leverage), or it will be expensive to purchase if solicited via an
annuity or a swap with a market-based counterparty; hence the
need for new instruments for DC investing.
Focus on the Wrong Objective
The second key point, as Merton (2014) argues, is that the goal of
retirement investors should not be to maximize wealth but rather
to maximize funded status, because this effectively puts the spotlight back on retirement income as the goal of investment decisions. Merton (2014) shows how assets regarded as safe in the traditional MPT context—T-bills—actually are risky when measured
from the perspective of annuity income units. This is shown in
figure 2 as the volatility of a T-bill, when measured in absolute
terms it is 0.5 percent per annum. However, when it is measured
relative to annuity (or real) income units, the relative risk rises to
15. 75 percent. As a result, assets considered safe from a CAPM
perspective are very risky from a RAPM perspective. Moreover,
Merton (2014) demonstrates that investment approaches adopted
by many DC funds and retail investors, especially target-date
funds (TDFs), actually are risky approaches from an individual
retirement income perspective.
Muralidhar (2011) has a similar and more-extensive critique of
Inability to Forecast Future Returns
TDFs, especially because U.S. regulators provide safe-harbor pro-
tection to DC pension plan sponsors if they offer TDFs. In effect,
TDFs offer no surety of any retirement outcome and only guaran-
tee that the asset allocation, focused on the basic assets described
above, will change with time. And for this assurance, investors pay
a reasonably high fee. However, these portfolios are considered safe
because they rotate more assets into nominal bonds, which are
considered safe because they offer some guarantee of principal (or
wealth). However, with volatile interest rates, a guarantee of nomi-
nal wealth is no guarantee of nominal or real retirement income.
Merton (2014) argues for DC funds to adopt a dynamic approach
that allocates to various assets to ensure some basic level of retire-
ment income (i.e., avoiding the purchase of deferred annuities and
instead investing in the hedging portfolio as well as equities early
in life to grow the asset pool—a dynamic income hedging strat-
egy). But this strategy has some risk because the basic retirement
income hedging instrument does not exist in markets.
The next major challenge is to forecast future asset returns. Recall
that even with just three assets (stocks, bonds, and commodities),
MVO requires a total of nine forecasts: three expected return forecasts, three volatility forecasts, and three correlation forecasts. As
the number of assets rises, the number of forecasts required rises in
a nonlinear fashion (because of correlations), thereby increasing
the possibility of a garbage-in garbage-out mistake. The track
record of even professionals in this regard is poor. The current
funding struggles in DB pensions reflect our inability to forecast
these variables. For example, many funds assumed that the
expected returns for stocks for the 2000–2010 decade would be in
the 8–9-percent annualized return range; in many markets, equity
returns actually were negative for this decade. Therefore, to base
the retirement prospects of a large swath of the population on a
process, with its inherent flaws, that is based on poor forecasts of
returns raises some serious concerns for retirement security.
Figure 2: Monthly Return of T-Bills in Absolute vs. Real Annuity Income Units
Note: The monthly returns are estimated by assuming that a fixed amount of money is invested in a one-month T-bill in the beginning of May 2005 and rolled over. The monthly
returns in USD are basically the yields of one-month T-bills. The “relative to real annuity income units” is the real income (annuity) amount that the T-bill portfolio can purchase.
The real income amount is estimated with the assumptions that the investor retires on April 2024 (after twenty years), and receives the same real annual income for twenty years
after retirement. The discount rate that is used in the estimation of the real income amount is thirty-year Treasury inflation-protected securities (TIPS) yields.
Data Source: Bloomberg
Absolute Return Relative to Real Annuity Income Units