NOT FOR QUOTATION UNOBSERVED COMPONENTS BUSINESS CYCLES FOR NEW ZEALAND. WHAT ARE THEY, AND WHAT MIGHT DRIVE THEM?

Bivariate, classical business cycle work for New Zealand has established that regional cycles exist, relative to an aggregate cycle (Hall and McDermott, 2007a). But despite the apparent robustness of the associated bivariate drivers, the work of Baxter and Kouparitsas (2005) suggests that such relationships are not always maintained when multivariate methods are used. Multivariate methods are more advanced for examining growth cycles than classical cycles. Using unobserved components methodology (e.g. Watson and Engle (1983), Kouparitsas (2001, 2002), Norman and Walker (2007), and Hall and McDermott (2008)), we first establish a New Zealand common cycle from economic activity data for 14 regions, and assess the extent to which the idiosyncratic/region-specific cycles are additionally important. We then aggregate the 14 regions to 5 regions, estimate a similar common cycle, and present the corresponding idiosyncratic cycles. This level of aggregation also provides us with an opportunity to assess the relative strengths of influence on the common cycle of monetary policy variables and several external shock variables; also the extent to which shocks in turn affect idiosyncratic cycles and lead to regional spillover effects. Our preliminary results suggest: (1) that structural breaks play an important role; (2) that Ne