Participants during the financial crisis: Total returns

Transcription

1 Participants during the financial crisis: Total returns Vanguard research November 2011 Executive summary. For the period, the typical defined contribution (DC) plan participant earned an average annual return of 3.76% and a cumulative return of just over 20%. In other words, for the typical participant, retirement wealth invested over the period grew by one-fifth because of investment results, despite the market decline. Portfolios selected by participants were more highly dispersed in terms of risk and return than professionally managed target-date funds or managed accounts. Authors* Stephen P. Utkus Shantanu Bapat Realized total returns. Over the period, the typical DC plan participant at Vanguard earned a 3.76% average annual total return and a cumulative return of 20.27%. Ninety-five percent of participants earned a positive total return over the five years. Over the shorter period, total returns were breakeven. * The authors would like to thank John Lamancusa for his research assistance. Past performance is not a guarantee of future results. Connect with Vanguard > vanguard.com

2 Dispersion of outcomes. Returns on participant-constructed portfolios are much more highly dispersed than those of professionally managed portfolios such as single target-date or managed account strategies. For example, five-year returns for single target-date investors ranged narrowly from 3.62% to 4.65% per year (for the 5th and 95th percentiles) with a mean of 3.93%. Among participants making choices on their own, five-year returns varied widely from 0.02% to 8.09% per year, with a mean of 3.76%. Portfolio risk (volatility) levels are also characterized by extreme values among participants making their own choices. [Note: All investing is subject to risk. Investments in target-date funds are subject to the risks of their underlying funds. The year in the fund name refers to the approximate year (the target date) when an investor in the fund would retire and leave the workforce. The fund will gradually shift its emphasis from more aggressive investments to more conservative ones based on its target date. An investment in a target-date fund is not guaranteed at any time, including on or after the target date. Diversification does not ensure a profit or protect against loss in a declining market.] Age and total returns. Among single target-date and managed account investors, equity exposure declines in a disciplined way over time, while among all other participants it is hump-shaped. For participants approaching retirement, ages 55 64, retirement wealth invested over the period grew by 21% for the average participant and by 24% among single target-date investors. Older single target-date investors not only realized higher returns but also had higher Sharpe ratios. 1 Implications. These results underscore the importance of evaluating DC plan performance and retirement wealth accumulation over longer time horizons. Despite the historic market shock of , DC account wealth accumulation was positive over five years because of investment results alone before considering the effect of contributions. This is also true for single target-date investors, including those nearing retirement. For plan fiduciaries, our results draw attention to the widely dispersed nature of participant outcomes. Strategies like target-date funds, managed accounts, and reenrollment into a qualified default investment alternative (QDIA) can be considered as ways to mitigate extreme portfolio choices. 1 The Sharpe ratio is defined as the average annual return of a participant portfolio, minus the risk-free (Treasury bill) rate over the period, divided by the standard deviation of the participant s portfolio. 2

3 Introduction DC plan participants endured a historic equity market shock during the period, when stock markets fell by more than half in response to the global financial crisis. In previous Vanguard research, we have shown that retirement wealth accumulation including investment performance and ongoing plan contributions was positive for most participants over the years surrounding the market decline. 2 In this paper we consider a narrower question: How did participant retirement accounts perform when measured solely in terms of investment results? This examines participant total returns over the period. 3 Total returns are the most common measure of investment skill or performance. In simple terms, they represent the results of investing $1 over a given period. They reflect solely the impact of investment decisions, independent of cash flows related to participant accounts (such as contributions and any loans or withdrawals). Participant total returns are comparable to widely published measures of investment performance, such as index returns and fund performance returns. Like individual fund performance results, but unlike index returns, participant total returns are net of all fees and expenses specific to a given participant s account. 4 After a brief review of methodology and the market environment, the paper presents several views of participant total returns and shows how they vary by portfolio strategy and participant characteristics such as age. In particular, the strategy section examines return differences among participants who held a single target-date fund, invested in a managed account advisory service, or held any other type of portfolio over the five-year period. The paper also presents various measures of the dispersion or distribution of participant returns. Participant asset allocations can be highly skewed, and returns accordingly are dispersed. Caveats Several caveats should be kept in mind while reviewing the findings of our analysis. First, realized total returns reflect investment results for a specific period, and are conditional to the market and investment conditions prevailing during that time. As noted below, the period was unusual in that the equity risk premium for U.S. stocks was negative, and was close to zero for international stocks. Future results can be expected to be quite different in a more normal period with a positive equity risk premium. Second, realized returns over the period reflect the impact of several portfolio characteristics, including asset allocation, subasset allocation, fund selection, rebalancing, and fees. Given this wide range of factors influencing portfolio outcomes, results across participants and results across different strategies should be expected to vary over future periods. In other words, this period s underperformance could be a harbinger of the next period s outperformance, and relative results should be judged over time. Finally, it s important to emphasize that the total returns shown here reflect investment performance only. Growth in retirement wealth would be even higher than we are showing here for any given period among those making ongoing contributions. Methodology Our data is drawn from Vanguard s recordkeeping systems encompassing more than three million DC plan participants. Total returns were calculated each month for participant accounts over the period from December 2005 to December Total returns are net of all fees and expenses associated with the investment options held by the participants, as well as any separate account-based recordkeeping charges such as per-capita recordkeeping fees. 2 How America Saves, For an analysis of the effects of 2011 market volatility on participant portfolios, see Clark and Young, Returns that are weighted according to the unique cash flow patterns of participants accounts, such as ongoing contributions, are known generally as cash-flow-weighted returns or, in Vanguard recordkeeping terminology, personal rates of return. Over the period, personal rates of return were slightly higher than total rates of return. See How America Saves, 2011, Figures Total returns are calculated for Vanguard participant portfolios from cash flow to cash flow, and then linked over time, in accordance with global investment performance standards. 3

4 When presenting one-, three-, and five-year returns, we use a dataset of all participants with corresponding one-, three-, or five-year account histories, as well as complete demographic and asset allocation data. This dataset is referred to as the all participants sample, and includes 2.62 million participants with a one-year history, 1.98 million participants with a three-year history, and 1.43 million participants with a five-year history. 6 Finally, company stock as a portfolio holding creates a higher degree of volatility in returns, and the most extreme return outcomes positive and negative are associated with concentrated company stock positions. Approximately 12% of the five-year dataset includes participants with a concentrated position in employer securities that is, a position that exceeds 20% of their account balance. Because they are only a small portion of the sample, their results are included within the totals for all participants. Market environment During the period, global stock prices increased steadily through 2007, reaching a peak in October of that year (Figure 1). Stock prices then fell precipitously during the bear market of which was similar in scope to other historic market declines, such as the bear market. Stock prices partially recovered through the end of In terms of average annual returns, U.S. stocks gained 3.0% per year over the five-year period, while U.S. bonds returned 5.8% (Figure 2). In other words, relative to U.S. bonds, the U.S. equity risk premium was a negative 2.8% per year over five years. International stocks generated a 5.3% annualized return, just shy of U.S. bond market returns. Figure 1. Market performance Cumulative total return for select market indexes 180 Index value Dec-05 Jun-06 Dec-06 Jun-07 Dec-07 Jun-08 Dec-08 Jun-09 Dec-09 Jun-10 Dec-10 Treasury bills U.S. bonds U.S. stocks International stocks Note: See figure 2 for corresponding indexes. All values set at as of month-end December Past performance is not a guarantee of future returns. The performance of an index is not an exact representation of any particular investment, as you cannot invest directly in an index. Sources: Citigroup, Barclays, MSCI. 6 One- and three-year returns are not materially different if they are presented only for the five-year panel of 1.43 million participants. Presenting returns for only the five-year panel leads to an inherent bias in the data, as it reports on results only for participants with five years or more of plan tenure. 4

5 Figure 2. Market returns and volatility For the period ended December 31, 2010 Annualized Cumulative Volatility total returns total returns (s)* 60 Asset class Index 1 year 3 years 5 years 1 year 3 years 5 years months Inflation Consumer Price Index (Urban) 1.5% 1.4% 2.1% 1.5% 4.4% 10.9% 3.0% Cash reserves Citigroup T-bill Index 0.1% 0.7% 2.3% 0.1% 2.1% 12.0% 0.6% U.S. bonds Barclays U.S. Aggregate Bond Index U.S. stocks MSCI U.S. Broad Market Index International stocks MSIC All Country World Index ex-us * Annualized standard deviation of monthly returns over the 60-month period ended December 31, Bonds offered the highest cumulative return over , with a dollar at the beginning of the period growing to $1.33. International stocks came close to matching these results, with $1.00 growing to $1.29 over five years. Meanwhile, U.S. stocks meaningfully underperformed U.S. bonds, with a dollar invested in U.S. stocks growing to $1.16. Figure 3. Participant total returns Defined contribution plan participants, for the period ended December 31, % 20.27% Participant total returns The typical Vanguard DC plan participant earned an average annual return of 3.76% and a cumulative return of 20.27% over the period (Figure 3). In effect, for the average participant, retirement wealth over the five-year period grew cumulatively by one-fifth because of investment results. Results for the shorter period reflect the fact that the beginning of the period occurred close to the peak of U.S. stock prices. Over the period, the average participant essentially broke even in investment returns, realizing a total return of 0.11% per year and a cumulative total return of 0.33%. On a one-year basis, the average participant earned a 12.22% annual return, reflecting the recovery of stock prices during % 12.22% 12.22% 3.76% 0.11% 0.33% Average annual total returns 1 year 3 years 5 years Average cumulative total returns Note: Includes 2,623,615 participants for the 1-year period; 1,979,116 participants for the 3-year period; and 1,432,404 participants for the 5-year period. Source: Vanguard,

6 Dispersion of returns One characteristic of participant portfolios is the tendency of many to adopt unconventional or even extreme investment allocations. Some participants may invest their entire portfolio in equities or invest exclusively in low-risk money market or stable value assets. Participant portfolios may also be concentrated in a single fund, in a certain management style, in a specialty asset class, or in company stock. As a result, participant total returns tend to be dispersed or distributed over a wide range. These statistics illustrate the wide variation in portfolio construction among individual participants. They also demonstrate that, over the five-year period, 95% of participants earned a positive investment return on their account, while over the three-year period, about half did and half did not. Figure 4. Distribution of participant total returns All DC plan participants for the period ended December 31, 2010 Over the five-year period, among all participants in our sample, the average participant portfolio gained 3.76% per year, while the 5th percentile portfolio earned 0.00% per year, and the 95th percentile portfolio gained 8.02% per year (Figure 4). On a cumulative basis, the 5th percentile participant saw a $1.00 investment stay at $1.00, while the 95th percentile participant saw a $1.00 investment grow to $ % 20.81% 15.17% 12.22% 9.89% 8.02% Similarly, from 2007 to 2010, the 5th percentile participant portfolio lost 5.13% per year, while the 95th percentile portfolio gained 5.65% per year. On a cumulative basis, the 5th percentile participant saw a $1.00 investment fall to $0.85, while the 95th percentile participant saw a $1.00 investment grow to $ % 10.00% 5.65% 4.62% 3.76% 1.69% 2.66% 0.68% 0.11% 0.00% 1.64% 5.13% 1 Year 3 Years 5 Years Reading a box-and-whisker graph: Top of line represents 95th percentile; top of box, 75th; diamond, average; bottom of box, 25th, and bottom of line, 5th. Note: Includes 2,623,615 participants for the 1-year period; 1,979,116 participants for the 3-year period; and 1,432,404 participants for the 5-year period. Source: Vanguard,

7 Impact of target-date funds and managed accounts Participant portfolio strategies in DC plans are being reshaped by two innovations: target-date funds and managed account advisory services. These innovations have introduced professionally managed approaches to portfolio construction in DC participant accounts. We separately calculated five-year total returns for participants investing in a single target-date fund or through a managed account service, and compared them with the returns for all other participants. Under our approach, the target-date or managed account participant had to maintain this strategy over the full 60-month period. For the five-year period, we observed 16,869 participants in a single target-date fund, and 6,685 participants in a managed account advisory service. They are compared with 1.41 million participants who are invested otherwise. 7 Both target-date funds and managed accounts have contributed to a substantial reduction in extreme portfolio outcomes (Figures 5 and 6). Over the five-year period, single target-date investors fell within the narrowest range of returns from 3.62% per year for the 5th percentile portfolio to 4.65% for the 95th percentile portfolio. Managed account returns were somewhat more dispersed, ranging from 2.20% per year at the 5th percentile to 5.06% at the 95th percentile, but still relatively compact compared to all other participants. Returns for all other participants ranged from 0.02% for the 5th percentile investor to 8.09% for the 95th percentile investor. Figure 6. Distribution of five-year returns by strategy All defined contribution participants, five-year annualized total returns for the period ended December 31, % 8.09% Figure 5. Distribution of five-year returns by strategy All defined contribution participants, five-year annualized total returns for the period ended December 31, 2010 Annualized five-year returns Single target-date Managed All other fund account participants Mean 3.93% 3.65% 3.76% 4.65% 5.06% 4.64% 4.22% 3.62% 3.65% 3.76% 3.08% 2.66% 2.20% 5th percentile th percentile th percentile (median) th percentile th percentile % 0.02% Source: Vanguard, % Single target-date fund* Managed account All other participants Reading a box-and-whisker graph: Top of line represents 95th percentile; top of box, 75th; diamond, average; bottom of box, 25th, and bottom of line, 5th. * Single target-date fund: 95th percentile is 4.65% ; 75th, 3.90%; mean, 3.93%; 25th and 5th, 3.62%. Note: Includes 16,869 single target-date investors, 6,685 managed account investors, and 1,408,850 other participants over the 5-year period. 7 Over the five-year period, participants who were in single target-date funds or a managed account advisory service for less than the full period, or participants with a partial position in a target-date fund, are included in all other participants. 7

8 Dispersion of risk and return characteristics Differences among these three investment approaches are particularly apparent when examining both return and risk measures. To present these differences graphically, we created a random sample of 1,000 participants holding a single target-date fund, a managed account, or any other strategy over five years. 8 Single target-date investors fall in a narrow band of risk and return because virtually all the target-date portfolios in our sample are a specified combination of indexed U.S. equities, international equities, and U.S. bonds (Figure 7, Panel A). In fact, the lack of dispersion in returns for single target-date investors creates a misleading visual effect in this graph. All three graphs have the same number of observations. But in the target-date panel, because most of the points are identical and fall on top of one another, there appear to be dramatically fewer observations than in the other charts. Results for managed account investors are more dispersed than those for single target-date investors, reflecting the fact that the advice service varies asset allocation not only by age but also by other factors such as risk tolerance and prior holdings (Figure 7, Panel B). Managed account portfolios include active manager risk and, in some cases, limited exposure to company stock. Some managed account strategies may also take non-plan assets into account in which case the plan results shown here may be tailored around those non-plan holdings. 9 The greatest dispersion is among all other participants (Figure 7, Panel C). Why do we see such widely dispersed portfolio results? Reasons include asset allocations overweighted to specific subasset classes; active manager risk exposure; concentrated positions in employer stock; trading and rebalancing activity (or a lack thereof) over time; and differences in recordkeeping and investment fees across plans and fund options. For example, more dispersed total returns might reflect the fact that a participant is overweight in small-cap or emerging market stocks; has a large position in an actively managed equity fund; has all of the portfolio in money market or stable value assets; or has his or her portfolio concentrated in employer securities. 10 In reviewing these measures of dispersion, it s important to recall that during this period the equity risk premium was negative. In a more typical period, in which equity returns are higher than fixed income returns, it is likely that all measures of return will be more diffuse. For example, in such an environment, single target-date fund investors with high equity allocations will have higher returns than single target-date fund investors with low equity allocations unlike the current environment, in which equity returns have been lower than fixed income returns. That said, it s likely that, even in a different risk environment, the relative ranking of these results would remain the same, with single target-date investors having the more compact results, managed account investors having a somewhat more dispersed cloud around the target-date results, and all other participants having more scattered results. Finally, one important point to mention here is that the number of single target-date holders has grown significantly over the five-year period. These five-year results are backward-looking and reflect a period when, in our sample, only 1% of participants were in a single target-date fund for all five years. By comparison, at the end of 2010, one-fifth of all Vanguard participants were invested in a single target-date fund. Over time, the dispersion of all participant outcomes is declining as a result of strategies like target-date funds and managed accounts as well as traditional balanced funds. 8 Because administrative processing of accounts can lead to extreme returns, both positive and negative, we eliminated 0.5% of the top and bottom of both the risk and return distributions i.e., a total tail reduction of 2% leaving 980 observations in each panel of Figure 7. 9 The managed account service may also take into account other individual factors such as other benefit plans and plan contribution rates. 10 The dispersion for all other participants is even greater than the chart indicates because approximately 6% of participants in this group hold a balanced portfolio, which would have highly predictable results similar to those of target-date funds. Removing this group would make the remaining results even more diffuse. 8

9 Figure 7. Risk and return characteristics with benchmarks Defined contribution participants for the five-year period ended December 31, 2010 A. Single target-date participants 15% Five-year annualized total return 0% U.S. bonds U.S. stocks Non-U.S. stocks 15% 0% 10% 20% 30% 40% Five-year annualized standard deviation B. Managed account participants 15% Five-year annualized total return 0% U.S. bonds U.S. stocks Non-U.S. stocks 15% 0% 10% 20% 30% Five-year annualized standard deviation C. All other participants 15% 40% Five-year annualized total return 0% U.S. bonds U.S. stocks Non-U.S. stocks 15% 0% 10% 20% 30% Five-year annualized standard deviation 40% Participants Benchmarks Note: Includes random sample of 1,000 participant accounts drawn from respective samples for single target-date, managed account, and all other participants. Excludes 0.5% top and 0.5% bottom outliers for both risk and return, for a net sample of 980 observations. Source: Vanguard,

10 Regardless of the measure of dispersion, some participants more extreme portfolio construction strategies resulted in exceptionally positive results over the period and some resulted in very poor results. For many others, investment results were scattered, seemingly unpredictably, across the risk-return space. For plan fiduciaries, an important question to weigh is whether such dispersion of outcomes reasonably reflects individual participants desires for portfolio customization or their lack of skill at portfolio construction. Comparing average returns Over the five-year period, the average annual return for a single target-date participant was 3.93%, compared with 3.65% for the typical managed account participant and 3.76% for all other participants. This represents a relative performance advantage for the single target-date investor of 28 basis points compared with a managed account investor, and 17 basis points compared with all other participants. The range of returns reflects meaningful portfolio differences in asset allocation, subasset allocation, fund selection, rebalancing, and fees. For example, the target-date funds in our sample are overwhelmingly passive or indexed, while the managed account portfolios and portfolios of all other participants generally include active management for both equities and fixed income. Participants in each group also differ demographically. Single target-date fund participants tend to be younger, managed account participants tend to be older, and all other participants are distributed across many age groups. The return results for all other participants are particularly affected by concentrated company stock exposure. Single target-date investors have zero exposure to company stock, and some managed account participants may have limited exposure, but the all other participants group includes some investors with concentrated positions. To adjust for this factor, we estimated an ordinary least squares (OLS) regression of five-year returns, controlling for company stock and the type of investor (single target-date, managed account, all other). In this regression model (Figure 8), the difference in returns between managed account investors and investors in a single target-date fund remains at 28 basis points over five years, meaning that differences between the two strategies are unrelated to idiosyncratic company stock exposure. However, using this regression, the difference in returns between single target-date investors and all other participants widens to 40 basis points. In other words, part of the relative advantage of all other participants was related to exposure to nonsystematic or idiosyncratic risk of company stock, which contributed to returns over this period. Once we put target-date funds and all other participants on an equal, diversified portfolio footing by removing company stock differences, the return disadvantage for all other participants is 40 basis points. Figure 8. Regression adjusting for single-stock risk Five-year annualized returns relative to single target-date fund 0.5% 0.0% 0.5% Unadjusted 0.28% 0.28% Managed account Adjusted 0.17% 0.40% All other participants Note: Unadjusted effects are differences in five-year mean returns. Adjusted results are from OLS regression of five-year returns against fraction of portfolio in company stock and dummy variables for investment strategy. Source: Vanguard,

11 Return variation by age Over the five-year period, there were minor differences in returns among different age groups (Figure 9). This stands to reason, given the quite narrow differences in five-year returns in stock and bond market indexes. For example, participants under age 25 earned the highest median return of 4.18%; the lowest was 3.71% among participants ages The differences in returns are similar for three-year results, but widen over one year. For example, comparing one-year returns for those under age 25 with returns for those age 65 and older, the younger group gains an advantage of around 4 percentage points. Meanwhile, of note is the fact that the average participant approaching retirement, ages 55 64, earned a median return of 3.82%. This is equivalent to a cumulative growth of 21% in retirement wealth invested over the five-year period related solely to investment results. Another notable feature is that returns across income levels and account balances are also largely similar. In other words, there are not large return differences among more and less affluent participants. Age is an increasingly important determinant of participant investment strategies, particularly with regard to target-date and managed account advisory services. As a result, we examined our three participant strategies by age and various portfolio characteristics. Portfolio characteristics include equity exposure over the period, annual total returns, annual portfolio standard deviations, and Sharpe ratios (a measure of return per unit of risk). For each age group, we report on median results for the typical participant. In reviewing these results, keep in mind that results for all other participants are much more dispersed around the median than results for single target-date or managed account investors. Figure 9. Total returns by participant characteristics Defined contribution plan participants for the period ended December 31, year 3 years 5 years All participants (mean) 12.22% 0.11% 3.76% All participants (median) Age < % 1.03% 4.18% Household income <$30, % 0.49% 3.83% $30,000 $49, $50,000 $74, $75,000 $99, $100, Gender Male 12.53% 0.07% 3.84% Female Job tenure (years) % 0.33% 3.44% Account balance <$10, % 0.76% 3.79% $10,000 $24, $25,000 $49, $50,000 $99, $100,000 $249, $250, Note: Includes 2,623,615 participants for the 1-year period; 1,979,116 participants for the 3-year period; and 1,432,404 participants for the 5-year period. Average returns within category. Source: Vanguard,

12 Age profile of equity exposure. By design, for both single target-date and managed account strategies, median equity exposure follows a disciplined, declining profile over time (Figure 10, and Figure 11, Panel A). The reduction in equity exposure is steeper among single target-date investors than among managed account investors, but both age profiles are downwardsloping over time. Meanwhile, the equity allocation for all other participants is characteristically humpshaped by age. Younger participants tend to be less sophisticated or knowledgeable, and so tend to overweight conservative investments such as money market and stable value holdings. 11 Many younger investors likely enrolled during volatile market conditions, and so may have adopted more conservative allocations as a result. As shown here, and as demonstrated in other Vanguard research, young target-date fund investors maintain a meaningful commitment to equities, and do not demonstrate the overreaction to market volatility of other younger investors. 12 Realized returns. At younger ages, particularly among participants under age 45, single target-date investors and all other participants realized similar returns (Figure 10, and Figure 11, Panel B). But these returns likely arose from quite different strategies. Among all other participants, results were strongly influenced by stable value and money market returns; among target-date investors, results were more influenced by bond and international stock returns. Figure 10. Investment characteristics by age and investment strategy Defined contribution plan participants for the five-year period ended December 31, 2010 A. Equity exposure < All Single target-date fund 89.0% 90.0% 89.0% 76.0% 61.0% 40.0% 76.0% Managed account N/A All other participants All participants B. Annualized return Single target-date fund 3.79% 3.78% 3.62% 3.90% 4.42% 4.65% 3.90% Managed account N/A All other participants All participants C. Annualized standard deviation Single target-date fund 17.1% 17.2% 17.1% 14.8% 12.0% 8.4% 14.8% Managed account N/A All other participants All participants D. Sharpe ratio Single target-date fund Managed account N/A All other participants All participants Note: Includes 1,432,404 participants for the five-year period (16,869 single target-date, 6,685 managed account, and 1,408,850 other participants). Results for managed accounts under age 25 are excluded because of a neglible sample size. Medians within category. Source: Vanguard, Some participants may be in these conservative asset allocations because of default policies of the plan sponsor. Although many plans moved to a QDIA during the period, some portion of participants may have been defaulted to a conservative option or been indirectly influenced in making their own choice by the default designation in their plan. 12 Ameriks and Utkus,

13 Figure 11. Investment characteristics by age and investment strategy Defined contribution participants for the five-year period ended December 31, 2010 A. Median monthly equity allocation 100% B. Median annual total return 5.0% Equity allocation 80% 60% 40% 20% Five-year annualized return 4.0% 3.0% 2.0% 1.0% 0% < Age 0% < Age C. Median annual standard deviation 20% D. Median Sharpe ratios (all participants) 0.30 Annualized standard deviation 15% 10% 5% Sharpe ratio % < Age 0.00 < Age band Single target-date participants Managed account participants All other participants Note: Includes 1,432,404 participants for the five-year period (16,869 single-target date, 6,685 managed account, and 1,408,850 other participants). Results for managed accounts under age 25 are excluded because of a negligible sample size. Medians within category. Source: Vanguard, Among older participants, ages 55 64, single targetdate investors realized higher returns than other participants. The reason for this result is two-fold. First, target-date funds have a lower equity risk exposure than the investments older participants choose on their own (Figure 10, Panel A). Second, target-date funds use bonds for their fixed income holdings rather than stable value and money market funds, and also diversify stocks internationally, which were relatively beneficial strategies over the period. It is worth noting that the median five-year return for single target-date investors approaching retirement was 4.42%, equivalent to a cumulative return of 24% over the period. In other words, among single targetdate investors nearing retirement, retirement savings invested over the period grew by nearly a quarter, despite the market volatility. This growth is related to investment results alone and does not include the benefit of additional contributions. 13

14 Among all age groups, the managed account strategy generated somewhat lower median returns than a single target-date strategy over the period. This difference is attributable to the factors noted above, including asset allocation, subasset allocation, fund selection and rebalancing policies, and fees. For participants ages 55 and older, the managed account strategy outperformed the strategies of all other participants but underperformed the single target-date strategy. Portfolio risk and Sharpe ratios. For single target-date and managed account strategies, the five-year annualized standard deviation of returns is virtually identical by age (Figure 10, and Figure 11, Panel C). For all other participants, the risk profile reflects the typical hump-shaped exposure to equities. At younger ages, all other participants had somewhat higher Sharpe ratios reflecting the benefit of conservative assets like money market funds and particularly stable value assets during this volatile period (Figure 10, and Figure 11, Panel D). Participants ages appear to have benefited from the higher bond allocations and internationally diversified equity exposure in target-date funds. Pre-retirees in single target-date funds had somewhat better risk-return results over this period than their peers in that age group, whether those peers were in a managed account service or investing on their own. Implications Despite the historic equity market decline of , most participant portfolios earned positive investment returns over the period, and retirement wealth invested over the period grew by one-fifth for the average participant because of investment results alone. DC plan participants approaching retirement, including older single target-date investors, had similar or better results. These findings underscore the fact that retirement wealth can continue to grow over longer time horizons, despite extreme market declines in the short run. This account growth is related only to investment results and does not consider the additional benefit of ongoing contributions. One quite distinctive characteristic of participant portfolios is the tendency for self-directed portfolios to be highly dispersed in terms of both risk and return. By comparison, single target-date and managed account strategies result in a much narrower range of investment outcomes, consistent with their more disciplined investment approaches. For plan fiduciaries, the question is whether such highly dispersed outcomes are an inevitable result of participants desires to customize their portfolios or whether they reflect participants lack of skill at portfolio construction. Sponsors concerned about the dispersion of portfolio outcomes will want to consider not only target-date funds and advice services like managed accounts, but also the strategy of reenrollment into a QDIA. 13 It is again worth emphasizing that historic returns reflect investment results from a given set of market and investment conditions. Portfolio returns have to be evaluated in a variety of market settings, and are just one of the factors sponsors and participants should consider in evaluating investment portfolios over time. 13 See Burns, Utkus, and Combs, 2008; and Utkus and Mottola,

15 References Ameriks, John A. and Stephen P. Utkus Generations: Key drivers of investor behavior. institutional.vanguard.com Burns, David M., Stephen P. Utkus, and Ann L. Combs, Improving plan diversification through reenrollment in a QDIA. institutional.vanguard.com Clark, Jeffrey W. and Jean A. Young, Participation reaction to the August 2011 market sell-off. Vanguard Center for Retirement Research. institutional.vanguard.com How America Saves 2011: A report on Vanguard 2010 defined contribution plan data. Vanguard Center for Retirement Research. institutional.vanguard.com Utkus, Stephen P. and Gary R. Mottola, Reenrollment and target-date funds: A case study in portfolio reconstruction. Vanguard Center for Retirement Research. institutional.vanguard.com Utkus, Stephen P. and Jean A. Young The great recession and 401(k) plan participant behavior. Vanguard Center for Retirement Research. institutional.vanguard.com 15

16 P.O. Box 2600 Valley Forge, PA Connect with Vanguard > vanguard.com > global.vanguard.com (non-u.s. investors) Vanguard research > Vanguard Center for Retirement Research Vanguard Investment Counseling & Research Vanguard Investment Strategy Group > For more information about Vanguard funds, visit vanguard.com or call to obtain a prospectus. Investment objectives, risks, charges, expenses, and other important information about a fund are contained in the prospectus; read and consider it carefully before investing. An investment in a money market fund is not insured or guaranteed by the Federal Deposit Insurance Corporation or any other government agency. Although a money market fund seeks to preserve the value of your investment at $1 per share, it is possible to lose money by investing in such a fund. An investment in a stable value fund is neither insured nor guaranteed by the U.S. government. There is no assurance that the fund will be able to maintain a stable net asset value and it is possible to lose money by investing in the fund. All investing is subject to risk. Foreign investing involves additional risks, including currency fluctuations and political uncertainty. Past performance is no guarantee of future results. Investments in bond funds are subject to interest rate, credit, and inflation risk. Diversification does not guarantee a profit or ensure against loss in a declining market The Vanguard Group, Inc. All rights reserved. Vanguard Marketing Corporation, Distributor. CRRTRP