Time Diversification and Horizon-Based Asset Allocations

Vanguard Investment Counseling & Research

Executive summary. Time diversification, the phrase used to refer to the concept that investments in stocks are less risky over longer periods than shorter ones, has been the subject of spirited debate for decades. While the foundation of the time diversification debate was laid much earlier, three works published successively in 1994Mark Kritzmans article What Practitioners Need to Know . . . About Time Diversification, Jeremy Siegels book Stocks for the Long Run, and Paul Samuelsons article The Long-Term Case for Equities And How It Can Be Oversoldseem to have set the inflection point for the discussion that continues today. Some of the finest investment minds have participated on both sides, without providing a conclusion. Over the last few years the growing acceptance of life cycle investment products, such as target retirement mutual funds, has renewed interest in the topic. The objective of this paper is not to prove or disprove time diversification, but to evaluate whether the concept must be valid for a horizon-based asset-allocation framework to be viable and appropriate. Our findings suggest that there is little evidence to support the notion that time moderates the perceived volatility inherent in risky assets. However, we would expect the risk-reward relationships of the past to prevail in the future, and if that is the case, a longer investment horizon may support a willingness and ability to assume the greater uncertainty of equitycentric asset allocations. This may be true particularly for younger investors for whom the allocation to human capital and the risk posed by the erosion of purchasing power by inflation can reasonably be assumed to be greatest.

Author Donald G. Bennyhoff, CFA

Connect with Vanguard > www.vanguard.com

The debate: An overview

While this paper is not intended to provide an exhaustive review of time diversification literature, a brief summary of some of the more notable contributions is in order. Kritzman defined time diversification as the notion that above-average returns tend to offset below-average returns over long time horizons. However, he points out that while the annualized dispersion of returns moderates toward the expected mean, the dispersion of terminal wealth increases as investment horizon increases. He suggests that although the probability of losing money in stocks is lower over longer investment horizons than shorter ones, the size of the potential loss increases. In Stocks for the Long Run, Siegel stated that for investment horizons of at least 15 years, historical returns suggested that stocks produced positive real returns in excess of both bonds and Treasury bills (Figure 1). In addition, Siegel added that it is little known that in the long run, the risks in stocks are less than those found in bonds and even bills! [Siegels emphasis.] The books conclusions supported the commonly held notions that younger investors should favor a portfolio heavily weighted in stocks to capitalize on the equity risk premium relative to bonds and Treasury bills and that, over long enough horizons, this equity risk premium was reliable. Samuelsons The Long-Term Case for Equities And How It Can Be Oversold was published later in 1994 and rejected the premise that the risk of stocks decreased over longer time horizons. Not one to mince words, he stated that it is an exact theorem that the investment horizon can have no effect on your portfolio composition. Relying on utility theory, he said that investors want to maximize the utility of wealth, rather than expected return or terminal wealth. That is, investors should be interested in what happens to their wealth over time, not just at a point in time (such as at retirement). The practice of rationalizing high equity allocations for investors with longer investment horizons seems to have been common enough in the financial community to be considered oversold in Samuelsons opinion.

Figure 1. Minimum periodic annualized real returns, 19262006

10

0 Real return (%)

10

20

30

40
1 year 5 years 10 years 15 years 20 years 25 years 30 years

Equities

Bonds

Cash

Note: For stock market returns, we use the Standard & Poors 500 Index from 1926 to 1970, the Dow Jones Wilshire 5000 Index from 1971 through April 22, 2005, and the MSCI US Broad Market Index thereafter. For bond market returns, we use the Standard & Poors High Grade Corporate Index from 1926 to 1968, the Citigroup High Grade Index from 1969 to 1972, the Lehman LongTerm AA Corporate Index from 1973 to 1975, and the Lehman U.S. Aggregate Bond Index thereafter. For returns on cash investments, we use the Citigroup 3-Month Treasury Bill Index. Source: Vanguard Investment Counseling & Research.

Past performance is no 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.

After these works appeared, a number of academics and practioners joined the debate on either side, adding additional layers to this increasingly complex topic. Bodie (1995) used option pricing theory to illustrate how the cost of insuring against a stock return below the risk-free rate increased, rather than decreased, with longer contracts. Since higher option premiums suggested higher perceived risk for longer contracts, he concluded that time diversification was not evident.

A Note on Risk: Investments are subject to market risk. Investments in bonds contain interest rate, credit, and inflation risk. Target retirement funds are subject to the risks associated with their underlying funds.

2 > Vanguard Investment Counseling & Research

Reichenstein and Dorsett (1995) modeled two sets of return projections, one based on the random walk assumption that is common among detractors of time diversification and another based on mean reversion, common among its supporters. Both models supported the notions that it is reasonable for investors with longer investment horizons to have larger allocations to risky assets (such as stocks) and that a portfolios relative risk depends upon the length of the holding period.

Samuelsons utility argument is based on the standard finance assumption that investors are always risk averse and that therefore risk tolerance does not vary with wealth. However, he also notes that this may not always be the case and concedes that people may be less risk-tolerant in absolute terms when they face poverty than when they are affluent. The work of Kahneman and Tversky (1979) also concluded that investors do not display constant risk aversion, but rather, behave as if they are risk averse in particular settings but not in others. Kritzman suggested that irrationality was one reason that an investor might adopt a time-based investment strategy, but behavioralists might note that investors are neither irrational nor rational, but normal (Statman, 2005). Investors commonly suffer from the behavioral bias of recency, embracing recent performance in their return expectations. To the degree that those recent returns are negative, they may find it tempting to change allocations to moderate the portfolios exposure to whatever asset class is expected to continue to underperform.

Evaluating the assumptions

Opinions on time diversification depend on the assumptions used to frame them. Fisher and Statman (1999) evaluated not only the time diversification issue, but the assumptions often used to rebut the concept: Stock market returns follow a random walk pattern. Risk aversion does not change as wealth changes. Investors future wealth depends only on their investment portfolios. Fisher and Statman said that the mathematical truth of [Samuelsons] argument is that time diversification is false under specific assumptions. Since acceptance of these assumptions is not universal, then the debate will likely continue. But does it have to? While all three are strongly opposed to the suggestion that time diversifies risk, Kritzman, Samuelson, and Bodie also acknowledge that other considerations may justify incorporating an investment horizon when considering the risk postureand therefore the equity allocationof a portfolio. For example, contrary to the random walk assumption, an investor may believe that stock market returns follow a mean-reverting pattern, where periods of poor returns are followed by periods of better returns. Though a debatable topic itself, mean reversion would suggest that a longer investment horizon would better enable investors to weather adverse market outcomes if experienced at the outset of the investment strategy.

The returns for asset classeswhether for the short term or longerare uncertain, and as such, investors willing to bear this uncertainty should be compensated for doing so. This compensation is determined not for them, but by them: The collective actions of market participants, using pricing mechanisms founded on very different objectives, investment horizons, levels of risk aversion, and valuation processes, determine the risk premium at any particular time. It is this process in particular that makes it likely that the risk premium relationships established in the pastlong-term bonds yielding more than shortterm bonds of comparable quality, for example, or the expected returns for stocks being greater than for either bonds or cashshould reasonably be expected to hold for the future.

Vanguard Investment Counseling & Research > 3

The last assumption, that investors future wealth is determined by the return on their portfolios alone, may serve in theory but fails in practice. Few investors inherit a portfolio and rely solely on its returns for growth. For most, future wealth will be the result of diligent savings and investment over a lifetime. The concept of human capital, an asset whose returns take the form of wages and other compensation, plays an important role and will be addressed later in this discussion. The belief that time heals all wounds is certainly an illusion when it comes to investing, but it may benefit investors, nonetheless. If the risks borne by diversified equity investors are to be rewarded, as they have been historically, then the largest portion of the payoff for that risk might reasonably be expected to follow periods of market difficulty, when the risk of loss may be perceived to be great. Over time, risk premiums will vary with the current, as well as the expected, investment environment, and lower current prices suggest higher risk premiums and higher relative future returns. While it may be very difficult to capitalize on these shifting risk premiums through tactical allocation strategies (Tokat and Stockton, 2006), a cogent strategic asset allocation can serve as a logical anchor for investors, helping to keep them from reallocating their assets in response to volatility. The belief that a sell-off in equities is a short-term event in an otherwise long-term investment horizon may reinforce investors convictions in their strategic asset allocations, and may help them overcome a reluctance to rebalance their portfolios.

Is there an inverse relationship between investment horizon and risk, as some might suggest?
If one defines risk by the standard deviation of returns, then the measure of risk will diminish with longer horizons. The standard deviation of annualized holding period returns (see Table 1) for any one-, three-, or five-year period would reasonably be expected to be greater than those for any 15-, 20-, or 30-year period. That being said, while the deviation in mean holding period returns will moderate considerably as the horizon increases (Table 1, middle panel), the yearto-year volatility remains high and unaffected by the holding period (Table 1, lower panel). For investors to realize a 30-year holding period, they must first endure thirty one-year periods. So which better describes the volatility experienced over the average 30-year period: the 1.38% standard deviation for the full 30 years, or the 18.55% average standard deviation year by year? In large part, it is this disconnect between the expected lower risk of an investment in stocks over the long run and the expected higher risk of such an investment in the short run that creates doubt and can foster poor decision-making under stress. An allocation based on asset-class return expectations that seemed so reasonable at the onset of an investment may seem less palatable when the market turns down. A closer look at historical returns illustrates that year-to-year returns are commonly much different from those suggested by the past (Figure 2).

4 > Vanguard Investment Counseling & Research

Table 1. Average annualized holding period returns and standard deviations for U.S. stocks, 19262006
Average annual returns (%) for various holding periods
1 year 3 years 5 years 10 years 15 years 20 years 25 years 30 years

10.45

10.96

10.41

11.12

11.23

11.42

11.44

11.30

Standard deviation of average annualized returns (%) for various holding periods
1 year 3 years 5 years 10 years 15 years 20 years 25 years 30 years

20.20

11.53

8.45

5.20

4.39

3.34

2.30

1.38

Average standard deviation of annual returns (%) for various holding periods
3 years 5 years 10 years 15 years 20 years 25 years 30 years

18.41

19.60

19.15

18.68

18.51

18.47

18.55

Note: Stocks represented by the S&P 500 Index through 1970; the Dow Jones Wilshire 5000 Index from 1971 through April 22, 2005; and the MSCI US Broad Market Index thereafter. Source: Vanguard Investment Counseling & Research.

Figure 2. Returns are rarely average: Annual stock and bond returns, 19262006
30 25 20 Bond returns 15 10 5 0 5 50 40 30 20 10 0 10 20 30 40 50 60 Stock returns
Represents

Only 3 years between 1926 and 2006 had both an average stock return and an average bond return: 1926: 11.6% stock, 5.4% bond 1968: 11.0% stock, 4.5% bond 1992: 9.0% stock, 7.0% bond 19262006 bond return5.2% annual average return range (3% to 7%) achieved in 21 of 81 calendar years (26%)

19262006 stock return10.5% annual average return range (8% to 12%) achieved in 6 of 81 calendar years (7%)

each calendar year from 1926 through 2006 (81 points=81 years) plotted at the intersection of that years stock return and that years bond return. The vertical shaded area contains all years whose stock return was between 8% and 12%. The horizontal shaded area contains all years whose bond return was between 3% and 7%.

For stocks, returns reflect the S&P 500 Index through 1970 and the Dow Jones Wilshire 5000 Index thereafter. For bonds, returns reflect the Ibbotson Intermediate-Term Government Bond Index through 1972 and the Lehman Intermediate U.S. Treasury Index thereafter. Source: Vanguard Investment Counseling & Research.

A Note on Risk: Past performance is no 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.

Vanguard Investment Counseling & Research > 5

Other factors may warrant the use of an investment time horizon

Is it reasonable and appropriate for younger investors, or for other investors with longer investment horizons, to allocate a large portion of their portfolios to risky assets, particularly equities? For most investment strategies based on age or time, the longer the time horizon, the larger the relative weight of equities in the portfolio. Some horizon-based funds maintain a more or less static allocation; others, such as target retirement funds, moderate the equity allocation in a predictable manner as time passes and the target horizon approaches (Figure 3). Indeed, the investment horizon is considered by leading investment and financial planning professional associations to be a key factor in developing investment policy statements and asset allocations.1 It is also often incorporated in the portfolio construction of the mutual funds being offered in defined contribution plans. The horizon fund model for portfolio construction seems reasonable. To the extent that the historical perspective is representative of the risk-reward relationships of asset classes in the future, an investors retirement contributions could benefit greatly from a significant equity weight.2 Of course, while investors with longer investment horizons may have more time to recover from market setbacks, they also have more time to experience them, and that experience may be more negative than history would suggest. Not every investor is equally preparedeither emotionally or financially to contend with the uncertainty that comes with seeking returns above the risk-free rate.

Figure 3. Moderating equity allocations over time

100%

80 Equity allocation

60

40

20

0
25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

Age

Vanguard State Street Global Advisors Barclays Global Investors Fidelity

AlianceBernstein JP Morgan T. Rowe Price Schwab American Funds

Note: Allocations approximate the companies asset-allocation strategy. The current allocations of the funds may be different because some firms will modify their allocations in response to changing circumstances. Source: Vanguard and company websites.

1 For example, the CFA Institute and the Certified Financial Planner Board of Standards, Inc. 2 This is not to suggest that historical returns should be considered indicative of future returns, but to highlight that, absent philosophical or structural shifts in the financial markets, the rationale for risk premiums and the historical relationships between asset classes might be reasonably expected to continue into the future.

6 > Vanguard Investment Counseling & Research

The importance of human capital Often, a very long investment horizon suggests that an investor is in the early stages of building wealth. In that case, the investors portfolio is typically small relative to his future wealth, and the portion of his portfolio represented by capital contributions is large. As a result, in the early stages of investing, portfolio growth depends more on contributions than market returns.

An investor with a larger commitment to equities may find it easier psychologically to contend with a market sell-off when his portfolio is relatively small and is accumulating value than later in life when the portfolio is larger, particularly if the assets are being eroded by spending. For example, a 25% market correction would likely be painful for younger and older investors alike. However, while the younger investor, presumably with less wealth accumulated, could readily replace the lost capital with earned income, a similar loss by an investor in retirement might be life-altering, assuming that their equity allocations were the same. As previously defined, human capital represents an asset of uncertain value, but one whose potential value is generally assumed to be greater earlier in life than later, as retirement approaches. In addition, it is generally accepted that labor flexibilitythat is, peoples ability to choose where, how often, and how long to workis greatest when they are in the earlier stages of their careers. This flexibility may affect peoples investment and consumption habits, and may induce them to assume more investment risk, at any age in the life cycle (Bodie, Merton, and Samuelson, 1992). At retirement, if a person stops earning wages, the human capital asset expires, and labor flexibility is limited.

In this framework, human capital is often characterized as a bond-like asset whose value reflects the discounted value of residual lifetime income. (There are certainly exceptions to this generalization in which human capital may be more equity-like, such as for an investment banker whose compensation may ebb and flow with the tide of the equity markets.) If human capital is bond-like, then a younger investor may well be considered to be underweighted in equities, even if her investment portfolio consists entirely of equities. As the younger investor matures and human capital is paid down, the allocation to equities can reasonably be expected to moderate as well. This framework is consistent with the general asset-allocation changes in horizon-based investment strategies over time. More tangibly, earned income from human capital provides capital for investment, converting a theoretical asset into material wealth. As a result, earned income can provide an alternative source for wealth creation, above and beyond investment returns. Declines in the value of a portfolio may be offset though additional capital contributions from earned income. While this does not erase the loss itself, the reallocation of income from consumption to savings can erase the effect of the loss on the portfolios value. Without incomehuman capital the loss can only be erased through market returns, which are beyond the investors control. As a result, knowing that additional capital may be available if performance falters may provide some psychological support for younger investors who should know that losses of 20%, 50%, or even 75% or more are possible even in a well-diversified equity portfolio.3

3 While the worst annual loss for the Standard and Poors 500 Index was 45% in 1931, the cumulative loss for the index for the four years ended 1932 exceeded 80%.

Vanguard Investment Counseling & Research > 7

Table 2. Normal and inflation-adjusted returns, 19262006

Nominal Average annual return Percentage of years with negative returns Highest annual loss Average annual return Real (inflation-adjusted) Percentage of years with negative returns Highest annual loss

19262006 total returns

100% Treasury bills 100% bonds 100% stocks

3.8% 5.2 10.5

0% 9 30

0.0% 2.3 43.1

0.8% 2.1 7.1

35% 38 35

15.0% 14.5 37.3

Note: Stocks represented by the S&P 500 Index through 1970; the Dow Jones Wilshire 5000 Index from 1971 through April 22, 2005; and the MSCI US Broad Market Index thereafter. Bonds represented by the Ibbotson Intermediate-Term Government Bond Index through 1972 and the Lehman Intermediate U.S. Treasury Index thereafter. Cash represented by the Citigroup 3-Month Treasury Bill Index. Sources: Ibbotson Associates and Vanguard Investment Counseling & Research.

The impact of inflation Another compelling reason to correlate longer time horizons to larger equity allocations is inflation. As Siegel noted, historically, equities have provided higher real returns than bonds or cash alternatives. If the purpose of investing is to provide capital to fund future expenditures, such as where the portfolios assets must replace a substantial portion of wages during retirement, then time horizon should be an important consideration (Table 2).

shorter-term planning, the emphasis should be on the preservation of principal, because the inflationary impact is likely to be minor even if the liabilitys inflation rate is significantly underestimated.

Conclusion
The debate over time diversification has been longrunning and remains unresolved. However, there is little empirical evidence to support the claim that time moderates the risks inherent in risky assets. In actuality, a longer investment horizon increases the magnitude of potential outcomes, both negative and positive. That being said, other factors may warrant the consideration of an investment time horizon in the asset-allocation process. Risk premiums reflect the compensation for undertaking risk, on average, and it seems reasonable that the historical risk-return relationships will prevail over longer periods. This is not to suggest that the level of compensation received in the past will be received in the future or that an investor should expect to receive it over each and every time period measured; although no 20-year historical period has produced a negative cumulative return, there have been numerous annual periods within them that have.

For most investors, these liabilities are not of fixed value, but are subject to inflation. In addition, the inflation-rate assumption for the liability (or stream of liabilities) should often be one based on an investors unique circumstances, rather than that suggested by a common aggregate measure such as the Consumer Price Index (CPI). If the investors liabilities inflate at a rate well in excess of the CPI, or are underestimated, then the real returns from lower-risk assets alone such as money markets, short-term bonds, or Treasury inflation-indexed securitiesmay not preserve the required purchasing power of the portfolio. Therefore, for long-term planning, where the value of the liability stream is less certain, an allocation to assets with higher risk and higher real returns, such as equities, may be preferable and may increase the probability of maintaining the portfolios purchasing power. For

8 > Vanguard Investment Counseling & Research

Maybe more important for those portfolios with longer investment horizons, the historical equity risk premium has contributed to higher historical real returns for stocks than for bonds and cash, which may be critical for preserving the purchasing power of the portfolio. If the purpose of investing is to provide capital to fund future expenditures, then time horizon should be an important consideration. In other words, for investors with a long investment horizon, inflation risk may be of greater concern than capital risk, while for shorter-term investors, the opposite may apply. The availability of human capitalthrough wages or other compensationis also a significant factor for assessment. In essence, human capital provides a means for portfolio reparation, should the expected gain result in loss instead. Without human capital, such as in retirement, it would seem reasonable that, while investors might be willing to assume some risk, they would likely recognize that the downside risk in bonds or cash has been considerably less than for stocks. As a result, they might tend to moderate their exposure to risky assets. Regardless of the actual returns of various historical holding periods, the potential for a negative outcomeparticularly after considering the uncertain effect of future inflationremains. For investors with portfolios that have longer-term holding periods and incorporate historical risk and return relationships in their construction, understanding that significant losses are possibleand, arguably, probableis essential for maintaining a well-considered asset-allocation strategy over the expected investment horizon. Prudence would suggest that investors preferring more certain nominal outcomes should limit their exposure to risky assets. Conversely, a longer investment horizon may support a willingness and ability to assume the greater uncertainty of equity-centric asset allocations, particularly for younger investors for whom the allocation to human capital and the risk posed by the erosion of purchasing power by inflation can reasonably be assumed to be greatest.

References
Bodie, Zvi, 1995. On the Risks of Stocks In the Long Run. Financial Analysts Journal 51(3): 1821. Bodie, Zvi, Robert C. Merton, and William F . Samuelson, 1992. Labor Supply Flexibility and Portfolio Choice In a Life Cycle Model. Journal of Economic Dynamics and Control 16: 447 . Fisher, Kenneth L., and Meir Statman, 1999. A Behavioral Framework for Time Diversification. Financial Analysts Journal, May/June: 89-90. Kahneman, Daniel, and Amos Tversky, 1979. Prospect Theory: Analysis of Decision Making Under Risk. Econometrica 47(2): 263291. Kritzman, Mark, 1994. What Practitioners Need to Know . . . About Time Diversification. Financial Analysts Journal, January/February: 1418. Reichenstein, William, and Dovalee Dorsett, 1995. Time Diversification Revisited. Charlottesville, Va.: The Research Foundation of The Institute of Chartered Financial Analysts, p. 46. Samuleson, Paul A., 1994. The Long-Term Case for EquitiesAnd How It Can Be Oversold. Journal of Portfolio Management, Fall: 1718. Siegel, Jeremy J., 1994. Stocks for the Long Run. New York: McGraw-Hill, p. 29. Statman, Meir, 2005. Normal Investors, Then and Now. Financial Analysts Journal, March/April: 3136. Tokat, Yesim, and Kimberly A. Stockton, 2006. A Primer on Tactical Asset Allocation: Strategy Evaluation. Valley Forge, Pa.: Vanguard Investment Counseling & Research, The Vanguard Group.

Vanguard Investment Counseling & Research > 9

P .O. Box 2600 Valley Forge, PA 19482-2600

Connect with Vanguard > www.vanguard.com

For more information about Vanguard funds, visit www.vanguard.com, or call 800-662-2739, 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.
Vanguard, Connect with Vanguard, and the ship logo are trademarks of The Vanguard Group, Inc. CFA is a trademark owned by CFA Institute. All other marks are the exclusive property of their respective owners.

Vanguard Investment Counseling & Research

E-mail > research@vanguard.com
Ellen Rinaldi, J.D., LL.M./Principal/Department Head John Ameriks, Ph.D./Principal Joseph H. Davis, Ph.D./Principal Francis M. Kinniry Jr., CFA/Principal Roger Aliaga-Diaz, Ph.D. Frank J. Ambrosio, CFA Donald G. Bennyhoff, CFA Maria Bruno, CFP Scott J. Donaldson, CFA, CFP Michael Hess Julian Jackson Colleen M. Jaconetti, CPA, CFP Karin Peterson LaBarge, Ph.D., CFP Christopher B. Philips, CFA Liqian Ren, Ph.D. Glenn Sheay, CFA Kimberly A. Stockton David J. Walker, CFA

Investment products: Not FDIC-insured No bank guarantee May lose value

2008 The Vanguard Group, Inc. All rights reserved. Vanguard Marketing Corporation, Distributor. ICRTD 012008