Accounting: The inflation challenge
Stephen Bouvier looks at the complications pension plan sponsors face when calculating the pension liability of their schemes under IAS 19
Inflation plays a central but surprisingly overlooked role when sponsors calculate a pension liability under International Accounting Standard 19 (IAS 19). The overall principle is that they must make an assumption about what they expect pension indexation to be in future. So in the UK where pensions are indexed to, for example, the inflation as measured by the retail prices index (RPI), they will need to make an assumption for what inflation, as measured by the RPI, will be in future.
So what is inflation? The simple answer is that it is a surprisingly wide spread of measures and outcomes that turns in no small part on a scheme’s rules. “There is more than one index that can be used to measure it,” says Alex Waite, a UK-based partner with Lane Clark Peacock. “Many pension schemes in the UK have linked pensions to the retail prices index. That tends to give higher indexes than the consumer prices index [CPI]. It has also led to a legal lottery where two similar schemes can be giving very different increases.
“There may actually be judgments that trustees need to be making about the level of protection that should be given. The wording of schemes’ rules is not always clear on this point. We’ve seen instances where one set of lawyers will say it is RPI and another will say it is CPI, which means you need to get a senior barrister’s opinion. It might sound picky, but the difference between the two is about 1% a year. That can take 20% off your liability, and for schemes with a deficit the shift from RPI to CPI could eliminate a 20% deficit.”
But even with the starting point for building the inflation component of the defined benefit (DB) liability established, it comes as perhaps no surprise to learn that application of IAS 19’s requirements is equally diverse. The starting point under the standard is paragraph 75. This says that assumptions should be unbiased and mutually compatible.
Paragraph 76 of the standard directs sponsors to use a best estimate. On the face of it, that could be central bank rates or some other rate which the entity says is a best estimate. Equally, paragraph 80 of IAS 19 expects them to base financial assumptions on market expectations. And inflation is, after all, a financial assumption – as 76(b)(ii) signals. It is this, some practitioners argue, that pushes them into thinking that market expectations are driven by yields on inflation-linked bonds.
And there is yet more complexity. In paragraph 79, IAS 19 suggests that where there is a deep market in inflation-linked bonds – at least, arguably, there is in the UK – preparers must use real discount rates. And real rates, the thinking goes, are the same as the difference between nominal discount rates and break-even inflation, which effectively points to the use of break-even inflation.
This compares with practice in the US where, to the extent that they make an assumption, practitioners there tend to look at long-term inflation. There is yet more diversity in mainland Europe where it is possible to observe German companies, say, plump for the European Central Bank (ECB) target of 2% on the basis that it is a long-term target for future price growth.
But in the UK, the landscape is somewhat more complex. First, inflation is a particularly relevant assumption in the UK because of inflation-linked benefits. UK actuaries look at the difference in yield between fixed and index-linked government bonds. The main principle is that the difference in yields represents the rate of inflation required to give the same overall return. This is known as the ‘break-even’ inflation rate (BEIR) and will vary from time to time.
“Arguably, that reflects market expectations,” says Simon Robinson, a consultant actuary at Aon Hewitt. “But if you step back from that, investors are prepared to pay a premium for index-linked investments as they offer protection against future inflation volatility. For example, if break-even inflation was 3%, if there is an inflation risk premium in there, the market might be expecting inflation of, say, 2.7%.
Tim Marklew, a partner in LCP’s UK practice, shares the view that the position is complex. “IAS 19 requires that companies set their assumptions consistent with current market conditions. So if, for example, investors generally expect high inflation, inflation-linked bonds should cost more, break-even inflation would, as a result, be high, and the company should assume higher inflation to be consistent with market prices.
“But there are other reasons why index-linked bonds may be relatively expensive. Investors may generally prefer index-linked bonds because of the inflation protection that they give, so investors may prefer inflation-linked bonds to fixed-interest bonds, even if they are expecting inflation to be a bit less than it would need to be for them to break even.
“For that reason, it’s entirely reasonable – and common market practice – for companies to make an adjustment to break-even inflation to reflect this – an ‘inflation-risk premium’. That gives a lower inflation assumption. It’s a matter of some debate of how big that market premium should be.”
Meanwhile, Robinson says he can see how all of these possible rationales fit into IAS 19. “You are supposed to use best estimates and market rates, and it is not clear to me which has priority, market rates or best estimate. In the UK, we use market rates, often with a risk-premium adjustment.
“So if you then look at paragraph 76, actuarial assumptions are an entity’s best estimate. It goes on to say that they include inflation. But, there again, IAS 19 also talks in paragraph 80 about market expectations.
“And paragraph 79 says that an entity determines the discount rate and other financial assumptions in nominal terms and that best estimates in real terms are more reliable. For example, in a hyper-inflation economy, or where the benefit is index-linked and there is a deep market in currency and term.
“But the key point is this: I don’t think IAS 19 is clear on this point and is open to interpretation. And certainly in practice. I think there is divergence in practice, with different companies setting inflation in various ways.”
Nonetheless, there remains the concern that the inflation-risk premium is fairly arbitrary. If you start with RPI of 3.2, you then have the challenge of finding evidence for that next step of knocking off 30bps. Critics argue that such objective market evidence does not exist.
Finally, notes Robinson, there is the rarely considered mismatch within IAS 19 over the way inflation is accounted for on both the asset and liability sides. “If the plan holds some ‘matching’ assets, which is increasingly the case, implicit in the price of the assets such as inflation-linked bonds is an assumption about future inflation,” he says.
“Anyone placing a value on those bonds is implicitly assuming or accepting that future level of inflation. The question is then whether the assumption implicit in the asset valuation is compatible with the inflation assumption used in the defined benefit obligation.
“Many entities with hedged investment strategies don’t allow for this because their investment strategy is hedged against inflation. They will argue it is appropriate to use effectively the same rate for their liabilities as for their assets.”