Professionally managed allocations and the dispersion of participant portfolios

Transcription

1 Professionally managed allocations and the dispersion of participant portfolios Vanguard research August 2013 The growing use of professionally managed allocations in defined contribution (DC) plans is reducing the dispersion in portfolio outcomes among participants. Target-date funds are the dominant type of professionally managed allocation, but the category also includes traditional balanced options and managed account advisory services. Authors: John A. Lamancusa Stephen P. Utkus Jean A. Young Dispersion of returns. From , a period encompassing the global financial crisis, the median Vanguard participant earned an average annual total return of 2.3%, equivalent to a cumulative return of just less than 12%. However, the average annual return ranged from 1.7% per year at the 5th percentile to 6.3% per year at the 95th percentile, a range of just more than 8 percentage points. 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 target date funds is not guaranteed at any time, including on or after the target date. Connect with Vanguard > vanguard.com

2 Professional allocations and returns. The expanded use of professionally managed allocations is leading to reduced dispersion in participant outcomes. For participants in a single target-date fund during the period, realized average returns ranged from 1.7% per year at the 5th percentile to 3.7% per year at the 95th percentile, a range of 2 percentage points. For participants in a single balanced fund or in a managed account advisory service, the difference between the 5th and 95th percentiles was about 2 and 3 percentage points, respectively. Professional allocations and risk. Risk outcomes are also less dispersed for participants with professionally managed allocations. Risk levels among singletarget-date fund investors were less than half as variable as risk levels for participants making their own choices. Implications. The rapid growth in professionally managed allocations most notably, target-date funds, but also traditional balanced options and managed account advice services is contributing to a reduction in extreme risk and return outcomes for participants. It is also gradually mitigating concerns about the quality of portfolio decision-making within DC plans. Plan sponsors should consider greater adoption of these strategies to encourage better risk control and investment discipline in DC participant portfolios. Actions to consider include: offering target-date funds within the plan s investment menu, either on a voluntary or default basis; promoting their use in educational programs; introducing a managed account advisory service for participants seeking a customized portfolio strategy; and considering reenrollment into a plan s designated default option. 2

3 Background This paper continues our research on professionally managed allocations and their impact on participant portfolio construction in defined contribution (DC) plans. Professionally managed allocations are participant accounts where 10 of the balance is invested by a professional money manager. A single target-date fund is the most common type of professionally managed allocation, but the category also includes traditional balanced funds and a managed account advisory service. In effect, by choosing a professionally managed allocation (or by being defaulted into one), participants are delegating the complex task of portfolio construction to a third-party money manager selected by the plan sponsor. As of 2012, 36% of Vanguard participants were invested in professionally managed allocations, with a single target-date fund representing the largest part of that group at 27%. We estimate that more than half of participants will be in such allocations within five years, principally because of the growing use of target-date funds. 1 One reason for the growing interest in professionally managed allocations is their designation as qualified default investment alternatives under the 2006 Pension Protection Act. However, voluntary investment in these options by participants is also an important factor underlying their expansion. This paper examines DC participant portfolio outcomes for the five-year period ended December 31, We confirm two main findings from our prior papers on this topic. 2 First, in the period which includes the years of the global financial crisis the typical participant earned a small but positive investment return. (When contributions are also included, median account balances grew by double-digit rates. 3 ) However, returns were highly dispersed, with the difference between the 5th and 95th percentile spanning some 8 percentage points per year. Second, risk and return outcomes for participants in professionally managed allocations were substantially less dispersed than for participants making their own choices. We believe that this reduction in dispersion of outcomes is attractive to plan fiduciaries, as it demonstrates both improved investment discipline and risk control in participant portfolios. This result also is gradually mitigating concerns about the quality of investment choices being made by inexperienced plan participants. After a brief review of methodology and the market environment, the paper presents several measures of participant-realized returns and risk characteristics over various periods. Caveats As in our prior papers, there are several caveats to consider when reviewing our results. First, realized total returns reflect investment results for a specific period, and are conditional on the market and investment conditions prevailing during that time. The three-year and five-year periods ended December 31, 2012, were quite different in terms of the realized equity risk premium. 4 Second, total returns over the period reflect the effect of several portfolio characteristics, including asset allocation, sub-asset allocation, fund selection, rebalancing, and fees. Given the range of factors influencing portfolio outcomes, results across participants, and results across different strategies, should be expected to vary over future periods. 1 Vanguard, Utkus and Bapat, 2011; Utkus and Bapat, Vanguard, The equity risk premium here is defined as is the difference between the return on and the return on government or high-grade bonds. 3

4 Methodology Our data is drawn from Vanguard s recordkeeping systems encompassing more than 3 million DC plan participants. Total returns were calculated monthly for each participant account over the 60 months ended in December Total returns reflect investment performance only, and do not include the effect of ongoing contributions. Total returns are net of all fees and expenses including investment expenses of the fund options held by the participants, as well as any separate account-based recordkeeping charges (such as per capita recordkeeping fees). When presenting one-, three-, and five-year returns, we utilize a dataset of all participants with corresponding one-, three-, or five-year account histories. This data set includes 2.99 million participant accounts with a one-year history, 2.34 million participants with a three-year history, and 1.58 million participants with a five-year history. 6 Also, company stock as a portfolio holding creates a higher return volatility. The most extreme risk and return outcomes either negative or positive are typically associated with concentrated company stock positions. Just less than 1 of the five-year data set includes participants with a concentrated position in employer securities defined here as a position in company stock that exceeds 2 of the participant s account balance. Because participants with concentrated holdings 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 markets at first fell precipitously during the global financial crisis (Figure 1). Stock prices subsequently recovered, and by year-end 2012, stock investments were almost 12% higher than the level at year-end 2007 (as calculated on a total return basis, including reinvested dividends). International improved but remained more than 1 below their December 2007 level. Bond investments rose steadily due to both reinvested interest income and rising bond prices (due to falling interest rates during the period). As a result, average annual returns were quite different during the three- and five-year periods ended December 31, 2012 (Figure 2). During the five-year period, returned 2.27% per year while bonds returned 5.99% a negative equity risk premium of 3.72% per year. During the three-year period, gained 11.34% compared with 6.26% for bonds, a positive equity risk premium of 5.08%. bonds offered the highest cumulative return during the five-year period, with a dollar at the beginning of the period growing to $1.34. A dollar invested in grew to only $1.12, while $1.00 invested in international fell to $0.86 during the period. Three-year cumulative returns for were much higher than for bonds. 5 Total returns are calculated for Vanguard participant portfolios from cash flow to cash flow, and then linked, in accordance with global investment performance standards. 6 One- and three-year returns are not materially different if they are presented only for the five-year panel of 1.58 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 1. Market performance, Cumulative total return for select market indexes 160 Index value Dec 07 Jun 08 Dec 08 Jun 09 Dec 09 Jun 10 Dec 10 Jun 11 Dec 11 Jun 12 Dec 12 Treasury bills bonds International Note: Indexes represented by the following benchmarks: Cash reserves Citigroup T-bill Index; bonds Barclays Aggregate Bond Index; MSCI Broad Market Index; and international MSCI All Country World Index ex-us Index. All values set at 100 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. Figure 2. Market returns and volatility For the period ended December 31, 2012 Annualized Cumulative Annualized total returns total returns volatility* Asset class Index year years years year years years months months Cash reserves Citigroup T-bill Index 0.09% 0.08% 0.44% 0.1% 0.3% 2.2% % bonds Barclays Aggregate Bond Index MSCI Broad Market Index International MSCI All Country World Index ex-us Index Note: Indexes represented by the following benchmarks: Cash reserves Citigroup T-bill Index; bonds Barclays Aggregate Bond Index; MSCI Broad Market Index; and international MSCI All Country World Index ex-us Index. * Standard deviation of monthly returns over the 36- or 60-month period ended December 31, Sources: Citigroup, Barclays, MSCI, Vanguard calculations. 5

6 Participant total returns The average Vanguard DC plan participant earned an average annual return of 2.3% and a cumulative return of 11.9% over the period (Figure 3). In effect, for the average participant, investment results alone boosted retirement wealth by just less than12%. Figure 3. Participant total returns Defined contribution plan participants, for the period ended December 31, % During the three-year period ended 2012, the average participant realized an average annual return of 7.8% and a cumulative three-year return of 25.1%, reflecting the sharp increase in stock prices over the period. On a one-year basis, the average participant earned 12.4% during % 12.4% 7.8% 11.9% Dispersion of returns One of the features of participant-constructed portfolios is the tendency of many participants to adopt unconventional or even extreme investment allocations. Some participants may invest their entire portfolio in equities, while others invest exclusively in low-risk assets, such as money market or stable value funds. Participant portfolios may also be concentrated in a single fund or 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. Over the five-year period, while the average participant gained 2.3% per year, a participant at the 5th percentile earned 1.7% per year, while a participant at the 95th percentile gained 6.3% per year a difference of 8 percentage points per year (Figure 4). On a cumulative basis, a $1.00 investment fell to $0.92 for a participant at the 5th percentile, but rose to $1.36 for a participant at the 95th percentile. Similarly, over a three-year period, the 5th percentile participant gained 1.6% per year, while the 95th percentile participant rose 11.5% per year. On a cumulative basis, the 5th percentile investor saw a $1.00 investment grow to $1.05, while the 95th percentile investor saw a $1.00 investment grow to $1.39. These statistics illustrate the wide variation in portfolio construction across participants. They also demonstrate that during the five- and three- year periods, very few participants had a negative return. In particular, we found that only 1 in 10 participants had a negative five-year return and only 1 in 100 had a negative 3-year return. Note: Includes 2,991,291 participants for the 1-year period; 2,337,819 participants for the 3-year period; and 1,583,839 participants for the 5-year period. Past performance is not a guarantee of future results. Source: Vanguard, Figure % Average annual total returns 1 year 3 years 5 years Average cumulative total returns Distribution of participant total returns Defined contribution plan participants for the period ended December 31, % 17.6% 15.6% 12.4% 9.2% 11.5% 9.2% 7.8% 6.8% 1.5% 1.6% 1 year 3 years 5 years 6.3% 3.3% 2.3% 1.2% -1.7% 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 2,991,291 participants for the 1-year period; 2,337,819 participants for the 3-year period; and 1,583,839 participants for the 5-year period. Source: Vanguard,

7 Effect of professionally managed allocations Participant portfolio strategies in DC plans are being reshaped by the growth of professionally managed allocations including target-date funds, traditional balanced funds, and managed account advisory services. Target-date funds are being offered within plans to streamline portfolio choice by participants; they are also a common default fund for automatic enrollment. Traditional balanced funds, whether offered on a stand-alone basis or as part of a riskbased series of life-cycle funds, play a similar role in a smaller fraction of plans. Managed account advisory services are also being increasingly offered within plans, typically on a voluntary choice basis. Managed accounts provide customized portfolio recommendations, typically for an additional fee, with a third-party advisor taking discretionary control over the participant s account. For each type of professionally managed allocation, we separately estimated portfolio return and risk characteristics over three- and five-year periods ended December 31, Under our approach, a participant in a particular strategy (i.e., in a singletarget-date or balanced fund or in the managed account service) had to maintain this strategy over the full three- or five-year period. Participants in these three types of allocations were compared with all other participants in essence, participants making their own portfolio construction decisions. 7 Examining returns over the five-year period, outcomes for single-target-date investors ranged from 1.7% per year for the 5th percentile to 3.7% for the 95th percentile, a difference of approximately 2 percentage points (Figure 5). For the single balanced fund and managed account participants, the 5th-to-95th percentile difference was approximately 2 and 3 percentage points. (Singlebalanced-fund investors also had generally higher returns over this period because of their larger fixed income exposure.) Figure 5. Distribution of five-year returns by strategy Defined contribution plan participants, five-year annualized total returns for the period ended December 31, Single target-date fund 3.7% 2.3% 2.2% 1.7% 1.7% Single balanced fund 4.3% 3.5% 3.3% 3.1% 2.3% Managed account 3.6% 2.4% 1.9% 1.2% 1.2% 0.2% All other participants 6.4% 3.3% 2.3% -1.9% 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: For the five-year period, includes 24,889 single target-date fund participants, 35,992 single balanced fund participants, 28,848 managed account participants, and 1,494,110 all other participants, for a total sample of 1,583,839 participants. Source: Vanguard, By comparison, among all other participants, those making choices on their own realized returns ranged from 1.9% per year for the 5th percentile to 6.4% for the 95th percentile, a difference of more than 8 percentage points. (Keep in mind that these differences would be wider in a period with strong equity returns.) 7 Over the three-year period, the sample size includes 90,171 single-target-date fund participants, 60,892 balanced fund participants, 66,822 managed account advisory service participants, and 2,119,934 all other participants. Over the five-year period, the sample size includes 24,889 single-target-date fund participants, 35,992 balanced fund participants, 28,848 managed account advisory service participants, and 1,494,110 all other participants. 7

8 Dispersion of outcomes over three years Differences among participant portfolios are particularly apparent when examining both return and risk outcomes in scatterplots. For ease in presentation, we created a random sample of approximately 1,000 participants for each group of investors. 8 During the period, which was characterized by a positive equity risk premium, outcomes for single-target-date investors were distributed among major market indexes (Figure 6, Panel A). These results are consistent with the fact that most of the target-date portfolios in our sample are a specific combination of indexed equities, international equities, and bonds. In the targetdate scatterplot (in Panel A), younger participants (represented by blue dots and in long-dated portfolios) are to the right of the chart; older participants (represented by purple dots and in near-dated portfolios) are to the left. The figure includes just less than 1,000 observations, although there appear to be far fewer. The reason is that while there are many observations in our sample, they are all 10 invested in a limited set of target-date portfolios, which means that portfolio outcomes are also limited. For example, if a plan offered 12 target-date options, then 1,000 participants invested solely in a single target-date fund would have 12 outcomes, not 1,000. The results for single balanced fund investors reflect the fact that most balanced funds have similar equity allocations, typically around 5 to 6 of assets (Figure 6, Panel B). Again, although there are nearly 1,000 participants in the scatterplot, their portfolios are confined to the investment results of the limited number of balanced funds offered by the plans in our data set. Managed account investors are more dispersed. A managed advice service varies asset allocation not only by age but also other factors such as risk tolerance and prior holdings (Figure 6, Panel C). We refer to this result as the cloud of customization. Managed account portfolios also include active manager risk and, in some cases, limited exposure to company stock, contributing to the cloud effect. Some participants may also incorporate their other assets in the service in which case the plan results shown here may be tailored to fit these other holdings. The managed account service may also consider other individual factors such as other benefit plans and plan contribution rates. Again, younger participants represented by blue dots cluster to the right of the cloud, and older participants represented by purple dots cluster to the left of the cloud. The greatest dispersion of risk and return outcomes is among all other participants (Figure 6, Panel D). Here age groups are scattered throughout the chart. Why do we see such dispersed portfolio results? One reason is differences in equity exposure. Outcomes on the far left are for portfolios invested exclusively in conservative options such as money market and stable value funds; outcomes on the far right represent all-equity portfolios. Other reasons for the dispersion of results include asset allocations overweighted to specific sub-asset 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. 8 Because certain types of recordkeeping processes contribute to erroneous and extreme returns, both positive and negative, we eliminated half a percentage point from the top and bottom of both the risk and return distributions i.e., a total tail reduction of 2% leaving 980 observations in each figure. 8

9 Figure 6. Risk and return characteristics, Defined contribution plan participants for the three-year period ended December 31, 2012 A. Single-target-date participants B. Single balanced fund participants 2 2 Three-year annualized total return bonds Non- Three-year annualized total return bonds Non- 5% Three-year annualized standard deviation 5% Three-year annualized standard deviation C. Managed account participants D. All other participants 2 2 Three-year annualized total return bonds Non- Three-year annualized total return bonds Non- 5% Three-year annualized standard deviation 5% Three-year annualized standard deviation Younger than 35 Ages 35 to 55 Older than 55 Note: Includes 1,000 random samples of participant accounts drawn from respective samples. Excludes 0.5% top and 0.5% bottom outliers for both risk and return, for a net sample of 980 observations. Past performance is not a guarantee of future results. Source: Vanguard,

10 Dispersion of outcomes over five years In contrast to the upward-sloping results for the three-year period, outcomes for the full five-year period, , are generally downward-sloping, reflecting the negative equity risk premium prevailing during that period. Even in this quite different market environment, similar patterns hold. Single-target-date fund holders (Figure 7, Panel A) span the risk/return spectrum, reflecting the age-based pattern of target-date investing. They are situated among major market indexes in a disciplined way. Single balanced fund investors (Figure 7, Panel B) are again clustered in a smaller group, reflecting the narrower range of allocations for those funds. The results for managed account advisory participants (Figure 7, Panel C) again demonstrate the cloud of customization. During this period, the data points at the extremes of the cloud represent the effects of customization of as well as sponsor fund selection. At the far left are accounts dominated by fixed income holdings, while at the far right are accounts with large positions in active international equities. Finally, the results for all other participants reflect yet again the greater dispersion in outcomes for participants making their own allocation choices (Figure 7, Panel D). Over time, the dispersion of all participant outcomes is declining as a result of the rapid growth in the use of target-date funds. Traditional balanced fund options and managed account advisory services are also expected to continue to grow, although at a slower pace. Yet for plan fiduciaries, these results underscore the wide dispersion in results that exists when participants make their own choices. An important question for fiduciaries to consider is to what extent such dispersion reasonably reflects participant desires for portfolio customization or reflects a lack of skill at portfolio construction. 10

11 Figure 7. Risk and return characteristics, Defined contribution plan participants for the five-year period ended December 31, 2012 A. Single-target-date participants 2 B. Single balanced fund participants 2 Five-year annualized total return bonds Non- Five-year annualized total return bonds Non- 5% Five-year annualized standard deviation 5% Five-year annualized standard deviation C. Managed account participants D. All other participants 2 2 bonds Five-year annualized total return bonds Non- Five-year annualized total return Non- 5% Five-year annualized standard deviation 5% Five-year annualized standard deviation Younger than 35 Ages 35 to 55 Older than 55 Note: Includes 1,000 random samples of participant accounts drawn from respective samples. Excludes 0.5% top and 0.5% bottom outliers for both risk and return, for a net sample of 980 observations. Past performance is not a guarantee of future results. Source: Vanguard,

12 Risk dispersion measures We also calculated a series of metrics examining participant risk exposures and the dispersion of those risk exposures across participants. The mean volatility level (Figure 8, Panel A) of participants in our different groups varies predictably. The average risk level for the single balanced fund options was lowest among the four groups, at 8% per year over three years and 11% per year over five years. Average risk levels were somewhat higher for managed account and all other participants, at 12% or 15% depending on the period. They were highest among single-target-date investors, at 13% or 16% depending on the period, because young investors with higher equity allocations dominate this group. How much do risk levels vary from participant to participant? The standard deviation of portfolio risk levels (Figure 8, Panel B) measures the dispersion of risk across participants in a given group. Among professionally managed allocations, the standard deviation in the value of risk levels was 2% or 3% over each period. However, among participants making choices on their own, the standard deviation was 6% three times as high as the professionally managed allocations. Figure 8. Risk and risk dispersion measures A. Mean portfolio risk Single target-date fund 13% 16% Balanced fund 8 11 Managed account advisory service All other participants B. Standard deviation of portfolio risk Single target-date fund 3% 3% Balanced fund 2 2 Managed account advisory service 2 2 All other participants 6 6 C. Coefficient of variation of portfolio risk* Single target-date fund 2 17% Balanced fund Managed account advisory service All other participants * The standard deviation of portfolio risk (Panel B) divided by its mean (Panel A). Source: Vanguard, The coefficient of variation (Figure 8, Panel C) scales the standard deviation across participants by the mean risk level and permits normalized comparisons across our groups of investors. Over the three- and five-year periods, the coefficient of variation for all other participants is nearly three times that of single-target-date investors. In other words, participants constructing portfolios on their own tend to choose risk levels that are three times as extreme as those selected by a target-date fund manager. These higher risk levels reflect the factors noted earlier in this report more extreme asset allocations, differences in sub-asset class and fund selection, concentrated fund or single stock risk, plus the effects of rebalancing, fees, and other factors. 12

13 Implications One of the distinctive characteristics of participant portfolios is the tendency for self-directed portfolios to be highly dispersed in terms of both risk and return. By comparison, professionally managed allocations result in a narrower range of investment outcomes, consistent with a more disciplined investment approach. Target-date funds are the dominant type of professionally managed allocations, but the category also includes traditional balanced funds and managed account advisory services. For plan fiduciaries, the question is whether such dispersed outcomes are an inevitable result of participants desires to customize their portfolios or whether participants appear to lack skill at portfolio construction. Sponsors concerned about the dispersion of portfolio outcomes will want to consider a number of strategies: offering target-date funds (or traditional balanced funds) on a default or voluntary basis within the plan investment menu; offering a managed account advisory service and promoting its use; and using the strategy of reenrollment into a qualified default investment alternative (QDIA) as a way to make rapid changes to participant investment holdings. 9 More broadly, the growing use of target-date funds and other professionally managed allocations is changing the investment decision-making within DC plans. Participants are increasingly not making complex investment choices on their own, but are instead delegating this responsibility back to the employer and, in particular, to the employerdesignated money manager or third-party advice provider. Such a development is resulting in improved patterns of diversification in participant portfolios, and is gradually mitigating concerns about errors in decision-making by inexperienced plan participants. 9 See Vanguard, 2012b, and Mottola and Utkus,

14 References Utkus, Stephen P. and Shantanu Bapat, Participants during the financial crisis: Total returns Vanguard Center for Retirement Research, Malvern, PA. com/vgapp/iip/site/institutional/researchcommentary/ article/invresduringcrisis Utkus, Stephen P. and Shantanu Bapat, Target-date funds and the dispersion of participant portfolios. Vanguard Center for Retirement Research, Malvern, PA. pdf/dispp.pdf Utkus, Stephen P. and Gary R. Mottola, Reenrollment and target-date funds: A case study in portfolio reconstruction. Vanguard Center for Retirement Research, Malvern, PA. institutional.vanguard.com/vgapp/iip/site/institutional/ researchcommentary/article/retrestdf. Vanguard, Improving plan diversification through reenrollment in a QDIA. Vanguard Strategic Retirement Consulting, Malvern, PA. institutional.vanguard.com/vgapp/iip/site/institutional/ researchcommentary/article/invcomreenrollqdia. Vanguard, How America Saves 2013: A report on Vanguard 2012 defined contribution plan data. Vanguard Center for Retirement Research, Malvern, PA. HAS13.pdf. 14

15 Page intentionally left blank. 15

16 P.O. Box 2600 Valley Forge, PA Connect with Vanguard > institutional.vanguard.com Vanguard research > Vanguard Center for Retirement Research Vanguard Investment Strategy Group > 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. All investments, including a portfolio s current and future holdings, are subject to risk. Bond funds are subject to the risk that an issuer will fail to make payments on time, and that bond prices will decline because of rising interest rates or negative perceptions of an issuer s ability to make payments. Investments in or bonds issued by non- companies are subject to risks including country/ regional risk and currency risk. Diversification does not ensure a profit or protect against a loss. Past performance is not a guarantee of future results The Vanguard Group, Inc. All rights reserved. Vanguard Marketing Corporation, Distributor. CRRDPPP