Single-premium immediate annuities (SPIAs) have been out of favor in the current low-interest-rate environment. But my new research indicates that SPIAs still offer especially attractive opportunities for retirees. One of the key reasons is that typical advisor clients will, on average, live longer than the overall population.
Background
Sometimes research in one area leads to findings in another, and this was the case in this investigation. I've recently been researching the retirement system in the U.K., where I live part of the year. This system is undergoing major changes, including the removal of rules requiring that accumulated retirement savings be turned into annuities. Part of my research involved comparing SPIA pricing in the U.K. and U.S., which yielded the surprising result that SPIAs are much more attractively priced in the U.S. Payout rates here for comparable products are 12% to 17% higher.
There is no obvious reason for the differences. Government bond rates are virtually the same in the U.S. and U.K., and there is no reason to expect that annuity purchasers in the U.K. will live significantly longer than their U.S. counterparts. An email exchange with U.K. actuary David Blake, director of the Pensions Institute at the Cass Business School, indicated that pricing differences might be due to more stringent capital requirements for British insurers, better opportunities for U.S. companies to earn attractive credit spreads, U.S. insurers not being as up-to-date with longevity estimates and different competitive environments because of forced annuitization in the U.K.
Regardless, the U.S. buyer is getting a much better deal than the U.K. buyer. I wanted to investigate whether U.K. buyers are getting a bad deal or U.S. buyers are getting an especially good deal.
Mortality assumptions
To evaluate pricing from a buyer's perspective, I needed to make assumptions about mortality rates. Annuity purchasers in the U.S. tend to be more upscale than the general population, and annuity buyers are typically healthier. Both those factors increase expected longevity. The Society of Actuaries (SOA) gathers mortality data for annuity buyers from the insurers who sell these products and publishes both mortality tables based on this data, and recommended mortality improvement scales for projecting future mortality.
As for general population mortality, the Central Intelligence Agency keeps international comparisons of mortality statistics, and these life expectancies reflect the current average ages at death. The Social Security Administration (SSA) also publishes general population longevity estimates and these are greater than the CIA numbers because they are forward-looking and reflect projections of future mortality improvements. It also makes a difference whether mortality is projected from birth or from a later age with earlier deaths excluded.
The following chart compares estimates of general population mortality with SOA projections for annuitants.
Comparative Life Expectancy Estimates--United States
|
Starting age
|
Source |
Males |
Females |
From birth |
CIA Fact Book |
77.1 |
81.9 |
From age 65 (projected) |
SSA |
84.3 |
86.6 |
From age 65 (annuity projected) |
SOA |
88.8 |
90.3 |
Sources: noted in article
These SOA estimates are greater than those I typically see in research on various retirement products (including my own past research), which may indicate that the value added by purchasing annuities has been underestimated. These SOA male/female life expectancies indicate an estimated age of 94 for the last-to-die for a 65-year-old couple.
Other research indicates that, for typical clients of advisors, a significant share of additional longevity may reflect socioeconomic status. This 2014 Brookings paper by Barry Bosworth and Karen Burke, based on Health and Retirement Study data, provides the most recent findings and also summarizes earlier research. This chart shows the effects of college education and career earnings on mortality.
Relative rates of mortality for higher socioeconomic status
Gender |
Women |
Men |
Age |
50 - 74 |
75 + |
50 – 74 |
75 + |
College and above |
0.54 |
0.79 |
0.57 |
0.74 |
Household earnings--top quintile |
0.58 |
0.69 |
0.65 |
0.73 |
Source: Bosworth and Burke from HRS
For example, women aged 50-74 with at least a college education exhibited annual death rates that were .54 times the average for the general population. Because education and earnings are highly correlated, we would not expect the effects to be additive. These differences would add about three years to a 65-year-old’s expected remaining lifetime.
Being a typical, upscale advisor client adds significantly to expected longevity, even before considering health status.
The Money's Worth measure
We can use this mortality data to develop a buyer's evaluation of annuity pricing by applying what is known as a "Money's Worth" measure. Over the past 20 years, economists in the U.S. and U.K. have used this measure to evaluate annuity pricing. For a given annuity, the Money's Worth is the present value of annuity payments over the expected remaining lifetime divided by the annuity purchase price. If the interest rate used in the present-value calculation reflects yields on bond investments an individual could purchase, the Money's Worth ratio can be thought of as a comparison of bonds versus SPIAs. However, this measure understates the value SPIAs provide, because it does not include the extra value provided in the form of longevity insurance.
The interest rates that researchers have used in present-value calculations have typically been long-term government bond rates, to match the expected duration of annuity payments. Remaining lifetimes have been based on the type of analysis outlined above for those who purchase annuities rather than for the general population.
We might expect the Money's Worth ratio to be less than 1, because insurers need to earn margins for risk and profit. This 2011 paper by Professor Ian Tonks of Bath University summarized findings from numerous studies in many countries and, for SPIAs that pay a fixed amount each year, reported Money's Worth ratios ranging from 0.89 to 1.00, with clustering in the 0.97 range. Tonks concluded that, based on the small margins, annuities provide attractive values for purchasers even before considering the added value of longevity insurance. Unfortunately, virtually all the studies on Money's Worth ratios predate the low-interest-rate environment we have experienced since the Great Recession, so updates are needed.
Updated Money’s Worth data
In the chart below, I have provided a Money's Worth calculation, updated to current conditions, based on level-payment annuities for 65-year-old males and females with rates from Income Solutions®, an annuity-shopping service that sells direct via Vanguard. I used level-payment SPIAs because these are more popular products than SPIAs with payments that step up by a set percentage each year or SPIAs that provide payments that adjust based on actual inflation. For an interest rate, I used the 20-year Treasury, which yielded 3.23% in late April 2014, and also included a separate analysis where I added the 15+-year corporate spread (increasing the yield up to 4.79%) to recognize that investors may be able to obtain more attractive fixed-income yields than Treasurys offer.
I've also included a rough estimate of a current U.K. Money's Worth ratio, based on the payout rate differentials cited at the beginning of this article. (The male/female Money's Worth differential for the U.K. reflects uniform annuity rates for males and females – an EU requirement – though expected mortality differs. The gender-distinct rates used for this analysis apply in the U.S. except under federally regulated programs.)
Annuity Money's Worth Calculations per $100,000 invested
Gender and purchase age |
Male 65 |
Male 65 |
Female 65 |
Female 65 |
Monthly SPIA payment |
$590.55 |
$590.55 |
$564.05 |
$564.05 |
Interest rate |
3.23% |
4.79% |
3.23% |
4.79% |
Life expectancy |
88.8 |
88.8 |
90.3 |
90.3 |
Money's Worth |
1.13 |
0.96 |
1.13 |
0.95 |
Money's Worth (U.K. est.) |
0.97 |
0.82 |
1.01 |
0.85 |
Sources: Income Solutions®, St. Louis Fed FRED, SOA, Author's calculations
Based on this limited evidence, it appears that SPIA rates are very attractive in the U.S. compared to bonds. This is the case even after adding a substantial spread margin to the Treasury yields. This result agrees with other research that has evaluated SPIAs versus bonds by comparing retirement outcomes (see my June 2012 article in Advisor Perspectives). The estimated rates for the U.K. are more in line with the historical ratios reported by Ian Tonks.
SPIAs in the U.K. offer good value, but SPIAs in the U.S. offer excellent value.
Arguments against SPIAs
Despite this good news for SPIAs, it's no secret that the product is not very popular in the U.S. It's worth considering my research in light of the most common arguments against purchasing SPIAs.
It's better to wait until rates rise before purchasing a SPIA. It is likely that interest rates will be higher in a few years, and SPIA rates will be more attractive then. The problem with this argument is that the long end of the yield curve, which is incorporated in current SPIA pricing, already reflects the market expectation for future rate increases. If waiting were to be a winning strategy, rates would have to increase by more than the market already expects. Parking money at near-zero short-term rates while waiting for rates to increase involves a sacrifice that that will not be recovered unless future rate increases exceed market expectations.
Purchasing a SPIA risks a loss to heirs in the event of an early death. This argument may not hold up if one also considers the expense side of the retirement equation. If a SPIA is purchased to provide income for basic living expenses not covered by Social Security or pensions, those expenses will end at death at the same time the SPIA income stops. So the SPIA actually reduces risk by matching basic living expenses regardless of the length of life. This particular anti-SPIA bias reflects a narrow-framing behavioral bias. Broadening the framing to include retirement expenses clarifies the picture.
Purchasing a SPIA results in a loss of flexibility. Again, if the SPIA is purchased to provide income for basic living expenses, one does not actually have the flexibility to forgo basic food and shelter expenses. So the flexibility one holds onto by living off savings may be an illusion.
Purchasing a SPIA reduces expected bequest values. Most retirees hold mixed allocations of bonds and stocks. If SPIAs are substituted for bonds in the allocation, there is no loss of expected bequest if the Money's Worth ratio, based on current bond investment opportunities, is 1 or greater. A Money's Worth Ratio of 1 means that an individual will achieve the same investment outcome from SPIAs and bonds if one lives to life expectancy.
Other behavioral biases may get in the way of fairly assessing the benefits offered by SPIAs. These include overconfidence about investment prowess, problems understanding economically equivalent asset-income tradeoffs, reluctance to trade savings for an income stream and the framing issue mentioned above. Advisors can help clients overcome these biases, assuming, of course, that they are not afflicted with the same biases themselves.
Joe Tomlinson, an actuary and financial planner, is managing director of Tomlinson Financial Planning, LLC in Greenville, Maine. His practice focuses on retirement planning. He also does research and writing on financial planning and investment topics.
Read more articles by Joe Tomlinson
