SOCIAL SECURITY IN ASIA AND THE PACIFIC. Southeast Asia s social security systems: Need for a system-wide perspective and professionalism

Transcription

1 SOCIAL SECURITY IN ASIA AND THE PACIFIC Southeast Asia s social security systems: Need for a system-wide perspective and professionalism Mukul G. Asher Public Policy Programme, National University of Singapore This paper provides an assessment of the social security arrangements in Southeast Asian countries, and suggests broad directions for reform. Noting the absence of functioning multitier social security systems in these countries, the analysis highlights wide variations in various characteristics of systems in Southeast Asia. The paper underscores the need to enhance professionalism in the performance of core functions of the provident and pension fund organizations in these countries; and to take a system-wide perspective. It recommends constituting a Provident and Pension Fund Authority (PPFA) in each of the Southeast Asian countries to undertake the above tasks. There are five core functions which any social security organization must perform (Ross, 2000). These are reliable collection of contributions, taxes, and other receipts, including any loan repayments; payment of benefit for each of the schemes in a timely and correct way; securing financial management and productive investment of provident and pension funds assets; maintaining an effective communication network, including development of accurate data and record-keeping mechanisms to support collection, payment, and financial activities; and production of financial statements and reports that are tied to providing effective and reliable governance, fiduciary responsibility, transparency, and accountability. A well-designed multitier social security system comprises several such organizations which collectively need to provide, on a sustainable basis, adequate, equitable, and fiscally affordable retirement income to as high a proportion of the population as feasible. This suggests that the needs of older people must be balanced against the needs of others. Each country will need to evolve its own trade-offs and consensus 71 International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

2 Table 1. Southeast Asia: Selected economic, demographic, and labour force indicators Country Total GNI a population (US$ (million) billion) Per capita GNI a (US$) Average annual rate of population growth Average annual rate of labour force growth Total Proportion fertility rate of population (TFR) b above 60 Elderly Life dependency expectancy ratio at birth (EDR) c Indonesia Malaysia , Philippines , Singapore 4 d , Thailand , Sources: World Bank, 2002 (Table 1, p. 232); Heller, 1997 (Table 1): his estimates are based on unspecified statistical sources of the World Bank; World Bank, 1994 (Tables A1 and A2, pp ); Bos et al., 1994 (Table 10, pp ). a GNI = Gross national income. b TFR is the average number of births per woman in the population. A TFR of 2.1 maintains a stable population, assuming no net migration takes place. c EDR is defined as the ratio of persons 60 years and above to persons aged d Singapore s resident population (i.e. citizens plus permanent residents) was 3.2 million, the difference constituting foreigners working in Singapore and their dependants. 72 among the generations. Thus, social security reform involves enhanced professionalism with which the five core functions mentioned above are performed by each concerned organization; and rebalancing of various elements of the social security system to attain its objectives. It is in the above context that this paper assesses the social security arrangements in five Southeast Asian countries, namely Indonesia, Malaysia, the Philippines, Singapore, and Thailand. These countries are still coping with the aftermath of the 1997 economic crisis, particularly with diminished medium-term growth prospects. The 1997 crisis also underscored the need not to rely on economic growth alone to provide economic security, but consciously to develop social security institutions and systems (Asher, 2002). The demographic trends in Southeast Asia suggest rapid individual and population ageing, resulting in a significant reduction in growth of the labour force (Table 1). In conjunction with increasing urbanization and industrialization, and a shift towards more individualistic lifestyles and preferences, the demographic trends suggest increased reliance on formal social security systems in the region (Asher, 1998). The rest of the paper is organized as follows. The following section provides an overview of the main characteristics of the social security systems in Southeast Asia. This is followed by a discussion of reform directions in the third section. The final section provides concluding remarks. International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

3 Table 2. Key provident and pension fund organizations in Southeast Asia Country Indonesia Malaysia Philippines Singapore Thailand a Organization PT Jamsostek and PT Taspen Employees Provident Fund (EPF) Government Pension Fund, Malaysia (GPFM) Social Security System (SSS) Government Service Insurance System (GSIS) Central Provident Fund (CPF) Government Pension Fund, Singapore (GPFS) Social Security Organization (SSO) Government Pension Fund, Thailand (GPFT) Source: Information obtained from official sources in individual countries. a Thailand has mandated its state enterprises to set up provident funds, but it is not mandatory for the employees to join. By 2000, 54 out of 64 enterprises had set up provident funds covering 211,000 workers, with contributions varying between 3 and 15 per cent. Thailand also has a large number of private sector provident funds. As of 1999, these covered 1.03 million workers in 4,005 enterprises, with total balances of billion baht, or US$ 5 billion (Jirapon, 2000). Main characteristics 73 The social security systems in Southeast Asia are in the early stages of development, with none of the countries currently having a functioning multitier system. The five countries covered in this paper, however, exhibit considerable heterogeneity with respect to the various characteristics, as evidenced from the discussion below. Philosophy and governance structure Table 2 identifies key provident and pension fund organizations in the five Southeast Asian countries. The Philippines and Thailand have accepted the principle of social insurance for both private and public sector employees, although, in the case of Thailand, the private sector pensions will not begin till the year The maximum pension benefit will be 30 per cent of the average wage of the employee over the last 60 months. In the other three countries, only the civil servants and politicians continue to benefit from defined benefit pensions, with inflation, longevity, and survivors protection largely or wholly financed from the government budget (Asher, 2000), while the private sector employees rely on mandatory savings with little or no protection against various risks. The existence of separate government pension funds and payment of pension from current budgetary revenue implies that government pensions are financially International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

4 Table 3. Social security coverage of private sector workers in Southeast Asia (2000) Country Malaysia (EPF) Philippines (SSS) Singapore (CPF) Thailand (SSO) Contributors as percentage of labour force 54.7 a b 58.1 c 19.5 d Source: Information obtained from official sources in individual countries. a Includes 0.6 million foreign workers. b Membership of the SSS is 23 million but the active contributors are 6-8 million. c Foreign workers are 25 per cent of the labour force and are excluded. d The SSO coverage is overstated as the figure refers to members rather than active contributors. If the provident funds of state-owned enterprises are included, the coverage may increase to 25 per cent. 74 more secure, and that a portion of the future tax revenue has already been allocated for the payment of future pensions. Indonesia and Malaysia exhibit marked dualism in their social security systems, which is also evident to a lesser degree in the Philippines and Thailand. Singapore has been progressively reducing the categories of civil servants eligible for pension (Asher, 2000). As a result, only high-level officials and politicians effectively remain eligible. Malaysia and Singapore s national provident funds are essentially mandatory savings schemes which are publicly managed. The EPF of Malaysia and the CPF of Singapore are statutory boards. In both cases, securing the services of independent-minded and competent personnel has been a challenge. Malaysia s somewhat open political system and society, however, do provide a degree of transparency and accountability (Asher, 2002). Indonesia s provident and pension fund organizations do not have a statutory status, but are organized as private limited companies, and therefore pay income taxes. The usefulness and timeliness of the information made available by the respective organizations in five countries vary considerably. There is a fairly long delay in publication of annual reports by the organizations in four countries, while such a report for Indonesia is usually not readily available. Only Malaysia s EPF and Thailand s GPF provide relatively detailed information in their respective annual reports. Singapore s CPF has the capability to provide considerably more information, but it appears to regard information as a strategic good rather than a public one. The CPF s website ( however, does contain timely and useful information about various schemes, and encourages use of the Internet for interactions with the CPF by members and employers. Indonesia s organizations need to make their annual reports both more timely and relevant. International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

5 Coverage In all five countries, the coverage of civil servants is near universal. But as shown in Table 3, the coverage of private sector employees varies considerably. The city State of Singapore may be characterized as having high coverage. Foreign workers, accounting for nearly a quarter of the labour force, have not been permitted to be a part of the CPF since Malaysia may be characterized as having achieved moderately high coverage. Its foreign workers (with significant exceptions) have been required to contribute to the EPF since 1998, albeit at lower rates. Thus, Singapore and Malaysia have moved in opposite directions over coverage of foreign workers since the 1997 economic crisis. The Philippines and Thailand have moderate coverage. The data for Indonesia are not available, but the coverage by Jamsostek at around 10 per cent of the labour force is fairly low. The above figures suggest a fairly close correlation in various countries between the extent of the informal economy and the coverage of the labour force. The experience of Singapore and Malaysia suggests that rapid employment growth in the formal sector is essential for reasonable coverage by mandatory schemes. Globalization, however, may affect formal employment patterns in these two countries, making one-off contractual employment and flexible work patterns more widespread (Holzmann et al., 1999). If this occurs, the coverage may be adversely impacted. The data also suggest that the acceptance of the social insurance principle does not necessarily ensure satisfactory coverage. Except for Singapore, and perhaps Malaysia, the countries are unlikely to be able to rely on formal sector expansion alone, and therefore need to formulate explicit strategies for social security protection for those in the informal economy. This issue received some attention during the 1997 crisis, but the programmes implemented have not been long term and do not constitute a coherent strategy for expanding coverage (World Bank, 2000, esp. ch. 6; Ortiz, 2001). 75 Contribution rates Table 4 provides data on the contribution rates for key pension and provident fund institutions in Southeast Asia. Malaysia and Singapore do not require civil servants on pensions to contribute. In both countries, the contribution rate for others is substantially higher than the international benchmark of 10 to 15 per cent. The pre-retirement withdrawals in these countries are, however, high owing to multiple roles assigned to the provident International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

6 Table 4. Contribution rates for key pension and provident fund institutions in Southeast Asia (mid-2002) Country Rate (%) Wage ceiling Indonesia Jamsostek PT Taspen Yes No Malaysia 23.0 No Philippines SSS GSIS Yes Yes Singapore 36.0 a S$ 6,000/month Thailand SSO GPF B 15,000/month Yes Source: Information obtained from official sources in individual countries. a Applies to those below 55 years of age; lower rates apply to those above 55 years. 76 funds, 1 reducing the resources available for retirement. In Singapore, such withdrawals averaged 75.2 per cent of the contributions during ; while the corresponding figure for Malaysia in the same period was 42.1 per cent. In contrast, the contribution rates for private sector workers are relatively low in Indonesia, the Philippines, and Thailand. The contribution rate for government employees is also quite low in Indonesia and Thailand, implying a substantial fiscal burden. Administration costs Administration costs refer to the costs of provident or pension fund organizations, while the compliance costs refer to the costs to employers and the employees of complying with various relevant requirements and regulations. Both these components are important, as improved efficiencies (or malpractices) in either directly enhance the resources available for retirement (James, 2000). The data on compliance costs are not available for the Southeast Asian 1. The CPF acts as the main mortgage financing mechanism in Singapore and as an instrument for financing healthcare. Malaysia also has provisions for withdrawals for housing and healthcare, but they are not as extensive as in Singapore. Both countries also permit approved investments to be made from the provident fund balances (CPF Board, 2000; EPF Board, 2000). International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

7 Table 5. Administrative costs for selected provident and pension funds in Southeast Asia Country Year Administrative costs/ contributions (%) Administrative costs/ assets (%) Malaysia (EPF) Philippines (SSS) Singapore (CPF) Thailand (SSO) Source: Information obtained from official sources in individual countries. countries. Table 5 provides a static view of the administration costs of selected organizations in Southeast Asia. The data indicate that the administration costs in Singapore and Malaysia are fairly low, and are comparable to international benchmarks. The costs are moderately high for Thailand, and inordinately high for the Philippines. In Indonesia, administration costs for PT Taspen and PT Astek (now Jamsostek) as a proportion of contributions were 7.0 per cent and 11.7 per cent respectively in 1994 (Leechor, 1996). More recent data, however, are not available. Without deeper analysis, it is difficult to explain the importance of various factors, such as the contribution rates, wage levels, coverage, asset size, productivity of human resources, information technology policies, and others (James, 2000). This is an underresearched area in Southeast Asia. While low contribution rates are a factor in high costs in the Philippines and Thailand (and Indonesia as well), their costs are too high to be accounted for by this factor alone. Anecdotal evidence suggests non-trivial compliance costs in these three countries. This is particularly the case for members availing themselves of various benefit schemes. In some countries, non-trivial costs in completing withdrawal of accumulated balances in provident funds at the time of retirement put further strains on retirement finances. The above discussion suggests that improving administration and compliance efficiency is a potential source of enhanced financial security in old age in at least Indonesia, the Philippines, and Thailand. 77 Investment policies and performance In the Southeast Asian countries, the social insurance schemes are demographically immature, resulting in accumulation of balances. In the mandatory savings provident fund schemes, accumulation of balances during the working years of the individual is drawn down in the retirement years. International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

8 Table 6. Total assets of provident/pension funds in Southeast Asia Country Year Total assets % of GDP Indonesia 1999 Jamsostek Rp 8.9 trillion (US$ 1.1 billion) 0.9 PT Taspen Rp 9.6 trillion (US$ 1.2 billion) 1.0 Employer pension funds a Rp 20.5 trillion (US$ 2.6 billion) 2.0 Financial institutions funds Rp 0.6 trillion (US$ 0.1 billion) 0.1 Total assets Rp 39.6 trillion (US$ 5 billion) 4.0 Malaysia 2001 M$ billion (US$ 50.4 billion) 57.3 Philippines 2000 SSS P billion (US$ 3.6 billion) 4.5 GSIS P billion (US$ 4.0 billion) 5.0 Total assets P billion (US$ 7.6 billion) 9.5 Singapore 2001 S$ 92.2 billion (US$50.7 billion) 60.1 Thailand SSO GPFT B billion (US$ 3.0 billion) B billion (US$ 3.6 billion) Total assets b B billion (US$ 11.9 billion) 9.3 Source: Information obtained from official sources in individual countries. a Numbering around 350. b Includes assets of provident funds of state-owned enterprises, and others. Since the wage level and employment history determine contributions, and since there are significant wage inequalities in these countries (Rao, 1999), the accumulated balances are also likely to exhibit considerable inequality, especially between male and female members. Gender issues are integral to any defined contribution scheme. Thus, while social insurance schemes by definition provide lifetime annuities, and therefore cover at least the longevity risk (i.e. that a person may live much longer than expected after retirement), no such automatic protection exists in the provident fund schemes. Therefore, in such schemes, the financial products such as annuities must be accessible and affordable. The focus therefore should be on both the accumulation and the decumulation phases in the provident fund schemes. The latter phase has received scant attention in Southeast Asia. The ratio of total provident fund and pension fund assets to GDP ranges from a low of 4.0 per cent in Indonesia to a high of 60.1 per cent in Singapore. This ratio is 57.3 per cent in Malaysia, 9.5 per cent in the Philippines and 9.3 per cent in Thailand (Table 6). International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

9 International experience and econometric research have found little evidence that mandatory retirement savings schemes necessarily and directly increase national savings (Barr, 2000; Orszag and Stiglitz, 2001). The mandatory savings, however, do alter the composition of savings, particularly towards longer-term savings. The beneficial effects could arise if such contractual savings are translated into economically productive assets through financial and capital markets, which in turn could help increase the trend rate of economic growth (Barr, 2000; Ribe, 1994). Ultimately, it is the economic growth over time which determines resources available in any given year, which in turn can be divided among the elderly population and the rest. It is in this sense that pension reforms, financial and capital market reforms, and corporate governance are linked. 2 The investment policies and performance of the provident and pension funds vary considerably among the five countries as evidenced from brief country descriptions below. Malaysia. The EPF began investment diversification away from government securities only at the beginning of the 1990s. In 1991, 73.6 per cent of the total balances of M$ 52.0 billion were invested in high-yielding government securities, 13.3 per cent in money markets, 11.0 per cent in debentures and loans, 2.1 per cent in equities and 1 per cent in property and other assets. However, by 2001, the respective proportions were 37.0, 19.0, 21.6, 21.7, and 0.7 (data supplied by the EPF Board). Thus, the share of government securities has declined substantially, while the share of equities has increased significantly, during the 1990s. The EPF s equity investments alone account for about 5 per cent of total market capitalization of the Kuala Lumpur Stock Exchange. This provides potential leverage to the EPF to help improve corporate governance, but also considerable responsibility to use its market power judiciously. The EPF Board thus has been able to diversify its investment portfolio considerably. However, all of the investments are domestic. Financial industry sources in Malaysia estimate that if current trends continue, the EPF assets will be around M$ 1,000 billion by Domestic financial and capital markets are likely to remain too limited to absorb such large amounts productively. Gradual steps towards international diversification are therefore likely to be necessary. This will require a different type of investment management skills. Moreover, as domestic markets become more sophisticated e.g., derivative instruments may become more widespread For detailed discussion of the bidirectional relationship between pension fund savings and capital markets, see Charlton and McKinnon, 2001, pp International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

10 there will be a need to enhance the skills of the existing investment staff. Malaysia s EPF has, over a 40-year period (1961 to 2001), provided an annual real rate of return of 3.34 per cent. 3 Since these returns are net of investment management fees, they may be considered satisfactory. 4 The EPF permits lump sum withdrawal of accumulated balances. Thus, individuals need to plan the rate of decumulation of the balances on their own. As annuity markets are not well developed in Malaysia, the resulting high costs act as a further deterrent. It was due to these factors that the EPF s efforts at encouraging members to purchase annuity have met with extremely limited success. The EPF members are therefore exposed to inflation and longevity risks. 80 Singapore. The investment policies and performance of the CPF are quite complex, with outcomes which are considerably at variance with what may be expected from a mandatory savings provident fund system. Under the CPF Act, all balances with the CPF Board must be invested in Singapore government securities. These securities are non-marketable, and interest on them is identical to what the Board pays to its members. The Board s interest on the ordinary account of the members is a weighted average of one-year fixed deposit and saving accounts in three domestically owned banks, but with a minimum guarantee of 2.5 per cent. 5 Most of the contributions to the CPF (ranging from 61.1 to 72.2 per cent) are in this account. So its interest rate dominates the return on balances. The other two accounts, Medisave Account (between 16.7 and 22.2 per cent of the contributions are channelled to this account) and Special Account, are credited a 1.5 percentage point higher interest rate than is credited in the ordinary account (CPF Board, 2000). The housing mortgage rate is 0.1 per cent higher than the interest on the ordinary account. This is one of the ways in which the CPF acts as a substitute for the mortgage finance market. The above suggests that the interest rate paid to members is administratively determined. During , the average annual real interest rate (nominal rate less inflation rate as measured by the GDP deflator) on CPF balances (amounting to S$ 92.2 billion in 2001) was 1.83 per cent; while for 3. Calculations by the author based on official data supplied by the EPF. The calculations are available on request. 4. If the shorter period, including the 1997 crisis, is taken, the annual real return was actually higher, at 4.20 per cent for the period. 5. Assuming a long-term average inflation rate of 3 per cent, the 2.5 per cent nominal rate guarantee implies a negative real rate of return. Such a guarantee therefore has minimal costs. For a discussion of various types of provident and pension fund guarantees and their implications, see Gillion et al, 2000, pp International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

11 the period it was only 0.82 per cent. 6 Such low rates essentially negate the power of compound interest made possible by mandatory savings. It is essential to analyse the use of the CPF funds borrowed by the government (Asher, 2002). As the government has been generating persistent budget surpluses, the borrowed funds are not needed to finance government expenditure. These funds are turned over to the Singapore Government Investment Corporation (SGIC) and other investment companies of the government. They are registered as private companies, and statutory provisions ensure that they are not subjected to scrutiny by the parliamentary, public, or any other authority. It is believed that nearly all the CPF balances are invested abroad by these companies. Thus, international diversification is the norm in investing CPF balances. The above arrangement thus does not result in transparency or accountability in the ultimate investment of CPF funds. To the extent SGIC and other agencies obtain higher returns than what is paid to CPF members, and indications are that this is indeed the case, there is an implicit tax on CPF wealth which is large, recurrent, and regressive. The above summary of investment policies and practices has transformed the nominal defined contribution (DC), fully funded (FF) nature of the CPF scheme to a notional defined benefit (NDB) scheme financed on a pay-as-you-go (PAYG) basis. It is NDB because interest paid on balances bears no resemblance to investment returns. It is PAYG because the government bonds will be serviced by the future taxpayers, whose consumption will decline to finance that of the elderly population. This is similar to the PAYG mechanism. The above discussion suggests that the substance of a social security system cannot necessarily be inferred from the formal nature of the system. Details of design and implementation do matter. Unlike Malaysia and Singapore, time series information on investment performance is not available for the other three countries. 81 Thailand. All provident and pension fund assets are currently required to be invested domestically. But the GPF is in the process of seeking approval to invest abroad. The limitations of domestic financial and capital markets will increasingly become relevant as funds accumulate. 6. There are also insurance funds which are invested in a diversified manner by the CPF Board. These funds amounted to $3.2 billion in The annual rate of return on them was 3.24 per cent for the period and 2.74 percent for the period. These returns are not credited to the members account. Details of calculations are available on request from the author. International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

12 In the year 2000, the nominal rate of return for the GPF was 6.41 per cent, while the 1999 return for the SSO was 5.43 per cent. As Thailand has had low inflation rates in recent years (1.6 per cent in 2000), the real rate of return may be considered satisfactory. Whether such returns will be obtained over a longer period remains to be seen. The asset allocation of the GPF as of June 2001 consisted of government or government-guaranteed bonds (nearly two-fifths of the total), bank deposits (28.3 per cent), but only about 8.0 per cent in stock and unit trusts ( The state enterprise provident funds have also invested conservatively. Thus, in 1999, nearly a third of their balances were in bank deposits, while less than one-sixth of the funds were in stock and unit trusts (Jirapon, 2000). Most of the SSO s assets are also invested in bank deposits. 82 Philippines. The average rate of return on investments by the SSS was 8.4 per cent in 2000, less than the Treasury bill rate of 10 per cent, and only marginally higher than the inflation rate of 8.4 per cent (Philippines Social Security System, 2000). Indeed, in recent years, the expenditure on benefits has exceeded the contributions plus operating expenses of the SSS. Its dependence on investment income will, however, become unsustainable as the system matures. Data for GSIS are not available. Indonesia. No reliable data are available on investment policies and performance of various provident and pension funds in Indonesia. Before the 1997 crisis, only the employer pension funds were believed to be providing satisfactory returns, and even for them comprehensive data were unavailable (Leechor, 1996; Purwoko, 1996). The situation is unlikely to have altered since then. Most of the provident and pension fund assets of PT Taspen and PT Jamsostek are believed to be in bank deposits, rather than invested on the financial and capital markets. Low returns and high administration costs pose severe challenges to public social security institutions in Indonesia. Financial adequacy during retirement It is conventional to evaluate financial adequacy by estimating the replacement rate at the time of retirement, and the rate during retirement. The latter depends on the extent of indexation. In general, wage indexation is preferable to price indexation for financial adequacy but requires higher contribution (or tax) rates. Replacement rate usually refers to the proportion of the last drawn (or suitably averaged) salary which is available to the retiree as a monthly income. Experts recommend around two-thirds of the last International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

13 drawn salary (maintained in real terms) for middle-income individuals, and higher for low-income workers. It is important to recognize that these replacement rates refer to the combined sources or tiers from which retirement income is derived. In general, attempting to obtain (or provide) the requisite replacement rate from just one source or tier is likely to be an extremely risky strategy. In addition, survivors benefits are an important element of financial adequacy. In the five Southeast Asian countries, the civil servants and polititians covered by the government pensions have generally been able to obtain satisfactory replacement rates (between 50 and 100 per cent). They also enjoy some protection against longevity risk. This protection varies. Ad hoc changes in pension benefits are the norm in Malaysia, the Philippines, and Singapore. In Thailand, pension benefits are adjusted concomitantly with civil service salary changes. In Indonesia, Malaysia, the Philippines, and Thailand, retired government pensioners receive medical care at no cost or at highly subsidized rates. In contrast, except for a small proportion in occupational pension schemes of large companies, the replacement rates for private sector workers are inadequate (ranging from 10 to 40 per cent). Characteristically, none of the provident and pension funds, however, provide estimates of the expected replacement rates, a gap which they need to fill urgently. Moreover, none of the provident funds require mandatory annuity to be purchased from the accumulated balances. Singapore has instituted a minimum sum scheme under which S$ 80,000 will have to be set aside by July 2003, with the amount equally divided between cash and pledging of property. Annuity purchase is only one of the three options for phased withdrawals of the cash component of minimum sum, the other two being fixed deposit with an approved bank, or leaving it with the CPF Board. In spite of annuity approved by the CPF Board providing satisfactory returns, they are not popular, however (CPF Board, 2000). This suggests a need for improving financial literacy of the members. Many members may not even have the minimum balances, and therefore the role of the minimum sum scheme, while in the right direction, is unlikely to be major. The other countries do not even have this type of scheme. The above suggests that inflation and longevity risks along with the absence of survivors benefits constitute major limitations of the social security schemes for private sector workers in Indonesia, Malaysia, and Singapore. There is an absence of any tax-financed redistributive first tier in these countries; and only a beginning of schemes which may be characterized as belonging to the tax-advantaged voluntary savings third tier dedicated to retirement. The private sector workers, even those who are cov- 83 International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

14 ered, therefore face considerable financial insecurity in old age. For those not covered by the formal schemes, the financial insecurity is likely to be even greater. Reform directions The discussion in the previous sections strongly suggests the need for reforms in each of the countries involving five core functions of provident and pension fund organizations; and a better balance among various elements of the social security system. The precise strategy, tactics, sequencing, timing, and pace of reforms in each country will depend on the political economy, initial conditions, institutions, regulatory capabilities, and policy objectives. The need for reforms is particularly urgent in the following areas. Governance and regulatory structure 84 The provident and pension fund organizations in all five countries need to give increased weight to fiduciary responsibility, transparency, accountability, and information disclosure. The annual reports of these organizations, even when they are available (and their websites), provide neither timely nor detailed data comparable to international norms for such organizations. In particular, data on investments and on the financial position of members requires significant improvement. The civil service pension schemes in these countries are operated with little disclosure. Independence of the provident and pension fund organizations has not been established in practice. And there is no social security authority which takes a system-wide view, and which enforces international norms of pension fund governance. Administration and compliance The discussion in the previous section has provided strong evidence that except for Malaysia and Singapore, the countries need to give high priority to record keeping, collection, and payment of benefits without undue delay or side payments. Providing a unique identification number to each member (and not just to the employer) has often proved useful in realizing the benefits of individual accounts, minimizing multiple accounts owned by the same individuals, and ensuring portability of benefits. Empowering and educating members, including in the financial aspects of retirement planning, is another aspect of administration which requires attention. International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

15 Improving efficiency of saving investment process The requirement is that there should be effective investment diversification; including, where appropriate (as in Malaysia and Thailand), steps towards international diversification; a shift away from the administered to market interest rates; and development of financial and capital markets to ensure increased supply of investible assets. It should, however, be noted that inflation-indexed bonds, advocated by some to develop the annuity markets, do not provide gains to the society as a whole, but are mere risk-shifting, redistributive instruments. Countries with weak fiscal positions, such as Indonesia and the Philippines, should particularly avoid use of such instruments. The extent of individual choice in the investment process depends on several factors (Heller, 1998; Charlton and McKinnon, 2001; Harris and Jones, 1999). In the context of Southeast Asian countries, a decentralized investment regime of the type adopted by Hong Kong s Mandatory Provident Fund (MPF) or by Chile is unlikely to be appropriate. This is because of the low level of financial literacy on the part of the members; high transaction costs due to limited development of financial and capital markets; and relatively underdeveloped regulatory and legal regimes. Some of the large provident and pension funds in the region (such as EPF, CPF and GPF) may consider using their expertise and financial power to give a choice of a limited number of portfolios (including a default option) to their members to allocate balances among them. The investment mandates may be given to different fund managers, as well as undertaking some of the functions within each organization. This may provide information and track record on investment performance, minimize transaction costs, and ensure that returns are credited to the members accounts. 85 Others As the social security programmes for those in the informal economy are usually financed from the budget, fiscal reforms, including more effective public expenditure management, will be essential. High-income countries such as Singapore should consider introducing a tax-financed redistributive first tier. Immediate prospects for the introduction of such a tier, however, are not encouraging. Indonesia and Malaysia need to urgently consider whether their current system of financing civil service pensions is fiscally and politically sustainable. Indonesia may consider how the share of its contributory component administered by PT Taspen can be enlarged to reduce the fiscal burden. International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

16 Concluding remarks 86 This paper has argued for a need to enhance professionalism in the performance of the five core functions of provident and pension fund organizations; and for taking a system-wide perspective in developing social security systems in the Southeast Asian countries. The analysis suggests that the current practice, essentially centring on self-regulation by each provident or pension fund organization, has outlived its usefulness. The Southeast Asian countries need to become more receptive to lessons drawn from pension reform experienced in other regions, particularly Latin America (Asher, 2001). More open and participatory public debate is needed about the future of social security systems, particularly in Indonesia, Malaysia, and Singapore. Each of the five countries covered in this paper is therefore strongly urged to consider constituting a Provident and Pension Fund Authority (PPFA) with a mandate to supervise and regulate all concerned organizations, including civil service pension schemes. The main objectives of the PPFA would be to make the social security systems in these countries sustainable by enhancing professionalism and ensuring system-wide development. Social security reform, therefore, is best regarded as a process with a medium time frame. Effective government will remain critical for the success of this process in each country. The longer the initiation of the social security reform process is delayed in these countries, the more challenging the task will become. Bibliography Asher, M. G The future of retirement protection in Southeast Asia, in International Social Security Review, Vol. 51, No Reforming civil service pensions in selected Asian countries (paper presented at the World Bank Social Security Workshop). Washington, DC, World Bank Social security systems in Southeast Asia: Are they sustainable?, in D. E. Andersson and J. P. H. Poon (eds.), Asia-Pacific transitions. New York, Palgrave Social security institutions in Southeast Asia after the crisis, in M. Beeson (ed.), Reconfiguring East Asia. London, RoutledgeCurzon. Barr, N Reforming pensions: Myths, truths, and policy choices (WP/00/139). Washington, DC, International Monetary Fund, Fiscal Affairs Department. Bos, E. et. al World population projections ( edition). Baltimore, MD, Johns Hopkins University Press. International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002

17 Charlton, R.; McKinnon, R Pensions in development. Aldershot, Ashgate. CPF Board (various years). Annual Report. Singapore, Central Provident Fund. EPF Board (various years). Annual Report. Kuala Lumpur, Employees Provident Fund. Gillion, C.; Turner, J.; Bailey, C.; Latulippe, D Social security pensions: Development and reform. Geneva: International Labour Office. Harris, D.; Jones, S. P Pensions and the consumer: Lessons from overseas, in P. Cartwright (ed.), Consumer protection in financial services. London, Kluwer Law International. Heller, P. S Ageing in the Asian tigers : Challenges for fiscal policy (IMF Working Paper WP/97/143). Washington, DC, International Monetary Fund Rethinking public pension reform initiatives (IMF Working Paper WP/98/61). Washington, DC, International Monetary Fund, Fiscal Affairs Department. Holzmann, R. et al Pension systems in East Asia and the Pacific: Challenges and opportunities (24 September). Washington, DC, World Bank, Social Protection Unit. James, E Administrative costs and the organization of individual account systems: A comparative perspective (paper presented at the Third APEC Regional Forum on Pension Fund Reform, March). Bangkok. Jirapon, K Social security system in Thailand: Past experiences, obstacles and ways to reform (paper presented at the Third APEC Regional Forum on Pension Fund Reform, March). Bangkok. Leechor, C Reforming Indonesia s pension system (Policy Research Working Paper No. 1677). Washington, DC, World Bank. Orszag, P. R.; Stiglitz, J. E Rethinking pension reform: Ten myths about social security systems, in R. Holzmann and Joseph E. Stiglitz (eds.), New ideas about old age security. Washington, DC, World Bank (including comments by R. Holzmann, E. James, A. Borsch-Supan, P. Diamond, and S. Valdés-Prieto). Ortiz, I Conclusion: Social protection in Asia and the Pacific, in I. Ortiz (ed.), Social protection in Asia and the Pacific (chapter 17). Manila, Asian Development Bank. Philippines Social Security System (various years). Annual report. Manila. Purwoko, B Indonesian social security in transition: An empirical analysis, in International Social Security Review, Vol. 49, No. 1. Rao, V. V. B East Asian economies: Trends in poverty and income inequality, in Economic and Political Weekly, Vol. 34, No. 18. Ribe, F Funded social security systems: A review of issues in four East Asian countries, in Revista de Análisis Económico, Vol. 9, No. 1. Ross, S. G Building pension institutions: Administrative issues (proceedings of the Third APEC Regional Forum on Pension Fund Reform). Bangkok. 87 International Social Security Association, 2002 International Social Security Review, Vol. 55, 4/2002 Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 01248, USA

18 Thailand Office of Social Security (various years). Annual Report. Bangkok, Ministry of Labour and Social Welfare. World Bank Averting the old age crisis: Policies to protect the old and promote growth. New York, Oxford University Press East Asia: Recovery and beyond. Washington, DC World Development Report New York, Oxford University Press. 88 International Social Security Review, Vol. 55, 4/2002 International Social Security Association, 2002