The Treasury

Global Navigation

Personal tools


A Comparison of the NZTM and FPS Models of the NZ Economy - WP 03/25

2.2  Recent changes to model specification for the Treasury model

There have been a number of developments made to the Treasury model since it was updated by Murphy (1998). These developments reflect changes in judgements about the appropriate model structure for the New Zealand economy and changes in judgements about how the economy evolves in response to shocks. This section summarises the key changes to the NZ Treasury model structure.

2.2.1  Inflation-targeting rule

When NZM was last updated (Murphy, 1998), the RBNZ used the Monetary Conditions Index (MCI) as its monetary policy tool. As a consequence the NZM model incorporated the MCI as the policy instrument. In addition, given the structure of NZM, the monetary reaction function needed to be specified as a contemporaneous price-level-targeting rule. Since then, the RBNZ has adopted a cash rate as its primary instrument for operating monetary policy and the policy target has changed to keep CPI inflation within a 1 to 3 percent range on average over the medium term.

Drew and Hunt (2000) compared the dynamic responses of FPS to a demand shock under both a price-level targeting rule and an inflation-targeting rule. The key insight from this comparison is that under the price-level rule, an extra cycle was observed with monetary policy needing to be tighter for longer. The variability of inflation and the variability of output are greater under price-level targeting than under inflation targeting because base level drift is not accepted under price-level targeting.

Since the core structure of NZM was based on price levels, it was not possible to formulate an inflation targeting monetary policy rule. This deficiency of NZM led to the development of a relative price structure in NZTM that allows the monetary policy rule to be specified as an inflation target.

2.2.2  Demand-pull framework

In NZM, firms were able to adjust their prices in the medium term to maximize their profits for given input costs. In other words, the key determinants of the inflationary process were cost factors including unit labour costs, the world price of exports and of imports, and the exchange rate. The cost-push approach to modelling the inflation process was commonly adopted in the past in RBNZ models (see Clements, Hansen and Hames, 1986). This paradigm was particularly appropriate for the New Zealand economy for the period prior to the mid 1980s and early 1990s product and labour market reforms. Product markets were protected from international competition by import licensing and tariffs. The wage setting system was a highly regulated system and relatively strong unions could exert influence on the outcome of annual wage movements.

However, the conventional cost-push model for inflation became less appropriate in the New Zealand context after the removal of border protection, price controls, the introduction of both the Reserve Bank Act (1989) and the Employment Contracts Act (ECA, 1991). The ECA made compulsory unionism illegal and led to the decentralisation of the institutional structure of the New Zealand labour market. The ECA achieved a competitive wage setting system based on bargaining at the individual firm level[2]. In addition, using Granger-causality tests Hampton (2001) found that over the 1990s prices have tended to lead wages, rather than the other way round.

The key feature of the Reserve Bank Act (1989) is that monetary policy has the primary objective of controlling inflation. As a result, the Reserve Bank succeeded in reducing price uncertainty and lowering inflation expectations in the 1990s. As a consequence, price and wages are now more responsive to domestic demand factors.

In NZTM, inflation is modelled as a combination of a demand-pull process and a cost-push process with relatively more weight placed on the first channel. In a demand-pull process, inflation pressures develop when demand pressures exceed the economy’s potential to supply. The output gap is used to proxy demand pressures for the economy.

2.2.3  The determination of the real exchange rate

In NZM, the evolution of the real exchange rate and its impact on the dynamic adjustment of the model was unclear. With a three-sector economy, NZTM has two relevant real exchange rates. One is the relative price of imports to non-tradables. The other is the relative price of exports to non-tradables. In a steady state, the framework of the determination of the equilibrium exchange rate is similar to that developed by Edwards (1987, 1989). The equilibrium exchange rate is reached when the economy attains both internal and external equilibrium. Internal equilibrium means that the non-tradable goods market clears in all periods. External equilibrium means that current account balances are consistent with long-run sustainable capital flows.

In NZTM, the long-run sustainable capital flows are determined by the desired ratio of net household wealth to GDP. In theory, this desired ratio could be derived using a full inter-temporal optimisation framework and is a function of various structural factors such as the demographic structure and the stage of a country’s economic development. However, this steady state ratio is set exogenously in NZTM.

The short-run dynamics of cyclical fluctuations in demand and prices causes the economy to deviate from the balanced growth path. Responding to both internal and external imbalances, the real exchange rate would be required to deviate from its steady state value so that the economy is able to converge back to its balanced growth path. Furthermore, financial agents are forward looking so the real exchange rate will adjust immediately in response to an anticipated future shock. The short-run dynamics of the real exchange rate are also influenced by interest rate differentials.

2.3  Recent changes to model specification for FPS

The FPS model has been used to generate the projections in the Reserve Bank’s Monetary Policy Statement since the middle of 1997, as described in Drew and Hunt (1998). The quarterly use of the model to generate published forecasts has meant continuous scrutiny of the FPS calibration, and some significant changes have been made since 1997.

A number of changes have been made in the way inflation process is modelled. Inflation is generated through two sources: excess demand and inflation expectations. The lags through which excess demand feeds into inflation have been shortened and inflation is less sluggish than in the initial model calibration. This feature also means that the monetary policy is able to stabilise inflation more effectively.

There has been a reduction in the response of inflation expectations to cyclical deviations of inflation from its target. This is consistent with the stability of surveyed inflation expectations since the temporary spike in CPI inflation during 2000.

The key instrument for monetary policy has changed from the yield curve to the 90-day rate. While long interest rates still influence the economy, the new specification makes short term interest rates less responsive to fluctuations in long term interest rates.

Another change is that the impact of the exchange rate on the economy has been muted. This means that when the real exchange rate deviates from its equilibrium, the impact on the economy is now smaller and slower This change was made in response to the surprisingly slow impact the weak New Zealand dollar had on New Zealand activity between 1998 and 2000.

Finally, based on recent trends there has been continued adjustment of some of the steady state ratios in the model. For example, the equilibrium exchange rate has been revised down somewhat after the weakness of the currency between 1998 and 2000, and the steady state capital to output ratio has also been lowered.

These adjustments to the dynamic parameters and steady state ratios have been made to bring the model’s initial design closer to the Bank’s evolving view of the economy. But these adjustments have been made to a core model structure that has remained quite stable over the last six years. Basdevant and Hargreaves (2003) describe some features of FPS that may have contributed to this stability, including the treatment of stochastic trends and the continuous review of dynamic parameters.


  • [2]Conway (1999) found evidence that the introduction of the ECA resulted in major changes in the wage bargaining environment and a significant reduction in wages for supermarket workers in the post-ECA period.
Page top