Deferred-income annuities (DIAs) have been receiving favorable press based on claims that they generate greater retirement income than traditional single-premium immediate annuities (SPIAs). DIAs have also been promoted by the Treasury Department, which has introduced new rules to facilitate their use. But I’ll present a contrarian view and demonstrate that retirement strategies built on SPIAs can outperform those that utilize DIAs.
DIAs have become a hot topic among advisors and researchers. Google searches on DIAs produce about 10 times as many hits as searches on SPIAs. The numerous articles and papers about DIAs are overwhelmingly favorable, while SPIAs receive mixed reviews. DIA sales of $2.7 billion for 2015 still lag SPIA sales of $9 billion, but have been growing faster.
My view on DIAs is a minority one, but I hold it for good reasons.
Background
A year ago, in an Advisor Perspectives article, I compared DIAs to a variety of rival products, including SPIAs, and concluded that no offering is one-size-fits-all and different products will best fit particular client needs. Here I’ll focus on particular situations that have been recommended for DIA use and demonstrate how alternative strategies utilizing SPIAs may actually work better.
DIAs can be used in two quite different ways to provide secure retirement income. The product can be used in the pre-retirement “red zone” when savings accumulations are particularly vulnerable to poor market performance. For example, a 55-year-old could purchase a DIA that will pay a specified lifetime income beginning at age 65. This would provide secure lifetime income insulated from stock market fluctuations prior to retirement. In 2014 the Treasury Department encouraged this approach by introducing new rules to allow 401(k) plan sponsors to include DIAs in target-date funds.
A second use of the product is to provide longevity insurance. An example would be a 65-year-old who purchases a DIA that pays lifetime income beginning at age 85 (a 20-year deferral period). This way, the retiree doesn’t have to worry about making savings last for an unknown future lifetime, but instead can focus on making savings last 20 years. In separate 2014 guidelines, Treasury named a new product class, qualified longevity annuity contracts (QLACs), which provide an exemption from the required minimum distribution (RMD) rule that withdrawals from tax-deferred accounts must begin at age 70.5.
There have been a number of different names assigned to this second version – longevity insurance, deeply deferred annuity, advanced-life deferred annuity (ALDA), and now the Treasury designation of QLAC. For the remainder of this article I’ll refer to the pre-retirement use as DIA and the late-in-life longevity version as QLAC.
DIA versus SPIA comparison
This comparison will be based on a 55-year-old female who wishes to dedicate $500,000 of an IRA or 401(k) to generate secure lifetime income beginning at age 65. Let’s look at two strategies she can use:
- Use the $500,000 to purchase a DIA that will provide level annual payments beginning in 10 years. Based on rates from CANNEX, a pricing information service, the annual payout would be $44,011 based on an average of the highest three companies. This particular DIA structure would refund the $500,000 premium if the purchaser died before age 65.[1]
- At age 55, dedicate $500,000 to either a long-term bond fund or a structured portfolio of long-term bonds with maturities ranging from 10 to 35 years (if available). At age 65, sell the bond investments and use the proceeds to buy a SPIA.
At first glance, these two strategies seem quite different. Strategy 1 locks in retirement income, while strategy 2 leaves the client exposed to the vagaries of the bond market for the next 10 years. With the prospect of rising rates, long-term bond investing under strategy 2 seems particularly vulnerable. But here’s the catch: these two strategies will produce essentially the same financial outcome!
To compare the two strategies, we need to estimate the payout rate for a SPIA 10 years hence. The current payout rates for a SPIA for a 65-year-old female based CANNEX rates is 6.18%. To recognize the potential impact of future mortality improvements, I’ve used the Society of Actuaries mortality projection scale for annuities and estimate a payout rate of 6.06%. With this information, we can calculate the nominal pre-tax interest rate one would need to earn on bonds during the 10 years before purchasing a SPIA under strategy 2 in order to match strategy 1 payments. The formula is (8.80/6.06) ^ .1 - 1 = 3.80%.
This leads to the question of whether it’s feasible to earn 3.8% from long-term bond investing. Examining SEC yields for Vanguard’s long-term bond funds shows a range of 2.35% for their long-term Treasury fund to 4.88% for their long-term corporate bond index, which invests most heavily at the lower end of the investment-grade scale – A and BAA. So 3.80% is a feasible target if one accepts modest credit risk.
But the equivalence of strategies 1 and 2 still leaves a question about interest rate risk. What happens to strategy 2 if interest rates rise, the bond portfolio suffers losses and there’s less money available to buy a SPIA in 10 years? The offset to this risk is that an increase in interest rates will increase SPIA payout rates. To get a bit technical, we can address interest rate risk by applying the concept of duration, which measures the timing of cash flows (coupons and principal) from bond investing and sensitivity to changes in interest rates. If the duration of the bond portfolio under strategy 2 matches the duration of expected DIA payments from strategy 1, the effect of changes in interest rates on the bond portfolio and SPIA rates under strategy 2 will exactly offset.
So that’s the technical discussion. In terms of what is going on conceptually, the decision to purchase a pre-retirement DIA is essentially a decision to lock in long-term bond yields. The purchase of annuities like SPIAs or DIAs has often been described as both fixed income investing and pooling longevity risk – those who die early subsidize those who live a long time. However, with the DIA used in strategy 1 there is minimal longevity pooling benefit between age 55 and 65 because mortality rates are low, and this example also assumes a refund of premium for deaths prior to 65. So there is really no additional mortality pooling benefit from buying a DIA at 55 versus waiting until 65 to buy a SPIA.
In terms of practical applications, I favor waiting to buy a SPIA at retirement over a pre-retirement DIA purchase. One issue is that people cannot predict when they will retire, so purchasing a SPIA at retirement offers more flexibility than the pre-retirement DIA strategy. Also with the SPIA strategy, pre-retirement planning can focus on accumulating sufficient savings for retirement. The focus on annuity purchase can be left for a time closer to retirement.
Pre-retirement DIAs have certainly received favorable publicity and Treasury support, but I don’t see them providing any financial advantage over waiting until retirement and purchasing a SPIA.
QLAC versus SPIA comparison
Now I’ll shift to longevity protection and the QLAC version of deferred-income annuities. I’ll use an example of a 65-year-old female with $500,000 in a tax-deferred account and compare annual income amounts that can be generated by purchase of a SPIA at 65 versus purchase of a QLAC at 65 that begins paying level income at 85 and income generated from a bond ladder during the 20-year deferral period. For the QLAC version, the $500,000 is split between bonds and the QLAC purchase to produce level lifetime income.
The following chart compares results for the QLAC strategy with two different SPIAs:
QLAC and SPIA Comparison (based on $500,000 investment)
|
Product choice
|
Annual pre-tax income
|
Payout rate
|
QLAC and Bond mix
|
Percent increase
|
|
QLAC with Treasury ladder
|
$25,232
|
5.05%
|
(12/88)
|
|
|
SPIA with cash refund
|
$28,482
|
5.70%
|
NA
|
12.9%
|
|
SPIA with no refund
|
$30,892
|
6.18%
|
NA
|
22.4%
|
Sources: CANNEX and author's calculations
Based on CANNEX rates for QLACs, a $100,000 premium at 65 will provide annual payments of $40,537 beginning at 85. Treasury rates for the bond ladder are assumed to average 1.75% based on current yields, and a 20 basis point expense assumption brings the yield to 1.55%. With some elementary algebra, we can show that a 12%/88% QLAC/bond mix applied to the $500,000 will generate level annual lifetime income of $25,232.
The second line of the chart is based on a SPIA that has a cash refund provision, so recovery of the original $500,000 premium is guaranteed, either as annuity payments or a cash refund to heirs. This approach provides about a 13% income increase over the QLAC with the Treasury ladder. The third line is based on a SPIA that makes annual payments until the annuitant dies with no refund for early deaths. This version pays about 22% more than the QLAC strategy.
The relationship between the three lines of the chart makes sense in that the QLAC strategy, which provides the lowest income, provides both liquidity and a death benefit during the first 20 years, both from the bond ladder. The no-refund SPIA provides the highest income but neither liquidity nor a death benefit. The cash refund SPIA is an in-between strategy that provides a refund on death but no liquidity. Viewed this way, the QLAC and SPIA strategies can be thought of as a continuum rather than as separate SPIA and QLAC strategies.
The choice of strategy depends on what’s important to the client. In making a decision, it will be important for the client to consider all retirement resources. For example, a client who has significant resources in addition to those needed to generate income may opt to maximize income with a no-refund SPIA because sufficient liquidity and bequest monies are available elsewhere. A more constrained client may opt for a QLAC strategy.
Substituting stocks for QLACS
In this recent article, Michael Kitces suggested an alternative strategy where long-term stock investing is utilized so neither QLACs or SPIAs are needed. The specifics are that the first 20 years of retirement are funded with fixed income investments as in the QLAC strategy discussed above, but the monies that would have been used to purchase a QLAC are instead invested in stocks and used to provide income after 20 years. Kitces demonstrated that, based on simulations using historical stock returns, this strategy almost always provides sufficient funds to replace the QLAC payments and often provides a substantial bequest as well.
But what happens if future average stock returns are lower than historical returns?
In the chart below, I show results under a variation of the Kitces strategy where I apply the QLAC strategy from the first chart and invest the QLAC money ($62,245) in stocks. I use Monte Carlo simulations and show a range investment balances at 85 based on different average stock return assumptions – 12% to represent the historical average, 7% for my conservative estimate of a lower return scenario and 5% for a pessimistic view. I’ve estimated that purchasing a SPIA at 85 to continue the income would cost $215,000, so the key result is how the balances generated compare with this benchmark.
Equity strategy--balances at age 85, compared to $215,000 needed for level
|
lifetime income
|
Assumed arithmetic average nominal return
|
|
Percentile
|
5%
|
7%
|
12%
|
|
90th percentile
|
$549,253
|
$851,365
|
$2,256,567
|
|
75th percentile
|
$322,911
|
$485,115
|
$1,316,897
|
|
Median
|
$175,621
|
$265,996
|
$732,994
|
|
25th percentile
|
$96,136
|
$146,501
|
$400,591
|
|
10th percentile
|
$55,850
|
$86,390
|
$232,633
|
Sources: CANNEX and author's calculations
Confirming the Kitces result, we see that, based on historical average returns, we’re OK even at the 10th percentile and produce substantial excess funds at the higher percentiles. However, the strategy is riskier based on lower average returns. The choice of whether to recommend a stock strategy and not use QLACs or SPIAs depends heavily on assumption about future equity market returns, as well as client flexibility to adjust if the strategy turns out badly.
Conclusion
DIAs and QLACs have supplanted SPIAs in terms of press coverage and government support, but in many cases SPIAs still do a better job for clients. There may be instances where a DIA or QLAC is easier to sell than a SPIA, but the financial advantages may be illusory.
The more I learn about the DIAs and QLACs, the better I like SPIAs.
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.
[1] Although SPIAs, DIAs and QLACs are available with a variety of payment structures – inflation-adjusted, fixed annual step-ups and others -- I’ll use simple level payments throughout this article to provide straightforward illustrations of product differences.
Read more articles by Joe Tomlinson
