PAYG versus funding
This article addresses three interlinked issues which are relevant to the current debate relating to retirement income provision. First, we consider the benefits and disadvantages of pay-as-you-go social security pensions. Second, we assess why funding may be seen as a superior alternative. Third, we consider arguments about the long term viability of funding, referring inter alia to the influence of forms of regulation. The discussion raises a number of policy issues and choices, which are highlighted in the text. Overall, the conclusion is that a partial switch to funding is justified but equally a basic form of pay-as-you-go should be retained, not least given the need for a "safety net" against poverty in retirement. As background, Table 1 shows pension indicators for the G-7 countries, namely demographic indicators illustrating the ageing of the population, indicators of the future burden of pay-as-you-go schemes, indicators of the scope of private funding and of the returns to funding and pay-as-you-go.
1 The case for and against social security pensions
The justifications for the compulsory public provision of pensions, are mostly economic (as they overcome a form of market failure) while others are social (in that they promote a desired distribution of income). These include the following (see Johnson (1992)):
· overcoming the market failure of information inefficiencies, whereby governments help each individual to save for retirement by providing a base level of benefits that is at least the minimum all would want, without needing to gather information on the precise nature of each individual’s preferences;
· by making participation compulsory and offering life annuities as the only option, public social security pensions help overcome the market failure of adverse selection problems that may plague private annuity markets - the main private sector alternative;
· insurance companies may also be unable to cover inflation risks and may fail during severe recessions;
· overcoming market failures relating to free rider problems, whereby if individuals know society will not let them die in poverty, they will not save (a problem of moral hazard); public pensions force all members of society to contribute to this safety net;
· transactions costs of individual private pensions for the bulk of the population may be uneconomically high, particularly for low earners and
· pay-as-you-go social security provides a form of risk-sharing, pooling the risk of variable returns to human as well as non-human capital (i.e. wages as well as asset returns) across the generations (see Merton (1983)).
· the decline of more traditional means of caring for the elderly (the extended family), combined with paternalism, lead governments to try to overcome the problem that individuals may not cater for their own retirement due to myopia;
· governments may find pay-as-you-go social security a convenient means of redistributing income, thus catering for the lifetime poor;
These justifications reflect a mix of insurance‑based and welfare‑based justifications, whereby adverse selection, free riding, risk sharing, transactions costs and information inefficiency are clearly arguments for the superiority of government over private insurance markets, while substitution for traditional modes of care, paternalism and redistribution are more welfare‑based reasons. It should be noted that it is mainly the welfare-based arguments which justify pay-as-you-go as a means of financing. The insurance-based justifications make more of a case for public or compulsory provision, irrespective of whether financing is pay-as-you-go or funded. Moreover, as noted by Bodie and Merton (1992), James (1993) and World Bank (1994), none of these arguments gives any guidance on the appropriate level of benefits. If anything, they suggest that governments should only provide a minimum level of security common to all, and not a uniform high level of provision regardless of individual preferences and endowments.
As regards other economic benefits of social security, there may also be broader macroeconomic benefits, as long as the system is seen as sustainable, whereby the anticipation of stable incomes over the life cycle may contribute to the stability of aggregate demand, protecting the economy against cyclical instability and thus promoting investment. Expectations of stable income may also stimulate consumer demand, given the reduced need for precautionary saving. And social security has the social benefit of correcting for the “market failure” that the private sector is unable alone to redistribute income, to alleviate poverty, and to promote social welfare and stability.
Economic disadvantages of pay-as-you-go social security systems may affect both labour and capital markets. In labour markets, employers’ and employees’ social security contributions increase the gap between labour costs, which determine labour demand decisions and net wages, which influence labour supply decisions. For employers, if they operate in competitive markets and cannot merely raise the product price, there is consequently an incentive to substitute capital for labour - or shift production to another country. Indeed, there is evidence that the level of wage and non-wage costs together does have a direct link to the level of unemployment (see Balassa (1984)). Such effects may be aggravated if dismissal of workers is difficult, as in Germany (see Schlesinger (1985)).
As regards labour supply, if younger workers perceive there to be no relation between contributions and pensions, or if this is disregarded due to myopia, then effort or hours worked will decline, if not labour force participation per se. But these effects are reduced if contributions are seen as giving rise to credible entitlements in the same way as saving - or an actuarially fair funded scheme - would, and disappear if the implicit saving is felt to have an adequate rate of return.
For older workers, availability of benefits at a certain age may be a strong inducement to retire, particularly as evidence suggests that changes in benefits of up to 10‑15% have little effect on decisions. Second, early retirement may be induced in the case that public pensions pay more over the lifetime if benefits are claimed early, as is often the case. These effects may be compounded by similar features of private pension schemes. Third, there is little incentive to participate partially in the labour market when due to the curtailment of retirement benefits there are effective marginal tax rates of 100% on earned income. Note that early retirement incentives are often a deliberate feature of pay-as-you-go seen as a “painless way to cut unemployment among the young - or to cut the number of redundant employees”. This is an example of the lump of labour fallacy, that there are a fixed number of jobs in the economy. Evidence does suggest that early retirement is induced by social security. This effect may be sizeable; James (1993) points out that the labour force in OECD countries would have been 3-7% higher in 1990 if the 55-64 age-group had participated in the labour market at the same rate it had done in 1960, entailing considerable loss of potential output - and of contributions.
As regards capital markets, there is some evidence for the United States and in international cross-section (see Feldstein (1974, 1977, 1995a)) that unfunded social security pensions reduce aggregate saving and hence capital accumulation and growth. This can be justified theoretically by a life cycle framework, where individuals structure their lifetime saving and asset accumulation to maintain steady-state consumption. If social security provides a guarantee of income to maintain consumption after retirement, then there is a form of implicit wealth accumulation, and the need to save during the working life is lessened. Underlying Feldstein’s approach, and in contrast to the discussion of the labour supply, is a view that workers see contributions by themselves as a form of saving and not as a tax. A further mechanism inducing lower saving under pay-as-you-go social security is that those who are myopic and would have otherwise continued working till they die are now able to retire.
Feldstein’s results have been disputed; but what is clear is that if a social security system is structured so as to provide benefits to a generation in excess of its contributions, then there will clearly be a reduction in saving thanks to the wealth transfer. The “free pensions” provided to the first generations in social security schemes which have not contributed are examples of this, so long as the public sector did not run an offsetting surplus. This may account for clearer results on the negative effect of social security on saving for certain other countries with generous social security pensions such as Sweden, Italy and Japan (noted in Hagemann and Nicoletti (1989)) than Feldstein obtained for the United States. Moreover, as noted by James (1993), the conditions under which funding will have a positive effect on saving - namely, myopia, limited access to credit, and lack of credibility of the pensions scheme - are precisely those whose absence will lead pay-as-you-go to reduce saving. So a switch from pay-as-you-go to funding is unambiguously likely to raise current saving in an economic sense, with the private sector and/or the public sector saving more.
Despite the various disadvantages noted above, the main incentive for governments to switch from pay-as-you-go to funding is the difficulty of financing pay-as-you-go as the population ages without unsustainably high contribution rates or chronic fiscal deficits (see Roseveare et al (1996), IMF (1996), Davis (1997a), also Table 1). In a scenario of high contributions, returns to pay-as-you-go will fall sharply, worsening disincentives; in the case of fiscal deficits, investment may be crowded-out and the country ultimately face a financing crisis. We now turn to the relative advantages of funding, with a particular focus on this issue.
2 Benefits of switching to funding
Conceptually, funding - setting aside and investing premia from each generation so their pensions are paid from the return stream - has a number of advantages. It increases the actuarial fairness of the system, given a tighter link of benefits to contributions than for pay-as-you-go, and thus contributions are more likely to be seen as saving than taxation. Hence funding is likely to reduce distortions to labour and financial markets and to saving and may in itself reduce the overall economic impact of ageing, by boosting labour force participation, and also by potentially raising aggregate saving, thus increasing the stock of fixed capital and the output out of which future pensions are to be paid. Even if saving remains rather stable, its structure is likely to shift towards longer term instruments such as equities, which may be favourable to productive investment, as well as enhancing the development of capital markets (stimulating innovation, efficiency and liquidity) and hence fostering efficient resource allocation (Davis (1996a, 1996b)). By raising growth 'endogenously', such effects could help to provide the resources necessary to cater for the remaining liabilities of pay-as-you-go, and/or for those elements of pay-as-you-go that for social reasons it is considered necessary to retain (Holzmann 1997).
In this context, it is relevant to note that in a steady state (with no change in the size and age distribution of the population), a pay-as-you-go pension scheme provides a return similar to the growth of real average earnings plus population growth (see Aaron (1966)), while a funded scheme offers real capital market returns. The latter are generally higher, so long as the portfolio includes a sizeable proportion of equities (Table 1), although returns may themselves be affected by the scope of funding, and risks may be higher especially for defined contribution funds. Also population ageing is likely to reduce returns on pay-as-you-go well below the “steady state” of growth in population plus average earnings.
Funding allows for risk diversification via international investment of accumulated funds, thus reducing the vulnerability of the retired to the overall performance of the domestic economy, which may deteriorate as population ageing becomes more severe. And assets accumulated under funding, since they are a form of private property, may be more secure against future political developments in the light of population ageing than are promises made under pay-as-you-go. For social security pensions remain subject to the political risk that benefit promises will be reneged upon.
Funding can also be seen as a form of burden transfer in the light of ageing, and more generally as a buffer against the need to raise contribution rates at a potentially undesirable time in the face of deteriorating economic performance or demographic shocks. The OECD (1993) calculates that the maximum rise in contribution ratios required under pay‑as‑you‑go to eliminate unfunded pension liabilities in the EU-4 is 4.4‑11.9% of GDP, whereas for funding it suggests that a sustained increase of 1.1‑5.3% would suffice.
Holzmann (1997) notes three particular benefits that funding could offer to the EU per se. First, labour mobility between member states would be enhanced, thus allowing gains from specialisation to be fully realised, if there were a co-ordinated and funded pension scheme in the EU on a defined-contribution basis. Second, workers may be more willing to accept the adjustments to labour market conditions required to cope with globalisation if they also have a stake in capital market gains at an EU and global level. Third, partial funding of pensions could help protect the EU against the symmetric demographic shock to which it will be subject, if on the one hand the growth effects identified above are realised, or on the other a significant share of funds are invested internationally, thus allowing risk diversification.
Despite these arguments, a wholesale switch to funding would be unlikely to be desirable or even feasible, particularly because funding is unable to redistribute to retired persons facing poverty in the way societies typically prefer; and more generally funded pensions are often ill-suited to low income workers or those with broken career patterns. Also it may be optimal to provide both forms of retirement income provision as a means of risk diversification. This is because as noted pay-as-you-go and funding are subject to different risks which are to some degree independent of one another. Finally, as discussed below, a wholesale shift would probably be extremely costly, given the scope of existing commitments under pay-as-you-go that would still have to be honoured.
Note, however, that these arguments for a partial shift do not imply a need for comprehensive pay-as-you-go schemes providing high replacement ratios regardless of income and individual preferences. Rather, they may justify a basic level of social security to alleviate poverty, allowing pensions over and above this level to be funded. Such a form of specialisation for the two systems, with the unfunded element specialising in redistribution and the funded element in the provision of annuities, may help reduce the distortions to labour and financial markets induced by the unfunded element (World Bank (1994)).
A general problem that arises in policy discussion of funding in countries currently dependent on pay-as-you-go is that there may be major fiscal problems, which can spill over to political resistance. In effect, funded pensions do not relieve pressure on public finances in the short run, as existing pension promises need to be met and, usually, tax relief granted on contribution and asset returns, with little tax revenue from the initially low amounts of funded pension payments to offset these costs. Hence the need for a rather contractionary fiscal stance, and the likelihood of political resistance to generations in the transition being thereby forced to "pay twice" for pensions, once for the previous generation via pay‑as‑you‑go, and once for its own via funding. These points raise an important public policy issue of how a transition is to be financed and the burden distributed between generations. Pay-as-you-go as it stands clearly imposes too great a burden on future generations; but how much redistribution of such burdens is appropriate?
As noted in Holzmann (1997), rather than forcing the current generation to pay twice by budget financing of the transition, the polar opposite is to recognise the implicit government debt which is represented by the accumulated benefit obligation of pay-as-you-go, and convert it immediately to explicit debt. In this case the transition is financed largely by future generations. In this context, Feldstein (1995b) suggests such bond financing of the transition can help redistribute the burden between generations, so the future generations who will benefit from the efficiency gains of a more flexible labour market and financial market development, as stimulated by funding, will also pay some of the costs. However, given the scope of current accrued obligations under pay-as-you-go, typically well over 100% of GDP, this would seem not to be feasible without severe effects on financial markets and on confidence in the domestic economy.
Accordingly, OECD governments have preferred in current circumstances to focus largely on scaling back their benefit promises to current and future generations, implicitly “defaulting” on part of their pension obligations. As noted by Holzmann (1997), such a process of reform, by reducing the benefit obligation of pay-as-you-go, may facilitate a partial switch to funding - whether financed by borrowing or taxation - at a later stage. It typifies the process undertaken successfully by some Latin American countries, notably Chile, and which is under active consideration in Eastern Europe.
Turning to types of funding, one may distinguish partial funding of social security from private funding. Social-security funding was adopted at an early stage in Sweden, such that assets valued at around 33% of GDP have been accumulated. Similar schemes exist in Finland and Denmark. An advantage of a public scheme is that the labour mobility problems typical of voluntary occupational schemes can be avoided by such a compulsory funded social security scheme. Difficulties are that particularly if there is a degree of redistribution, contributions to a social security trust fund may be seen as taxes, thus engendering distortions to labour markets and other welfare losses. Moreover, a social security trust fund may face problems in investment (Thompson (1992)). A large trust fund may induce higher government consumption or even fiscal deficits, thus actually reducing national saving, and its management could be subject to political interference. Investment in government bonds, which is typical of such funds, has ambiguous consequences. It is likely to eliminate any benefit to national saving as a consequence of funding. Even if used to fund investment, finance may be diverted to unprofitable projects for political reasons. Also lack of international investment, which is typical of social security trust funds, leaves them dependent on the performance of the domestic economy.
Funding through occupational pension funds or individual arrangements avoids some of these difficulties. Benefits to saving arising from a switch from social security are more likely, as workers will perceive contributions as saving invested at market rates of return. Fund managers may focus on maximisation of return for a given risk, which will ensure efficient allocation of funds in the capital market. By being more able to invest internationally, they may avoid being constrained by limited investment opportunities in the home economy and reduce risk. Private pensions, notably defined contribution plans, are more capable of meeting individual preferences, while defined-benefit plans may provide intergenerational risk-sharing similar to pay-as-you-go. Company based defined contribution funds may have attractions relative to individual funds given lower administration costs and reduction in the incidence of adverse selection in annuities markets. It should, however, be noted that private pensions have some disadvantages, notably cost of regulation, administrative costs, vulnerability to market risks (notably for defined-contribution funds), inability to redistribute and, for defined-benefit funds, obstacles to labour mobility if portability is limited and possibly the need for compulsion at the level of the firm to make them viable. Final-salary defined-benefit funds may increase incentives of employers to lay off older workers, as the rate of their pension accruals increases as retirement age approaches.
3 The long term viability of pension systems
An important question that must be answered by those favouring funding is whether it is any more viable than pay-as-you-go, faced with coming demographic difficulties. This question can be addressed at a micro and macro level. One objection to all types of funding, taking a ‘closed economy’ view, is that it is actually equivalent to pay-as-you-go in that in each case pensions must be paid from domestic production, for pay-as-you-go from labour income and for funding from capital income (see Samuelson 1958). Funding may adversely influence the exchange rate and the current account if ex ante domestic investment is less than the increase in saving. Moreover, in the longer term extra saving generated by a switch to funding may reduce the interest rate, thus reducing the benefit of funding relative to pay-as-you-go. Next century, pension funds are likely to become sellers of assets as the population ages, which could depress asset prices (Schieber and Shoven 1994).
However, as noted above, funding may raise growth potential by reducing distortions to labour and financial markets relative to pay as you go, shifting the structure of saving towards longer term instruments such as equities and possibly raising saving per se (Holzmann 1997). Hence the future state of the economy may differ between the cases of funding and pay as you go. Moreover, the possibility of international investment leaves pensions less dependent on the performance of the domestic economy. Indeed, there are strong arguments that investment from funding should flow to countries with younger populations, whose investment needs exceed national saving. Conceptually, this allows a form of burden-sharing at a global level. Moreover, when these countries age in the next century, they may be willing buyers for assets sold by pension funds from OECD countries. For such international investment to occur, it is essential that portfolio regulations be suitably liberal.
At a micro level, other aspects of pension fund regulation may have an important role to play in ensuring funds remain viable. For defined benefit funds, adequacy of funding is essential; underfunding, as was rife in the US in the 1980s and in Japan now, may put benefits under threat (Smalhout 1996). For defined contribution funds, the issue is whether contributions are adequate given rates of return and administrative costs (Davis 1997c considers the UK situation). External as opposed to internal funding spreads the risks to an individual firm of the ageing and diminution of its workforce. Internal funding is also subject to liquidity risks when pensions are paid, which are absent for external funding. Adequate provision of internally-funded pensions, as in Germany, is likely to be particularly difficult for declining industries, as the worker/pensioner ratio falls (Frijns and Petersen 1992).
As regards regulations for external funding, defined contribution funds impose less of a burden on firms than defined benefit and are in that sense more sustainable - the risk to the worker is that ageing may reduce returns in the capital markets, and/or, as noted, they may make inadequate contributions (Samwick and Skinner 1993). For defined benefit funds, an important argument in favour of funding the pension benefits in advance (the Projected Benefit method) as opposed to as they accrue (the Accrued Benefit method) is that it ensures advance provision for the burden of maturity of the plan, when there are many pensioners and fewer workers, by spreading costs over the life of the plan (Frijns and Petersen (1992)). Such burdens would otherwise be severe owing to "backloading". More generally, taking account of future obligations instead of purely focusing on current liabilities is likely to permit smoother levels of contributions as the fund matures, which may be better for the financial stability of the sponsor. Where funding and taxation provisions encourage firms to delay funding; the risk to future pensioners is that firms will in due course find the burden too great, leading to curtailment of benefit promises (Schieber and Shoven 1994).
Again, portfolio regulations, if they make benefit provision expensive, may make the burden of providing company pensions much greater than would otherwise be the case. There is a clear advantage in terms of realised rates of return to having "prudent man rules" enjoining sensible diversification over portfolio restrictions forcing funds to hold assts in a certain proportion (which typically require holding of a high proportion of bonds). Davis (1997b) shows that pension fund sectors with prudent man rules typically earned more per annum over 1967-90 than those with portfolio restrictions. Meanwhile, estimates suggest that a 1% higher rate of return can reduce contributions by 2.5-3% of labour costs (EFRP 1996). The overall success of a system of privately-funded pensions may also be measured by the degree to which it can provide comprehensive cover. Experience typically shows that voluntary coverage leads to a focus on subgroups of the population (men, white collar, large companies, unionised, etc.). Compulsory provision as in Australia, Switzerland and Denmark can avoid this. It would facilitate job mobility by standardising terms and conditions. Notably for personal pensions, compulsory participation should help to avoid adverse selection problems which typify free markets in annuities. It could be argued that if funds are compulsory, then relative tax advantages are not needed. On the other hand, compulsion could also have an adverse effect on the corporate sector, since it would impose an unavoidable burden on companies, which in turn could affect international competitiveness of the economy. These effects would make measures to minimise costs, such as a prudent man rule and competitive fund management, all the more urgent. Also such schemes tend to be defined contribution, thus imposing greater risk on workers than would a combination of (voluntary) defined benefit schemes and social security.
Pay-as-you-go social security pensions are not without benefits; they may for example help to stabilise consumers' expenditure and alleviate poverty. But their inability to cope with population ageing, combined with their deleterious effects on labour and financial markets, are prompting an increasing interest in the alternative of funding pensions. Funding offers a number of advantages, notably less distortionary effects on markets and higher rates of return. But the inability of funding to redistribute as well as considerations of risk diversification mean that it is appropriate to retain at least basic pay-as-you-go. The difficulties of transition from pay-as-you-go to funding also tends to support a partial shift. Meanwhile private funding is arguably superior to public accumulation of vast "trust funds". It can be argued that the viability of funding is aided in particular by the scope it offers to invest internationally. But in addition, the regulations of funding and investment which are applied to funded pensions may have an important influence on retirement income security.
Aaron, H. J. (1966), “The social insurance paradox”, Canadian Journal of Economic and Political Science, 32, pp. 371‑77.
Andrews, E.S. (1993). Private pensions in the United States. OECD Series on Private Pensions and Public Policy. OECD, Paris.
Balassa, B. (1984), “The economic consequences of social policies in the industrial countries”, Weltwirtschaftliches Archiv, 2/84.
Besseling, P. J. and Zeeuw, R. F. (1993), “The financing of pensions in Europe: Challenges and opportunities”, Research Memorandum No. 111, Central Planning Bureau, The Netherlands.
Bodie Z (1991), "Shortfall risk and pension fund asset management", Financial Analysts Journal, May/June 1991.
Bodie, Z. and Merton, R. C. (1992), “Pension benefit guarantees in the United States; a functional analysis”, in ed. R. Schmitt “The future of pensions in the United States”, University of Pennsylvania Press.
Davis, E.P. (1995), ‘Pension Funds, Retirement-Income Security and Capital Markets, an International Perspective’, Oxford University Press.
Davis, E.P. (1996a), ‘Pension fund investments’ in ed. B. Steil, ‘The European Equity Markets’, Royal Institute of International Affairs
Davis, E.P. (1996b), ‘The role of institutional investors in the evolution of financial structure and behaviour’, in ed. M. Edey ‘The Future of the Financial System; Proceedings of a Conference’, Reserve Bank of Australia, Sydney, Australia.
Davis, E.P. (1997a). “Can pension systems cope? Population ageing and retirement income provision in the European Union.” Special Paper. Royal Institute of International Affairs.
Davis, E.P. (1997b).”Regulation of pension fund assets”, forthcoming, OECD.
Davis E.P. (1997c). “Private pensions in OECD countries, the United Kingdom”, OECD, Paris
Diamond, P. A. (1993), “Privatisation of social security; lessons from Chile”, NBER Working Paper No.4510.
EFRP (1996), ‘European Pension Funds, their Impact on Capital Markets and Competitiveness’, European Federation for Retirement Provision, Brussels.
Feldstein, M. (1974), “Social security, induced retirement and aggregate capital formation”, Journal of Political Economy, 82, pp. 905‑6.
Feldstein, M. (1977), “Social security and private saving; international evidence in an extended life cycle model”, in ed. M. Feldstein and R. Inman “The economics of public services”, International Economic Association.
Feldstein, M. (1995a), “Social security and saving, new time series evidence”, NBER Working Paper No. 5054.
M. (1995b), ‘Would privatising
social security raise economic welfare?’, NBER Working Paper
Franco, D. (1996), ‘The taxation of funded pension schemes and budgetary policy’, European Commission Economic Paper No. 117.
Frijns, J. and Petersen, C. (1992), “Financing, administration and portfolio management; how secure is the pension promise?”, in “Private pensions and public policy”, OECD, Paris.
Hagemann, R. P. and Nicoletti, G. (1989), “Ageing populations; economic effects and implications for public finance”, Department of Economics and Statistics Working Paper No. 61, OECD, Paris.
Holzmann, R. (1997), ‘On economic benefits and fiscal requirements of moving from unfunded to funded pensions’, Forschungsbericht 9702, University of Saarland
IMF (1996), “World Economic Outlook, May 1996”, International Monetary Fund, Washington D.C.
James, E. (1993), “Income security in old age; conceptual background and major issues”, Working Paper No. WPS 977, The World Bank, Washington D.C.
Johnson, P. (1992), “The taxation of occupational and private pensions in Europe; the theory and the practice”, forthcoming in ed. J. Mortensen, “The Future of Pensions in the European Community”, published by Brassey’s, London, for the Centre for European Policy Studies, Brussels.
Lannoo, K. (1996), ‘The draft Pension Funds Directive and the financing of pensions in the EU’, The Geneva Papers, 78, 114-26.
Merton, R. C. (1983), “On the role of social security as a means for efficient risk-sharing in an economy where human capital is not tradeable”, in eds. Bodie, Z., Shoven, J. and Wise, D., “Financial aspects of the US pension system”, University of Chicago Press.
OECD (1993), ‘Pension liabilities in the seven major industrial countries’, Organisation for Economic Co-operation and Development, Paris.
Roseveare, D., Leibfritz, W., Fore, D. and Wurzel, E. (1996), ‘Ageing populations, pension systems and government budgets: simulation for 20 OECD countries’, Economics Department Working Paper No. 168, OECD, Paris.
Samuelson, P.A. (1958). “An exact consumption-loan model of interest with or without the social contrivance of money”. Journal of Political Economy, 66, 467-82.
Samwick, A.A. and Skinner, J. (1993). “How will defined contribution pension plans affect retirement income?”. Mimeo. Stanford University.
Schieber, S.J. and Shoven, J. (1994). “The consequences of population ageing on private pension fund saving and asset markets”. NBER Working Paper No. 4665.
Schlesinger, H. (1985), “Die Finanzierung der sozialen Sicherheitssysteme bei geringerem wirtschaftlichem Wachstum und schrumpfender Bevölkerung”, from Band XXVIII “Veränderungen in der Arbeitswelt und soziale Sicherung”, Verlag Chmielorz, Wiesbaden.
Smalhout, J. (1996). “The uncertain retirement” Irwin, Burr Ridge, Illinois.
Thompson, L. H. (1992), “Social security surpluses”, New Palgrave Dictionary of Money and Finance, MacMillan, London.
World Bank (1994), “Solving the old age crisis”, The World Bank, Washington D.C.
Table 1: Pension indicators for the G-7 countries
(1) The elderly dependency ratio is the ratio of those aged 65 and over to those who are 15-64. Source: Bos et al, World Bank.
(2) Gross pension liabilities are the discounted present value of future pension expenditures. Source: Roseveare et al, OECD Working Paper No. 168.
(3) Net pension liabilities deduct the value of contributions from gross liabilities, and indicate the scale of adjustment to contributions required to pay
benefits at the current rate. Source: Roseveare et al (op. cit.)
(4) Source: Roseveare et al (op. cit.)
(5) Source: E P Davis, “Can pension systems cope?” (op. cit.)
(6) Source: National flow-of-funds balance sheet data
(7)-(9) Source: E P Davis “Pension funds” (op. cit.) The growth rate of average earnings proxies the return to pay-as-you-go, which may be compared
to the real return on equities and the real return on a balanced portfolio (50% bonds, 50% equities)
 The author is a senior economist at the Bank of England who is currently on secondment to the European Central Bank. He is also a Research Associate of the LSE Financial Markets Group, Associate Fellow of the Royal Institute of International Affairs and Research Fellow of the Pensions Institute at Birkbeck College, London. Views expressed are those of the author and not necessarily those of the institutions to which he is affiliated. The article draws on the author's book "Pension Funds, Retirement Income Security and Capital Markets", published in 1995 by Oxford University Press.
 Of course, in the next century, funded schemes would need to dissave as the population ages, although if pay-as-you-go obligations have been reduced, the effect on aggregate saving will be mitigated.
 See Feldstein (1977), (1995a).
 Funding may also increase overall economic efficiency and flexibility more generally by reducing the conflict between labour and capital, as with funding workers do not only focus their interest on high wages and safe employment. This may, for example, help wage moderation and reduce demand for job security provisions, as they would be seen as benefiting future incomes from capital in retirement.
 As shown in Table 1, the scope of funding, and correspondingly the size of funded sectors, differs sharply across the G-7 countries.
 On related issues, see Franco (1996).
 In this context Feldstein (1995b) shows that the conditions for funding to improve welfare even abstracting from demographics and distortions to labour markets are quite likely to hold. These conditions are: that the return on capital exceeds economic growth (so the return to funding exceeds that to pay as you go); that the return on capital exceeds the rate of time preference (the capital intensity of the economy is below the welfare-maximising level); and the rate of growth of the economy is positive (so there is a gain in extra retirement income which more than offsets the (given) costs of the transition).
 The Swedish ATP fund is an exception, being invested largely in private-sector debt instruments.
 The impact of institutional investors such as pension funds on the capital market is discussed in Davis (1996b).
 There are numerous barriers to international investment of private pension funds in the EU, usually imposed for ‘prudential’ reasons (see Davis (1995)). Lannoo (1996) discusses recent action by the European Union in this field.
 This is because younger workers can accept temporary underfunding of their pension rights during periods of market volatility while older workers continue to have a stable replacement rate.
 Adverse selection is reduced because a company based defined contribution scheme can offer an insurance company a pool of risks, whereas when an individual seeks an annuity instead of a lump sum, the company will fear that they have superior knowledge regarding their own life expectancy.
 In the Netherlands, for example, administrative costs of state pensions are 1% of contributions; company pensions 7% and personal pensions 24% (Besseling and Zeeuw (1993)). Diamond (1993) notes that US social security costs are between 3 and 12 times less than private pensions, partly owing to the natural monopoly in collection of social security contributions.
 Typical restrictions in company based defined benefit funds are that “deferred pensions are not indexed, workers’ rights are not vested for several years and possibility to transfer funds out of a scheme is limited. All of these may be overcome by regulation - as in countries such as the UK - at a cost of making defined benefit funds less attractive to firms.
 Inter alia to combat this, the Netherlands is now introducing a reform to make average-salary based calculations standard for defined-benefit funds in that country. With an average-salary base (upgraded in line with inflation), pensions accrue smoothly over the working life of the individual.
 Andrews (1993) points to the US railroad fund as an example of the plight a reserve funded scheme get into in a declining industry.
 In more detail, the 'wind‑up' definition of liabilities, the 'solvency' level at which the firm can meet all its current obligations, is known as the accumulated benefit obligation (ABO). Advance provision for salary increases up to retirement, as is normal in a final salary scheme, gives the projected benefit obligation (PBO). The indexed benefit obligation (IBO) assumes indexation after retirement (See Bodie (1991) for a further discussion of these concepts.)
 The facility with which funds of declining industries in the UK funded on a Projected Benefit basis (such as coal mining and railways) coped with maturity are a case in point.