Sunday, January 18, 2009

Pension Expected Return Assumptions - Hitting Back


Current pension accounting allows pension plans to use an estimate of expected future returns instead of actual returns to compute periodic pension costs. The logic behind this accounting is that, in the long run, a combination of company resources and investment earnings must equal plan obligations. Pension plan liabilities are long lived and therefore some stable long run estimate of corresponding asset returns should be used to determine the periodic cost to the business of meeting plan obligations. Accounting further recognizes that annual differences between actual and expected returns are inevitable. These are accumulated and amortized into periodic cost over time.

There is a natural asymmetry to how this accounting regime is viewed depending on market performance. In bull markets, most seem satisfied in the conservatism reflected by periodic pension costs being higher than current earnings might suggest. Yet in bear markets, critics become vocal about the potential manipulation and overstatement of income as a result of using positive expected returns when market returns are negative. The wider the gap the more vocal the critics and the more concerned the actuaries and auditors who have a professional tendency towards asymmetrical conservatism. A recent article reflects this view.


“That 8 percent annual return on investment you and your pension fund manager were banking on is now looking almost as optimistic as Madoff’s magic 12 percent, as deleveraging and deflation bite. With extremely low or negative interest rates and everyone from consumers to banks trying to shed debt and assets at the same time, what seemed like reasonable projections for a mixed portfolio of stocks, bonds and other assets are now substantially too high.”
It is a well accepted fiduciary notion that short term investment results are very noisy and carry little useful information to guide long term investment strategies. Yet, after market drops like 2008, misguided critics insist that plan sponsors should be lowering their expected return assumptions to accommodate what ever change in strategic assumption they would argue has been revealed by the market drop. In this case, presumably, overleveraging was explicitly factored into past return forecasts rather than being a source of an unexpected positive variance from past forecasts.

In fact, long term expected return assumptions as defined by FASB, properly based on long horizon strategic factors, should not be subject to frequent marginal shifts given; the likely estimation error in any such forecast, the infrequency of major structural market changes and the provisions of accounting to amortize actual to forecast return differences overt time. Further, unless the market principle of mean reversion has been suspended, it is fallacious to assume that forward expected return forecasts should be lower after a market collapse. Historically, periods of heavy losses have been followed by periods with above average returns. Diamond Hill calculated that if the S&P 500 only returned to its April 2000 level in the year 2016, that appreciation, coupled with the dividend yield, would result in an annualized total return of over 12%.

While we are not expressing a view as to whether the average pension plan’s expected returns accurately reflect future returns, it is illogical to presume that the passage of the last year has provided plan sponsors with significant additional insight on strategic factors that were embedded in their long term forecast assumptions. Though it is understandable from a risk aversion perspective, it is also illogical that pension actuaries and auditors should bias plan sponsors to lower long term expected returns after the market has seen its worst decline in 50 years. Given their biases, pension plan fiduciaries should not unduly allow their advisors to tactically influence their strategic pension accounting assumptions.

Links to this post:

Create a Link

<< Home