7 Conclusion
This paper has estimated a fiscal SVAR model for New Zealand to investigate the effects of discretionary fiscal policies on New Zealand's economic activity (building on the contribution of Claus et al., 2006). Its contribution over recent New Zealand literature is to explicitly include a feedback from the level of government debt in a manner that incorporates the government’s inter-temporal budget constraint, using the technique in Favero and Giavazzi (2007).
The fiscal multipliers from changes in government spending in New Zealand appear to be positive but small, estimated to be about 0.26 on impact. An unexpected one dollar increase in government spending would typically raise GDP temporarily by around 42 cents in the first year, at the cost of higher interest rates and lower output in the medium to long-run. The effect on GDP is less clear cut from a discretionary increase in taxes less transfer payments, but the results suggest that an unexpected one dollar increase in taxes would lower GDP on impact by 23 cents and have a similarly negative medium-term impact on GDP.
The model included the government's long-term interest rate, which is also likely to be an important variable in the wider economy linked to the cost of capital of firms and the borrowing rate for households. The results show a statistically significant and persistent increase in the nominal interest rates of approximately 7 basis points in response to a one percent increase in government spending. The corresponding impact on inflation is rather modest.
Past fiscal policy was analysed through a historical decomposition of the shocks in the model. This suggests that discretionary fiscal policy has had a generally pro-cyclical impact on GDP over 1998 to 2010, and a material impact on long-term interest rates. For example, the model suggests that discretionary changes in government spending dampened the long-term interest rate over 2000 to 2004 by 45 basis points on average and then exerted an upward contribution by an average 30 basis points over 2005 to 2010.
7.1 Future work
There are a number of possible extensions to the analysis conducted in this paper:
First, the approach adopted here only portrays average estimates of fiscal multipliers across the sample time period. Inferences for policy or forecasting will be appropriate so long as the main structural characteristics of the New Zealand economy remain unchanged. For example, the fiscal multipliers are likely to be different depending on the monetary regime, the capacity for monetary policy accommodation, the amount of fiscal space, and whether there is a financial crisis. The multipliers may also differ over the cycle depending on the underlying state of the economy (IMF, 2012). The model is linear so does not account for this possibility. It may be reasonable to assume that the linear estimates will be broadly reasonable so long as unemployment is within the range of historical experience. For the G7 economies, Baum et al. (2012) find that, on average, government spending and revenue multipliers tend to be larger in downturns than in expansions. Therefore, a useful extension would be to extend this framework to a time-varying VAR setting in order to investigate possible changes to the effectiveness of fiscal policy over time.
Second, a further extension of this work would be to add an external sector and exchange rate. The model did include an interest rate, and detected a positive response in the interest rate in response to fiscal expansion. To the extent that uncovered interest parity holds, changes in interest rates induced by fiscal policy shocks should also affect the exchange rate.
Net tax shocks have a counterintuitive output multiplier at certain horizons (ie, GDP increases in response to a positive net tax shock), consistent with the puzzle discussed in Fielding et al. (2011). Although we show that this result is highly sensitive to the choice of the elasticity of tax to GDP, it also raises the possibility that the model is misspecified and we should be conscious of this when interpreting these responses. It would be useful for further work to investigate alternative means of identifying tax shocks to check the robustness of this result.
