Following the 2008 financial crisis, targetdate funds (TDFs) were criticized for exposing investors nearing retirement to excessive equity allocations. Were those criticisms justified? How well do TDFs stack up against the venerable strategy of matching one’s bond allocation to one’s age? My research has yielded surprising answers to those questions and to the proper role of singlepremium immediate annuities (SPIAs) alongside TDFs.
TDFs are grabbing a bigger share of retirement savings portfolios and getting more attention from advisors. As clients approach retirement, they will need advice about what to do after retiring – whether to stay with the same TDFs or change strategies. Part of the evaluation will involve examining stock allocations and postretirement glide paths. Advisors may also recommend investments to supplement TDFs such as annuities. I'll use an example to show how allocations and glide paths can affect retirement outcomes, and I’ll show how outcomes can be improved by adding SPIAs to the investment mix.
Background
TDFs have become hugely popular since being named a qualified default investment alternative for 401(k) plans under the Pension Protection Act of 2006. Savings in these funds topped $500 billion in 2013's first quarter, and the rapid growth continues.
But since the stock market losses of 2008, the growth has been accompanied by controversy. For TDFs with a target maturity date of 2010 (two years from retirement as of 2008), losses ranged from 3.5% to 41.3% in 2008, according to Ibbotson®. This was devastating for many workers preparing for retirement.
Congress and Government Accountability Office launched investigations. Everyone involved began to learn about stockallocation glide paths and the difference between "to" funds (which gradually reduce stock allocations up to a target retirement date) and "through" funds (which continue to reduce stocks over the full course of retirement). It became apparent that, even within the "to" and "through" categories, various TDFs were using widely different stock allocations.
This commotion awakened clients and advisors to the importance of allocation strategies and glide paths. These fund characteristics are critical as clients approach retirement.
An example
I'll base this analysis on a 65yearold female, just retired, with $500,000 in a TDF. She wishes to take withdrawals of $20,000 in the first year of retirement with inflation increases each year thereafter – the classic 4% inflationadjusted withdrawal rate. These withdrawals and her Social Security will be needed for living expenses, and her primary focus will be retirement security
If the client decides to continue investing in a TDF, she can choose between funds with declining stock allocations over the course of retirement and funds that hold the stock allocation level. Based on information from Morningstar® on the ranges of stock allocations and glide paths offered in TDFs, I developed the following four hypothetical fund options:
 Declining glide path, aggressive allocation: 65%/35% stocks/bonds at age 65, declining linearly to 40/60 over 15 years.
 Declining glide path, conservative allocation – 30%/70% stocks/bonds at age 65, declining linearly to 20/80 over 15 years.
 Level glide path, aggressive allocation – 50%/50% stocks/bonds at age 65, and holding this allocation for life.
 Level glide path, conservative allocation – 25%/75% stocks/bonds at age 65, and holding this allocation for life.
I also compared these allocations to a rule of thumb called "your age in bonds," in which a client's portfolio holds 65% bonds at age 65 and increases the bond percentage by 1% each year thereafter.
To generate estimated retirement outcomes, I used a Monte Carlo projection method based on average returns after inflation of 5.90% for stocks and 1.00% for bonds, and I subtracted 0.20% for expenses. These returns are more conservative than historical averages reflecting the reasoning I described in a January 2013 Advisor Perspectivesarticle. My expense assumption is representative of lowcost target date funds such as those offered by Vanguard. I also assumed variable longevity based on a life expectancy of 22 years to age 87.
This analysis focuses on stock allocations and glide paths only. I will bring in SPIAs later. A fuller evaluation could also compare expense levels for individual TDFsmost charge more than I have assumed, particularly for active management strategies. Other considerations could include the potential value added from factors such as stock selection, tactical allocation shifts and the inclusion of alternative investments like commodities.
Comparison of results
I show projected bequests as positive measures and depleted savings during retirement as negative measures. Average bequests are higher than medians, particularly for heavy stock allocations, so I rely more on median bequests for comparison. Such bequest amounts could also be used to support discretionary spending during retirement. As a failure measure, I show the probability of failure, which is the standard often used in retirement planning research. It represents the probability of depleting one’s savings while still alive. I also show the average failure for those cases where the client depletes savings, which provides a way to gauge the magnitude of failure. For a failure measure that takes both probability and magnitude into account, I multiply the average failure by the probability of failure.
Results for different glide paths

Glide path

65% > 40%

30% > 20%

Level 50%

Level 25%

Age in bonds

Average bequest

$392,000

$223,000

$393,000

$225,000

$227,000

Median bequest

$296,000

$192,000

$295,000

$193,000

$194,000

Failure probability

14.1%

17.4%

12.7%

17.2%

17.5%

Average failure

$92,000

$72,000

$82,000

$72,000

$73,000

Avg. times Prob.

$13,000

$12,600

$10,500

$12,300

$12,800

Source: Author's estimates
There's a bit of irony in these results. Much of the complaining about 2008 TDF performance was that funds close to target retirement dates were too heavily invested in stocks. But these projections show that aggressive stock allocations perform better over the full course of retirement than conservative ones do.
The rule of thumb that stock allocations should decline with age also doesn't hold up relative to the other strategies. That rule of thumb, however, was conceived in an era of higher bond yields. Some investment professionals, including Jack Bogle, have modified this rule to include the present value of Social Security as a bondlike investment, which also argues for more aggressive stock allocations.
The glide path that achieves the best balance between positive and negative measures is the level 50% stock allocation. In some ways, these results confirm financial planner Bill Bengen’s research on the retirement withdrawals in the early 1990s. Bengen showed that level stock allocations in the 50%to75% range produced the best retirement outcomes. However, Bengen based his assumptions on historical returns and demonstrated nearzero failure rates, whereas my study assumes lower investment returns and demonstrates failure rates above 10% for a 4% withdrawal rate.
Additional tests
Can one achieve even better retirement outcomes by using even higher stock allocations? In this chart, I start with the best outcome from the previous chart (level 50%) and try some variations.
Outside the glide paths – comparison of results

Glide path

Level 50%

Level 40%

Level 60%

Level 75%

40% >60%

Average bequest

$393,000

$316,000

$472,000

$629,000

$377,000

Median bequest

$295,000

$254,000

$330,000

$380,000

$283,000

Failure probability

12.7%

13.6%

13.4%

14.2%

13.0%

Average failure

$82,000

$76,000

$95,000

$108,000

$82,515

Avg. times Prob.

$10,500

$10,400

$12,700

$15,300

$10,700

Source: Author's estimates
Not surprisingly, the average and median bequests increase with higher stock allocations, and the negative outcomes deteriorate. Going in the other direction to 40% stocks reduces bequests but barely improves the failure measures.
The rightmost column shows an unusual glide path that begins at 40% at age 65 and increases to 60% over the next 22 years (life expectancy). I had conjectured that, if level allocations worked better than declining allocations, perhaps increasing allocations would work best. My conjecture was incorrect, and the 50% level was the sweet spot for this example. Overall, the average stock allocation during retirement has a much bigger impact on performance than whether the glide path rises or falls.
These results depend on the combination of assumptions, client characteristics and withdrawal rate. They provide indications of what works best, but I would not recommend using them as prescriptions for actual clients.
Can SPIAs help?
This chart repeats the level 50% outcomes as a base case and shows outcomes for two different strategies in which SPIAs are introduced as an investment alternative along with stocks and bonds. Both SPIA strategies involve investing $125,000 (25% of savings) in an inflationadjusted SPIA and leaving the remaining $375,000 to be invested in stocks and bonds. The SPIA has a payout rate of 4.77%, based on rates from Income Solutions®, so it will provide an inflationadjusted lifetime income of $5,963 annually. The two strategies differ in the stock/bond allocations on the remaining $375,000.
Bringing in SPIAs – effect on outcomes

Base case

Strategy 1

Strategy 2

Glide path

Level 50%

Level 50% w/ 25% SPIAs

Level 67% w/25% SPIAs

Average bequest

$393,000

$321,000

$448,000

Median bequest

$295,000

$249,000

$303,000

Failure probability

12.7%

9.8%

10.5%

Average failure

$82,000

$55,000

$65,000

Avg. times Prob.

$10,500

$5,400

$6,800

Source: Author's estimates
Focusing first on failure outcomes, both SPIA strategies produce improvements compared to the base case. SPIAs have the biggest impact in cases where the SPIA payout rate is significantly higher than the withdrawal rate, because this lowers the withdrawal rate on remaining assets. In this example, $5,963 of inflationadjusted income will come from the SPIA, leaving $14,037 to come from withdrawals on the remaining $375,000, which translates to a withdrawal rate of 3.74% ($14,037/$375,000) on those remaining assets. For such cases, failure outcomes could be improved by further raising the SPIA allocation. I chose 25% for this example based on what might be practical to recommend to a client.
In Strategy 1, we see a bequest reduction compared to the base case. SPIAs are much like bond investments – adding a 25% SPIA allocation and leaving remaining $375,000 invested 50%/50% lowers the initial stock allocation on the total $500,000 to 37.5%. Strategy 2 attempts to correct the overall allocation back to 50%/50% (67% of $375,000 = $250,000), and now the bequest values end up above the base case. During the course of retirement, the present value of remaining SPIA payments will decline similar to a mortgage amortization, while the remaining balance of other savings will follow an independent path. Therefore, an initial allocation involving SPIAs will not continue at the same percentage of overall assets over the course of retirement without annual adjustments.
This SPIA amortization effect is explained in detail in Wade Pfau's August 6 Advisor Perspectivesarticle. His article also referred to research with Michael Kitces that analyzed SPIAs and glide paths in a different manner than I do here, and they reach somewhat different conclusions. Anyone with a strong interest in this subject should read their work as well as mine.
Implications for advisors
Advisors will increasingly encounter clients who approach retirement with substantial investments in TDFs. They will need to evaluate the suitability of the stock allocations and glide paths to meet retirement needs. This analysis has indicated that stock allocations toward the higher end of the range offered in TDFs provide better retirement outcomes, although client tolerance for shortterm volatility also has to be considered. Decreasing equity allocations over the course of retirement, although intuitively appealing, does not improve retirement outcomes. SPIAs (and likely other types of fixed annuities) can serve as bond substitutes and also reduce the risk of outliving one’s savings. Substituting SPIAs for bonds will support higher stock allocations without undue risk of failure.
Client characteristics such as age, marital status, required withdrawal rate, and other sources of income such as Social Security must factor into what works best in individual cases.
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.