In Search of the Liability Asset?
The January/February 2015 issue of the Financial Analysts Journal (FAJ) republished a 1988 article titled as above, but without the question mark. The authors, Richard Bookstaber and Jeremy Gold, wrote as Principals at Morgan Stanley. They argued that defined benefit pension funds should hold some stocks, not just bonds. They also argued that it would make sense to hold a derivative hedge against unfavorable changes in the surplus of assets over pension liabilities. These messages seem to be worth updating in light of our current understanding.
Before 1974 in the United States, pensions were based on private contracts between employers and their employees. These were promises that retired workers would receive predefined dollar payments. The employer’s contribution to any pension benefits, subject to qualification by the Internal Revenue Service, was tax deductible immediately, while if the pension were funded in advance, only the income received by a retired employee much later was taxed. In 1974, Congress enacted ERISA for greater protection of worker benefits. Now companies became fiduciaries subject to oversight by the US Labor Department. Insurance premiums were paid on the pension plan’s behalf to the Pension Benefit Guarantee Corporation.
At the time the article was written, the prevailing investment model focused on asset management. To their merit, the authors effectively argued that the focus should be broadened to one of managing pension surplus, influenced both by asset investment returns and by changes in liabilities resulting from changes in interest rates used to calculate their present values and from increases in worker compensation over time from both productivity gains and monetary inflation.
What is Especially Good
The article contrasted 1) pension obligations known with a high degree of certainty based on employee services and wages already in place, with 2) the less certain benefits likely to be received with further seniority and inflation in wages and living costs. In so doing, it departed from objective financial accounting as a basis for planning, including in the balance sheet both objective and subjective elements. The authors asserted that while it made sense to finance current contractually locked-in obligations with safe bonds, increases in future benefits would increase with growing productivity and with monetary inflation, and in the long run would be better financed by owning stocks. The locked-in obligations could be assured if invested in bonds, while the long-term increase in obligations would be hedged at least somewhat by holding stocks.
The authors pointed out that quantifying risks affecting both stocks and prospective obligations, together with their correlation, could be used as a foundation for reducing the risks to surplus even further, potentially by constructing either a dynamic or option-based hedge security.
What Should Be Updated
Irwin Tepper had shown in the Journal of Finance in 1981 that, subject to simplifying assumptions, if the employer and employee were considered jointly, tax considerations indicate that the employer should fully fund the future benefits (discounted to present values using a safe interest rate) and invest the pension fund entirely in bonds. Even though reality is more complicated, this tax effect appears too important to be ignored. Yet Bookstaber and Gold did ignore it in making their recommendation for holding stocks within the pension fund. The fact that most pension funds are managed with a heavy stock component is not a sufficient reason to discard at least some normative tilt toward bonds relative to the picture drawn by the authors.
Even more fundamentally, no allowance was made for differences in financial circumstance. An underfunded pension with no prospects for further contributions from a near-bankrupt employer surely should be more conservatively managed than one over-funded with ample backup financing capability from a prosperous employer. The article came close by discussing surplus, which I might call discretionary wealth, but did not clarify its role in determining appropriate risk aversion.
The article’s discussion of a hedging program for surplus strikes one as unrealistic. If it is intended to hedge long-term as well as short-term risks to surplus, in the absence of long-term option securities it must be realized through a series of dynamic hedging adjustments or by rolling over short-term options. But compounding a sequence of hedges individually intended to limit downside risk will not have the intended long-run effect. This is obvious by the central limit theorem, but in any case is readily verified by simulation. For example, a sequence of puts overlaid on bonds will eventually have a return probability distribution profile much like a mixture of cash and bonds, less fees. And in the case of custom hedges for surplus, the fees will be considerable.
Finally, we now know that the major problem with defined benefit pension plans was not that they were poorly invested but that they were underfunded. That is, the temptation for pension actuaries and consultants to allow improperly high discount rates for future obligations caused many corporations, states and municipalities to severely underestimate future needs. This is not much different from the problems many of us have in personal saving, so we should not be dismissive. Nevertheless, the search for better hedges was in some sense looking in the wrong place. Consequently, there has been a strong tendency to shift the burden, not just of risk-taking, but also of prudence, to the employee in the form of defined-contribution plans.