Can stock returns predict output? An application of threshold panel data model

• Main Authors: Yun Chiao, 喬勻
• Other Authors: Chien-Fu Chen
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/255t3x

碩士 === 國立東華大學 === 國際經濟研究所 === 96 === This paper employs generalized method of moments (GMM) to estimate the threshold panel data model and investigate whether stock returns can predict output growth from a sample of 37 countries (including 19 developed countries and 18 developing countries). We analyze the prediction ability of stock returns on output growth by taking different economic regimes (expansion and recession periods) into consideration. Furthermore, we also use the rolling approach to examine whether the relationship between stock returns and output growth change through time. Finally, we compare the empirical evidences from developed countries with those from developing countries. Our empirical results are as follows: First, whether in the developed or the developing countries, the stock returns cannot predict output in expansion period, but they can predict output significantly in recession period. Second, the stock returns of the developed countries can better predict output than those of the developing countries. Finally, the relationship between stock returns and output growth has recently become weak when the asymmetry of business cycle is included in regression, which means that the stock returns cannot predict output growth as well as before.