As pension funds in the United States have grown in assets to $7.66 trillion in 2018, according to Pensions & Investments, their investment decisions have attracted a lot of attention from both academics and practitioners. Performance evaluation is a daunting task because it is typically done within an asset-liability framework and driven by many actuarial assumptions regarding benefit payouts, contributions and accruals of liabilities as well as investment assumptions regarding individual investments and their correlation structures. As a result, the significant role of hidden assumptions in conventional models makes evaluation results ambiguous and often difficult to interpret.
In our research, Parry Wang from the $242.1 billion California State Teachers' Retirement System, West Sacramento, and I introduced a simple intuitive methodology designed to evaluate the contribution of any investment or asset class (such as alternative investments) to a pension fund's portfolio by incorporating two components. First, the framework tracks the dynamics of a pension fund's funding ratio, the most direct and relevant measure of the pension fund's ability to meet its obligations. Second, it incorporates an annual net benefit account, which is adjusted for inflation consistent with industry practice. The proposed framework is customizable to the pension fund's specific portfolio, its annual benefit payout and the choice of inflation adjustment.
We illustrate our methodology by considering an investment decision that involves a choice to allocate to commodity trading advisers. This example is particularly interesting due to its evaluation challenges but, as we show, this strategy can have a meaningful impact on pension funds' portfolios.
To simplify the modeling of pension liability, we consider a hypothetical mature pension plan that has reached a stage where demographics remain static and both contributions and benefit payments grow at the rate of inflation and wage inflation. For such a plan, it is assumed that benefits paid by the pension plan are based on salaries. It is also assumed that contributions would not be adjusted to reflect varying funding levels.
On the asset side, we use a proxy for a hypothetical pension fund portfolio that is constructed using a 60%/40% blend of stocks and bonds with an additional yield from private assets and CIO skill. We use the S&P 500 Total Return index as a proxy for stocks and the J.P. Morgan Global Government index as a proxy for bonds.
We assume that the additional yield is equal to 0.5% a year and the initial annual net benefit payout is equal to 4.5% of the fund's liabilities. We consider the time period from 2000, when the funding ratios of pension plans were close to 100%,to the end of 2018. The portfolio with trend-following allocates 10% of the original portfolio to the Societe Generale SG Trend index by reducing the stock exposure by 6% and the bond exposure by 4%. The SG Trend index is an index of the 10 largest trend-following CTAs that are open to new investments. The SG Trend index is reconstituted annually and free of biases inherent to publicly available CTA and hedge fund databases.
Table 1 summarizes the performance of stocks, bonds and the SG Trend index for the period between January 2000 and December 2018. We start in 2000 to coincide with the inception of the index.