作者: Gordon L. Clark
DOI: 10.1111/J.1467-8306.1982.TB01827.X
关键词: Autoregressive model 、 Time series 、 Moving average 、 Population 、 Lag 、 Geography 、 Statistics 、 Internal migration 、 Short run 、 Econometrics 、 Net migration rate
摘要: The dynamics of U.S. interstate migration are considered in the light conventional assumptions regarding regional growth and decline. In-and-out modeled separately using Box-Jenkins techniques, focusing upon underlying temporal processes (autoregressive, moving average, or some combination both) estimated parameters. Gross flows cross correlated for a selected group states to establish lead lag relationships. Analysis indicates that gross is subject distinct fluctuations over short run, each set closely synchronized time. Differences parameter estimates identified do, however, exist between growing declining states. Markovian models shown be susceptible significant error if labor actually moving-average rather than autoregressive structures. Similarly, net very misleading errors forecasting.