The Lost Decade : Rational Expectations in Uncertain Markets

The Lost Decade : Rational Expectations in Uncertain Markets Vanguard Investment Counseling & Research The 2 22 bear market was booked six years ago, but it s just now making a big dent in trailing returns. Authors Francis M. Kinniry Jr., CFA Christopher B. Philips, CFA As the high-flying 199s disappear from ten-year-return calculations, the 2 22 bear market is looming larger in the stock market s trailing performance. In the years ahead, trailing stock market returns could be lower than those of bonds and cash, contrary to expectations. The risk is that investors will base future asset allocation decisions on these date-dependent snapshots. The last ten years have merely provided the most recent demonstration that, while it is reasonable to expect riskier asset classes to outperform less-risky ones over the longer term, they do not always do so. And this fact is why it is reasonable to expect a risk premium when investing in stocks. Connect with Vanguard > www.vanguard.com

Introduction To judge from recent commentary, it is only now dawning on investors that stocks do not always outperform bonds, even over relatively long periods. For the ten years ended June 3, 28, the broad U.S. stock market returned 3.53%. The broad bond market returned 5.68%. 1 When the first decade of the new millennium comes to a close, the numbers could look worse for stocks. Have investors been living through a lost decade, as the Wall Street Journal recently put it? 2 Most investors acknowledge that stock returns can be especially volatile from year to year. Far fewer recognize that the same is true, though to a lesser extent, over ten- and even twenty-year periods. Table 1. Change in data series created by date shifts, expressed as percentage of the prior series Shift in years Time series 2 4 6 years 4% 8% 12% 15 years 3 6 8 2 years 2 4 6 An unreliable anchor The returns from any particular period are an unreliable anchor for long-term return expectations. Even over relatively extended horizons such as ten and twenty years, investment results can be highly date-dependent. At the end of 22, after the worst bear market in 7 years, the stock market was able to boast an average 8.74% annual gain over the previous decade. By the end of 24, the ten-year average had increased to 11.92%. Within two years, in other words, a measure widely regarded as a reliable benchmark for intermediate- to long-term stock market performance had changed by nearly 32 basis points. By the end of 26, though, the ten-year average was back to 8.63%. Table 1 highlights an important, if often overlooked, source of this long-term return volatility. Small shifts in the start and end dates of a time series produce significant changes in the composition of the data. For example, a two-year change in start and end dates alters 4% of the data points used in a tenyear data series, by eliminating 2% of the start points and adding 2% worth of new end points. A six-year shift eliminates enough start points and adds enough new end points to equal 12% of the original ten-year series. Notes on risk: Investments are subject to risk. Investments in bonds are subject to interest rate, credit, and inflation risk. 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. 1 Unless specified otherwise, all returns cited are based on these benchmarks: for stocks, the Standard & Poor s 5 Index from 1926 through 197, the Dow Jones Wilshire 5 Index from 1971 through April 22, 25, and the Morgan Stanley Capital International US Broad Market Index thereafter; for bonds, the Lehman Brothers U.S. Aggregate Bond Index; for Treasury bills, the Citigroup 3-Month Treasury Bill Index. 2 E.S. Browning, 28, Stocks Tarnished by Lost Decade, Wall Street Journal, March 26. 2 > Vanguard Investment Counseling & Research

The volatility in longer-term averages suggests that the most valuable information provided by trailing returns is not their means but the range of results. This is true for both historical and forwardlooking analysis. Equities as an asset class are highly unpredictable in the short run. Indeed, the distribution of annual returns for stocks is very wide, with a standard deviation of approximately 2% (illustrated in Figure 1). This means that, given a mean return of approximately %, investors can expect to see a one-year return between 3% and +5% about 95% of the time. This wide variability in yearly returns is what makes longer-term averages so volatile. Since 1926, tenyear returns for equities have displayed a standard deviation of approximately 5%. The mean annual return for all ten-year periods has been approximately %; therefore, investors can reasonably expect to see a ten-year return average between % and 2% about 95% of the time. Figure 1. Hypothetical distributions of 1- and -year returns -year 1-year 2 Standard deviations +2 Standard deviations 3% % % 5% % 15% 2% 3% 5% 1 Standard deviation Mean +1 Standard deviation Vanguard Investment Counseling & Research > 3

An unusual period of 18 years Much of this predictable volatility did not materialize from 1982 to 1999, which might explain why the recent return to normality has startled many investors. From 1982 to 1999, in fact, the stock market experienced only one down year, returning 2.4% in 1994. The result was a more or less one-directional return series that produced unusually stable ten-, fifteen-, and twenty-year returns. The historical norm is a negative return every four years, on average. Since 1999, the U.S. stock market has climbed back toward historical levels of volatility and beyond. From the start of 2 through 22, the market declined almost 4% (more than 14% a year), charting its worst three-year performance since the 193s. This decline has been one source of the recent uptick in the volatility in long-term return averages, as charted in Figure 2. A less-obvious contributor has been the replacement of the exceptional returns from 1995 to 1997 with strong, but lower, returns from later years. Consider some recent snapshots of historical performance: 1. Despite a two-and-a-half-year bear market, stocks returned 8.74% a year, on average, during the decade ended December 31, 22. 2. Over the next five years 23 through 27 stocks averaged a gain of 13.96% each year. The trailing ten-year average increased at first, reaching 11.92% at the end of 24. Figure 2. What bear market? Until lately, the averages obscured what happened in 2 22 Ten-year average annual returns for U.S. stocks: Return 12% 8 6 4 2.59% 11.92% 9.16% 23 24 25 26 Years ended December 31 8.67% Note: Stocks are represented by the Dow Jones Wilshire 5 Index from 1993 through April 22, 25, and the MSCI US Broad Market Index thereafter. The ten-year average declined over that three-year stretch even as the stock market rallied from the bear market s bottom because 1995, 1996, and 1997 dropped out of the calculation. These years produced an average gain of 29.49%. The market s mid- to late- 199s peak, in other words, has cast a long shadow. 3. But by the end of 27, the stock market s ten-year return had declined to an annual average of 6.29% a change of 563 basis points in the space of three years. 4 > Vanguard Investment Counseling & Research

Figure 3. Recent return patterns and headlines echo the past Changes in the S&P 5 indexed to as of December 31, 1969, S&P 5: Logarithmic scale 1, Up, but mostly flat, from 197 to 198 Down, but mostly flat, from 2 to April 27 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 21 23 25 27 Sources: Vanguard and Thomson Datastream. Year Found: The last lost decade The recent talk of a lost decade is déjà vu for seasoned investors, who remember the headlines proclaiming the death of equities after the difficult decade that ended in 1979. 3 In fact, these periods share many similarities, not only in their headlines but in the return patterns that led to the headlines. For example, in Figure 3 we chart the price of the Standard & Poor s 5 Index since 1969. Immediately apparent are a pair of weak periods bracketing the strong bull market of the 198s and 199s. Each of these two lost periods is characterized by a significant bear market: 1973 1974 in the first case and 2 22 in the second. In addition, each displays a steady march upward in stock prices following the trough of the bear market. And yet, in the 197s, that steady ascent wasn t enough to erase the bear market by the decade s end, and the same is likely to be true for the ten years ending 29. Should investors therefore pay attention to the lost headlines? Should they be paring down their equity holdings, or perhaps abandoning stocks altogether in favor of other asset classes? As Figure 3 shows, these would have been poor strategies to adopt at the end of the 197s. As for the current decade, only time will tell whether U.S. stocks will rebound into a bull market similar to that of the 198s and 199s, or will sag into a protracted bear market such as Japan has endured since the 9s or will experience something different from either. The view backward may get gloomier The math of trailing returns is straightforward and uncomplicated; however, the implications are significant. When short-term volatility leads to longer-term returns that deviate significantly from the historical average, investor psychology can play an important role in portfolio decisions. There s a risk that trailing returns will unduly influence investors expectations for future performance, potentially leading to suboptimal asset allocation decisions. Further, our analysis indicates that, because of the 2 22 bear market, trailing returns could remain subdued for another ten years or more. 3 Death of Equities, 1979, Business Week, August 13. Vanguard Investment Counseling & Research > 5

Of course, ten years may not be considered truly long-term, as many investors have investment horizons of well over twenty years. However, the significant disparity in returns from the late 199s to the first three years of the 2s also has implications for these longer periods. To evaluate the potential impact on twenty-year returns, we assumed that the market would deliver a normalized return averaging 9.5% 4 for each year from 28 through 219. As Figure 4 illustrates, based on our assumption, the ten-year trailing return would fall to an average 3.76% by the end of 29. Declining more slowly, the twenty-year average annual return would reach 6.59% by 219. Figure 4. What if stocks return only 9.5% a year? A look at the hypothetical possibilities Figure 4a. 1-year returns: Actual returns through 27; assumed 9.5% annual return thereafter 4 3 Total return (%) 2 2 Actual Hypothetical 3 1993 1995 1997 1999 21 23 25 27 29 211 213 215 217 219 Year Figure 4b. - and 2-year average annual returns: Based on actual returns through 27 and assumed return of 9.5% yearly thereafter 25 17.59% Total return (%) 2 15 8.74% 6.29% 3.76% 6.59% 5 Actual Hypothetical 1993 1995 1997 1999 21 23 25 27 29 211 213 215 217 219 -year average 2-year average Notes: These hypothetical illustrations do not represent returns on any particular investments. Historical returns are based on the Dow Jones Wilshire 5 Index from 1973 through April 22, 25, and the MSCI US Broad Market Index thereafter. 4 In this example, the hypothetical 9.5% return consists of an approximation of the historical real return for U.S. stocks (7.%) and an expected average annual inflation rate of 2.5%. The expected inflation rate is based on 28 217 long-run Consumer Price Index estimates in the Survey of Professional Forecasters, published on February 12, 28. This survey, currently conducted by the Federal Reserve Bank of Philadelphia, is the oldest quarterly survey of macroeconomic forecasters. 6 > Vanguard Investment Counseling & Research

Figure 5. Bonds could win: Hypothetical scenarios using actual and assumed returns Average annual returns based on historical results through 27 and assumed yearly returns of 9.5% for stocks and 5.% for bonds thereafter Return 12% 8 4 6.31% 3.76% years ending 12/31/29 5.87% 5.64% 5.65% 15 years ending 12/31/214 Bonds 6.59% 2 years ending 12/31/219 Stocks 5.51%.16% Historical average: 1926 27 Notes: This hypothetical illustration does not represent returns on any particular investment. Historical stock returns are based on the Dow Jones Wilshire 5 Index from 1999 through April 22, 25, and the MSCI US Broad Market thereafter. Historical bond returns are based on the Lehman U.S. Aggregate Bond Index. The immediate implication for investors is that, over this particular time period, the average reward for enduring the risk of the stock market could be similar to or even below the reward offered by the bond market. In fact, if we add a hypothetical bond return to our assumptions, we can get a sense of how such a situation might look. In our fixed income scenario, we assume that bonds return 5.% per year, comprising a 2.5% real return (which approximates the bond market s long-run average) and a 2.5% expected inflation rate. 5 When we combine this with our stock-return assumption, as shown in Figure 5, it appears possible that in 29 the trailing ten-year average returns for stocks might be little more than half those of bonds. In fact, in this particular scenario, the twenty-year stock return would exceed that of bonds by only 9 basis points annually, well below the long-term historical equity risk premium. While some might find it surprising that stocks can post returns similar to or even lower than those of bonds over such long periods, this scenario would not be unique in the history of the U.S. markets. In fact, as Figure 6, on page 8, shows, it is not especially uncommon to see stocks only modestly outperform bonds and T-bills, or even underperform them, for extended periods. The last time that the ten-year average annual return for stocks lagged the returns for bonds and T-bills was at year-end 1982, following a very difficult decade that included three bear markets 6 and three recessions. 7 Figure 7, on page 8, shows the effect of that difficult period on ten-year equity returns, which averaged 7.63% at year-end 1982. Only two years later, the picture changed dramatically: The ten-year annualized return more than doubled to 16.36% after the returns from 1973 1974 dropped off the calculation. 5 From 1926 through 27, the annualized real return from the fixed income market was 2.39%, based on these benchmarks: the S&P High Grade Corporate Index from 1926 through 1968, the Citigroup High Grade Index from 1969 through 1972, the Lehman U.S. Long Credit AA or Better Index from 1973 through 1975, and the Lehman U.S. Aggregate Bond Index from 1976 through 27. 6 Although there s no agreed-upon definition of a bear market, one generally accepted measure is an index decline of 2% or more over at least a two-month period. By this measure, based on returns for the S&P 5 Index, there were three bear markets in the ten years through the end of 1982: January 1973 October 1974 ( 48.2%), September 1976 March 1978 ( 19.4%), and January 1981 August 1982 ( 25.8%). 7 According to the National Bureau of Economic Research: November 1973 March 1975, January 198 July 198, and July 1981 November 1982. Vanguard Investment Counseling & Research > 7

Figure 6. Rolling - and 2-year excess returns of stocks over bonds and T-bills: 1926 27 Based on average annual returns 2 Annualized excess return (%) 15 5 5 1935 1941 1947 1953 1959 1965 1971 1977 1983 1989 1995 21 27 -year excess return over T-bills 2-year excess return over T-bills -year excess return over bonds 2-year excess return over bonds Notes: Stocks are represented by the S&P 5 Index from 1926 through 197, the Dow Jones Wilshire 5 Index from 1971 through April 22, 25, and the MSCI US Broad Market Index thereafter. Bonds are represented by the S&P High Grade Corporate Index from 1926 throguh 1968, the Citigroup High Grade Index from 1969 through 1972, the Lehman U.S. Credit AA or Better Index from 1973 through 1975, and the Lehman U.S. Aggregate Bond Index thereafter. Treasury bills are represented by the Citigroup 3-Month Treasury Bill Index. Figure 7. Short-term results have always shaped trailing stock market performance 1-year returns and -year average annual returns for U.S. stock market: 196 27 5 4 16.36% Total returns (%) 3 2 2 3 4 7.63% 196 1963 1966 1969 1972 1975 1978 1981 1984 1987 199 1993 1996 1999 22 25 1-year return -year average annual return Note: Stocks are represented by the S&P 5 Index from 1926 through 197, the Dow Jones Wilshire 5 Index from 1971 through April 22, 25, and the MSCI US Broad Market Index thereafter. 8 > Vanguard Investment Counseling & Research

To the extent that investors use trailing returns to frame their expectations, our analysis of how a normal return environment could affect these returns is cause for concern. In 219, if investors see that volatile stocks have produced twenty-year average returns that are just 1 percentage point higher than those from lower-risk bonds, will they pull back from stocks, potentially diminishing their portfolio s expected performance? The challenge for advisors and other investment professionals will be to put the historical results in perspective, presenting them as one somewhat unlikely path that stocks can follow, not as a reasonable basis for expectations of future performance. Conclusion That the future is uncertain is well understood. However, the knowledge that the outcomes of many endeavors particularly those related to investing are far from assured should not impede decisionmaking. Rather, this realization should promote a thorough evaluation of the risks to be assumed during the pursuit of rewards. Such an evaluation helps to reduce uncertainty and is part of what distinguishes investing from speculation. The stock market returns of the last decade have been both interesting and troublesome. From the viewpoint of an investment professional, the 2 22 bear market and the subsequent rally have served as a valuable reminder of principles as timely as they are timeless: First, strategic asset allocation and diversification remain the most readily available and effective risk management tools at our disposal; and second, investors must make realtime decisions without the knowledge of what will transpire. Indeed, the events of the last decade have created a potential hurdle for investors to overcome as they build their portfolios for the future. Given the poor returns, should stocks be underweighted? Should other investment strategies that promise higher expected returns take their place? It s important to note that without the benefit of hindsight, market tops and bottoms are never clear. The challenge for investors is that during a bear market, does a rebound signal a bottom and the beginning of a new bull market? Or is the rebound a brief reversal caught in the midst of a long-term systematic decline? In retrospect, we know with certainty which rallies were new bull markets and which were not. For example, during the worst bear market in history, the Great Depression, stocks rallied over a span of four years, returning an average of 3% annually. And yet history shows that the markets remained in a bear market, not establishing a new bull market until five years later. Did the rebound from 23 through 27 characterize the start of a new bull market or just the most recent false positive? Distinguishing between a bear market rally and a new bull market is extremely difficult in real time, which is why we advise our readers not to make too much of a specific time series. As a result, for most investors, the answer is to take comfort in a sound investment strategy. The last ten years have merely provided the most recent demonstration that, while it is reasonable to expect riskier asset classes to outperform less-risky ones over the longer term, they do not always do so. And this fact is why it is reasonable to expect a risk premium when investing in stocks. The period has also served as a reminder that over ten, fifteen, twenty years, and longer, the returns of a volatile asset class such as stocks can and should be expected to deviate from the long-term mean. This is especially likely if the period includes returns from the tail of the long-term distribution. A ten-year equity return averaging 2% a year, or even less, is well within the historical parameters illustrated in Figure 1 just as likely, in fact, as the 18.11% average for the ten years through 1999. Both results deviate from the historical mean by about 8 basis points. Although some may proclaim the last ten years a lost decade for stocks, the lessons learned will more than pay for themselves if they lead to smarter investment decisions. Vanguard Investment Counseling & Research > 9

P.O. Box 26 Valley Forge, PA 19482-26 Connect with Vanguard > www.vanguard.com CFA is a trademark owned by CFA Institute. The funds or securities referred to herein are not sponsored, endorsed, or promoted by MSCI, and MSCI bears no liability with respect to any such funds or securities. For any such funds or securities, the prospectus or the Statement of Additional Information contains a more detailed description of the limited relationship MSCI has with The Vanguard Group and any related funds. Standard & Poor s, S&P, S&P 5, Standard & Poor s 5, and 5 are trademarks of The McGraw-Hill Companies, Inc., and have been licensed for use by The Vanguard Group, Inc. Vanguard mutual funds are not sponsored, endorsed, sold, or promoted by Standard & Poor s, and Standard & Poor s makes no representation regarding the advisability of investing in the funds. E-mail > research@vanguard.com Contributing authors John Ameriks, Ph.D./Principal Joseph H. Davis, Ph.D./Principal Francis M. Kinniry Jr., CFA/Principal Roger Aliaga-Díaz, Ph.D. Donald G. Bennyhoff, CFA Maria A. 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. Kimberly A. Stockton David J. Walker, CFA Yan Zilbering Investment products: Not FDIC-insured No bank guarantee May lose value 28 The Vanguard Group, Inc. All rights reserved. ICRLD 1228