This Accomplishment Based Renewal is warranted because of the impressive methodological and substantive contributions made under the previous NSF award (SBR 95-20966) to macroeconomics, finance, labor economics, and agricultural economics. This renewal focuses on the connection between financial economics and forecasting. The research deepens and refiner that connection from an econometric perspective, as it advances: (1) measurement (e.g., new methods and results for both parametric and nonparametric estimation of asset return volatility) (2) models (e.g., for the sources of long memory in asset market information arrival, with implications for asset return volatility dynamics) (3) tools (e.g., for multivariate density forecast evaluation, calibration, and structural change detection) (4) substantive knowledge (e.g., can crash probabilities be assessed from options and related derivatives markets and successfully modeled as functions of observable fundamentals, resulting in a workable "early warning system").
The research on crash probabilities. fundamentals and derivatives markets is especially significant. For obvious reasons, a growing number of very recent papers focus on early warning systems for currency crises. Most of that literature focuses either on leading indicator methods, which convey information about crash probabilities even if those probabilities are not explicitly modeled, or explicit probability models such as the logistic. This project takes a different, complementary, and powerful approach, which makes use of the rich market-based information contained in options prices to infer market-assessed crash probabilities.
