Most retirees without annuities run the risk of either running out of savings or scrimping unnecessarily. An annuity provides income security, especially for the rising number of retirees starting off retirement with a lump sum rather than a traditional pension. Indeed, some research has found that retirees with annuities are happier and more satisfied with retirement than those without such income guarantees.1

For various reasons, however, annuities remain relatively unpopular with retiring workers. It is well known that annuities sold individually by insurance companies reflect the cost of marketing, profit margins and of providing lifetime income to purchasers who tend to live longer than others. These loading fees diminish the efficiency of retirement savings and likely discourage some retirees from buying annuities. Even when workers can obtain lower-cost annuities in a group market, however, the evidence suggests that few people buy them.

Finally, annuities cost more when interest rates are low, so timing the purchase is a risk. For example, the annuity a worker who retired in 2010 could buy would be only 80% of an annuity bought in 2007 and only 63% of an annuity purchased in 2000, even when all three retirees spent the same amount.2

Social Security provides longevity insurance — with the monthly payout rising for those who delay their claims. Many workers file a claim as soon as they can, which is age 62, often because they think they can't afford to wait without dipping into their savings. The following analysis explores the financial outcomes for workers who retire at 62 but wait until 70 to claim their Social Security benefits, using a portion of their savings to finance the eight-year interval.

Delaying Social Security benefits

In recent years, a number of studies have trumpeted the advantages of delaying Social Security retirement benefits.3 Retirees can start collecting Social Security benefits at age 62, but those who are willing and able to wait until later will collect higher monthly benefits for life. For workers born from 1943 through 1954, Social Security considers 66 as the "full retirement age," and taking the benefit before then reduces the monthly payout. A 62-year-old has a longer life expectancy than a 66-year-old, so the lower monthly benefit stretched out over the longer period is intended to be actuarially equivalent to the higher benefit taken over fewer years. The Social Security program also provides "deferred retirement credits" for people who wait even longer to claim benefits, up to age 70.

Social Security benefits are derived from the primary insurance amount (PIA). To determine the PIA, a three-tiered formula is applied to the average of the worker's highest 35 years of wage-indexed earnings. Workers who wait until full retirement age to start collecting Social Security benefits receive the equivalent of the PIA. Benefits are reduced by 6.67 percentage points of the PIA for each of the first three years before full retirement age and by an additional 5% for each year earlier than that. In 2015, a worker claiming benefits at age 62 receives 75% of the PIA, with the percentage rising to 80% at 63, 86.7% at 64 and so forth. For claims filed between birthdays, the benefits are adjusted on a pro rata basis.

For each year the claim is delayed beyond full retirement age, the benefit increases by 8%, plus any cost-of-living increases in Social Security benefits being paid in the meantime. In inflation-adjusted terms, a 67-year-old first-time claimant receives 108% of the PIA, and a 70-year-old claimant receives 132% of the PIA. Relative to the PIA, a benefit claimed at age 70 is around 75% greater in real terms than a benefit claimed at age 62.

Economists John Shoven and Sita Slavov developed a series of papers on the subject of the added benefits that retirees can realize over their lifetimes by deferring Social Security beyond age 62. Their 2013 paper focused on the advantageous effects of legislative changes to the program in the 1990s and early 2000s, and of the historically low-interest environment in 2013.4 They found that the advantages of deferring benefits beyond age 62 were minimal for both singles and couples born in 1930. For those born in 1951, however, they estimated lifetime gains from delaying benefits at up to $30,000 for single men, $50,000 for single women, $85,000 for single-earner couples and more than $100,000 for two-earner couples.

Tax implications of delaying Social Security benefits

Research by William Meyer and William Reichenstein compared expected lifetime Social Security benefits for singles and couples based on different claiming ages and including the effects of tax treatment on retirement income, demonstrating its importance in retirement decisions.5 They also considered the decision of when to claim Social Security benefits in the context of meeting retirement income goals using all retirement benefits and assets.

Their hypothetical workers started retirement with some combination of a $700,000 retirement fund and benefits from Social Security. Given retirement at age 62 as the baseline and specified assumptions for annual returns on assets and tax liabilities, they estimated annual income over a 30-year retirement. At the same after-tax income as used in the baseline, delaying retirement and tapping either Social Security or the retirement savings extend the period over which benefits can be claimed, in effect providing longevity insurance. The higher Social Security benefit the retiree gains by delaying the claim enables him or her to reduce annual withdrawals from retirement savings, thereby increasing the likelihood of meeting the income goal and extending the payout period.

Many older workers, however, don't have the option of working past age 62. Some cannot continue working because of health concerns or physical limitations, while others lose their jobs and cannot find suitable employment elsewhere. These people may have little choice but to tap assets in their retirement portfolio sooner rather than later. The inclination is often to take Social Security benefits first, both because they are available and to protect the accumulated retirement savings. For a variety of reasons, however, this is not always the most cost-effective option.

Not all income is created equal when it comes to taxes, as the research by James Mahaney and Peter Carlson points out. Social Security benefits are tax-free for filers whose "combined income" is less than $25,000, and only 50% of the benefit is taxable if the combined income falls between $25,000 and $34,000. Above $34,000, 85% of the benefit is taxable. For couples, the respective thresholds are $32,000 and $44,000. Combined income is the sum of adjusted gross income, nontaxable interest and one-half of the Social Security benefit. This tax treatment of Social Security suggests that the value of deferring the claiming of Social Security benefits may be greater than it appears by simply comparing present values on gross benefits claimed at different ages.

Combined income from retirement savings and Social Security can result in relatively high tax rates on Social Security benefits.6 In their example, Mahaney and Carlson assume that a couple in the 25% tax bracket receives an additional $1 of individual retirement account (IRA) income. In addition to being subject to a 25% tax, the extra $1 of IRA income can also make additional Social Security benefits subject to taxes. For example, if the $1 added to "combined income" results in the taxation of 85% of an additional dollar of Social Security benefits, it creates an additional $0.2125 ($1 x 0.85 x 0.25) tax on the latter benefits. Combined with the direct 25% tax on the IRA benefit, the combined tax rate on the additional dollar is 46.3%.

Because only half of Social Security benefits are included in taxable income, the retiree could receive larger amounts of Social Security benefits before triggering extra taxation than with IRA income. Working through the tax arithmetic, the authors show that the tax savings and effects on disposable income can be substantial. They also make the case that trading income from retirement savings early in retirement for higher Social Security benefits later can reduce administrative costs on the retirement funds.

Clearly, the interaction between the timing of claiming Social Security and spending of retirement savings is an important consideration for retirees trying to maximize their incomes. But the ability to defer Social Security depends on having other retirement assets on tap. This raises a series of questions about how much retirees need in order to take advantage of some of the options, how much they need to save in order to do so and how the answers to these questions vary across the earnings spectrum.

If retirees are seeking the most efficient way to buy longevity insurance — by increasing their retirement annuity income later in life — how can they benefit from tapping their retirement savings during retirement while deferring their Social Security claiming?

Using retirement savings as an income bridge

To explore the implications of using retirement savings as an income bridge for workers who retire at 62 but wait until 70 to claim Social Security benefits, we looked at a 1% sample of Social Security claimants who took benefits at age 62 in 2004 and who had at least 35 years of covered earnings. We recorded covered wages for each year and used the following assumptions for our scenarios:

  • Retirees started saving 6% of pay at age 21 and stopped at 61.
  • Workers allocated 60% of their savings to an index stock fund that paid returns commensurate with U.S. stocks and 40% to 10-year bonds held to maturity so their returns equaled annual yields on such assets (we used actual returns on stocks and bonds). They reallocated their accounts each year to maintain this balance between equities and bonds.
  • Administration fees on accumulating assets were 25 basis points per year.
  • Workers stopped working at year-end 2003 just before turning age 62.
  • Workers deferred their Social Security benefits until age 70 but withdrew income annually from their accumulated savings through age 69, with the withdrawn amounts equal to the Social Security benefits they would have received by claiming at 62.
  • People took these actions to increase total lifetime benefits over their retirement, obtain higher annuity income to protect against longevity risks or some combination of both.
  • Retirees had normal life expectancies.

The "bridge" income withdrawn from savings to replace foregone Social Security benefits reduced accumulated savings each year, raising the question of whether the delayed Social Security benefits more than offset the opportunity cost of spending accumulated savings early in the retirement period.

Figure 1 tracks the results of the experimental modeling of using savings to cover the income retirees would forego by waiting until 70 to claim Social Security benefits. The columns in the table reflect the median results for those in the even-numbered average indexed monthly earnings (AIME) deciles for the 1942 birth cohort who retired at age 62. The gains from deferring Social Security claiming are shown in present value terms at age 70.

Figure 1. Net gain from using retirement savings to bridge the gap from deferring Social Security from 62 to 70

Figure 1. Net gain from using retirement savings to bridge the gap from deferring Social Security from 62 to 70
Click to enlarge

Source: Developed by the authors

As shown in Figure 1, a single worker who waited until age 70 to start collecting Social Security and tapped savings to make up the shortfall in the meantime reduced his or her retirement savings by between 40% and 60% at age 70. The hit to savings was larger for lower-earning workers because Social Security is relatively more generous to them, so it requires a larger share of savings to offset the lost Social Security income.

In each case, the boost to the net value of lifetime Social Security benefits exceeded the reduction in savings. For the single worker in the sixth AIME decile, deferring Social Security benefits to age 70 achieved a $159,786 gain. An annuity paying the equivalent of the added Social Security benefits would cost $202,697 (assuming a 10% load). So in this case, the lifetime annuity gain was $58,148 or 40% of the reduction in savings at age 70 from financing the bridge income to offset the missing Social Security income from age 62 to 70.

One-earner couples need to withdraw more from savings than single people in order to replace both the worker's and the spouse's Social Security benefit during the bridge period. Still, the net potential gains are larger both in raw dollar differences and the additional annuity values the couples could realize over time. The net annuity gains from deferring Social Security — as valued by annuity pricing in commercial markets — were equivalent to 61% of the reduced savings at age 70 (from spending the savings for income from ages 62 to 70). For workers with at least a normal life expectancy at retirement, the higher lifetime annuities are more favorable than other available alternatives.

For the two-earner couple, the dollar magnitudes of the benefits derived from deferring Social Security were considerably larger than for the single worker but were comparable in relative terms. In this example and the others, we did not take into account the different tax treatments of income generated from tax-favored retirement savings versus Social Security.7

These retirees were born in 1942 and reached age 69 in 2011. By then, the equity markets had mostly recovered from the 2008 financial crisis. Figure 2 uses the same earnings histories used in Figure 1, but instead of using asset returns for the 1942 birth cohort, the figure shows real asset returns for retirees born three years earlier. These retirees turned 70 at the end of 2008, when equity prices had fallen dramatically. Some analysts believe that the steep drop in asset values hurt retirement income prospects for workers who planned to rely on their savings.8

Figure 2. Three-year earlier history of asset returns and net benefits from deferring Social Security benefits

Figure 2. Three-year earlier history of asset returns and net benefits from deferring Social Security benefits
Click to enlarge

Source: Developed by the authors

As shown in Figure 2, the gains from deferring Social Security and relying on retirement savings instead were greater for those who turned 70 when the stock market reached its lowest point in 2007 – 2008. These retirees were withdrawing assets from savings in the years leading up to the market decline, and thus had less savings to lose and more to gain from deferring Social Security. This strategy appears to offer retirees a considerable buffer against adverse effects from financial market downturns.

If using retirement savings as bridge income to enable retirees to defer Social Security benefits mitigates the consequences of market downturns, it likely would have the opposite effect on sustained market upswings. Figure 3 also uses the 1942 birth cohort's earning histories but applies asset returns from 13 years earlier. These retirees tapped into their retirement savings for bridge income at a time when equity prices were climbing steeply, and they turned 70 in early 1999 — when stock prices were nearing their peak before the dot-com bubble burst. They would have been considerably better off by starting Social Security benefits at 62.

Figure 3. Thirteen-year earlier history of asset returns and net benefit from deferring Social Security

Figure 3. Thirteen-year earlier history of asset returns and net benefit from deferring Social Security
Click to enlarge

Source: Developed by the authors

However, the net loss to the two-earner couple from tapping retirement savings at age 62 and deferring Social Security benefits until 70 does not mean that it was the wrong decision. Hindsight is not a proper basis for evaluating investment strategies. For example, workers who had invested 100% of their retirement savings in equities from 1991 through 1998, rather than a recommended 60-40 equity-bond split, would have accumulated nearly an additional 40% in retirement savings. But that doesn't mean that diversification would have been inappropriate. Indeed, virtually all the default lifecycle investment strategies that are popular in retirement savings plans today include some diversification of assets during workers' careers.

The scenarios in Figures 1, 2 and 3 demonstrate the financial market timing risks facing participants in 401(k) and other defined contribution plans in converting some or all their retirement savings into annuity income. What may be less clear is that this option eliminates the interest-rate risk associated with annuity purchases that was mentioned earlier.

For this approach to become more widely used and effective in providing additional longevity insurance protections to retirees, people need to learn about the options and potential benefits while they're still working. Employers with defined contribution plans could play a crucial role in educating workers and in structuring their plans to facilitate payouts during the bridge period before starting Social Security benefits. For this to be a long-term approach that could be incorporated into retirement planning, however, policymakers would need to resolve the projected financing shortfalls that Social Security is facing.


While policy analysts and policymakers often recommend that retirees acquire greater longevity insurance, their advice flies in the face of the widespread reluctance to purchase annuities. Analyses of the potential benefits of deferring Social Security benefits generally ignore the liquidity constraints that many people have — or perceive they have — during the recommended deferral periods. Many people are unaware of the potential lifetime benefits they might accrue or don't know how to bridge the income gap between when they stop working and when they file a claim for Social Security benefits.

If policymakers want to encourage annuitization, helping people defer their Social Security benefits has a number of economic advantages: 1) There are little or no commercial marketing costs or profit loads; 2) the benefits from deferral are incremented on a general life expectancy basis as opposed to those used to price commercial annuities; and 3) Social Security annuities are indexed for inflation over the remainder of recipients' lifetimes.


1. See "Annuities and Retirement Happiness," Towers Watson Insider, September 2012.

2. See McFarland, Brendan and Mark J. Warshawsky, "Balances and Retirement Income from Individual Accounts: U.S. Historical Simulations," Benefits Quarterly (Second Quarter 2010), pp. 36 – 40.

3. See Mahaney, James I., and Peter C. Carlson, "Rethinking Social Security Claiming in a 401(k) World," in John Ameriks and Olivia S. Mitchell, eds., Recalibrating Retirement Spending and Saving (Oxford University Press, 2008), pp. 141 – 167; Meyer, William, and William Reichenstein, "Social Security: When Should You Start Benefits and How to Minimize Longevity Risk?" Journal of Financial Planning (March 2010), pp. 52 – 63; and Shoven, John B., and Sita Nataraj Slavov, "Recent Changes in the Gains from Delaying Social Security," NBER Retirement Research Paper No. NB 13-04 (November 2013).

4. See Shoven, John B., and Sita Nataraj Slavov, "Recent Changes in the Gains from Delaying Social Security," NBER Retirement Research Paper No. NB 13-04, pp. 19 – 20 (November 2013).

5. See Meyer, William, and William Reichenstein, "Social Security: When Should You Start Benefits and How to Minimize Longevity Risk?" Journal of Financial Planning (March 2010), pp. 52 – 63.

6. See Mahaney, James I., and Peter C. Carlson, "Rethinking Social Security Claiming in a 401(k) World," in John Ameriks and Olivia S. Mitchell, eds., Recalibrating Retirement Spending and Saving (Oxford; Oxford University Press, 2008), pp. 141 – 167.

7. There are proposals elsewhere calling for Social Security to be taxed equivalently to benefits provided under tax-qualified plans. If this should come to pass, the difference between income taken from a tax-qualified retirement plan and Social Security would be significantly diminished, although in most cases not completely eliminated.

8. See Butrica, Barbara A., Karen E. Smith and Eric J. Toder, "Retirement Security and the Stock Market Crash: What Are the Possible Outcomes?" Center for Retirement Research at Boston College (November 2009), CRR WP 2009-30.