The Treasury

Global Navigation

Personal tools

7  Concluding comments

As the development of a macroeconomic model is an on-going process, the Treasury has embarked on a process of re-development of the core macroeconometric model. The main aim of the new model, the New Zealand Treasury Model, is to provide a theoretically consistent framework for forecasting. This paper has described the steady-state structure and the dynamic structure of this new model.

The model is founded on a well-developed production block with a macroeconomic balance framework on the determination of the equilibrium real exchange rate. The production block provides an estimate of potential output, which is then used to measure the output gap. Unlike the previous macroeconometric model (NZM), the process of inflation generation is based mainly on a demand-pull framework. Like other neoclassical general equilibrium models, the long-run equilibrium is characterised by a balance growth path and stock-flow consistency across all sectors.

The dynamic structure of the model is composed of two key elements. Firstly, some agents adjust partially to the long-run equilibrium, which indicates some inherent inertia in the system. Secondly, some agents in the financial market such as the monetary authority base their decision making partly on expectations. Therefore, modelling expectations formation plays a key role in the dynamic properties of the model.

One of the interesting findings from the shock experiments is that the dynamic path of the external export price shock is quite different from that of the import price shock. This result suggests that it is no longer appropriate to treat exportables and importables as a composite good when there are changes in the terms of trade.

Future developments on the model can be divided into two main areas: validation and policy simulation. Since the dynamic component of the model is based on calibration, it is important to validate the dynamic properties of the model by testing against other empirical models such as the SVAR model developed by Buckle et al (forthcoming). As with any macroeconometric model, the model needs to be continually updated in the light of new data. Hence, the specification of the model may evolve over time.

While the model development has focussed on how to inform the macroeconomic forecasts, there is potentially a role for the model in informing policy through examining the simulation of the model. This would likely involve refining the model in order to capture the greater complexity of how economic agents respond to policy changes.

Page top