On Markov error‐correction models, with an application to stock prices and dividends
提出马尔可夫误差修正模型，允许长期均衡偏离在不同状态下具有不同调整速度，并以美国股价和股息数据为例，展示了如何检验和设定该模型。
Abstract This paper considers Markov error‐correction (MEC) models in which deviations from the long‐run equilibrium are characterized by different rates of adjustment. To motivate our analysis and illustrate the various issues involved, our discussion is structured around the analysis of the long‐run properties of US stock prices and dividends. It is shown that the MEC model is flexible enough to account for situations where deviations from the long‐run equilibrium are nonstationary in one of the states of nature and allows us to test for such a possibility. An empirical specification procedure to establish the existence of MEC adjustment in practice is also presented. This is based on a multi‐step test procedure that exploits the differences between the global and local characteristics of systems with MEC adjustment. Copyright © 2004 John Wiley & Sons, Ltd.