Workshop on “Accounting for Implicit Liabilities”
OECD Paris, 4 June 2004
Main issues for discussion
Round 1: How to define the boundaries of implicit liabilities?
1.The Secretariat suggests organising the exchange of views around three main points dealing with, respectively, how to define the boundaries of measured implicit liabilities, how to address implementation issues for given agreed boundaries and, finally, how to integrate wider concepts of implicit liabilities in economic policy-making.
2.The current System of National Accounts (SNA93) only recognises pension obligations as liabilities when they are funded. This leads to noticeable differences across countries in the recording of these liabilities and many accounting anomalies. In some OECD countries, private sector standards are already more advanced in the recognition of such pension liabilities. A background note prepared by the IMF and the OECD Statistics Directorate describes these problems and put forward the proposal by the National Accounts statisticians to reform the current system.[1] Basically, the national accountants’ proposal leads to a wider concept of, so-called, ‘constructive liabilities’, i.e. obligations that result from an established pattern of past practice and have created a valid expectation that they will be fulfilled in the future. This definition, which is in line with the Generally Agreed Accounting Principles (GAAP), implies that the recognition of liabilities does not depend on whether a pension schemeis funded or not.
3.By contrast, public budget accountants tend to limit the recognition of liabilities strictly to those covered by a contractual relation, say, between employer and employees, which can not be changed retroactively. From their point of view, the latter is the only definition consistent with a government financial reporting on the basis of their interpretation of accrual accounting. A note prepared by the OECD Public Governance & Territorial Development Directorate develops this argument.[2] Along these lines, the experience in some OECD countries is that only the liabilities associated with pension system of public employees, for which there is a contractual relation, are included in the definition of public debt.
What are the pros and cons of debt definitions based on the concept of ‘constructive obligations’ vs. the concept of contractual obligations?What is the preferred concept, in particular for use in analysis of public finance?
Round 2: What are the implementation issues?
4.New accounting rules need to be implemented in a uniform way across countries and over time. In particular, there a risk that estimates would be too dependent on uncertain assumptions, such as for longevity or discount rates. While in the private sector accounting assumptions/practices are disciplined to some extent by shareholders and market valuations, a similar discipline would have to apply in the case of statistical offices for computing public liabilities.
5.In order to allow national accountants to compute implicit debts at the country level, each pension scheme (in particular public schemes) should be able to calculate their implicit debts. This implies a choice on mortality tables and pension benefits depending on each institutional framework. In doing this, public institutions could follow the example of private sector standards (cf. presentation on the UK-FRS17).
6.More broadly, rules have to be defined in order to accommodate changes in the value of the implicit debt driven by changes in longevity, as well as changes in the pension rules (e.g. eligibility, replacement rates). Finally, the incorporation of actuarial accounting will automatically increase the implicit contributions to social security. This may have implications for the computation of labour costs.
When defining implicit liabilities, how should exogenous assumptions be set so as to ensure comparability of statistics across time and countries and preserve the integrity of National Accounts?
How should changes in assessed liabilities, as a result of changed exogenous assumptions (e.g. discount rates, longevity) be dealt with? In particular, how should stock/flow consistency be ensured?
Should labour costs reflect implicit social security contributions?
Round 3: How should wider categories of implicit liabilities be considered in economic policy-making?
7.Wherever the “boundaries” of a rigorous definition are set, there will be factors that will not be well enough defined to fit within such boundaries (in the sense that they would weaken the analytical usefulness of the relevant series) but which may nonetheless have important long-term implications for policies. Liabilities under this heading are those that can change as a result of a change in policy. The clearest examples of this come from general pension schemes, the liabilities of which depend on parameters that governments can change, and indeed have changed: contribution rates; benefit rates; years of work required to qualify for a full pension; period over which pension levels are calculated; and indexation. As well, there are factors less easy to define like increases in ageing-related health expenditures and the fact that, in practice, the government is the social insurer of last resort.
What is the best way of taking account of such factors, in particular with an eye to achieving cross-country comparison?
1
[1].See background note on “Accounting for Implicit Pension Liabilities: Proposals from National Accountants for a Change of SNA93/ESA95”, OECD/STD, May 2004.
[2].See background note on “The recognition of pension liabilities in financial reporting”, OECD/GOV, May 2004.