Behavioral finance.

The NBER's Group on Behavioral Finance held its spring meeting in Cambridge on April 10. Project directors Robert J. Shiller, NBER and Yale University, and Richard H. Thaler, NBER and University of Chicago, organized this program:

Nicholas Barberis, Harvard University, Ming Huang, Stanford University, and Tano Santos, University of Chicago, "Prospect Theory and Asset Prices"

Discussant: Sendhil Mullainathan, NBER and MIT 

Robert J. Shiller, "Measuring Bubble Expectations and Investor Confidence" (NBER Working Paper No. 7008)

Discussant: Werner DeBondt, University of Wisconsin

Harrison Hong, Stanford University, and Jeremy c. Stein, NBER and MIT, "Differences of Opinion, Rational Arbitrage, and Market Crashes"

Discussant: Olivier J. Blanchard, NBER and MIT

William N. Goetzmann, NBER and Yale University, and Massimo Massa, INSEAD, "Index Funds and Stock Market Growth" (NBER Working Paper No. 7033)

Discussant: Andrew Metrick, NBER and Harvard University

Jeff Wurgler and Ekaterina Zhuravskaya, Harvard University, "Does Arbitrage Flatten Demand Curves for Stocks?"

Discussant: Randall Morck, University of Alberta

Allen Poteshman, University of Chicago, "Does Investor Misreaction to New Information Increase in the Quantity of Previous Similar Information? Evidence from the Options Market"

Discussant: Ming Huang, Stanford University

Barberis, Huang, and Santos propose a new framework for pricing assets, derived in part from the traditional consumption-based approach but also incorporating two longstanding ideas in psychology: Kahneman and Tversky's (1979) prospect theory and the evidence of Thaler and Johnson (1990) and others on the influence of prior outcomes on risky choice. Consistent with prospect theory, investors in the authors' model derive utility not only from consumption levels but also from changes in the value of their asset holdings. The investors are much more sensitive to reductions than to increases in wealth. Moreover, the investors' utility from gains and losses in wealth depends on prior investment outcomes: prior gains cushion subsequent losses, while prior losses intensify the pain of subsequent shortfalls. Studying asset prices in the presence of a representative agent with preferences of this type, the authors find that their model can explain the high mean, volatility, and predictability of stock returns. The agent's risk-aversion changes over time as a function of investment performance: this generates time-varying risk premiums, which in turn make prices much more volatile than underlying dividends. In combination with the agent's loss-aversion, the high volatility of returns generates large equity premiums.

Shiller presents evidence of changes in attitude among investors in the U.S. stock market, exploring two basic attitudes: bubble expectations and investor confidence. He produces five different time-series which indicate when investors expect a speculative bubble - an unstable situation in which an increase is anticipated only in the short run - and four different time-series which indicate when investors expect a negative speculative bubble. He also produces four different time series that indicate investor confidence. The time-series variation for these indicators is significant, and cross-correlations are generally positive. Finally, Shiller examines the behavior of the indicators and indexes through time and compares these indexes with other economic variables. One notable finding is the degree of high-frequency fluctuation, semester to semester, in the indexes.

Hong and Stein develop a theory of stock market crashes based on differences of opinion among investors. Because of constraints on short sales, bearish investors do not initially participate in the market, and their information is not revealed in prices. However, if other previously bullish investors have a change of heart and bail out of the market, then the originally more-bearish group may become the marginal "support buyers." In that way, more will be learned about their signals. Accumulated hidden information thus tends to come out during market declines. This helps to explain large movements in prices unaccompanied by significant news about fundamentals; negative skewness in the distribution of market returns; and increased correlation among stocks in a falling market.

Goetzmann and Massa analyze the relationship between index funds and asset prices. They find a strong positive correlation between daily inflows into index funds and stock market returns. There is a strong negative correlation between fund outflows and stock market returns, except for outflows from funds with a very high initial investment requirement. These effects may be interpreted in two ways: either investor supply and demand affects market prices, or investors condition their supply and demand on intraday market fluctuations. The authors conclude that the market reacts to daily demand. However, only negative reactions appear to be attributable to past returns. Using the average market-timing newsletter recommendation over the period, the authors find that investors appear to react to "expert" advice about the market. Bullish newsletter sentiment is associated with greater inflows, although outflows are not explained well by newsletter advice.

Because individual stocks do not have perfect substitutes, would-be arbitrageurs who aim to exploit relative mispricings face "arbitrage risk": the zero-net-investment portfolio that holds $1 long in the mispriced stock and $1 short in its closest substitutes is not riskless. In a simple model of the aggregate demand curve of a stock, this risk deters risk-averse arbitrageurs from flattening the curve at the efficient price. Consistent with this model, Wurgler and Zhuravskaya find that stocks that do not have close substitutes experience differentially higher price jumps upon inclusion into the S&P 500 Index. The results suggest that corrective price pressure is weakest, and other pricing anomalies are likely to be severest, among stocks without close substitutes.

A substantial body of stock price evidence has been (controversially) interpreted as supporting the claim that investors tend to underreact to single pieces or short strings of similar information and to overreact to long strings of similar information. Poteshman tests for this phenomenon in the S&P 500 options market under the assumption that investors subscribe to a general stochastic variance option pricing model. His principal finding is that, under the assumed model, investor misreaction to a current change in instantaneous variance is increasing (along a scale that ascends from underreaction to overreaction) in the quantity of previous similar changes in instantaneous variance. The associated one-day deviation in at-the-money S&P 500 Index option prices is on the order of 3 percent.