Pension promises: Hybrid plan accounting
Just like waiting for a bus, you can stand around waiting all day for an accounting research proposal to land in your inbox and all of a sudden you are deluged with them. Because, right now, staff at the International Accounting Standards Board (IASB) are sketching out an approach to tackle so-called hybrid pension plan accounting. The plan is for board members to evaluate the research and decide whether to launch a full-blown standard-setting project.
Now, as luck would have it, the European Financial Reporting Advisory Group (EFRAG), a technical expert body that advises the EU on accounting matters, has just released a discussion paper to deal with pension promises where the final benefit is the higher of a return on plan assets or a guarantee.
The EFRAG paper explores three alternative accounting approaches to deal with such pension benefits. One important feature of the pension promises under consideration is that the plan sponsor must hold the assets referred to in the plan.
To deal with the accounting challenges posed by these plans, EFRAG proposes three possible accounting alternatives: capped asset return; fair-value-based; or fulfilment value.
One of the shortcomings of the IASB’s pensions accounting rulebook, International Accounting Standard 19, Employee Benefits, is its binary approach to classification and measurement of pension promises. The standard dates back to a simpler time when it was easy to sum up the pensions landscape; benefit promises were either defined-benefit (DB) plans or defined-contribution (DC) plans, with no grey area.
IAS 19 requires pension scheme sponsors to decide whether the pension promise they want to account for is a DB promise or a DC promise. If it is a DC promise, they simply book an expense and move on. This is because IAS 19 treats a DC promise as a simple pledge to pay a pension contribution with no further risk.
However, If the sponsor continues to bear some sort of risk, perhaps because the promise has an embedded investment guarantee, then IAS 19 treats the promise as a DB promise. IAS 19 tells scheme sponsors that they must then apply the projected unit credit method to account for their obligation. This means that once they have netted off any plan assets, the sponsor must book the net asset or liability on their balance sheet.
Historically, this method has worked quite well with plain vanilla DB promises. However, it tends to fall apart with modern risk transfer, or hybrid, plans. Since the turn of the millennium, employers have attempted to reduce their exposure to potentially troublesome DB schemes by designing new benefit promises to reduce their overall balance sheet and cashflow risk.
One way of achieving this process of risk transfer is to offer workers a pension promise where they have can have the higher of an investment return or a minimum performance guarantee or floor.
In simple terms, the IAS 19 projected unit credit method struggles to measure this type of pension promise reliably. This is because it requires preparers to project the total liability forward at an asset-linked rate of return, but forces them to discount back to arrive at a net present value at an AA corporate bond rate.
Commentators have observed that the resulting number is total nonsense. Another way of looking at it is to say that the variability of the plan’s investment performance is reflected in its cashflows but not in the discount rate.
This is why the IASB and the EFRAG have ended up running two broadly similar, overlapping pensions projects. The most likely output from the IASB staff’s work is an analysis to help the board decide whether or not to undertake a full standard-setting project to amend IAS 19. So what exactly have the staff been up to?
During last December’s board meeting, they explained that the project’s scope is defined in terms of types of benefit, not types of plan. That is possibly enough to get them past the accusation that the project favours any one jurisdiction and its plan designs over another.
And, just like the EFRAG research effort, the IASB has made the unsurprising admission that the practice of projecting forward with an asset return and discounting back at a more conservative corporate bond yield produces a measurement mismatch – not to mention a number that has little credibility in the marketplace.
Moreover, at the December 2018 meeting of the IASB’s Accounting Standards Advisory Forum, it emerged that the staff’s thinking is focused on pension promises linked to a specific pool of assets – whether held by the plan sponsor, the plan itself, or neither.
Equally unsurprisingly, the IASB has landed on a possible approach that caps the rate of return on assets at the level of the discount rate. All of which makes it fairly easy to see how feedback from constituents in Europe to the EFRAG discussion paper could well inform the IASB’s thinking.
Among the jurisdictions – together with the US, Canada and Japan – that might have an interest in the outcome of the IASB and EFRAG research work is Germany.
The approach there, says Mercer’s chief actuary Thomas Hagemann, is to adopt what he calls “a rather simple and economically appropriate solution for such plans”. This approach involves German plan sponsors determining the defined benefit obligation (DBO) based on the minimum guaranteed return and, in the first step, ignoring the past performance of any underlying assets.
“If the fair value of the assets is higher,” Hagemann says, “we set the DBO to the amount of the fair value of the corresponding asset. This approach is well accepted in the market as it is easy to follow and straightforward. It is also economically appropriate, as the employee’s right to receive the higher of the return on plan assets and the minimum guaranteed return is reflected in the DBO.”
As for the three approaches in the EFRAG discussion paper, Hagemann says the capped asset return approach will lead to similar results to current German practice where the expected return on the plan assets is higher than the discount rate. In fact, he explains, it could even produce more accurate measurement because workforce turnover affecting unvested entitlements, or deaths without beneficiaries, are reflected in the DBO.
However, Hagemann adds, the approach is economically inappropriate where the expected return on the plan assets is lower than the discount rate. “In this case, the DBO is lower than the fair value of the assets, even if there will be no economic benefit for the employer. Only where the assets are plan assets will the asset ceiling lead to economically appropriate accounting.”
For these reasons, the Mercer actuary argues that the pension obligation under the capped asset return approach should be determined always by replacing the expected return with the discount rate, instead of simply capping the expected return at the discount rate.
All in all, he concludes, the capped asset return approach, with modifications, is the most viable approach of the three. The two other candidates, he warns, at least in Germany, would nullify the advantages of such plans and could no longer be applied with reasonable effort.