by James Picerno
Where do you go if you want to rub elbows with the movers and shakers of modern finance? One place is the annual Western Finance Association (WFA) meeting. This year, it was held at the Lowe∆s Coronado Bay Resort in San Diego. A great spot! But with more than 100 papers on finance and economics on tap, finding time to enjoy the locale was a tall order.
An equally tall order was deciding which papers to focus on. So many papers, so little space. But after reflecting on what I heardýmuch of it destined to appear in leading academic and professional journalsýone theme resonated. Global asset allocation strategies look convincing on paper. But what about the real world?
Two papers in particular shed light on that subject. The first examines historical equity premiums in stock markets around the world, utilizing performance data going back to 1921. A key finding: Historical equity premiums (excess returns in stocks over so-called riskless securities, such as Treasury bonds) look far less attractive on a global basis compared to the more encouraging numbers derived by examining a U.S.-only sample.
The second paper should persuade international investors committed to a buy-and-hold strategy to at least take a second look at momentum investing active management. Based on 15 years of data from overseas stock markets, the report found that trading in foreign markets every six months on a top-down basis can add value. Overall, the paper suggests that gains from momentum investing are both significant and recurring.
Perhaps the most pressing challenge in asset allocation is determining the relevance, if any, of historical performance when projecting future returns. The stakes are high. The case for buying stocks at the expense of bonds is easily justified, of course, if investors expect a continuation of the 7.3% equity premium (U.S. large-cap stock returns minus long-term T-bond returns) posted during 1926 to 1996.
According to data from Chicago-based Ibbotson Associates, of the 52 holding periods of 20-year duration during the aforementioned time period, there∆s just one instance (1929 to 1948) when large company stock returns lagged those of long-term government bonds. Though bonds did edge out stocks a bit more often when the data was sliced into 10-year holding periods, stocks nonetheless remained the superior performer by far on both an absolute and a frequency basis over that span.
What does the data tell us? Stocks always bounce back, at least eventually. Yes, there will be losing periods, perhaps years at a stretch, but patience and a generous allocation to stocks pays off over time.
This conclusion jibes ever so nicely with modern portfolio theory, which advises that higher return comes via higher riskýunsystematic risk, to be precise. After plugging historical performance numbers of stocks and bonds into the quantitative framework of the Capital Asset Pricing Model (CAPM), the resulting asset allocation is likely to be one of heavy equity weighting and low bond exposure for a long-term investment horizon.
There∆s at least one conundrum, however. If stocks appear safer than bonds over timeýbased on the fact that equities outperform bonds over the long haulýwhy do investors require a risk premium for investing in stocks? Shouldn∆t it be the other way around? Puzzling!
We∆re not the only ones who find it puzzling, either. Rajnish Mehra and Edward Prescott are puzzled, as well, or so they wrote more than a decade ago. In their paper, "The Equity Premium: A Puzzle" (The Journal of Monetary Economics, 1985), the pair noted the historical premium generated by stocks over bonds looked unjustified based on reasonable assumptions of investors∆ risk aversion.
Solutions to the puzzle have since been offered, though no single explanation has become widely accepted. Some say flaws in CAPM are to blame. Others argue the puzzle is the logical consequence of a market dominated by investors fixated on short-term performance. Still others say what seems a large equity premium is justified by low-probability events with potentially large impacts, such as a stock market crash.
Two professors with a slightly different take on this subject presented their thoughts at the recent WFA gathering. In an as-yet unpublished paper ("A Century of Global Stock Markets"), Will Goetzmann of the Yale School of Management and Philippe Jorion of the University of California at Irvine, sound a cautionary note. Goetzmann and Jorion stress that what investors expect in terms of a risk premium relies heavily on data from the U.S. equity market. That∆s a risky assumption to make, they add. Simply stated, the exceptionally positive history of U.S. capital markets isn∆t a random sample and, therefore, may cast an undeserved bullish aura on global equity markets.
"We find striking evidence in support of the survival explanation for the [U.S.] equity risk premium," Goetzmann and Jorion write. Reflecting on their study of foreign markets in the 20th century, they report the "United States had by far the highest uninterrupted real rate of appreciation, at about 5% annually. For most other countries, the median real appreciation rate was around 1.5%. This strongly suggests that estimates of equity premia obtained solely from the U.S. market are either biased upward by survivorship or reflect fundamentally different investor expectations about risk across markets."
Overall, the pair advises that the higher real rate of return in U.S. stocks relative to the other markets studied is "the exception rather than the rule."
During his presentation to a standing-room-only crowd at the WFA meeting, Goetzmann recalled the many challenges associated with reassembling long-term performance data for 39 markets based on monthly data going back, in some instances, to 1921. The findings cited in "A Century of Global Stock Markets" were "painstakingly" collected from an array of international sources, many with erratic reporting histories, Goetzmann said.
The biggest challenge was reconstructing performance histories from markets that closed, if not collapsed. The authors wrote that overall, the process is akin to "financial archaeology." It was well worth the effort, they state, because their findings provide "the first comprehensive long-run estimates of return on equity capital across a broad range of markets."
The paper expands on an earlier study co-written by Goetzmann, Stephen Brown of New York University, and Stephen Ross of the Yale School of Management. That paper, "Survival," was published in the July 1995 issue of The Journal of Finance. In it, the three professors suggest the true equity premium may be closer to zero when correcting for survival bias. A market∆s survival, as in the case of the United States, they explain, "will induce a substantial spurious equity premium."
What∆s beyond doubt is that the equity premium for stocks on an international basis is far below the comparable U.S. level, based on data published in "A Century of Global Stock Markets." The annualized, compound real (inflation-adjusted) return for the United States is 7.37% between January 1921 and the end of 1996, the paper reports. That∆s far above the negative 0.28% mean, annualized return for the comparable period for all 39 countries analyzed (including the U.S.), and the 1.51% for the 14 countries with continuous histories from the 1920s onward (see Table 1).
Goetzmann and Jorion warn in "Century" that "if we fail to account for the ślosers∆ as well as the świnners∆ in global equity markets, we are providing a biased view of history which ignores important information about actual investment risk."
If so, then there∆s a lot of bias going on in the real world, at least from Goetzmann∆s point of view. During a recent phone interview for this article, he stressed that he knows of no asset manager who takes into account the survival bias in U.S. data when making asset allocation decisions.
Do winners repeat? Professor K. Geert Rouwenhorst of the Yale School of Management says he found they do. In the paper he presented at the WFA meeting, Rouwenhorst shed new light on short-term momentum effects in international stock markets. In the short termýroughly six monthsý"winners repeat," he said.
His paper∆s findings suggest trading during six-month periods can generate superior returns relative to a buy-and-hold strategy. That∆s a "surprising result," Rouwenhorst told us during a recent telephone interview. He noted that "academics have been skeptical of technical analysis for a long time." Reassessing the underpinnings of that doubt may be warranted, he suggested, explaining that the examination of the momentum phenomena he studied represents the "simplest form of technical analysis."
In his paper, "International Momentum Strategies," Rouwenhorst wrote that international equity markets "exhibit medium-term return continuation. Between 1980 and 1995, an internationally diversified portfolio of past medium-term winners outperformed a portfolio of medium-term losers after correcting for risk by more than 1% per month."
As it turns out, Narasimhan Jegadeesh and Sheridan Titman reported a similar result in their 1993 paper in The Journal of Finance ("Returns To Buying Winners and Selling Losers: Implications for Stock Market Efficiency"). Rouwenhorst∆s paper, which will be published in an upcoming issue of The Journal of Finance, builds on the Jegadeesh and Titman paper by searching for comparable results in foreign markets.
In that quest, Rouwenhorst studied the monthly equity performance of 2,190 firms from 12 European markets during 1980 to 1995. The samples represented roughly 60% to 90% of each nation∆s market capitalization; returns were converted to D-marks. The basic methodology underlying the Rouwenhorst study: At the end of each month, all stocks were ranked in ascending order based on previous six-month performance. Equities with the lowest decile performance were put in the "loser" portfolio; those in the top decile were assigned to the "winner" portfolio. The 10 portfolios were equally weighted initially and held for six months. The average monthly returns over the 15-year period (1980 to 1995) can be found in Table 2.
This table shows the top-decile group outperformed the bottom decile by an average of 116 basis points per month. In addition, "this momentum in returns ... is present in all 12 markets in the sample," and "it holds across size deciles," although performance is stronger in smaller companies, Rouwenhorst wrote. "The outperformance lasts for about one year, and cannot be attributed to conventional measures of risk."
Rouwenhorst believes the momentum behavior he observed "points either toward a more serious misspecification of commonly used asset pricing models or a general tendency of markets to underreact to information."
Asked if his findings are in conflict with the efficient market theory (EMT), Rouwenhorst said "it is and it isn∆t." Elaborating, he said that adherents to EMT may explain away the excess returns generated by momentum investing as yet another risk factor. So, while a momentum-investment strategy may generate higher returns than does a buy-and-hold approach, those excess returns may be the appropriate reward for assuming extra risk, which, in this case, is engaging in momentum investing.
In short, there∆s no free lunch, or so EMT predicts. Taking that line of thinking a step further raises questions as to whether the conventional thinking about risk should incorporate the effects of momentum in asset pricing models, Rouwenhorst added.
Rouwenhorst believes that his paper∆s findings give aid and comfort to the market-inefficiency argument. This camp believes that investors underreact to information in the short and medium term. So, if new information about a stock becomes available, the stock movesýbut not by the amount warranted by the news. "If you believe the market underreacts to information, then you would think the findings in my paper are evidence for market inefficiency," Rouwenhorst told us.