{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

HongStein2007 - Disagreement and the Stock Market

HongStein2007 - Disagreement and the Stock Market - joumai...

Info icon This preview shows pages 1–20. Sign up to view the full content.

View Full Document Right Arrow Icon
Image of page 1

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 2
Image of page 3

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 4
Image of page 5

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 6
Image of page 7

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 8
Image of page 9

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 10
Image of page 11

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 12
Image of page 13

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 14
Image of page 15

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 16
Image of page 17

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 18
Image of page 19

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 20
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: joumai of Economic Perspectives—Volume 21, Number 2—Spring 2007—Puges 109—428 Disagreement and the Stock Market Harrison Hong and jeremy C. Stein ver the last 20 years, the field of behavioral finance has grown from a startup operation into a mature enterprise, with well-developed bodies of both theory and empirical evidence. On the empirical side, the benchmark null hypothesis is that one should not be able to forecast a stock’s return with anything other than measm‘es of its riskiness, such as its beta. This hypothesis embodies the familiar idea that any other form of predictability would represent a profitable trading rule and hence a free lunch to investors. Yet in a striking rejection of this null, a large catalog of variables With no apparent connec- tion to risk has been shown to forecast stock returns, both in the time series and the cross-section. Many of these results have been replicated in a variety of samples and have stood up sufficiently well that they are generally considered to be established facts. One prominent set of patterns from the cross-section has to do with medium- term momentum and post-earnings drift in returns. These describe the tendency for stocks that have had unusually high past returns or good earnings news to continue to deliver relatively strong returns over the subsequent six to twelve months (and vice-versa for stocks with low past returns or bad earnings news). Early work in this area includes jegadeesh and Titman (1993) on momentum and Bernard and Thomas (1989, 1990) on pest—earnings drift. Another well-established pattern is longer-run fundamental reversion—the tendency for “glamour” stocks with high ratios of market value to earnings, cashflows, or book value to deliver weak returns over the subsequent several years (and vice-versa for "value" stocks with low ratios of market value to fundamentals}. Standard references for this I Humison Hang is Professor of Economics, Princeton University, Princeton, NewI jersey. jeremy C. Stein is Moise Y. Sufru Professor of Economics, Harvard University, and Research Associate, National Bureau of Economic Research, sack in Cambridge, Massachusetts. Their e-muii addresses Isire (khougflprinceionedu) and ([email protected]), respectiveiy. 110 Journal cf Economic Perspectives value—glamour phenomenon include Fama and French (1992) and Lakonishok, Shleifer, and Vishny (1994). On the theory side, research has proceeded along two distinct fronts. First, one needs to explain what prevents rational arbitrageurs from eliminating these and other predictable patterns in returns. Work in this "limits to arbitrage" vein has focused on the risks and market frictions that arbitrageurs face. These include simple transactional impediments, like short—selling constraints, as well as a variety of other complications. Potential arbitrageurs face the risk that when they bet against a given mispricing, this mispricing may subsequently worsen, with the ultimate correction coming only much later (DeLong, Shleifer, Summers, and Waldmann, 1990; Shleifer and Summers, 1990). This risk is exacerbated by the fact that many of the most sophisticated w0uld-be arbitrageurs are professional asset managers, who act as agents when they invest other people's money. A professional manager has to worry that poor short-run performance will lead to withdrawals from his fund, causing that asset manager to become liquidity—constrained and unable to hang on to even those positions that in the long run are likely to be winners (Shleifer and Vishny, 1997). Although research on limits to arbitrage is far from played out, it is fair to say that a broad consensus is emerging with respect to the key ideas and modeling ingredients. This should not be too surprising, given that the relevant tools all come from neoclassical microeconomics: arbitrageurs can be modeled as fully rational, With no need to appeal to any behavioral or psychological biases. For example, the work on the liquidity constraints associated with delegated arbitrage can be thought of as embedding familiar theories from corporate finance into an asset—pricing framework. Much less consensus has been achieved on the second front, which seeks to explain the specific nature of the patterns of predictability. Even taking as given that rational arbitrage cannot correct all instances of mispricing, what is it about the behavior of other, presumably less rational investors that makes stock prices appear to underreact to certain types of information in the short run, but overreact in the longer run? Here it is by definition harder to proceed with0ut entertaining devia- tions from the standard rational-agent paradigm, which necessarily opens up a can ofworms. Different authors have taken very different approaches, including representative-agent models with rational beliefs but unconventional prefer- ences, such as those associated with prospect theory (Barberis and Huang, 2001); representative—agent models With standard preferences but biased beliefs (Barberis, Shleifer, and Vishny, 1998; Daniel, Hirshleifer, and Subrahmanyam, 1998); and a variety of heterogeneous-agent models. Excellent surveys of recent work in behavioral asset pricing include Hirshleifer (2001) and Barberis and Thaler (2003). In this paper, we do not attempt to be either as balanced or as comprehensive as these authors. Rather, we adopt the role of advocates, and argue in favor of one particular class of heterogeneous-agent models, which we call “disagreement” models. This category is fairly broad, encom~ passing work that has focused on the following underlying mechanisms: i) gradual information flow; ii} limited attention; and iii) heterogeneous priors, that is, Harrison Hang andjsremy C. Stein 1 I l differences in the (Bayesian) prior beliefs that investors hold. While these three mechanisms each have their own distinct features, both theoretically and in terms of empirical content, we argue below that they share important common elements. In particular, this class of models is at its heart about the importance of differences in the beliefs of investors.l Disagreement models have a number of attractive features. In our view, the most compelling is that they allow us to speak directly to the joint behavior of stock prices and trading volume. Indeed, we find it hard to imagine a fully satisfying asset-pricing model-win either the rational or behavioral genres——that does not give a front-and—center role to volume. Trading volume is extremely large across virtu- ally all developed stock markets, and many of the most interesting patterns in prices and returns are tightly linked to movements in volume. For example, high-priced glamour stocks tend to have higher volume than low-priced value stocks, all else equal. Also, controlling for a stock’s ratio of price to fundamentals, its future returns tend to be lower when it has higher trading volume—in other words, trading volume appears to be an indicator of sentiment. In what follows, we argue that disagreement models offer a natural framework for understanding these and other related phenomena. The Importance of Trading Volume To help motivate our argument, we begin with some basic facts about the raw magnitude of trading volume. We then document the pervasive tendency for higher volume to accompany higher price levels, both in the time series and the cross section. In 2005, the dollar value of trading volume on the New York Stock Exchange (NYSE) was $14.1 trillion; on NASDAQ, it was $10.1 trillion; on the London Stock Exchange, it was $5.7 trillion; and on the Tokyo Exchange, $4.5 trillion. Worldwide, the roughly 50 members of the World Federation of Exchanges accounted for a total of $51.0 trillion of trading volume. In recent years, turnover on the NYSE has averaged about 100 percent (it was 102 percent in 2005), meaning that the entire market value of a typical firm changes hands about once a year.2 In conventional ratiOnal asset-pricing models with common priors—even those that allow for asymmetries in information across traders (Grossman and Stiglitz, 1980; Kyle, 1985)—the volume of trade is approximately pinned down by the unanticipated liquidity and portfolio rebalancing needs of investors. However, these motives would seem to be far too small to account for the tens of trillions of 1As long as arbitrage by fully rational agents is limited, we do not need to assume that these disagreement mechanisms apply to all investors. [t is Sufficient that they apply to a significant subset of the universe, which may include both some individual investors as well as some professional money managers. 2 Source: World Federation of Exchanges website, (http://www.world-exchanges.org), and NYSE Fact Book, available at (http://ww.nyse.com). I 12 journal ofEeanomic Perspectives dollars of trade observed in the real world. This dissonance has led even the most ardent defenders of the traditional pricing models to acknowledge that the bulk of volume must come from something else—for example, differences in prior beliefs that lead traders to disagree about the value of a stock even when they have access to the same information sets. Nevertheless, the implicit view taken by the traditionalists is that while such disagreement generates trading activity, these trades are idiosyncratic and there— fore cancel each other out, with no consequences for prices. This implies that stock prices can continue to be analyzed in the usual representative-agent efficient- markets setting, with trading volume being left as a separate and effectively uncon- nected area of inquiry. Reflecting this “decoupling” point of view, many of the important first-generation papers on disagreement in financial markets focus on generating predictions for trading volume, but do not attempt to speak directly to questions of pricing (Varian, 1989; Harris and Raviv, 1993; Kandel and Pearson, 1995). In contrast to this view, many early economists believed that elevated trading volume could play a central role in generating speculative bubbles. Kindleberger and Aliber (2005) point out that classical ideas about bubbles by Adam SmithJohn Stuart Mill, Knut Wicksell, and Irving Fischer were based on the concept of "overtrading," the process whereby euphoric investors purchase shares solely in anticipation of future capital gains. Perhaps the classical economists were influ- enced by episodes such as the Smith Sea Bubble of 1720. Carlos, Neal, and Wandschneider (2006) document dramatic increases in turnover in the shares of the Bank of England, the East India Company, and the Royal African Company during that bubble. Their most detailed data pertain to trading in Bank of England stock. In the three years prior to the bubble, there were approximately 2,000 transactions per year in the Bank’s stock, while in 1720 —the year of the South Sea Bubble—nearly 7,000 transactions occurred. In a similar vein, accounts of the stock-market boom of the late 19205 such as Galbraith (1979) often emphasize the heightened level of trading volume in 1928 and 1929 as an important element of market dynamics. Here's one way to get a sense for the intensity of trading activity during this period. Since 1900, NYSE share volume has set an all-time record on 74 days, according to our calculations using data available at (http: //www.nyse.com). Of these 74 record—breaking days, ten were in 1928 and three were in 1929. After 1929, a new record was not set until April 1, 1968, following Lyndon johnson’s announcement that he would not seek re-election. Ofek and Richardson's (2003) analysis of the recent Internet bubble touches on the same themes, but with the benefit of much more complete stock-level data on trading volume. Figure 1 reproduces one of their main findings. The figure plots the average monthly level of turnover and prices, for an index of Internet stocks, as well as for the remainder of the nonnInternet stocks, over the period january 1997 to December 2002. As can be seen, the remarkable run-up in Internet stock prices was accompanied by an extraordinary explosion in trading volume. Monthly turnover in Internet stocks exceeds 50 percent in twelve Out of the 24 Disogreemnt and the Stock Market 113 Figure 1 Prices and Turnover for Internet and Non-Internet Stocks, 1997—2002 120 1 + Average Turnover: Other - 1,200 + Average Turnover: Internet -- Return Index: Other 100 - — Return Index: Internet - 1,000 80 - — 800 — 600 .3 -— 400 Monthly turnover (%) D" o 20 - I 200 (001 = 9661390)XQPUIUJn39‘d 1391143191“ Imba 0 r I I r I r r r t I'1""_""_0 sssssseeeeeeess assessses eeeeeeeeeeeeeeeefiseeeeeee Sources: The underlying data is from the Center for Research in Security Prices (CRSP) database. We use the same sample of Internet stocks as Ofek and Richardson (2003). Their sample is obtained from lists of “pure” Internet companies published by Morgan Stanley, available on Eli Ofek‘s home page at (http://pages.stern.nyu.edu/~eofek/). Notes: For each month, we divide the set of all common Stocks listed on CRSP into "Internet" and "all other” portfolios, and calculate average monthly turnover and price indices for these two categories. Our turnover and price-level indices are equal—weighted, but the results are qualitatively similar using market-capitalization weights. months preceding the Internet index‘s price peak in February 2000, with monthly turnover reaching as high as 101 percent in December 1998—an annualized rate of turnover of over 1,200 percent. By contrast, monthly turnover for non~Internet stocks is generally in the 10-15 percent range over the same period, and only once does it creep slightly above 20 percent.3 The common thread across these episodes is that trading volume appears to act as an indicator of investor sentiment. In other words, when prices look to be high relative to fundamental values, volume is abnormally high as well. This basic relationship turns out to be quite general, arising not only in dramatic bubble-like situations, but also in broader cross-sectional and time-series samples. Figure 2 illustrates the cross—sectional phenomenon. We begin with the uni- verse of the 1,000 largest stocks in the Center for Research in Security Prices (CRSP) database for each quarter in the period 1986 to 2005. We then compute 3 A caveat is that measured turnover in NASDAQ stocks tends to be a bit higher than in NYSE/AMEX stocks because of the dealer nature of the NASDAQ market This difference could affect our compar— isons because Internet firms disproportionately trade on the NASDAQ. However, if we adjust the data to account for exchange-wide mean levels of turnover, the results are very similar to those in Figure l. I 14 fournoi of Economic Perspectives Figure 2 Turnover in Value and Glamour Stocks, 1986—2005 Adjusted monthly turnover (9%) 03? {~39 093% 03? of)? 999 $990943 99%?“ do? do? do?) of? 0’63 «Sax 9% ‘94! 6&3? s s s s s s s s s s “r s s s s s s“ s“ «r s Source: The underlying data is from the Center for Research in Security Prices (CRSP) database. Notes: At any point in time‘ glamour stocks are the top 200 stocks by market—to—book ratio Out of the 1,000 largest stocks in the CRSP database. Value stocks are the bottom 200 stocks by market-to-book ratio out ofthis same universe. For each stock in each month, adjusted turnover is that stock’s turnover minus the average turnover of all stocks listed on the same exchange (either NYSE/AMEX or NASDAQ) for that month. This adjustment is done to eliminate differences in reported turnover that are due to the dealer nature of the NASDAQ market. For each month, we then plot average adjusted turnover for the set of glamour and value stocks. exchange-adjusted monthly turnover for each stock, defined as that stock’s turn— over minus the average turnover of all stocks listed on the same exchange (either NYSE/AMEX or NASDAQ]. Finally, we calculate average adjusted monthly turn— over for each of two portfolios: i) a portfolio of high-priced glamour stocks—those ranked in the upper quintile of the universe by market-to-book ratio; and ii) a portfolio oflow-priced value stocks—those in the lower quintile of the universe by market-to-book. As Figure 2 shows, glamour stocks tend to have significantly higher turnover than value stocks (see also Piqueira, 2006). This differential is largest during the run—up of the Internet bubble, but it is apparent in virtually every month in the sample. In particular, glamour stocks continue to display more turnover even in 2000—2001, when the Internet bubble is collapsing and the returns to glamour stocks fall far below those of value stocks. So the phenomenon captured in Figure 2 reflects more thanjust the so-called “disposition effect“ (Odean, 1998), whereby investors tend to be reluctant to sell stocks that have recently declined in value. Figure 3 considers the relationship between trading volume and prices in the time-series dimension. Here we go back to 1901, and plot for each year the annual Harrison Hang andferemy C. Stein I15 Figure 3 Market-Wide Stock Returns and Changes in Turnover, 1901-2005 140% - _ w--' S&P real return 120% - o + % Change in NYSE turnover 100% - 80% - 60% - 40% - 20% - 0%“ -20% - _w% _ WITIWITIWIWIHIIWIWI *0 Q ‘3') (3 53 q? *9 Q ‘0 (3 1’3 Q (’3 Q ‘0 Q (‘3 Q ‘0 Q *9 Ogle? \ N ‘L ‘L ‘b by by *9 *9 h {o '\ ’\ ‘b ‘b Sources; Data on the real (inflation-adj usted) level of the S&P index are from Robert Shiller‘s web-page, (http://www.econ.yale.edu/"-shi1ler). Data on NYSE turnover are from the NYSE Fact Book, available at (http://www.nyse.com). Notes: The figure plots year—to—year percentage changes in the 88:? index and year—to-year percentage changes in turnover. We omit the years 1914 and 1915 from the plot since the NYSE was closed for much of the latter half of 1914 due to the outbreak of World War 1. The correlation between the two series shown in the plot is 0.49. real return on the Standard and Poor‘s (38:?) 500, along with the percentage change in turnover on the NYSE from the preceding year. Thus we are now looking at how changes in price levels covary with changes in the level of trading volume; one advantage of this differencing approach is that it removes low-frequency time trends from the turnover data. The resulting relationship is visually striking, and is also confirmed statistically: the correlation between the two series is a highly significant 0.49. The connections between prices and trading volume documented in this section should serve to whet the reader's appetite for asset-pricing theories in which volume plays a central role. However, we have not yet addressed two key questions: First, what are the underlying mechanisms, either at the level of market structure or individual cognition, that give rise to disagreement among traders and hence to trading volume? Second, how do these mechanisms simultaneously generate mis— pricings of one sort or another? Or said differently, why do they not simply lead to trades that cancel each other out in terms of price effects, as implicitly assumed by the traditional model? The next two sections seek to answer these questions. Before preceeding, however, We should clarify the distinctions between our approach and the well—known model of DeLong, Shleifer, Summers, and Wald— I I 6 journal ofEconomic Perspectives mann (1990). In that earlier model, a group of smart-money arbitrageurs interacts with a group of noise traders who are subject to exogenous sentiment shocks. This model also generates trading volume, to the extent that the valuations of the noise traders shift relative to those of the arbitrageurs—that is, to the extent that equilibrium prices move closer to or further away from fundamental values. At the same time, the amount of volume generated by the earlier model is not always large. This result is most clearly seen by considering a scenario in which th...
View Full Document

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern