Determinants of government underfunded public pension liabilities in the OECD

Abstract

Underfunded government liabilities for public pensions constitutes a major expenditure in the management of social programmes in many countries, but to date has not attracted much attention from accountants as it does not easily fit within an accrual-based accounting system. This paper discusses major measurement problems associated with this liability and then examines determinants of variations in projected flow-based funding patterns among OECD governments. Alternative ‘behavioural persistence’ and ‘regression to the mean’ hypotheses about the determinants of underfunding practices are formulated and tested using an OECD data set describing the financial and socio-economic characteristics of government-sponsored public pension systems in these countries. Consistent with the behavioural persistence hypothesis, cross-sectional variations are found to be associated with the funding ratio and the rate of taxation required to keep government debt constant. Variations in underfunding practices across the sample are also sensitive to cultural differences in attitude towards public pension accountability between continental European and Anglo-American countries.

ACKNOWLEDGEMENTS

This research was funded by the International Consortium of Government Financial Management. Earlier versions of this paper benefited from comments received on earlier versions by Brian Craven, Dick Feenstra, Larry Kotlikoff, Ken Peasnell, Michael Mumford, Dieter Ordelheide, and participants at the 1996 EAA Conference, the 1996 EIASM Workshop on Accounting and Economics, the 1997 EAA Conference, the UK House of Commons Seminar on Unfunded Pension Liabilities in the European Union and the AOS Internationalising International Accounting Research Conference, Hong Kong.

Notes

¹The ramifications of ageing populations on the financial viability of PAYG public pension funds has been forecasted in the financial press (The Spectator, May 1995), by the European Union (e.g. Franco and Munzi, 1996), international economic bodies (e.g. OECD, 1996; World Bank, 1994; IMF, 1996) and the UK's House of Commons Social Security Committee (1996). Recently the president of the European Union alerted heads of European governments to the dangers of these practices to fiscal stability.

²According to the corporate financial perspective, corporate employer sponsors of defined benefit pension funds treat pension assets and liabilities as just another form of financial capital raising. Prior US-based empirical studies which incorporate this assumption into their analysis include Stone et al. (1987). Pesando (1992) questions the validity of this perspective because it implicitly assumes that corporate sponsor shareholders ‘own’ the surplus and that corporate pension plans should hold the most tax-disadvantaged assets. Ghicas and Tinker (1989) dispute the claim that employer shareholders own surplus.

³GASB 5 includes requirements for disclosure of pension information by US public sector pension funds and state and local government employers (GASB, 1984). Stone et al. (1987) examine the determinants of voluntary accounting policy choice. GASB 27 requires calculation of the projected benefit obligation by US state and local government employers (GASB, 1994). However, GASB 27 does not apply to the US Federal government's own employees' pension funds or to old-age security provision in general.

⁴Mitchell and Smith (1994) is the only study which recognizes the endogenous relation of required to actual pension contributions.

⁵The GASB (1984: 22) claimed that financial statements need to report on interperiod equity as a key element in demonstrating accountability in the governmental context, implicitly assuming that pension underfunding is undesirable (Marks et al., 1988: 157).

⁶Prior to the OECD (1996) study, there was little standardized, comparable evidence available about public pension underfunding practices across the OECD.

⁷This form of analysis is necessary to determine the sensitivity of public pension underfunding practices to the likely effects of demographic transition.

⁸This analysis provides empirical evidence which bears upon debates in recent public policy literature about whether cross-sectional variations exist among countries between the extent of underfunded public pension liabilities and other demographic-sensitive aspects of government public policy (e.g. World Bank, 1994; Davis, 1995, 1996; OECD, 1996).

⁹Mitchell and Smith (1994) provide similar empirical evidence on the analogous debate regarding cross-sectional variation in USA public sector underfunding.

¹⁰Individual behaviour related to public pensions is often at odds with that assumed by neoclassical economics (Weiss, 1991). For instance, people are prepared to pay additional taxes to secure the future of their children, but not for the elderly (Richter, 1992).

¹¹Empirical inspection of the underlying data set (e.g. plotting the predicted values against the residuals; skewness and kurtosis-coefficient statistics for the residuals) suggests that it is robust to the assumptions of linearity, normality and independence.

¹²Unlike the simultaneous equations model developed by Mitchell and Smith (1994), formulas for determining the generosity of pension benefit levels in public pension systems may or may not be linked to average national wages.

¹³These primary results are consistent with those obtained by Mitchell and Smith (1994) in examining state and local US government pension underfunding practices, although their behavioural persistence hypothesis is also consistent with both actual and required spending.