5 Historical analysis of the contributions to New Zealand business cycles: 1984 to 2002
Impulse response functions are useful for understanding macroeconomic dynamics that eventuate from a particular shock. However, many macroeconomic debates are centred on the causes of the business cycle. A typical example is the debate concerning the contributions of changes in interest rates and exchange rates to the New Zealand business cycle either through their impact on exports or domestic investment and housing demand.
The difficulty in resolving these debates typically concerns the accurate identification of interest rate and exchange rate changes that have occurred over and above the endogenous or automatic process of adjustment that takes place in response to shocks originating from other sources. In the context of this structural VAR model for example, a rise in import prices is associated with a fall in domestic GDP, falling interest rates and an appreciation of the exchange rate. To the extent that the simulation captures the actual New Zealand response to an import price shock, these responses represent the configuration of events that commentators would observe at various points in time during New Zealand’s history. However, these observations would not be sufficient to identify whether the observed exchange rate and interest rate changes were having a significant influence on the decline in GDP or whether they were simply endogenous outcomes in response to the common import price shock (or any other shock that generated a similar configuration).
Structural VAR models can be used to help resolve these debates by identifying the relative contributions of different shocks to the business cycle. The purpose of this section of the paper is to use the structural VAR model to identify the relative importance of international and domestic shocks, where domestic shocks include domestic interest rate and domestic climate shocks, to fluctuations in New Zealand GDP around trend during the past 15 years.
The business cycle represented here is the percentage deviation of real GDP from its trend level, which is also known as the growth cycle. Therefore, this part of the paper concentrates on contributions to detrended domestic GDP (
). Following equation (7), the structural VAR can be written in moving average representation to explain the impact of shocks to detrended GDP as follows:
(8)
initial conditions 
where 
is the ith impulse response associated with the jth shock, with 13 shocks in our system.
Figure 10 shows the contributions to detrended GDP () arising from foreign output (
), foreign interest rates (
), foreign equity returns (
), import prices (
), export prices (
), exchange rates (
), domestic interest rates (
), domestic equity returns (
), domestic demand (
) and climate (
) during the period 1983:1 to 2001:2. Each diagram in Figure 10 shows the contribution to detrended GDP (y) from one of the 10 variables listed above. The zero line in each diagram represents the point at which GDP is at trend and it also represents the point where the respective shocks are making a zero contribution to deviations in GDP from trend. For example, if the variable plotted lies above the zero line, it would imply that it had contributed towards moving GDP above trend.
At the beginning of the sample period, initial conditions will contribute a substantial proportion to deviations of real GDP from trend growth. But over time the contributions from initial conditions converge toward zero. Therefore, this analysis applies predominantly to the period since the mid 1980s.
It is apparent from Figure 10 that internationally sourced shocks (that is , , , , and ) in general tend to contribute more to detrended New Zealand GDP fluctuations than shocks that originate from the domestic economy (, , , ).
World output, world equity and the world interest rate disturbances have made large contributions to fluctuations in domestic GDP from trend. In general, fluctuations in domestic GDP around trend have been positively correlated with these shocks
Contributions from export and import prices shocks to real GDP fluctuations vary over the sample period, highlighting the importance of separating out the two variables rather than using a portmanteau terms of trade variable. Contributions from export price shocks have tended to have a relatively long cycle, whereas import prices, at least until the mid 1990s, tended to be more volatile.
Exchange rates contribute very little to the deviations in New Zealand GDP from trend over the sample period. The relatively low importance of shocks from exchange rates is in marked contrast to the attention they receive from economic commentators and highlights the importance of statistical procedures that can identify these contributions.
Shocks from domestic demand and domestic equities have also tended to contribute less to fluctuations in GDP from trend compared to the contribution from foreign shocks. Throughout the 1980s and 1990s contributions from domestic demand shocks have been very small. Domestic equity shocks have tended to have relatively short cycles.
Domestic interest rates have generally been countercyclical. An exception was the 1998 recession when they partially contributed to GDP being below trend. The contribution of domestic interest rates to the 1998 recession was however smaller than the contributions from export prices and climate.
In contrast to interest rates and exchange rates, climate shocks have attracted less interest as sources of expansions and recessions. Climate has also been largely ignored in previous macroeconomic modelling. It is apparent from Figure 10 that in New Zealand climate is an important contributor to the New Zealand business cycle. In particular, climate appears to have been the dominant source of domestic shocks. Climate shocks made significant contributions to the 1991 to 1993 recession and the 1998 recession. Climate also contributed positively to above trend domestic GDP in the mid-1980s and the mid-1990s.
Throughout most of the 1990s, internationally sourced shocks generated a substantial amount of the deviations from trend in New Zealand’s GDP. This highlights the importance of international linkages in understanding the evolution of detrended domestic GDP (y). The most noticeable period where domestic shocks dominated was in the mid-1980s, between 1995 and 1997, and the 1998 recession. The contribution of climate to the 1998 recession was especially important, as illustrated by Figure 10.
