Venue: Room 418, School of Economics, Yuquan Campus
Speaker: Zhipeng Liao
Zhipeng Liao received his Ph.D. from Yale and is an Assistant Professor in Economics at UCLA. His research develops methods to use data to select among different economic models, and to make inferences from nonstationary time series data, and to make robust inferences from nonparametric models. His work has been published in the Econometric Theory, the Journal of Econometrics, the Quantitative Economics and the Review of Economic Studies.
Abstract:
This paper proposes a test for the conditional superior predictive ability (CSPA) of a family of forecast methods with respect to a benchmark. The test is functional in nature: Under the null hypothesis, the benchmark's conditional expected loss is no more than those of the competitors, uniformly across all conditioning states, and, under the alternative, there exists some competing method that outperforms the benchmark in certain states. By inverting the CSPA tests for a set of benchmarks, we obtain confidence sets for the uniformly most superior model. The econometric inference pertains to testing a system of conditional moment inequalities for dependent data, and we justify its asymptotic validity using uniform nonparametric inference methods based on time-series strong approximation. The usefulness of the proposed method is demonstrated in empirical applications on volatility and inflation forecasting.