Journal Issue
Share
Article

Turkey: Selected Issues and Statistical Appendix

Author(s):
International Monetary Fund
Published Date:
February 2000
Share
  • ShareShare
Show Summary Details

II. The 1999 Social Security Reform20

A. Introduction

34. There has been a broad consensus in recent years that Turkey’s pay-as-you-go social security system is unsustainable. The generosity of benefits relative to contributions has resulted in the system having a large deficit (equivalent to around 3 percent of GNP), notwithstanding the country’s favorable demographic situation. In the absence of reform, the deficit of the social security system was set to increase sharply in the next few years, further widening the already large public sector deficit.

35. The landmark social security reform legislation approved by the Turkish government on September 8, 1999 is designed to address this issue. Simulations done by both the authorities and the World Bank show that the reforms will be able to stop the trend deterioration and indeed reduce the deficit of the social security in the short- to medium term, although the long-term impact of the reforms as yet remains unclear. Further reform is, nonetheless, likely to be required to eliminate the deficit in the social security system in the long run. This chapter outlines the key elements of the recent reforms, and discusses the additional measures that the authorities are contemplating to reform the system further.

B. The Old Social Security System21

36. Pension benefits for old age, disability and survivorship as well as health and other social benefits are provided in Turkey through three separate funds: the Sosyal Sigortalar Kurumu (SSK), covering private workers and non-civil servant public sector workers; the Emekli Sadigl (ES) covering civil servants; and Bag Kur (BK) covering the self-employed, permanent agricultural workers, elected local government officials, housewives and the unemployed. Together, those insured for pension benefits for old age, disability and survivorship by the three funds numbered 12.4 million people in 1998 (close to half the total employed population), and benefits were paid to a further 5.6 million people.

37. The balance of the social security system has deteriorated significantly in recent years as shown in Table 1. This is primarily due to a number of structural flaws in the system, including:

Table 1.Turkey: Overview of Major Pension Indicators, 1992–99
19921993199419951996199719981999 1/
SSK
Number of contributors (in millions)4,2744,5925,1885,6455,9246,1656,8767,120
Percentage growth7.413.08.84.94.111.53.5
Number of pensioners1,8571,9992,1752,3382,5402,7322,9313,157
Percentage growth7.68.87.58.67.67.37.7
Contributors/pensioner2.302.302.392.412.332.262.352.26
Average annual pension (in millions of TL)15244075143301519871
in US$2,1212,1951,3321,6411,7661,9831,9942,079
Percentage growth3.5-39.323.27.612.30.64.3
Average annual contribution (in millions of TL)710131947104182314
in US$953883424420584685700749
Percentage growth-7.4-52.0-1.039.217.32.37.0
Pension/contribution (avg.)2.22.53.13.93.02.92.82.8
Balance of SSK (percent of GNP)-0.2-0.4-0.8-1.1-1.0-1.3-1.0-1.3
Emekli Sandigl
Number of contributors1,8501,8961,8961,8801,9642,0362,0722,109
Percentage growth2.50.0-0.84.53.71.81.8
Number of pensioners7818289019521,0481,1141,2191,273
Percentage growth6.08.85.710.16.39.44.4
Contributors/pensioner2.372.292.101.971.871.831.701.66
Average annual pension (in millions of TL)19314989177333554921
in US$2,7382,8241,6361,9402,1762,1962,1292,197
Percentage growth3.1-42.118.612.20.9-3.03.2
Average annual contribution (in millions of TL)814234594182326556
in US$1,1561,2337789831,1611,2011,2531,327
Percentage growth6.7-36.926.418.23.44.35.9
Pension/contribution (avg.)2.42.32.12.01.91.81.71.7
Balance of ES (in percent of GNP)-0.6-0.7-0.6-0.7-1.1-0.9-1.0-1.2
Bag-Kur
Number of contributors2,8912,8712,7002,7692,6502,6572,9083,147
Percentage growth-0.7-6.02.6-4.30.39.4S.2
Number of pensioners5896466877398039571,1041,189
Percentage growth9.76.37.68.719.215.47.7
Contributors/pensioner4.914.443.933.753.302.782.632.65
Average annual pension (in millions of TL)713282985180296494
in US$1,0701,1489346351,0421,1841,1351,180
Percentage growth7.3-18.6-32.064.113.6-4.14.0
Average annual contribution (in millions of TL)1247154469109
in US$196151120153180290266260
Percentage growth-22.9-20.227.517.560.9-8.5-2.0
Pension/contribution (avg.)5.57.67.84.15.84.14.34.5
Balance of BK (in percent of GNP)-0.1-0.2-0.3-0.3-0.5-0.5-0.9-0.8
Total of pension funds
Number of contributors9,0159,3599,78410,29410,53810,85811,85612,376
Percentage growth3.84.55.22.43.09.24.4
Number of pensioners3,2273,4733,7634,0294,3914,8035,2545,619
Percentage growth7.68.47.19.09.49.46.9
Contributors/pensioner2.82.72.62.62.42.32.32.2
Weighted average annual pension (in millions of TL)14244070140284480803
in US$2,0792,1501,3321,5281,7321,8731,8451,916
Percentage growth3.4-38.014.713.48.2-1.53.8
Weighted average annual contribution (in millions of TL)5E122148104130303
in US$752729409451590685690723
Percentage growth-3.0-44.010.330.916.10.74.8
Pension/contribution (avg.)2.82.93.33.42.92.72.72.6
Overall balance of three funds (percent of GNP)-0.9-1.3-1.7-2.1-2.6-27-2.9-3.3
Sources: Undersecretariat of the Treasury, Stat Planning Office; and Fund staff calculations.

Projections.

Sources: Undersecretariat of the Treasury, Stat Planning Office; and Fund staff calculations.

Projections.

(i) No minimum retirement age. Rather, pension eligibility in all three funds was determined by the number of years worked (20/25, women/men respectively) and period of contribution (see Table 2). For BK and ES, the number of years worked needs to be equal to the period of contribution. As long as they had worked for 20/25 years, SSK participants could, however, contribute for just a 14-year period. This meant that people who had started working and contributing to the system at the age of 18 (the minimum employment age) could retire as early as at 38/43 years of age in all three funds. And indeed the average age for new retirees was 47/50 for SSK and ES and 57 years for BK.22 This contrasts with a life expectancy at birth of 71/67 and at age 60 of 79/77 (World Bank, 1999).23

Table 2.Turkey: Changes to the Basic Features of the Social Security System under the September 1999 Reform
Contribution RateEligibility for Retirement 1/Reference Period for BenefitsBenefits Determination for Regular Pension
FundCoverageOldNewOldNewOldNewOldNew
Sosyal Sigortalar Kurumu (SSK)Private sector employees and public sector workers except civil servants33.5 percent

20 percent for pensions; 12 percent for healthcare, and 1.5 percent for other benefits 2/
Unchangedpension 5.000 days (13.9 years)



Partial Pension 3,600 days (10 years) if age

55 – men;

50 – women
Age:

58 Female/60 Male and



Regular Pension 7000 days of (19.4 years) Partial Pension 12.5 years
Last 5 years for those paying at less than the contribution ceiling; last 10 years for those paying at the contribution ceiling 2/Full contribution period.60% for 5,000 days plus 1% for each additional 240 days 1/3.5 percent accrual rate for first 10 years; 2 percent for next 15 years; 1.5 percent fox each additional year
Emekli Sandiǧl (ES)Civil servants35 percent

combined for pensions and health care
Unchanged25 years – men

20 years – women
58 Female/60 Male and



25 years
Last month’s wage; last six months wage if the worker has recently been promotedUnchanged.75% plus 1% for each additional yearUnchanged
Baǧ-Kur (BK)Self-employed urban workers and farmers32 percent

20 percent pensions; 12 percent healthcare
35 Percent

20 percent pensions; 15 percent healthcare
Regular pension

9,000 days (25 years)-men 7,200 days (20years)-women



Alternative

15 years if age 55-men, 50-women
Age:

58 Female/60 Male and



Regular Pension

9000 days of (25 years)



Partial Pension

15 years
Last 1 yearFull contribution period70% plus 1% for each additional year3.5 percent accrual rate for first 10 years; 2 percent for next 15 years; 1.5 percent for each additional year

For new entrants. Days (years) refer to the minimum contribution period.

The wage on which contributions are based was previously capped at 1.8 times the minimum pensionable wage. Following the reform, the maximum contribution ceiling has been raised to a level close to 3 times the statutory minimum wage.

For new entrants. Days (years) refer to the minimum contribution period.

The wage on which contributions are based was previously capped at 1.8 times the minimum pensionable wage. Following the reform, the maximum contribution ceiling has been raised to a level close to 3 times the statutory minimum wage.

(ii) A high replacement ratio. The contribution rates in these schemes are high at 30 to 35 percent of the pensionable wage (including pension and health insurance contributions), with costs being shared between the employee and the employer. However, the absolute level of pension was not particularly high because the pensionable wage is well below the actual wage (see below). Moreover, because of the limited time over which contributions were required to be made, benefits were extremely generous relative to contributions, with an average replacement ratio of the order of 80 percent for the three funds. Contributors were able to reach a relatively high replacement ratio in the initial years of contribution. In the case of SSK, for example, for each of the first 14 years of contribution, the average replacement ratio was 4.3 percent, compared to just 1.5 percent for each additional year of contribution. This sharp decline in the marginal benefit of continuing to contribute to the system provided an incentive for people to retire as soon as they fulfilled the minimum contribution period.

(iii) A short reference period for the payment of pensions. An additional source of weakness has been the short reference period for determining the retirement pension. In SSK, the reference period for pensions was the 5 most recent years for those making contributions at the lower pension scale and the most recent 10 years for those making contribution at the upper scale. In the case of ES, pensioners were eligible to receive 75 percent of their actual wage on retirement (or the salary over the last six months if the civil servant was recently promoted). For BK, the reference period for the pension payment is the year preceding retirement.

38. A number of other weaknesses were evident in the social security system. Supplementary social assistance payments made by the funds were (and still are) untargeted, leading to wasteful expenditure. There are also significant compliance problems in SSK and BK, in large part due to administrative problems. In particular, the absence of a unique contributor identification number, makes the task of monitoring workers’ records difficult as they change jobs. There is also very little coordination of social security contributions and tax payments.

39. The implication of these problems can be seen in the indicators outlined in Table 1. The ratios of contributors to pensioners (dependency ratio) for the three funds have declined from 2.8 in 1992 to a projected 2.2 this year. The decline in the dependency ratio of ES and BK is particularly sharp: from 2.4 to 1.7 and from 4.9 to 2.6, respectively, between 1992 and 1999.24 The policies outlined above have also meant that the average level of contributions in ES and BK has not kept pace with average pension payments.

40. Financially, this has led to a dramatic increase in the deficit of the social security system in recent years. Starting from a position of approximate balance in the late 1980s, the three funds registered a total deficit of 2.9 percent of GNP in 1998. Moreover, the deficit is projected to worsen to 3.3 percent of GNP this year.

C. The September 1999 Reforms

41. On September 8, 1999 a new law governing social security was enacted. The law included a number of major changes, including:

  • the introduction of a minimum retirement age;

  • increases in the minimum contribution periods for pensions, and changes in the replacement ratio;

  • basing pensions on indexed lifetime earnings, rather than earnings in the years immediately before retirement;

  • increasing the retirement pension in line with changes in prices;

  • a significant increase in the ceiling on contributions;

  • the introduction of copayments for medical insurance; and

  • the introduction of unemployment insurance

The minimum retirement age

42. The absence of a minimum retirement age was one of the most glaring deficiencies of the pre-reform system. Under the new system, the minimum retirement age for new entrants to the schemes is fixed at 58 for women and 60 for men. For existing members who have contributed for less than ten years, the minimum retirement ages will be 52 and 56. For those who have been in the system for more than ten years, there will be a transition period, with the retirement age being set at the current (implied) levels for those very close to retirement, and gradually increasing as time already spent in the schemes diminishes.

Increases in contribution periods and changes in replacement ratios

43. The minimum retirement age will be supplemented by a minimum contribution period to establish entitlement to a full pension. For new entrants, this will be set at 7000 days (19.4 years) for SSK and at 9000 days (25 years) for ES and BK. As in the case of retirement ages, there will be a gradual increase in the minimum contribution period for those already in the system. The amount of the pension received will be graduated according to length of contribution, with the replacement rate (the annual pension as a proportion of average lifetime earnings) set, in the case of SSK, at 3.5 percent for each 360 days for the first 10 years, 2 percent for the next 15 years and 1.5 percent for each subsequent year (for those who carry on contributing after the minimum period). This system has also been supplemented by explicit provision for a partial old age pension for those who have reached the minimum retirement age but have not made sufficient contributions to be eligible for a full pension, though 12½ years of contributions (10 years in the case of current members of SSK) will be necessary to get even this.

Basing pensions on lifetime earnings

44. One of the features of the pre-reform system which provided the greatest scope for abuse was the fact that pensions were based on earnings in the last year or few years before retirement rather than on lifetime earnings, which had been the basis for contributions. Since employers and employees both contribute to the schemes, there had been incentives for them to collude in underreporting wages until close to the end of the contribution period, and then raise reported wages so that the pension benefit would be maximized.

45. Under the new system, the reference period for pensions in SSK and BK will be lifetime earnings, with earnings in earlier years being indexed to take into account inflation and real growth. Again there will be a transitional period during which the old system will be used for part of the calculation of pensions (with the calculation based on wages at the time the system changed) and the new system for the remaining part, with the relative weights of the two depending on the time the employee is a member under each system. For ES, the old system will continue to operate.

Indexation of pensions to the consumer price index

46. Another important measure was the indexation of pensions to the consumer price index. In the past, the system of indexation implied that pensions were increased on a discretionary basis, typically in line with civil servants’ wages. This led to large swings in the level of pensions, with pensions declining in periods of fiscal austerity and increasing as the fiscal stance was loosened. Under the new arrangements, pension increases will be made on the basis of a transparent and financially conservative rule. Specifically, pensions will be increased monthly on the basis of the previous months inflation for SSK and BK, with a similar indexation mechanism envisaged for ES after necessary amendments to the civil servants employment laws have been made.

An increase in the ceiling on contributions

47. Despite the fact that the replacement rate for pensions in Turkey was high under the old system and remains high under the new one (by way of comparison, the average rate in OECD countries is 45 percent) actual pensions received have been relatively low. The main reason for this is that the wage which formed the basis for calculating both contributions and pensions is capped at a level well below the actual wages of many participants. Under the new system, the lowest wage on which contributions and pensions will be based will be set at a level close to the statutory minimum wage for workers, and the highest contribution level will be set three times higher than this.25 The Council of Ministers also has authority to raise the ceiling further, to five times the lower limit. The result should be both an immediate increase in contributions and, over the longer-term, a higher level of pensions for those who contribute more.

Medical insurance reforms

48. Health insurance is also covered by the major social insurance schemes, and part of total contributions are for health insurance.26 While health contributions have been broadly in line with spending on health in SSK, in BK a large deficit has emerged. To reduce this, in BK the health insurance contribution rates were increased from 12 percent (for the self employed) and 10 percent (for farmers) to 20 percent of the pensionable wage. In addition, in all three schemes, a system of copayments for medical equipment and prostheses was introduced for the first time, with copayments being set at 20 percent of the cost for working members of the schemes and 10 percent for pensioners. Finally, a waiting period was introduced in SSK and BK before members become eligible for medical benefits.

Unemployment insurance

49. The new legislation provides for the introduction of unemployment insurance in SSK, the largest scheme. There will be mandatory additional contributions of 7 percent of the pensionable wage (3 percent being paid by the employer and 2 percent each by the employee and the government). Those who have paid insurance premia for at least 600 days will be entitled to unemployment benefits equivalent to 50 percent of the last four months’ insurable wage for a period of six months. Those who have contributed for longer will be entitled to unemployment insurance for a longer period. The unemployment insurance scheme will begin operation in June 2000, with the mandatory contributions to it replacing contributions to a mandatory savings scheme. However, the existing system of lump-sum redundancy payments will be retained alongside the new unemployment insurance.

D. Fiscal Implications of the Social Security Reform

50. In the absence of reforms, both the World Bank’s and the authorities simulations show approximately the same deterioration in the balance of the deficit of the social security system in the coming years: from 2.6 percent of GNP in 1998, to 4.4 percent in 2005, 7.6 percent in 2020, and 14 percent in 2050 (Chart 1). The two models, however, diverge in their projections of the fiscal implication of the recently introduced reforms. In particular, the World Bank’s model shows the reforms delivering a sharper improvement in the deficit of the social security system in the short- to medium-term. However, it also suggests that in the long run, a growing deficit will emerge, which will reach some 5 percent of GNP in 2050. In contrast, the authorities’ model shows a more modest improvement in the deficit in the short- to medium term, but projects that the deficit will remain contained even in the long-run.

51. The divergence in the results is in large part explained by the different assumptions underlying the two models, including in the mortality rate (which the authorities model assumes: will remain constant, while the World Bank model projects a decline), differences in GDP projection and the assumption as to what age people begin contributing into the system. There are as well a number of inherent difficulties in modeling the impact of the reforms. For example, it is unclear if the effective barrier to early retirement in the long-run will be the retirement age or the minimum contribution period. Relatedly, it is also unclear whether the fact that the marginal replacement ratios continue to decline after the initial contribution period, albeit more gradually than before, will continue to be a disincentive for continued contribution into the system. These complications as well as the large degree of uncertainty that accompanies all such long-term projections suggests that it is best to consider the two projections as the lower and upper bound of the likely fiscal effects of the recent reforms.

E. The Future of the Social Security System

52. The social security reform enacted in September 1999 is intended to be merely the first stage in a more comprehensive overhaul of the social insurance system. Beginning in 2000, the government is planning:

  • the introduction of supplementary individual pension schemes and preparation of a regulatory framework for these;

  • measures to increase the coverage of the schemes and compliance with them;

  • integration of the social insurance functions of the three schemes under a single agency, and separation of the health and insurance functions of SSK; and

  • the introduction of full healthcare coverage by means of a system of general health insurance.

53. These are ambitious goals, and it remains to be seen whether the government can indeed move rapidly towards them. The intention to introduce supplementary individual pension schemes reflects a very different philosophy of insurance from that currently operating in Turkey. At present, the system is one where benefits are clearly defined, contributions are also established by law, and the state is responsible for making up any differences between expenditures and revenues in the schemes. Individual pension schemes generally involve defined contributions, but are “fully funded”, which is to say that the participants get back their contribution plus the accumulated return, whatever that is. There is a case to be made for both systems, but the government will need to consider how it will integrate a voluntary defined contribution system with the mandatory defined benefit scheme which provides the basic safety net in Turkey. There are also many regulatory issues to be worked out. If the three schemes are indeed to be brought under the control of a single agency, then the authorities will need to consider ways to harmonize their contributions and benefits. Finally, the details on the government’s plans on healthcare have yet to be worked out and the government is also likely to be confronted with the problem of rising demand for healthcare from those already insured.

54. These outstanding issues, together with the still unresolved questions about the long-term solvency of the current system, are likely to complicate plans for further reforms. They are also likely to keep social security and healthcare reform in a prominent place on the political agenda in Turkey for some time to come. Nevertheless, the authorities’ achievements in 1999 have been impressive. They have taken a system heading rapidly towards bankruptcy and given it a new lease of life.

Prepared by Christopher Jarvis and Abebe Aemro Selassie.

This section draws on Appendix II of the 1998 Selected Issues Papers (SM/98/196, 7/23/98) entitled Reforming Turkey’s Social Security System, prepared by David W. H. Orsmond and Jeffery R. Franks.

The higher retirement age of BK is explained by the fact that the fund was only created in 1972 as well as the fact the self-employed tend to join the system later in life.

World Bank, “Turkey-Country Economic Memorandum: Structural Reforms for Sustainable Reform, Mission Aide-Memoire (November 1999).

In BK’s case this is partly because the scheme is relatively new, so that in the early years few of its members had been in the scheme long enough to retire.

There is no longer a formal link between the contribution level and the statutory minimum wage.

In the cases of SSK and BK health insurance contributions are identified separately; in the case of ES, a single payment is made, covering both pension contributions and health insurance contributions.

Other Resources Citing This Publication