Making Retirement Income Last a Lifetime

Abstract

To enable investors to spend down the assets in their defined contribution accounts more easily, the authors propose a decumulation benchmark comprising a laddered portfolio of TIPS for the first 20 years (consuming 88 percent of available capital) and a deferred life annuity purchased with the remaining 12 percent. This portfolio can be used directly by the investor (akin to indexing) or as a benchmark for evaluating the performance of a more aggressive strategy.

In the field of personal finance, a great deal of attention has been paid to asset accumulation, but much less attention has been paid to asset decumulation, which is the planned spending down of one’s accumulated savings in retirement. We designed a prototype strategy for post-retirement investing and used the cash flows from that strategy as a decumulation benchmark. Because this benchmark is most likely to be applied to a defined contribution (DC) savings plan, we call it the DCDB (defined contribution–decumulation benchmark). We believe that a well-engineered DC plan should be experienced by the participant in much the same way as the participant experiences a defined benefit plan.

Our benchmark is intended to embody the lowest-risk strategy available for converting accumulated capital into post-retirement income while satisfying two essential conditions: The strategy must protect the investor against longevity risk and be appealing enough that it is likely to be used by a broad cross section of investors. Immediate life annuities achieve the first condition but not the second. The apparent reason that investors shy away from immediate life annuities is that the loss of liquidity from transferring one’s capital irrevocably to an insurance company is too onerous. Thus, the benchmark strategy preserves most of the investor’s liquidity while achieving the goals of longevity protection and minimal investment risk.

The strategy that forms our benchmark is to buy, with most of one’s capital, a portfolio of laddered Treasury Inflation-Protected Securities (TIPS) out to the latest TIPS effective maturity date, currently about 20 years. The remainder of the capital is used to buy a deferred annuity that begins its payout when the cash flows from the TIPS ladder end. The proportions invested in each asset class—TIPS and a deferred annuity—are set in such a way as to make the first deferred annuity payout equal to the last TIPS payout (plus an allowance for inflation). The resulting benchmark differs from ordinary benchmarks by consisting of a set of future cash flows produced by a given amount invested.

As of 30 September 2010, a single 65-year-old male who invests $100,000 in the benchmark portfolio can expect to receive a first-year payment of $5,118, increasing at the U.S. Consumer Price Index (CPI) rate until Year 20. Thereafter, the deferred life annuity pays $7,332 (in today’s money) annually until the participant dies.

This schedule of expected cash flows can be compared with the those from other post-retirement investment strategies to determine which one a given investor might prefer. We evaluated three alternatives to the benchmark strategy: an immediate, real life annuity purchased from an insurance company; a target-date portfolio of risky assets; and an immediate, nominal life annuity purchased from an insurance company.

The immediate real annuity pays $4,856 in the first year, almost exactly the same as the first year’s payout in the DCDB. The cash flows from both strategies inflate at the CPI rate until the 21st year, when the DCDB stops inflating but the inflation-indexed annuity continues to inflate. Thus, over any life span, the inflation-indexed annuity either matches or dominates the DCDB if the investor does not care about liquidity or counterparty risk. But most investors are strongly averse to such risks, potentially tipping the choice to the benchmark strategy.

The target-date portfolio produces cash flows that cannot be accurately forecasted. Given today’s low yields, however, these cash flows are likely to be much lower than those from the benchmark in the initial years. Over time, however, the target-date portfolio should yield more than the benchmark owing to growth in earnings and dividends. The participant must thus decide which cash flow pattern she prefers: front-loaded (the benchmark) or both more uncertain and more back-loaded (the target-date fund).

The immediate nominal annuity pays $6,811 (nominal) every year. The participant must weigh this certainty—and high current income—against the likelihood that inflation will erode his purchasing power unacceptably in later years.

The benchmark strategy is not only useful as a measuring stick for evaluating alternatives but is also potentially an investment in itself, akin to indexing. Such an investment would have to be offered by an entity that can issue or sell deferred annuities as well as conventional investment portfolios.

Notes

¹ Waring and Siegel (2007) estimated that $1,802,431 in savings would be required to generate, from age 65 to age 105, an annual income of $100,000 (in real terms) that could have been purchased in the immediate annuity market for $1,180,000; (1,180,000/1,802,431) – 1 = –34.5 percent.

² Although closing and terminating a plan sound the same, closing usually refers to closing the plan to new contributions (or new participants, depending on the situation) while allowing existing contributions to remain in the plan, grow, and be paid out to beneficiaries. In a terminated plan, the contributions, along with some or all of the investment return earned, are returned to the employee, usually in the form of a DC plan balance; the employee then becomes responsible for investing the money and generating retirement income.

³ Save More Tomorrow is a registered trademark of Shlomo Benartzi and Richard H. Thaler.

⁴ Although the literature on sustainable spending rates for individuals does come from classic finance and is mostly simulation based (see the bibliographies in Dybvig and Liu 2010; Scott, Sharpe, and Watson 2009), this literature is only marginally relevant because it focuses on minimizing the probability of running out of money while still alive. In line with Bodie, McLeavey, and Siegel (2008), we do not want to minimize the probability of shortfall—we want to eliminate it.

⁵ In the context of a U.S. DC plan, a default investment is one that is purchased for the participant who makes no explicit investment choice. A list of acceptable default investments (called qualified default investment alternatives, or QDIAs) is maintained by the U.S. Department of Labor, which accords “safe harbor” status to such investments. The purpose of a default investment is to provide an executable investment strategy for an individual client who, owing to skill level or current circumstance, is not in a position to make the complex calculations and investment decisions to create a portfolio that provides either (1) lifetime income or (2) a combination of income and expected capital appreciation that the client judges to be superior to a lifetime income portfolio.

⁶ Say, by imagining a stand-alone insurance company with infinite reserve capital that holds inflation-indexed Treasury bonds with indefinitely long maturities as collateral and issues against that collateral an inflation-indexed life annuity to anyone who wants one, at actuarially fair prices with no (excess or economic) profit. That would do it.

⁷ See Charupat, Huang, and Milevsky (forthcoming 2012).

⁸ DB plans do not need our decumulation strategy because the annuitized or tontinelike payout of a DB plan eliminates the need for a separate strategy, but they could use our benchmark as a measuring stick.

⁹ Other factors that we do not address here include taxes, health, and medical insurance status (e.g., Medicare coverage). The key to the DCDB is that it is a starting point from which to make better-informed individual decisions.

¹⁰ Just as we finished writing this article, the U.S. Treasury issued a 30-year TIPS. But the issue was not very large, and how committed the Treasury is to further issuance of TIPS at this maturity is unclear. In the past, 30-year TIPS have been issued and then discontinued. Thus, our argument that a non-TIPS solution is required beyond the 20th year of the strategy is likely to remain valid for some time.

¹¹ We based our calculations on the mortality rates in the 2010 IRS Generational Mortality Table.

¹² A more fully general notation would allow for a nonzero inflation rate in the second period—say, ($5,022, 1.91 percent, 20, 1 percent) if a deferred annuity could be found that provides cash flows inflating at 1 percent a year.

¹³ Thus far, no annuity money has ever been lost by an annuitant.

¹⁴ One innovation is for the government to issue self-liquidating (amortizing) TIPS (see Goldsticker 2010). This innovation would change and improve our strategy.

¹⁵ We calculated the annuity rate by using 2010 annuity factors.