巴菲特指標於亞洲股市之實證研究

Abstract

本文以巴菲特指標（股票總市值對國內生產毛額比率）搭配常態分配單尾信賴區間與坎泰利不等式兩種門檻值建構股市修正預測模型，實證分析台灣、日本、韓國、中國、印度、馬來西亞、新加坡、泰國、及澳洲等九個亞洲國家股市，研究期間自1985至2020年不等，探討巴菲特指標在亞洲國家股市是否具有修正預測的能力。實證結果發現，採用常態分配單尾信賴區間門檻值的修正預測模型對於台灣、韓國、新加坡等亞洲四小龍國家預測結果最好，並具統計顯著性，而採用坎泰利不等式門檻值的修正預測模型雖產生的修正訊號次數較少，但在多數亞洲國家達100%的預測準確率，對於股市修正也有預測能力。整體而言，巴菲特指標於亞洲國家股市的應用上，可作為捕捉股市修正的參考依據。

This paper uses the Buffett Indicator (the ratio of total stock market value to gross domestic product) to construct correction forecasting models with two thresholds: (i) one-tail confidence interval based on a normal distribution; and (ii) Cantelli`s inequality. We empirically analyze nine stock markets in Asian countries, including Taiwan, Japan, South Korea, China, India, Malaysia, Singapore, Thailand, and Australia, and the time period ranges from 1985 to 2020. This paper aims to observe whether the Buffett Indicator has the ability to predict equity market corrections in Asian stock markets.\nThe empirical results show that the correction forecasting model using the threshold of one-tail confidence interval based on a normal distribution can successfully predict corrections in the stock markets of Taiwan, South Korea, and Singapore, which reaching a high degree of statistical significance. On the other hand, although the correction forecasting model using the threshold of Cantelli’s inequality generates fewer correction signals, it still has high forecasting accuracy and predictive ability to accurately forecast corrections. Overall, the Buffett Indicator is useful in Asian stock markets to capture equity market corrections.