密度预测评估的数据驱动平滑检验张玉鹏.pdfVIP

Density Forecast Evaluation Using Data-Driven Smooth Test∗ Sung Yong Park† Yupeng Zhang‡ November 26, 2009 Abstract In this paper, we develop a data-driven smooth test for out-of-sample density forecast evaluation and provide simulation evidence on its finite sample performance. Relative to Hong, Li Zhao(2007) and Corradi and Swanson (2006, section 3.3),we derive our tests in an environment where the finite sample properties of the estimators on which the forecasts may depend are preserved asymptotically, our tests can be computed easily and accommodate general estimation procedures. Similar to Hong and Li(2005)’s paper, our paper is a joint test for uniformity and serial independence. Keywords: density forecast evaluation; Neyman’s smooth test; data driven procedure; probability integral transform. JEL classification: C12; C15; C52; C53. ∗We are grateful to However, we retain the responsibility for any remaining errors. †The Wang Yanan Institute for Studies in Economics, Xiamen University, Xiamen, Fujian 361005, China. Tel.: +86-592-2181675; fax: +86-592-2187708. E-mail: sungpark@ ‡The Wang Yanan Institute for Studies in Economics, Xiamen University, Xiamen, Fujian 361005, China. E-mail: yupengzhang514@ 1 1 Introduction Comparing with the point forecast and interval forecast, density forecast provides a com- plete description of the uncertainty associated with the forecast of a variable. It is, therefore, not surprising that predictive density evaluation has received considerable attention in eco- nomics and finance over the last few years. For instance, there has been increasing interest in evaluating forecasting mo