Recently developed methods in the analysis and measurement of latent f
actor models for time series are utilised to study international busin
ess cycles and their relationship to international stockmarket price b
ehaviour. An advantage of these methods is that the duality properties
between time domain and frequency domain approaches for investigating
the properties of time series can be exploited to identify and model
business cycles. The empirical results show that the six countries stu
died, which include the United States, Australia, Canada, United Kingd
om, Germany and Japan, exhibit coherent national business cycles, alth
ough these cycles are not all alike. It is also found that internation
al coherence in economic activity has increased in the flexible exchan
ge rate period, although it is not as strong as it is for the national
business cycles. The coherence between stock-market prices and busine
ss cycles is not strong, both nationally and internationally, but inte
rnational stockmarkets appear to show greater mutual coherence than do
the corresponding economies.