Report No. 49482-PH PHILIPPINES FOSTERING MORE INCLUSIVE GROWTH May 17, 2010 Human Development Sector Unit East Asia and Pacific Region Poverty Reduction and Economic Management Unit East Asia and Pacific Region Background Papers Document of the World Bank REPUBLIC OF THE PHILIPPINES FISCAL YEAR January 1 - December 31 CURRENCY EQUIVALENTS (Exchange Rate Effective May 10, 2010) Currency Unit = Philippine Peso (PhP) PhPl.OO US$0.022 US$1.00 = PhP45.045 Vice President: James W. Adams (EAPVP) Country Director: Bert Hofman (EACPF) Sector Directors: Emmanuel Jimenez (EASHD) Vikram Nehru (EASPR) Task Team Leaders: Jehan AruJpragasam (EASHS) Ulrich Uichler (EASPR) ACRONYMS AND ABBREVIATIONS 4Ps Pantawid Pamilyang Pilipino Program NAPC National Anti-Poverty Commission ADB Asian Development Bank NCO Non-communicable Disease ARMM Autonomous Region of Muslim Mindanao NCR National Capital Region APIS Annual Poverty Indicators Survey NDCC National Disaster Coordinating Council ASEAN Association of Southeast Asian Nations NEDA National Economic and Development Authority BAS Bureau of Agricultural Statistics NGO Non-government Organization BCG Bacillus Calmette-Guerin NNS National Nutrition Survey BPO Business Process Outsourcing BTR Bureau of Treasury NDHS National Demographic and Health Survey CALABARZON Cavite, Laguna, Batangas, Rizal, and Quezon NFA National Food Authority CAR Cordillera Autonomous Region NHTSPR National Household Targeting System for Poverty Reduction CCT Conditional Cash Transfer NIA National Income Accounts CHD Centers for Health Development NRA Nominal Rates of Assistance CHED Commission on Higher Education NSCB National Statistical Coordination Board DepED Department of Education NSO National Statistics Office DHS Demographic and Health Survey OECD Organization for Economic Cooperation and Development DOH Department of Health OFWs Overseas Filipino Workers DOLE Department of Labor and Employment OOP Out-of-Pocket DPT3 Diphtheria-Pertusis-Tetanus Vaccine 3 PAP Program! Activity/ Project DSWD Department of Social Welfare and Development PESFA Private Education Student Financial Assistance EAP East Asia and the Pacific Phil health Philippine Health Insurance Corporation EPL Employment Protection Legislation PMT Proxy Means Test FAC Family Aecess Card PPP Purchasing Power Parity FAOSTAT Food and Agriculture Organization Statistics SAE Small Area Estimates FDI Foreign Direct Investment S8M School-Based Management FIES Family Income and Expenditure Survey SIP School Improvement Plan FIVIMS Food Insecurity and Vulnerability Information SNPLP Study Now Pay Later Plan Mapping System FSP Food-for-School Program SOCCSKSARGEN South Cotabato, Cotabato, Sultan Kudarat, Sarangani and General Santos City GDP Gross Domestic Product SUCs State Universities and Colleges GNI Gross National Income SY School Year GNP Gross National Product TESDA Technical Education and Skills Development Authority HH Household TFP Total Factor Productivity HNP Health, Nutrition and Population TFPG Total Factor Productivity Gro\\>1h liP Individually Paying Program TFR Total Fertility Rate ILO International Lahor Organization TNP Tindahan Natin Program IMR Infant Mortality Rate TVET Technical and Voeational Education and Training KALAHI- Kapit-Bisig Laban sa Kahirapan Comprchensive UN United Nations CIDSS and Integrated Delivery of Social Services LGU Local Government Unit UNDP United Nations Development Program LFS Labor Force Survcy USAID United States Agency for International Development M3 Domcstic Liquidity WB The World Bank MDGs Millennium Development Goals WHO World Health Organization MlMAROPA Mindoro Oriental, Mindoro Occidental, Marinduque, Romblon, and Palawan MMR Maternal Mortality Rate MOOE Maintenance and Other Operating Expenses REpUBLIC OF THE PHILIPPINES Fostering More Inclusive Growth Table of Contents BACKGROUND PAPERS PAGES CHAPTER I: INTRODUCTION AND MACROECONOMIC OVERVIEW ..................................................... 1 CHAPTER II: POVERTY IN THE PHILIPPINES .................................................................................... 17 CHAPTER II - ANNEX I: MEASURING POVERTY IN THE PHILIPPINES ............................................... 43 CHAPTER Il- ANNEX II: SIMULATING THE IMPACT OF INFLATION ON POVERTy ............................ 53 CHAPTER III: THE SECTORAL AND REGIONAL PATTERNS OF GROWTH IN THE PHILIPPINES ........... 57 CHAPTER IV: REVISITING AGRICULTURAL GROWTH AS A PATHWAY OUT OF RURAL POVERTy .... 69 CHAPTER V: ENHANCING THE CONTRIBUTION OF MANUFACTURING TO GROWTH AND EMPLOyMENT ................................................................................................................................ 82 CHAPTER V - ANNEX: METHODOLOGY FOR CALCULATING POTENTIAL DISTORTIONS THAT AFFECT THE RELATIVE PRICE OF LABOR ...................................................................................................... 99 CHAPTER VI: POVERTY AND THE LABOR MARKET ...................................................................... 102 CHAPTER VII: HEALTH, NUTRITION, AND POPULATION ............................................................... 122 CHAPTER VIII: EDUCATION AND INCLUSIVE GROWTH ................................................................ 141 CHAPTER IX: SOCIAL PROTECTION AND INCLUSIVE GROWTH .................................................... 159 ACKNOWLEDGNIENTS This report was prepared by a team led by Jehan Arulpragasam (EASHD) and Ulrich Lachler (EASPR) with the much appreciated contribution of Minna Hahn Tong (Consultant) in drafting and editing. The team also comprised Michael Alba (Consultant), Eduardo Banzon, Fabrizio Bresciani, Karl Chua, Lynnette Dela Cruz Perez, Tomoki Fujii (Consultant), Nobuhiko Fuwa (Consultant), Swati Ghosh, Timothy Johnston, Aart Kraay, Xubei Luo, Rosechin Olfindo, Laura Pabon, Claudio Raddatz, Silvia Redaelli, Iamele P. Rigolini, Rashiel Velarde and Soonhwa Yi. Able team assistance was provided by Kristine San Juan-Ante and Melanie Esteban. The peer reviewers are Elizabeth King, Humberto Lopez, and Pierella Paci. The principal authors responsible for each chapter in the report are as follows: · Chapter I Karl Chua, Ulrich Lachler · Chapter II - Rashiel Velarde, Tomoki Fujii, Ulrich Lachler · Chapter III Claudio Raddatz, Iamele P. Rigolini, Karl Chua, Ulrich Lachler · Chapter IV Fabrizio Bresciani, Nobuhiko Fuwa · Chapter V Aart Kraay, Swati Ghosh, Soonhwa Yi, Ulrich Lachler · Chapter VI Silvia Redaelli, Xubei Luo, Laura Pabon · Chapter VI Timothy Johnston, Eduardo Banzon · Chapter VIn Lynnette Dela Cruz Perez, Michael Alba · Chapter IX Jehan Arulpragasam, Rosechin Olfindo, Rashiel Velarde The preparation team would like to thank the Philippine authorities for their support in providing logistical assistance and for facilitating timely access to data and information. A draft report was submitted to the government for review in early February 2010, and the comments subsequently received from the Department of Social Welfare and Development and from the National Economic and Development Agency have been incorporated into the final version. CHAPTER I INTRODUCTION AND MACROECONOMIC OVERVIEW 1.1 After two decades of stagnation, the Philippine economy began to grow again in the mid-2000s. During 1980-99, the gross domestic product (GDP) of the Philippines exhibited no growth in per capita terms, in marked contrast to the East Asia and Pacific region as a whole which recorded an annual average per capita growth of 6.5 percent during this period. This period of stagnation came to an end in the 2000s, with per capita growth accelerating to an annual average rate of 3.5 percent between 2003 and 2006 and reaching a record high of 5.4 percent in 2007. Though still modest by East Asian standards, this resurgence of growth represented a major improvement by historical standards. The Philippines had not seen such levels of growth since the 1970s. Even before the global economic downturn of 2008 and 2009 set in, however, there were concerns that such growth would be difficult to sustain. These concerns revolved around the economy's dependence on an exceptionally favorable external environment which was fueling economic booms worldwide and the prevalence of various long-term growth constraints, particularly the lack of adequate infrastructure coupled with low overall investment levels, the persistence of high unemployment and emigration rates that reflected a lack of domestic opportunities, and governance weaknesses that undermined the investment climate. An inability to sustain the tax revenue effort and a high public debt ratio, despite the fiscal reforms introduced in the first half of this decade, also contributed to investor uncertainty. 1.2 This improved economic performance has not translated into greater progress in poverty reduction. The rate of poverty reduction was slower in the Philippines than elsewhere in East Asia during the 1980s and 1990s, in part due to the much slower rate of growth in the Philippines over this period. 1 A surprising development, however, is that as economic growth accelerated, the share of the population living below the national poverty line increased from 30.0 percent in 2003 to 32.9 percent in 2006, according to official poverty estimates? The poverty level has now returned to where it was at the end of the 1990s. This observation is corroborated by perception surveys (e.g., Social Weather Stations3 reports) which indicate that poverty and hunger have increased in recent years, as well as by the slow progress made toward some key Millennium Development Goal (MDG) targets, such as universal access to primary education and improved maternal and reproductive health. I The Philippines also started out with lower poverty levels and thus had already reaped the relatively easy gains to be had at the beginning of the poverty reduction process. 2 Other poverty estimates based on different calculation methods yield similar increases in poverty incidence over this period. For example, World Bank calculations yield an increase from 31.l percent to 32.9 percent over this period, while Balisacan (2008) calculated an increase from 26.0 percent to 28.1 percent using consumption-based measures. See also the papers presented at the National Conference on Imperatives for Poverty Reduction Amidst and Beyond the Global Economic Crisis, held in Manila on March 31, 2009. 3 The Social Weather Stations (SWS) is a private non-stock, nonprofit social research institution that collects quarterly perception surveys, including self-rated poverty and hunger (www.sws.org.ph). 1 1.3 The lack of progress in reducing poverty despite the acceleration in economic growth indicates that growth has not been sufficiently inclusive. The apparent combination of economic expansion and rising poverty flies in the face of the strong empirical regularity that has been observed worldwide between poverty reduction and positive economic growth. 4 However, a number of factors may have weakened the link between poverty reduction and growth in the Philippines, including unbalanced sectoral growth, an unequal distribution of human capital, and changes in relative asset and output prices that have hurt the poor. All of these factors involve a worsening distribution of income or consumption, in an environment that had already been exhibiting high degrees of inequality by regional standards. 5 1.4 This report analyzes the aspects of the growth process that may have contributed to a less egalitarian structure of income and consumption and examines various policy options for making growth more inclusive. The strategic framework for this report views inclusive growth as depending on two critical components: improved income opportunities and greater market participation, especially for the poor. This means that policies to foster more inclusive growth must focus on both strengthening the fundamentals to accelerate economic growth, while raising labor mobility-thereby facilitating the movement of workers to better employment opportunities-and building human capital-thereby enabling those workers to seize the improved job opportunities available to them. 1.5 The remainder of this chapter provides a brief overview of macroeconomic developments in the Philippines over the last decade. As context for the report, the discussion points to several important areas of vulnerability that should be taken into account in designing policies that aim to foster economic growth or reduce inequality.6 A. Recent Economic Developments Economic growth 1.6 Philippine GDP growth had been rising since 2001 but then slowed again in 2008. GDP growth reached an average of 5.4 percent per annum during 2003-06 and peaked at 7.1 percent in 2007, then it slowed to 3.8 percent in 2008 as the food and fuel 4 An important data problem should be recognized in this context: whereas the National Income Accounts (NIA) data indicate rapid growth in aggregate real income, GDP, and consumption since 2000, the Family Income and Expenditures Survey (FIES) data indicate declines in real per capita income over that same period. While the latter fmding casts doubt on the magnitude of GDP growth experienced since 2000, circumstantial evidence indicates that per capita GDP growth has been positive, which means that the observed rise in poverty must have been associated with a worsening distribution of income or consumption. These issues are discussed further in Chapter II. 5 The World Bank's Development Data Platform identifies the Philippines as having the highest Gini index among the East Asian middle-income countries for the period 2004-06. This database shows the Philippines with a Gini index of 0.44, followed by China (0.42), Indonesia (0.39), Malaysia (0.38), Thailand (0.42), and Vietnam (0.38). 6 For a more detailed discussion of recent economic developments, see the World Bank's Philippines Development Report 2009 and the World Bank's Philippines Quarterly Updates. 2 crisis and global financial crisis began to take their toll. On the production side, the country's growth performance was driven by the services sector, which comprises more than half of GDP and employs about half of the labor force (Figure 1.2). Within the services sector, several sub-sectors performed remarkably well: (i) transportation, communication, and storage; (ii) private services, which include the fast growing business process outsourcing (BPO) industry; and (iii) financial services. Growth in industry was driven by manufacturing and the utilities sub-sectors. Agriculture was the least dynamic sector, contributing relatively little to the acceleration of growth after 2001. Figure 1.1: Per Capita GDP Growth, Figure 1.2: Contribution to GDP Growth, 1997- 1997-2007 . 2007 12.0 ~ 8,0 ~ .t::. 10.0 rfl ;; '6.0 'i 8.0 i... e CI 6.0 CI 4.0 :s 4.0 Do. c 2.0 CI ~ g. (.) .. 2.0 0.0 -2.0 - CI 0 c .2 0.0 'S -2.0 ~ ~ -4.0 J 'E: e 0 -4.0 1997 1999 2001 2003 2005 2007 (.) 1997 1999 2001 2003 2005 2007 Year Year III Philippines .EastAsia IIIAgriculture · Industry IJService Source: World Bank Source: National Statistical Coordination Board (NSCB) Figure 1.3: Contribution to GDE Growth, 1997- Figure 1.4: Investment-to-GDP and GDP 2007 Growth 0, 1997-2007 i:i 10.0 25.0 7.0 ~ .t::. ~ 8.0 5.0 6.0 20.0 CI w 4.0 3.0 C CI 2.0 .s c 15.0 1.0 o 0.0 ~ -2.0 ~ 'E: 10.0 e o -4.0 1997 1999 2001 2003 2005 2007 (.) 1997 1999 2001 2003 2005 2007 Year Year _Investment (percent of GOP) (Ihs) IIIConsumption · Government IJNet export --GOP growth (rhs) Source: NSCB Source: NSCB 3 1.7 The liberalization of several service sector industries, most notably telecommunications, was one of the main catalysts of growth in the last decade. Between 1996 and 2006, the communications industry-in particular, mobile communication and internet services-grew rapidly, contributing to the equally strong growth of the BPO industry. Currently, the BPO industry is estimated to generate about 4 percent ofGDP. Growth in the financial sector, which averaged 12 percent between 2004 and 2007, can be traced mainly to hefty earnings from investments in government paper, which replaced private lending as the main source of bank income. From 2001 to 2007, private credit as a share of GDP fell from around 40 percent of GDP to less than 30 percent ofGDP. Bank earnings were also boosted by the growing inflow of remittances. 1.8 On the expenditure side, economic growth since 2001 was driven mainly by private consumption, in part supported by fast-growing remittances (Figure 1.3). Both private investment and government spending remained largely subdued over much of this period by uncertainties associated with recurrent political crises and the deteriorating fiscal position after 1999. With the fiscal consolidation of 2005-2006, government spending and private investment began to expand again, providing a modest boost to growth. 1.9 The low level of investment has been a major constraint to sustained growth in the last decade. The share of investment to GDP declined from 25 percent of GDP in 1997 to less than 15 percent in 2005 (Figure 1.4). This decline in total capital formation can be traced to a sharp fall in private investment. In particular, purchases of durable equipment fell by about 50 percent in real terms. Though slightly less severe, the 25 percent decline in public infrastructure investment has also been an impediment to private sector investment. 1.10 From a simple growth accounting perspective, the growth acceleration of the 2000s has largely been the result of a pickup in total factor productivity (TFP) growth. 7 Growth in output per worker rose to an estimated 2.3 percent per year in 2000- 07. Given the weakness in fixed Table 1.1: Philippine Growth Accounting, 1961-2007 investment, the accumulation of physical capital made virtually Contribution to Growth of: Growth in no contribution to growth, while Output per Physical Human TFP TFP Cycl. the accumulation of human Worker Capital Capital Adjusted* capital contributed a steady but 1961-06 1.1 0.8 0.4 -0.1 -0.1 modest 0.3 percentage points to 1961-79 2.4 1.2 0.4 0.7 0.5 annual growth. TFP growth, on 1980-89 -0.9 0.8 0.4 -2.0 -2.2 the other hand, jumped to 1.9 1990-99 0.1 0.4 0.4 -0.7 0.1 2000-07 2.3 0.1 0.3 1.9 1.1 percent per year-stronger even * Cyclically adjusted. than in the rapid-growth decades Souree: Bosworth and Collins (2003) through 2000; World Bank staff of the 1960s and 1970s (Table estimates after 2000. 7 TFP growth estimates are residuals that combine many types of economic effects-technological progress, for example, or greater efficiency due to better allocation of factors of production within sectors, movement of factors between sectors, improvements in the quality of factors of production, or changes in the utilization rate of factors of production. 4 1.1). Corrected for changes in capacity utilization, which has been increasing in the same period, the underlying rate ofTFP growth was a more modest 1.1 percent per year. Labor market 1.11 Labor force participation has remained fairly stable over the past decade, while the unemployment rate has shown little improvement. In the last ten years, the labor force participation rate has remained around 64 to 67 percent of the working population. Meanwhile, the labor force has been growing at slightly above the Figure 1.5: Unemployment Rates, 1997-2007 population growth rate of about 2.3 14.0 percent. Under the old definition of the 12.0 labor force and unemployment, 8 ~ 10.0 unemployment in 1997 was 8.7 e.... 1: 8.0 percent. This worsened in 1998 during ~ 6.0 the height of the East Asian financial (I) a. 4.0 and EI Nino crises, which saw the unemployment rate jump to 10.2 2.0 percent before improving moderately to 9.8 percent in the following year. 1997 1999 2001 2003 2005 2007 Despite higher gro",th in succeeding Year years, the unemployment rate hardly Source: National Statistics Office (NSO) fell. Rather, under the new definition, it only declined from 6.6 to 6.5 percent between 2001 and 2007,9 while the underemployment rate increased from 17 to 20 percent, leading to the characterization of this period as one of "jobless growth" (Figure 1.5). 8 The old defmition counts as part of the labor force unemployed workers who, believing that no work is available, give up looking for work. The old definition was used until 2006. 9 Under the new definition of the labor force and unemployment, the average unemployment rate is about 8 percent. 5 1.12 The composition of employment has not changed much in the last decade. Wage and salaried workers who are mostly employed in the formal sector account for about half of employed workers. Many Figure 1.6: Composition of Employment, informal workers are either working 1997-2007 on their own account (i.e. self- employed) or are unpaid. Workers on - 100% ] own account comprise 37 percent of ~ " 80% employed workers, while unpaid workers comprise the remaining 13 -w~ 60% j: percent of employed workers (Figure 40% , 1.6). The share of workers in the '0 agricultural and industrial sectors C 20% 1 continued to decline in the last decade, ~ 0% +-~~~~~~~~~~~~ CI) a... 1997 1999 2001 2003 2005 2007 with the balance being absorbed by the services sector which now comprises Year about 50 percent of total employment IIWage and salary _Own account CUnpaid family worker compared to 44 percent in 1997. Source: NSO 1.13 Since 2000, the number of Filipino overseas workers has increased rapidly. Between 2006 and 2008, over one million Filipinos were deployed annually on a contractual basis. In 2007, the government estimated that around 5 million Filipinos were working overseas under fixed-term contracts, in addition to about 4 million permanent overseas workers and migrants. The number of Filipino overseas workers in high-skilled jobs has also grown in the last decade: over 40 percent of Filipino overseas workers are in high-skilled jobs such as health care, accounting, and information technology, compared to less than 30 percent a decade ago. Geographic and occupational diversification is also characteristic of recent waves of migration and offshore employment. From the traditional construction workers going to Middle East destinations in the 19705 and 19805, many workers are now venturing into Europe, Canada, and other countries in East Asia as highly skilled workers. 6 The external sector 1.14 The Philippine external Figure 1.7: Current Account Balance and account has undergone a major Remittances through Banks, 1997-2007 transformation in the last ten years. From recurring current account 15.0 deficits, the Philippines gradually 10.0 improved its current account balance beginning in 1999. It achieved a surplus in 2003 and sustained it for the 0.0 +..II,..,.................,..,.,rI&-rJll-,..-..,--.,...o.,.- next five years, with the current -5.0 account surplus reaching a historic -10.0 high of 4.9 percent of GDP in 2007 1997 1999 2001 2003 2005 2007 (Figure 1.7). Year 1.15 Driving the fast-growing Ell Current account balance (% of GOP) · Remittances (billions of dollars) current account balance are the large inflows of remittances. As Source: Bangko Scntral ng Pilipinas (BSP) shown in Figure 1.7, remittances jumped from US$7.4 billion in 1998 to almost US$16.4 billion in 2008, representing an annual average increase of 11 percent. About 42 percent of total remittances come from the United States and about 25 percent of remittances come from the Middle East. Remittances from the Middle East doubled in the last five years and benefited from economic expansion triggered by higher oil prices. 1.16 A second, smaller driver of the improving current account balance has been the surge in electronics exports over the past decade. Comprising 59 percent of total Philippine exports in 2009, semi-conductors and other electronic exports have replaced traditional Philippine exports such as garments, minerals, and agricultural products as the main export products. With equally strong growth in imports driven by electronic parts and petroleum products, the trade balance has been exhibiting a fairly stable deficit of 7-8 percent ofGDP. 7 1.17 The capital and financial accounts have been less stellar than the current account. Foreign direct investment (FDI)-a major source of growth in the East Asia region-has been patchy in the Figure 1.8: Direct and Portfolio Investments, Philippines. Between 1999 and 2008, 1997-2007 the Philippines on average received 6.0 US$l billion a year in direct ..,. 4.0 investments, compared to Indonesia's CIJ US$3 billion, Vietnam's US$4 billion, ::::> c: 2.0 and Thailand's US$8 billion. A weak .2 investment climate, characterized by 0.0 iii macroeconomic instability and poor -2.0 governance for much of the last -4.0 decade, is seen as the mam 1997 1999 2001 2003 2005 2007 impediment to higher investment. Year Portfolio investment has also been II Direct investment · Portfolio investment weak and for the most part volatile, averaging about US$l.2 billion over Source: BSP the same period (Figure 1.8). 1.18 The strengthened balance of payments position enabled the Philippines to build up its international reserves. From US$8.8 billion (2 months of imports) at the end of 1997, gross international reserves gradually improved to reach US$18.5 billion (3.8 months of imports) in 2005 and jumped to almost US$38 billion (5.7 months of imports) by the end of2008. These increases have placed the Philippines in a much better position to withstand shocks from the global financial turmoil and higher commodity prices. Monetary and exchange rate policy 1.19 A maj or shift in policy regime occurred in the late 1980s when the international capital account was progressively opened. The shift culminated in the full convertibility of the peso in 1992 and the opening of the new central bank, the Bangko Sentral ng Pilipinas (BSP), in 1993. The BSP placed emphasis on inflation- targeting and shifted to a managed float of the peso. In practice, however, there was some ambivalence with inflation-targeting, so a mixture of inflation and exchange rate targeting was pursued. These reforms led to a tempering of inflation, which stayed below double digits in the last fifteen years and averaged 5.8 percent (Figure 1.9). 1.20 The exchange rate was fairly steady until the East Asian financial crisis of 1997-98. Between 1993 and 1997, the exchange rate moved within the 25-27 peso band. When the East Asian financial crisis struck the Philippines, the BSP defended the peso by drastically increasing interest rates (Figure 1.10), but the defense proved futile as the BSP was forced to float the peso. Within months, the peso fell by 50 percent to about PhP 40 per US$I. Subsequent macroeconomic and political developments led to further 8 depreciation of the peso to about Php55 per US$l in the mid-2000s. It appreciated in more recent years (ranging within Php44.5 and Php47.6 per US$l), supported by remittances inflows and a weakening dollar (Figure 1.9). Figure 1.9: Inflation and Exchange Rates, Figure 1.10: Central Bank Policy Rates, 1997-2007 1997-2009 10,0% 60,0 40.0 , - - - - - - - - - - - - - , -. ;:;e 8,0% 50.0 ~ 30.0 -. ;:;e ... S - '" 6.0% -; cu a:: 40.0 rn a:: :::I .. ..II: C I'll J:I 20.0 - c: 4.0% .z:. 10.0 a.. CD ~ .E 20% 30.0 .E o!I :>. 0.0 ..... m ('I) l!) ..... .S:! '0 m C: m I C 9 c: 9 c: 0 C: 0I c a.. ..., til ..., til ..., til ..., til ..., til ..., CI! 1997 1999 2001 2003 2005 2007 --RPrate --RRPRate --Inflation rate (Ihs) - - Exchange rate (rhs) --Interbank Call Loan Rate Source: BSP Source: BSP 1.21 In 2000, the BSP refined its monetary policy stance by progressively moving from a range target to a point target to achieve certain inflation outcomes. As a result, inflation declined between 2000 and 2003, averaging 4.3 percent. With the escalation in fuel prices beginning in 2004 and the rise in food prices in 2008, inflation jumped to about 8 percent. Prudent monetary policy has kept the inflation rate below double digits, averaging only 3.2 percent in 2009. Fiscal policy 1.22 The country's fiscal situation weakened with the onset of the East Asian financial crisis in 1997-98. In addition to weak corporate earnings in the years following the crisis, tax policy changes granting very generous but often redundant fiscal incentives and the shift to non-indexed specific taxation for excise products gradually eroded the tax base. Weaker tax administration was also to blame for the fall in tax effort. From its peak of 17 percent of GDP in 1997, the tax effort gradually fell to 12.4 percent of GDP in 2004, the lowest since 1986. 1.23 The fall in tax effort translated into rising deficits and debt and a spending contraction (Figures 1.11 to 1.14). By 2003, the Philippines was on the brink of a fiscal crisis. The consolidated public sector deficits reached over 5 percent of GDP and the public debt rose above 100 percent. The debt of the national government was nearly five times its revenues, and by 2004, national government interest payments reached 5.4 percent of GDP. The growth of the public debt over this period reflects not only past deficits but also the calling of numerous contingent claims on the government and bailouts of government-owned corporations. The rising public debt burden led to a number of adverse consequences, including fewer resources for human development and 9 infrastructure, reduced investor confidence and private investment, and an increased vulnerability to shocks. Figure 1.11: Revenue, 1997-2007 Figure 1.12: Expenditure, 1997-2007 20.0 25.0 c:- i§' 18.0 g 20.0 ~ '0 '0 16.0 ~ 15.0 ~ ~ 14.0 ~ 10.0 ;:: "'0 ;:: ~ ~ 12.0 ! 5.0 10.0 1997 1999 2001 2003 2005 2007 1997 1999 2001 2003 2005 2007 Year Year IIRevenue effort .Tax effort · Interest payment IIPrimary expenditure Source: Bureau of Treasury (BTR) Source: BTR Figure 1.13: Deficits, 1997-2007 Figure 1.14: Debt, 1997-2007 2.0 100.0 c:- o 0.0 c:- ~ o ~ 80.0 '0 '0 ;:,!l rL J!! 'i) \I: -2.0 -4.0 - ;:,!l ~ .c 0 G> 60.0 IJ G> 0 -6.0 40.0 J 1997 1999 2001 2003 2005 2007 1997 1999 2001 2003 2005 2007 Year Year liNG deficit .Public sector deficit liNG debt · Non-financial public sector debt Source: BTR Source: BTR 1.24 Actions to strengthen tax administration were initiated in the revenue agencies between 2003 and 2004, which slowed but did not stop the decline in tax effort. In 2004, the national government deficit declined as civil service wages were not adjusted for inflation and transfers to local government units were limited to 2003 levels, reflecting the re-enactment of the 2003 budget. However, power sector deficits increased sharply from 0.2 percent of GDP in 2001 to almost 2 percent in 2004, adding further pressure on the consolidated public sector deficit. 10 1.25 After the 2004 general elections, the government began to address the fiscal crisis seriously. The president declared a state of fiscal crisis in August 2004 and asked Congress to enact a number of tax policy reforms. 1O The reforms significantly reduced fiscal and macro vulnerabilities and brought back investments. Primarily as a result of policy reforms, the tax effort improved significantly from 13 percent of GDP in 2005 to 14.3 percent of GDP in 2006. Consequently, the deficit fell to 2 percent of GDP from above 3 percent of GDP a year earlier, and borrowing costs were halved from close to 500 basis points in 2004. Higher tax revenues permitted the government to increase real spending in 2007 for the first time in seven years. The improving fiscal balance, aided by the appreciation of the peso, helped trim down public debt to 56 percent of GDP by the end of2007. 1.26 However, the improvement on the fiscal front was short-lived. In 2007, the tax effort fell to 14 percent of GDP and remained at 14.1 percent of GDP in 2008. With higher spending in 2008 to mitigate the impact of the food crisis, the national government deficit also deteriorated, while the national government debt-to-GDP started to increase. B. Macroeconomic Prospects 1.27 Economic growth in the Philippines has slowed since 2008. Growth slowed to 3.8 percent in 2008, as higher oil and food prices and the onset of the worst financial crisis in decades reduced real incomes and slowed the growth of private consumption, investments, and exports. The economy slowed further to 0.4 percent in the first quarter of 2009 (Figure 1.15). Inflation, which had been falling over the past few years and averaged only 2.8 percent in 2007, rose sharply in the first half of 2008, peaking at 12.5 percent in August 2008. It then receded almost as quickly to below 4 percent by May 2009 (Figure 1.16). 1.28 Although macroeconomic prospects for 2009 and 2010 have become less favorable, the economy is in a stronger position than it was at the time of the East Asian financial crisis. The financial and fiscal reforms over the last decade have helped improve economic fundamentals and investor confidence. Nevertheless, the Philippines remains vulnerable to the slowdown in the global economy. Growth in 2009 is expected to decline to around 1.4 percent, before rising moderately again in 2010 (Table 1.2). This is due to weaker domestic demand on account of lower real incomes, rising unemployment and underemployment, slowing remittances, and falling exports of key products, partiCUlarly electronics. Moreover, these growth projections exhibit downside risk, especially if the recovery is drawn out. In addition, there is a danger that ari early 10 These refonns included: a law increasing the excise tax on cigarettes, tobacco, and alcohol; the Attrition Act for revenue agencies; and the Comprehensive Tax Refonn Package of 2005. The tax refonn package included the removal of exemptions from the Value Added Tax (VAT) of oil, power, medical, and legal services; an increase in corporate income tax from 32 percent to 35 percent, reverting to 30 percent in 2009; and authority for the president to increase the V AT rate from 10 to 12 percent in January 2006 upon satisfaction of certain conditions prescribed by law. Certain excise, franchise, and common carrier taxes were repealed in favor of the VAT. 11 start of the election campaign period could discourage domestic investment and the inflow ofFDI, as investors shift to a "wait-and-see mode." 11 Figure 1.1 5: GDP and GNP Growth, Figure 1.16: Contribution to Overall Inflation, 2006-2009 2007-2009 % 4,0 15 " ,60. 3,0 11 J ~ ilHUilliHI 2.0 40. 1,0 -1" ' 0,0 -5 - 20. -9 -1,0 -13 -2.0 0. -17 -3.0 -21 ~ ~ ~ § ~, ~ -25 -20. a 0 a cJ-, 0 a 8 i~i~i~i~ii~~i Source: NSCB Source: BSP 1.29 Over the medium to long term, economic growth is projected to recover to between 4 and 5 percent, but its sustainability will depend on further progress in structural reforms. In particular, turning around the low investment rates and sustaining high productivity are essential for achieving this target. Domestic investment as a share of GDP is expected to fall in 2009 and 2010 due to tighter global credit, financing conditions, and domestic concerns surrounding the approaching election period, followed by a gradual recovery in response to future improvements in governance and the investment climate. 1.30 The current account is projected to remain in surplus over the near term. Under the current assumption that the global economy may take up to two years to recover, exports are expected to contract significantly before improving, while the import bill would also contract given the high content of electronics parts and falling commodity prices. Slower exports and imports could lead to a lower trade deficit in the near term. Exports of manufactured goods, dominated by highly cyclical electronics, remain vulnerable to global demand conditions. While the growth of remittances is expected to slow significantly, they are projected to remain fairly resilient because a sizable share of the migrants is employed in sectors that are less vulnerable to cyclical downturns, such as health care. 1.31 The global uncertainties and risks are likely to pose a threat for foreign investors. Net inflows of direct investments are expected to remain modest in 2009 and 2010. Over the medium term, attracting significant FDI inflows will depend on further II Another source of downside risk is the advent of natural catastrophes, such as typhoons Ondoy and Pepeng, which struck the Philippines in September 2009. These created considerable damage and may slow down or raise overall economic growth further in 2009 and 2010 by a few basis points, depending on whether the supply-side disruptions outweigh the reconstruction-driven stimulus on the demand side. 12 improvements in the investment climate. Consistent with global financial volatility, net portfolio investments are also expected to be minimaL The net outflow in 2008 could be reversed if the interest rate differential rises in the coming months and if global prospects become less pessimistic. Overall, however, the combined strength of the current account and capital and financial accounts would still contribute to further reserve accumulation in the medium term. Table 1.2: Macroeconomic Prospects (as of September 2009) (as percent of GDP, unless stated otherwise) Actual Proi!:;ctions 2006 2007 2008 TID 2009 II 2009 2010 2011 2012 2013 Growth, inflation and uncmplorment (percent) 51 Gross domestic product 5.3 7.1 3.8 1.0 1.4 3.1 4.5 61 InfIathn (pernd average); 2000 base year 6.2 2.8 9.3 3.4 2.9 4.5 4.0 Savings and investment 51 Gross nathnal saving<; 19.1 20.3 17.8 20.9 17.0 16.6 16.3 51 Gross domestic investment 14.5 15.4 15.2 15.7 13.6 13.8 14.0 Public sector 7/ Nathnal govermnent balance (GFS basis) II -1.4 -1.7 -1.5 -2.9 -3.8 -3.1 -2.7 -2.6 -2.4 71 Nathnal govermnent balance ZI -1.1 -0.2 -0.9 -2.7 -3.2 -2.9 -2.4 -2.3 -2.1 Total revenue 21 7/ 16.2 17.1 16.2 9.5 15.0 15.2 15.3 15.2 15.0 Tax revenue 21 14.3 14.0 14.1 8.5 12.8 13.3 13.6 13.6 13.6 7/ 71 Total spending 21 17.3 17.3 17.1 12.3 18.2 18.0 17.7 17.4 17.1 81 NathnaIgovermnent debt 63.9 55.8 56.9 54.5 59.6 59.4 59.2 59.1 58.5 Consolidated non- financial public sector balance 0.2 0.7 0.4 -2.1 -2.0 -0.9 0.1 0.5 Consolidated non-finaucial public sector debt 73.9 61.1 60.0 62.0 61.8 61.0 60.5 60.1 Balance of payments 71 Merchandise exports 15.6 6.4 -2.6 -30.3 ~19.0 3.0 5.1 5.5 5.8 71 Merchandis~...~orts 10.9 8.7 5.0 -30.8 -18.0 6.5 7.0 7.0 7.0 71 Remittances (expressed in US$) 19.4 13.2 13.7 3.7 4.0 5.5 8.0 9.0 9.5 Current account balance (percent of amrual GDP) 4.5 4.9 2.5 5.2 3.4 2.8 FOI (billio1l'i of dollars) 2.8 -0.6 1.3 0.8 0.4 0.5 PortlOw Investment (billions ofdollars) 3.0 4.6 -2.6 0.4 0.1 0.5 International reserves 61 Gross official reserves 3/ 23.0 33.8 42.5 42.3 44.2 45.1 46.7 48.0 7/ ('rross official reserves (roontbs ofimports) 4.2 7.8 7.7 7.6 51.29 45.8 42.5 41.0 35.7 154.0 206 340.0 350.0 350.0 61 II Excludes privatization receipts (these are treated as financing items, in accordance with GFSM) and includes CB- BOL restructuring revenues and expenditures 21 Government definition 31 Includes gold 4/ World Bank definition 5/ Year-to-date (YTD) for 2009. is as of June 6/ YTO for 2009 is as of September 7/ YTD for 2009 is as of August 81 YTD for 2009 is as of July 13 Source: World Bank, Philippine Quarterly Update, November 2009 1.32 The public sector's fiscal position is expected to remain manageable, but fiscal risks could materialize. A deficit of about 3.8 percent of GDP is projected for 2009, in part reflecting the implementation of the government's Economic Resiliency Plan in response to the global crisis. Maintaining tax collections will become more of a challenge this year due to several policy measures already taken and lower imports. 12 In the absence of new measures, tax revenues as a share of GDP are projected to decline well below the ratio achieved in 2008. Under these circumstances, reforms to improve tax policy and administration will be critical for keeping the deficit within sustainable levels. 1.33 Public debts are broadly sustainable, and the public sector debt-to-GDP ratio is projected to fall over the medium term, despite the global recession. The consolidated non-financial public sector debt is projected to increase to 62 percent of GDP in 2009 before gradually falling again in 2010. C. Medium-Term Macroeconomic Challenges 1.34 In the medium term, the Philippines faces two key macroeconomic challenges that need to be overcome before sustained and inclusive growth can be achieved. Specifically, the country needs to (i) improve public finances and (ii) strengthen the investment climate. 1.35 Improving public finances is crucial for ensuring macroeconomic stability and generating enough fiscal space to permit greater spending on public infrastructure and human development, both of which have been neglected over the past decade. Halting the decline in the tax ratio and improving the equity of taxes are especially necessary in this context. Important tax policy measures include the rationalization of excise taxes and fiscal incentives. In addition, there is scope to consider raising excise taxes on petroleum products, which have not been adjusted since 1996. In the area of tax administration, there is room for increasing revenues by expanding the tax base and improving compliance and collection, especially from the country's largest taxpayers. 13 Strengthening governance in the public sector and reducing corruption are also important for boosting investor confidence and improving service delivery. 1.36 Increasing the total investment rate is another critical step for achieving sustained growth in the Philippines. The strengthening of public finances, as suggested above, has already proven effective in boosting investor sentiment in the past. Other important steps for improving the investment climate include the acceleration of ongoing governance reforms and the removal of unnecessary constraints on doing business, particularly the provision of better access to credit to small and medium-size enterprises. Public investment in infrastructure, guided by increased transparency and accountability, 12 These policy measures included the 5 percentage point reduction in the corporate income tax rate, a full year impact of the personal income tax threshold increase, and tax exemptions arising from personal equity and retirement accounts. 13 Specific policy options for raising public revenues are discussed in World Bank (2010), Philippines Discussion Note on improving tax policy and administration. 14 could be ramped up and used as a catalyst for attracting more private sector investment. Strengthening the investment climate would help buoy growth and increase employment in the short term, enabling the country to better weather the impact of the global recession and providing the foundation needed to sustain higher levels of growth over the medium term and accelerate poverty reduction. Box 1.1 below provides an example of how a better investment climate has boosted growth in the past. 1.37 The current economic downturn has re-focused attention on the need to restore growth and economic recovery, but ensuring broad-based growth is just as critical. While meeting the two macroeconomic challenges described above should help significantly in restoring growth, it is also important to ensure that growth is inclusive. As emphasized in subsequent chapters, broad-based participation in the growth process remains critically important to make sure that the benefits of growth are shared by the poor and most vulnerable and that the faster growth becomes reflected in lower poverty rates. It also provides a stronger platform for future growth and competitiveness. Box 1.1: The Rise of the BPO Industry The business process outsourcing (BPO) industry is a prime example of a dynamic industry which has benefited tremendously from improvements in the investment climate and liberalization of the telecommunications industry in the mid 1990s. Prior to this, the industry was virtually a monopoly, with one major player controlling 90 percent of the market and owning the country's only transmission line. Telephone lines were hard to obtain, with waiting times that reached up to 10 years, and service quality was low. At the same time, macroeconomic instability and unreliable power supply were hindrances in attracting foreign direct investment. This changed in 1993, when the Ramos Administration liberalized key industries such as water, oil, and telecommunications and embarked on a plan to restore continuous power supply and improve the fiscal position of the country. As a result, investments flowed in at higher rates and spawned new non-traditional and non-captured industries such as BPOs, which benefited from the much lower telephone and internet rates. The BPO industry has grown at impressive rates in the last decade, averaging 30 percent annually, and now contributes 4 percent of gross value-added and employs nearly 400,000 people. The industry has also contributed to the rapid growth of the property sector and to some extent to retail trade and private services. The BPO industry is led by the contact centers, which account for 70 percent of total BPO revenues and 60 percent of total employment. In recent years, BPO growth has shifted from contact centers to more knowledge-based businesses such as back office systems development and accounting. Growth in these non-voice BPOs ranged from 40 to over 100 percent. The growth potential of the industry is tremendous. Recent supply bottlenecks (mainly the availability of qualified BPO workers) and the global economic slowdown are slowing growth in the industry. Nevertheless, as long as the overall investment climate improves further, input costs for communications and labor services remain competitive, and an abundant supply of qualified workers is available, the industry is expected to grow and contribute significantly to the economy over the medium term. The success of the BPO industry highlights the large dividends that can be earned from improving the . investment climate. Going forward, non-traditional and non-captured industries with very strong potential for growth, such as tourism and the outsourcing of manufacturing (especially electronics), could grow significantly under appropriate infrastructure and governance reforms. 15 References Balisacan, Arsenio M. (2008). "Poverty Reduction: What We Know and Don't," University of the Philippines Centennial Lecture Series. Bangko Sentral ng Pilipinas (BSP), http://www.bsp.gov.phistatistics/overview.asp. varIOUS years. Bosworth, Barry and Susan Collins (2003). "The Empirics of Growth: An Update." Washington, D.C.: The Brookings Institution. Bureau of the Treasury, http://www.treasurv.gov.phistatdata.html. various years. National Statistical Office (NSO), http://www.census.gov.phi. various years. Philippine National Statistical Coordination Board (NSCB), http://www.nscb.gov.phi. various years. Social Weather Stations, http://www.sws.org.phi World Bank (2010), Philippines Discussion Note No.3, "Improving Tax Policy and Administration," manuscript. 16 CHAPTER II POVERTY IN THE PHILIPPINES A. Introduction 2.1 The Philippines made progress in the fight against poverty during the 1980s and 1990s, although the progress was relatively slow. Using the $L25-a-day income threshold measure of poverty, it succeeded in reducing poverty from around 30 percent in the early 1980s to just over 22 percent at the end of the 1990s. Though significant, the decline of poverty in the Philippines over this period was quite weak compared to other countries in the region (Figure 2.1). 2.2 As a result, the Figure 2.1: Evolution of Poverty in the Philippines and Other East poverty rate for the East Asian Countries, 1980s to Mid-2000s Asia and Pacific region as a whole is now below the 100.0 Philippine rate, even ~ '2 g! .2 80.0 though it was nearly twice ~~!i 60.0 >- C::J as high just two decades I\ICI)Q. ,,:20 400 ago. This notable · CJ Q. '?c .... . difference in the pace of It)-O 20 ~ ~ o. poverty reduction was ~ 0.0 +----r--.,.......:::=~~_r-~-'___; mainly due to an equally Early Late Early Late Early Mid notable difference in the 80s 80s 90s 90s 2000s 2000s pace of per capita -East Asia & Pacific - Philippines economic growth, which averaged 0.1 percent in the Note: The evolution of poverty in East Asia is strongly influenced by China, which weighs heavily in the regional average. Philippines over this Source: World Bank, Development Data Platform. period, compared to 6.5 percent in the rest of the region. 14 2.3 Progress in poverty reduction was interrupted by the East Asia financial crisis in 1997-98. Although the Philippine economy recovered after the crisis, with GDP growth reaching a high of 7.1 percent in 2007, its poverty indicators remained unchanged or worsened. The poverty headcount ratio, or proportion of the population with incomes below the national poverty line, remained about the same between 2000 and 2003 then increased between 2003 and 2006 (Table 2.1). The increase in poverty over this period has been confirmed by different sources, using different methodologies. 15 The poverty 14 As noted earlier, the Philippines started out with lower poverty levels in the early 1980s and thus had already reaped the relatively easy gains to be had at the beginning of the poverty reduction process. 15 The poverty headcount indicators in Table 2.1 represent World Bank staff calculations based on FIES data., with poverty lines adjusted in line with the consumer price index at the provincial level to render them comparable across time. These indicators differ from the official indicators, which show a decline in the overall poverty headcount indicator between 2000 and 2003. However, both measures coincide in showing a marked increase in poverty between 2003 and 2006. The World Bank's $1.25-a-day threshold indicator 17 gap, which measures how far households lie below the poverty line, also increased from 2000 to 2006. Poverty severity, which takes into account the poverty gap but places a higher weight on households that are further away from the poverty line, also increased. The global food price crisis in 2007 and the global financial crisis that followed in 2008 are likely to have raised poverty levels further. 16 Table 2.1: Poverty Estimates Based on Poverty Lines at 2006 Prices, 2000-2006 Poverty Headcount Poverty Gap Poverty Severity ( of population) (of poverty line) ( of poverty line) 2000 2003 2006 2000 2003 2006 2000 2003 2006 Urban 16.8 17.3 19.5 4.6 4.7 5.4 1.8 1.8 2.2 Rural 44.7 44.4 45.9 14.0 14.4 14.6 5.9 6.3 6.2 Philippines 31.0 31.1 32.9 9.4 9.6 10.0 3.9 4.1 4.2 Poverty headcount based on official estimates ------------------------------------------- Philippines 33.0 30.0 32.9 Note: The estimates of poverty incidence are calculated on the basis of per capita income. For details on how the indices were computed, please refer to Annex 1. Source: World Bank estimates based on FIES 2000,3003,2006. 2.4 In addition to the population that lies below the national poverty line, a large proportion of the population is clustered just above the poverty line. Raising the poverty threshold to the international $2-a-day poverty line reveals that an additional 19.4 percent of the population in 2006 would be classified as poor. According to the Medium- Term Philippine Development Plan 2004-2010, 22 percent of the country's population is chronically poor, while 32 percent moves in and out of poverty (National Economic and Development Authority (NEDA), 2004). The large proportion of vulnerable population or "near poor" makes the challenge of poverty reduction even more daunting. 2.5 Progress on the non-income dimensions of poverty reduction, as measured by the MDG indicators, has been mixed. Considerable progress has been made in areas related tochild mortality, gender equality, and water supply and sanitation. However, the Philippines has fallen behind in critical areas such as universal access to primary education, with the net enrollment rate at the elementary level declining since 2002 and net enrollment rates at the secondary level leveling out to between 58-60 percent over the past five years. Based on progress to date, the MDGs for maternal health and reproductive health are unlikely to be achieved by 2015. The maternal mortality rate in 2006 was 162 deaths per 100,000 live births, which is high for a country at the Philippines' level of development. Malnutrition also remains a major issue, with the and Balisacan (2008), both of which use consumption-based measures, also show increasing poverty in this ~eriod. 6 The updated poverty figures are scheduled to become available in early 2011. World Bank simulations using a micro-macro general equilibrium model indicate that the poverty rate will be 0.9 and 1.5 percentage points higher in 2009 and 2010 as a result of the global crisis than it would have been in the absence of the crisis. 18 Philippines being among 20 countries in the world with the highest burden of child malnutrition. 17 2.6 The Philippines exhibits a very unequal, and possibly worsening, distribution of income and consumption. The World Bank's Development Data Platform currently identifies the Philippines as having the most unequally distributed income (or consumption) among the East Asian middle-income countries, whether measured by the Gini coefficient or the relative shares earned by the richest and lowest quintiles of the population (Table 2.2). The evidence on the evolution of income distribution is less clear. Table 2.2 indicates no change in the Gini coefficient for the Philippines since the early 1990s and a modest increase in the relative quintile share, while the rest of the region exhibits declines in both. Meanwhile, the FIES household survey data indicates that overall income inequality has been declining since 2000. However, as discussed below, the national income accounts and associated circumstantial evidence suggest that the distribution of income has worsened over the last decade once the richest households, which are usually under-represented in household surveys, are taken into account. Table 2.2: Inequality in Selected East Asian Countries, 1990-2008 Gini coefficients Bi-Lo quintile shares* (averages, ) (ratios; latest year) 1990-95 2002-0S 1990-95 2002-08 China n.a. 42 n.a~ 8.4 Indonesia n.a. 39 n.a. 6.7 Malaysia 49 38 12.2 6.9 Philippines 44 44 8.3 9.0 Thailand 45 42 9.4 8.0 Vietnam 36 39 5.6 6.4 Sample Average 44 41 S.9 7.6 *The Hi-Lo quintile shares refer to share of total income (or consumption) received by the richest quintile divided by the share of the poorest quintile; n.a. = not available. Source: World Bank Development Data Base; based on household income (or. in some cases, consumption expenditure) 2.7 Non-income indicators also reveal a high level of inequality in the Philippines. Large inequalities in health and education outcomes and in access to services persist across regions and income groups. In health, for example, one child in a thousand died between ages 1-5 among the wealthiest quintile of the population, compared to 25 per thousand in the poorest quintile and 33 per thousand in the Autonomous Region of Muslim Mindanao (ARMM). Similar disparities can be seen in education, with 82.9 percent of children 16 years or younger in the poorest quintile being in school compared to 98.0 in the richest quintile. In the ARMM, only 78 percent of children 16 years or younger attended school, compared to the national average of 90.2 percent and the National Capital Region (NCR) average of 943 percent. These \7 See NEDA-VNDP, Philippines Midterm Progress Report on the Millennium Development Goals, 2007. 19 discrepancies indicate a highly unequal distribution of human capital that becomes reflected in unequal earnings. B. Why Has Poverty Not Been Declining in Recent Years? 2.8 Broadly speaking, increases in poverty must be due to declines in average real income (or consumption), a worsening distribution of income, or a shift of the poverty line. That is, standard measures of poverty are determined by (i) the mean level of real income, (ii) the distribution around that mean, and (iii) where the poverty line is drawn. The poverty headcount ratio in Table 2.1 is based on a constant poverty line. The increase in poverty observed between 2003 and 2006, therefore, would either have to be due to an absence of growth (resulting in falling incomes and consumption), a significant worsening of the distribution of income, or both. Compelling evidence can be found in favor of both explanations. 18 2.9 The absence of more dynamic economic growth, coupled with high degrees of income inequality, partly explains why poverty failed to decline since 2000. As noted at the beginning of this report, the absence of economic growth has a long-standing history in the Philippines and the growth episode registered since 2000 has been modest by regional standards (Table 2.3). Once the data biases discussed below are accounted for, economic growth in the Philippines is likely to have been even more modest than is indicated in the National Accounts. At the same time, the relatively high degree of income inequality exhibited by the Philippines has the effect of reducing the income elasticity of poverty, posing a further barrier to faster poverty reduction. 19 18 This conclusion is similar to one arrived at in the World Bank's last Philippines Poverty Assessment (World Bank, 2001b). After contrasting the evolution of poverty in the Philippines and other middle- income countries in East Asia over 1985-1997, it states (para. 2.6), "This begs the question: has slower poverty reduction been on account of slower growth in the Philippines or rather due to growth being less pro-poor? The evidence based on household survey data suggests that probably both factors have been at work." 19 To appreciate this difference, note that it would take 20 years to cut the Philippine poverty rate in half with an annual per capita income growth rate of 2.5 percent and the growth elasticity of -1.3 estimated for the Philippines. The time to achieve this target would be reduced to 10 years with an annual growth rate of 5 percent, leaving everything else the same, or to 12 years if only the growth elasticity were -2.3 (which is the estimated average elasticity for all East Asian countries). 20 Table 2.3: Per Capita GDP Growth in East Asia and the Pacific Q, 1980-2007 1980-1989 1990-1999 1980-1999 China 8.3 8.9 8.6 9.4 Indonesia 4.2 3.2 3.7 3.6 Korea, Rep. 6.6 5.3 6.0 4.8 Malaysia 3.0 4.5 3.8 3.5 Philippines -0.6 O.S -0.1 3.0 Thailand 5.7 4.2 5.0 4.1 Vietnam 2.2 5.4 4.3 6.4 East Asia & Pacific 6.0 7.0 6.5 8.0 World 1.4 1.2 1.3 2.0 Source: World Bank, DDP database, based on national accounts data. 2.10 The failure of poverty to decline after 2000 also appears linked to a worsening distribution of income. As argued below, the National Accounts-based figures are likely to have over-estimated income growth, while the FIES household survey-based figures are likely to have under-estimated consumption growth and poverty reduction. However, the corroborating evidence indicates that, on balance, aggregate economic growth was positive and that poverty did not decline. That is, data on the growth of exports, corporate profits and fiscal revenues and enterprise surveys all point toward positive economic growth during 2003-06. Meanwhile, the Social Weather Station reports, which regularly monitor perception data, corroborate the findings from the survey-based data that poverty and the incidence of hunger has been increasing in that period. These findings imply (as per paragraph 2.5) that the modest growth that did take place during 2003-06 must have been associated with a deteriorating distribution of income. _ _ _ _ _T_a;,.;.,b.;. .l. .:. e. .:. 2,.;. . 4L Estimates of the Growtb:!=lasticity of Poverty, 1980-20'-0.-...:6_ _ _ __ Source Years World E.Asia China Indonesia Philippines Thailand Vietnam Besley Varies by -0.73 -1.06 -0.60 -1.12 -0.70 -1.72 and country n.a. Burgess ( 1980-1998) (0.24) (0.25) (0.14) (0.38) (0.12) (0.48) (2003) World -2.12 -1.20 -2.60 -1.85 -5.15 -2.13 1990-2000 n.a. Bank'" (0.42) (0.14) (0.74) (0.21 ) (0.46) (0.10) World -2.19 -1.29 -1.85 -1.27 -4.55 -3.04 2000-2006 n.a. Bank'" (0.34) (0.07) (0.36) (0.45) (0.81) (0.18) Note: Standard errors in parentheses; n.a. = not available. *World Bank estimates using data from World Bank (2008). 2.11 Several factors have contributed to the worsening distribution of income. They include: (i) an unequal sectoral distribution of growth, which slows progress in poverty reduction when the poor are concentrated in the sectors that are stagnating or contracting; (ii) an unequal spatial or regional distribution of growth, which has a similar effect as the unequal sectoral growth when the poor are concentrated in declining or 21 stagnating regions and population mobility is limited; (iii) strong demographic pressures, reflecting relatively high population growth rates; (iv) declines in the relative price of labor provided by the poor, and (v) an unequal distribution of human capital and access to social services and, in some cases, a decline of public spending on social services that mainly benefit the poor. These factors are described in more detail below. C. The Puzzle of Diverging Per Capita Income Measures in the Philippines 2.12 While the National Accounts point to a significant improvement in GDP growth over 2000-06, the Philippines household survey data indicates that average real income has steadily declined during this period. The FIES, from which the poverty indicators are drawn, indicate that average real per capita consumption declined annually by an average rate of 1.4 percent between 2000 and 2006 (Table 2.5). This finding is consistent with the commonly observed negative relationship between poverty and consumption or income growth, and it squarely attributes the absence of progress in poverty reduction to an absence of growth. The National Income-based data, however, indicate that per capita consumption and GDP have been growing by well over 2 percent per annum during this period, so the failure of poverty to decline must be due to a substantial deterioration in the distribution of income. That is, it would mean that the benefits of growth since 2000 have accrued mostly to a sub-sample of mainly richer households that is not fully captured by the FIES. Table 2.5: Annual Growth in Per Capita Income and Expenditures in the Pbilippines Q, 1985-2006 1985-1997 2000-03 2003-06 2000-06 Household Survey-based growth rates Expenditures per capita 2.7 -0.9 -0.5 -0.7 Income per capita n.a. -1.7 -1.2 -1.4 National Accounts-based growth rates Consumption per capita 1.7 2.3 3.5 2.9 GDP per capita 1.4 1.7 3.6 2.7 Note: n.a. = not available. Sources: Household survey-based figures are based on the FIES; figures up to 1997 are from World Bank (200tb), Philippines Poverty Assessment; figures as of 2000 are from the Philippines National Statistics Office. The National Accounts-based growth rates are from the World Bank Development Database. 2.13 Ascertaining the nature of income and consumption growth during periods of rising poverty has important policy implications. If the survey-based findings that average real per capita consumption and income have been falling are taken at face value, the main challenge facing policy-makers is to bring about faster growth. If the National Accounts-based findings that average per capita income has been increasing are taken at face value, the accompanying rise in poverty indicates that policymakers also have to be 22 concerned about the distributional consequences of growth and ensure that it is sufficiently broad-based. 2.14 The divergence of GDP and survey-based incomes is not due to diverging national income and output patterns. One conceivable explanation for this phenomenon is that the growth patterns of gross national income (GNI) and GDP diverged over the period in question?O However, this does not appear to have been the case, as the evolution of GNI closely mirrors the pattern followed by GDP (Figure 2.2). Figure 2.2: Evolution of GNI versus GDP . in the Philippines, 2000-2008 400.0 1 iIIt 350.0 (/) :::l it) 300.0 co - (1) ~ 250.0 c ~ c 200.0 0 U 150.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 -Atlas GNI per capita (US$. cons 1985 prices) - Atlas GDP per capita (US$. cons 1985 prices) Source: World Bank estimates based on National Income Accounts. 2.15 The divergence in the survey-based and National Accounts-based data is not unique to the Philippines. As noted by Deaton (2005) in his comprehensive review of cross-country and inter-temporal relationships between survey and national accounts estimates of consumption expenditures per capita, "National Accounts estimates of consumption [levels] are typically, although not always, larger than survey-based estimates, and there is a tendency, both across countries and over time within important countries, for the NA estimate of consumption to grow more rapidly than does the survey-based estimate." Moreover, the differences in levels and growth rates are not trivial: Deaton finds that the ratio of survey-based consumption levels to National Accounts-based consumption averages 86 percent worldwide (82 percent in East Asia and the Pacific countries). Furthermore, while the annual growth of survey-based real consumption averaged between 1.9 and 2.3 percent over 1990-2000, the National Accounts yielded an average growth rate between 3.8 and 4.5, which is twice as large. 20One possible scenario that could have resulted in a rising GDP and falling GNI would have been the onset of negative remittance growth combined with growing non·factor transfers. Since 2000, however, remittance growth actually increased while non·factor transfers declined as the interest bill on the external debt declined, causing per capita real GNI to grow faster than per capita real GDP (2.8 percent vs. 2.2 percent) during 2001-08. 23 Finally, it should be noted that there is no general presumption of greater accuracy in favor of either the surveys or the national accounts,21 as both exhibit significant biases. 2.16 What is unusual about the Philippines' experience after 2000 is that the survey-based data exhibits lower growth (and rising poverty), just as the National Accounts exhibit a significant increase in growth. Deaton and Dupriez (2009) noted that while the survey-based and national accounts-based consumption figures may diverge significantly from each other, their year-to-year variation is highly correlated. In a separate cross-country analysis using data for Latin America and the Caribbean, Perry et al (2006) reached the same conclusion when they regressed income growth according to the household surveys on income growth according to the national accounts, which yielded the following estimated regression line: y = 0.977x - 0.864. That is, the estimated slope of this regression line is very close to 1, so changes in national accounts-based growth are generally associated with equal changes in survey-based growth. In contrast, the recent experience of the Philippines suggests a negative relationship between both variables; that is, as the National Accounts show an increase in average annual per capita consumption growth from 1.7 percent over 1985-97 to 2.9 percent over 2000-06, the average survey-based per-capita growth rate declined from 2.7 percent to -0.07 percent. 2.17 The divergence in the survey-based and National Accounts-based growth pattern is most pronounced in the National Capital Region (NCR). The household survey-based income growth figures and the National Accounts-based GDP growth figures also exhibit a large divergence across the 16 regions in the Philippines: both sets of figures are uncorrelated, yielding an R-squared of less than 0.001. This lack of correlation is especially pronounced for the NCR, which shows an overall decline in survey-based per capita income of 20.6 percent between 2000 and 2006 and an overall increase in GDP per capita of26.7 percent over the same period?2 2.18 The FIES most likely underestimates the growth of real household consumption and poverty reduction since 2000. Deaton and Dupriez (2009) notes that the main source of bias in household surveys worldwide is that they are not fully representative because the more affluent households are typically under-represented in survey samples and also more reluctant to disclose their full incomes?3 This means that true income growth is likely to be higher than the data suggest. One telling sign that the 21 As Deaton (2005) points out, there is no reason to presume that both data sources should yield the same figures considering that the National Accounts are designed to track money (or macroeconomic aggregates), while the household surveys are designed to track people and estimates of poverty. 22 The large divergence in the NCR growth figures may be reflecting the rapid growth of profits in the Filipino corporate sector, which is largely concentrated in Manila and unlikely to be reported in the household survey figures for reasons discussed in paragraph 2.17. When NCR is excluded from the regressions, the cross-regional correlation is positive, but still weak, with an R-squared ofO.OS. 23 This sampling bias is particularly important for countries with very unequal distributions of income, such as the Philippines. As noted in World Bank (2005), pg. 16, the Philippines continues to stand out as a country with an unusually high concentration of family control over firms. In the mid-1990s, the top 15 families controlled over 50 percent of the stock market capitalization in the Philippines, which represents an ownership share equivalent to roughly 17 percent of Philippine GDP. Just leaving out these 15 families from the household survey would result in a major omission in the estimation of average household income figures. 24 household surveys in the Philippines may not be capturing the top end of the income distribution is that according to the FIES data, total household income of the top decile in 2006 was only Php620,000 (about US$12,400), or Php124,OOO (about US$2,700) in per capita terms. This would mean that the top decile earned just a little above the minimum wage,24 which is roughly equal to the expected income of a new graduate entering the labor force. Even though the FIES survey data may have underestimated household consumption growth, however, the finding that poverty incidence has increased during that period is corroborated by independent perception surveys, such as the Social Weather Station reports, and by the decline in average real wages. 2.19 The National Accounts, in turn, are likely to be over-estimating real consumption and GDP growth, primarily on account of measurement conventions in the compilation of national income accounts. These conventions exclude most non- exchanged services, such as home food preparation or home repairs, and the failure to capture intermediate consumption. As people become richer, they tend to substitute market-bought services for home-produced services (that are not captured in the national accounts), which biases estimated growth upward. While household surveys also tend to exclude non-exchanged services, Deaton (2005) identifies three important non-exchanged items of consumption that are included in the national accounts but not in the surveys: imputed rents to homeowners, indirectly imputed financial services, and consumption by non-profit institutions serving households. Their inclusion in the compilation of the national accounts but not in the survey data contributes further to the divergence in the national accounts-based and survey-based expenditure data. In addition, Medalla and landoc (2008) have pointed to shortcomings in the quality of the National Accounts in the Philippines, which also appear to have contributed to an overestimation of GDP growth. As noted earlier, however, even though the National Accounts-based data may have over-estimated economic growth since 2000, the overall evidence suggests that growth has nevertheless been positive during this period, which means that its limited impact on poverty gives cause for concern. C. Basic Poverty Profile 2.19 Poor Filipinos differ from the rest of the population along various dimensions. Table 2.6 shows how the poor are distributed spatially and sectorally relative to the population as a whole. It also shows how poor families differ from the average Filipino family in terms of household characteristics and access to basic services. In particular, the poor are concentrated in rural areas and in agriculture, have less access to basic services, lower levels of education, and larger families. 24In 2006, the daily minimum wage in NCR was around PhP340 for non-agricultural workers; in 2008, it was around PhP375 (Source: National Wages and Productivity Commission). 25 Table 2.6: Profde ofthe Poor in the Philippines, 2006 Out of 100 Filipinos ··· Out of 100 POOR Filipinos ··· 51 live in rural areas 71 live in rural areas 18 do not have access to electricity 40 do not have access to electricity 52 do not have their own water source 78 do not have their own water source 85 belong to male-headed households 90 belong to male-headed households 52 belong to families with more than five 71 belong to families with more than five members* members* 35 belong to families whose head works in 59 belong to families whose head works in agriculture agriculture 44 belong to families whose household heads are 55 belong to families whose household heads informal workers are informal workers 15 belong to families with unemployed 8 belong to families with unemployed household heads household heads 3 belong to families whose household heads did 6 belong to families whose household heads not attend school did not attend school 44 belong to families whose heads did not reach 66 belong to families whose heads did not high school reach high school '" The average family size at the national level is five. Source: World Bank estimates based on the 2006 FIES. Geographic distribution 2.20 While rural areas continue to be home to most of the poor, the gap between rural and urban poverty has declined. The share of the poor population living in rural areas has declined since 2000, but rural poverty still comprised 70.8 percent of all the poor in 2006 (Table 2.7). The rural-urban shares of the overall population have not changed much since 2000, as the population has remained split almost equally between rural and urban areas. However, increases in the poverty incidence in urban areas as well as in the share of the poor population living in these areas have reduced the gap between urban and rural poverty. The poverty gap and poverty severity in urban areas have also worsened since 2000. Table 2.7: Rural aud Urban Poverty in the Philippines, 2000-2006 Poverty Headcount Share of the Poor Population Share Change in Per ( of population) (of poor population) ( of population) Capita Income 0 2000 2003 2006 2000 2003 2006 2000 2003 2006 2000-2006 Urban 16.8 17.3 19.5 26.8 27.3 29.2 49.4 49.1 49.3 -11.0 Rural 44.8 44.4 45.9 73.2 72.7 70.8 50.6 50.9 50.7 -1.4 Source: World Bank estimates based on FIES data. 26 2.21 Given the stationary character of the population since 2000, rural-urban migration is unlikely to have played a significant role in the recent changes in the overall poverty rate. This is confirmed by Table 2.8, which decomposes the change in the overall poverty rate into a population-shift effect, an intra-regional effect, and an interaction effect, using the technique described in Ravallion (1994). It shows that most of the changes in poverty can be attributed to changes in headcount poverty within urban and rural areas, while shifts of the population across both areas have had limited effects on the overall poverty rate, particularly between 2003 and 2006. Table 2.8: UrlJan-Rural Povel'ty Decomposition, 2000-2006 , Change between 2000-2003 t Change between 2003-2006 I Absolute Percentage II Absolute Percentage I change change I change change Change in poverty headcount 0.13 100.00 1.78 100.00 Total intra-regional effect 0.10 74.65 1.83 103.12 Population-shift effect 0.03 26.10 -0.06 -3.21 Interaction effect -0.00 -0.75 0.00 0.08 Intra-regional effects Urban 0.25 193.69 1.08 60.56 Rural -0.15 -119.04 0.76 42.56 Source: World Bank estimates based on FIES data. 2.22 The regional distribution of poverty in the Philippines has not changed much since 2000. As shown in Table 2.9, the ARMM, Caraga, and Bicol remain the poorest regions of the country. Except for Davao, the Mindanao regions had poverty headcount rates above 40 percent in 2006. Although the Zamboanga Peninsula, Northern Mindanao, and Caraga have recorded improvements in the incidence of poverty, they continue to account for a disproportionate share of total poverty. For example, the Mindanao regions together account for 32 percent of total poverty but only for 23 percent of the total population. In contrast, the NCR, home to about 15 percent of the total population, accounts for less than 5 percent of total poverty. However, poverty incidence in the NCR has risen steadily since 2000. 27 Table 2.9: Poverty bl'. Administrative Region in the Phili~~ines, 2000-2006 Change in Poverty Headcount ~ Share of the Poor Population Share Per Capita ( of population) ~ (of poor population) ( of population) Income 0 2000 2003 2006 ~ 2000 2003 2006 2000 2003 2006 2000-06 ---------------~--- Region I IIoeos 30.3 28.5 32.7 I 5.2 4.8 5.3 5.3 5.3 5.3 -8.8 ! Region II Cagayan 25.5 25.2 25.6 2.9 2.8 2.6 3.5 3.4 3.4 3.22 Region III - Cen. Luzon IS.l 17.3 20.6 I 4.9 6.1 6.9 1O.l ILl 11.0 2.9 Region IV - Sou. Luzon 22.3 25.1 27.2 I 10.6 12.9 13.0 14.7 16.1 15.8 -9.4 Region V- Bieol 50.8 49.2 51.1 10.3 9.6 9.6 6.3 6.1 6.2 1.1 Region VI W. Visayas 42.5 39.8 38.6 I I ll.l 9.7 9.0 8.1 7.6 7.7 -2.1 Region VII - C. Visayas 36.5 34.9 35.3 I ! 8.5 8.2 8.0 7.2 7.4 7.5 5.5 Region VIII - E. 49.2 47.2 48.6 7.5 7.2 7.1 4.7 4.7 4.8 3.6 Visayas Region IX - Zamboanga 48.0 51.0 45.4 6.3 6.0 5.1 4.0 3.6 3.7 13.7 Region X 39.3 44.3 43.2 4.6 6.4 6.0 3.7 4.5 4.6 -0.9 N.Mindanao Region XI Davao 38.2 36.5 36.6 i 8.5 I 5.7 5.3 6.9 4.9 4.7 -5.0 ! Region XIl- 46.9 40.8 40.8 5.3 5.7 5.3 3.5 4.3 4.3 0.7 SOCCSKSARGEN Region XlII - NCR Region XIV - CAR 5.8 32.4 6.9 32.6 10.4 34.5 1 I 2.7 1.9 3.0 1.8 4.2 1.8 i 14.4 1.8 13.5 1.7 13.2 1.7 I -20.6 9. I Region XV ARMM 53.5 51.0 61.9 I 5.2 5.3 6.4 I 3.0 3.3 3.4 I -16.8 Region XVI - Caraga 51.4 55.2 52.6 I 4.5 4.6 4.2 I I I 2.7 2.6 2.6 I I 8.4 Philippines 31.0 31.1 32.9 I 100.0 100.0 100.0 I 100.0 100.0 100.0 I -8.2 Source: World Bank estimates based on FIES data. 2.23 The regions with the lowest poverty incidence are those that are economically most active. As shown in Figure 2.3, the NCR contributed 37.2 percent of national GDP in 2006 but had the lowest regional poverty headcount rate at 10.4 percent. Southern Luzon, with the second highest contribution to national GDP, also had one of the lowest poverty headcount ratios. However, it is noteworthy that the regions that experienced the greatest reductions in poverty between 2000 and 2006 were not necessarily the ones that enjoyed the fastest economic growth during this period. 28 Figure 2.3: Poverty and Regional Contribution to National GDP in the Philippines, 2006 Re&iGnal Contrilt.ltion to National GD~, 2006 Share in percent I~) Sources: Poverty rates are World Bauk estimates based on FIES 2006; Regional GDP data are from NSCB. 2.24 Most of the change in the national poverty rate since 2000 is attributable to intra-regional rather than inter-regional effects. To detennine how the changes in poverty across regions contributed to the overall increase in the national poverty rate, Table 2.10 decomposes the overall change in poverty into intra-regional effects and a population-shift effect. It shows that the increase in national poverty between 2003 and 2006 was concentrated in Central Luzon, Southern Luzon, NCR, and ARMM. The changes in the headcount index in these four regions combined explain almost 90 percent of the total change in the national poverty rate over that period, while the population-shift effects are fairly insignificant. This means that poverty has been increasing because some regions have grown significantly poorer and not because the distribution of people has shifted toward poorer regions. 29 Table 2.10: Poverty Decomposition by Administrative Region, 2000-2006 Change between 2000-2003 I Change between 2003-2006 Absolute Percentage i Absolute Percentage change change I change change Change in poverty headcount 0.13 100.00 1.78 100.00 Total Intra-regional effect Population-shift effect 0.16 -0.10 125.43 -75.48 I l.72 0.12 96.66 6.77 Interaction effect 0.06 50.06 -0.06 -3.43 Intra-regional effects I Region III - Central Luzon 0.22 174.83 I 0.34 19.03 ~;~:~: Region IV - Southern Luzon 0041 0.36 20.08 Region XIII - NCR 0.16 I 0.47 26.29 Region XV - ARMM -0.08 -58.95 i 0.42 23.36 All Other Regions -0.55 Source: World Bank estimates based on PIES data. -434.42 I 0.13 7.90 Employment and income 2.25 Poverty is much more prevalent among agricultural households than among non-agricultural households. In 2006, poverty incidence among agricultural households-defined as households in which income earned from agricultural activities is greater than or equal to income earned from non-agricultural activities-was three times higher than poverty incidence among non-agricultural households (Table 2.11). While households that derive most of their income from agriculture account for a quarter of the population, they account for half of all the poor. Poverty incidence among households whose family heads were employed in mining and quarrying was almost as high as in agriculture, while poverty rates were lowest among households whose family heads were employed in finance or public utilitics. Table 2.11: Distribution of Poverty b~ Sector, 2000-2006 Poverty Headcount Sbare of the Poor Population Share 2000 2003 2006 2000 2003 2006 2000 2003 2006 Agricultural Indicator of Household* Agricultural Household 61.9 63.2 65.4 53.2 53.7 50.2 26.6 26.4 25.3 Non-Agricultural 19.8 19.6 21.9 46.8 46.3 49.8 73.4 73.6 74.7 Household Sector of Employment ofHousehold Head Agriculture 53.1 53.7 55.4 62.4 64.4 59.4 34.6 33.3 31.4 Industry 25.7 25.2 28.7 13.3 12.5 12.6 16.4 17.2 16.3 Services 15.7 15.5 18.9 16.7 17.3 20.3 34.5 36.6 36.9 Not Employed 16.2 14.3 17.0 7.6 5.8 7.7 14.4 12.9 1504 Philippines 31.0 31.1 32.9 100.0 100.0 100.0 100.0 100.0 100.0 *Agricultural households are defined as those households in which the total income earned from agricultural activities, including family sustenance activities in agriculture and crops received as gifts, are greater than or equal to income earned from non-agricultural activities. Source: World Bank estimates based on PIES data. 30 2.26 The poor derive most of their income from wages and rely more on domestic remittances. Table 2.12 provides a further disaggregation of households' major source of income in 2006. While both poor and non-poor households derive most of their income from wages, wages comprise a bigger share of the non-poor's total income (45.2 percent) compared to the poor who are more diversified between wage incomes (39.4 percent) and entrepreneurial incomes (36.2 percent). Poverty incidence among households living on assistance from abroad (foreign remittances) is only 6.3 percent. In contrast, poverty incidence among households living mainly on domestic remittances is 43.6 percent, which is much higher than the national average of 32.9 percent. Poor households are more dependent on domestic remittances (5.4 percent of total income), while non-poor households are more dependent on foreign remittances (12.5 percent of total income). Table 2.12: Poverty b~ Major Source of Income 2 2006 Share of Total Income 0 Poverty Share of Population Among Among Major Source of Income ALL Headcount Poverty Share Non-Poor Poor HHs HHs HHs Wage Income 50.1 45.2 39.4 44.7 Wage/salary from agric. activity 66.5 20.7 7.9 1.9 13.0 3.0 Wage/salary from non-agric. 22.1 36.5 42.2 43.3 26.4 41.7 activity Entrepreneurial Income 31.0 21.5 36.2 22.9 Crop farming and gardening 61.5 31.6 13.1 4.1 16.9 5.4 Livestock and poultry raising 47.1 l.l 0.6 0.7 1.8 0.8 Fishing 64.8 8.8 3.5 0.8 5.0 1.2 Forestry and hunting 78.6 1.2 0.4 0.1 0.8 0.1 Wholesale and retail 24.0 6.9 7.4 8.4 5.7 8.2 Manufacturing 36.6 1.4 1.0 0.9 1.2 0.9 Community, etc. services 18.8 0.9 1.3 1.9 0.9 1.8 Transport and communication 30.9 3.4 2.9 2.6 2.5 2.6 Mining 63.7 0.4 0.2 0.1 0.3 0.1 Construction 13.8 0.1 0.2 0.4 0.1 0.4 Entrepreneurial activity n.e.c. * 20.1 0.3 0.4 0.8 0.3 0.7 Net share of crops and others 39.7 0.9 0.6 0.7 0.7 0.7 Other Sources 18.5 33.4 24.0 32.4 Assistance from abroad 6.3 2.1 8.5 12.5 1.6 11.5 Assistance from domestic source 43.6 6.8 4.0 3.1 5.4 3.3 Rental of lands and other 11.8 0.2 0.4 1.0 0.1 0.9 properties Interest from banks / loans 19.6 0.0 0.0 0.1 0.0 0.1 Pensions and retirements benefits 13.6 1.1 2.0 3.0 0.6 2.7 Dividend from investments 0.0 0.0 0.0 0.2 0.0 0.1 Rental value of owner-occup. 23.2 1.8 2.0 9.9 7.9 9.7 dwell. Income from fam. sustenance 77.8 0.9 0.3 0.7 4.7 1.1 activity Gifts 30.6 1.6 1.3 2.8 3.6 2.9 Others 25.5 0.0 0.0 0.1 0.1 Philippines 32.9 100.0 100.0 100.0 100.0 100.0 31 * n.e.c. = not elsewhere classified. Source: World Bank estimates based on FIES data. 2.27 The majority of the poor are engaged in the informal sector.25 According to the 2006 Labor Force Survey (LFS) and FIES data, nearly 60 percent of poor workers are in the informal sector. Poor households tend to be headed by individuals who are working without pay in their own family-operated businesses or who are self-employed without any employees, which is not surprising since this category includes many poor rural farmers. The fact that the poor are concentrated in the low-productivity agricultural sector, which can be seriously affected by natural disasters, and in informal jobs indicates that the poor face higher vulnerability from income and labor market shocks than the non-poor. 2.28 Poverty in the Philippines appears to be correlated more with underemployment than with unemployment. In 2006, only 17 percent of the unemployed were poor. The fact that the non-poor are disproportionately over- represented among the unemployed is not surprising, since the poor cannot afford to stay out of work for long periods (Esguerra, et.al., 2002). In contrast, Figure 2.4 shows that poverty incidence is correlated with underemployment. Fij!ure 2.4: Poverty Incidence and Unem Jo mentlUnderem 10 ment 2006 45.0 45.0 ~ e... 40.0 35.0 - C 40.0 35.0 · .! ~ ... 30.0 rl. ... 30.0 25.0 C 25.0 C III III ~ 20.0 ~ 20.0 o Q. 15.0 · Q. 15.0 ~ E 10.0 III C 10.0 l' ::::l 5.0 -g 5.0 0.0 +-~... ..,~~--r-...... ----,------, ::::l 0.0 +--~---,~_--,-_ _-.-_-----, 0.0 20.0 40.0 60.0 80.0 0.0 20.0 40.0 60.0 80.0 Poverty Incidence (% of population) Poverty Incidence (% of population) Note: Each observation corresponds to the combination of poverty incidence and unemployment/underemployment observed in a particular region. Source: World Bank estimates based on FIES and LFS 2006 Household characteristics 2.29 Poverty rates are higher for households with less-educated household heads. 26 The poverty rate of 65 percent among household heads without education (no 25 Informal sector workers as used here include (i) infonnal sector operators who are either self-employed without any paid employee or employer in own-family operated fann or business (NSO, 2009), and Oi) workers who do not receive wages from own family-operated fann or business. 26 The household head is defined as "the person responsible for the care and organization of the household. He/She usually provides the chief source of income for the household. In households consisting of two or 32 grade completed) was almost twice as high as the national poverty rate of 33 percent in 2006 (Table 2.13). Although the overall level of education in the Philippines has been rising, the share of poverty in less-educated households remains large. In more than two- thirds of poor households, the household head had no high school education. In terms of gender and age, unlike in many other developing countries, female-headed households have a much lower poverty rate than male-headed households in the Philippines. Poverty rates are highest among households whose head is between 30 and 39 years of age, while the share of the poor is highest among households with heads between 40 and 45 years of age. Table 2.13: Poverty by Gender and Education Level of Housebold Head, 2000-2006 Poverty Headcount Share of the Poor Population Share 2000 2003 2006 2000 2003 2006 2000 2003 2006 Gender Male 32.9 33.3 35.2 91.2 92.4 90.7 85.8 86.4 84.7 Female 19.2 17.5 20.0 8.8 7.6 9.3 14.2 13.6 15.3 Educational Attainment No grade completed 58.2 62 65.2 6.9 5.6 4.8 3.7 2.8 2.4 Some elementary 49.9 52.1 52.2 35.3 36.8 34.9 22.0 22.0 22.0 Elementary completed 38.3 38.8 43.8 26.5 25.6 26.2 21.5 20.5 19.6 Some high school 34.2 33.7 37.0 13.3 13.9 14.4 12 12.9 12.8 High school completed 21.2 20.3 22.8 14.2 13.9 15.2 20.8 21.3 21.9 Some college 9.5 9.7 10.8 3.3 3.5 3.8 10.7 11.2 1I.5 College completed 1.7 2.5 2.4 0.5 0.7 0.7 9.1 9.2 9.5 Postgraduate 0.0 1.0 0.0 0.0 0.0 0.0 0.3 0.2 0.2 Philippines 31.0 31.1 32.9 100.0 100.0 100.0 100.0 100.0 100.0 Source: World Bank estimates based on FIES data. 2.30 As in many other countries, larger households in the Philippines tend to be poorer. Table 2.14 shows that poverty incidence steadily increases with household size, with nearly half of households with seven or more members living below the poverty line. The number of children in the households seems to be a particularly important determinant of poverty in the Philippines. In 2006, more than half of households with three of more children were poor, and nearly two-thirds of poor households had three or more children. The link between household size and ~overty is particularly significant in the Philippines due to its high population growth rate, 7 as discussed in later chapters. more unrelated persons sharing the same cooking facilities and meals, the head is usually the oldest male or female in the group regarded as such by the other members" (NSCB, 2003b). 27 The annual population growth rate in the Philippines is 2.0 percent over the period 2000 to 2007, compared to an average rate of 0.8 percent in East Asia and the Pacific and an average rate of 1.2 percent for lower middle-income countries worldwide (NSO 2008; World Bank Development Data Platform). 33 Table 2.14: Poverty b;r Household Size and Number of Children, 2000-2006 Poverty Headcount Share of the Poor Population Share 2000 2003 2006 2000 2003 2006 2000 2003 2006 Household size 1 member 7.6 6.3 4.6 0.2 0.2 0.1 0.7 1.0 1.0 2 10.8 9.7 10.6 1.2 1,4 1,4 3,4 4,4 4,4 3 12.9 13.5 14.2 3.4 4.5 4.3 8.2 10,4 9.9 4 17.0 19.3 19.8 8.0 10.7 10.2 14.7 17.2 16.9 5 24.3 26.7 29.7 14.8 16.7 17.5 18.8 19.4 19,4 6 33.5 35.3 38.2 19.4 18.9 19.7 18.0 16.7 17.0 7+ members 45.4 47.9 49.1 53.0 47.6 46.8 36.2 30.9 31.4 Number of Children ochild 11.2 10.7 11.7 6.6 6.2 6.8 18.3 17.9 19.2 1 15.0 15.2 18.3 8.7 9,4 11.1 17.9 19.1 20.0 2 23.6 24.3 26.7 14.8 16.8 17.6 19,4 21.5 21.6 3+ children 48.8 50.8 54.2 69.9 67.7 64.5 44.4 41.5 39.1 Philippines 31.0 31.1 32.9 100.0 100.0 100.0 100.0 100.0 100.0 Source: World Bank estimates based on FIES data. 2.31 Poor and non-poor households also differ significantly in their spending patterns. As shown in Table 2.15, the poor tend to spend proportionately more on food than the non-poor. The share spent on cereals is particularly high for the poor, accounting for one quarter of their total expenditures. The hike in food prices during 2008 is therefore likely to have affected the poor most severely-see Chapter Annex II on the impact of inflation on poverty. Next to food, households spend much of their incomes on housing rental, which account for about 13 percent of an average household's total expenditures but a much smaller share of poor households' total expenditures since, compared to the non-poor, a greater proportion of the poor either own the house and lot they live in or do not pay rent to the owner of the lot. While the average share of household spending devoted to medical care is less than 5 percent of total household spending and the share for education is only about 3 percent, poor households devote an even smaller share of their total expenditures to health and education-about half that of the average Filipino household. 34 Table 2.15: Household EX2endituresz 2000-2006 2000 2003 2006 Non- All Non- All Non- All Poor Poor Poor Poor HHs Poor HHs Poor HHs AnnualHH Expenditures 143,191 48,726 118,838 130,091 47,020 108,950 129,264 47,664 107,343 Annual Per Capita 30,091 7,920 23,220 29,234 7,984 22,621 29,228 8,076 22,270 Average Family Size 5 6 5 4 6 5 4 6 5 Annual BH Expenditures in 2000 PhP Food 12,414 4,997 10,116 11,946 4,888 9,750 11,352 4,855 9,215 Non-food 17,677 2,923 13,105 17,288 3,096 12,872 17,876 3,221 13,055 Fuel, Ligh~ & Water 1,874 516 1.453 1,912 521 1,479 2,248 583 1,700 Transport &Comm. 2,197 232 1,588 2,287 266 1,658 2,548 338 1,821 Education fees 1,353 156 982 1,258 147 912 1,385 141 976 Medical fees 609 85 447 680 103 501 910 113 648 Taxes 720 12 501 691 14 481 536 13 364 Shares of Total Expenditure 0 Food 41.3 63.1 43.6 40.9 61.2 43.1 38.8 60.1 41.4 Non-food 58.7 36.9 56.4 59.1 38.8 56.9 61.2 39.9 58.6 Fuel, Ligh~ & Water 6.2 6.5 6.3 6.5 6.5 6.5 7.7 7.2 7.6 Transport &Comm. 7.3 2.9 6.8 7.8 3.3 7.3 8.7 4.2 8.2 Education fees 4.5 2.0 4.2 4.3 1.8 4.0 4.7 1.7 4.4 Medical fees 2.0 1.1 1.9 2.3 1.3 2.2 3.1 1.4 2.9 Taxes 2.4 0.2 2.2 2.4 0.2 2.1 1.8 0.2 1.6 Source: World Bank estimates based on FIES data. 2.32 Poverty is not simply a lack of income but is also associated with poor housing conditions, limited access to infrastructure and basic services, and lack of assets. In 2006, only 48 percent of poor households lived in a house with walls made of predominantly strong materials, compared to 84 percent of non-poor households. The poor live in much smaller houses, with an average floor size of37.5 m2 versus 62.3m2 for non-poor dwellings, and the houses of the poor have worse sanitary facilities, less access to electricity, and inferior water sources. Such inequalities also persist across regions, with the least prosperous regions-mostly areas in Mindanao-having the least access to infrastructure and basic services (Table 2.16). The poor also have fewer assets than the non-poor. For example, only 40 percent of the poor possess a television, compared to 85 35 percent of the non-poor. The same applies to ownership of other items such as radios, refrigerators, telephones, and cars. Table 2.16: Access to Basic Services, 2006 Proportion 0 of the Proportion 0 of population Telephone Region I Island Group concrete & without access to density* asphalted Electricity Water Source roads Region XIII NCR 2.3 ]0.9 100.0 31.1 Region XIV CAR 21.2 23.3 35.7 5.5 Region I - Ilocos Region 9.7 6.8 90.0 3.8 Region II - Cagayan Valley 19.1 16.4 69.5 1.4 Region III - Central Luzon 5.5 4.2 87.2 4.5 Region IVA CALABARZON 7.7 14.5 85.8 10.2 Region IVB MIMAROPA 38.0 19.9 46.1 10.2 Region V- Bicol 27.9 26.9 72.2 2.3 Region VI - Western Visayas 23.1 33.6 75.6 6.3 Region VII - Central Visayas 21.5 27.3 85.7 7.4 Region VIII - Eastern Visayas 26.2 19.3 81.3 3.7 Region IX - Zamboanga Peninsula 34.1 34.5 68.6 1.1 Region X - Northern Mindanao 23.8 18.1 69.5 3.8 Region XI - Davao 24.2 19.9 62.9 7.9 Region XII - SOCCSKSARGEN 30.8 19.5 62.4 2.1 Region XV - ARMM 50.1 64.9 46.3 5.4 Region XVI - Caraga 20.3 16.9 n.a. 1.0 Island groups Luzon 11.2 13.0 71.5 11.7 Luzon (excluding NCR) 13.9 13.6 69.4 5.6 NCR 2.3 10.9 100.0 31.1 Visayas 23.2 27.8 80.3 6.1 Mindanao 30.3 28.0 62.2 3.6 Philippines 18.0 19.4 71.5 8.3 * Number of telephone lines installed per 100 persons; n.a. = not available. Sources: 2006 FIES; NTC; DPWH. D. Key Correlates of Poverty 2.33 To identify the key correlates of poverty, per capita household income and expenditures were regressed on a number of demographic, education, employment and sector indicators. Table 2.17 reports the regression results for household incomes,28 both for urban and rural areas using data from FIES 2006. The results show the partial correlation between incomes and other variables but do not necessarily indicate a causal relationship between the dependent and explanatory variables. 28 The regression using household expenditures as the dependent variable yields almost identical results. 36 Table 2.17: Household Income Functions for Urban and Rural Areas, 2006 (Dependent Variable: Log of Household Income) Explanatory Variable. Constant Coef. URBAN ··e. 9.542 ** (0.102) I Coef. RURAL ··e. 9.488 ** (0.076) Demographic indicators I Log of household size -0.393 ** (0.035) I -0.254 ** (0.030) Log of household size squared -0.018 * (0.012) I -0.067 ** (0.010) Ratio of children in household -0.714 ** (0.024)! -0.628 ** (0.020) Ratio of elderly in household -0.256 ** (0.032) I -0.339 ** (0.023) Female household head 0.091 ** (0.013)! -0.012 (0.013) Log of household head's age 0.196 ** (0.022) I 0.186 ** (0.018) Education attainment indicators I Some elementary 0.226 ** (0.051) i 0.109 ** (0.016) Elementary completed 0.379 ** (0.051) I 0.236 ** (0.017) Some high school 0.470 ** (0.051) i 0.326 ** (0.019) High school completed 0.702 ** (0.051) I 0.461 ** (0.018) Some college 0.936 ** (0.051) I 0.704 ** (0.022) College completed 1.381 ** (0.052) I 1.189 ** (0.027) Postgraduate education 1.781 ** (0.106) i 1.731 ** (0.158) Employment status indicators I Worked for private household -0.234 ** (0.033) j -0.330 ** (0.040) Worked for private establishment -0.117 ** (0.015) i -0.161 ** (0.016) Worked for government/government -0.020 ** (0.020) I 0.084 (0.023) corporation I Self-employed without any employee -0.162 ** (0.015) i -0.142 ** (0.015) Employer in own family-operated farm 0.270 ** (0.025) I 0.1 05 ** (0.019) Worked with pay in own family-operated -0.056 (0.103) I -0.269 ** (0.097) business i Worked without pay in own family- 0.041 (0.058) I -0.054 (0.069) operated business III Sector indicators Non-agricultural household 0.505 ** (Om 7) I 0.341 ** (0.008) Number of observations 17,267 III 21,216 R-squared 0.495 . 0.490 Notes: ** denotes signifieance at 0.01 level; '" denotes significance at 0.05 level; s.c. = standard error. Source: World Bank estimates based on FIES 2006. 2.34 Demographic characteristics are among the most important correlates of household incomes and expenditures. Larger households and households with higher ratios of children and elderly people tend to have lower per capita incomes and expenditures. Education is also an important correlate: all the education dummy variables are positive and significant in both urban and rural areas. Higher educational levels are associated with higher levels of per capita income and expenditures. For the same level of education, the correlation coefficient for urban areas is higher than for rural areas, suggesting that the returns on education may be higher in urban areas. Compared to the reference case of no education attainment, per capita income is 46.1 percent higher in 37 urban areas and 26.6 percent higher in rural areas when the household head has completed elementary schooL When the household head has completed high school, per capita incomes are 101.9 percent and 58.6 percent higher. 2.35 The employment status of the household head is also an important correlate. Compared with the reference case of no job or business, per capita income and expenditures in the household tend to be much lower when the household head is working for a private household or is self-employed without any employee. Other things being equal, non-agricultural households tend to have higher per capita incomes than agricultural households by 65.7 percent and 40.6 percent in urban and rural areas, respectively. 2.36 Although these regression results do not point to a causal link between policy instruments and poverty outcomes, they have identified areas of vulnerability that deserve further attention. In particular, the importance of the demographic indicators, coupled with the Philippines' relatively high population growth rate, suggests that greater attention should be devoted to population and maternal health policies. Policies to raise economic productivity and remove distortions in the labor market, which must absorb the fast growing workforce, also deserve close attention. In a similar vein, the high correlation of education indicators with household spending (and earnings) capacity suggests that social sector policies and education policy, in particular, require a closer look. Finally, the strong positive association between agriculture sector-related activities and poverty indicators suggests that special attention should be devoted to policies affecting the agriculture sector, as well as to policies that would facilitate a smoother labor transition to the manufacturing sector. E. Policy Implications and Conclusions 2.37 Better data on growth and poverty changes is needed to help inform policymaking. As discussed earlier, data constraints make it difficult to determine whether the recent increase in poverty has been caused by inadequate growth or by growth that is not sufficiently inclusive. If the FIES data is taken at face value, the main obstacle to poverty reduction in the Philippines has simply been the absence of growth. If the NIA data is taken at face value, the main problem has not been a lack of growth but instead a failure in ensuring that such growth is shared more broadly. Both data sources appear to have significant shortcomings. The system of national statistics in the Philippines needs to be strengthened urgently so policymakers have timely and accurate data that will enable them to identifY and respond quickly to adverse changes in poverty or other social indicators. 2.38 Given the preceding findings, an effective strategy for attacking poverty will involve efforts both to accelerate growth and to ensure that growth takes place in an inclusive manner. With regard to inclusiveness, the poverty profile presented above identified various factors that inhibit the poor from being able to take advantage of better income opportunities and participate in the market economy. These factors include particular demographic characteristics (i.e. large household size and dependency ratios 38 associated with fast population growth), limited human capital (i.e. low education attainment levels), and sector and type of employment (i.e. dependence on agriculture and the informal sector), which are strongly correlated with a higher incidence of poverty. 2.39 The remainder of this report explores the different factors that may have contributed to the increase in poverty since 2000 and how policymakers can address them to make growth more inclusive. The subsequent chapters are divided into two thematic sections, with the first focusing on enhancing the income opportunities of the poor and the second focusing on assisting households to participate in markets. The chapters in the first section analyze productivity growth in the agriculture and manufacturing sectors and the prevalence of distortions and segmentation in the Philippine labor market, with the aim of identifying policies that can raise labor productivity and eliminate labor market distortions, especially those that tend to exclude the poor. The chapters in the second section examine public actions and outcomes in the health, education, and social protection sectors, with the aim of identifying policies that could improve human capital as well as help the poor manage shocks that could threaten their livelihoods. 39 References Asian Development Bank (2005). Poverty in the Philippines: Income, Assets, and Access. Asian Development Bank. Balisacan, A. M. (1995). "Anatomy of Poverty during Adjustment: The Case of the Philippines." Economic Development and Cultural Change. 44(1): 33-62. (2000). "Growth, Redistribution and Poverty: Is the Philippines an Exception to the Standard Asian Story?" Journal of the Asia Pacific Economy. 5(1/2): 125-140. (2008). "Poverty Reduction: What We Know and Don't," University of the Philippines Centennial Lecture Series. January 2008. Balisacan, A.M. and S. Fujisaki (1999). "Causes of Poverty: Myths, Facts and Policies Philippine Study." Balisacan, A.M. and N. Fuwa (2004). "Going beyond Crosscountry Averages: Growth, Inequality and Poverty Reduction in the Philippines." World Development. 32(11): 1891 1907. Besley, T., and R. Burgess (2003). "Halving Global Poverty." Journal ofEconomic Perspectives. 17(3): 3-22. Chen, S. and M. Ravallion (2008). "The Developing World is Poorer than We Thought, but No Less Successful in the Fight against Poverty." World Bank Policy Research Working Paper No. 4703. Datt, G. and M. Ravallion (1992). "Growth and Redistribution Components of Changes in Poverty Decomposition with application to Brazil and China in 1980s." Journal of Development Economics. 38: 275-295. Deaton, Angus (2005), "Measuring Poverty in a Growing World (or Measuring Growth in a Poor World)", The Review of Economics and Statistics, Vol. LXXXVIII, No.1, February, pg. 16. Deaton, Angus and Olivier Dupriez (2009). "Purchasing Power Parity Exchnage Rates for the Global Poor," Princeton University Working Paper. Esguerra, 1., M. Ogawa, and M. Vodopivec (2002). "Options for Public Income Support for the Unemployed in the Philippines." Employment Policy Primer. Foster, lE., 1 Greer, E. Thorbecke (1984), "A Class of Decomposable Poverty Indices," Econometrica 52, pp.761-766. Hutchcroft, Paul D. (2008). "The Arroyo Imbroglio in the Philippines." Journal of Democracy Vol. 19, No.1, pp. 141-55. 40 Mangahas, M. (1995). "Self-rated poverty in the Philippines, 1981-1992." International Journal of Public Opinion Research 7(1): 40-52. Mangahas, M. and L.L. Guerrero (2002). "Self-Sustained Quality of Life Monitoring: The Philippine Social Weather Reports." Social Indicators Research 60: 123-144. Medalla, Felipe M. and Karl Robert L. Iandoc (2008). "Philippine GDP Growth after the Asian Financial Crisis: Resilient Economy or Weak Statistical System?" University of the Philippines School of Economics Discussion Paper No. 0802. May 2008. National Economic and Development Authority (2004). Medium Term Philippines Development Plan 2004-2010. National Statistical Coordination Board (2003a). "Approving the Proposed Methodology for Computation of Provincial Poverty Statistics." In http://www.nscb.gov.ph/resolutions/20031l.asp. (2003b). "Approving and Adopting the Official Concepts and Definitions for Statistical Purposes of the Selected Sectors: Prices, Population, Housing, and Tourism," in http://\\'WW.nscb.gov.ph/resolutions/ 2003/ lIAnnex.asp. (2005). "Official Poverty in the Philippines: Methodology and 2003 Estimates." (2009). "Informal sector operators counted at 10.5 million - Results from the 2008 Informal Sector Survey." (2008). "2006 Official Poverty Statistics (posted OSMarch 2008)." In http://WVvw.nscb.gov.ph/poverty/default.asp. (2007). "Annual Per Capita Poverty Thresholds by Province, 2006 - 2007 (preliminary estimates as of 02 March 2007)." In http://www.nscb.gov.ph/poverty/2006-2007/pov_th_07 .asp. National Statistics Office (2008). "Population Census 2007 Press Release." National Wages and Productivity Commission, website www.nwpc.dole.gov.ph. Pemia, E. (1990). "Economic Growth Performance of Indonesia, The Philippines, and Thailand: The Human Resource Dimension." ADB Report No. 48. Perry, Guillermo, Omar Arias, Humberto Lopez,; WilJiam Maloney, and Luis Serven (2006). "Poverty Reduction and Growth: Virtuous and Vicious Circles. The World Bank," Washington DC. 41 Ravallion, M. (1994). Poverty Comparisons. Fundamentals of Pure and Applied Economics 56. Harwood Academic Publishers. (2001). "Growth, Inequality and Poverty: Looking Beyond Averages." World Development 29(11): 1803-1815. Ravallion, M. and S. Chen (2002). "Measuring Pro-Poor Growth." Economic Letters 78: 93-99. World Bank, Development Data Platform. Available in http://go.worldbank.orgIK2XOMZ5DCO. (2001a). "Philippines Poverty Assessment Volume I: Main Report." Poverty Reduction and Economic Management Unit, East Asia and Pacific Region. The World Bank. (2001 b). "Philippines Poverty Assessment Volume II: Methodology." Poverty Reduction and Economic Management Unit, East Asia and Pacific Region. The World Bank. (2008). "East Asia: Testing Times Ahead." East Asia & Pacific Update. April 2008. The World Bank. (2005) The Philippines Towards a Better Investment Climate for Growth and Productivity. Poverty Reduction and Economic Management Unit, East Asia and Pacific Region. The World Bank. 42 CHAPTER II - ANNEX I MEASURING POVERTY IN THE PHILIPPINES Measuring poverty is inherently difficult because of its multi-dimensionality. It is characterized not only by a lack of income but also by the lack of other important dimensions of well-being, such as health and education. It is a challenge, therefore, to agree on which dimensions to incorporate and how to incorporate them in analyzing poverty in any given country. A simpler approach is to focus on a scalar measure that is of particular importance, and such measures will be the focus of this analysis. Income and consumption measures are two of the most predominantly used measures for this purpose. They are objective and relatively easy to interpret. Consumption is often preferred to income for three reasons: (i) income fluctuates more than consumption because people smooth consumption by loans and savings; (ii) income may become negative, whereas consumption is always nonnegative; and (iii) income tends to get underreported, particularly by the rich-this is also a problem for consumption, but the extent is generally believed to be less. Meanwhile, in most low- and middle-income countries where access to financial markets is limited, the use of income-based poverty measures is common. The FIES contains both income and consumption measures. Despite the several advantages of using consumption-based poverty measures, this report uses the income- based definition of poverty for 2006 as the point of reference because official poverty statistics published by the NSCB are based on income and have already been widely used. Hence, providing the poverty profile that is consistent with the latest official statistics (2006) is useful for those who are already familiar with them. 29 Box Al below describes the official methodology for computing poverty estimates in the Philippines. The current methodology used for setting the poverty threshold for official statistics is a result of two major revisions to the original methodology that was created in 1985-the 1992 and 2003 revisions. 3o These changes resulted in lower poverty lines and thus a lower poverty incidence, as well as breaks in the comparability of poverty data (Figure AI). For instance, the 1992 revision brought down the original poverty estimate in 1985 from 59 percent to 44.2 percent, while the 2003 revision resulted to a change in poverty estimates in 2000 from 33.7 percent to 28.4 percent. Changes in methodology also changed the trend of poverty incidence. The 2003 revision, for instance, revealed a 0.6 percentage point reduction in poverty among families between 1997 and 2000, while the old methodology showed a 1.9 percentage point increase in poverty. A recent refinement of the 2003 methodology also resulted in the lowered poverty incidence. Hence, great caution needs to be exercised when interpreting and comparing official poverty estimates 29 For consumption-based poverty estimates, see Balisacan (1999). Income-based poverty and consumption-based poverty are likely to be very similar because of the high correlation between income and consumption. The correlation coefficient between per capita income and per capita expenditure based on the 2006 FIES, for example, is 0.9. 30 The NSCB (2005) provides details on these changes. 43 across time. Official poverty thresholds and incidence for 2000 to 2006 are provided in (Table A4). Box At: Official Methodology for Estimating Poverty in the Philippines In accordance with NSCB Resolution No.1, Series of 2003, "Approving the Proposed Methodology for Computation of Provincial Poverty Statistics," the official methodology for estimating poverty starts with the computation of the food threshold, which refers to the minimum cost of the food items that satisfY nutritional requirements for economically necessary and socially desirable physical activities. The nutritional requirements are determined by the Food and Nutrition Research Institute (FNRI). Currently, these are based on 100 percent adequacy for the recommended dietary allowance (RDA) for protein and energy equivalent to an average of2,000 kilocalories per capita, and 80 percent adequacy for the other nutrients .. These nutritional requirements, found in the 1989 RDA for Filipinos for Energy and Specific Nutrients, were computed by the FNRI based on the age and sex population structure of each region using the results of t1:te 1990 Census of Population and Housing. The food threshold is also sometimes referred to as the subsistence threshold or the food poverty line. The provincial food thresholds, which are computed separately for urban and rural areas, are determined by using regional menus (with separate menus for urban and rural areas) and provincial prices in accordance with NSCB Resolution No.1, Series of 2003. For food items in the menu that come in different varieties-e.g., rice, which could be ordinary, special, NFA rice, etc.-and hence with different prices, the price of the cheapest variety available in the province is used in the computation of the provincial food threshold. For this reason, the price of special rice is not used since it will violate the minimum income/expenditure criterion. The NFA rice, on the other hand, suffers from the issue of accessibility, especially in poor areas, even though it is the lowest-priced variety of rice. Taking into account these concerns on the use of special and NFA rice, the price of ordinary rice is used as a compromise between the high-priced special rice and the low-priced but possibly inaccessible NFA rice. The expenditure pattern of households within the 10-percentile band around the provincial urban/rural food threshold as indicated by the ratio of food expenditure (FE) to total basic expenditure (TBE), or the FEITBE ratio, is then used to derive the provincial urban/rural poverty threshold. The poverty threshold or poverty line, also computed separately for urban and rural areas, refers to the cost of the basic food and non-food requirements. The basic non-food requirements cover the non-food expenditure items of the total basic expenditures. Thus, the poverty line may be viewed as the minimum income required to buy basic food requirements and other non-food basic needs. The food and poverty thresholds are then used to estimate the subsistence and poverty incidence, respectively, using the income approach. The subsistence incidence refers to the proportion of families (or population) with per capita income less than the per capita food threshold to the total number of families (population). On the other hand, the poverty incidence refers to the proportion of families (or population) with per capita income less than the per capita poverty threshold to the total number of families (population). Provincial food and poverty thresholds are updated whenever a new FIES comes out to account for actual food and non-food exoenditures of households. Source: National Statistical Coordination Board (NSCB). 44 Figure AI: Official Poverty Estimates in the Philippines, 1985-2006 1/1 70.0 60.0 j- ~ 50.0 's «I 40.0 u.. L- o 0_ a..~ .... - ~~.~ j - o 10.0 (I) u 0.0 +----r~-____,__- ...-.......,,___-...,._-__,_-____,__-___,,_____, c (I) 1985 1988 1991 1994 1997 2000 2003 2006 " '0 c Year -1985 Methodology -1992 Methodology Source: NSCB Besides the methodological changes in computing poverty indices, there have also been changes in the master sample design from which survey households are selected, which affected the representativeness of the surveys. Beginning in July 2003, the NSO employed the 2003 master sample (MS) design in conducting household surveys. The new master sample shifted the sample domains to regions, unlike past surveys which used urban and rural areas of each province as the principal domains for the survey. Hence, the master sample is no longer representative at the level of urban and rural areas nor at the provincial level. While the NSCB still reports provincial-level statistics, the official statistics may not be very reliable at the level of urban and rural areas or at the provincial leveL The provincial poverty statistics derived from FIES 2003 and 2006 in this report also suffer from the same problem. As mentioned earlier, this analysis constructs the poverty statistics that are consistent with the official estimates for 2006. Thus, the official poverty lines in 2006 are used to benchmark the estimates for prior years 2000 and 2003. Unlike in the past, the NSCB only published the overall provincial poverty line for the 2006 fmal poverty estimates, even though they were based on provincial poverty lines defined at the level of urban and rural areas. Hence, the analysis uses poverty lines for urban and rural areas in each province that will replicate the official estimates. While the overall provincial poverty lines and poverty estimates are very similar to the official ones, they are not identicaL It should also be noted that the use of poverty lines that are differentiated by province raises issues of non-comparability across space. The purpose of this exercise is to mainly address the question on what the real trend in poverty reduction in the Philippines has been in recent years. The last section of this Annex discusses further details on income poverty lines used in this paper. For comparing poverty between 2006 and earlier years, poverty lines have been adjusted using the consumer price index at the provincial level, instead of using the official nominal poverty lines for earlier years. This is because the underlying menu changes every time a poverty line is set, making the official statistics not directly comparable across time (see Box AI). This issue is particularly important when the change in the 45 menu involves changes in the quality of food but that is not necessarily reflected in its nutritional content. 3l Note, however, that the use of the consumer price index to render the poverty estimates comparable across time in this report does not suggest that this is the best price index to use in adjusting the poverty lines. The consumer price indices used in the analysis reflect the food bundle for average consumers in each province, which do not necessarily reflect the consumption basket of a typical poor household. Ideally, one should use the food bundle for households around the poverty line. Since the Philippines uses regional menus in constructing the poverty lines (refer to Box AI), this entails updating the poverty line for all 17 regions which merits a separate exercise on its own. Throughout the report, the headcount index (used interchangeably with poverty incidence in this report), poverty gap, and poverty severity are used, which belong to the Foster- Greer-Thorbecke (FGT) class of poverty measures with parameter value of 0, 1, and 2 respectively. Formally, a FGT measure of poverty Pa is defined as where are a, N, Yi and z are parameter, population, individual i's income or consumption, and poverty line,32 where the minimum food and basic non-food needs are just satisfied at the poverty line. The headcount index (PO) refers to the proportion of the population with per capita income less than the per capita poverty threshold. Poverty gap (PI) is the total income shortfall of the poor expressed in proportion to the poverty line. Poverty severity (P2) is the total of the squared income shortfall of the poor expressed in proportion to the poverty line. Compared to PI, P2 is more sensitive to the income distribution among the poor-the worse the income distribution, the higher P2 would be. The Decomposition of Poverty Change between 2000 and 2006 As noted earlier, the decomposition analysis is path dependent. A question arises whether the analysis would provide a consistent picture for the period spanning 2000-06, rather than analyzing the 2000-03 period and the 2003-06 period separately. The analysis indicates that falling income is indeed the most important factor contributing to increased poverty, while the redistribution effect tends to decrease poverty, as shown in Table Al and Figure A2. 31 The menu is set at the point that is 100 percent adequate for energy and protein and 80 percent adequate for vitamins and minerals. (See http://www.nscb.gov.phitechnotes/poverty/compilation.asp for further discussion. ) 32 For further discussion, see Foster, Greer, and Thorbecke (1984). 46 Table AI: Growth and Redistribution Decom,eosition of Poverty Changes Change between 2000 and 2006 i I Actual ! Growth Redistribution Residual change ! component component component Urban 2.70 5.28 -l.90 -0.68 Rural 1.19 l.00 -0.12 0.31 Philippines 1.90 4.47 -2.67 0.10 Source: World Bank estimates based on FIES data. Figure A2: Growth Incidence Curves, 2000-2006 Total (years 2000 and 2006) Urban 95% confidence bounds Mean gro\\th rate ,2 .2 :,;? !l e -,6 ~ eOJ) ca ::l -1.4 " " « -2,2 -3 10 20 30 40 50 60 70 80 90 100 10 m w ~ ~ W M W W ~ Expenditure percentiles Expenditure percentiles Rural I I I I 10 m w ~ ~ W M W W ~ Expenditure percentiles Source: World Bank estimates based on FIES data. It is worth noting that mean per capita incomes in real terms have decreased between 2000 and 2006 in all quintiles, as shown in Table A2. The steepest drop occurred in the highest quintile, leading to a lower level of inequality overall in the Philippines. 47 Table A2: Mean Per Capita Incomes in Real Terms (2006=100), 2000~2006 Mean Per Capita Income Income Quintile 2000 2003 2006 I-Lowest 8,585 8,491 8,464 2 14,996 15,049 14,551 3 23,527 23,436 22,497 4 38,957 38,242 36,895 5 - Highest 109,195 100,526 96,783 Philippines 39,051 37,146 35,836 Source: World Bank estimates based on FIES data. Adjustment of Poverty Lines In order to adjust the poverty lines taking into account the difference between urban and rural areas, two sources have been used: the official estimate of poverty at the provincial~ level and the preliminary release of the official poverty line for urban and rural poverty lines. First, the "preliminary" poverty lines for urban and rural areas in each province were constructed by changing them while holding the ratio fixed. Using these poverty lines can ensure that the resulting poverty rate for each province coincides with the official poverty rate up to the published digit. However, due to rounding errors, the national~level poverty estimate was slightly lower with these poverty lines. Hence, the poverty line has been scaled up in the analysis so that the resulting poverty ratio at the national level coincides with the published official statistic. Table A3 below shows the preliminary release of poverty line as well as the ones used in this analysis. The provincial~level poverty line (poverty line for urban and rural areas weighted by the population) for this study is very close to the official poverty line. Furthermore, the resulting poverty rates are very close to the official statistics, with the difference between official estimates and the estimates from this analysis being within 1 percentage for all the provinces except for Camiguin, Aurora, Guimaras, and Cotabato. In these four provinces, the number of observations is very small (at most 80), so the difference is still relatively small compared with the standard errors associated with the estimates. Therefore, while the estimates from this analysis and the official estimates are not identical, the difference is practically negligible. 48 Table A3: Adjusted Provincial Poverty Lines (pL) and Poverty Estimates, 2006 Official : Official Preliminary PL : PL /a - Ib . PL for this study Poverty Headcount ( of population) --------.----- (PhP per capita p'er year) ------------- This Province Total Urban Rural Total : Urban Rural Total ~ Official Difference Study Abra 17900 17468 17468 I 17900 17900 60.7 60.7 0.0 Agusan del Norte 13986 14964 13059 13427 I 15587 13603 14171 40.0 40.1 -0.1 Agusan del Sur 14544 17358 13853 14617 i 17271 13784 15010 56.2 56.2 0.0 Aldan 15150 16980 13520 14095 I 18251 14532 15357 52.0 51.8 0.2 Albay 16128 17665 13915 14969 I 19033 14992 16351 46.2 46.2 0.0 Antique 14650 13776 12605 127241 15861 14513 14722 51.6 51.6 0.0 Basilan 13255 15712 13006 13335 i 15618 12928 13507 43.7 43.9 -0.2 Bataan 15538 16486 14270 15627 i 16392 14189 15545 10.5 10.3 0.2 Batanes 14970 15367 15367 l 14970 14970 0.0 0.0 0.0 Batangas 19616 18428 17222 17583 20559 II' 19213 19761 30.7 30.7 0.0 Benguet 17483 17094 15661 16086 18579 17021 18036 11.1 11.1 0.0 Bohol 13610 13845 12317 12541 I 15025 13367 13833 46.9 46.7 0.2 Bukidnon 12186 14791 13223 13410 i 13441 12016 12313 37.2 37.3 -0.1 Bulacan 17768 17069 15844 16895 I 17951 16663 17709 13.4 13.5 -0.1 Cagayan 12928 14966 11575 12169 15899 12297 13144 23.1 23.1 0.0 Camarines Norte 14854 17704 13691 14928 17616 13623 14601 49.3 49.2 0.1 Camarines Sur 14634 17055 13017 13737 I 18169 13867 14770 49.9 49.9 0.0 Camiguin 16145 13676 15005 14387 i 15347 16839 16412 42.1 39.5 2.6 Capiz 14242 14372 12872 13068 15663 14028 14279 30.2 30.2 0.0 Catanduanes 13654 21980 13527 14230 1 21090 I 12979 14532 46.8 46.3 0.5 Cavite 18718 17293 18332 17595 i 18397 19502 18473 11.2 11.2 0.0 I Cebu 13960 13927 11645 125241 15524 12980 14701 28.2 28.2 0.0 Davao 15753 17822 14992 15535 18072 15202 16249 44.8 44.9 -0.1 Davaode Sur 14452 17314 12732 14350 I 17437 12822 15133 27.4 27.4 0.0 Davao Oriental 13741 14932 12474 13294 I 15434 12893 13578 48.8 48.8 0.0 Eastern Samar 13873 13704 13257 13291 14304 13838 13905 51.9 51.9 0.0 Ifugao 15556 25240 15115 15462 I 25393 15207 15911 40.3 40.3 0.0 11ocos Norte 16024 16869 14263 14436 I 18725 15832 16538 21.2 20.9 0.3 lIocos Sur 16922 14940 14538 14624 ! 17288 16822 16934 32.6 32.6 0.0 Iloilo 14810 13376 14157 14031 I 14119 14943 14683 30.4 30.4 0.0 Isabela 14124 15060 13079 13467 I 15795 13717 14335 30.7 30.9 -0.2 Kalinga 15031 17246 15237 15415 i 16816 14857 15212 51.9 51.9 0.0 La Union 16372 16714 15338 15400 i 17769 16306 16595 32.6 32.7 -0.1 Laguna 17724 16577 15194 16053 I 18303 16776 17989 13.2 13.1 0.1 Lanao del Norte 15225 16213 13944 14410 17130 II' 14733 15585 52.2 52.0 0.2 Lanao del Sur 16567 16419 14983 15416 17645 16102 16454 58.5 58.8 -0.3 Leyte 13919 12923 12383 12468 ! 14427 13824 13991 47.3 47.2 0.1 Maguindanao 15556 14955 14576 14647 i 15883 15481 15527 69.3 69.3 0.0 Manila 20868 19621 19621 20868 ii' 20868 11.0 11.0 0.0 Marinduque 14041 13395 13395 14041 14041 50.6 50.6 0.0 Masbate 14248 16402 14542 14772 15820 14026 14319 59.5 59.7 -0.2 Misamis Occidental 14555 15859 13073 13846 i 16671 13742 15149 56.3 56.5 -0.2 Misamis Oriental 14787 15076 12875 13646 i 16337 13952 15423 37.5 37.9 -0.4 Mountain Province 16785 15269 16591 16499 i 15534 16879 16738 50.4 50.4 0.0 49 Negros Occidental 13975 13532 14468 14316 13210 14123 13750 42.0 42.1 -0.1 Official Official Preliminary PL PL for this study Poverty Headcount PL/a Ib ( of population) ----------------- (PhP per capita year) ------------ This Province Total Urban Rural Total Rural Total Official Difference Study Negros Oriental 12159 11777 11507 11532 12417 12133 12199 48.1 48.0 0.1 Cotabato 13315 15202 12671 13045 15517 12933 13182 34.6 34.6 0.0 Northern Samar 14275 20254 13932 14588 19819 13633 15398 61.1 61.3 -0.2 Nueva Ecija 17830 17572 15051 16201 19339 16564 17798 37.7 37.7 0.0 Nueva Vizcaya 14325 15870 12998 13457 16894 13836 14525 16.7 16.7 0.0 Occidental Mindoro 14219 15254 13431 14064 15422 13579 14555 57.0 56.8 0.2 Oriental Mindoro 16723 16735 15556 15777 17738 16489 16776 55.1 55.0 0.1 Palawan 13850 14018 12488 12766 15208 13548 14151 49.3 49.2 0.1 Pampanga 17243 17603 15261 16812 18054 15652 17376 10.8 10.8 0.0 Pangasinan 15656 15816 14879 15333 16149 15192 15660 35.0 35.0 0.0 Quezon 16125 17030 14854 15137 18142 15824 16145 47.7 47.9 -0.2 Quirino 14665 17451 13950 14924 17148 13708 14706 22.4 22.3 0.1 Rizal 17464 16552 16000 16496 17523 16939 17469 8.9 9.1 -0.2 Romblon 13832 14378 12162 12491 15922 13468 13940 51.7 51.4 0.3 Samar (Western) 13869 14168 13096 13191 14896 13769 13832 47.6 47.7 -0.1 Siquijor 12733 11226 11226 12733 12733 21.5 21.5 0.0 Sorsogon 15687 19056 13572 14631 20431 14552 16400 55.3 55.3 0.0 South Cotabato 15431 15530 13235 14136 16953 14447 15645 37.3 37.3 0.0 Southern Leyte 13998 12886 12886 13998 13998 36.0 36.0 0.0 Sultan Kudarat 13036 15934 12270 12990 15990 12313 13057 47.4 47.3 0.1 Sulu 15651 16525 13207 14011 18459 14753 15344 52.2 52.3 -0.1 Surigao del Norte 16961 17865 14451 15183 19957 16143 17345 60.2 60.0 0.2 Surigao del Sur 15264 16795 13533 14243 17999 14503 15907 55.1 55.1 0.0 Tarlac 16463 18198 14580 15810 18950 15182 16647 27.6 27.5 0.1 Tawi-tawi 14765 16473 13403 13803 17621 14337 14887 78.2 78.0 0.2 Zambales 16685 15904 13649 14746 17995 15444 16818 28.9 28.8 0.1 Zamboanga del Norte 13947 15108 12992 13183 15984 13745 14128 67.5 67.7 -0.2 Zamboanga del Sur 12741 16309 12851 13364 ! 15549 12252 13251 33.8 33.8 0.0 NCR-2nd Dist. 20085 19041 19041 20085 I' 20085 9.5 9.5 0.0 NCR-3rd Dist. 20908 18567 18567 20908 20908 12.8 12.8 0.0 NCR-4th Dist. 20582 19523 19523 i 20582 20582 9.2 9.2 0.0 Aurora 16275 15761 15319 15358 II 16702 16234 16370 36.8 32.8 4.0 Biliran 12028 12015 12501 12409 11646 12117 11836 42.2 41.8 0.4 Guimaras 14811 15837 15425 15445 I 15187 14792 14879 39.6 38.3 1.3 Sarangani 13746 16099 13044 13483 I 16413 13298 13799 52.0 52.0 0.0 Apayao 17837 17482 15712 157121' 19846 17837 18051 63.1 62.6 0.5 Compostela Valley 15822 14787 13812 14019 16689 15588 15835 47.1 47.4 -0.3 Zamboanga Sibugay 12188 14731 11911 12083 I 14859 12015 12690 40.5 40.5 0.0 Isabela City 14115 17108 12148 12654 i 19083 13551 14782 51.9 51.4 0.5 Cotabato City 17335 15649 15649 I 17335 17335 44.1 43.0 1.1 I Philippines 16058 16925 13753 15317 I 17991 14474 16208 32.9 32.9 0.0 Sources: la NSCB (2008); Ib NSCB (2007) 50 Table A4: Official Poverty Estimates, 2000-2006 Region! Annual Per Capita Poverty Threshold Poverty Incidence Magnitude Province (in Pesos) Among the Population (%) Rank of Poor Populatiun 2000 2003 2006 2000 2003 2006 2006 2000 2003 2006 PHILIPPINES 11,458 12,309 15,057 33.0 30.0 32.9 25,472,782 23,836,104 27,616,888 NCR 15,722 16,737 20,566 7.8. 6.9 10.4 17 860,934 742,549 1,156,313 1st District (Manila) 16,218 17,223 20,868 7.3 5.5 11.0 79 120,663 90,446 198,391 2nd District (E. Manila) 15,727 16,715 20,085 6.0 8.6 9.5 82 229,301 202,197 365,169 3Td District (CAMANAV A) 15,090 16,298 20,908 12.1 6.9 12.8 76 304.583 261,328 325,964 4th District (S. Manila) 16,359 17.137 20,582 6.7 6.5 9.2 83 206.387 188,578 266,789 Region I 12,1i87 13,281 15,956 35.3 30.2 32.7 13 1,452,222 1,262,799 1,464,245 Iloco. Norte' 13,143 12,893 16,024 22.8 24.6 21.2 72 115,116 120,945 112,835 lIocosSur 13,515 12,824 16,922 35.2 28.4 32.6 60 194,881 154,922 184,397 La Union 12,978 . 13,356 16,372 ; 38.4 30.2 32.6 59 253,382 198,307 229,739 Pangasinan 12,363 13,412 15,656 37.0 31.7 35.0 56 888,844 788.625 937,274 Region II 11,128 11,417 13,791 30.4 24.5 25.5 14 821,294 659,666 721,036 Batanes 15,264 12,279 14,970 18.1 9.0 - 2.535 1,459 - Cagayan 10,209 10,320 12,928 27.0 21.4 23.1 69 252,930 196,014 227,454 Isabel_ 11,616 11,808 14,124 34.6 30.1 30.7 ; 62 424,580 372,429 396,608 1\ueva Vizcaya* 1I,611 11,880 14,325 22.2 12.1 16.7 73 81,696 44.502 61,153 Quirino' 10,713 12,463 14,665 38.2 29.2 22.4 70 59,555 45,262 35,821 Region III 13,760 14,378 17,298 21.4 17.S 20.7 16 1,695,227 1,535,784 1,914,590 Aurora* 11,405 12,898 16,275 33.6 39.0 36.8 54 59,985 66,417 66,701 Bataan* 12,434 13,607 15,538 12.1 13.9 10.5 81 68,659 79,841 62,022 BuIacan 13,882 15,027 17,768 7.5 12.3 13.4 74 147,812 307,762 358,012 Nueva Ecija 14,750 14,394 17,830 32.7 27.1 37.7 50 532,961 484,106 662,742 Pampanga 14,698 15,148 17,243 18.2 14.7 10.8 80 331,739 289,106 234,820 Tarloe 12,578 13,866 16,463 33.6 18.4 27.6 67 360.109 208,104 328,428 Zambales 12,733 12.754 16,685 28.2 15.5 28.9 65 193,962 100,447 201,864 Region IV-A 13,670 14,720 17,761 19.1 18.4 20.9 15 1,697,033 1,899,827 2,210,756 Batangas 15,192 15,957 19,616 25.8 30.4 30.7 61 440,603 602,557 618,297 Cavite 14,742 16,150 18,718 13.0 12.5 11.2 77 244,712 300,636 287,292 Laguna 12,937 13,921 17,724 10.8 10.6 13.2 75 207,184 236,460 297,648 Quezon 12,501 13,349 16,125 39.3 39.8 47.7 32 668,237 660,224 829,802 Ri7.al 13,676 13,903 17,464 8.1 4.9 8.9 84 136,296 99,950 177,718 Region IV-B 12,013 12,402 14,800 45.3 48.1 52.7 2 1,032,123 1,163,867 1,400,417 Marinduque 11,553 11,781 14,041 52.7 47.4 50.6 25 113,553 101,271 118,365 Occidental Mindoro 11,745 12,522 14,219 46,2 50.5 57.0 ]0 176,790 203,741 263,965 Oriental Mindoro 13,510 13,813 16,723 48.9 44.3 55.1 15 340,690 321,441 422,103 Palawan 11,163 11.591 13,850 31.9 52.0 49.3 29 230,174 419,389 445,097 Romblon 10,758 11,769 13,832 64.4 43.6 51.7 23 170,917 118,026 150,889 Region V 11,375 12,379 15,015 52.6 48.5 51.1 4 2,540,660 2,332,719 2,643,799 A1bay 12,144 12,915 16,128 48.4 42.7 46.2 39 553,629 464,510 552,881 Camarine. Norte 11.505 12,727 14,854 57.3 55.5 49.3 28 301,147 269,604 256,708 Camarines Sur 11,054 11,873 14,634 47,2 47.1 49.9 27 765,373 750,674 846,030 Catanduanes 11,587 11,815 13,654 51.9 36.8 46.8 38 117,740 76.609 105,075 M_sbate 11,019 12,504 14,248 70.2 63.4 59.5 8 482,818 470,670 487.672 Sorsogon 11,146 12,452 15,687 51.4 43.5 55.3 13 319,952 300,652 395,434 Region VI 11,314 12,291 14,405 44.5 39.2 38.6 9 2,773,352 2,374,772 2,491,535 Aldan 11,527 11,980 15,150 42.9 41.8 52.0 19 190,470 173,340 242,249 Antique 10,938 11,377 14,650 45.9 48.9 51.6 24 208,169 232,602 255,321 Capiz 10.536 11,298 14,242 47.2 29.2 30.2 64 328,635 194,558 205.168 Guimaras 10,759 11,694 14,811 28.3 49.5 39.6 49 37,838 66,944 59.001 Iloilo 12,122 13,221 14,810 36.9 38.6 3004 63 695,280 708,899 609,625 Negros Occidental 11,126 12,131 13,975 50.2 39.5 42.0 45 1,312,961 998,429 1,120,171 51 Regionl Annual Per Capita Poverty Threshold Poverty Incidence Magnitude Province (in Pesos) Among the Population (%) Rank of Poor Popnlation 2000 2003 2006 2000 2003 2006 2006 2000 2003 2006 Ree:ion VII 9,659 9,805 13,390 36.2 28.3 35.4 11 2,016,910 1,652,316 2,213,167 Bohol 9,762 10,032 13,610 56,7 34,9 46,9 37 590,926 375,277 532,711 Cebu 9,914 10,222 13.960 29.5 2\,\ 28,2 66 973,490 746,100 1,077,492 Negros Oriental 8,981 9,017 12,159 37,1 43.4 48,1 31 427,509 502,825 584,695 Siquijor* 8,892 9,767 12,733 32,7 37.4 21.5 71 24,984 28,114 18,269 Ree.ion VIII 9530 10804 13,974 45.1 43.0 48.5 5 1,649,582 1,619731 1,947,323 Biliran 9,858 11,144 12,028 43.4 55,6 42,2 43 58,135 77,193 66,781 Eastern Samar 9,108 11,025 13,873 55,6 4\,\ 51.9 22 203,104 159,184 224,755 Leyte 9,447 10,600 13,919 40,8 42.3 47.3 35 680,536 692,391 846,526 Northern Samar 8,898 9,945 14,275 49,5 40.8 6\'\ 5 240,228 215.859 339,232 Southern Leyte 9.459 10,668 13,998 35.1 41.2 36.0 55 116,738 147,484 131,172 Western Samar 10,338 11,675 13,869 52.4 45,9 47,6 33 350,841 327,620 338,857 Region IXb 9,128 10,407 13,219 I 44.8 49.2 45.3 6 1,257210 1.427,722 1404,098 Zamboanga del Norte 9,417 10,871 13,947 53.3 68,5 67.5 3 433,091 573,506 614,876 Zamboanga del Sure 8,975 10,310 12,741 414 38.8 33.8 58 824,119 571,833 527,005 Zamboanga Sibugay d 9,580 12,188 d 50.3 40.5 46 d 256,705 223,429 Isabela City' e 10,429 14,115 e 33.5 51.9 21 e 25,677 38,788 Region X 10,509 11,605 14,199 43,8 44,0 43,1 I 7 1,582,225 1567,963 1663,283 Bukidnon 9,201 11,083 12,186 41.0 42,9 37,2 53 449,647 460,292 438,293 Camiguin 12.155 12,109 16,145 57,0 39,7 42,1 44 41,465 29,420 35,265 Lanao del Norte 11,296 12,103 15,225 54,2 54,0 52,2 17 426,347 404,674 416532 Misamis Occidental 10,184 11,711 14,555 53.0 54.3 56.3 11 260,764 263,398 294,806 ~fisamis Oriental 11,176 11,594 14,787 34.8 34.8 37.5 51 404,002 410,180 478,387 Region XI 10,278 11~~99 14,942 33~~ 34.7 36,6 10 1,231,277 1.346,269 1450,542 Davao del Nortef 10,566 11,833 15,753 46.5 36.8 44,8 40 642,900 287,572 353,366 Davao del Sur 9,987 11,470 14,452 21.9 28.9 27.4 68 412,442 591,886 568,808 Davao Oriental 9,906 10,580 13,741 40.8 47,9 48.8 30 175,934 2\0,903 227,287 Compostela Valley g 11,422 15,822 g 419 47.1 36 g 255,909 301,081 Region XII 10458 11,328 14,225 46,8 38.4 40.8 8 1,595474 1.319,563 1,482130 North Cotabato 9,990 10,972 13.315 50,1 32,1 34,6 57 511,353 317,424 350,178 Saranggani 10,419 10,846 13,746 52.3 510 52.0 18 220,079 241,641 271,713 South Cotabato 10.686 11,741 15,431 39,1 31.8 37,3 52 469,874 380,204 469,717 Sultan Kudarat 10,544 10,870 13,036 56.4 49.4 47.4 34 344,172 296,215 309,887 Cotabato City' 12,670 \3,805 17,335 31.3 48,3 44,1 41 49,997 84,079 80,636 CAR 13,071 14,033 16810 37.7 32.2 34.5 12 537975 445'036 506823 Abra 13,426 14,654 17,900 57.6 50.2 60.7 6 113,326 100,013 128,614 Apayao 11,368 12,256 17,837 34.1 23,2 63,1 4 28,770 22,815 67,907 Beoguet 14,014 14,447 17,483 18.8 15,0 ILl 78 122,178 89,132 71,190 lfugao 11,809 13,148 15,556 64,1 35.5 40,3 47 113,719 60,226 69,605 Kalinga 11,652 13,284 15,031 45.7 52,0 51.9 20 83,844 93,693 94,995 Mt. Province 15,122 14,855 16,785 57.1 57.0 50.4 26 76,137 79,157 74,512 ARMMh 12,199 12.733 15,533 60.0 52.S 61.8 1 I 1,652,890 1,373620 1778,262 Basilan*h 9,509 10,987 13,255 39,1 42.0 43,7 42 123,825 101,504 118,183 Lanao del Sur 13.892 13,702 16,567 61.6 44.6 58,5 9 432,307 301,215 442.338 Maguindanao 11,906 12,322 15,556 65.1 68.1 69.3 2 536,479 527.225 596,464 Sulu 11,672 13,473 15,651 63.3 53.5 52.2 16 397,119 315,635 310,140 Tawi-tawi 12,003 11,707 14,765 57.2 40.2 78,2 1 163,160 128,041 311,137 Caraga 10,903 11,996 15,249 51.2 54.0 52.6 3 1,076,395 1.11 1,901 1,168569 Agusan del Norte 10,933 11,460 13,986 46.3 40,0 40.0 48 259,475 219,514 236,297 AI,,'llsan del Sur 11,017 12,150 14544 60.1 60.3 56,2 12 359,215 337,889 334,069 Surigao Del Norte 11,160 12,998 16,961 51.3 59,8 60,2 7 232,065 277,763 309,540 Surigao Del Sur 10,421 11,227 15,264 45,8 57.1 55.1 14 225,640 276,735 288,664 Notes: a - No CVs were computed SInce only one sample household was claSSIfied as poor In 2003 and none In 2006; b - 2000 estimates do not include Isabela City; c - 2000 estimates still include Zamboanga Sibugay; d - No separate estimate yet; still included in Zamboanga del Sur; e - No separate estimate yet; still included in Basilan; f - 2000 estimates include Compostela Valley; g - No separate estimate yet; still included in Davao del Norte; h - 2000 estimates include Isabela City. I. Zamboanga Sihugay (Region IX) and Compostela Valley (Region XI) are new provinces created under EO 36 and EO 103, 2, Isabela City (Region IX) and Cotabato City (Region XII) have been separated from their respective mother provinces· BasiIan and Maguindanao (both ARMM) under the present regional configuration. * Coefficient of Variation (CV) of 2006 poverty incidence is greater than 20 percent Source: NSCB (2008) 52 CHAPTER II - ANNEX II SIMULATING THE IMPACT OF INFLATION ON POVERTY The recent price hike in food prices hurt net purchasers of food across the globe. Although the Philippine population spends a large amount of their income on food, a majority of the poor are in the agricultural sector producing food, so the net impact of the food price increase on poverty is not obvious. It is, therefore, useful to quantify the impact of price changes on poverty in the Philippines. To simulate the effect of inflation on poverty, a direct impact analysis can be undertaken using information from the latest FIES. The FIES, conducted every three years by the NSO, contains household level data on family income sources, such as income from agricultural activities and transportation, as well as detailed expenditure items, including spending on all food items, rice, and fuel. Consumer price indices (CPI), farmgate and retail prices of rice at the regional level are also used to capture the spatial heterogeneity in inflation. Regional CPI data, collected monthly by the NSO, have separate indices for food, beverage, and tobacco and fuel, light, and water. Regional farm gate and retail prices of rice are available monthly from the Bureau of Agricultural Statistics. The price changes calculated from these data used in the simulation are reported in Table Bl. Table B1: Price Increase between 2007 and 2008 for Select Goods Q Food,beverage Light, Fuel Regular Region Overall & tobacco & Water Milled Rice Ilocos 11.4 16.3 9.0 65.1 Cagayan Valley 14.2 17.7 19.6 66.7 Central Luzon 14.3 17.8 14.2 60.7 Calabarzon 9.3 14.2 2.6 65.8 Mimaropa 14.2 18.9 15.7 55.6 Bicol 13.7 18.6 10.2 60.8 Western Visayas 14.4 20.4 16.5 44.2 Central Visayas 14.9 20.1 12.4 60.0 Eastern Visayas 20.0 27.0 19.0 61.8 Zamboanga Peninsula 20.3 26.0 18.6 52.4 Northern Mindanao 15.3 20.7 12.0 65.5 Davao 14.9 21.8 7.1 53.7 SOCCSKSARGEN 16.3 21.1 15.4 34.6 NCR 8.6 13.1 (4.5) 67.7 CAR 11.9 18.3 5.4 66.7 ARMM 17.2 21.8 23.5 56.7 22.2 30.6 17.0 59.3 Philippines 12.3 17.8 5.5 58.3 Note: Year-on-year inflation based on prices in July. Sources: NSO and BAS. 53 In using the FIES 2006, income and spending patterns of households now are assumed to be as they were in 2006. The analysis also assumes that there is no substitution between goods, which implies that (i) households do not switch less expensive commodities with more expensive ones and (U) the composition of the menu for the poverty line remains unchanged. This is admittedly a strong assumption and so is used in assessing the· impact of shocks or reforms only in the short term, before economic agents are able to make adjustments and behavioral changes. In the face of limited data, this assumption could be useful for providing an idea of the upper bound of the impact of inflation on poverty. To simulate the impact of inflation on poverty, suppose that the poverty line is initially given by z. If the price of everything increased by x, then the amount of money one needs to maintain the consumption at the poverty line is z(l +x). Now suppose instead that only the price of food increased by x. Then, under the assumption of no substitution, the amount of money one needs after the increase in the food price will be zs(l+x)+z(l-s), where s is the share of food expenditure in the poverty line. In a similar manner, the poverty line after the increase of other goods, such as fuel, can be computed. When one sells food, one benefits from higher food prices. The net effect of price increases in specific commodities, say rice and fuel, will depend on how much people earn from and spend for them. Hence, it is also important to take into consideration this "income effect." Again, the substitution between production processes shall be ignored so that the agricultural output remains unaffected by higher food prices. Under this assumption, the new income level after the food price increase can be calculated. That is, suppose that initial income level is initially given by y. If the price of food increased by x, then the income one receives after the inflation is yt(1 +x)+y(1-t), where t is the share of income from selling food. Capturing the income effect is a challenge because the exact share of income from selling food, t, is not in the FIES. The analysis then used the share of income from agricultural activities (farming, fishing, and poultry-raising) and income from transportation to total income in the FIES to approximate the income effect of price increases in all food commodities and fuel, respectively. Table B2 shows the income and spending patterns of the population by income quintile in 2006. The bottom quintile got 30 percent of their earnings from agricultural entrepreneurial activities, which includes farming and gardening, fishing, and poultry-raising, where the upper two quintiles got less than 5 percent of their earnings. Meanwhile, food comprises 60 percent of the bottom quintile's total spending, and about one-third (21.4 percent) of their food expenditure is spent on rice. In general terms, Filipino households spend more on food than they earn from food production, which suggests that high food prices will have a negative impact on overall poverty. Transportation, both as a share of income and expenditure, is higher for the upper income groups so that increases in the price of fuel are expected to hurt them more. 54 Table B2. Income and Expenditure Shares by Quintile, 2006 Income Shares 0 Expenditure Shares 0 Income Quintile AgricuIture- Fuel, Light Transportation Food Transportation related & Water I-Lowest 31.3 1.9 63.1 6.7 3.1 2 22.1 3.0 57.4 7.1 4.1 3 12.7 3.5 5l.8 7.8 4.9 4 7.0 3.1 44.5 8.3 5.6 5 -Highest 3.3 2.2 31.3 7.5 6.5 Philippines 8.1 2.6 41.4 7.6 5.7 Source: World Bank estimates based on FlES 2006. Table B3 shows the simulated impact of inflation on poverty. The effect of inflation was examined both with and without the income effect of price increases (0 on all food commodities, (ii) on fuel only, and (iii) on both food and fuel. In all three scenarios, the effect of a 10 percent price increase was simulated as well as the actual price change between July 2007 and July 2008, when food inflation was highest. Table B3. Simulated Impact of Inflation on Poverty No Income-effect With Income-effect Actual Price Actual Price 10 Price 10 Price Increase Increase Increase Increase 2007-2008 (July) 2007-2008 (July) Price Increase in ALL FOOD Items Poverty Incidence ( of popUlation) 38.6 36.1 36.5 34.9 Increase in Magnitude (million) 4,774 2,681 3,027 1,717 Price Increase in FUEL only Poverty Incidence ( of population) 33.4 33.3 33.2 33.2 Increase in Magnitude (million) 0.398 0.359 0.255 0.228 Price Increase in FOOD & FUEL Poverty Incidence ( of population) 39.0 36.5 36.8 35.2 Increase in Magnitude (million) 5,083 2,989 3,261 1,922 Increase in Income gap (percentage point) 3.0 1.3 1.6 -1.0 Note: Official poverty estimate in 2006 is 32.9 percent of the population, equivalent to 27.6 million poor people. The income gap of the poor is estimated at 30.5 percent of the poverty line. Source: World Bank estimates based on the FIES 2006 data. The increase in food prices has a large effect on poverty because the price change was particularly large between 2007 and 2008 (18 percent). If the income effect is not taken into account, poverty would have increased from 32.9 percent of the popUlation to 38.6 percent, or an additional 4.8 million poor people (Table B3). The increase in the price of rice alone would account for the additional 3.9 million people who will fall into poverty. The effects of the increase in energy prices alone appear relatively limited, with a 10 percent increase in fuel prices raising poverty to 33.3 percent, or an additional 360,000 poor people. However, its compounding effect on food prices would mean an additional 5.1 million people who will fall into poverty. The 55 extent of the impact is shown by the increase in the income gap of the poor: from 30.5 percent at baseline up to 33.5 percent after the food and fuel crises. If the offsetting effects of inflation are taken into account, the impact on poverty is reduced. The increase in poverty resulting from the actual increase in food prices is higher among households in the urban areas compared to those in rural areas and also higher among non-agricultural households compared to agricultural households (Table B4). Accounting for possible income effects, the income gap of the poor is estimated to increase by 1.6 percentage points due to the combined effects of the food and fuel price increases. Table B4. Simulated Impact ofInflation among Various Population Groups Poverty Incidence after Actual Price Official Increases in Poverty Population Incidence All Food Fuel Food & Fuel Group Items Prices Only Combined ------------ ( of population) - - - - - - - - - By Agricultural Indicator Agricultural HH 65.4 66.8 66.1 67.5 Non-Agricultural HH 21.9 26.3 22.1 26.4 By Urban/Rural Area Urban 19.5 23.2 19.7 23.3 Rural 45.9 49.5 46.3 49.9 Philippines 32.9 36.5 33.2 36.S Source: World Bank estimates based on the FlES 2006 data. Recall, however, that these calculations are made under the assumption of no substitution. If substitution is taken into account, the negative effect on poverty incidence would have been smaller. People can, for example, switch from expensive food to less expensive food without necessarily reducing the nutritional content. However, this does not mean that the current rapid inflation is of no concern. As this exercise shows, the rapid increase in food prices would severely hurt the net purchasers of food, which means that the urban poor are the most at-risk from rising food prices. 56 CHAPTER III THE SECTORAL AND REGIONAL PATTERNS OF GROWTH IN THE PHILIPPINES A. Introduction 3.1 The link between growth and poverty reduction appears to be relatively weak in the Philippines. As discussed in Chapter II, poverty has stopped declining in the Philippines over the past several years, even though economic growth has accelerated. The preceding chapter offered several potential explanations for the weak link between poverty reduction and growth, including the sector and spatial composition of growth, the evolution of relative output and factor prices, and changes in public spending and policies. 3.2 This chapter analyzes the extent to which sectoral and regional patterns of growth in the Philippines may have reduced the impact of growth on poverty reduction in recent years. The analysis reveals that sectors that are intensive in the use of unskilled labor have grown more slowly than sectors that do not use this factor intensively. As a result, while overall positive GDP growth may have tended to reduce poverty, the uneven pattern of growth across sectors worked in the opposite direction, resulting in a slight increase in poverty since 2000. The analysis also shows that the regional differences in GDP growth do not appear to have been a major contributing factor to the increase in poverty. The chapter ends by highlighting several key issues emerging from this analysis, which are then examined further in the subsequent chapters. B. The Link between Sectoral Growth Patterns and Poverty Reduction 3.3 While it is possible for growth to raise poverty if income distribution worsens enough at the same time, most cross-country evidence suggests that economic growth is good for the poor. The contribution of each sector to overall GDP growth depends on the sector's size and its rate of growth. It follows, therefore, that a sector's impact on poverty reduction varies with the magnitude of that sector's growth performance. In addition, Loayza and Raddatz (2006) has shown that if a country's growth performance is biased toward unskilled labor-intensive growth, it could gain from a further reduction of poverty beyond the direct impact exerted through the average GDP growth rate, since the poor are found mainly among the unskilled. 3.4 The sectoral composition of growth can have important implications for poverty reduction. If labor markets are strongly segmented across sectors, growth will have a larger impact on poverty alleviation if it occurs in the sectors where most poor people work. However, even without labor market segmentation, uneven sectoral growth may result in changes in relative prices (both in product and factor markets) that can either favor or hurt the poor. For example, the poor would benefit if rapid growth in agriculture led to a reduction in the relative prices of foods that are consumed more intensively by the poor. The poor also stand to gain the most from increased wages for unskilled labor since they tend to be less endowed with skills. A growth pattern that is biased toward sectors that are most intensive in the use of unskilled labor would put upward pressure on the unskilled wage economy-wide, raising the incomes of the poor 57 and thereby reducing poverty more than would be the case if all sectors were growing at the same uniform rate as GDP (Box 3.1). Box 3.1: Link between the Sector Composition of Growth, the Unskilled Labor Wage, and Poverty Using a simple multi-sectoral general equilibrium model, Loayza and Raddatz (2006) obtain the following expression for the evolution of real wages associated with the income of the poor: A OJ = L 2 Sj Yj+ -- A (& -1)L (I -1} Y ~ 2 i A i' where 0) is the real wage, Ii is sector j's share of total 1=1 & H Si unskilled employment, Si is the same sector's share of total value added, e 0 is the elasticity of substitution across sectors in the production of GDP, y is the sector's value added in per capita terms, and hats denote rates of growth of the underlying variable. Assuming that the change in the headcount index, h, or any other poverty measure is a linear function of changes in unskilled wages, this expression translates into a relationship between the composition of growth and poverty alleviation: h== 80 ~ (tSI + I-I .;J +8 (tUI 2 I-I Sj)' ;J, where the different 9's represent parameters that are estimated using cross-country data. The fIrst sum on the right-hand side of the equation is the rate of growth of total value-added, the second term, which multiplies the share-weighted growth of each sector by the ratio of the sector's share in employment to the sector's share in value added, corresponds to labor-intensive growth. The coefficient that multiplies the share-weighted sectoral growth in the second term, which is the ratio of a sector's share of unskilled employment to its share in total GDP minus one, is a measure of the unskilled labor intensity of the sector. Assuming that the elasticity of substitution in demand across sectors is greater than one (or negative which is possible when there are more than two sectors), growth in an unskilled labor-intensive sector will contribute to an increase in wages and a decline in poverty above and beyond its contribution to overall GDP growth. Source: Loayza and Raddatz (2006) and Raddatz (2008). C. The Sectoral Pattern of Growth and Employment 3.5 The Philippines' growth in recent years has been driven primarily by manufacturing, wholesale and retail trade, and transport and communications. A sector's potential impact on wages and poverty reduction depends on its growth rate and size33 -which together determine its contribution to overall GDP growth-as well as on its unskilled labor intensity ratio. The manufacturing, wholesale and retail trade, and transport and communications sectors have accounted for a sizable share of GDP in recent years, with their combined value- added during 1997-07 averaging just under 45 percent of GDP. The manufacturing and wholesale and retail trade sectors have each contributed to around 20 percent of GDP growth over the last decade, while the transport and communications sector contributed to around 13 percent of total GDP growth (Figure 3.1). Of these sectors, only retail and wholesale trade 33 There was a weak negative correlation between size and growth of sectors during the period 1997-07-that is, larger sectors grew less, which is consistent with standard convergence mechanisms. However, these differences did not materially affect the relative contribution of the different sectors to overall growth. 58 exhibits a positive unskilled labor-intensity ratio, meaning that it is intensive in the employment of unskilled labor. 34 Figure 3.1: Sector Contributions to GDP and Employment Growth, 1997-2006 Sector Contribution to GDP Growth Sector Contribution to Employment Growth Contribution to Contribution to growth employment % growth 25 % Traoo 20 Ag 1~ Trans otI! ser 10 15 Fin Trans 10 5 um Realest Cons () o 5 10 15 20 o 10 20 30 40 50 Share of Share of GDP employment % % Source: Raddatz (2008) 3.6 Agriculture has made a much smaller contribution to overall GDP growth than would be expected given its fairly large share ofGDP. Agriculture comprises about 15 percent of GDP. With growth of 3.2 percent per year, it has performed less dynamically than manufacturing (which grew on average by 3.6 percent), trade (5.5 percent), and transport and communications (8.2 percent). As a result, the agriculture sector's overall contribution to total GDP growth was around 13 percent-similar to the contribution of the transport and communications sector, which accounts for a much smaller share ofGDP. 3.7 However, the agricultural sector has been one ofthe most important contributors to employment growth, particularly for unskilled labor. Agriculture has a large positive unskilled labor intensity ratio and has made the largest contribution to unskilled labor employment compared to other sectors (Figure 3.2). In contrast, manufacturing--one of the two sectors that contributed most to GDP growth over the last decade-contributed little to employment growth. Specifically, the manufacturing sector contributed around 20 percent of total GDP growth during 1997-07, growing on average by 3.5 percent per year (and 4.6 percent per year since 2003). However, its contribution to employment creation during this period wa,s only 4.5 percent of total growth in employment, less than 2 percent of total growth in unskilled employment, and only 6 percent of total growth in skilled employment. 34 The unskilled labor intensity ratio is defmed as the ratio of a sector's share of unskilled employment to its share in total GDP, minus one. In other words, ri,t = (lj,{ ~ 1J, where I i,~ is a sector's share of total unskilled Sf,! employment, and s 1.1 is the same sector's share of total value added. 59 Figure 3.2: Sector Contribution to Skilled and Unskilled Labor Employment and Employment Growth, 1997-2006 Percent 50 _ contribution to employment of unskilled labor _ contribution to employment growth of skilled labor 40 30 20 Ag Man Util Cons Trade Trans Fin Real estataJth ser Source: Raddatz (2008) 3.8 Sector employment growth and labor movements are weakly correlated with labor productivity. Sectors that have grown faster in absolute tenus have experienced a larger expansion of employment, indicating that at least part of the sectoral growth has been associated with employment generation. However, the relationship between employment changes and output growth per worker-a rough measure of labor productivity-is much weaker. That is, labor does not appear to have moved into sectors where the average labor productivity has been increasing faster (Figure 3.3), which suggests a misallocation of resources. This lack of overall correlation seems to be driven mainly by the flow of skilled workers, which appears to be unrelated to the growth of value added per worker (Figure 3.4). Figure 3.3: Sector Employment Growth and Labor Productivi ,1997-2006 Overall real per worke r VA growth (%) Mining U tiliti ·· Trans Man 2 Other ter Fill Agr Real est Trade ·1 Const ·2+---~~--~----~----~--~----~----,----. ·2 ·1 6 Overall employment growth % Source: Raddatz (2008). 60 Figure 3.4: Correlation between Labor Movements and Labor Productivity, 1997-2006 Relation with Unskilled Labor Relation with Skilled Labor Overall Overall real per real per worker ""rker VA growth VA growth (%) (%) 6 6 Mining 5 Mining Utilities Utilities Trans Trans Man vuoer ser Fin 2 Agr Real est Trade ·1 Const Const ·2+i----._--~--~----~--~----._--~ ·2 ·1 0 ·2 +------.-------.-------.------...-------, ·2 4 6 Unskilled employment growth Skilled employment growth (%) (%) Source: Raddatz (2008) D. Implications for Poverty Reduction 3.9 The sectoral pattern of growth in the Philippines has been biased against the unskilled labor-intensive sectors. Finance and mining, which have been the fastest-growing sectors in recent years (although they account for a small share of total GOP), exhibit the most negative unskilled labor intensity. Manufacturing, real estate, and utilities, which have also grown fairly rapidly, also exhibit negative unskilled labor intensity. Indeed, as shown in Figure 3.5, most sectors have a median unskilled labor intensity ratio below zero. Although the negative unskilled labor intensities of the manufacturing, real estate, and utilities sectors are smaller in absolute size than the high positive unskilled labor ratio of the agriculture sector, the three sectors combined have outweighed the contribution of the agricultural sector due to the latter's relatively weak growth performance. Figure 3.5, which also summarizes the evolution of per worker GOP growth and labor-intensive growth during 1997-2006,35 indicates that GOP growth per worker was positive in all but two years, while labor-intensive growth per worker (and in per capita terms) was negative during most of this period. Therefore, while overall GOP growth has tended to reduce poverty, the pattern of sectoral growth has been tilted against the employment of unskilled labor, tending to raise poverty. 35 Similar results are obtained with per capita growth rates. 61 Figure 3.5: Labor Intensity Index and Growtb Unskilled Labor Intensity Index by Sector GDP Growtb vs. Labor-Intensive Growtb ~,re'l Real estate Finance Trans Trade ~ j · ,illl 8 I> 4 ·· i Const Utilities Manu Mining A;Jr ·1,5 -to ·0,5 I 0,0 0,5 to 1.5 - ~ 2,0 2,5 -4 Unskilled Labor Intensily Note: e> 0, Hi '< 0 and H;" > 0, for i = Rand P. These second order assumptions simply say that the impact of an increase in per capita GDP on poverty reduction tends to diminish as the region becomes richer. With this simple model, the evolution of a country's GDP and overall poverty head count ratio are determined by the evolution of the regional per capita GDP levels, as follows: or, expressed in vector notation: (dH, dG)' = [A] (dG R , dG p )'. The matrix A in this expression is signed[~ ~]. and its determinant is given by 6. = OR*Op (H R' - Hp') > 0, since HR' > Hp' as long as G R > Gp. By inverting the previous expression, it is possible to determine what the regional growth patterns would have to look like in order for both the overall poverty headcount ratio and the overall per capita GDP level to be increasing. That is, (dG R , dG p )' = [Ar 1 (dH, dG)" where the inverse matrix, [Ar\ is now signed [~ ~].. From this expression, the only way that Hand G can both increase (i.e., dH > 0 and dG > 0) is if the rich region becomes richer and the poor region poorer (Le., dG R > 0 and dG p < 0). As discussed in the main text, such a pattern seems to have characterized the regional pattern of growth in the Philippines during 2000-06. 3.12 The uneven pattern of per capita growth across the regions may have contributed to the increase in poverty observed during 2003-06. Some regions have exhibited increased per capita income growth since 2000, while others have reported declines. It is conceivable, therefore, that overall poverty stopped declining during 2000-06 due to an uneven pattern of growth. When aggregate GDP growth is positive, however, this can only happen if the overall income distribution across regions worsens. That is, the pattern of growth would have to be such that the rich regions are becoming richer on average, while the poor regions are becoming poorer; see Box 3.2 for an analytical justification. Such a pattern is observed on the right hand side of Figure 3.7, which shows that the refcions with a higher per capita GDP in 2000 tended to exhibit faster GDP growth during 2000-06. 0 40 It turns out, however, that the correlation between the poverty headcount ratios and per capita GDP growth across regions is insignificant. That is, the regions with the highest poverty headcount ratios have generally not tended to exhibit slower growth than the regions with lower headcount ratios. This result is largely attributable to the statistical divergence problem identified earlier (paras. 2.14-2.19 in Chapter 2), given that the poverty headcount ratios are derived from the FIES household surveys and the per capita GDP figures are derived from the national accounts. 64 Figure 3.7: Poverty and Economic Performance in the Philippines, hy Region Richer regions exhibit less poverty.. ·.. and faster GOP growth. Iii Iii 0 0 40,000 1/1 40,000 1/1 CI! 0 0

::; 200 .2 If Coo -- w"" E '" .. .... 150 ~ 100 .., " .: 50 o N '" o en .... .. '" '" '" '" ........ .. o o .. o '" .... '" '" N N 83 5.3 The stagnation of manufacturing value-added in the Philippines is partly responsible for the sector's weak Figure 5.3: Evolution of Labor Productivity in contribution to employment growth. Since Industry,1980-2006 the mid-1980s, employment in industry 400.0 1 doubled in the Philippines but tripled in other i East Asian countries, as depicted on the right- hand side of Figure 5.2. Labor productivity in g 300.0 1 the Philippines' industry sector declined during ~ . 200.0 1'. the 1980s, and even though labor productivity ; ~ has been on an uptrend since the early 1990s, ~ 100.0 the rate of productivity growth has been slow compared to that observed in the other East Asian countries (Figure 5.3); just serving to 1980 1985 1990 1995 2000 2005 Year recover the productivity level reached in the early 1980s by 2008. This suggests that, on Source: World Bank, World Development Indicators 2008. average, Philippine industry has been investing less than its regional counterparts and/or has been slower to adopt new technologies. 5.4 Not all manufacturing sub-sectors have been equally stagnant. Electronics, transport, industrial chemicals, and machinery Figure 5.4: Value-added Growth in Manufacturing, have experienced rapid· value-added 1999-2005 growth since the East Asian crisis Percent (Figure 5.4). Together, these sectors 15 accounted for over 98 percent of the total growth in manufacturing value- added during 1999-05.47 By the same 5 -5 ,,.,dll token, employment growth has also -15 been dynamic in some subsectors: the FDI-driven boom in the electronics and -25 transport equipment industries, for example, raised employment by 18 percent over the period of 1999-05. -35 g:;; 1l-§l ~ * § 2 "8 E "8 ~ ~ ~ ~ ~ ~ ~ !l ~~ ~s; § l ~o~ ~ ro LL~ ~ E]l £ E ~ -g E616 <5 U ~ C' ~ ~ ~~o~ ~o~~ 0g Output growth in capital-intensive ~ I- -5 ro ~.<:t mach Other chern Paper I!lnng "'f'IT Pet Ref -50 -100 + - - , . - - - - , - - , - - , - - - , - - - , - - , - - , . - - - - , - - , - - - , -80 -60 -40 -20 0 20 40 60 80 100 120 140 Growth in value added 1999-05 % Source: World Bank calculations based on NSCB, Philippines Statistical Yearbook, various years. 5.7 Total factor productivity growth responds to competition. The largest increases in TFP levels within manufacturing since 1999 have occurred in the sub-sectors that faced greater competition. The degree of competition is found to be among the most important factors influencing the productivity of firms. Competition forces the less efficient firms that use more resources than required under existing technologies to restructure or exit the market, thus freeing up resources to be reallocated to more productive activities. Competition can also affect dynamic efficiency (i.e. efficiency under conditions of technical change) by influencing the rate of innovative activity and the rate of technological diffusion. 48 A good measure of the degree of competition in a sector is the markup of prices over costS.49 Sectors that had lower markups in 1999--or faced greater competition-have shown greater increases in TFP levels during 1999- 05, although the relationship is not very strong (Figure 5.7). 48 Some theoretical work posits that innovation is fostered by high concentration or market power and large finn size because of economies of scale and scope, and because fInns with greater market power can appropriate the returns from innovation more easily, giving them more incentives to innovate. However, much empirical work suggests that a low degree of competition (high concentration) is not conducive to economic activity (see, for example, OECD 1996). 49 Distinguishing markups from scale economies is difficult. In general, the ratio of revenues to costs is equal to the ratio of markups to scale economies. Throughout this chapter, the crude identifying assumption of a constant returns production function is imposed, since comparable sectoral data for a long period that would allow for an estimate is not available. It should also be noted that the level of markups consistent with competitive forces may differ significantly across sectors and industries. Some market structures may be compatible with both a positive mark-up and strong competition. In particular, in industries/sectors where goods are differentiated or technologically sophisticated, higher markups are likely. . 86 Figure 5.7: Markups and Total Factor Productivity Markups in 1999 and 2005 Relationship between Markups and TFP IIJBrkups Degree 2005 of markup in 1999 1.7 1.7 1.6 1.6 Cement Beverages 1.5 1.5 1.4 er 1.4 Other t3 Beverages 1,3 Leather 1.2 Cement t2 1.1 1.1 to 1.0 0.9 -100 -50 o 50 100 150 200 to 1.1 12 1.3 1.4 1.5 1.6 1.7 TFP growth 1999-2005 IvIarkups 1999 % Source: World Bank calculations based on NSCB, Philippines Statistical Yearbook, various years. 5.8 While the degree of competition has increased in recent years, the overall degree of competition in the Philippines' manufacturing sector remains fairly low. Markups have declined since 1999 in almost every sector, suggesting an across-the-board increase in market competitiveness. However, on average, the extent of markups in the Philippines' manufacturing sector remains high compared to other countries (see, for example, Aldaba 2005). Furthermore, the fact that the relationship between the markup level and TFP changes is relatively weak may indicate that the degree of competition is still not sufficiently high in most sectors to make a significant difference to productivity. 5.9 TFP levels and investment in manufacturing are also found to be significantly affected by other investment climate factors. An analysis based on firm-level data in the Philippines (ADB and World Bank, 2005) found that both TFP levels and investment depend on several other important aspects of the investment climate, as described in Box 5.1 below. In particular, as Table 5.1 shows, the analysis found that firms' productivity and investment decisions in the manufacturing sector are significantly affected by: · Governance-related factors. In particular, reductions in bribes (measured by expenses paid in administrative bribes as a percentage of total sales) are found to be associated with higher productivity and a higher likelihood of investing. · Security-related factors. Reductions in civil unrest, crime, and disorder are found to be associated with higher productivity. · Infrastructure quality. The number of days of power outages significantly reduces TFP and the probability of investing. · Labor issues. Labor strikes adversely affect firms' decision to invest in physical assets, while firms that are over-staffed (i.e. constrained by labor regulations from rationalizing employment) tend to be less productive than firms that are not over-staffed. 87 Box 5.1: Firms' Perceptions of Major Investment Climate Constraints in the Philippines Macroeconomic instability and weak governance are perceived as major business obstacles by the largest number of ftrms. Data from the 2005 investment climate survey revealed that ftrms were most concerned about macroeconomic instability and governance. The concerns about macroeconomic instability likely reflected the high public debt and fiscal deficit, both of which have declined progressively since the survey was undertaken. Among the other top six obstacles cited, three dealt with weak governance: corruption, regulatory uncertainty, and security. Corruption was seen as a "major" or "severe" obstacle to more than a third of the firms surveyed, and regulatory uncertainty to a quarter of the ftrms surveyed. Firm Perceptions Of Severe Constraints Cost as percentage of sales (% affirms finding issue a major or se\ere contrsint) Telecorr'l"'r'l..lnicaoons 300 VVorker skills and education Access to fine:ncing 1 :. Crime and security CI Unreliable infrastructur.: LiCensing and permits I [J Philippines 25.0 j I · Bribes I Access to Land Transportation 20.0 WContract enforcement 'I I!I Regulatory cost .....--J , Cost of Financing Anti-corrpetitive practices " ~ 15.0 Labor regulations Tax administration ... Q-ime, theft and disorder 10.0 Regulatory policy uncertainty Tax rates 5.0 Bectrk:.ity CorruptiOn I\Aa;croeconomc instability 0.0 Philippines indOllesia China Brazil o 10 20 30 Compared to fIrms in other countries of the region such as China and Indonesia, ftrms in the Philippines are more concerned about their infrastructure and labor regulations. Firms in the Philippines suffer higher losses as a share of sales from various unproductive expenses related to these above-mentioned factors. For example: · Losses caused by weaknesses in infrastructure service delivery and by delays in shipment costs are three times higher for firms in the Philippines than those in China. High levels of congestion, the poor condition of large parts of the road network, and inadequate connectivity reduce the efficiency of the road network in promoting growth. The power sector also faces problems with inadequacy. · Concerns about crime rate and security are higher in the Philippines and are reflected in higher losses from crime and higher firm expenditures on security compared to Indonesia and China. This may be correlated with Philippines' lower policing expenditures compared to Malaysia, Thailand, and Korea. · Firms in the Philippines also have less flexibility in staff reshuffling amidst excess supply of labor due to government overregulation of the labor market. Dismissing a worker in the Philippines is harder than in some neighboring countries. This inflexibility makes it difficult for enterprises to make adjustments that are crucial to their survival and growth. Doing Business: Labor Regulation Indexes (Higher value means more restrictive) Ill! Difficulty Of Hiring · Rigidity Of Hours 88 Table 5.1: Correlates ofInvestment Climate with Manufacturing Firms' Performance Multivariate Regression Analysis a b c (In TFP) Pro (invest=l) Pro (invest=l) Corruption and Governance -0.015** -0.008** -0.025** Bribes (0.004) (0.002) (0.008) Security and crime -0.018* -0.007 -0.054 Civil unrest (0.010) (0.004) (0.055) Infrastructure -0.004** -0.002* -0.005* Length of power outage (0.001) (0.001) (0.003) Labor problems & regulation 0.000 -0.003** -0.012** Strikes (0.003) (0.001) (0.005) Inflexibility of labor regulations -0.051 0.009 0.038 (0.063) (0.030) (0.089) ** indicates coefficient is significant at 10 percent, * indicates coefficient is significant at 5 percent, and robust standard errors are given in parentheses. Note: TFP is derived as a residual series of a Cobb-Douglas production function estimate. (a) Estimation using OLS controlling for industry, location, year, and firm's attributes such as exporters, foreign owned, part of a larger holding company, managers' education, and over pessimism (see Kaufmann and Wei, 1999). (b) Linear probability model (LPM) regression controlling for variables in (a) and lagged of log assets to control for size. (c) Probit controlling for variables in (a) and lagged of log assets to control for size. (d) Z-statistics presented for standardized Probit coefficients. Source: World Bank (2005). 5.10 Among these factors, TFP is most affected by governance, especially bribes or administrative corruption. The propensity of firms to invest is most adversely affected by infrastructure problems, followed by governance/corruption factors. Table 5.2 presents standardized coefficients50 that allow for a comparison of the relative importance of each explanatory variable. 50 Standardizing the dependent and independent variables into variables with zero mean and a variance of unity allows a comparison across the coefficient of variables since the standardized coefficients represent the impact of a 1 standard deviation change in the explanatory variable on the dependent variable measured in teons of standard deviation. 89 Table 5.2: Relative Importance of Investment Climate Factors in Manufacturing Firms' Performance Standardized coefficients e a b c (In TFP) Pro (invest=l) Pro (invest=l) Corruption and Governance -0.067** -0.041 ** -0.033** Bribes (0.004) (0.012) -3.05 Security and crime -0.062* -0.009 -0.004 Civil unrest (0.018) (0.006) -0.98 Infrastructure -0.020** -0.026* -0.049* Length of power outage (0.011) (0.015) -1.69 Labor problems & regulation -0.003 -0.031** -0.048** Strikes (0.024) (0.0006) -2.52 ** indicates coefficient is significant at 10 percent, * indicates coefficient is significant at 5 percent, and robust standard errors are given in parentheses. Note: TFP is derived as a residual series of a Cobb-Douglas production function estimate. (a) Estimation using OLS controlling for industry, location, year, and firm's attributes such as exporters, foreign owned, part of a larger holding company, managers' education, and over pessimism (see Kaufmann and Wei, 1999). (b) Linear probability model (LPM) regression controlling for variables in (a) and lagged of log assets to control for size. (c) Probit controlling for variables in (a) and lagged of log assets to control for size. (d) Z-statistics presented for standardized Probit coefficients. (e) Except for the probability to invest, all variables are standardized with mean zero and variance equals to unity. For (d), standardized coefficients are from the ordinary Probit multiplied by the ratio of standard errors of explanatory variable and the latent variable (variance of the latent variable is given by W(X'X)P + 1, where Pis veetor of Pro bit eoefficients). Source: World Bank (2005). 5.11 Notably, the Philippines has relied much more on the private sector for capital investment needs compared to other countries in the region (Figure 5.8). It is equally noteworthy, however, that the amount of spending on fixed capital formation is notably less in the Philippines than the average for the East Asia and Pacific region. Furthermore, this gap has been increasing since the East Asia financial crisis, raising questions about the sustainability of economic growth in the Philippines. 1------.. -.. -.---ft'!.r.:~_~.8: Pull.~!~_.~!!d Priyate Fi!~cLCap!!~.¥~rmatio,!!.!s of GDP I East Asia & Pacific The Philippines 40 35 35 30 Q. 30 Q. 25 c c;, c c;, .... 25 0 .... 0 20 .... 20 c .... c OJ OJ 15 1.1 "- 15 1.1 "- Q. OJ 10 OJ Q. 10 Private 5 5 Sector 0 0 0 N <:t <.D 00 0 N <:t <.D 00 N <:t <.D 00 00 00 00 en en en en en 00 en en en en .-i en en .-i en en en .-i .-i en .-i .-i 8 0 0 0 0 0 0 0 0 N 00 00 00 en en <:t <.D 00 0 en N 00 00 en en en en en en .-i .-i en en .-i en en en en <:t <.D 00 0 0 0 N 0 0 <:t <.D 0 0 0 0 .-i .-i .-i .-i N N N N .-i .-i .-i .-i .-i .-i .-i N N N N 1-_-------------______ _ _ _ _ ···· _ - - - - _ . _ . _ - - 1 ._._ _ _ __ Source: World Bank, Development Data Base. 90 B. The Potential for Enhancing Employment Generation in Manufacturing 5.12 Some of the investment climate factors that influence productivity and investment also affect overall employment growth in manufacturing. In particular, employment growth appears to be adversely affected by infrastructure problems and labor regulation (Table 5.3). Table 5.3: Correlates ofInvestment Climate with Employment Growth in Manufacturing (multivariate regression analysis) Corruption and Security and Infrastructure. Labor problems and governance crime regulation Employment -0.196 -0.019 -0.086* 0.0132 (strikes) growth a (0.148) (0.212) (0.053) (0.208) -2.756* (lab. reg.) (1.523) * indicates coefficient is significant at 10 percent level. Robust standard errors are given in parentheses. Source: World Bank (2005) 5.13 In addition, it appears that market distortions may be constraining the movement of labor and employment growth in the manufacturing sectors. Such distortions could be forces that prevent the marginal products of labor and capital from being equalized across sectors, as would be the case if factors were freely mobile across sectors and absent any distortions. The lack of such labor mobility is apparent from Figure 5.9, which shows wide variation in the marginal product of labor (or output produced per additional worker) among sectors in manufacturing, but these do not appear to be converging over time. Furthermore, employment growth has not been notably faster in the sectors exhibiting high marginal labor products. Figure 5.9: Marginal Products of Labor in Manufacturing The marginal products of labor vary widely and have Employment growth has not been faster in sectors not converged over the period 1999-2005 with a high marginal labor product Marginal Employment growth product of (1999·2005, labor average annual (2005) %) 600 '0 Electronics 500 400 · Transport Plasttqect Mach Metal casting Other chemical F~~nts Glass Metal ProdUtf~er 300 200 · -5 R '-0.006 100 · -'0 NMetal min Beverages Cement '00 200 300 400 500 ·,5 +---.----.----.----.----.----.----.------. 50 100 150 200 250 300 350 400 Marginal product of labor (1999) Marginal product of labor (1999) Source: World Bank calculations based on NSCB, Philippines Statistical Yearbook, various years. 5.14 Much of the variation in the marginal product of labor across sectors reflects differences in the average output per worker. As described in the Chapter Annex, inter· 91 sectoral differences in the marginal product of labor can be decomposed into (i) differences in the labor share in total revenues across sectors, (ii) differences in the markup over costs across sectors, and (iii) differences in the average productivity of labor (i.e. output per worker) across sectors. As the left-hand side of Figure 5.10 shows, sectors with higher shares oflabor (i.e. the more labor-intensive sectors) are not the ones that exhibit higher marginal products of labor. Meanwhile, as shown in the right-hand side of Figure 5.10, the sectors with higher marginal products of labor only exhibit slightly higher markups. The bulk of the inter-sectoral differences in the marginal product of labor therefore appear to reflect differences in the average product of labor (Figure 5.11). Figure 5.10: Marginal Products of Labor in Manufacturing Sectors with a high marginal product of labor are not the most labor-intensive ones ..· Marginal Marginal product of labor product of labor (1999·2005) (av 1999-2005) 400 400 Cement Cement 350 350 300 Other chemicals R':0.37 300 Other chemicals ::J Beverages Glass 250 200 Bev ages Rl= 0.22 150 100 Garments ~ 100 50 o 10 20 30 40 50 1.0 1.1 1.2 1.3 1.4 1.5 Average markup (1999·2005) Source: World Bank calculations based on NSCB, Philippines Statistical Yearbook, various years. Figure 5.11: Output per Worker in Manufacturing Instead, a higher marginal product largely ... and sectoral differences in the average product reflects a higher average product of labor... of labor are not correlated with TFP Marginal Output produc1 of labor per worker (a ·. 1999·2005) (av.1999-2005) 400 3000 Cement 350 2500 emant 300 Other chemicals R'·0.75 2000 250 Tobacco 200 Glass 1500 Bevers\!!lRer chemicals Transpo Nf rrous met Nferrou$ met 150 m 1000 Tobacco 100 50+-----~----~----~----~----~--__. 500 1000 1500 2000 2500 3000 Output per worker (a ·. 1999·2005) ·500+-----.-----~--~----~----_.----~ o 10 15 20 25 Total factor productivity (average 2005) Source: World Bank calculations based on NSCB, Philippines Statistical Yearbook, various years. 92 5.15 These inter-sectoral differences in output per worker, in turn, mainly reflect differences in the ratio of capital to labor. Output per worker can vary across sectors due to differences in TFP and/or differences in factor accumulation. 51 As shown on the left-hand side of Figure 5.12, the relationship between output per worker and the capital-labor ratio is very strong, while the relationship between output and TFP levels was shown earlier in Figure 5.11 to be very weak (and, somewhat surprisingly, negative), suggesting that the inter-sectoral differences in output per worker mainly reflect differences in factor accumulation-i.e. the growth of the capital-to-labor ratio. Figure 5.12: Capital-Labor Ratios and Factor Market Distortions Sectors with high average products of labor ... reflecting distortions that make labor more have high capital-labor ratios ... expensive and reduce employment in these sectors Output per worker (av, 1999-2005) 1BOO Tobacco 1800 Beverages 1400 Other chemicals 1200 1000 BOO 600 400 200 0 o 100 100 200 300 400 500 Capital per worker (av, 1999-2005) Note: The horizontal bars on cement and petroleum have been omitted from the figure so as to make the differences between the bars ofthe other sectors visible. Source: World Bank calculations based on NSCB, Philippines Statistical Yearbook, various years. 5.16 In many manufacturing sub-sectors, capital-labor ratios are far higher than what would be expected after accounting for differences in technology. This may be indicative of the existence of policy distortions that affect the relative price of labor and keep employment artificially low. 52 This conclusion may need to be tempered by the possibility that differences in 51 Consider a standard neo-classical production function with constant returns of scale, Y = A *F(K,L}, where Y denotes real output, A is total factor productivity, K refers to the capital stock, and L is the quantity of labor. Then the rate of labor productivity growth, dlog(Y/L)/dt, is equal to the rate of growth of the capital-labor ratio, dlog(KlL)/dt, and the rate of total factor productivity growth, dA/A. 52 Assume that output in each manufacturing sector is determined by a simple production function with constant returns to scale and two homogenous factors of production, Capital (K) and Labor (L). In such a framework, profit- maximizing firms in each sector will employ labor and capital until the point where the capital-labor ratio multiplied by the inverse of their respective factor shares in each sector s) is equal to the economy-wide wage-rental ratio multiplied by sector specific factor distortions, or [K/Ls]*[aL/aKs] = [w/r]*~s, where ~s refers to the factor distortion in sector s. In the absence of factor market distortions (meaning that ~s 1) and assuming that the wage-rental ratio is the same across sectors, the variations in KIL across sectors are expected to reflect only variations in the relative factor shares (adaK)' Any variation in KIL in excess of what can be explained by variations in (aL/ad constitutes 93 capital-labor ratios across sectors may reflect differences in the quality of labor. 53 If the quality of labor varies significantly across sectors (e.g., with some sectors employing a higher proportion of skilled versus unskilled labor), the more skill-intensive sectors should exhibit a higher average wage-rental ratio (and higher capital-labor ratio), even in the absence of factor price distortions that raise the price of labor. 5.17 It appears that the level of factor market distortions varies significantly across manufacturing subsectors. The right-hand side of Figure 5.12 above shows the level of factor market distortions calculated for each manufacturing sector as a proportion of the economy-wide wage-rental ratio. This calculation reveals large variations across sectors, at least part of which can be attributed to differences in the magnitude of the sector-specific factor price distortions. Ranking the sectors according to the calculated size of the factor market distortions indicates that cement and petroleum products are by far the most distorted sectors, followed by rubber, glass, iron and steel, and non-ferrous metals. 5.18 Building upon this last finding, the remainder of this section briefly discusses some policies that are likely to have limited competition in the manufacturing sub-sectors. It focuses on those sub sectors exhibiting the greatest prevalence of factor market distortions in Figure 5.I2-namely glass, cement, and petroleum products. 54 Labor market distortions, which directly interfere with labor hiring/firing processes, are addressed separately in Chapter VI. 5.19 The manufacturing sector in the Philippines exhibits a largely oligopolistic market structure. Most sub-sectors show a four-firm concentration ratio above 70 percent. Moreover, the market structure has become increasingly concentrated, as the average four-firm concentration ratio across all sub-sectors rose from 71 percent in 1988 to 81 percent in 1998. This increasingly concentrated market structure, however, did not necessarily translate into greater market power: while price-cost margins have traditionally been quite high in the Philippines' manufacturing sector, these margins remained fairly stable over the 1990s, despite the increase in firm concentration ratios. As others had already recognized, domestic market structure is not a good indicator of market power in the presence of trade liberalization. prima facie evidence that Industry 10.32 70.38 19.3 100.0 <:> N Services 6.22 5.89 87.89 100.0 Total 35.16 16.0 48.84 100.0 Note: Share of workers employed in both periods, panel observations only. Source: Matched LFS-FIES data 2003, 2006. 6.26 Significant differences in wages across different geographic areas also suggest regional segmentation in the labor market. As noted above, individuals employed in the NCR earn the highest wages, while the wage in Visayas is 40 percent lower than in the NCR, ceteris paribus. Rather than declining, wage gaps have actually widened over time. Even within regions, wages vary widely across sectors. For example, in Luzon, workers in the industry sector and services sector earn 46 percent and 24 percent higher daily wages, respectively, than those in the agricultural sector. 6.27 Another possible sign of labor market segmentation is the rising unemployment rate among skilled labor. The overall ratio of skilled to unskilled unemployment rose from 2.3 in 2003 to 2.7 in 2007, with skilled individuals aged 25-34 experiencing the highest increase in unemployment. The combination of greater relative demand for skilled labor, as discussed above, and higher unemployment rates among skilled labor points to an excess supply of skills. Although an excess supply of skilled labor should depress their wages, especially for younger cohorts, the skills premium in the wages paid to younger workers has increased (Figure 6.9). 115 These trends suggest the existence of Figure 6.9: Changes in the Skill Premium and Relative labor market frictions that result in the Supply of Skills, 2003-2007 rationing of jobs in growing sectors 0.35 and prevent wages from falling to their :>.3 market level. It could also indicate the 0.2.') existence of a skills mismatch, in ... <:> :1.2 which the skills supplied do not match ~ 0.15 those demanded by employers, or it 8 N :J.1 .. Relat ve suppy 0.05 could reflect an increase in the ~ """ Skill proOlium iii 0 .t:: reservation wages of younger, skilled " ;;<: ..(lOS workers. ·:).1 ..O.lS 6.28 The mobility of labor may be Age-group hindered in part by labor market rigidities stemming workers individuals with higher education or from Note: Skilled unskilledare defined asrepresented by individuals with only above, while workers are employment protection primary or no education. Relative supply is the ratio of skilled to legislation.Although meant to protect unskilled individuals actively participating in the labor market. The skill premium is the ratio of skilled to unskilled workers, the strictness of employment Source: LFS, October round 2003, 2007 average wage. protection legislation and level of enforcement, interacting with other labor market institutions such as minimum wages, have been found to affect labor costs, which in turn affects job creation and job destruction rates (Boeri et aI, 2008). Compared to other countries in the region, the conditions placed on the use of fixed- term contracts are more restrictive in the Philippines. As shown in Table 6.10, the Philippines, along with Indonesia and Vietnam, has the least flexibility in contractual arrangements. According to the Philippines Labor Force Survey, 80 percent of workers were employed under a permanent contract in 2007. Furthermore, the distribution of workers across different contract arrangements-permanent, temporary, and working for multiple customers or employers-has remained fairly stable: the share of permanent workers changed by less than 2 percentage points between 1997 and 2007. 116 Table 6.10: Fixed-Term Contract Regulations in Selected East Asian Countries Country Regulation Limited duration (3 years; 5 years for highly specialized employees) Japan several renewals are possible; exemptions exist for specified projects. Korea Limited duration (2 years); exceptions exist for specified projects or tasks. Most flexible Malaysia No limit on duration; contract must be in writing if longer than 1 month. No restrictions; regulations cover every establishment employing at least Thailand 20 employees. Singapore Specified work or specified period of time. Moderate Overwhelming majority of employment relationships is permanent; China contract for defmed tasks must be in writing. Cambodia Limited duration (2 years); defined tasks; contract must be in writing. Indonesia Limited duration (3 years); no renewals; temporary work. Most restrictive Specified projects; limited duration depending on the defmed categories; Philippines no renewals. Vietnam Limited duration (1-3 years); 36-month extension; specific/seasonal work Source: ILO-LABORSTAT,2008. 6.29 Minimum wages in the Philippines are among the highest in the region. The average level of minimum wage for 2007 was set at US$494 PPP. Higher minimum wage values have been registered in only two other more developed countries in the region, Taiwan and Korea (Table 6.11). The Philippine level is especially striking considering that only 20 percent of countries worldwide have monthly minimum wages exceeding U8$500 PPP. The Philippines also ranks first regionally with a minimum wage being 50 percent higher than GDP per capita and 90 percent of the average wage of salaried workers. The objective of the minimum wage is usually to improve the welfare of low earners. In a competitive labor market, however, minimum wages contribute to job rationing and unemployment if set above the market-clearing wage level. Moreover, minimum wages will price out workers whose marginal productivity falls below the wage floor-typically younger, unskilled workers-inducing firms to substitute unskilled labor with relatively more productive skilled labor. 73 Preliminary evidence on the employment impact of minimum wages suggests a negative relationship, but requires further analysis. 73 One way to mitigate this adverse effect on unskilled workers is through the implementation of capacity-building and entrepreneurship programs. An example of such a program is the Self-Employment Assistance Kaunlaran (SEA-K), being implemented by DSWD in coordination with local government units. 117 Table 6.11: International Comparison of Minimum Wages Level of the minimum wage (2007 or latest) MW/ MW/GDP Country PPP (US$) average per capita 0 wages 0 424 150.6 90.8 Nepal 133 132.4 Cambodia 156 103.8 Bangladesh 69 63.6 Vietnam 120 55.7 58.5 India 113 50.9 22.8 China 204 46.3 37.5 Thailand 304 46.2 56.0 Indonesia 142 45.8 64.0 Korea 815 39.4 28.9 Lao 65 38.0 Taiwan 955 38.0 36.7 Sri Lanka 122 36.0 Source: ILO-Global Wage Report 2008/09, "Minimum wages and collective bargaining: Towards policy coherencc." F. Policy Implications and Conclusions 6.30 Improving income-earning opportunities, especially for the poor, is a key component of fostering inclusive growth. Since the poor derive most of their incomes from wage work and self-employment, the impact of growth on poverty depends on the extent to which growth can generate employment and good earning opportunities. Improving income opportunities for the poor requires removing constraints to growth and productivity in the labor- intensive agriculture sector where the poor are concentrated, as well as removing constraints to employment generation in well-performing sectors of the economy. 6.31 The Philippines will also need to tackle the demographic challenge to relieve pressure on the labor market. The country's relatively high population growth is already placing tremendous pressure on the labor market to generate more jobs. If the working-age population continues to follow the same trajectory, the insufficiency of wage-earning opportunities will likely lead to even greater unemployment and underemployment and hence poverty. 6.32 Beyond creating opportunities, it is critical to ensure that workers have the ability to move to those opportunities. The preceding findings indicate a need to review labor market policies and consider removing the policy-induced distortions and rigidities that inhibit labor mobility. Further analysis is needed to understand the causes of labor market segmentation and rigidity and their impact on mobility and employment generation. The factors that limit mobility between jobs as well as physical mobility between different geographic areas should be examined, and such distortions must be removed to maximize flexibility and allow the labor market to function efficiently. The recent debate in OECD countries about the optimal configuration of flexible labor legislation and secure social protection-the so-called flexicurity model-provides useful insights for potential new approaches to tackling the problems of the Filipino labor market. The international evidence suggests that efficiency gains might be 118 achieved by strengthening social protection-i.e. unemployment insurance, wage subsidies, safety nets-in a way that is inclusive and mobility.,.friendly. Measures to strengthen social protection are discussed in Chapter IX. 119 References Ashenfelter, O. and Smith, R.S. (1979). "Compliance with the Minimum Wage Law." Journal ofPolitical Economy, Vol. 87(2): 333-50. BLES (2004). "Statistics on non-regular workers." Labstat Updates Bulletin. Vol. 8(21). Manila, Philippines. (2008). "Concept and measurement of employment in the informal sector (IS)." Labstat Updates Bulletin. Vol. 12( 17). Manila, Philippines. Bloom, D.E. and Canning, D. (2008). "Global demographic change: dimensions and economic significance." Population and Development Review, Vo1.34, Supplement: 17-51. Bloom, D.E. and Williamson, J.G. (199S). "Demographic transitions and economic miracles in emerging Asia." World Bank Economic Review, V01.12(3): 419-455. Boeri, T. and van Ours, J (2008). The Economics of Imperfect Labor Markets. Princeton University Press. Boeri, T. et al. (2008). "Labor regulations in developing countries: a review of the evidence and directions for future research." SP Discussion Paper No.OS33. Washington, DC: The World Bank. Botero, J.C. et al. (2004). "The Regulation of Labor." Quarterly Journal o/Economics, Vol. 119(4): 1339-1382. Fields, G.S. (1994). "The unemployment effect of minimum wages." International Journal 0/ Manpower, Vo1.l5(2,3): 74-S1. ILO (2008). Global Wage Report 2008/09. "Minimum wages and collective bargaining. Towards policy coherence." (2009). "Update on minimum wages development" http://www.ilo.org/wcmsp5/groups/public/---ed normln- re1confldocuments/meetingdocument/wcms 101713. pdf. website www.ilo.org Lanzona, L and Felipe, J. (2006). "Labour laws, and Economic Policies in the Philippines." Chapter 7 in Felipe, J. and Hasan, R. (eds;), Labour Market Issues in Asia: Issues and Perspectives. London: Palgrave Macmillan for the Asian Development Bank, 2006. Lee, R. and Mason, A. (2006). "What is the Demographic Dividend?" Finance and Development Vo1.43(3). 120 National Statistics Coordination Board, website www.nscb.gov.ph. Orbeta, A. C. Jr. (2009). "Overview if the labor market and social protection in the Philippines" Mimeo. Manila, Philippines. Saget, C. (2008). "Fixing minimum wage level in developing countries: common failures and remedies." Internatinal Labor Review, Vol. 147(1). Shorrocks, A.F. (1978). "The measurement of mobility." Econometrica, Vol. 46 (5): 1013-1024 OECD (2009). "Is Informal Normal. Towards more and better jobs in developing countries." JUtting, 1. and Juan R. de Laiglesia, J.R. (eds). Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat, "World Population Prospects: The 2008 Revision." In http://esa.un.org/unp. UNDP (2008) in page 114. World Bank (1984). "World DeVelopment Report." Washington, DC: The World Bank. World Bank (2009). "Doing business." Washington, DC: The World Bank. (2009). "Hiring and firing regulation." In Indonesia Job report. Jakarta, Indonesia. (2009). "Minimum Wages." In Indonesia Job report. Jakarta, Indonesia. 121 CHAPTER VII HEALTH, NUTRITION, AND POPULATION A. Introduction 7.1 Improvements in health, nutrition, and population (HNP) can contribute to inclusive growth and poverty reduction through several mechanisms. First, lower fertility rates can reduce demographic pressures, contribute to improved growth in per capita incomes, increase per capita investments in human capital, and reduce poverty through smaller family sizes. Second, better nutrition improves cognitive function and productivity and reduces child and maternal mortality (Black, et aI., 2008). Third, improved health increases productivity and reduces loss of income from morbidity and mortality. Maternal death, which is a tragedy in itself, also contributes to child poverty, malnutrition, and reduced educational attainment by children. Fourth, social insurance or government subsidies for public services can provide "risk protection" to prevent individuals from falling into poverty (Doorslaer, et aI., 2007). To achieve these gains, however, improvements in health and nutrition must be inclusive-that is, they must be shared by the rich and poor and among regions within the country. Thus, improving HNP outcomes for the poor and for "lagging" regions in the Philippines is key both to attaining the Millennium Development Goals (MDGs) and to achieving inclusive growth. 7.2 At the same time, inclusive growth and poverty reduction contribute to improved health, population, and nutrition outcomes. Household incomes and socioeconomic status are among the most important determinants of health and nutrition status. Better-off households tend to prefer fewer children and invest more in their children, creating a "virtuous cycle" of increased incomes and improved health and nutrition. Growth also contributes to increased resources available for health, which can in turn contribute to increased supply and quality of health services. Furthermore, inclusive growth generates more resources for improving education, especially for girls, which is strongly correlated with improved HNP outcomes. In addition, growth and poverty reduction contribute to improved environmental health, including water and sanitation, and better living conditions reduce the risk of communicable diseases such as tuberculosis. Conversely, however, the negative externalities of growth need to be mitigated, since industrial and traffic pollution can have a negative effect on health, particularly for the poor. 7.3 This chapter seeks to address two key questions regarding the linkages between HNP and poverty reduction in the Philippines. First, is there evidence that inadequate investments and progress in HNP are contributing to the relatively slow progress in poverty reduction in the Philippines? Second, are the resources generated by economic gro\\'ih being used effectively to improve health and nutrition for the poor and to provide risk protection against the impoverishing effects of illness? After a brief overview of the health sector, this chapter reviews key indicators and recent trends in HNP. Next, it briefly examines equity of HNP outcomes, access to and utilization of health services by the poor, and expenditures by the poor on health. It then analyzes the trends and patterns of spending on HNP and the roles of government, the National Health Insurance Program, and the private health sector in financing and providing key HNP goods and services for the poor. The chapter concludes with 122 recommendations that could contribute to inclusive growth and improved health outcomes for the poor. B. Background on the Health Sector 7.4 The Philippines health system is a heterogeneous mix of public and private sectors. Public health facilities provide curative and preventive public health services, including primary health care, immunization and vitamin A distribution, and tertiary hospital services. Since the decentralization of 1992, the Department of Health (DOH) has assumed responsibility for public health commodities such as vaccines and tuberculosis drugs and for a network of 70 "retained" DOH hospitals, and it provides technical support to local governments through regional Centers for Health Development (CHDs). The DOH is also responsible for policy and regulation of the sector. Responsibility for financing and management of local hospitals and primary care is devolved to local government units (LGDs), with provinces typically responsible for managing provincial and sub-provincial hospitals and with cities and municipalities responsible for public health programs, city and municipal hospitals, and barangay health stations. The private sector includes a range of private hospitals-particUlarly in urban areas--doctors and midwives in private practice, and large private health education and pharmaceutical sectors. 7.5 The Philippines has a national health insurance program in place. In 1995, the government expanded its formal sector social health insurance program (Medicare) into a national health insurance program. It established the Philippine Health Insurance Corporation (PhiIHealth) to manage the program, which is financed through both contributions from formal sector workers and employers, individual or group contributions from the informal sector, as well as central and local government subsidies for the poor. 7.6 The country has been implementing a health sector reform agenda or the FOURMULA Oue for Health. The reform agenda promotes the improvement of health outcomes and the attainment of the MDGs for health through increased public investment in health care delivery, including investments to support expanded facility deliveries, immunization rates, and treatment rates of tuberculosis, particularly for the poor. It also calls for universal social health insurance coverage and enhanced insurance benefits, as wen as the use of performance-based financing to support local level service delivery and improved performance of public hospitals. The reform agenda also aims to strengthen regulatory capacity to assure access to quality medicines and other health goods and services by the poor. C. Key Indicators and Recent Trends Demographic trends and poverty 7.7 Although population growth and fertility rates in the Philippines have declined modestly in recent decades, they are now among the highest in the region. As discussed in Chapter 6, the population of the Philippines was estimated at 88.6 million in 2007, an increase from 76.5 million in 2000. Population growth rates declined slowly from over 3 percent in the 1970s to about 2.3 percent annually in the past decade, compared to an average population growth rate of 1.6 percent for Southeast Asia (Figure 7.1). Total fertility rates have also declined 123 from 4.1 births per woman on average in 1991 to 3.3 births in 2006 (NDHS 2008). However, this fertility decline is among the slowest in Southeast Asia. Figure 7.1. Annual Population growth Rate of Select ASEAN Countries 1950-2000 +---~~~--------------~~~~~~~------~ ___ ~i~~ _Thailand -e-- VI9t Nam +-----------------------------~~----~~--~ -*-I~~ _ SEAsia 7.8 Poor women in the Philippines have three times as many children as rich women- one of the widest disparities in the East Asia region. While women in the wealthiest quintile have only two children on average, poor women average nearly six births (NDHS 2003). Fertility among poor women has declined only slightly from 6.5 children in 1998. In contrast, the fertility rate for women in the poorest quintile is 2.2 in Vietnam, 3.0 in Indonesia, and 4.7 in Cambodia (see Table 7.2 below). Disparities can also be seen across rural and urban areas of the Philippines: women in rural areas have 3.8 children on average compared to 2.8 children for women in urban areas (NDHS 2008). 7.9 Use of family planning has stagnated in the past decade, with a persistent nnmet need for family planning and inadequate access to family planning services for poor women. The percentage of married women using family planning has remained static at around 50 percent since 1995, although the use of modern methods has increased to over 35 percent of married women (NDHS 2008). In 2003, only 24 percent of women in the poorest quintile were using family planning, compared to 58 percent in Vietnam and 49 percent in Indonesia (see Table 7.2 below). Over half of pregnancies in the Philippines are unintended. While poorer women typically want more children than wealthier women, women in the poorest 40 percent of the population have between 1.5-2.1 more children than they desire (Guttmacher, 2009). Moreover, poor households bear the heaviest financial burden for family planning, with the lowest income quintile spending 5.6 percent of household health expenditures on contraceptives compared to less than one percent for the highest income quintile (FIES 2006). 124 7.10 The slow decline in population growth and the continued high fertility among poor women have contributed to, slow progress in poverty reduction. It can be difficult to disentangle the extent to which high fertility and population growth is a cause or consequence of slow progress in poverty reduction. However, a convincing range of both macro and micro evidence from the Philippines suggests that continued high fertility among the poor has hindered progress in reducing poverty (Racelis, 2008). At the macroeconomic level, simulation models using economic and demographic data for the Philippines showed that higher population growth lowered GDP per capita (Orbeta, 2002). Similar models also found that while expenditures on human capital rise with population growth, the increases are not sufficient to maintain per capita expenditure levels (Orbeta, 2002). The growth of provincial per capita income is negatively correlated with the proportion of young dependents in the province (Mapa, Balisacan and Briones, 2006). At the household level, large family size is strongly correlated with higher poverty incidence, lower savings and asset accumulation, and reduced per capita household expenditures for education and health (Orbeta, 2002; Orbeta, 2005; Racelis 2008). 7.11 Depending on future fertility trends, the country's population in 2050 could be as low as 127 million or as high as 160 million. Racelis (2008) undertook alternative population projections based on differing assumptions as to when the Philippines will attain a replacement total fertility rate (TFR) of 2.0, together with assumptions about continued improvements in life expectancy. Under the "medium" scenario, which assumes that replacement fertility is reached by 2040, the population would reach nearly 120 million by 2025 and 150 million 2050. The "high" scenario cited above assumes that a TFR of2.0 will not be reached until 2050. In contrast, if intensive family planning efforts and changes in fertility preferences were to result in replacement fertility being reached in 2020 or 2025, the population in 2050 would stabilize at around 127 million or 133 million, respectively. 7.12 These alternative popUlation scenarios have significant implications for public expenditure requirements, employment, human capital investment, and environmental management. As discussed in Chapter 6, a larger population requires that the economy generate more jobs. It also requires that the government build more schools and hire more teachers. Increased public and private expenditures for health care are also necessary, and as the population ages, the burden of health expenditures increases further. For health, the growth scenarios above suggest that health expenditure will need to at least double, with an even higher burden for the high-growth scenarios (Racelis, 2008). Population pressures can also increase environmental degradation and may push more people into areas that are more prone to natural disasters. Epidemiological trends and progress toward the MDGs 7.13 The Philippines has performed relatively well in reducing infant and child mortality and is currently "on track" for these MDGs. The MDGs for health call for reducing the infant mortality rate (IMR) and child mortality rate (CMR) by two-thirds from 1990 to 2015. As shown in Table 7.1, the Philippines has achieved a steady reduction in infant and child mortality over the past 15 years, and if current trends continue, these MDG goals are likely to be met The progress in child health is likely attributable to factors such as improved vaccination coverage 125 and other preventive interventions such as vitamin A distribution. Overall measles immunization coverage increased from 80 percent in 2003 to 84 percent in 2008, and coverage of DPT3-three doses of vaccine against diphtheria, pertussis, and tetanus, commonly used as a measure of health service availability-increased from 79 percent to 86 percent. However, reaching the MDGs for infant and child health cannot be taken for granted. The recently released 2008 DHS data show that while IMR and CMR continued to decline from 2003 to 2008, the IMR decline was relatively small. Table 7.1: Progress toward Health, Nutrition, and POl!ulation MDGs MDG Baseline Level Current Target 1990 or year 2002/2003 Level by closest to 2008 2015 1990 Eradicate extreme poverty and hunger Prevalence of malnutrition among 0-5 year-old 34.5 27.6 n.a. 17.3 children ( underweight) Reduce child mortality Under 5 mortality rate (per 1,000 children) 80.0 40.1 34.0 26.7 Infant mortality rate (per 1,000 live births) 57.0 29.0 25.0 19.0 Improve maternal health Maternal mortality rate 209.0 172.0 162* 52.2 Increase access to reproductive health services Prevalence of men and women/couples practicing 40.0 48.9 50.7 70.0 responsible parenthood HIV prevalence <1 <1 <1 <1 Halt and begin to reverse the incidence of malaria and other diseases Malaria morbidity rate (per 100,000 population) 123.0 48.0 n.a. 24.0 Provide basic amenities Proportion of families with access to safe drinking 73.7 80.0 86.8 water '" 2006 Family Planning Survey. Sources: National Demographic and Health Surveys (1998, 2003, 2008) 7.14 The Philippines is currently not on track to reaching the MDGs for maternal mortality reduction and reproductive health. The health MDGs call for reducing the maternal mortality rate (MMR) by three-quarters from 1990 to 2015. The current estimate for MMR is 162 deaths per 100,000 live births, which is high relative to the country's level of development. The slight increases in the percentage of births attended by health professionals and taking place in health facilities from 2003 to 2008 indicate the difficulties in reaching the MDG on maternal mortality reduction. 7.15 Child malnutrition rates continue to be very high, particularly among the poor. Despite the declining prevalence of malnutrition, the Philippines remains one of 20 countries in the world with the highest burden of malnutrition (Black, et aI., 2008). In 2003, the prevalence of underweight children was 25.6 percent among 6-10 year oIds, while the prevalence of stunting was 35.8 percent-among the highest rates in the region. Furthermore, disparities in malnutrition 126 persist across income groups, with 42 percent prevalence of underweight children aged 6 to 10 among the lowest income quintile families compared to 10 percent for the highest income quintile. Similarly, the stunting prevalence of the poorest quintile was 59 percent compared to 16 percent for the richest quintile (NNS 2003). 7.16 The Philippines is undergoing an epidemiological transition, with a growing burden of non-communicable diseases (NCDs) and injuries and an increasing proportion of elderly in the population. NCDs are increasingly becoming the main burden of disease in the country. The prevalence of hypertension among adults aged 20 and above increased from 21.0 percent in 1998 to 22.5 percent in 2003, while the percentage with elevated total cholesterol more than doubled from 4.0 percent to 8.5 percent. The prevalence ofNCDs is expected to rise due to the increasing share of fats in the Filipino diet and the increasing rate of obesity, which rose from 17 percent of the population in 1993 to 24 percent in 2003. The increasing prevalence of NCDs affects the poor both in terms of treatment costs for these chronic conditions and in terms of reduced productivity, morbidity, and premature mortality. The increasing proportion of those aged 65 years and above further compounds this burden, with estimates of per capita health expenditures by population age group showing relatively high per capita spending for this age group (Racelis, 2008). D. Equity in health outcomes and utilization of HNP services 7.17 The overall positive trends in infant and child mortality mask significant differences between rich and poor and among regions in Philippines. In 2003, infant mortality among the poorest quintile was more than twice as high as among the richest, while child mortality (ages 0- 5) was more than three times higher among the poorest quintile-an increase compared to 1997 (see Table 7.2 below). Such disparities are likely due in part to disparities in immunization coverage and access to preventive measures. While immunization coverage exceeded 90 percent among the richest quintile in 2003, less than 70 percent of the poorest were vaccinated against measles. The 2008 data are not yet available by income quintile, but wide geographic disparities in DPT3 coverage can be seen between the National Capital Region (NCR) at 89 percent and the Autonomous Region of Muslim Mindanao (ARMM) at 41 percent. Children of parents with no education have DPT3 coverage of 36 percent, compared to 93 percent among children of college graduates (NDHS 2008). For other preventive and child health care indicators such as vitamin A coverage or medical treatment of acute respiratory infections, the gap between the richest and poorest quintiles was about one-third in 2003. 7.18 Neonatal mortality-which is influenced by the quality of care, fertility, and maternal nutrition-remains high among poor and relatively high among middle income groups. Neonatal mortality was 21 per thousand among the poorest in 2003, compared to l3 per thousand among richest quintile (Gwatkin, et aI., 1998). Moreover, neonatal mortality showed limited improvement in the decade preceding the 2003 survey, most likely due to limited progress in increasing access to and improving the quality of facility-based deliveries and post- natal care and in reducing fertility. As illustrated by Figure 7.2, the regional disparities in facility-based deliveries are striking, with 68 percent of deliveries in· the NCR being facility deliveries compared to 15 percent in the ARMM. Similarly, 87 percent of deliveries in the NCR were attended by a health professional compared to 19 percent in the ARMM. 127 Fi ure 7.2: Birth Deliveries in a Health Facili Regional Penentage delivered in a Health fatUity 0·15 15.1.25 25.1·40 40..1-50 50.1.68.5 Source: Data from NDHS 2008. 7.19 From a comparative perspective, while coverage of child health services for the poor is comparable to similar countries in the region, coverage of reproductive health services is worse in the Philippines. Table 7.2 shows outcomes and service coverage for maternal and child health, comparing the richest and poorest quintiles of the population for the Philippines with Vietnam, Indonesia, and Cambodia. The table shows that coverage and utilization of curative and preventive child health services by the poor in Philippines increased modestly between 1998 and 2003. While coverage needs to improve further, the use of public health services by the poor is comparable to that in Indonesia and Vietnam, although the use of private services is lower. Total fertility rates and utilization of reproductive health services by the poor showed limited improvement in the Philippines: the percentage of poor women delivering at home declined from 91 percent in 1997 to 89 percent in 2003, but the difference is not statistically significant. The Philippines has worse coverage of medically assisted deliveries than 128 Indonesia or Vietnam. The utilization of family planning services by poor women is also significantly worse in the Philippines compared to Indonesia and Vietnam. 129 Table 7.2: Comparison of Maternal and Child Health Outcomes and Services for the Lowest and Hi2hest Income Quintiles in the Philippines and Selected Countries in East Asia Philippines Philippines Vietnam........ Indonesia .. Cambodia.. "'''',' ! 1998 2003 2002 2002/3 2000 Indicators Low High Low High Low High Low High Low High Child illlness and mortality I"""""'" Infant mortality rate 48.8 20.9 42.0 19.5 39.3 13.8 60.6 17.1 109.7 50.3 " 1 Under-five mortality rate 79.8 29.2 66.3 20.6 52.9 15.8 76.8 21.9 154.8 63.6 1 "'" Prevalence offever (%) 26.4 19.4 28.0 17.8 30.9 18.4 24.6 20.5 33.9 36.3 I Prevalence of diarrhea (%) 8.8 4.9 13.1 9.4 18.2 3.8 10.0 8.2 19.5 16.0 Prevalence of acute res piratory infection (%) ,..... 15.3 9.1 14.6 5.8 23.7 14.0 8.3 5.2 18.3 19.4 " " I······ Full basic coverage (BCG, measles, DPT) 59.8 86.5 55.5 83.0 44.3 92.3 37.1 64.7 28.6 67.7 Meas les coverage ..... 67.6 92.2 69.7 89.4 64.4 97.8 59.5 84.9 43.9 81.8 1 Maternal and reproducthe Health ...... Total fertility Rate 6.5 2.1 5.9 2.0 2.2 1.4 3.0 2.2 4.7 2.2 Antenatal visits to medically- trained personnel 73.1 97.8 72.3 96.6 67.7 99.5 77.6 99.1 79.6 21.9 ... Delivery by medically trained personnel 21.2 91.9 25.1 92.3 58.1 99.7 39.9 93.6 14.7 81.2 Delivery by a doctor 7.1 75.8 8.6 73.2 29.2 91.4 0.7 0.8 0.4 8.8 Delivery in a public facility 7.1 26.3 9.2 31.6 43.2 94.8 3.6 14.7 1.7 36.0 Delivery in a private facility 1.6 52.5 ....... 1.2 45.5 4.7 3.9 7.3 66.6 0.1 11.0 Delivery at home 91 21.2 88.7 22.6 51.8 1.1 87.5 17.9 97.1 52.2 Contraceptive prevalence among women 19.6 29.4 23.8 35.2 57.9 51.6 48.6 58.1 12.5 25.4 .. .... ....... Treatment of childhood illness in public facility (% ) Fever 25.6 5.4 38.7 11.6 30.6 19.3 32.0 22.8 10.8 18.1 Acute respiratory infection 37.4 11.9 42.8 14.6 50.5 23.1 39.8 19.6 12.1 22.3 Treatment of childhood illness in private facility (%) , Fever 6.5 36.0 6.2 43.8 11.5 43.1 ..... 7.1 48.1 18.1 27.3 Acute respiratory infection 9.4 60.8 7.4 60.7 13.1 54.60 8.5 54.2 19.6 29.4 Note: "Low" corresponds to the lowest mcome qumttle, whIle "hIgh' represents the highest mcome qumtIle. Data by income quintile for the Philippines 2008 NDHS are expected to be available in August 2009. Source: Gwatkin, et al. (2008). 7.20 The effects of inadequate access to water and sanitation services and the impacts of air and water pollution are felt most by the poor. Based on the 2003 National Demographic and Health Survey, only 74 percent of the population has improved water supply and sanitation, 11 percent has improved sanitation but no improved water supply, while 10 percent has improved water supply but no improved sanitation. Although only 5 percent of the population at the national level has no improved water supply and sanitation, 28 percent of the population in 130 ARMM has no improved water supply and sanitation compared to a mere 0.23 percent in the NCR (Arcenas, 2008) As a result, it is the poor, particularly those living in ARMM, who are most likely to suffer from illnesses related to having no improved water supply and sanitation. Water pollution and lack of access to improved water supply and sanitation are estimated to account for 17 percent of all reported illnesses. Similarly, the poor are most likely to be affected by illnesses related to air pollution, since they are more likely to use fuel wood and are more exposed to traffic pollution with their use of open-air 'jeepneys" and tricycles. Furthermore, a survey of children in rural towns in the Visayas islands found that 21 percent of 2,861 children living in rural areas had elevated levels of lead in their blood (Riddell, et aI., 2007). E. Expenditure Issues Protecting against the impoverishing effects of illness 7.21 Aside from having negative impacts in terms of morbidity, mortality, and lost wages, illness can also contribute to poverty through the cost of health care expenditures. Table 7.3, which shows the percentage of non-food expenditures for health, indicates that the percentage of out-of-pocket (OOP) expenditures for health has been Table 7.3: Percentage of Health Expenditure increasing steadily for all income categories in Total Household Nonfood Expenditures, since 2000. The table also shows that wealthy 2000-2006 households spend a greater percentage of their Income quintiles 2000 2003 2006 nonfood income on health than poorer 2.7 3.3 3.5 households. The wide disparity in per capita 2 3.2 3.5 3.7 spending for medical care by the highest income 3 3.6 3.6 4.3 quintile (Php3,138 per capita), which spends 4 3.5 3.9 4.9 more than 25 times that of the lowest income 5 3.4 4.1 5.5 quintile (Php127 per capita), points to non-use of Total 3.4 3.9 5.0 health care among the poor and to greater Note: Calculated at household level. reliance of the poor on publicly provided Source: FIES 2000, 2003, 2006 services (FIES 2006). 7.22 Catastrophic health expenditures have also been rising. In 2006, about 3 percent of Philippine households spend more than 25 percent of nonfood expenditure on health, and 1.2 percent spent more than 40 percent. This incidence of "catastrophic expenditure" is much lower than in China and Vietnam (11 percent and 15 percent, respectively) but higher than countries that have achieved nearly universal health insurance coverage, such as Thailand, Taiwan (China) or Hong Kong (SAR) (Van Doorslaer, et aI., 2007; FIES 2006). The percentage of catastrophic expenditure has increased steadily for all income categories in the Philippines, but a smaller percentage of poor households spend more than 25 percent or 40 percent of non-food expenditures on health compared to rich households (Table 7.4). 131 Table 7.4: Percentage of Households Spending More Than 25 and 40 of N on r.00 d Expen d'ture on HeaIth 2000-2006 I I, 25 nonfood consumption 40 nonfood consumption Income quintiles 2000 2003 2006 2000 2003 2006 I 1.32 1.43 1.81 0.38 0.49 0.58 2 1.68 2.02 2.11 0.58 0.69 0.79 3 2.16 2.07 2.88 0.85 0.75 1.18 4 2.41 2.55 3.61 1.08 0.97 1.71 5 2.98 2.97 4.43 1.29 1.16 1.8 Total 2.11 2.21 2.97 0.84 0.81 1.21 Source: FIES 2000, 2003, 2006. 7.23 The relatively low incidence of catastrophic health expenditures by the poor is likely due to a combination of government subsidies for public sector health services and non-use of health services by the poor. It seems that the poor are accessing government-financed public health services, particularly vaccine and tuberculosis control services, as primary health services at the local level. National government budgetary support for public health commodities has increased in recent years, with the allocation for vaccines rising from Php316 million in 2006 to Php843 million in 2009 and the allocation for tuberculosis control rising from Php 139 million to Php 1.3 billion over the same period, Of the four main sources of health financing, national and local government health expenditures are evenly distributed across all income quintiles, although these should be skewed in favor of the lower income quintiles (Figure 7.3), The low catastrophic health expenditures of the poor may also be due to the non-use of health services. The large overall share of out-of-pocket spending for health services (55 percent in 2007) points to inadequate subsidies for a wide range of necessary health services, which results in the poor forgoing such services. Figure 7.3: Health Expenditures by Financing Agent and by Income Qnintile, 2003 Health expenditures by financing agent and by income quintile, 2003 o Out of pocket oPhilHealth i _ Local govemment .i III National gover~l11~ Source: Racelis (2008). 132 7.24 Despite reportedly high coverage, social health insurance has been making a limited contribution to improving access to services for the poor. The national health insurer, PhilHealth, has estimated the coverage of social health insurance at 76 percent for 2008 (philHealth Stats and Charts 2008). However, preliminary data from poor households identified by the Department of Social Welfare and Development in 2008 revealed that only 26 percent of identified urban poor households and 20.5 percent of identified rural poor households have social health insurance coverage. Furthermore, a review of admission records of 12 DOH managed hospitals showed that although the percentage of admitted patients who were PhilHealth members increased from 2005 to 2007, the reported percentage in 2007 was only 27 percent (Lavado, 2008). Notably, PhilHealth has implemented primary care benefits solely for its poor or indigent members and has introduced outpatient benefits for tuberculosis and malaria care and normal deliveries for all members. Internal PhilHealth studies report that members with the highest support value in terms of percentage of hospital bills covered by PhilHealth are indigent members who were admitted in the ward or service beds in government hospitals. Nonetheless, it appears that social health insurance has not decreased the share of medical spending in total household spending among poor households (APIS 2004). Public and private health expenditure: concerns with equity and efficiency 7.25 Total spending on health is low compared to countries with similar levels of income. Preliminary estimates by the World Health Organization (WHO) of the country's National Health Accounts show that total health expenditure as a percentage of GDP remained relatively unchanged at about 3.4 percent for most of the decade up to 2005 but then increased to reach an estimated 3.9 percent in 2007. This represents a per capita increase from about US$39 in 2005 to US$63 in 2007. Despite these increases, the Philippines still spends less on health as a share of its GDP and less on health per capita than other countries with comparable levels of income (Figure 7.4). 133 Figure 7.4: Government Health Spending in the Phili ines and Other Countries GOVERNMENT HEALTH SPENDING VS INCOME, 2004 .Sampa · .lVIalaysia · Viet .China · .Philippines .Indonesia .Lail"PeOple's Democratic Republic 10 100 250 1000 2500 1000025000 GOP per capita Source: World Development Indicators Note: GDP per capita in current US$; Log scale 7.26 Although central government spending on health has increased in recent years, total government spending on health remains low. After declining in real terms for nearly a decade, the DOH budget has increased from Phpl1.3 billion to Php23.7 billion over the past three years. General government expenditure on health increased from 6.0 percent in 2002 to 6.8 percent in 2007, which is comparable to the levels in Cambodia, Indonesia, and Vietnam. However, countries such as China and Thailand spent more than 10 percent of general government expenditures on health. 7.27 Local government spending has stagnated in real terms over the past decade, which has important equity implications since the poor depend on health services provided by LGUs. Local governments spent a total of Php19.9 billion for health, nutrition, and population control in 2007. Although this represents a modest nominal increase from Php16.5 billion in 2002, it was a stagnant contribution in real terms. As a share of total LGU spending, the contribution for health declined from 12 percent in 2002 to 9.S percent in 2007. Cities and towns are even spending less, with only 7.S percent and 7.7 percent, respectively, of their total 2006 expenditures going to health. In contrast, the provinces spent 17 percent of their total expenditures for health. Indeed, a number of provinces including Camiguin, Negros Oriental, Catanduanes, Romblon, Eastern Samar and Capiz are spending more than 30 percent for health. Interestingly, these provinces do not host any DOH-managed hospitals (DOF, 2008). 7.28 Salary expenditures consume most of LGU health financing, while pharmaceutical and medical supplies, other operating expenses, and capital investments are systematically underfinanced. Data on the composition of LGU spending is limited, but a study in one of the provinces showed that salaries consume more than half of LGU health spending and can reach up to 81 percent (Lavado 2008). This results in systematic under-financing of MOOE and typically little or no investment in capital outlays or infrastructure maintenance. The efficiency 134 of LGU health expenditures are further reduced by the devolved arrangement which allows municipalities to set up small municipal hospitals and allows provinces to retain under-staffed district hospitals. Both types of hospitals are usually under-utilized and often bypassed by the local population. Although the DOH has encouraged LGUs to develop facility rationalization plans, it has been politically difficult to close underused facilities. 7.29 Although expenditures on pharmaceuticals comprise a large percentage of health spending, generics policy and other competition promotion reforms are starting to take effect. Spending for drugs and medicines is the largest expenditure item for household health spending, accounting for 45 percent among all households and increasing to 52 percent for households in the lowest income quintile (FIES 2006). The total pharmaceutical market was estimated at Php94 billion in 2006, accounting for 43 percent of total health spending. The national government has focused on implementing a generics policy and other competition promotion policies to bring down drug prices and improved affordability. Local governments continue to procure drugs for both primary care and hospital services, but the transparency of procurement needs to be improved. Efforts to introduce transparency measures, including electronic drug procurement in both national and local government procurement, are ongoing. 7.30 The share of OOP expenditure is high, representing at least half of total health spending over the past decade. Preliminary data suggest that the OOP percentage has increased despite the reported expansion of social health insurance coverage (FIES 2006; WHO, 2009). This is primarily due to the lack of a comprehensive benefit package that addresses the main reasons for out-pocket expenditure-particularly spending for drugs and medicines-and the continued insistence on retaining inpatient benefits that tolerate balance billing and unlimited co- payments. As mentioned above, social health insurance has not lowered the percentage of medical spending as a share of total household spending, as it is roughly similar at 2.8 percent and 2.7 percent for households with and without PhilHealth membership, respectively (APIS 2004). 7.31 PhilHealth revenues and expenditures have grown steadily in the past decade but it has yet to assert itself and properly leverage its purchasing power. The premiums collected by PhilHealth have increased from PhP8.6 billion in 2000 to PhP25.6 billion in 2008. Benefits payments have also increased at a lower scale, increasing from PhP6.8 billion in 2000 to PhP18.2 billion in 2008. The widening gap between premiums collected and benefit payments increased its equity, including reserves, to over PhP80 billion by the end of 2008. PhilHealth has yet to assert itself as it continues to pay over 70 percent of benefits to private hospitals and tolerates a voluntary component or the individually paying program (IPP), which has severe adverse selection. In 2008, IPP members paid PhP 1.2 billion in premiums while being provided PhP2.5 billion worth of benefits (Phil Health Stats and Charts 2008). PhilHealth has also delayed reforming its inpatient benefits and associated· provider payment system of itemized fee for service and introducing primary care and outpatient pharmaceutical benefits. 135 F. Accountability Issues 7.32 The poor rely predominantly on the public sector for health services, but the quality of care in public hospitals has been problematic due in part to lack of clear performance benchmarks. The DoH used to allocate 60 percent or more of its budget to hospitals, but this share decreased to 35 percent in 2008 as budget increases were allocated to public health measures (Lavado, 2009). Despite the traditionally large share in the budget allocation, public hospitals continue to suffer from inefficiencies, uneven quality, and inadequate funding for medicine and maintenance. To address the funding shortfall, public hospitals have formal user fees, including revolving drug funds. A major issue is that clear performance benchmarks have not been established between the Department of Health and the DoH-managed and LGU-run hospitals, resulting in ambiguity over what services the hospitals are accountable for, which in turn affects public health service delivery. 7.33 Insufficient monitoring and evaluation of service provision have also weakened accountability in public health service provision, although the current reform agenda attempts to address this problem. The health sector reform agenda calls for the use of performance-based financing to support local service delivery and improved performance of public hospitals. The government is also strengthening its regulatory capacity to ensure that the poor have access to quality medicines and other health goods and services. G. Policy Implications and Conclusions74 7.34 Given the implications for inclusive growth and health outcomes for the poor, a key priority is to strengthen the focus on family planning. For the middle and upper classes, private sector and social marketing could help encourage better family planning practices. For the poor, enhanced public service provision will be critical. 7.35 To address the disparities in access to health services, increased public spending for properly targeted interventions is needed. It should be recognized that the macro environment will remain constrained, which highlights the need to improve the efficiency and equity of existing spending. In particular, greater investment in vaccination coverage, access to health professionals and facilities for deliveries, nutritional status of children, and water supply and sanitation will be needed for poorer areas. Financing and appropriate incentives should be provided to local governments to increase expenditures on health and to upgrade the quality of, and access to, essential maternal and child health services. 7.36 Access to health services could be improved by expanding coverage of the National Health Insurance Program. In general, reform efforts should seek to increase mobilization and pooling of funds for health-including through a shift toward universal insurance coverage-and improve risk protection for the poor. Membership coverage of the national insurance program could be expanded by reviewing the voluntary component and expanding government-subsidized enrollment of the poor. In addition, deeper risk protection, another priority area for the national insurance program, could be achieved by improving the benefit package-in particular, through 74 It should be noted that these policy recommendations do not attempt to cover all issues for health reform but rather emphasize reforms that could contribute to inclusive growth and improved health outcomes for the poor. 136 revising the current inpatient benefit and provider payment schemes, and by introducing primary care and outpatient drug benefits. 7.37 To ensure better quality service provision, greater efforts are needed on a number of fronts to improve accountability in the health system. First, the services provided by public hospitals and LGU-provided services could be improved by introducing performance benchmarks, holding hospital management accountable through explicit contracts, and scaling up performance-based financing. Second, access to essential drugs and quality medicines could be increased by strengthening regulation, increasing procurement transparency, expanding government subsidies, and promoting further competition. Finally, the overall stewardship of the health sector could be enhanced by improving the monitoring and evaluation of the sector and the regulation of the private health care sector. 137 References Arcenas, Agustin (2009). "Sustainable Sanitation in East Asia: Philippine Program Mid-Term Review" inception report. Black, Robert, Lindsay Allen, Zulfiqar Bhutta, et al. (2008). "Maternal and Child Undernutrition: Global and Regional Exposures and Health Consequences." The Lancet. January 2008. Daniels, M.C., and L.S. Adair (2004). "Growth in young Filipino children predicts schooling trajectories through high school." Journal ofNutrition. 2004, Vol. 134 (pp. 890-897). Department ofFinance, Bureau ofLocal Government Finance (2007). "Statement of Income and Expenditures 2005-2007." http://www.blgf.gov.ph Department of Health (2008). "Racing Toward our Health Goals: Proposed Budget CY 2009." Manila. Food and Nutrition Research Institute (2003). Philippines: National Nutrition Survey. Doorslaer, Eddy van, et al. (2007). "Catastrophic Payments for Health Care in Asia," Health Economics. Food and Nutrition Research Institute, 2003 National Nutrition Survey. Guttmacher Institute (2009), "Meeting Women's Contraceptive Needs in the Philippines," In Brief, 2009 Series, No.1. Gwatkin, Davidson, et al. (2007). Socio-Economic Difference in Health, Nutrition, and Population Within Developing Countries. The World Bank. Gwatkin, Davidson, et al (2007). "Socio-Economic Difference in Health, Nutrition, and Population: Philippines, 1998,2003" (unpublished). The World Bank. Herrin, Alejandro (2008) "Asian-Pacific Health Financing Profiles: Philippines," in Health Financing Note, East and Pacific Region: Volume II: Country Health System Profiles (draft). Washington, DC: The World Bank. Herrin, Alejandro and Rachel H. Racelis (2006). "Equity and Financing in Healthcare" presentation to the 7th Health Research for Action National Forum (September 28-29,2006), Philippines. Horton, S. et al. "Micronutrient Supplements for Child Survival", www.copenhagenconsensus.com Lavado, Rouselle and Emlyn Cabanda (2009). "The Efficiency of Health and Education Expenditures in the Philippines" (draft). 138 Lavado, Rouselle (2009). "DOH-Retained Hospitals Amidst Health Sector Reform" presentation made to the DoH (February 2009), Philippines. Lavado, R. (2008), Background Paper on the Delivery of Local Government Health Services unpublished. Leiberman, Sandy, and Adam Wagstaff (2008). "Health Financing and Delivery in Vietnam: The Short- and Medium-Term Policy Agenda." Washington, DC: The World Bank. Manasan, Rosario (2006). "Financing the Millennium Development Goals: The Philippines." National Statistics Office (2004). Philippines: Annual Poverty Indicators Survey 2004. National Statistics Office (2007). Family Planning Survey 2006. National Statistics Office (2006). Philippines: Family Income and Expenditure Survey 2006. National Statistics Office (NSO) and Macro International (2003). Philippines Demographic and Health Survey (NDHS) 2003. National Statistics Office (NSO) and USAID (2009). Philippines National Demographic and Health Survey 2008: Preliminary Report. Orbeta, Aniceto Jr. C., 2002. "Population and Poverty: A Review of the Links, Evidence and Implications for the Philippines," Discussion Papers DP 2002-21, Philippine Institute for Development Studies. Orbeta, Aniceto Jr. c., 2005. "Poverty, Vulnerability and Family Size: Evidence from the Philippines," Discussion Papers DP 2005-19, Philippine Institute for Development Studies. Phil Health and DOH (2008), "Health Care Financing Strategy of the Philippines: 2009-2020: The Path Towards a Social Health Insurance Model." Draft Working Paper. Phil Health (2008). "Phil Health Stats and Charts 2008." www.philhealth.gov.ph Racelis, Rachel (2008). "Population and Health Expenditures in the Philippines: Projections from 2005 to 2050" (draft). University of the Philippines (financed by the World Bank). Riddell, Travis et aL (2007). "Elevated blood-level levels among children living in the rural Philippines." Bulletin o/the World Health Organization. September 2007. Rivera, Teddie (2008). "Profile of the Philippine Pharmaceutical Industry 2008" (draft), financed by the World Bank. 139 Solon, Orville etal (2008). "Associations between Cognitive Function, Blood Lead Concentration, and Nutrition among Children in the Central Philippines." Journal 0/Pediatrics. February 2008. Van Doorslaer, et aI, (2007). Catastrophic Payment for Health Care in Asia, Health Economics 16:1159-1184 Victora, Cesar, Linda Adair, Caroline Fall, et al. (2008). "Maternal and Child Under-nutrition: Consequences for Adult Health and Human Capital." The Lancet. January 2008 World Bank (2004). The Millennium Development Goals/or Health: Rising to the Challenges. Washington, DC: The World Bank. World Bank (2006). The Philippine Environment Monitor 2006. Manila, Philippines. World Health Organization, (2009) National Health Accounts, Estimates for the Philippines. 140 CHAPTER VIII EDUCATION AND INCLUSIVE GROWTH A. Introduction 8.1 Extensive empirical research has proven that investments in education, especially of the poor, are critical to sustained and broad-based economic growth. Education investments are a vital part of the virtuous circle of high and sustained economic growth, impressive gains in poverty reduction, large capital investments that are associated with more advanced and productive technologies, a more competitive domestic economy that is more closely integrated with the globalizing world, and strong institutions and a politically-empowered citizenry. In particular, historical evidence shows that countries that exhibited consistently strong economic performance in the second half of the 20th century and concomitantly drastically reduced their numbers of poor people also pursued a strategy of balanced upgrading for all levels of their education systems. Some examples are the very neighbors of the Philippines, such as Hong Kong, South Korea, Singapore, and Taiwan. 8.2 Filipinos have a good intuitive understanding of this education-economic growth- poverty reduction nexus. For this reason, education is highly valued in the Philippines by the rich and the poor alike. Viewing education as a means to get a high-paying job, many poor families pool resources from networks of extended relations to put a child through high school if not college. The child is then expected to help other siblings through school and thereby lift the family out of poverty and deprivation. 8.3 The belief that education and poverty are negative correlates is borne out by data from the 2006 FIES. As shown in Table 8.1, which classifies households by educational attainment of the household head, Table 8.1: Poverty Incidence and Share in Poor both the percentage of the poor among Households, by Educational Attainment of Household households in a given category Head, 2006 (poverty incidence) and the percentage of poor households who Educational Attainment Poverty Share in Total belong to an educational attainment of Household Head Incidence Poor Households category (the share in the total number ~N::-:::-o--:--:----------:-"-::--------::-::--- 53.9 5.7 of poor households) generally decline Some elementary 43.3 36.3 as the educational attainment of the Elementary graduate 35.3 25.4 household head rises. In other words, Some high school 29.5 13.9 the education of the household head is High school graduate 17.2 14.4 a key indicator of household poverty Some college 8.1 3.7 status. Moreover, other survey results Co llege graduate 1.6 0.6 suggest that lack of education increases the likelihood of a worker All households 26.9 100.0 being employed on less than a fu11- Source: Basic data from 2006 FIES. time basis (visible underemployment) or at a wage less than his or her target income (invisible underemployment). 141 8.4 The recent improvements in economic performance have presented an opportunity for the Philippine government to invest in education and human resources. As discussed in previous chapters, per capita GDP growth in the Philippines accelerated in recent years, reaching a record high of 5.4 percent in 2007. This growth represented an opportunity for the government to invest in education and human resources as a strategy for sustaining the growth momentum, especially in view of the long-run underperformance of the country's economy over the last half century. A key question then for the Philippine education sector is: did this recent period of economic growth yield an education dividend in the form of a more efficient or higher quality educational system, if not one that is more affordable for or accessible to children from poor and disadvantaged families? 8.5 This chapter assesses education outcomes among the poor and the performance of the education sector. The chapter begins by reviewing access to education and outcomes for the poor. It then examines expenditure issues, looking at resource allocation and utilization issues that facilitate or constrain sectoral performance as well as assessing the efficiency of resource utilization. The chapter also discusses accountability measures in basic education. Finally, it concludes with recommendations on a number of education reform initiatives. B. The Accessibility and Outcomes of Education for the Poor Basic education 8.6 Gaps in the accessibility of Figure 8.1: Percentage of Children in School, 2004 schooling persist between poor and 100.0 non-poor families and poor and non-poor areas, contributing to greater inequality. The proportion of - 95.0 i 90.0 f 85.0 children 16 years or younger who are ~ in school increases by expenditure 80.0 quintile, from 85.0 percent in the 75.0 poorest quintile to 98.1 percent in the AIIHHs 1 (Poorest) 3 5 (Richest) richest (Figure 8.1). Because the rate Expenditure Quintile of increase decreases over successive quintiles, the largest adjacent-quintile Source: Basic data from 2004 APIS. difference is between the first two quintiles at 5.7 percentage points, implying that the drop-off in schooling opportunities is most severe for children in the poorest quintile. Gaps in access are also evident between poor and non-poor areas: as shown in Figure 8.2, children in the Mindanao regions of Zamboanga, SOCCSKSARGEN, and the Autonomous Region of Muslim Mindanao (ARMM) find schooling relatively less accessible. In the ARMM, only 82.9 percent of children 16 years or younger attended school, compared to the national average of 91.0 percent and to 94.7 percent in the National Capital Region (NCR). 142 Figure 8.2: Percentage of Children in School, 2004 95.0 93.0 91.0 89.0 i 87.0 e 85.0 ~ 83.0 81.0 79.0 77.0 75.0 ...... +-L..,-.......II....,-J.....,....,---~.,...-...,...... ......II....,-J.....,........., II....,-J.....,........,.---~.,...-...,...... Region Source: Basic data from 2004 APIS. 8.7 As a consequence of differences Figure 8.3: Elementary School Completion Rate, 2004 in access to schooling, educational outcomes are heavily biased against the 60.0 50.0 1 poor. Figure 8.3, which shows the 1: 40.0 proportion of 11-13 year old children in CD each expenditure quintile who graduated e 30.0 CD D... 20.0 from elementary school, indicates that completion rates are quite low. Less than 10.0 half of children in the first three quintiles 0.0 ~ graduated from elementary school, and AIIHHs 1 3 5 (Richest) even in the richest quintile, only 55.6 (Poorest) percent of children did so. The figure also Expenditure Quintile shows that completion rates increase at a Source: Basic data from 2004 APIS. decreasing rate across quintiles, so the largest adjacent-quintile difference is between the first two (poorest) categories. The situation is even worse among 15-17 year olds: less than half of those teenagers in each expenditure quintile graduated from high school, with only 10.6 percent in the poorest quintile (Figure 8.4). The staggering 31.7 percentage point difference in completion rates between the richest and the poorest quintile implies that a 15-17 year old teenager in the richest quintile was 4 times more likely to have graduated from high school than a peer in the poorest quintile. 143 Figure 8.4: Secondary School Completion Rate, 2004 45.0 40.0 35.0 30.0 1: 25.0 ~ 20.0 ~ 15.0 10.0 5.0 0.0 All HHs 1 (Poorest) 3 5 (Richest) Expenditure Quintile Source: Basic data from 2004 APIS. 8.8 Differences in education outcomes can also be seen across different geographic regions of the country. Figures 8.5 and 8.6 suggest that elementary and secondary completion rates share similar features in terms of regional distribution. Better outcomes are concentrated in regions north of MIMAROPA (including CAR), which have completion rates at least as high as the national average, and the worst-performing regions are the Zamboanga Peninsula and ARMM. In other words, the geographic distributions of completion rates indicate that schooling in the first two education levels is more accessible and efficient in the Luzon regions (except Bicol and MIMAROPA) and much less so elsewhere in the country, but particularly in two Mindanao regions. Figure 8.5: Elementary Completion Rate, 2004 SOO 45.0 1 .. c G> 40.0 ... u Q) 35.0 Q. 30.0 25.0 20.0 I I Region Source: Basic data from 2004 APIS. 144 Figure 8.6: Secondary Completion Rate, 2004 1 - c G) ... (,) 350 30.0 25.0 20.0 a G) 15.0 a. 10.0 5.0 0.0 (/) ~ (/) >- c: Z :> :> .8 (/) c: Cl '0 ~ « () ~ c: :c m ~ aJ ~ E Cl.. >- c: m « c: .... c: .... "iii .... m ~ :x: (1) ::2 (1) N E * ...J ~ 'E (J) ~u - - 2 7 8 9 10 11 12 13 14 15 16 17 o 7 8 9 10 11 12 13 14 15 16 17 Age in years Age in years -NCR lIocos - Cagayan Valley -Central Luzon -Western Visayas ""~-Central Visayas -CALABARZON -Bicol -CAR --Eastern Visayas -MIMAROPA Source: Basic data from 2004 APIS. Source: Basic data from 2004 APIS. 146 Figure 8.8c: Age-specific Education Deficits, by Regions in Mindanao 10 1 al .5 8 '0 0 .c 6 u III 00- 0 4 ~ 111 CI) 2 >- 0 2 3 4 5 6 7 8 9 10 11 Age in years - N o Deficit - Zamboanga Peninsula - Northern Mindanao -Davao -SOCCSKARGEN --ARMM ¥~~Caraga Source: Basic data from 2004 APIS. Tertiary Education 8.11 Enrollments in colleges and universities have increased dramatically since the 1970s but have not performed as well more recently. Between SY1970-1971 and SY2005-2006, tertiary enrollment expanded at an average annual rate of3.83 percent (Figure 8.9), outpacing the growth rate of the 15-24 year old popUlation by 1.45 percentage points. This trend can be attributed mainly to rapid expansion of enrollments in state universities and colleges and suggests that colleges and universities, once the province of scions of elite and wealthy families, became more accessible to the middle class. After reaching a threshold value of around two million in SY1998-1999, Figure 8.9: Tertiary Enrollment, 1970-2004 however, enrollment 2.5 increases have since .. j 2.3 declined to 1.23 percent per 2.1 year, a pace 0.67 1.9 1.7 ]! c: percentage points slower CI) than that of the 15-24 year .E 1.5 old population. In fact, ... 1.3 '0 c: w 1.1 annual growth rates of 0.9 tertiary enrollment were 0.7 negative between SY2002- 0.5 II I 2003 and SY2004-2005. The lackadaisical trend smce 1999 indicates that School Year further progress may be difficult to achieve unless family incomes rise Source: Orbeta (2008) based on NSCB, Philippine Statistical Yearbooks of various years, as reported by thc Department of Education, Culture, and Sports, significantly with sustained and the Commission on Higher Education. economic growth or tertiary 147 education is heavily subsidized by the government. However, such subsidies are regressive, since benefit incidence analyses undertaken by Manasan and Cuenca (2007) suggest that scarce government resources have greater impact on the poor if allocated to primary schools. 8.12 Higher education, which is important for economic growth, remains out of reach for most of the poor. Figure 8.10, which presents attendance rates in tertiary institutions by income decile, shows that the attendance rate was only 24 percent in the poorest category but 46 percent in the richest. As explained by human capital theory, this disparity is due to the high indirect costs of schooling and to an inherent market failure in education financing. As soon as they are of legal age to join the labor force (if not earlier), children from poor families are sent out to work to help provide for the family. At the same time, bright but poor students are unable to mortgage their talents and potentialities to get loans that would finance their college educations due to capital market imperfections. Figure 8.10: Attendance Rate in Tertiary Institutions, 2006 8.13 The situation is not helped 50.0 J by the fact that the unit cost of 45.0 J 40.0 ; operations of colleges and ... 35.0 -1 c 30.0..j universities remains high and ~ 25.0 highly variable. As reported in ~ 20.0 15.0 Orbeta (2008), the costs of operation 10.0 per student (in nominal PhP) ranges 5.0 0.0 +""o....,...-""""",,--"'''-''''''''-'''''''''-r''''''''''''''-r''''''''''---'''''--,'''''''' from just under PhP8,000 in a medium private sectarian college outside Metro Manila to over PhP165,000 in a large private Income Decile sectarian university in Metro Manila. Source: 2006 FIES and LFS. If colleges and universities charge the full cost of their operations, tertiary education would be beyond the means of the poor. Even if the average tuition fees of PhP7, 190 and PhPl,908 for public and private schools, respectively, which were reported from a 1995 survey by the Fund for Assistance to Private Education (F APE), remained constant, a college education would still be barely within the means of poor families. 8.14 Although financial assistance programs are available, coverage is quite limited. Programs such as the Private Education Student Financial Assistance (PESF A) program and the Study Now Pay Later Plan (SNPLP) can help alleviate some of the financial burden for low- income families. 76 However, in SY2005-2006, only about 66,000 students-less than 3 percent of the 2.5 million college students-availed of these CHED-administered programs. While institution-based scholarships such as the Socialized Tuition Fee Assistance Program (STFAP) 76 PESFA provides financial assistance to incoming college freshmen who enroll in CHED-designated priority courses and whose family income does not exceed PhP120,OOO per year. SNPLP is a loan program for college students whose family income is no more than PhP150,OOO per year. However, the value of a PESFA grant or SNPLP loan is set to a maximum of only PhP7,250 per semester, which is not enough to cover the direct costs of schooling such as tuition and other school fees, books and study materials, and stipends. Recipients of SNPLP loans are also required to start repaying them at 6 percent interest 13 months after graduating. 148 of the University of the Philippines are also available, the strict entry requirements of the schools as well as the limited number of grants do not increase the number of beneficiaries significantly. 8.15 For the poor who do get to college, the future is not made much brighter by the low quality of tertiary institutions in general. Data on passing rates in professional licensure exams point to uniformly low but highly variable passing rates. For example, the passing rate for PBET went from 18 percent in 1995 to 36 percent in 2000 to 27 percent in 2005, while that for accountancy varied from 21 percent in 1995 to between 14 and 19 percent during 1990-2000 to 25 percent in 2005. In nursing, the proportion of examinees passing the licensure exam declined from 65 percent in 1985 to 52 percent in 2005. In civil engineering, the statistics were 45 percent in 1985,28 percent in 1990, and 35 percent in 2005. Technical and vocational education and training 8.16 Enrollment in technical and vocational education and training (TVET) programs has increased at a furious pace in recent years. Between 2002 and 2006, enrollment in TVET programs and the number of their graduates increased at compounded annual growth rates of 9.96 percent and 6.85 percent, respectively (Figure 8.11). However, relative to the size of the TVET constituency as may be estimated by the unemployed labor force, the magnitudes served may not yet be very significant, given that they are still below the level of tertiary enrollment. 8.17 Evidence suggests that Figure 8.11: Enrollment and Graduates in TVET Programs, TVET programs may be less 2002-2006 accessible to the poor. Ascertaining 2,000,000 the accessibility of TVET programs 1,500,000 for the poor is difficult, given the 1,000,000 heterogeneity of the target clientele. 500,000 Figure 8.12, however, provides a telling statistic: graphing regional o poverty incidence and regional share 2002 2003 2004 2005 2006 Year in total enrollment in TVET Ill! Enrollment Ill! Graduates programs, the figure shows that the two indicators are negatively Source: Lanzona (2008), based on data from TESDA. correlated. In other words, smaller shares of TVET resources were allocated to regions that needed them more. 149 Figure 8.12: Poverty Incidence and Share in TotaJ Enrollment in TVET Programs, by Region, 2006 60.0 - e c 50.0 40.0 30.0 8: 20.0 10.0 0.0 +"".......,..-..,......-...,... Source: Lanzona (2008), based on data from TESDA and 2006 FIES. c. Expenditure Issues Overall education spending 8.18 Education expenditures could have more fully benefitted from recent economic growth. As mentioned earlier, the economic performance of the Philippines improved significantly in the mid- to late-2000s. However, government expenditures on education as a proportion of GDP fell from 3.4 percent in 2002 to 2.4 percent in 2008, as did the ratio of public spending on basic education to GDP from 2.9 percent to 2.1 percent (Figure 8.13). Over this same period, the shares in national government outlays for the different levels of education did not change significantly (Figure 8.14). Over four-fifths of these national government funds went to basic education, less than 15 percent to higher education, and less than 2 percent to TVET activities. Figure 8.13: Government Expenditnres on Education as a Figure 8.14: Percent Shares on National Government Percentage ofGDP, 2002-2008 Expenditnres on Education, 2002-2008 4.00 100% 1 -I 3.00 2.00 f==============aiiSiIiI; - 80% 60% 40% 1.00 20% 0.00 ~~:::;:::::=~=::::;:::=~~~=~ 0% 2002 2003 2004 2005 2006 2007 2008 2002 2003 2004 2005 2006 2007 2008 - Total government - National government ~ Basic education · Higher education Local government - Basic education TVET III Others Source: DepEd and Department of Finance. Source: DepEd and Department of Finance. 150 8.19 From an international perspective, the Philippines has fallen further behind most of its neighbors in the region. As shown in Tables 8.2a and 8.2b, the Philippines has not compared favorably with its Asian neighbors in terms of national government spending on education. Only Indonesia has shown lower national government expenditures on education as a percentage of GNP and as a percentage of total government expenditures. Table 8.2a: National Government Expenditures on Table 8.2b: National Government Education as a Percent of GNP, 1985-2003 Expenditures on Education as a Percent of Total National Government Expenditures, 1985 1990 1995 2000 2003 1995-2003 Philippines 1.35 2.90 3.15 3.28 2.77 1995 2000 2003 Indonesia n.a. 1.04 n.a. 1.01 1.28 Philippines 16.57 17.12 17.92 Indonesia n.a. 5.37 8.45 Malaysia 6.61 5.45 5.00 6.38 7.91 Malaysia 20.94 23.70 25.51 Singapore (as of 4.40 3.01 2.98 3.98 4.23 Singapore (as of2001) 18.89 21.03 19.44 2001) Thailand 23.03 25.83 24.19 Thailand 3.79 3.59 3.59 4.55 4.13 Source: Manasan and Cuenca (2007). Source: Manasan and Cuenca (2007). Basic education spending 8.20 DepEd spending on basic education has been somewhat erratic over the past several years. Perhaps reflecting the combined effects of low budgetary appropriations, poor budget executions, and late or insufficient releases of funds from the Department of Budget and Management (DBM), DepEd spending in real terms behaved rather erratically between 2002 and 2008. It fell from PhP115.2 billion in 2002 to PhPI07.5 billion in 2005, rose to PhP122.4 billion in 2007, then declined again to PhP112.9 billion in 2008, a level even lower than in 2002 (Figure 8.15). Because public school enrollment remained at 18-19 million during the period and did not increase significantly, except in 2003 and 2008, a similar roller-coaster pattern was observed for per student DepEd spending. Per student spending by DepEd in 2008 was the lowest level observed for the entire period, at PhP5,749 in 2002 prices. Figure 8.15: DepEd Expenditures, 2002-2008 II) 125,000 ~ 6,600 0 II) 120,000 6,400 II) 0 I\) 11. 115,000 6,200 II) I\) N 6,000 11. C> 110,000 N C> N 5,800 C> C> c 105,000 5,600 N ~ 100,000 5,400 i 2002 2003 2004 2005 2006 2007 2008 Year - DepED expenditures (left axis) --OepEO expenditures per student (right axis) Source: DepEd. 151 8.21 Large regional variations in national government per pupil spending have persisted, while local government spending comprises a small proportion of basic education expenditures. Even at the peak of national government spending on basic education in 2007, estimates of this spending ranged from PhP5,517 and PhP5,405 in Central Visayas and NCR, respectively, to PhP7,403, PhP7,495, and PhP8,511 in Cagayan Valley, I1ocos, and CAR, respectively (Figure 8.16). In the NCR, LGU spending in education usually compensates for the national government shortfall, as indicated by the data for 2006. Other regions are not as fortunate, since their LGUs are unable to generate revenues as large as those of the NCR. Local governments fund a very small proportion of basic education expenditures-about 7-8 percent over the years, with a peak of 8.4 percent in 2006. Local government spending on basic education-primarily from the Special Education Fund, sourced from 1 percent of the LGU real estate tax income-has been stable at 0.2 percent of GDP. The share of the Special Education Fund in LGU spending has increased recently, reflecting a reduced dependence on the Internal Revenue allotment to finance local education. As a percentage of LGU expenditures on education, spending on basic education fell from 82 percent in 1997 to 70 percent in 2005. Figure 8.16: National Government Expenditnres per Stndent and Local Government Expenditnres per Student, by Region Q 10000 Q Q 8000 N 6000 .5 4000 II) 0 2000 II) III a. 0 c:: 0 II) 8 ~ >. c: 0 z « 0 a. (5 t.l ... c: III II) ro .... O'l r.:: 0 en « O'l c: eli III c: Q) > ro « c:: ro g m N >. ~ z ~ N 0 in 1i) ro 1i) ro ..c: ro ~ 0 ro > ...J 0:: c:: II) ro t 0 (/) ~ :> 0 ro m co c: .... « « UJ .Q E 0 z 0 0 0 ~ ro c >. O'l III ro 0 0 « :s:E li! C III ro N 0 en 0 0 !liNG 2007 IILG 2006 liNG 2006 Source: DepEd. Technical and vocational education spending 8.22 The Technical Education and Skills Development Authority (TESDA) receives approximately 2.1 percent of the national education budget. Peano, et a1. (2008) estimated the national expenditure for TVET activities to be approximately PhPll.1 billion (roughly US$216 million) or 0.3 per cent of GDP in 2002. Historically, TVET has relied heavily on government funds since the private sector, specifically the companies, are quite reluctant to invest in training. A breakdown of expenses indicates that much of these government expenditures have been devoted to government direct delivery of training. The share of LGU expenditures going to TVET increased from 1 percent in 1997 to 2.7 percent in 2003. Tertiary education spending 152 8.23 State universities and colleges (SUes) are financed largely by national government subsidies, although this dependence has been declining. Data from 1998 to 2004 indicate that SUCs received an average annual allocation of about 14 percent of the total government budget for education-related organizations. More recent data on SUC finance shows a declining dependence on subsidies and an increasing reliance on internally-generated funds, with the share of government subsidies falling to about 84 percent in 2002 and down further to about 73 percent in 2006. The declining dependence on government subsidy may be attributed to the policy on limited financial autonomy.77 At the same time, the contribution from tuition fees and other income sources grew to as much as 11 percent and 14 percent, respectively, in 2006. The share of grants and donations remains negligible and erratic, increasing from 0.5 percent in 2002 to 2.4 percent in 2004 then falling to 1.2 percent in 2006. LGU expenditures on education have gone increasingly to tertiary institutions, with the share increasing from less than 1 percent in 1997 to 7 percent in 2005 due to the establishment ofLGU-funded universities and colleges. Efficiency of resource utilization 8.24 Although the budget allocation for recurrent expenditures has increased, inefficiencies in the utilization of resources persist. In the last two years, DepEd has commendably increased its budget allocation for recurrent expenditures, which enables it to implement interventions aimed at improving teaching and learning. Nonetheless, it appears that resource management and utilization remain inefficient. To establish the extent of delay in executing a particular budget item and identify bottlenecks, a World Bank Budget Execution Study of large items in the 2007 DepEd budget tracked the timing of fund releases from the Department of Budget and Management (DBM) to the DepEd Central Office or Regional Office and eventually to the frontline units. The study found that slow execution of some priority programs hampers the Department's ability to deliver quality education services efficiently. Figure 8.17 shows an erratic pattern of budget execution in four priority program/activity/project (PAP) items. Analyses carried out thus far indicate that much of this erratic execution pattern is due to various institutional and managerial weaknesses. 77 The Higher Education Act of 1997 (RA 8292) gave State Universities and Colleges limited financial autonomy, allowing them to retain all generated income from tuition and service fees from auxiliary services. 153 Figure 8.17: Utilization Rate for Selected DepEd p rOl!rams 137.8% 140.0% 120.0% 103% 103% ~ 100.0% - - .~ ~ 80.0% C ~ 60.0% 55% 53% ., D.. 47.4% 44.0%4 2. % 44.8°'------ r-- ':"" 40.0% 20. % 20.0% H 0.0% I :: 0.6% 0.0% Deployment of School Furniture Textbooks (MOOE) School Building New Teachers (PS) Program (CO) Program (CO) []2005 ~2006 02007 Source: DepEd Budget Division. D. Accountability Issues - A Focus on Basic Education 8.25 In the past, weak governance greatly hampered the performance of the basic education sector in the Philippines. There was limited provision for disadvantaged communities, little empowerment of local school communities, and favorable treatment of selected communities that received strong political support. With the slow implementation of decentralization, decision-making in education remained largely regulated by central prescriptions, and no clear accountability structure was in place. DepEd's existing structures and agencies, including regions, divisions, and schools, lacked the capacity to operate successfully in a more decentralized environment. Moreover, resource management was inefficient, and schools were unable to respond flexibly to local needs. 8.26 Recognizing these shortcomings, successive DepF;d administrations have progressively refined a set of reform strategies that have culminated in the school-based management initiative. The reform strategies have aimed to make the system more responsive to local needs and demonstrate greater commitment to improved outcomes. By 2005, there was widespread consensus on the nature of reforms that would empower local communities to improve schools and monitor results. Efforts were initiated to decentralize the administration of education to the grassroots level through school-based management (SBM), which was piloted and implemented in public elementary and secondary schools in 2000 and is now being rolled out on a phased basis nationwide through the government's Basic Education Sector Reform Agenda. The SBM model has forged partnerships between schools and communities that chart plans to improve learning outcomes and school performance, mobilize resources for reforms, and foster a culture of ownership and responsibility among local stakeholders, thus enhancing accountability. Support for SBM has been validated by the international literature on school improvement and by successful local pilots on SBM. 154 8.27 SBM has been operationalized by decentralizing decision-making authority from central, regional, and division levels down to individual schools. The aim has been to unite school heads, teachers, students and their parents, the LGUs, and the community at large in promoting effective schools and achieving desired learning outcomes. In SBM, decision-making is made by all those who are closely involved with resolving the challenges of the individual schools, so that the specific needs of the students can be addressed more effectively. The premise is that the people who have the most to gain or lose and the people who have the best information about what actually goes on in schools are the ones most able to: (a) make the right decisions about how schools should use scarce resources to deliver essential services, (b) monitor how effectively and efficiently these resources are being utilized, and (c) evaluate the impact of public policy and programs. In effect, the SBM framework acknowledges that since needs and problems vary among schools, central authorities cannot know the right solution for every school. It thus recognizes that involving local stakeholders in addressing local problems is key to improving schools and even to mobilizing much-needed resources. 8.28 A crucial component for fostering accountability in SBM is the system of checks and balances. Under this approach, decisions made collaboratively are affirmed by other bodies, implementation processes are duly reported and funds are accounted for, and results and outcomes are assessed and reported relative to previously set and agreed upon standards and targets. The accountability system has three components: (a) A School Improvement Plan (SIP) is developed through a collaborative process involving all stakeholders, including the school head, teachers, students, representatives from parent-teacher-community associations, LGUs, NGOs, civil groups, and other interested parties. The plan is confirmed by the school and community in a public assembly that also establishes their commitment to its implementation. At the end of every school year, a School Report Card, which assesses implementation and outcomes, is made at a public assembly, after which a review of the SIP is made as a basis for the next year's plan. (b) A School Operating Budget is drawn up, and funds are allocated, mobilized, and committed by the different stakeholders according to the priorities of the SIP. At the end of every school year, inflows and outflows of financial resources are reported at the same public assembly where the School Report Card is presented. (c) The performances of the school and students are tracked, monitored, and reported by key stakeholders, and become the basis of the School Report Card. 8.29 In schools where the SBM model has taken root, stakeholder participation has made basic education delivery more responsive to the needs of the poor and disadvantaged sectors, such as indigenous peoples. Options for enhanced Alternative Learning Systems have even been explored. Spillover effects can also be seen: the collective responsibility that local stakeholders have developed to harness resources for education has expanded to provide for health, nutrition, and early childhood development needs. By giving schools and the communities they serve the right to prioritize resource use and by instituting accountability mechanisms for both inputs and outcomes at the school level, the SBM model has engendered participative monitoring and evaluation by the local community, thus providing strong assurance for relevance, flexibility, transparency, advocacy, and accountability. 155 8.30 The benefits of the SBM approach also go beyond community empowerment. The results of a survey of teachers and principals in SBM pilot schools indicate that teacher motivation and enthusiasm are higher as a result of their greater involvement in planning and managing school improvement. In-service training of teachers linked to SBM has also sparked teacher interest in continuous learning and professional development. Principals' efforts to fully involve teachers in identifYing problems and needs in the schools and improving channels of communication have led to appropriate solutions. At the same time, marked shifts in school- community partnerships have taken place in several school divisions. Creative, varied, and cost- effective responses to school-specific problems and situations were undertaken in the SBM pilot schools. Strong support from parents, communities, LGUs, and other stakeholders in school planning, management, school activities, and meetings appears to have increased student attendance, produced better school improvement plans, and mobilized more resources for schools. These stronger school-community partnerships have also catalyzed the development of the communities outside the schools. 8.31 The SBM model appears to have had good payoffs in terms of student performance. An initial assessment of the effects of SBM on .student performance in the Philippines, undertaken by World Bank senior evaluation officers in 2008, estimated the effects of the SBM model on student performance using the administrative dataset of the SBM pilot schools in the 23 poorest public school divisions of the country over the three-year period from 2003 to 2005. The results suggest that the introduction of SBM had a statistically significant, albeit small, overall positive impact on average school-level test scores in these pilot school divisions. The improvement in schools that received SBM early was significantly higher in science and English and on composite test scores. A simple difference-in-difference comparison between SBM and non-SBM schools also showed that test scores increased more rapidly among SBM schools. E. Policy Implications and Conclusions 8.32 To help increase school enrollments, particularly among the poor, demand-side interventions and scholarship programs should be expanded. Conditional cash transfer programs, which provide transfers to families who fulfill certain criteria such as school attendance for children, can be an effective demand-side intervention to help improve the human capital of the poor and thereby lower inequality. An effective targeting system is critical to the success of such programs. To help reduce the financial burden of schooling for poor households, effective scholarship programs and student loan programs--especially for the poorest income quintile-should also be expanded. 8.33 The level and efficiency of education expenditures need to be improved to enable the provision of quality education services around the country. Greater real per capita spending in basic education is needed to help reduce inequality between the poor and non-poor. At the same time, the limited public funding for tertiary education should be used more strategically. In general, new public investments should be better targeted to benefit the poorest. Special attention is needed to ensure that adequate resources reach schools in a timely manner-the continued decentralization of resources to schools, for example for school maintenance and other operating expenses (MOOE) and school-based management grants, would help in this regard. The equity- and poverty-based formula for MOOE allocation should be applied immediately. 156 8.34 Further involvement of local communities and the private sector would also help improve education service delivery. Given the achievements of SBM to date, it should be implemented as quickly as possible, which would require changes in attitudes and actions at all levels, including the need to move away from regulation by central prescriptions. Where available, private sector capacity should also be utilized in cases where it has been shown to be more effective but under an improved quality assurance framework. 8.35 To help address the problem of skills gaps, which is critical for raising economic growth, higher education and TVET should be made more responsive to the needs of the labor market. As a starting point, the quantity and quality of information on the labor market need to be improved, with better and more complete firm and labor force surveys. Close linkages between post-secondary and tertiary education and industries, for example through greater collaboration in curriculum design, training, and research and development, would also help make education more responsive to the demands of employers. In addition, the skills supply system needs to be better articulated through strengthened skills certification and through an education and training quality assurance and accreditation system. 157 References Commission on Higher Education (the Philippines). Statistical Bulletin. Various years. Department of Education. Statement of Appropriations, Allotments and Obligations. Various years. Department of Education. Basic Education Information System. Various years. Department of Finance, website www.dof.gov.ph Lanzona, L (2008). "Technical, Vocational Education and Training in the Philippines." Prepared as background study for this report. Manasan, R. and J. Cuenca. (2007). Benefit Incidence of Public Spending in Education. PIDS. Makati City. National Statistics Coordination Board, Philippine Statistical Yearbook, various years. National Statistics Office (2004). Philippines: Annual Poverty Indicators Survey 2004. National Statistics Office (2006). Philippines: Family Income and Expenditure Survey 2006. National Statistics Office (2006). Philippine Labor Force Survey 2006. Orbeta, A. (2008). "Background Paper on Higher Education." Prepared as background study for the Philippine Skills Report 2009. Peano, Serge, Benjamin Vergel de Dios, David Atchoarena, and Ursula Mendoza (2008). "Investment in Technical Vocational Education and Training (TVET) in the Philippines." United Nations Educational, Scientific and Cultural Organization, International Institute for Educational Planning. Technical Education and Skills Development Authority (2008). TESDA Statistics. Various years. 158 CHAPTER IX SOCIAL PROTECTION AND INCLUSIVE GROWTH A. Introduction 9.1 Social protection has been viewed as a way to promote equity and growth by assisting the poor to manage risk and safeguard their human capital. In the face of health, labor, price, and other idiosyncratic and macroeconomic shocks, poor households are bound to adopt coping strategies that can impair households' human capital or hinder further investment, leading them to a vicious cycle of increased destitution. For instance, falling income among poor households leads to increased child malnutrition with long-lasting consequences on child growth, cognitive and learning ability, and schooling attainments (Alderman et al, 2006; Ferreira and Schady, 2008); higher dropout rates with long-term consequences for the children's labor market performance (Ferreira and Schady, 2008); and selling productive assets (livestock and land) which impedes households' recovery from the shock (Fafchamps et aI., 1998; Carter et aI., 2004). 9.2 The lack of progress in poverty reduction and the human and social impacts of the recent food, fuel, and financial crises, as well as the typhoon disaster that followed, have underscored the need for social protection to assist the poor and vulnerable. As in other countries in East Asia and the Pacific, the negative effects of the 1997-98 East Asian financial crisis drew greater attention to the problem of vulnerability in the Philippines and the need to protect poorer households from shocks through social safety net programs and other measures. Since then, rising poverty incidence, the adverse impacts of the food and fuel price shock in 2008, and the socioeconomic effects of the global economic crisis on the Philippines, and the disaster that followed the onset of two typhoons in September and October 2009, have renewed the attention to social protection services. If properly designed and implemented, social protection can help promote inclusive growth by preventing the poor from falling into vicious cycles of increased destitution. Without such support, households may respond to shocks by engaging in negative coping strategies such as pulling children out of school or selling productive assets, which may provide short-term relief but will ultimately push them further into poverty. 9.3 The large proportion of the popUlation living near the poverty line also points to the importance of social protection. Raising the poverty threshold to an international poverty line of US$2 per day reveals that an additional 19.4 percent of the population in 2006 would be classified as poor.78 The National Anti-Poverty Commission (NAPC) and the National Statistical Coordination Board (NSCB) estimated that around 45 percent of Filipinos face the risk of falling into poverty. In the Medium-Term Philippine Development Plan 2004-2010, the government also reported that while 22 percent of the population is chronic poor, 32 percent moves in and out of poverty (NEDA, 2004). Shocks that commonly push vulnerable households into poverty in the 78 World Bank estimates using the Family Income and Expenditure Survey in 2006. 159 Philippines are those related to health, employment, natural disasters, civil unrest, and food prices (World Bank, 2001). 9.4 Although the Philippines has numerous social protection programs in place, their effectiveness has been compromised by targeting, financing, and management problems. The largest share of social protection resources is dedicated to programs aimed at addressing price and income shocks, such as price subsidies, food transfers, and cash transfers, including those that support livelihood and other community-based programs. Other major programs focus on protecting households from health shocks, natural disasters, and labor market shocks. Although the attention to social protection is commendable, the lack of a legitimate and functional system for targeting has undermined the effectiveness of programs in helping the poorest and most vulnerable households. These efforts have also been compromised by the limited resources for social protection and by weak policy and institutional coordination. Recently, the government has increased its focus on reforming and reinforcing its social protection system and has taken some positive steps in addressing the major lapses in social protection policies. To assist in the establishment of a common framework for a national social protection strategy, the government issued an operational definition of social protection and the types of programs covered therein (Box 9.1). Box 9.1: Definition of social protection in the Philippines "Social protection constitutes policies and programs that seek to reduce poverty and vulnerability to risks and enhance the social status and rights of the marginalized by promoting and protecting livelihood and employment, protecting against hazards and sudden loss of income, and improving people's capacity to manage risks. The components of social protection are: · Labor market programs. Measures aimed at enhancing employment opportunities and protection of the rights and welfare of workers. Employment-enhancing measures include trade policies and skills development and training. Labor protection includes compliance with labor standards such as minimum wages or health and safety in the workplace. · Social insurance. Programs that seek to mitigate incom.e risks by pooling resources and spreading risks across time and classes. These are designed in such a way that beneficiaries pay a premium over a given period of time to cover or protect them from loss of income and unemployment as a result of illness, injury, disability, retrenchment, harvest failure, maternity, old age, etc. This component includes micro- and area- based schemes to address vulnerability at the community level (such as micro-insurance, agricultural insurance, and social support funds). · Social welfare. Preventive and developmental interventions that seek to support the mmlmum basic requirements of the poor, particularly the poorest of the poor, and reduce risks associated with unemployment, resettlement, marginalization, illness, disability, old age, and loss of family care. Social welfare and assistance programs usually comprise direct assistance in the form of cash or in-kind transfers to the poorest and marginalized groups, as well as social services including family and community support, alternative care, and referral services. · Social safety nets. Stop-gap mechanisms or urgent responses that address effects of economic shocks, disasters, and calamities on specific vulnerable groups. These are measures that specifically target affected groups with the objective of providing relief and transition. Measures include emergency assistance, price subsidies, food programs, employment programs, retraining programs, and emergency loans." Source: NEDA, www.neda.gov.ph. 160 9.5 This chapter reviews the current challenges in providing social protection in the Philippines and lays out priorities for improving the design and implementation of programs. The chapter begins by reviewing the major social protection programs currently under implementation and social protection expenditures, assessing the .level, allocation, and efficiency of spending. It then assesses the efficiency of programs in terms of coverage and targeting performance. Accountability issues are also discussed. Finally, the chapter concludes with some policy recommendations for improving the effectiveness of social protection programs in reaching the poor and vulnerable. B. Overview of Social Protection Programs and Expenditure Issues 9.6 Government spending on social protection has increased in response to the food price shock in 2008. When the sharp spike in global rice prices combined with increases in fuel prices hit the Philippines in mid-200S, the government increased its spending on social protection to nearly three times the spending in 2007. The government increased social Table 9.1: Government Spending on Social protection spending through programs that Protection Programs, 2007-2008 already had allocated additional funds in the 2007 2008 Total Social Protection 2008 budget. In addition, more funds were ( PhP, millions) 28,214 80,418 allocated in the course of the year to a series of of national budget 2.4 6.6 safety net programs including food, nutrition, breastfeeding, and other social assistance Source: Manasan (2009). programs. As a result, the total amount of government spending on social protection rose from PhP28.2 billion in 2007 to 80.4 billion in 2008, where the latter accounted for 6.6 percent of the national budget and 1.1 percent of GDP (Table 9.1). 9.7 However, government spending on social protection during non-crisis times is relatively low. The "correct" amount of spending on social protection is debatable and depends on numerous factors such as poverty levels, the need for social protection programs, and the ability of the government to meet those needs given overall resource constraints. Nevertheless, social protection spending in the Philippines for a relatively stable year-meaning a non-crisis year such as 2007-seems quite low at 0.4 percent of GDP, especially compared to spending in neighboring countries at similar levels of economic development (Figure 9.1). A cross-country analysis of patterns of government social protection spending from 1972 to 1997 shows that, as a percentage of GDP, Malaysia spent 1.2 percent, Indonesia 0.9 percent, Thailand 0.6 percent, and Singapore 0.5 percent. Meanwhile, social protection spending in East Asia and the Pacific (EAP) is, on average, about 2.4 percent of GDP, which is relatively lower compared to Latin American Countries (LAC), which is 2.9 percent (Besley et aI, 2003). Another study covering 87 developing and transition countries during 1996-2006 shows that mean spending on safety nets is 1.9 percent ofGDP (Weigand and Grosh, 2008). 161 Figure 9.1: Government Spending on Social Protection, 2007 3.0 D.. g 2.0 '0 1.0 'ifl. 0.0 Latin South Malaysia Indonesia Thailand Singapore Philippines* American Asian Countries Countries Sources: Besley, et aL (2003) and Manasan (2009). 9.8 The largest share of social protection spending in the Philippines is directed toward programs to address price and income shocks. Social protection programs such as price subsidies, food transfers, and cash transfers-including those that support livelihood and other community-based programs-account for the largest share of government spending on social protection. These programs accounted for 84 percent of government social protection spending in 2007 and expanded to 93 percent in 2008 (Table 9.2). The imperative to protect poorer households from the short-term price and income shocks called for measures to increase transfers and subsidies to these households as quickly as possible. Programs aimed at protecting households from health shocks, natural disasters, and labor market shocks account for the rest of social protection spending. Table 9.2: Estimated Government Spending on Social Protection Programs, 2007 and 2008 2007 2008 Programs by Shock Pesos Pesos dist. dist. {millions} (millions} Price and Income Shocks 23,659 83.9 74,832 93.1 NF A rice price subsidy 16,200 57.4 60,880 75.7 Food transfers 4,098 14.5 4,347 5.4 Cash transfers 50 0.2 6,197 7.7 Livelihood programs 251 0.9 160 0.2 Community-based programs 1,962 7.0 2,348 2.9 Programs for special needs 1,098 3.9 900 l.l Natural Disasters 1,265 4.5 986 1.2 Disaster relief 263 0.9 326 0.4 Disaster rehabilitation 1,002 3.6 660 0.8 Health Shocks 2,100 7.4 2,200 2.7 Health insurance for indigents 2,100 7.4 2,200 2.7 Labor Market Shocks 1,188 4.2 2,401 3.0 Skills enhancement programs 510 1.8 1,350 1.7 Labor and employment assistance programs 678 2.4 1,051 1.3 Total 28 2214 100.0 80!418 100.0 Source: Manasan (2009). 9.9 The rice price subsidy administered by the National Food Authority (NFA) has been the most significant transfer program to date, consistently capturing a substantial slice of government social protection spending. The NF A rice subsidy has been a mainstay in the government's portfolio of social protection interventions. In 1998, the government's response to the East Asian financial crisis focused primarily on ensuring an adequate supply of rice, so in that year, NF A rice releases jumped to over 20 percent of total consumption, more than double 162 the 1997 level of 8 percent. Two years later, the NF A rice subsidy accounted for around 97 percent of total government spending on social protection (Manasan, 2006). Although its share has declined since then, the program still accounts for the largest share in spending. During the food and fuel price shock in 2008, the government tried to protect the poor from high rice prices by releasing NFA rice stocks, especially to areas that showed the biggest price increases. At the same time, it sought to build its inventory further, mainly through large-scale buying from international rice market. Government spending on the NFA rice subsidy in 2008 reached around PhP61 billion, accounting for 76 percent of government social protection spending during that year and nearly four times the spending in 2007. 79 9.10 Other food transfers also comprise a major share of social protection spending. Other food transfers are mostly in the form of supplementary feeding programs undertaken in schools, such as the Food-for-School Program (FSP) and the Malusog na Simula Yaman ng Bansa (see Box 9.2 for description). These programs provide meals for children at school to encourage their enrollment and to improve their nutritional status and ability to pay attention in class. They can also be given as transfers in the form of food to take home. Such programs accounted for 14.5 percent of total expenditures on social protection in 2007, but the share relative to other social protection programs fell to 5.4 percent in 2008 (Manasan, 2009). 9.11 Cash transfer programs are becoming more important in social protection expenditures. Cash transfers such as the Pantawid Pamilyang Pilipino Program (4Ps), Pantawid Kuryente, and Tulong Para Kay Lolo at Lola accounted for approximately 8 percent of the social protection expenditures in 2008, which was significantly higher compared to the spending level in 2007 (Manasan, 2009). Cash transfers increased substantially in 2008 as the government relied on the cost-effectiveness and flexibility of cash to deliver quick assistance to help poorer households cope with the food and fuel crisis. The Pantawid Kuryente and Tulong Para Kay Lolo at Lola were introduced only in 2008, specifically for the food and fuel crisis (Box 9.2). The 4Ps started as a pilot conditional cash transfer program in 2007 but was accelerated in 2008 to address the food and fuel crisis. 9.12 Community-based programs, which are considered to be major poverty reduction projects of the government, comprise a relatively smaller share of social protection spending. The government allocated 7 percent of its social protection spending in 2007 for community-based programs, which mostly covers the Kapit-Bisig Laban sa Kahirapan- Comprehensive and Integrated Delivery of Social Services (KALAHI-CIDSS), a community- driven development project, which aims to reduce poverty by empowering communities and improving local governance. Although the spending increased in absolute terms in 2008, its relative importance compared to other social protection programs declined during the food and fuel crisis. It accounted for only 3 percent of total social protection spending in 2008 (Manasan, 2009). 9.13 Notwithstanding the need to promote and protect human capital, labor market programs only account for the third largest share of total social protection expenditures. 79 These figures refer to the total financial cost ofNF A operations to the national government (operational and off- budget subsidies). However, they do not include the tax expenditures, which are implied subsidies provided by the national government to cover the tariff imposed on NFA imports of rice (Manasan, 2009). 163 The Department of Labor and Employment (DOLE) and Technical Education and Skills Development Authority (TESDA) implement most of the labor market programs, which altogether accounted for 4 percent of social protection spending in 2007 and 3 percent in 2008 (Manasan, 2009). These programs include: (i) employment facilitation and job search assistance such as the Public Employment Service Offices (PESOs), the electronic job matching system also known as Phil-JobNet, job fairs, and vocational guidance and employment counseling; (ii) direct employment generation services such as the Special Program for Employment of Students (SPES), Work Appreciation Program (WAP), Kasanayan at Hanapbuhay (KasH), and Tulong Alalay sa Taong May Kapansanan; and (iii) livelihood and entrepreneurship through the Self- Employment and Entrepreneurship Development (PRESEED) program (Orbeta, 2009). As part of its mandate, DOLE is also responsible for wage regulations (e.g., minimum wage, severance payments) and labor standards to protect workers and promote their welfare. Box 9.2: Key food and cash transfer programs in the Philippines Food transfers are intended to provide access to food for vulnerable and food-insecure households. They are designed to address food deprivation and malnutrition, .increase purchasing power, and/or provide livelihoods. Food-based transfers augment the real income of households, but unlike cash transfers, they tend to have a greater impact on food consumption because of the transaction costs involved in converting food to cash. · Food-for-School Program (FSP) is a conditional in-kind transfer that aims to mitigate hunger of poor families and improve school attendance by providing food subsidies to public school children enrolled in Grade I, pre-school, and day care. Launched in November 2005, the FSP provides a daily ration of 1 kilogram of iron-fortified rice per day to families to fulfill daily food requirements, to eventually help improve children's nutritional status. · Malusog na Simula Yaman ng Bansa Program is a breakfast feeding program that aims to improve the active learning capabilities of school children by providing breakfast to Grade I pupils in selected schools in the form of specially formulated noodles enriched with eggs. The program was launched in 2005. Cash transfers aim to increase the real incomes of poor and vulnerable households. They include provision of assistance in the form of cash and other instruments almost like cash that can be used to transfer resources to the poor or those who, in the absence of the transfer, face a probable risk of falling into poverty. Cash transfers can also be provided subject to compliance with specific conditions related to education and/or health (i.e. conditional cash transfer). · Pantawid Pamilyang Pilipino Program (4Ps) is a conditional cash transfer program that aims to break the cycle of poverty by addressing the issue of low educational achievement, high levels of maternal and infant mortality rates, high malnutrition, and high rates of child labor among the poor. It provides cash transfers to qualified households ofPhP500 per household per month as a health grant and PhP300 per school child aged 6-14 years in the household (up to 3 children) per month as an education grant, on the condition that they send their children to school and to health centers for check-ups and vaccinations. Initiated in 2007 and rolled out on an accelerated basis in 2008, the 4Ps is targeted to benefit around 700,000 households for five years. · Pantawid Kuryente Program provides a one-time cash subsidy of PhP500, targeted to benefit 6 million · households that utilized 100 kilo-watt hours or less electricity in May 2008 (lifeline consumers). It aimed to cushion the impact of the high cost of electricity in 2008 and was funded from oil VAT collection. · Tulong Para Kay Lolo at Lola Program provides a one-time cash subsidy of PhP500 to one million senior citizens aged 70 and above who have no regular income and are not covered by social security or any other government benefit. The program was also funded from oil VAT collection. Source: Manasan (2009) 9.14 Social protection against health shocks refers for the most part to the health insurance for the indigents. As described in Chapter VII, the Philippine Health Insurance Corporation (PhiIHealth) administers the Indigent Program, which is a non-contributory 164 insurance program intended for indigents or the marginalized sector. The health insurance premium for a year's coverage is paid jointly by the national government and local government units (LOUs) where the indigent member resides. Alternatively, the premium contributions for enrolled indigents may be paid through a partnership between the national government and other sponsors such as philanthropic private corporations or private individuals. In 2008, the national government allocated PhP2.2 billion for the Indigent Program, which accounted for about 3 percent of total government spending on social protection (Manasan, 2009). 9.15 Given the number of occurrences of natural calamities in the Philippines, the national government spends relatively less on assistance programs aimed at mitigating the impacts of natural disasters. Due to the country's geographical location, the Philippines is highly prone to natural disasters such as earthquakes, volcanic eruptions, tropical cyclones, and floods, making it one of the most disaster-prone countries in the world (World Bank and NDCC, 2004). In September/October 2009, the country was struck by two strong typhoons that caused damages and losses estimated at 2.7 percent of ODP. In 2008, national government allocated only 1.2 percent of total social protection expenditures for natural disasters. The most important of these programs in terms of budget allocation is disaster relief and rehabilitation. Most funds for disaster relief and rehabilitation went to shelter assistance (73 percent) followed by relief assistance (21 percent), with the rest going to administrative costs, cashlfood-for-work, and livelihood assistance (Manasan, 2009). It should also be noted that social protection programs to mitigate shocks from natural disasters rely on a reactive approach rather than a more effective proactive approach, wherein disasters are avoided through land-use planning, construction, and other preventive measures. Existing programs tend to focus on post-disaster relief and short-term preparedness (i.e. forecasting, evacuation) rather than on mitigation or post-disaster support for economic recovery (World Bank and NDCC, 2004). 9.16 The allocation of funding across social protection programs is not yet considered to be optimal. As described above, the social protection system in the Philippines has relied heavily on in-kind or food transfers, with rice and food subsidies accounting for 70-80 percent of social spending in 2007 and 2008 (Table 9.2). Even though they involve higher implementation costs, food transfers continue to be the main social protection instrument. At the same time, resources for programs protecting the poor from shocks related to health, employment, and natural disasters have been limited, oftentimes leaving the poor to rely on their own devices when such shocks occur. However, the recent shift toward CCT programs, which provide social assistance and encourage formation of human capital as a means of breaking the intergenerational cycle of poverty, is a promising development. 9.17 Lack of policy coordination in social protection also reduces the efficiency of sector spending. As described above, the existing social protection system has a range of subsidy and transfer programs that are administered across multiple agencies and in an ad hoc manner. This fragmentation of programs reduces the overall efficiency of spending on social protection. It also prohibits the adoption of a systematic strategy for addressing the risks and vulnerabilities of the population and makes it difficult to develop a more coherent social protection response to crises. 165 C. Coverage and Targeting of Social Protection Programs 9.18 The development impact of the most significant social protection programs is severely limited by low coverage of the poor. International experience shows that directing resources toward the poor or vulnerable can increase the benefits that can be achieved within a given budget or can achieve a given impact at the lowest cost (Grosh, et aI., 2008). However, a number of social protection programs in the Philippines, particularly those that comprise the largest portion of government spending, are characterized by exceptionally high leakage rates- that is, a large proportion of those reached by the programs are classified as non-poor. As shown in Table 9.3, leakage rates for the major social protection programs range from 20 to 72 percent. Design weaknesses and lack of a systematic method to target program beneficiaries often lead to undercoverage of the intended beneficiaries and inclusion of non-eligible individuals or households. Table 9.3: Estimated Leakage Rates of Selected Social Protection Programs Selected Programs Intended Beneficiaries Leakage Rate · Pantawid Kuryente Program Households with electricity consumption of no more than 72 100kwh (lifeline power consumption level) in May 2008 · Food-for-School Program Poor households in selected geographic areas with public school children enrolled in accredited day-care centers, 59-62 pre-school, and Grade I · Tulong Para Kay Lolo at Persons 70 years old or older with no income or 61 LolaProgram retirement benefits · PhiIHealth Indigent Program Indigent households 40-50* · NFA rice price subsidy Poor households nationwide 41* * World Bank estimates. Source: Manasan (2009) 9.19 The rice price subsidy, which consumes the largest share of government spending on social protection, is poorly targeted and undercovers the poor. The NF A rice intervention is a universal consumer price subsidy which by design also benefits the non-poor. As shown in Figure 9.2, the poorest income decile consumes only 14 percent of the total NFA rice subsidy, while around 2 percent of the subsidy goes to the richest income decile. In addition, the geographic distribution ofNFA rice is not sensitive to poverty incidence. For instance, the shares of NF A rice in total rice consumption in Western Visayas (5.9 percent), ARMM (7.1 percent), and Cagayan Valley (9.9 percent) in 2006 were low, although poverty incidence in these regions was high (Table 9.4). 166 Figure 9.2: Incidence ofNFA Rice Subsidy by Table 9.4: NFA Rice Distribution and National Income Deciles Poverty Incidence 0 NFA rice, of 10th Poverty Regions total incidence consumption 9th 8th Philippines 20.4 32.9 National Capital Region 40.5 lOA 7th Iloeos Region 13.1 32.7 Cagayan Valley 9.9 25.5 6th Central Luzon 2004 20.7 Southern Tagalog 12.0 36.8 5th Bieol Region 24.5 51.1 Western Visayas 5.9 38.6 4th Central Visayas 295 3504 Eastern Visayas 19.6 48.5 3rd Western Mindanao 30.3 45.3 Northern Mindanao 26.5 43.1 2nd Caraga Region 20.0 52.6 1st Southern Mindanao 2704 36.6 Central Mindanao 11.0 40.8 ARMM 7.1 61.8 o 5 10 15 20 CAR 24.3 34.5 ARMM Autonomous Region of Muslim Mindanao CAR Cordillera Autonomous Region Source: World Bank estimates using FIES 2006. Source: World Bank estimates using FIES 2006. 9.20 Although attempts at improving the targeting of NFA rice distribution have increased coverage of the poor, the leakage rate remains high. Over the years, numerous attempts have been made to reduce leakage rates in NF A rice distribution by introducing mechanisms that will better target poor households. For example, in 2005, the NFA implemented the Tindahan Natin Program (TNP), a small rolling store where NFA rice is sold and only eligible TNP household beneficiaries identified by the Department of Social Welfare and Development (DSWD) may purchase food items. However, given the geographic distribution of the TNP stores across the country as well as the number of beneficiaries served by these stores, the leakage rate of the TNP is estimated at 66 percent. In 2008, the government issued Family Access Cards (FACs) to families with income below the food threshold (PhP5,000 per month) that entitle them to buy two kilograms of NF A rice a day from accredited NF A outlets. As the FAC targeting approach is an unverified means test, it is estimated that the leakage rates are 57 percent relative to overall poverty incidence (Manasan, 2009). 9.21 Meanwhile, the targeting mechanism used for the FSP shows higher leakage rates than if FSP had directly targeted the poorest municipalities. The FSP is among the limited number of social protection programs that employ geographic targeting to select beneficiaries. The FSP made use of the Food Insecurity and Vulnerability Information Mapping System (FIVIMS) to identify the areas to be covered by the program. 80 While geographic targeting is administratively simple and inexpensive to implement, FSP's targeting method can still be improved to increase the program's efficiency and effectiveness (Manasan and Cuenca, 2007). 80 See Manasan and Cuenca (2007) for details on FIVIMS. 167 An analysis of the targeting scheme used for the first cycle of the program in SY 2005-2006 found that the ranking of municipalities according to their income class did not correlate well with their ranking according to small area estimates (SAE) of poverty incidence. A counterfactual simulation indicates that if FSP directly targeted the municipalities according to SAE, the leakage rates and undercoverage rates would have been reduced, while the share of benefits going to the poor would have increased (Figure 9.3). Figure 9.3. Food-for-School Program Performance Using Alternative Targeting Methods DepEd Component DSWD Component 100 r-------------------------~ 100 r-------------------------~ 80 80 60 60 40 40 20 20 o Leakage rate Undercoverage % benefits that o rate go to poor Leakage rate Undercoverage % benefits that rate go to poor .Using FIVIMS (existing) [JUsing SAE (alternative) .Using FIVIMS (existing) [J Using SAE (alternative) Source: Manasan and Cuenca (2007) 9.22 The emergency cash subsidies that were intended to soften the impact of the food and fuel crisis on incomes of the poor benefited less than 10 percent of the poorest households. As described in the previous section, during the food and fuel crisis in 2008, the government released funds in the form of one-time cash subsidies such as the Pantawid Kuryente Program and Tulong Para Kay Lolo at Lola. The Pantawid Kuryente Program is essentially a self-targeted program, as the beneficiaries are identified based on the amount of electricity they consume. However, the lifeline power consumption level (no greater than 100 kilowatt-hours) does not appear to be effective in distinguishing poor from non-poor households. In fact, only 8 percent of the households that belong to the poorest income decile are lifeline power consumers (Figure 9.4). The Tulong Para Kay Lola at Lola program used categorical targeting, basing eligibility on age (70 years and above) and adding other eligibility criteria such as not receiving pensions or regular income. This group, however, only accounts for 9 percent of the poorest households, resulting in a high program leakage rate of 61 percent. 168 9.23 Lack of a systematic method to identify members of the PhilHealth's Indigent Program has resulted in higher enrollment of poor households than the recorded poverty incidence. Although the national government and the LGUs jointly Figure 9.4: Distribution of Beneficiaries of One-Time Casb pay the insurance premium in Subsidies 0 PhilHealth's Indigent Program, 20 ~--------------------------------------------------------------~ indigents enrolled in the program 15 are identified by local social 10 welfare officers using fairly ad hoc schemes. As a result, 23-44 percent 5 of provlllces have enrolled o beneficiaries in excess of about 64- 18t 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 78 percent of the actual number of · Pantawid Kuryente IJTulong Para Kay Lola at Lola poor households based on the Family Income and Expenditure Source: World Bank estimates using FIES 2006. Survey (FIES) in 2006. Despite the over-enrollment in the program, it is estimated that at least 1.8 million poor households, or 38 percent of total poor households, are not yet covered by the program. 9.24 The proxy means test targeting methodology used for the 4Ps CCT program appears to have resulted in higher coverage of the poor and potentially higher poverty impact. As described in Box 9.3, beneficiaries for the 4Ps were selected using proxy means testing (PMT), which appears to have performed well in pilot areas. It is estimated that 87 percent of households classified as poor were identified as potential program beneficiaries. The 4Ps is expected to increase the total incomes of poor and eligible households by an average of 23 percent and to reduce poverty incidence in targeted areas by 6.1 percentage points. sl The 4Ps is also expected to reduce the poor's income gap by 7.6 percentage points and poverty severity in beneficiary areas by about 5 percentage points. Although these estimates are preliminary, they are consistent with the results of impact evaluations of comparable CCT programs in countries such as Mexico and Colombia. 81 World Bank staff simulations based on the total cash transfer (health plus education component), computed according to the actual demographic composition of potentia:! beneficiary households and per capita income predicted using the PMT. 169 Box 9.3: National Household Targeting System for Poverty Reduction The National Household Targeting System for Poverty Reduction (NHTS~PR) uses a Proxy Means Test (PMT) approach to select poor beneficiaries for program access. It is used by the Pantawid Pamilyang Pilipino Program (4Ps), a conditional cash transfer (CCT) program. During the pre~pilot and initial phase of the roll~out to 360,000 households in 2008, targeting included two steps: (i) selection of poor provinces based on poverty incidence according to the Family Income and Expenditure Survey (FIES) in 2006 and selection of municipalities based on poverty incidence according to Small Area Estimates in 2003 and Oi) assessment of households through the application of PMT, in which household incomes are predicted on the basis of demographic, educational, and socio~economic characteristics. The information for estimating the PMT is gathered in a two~page questionnaire with relevant variables that predict household income. The PMT model was estimated using the FIES and Labor Force Survey. The information for calculating the PMT is collected using a number of data~collecting strategies such as survey sweeping (i.e. all residents in a given area are surveyed), on~demand application, or a combination of the two methods depending on the poverty incidence and the urban/rural classification of the areas to be surveyed. The information is processed in a standardized way using software developed for that purpose and is routinely validated. A PMT model is then calculated to classifY households as poor or non~poor based on the provincial poverty thresholds established by the National Statistical Coordination Board for 2006. This targeting method has been used in other middle~income countries such as Mexico, Colombia, Chile, and Costa : Rica, where reported income is hard to verifY as a large number of people work in the informal sector and there are few databases available to cross~check incomes or assets. Source: World Bank (2009). 9.25 The PMT targeting methodology could help improve the poverty reduction impact of existing social protection programs such as the NFA rice price subsidy. Estimates show that the implicit subsidy for NF A rice in 2008 was PhP32.1 billion,82 and given the high leakage rate of the NFA rice price subsidy, poor households only get an annual transfer of PhPl,843 (US$O.1l1day) which is equivalent to 15 percent of their spending on rice. Using the PMT methodology would enable the government to move away from untargeted assistance that covers all households to one that specifically targets poor households only, enhancing the poverty reduction impact of the program. Assuming the same program budget, such an improvement in targeting would increase the size of the transfer to poor households nearly fourfold and reduce poverty incidence and the income gap by 4.7 and 3.1 percentage points, respectively (Table 9.5).83 Targeting rural poor households only could reduce poverty incidence by as much as 6.4 percent and the income gap by 4.2 percent. 82 World Bank staff estimates based on the difference between the international price of rice and NFA rice at the height of the rice crisis in 2008 (World Bank, 2008). 83 This estimate assumes complete accuracy ofthe PMT, but there may be some exclusion errors in practice. 170 Table 9.5: Impact of Improving Targeting of NFA Rice Distribution Baseline Mean Household Poverty Incidence ( of Income gap Poverty Severity ( (HH) Income the population) ( of poverty line) of poverty line) All Households 172,730 32.9 percent -30.5 percent 4.2 percent Impact of Transfer* to Change in HH Percentage point change in the national- Overall Poverty Income** 0 Poverty Incidence Income Gap Poverty Severity All Households 1.1 -1.4 -0.8 -0.4 All Poor Households 10.9 -4.7 -3.1 -1.3 Rural Poor Households 16.7 -6.4 -4.2 -1.7 Decile 1 Households 57.1 -1.7 -2.8 -0.9 * Assumes equal distribution of the PhP32.1 billion across households in the specified group. ** Compared to actual household income of the cohort group. Source: World Bank (2008). 9.26 To improve the targeting of its social protection programs, the government is putting in place a national household targeting system that uses the PMT methodology. As part of its social protection reform agenda, the government is implementing the National Household Targeting System for Poverty Reduction (NHTS-PR), which uses the PMT methodology in selecting the poor. The NHTS-PR is intended to be used for targeting the beneficiaries of key social protection programs of DSWD as well as other government agencies. By using the PMT methodology in the development of the NHTS-PR, benefits to poorest households can be increased by reducing leakage of existing social protection programs, hence, having a greater impact in reducing poverty incidence and severity, which also allows for cost- efficiencies. D. Accountability Issues 9.27 Limited institutional capacity and other factors pose a constraint to improving accountability in social protection programs. Clear governance structures, robust control, and accountability mechanisms are necessary for the effective and efficient delivery of social protection programs. International best practice shows that to minimize losses in administering social protection programs, adequate systems to prevent, detect, and deter error, fraud, and corruption must be in place. However, in the Philippines as in other developing countries, limited capacity and resources, the lack of accurate computerized systems for auditing, and the large share of the informal economy make it difficult to improve accountability. In recent years, some developing countries have made significant steps in implementing mechanisms to control the use of funds in social protection programs. Latin American countries, for instance, have embarked on CCT programs designed to minimize incentives and opportunities for error, fraud, and corruption, in which administrative procedures are in place, institutional responsibilities are aligned, and transparency and communication are used well. 9.28 The 4Ps CCT program is paving the way for a modernized social protection system that integrates control and accountability measures. The design of the 4Ps draws heavily on international experience that establishes solid structures for ensuring adequate transparency, accountability, and governance in program implementation. For example, an objective and transparent methodology is used to select the geographic areas where the 4Ps will be implemented as well as to select beneficiaries, and the PMT results are subject to community- level verification. The 4Ps has also established a simple and accessible grievance and redress 171 system for program beneficiaries and non-beneficiaries. Moreover, the payment system for 4Ps does not require the services of intermediaries, as cash transfers are paid directly into the designated accounts of beneficiaries through electronic banking. Routine monitoring by the project management office and government and by an independent committee helps ensure strong project oversight. Table 9.6 summarizes the common risks and major issues associated with a social protection program and how the 4Ps structure mitigates such risks. Table 9.6: Improving Accountability Through CCT Risks Accouutability Measure Geographical selection Provincial selection based on FIES and municipal selection based on SAE. Targeting PMT objectively determines beneficiary list; public validation of beneficiary lists through social audit; grievance redress system captures and resolves complaints; database updated regularly; monitoring and spot checks . Compliance monitoring . Simple MIS forms designed; well-staffed MIS unit at both national and regional levels; multiple municipal links in high population LGUs to collect and maintain compliance forms; monitoring and spot checks. Payment Monitoring of reports of cash cards lost or stolen; education for recipients on use of the ATM and to discourage the use of intermediaries; grievance and redress system; monitoring and spot-checks; internal audit services' semi- annual review of the program. Lack of institutional clarity Establishment of inter-agency advisory committee comprising all relevant departments at each level of government; signing ofMOA between DSWD and participating LGUs to ensure clarity of roles and responsibilities. Political abuse of power Pre-election freeze on new registrations; prohibition on collection of fees by LGUs in Operations Manual and orientation sessions; monitoring and spot- checks; grievance and redress system. E. Policy Implications and Conclusions 9.29 Better targeting is critical for ensuring that social protection programs benefit the poor and vulnerable. Programs should be targeted wherever and whenever possible so that more generous benefits can be provided to the poor and vulnerable who really need the assistance. A standardized and transparent household targeting system can also improve the governance, transparency, and credibility of programs. While no program is immune from errors and leakages, other social protection programs could benefit from the National Household Targeting System for Poverty Reduction. The PMT methodology is one of the most advanced systems for targeting, and once in place, such a system could be used to target various social protection programs. Other targeting methods may also be used, as PMT tends to be unresponsive in the short term to changes in the need for assistance and could still miss some vulnerable households (see Ravallion, 2003 and 2008). Other targeting methods include simpler categorical targeting (e.g., elderly, disabled), geographical targeting (universal coverage in areas with high poverty incidence), and self-targeting, where limited assistance is offered to everybody but at some costs to households, such that only the neediest ones would participate. 9.30 The delivery of social protection services could also be enhanced through improved policy coordination. Harmonizing interventions to avoid overlaps and gaps in coverage is critical for ensuring the efficiency of social protection spending. A key step in this direction was the passing of the operational definition of social protection in 2007 as described above. Another important development was the creation of the inter-agency National Social Welfare and 172 Protection Program in 2008. Currently, this inter-agency body is undertaking an assessment of the existing social protection programs, with a view to scaling up and reallocating resources to the most effective and efficient ones. In 2009, the government adopted a Social Protection Framework that provides an inventory and typology of major programs. Further analysis leading to an overall operational social protection strategy will be critical for taking these efforts forward. An important positive step taken in this direction is the establishment of the Social Protection Sub-Committee (SPSC) under the Social Development Committee (SDC) of NEDA, chaired by DSWD. Among others, the sub-committee is responsible for coming up with a social protection strategy that is consistent with national development priorities and ensuring that institutional arrangements are in place to implement the strategy. 9.31 The 4Ps could serve as the potential backbone of a modern and more consolidated social protection system for the Philippines. As described above, the Philippines has had a proliferation of transfer programs that suffer from high under-coverage and leakage rates. The design of the 4Ps and its PMT targeting system are based on best practice examples from other countries which have proven to have good targeting outcomes. Such CCT programs have also been found to be effective demand-side interventions for building human capital. The 4Ps is therefore a promising backbone for the social protection system and, if successful, could potentially replace other subsidy and transfer programs as a flagship social protection and poverty reduction program for the country. 9.32 Finally, to help ensure successful implementation, it will be important to enhance the capacity of government agencies involved in social protection. The implementation capacity of key agencies is a key consideration in designing social protection programs. For example, the 4Ps requires significant administrative capacity in the DSWD. To roll out the program successfully, extensive training will be needed on all the operational steps of the program, including on how to operate effective monitoring and evaluation systems, for staff at the central and local levels. Enhancing the capacity of DSWD to implement the 4Ps, as well as the National Household Targeting System for Poverty Reduction (NHTS-PR), could lead to the emergence of a strong institutional leader in the social protection sector. 173 References Alderman, H. Hoddinott, J. and Kinsey, B. (2006). "Long Term Consequences on Early Childhood Malnutrition." Oxford Economic Papers, 58(3), 450-74. Besley, T., Burgess, R. and Rasul, I. (2003). "Benchmarking Government Provision of Social Safety Nets." Social Safety Net Primer Series, May 2003. Carter, M., Little, P., Morgues, T. and Negatu, W. (2004). "Shocks, Sensitivity and Resilience: Tracking the Economic Impacts of Environmental Disasters on Assets in Ethiopia and Honduras." BASIS Research Program on Poverty, Inequality, and Development, USAID, Washington, DC. Department of Social Welfare and Development (2009). "Pantawid Pamilyang Pilipino Program." Available at: http://W\\lw.dswd.gov.ph/ Esguerra, J., Ogawa, M. and Vodopivec, M. (2002). "Options of Public Income Support for the Unemployed in the Philippines." Social Protection Discussion Paper Series No. 0204. Fafchamps, M., Udry, C. and Czukas, K. (1998). "Drought and Saving in West Africa: Are Livestock a Buffer Stock?" Journal of Development Economics 55(2), 273-305. Ferreira, F. and Schady, N. (2008). "Aggregate Economic Shocks, Child Schooling, and Child Health." World Bank Policy Research Working Paper No. 4701. Grosh, M., del Ninno, C., Tesliuc, E. and Ouerghi, A. (2008). For Protection and Promotion: The Design and Implementation ofEffective Safety Nets. World Bank, Washington, DC. Lemieux, T. and Milligan, K. (2008). "Incentives Effects of Social Assistance: A Regression Discontinuity Approach." Journal of Econometrics 142(2),807-28. Manasan, R. (2006). "Review of Government Programs and Spending Priorities for Social Welfare Social Protection and Social Development: Phase I." Mimeo. Manasan, R. and Cuenca, J. (2007). "Who Benefits from the Food-for-School Program and Tindahan Natin Program: Lessons in Targeting." PIDS Discussion Paper Series No. 2007-10. Manasan, R. (2009). "Assessment of Social Protection Programs in the Philippines." Mimeo. Moffitt, R. (2003). Means-Tested Transfer Programs in the United States. NBER Conference Report. Cambridge, MA. National Anti-Poverty Commission and National Statistical Coordination Board (2006). Assessment of Vulnerability to Poverty in the Philippines. Quezon City. 174 National Economic and Development Authority (2007). Social Development Committee Cabinet Resolution No.1, Series of2007. (2004). Medium-Term Philippine Development Plan, 2004-2010. Manila, Philippines. National Statistics Office (2006a). "Labor Force Survey" public use file. (2006b). "Family Income and Expenditures Survey" public use file. Orbeta, A. (2009). "Overview of Labor Market and Social Protection in the Philippines." Mimeo. Ravallion, M. (2003). "Targeted Transfers in Poor Countries: Revisiting the Trade-Offs and Policy Options." World Bank Social Protection Discussion Paper No. 0314. Ravallion, M. (2008). "Bailing Out the World's Poorest." World Bank Policy Research Working Paper No. 4763. Reyes, C., Manasan, R., Orbeta, A., de Guzman, G. (1999). "Social Impact of the Regional Financial Crisis in the Philippines." PIDS Discussion Paper No. 99-14. Reyes, C. (2002), "The Poverty Fight: Have We Made an Impact?" PIDS Discussion Paper 2000-20. Stephens, M. (2009). "Governance and Anti-Corruption in the Pantawid Pamilyang Pilipino Program." Social Welfare and Development Journa/3(1). Weigand, C. and Grosh, M. (2008). "Levels and Patterns of Safety Net Spending in Developing and Transition Countries." Social Protection Discussion Paper No. 0817. World Bank (2009). "Social Welfare and Development Reform Project," Project Appraisal Document, October 2009. World Bank (2008). "Food Crisis Response Development Policy Operation Under the Global Food Crisis Response Program to the Republic of the Philippines," Program Document, December 2008. World Bank (2001). Philippines Poverty Assessment. Washington, DC. World Bank and National Disaster Coordinating Council (2004). "Natural Disaster Risk Management in the Philippines: Enhancing Poverty Alleviation Through Disaster Reduction." Available at siteresources. worldbank.orglINTEAPREG TOPENVIRONMENTIResourceslPH_Disaster_ Risk_ Mgmt.pdf 175