Abstract
On the basis of data from the Canton of Zurich for the period from 1880 to 1983 this paper tests three models of political control: the ‘crime control’, ‘conflict’, and ‘economic crises’ models. It is suggested that each of the models might be valid for a particular sub‐period. The identification of sub‐periods is based on the idea of Kondratiev cycles. For each sub‐period the effects of crime rates, the frequency of strikes and of bankruptcies on the number of police personnel and the severity of convictions are estimated by means of ARIMA modeling. The results show different patterns for each sub‐period. For the period from 1880 to 1933 growth rates of the indicators of political control are best explained by the frequency of strikes. For the second period there is a strong direct effect of economic crises on the extent of political control.