Testing Business Cycle Asymmetries Based on Autoregressions With a Markov-Switching Intercept
研究了马尔可夫转换截距自回归过程中的两种不对称性（深度和陡度），推导了偏度公式，指出先前检验可能导致错误结论，并发现美国失业率具有陡度特征。
Abstract Implications of two concepts of asymmetry—deepness and steepness—are investigated for autoregressive processes with a Markov-switching intercept. The formulas for the skewness of these processes and the skewness of the first differences of these processes are derived. The parameter restrictions leading to nondeepness and nonsteepness are presented for the special case of a first-order autoregression and two states. It is shown that these restrictions imply that previous tests for asymmetries in autoregressive processes with a Markov-switching intercept can lead to wrong conclusions. In an empirical application of the developed tests, the U.S. unemployment rate is found to be steep. Keywords: : DeepnessRegime switchingSkewnessSteepness