cash flow would be generatedto expand the system when needed, but this will require a radical change in management culture inthe SEBs. THEROLEOFENERGYTRADE 6.40 The Agreement on the South Asian Free Trade Area was signed on January 6, 2004, to enter into force on January 1, 2006. However, progress seems somewhat troubled as Bangladesh on August 8, 2005, "once again outright rejected an Indian proposal for signing Free Trade Agreement (FTA) with her, urging the counterpart to sign the proposed South Asia Free Trade Agreement (SAFTA) for boosting the regional trade and commerce."84Counterbalancing this, on August 15, 2005, the Indian Express announced plans for promoting trade between India and Pakistan via a free trade area in Kashmir. "In a radical proposal to end the current Indo-Pak conflict over water resources of J&K (Jammu and Kashmir), Burh is calling for joint development o f the power potential of the Indus waters that run through the state. Instead of separately developing hydel power in their own parts o f J&K and raising suspicions across the border, Burkiproposesjoint generation of hydel power for use in both parts o f J&K and selling the surplus to northern Pakistan and India through a common electric grid. Such an approach, according to Burki, does not involve either a renegotiation of the Indus Waters Treaty or a reduction o f water flows to either India or Palustan. It needs a mutually satisfactory reinterpretation of the treaty and negotiating a subregional agreement on hydelpower generation and di~tribution."~~ 6.41 The potential for mutually advantageousenergy trade in South Asia i s considerable. India i s short o f indigenous gas and i s actively importing expensive LNG. Bangladesh has substantial reserves estimated by the Oil and Gas Journal in 2005 at 10.6 trillion cubic feet (Tcf) and net proven reserves estimated by Petrobangla in 2004 at 15.3 Tcf, compared with a 2003 production (and consumption) of 420.2 billion cubic feet. At that rate based on Petrabangla's estimate, 83 The case for privatization has been recently summarized by Kikeri andNellis (2004). 84 http://www.bilaterals.org/recherche.php3?recherche=SAFTA. 85 http:Nwww.bilaterals.org/article.php3?id~article=2487&v~~recherche=SAFTA. 120 reserves would last 36 years, although if demand grows at a projected 6 percent p.a. the reserves would only last 19 years. In addition to proven reserves, the U.S. Geological Survey has estimated that Bangladesh contains 32.1 Tcf o f additional "undiscovered reserves," which would increase the ratio o f reserves to use to over 100 years or nearly 50 years at a growth rate o f 6 percent.86On the other hand, Bangladesh has been suffering from gas shortages, despite these abundantgas reserves. The immediate cause i s a combination of underpricing, poor collectionrates, and theft, whch have created serious financial shortfalls, and hence an inability to finance the requiredmaintenanceand investment. The deeper cause is poor managementand extensivepolitical interference in tariff setting and reform efforts. The fundamental problem is one of pervasive corruption, inturn sustainedby a lack ofpoliticalaccountability. 6.42 Bangladesh is short o f foreign exchange and government revenue, both o f which would be significantly enhancedby exporting gas to India, ideally by pipeline to the Delhi area where electricity demand i s high(incomparison to the adequate supply o f coal-fired generation near the India-Bangladesh b~rder).~' Findingan outlet for exported gas priced inforeign exchange would encourage more gas exploration and development. It i s therefore particularly perverse that both major political parties are officially committed to reserving gas for domestic use until "proven reserves will cover 50 years o f domestic demand."88 One confidence-enhancing step might be to allow transit gas from Burma to Indiavia Bangladesh, and this i s under discussion. THECASEFORA SOUTHASIA ENERGY CHARTER 6.43 Energy trade requires costly infrastructure that can be stranded without continuing cooperation, so the case for some legally binding treaty underwriting assurances on continued cooperation i s strong. The European Energy Charter signed on December 17, 1991, might form a useful model to promote such trade. The 51 signatories agreed to cooperate under a legally binding Energy Charter Treaty "designed to promote east-west industrial cooperation by providing legal safeguards in areas such as investment, transit and trade."89The proposed gas pipeline from Burma to India might encourage the signing o f an Energy Charter and if Bangladesh believed that Burmese gas could supplement domestic gas in future if local supplies proved inadequate, thenthe deadlock over Bangladesh gas exports mightbe broken. 6.44 Similarly, India would then have access to two additional sources o f gas (although both coming through Bangladesh) and would feel greater security o f supply. A similar transit proposal to deliver gas from Iran through Palustan to India would have the further advantage o f securing additional gas to meet Palustan's rapid growth in demand for power generation, although again mistrustbetween the two countries (and United States hostility to trade with Iran) have hampered progress. Again an Energy Charter guaranteeing security of transit might break the deadlock. According to the EIA, Palustan could earn US$600 million p.a. in transit fees from this US$3 billionproject.g0 6.45 Apart from the obvious benefit that India should enjoy cheaper additional supplies of gas than from LNG imports and more diverse and thus more secure sources o f supply, there are further benefits from the kinds of agreements, treaties, and contracts needed to underwrite gas imports. Most obviously (and this i s also a potential benefit from LNG imports), the gas would be pricedinforeign exchange at market levels, and would provide pressure to set fuel and electricity 86http://www.eia.doe.gov/emeu/cabs/bangla.html. 87 Itis cheaper to move gas thantransmit electricity, hence the preference for a pipeline. http://www.eia.doe.gov/emedcabs/bangla.html. 89http://europa.eu.int/scadplus/printversion/en/lvb/l27028.htm. http://www.eia.doe.gov/emeu/cabs/pa!&tan.html. 121 prices at market levels. Second, and perhaps more important in the longer run, it ought to encourage a more commercial approach to contracts in the energy industries, and that would provide the kindo f comfort that foreign private investors seek. That inturn should lower the cost o f capital, which in capital-intensive industries such as electricity, feeds straight through to consumer (and creditor) benefits. Third, gas (particularly pipeline gas) has lower C02emissions per unit of electricity generatedthan coal or oil. At some stage this will have a cashbenefitwhen emissions trading is extended more widely. Finally, creating a regional gas market with proper pricing would encourage private investors to develop indigenous gas fields, and mighteventually leadto gas-on-gas competition (as occurred inthe UnitedStates andBritain) to mutual benefit. 6.46 The other major potential source o f mutually beneficial trade relates to the more efficient exploitation of hydro resources, particularly in Nepal and Bhutan, where domestic electricity demand may be inadequate to justify large dams, but where exporting electricity to India would earn valuable foreign exchange and relieve India's power shortages in a zero-emissions way. Nepal has large untapped hydro potential, estimated by EIA at 43,000 MW (of which 244 MW hasbeen developed)." Promisingly (at least from the viewpoint o fregional energy cooperation):* inOctober 2002 "Australia's Snowy Mountains Hydro signed a MemorandumofUnderstanding for the development of the 750-MW West Seti hydroelectric dam. It is scheduled for completion in 2005 and will export power primarily to India. Renewable power sources are increasing in Nepal through rural electrification programs which aim to lessen the disparityinelectricity access between rural (30 percent) and urban (90 percent) areas. The overall quality o f Nepal's electricity infrastructure, however, is low and i s frequently a target for attack by Maoist rebels."93 In2003, Nepal's industrial sector was reported as losing US$25 million annually, 4.4 percent of its gross output value. Ifhydropower were developed for export, the annual contribution to GDP couldrise from $96 million inthe early years upto US$1.51billion by 2027. By 2010royalty earnings at 10 percent o f electricity sales would yieldUS$46 millionp.a. and rising (USAID, n.d.). 6.47 Bhutan'shydropowerpotential is estimated by EIA at 30,000 MW (of which 16,000 MW are safe and exploitable). Bhutan and India have been actively cooperating in its exploitation since the signing of the Jaldhaka agreement in 1961, and hydroelectric exports are the largest single source of foreign exchange, demonstrating the value of such regional cooperation. According to Bhutan News Online:4 "the Chukha Hydro Power Corporation has been earning more than 40 percent of the national (government) revenue o f Bhutan." Exports in 1995-6 were 1,564 GWh, and are projectedto rise to 6,400 GWhin2006 when the 1020 MW run-of-river Tala Hydroelectric Project Authority i s commissioned. "India's Tata Power Company and the Power Grid Corporation of India Ltd. have formed a partnership to construct the 1,020-MW Tala hydropower project inBhutan and a 750-mile transmission line to export power produced by the Tala project to New Delhi and surrounding areas o f India."95 91 '* http://www.dams.orgfl