International accounting standards have served pensions accounting well, but it is time for an update, argues Andrew Lennard

If you asked five pensions professionals their views on current accounting standards, you would easily generate a list of 20 issues. While FRS17 in the UK and IAS19 internationally have done much to increase awareness of pensions and their impact on companies’ financial health, no one would claim these standards are perfect.

Some of the points can be addressed through relatively small amendments and the Accounting Standards Board has tackled some of the disclosure issues. But a thorough overhaul is necessary. Given that financial reporting is now international, that can only be carried out by the International Accounting Standards Board, and it is indeed committed to a fundamental review of pensions

To provoke discussion and provide ideas ahead of that review, the ASB has taken a fresh look at pensions accounting. The project needed to be international and so the ASB has worked with European colleagues under the Pro-active Accounting Activities in Europe (PAAinE) initiative. It was essential to adopt a principles-based approach that will work anywhere in the world. One result of this is that the principles can be applied without relying on the distinction between defined benefit (DB) and defined contribution (DC) plans, which is particularly troublesome to apply to the hybrid plans that are becoming increasingly common.

If an employee has provided services and so is entitled to a pension, there must be a liability. It is more difficult to pin down precisely what that liability is. Applying the accepted accounting definitions, there must be a present obligation - which may be either legal or constructive. There are many uncertainties surrounding that obligation: whether unvested benefits will vest in due course, for instance, and the longevity of the employees (and their spouses). But most of these uncertainties affect the amount of the obligation rather than whether a liability actually exists. Accordingly the recorded liability should include these factors in the estimate of its size. However, where there is simply an expectation that further benefits will be offered but genuine discretion as to whether they will be paid, there is no present obligation and hence no liability.

But how does that principle apply where the amount of pensions is related to final salaries? Does an employer have a present obligation to increase salaries, and hence to pay the higher pensions that such increases will lead to? Views on this may vary depending on whether the increases in question relate merely to inflation or also include promotion, and whether the liability is seen as the aggregate of amounts owed to individuals or as a liability to the workforce as a whole. It may be possible to freeze the salary of an individual employee, but impossible not to increase salaries in aggregate.

Current pensions standards deal with future salary increases similarly to longevity - an estimate of the effect is included in the reported liability. But many (including a majority of ASB members) disagree with that: they believe that there is no compulsion to increase salaries and so the present liability should be restricted to that which relates to current salaries.

A change on this point would have two effects: it would reduce the amount of the reported liability and hence any reported deficit in the plan. The second effect is that, when salaries are increased, the increase in the amount of the pension already earned would have to be recognised as an expense in the period in which the salary is increased. Under current standards, the salary of a 58-year-old can be increased by 50%, thus increasing the value of 30 years’ service by a similar amount and the effect will simply be reported as an actuarial gain or loss. If the standards were to change, it might provide an incentive to offer bonuses or other non-pensionable
payments to employees rather than increases in pensionable salaries.


If there is a liability, whose liability is it? In many cases it may be that of an employer, or it may be passed to a third party pension provider. Or, as is commonly the case, it may be passed to a company pension plan. In the last case the liability is that of the plan, and the employer need account only for the effect of any guarantee it has given - typically the amount by which the liability exceeds the plan’s assets.

But this assumes that the plan is genuinely independent of the employer. Today’s pensions standards do not require much heart searching on this point: they give a blanket exemption from consolidation requirements. But the logic of this is far from clear, so we are suggesting that the normal consolidation rules should apply - that is, where the plan is controlled by the employer it should be consolidated.

In many cases - including the majority of UK plans - the control test will not be met and so consolidation will not be required. The key consideration is that the plan is controlled by trustees who are obliged to act in the interests of plan members and cannot accede to the employer’s wishes unless it is in members’ interests. But the suggestion might require pension plans in other countries - including those relating to UK employers - to be consolidated, and so their total liabilities and assets would appear on the employer’s consolidated balance sheet.

One of the improvements made by current pensions standards is that the liability (like the assets) is no longer measured on an actuarial basis. An actuarial perspective is useful for assessing the extent to which liabilities are funded, but they do not say how big the liability really is.

In theory the liability to pay pensions could be stated at a buyout amount, but this would often overstate its real burden. Instead the cash flows that will be incurred in settling the liability should be discounted. But it is not clear why current standards use a corporate bond rate. It seems to have something to do with risk - but the risks are so large and variable that information on them is best provided by supplementary disclosure and so, we suggest, the liability should be discounted at a risk-free rate. This would increase, perhaps significantly, the amount of the reported pension liability and any deficit.


Some changes in the way in which income and expenses relating to pensions are reported are also suggested. For example, the finance cost of pensions is reduced under current accounting standards by the expected return on assets. But nowhere else in financial reporting do we report what was expected to happen: we report what actually did occur. So we propose that the total actual return (dividends and value changes) should be reported as a financing item. This will inevitably be more volatile than the expected return: expected returns are never negative, but actual returns (as in recent weeks) may be. The financing section would also include the effect of changes in the discount rate. These two changes will reduce the amounts that are reported as ‘actuarial gains and losses’.

The financial reports of pension plans have not received much attention from accounting standard setters recently. The international standard is old and allows many options. We suggest that a fresh look is appropriate and, in particular, that the balance sheet of a pension plan should include the liability to pay future pensions and that the extent to which the plan is dependent on the employer’s guarantee should be highlighted. The financial statements should also include a discussion of the relationship with the employer and show the plan’s transactions with it. These ideas will greatly enhance transparency but are likely to be controversial.

The paper ‘The Financial Reporting of Pensions’ may be obtained from Comments are requested by 14 July. A report taking the responses into account will be submitted to the IASB.