Thailand Country Development Partnership for Social Protection Phase II Issues and Directions for Social Protection April 9, 2006 Paper presented at the Workshop on the Country Development Partnership for Social Protection II (CDP-SP II), Amari Watergate Hotel, Bangkok, February 23, 2006. CONTENTS CONTENTS ........................................................................................................................ 2 ACRONYMS ...................................................................................................................... 3 EXECUTIVE SUMMARY ................................................................................................ 4 A. INTRODUCTION ......................................................................................................... 6 B. ECONOMIC AND SOCIAL CONTEXT ...................................................................... 8 C. SOCIAL PROTECTION POLICIES, POVERTY AND INEQUALITY ................... 10 1. Poverty and Inequality .............................................................................................. 11 2. The Labor Market ..................................................................................................... 13 3. Education and Training ............................................................................................. 15 4. Labor Market Programs ............................................................................................ 17 Employment Services and Labor Intermediation ...................................................... 17 Unemployment Insurance ......................................................................................... 17 Occupational Safety and Health ............................................................................... 18 5. Informality and Employment Programs.................................................................... 18 6. Social Pensions and Community Savings ................................................................. 20 7. Social Assistance ...................................................................................................... 22 8. The Administrative Challenge: Agency Coordination and Decentralization ........... 24 9. Statistical Requirements............................................................................................ 26 D. SUMMARY AND RECOMMENDATIONS .............................................................. 26 A. An integrated strategy and coordinated implementation that would also include the informal sector. ............................................................................................................. 28 B. Balanced Regional Development ............................................................................. 29 C. More Effective Decentralization .............................................................................. 30 D. Further Analysis and Statistics................................................................................. 31 REFERENCES ................................................................................................................. 33 2 ACRONYMS APO Autonomous Public Organization CDP-SP Country Development Partnership for Social Protection LFS Labor Force Survey MOL Ministry of Labor MOLSW Ministry of Labor and Social Welfare MSDHS Ministry of Social Development and Human Security NGO Non-Governmental Organization NSO National Statistical Office OSH Occupational Safety and Health RTG Royal Thai Government RVA Risk and Vulnerability Assessment SES Socio-Economic Survey SSO Social Security Office TAO Tambon Administrative Organization TDRI Thailand Development Research Institute THB Thai baht UI Unemployment Insurance WCF Workmen’s Compensation Fund 3 EXECUTIVE SUMMARY1 The Country Development Partnership for Social Protection (CDP-SP) has provided new insights into social protection issues in Thailand. The CDP-SP, a collaborative effort between the Royal Thai Government (RTG), World Bank, and other partners, helped deepen the understanding of poverty and vulnerability in Thailand and pointed to areas for further action. The Risk and Vulnerability Analysis (RVA), for example, found that formal social protection has very limited coverage and only for a few: as low as 20% even among those in the top income decile are covered, and practically none from the lowest decile. Additional examination of social insurance and social assistance programs suggested that some are regressive. Although these two findings are subject to statistical uncertainties, they do suggest that any expansion of programs should be carefully assessed in advance. The findings also indicate that the programs will only affect a few and therefore cannot be considered a significant substitute for broader and more integrated approaches. Thailand, a middle-income country with an established record on social development, rightly aspires to increase the coverage and effectiveness of formal social protection. Improvements in formal social protection programs are feasible and desirable, and the CDP-SP results can provide guidance for many sectoral programs such as employment services, unemployment insurance (UI), and occupational safety and health (OSH). The CDP-SP also pointed to the need for capacity building as well as better statistical systems, including wide dissemination of existing statistics and the introduction of panel data. Given the extent of vulnerability across the population, improving only the internal efficiency of sectoral programs—that is, strengthening their performance without even expanding these programs—could increase the security and incomes of many people. At the same time, the expansion of programs such as UI or OSH should be assessed both from an actuarial as well as a benefit impact perspective. One measure stands out: the social pension that can, after additional examination, be a candidate for becoming a universal benefit under certain conditions (for example, that the beneficiary should not be in receipt of another pension and meets some income and asset tests). However, while the pursuit for broader and better formal “second generation” social protection programs continues, Thailand faces a partially unfinished agenda on “first generation” issues. The expansion of specific social protection programs can have only a limited impact compared to more systemic interventions that would (a) provide greater employment opportunities for work in the wage sector through education, thereby reducing informality; (b) reduce regional disparities by more targeted regional allocations of public funds, thereby reducing the variability of farm incomes; and (c) ensure that decentralization is not just confined to decentralization of budgets but also covers functions and that regional authorities have the capacity to undertake their increased or new responsibilities. 1 This paper was prepared by P. Zafiris Tzannatos (Adviser, World Bank Institute), under the guidance of Ana L. Revenga and Minna J. Hahn, with support from Kwanchai Niyomthamkit. 4 Most important of all, however, is the need for a comprehensive understanding of the dynamics of social insurance and social assistance mechanisms—both formal and informal, public and private—in the context of an integrated strategy for human development and social protection. This need is bound to become increasingly important for Thailand over time, first because many of the actors involved in social protection are outside the conventional agencies dealing with labor issues or social development. Thus, there will be an increasing need for greater coordination of policies, programs, and the involved stakeholders, including various government agencies and the private sector. Second, although Thailand has been able to respond quickly to short term crises such as the 1997 financial crisis and the tsunami in 2004, it still faces periodic floods and droughts and the uncertainty surrounding avian influenza. These factors and the unfinished agenda require a comprehensive strategy, not just for specific sectoral programs, but for broader human development and social protection as a whole. 5 A. INTRODUCTION This note reviews the issues covered by the Country Development Partnership for Social Protection (CDP-SP) to draw implications for Thailand’s social protection strategy going forward. The CDP-SP was a multi-year rolling reform program implemented by the Royal Thai Government (RTG) in collaboration with the World Bank and other partners (see Box 1). Focusing on priority reform areas within Thailand’s rapidly evolving social protection agenda, the CDP-SP aimed to help the RTG meet the goal of the Ninth National Economic and Social Development Plan to create a sustainable and “sufficiency” economy using a balanced socio-economic approach.2 The CDP-SP pointed to some important lessons for social protection policies and programs in Thailand. As a result of the CDP-SP program, much more is known about the extent and patterns of poverty, inequality and vulnerability and the differences between formal and informal sectors, incidence of public social spending, characteristics and performance of various formal social protection programs (and others – such as in education and health) and technical/administrative and policy constraints including the challenges of decentralization. In addition, the program’s findings included international benchmarks for institutional improvements in the administration of key agencies operating in the areas of social insurance, social assistance, and safety nets. The most recent phase of the CDP-SP (CDP-SP II) also indicated some promising policy directions for the future. This note, therefore, supplements the findings and recommendations of the CDP-SP II by casting them in a broader social protection framework.3 Using the CDP-SP II as a starting point, the objective of this paper is to derive policy and program recommendations for social protection. The paper summarizes what we have learnt and identifies some strengths and challenges for moving forward. It has some recommendations what to do (and key among them is the development of an integrated humand development and social protection strategy) and how (and critical for this is to improve agency coordination and the effectiveness of decentralization). It is structured as follows: the next section provides a brief overview of Thailand’s economic and social strengths and shows that significant and sustainable gains have been achieved even in the absence of formal social protection programs. Section C then presents the key recent chenages in poverty and inequality and the role of the labor market, education, employment programs and social assistance taking into account the constraint imposed by the large informal sector and the limits of decentralization. The findings are then used to derive recommendations (section D) and conclusions (Section E). . 2 Royal Thai Government (2001), p. 1. 3 The findings and outcomes of the activities supported by CDP-SP II are summarized in a companion note “Country Development Partnership for Social Protection: Summary of Activities and Outcomes” (2006). 6 Box 1. The Country Development Partnership for Social Protection The Country Development Partnership for Social Protection (CDP-SP) was a multi-year rolling reform program, implemented through a series of technical assistance and capacity building activities. The CDP-SP aimed to help Thailand develop a balanced social protection strategy which would complement existing informal risk coping mechanisms with a more diverse set of public and market-based risk management mechanisms. It was not intended to be a comprehensive program of reform in social protection; instead, it provided an evolving framework for reforms in selected priority areas as identified by the Royal Thai Government (RTG). The specific objectives were to:  Close the equity gap by focusing services more effectively on those most in need —the poor, the vulnerable, the jobless;  Close the coverage gap of programs so that the benefits of these programs reach a critical mass of intended beneficiaries;  Improve the efficiency and cost effectiveness of programs through application of performance- based management approaches; and  Search for a more productive balance between the public and the private sector in the delivery of services. Following the successful completion of the initial phase (2001-2002), CDP-SP II was launched in 2003 to consolidate the analysis undertaken in the first phase. CDP-SP II was designed to (i) extend and complement the activities and strengthen the reforms carried out in the first phase; (ii) apply the technical and analytical tools developed; and (iii) prepare plans to implement reliable service improvements. Activities were implemented by the Ministry of Labor (MOL), Ministry of Social Development and Human Security (MSDHS), Ministry of Finance (MOF), and the National Statistical Office (NSO) in collaboration with the World Bank and other partners. In each priority area, CDP-SP II sought to improve the targeting, efficiency, cost effectiveness, and equity of social protection services. The program addressed the areas of: social assistance, employment services, unemployment insurance, social protection data, occupational safety and health and Workmen’s Compensation, and social protection for informal workers. Technical assistance and capacity building support, financed by grant resources from the Asia-Europe Meeting (ASEM), were provided by local and international researchers and practitioners who drew upon good practices in the design, management, and delivery of social services and social policies. The CDP-SP appears to have made a significant contribution in the areas it covered . Although the CDP-SP did not address the full range of public efforts to reduce poverty—nor was it intended to do so—it constituted an important step toward understanding poverty and vulnerability and the policy and implementation challenges faced in social protection. The partnership also helped build capacity to address these challenges in the selected priority areas. The CDP-SP is expected to serve as a strong platform for further reforms, and if the activities maintain their momentum, they could be expanded in a way that would contribute to the creation of a comprehensive national approach to social protection. 7 B. ECONOMIC AND SOCIAL CONTEXT Thailand has adopted successful economic policies for a long time, and the National Plans have successively placed an increasing emphasis on the social sectors. The result has been the creation of a successful economy that is resilient to shocks. The growth-oriented policies, including fiscal and trade policies, have over time yielded high economic growth rates which have enhanced opportunities and increased the potential of poor people. Thailand’s growth rates have also contributed to dramatic reductions in poverty over the past 30 years. In the 1980s, Thailand was one of the fastest growing economies in the world, and the poverty rate decreased from 33% in 1988 to 12% in 1996, lifting nearly 11 million people out of poverty.4 Although poverty rates increased sharply following the financial crisis (to 16% by 1999), they have now reverted to a lower rate (11%). Today, there are around one million people fewer living in poverty than in 1996, and 5.6 million fewer than at the peak of the increase in poverty in 2000.5 In addition to pursuing sound macroeconomic policies for growth, Thailand implemented a comprehensive response to the financial crisis. The Government moved swiftly to increase budget support for existing programs in health and education, expand job creation through civil works, extend formal sector social insurance, and expand the benefits provided under various social assistance schemes to the indigent. More recently, Thailand has embraced wide-ranging fiscal reforms. The economic recovery since the financial crisis has brought back fiscal surpluses since FY2002/03. Fiscal reforms have related to accountability and transparency, changes in financial management and reporting, and service delivery. Public expenditure management is also being reformed to become more performance-based and transparent. Thailand is also moving toward greater decentralization of government activities to local administrations. Although the effects of the recent policy changes are not yet fully known, decentralization can result in increases in allocative efficiency, because lower levels of government have better information regarding local needs and preferences. Decentralization can also increase productive efficiency by resulting in greater accountability at the local level and increased community participation in the decision- making process. A series of different decentralization laws have already been enacted, and the share of local revenue relative to total government revenues increased from 4 Thailand had nearly 17 million poor people in 1988, compared to 7 million in 1996. 5 World Bank (2005b). 8 22.8% in fiscal year (FY) 2004 to 23.5% in FY 2005 and is expected to reach 24.1% in FY 2006.6 Together with devolution of administration, greater emphasis is being placed on strengthening the participation of civil society and the private sector in the provision of social assistance. One distinctive characteristic of Thai society is the value and emphasis placed on family self-reliance and mutual community-based support. For example, Thai families, especially in the lower-income groups, took concrete steps to cushion the effects of the financial crisis and protect their social assets, even as their incomes contracted more than those of better-off families. One informal support mechanism is migration: more than half of households benefited from migrant remittances following the financial crisis. Among receiving households, remittances amount to around one-third of household income in the Northeast, Center, and North. The poverty headcount in households without remittances is about 5% higher than in households without remittances. This strengthening of the role of the civil society is not just a symbolic declaration, as Thailand has an institutionalized approach for community empowerment . The Constitution provides the legal foundation for community empowerment and includes provisions on the decentralization of power to the local level, the right of access to information, and the right to participate in the decision making process at all levels. Complementing the Constitution, the Eighth Plan (1997-2001) emphasized an agenda of “people-centered development” focusing on the poor and under-privileged. The Ninth Plan (2002-2006) has taken this poverty focus as a more central strategic concern and espouses four main objectives: (i) poverty alleviation, (ii) good governance, (iii) adherence to the new economy and self-sufficiency, and (iv) strengthening of the social foundation through the enhancement of quality of life and protection of natural resources. Although it recognizes the role of informal social protection mechanisms, the RTG does not rely solely on them, but has developed a wide-ranging social assistance network. The government safety net system includes cash and in-kind as well as income- generation programs for needy families, poor village communities, the elderly without any means of support (social pension), children (through lunches, scholarships, and children’s allowances), the disabled or sick (through low-income health cards or participation in the 30-baht health scheme), and job seekers (through rural job creation and public works programs). Aggregate government spending on anti-poverty measures has increased significantly in real terms over time, and the coverage of formal social insurance continues to expand. Public expenditures in this area have been increasing for many years, for example from THB4.5 billion in 1993 to THB21.1 billion 1999.7 Thailand is also continuing its prudent tradition of gradually sequencing the introduction of social insurance to new groups as economic conditions require and allow. For example, the Workmen’s Compensation Fund (WCF) was introduced in 1974 to provide social 6 However, this percentage falls short of the originally planned 35% for the current year. 7 In constant 1988 Baht. FY2005/06, the RTG has allocated Baht 70 billion (or USD 1.75 billion) towards anti-poverty measures, representing approximately 5% of the total budget spending. 9 security to workers who were injured or sick due to work-related activities. Its initial coverage was restricted to workers in private enterprises in Bangkok with 20 or more employees. In 1988, the coverage of insured persons was extended nationwide. In 1994, workers in small private enterprises (10-19 employees) were brought under the scheme, as were those in micro enterprises (1-9 employees) in 2002. Today, as many as 40% of workers in the private formal sector are covered, and the range of benefits has been expanded to include sickness, maternity, invalidity, death, old age, and unemployment benefits as well as child allowances.8 C. SOCIAL PROTECTION POLICIES, POVERTY AND INEQUALITY Despite the positive developments discussed above, some issues remain that, when addressed properly, can contribute further to Thailand’s economic and social development through a more effective social protection system. Social protection is examined below in a broad sense that includes:  protection afforded by greater employment opportunities;  education and skills development programs than enable citizens to take advantage of labor market opportunities;  formal employment-based insurance and social assistance schemes, as well as informal mechanisms; and  broader policies that support regional development and decentralization. These areas have also emerged from the findings of the CDP-SP and are identified in the Ninth Plan,9 which notes: “Thailand’s development trajectory over the past four decades clearly indicates imbalanced development. While success measured in terms of quantitative indicators has been achieved, improvements in quality of life seem to lag far behind. This can be explained by weaknesses in Thailand’s economic, political and administrative management systems that are centralized and inefficient…At the same time, the quality of education among Thai people has not been significantly upgraded. Therefore, the Thai people cannot adapt themselves to modern technology… [and this] gives rise to continuing erosion of Thailand’s international competitive position…Meanwhile, the widening income gap, increased poverty, and natural resources and environmental deterioration have contributed to increased social conflict and tension.” The increased poverty and widening inequality are addressed immediately below as they constitute the key outcomes of the government’s efforts in the area of social policy. Then, we examine, 8 However, the coverage and adequacy of benefits remains low, as discussed below. The speed at which the extension of benefits should increase needs to be further assessed from a fiscal and equity perspective that also takes into account administrative constraints (such as decentralization) and the role of informal social protection mechanisms. 9 Royal Thai Government (2001), p. 2. 10  recent developments in and the functioning of labor market as well as education and training – both constitute the backbone for further increases in labor productivity on the labor demand side, and more sustainable employment and income security (on the labor supply side;  the key employment programs assessed by CDP-SP in the broader context of widespread informality in Thailand. These programs include employment services, unemployment insurance and occupations safety and health;  non-contributory pensions, social assistance and community savings that were also examined by CDP-SP and can serve as bridges between the formal and informal social protection systems;  finally, the administrative challenge for policy implementation that can arise from lack of coordination between the various ministries and agencies involved in social protection, ineffective decentralization and absence of statistical information. 1. Poverty and Inequality Agriculture remains a critical factor for factor for reducing poverty. The rise in agricultural incomes has contributed to a large extent to the recent reduction in poverty. Under the recently revised poverty line, poverty incidence in terms of headcount has fallen from 21.3% in 2000 to 11.3% in 2004.10 Agricultural incomes have risen by 14% between 2000 and 2004, higher than the rise in salaries (8%) and non-agricultural incomes (6%).11 Despite progress in poverty reduction, household income inequality and regional Poverty rates by Region, 2000-2004 disparities remain large. Income inequality decreased slightly since 2000, but remains 40 Rate (%) persistently high, with the Gini coefficient at 20 approximately 0.5 since 1996. The headcount 0 ratios in the Northeast and North are nearly three Northeast North South Central Bangkok times the headcount ratio in the Central region. Region The headcount poverty rates are near zero in Bangkok, but reach nearly 50% in the Northeast 2000 2004 and the South. The Northeast, which is the most populous region, is home to over half of Thailand’s poor,12 and the 17 poorest provinces (out of 76) account for nearly two-thirds of all poor people. A mere 10% of the poorest villages in the country account for nearly one-third of all poor people, while slightly more than one-quarter of the poorest villages account for practically all of the Thai poor. The decline in national poverty rates was mainly caused by the reduction in the number of poor in the Northeast. As the majority of poor people reside in rural areas 10 The headcount poverty rate based on the old poverty line was 14.2% in 2000 (compared to the new value of 21.3). See World Bank (2005b). 11 World Bank (2005b). 12 World Bank (2005b). 11 and are engaged in agricultural activities, rural/agricultural incomes in the Northeast are critical for poverty reduction. The headcount poverty in the Northeast was reduced from 35% of the population in 2000 to 17.2% in 2004, primarily because of increases in farm incomes (Figure 2). Nonetheless, disparity among regions has increased over time.13 The financial crisis exacerbated regional disparities, as poverty rates actually decreased in Bangkok and the North but increased everywhere else. While the Thai economy continues to grow rapidly, this growth is driven to a considerable extent by the growth pole of the extended Bangkok area which houses around 6.3 million people or about 17 times the number of inhabitants of Samut Prakan, the second largest city, where around 400,000 people live. Even by East Asian standards, Thailand’s urbanization rate is very low. In a comp arative context, Thailand’s urbanization was on par with its income level in 1963, but at its rate of 33% today, it is 20% lower than other countries of its income level. Much of the vulnerability of people living in rural areas, especially Northeast farmers, comes from exposure to weather shocks. Currently, the low asset base and limited access to well-developed insurance and credit markets leaves rural communities ill-equipped to deal with weather shocks. Their weather risk management is inefficient, with negative implications for economic and social development.14 The effects of crisis are complex, but are borne disproportionately by the poor. The incomes of those educated at secondary education and above increased immediately after the financial crisis, while those educated at below that level saw their incomes decline. Those with primary education experienced a decline in income by 4% and the illiterates by as much as 16%. An understanding of these distributional impacts of shocks can help guide policies on how incomes may be maintained. Although the striking geographical patterns of poverty in Thailand imply that geographical targeting can be a powerful tool, the distribution of RTG expenditures on anti-poverty programs tends to be regressive. Thailand’s poverty pattern shows a close overlap of poor areas and areas with large populations, in contrast to the more common pattern in other countries of high poverty in sparsely populated areas. Poverty in Thailand is primarily a rural problem, with about 90% of the poor located in rural areas. However, wealthier provinces receive a slightly higher per capita proportion of public expenditures than poorer provinces, and budget allocations to the Northeast are lower than to other regions. Concerns have been expressed that a large portion of the RTG’s expenditure on poverty programs is wasted due to poor targeting, selection of the wrong types of poverty interventions, and inefficiency in execution of poverty alleviation projects. However, due to insufficient monitoring and evaluation of the effectiveness of 13 These regional disparities follow the different characteristics of each region that in turn imply different responses of poverty to regional economic growth. The Northeast stands out for a growing service sector; the North for a growing industrial sector; the South for the resilience of agriculture; the Center represents the classic case of economic growth through industrialization; and Bangkok generates income foremost through services. For the poverty gap, the Center ’s elasticity is lowest at 2.5, compared to 2.6 for the Northeast, 3.2 for the South, 3.3 for the North, and 3.7 for Bangkok. See World Bank (2005a). 14 Insurance instruments that can address this issue are discussed below. 12 government poverty programs in achieving their objectives, little is actually known about how much RTG spending on poverty reduction actually reaches the poor and the extent to which it reduces poverty. 2. The Labor Market Although education and labor market indicators have improved, labor force participation rates have declined. The salient, albeit very aggregate, features of the labor market are shown in Table 1. The shares of workers qualified at post-primary and university levels have doubled in the last 15 years. Employment in manufacturing and services increased by 50%, while the share of agriculture declined by nearly two-thirds and the share of the informal sector by one-fifth. One trend for concern is the significant decline in female labor force participation rate over time, which is rather unique in the world. This decline in labor force participation rates applies also to men, although to a lesser extent.15 The decline can be positive if, for example, it reflects increasing enrollments in education. However, this issue needs to be better understood, as it has implications for both raising funds for paying employment benefits through payroll contributions as well as extending the coverage of social protection. Table 1: Labor Market Indicators Percent Index 2005 Indicator 1990 1995 2000 2005 (1990=100) Education Primary education or less 81 74 66 62 77 Above primary/below university 14 20 26 29 207 University 5 6 8 9 180 Shares of Employment Agriculture 63 52 48 43 68 Manufacturing 10 13 15 15 150 Services 27 35 37 42 156 Informal Sector 71 64 60 56 79 Labor Force Participation Rate Male LFPR 88 83 81 82 93 Female LFPR 76 69 65 65 86 Ratio of F/M 0.86 0.83 0.80 0.79 92 Ratio of F in the LF 0.46 0.45 0.45 0.44 95 Note: Between 1994 and 2004, the labor force has grown from 32.5 million to around 36 million. 1.5 million of the increase originated in the Northeast, where employment increased from 10.5 million to around 12 million. Source: Based on LFS, 3rd round, quoted from Falter (2005) 15 Male labor force participation rates usually decline over time as education enrollments rise and the availability of pensions increases. 13 The labor market in Thailand seems quite flexible but also skills constrained. Minimum wages have been applied flexibly (Box 2), and their ratio to average wages has decreased over time (Figure 1). Figure 1:Ratio of minimum to average Employment regulations do not seem wages to impose significant constraints on labor deployment in the economy at 1 large nor be associated with high 0.8 Urban Ratio labor costs of hiring or firing 0.6 16 0.4 Private Urban workers. However, the educated 0.2 saw their incomes rise even amidst 0 the financial crisis, and 1985 1990 1995 2000 2005 manufacturing faces critical skills Year shortages (see below). Both findings point to the importance of improving the workforce development system, including education and training. Despite the apparent flexibility in the labor market, the employment elasticity to output has decreased significantly over time and is lower even than in South Korea , where incomes are much higher than in Thailand.17 Unemployment has traditionally been low in Thailand (generally around 2% in normal times) and was only 1.3% in the last quarter of 2005. However, the value of employment elasticity (the ratio of employment growth to output growth) is only around 0.2, which indicates that annual output growth of 5% will increase employment by only 1%. In a regional context, the employment elasticity is around 0.4 in Malaysia and Indonesia and exceeds 0.6 in Singapore and the Philippines, where it has been increasing over time. The slow rate of employment creation can put a strain in a labor surplus economy like Thailand’s, which is still characterized by a high level of informality as described below18. 16 The values of the Employment Rigidity Index for Thailand are in general lower than or similar to those observed in other regional economies. See World Bank (2006), Doing Business in 2006. 17 Asian Development Bank, quoted from The Economist (January 14, 2006). 18 The low employment-output elasticity in Thailand may be the result of fast productivity increases. This is not an undesirable outcome to the extent that it is not the result of policies that bias the price of capital downwards (i.e. induce more investment that would have taken place under undistorted prices) or the price of labor upwards (for example, through undesirable labor regulations). The labor analysis underpinning CDP-SP did not dwell on this issue which now needs some further investigation. 14 Box 2. Minimum Wages in Thailand Minimum wages were first introduced in Thailand in 1973 . Three (and for an intermediate period, four) different wages were prescribed and applied to different zones until 1998. Following the 1998 labor protection law, minimum wage administration is now decentralized, and each province can set its own rate subject to a national floor determined by the National Wage Committee made up of representatives from employers, workers, and government. The determination of minimum wages has increasingly become more flexible. Minimum wages were originally defined as "the payment sufficient for the employee and two additional family members to dwell in society." In 1976, the criterion for determining its level changed to the needs of the employee alone, no longer taking into account the two additional family members. The basic criterion of determination of the wage rate is now defined even more narrowly as "the wage that is sufficient for a newly-recruited unskilled worker to make a living in his/her community." Minimum wages are adjusted according to the inflation rate, costs of production, changes in productivity, and prices of goods and services, as well as the employer's ability to pay. Changes take place when either of the social partners demand it or by governmental order. Their relative value has declined over time, but compliance has increased. In 1990, they stood at more than 50% of average wages but dropped to 42% by 2000 and further to 40% by 2004. The increasing share of the formal sector has brought more workers under the purview of the legislation, and minimum wage compliance for urban workers aged 20-50 has increased from around 70% in 1990 to more than 80% in 2000. Overall, compliance rates now are around 85% in Bangkok but much lower in the North at 40%. However, in relative terms, minimum wages are higher in the Northeast and North, where they correspond to around 80% of average wages compared to no higher than 60% in Bangkok. Among urban workers, compliance increases with age: from 54% for those below the age of 20 to 87% for those aged 25 to 49. The share of those paid exactly the minimum wage is around 5% of all workers but increases to 10% for workers on daily rates, reaching 26% for those in manufacturing. Manufacturing is also the sector that has the highest social insurance coverage (92%) compared to commerce (40%) and construction (18%). The overall effects of minimum wages in Thailand upon the aggregate labor market are probably small, which has been facilitated by the gradual change of its determination from the needs of a family to those of young unskilled workers. The practice of freezing increases in minimum wages during periods of difficulty and the application of three regional rates have also allowed its flexible application. The only sector where they may have an effect is manufacturing. As manufacturing remains an important employment sector and is increasing over time, controlling wage growth may be the last resort if education and skills do not increase fast enough to induce productivity gains. Skills shortages in the Thai manufacturing sector are pervasive across firms. Source: Falter (2005); World Bank (2006). 3. Education and Training Rising school enrollments have led to a better-skilled labor force across the country, but the outcomes of the education system can improve further. Overall, employment has become more educated and more waged across all regions. The share of paid workers has increased across all sectors, including agriculture (from 4% to 9% since 1980). However, education enrollment rates in basic education remain low. The primary education net enrollment rate in Thailand was 86% in 2002 compared to an average 92% in the East Asia region. 15 In the manufacturing sector, skills shortages are pervasive across firms. 19 Surveys confirm that job applicants lack appropriate basic and technical skills, and skills shortages are felt equally by exporters and non-exporters and by domestic as well as foreign-owned firms. Ratings of workers by managers in the areas of information technology and English skills are lower in Thailand than in Malaysia and other comparator countries. The skills shortages are leading managers to hire less-than- qualified workers and, as a result, pay them lower wages than if the workers had the appropriate skills. The shortage of skilled labor leads firms to pay large wage premiums to tertiary education graduates and to workers who receive training in technical skills. Production costs increase as a consequence, and from a social protection perspective, it reduces the ability of less-qualified workers to receive higher earnings compared to what they would have been paid if they had better education and training. It also exacerbates inequalities in the labor market. Vocational education can be a significant avenue for increasing employability and productivity in Thailand, especially among the poorer segments of the population. While in many countries the labor market outcomes for workers with vocational education are worse than those with general education, workers with vocational education account for one-quarter of monthly salaried workers across Thailand and one-third in the Northeast. The wage premium for vocational education in the Northeast is 200% larger than upper secondary education and only 50% smaller than university education. The corresponding differences in Bangkok are 60% and 180%, respectively. In the Northeast, labor demand for vocational education has been increasing relative to upper secondary education. The lunch program of the Ministry of Education can be helpful in attracting children to schools while providing in-kind social assistance to poor families. Although the cost of meals is pegged at the relatively low level of about 5 baht, the program covers more than half of all children attending primary school.20 The process of selecting beneficiaries for these programs is somewhat arbitrary, with teachers in each school deciding which families are poor and which children are eligible to join the programs. Nonetheless, the program appears progressive. As many as 66% of children aged 6-10 in the bottom two income quintiles receive free school lunch benefits, while coverage for the top income quintile is much lower at 28%. In contrast, the government scholarship and student loans programs cover less than 5% of the eligible population and are regressive: more than half of all scholarship recipients and two-thirds of student loan recipients belong to the richest three quintiles.21 19 World Bank (2006). 20 Based on SES 2000 data, quoted from Deolalikar (2005). 21 Deolalikar (2005). 16 4. Labor Market Programs Employment Services and Labor Intermediation The need for fair and transparent government involvement in this area remains. The RTG is involved both as a provider of employment services and a supervisor and approver of visas and work permits. Studies conducted under CDP-SP II found that “irregularities, exploitation, abuse, and corruption” affect both Thai job seekers as well as foreign workers in Thailand. A number of useful measures could benefit both groups of workers. For example, having the right policy framework (i.e. what kind of employment foreign workers should be allowed to do) can indicate growth areas of skills that the nationals can usefully pursue. In terms of administration, greater transparency is critical for the welfare of workers as well as for increasing the efficient allocation of labor. For example, it can reduce the time between the application for and granting of a permit to a foreign worker, especially for skilled labor which is in short supply in Thailand. As far as MOL operations are concerned, the role of public employment services can expand following the trends of the modernizing economy. The shares of vacancies for higher skilled workers are increasing, and this is the area where employers report the greatest shortage. Thus, public employment services should adjust their operations accordingly, ideally with the involvement of private services. In general, improvements in the education and training systems can, however, have a much greater impact on worker mobility than under an expansion of public employment services. This does not mean that public employment centers should not be professionalized and computerized. It implies that complementarities between private and public services should be explored and that information on job applications and vacancies generated by these services should be provided to the education and training authorities. The information should then be used by them for education planning and provision of training services. In this regard, strong links between the MOL and other Ministries are critical. Unemployment Insurance The CDP-SP provided a useful initial assessment of the early experience with UI but the right rates of payroll contributions and effects on the behavior of workers would need to be assessed further as the program matures. In 2004, the UI benefit was paid to 50,000 beneficiaries. However, during January-November 2005, the number of beneficiaries increased more than fourfold to 220,000. This growth may be genuine in the sense that it reflects greater knowledge of workers about the scheme.22 It may also reflect changing behavior by workers or their employers (the latter because they face less resistance to firing). Thus, the rates may be appropriate at this early stage of implementation of the UI program but would need to be reassessed periodically. The effect of benefits on the incidence and duration of unemployment would also have to be 22 Note, however, that unemployment has declined over time from a year-round average of 2.2% in 2003 to 2.1% in 2004 and from 1.5% in the 4th quarter of 2004 to 1.3% in September 2005 (World Bank, 2005b). 17 examined. Finally, any extensions to those employed in small establishments (employing fewer than 10 employees) should be carefully considered in view of the regressive nature of the program and the more limited ability of small enterprises to cope with payroll contributions. Occupational Safety and Health The current coverage and efforts in this area are small and their expansion should take into account the fact that OSH involves many agencies outside the MOL. The MOL has developed its Occupational Safety, Health, and Working Environment (OSHE) Master Plan for 2002-2006, with a target of reducing work-related injuries by 10% per year. However, additional agencies involved in this area include, for example, the Ministry of Industry and the Ministry of Public Health. These ministries also have their own regulations, inspections, and promotions. Measures additional to inspections (“sticks”) can also be used and accompanied by actuarial assessments of the level of contributions (“carrots,” in the form of reduced contributions from establishments that improve their OSH record). In addition to enforcement, public promotional campaigns—especially with the participation of socially responsible employers—can promote practices for safer workplaces and help improve Thailand’s OSH indicators. Activities could include training, dissemination of information, technical and advisory services to establishments, research and development, and focused support to high-risk industries and to small and medium-sized enterprises (given the large share of the labor force employed by them). The benchmark analysis of organizational structures in other countries conducted under CDP-SP II provides a useful platform for addressing OSH more effectively and efficiently. 23 As far as contribution rates by employers to the WCF are concerned, they can be “experience rated” (i.e. set according to the type of industry and the employer’s safety record ), but care should be exercised to avoid accumulating surpluses beyond what is actuarially justified. 5. Informality and Employment Programs Although informality is declining, it remains high and thus presents a serious challenge for expanding coverage of social insurance. Although the informal sector is difficult to define and national conventions for measuring it vary, Thailand has an exceptionally large share of informal workers, even more so when its per capita income is taken into account. For example, the share of informal workers in Thailand is higher than in the Philippines, which has less than half the per capita income of Thailand (Table 2). 23 The RTG is considering the establishment of an Autonomous Public Organization (APO) to be in charge of OSH promotion under MOL. The international review conducted under CP-SP II examined APOs in Australia, Canada, Germany, Japan, Korea, Malaysia and the United States and suggested a functioning APO be developed and managed on a tripartite basis. 18 Table 2: Share of Informal to Total Employment and Per Capita Incomes, East Asia Region Country Income Share of informal to total labor force Thailand 2,540 53 Philippines 1,170 43 Malaysia 4,650 31 Taiwan, China 14,630 20 Indonesia 1,140 19 Mongolia 590 18 Hong Kong, China 26,810 17 Vietnam 550 16 China 1,290 13 Singapore 24,220 13 Source: World Bank (2006) High informality acts as a barrier to the expansion of social insurance and limits the role of labor market programs. For example, although employment services can address significant information gaps in the labor market and facilitate labor mobility, they may have limited impact. Placements by employment offices in Thailand account for about 5% of job placements,24 mainly for low-educated, low-skill workers. Similarly, although the introduction of the unemployment insurance (UI) program was probably justified, it still covers only a fraction of the labor force and is likely to be regressive, even among wage employees insured with the scheme. The UI program was introduced in January 2004. The program covers only formal sector workers in establishments with 10 or more employees. It sets low levels of contributions: 0.5% of monthly wages by employers and employees and 0.25% for the government. The compensation is time-bound—maximum six months for the involuntarily discharged—at a replacement rate of 50% of earnings. For those who voluntarily quit, the replacement rate is 30% and for a maximum period of three months. However, a simulation analysis of the UI scheme indicates that the share of the bottom two per capita monthly wage quintiles in UI contributions is likely to be much greater than their share in payouts (15% versus 4%) compared to the top wage quintile (59% versus 44%).25 This difference occurs because unemployment rates are lower, and unemployment duration smaller, for workers working in smaller establishments than in larger establishments. As in the case of UI, increased attention to occupational safety and health (OSH) issues is needed, but again it can only benefit few workers under present labor market conditions. Thailand faces high rates of work-related injuries and illness,26 but since nearly two-thirds of the labor force operates in the uncovered sector, many work- 24 Each year, there are around 700,000 job seekers in Thailand. 25 Deolalikar (2005). 26 Work-related injury and disease statistics in Thailand are generated from only around 20% of the working population and are designed mainly for internal administration of the WCF, not to monitor and evaluate work-related injuries and diseases. 19 related injuries and diseases remain unreported. The numbers of employers and insured workers covered by the scheme were around 80,000 and 7 million, respectively, in 2003. OSH regulations are enforced mainly through inspectors at the MOL who number only around 700 nationwide, and fewer than half of them are trained safety inspectors. Each year, about 5-6% of over 400,000 establishments are normally inspected, although not only for OSH but also for other labor issues such as labor relations, female workers, and child laborers. The Provincial Labor Offices are not able to collect and disseminate information efficiently as they lack specialized staff and software and have no systematic procedures for labor data storage (database system).27 Thus, the conventional notification systems need to be supplemented from other sources such as household surveys. Overall, as a consequence of the large share of informality, only a very small part of the population is covered by social Proportion of People with Social Insurance by Expenditure Decile insurance, even among the better off. Although the percentage of the 25.00 population covered by social insurance 20.00 provisions has been increasing, it still 15.00 2002 (%) 2004 10.00 remains below 20% for those in the top 5.00 two income deciles. Less than 10% of 0.00 the population in the sixth decile and 1 2 3 4 5 6 7 Expenditure decile 8 9 10 lower, and practically none in the lowest Source: SES quoted from Chandoevwit, Siamwalla and Sutummakit (2006) income decile, are insured under the public social insurance system. 6. Social Pensions and Community Savings One of the most prominent forms of social protection in Thailand is the relatively extensive social pension program. Social pensions reached 66,500 villages and 320,000 old persons in 1998. The beneficiaries are individuals over 60 years of age who are economically inactive and live alone or have no one to take care of them. According to the latest count in 2003, the number of beneficiaries had increased to 400,000 out of approximately 875,000 poor old-age persons. The program pays a modest monthly pension of THB200-300, and its overall budgetary costs are around THB1.2 billion. Although the program is progressive overall, only 13% of the elderly in the poorest quintile are covered, and 60% of beneficiaries are non-poor, including 5% from the top quintile. The reasons for this disparity between the percentages of poor and non-poor elderly receiving social pensions are not very clear. It would be helpful to supplement these findings of incidence analysis with additional examination of administrative and institutional factors that may be responsible. The CDP-SP II also examined community-based saving initiatives—with which Thailand has some experience—in relation to pensions. A background study provided significant insights on the willingness of some communities to save for participation in a “defined contribution” pension scheme. It also explored what the retirement age should 27 See TDRI (2005). 20 be, how the funds should be managed, the appropriateness of regulations, the role of the community, the need for government to “top up” the funds or issue guarantees, and other issues. The study also identified an interest in a “defined benefit” scheme funded on a pay-as-you-go basis. In return for some regular contributions, members could receive benefits for health care, disability, maternity, child support, death, income replacement, and a minimal pension. Another background study showed some community preference for the pension system for informal workers to be based on voluntary individual accounts with defined contributions. For rural informal workers, the system would take advantage of the existing savings culture and use community mechanisms to collect contributions and pay benefits. For urban informal workers, the system could be linked with professional or occupational associations for collecting contributions and paying benefits and/or use some community mechanisms. In practice, however, workers might be reluctant to join due to concerns about the returns on investments of the fund, the community’s capacity to manage the funds itself, information requirements, and other factors. The eligibility criteria and coverage of the current social pension scheme can be broadened and more focused on the poor. Focusing on the poor and reducing benefit payments to the non-poor (who currently account for 60% of beneficiaries) would allow more coverage than the current 400,000 among the 875,000 elderly poor. Importantly, the SES 1998 and SES 1999 data indicate that the incidence of income poverty among the elderly who live alone is almost negligible, while the incidence of poverty among the elderly rises sharply with their household size. This finding suggests that the appropriate target group for the social pension scheme should be the elderly who live in large households, possibly containing many dependents. In addition, some international evidence suggests that social pensions are effective when the recipient lives in a poor family—often the result of old people being predominantly female (often widows) who do not see their needs in isolation of those of their offspring.28 As the population ages, old age security will become a critical social protection issue in Thailand. While government and voluntary schemes (see below) can each play a critical role, both will need to be considered and integrated in a broader system that would include occupational pensions and an expanded and better-targeted social pension program. As the experience of many economies in demographic transition suggests, pension reforms are among the most complex and involve political economy considerations. Options for creating a savings scheme for informal sector workers are being studied by the Fiscal Policy Office (FPO) of the MOF, but this analysis is hampered by inadequate data and operational understanding of the issues involved. It will be important for Thailand to start defining its old age security approach and policies early in order to integrate its formal and informal systems better and avoid taking measures amidst the pressure that the demographic transition will create in the future. 28 For example, the presence of a grandmother in a poor household can relieve some of the household and childcare burden from the mother, who in turn can work while the social pension of the grandmother would be spent on family needs. 21 7. Social Assistance The above examination of the poverty, labor market programs and and regional inequality indicates the critical role that informality and the rural sector play for income security and social protection. An analysis of the effects of income changes confirms the dominant role of farm incomes and the marginal role of transfers in poverty reduction. Based on simulations, the analysis examines how a 10% across-the-board decline of a specific source of income (i.e. wages) affects the poverty headcount index.29 As Table 3 shows, the biggest effects on poverty arise from changes in farm incomes and the smallest ones from transfers. This analysis reinforces the message that policies in Thailand should take into account regional and agricultural issues and concentrate for the time being on the informal sector. Table 3: Increases in Headcount Poverty Associated with Reductions in Different Sources of Incomes Simulation Simulated Reduction in Income Headcount Increase 1 Wage Income 0.7 2 Enterprise Income 0.1 3 Farm Income 1.2 4 Transfers 0.2 Source: Alwang and Siegel (2002), Table V.2 Social assistance is extremely limited. Although the increasing shift away from institutionalized care and toward community care is a step in the right direction, Thailand has few residential facilities. The MSDHS provides social assistance through its 25 provincial centers for 8,000 orphans and disadvantaged children; 20 centers for 3,000 elderly; 11 centers for 8,000 beggars and the homeless; 9 rehabilitation and training centers for 700; and 11 training centers for 9,500 disadvantaged women. Aside from family and community help and support, the poor and the indigent have little to rely upon. Furthermore, those outside the formal employment sector receive benefits only from the RTG’s universal health insurance scheme. These persons include private employees in agriculture, the self-employed, those out of the labor force, and often those employed in smaller establishments. The limited effects of formal social protection are caused mainly by the limited coverage of programs, but are also compromised in some cases by poor targeting. In addition to employment-based programs of UI and OSH, Thailand’s social assistance includes residential care, social pensions, universal health insurance, “village funds,” and school meals. However, many program beneficiaries are not necessarily poor, as discussed below. Box 3 also provides a summary of CDP-SP II findings regarding social assistance. 29 The decline is simulated by subtracting 10% of the wage income base from total household income for all households and examining the resulting change in distribution. 22 Aware of these limitations, the Government has recently introduced some large and ambitious programs such as the “30-baht universal health coverage scheme” and the “one-million baht village/urban fund”.30 The widespread availability of the generally affordable 30-baht health scheme has resulted in virtually no significant differences in non-coverage rates across different income groups. The 30-baht health scheme also appears to be progressive. Coverage of the scheme in 2002 was 86% of the bottom quintile compared to 44% of the top quintile. Taking into account other forms of health insurance brings the coverage rate to 87%, more or less evenly distributed across the various income deciles. Box 3. CDP-SP II and Social Assistance The CDP-SP II identified many issues in the area of social assistance and provided recommendations which are summarized below. Findings 1. Relatively little useful information is available for measuring the effectiveness of programs both in terms of their coverage and adequacy. 2. Limited systems exist for tracking demographic changes by region and expansion or contraction of services by region. 3. Programs offered by MSDHS have gaps in terms of coverage and may not be cost-effective. 4. Benefits are often paid to the non-poor, including those in the richest decile of the income distribution. 5. Existing programs miss some critical social areas, such as social misconduct. 6. While placing increased emphasis on community care is appropriate, coverage of institutionalized care is at present very limited (e.g., too few centers serving too few poor people). 7. Although decentralization is moving service provision to the regions, local authorities have limited capacity and understanding for undertaking their new functions, and central ministries need to increase their role as promoters, supporters, regulators, and advisers. 8. The effectiveness of programs and policies is reduced by lack of clear definitions of the target groups and adequate instruments for identifying their problems and correct policy options. Recommendations 1. Develop information systems and monitoring mechanisms to identify the demand and supply of social assistance programs. 2. Introduce impact evaluation systems for social assistance programs to measure their effectiveness and efficiency. 3. Enhance capacity through training, assistance, and support to local administration organizations in the delivery of social assistance. 4. Expand existing pilot programs, such as the Key Performance Indicators. 5. In view of decentralization, build capacity to monitor and evaluate the performance of welfare centers providing residential care services to children, disabled, and the elderly. Some areas for improvement in the “One Million Baht Village Fund” have been identified.31 The Village Fund has a coverage rate of only 10% in rural areas, although the scheme was never meant to assist everyone in every village—only the neediest households. However, the SES data indicate that the scheme has no progressivity: 30 However, as Deolalikar (2005) notes, “some of the universal programs are not really universal and miss the poor and the potentially vulnerable for reasons not well understood. More than 7 million people (13 percent of the population) do not have any health insurance coverage. The reasons why so many Thais do not have health insurance coverage are not clear.” 31 Under the one-million baht village fund program, each village or urban area in the country is allocated a budget of one million baht as a revolving fund to be utilized within the village for common as well as individual household activities. 23 roughly 8-12% of all rural consumption quintiles benefit equally from it. This information suggests that the mechanism for allocating funds among households within a village could be improved, with more attention paid to the poverty status of a household or a household’s need for funds.32 In addition, some focus groups who met in the context of CDP-SP II activities indicated that the scheme “discourages informal saving cooperatives and encourages indebtedness of the household.”33 Two other large-scale programs to be considered as part of broader social protection, although formally outside the purview of the MOL and MSDHS, are the main school feeding programs. The free school lunch program and a supplementary food (school milk) program are administered by the Ministry of Education. The school lunch program was originally intended to assist undernourished children (as measured by weight-for-age) in village primary and pre-primary schools in obtaining adequate food. Over time, the program was expanded to cover all primary and pre-primary schoolchildren from poor families, whether undernourished or not, and it currently covers 54% of all children attending school. 8. The Administrative Challenge: Agency Coordination and Decentralization Even if the right programs have been identified, their impact would depend on effective implementation. Effective implementation requires, first, the coordination of the many agencies involved in offering programs in the area of social policy. These agencies include Social protection policies and programs involve the Mo Labor, Mo Social Development and Human Security, Mo Finance, Mo Information and Communication Technology, Mo Education, Mo Public Health, National Statistics Office, National Housing Authority as well as regional offices of the above and the private sector. These actors can be brought together and their roles and responsibilities be better defined though the preparation of an integrated “Human Development and Social Protection Strategy”. Their actions can be integrated and disciplined though performance based budgeting (instead of unit-cost budgets). Second, common to the success of many social programs in Thailand today is their dependence on the outcomes of decentralization. Therefore, the transfer of budgets and functions from the central government to local authorities needs to be smooth and effective. The labor market must also continue to perform well. Programs that are introduced prematurely and are over-designed (in terms of coverage or the conditions they prescribe) can act as barrier to economic and social development as they can increase costs and reduce the impact of benefits. Fiscal decentralization is proceeding, but more slowly than initially planned. The share of local government revenue to national government revenue increased from 22.8% in FY 2004 to 23.5% in FY 2005 and is budgeted to reach 24.1% in FY 2006. However, the share of local government revenue remains well below the initial target of 35% set in the Decentralization Act (1999). 32 Deolalikar (2005). 33 Chandoevwit, Siamwalla, and Sutummakit (2005). 24 The transfer of some social sector functions to local governments, such as education, has been delayed. The intended progress may have been overestimated in the Decentralization Act of 1999, due in part to the need to revise laws and regulations. Furthermore, although grassroots strengthening invariably takes time, capacity at the local level remains low. The process of capacity building for service management at the local level has been slow and affected by the slow process of decentralization of education personnel. During FY 2003-2005, only 4,500 personnel were transferred to local governments, accounting for 0.4% of total civil servants. Notably, budgetary reforms and government decentralization may have reduced the targeting of old age pension. For example, the transfer of benefits administration to the Tambon Administration Organization (TAO) in 2002 appears to have exacerbated existing leakages of the old age monthly allowance scheme. The proportion of beneficiaries from the poorest two deciles declined, while the proportion of the richest two deciles increased, and the new recipients are distributed more evenly across all expenditure deciles.34 According to the relevant CDP-SP II study, the cause may not be because the TAO selected new recipients to replace pre-existing ones but because, when allocating their own revenues collected at the local level, some TAOs concentrate on the old in their own area. People living in outside areas may thus be deprived of cash allowances that were previously provided by the central government budget. One unintended effect of decentralization may be that otherwise eligible beneficiaries lose benefits, because they no longer satisfy the residence criterion in the area where services are provided. As in the case of the old age monthly allowance scheme, certain local authorities may be reluctant to support elderly from other areas or provinces. For example, since Thailand has only 20 public home cares for the elderly across the 76 provinces, provincial governments with these services have an incentive to accept elderly only from their own province. Therefore, people living in provinces without public home care for elderly might have difficulty accessing these services elsewhere. Finally, the transfer of the administration of public home care for elderly from central to local government may have increased its costs. The administration of this activity, originally undertaken by the Department of Public Welfare, was passed on to the provincial governments in 2003. Looking at changes of unit costs over time, a CDP-SP II study found that in Chiang Mai, for example, the cost of care increased fourfold from nearly THB27,000 in 2002 to THB118,000 in 2005. No immediate explanations are available for this sharp increase in unit costs. Although this evidence cannot be generalized across the many programs offered in different regions, this issue should be examined further in future evaluations of programs and the effects of decentralization. 34 Chandoevwit, Siamwalla and Sutummakit (2005): Figure 3.5. 25 9. Statistical Requirements Informed policy design and evaluation of programs requires better coordination of statistical collection among agencies is needed to build on complementarities and foster cross-fertilization. The collection of statistics in Thailand is decentralized, and individual agencies usually focus on data for their own use. At times, there is duplication among the different agencies involved with the same issue and, more importantly, conflicting results due to the different concepts, coverage, and collection procedures used. As a result of efforts under CDP-SP II to develop Thailand’s Data Directory, it was found that there are over 200 departments in 20 ministries producing 20,000 data series. The pilot development of the Labor and Social Welfare Statistical Master Plan, also under CDP-SP II, represents a positive step toward improving coordination among agencies. In addition, making administrative and survey data available to the wider public will help contribute to a better understanding of social protection issues and increase accountability. The availability of panel data would help inform the understanding of the dynamics of poverty and vulnerability. Thailand currently does not have panel data, which can provide valuable information not only during periods of unexpected crises but also for more regularly recurring adversities such as droughts and floods. Analysis of panel data can contribute to a better understanding of poverty and vulnerability as well as enable the timely introduction of appropriate policies and appropriate rate setting for raising funds and benefit payments. D. SUMMARY AND RECOMMENDATIONS The Ninth Plan calls for a more balance economic and social development that would reduce poverty and narrow inequality through more accessible and better education, fiscal reforms, increased decentralization, improved administration and management and greater civil society engagement. From Thailand’s historical record, these aspirations should be within reach given, following, on the one hand, the continuation of sound economic and social policies that have been adopted in the past and led to poverty reduction, human development, employment creation, low unemployment and, on the other hand, a gradual approach to the introduction and expansion of social programs based on relatively good statistics and analysis. Thailand’s gains can further consolidated and enhanced by the findings of the CDP- SP. These findings indicate that Thailand needs to expand its programs, although some are more promising than others, while a few have significant issues to be addressed before they are considered further. The CDP-SP provided valuable insights for reforms in the areas addressed by the program. More specifically, the program analyzed dimensions of poverty and vulnerability in both the formal and informal sectors. It also reviewed nearly all formal programs in the social protection area, providing more information about their characteristics and outcomes. The CDP-SP II also examined the 26 technical and administrative constraints in the Ministries concerned in view of the Government’s desire to make more effective use of public funds, decentralize delivery, and increase coverage of programs. Finally, it set international benchmarks for improvements in the administration of some areas of social insurance, social assistance and safety nets by the key agencies concerned. All this provides useful information on labor market and social indicators including minimum wages, regional dimensions of poverty and inequality, coverage of formal social protection among the poor and also the population at large, and the way incomes from various sources can help the poor mitigate shocks. On the positive side, Thailand does not face the issue of “which program to introduce” but “which existing programs to expand and under what conditions.” In addition to its historic approach to gradualism, Thailand has the advantage of having experience with nearly all aspects of social benefits and programs. Although many programs are small, it is difficult to identify a benefit being offered elsewhere but not in Thailand or a vulnerable group that is excluded completely, even if its coverage is low. The multiplicity of already existing social protection programs provides a laboratory for examining the properties of each one before they are expanded or contracted. However, many aspects of social protection remain unexplored, even for “first generation” policies that need be completed before expanding more ambitious programs. The analytical work undertaken under the CDP-SP II focused mainly on promising “second generation” issues (i.e. cash transfers, unemploym ent insurance, occupational safety and health) that are increasingly important as Thailand moves from middle- to high-income status. However, some more basic “first generation” issues remain unresolved. For example, in the area of health, who are the more than 7 million uninsured, and why? Since education enrollment continues to be less than universal and lower than the regional average, how can synergies between formal social assistance and education be improved? Furthermore, Thailand still has some structural characteristics that can hamper the outcomes of economic and social policy. These characteristics can adversely affect the momentum of employment creation, the stabilization of farm incomes and the achievement of greater regional integration. Finally, little is known about expenditure leakages or the link between budgeted expenditures and actual disbursement. The key challenge in translating the individual program findings of the CDP-SP II into action is to put them in the context of both Thailan d’s current developmental stage and its desired developmental goals. Rapid progress in addressing key challenges, such as the extension of social insurance in the labor market or social assistance to the needy groups, may be desirable. However, progress can be compromised by both economic and administrative constraints. The issue is how to prioritize the objectives of social protection and sequence efforts to improve administrative and institutional capacity to achieve these objectives. For example, should more resources be spent immediately to improve the geographical coverage of labor inspections or the centers serving orphans? What should the right mix of programs be? 27 Nevertheless, it is important to maintain the momentum of the CDP-SP II program going forward. Social sector reform is always “work in progress.” The progress and gains made by Thailand can be enhanced by developing and implementing a national strategy paying particular attention to the informal sector and certain regions, improving the administration of policies – including decentralization but also greater coordination among agencies, and improving the statistical systems needed for informed policy analysis, including the preparation of annual reports (such as regional public expenditure reviews at the regional level) that would monitor and evaluate trends in social protection. This report proposes the following key areas for further consideration by the RTG. A. An integrated strategy and coordinated implementation that would also include the informal sector. The first order of priority is to increase the internal efficiency of existing programs. The CDP-SP provided lessons that, if studied further, can help to achieve additional results with the same resources, for example, through better targeting. This can be done relatively easy as it does not require much additional effort and falls within the purview of the individuals agencies that offer these programs. To increase the effect of individual programs, it would be necessary to develop and implement an overall strategy for human development and social protection: a more integrated approach can exploit synergies between different programs and, while program outputs may be the same, the outcome of social protection policies can increase. Although Thailand’s record provides reassurances that it can avoid the chances of “doing well the wrong thing”, poverty and vulnerability can be reduced further in the presence of a national strategy. This need is probably greater in Thailand than elsewhere. Although Thailand has an impressive development record, it still faces some structural issues such as regional inequality as well as periodic floods and droughts and the uncertainty surrounding avian influenza. A national strategy should take into account national measures that would increase regional integration and lower variability in farm incomes. With respect to the social sectors, the policies and programs of all agencies involved should complement each other in a way that increases the impact of public expenditures while crowding in private funds and efforts. To be successful, a national strategy needs to pay more attention to the informal sector: While the expansion of “second generation” social protection programs is important, it should be recognized that their effect on the employment and social protection outcomes for the large majority of workers is marginal. Even if such programs are designed appropriately, their impact is likely to be limited given the large proportion of workers in the informal sector. For example, the UI scheme excludes about two-thirds of workers in Thailand who are own-account and unpaid family workers and employees in enterprises with less than 10 employees. Although coverage may expand in the future, for current purposes, the welfare of the large majority of Thai workers must rely on other interventions. Similarly, OSH covers only 7 million workers in a labor force of more than 36 million. Expanding these programs would involve significant effort and 28 cost and would not be able to produce the same scale of results that would arise from an expansion of education and skills creation, successful regional and agricultural development, and a balanced decentralization process. Broadening access to social protection should be based on the gradual expansion of formal mechanisms and take into account Thailand’s significant social capital and community-based mechanisms. Decisions for the expansion of formal mechanisms should be evidence-based decisions. Taking advantage of informal mechanisms already in place calls for a multi-pronged effort that would increase the public/private and formal/informal synergies. However, it should be also recognized that many communities do not have the capacity to evaluate and implement the best options available to them. Prematurely relieving the responsibility of government to assist them can have adverse effects. It is not just enough to share common objectives but work independently of each other towards them, and agency coordination should therefore increase. They can accomplish more by cooperation that would complement each other’s efforts and by sharing information. Although the capacities of the MOL, MSDHS, and other involved agencies were enhanced as a result of the CDP-SP, their capacities to design the right policies at the broader national level and to coordinate their policies with the many other agencies involved in social services require further strengthening. B. Balanced Regional Development Although mainly outside the mandate of the MOL and MSDHS, regional assistance could be directed to poorer areas, especially the Northeast. The more successful regional development is, the less the reliance on formal social protection will be. As noted above, the recent reductions in poverty have been driven primarily by reductions in the Northeast. Yet the Northeast receives less public funds per person than other regions, and given that the share of the poor living in the Northeast exceeds the share of the poor living elsewhere, it receives even less public funds per poor person than other regions. Given that the Northeast has the highest population density outside Bangkok and vicinity, reaching the poor in the Northeast also means reaching a large number of poor. Evidence also suggests that public expenditures appear on average to be spent more efficiently in the Northeast than in other regions, at least as far as secondary education is concerned. Insurance against natural disasters and weather adversities could be also considered. Given the clear effects of changes in farm incomes on increasing vulnerability, measures to reduce this variability are important. For example, index-based weather insurance, if well-administered, can be inexpensive and would allow for standardization, avoiding the need to draw up and monitor individual contracts. It can be supplied by the private sector with little or no government subsidies, as it avoids the incentive problems of crop insurance programs related to asymmetrical information. It is affordable for poor and rich farmers alike and accessible to agribusiness and other sectors. 29 Economic policies would have to strike a balance between supporting lagging regions to reduce poverty and achieve economic integration and tackling growth constraints in prosperous regions that underpin the country’s economic growth. The costs of ignoring competitiveness of the urban centers today can be very high, while neglecting the needs of people remaining in poor areas and villages is costly in human and economic terms. Urban centers would benefit from improved competitiveness, peri- urban areas from upgrading of services, and rural areas from off-farm diversification and farm productivity growth. C. More Effective Decentralization One major theme emerging from the CDP-SP II results is that good policies can be hampered by poor administrative arrangements, and in this respect perhaps the most important issue is decentralization. Delays in improving the capacity of the regions to undertake their new functions are highly problematic. Such delays may have much more important adverse effects than, for example, keeping the operations of the Ministry of Labor (MOL) and Ministry of Social Development and Human Security (MSDHS) at their current levels. Analysis under the CDP-SP II showed that much of social protection comes from stabilizing farm incomes rather than increasing benefits or even wages. Thus, although important, the formal programs of the Ministries cannot be expected to have the same impact as measures that affect the population at large. The administrative challenge of decentralization can only increase in the future as Thailand integrates its human development and social protection agenda. A more comprehensive approach to the social sectors would require the coordination of more diverse agencies beyond only the MOL and MSDHS. These agencies include–but are not limited to—the Ministry of Education, Ministry of Public Health, Ministry of Information and Communication Technology, Ministry of Interior, National Economic and Social Development Board, National Housing Authority, National Statistical Office (NSO), and various other government organizations (i.e. for veterans or women), as well as the Ministry of Finance, which would play a more coordinating role. This coordination will be necessary not only at the national level but also at the regional level due to decentralization, with increasing participation of non-formal providers of social services with communities, villages, and families. Thus, decentralization needs to become quickly operational in the broader context of a national strategy for human development and social protection. Decentralization needs to better define the functions and responsibilities of the sub-national agencies involved. The CDP-SP II warns that decentralization may have increased costs, at least in certain areas of social assistance. Developing criteria to disburse the funds and clarifying the flow of funds from the central government to the regions is particularly important, as significant financial resources are already being transferred from the central to the regional authorities. Ensuring the successful move from the current unit-cost budget system to a performance-oriented and forward-looking system is also critical. It would help regional development inasmuch as it is expected to help the Northeast and other deprived areas receive relatively more funds. 30 While the Government’s decentralization reforms are moving in the right direction, care should be taken to synchronize the introduction of new policies with the ability to implement. Decentralization should synchronize the transfer of revenues with the transfer of functions and be accompanied by capacity building. Experience in other countries has shown that it is more effective to decentralize expenditure responsibility before decentralizing resources. The focus in Thailand remains on the revenue side, in part because of the target set by the Constitution on increasing local revenues. Nonetheless, the central government has not always made the budgetary allocations on time, with the expected adverse consequences on social service delivery at the regional level. Improvements in central government budgetary allocations and disbursements would also be needed so resources are available to fulfill new responsibilities in a timely manner. If not done properly, decentralization can lead to increased costs as well as greater regional disparities and inequities. Efforts to decentralize central government functions have been underway in many developing countries since the 1980s. While studies have shown the positive effects of decentralization in some cases, studies in Vietnam and Argentina revealed that central governments were more effective than sub-national governments in targeting poorer areas.35 Other evidence indicates that central governments are generally better at redistributing resources to poor provinces than provinces are to poor areas. Thus, decentralization involves the basic consideration of which level of government could most efficiently and effectively deliver and finance public services. Although the results of decentralization in the initial stages as studied by CDP-SP II may be mixed, they do not imply that decentralization should be abandoned. Rather, the results underscore the importance of doing decentralization properly, with transfer of functional responsibilities from the central to regional authorities accompanied by capacity building. To be effective, constituents in local communities should be kept informed and have a voice in the decisions that affect them so authorities can held more accountable and improve public service provision. D. Further Analysis and Statistics Though the CDP-SP covered a wide array of issues, policies and programs, further probing is required into some additional areas. The gender dimension received little attention under the analytical work to date, as have some other groups (i.e. child laborers and domestic workers) and programs (such as credit programs). Even within areas addressed under the CDP-SP II, more in-depth analytical work could be undertaken to explore unanswered questions on program coverage and impacts. For example, why are so many non-poor old persons receiving the social pension, and how would poverty be affected if coverage were to be expanded to those who live with families? Additional areas worth exploring further include: an examination of public expenditures not only 35 Ravillion (1998); Van De Walle (2001). 31 from the perspective of coverage and benefit incidence, but also in terms of leakages and quality of services; a more complete mapping of the informal sector that would carefully assess differences between agricultural and non-farm rural employment, a better understanding of the interplay between formal social insurance and social assistance programs with comparable informal mechanisms; and the usefulness of previously used mechanisms (such as public employment programs or social funds) as stand-by measures to be activated should an adversity reoccur in the future. Interventions would benefit from improvements in the information systems for social protection, insurance, and analysis, including the availability of panel data. The new statistical act provides that the NSO will be responsible for the overall coordination of statistics in Thailand. This does not imply that line ministries and other agencies should be relived for the responsibility to collect and analyze relevant information. However, in order to do so, the agencies (such as MoL, MSDHS, MoE) would need to have the necessary capacity to fill information gaps in their relevant as well as the ability to undertake analysis that would ensure the relevance , proper targeting and monitoring and evaluation of the programs they offer. 32 REFERENCES Alwang, J. and P. Siegel (2002), Risk and Vulnerability Assessment for Thailand. World Bank: Social Protection Department (HDNSP). CDP-SP (2005a), Brief on Outputs/Outcomes of the Country Development Partnership for Social Protection. CDP-SP (2005b), Study on the Development of Workers Compensation Fund (WCF) Simulation Modeling and Actuarial Projections. CDP-SP (2005c), Executive Summary of a Pilot Project on Developing Indicators for Measuring the Effectiveness of Activities for Children, Older Persons, and Persons with Disabilities. CDP-SP (2005d), Executive Summary of a Participatory Survey on the Suitability of Using Community Mechanism to Mobilize Informal Workers’ Savings for Welfare and Retirement. CDP-SP (2005e), Executive Summary of a Social Statistical Master Plan (on Labour Statistics and Social Security Statistics). CDP-SP (2005f), Executive Summary of Research on the Permission Process for the Export of Job Seekers Overseas under the Act on Employment and Job Seeker Protection 1985. Country Development Partnership for Social Protection: Summary of Activities and Outcomes (2006). Chandoevwit. W., A. Siamwalla and N. Sutummakit (2005), Risk and Vulnerability in Thailand: A Quantitative and Qualitative Assessment. Thailand Development Research Institute (TDRI). December. Chaninat et al. (2005), Summary of Report on The Administrative Procedure and Administrative Order under the Working of Aliens Act B.E. 2521. Submitted to the Ministry of Labour, Royal Thai Government. Chulalongkorn University (2005a), Analysis of Data Structure of Each Government Offices. Report Submitted to The National Statistical Office Ministry of Information and Communication Technology, Royal Thai Government.Faculty of Commerce and Accountancy. June. Chulalongkorn University (2005b), The Design and Development for Common Sampling Frame of Business Establishments in Thailand: Phase II” by Research Unit, Faculty of 33 Commerce and Accountancy. Submitted to the National Statistics Organization, Ministry of Information and Communication Technology, Royal Thai Government. Deolalikar, A.B (2005), Social Risks and Responses in Thailand: Analysis of Household and Individual Survey Information. University of California: Riverside, CA, USA. December. Ellis, A.F. (2005), International Experience of Enforcement of OSH Law and Regulations. August. Falter, J.M. (2005), Minimum Wages and the Labour Market: The Case of Thailand. Kogi, K. (2005), International Experiences in Autonomous Public Organizations in the Field of Occupational Safety and Health. Ravallion, M. (1998), Reaching Poor Areas in a Federal System. The World Bank: Policy Research Working Paper, 1901. Rinehart, R.D (2005), Design and Administration of OSH Information Systems in Nine Countries” Report Submitted to the Department of Labor Protection and Welfare, Ministry of Labour. Royal Thai Government. August. Royal Thai Government (2001), Summary of the Ninth National Economic and Social Development Plan 2002-2006. Royal Thai Government. (2005), Results-Based Summary of Sub-Activities Financed by CDP-SP II. Ministry of Social Development and Human Security, Office of Permanent Secretary. February Siam Development Institute (SDI) (2005), Executive Summary on Preliminary Feasibility of Savings and Pension System Options for Informal Workers in Thailand. Report Submitted to the Ministry of Finance, Bureau of Policy on Savings, Fiscal Policy Office, Royal Thai Government. Thailand Development Research Institute Foundation (TDRI) (2005), The Study on the Development of the Capacity of Provincial Labor Information Centers. Report Submitted to Ministry of Labour, Royal Thai Government. December. Van De Walle, D. (2001), The Static and Dynamic Incendience of Vietnam’s Public Safety Nets. The World Bank. World Bank (2001), Thailand Social Monitor. November. World Bank (2002), Thailand: Country Development Strategy, Social Protection. Human Development Sector Unit, East Asia and Pacific Region. Report No. 24378. June 2002. 34 World Bank (2003) ASEM Grant for Strengthening Social Protection for the Working Population, the Poor and the Vulnerable in Thailand: Grant No.TF050975. June 13. World Bank (2005a),Thailand: Northeast Economic Development Report. Poverty Reduction and Economic Management Sector Unit, East Asia and Pacific Region. Joint Report of Thailand’s National Economic and Social Development Board and the World Bank. June. World Bank (2005b), Thailand Economic Monitor. November. World Bank (2006), Doing Business in 2006: Creating Jobs. 35