This Selected Issues paper highlights that the authorities in the Republic of Korea recognize the pension policy challenges ahead, and a first wave of reforms has already been launched. Despite the reforms to date, much remains to be done. Without further reforms, the public pension systems in Korea are not financially sustainable. This paper considers options for moving to a funded first-pillar pension system. The main results show that a sustainable, funded pension system can be achieved in Korea with reasonably modest changes to key parameters and extra financing.


This Selected Issues paper highlights that the authorities in the Republic of Korea recognize the pension policy challenges ahead, and a first wave of reforms has already been launched. Despite the reforms to date, much remains to be done. Without further reforms, the public pension systems in Korea are not financially sustainable. This paper considers options for moving to a funded first-pillar pension system. The main results show that a sustainable, funded pension system can be achieved in Korea with reasonably modest changes to key parameters and extra financing.

I. Options for Pension System Sustainability in Korea1

A. Introduction

1. Korea’s population will age very rapidly in the coming decades, putting considerable pressure on age-related expenditures, particularly pensions. This sweeping demographic change reflects the effects of rapid industrialization, urbanization, and income growth on fertility rates and life expectancy. As a result, Korea will become an “aged society” in two decades, and current pension systems will come under increasing financial strain. Health expenditures will also rise with the demographic transition, but these are of a lesser order of magnitude and will not be addressed here.

2. The authorities recognize the pension policy challenges ahead, and a first wave of reforms has already been launched. As in many other OECD countries, Korea’s defined benefit “first pillar” system is not fully funded, in large part due to overly generous benefit promises made in the past.2 In recognition of the funding problem, the replacement ratio (pension benefits as a percentage of income at retirement) has been lowered, the retirement age and contribution rates for some are being increased gradually, coverage has been expanded to achieve universality and broader funding, and a series of operational and administrative reforms have been enacted. The government reviews the performance of the public pension systems every five years—the next review is in 2003, which will likely become a major issue for the newly-elected government. This paper aims to contribute to the debate on pension reform options in the context of that review.

3. Despite the reforms to date, much remains to be done. Without further reforms, the public pension systems in Korea are not financially sustainable.3 Although pension system assets are rising, they will peak near the end of the next decade due to mounting pension expenditures, and will be fully exhausted some 15–20 years later. In terms of annual flows, the current surplus of 2 percent of GDP will switch to a deficit of some 8 percent of GDP once the pension system matures; i.e., when the initial young cohort reaches retirement age. This outlook is robust to changes in key demographic and macroeconomic projections.

4. This paper considers options for moving to a funded first-pillar pension system. Adopting a smoothing approach and using a partial equilibrium simulation model, it looks at combinations of changes to contribution and benefit parameters, and more aggressive portfolio investment that would put Korea’s public pension systems on a sustainable path within a 60-year time horizon. The parametric reform options include: (i) a reduction in the replacement ratio of as much as 30 percent (in 10 percent increments) and (ii) an increase in the contribution rate of up to 100 percent (in 33⅓ percent increments).4 An increase in the retirement age is already planned for the period in question, and is not considered as a candidate for reform in this paper. A second-pillar system based on the current “retirement allowance” system is discussed only briefly in terms of the sequencing of broader reform efforts. That system is also underfunded and is the subject of numerous reform proposals.

5. The main results show that a sustainable, funded pension system can be achieved in Korea with reasonably modest changes to key parameters and extra financing. The paper does not come to a definitive view on the optimal reform-extra financing package. Rather, the results provide different combinations of parameter reforms, including benefit reductions, contribution rate increases, and equity allocations in the investment portfolio that would achieve given levels of additional financing in terms of GDP. For example, completely eliminating any additional financing need would require lowering the replacement ratio by 30 percent (20 percent) while increasing the contribution rate by two-thirds (100 percent), or some intermediate combination. Alternatively, generating a 1 percent of GDP annual financing requirement would require lowering the replacement ratio by 30 percent (10 percent) while increasing the contribution rate by one-third (100 percent), or some intermediate combination. The results suggest that pension system sustainability can only be achieved through a combination of reforms—choosing only one type of reform will generally not suffice.

6. Pension reform will also require attention to a host of institutional issues. These include, but are not limited to: how to sequence the first- and second-pillar reforms to make them politically feasible (progress on the second-pillar reforms should arguably come first), the public-private split in first-pillar pension provision (the paper is agnostic as to what extent a reformed first-pillar system should be public or private, although some key tradeoffs of that decision are discussed), and how and where to manage and invest the build-up of the large stock of assets necessary for sustainability. In addition, more work is required to assess the macroeconomic effects of pension system reforms.

7. The remainder of the paper is organized as follows. Section B presents the demographics, a brief history of the evolution of retirement income in Korea, and the associated pension system challenges. Section C presents the model, simulation results, and sensitivity analysis, Section D briefly touches on some operational and institutional issues that would need to be addressed in the context of the types of reforms under consideration. Section E concludes.

B. Demographics, the Retirement Income System, and Looming Challenges5

8. Support for the elderly in Korea has traditionally been provided by the family. This was the norm in a largely rural society with a typical family size of over five children and relatively short life spans. In Korea’s Confucian culture, children were largely responsible for the care—broadly defined to include any retirement income—of their elders. The idea of preparing for one’s own retirement was virtually unknown. Thus, a central challenge for policymakers in the context of rapid development and urbanization has been how to move to a modern pension system.

Demographics and the Evolution of Retirement Income

9. Korea’s still-young population reflects historical factors. The average age of the population is only 33 years—the third lowest in the OECD—compared with the OECD average of 38 years. The elderly dependency ratio, defined as the population over 65 as a percentage of the working age population, is around 10 percent, one-half the OECD average However, these apparently favorable statistics reflect relatively high birth rates and low life expectancy in the past, and are not indicative of future demographic trends.

Speed of Aging in Selected OECD Countries

article image
Source: OECD (2001).

10. With its rapid industrialization and the associated behavioral change, Korea is in the midst of sweeping demographic changes. The change from an agrarian economy to an industrial power in less than fifty years has been accompanied by a marked fall in female fertility—by almost four children for each woman—and the largest increase in life expectancy among OECD countries—27 years. As a result, the aging of Korea’s population will be among the fastest in the world. The population will peak at 52–53 million in the 2020s, and decline to current levels by 2050 (NSO, 2002). As a result, Korea will move from an “aging society” (defined by the United Nations as one where the population over age 65 is 7–14 percent of the total) to an “aged society” (where the population over age 65 is over 14 percent of the total) in only 22 years. This represents a small fraction of the transition time in most other OECD countries. Sensitivity analysis involving key demographic parameters suggests that this result is fairly robust.

11. With the rise of incomes and urbanization, the family-centered model has begun to break down. The most important change has been the dissolution of the multi-family household and the breakdown of the associated support system. The percentage of three-family households roughly halved from 1970 to 11 percent in the mid-1990s. While transfers from children still constitute the main source of retirement income, surveys show that these have fallen sharply, from 64 percent of retirement income in 1988 to 44 percent in 1994, replaced largely by increases in income from the government-mandated retirement allowance system. As of 1994, public and private pensions combined accounted for only 4 percent of retirement income for Koreans, compared with over 80 percent in other OECD countries.

12. As Korea moved away from the family-based support system, a number of private and public retirement schemes were introduced. Occupational schemes appeared in the early 1960s along with the retirement allowance system. The National Pension Scheme (NPS) and individual pension accounts appeared roughly one generation later.

  • Occupational schemes were introduced for civil servants in 1960, military personnel in 1963, and private school teachers in 1975. Benefits are relatively generous, with a replacement rate of 76 percent, and portable across schemes. However, these schemes cover only 4 percent of the workforce, a share that is falling due to declining government employment

  • The retirement allowance is a mandatory payment to departing employees equivalent to at least one month’s wage for each year of employment. It was established in 1961 for firms with more than 30 employees, and has since been expanded to firms with more than five employees. Over one-quarter of the workforce is entitled to the retirement allowance. Collective bargaining agreements typically specify the actual compensation, which is often more generous than the required minimum.

  • Following a 15-year delay stemming in part from the economic crises of the 1970s, the National Pension Scheme (NPS) came into effect in 1988.6 The NPS is a social insurance scheme with pension, insurance, and income redistribution elements. It is intended to cover workers not falling under any of the other public schemes.

  • Individual private pension accounts were introduced in 1994, and enjoy generous tax concessions. However, the number of accounts (one-tenth of the working population) and the average balances (one-half the average wage) remain low, due to low public awareness of—and hence demand for—pensions, as well as insolvency concerns for the institutions (insurance companies and investment trusts) managing these accounts.

Initial Reform Efforts

13. The shortcomings of the public pension system—fragmentation, inadequate coverage, and financial unsustainability—became increasingly apparent during the 1990s.7 The first major reform of the NPS began in December 1998, and included: expanding the coverage to include the urban self-employed, employees in workplaces with less than five workers, non-income earners, and foreign nationals;8 lowering the average replacement ratio for workers with a 40-year history from 70 percent to 60 percent; and increasing the retirement age in five one-year steps every five years beginning in 2013, which will raise the retirement age from 60 to 65; and increasing the contribution rate of the self-employed from 3 percent to 9 percent, which will be done in one percentage point steps each July from 2000–05. Reforms in the other public pension systems took place more or less concurrently.

14. Administrative and operational reforms followed in 2000, including a ban on obligatory, subsidized lending to the government. The size of the main operational committees was expanded to ensure a majority for representatives of insured persons groups—previously, government representatives and investment specialists held a majority of seats. Importantly, the “compulsory deposit” of National Pension Fund (NPF) resources with the government, which had received below-market rates of interest, was phased out by June 2001. In addition, an external manager for the NPF was hired, management of a small part of NPF assets was outsourced to private firms, and the NPF was allowed to invest overseas as well as in futures and options.

National Pension Fund Asset Allocation

(in percent)

article image
Source: National Pension Corporation (2001).

Long-term Growth and the Pension Balance

15. GDP growth will decline in the coming decades owing largely to demographic trends and productivity convergence. According to the OECD and the Korea Development Institute, annual employment growth will fall steadily to zero in 2020 and to -0.7 percent by 2050. This reflects broadly unchanged male labor force participation rates, a 20 percentage point rise in female participation rates to levels more in line with OECD averages, and population aging, which dominates the effect on labor force growth of higher female participation. Regarding labor productivity, the projections broadly follow OECD (2001) in assuming that Korean productivity converges to the one-half of the U.S. level, from less than one-third at present and one-tenth a generation ago.9 This implies that the growth of labor productivity declines gradually from over 3 percent in the first decade of the century to 1¾ percent by 2050. As a result, the OECD-KDI study assumes that GDP growth falls from over 5 percent this decade to 1–1½ percent in the final decades of the projection period. Although this growth assumption is low compared with most estimates of potential growth in Korea, which undoubtedly refer to a much shorter time horizon, the assumed rate is not crucial for the analysis presented below.10

16. The rapid aging of the population and the associated maturing of the NPS will result in a sharp deterioration in the pension balance. From a surplus of about 2 percent of GDP in 2002, the public pension balance declines steadily to a surplus of ¾ percent of GDP by 2020, falls sharply thereafter to a deficit of around 7 percent of GDP by 2045, and levels off at around 8 percent of GDP by 2060. Most of the deterioration comes from the NPF. The occupational schemes currently have a small overall deficit, which declines steadily before leveling off at 2–2½ percent of GDP around 2030. The projected deterioration of the pension balance in Korea is the largest among the OECD countries, although net pension outlays as a percent of GDP would be in the middle range of OECD countries by the end of the projection period.


Pension System Balance

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A001

Sources: KDI (2002) and author’s calculations.

C. Simulations

17. The analysis of pension system financial sustainability is undertaken using an approach that smoothes changes in contribution rates and in future consumption across generations. Such an approach is attractive owing to the 10 percentage point of GDP deterioration in the pension balance noted above. Given a swing of such magnitude, the role for public policy is to minimize the reduction in the consumption of future generations, who would otherwise face sharp increases in taxes to finance future pension outlays (Gruber and Wise, 2001). In essence, the idea behind the model is to use politically feasible reforms and market-based assumptions on rates of return to construct a 60-year reform-investment plan that puts the pension system on a sustainable financial path and minimizes the variance in contribution rates and consumption across generations.11

18. A summary of the model—which is formally presented in the Annex—is as follows:

  • Objective: given the demographic and macroeconomic framework described above and the pension-related parameters described below, the model calculates a constant annual resource requirement (as a percent of GDP) that is consistent with full funding by end-2060.12 Full funding is defined as an end-period asset level that is sufficient to generate returns to finance all future pension expenditures without requiring any additional non-pension system resources thereafter. Thus, the model generates self-sustaining paths for pension system assets under various reform scenarios.

  • The simulations assume that any pension-related surpluses are “lock-boxed.” This means that any pension surplus is invested and that there is no diversion, for example, into the government’s general account. In addition, the government is assumed not to finance deficits in the occupational pension plans out of general revenues, which it is doing at present—those amounts are captured under the aggregate pension balance.13

  • Baseline pension-related parameters are pension system assets of 15 percent of GDP at end-2001; a real interest rate on debt of 3 percent;14 and an equity premium of 300 basis points.15

  • The model considers two types of parameter reforms: a reduction in the replacement ratio and an increase in the contribution rate. For the replacement ratio, step reductions of 10, 20, and 30 percent over 10, 20, and 30-year time periods, respectively, are considered. These reforms are assumed to begin in 2010 and would reduce the replacement ratio to 42 percent in the 30 percent case, which is broadly consistent with a number of reform proposals currently on the table.16 For the contribution rate, increases of 33, 67, and 100 percent are considered over periods of 10, 20, and 30 years, respectively, also beginning in 2010. These magnitudes take into account that contribution rates in Korea—9 percent for the NPS and 13–15 percent for the occupational plans—are relatively low by OECD standards (see text table), and were chosen to broadly match the impact of the step reductions in the replacement ratio. Indeed, as will be seen below, the presentation of all possible parameter reform scenarios results in a matrix of residual resource requirements that is nearly symmetric.

  • The sensitivity of the results is tested by varying the equity premium by 100 bp in both directions, which could equivalently be interpreted as lower or higher real interest rates.

Pension Contribution Rates in OECD Countries, 1995

(As a percent of average earnings)

article image
Source: Blondal and Scarpetta (1998).

The sequence of the analysis is to first derive the path for pension system assets under the baseline scenario. The effects of adopting a more aggressive investment portfolio are then presented as well as the resulting resource requirements needed to fund the system. The magnitude of the residual resource requirement suggests that parameter reforms will be needed to bring the funding requirement down to a more politically feasible level. Parameter reforms are then added, and their effect on the pension balance is assessed. Finally, the tradeoffs between all parametric reforms, more aggressive portfolio investment and additional financing, as well as a sensitivity analysis, are presented.

Financial Implications of the Current Pension System

19. Under the baseline scenario, pension system assets will peak in the mid-2010s, and will be fully exhausted two decades later; a more aggressive investment strategy will only marginally postpone the depletion date. Assuming a debt-only portfolio,17 the value of pension system assets will rise to just over one quarter of GDP in the middle of the next decade, and turn negative in 2033. (The NPF will run surpluses for another 30 years or so.) Allowing up to 75 percent of the pension portfolio to be invested in equities only pushes out the peak year and the year in which the funds are depleted by five years, with system assets peaking at 35 percent of GDP. The implications are twofold: first, the deterioration of the public pension balance outweighs the positive effects of the higher investment returns; and second, that an aggressive investment strategy alone will not ensure pension system sustainability.

Pension System Performance Baseline Scenario

article image

20. Without parametric reforms, funding the pension system would require additional financial resources of 2¼–5¾ percent of GDP per year. The results appear in the upper-left cell of Table I.1.a. The high end of the range represents the additional resources required if no equity investment were undertaken, while the low end reflects a 75 percent equity allocation.18 These results are somewhat sensitive to the assumption for the equity premium. For example, for the 50 percent equity allocation, reducing (increasing) the equity premium by 100 basis points increases (reduces) the annual resource requirement to fund the system by around ½ percent of GDP. It is doubtful whether sustained fiscal surpluses of 2¼–5¾ percent of GDP are politically feasible, suggesting the need for further reforms to the public pension system to reduce its cost.

Reform Scenarios

21. Implementation of parametric reforms can substantially improve the pension balance. Assuming a zero allocation to equity in the portfolio, each increment in the reforms to the replacement ratio and contribution rate outlined above (i.e., 10 percent reductions in the replacement ratio and 33 percent increases in the contribution rate) would improve the pension balance by around 1 percentage point of GDP by the end of the projection period. Thus, adoption of the strongest reforms to both parameters would reduce the pension balance deficit to 2 percent of GDP by the end of 2060.

Pension Balances at end-2060 Under Reform Scenarios

(in percent of GDP)

article image

22. Combined with parametric reforms, moderate allocations to equity in the portfolio provide a number of options to reduce the annual resource contribution necessary to fund the pension system to less than 1 percent of GDP (Table I.1.a). Assuming an equity allocation in midpoint of the 25–50 percent range, the resource requirement could be reduced to under 1 percent of GDP per year by implementing a 30 percent reduction in the replacement ratio combined with a one-third increase in the contribution rate, a 10 percent reduction in the replacement ratio combined with a doubling of the contribution rate, or any combination in between. The elimination of any additional resource requirement could be achieved by increasing the strength of both parametric reforms by one additional increment. The stock of assets at end-2060 consistent with funding the system would fall by around one-half in these scenarios compared to the baseline. Regarding sensitivity, a 100 bp decrease (increase) in the equity premium would raise (lower) the annual resource requirement by roughly ½ percent of GDP (Tables I.2).

Table I.1.

Pension System Baseline and Reform Scenarios

(in percent of GDP)

article image
Note: The northeast, southwest, and southeast elements of each cell correspond to pension portfolio equity allocations of 0, 25, 50, and 75 percent, respectively. Negative figures in Table I.1.a indicate an excess of resources.Source: Author's calculations.
Table I.2.

Sensitivity Tests

(in percent of GDP)

article image
Note: The northeast, southwest, and southeast elements of each cell correspond to pension portfolio equity allocations of 25, 50, and 75 percent, respectively. Negative figures indicate an excess of resources.Source: Author's calculations.

D. Institutional and Operational Issues

23. The reforms to the public pension systems currently under consideration, which are not limited to those discussed above, suggest a number of institutional and operational issues that will need to be addressed. These issues will tend to increase in importance as the system—in particular the NPS, given its predominance—matures and as pension assets increase.

24. Attention will need to be paid to the sequencing of pension reforms, including in combination with the second pillar system. While there is consensus among Korean economists as to the source of the current sustainability problem, there is no consensus on the solution. Time will be needed to build momentum for the next round of reforms. On the link with the second pillar system, it will be important to establish a credible transition from the current, largely unfunded retirement allowance system to a joint employer-employee funded defined contribution plan in order to lay the groundwork for the reforms to the first-pillar system described above.19 With that would come a welcome development in the investment culture and investor sophistication that would hopefully allow the kinds of reforms under discussion to become politically feasible.

25. The public-private split of any first-pillar system will need to be resolved. As noted above, Feldstein and Samwick recognize that, in principle, pension contributions in their framework could be collected and invested by the government, although they preferred a decentralized system. This sentiment has been echoed by a U.S. expert panel on social insurance (Diamond, 1999), who argue that adoption of a private model would; (i) instill ownership in the retirement system on the part of citizens; (ii) offer individuals more choice regarding their investments, thereby better aligning their portfolios with individual preferences; (iii) reduce the influence of the government over both the investment of the funds and the workings of corporations whose stocks were held; and (iv) lessen the temptation for the government to spend the accumulated funds. That said, the public system still enjoys considerable support in Korea. The public-private issue will need to be debated against the backdrop of an increasingly sophisticated financial system and an increasingly financially sophisticated populace. However that debate is resolved, the need for appropriate firewalls and high transparency remain regardless of who manages pension system assets.

26. The issue of a passive or an active investment approach would need to be addressed. The passive, or indexing, investment approach is based on efficient markets theory and would entail a strategy of trying to mimic—rather than outperform—a benchmark index. This approach would carry a number of advantages: operating costs tend to be lower since there is lower turnover; the need for investment advisors is lower since the main objective is to replicate (not beat) the performance of a selected benchmark; and the system would be more transparent and therefore less prone to political interference. If a more active approach is chosen, consideration will need to be given to the issue of goal and instrument independence, with the owner of the funds setting the former and the managers having the latter. Management expertise will, of course, be a much more important issue in an active system, as will the issue of the appropriate level of competition amongst asset managers.

27. There are clear benefits from investing pension assets abroad, although this remains a delicate issue in Korea. The main advantage of pension funds investing abroad would be to diversify the portfolio; indeed Korea, like most other countries, suffers from a “home country bias” in its public pension asset allocation. Local markets are not perfectly correlated with international markets, so investing abroad would improve the risk-return tradeoff. Moreover, the size of Korea’s bond and equity markets remains modest by international standards, and managing and purchasing assets of the size discussed above- often over 100 percent of GDP—could have distortionary effects on the market. Furthermore, international investment would hedge against the high long-run correlation between real returns on human and physical capital in Korea. Foreign investment of pension funds would lead to foreign currency risk, which would need to be managed, though arguably not fully hedged over the long run.

E. Conclusions and Further Research

28. Sustained and relatively modest reforms should be sufficient to ensure the financial sustainability of Korea’s first-pillar pension system. These reforms include a phased reduction in the replacement ratio of up to 30 percent, a phased increase in the contribution rate of up to 100 percent, and the adoption of a significant—but still minority—allocation to equity in the pension asset portfolio. The results also hinge on “lock-boxing” the ongoing pension system surpluses, but not on continuing to finance deficits in the occupational plans out of general revenues. Under the assumption that achieving pension system sustainability would require additional fiscal resources (say, on the order of 1 percent of GDP per year), this would have implications for medium- and long-term budgeting. When these reform measures in the first pillar system are combined with the necessary reform of the current retirement allowance system into a private savings plan, the sum of the replacement ratios for the first and second pillar systems would be on a par with comparator countries.

29. Additional work should verify to what extent the results obtained here are model dependent. Possible alternative specifications could include: a rolling smoothing period, varying the end-period objective, letting the fiscal savings rule vary over time, and consideration of a longer projection period. Whatever approach is taken, the margins of error involved in compounding and in adopting realistic assumptions on rates of return and equity premia argue for regular reassessment and recalibration of the system.

30. Finally, the macroeconomic effects of pension reform strategies need to be explored further. The partial equilibrium framework employed above is sufficient to identify the “first round” effects and trade-offs involved in moving to a financially sustainable pension system. While useful, this approach arguably falls short of a definitive assessment of the funding issues. Three key relationships that would need to be modeled in a computable general equilibrium model in the context of pension reform would be: the effects on national savings, or, alternatively, the amount of Ricardian offset; the response of labor supply to any tax changes and pension wealth; and the effects on the composition of the current account, particularly from investing large amounts of assets abroad. Ideally, a general equilibrium analysis would also capture the interaction of the defined-contribution and defined-benefit schemes under consideration, which would lead to a better informed investors and pensioners and more precise tools for policymakers.


  • Blöndal, Sveinbjörn, and Stefano Scarpetta, 1998, “The Retirement Decision in OECD Countries,OECD Economics Department Working Paper No. 202.

    • Search Google Scholar
    • Export Citation
  • Dang, Thai Than, Pablo Antolin and Howard Oxley, 2001, “Fiscal Implications of Ageing: Projections of Age-Related Spending,OECD Economics Department Working Paper No. 305.

    • Search Google Scholar
    • Export Citation
  • Edleson, Michael E., 1993, Value Averaging, Independent Publishing Corp., Chicago, p.15.

  • Feldstein, Martin, and Andrew Samwick, 1997, “The Economics of Pre-funding Social Security and Medicare Benefits,” in B. Bernanke and J. Rotemberg (eds.) NBER Macroeconomics Annual 1997, pp. 115148 (Cambridge, Massachusetts: MIT Press)

    • Search Google Scholar
    • Export Citation
  • Feldstein, Martin, and Andrew Samwick, 1999, “Maintaining Social Security Benefits and Tax Rates through Personal Retirement Accounts: An Update based on the 1998 Social Security Trustees Report,NBER.

    • Search Google Scholar
    • Export Citation
  • Gruber, Jonathon, and David A. Wise, 2001, “Different Approaches to Pension Reform from an Economic Point of View,NBER-Kiel Institute Conference.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2002, Global Financial Stability Report (June), pp. 1011.

  • Kong, Qingying, 2001, “Options for Pension Reform in Korea,” in Korea: Selected Issues, IMF Country Report No. 01/101 (Washington, D.C.: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Korea Development Institute, 2002, “Economic Impact of the Aging Population and Countermeasures,unpublished mimeo.

  • Lee, Hyehoon, 2002, “Aging Impacts on the Structure of the Korean Economy,Korea Development Institute, unpublished.

  • Mackenzie, G. A., Philip Gerson, Alfredo Cuevas, and Peter S. Heller, 2001, “Pension Reform and the Fiscal Policy Stance,IMF Working Paper WP/01/214 (Washington, D.C.: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Moon, Hyungpyo, 2002, “The Korean Pension System: Current State and the Tasks Ahead,Korea Development Institute, unpublished.

  • National Pension Corporation, 2001, National Pension Scheme in Korea, Ministry of Health and Welfare, Korea.

  • National Statistical Office, 2002, “Estimated Population,available on

  • Organization for Economic Cooperation and Development (OECD), 2001, OECD Economic Surveys: Korea, Chapter 3.

  • World Bank, 1994, Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth, Oxford University Press.

  • World Bank, 2000, Republic of Korea: The Korean Pension System at a Crossroads (Washington, D.C.: The World Bank)

ANNEX The Model

The model is a discrete time, finite horizon optimization problem. The planner is given time paths for real GDP and pension expenditures as well as parameter values for the real riskless interest rate, equity premium, and portfolio allocation. Subject to these, she chooses a time invariant government savings rule so that the end-period pension fund asset stock is sufficient to generate resources to fund pension outlays for all subsequent years without reducing end-period principal. The model must be solved using algorithmic methods.20

Formally, the model is as follows:



Subject to:

At11+gt+X1+Y1+Z2Pt+itAt¯2 =Atit =(1α)rt+αet =(1α)rt+α(rt+st) =rt+αsti50A50* =P50A1 =15


At = end-period pension system assets in year t

At¯ = the average pension system assets in period t

Xt = the fiscal policy rule (i.e., surplus excluding pension balance) in period t

Yt = the additional fiscal surplus needed to fully fund the system

Z2 = a one-off injection into the pension system in period 2

Pt = net pension balance (contributions less outlays) in period t

gt = real GDP growth rate in year t

it = real riskless (government bond) rate of return in period t

α = share of portfolio allocated to equities

et = real return on equities

st = equity premium (in percentage points)

Scenario values:

α = {0,.25,.5,.75}

s = {2,3,4}

All variables are real, and upper-case denotes variables expressed as a percentage of GDP. Time is indexed so that 1,2,3… correspond to 2001, 2002, 2003…:

The model was also used to simulate the “stock solution” to the problem where the objective is to choose YI* with Yt* = 0 for t = 2, 3, 4, …, 60. In these scenarios, the question is what about of resources is needed up front (i.e., in period 1 only) to fund the pension systems.


This paper was prepared by Paul F. Gruenwald (Resident Representative in Korea).


World Bank (1994) pension nomenclature will be used throughout. “First pillar” refers to a publicly mandated, defined benefit system with the main goal of reducing poverty. “Second pillar” refers to a mandatory, defined contribution, privately managed system. A voluntary, private “third-pillar” system is not discussed.


The public pension system is defined here as the National Pension System plus the occupational public pension schemes for veterans, teachers, and civil servants.


The parametric reform measures in this paper are broadly consistent with the recommendations of the Pension Reform Task Force—see Kong (2001) for details.


This section draws on OECD (2001).


For a history of Korea’s public pension system as well as recent reforms, see National Pension Corporation (2001).


See World Bank (2000) for a detailed discussion of these issues.


The system’s initial coverage included workers in firms with 10 or more employees, which was extended in 1992 to workers in firms with five to nine employees.


This was labeled the “convergence” scenario in the OECD-KDI study. A “no convergence” scenario, where labor productivity in Korea is assumed to grow in line with the U.S., which would be less than in the convergence scenario, was also presented, but is not discussed here.


Lee (2001) finds that changes in labor productivity growth, and hence output growth, have only a marginal impact on the pension balance: a 0.5 percentage point reduction (increase) in annual labor productivity growth resulted in a 15 percent decline (increase) in both pension-related revenues and expenditures by 2050, leaving the pension balance largely unaffected, as the reduction in pension contributions is offset in percentage terms by reduced benefits.


The methodology is based on Feldstein and Samwick (1997), who study the transition path from a partially-funded government-run pension system to a private, fully-funded plan for the United States. Their approach exploits the difference between the historical returns to equity and the implicit return in the current pension system to finance the transition to a fully-funded pension system. In a related model (Feldstein and Samwick, 1999), mandatory individual retirement accounts are set up and invested in a 60–40 equity-bond mix.


The time frame is, of course, arbitrary. The adoption of 2060 as the end year accords with the latest pension projections of the Korea Development Institute. Feldstein and Samwick use a 75-year projection period.


Social security funds as defined in the consolidated fiscal balance released by the government exclude the government employees pension fund and the military pension fund.


The aging model for a stylized OECD country assumes a 4 percent real interest rate (Dang et al., 2001).


The 10-year average equity premium is 340 basis points (bp) for the United States, 280 bp for Germany, and 360 bp for Japan (IMF, 2002). During 1926–91, the equity premium in the United States was 700 bp (Edleson, 1993).


Moon (2002) argues that benefit reduction is more defensible than contribution increases since Korea’s first-pillar replacement ratio is 60 percent compared with 40 percent in the U.S., U.K., and Canada, and in light of the (implicit) 25 percent replacement ratio in the second-pillar system. He notes from Schmitt (1985) that the aggregate replacement ratio for an OECD country should be 55–70 percent.


This assumption is intended to approximate the end-2001 composition of the NPF, where equity investments comprised less than 10 percent of the portfolio.


Alternatively, a one-time up-front funding of the pension system would require resources of some 200 percent of GDP for equity allocations in the 25–50 percent range (Table I.1.b, upper-left cell).


This has become particularly pressing since members of President-elect Roh’s transition team have come out against promoting the development of a second-pillar system.


Simulations for the present model were solved using the “Goal Seek” function in Excel.