Journal of Financial Econometrics - current issue

The JFEC Invited Lecture at the 2008 SoFiE Conference

Garcia, R., Ghysels, E., Renault, E. — 2009-06-24

Inference on Risk-Neutral Measures for Incomplete Markets

Kaido, H., White, H. — 2009-06-24

This paper proposes an econometric framework to estimate market risk prices associated with risk-neutral measures Q under incomplete markets. We show that, under incomplete markets, the market price of risk is not point-identified but is instead identified as a bounded subset of an affine subspace. On the other hand, a structural assumption fully identifies diffusion coefficients for the data-generating probability measure P. We apply Kaido and White's (2008, Discussion Paper, University of California, San Diego) two-stage extension of Chernozhukov, Hong, and Tamer's (2007, Econometrica, 75(5), 1243–1284) partial identification framework to construct a set estimator and confidence regions for the identified set of market risk prices and to test hypotheses. We apply our results to study international risk sharing and risk premiums for market cap range indexes.

A Dynamic Asset Pricing Model with Time-Varying Factor and Idiosyncratic Risk

Glabadanidis, P. — 2009-06-24

This paper uses a multivariate GARCH model to account for time variation in factor loadings and idiosyncratic risk in improving the performance of the CAPM and the three-factor Fama–French model. I show how to incorporate time variation in betas and the second moments of the residuals in a very general way. Both the static and conditional CAPM substantially outperform the three-factor model in pricing industry portfolios. Using a dynamic CAPM model results in a 30% reduction in the average absolute pricing error of size/book-to-market portfolios. Ad hoc analysis shows that the market beta of a value-minus-growth portfolio decreases whenever the default premium increases as well as during economic recessions.

Measuring Event Risk

Nyberg, P., Wilhelmsson, A. — 2009-06-24

This paper decomposes the popular risk measure Value-at-Risk (VaR) into one jump- and one continuous component. The continuous component corresponds to general market risk and the jump component is proportional to the event risk as defined in the Basel II accord. We find that event risk, which is currently not incorporated into most banks' VaR models, comprises a substantial part of total VaR. It constitutes 30% of the risk for a portfolio of small cap stocks but less than 1% for a portfolio of large cap stocks. The national supervising agency in each membership country is advised by the Basel rules to add an additional capital charge to a bank whose models do not capture event risk. The large variation in event risk, also found across 10 individual stocks, suggests that an approach that varies the capital surcharge, based on the type of asset, should be used by the supervisors.

Using High-Frequency Transaction Data to Estimate the Probability of Informed Trading

Tay, A., Ting, C., Tse, Y. K., Warachka, M. — 2009-06-24

This paper applies the asymmetric autoregressive conditional duration (AACD) model of Bauwens and Giot (2003) to estimate the probability of informed trading (PIN) using irregularly spaced transaction data. We model trade direction (buy versus sell orders) and the duration between trades jointly. Unlike the Easley, Hvidkjaer, and O'Hara (2002) approach, which uses the aggregate numbers of daily buy and sell orders to estimate PIN, our methodology allows for interactions between consecutive buy-sell orders and accounts for the duration between trades and the volume of trade. We extend the Easley–Hvidkjaer–O'Hara framework by allowing the probabilities of good news and bad news to vary each day. Our PIN estimates can be computed daily as well as over intraday intervals.

A New Look at the Forward Premium Puzzle

Gospodinov, N. — 2009-06-24

This paper analyzes the sampling properties of the widely documented large negative slope estimates in regressions of future exchange returns on current forward premium. We argue that the abnormal behavior of the slope estimators in these regressions arises from the simultaneous presence of high persistence, low signal-to-noise ratio, strong endogeneity, and an omitted variable problem. The paper develops the limiting theory for the slope parameter estimators in the levels and differenced forward premium regressions under some assumptions that match the empirical properties of the data. The asymptotic results derived in the paper help to reconcile the findings from the levels and difference specifications and provide important insights about the time-series properties of the implied risk premium.