Essays in business cycle measurement.

This dissertation is concerned with the issue of economic fluctuations; the following related topics are analysed: co-integration and the NAIRU hypothesis: the theoretical implications of different classes of models, some implying that the NAIRU is a structural parameter that can only be influenced by supply-side measures, others that the attainable level of unemployment is a function also of demand variables, are first discussed; co-integration techniques (the Engle-Granger and the Johansen procedure) are then used to test the NAIRU hypothesis; the more powerful maximum likelihood method developed by Johansen shows that the unemployment rate is co-integrated with both supply and demand variables only as well as a combination of the two; supply versus demand shocks as the driving force of business cycles: using two measures of productivity growth (the Solow residual and the dual residual from the cost function), competing theories of the cycle are tested in a number of OECD countries; the issue of market structure and its relevance to explain economic fluctuations is also addressed; the empirical evidence refutes the "stronger" real business cycle (RBC) hypothesis that denies the role of demand shocks; aggregate versus sectoral shocks: their relative importance in the UK economy is evaluated by estimating a vector autoregression (VAR) of the output growth rates of 19 industrial sectors and doing a factor analysis on the innovations; the one-factor model performs quite well when applied to the British data implying that there is an aggregate shock that can account for a high percentage of the fluctuations of output over the cycle; the "seasonal cycle" in the UK economy: the quantitative importance of seasonal fluctuations and the existence of a "seasonal cycle" whose main features are very similar to those of the conventional business cycle are documented by running regressions with seasonal dummies and band spectrum regressions; a one-sector, neo-classical model of capital accumulation in which seasonal preferences are explicitly incorporated (the coefficient of risk aversion depending on the season s) is then set up; the model is not rejected by the data, confirming that seasonality is a feature to be explained within the economic model