Asset pricing.

The NBER's Program on Asset Pricing, directed by John Y. Campbell, held its spring meeting in Cambridge on May 13. Dimitrios Vayanos and Jiang Wang, NBER and MIT, organized the program and chose the following for discussion:

Torben G. Andersen, Northwestern University, Tim Bollerslev, NBER and Duke University, Francis X. Diebold, NBER and University of Pennsylvania, and Paul Labys, University of Pennsylvania, "The Distribution of Exchange Rate Volatility" (NBER Working Paper No. 6961)

Discussant: Ludger Hentschell, University of Rochester 

Wayne E. Ferson, NBER and University of Washington, and Andrea Heuson and Tie Su, University of Miami, "How Much Do Expected Stock Returns Move over Time? Answers from the Options Market"

Discussant: John C. Heaton, NBER and Northwestern University

Martin Lettau and Sydney Ludvigson, Federal Reserve Bank of New York, "Consumption, Aggregate Wealth, and Expected Stock Returns"

Discussant: Kenneth R. French, NBER and MIT

Peter Bossaerts, California Institute of Technology, "Learning-Induced Securities Price Volatility"

Discussant: Lars P. Hansen, NBER and University of Chicago

Michael W. Brandt, University of Pennsylvania, and Pedro Santa-Clara, University of California, Los Angeles, "Simulated Maximum Likelihood Estimation of Multivariate Diffusions with an Application to Interest Rates and Exchange Rates with Stochastic Volatility"

Discussant: Matthew P. Richardon, NBER and New York University 

Lubos Pastor, University of Chicago, and Robert F. Stambaugh, NBER and University of Pennsylvania, "Comparing Asset Pricing: An Investment Perspective"

Discussant: Ravi Jagannathan, NBER and Northwestern University

Using high-frequency data on Deutschemark and yen returns against the dollar, Andersen, Bollerslev, Diebold, and Labys estimate daily exchange rate volatility and correlation for an entire decade. The authors characterize their joint distribution both unconditionally and conditionally. They find high contemporaneous correlation both across volatilities and between correlation and volatilities. There is also pronounced and highly persistent variation in both volatilities and correlation over time.

Ferson, Heuson, and Su estimate the predictability of stock returns implied by option prices, the "option-implied R-squares." They use monthly data from 1975 to 1997 for 15 large common stocks and from 1986 to 1997 for the Standard & Poors 500 Index. The authors estimate that the implied volatility horizon is one month for one-month options on individual stocks. The option-implied R-squares for individual stocks average 20 percent, which is higher than the regression R-squares on a standard set of marketwide lagged instruments. The implied volatility horizon for the S&P 500 Index is shorter than one month, and the option-implied R-squares is indistinguishable from zero.

Lettau and Ludvigson study the role of detrended wealth in predicting stock returns. They define a transitory movement in wealth as one that produces a deviation from its shared trend with consumption and labor income. Using quarterly U.S. stock market data, they find that these trend deviations in wealth are strong predictors of both real stock returns and excess returns over a Treasury bill rate. They also find that this variable is a better forecaster of future returns at short and intermediate horizons than is the dividend yield, the earnings yield, the dividend payout ratio, and several other popular forecasting variables.

Bossaerts provides a mathematical foundation for the empirical observation that volatility accompanies large movements in security prices: significant downward price corrections almost never seem to occur in an orderly way (they are called "panics") and, likewise, substantial price run-ups are equally erratic (they are called "bubbles"). What Bossaerts refers to as "market beliefs" actually lead to lower bounds on the volatility of securities prices as a function of the mean return. He documents the presence of a clear, U-shaped pattern in the relationship between return and volatility for the daily returns of the Dow Jones and S&P 500 indexes since 1883. His explicit calculation of the theoretical bounds suggests that the volatility marginally may have been too low for moderate price increases.

Brandt and Santa-Clara develop a new econometric method for estimating the parameters of a diffusion from discretely sampled data. Compared with existing estimators of continuous time models, their method is particularly effective for multivariate diffusions, is very transparent and adaptive, and inherits all the desirable asymptotic properties of the unattainable maximum likelihood estimator. Using this method, the authors estimate a new continuous time model of the joint dynamics of interest rates in two countries and of the exchange rate between the two countries' currencies. Their empirical results offer several new insights into tile dynamics of exchange rates.

Pastor and Stambaugh investigate the portfolio choices of investors who optimize mean variance by using sample evidence to update prior beliefs centered on either risk-based or characteristic-based pricing models. With dogmatic beliefs in such models and an unconstrained ratio of position size to capital, optimal portfolios can differ across models to economically significant degrees. The differences are reduced substantially by modest uncertainty about the models' pricing abilities: when the ratio of position size to capital is subject to realistic constraints, the differences in portfolios across models become even less important or, in some cases, nonexistent.