Studies the distinction between I. Fisher's implied data-generating process for Monte Carlo cycles and the Markov process. Nonparametric tests for constant hazard expansions and contractions; British business cycle for the period 1854-1938 and 1952 to 1992; Testing for duration dependence.