• Home
• International Trade
• New Zealand Tariffs
• Post-2005 Tariff Review
• Import Tariffs beyond 2005
• 5. Alternative Tariff Regimes

• New Zealand Tariffs

• Post-2005 Tariff Review

• Media Releases
• Key Decisions
• Ministerial Report
• Import Tariffs beyond 2005
• Terms of Reference
• Archive

[Released under the Official Information Act]

• 5.1 Approach
• 5.2 Free Trade in 2010 - Scenario Specification
  • 5.2.1 Macroeconomic Results
  • 5.2.2 Industry Level Effects
  • 5.2.3 Regional Effects
• 5.3 Historical Tariff Changes
  • 5.3.1 Dynamic Effects
• 5.4 Conclusion
• 5.5 Appendix: The Business as Usual Scenario

5.1 Approach

The way in which tariff scenarios are investigated with a general equilibrium model is by measuring the effects of an exogenous shock - a change in tariffs - relative to a "business as usual" scenario. The BAU scenario is intended to represent a plausible picture of the economy at some future point in time. Our horizon in this case is 2010. The BAU scenario is not intended to be a forecast as a general equilibrium model is not designed for forecasting. Its strength is in measuring relative comparative static effects rather than absolute effects; for example the change in imports when tariffs are reduced, rather than the starting level of imports.

A brief comparison of the main macroeconomic features of the BAU scenario in relation to 2001 are presented in Section 5.5.

In the sections below we look firstly at the effects of reducing all tariffs to zero by 2010. At the industry level the focus is on the TCF industries. To put this into perspective we then look at the effect of moving tariffs back to what they were around 1986/1987. This is not intended as a plausible scenario for 2010. It is merely to demonstrate what 2010 might look like if there had been no changes in tariffs since1986/1987. We then present a brief discussion on how the results from the general equilibrium model can inform tariff policy after the current freeze ends on 2005.

The obvious criticism of this approach is that the 1995/1996 inter-industry data used in the model already incorporates some of the effects that we are trying to model. Hence the simulation is likely to under-estimate the effects of the reduction in protection that have occurred since 1986/1987.

5.2 Free Trade in 2010 - Scenario Specification

The following three scenarios are examined:

Run 1 - business as usual.

Run 2 - zero import tariffs; wage rates adjust to ensure the BAU level of employment, no change in real government consumption, personal income tax rates adjust to maintain fiscal balance, the real exchange rate adjusts to maintain the BAU current account balance.

Run 3 - as in Run 2 with real wage rates set at the BAU level so that employment is determined endogenously by the model.

5.2.1 Macroeconomic Results

The results of the simulations are set out in Table 5.1 - Table 5.3. Looking firstly at Table 5.1, it is clear that the macro-economy is not sensitive to the complete abolition of tariffs. This is because the size of the shock is much smaller than it seems. In the BAU scenario the revenue from tariffs is approximately $314m and the mean rate of tariff on all manufactured goods combined (excluding petrol) is barely 1%.

Under zero tariffs GDP is projected to be 0.1% higher than in the BAU scenario. The increment in private consumption is imperceptible. With the removal of tariffs potentially causing a reduction in government revenue, personal income tax rates need to rise slightly to maintain fiscal balance. The mean income tax rate on households rises from 22.5% to 22.8%. This largely offsets the 0.35% rise in real wage rates, leading to static real private consumption. Other options for maintaining fiscal balance could also be analysed - for example lower consumption or lower welfare expenditure.

Removing tariffs does lead to lower prices throughout the economy, with the GDP deflator falling by one half of a percent. This improves international competitiveness such that exports rise by around 0.7%, which indeed is necessary to finance the increased demand for imports (as the external balance is prevented from deteriorating).

The Run 3 results are marginally more optimistic, simply because the reduction in prices is allowed to flow through into lower nominal wage rates, with real wage rates held constant. This reinforces the improvement in competitiveness, as manifested by the 0.2% rise in employment - about 3,500 FTE jobs.

For longer run horizons (of at least five years) the fixed employment assumption for closing the labour market is used more often by general equilibrium modellers, than the fixed real wage rate assumption. This is partly because use of the fixed employment assumption makes it easier to isolate the allocative effects of a policy change. It is also argued that the level of employment is ultimately determined by labour supply and demand, not by border protection policy - a reasonable enough stance, but not one which should be pursued too vehemently as it implies that no policies can ever affect the level of employment in the long run.

The two labour market closure assumptions are at opposite ends of a continuum. Undoubtedly a more realistic assumption would lie somewhere in the middle. However, for this particular policy shock, the difference in results is not material.

Table 5.1: Macroeconomic Results

5.2.2 Industry Level Effects

Industry specific results for the TCFL industries are shown in Table 5.2 while results for other manufacturing industries are shown in Table 5.3.

Not surprisingly the largest changes are in the clothing and footwear industries. The mean tariff rates for these industries in the BAU scenario are 14.2% and 13.5% respectively. With the removal of this protection the shares of the market for clothing and footwear held by domestic manufacturers fall by 17% and 22% respectively.

Gross output in these two industries is lower by only about 8%. The reason that the fall in market share does not translate directly into an equivalent fall in output is that both industries also export. For Clothing the export-output ratio is about 39% while for Footwear the ratio is 52%. In addition there is an overall expansion in the demand for clothing and footwear (of around 4% and 5% respectively) as the prices of these goods decline relative to the prices of all other goods and services available in the economy.

Table 5.2: Results for TCF Industries

Note that the projected declines in output of 8% does not mean that output necessarily falls over time, merely that output under zero tariffs would be 8% lower than under the current tariff regime. Spread over a five year period for example, the difference in the annual growth rate would be approximately 1.7%.

For the four TCFL industries combined the total fall in employment is around 1300 FTE positions, or about 4%. This is because the Clothing and Footwear industries account for a higher proportion of TCFL employment than they do of TCFL output.

In each of the individual industries the fall in employment is almost exactly the same as the fall in output. The small differences reflect a marginal increase in capital-labour ratios in Run 2 as wages rise relative to the cost of capital, and a marginal decrease in capital-labour ratios in Run 3 where wage rises are constrained.

Arguably one might expect an improvement in total factor productivity as protection is lowered. Evidence on this is difficult to obtain, but we examine this possibility in a later section.

For the other industries (Table 5.3) there is no change in gross output which exceeds 1%, positive or negative. A few industries see a decline in the market share as import penetration rises, but nonetheless experience small increases in output. Examples are Wood Products and Paper Products.

The industry that experiences the largest loss in market share to imported goods (outside of the TCFL industries) is Other Manufacturing, which includes furniture manufacture. Its market share drops by around 1% which is the main cause the 0.6% decline in industry output.

Table 5.3: Results for Other Industries

5.2.3 Regional Effects

Section 2 presented the changes in employment in manufacturing industries over the period 1981 to 2001 by region, ethnicity, sex, household type and family income decile. With a few exceptions, little correlation with changes in tariffs was evident. This is likely to be because of other events which occurred simultaneously - changes in monetary policy and fiscal policy, the downturn in Asian economies, new employment law and so on.

The following tables show the changes in manufacturing employment from Run 2 on a regional basis. The changes are distributed over the regions pro rata, on the basis of the regional distribution of manufacturing employment in 2001. The implicit assumption therefore is that changes in employment in a given industry are proportionately spread over all regions where that industry exists. Thus any observed changes in regional employment are purely the result of differences in industry composition between regions, not to regional differences within industries.

Probably the best way to interpret the results is as a potential indication of regional susceptibility to changes in tariffs - other things equal. In a given industry less profitable firms will generally be the first to contract or close if they lose protection provided by import tariffs. Hence if there are significant differences in relative firm efficiency within an industry, the regional impacts of changes in that industry could be far from uniform

The large spike is in Mackenzie District (the bar between the Timaru and Waimate labels), but the absolute numbers involved here are very small.

Excluding Mackenzie District, the districts of Horowhenua, Carterton, and South Wairarapa, and to a lesser degree Waitakere City are where the biggest changes occur. However, none of the changes exceed 1.0 percentage point, although again the small size of the industry in places such as Carterton and South Wairarapa means that the closure of a single operation could lead to reductions in employment of much more than 1%. Similar results occur for Run 3.

Overall the impression gained from Figure 5.1 is that in all but a few regions, any regional dislocation caused by further tariff reductions is unlikely to be noticeable above the effects of other economic developments that can be expected to occur, even if differences in intra-industry efficiency mean that firm closures are more uneven than implied by a simple pro rata regional allocation of changes in employment. This continues the historical trend evident in Section 2 that tariff reform has been, and will continue to be a minor contributor to the dislocation costs associated with improving the competitiveness of the New Zealand economy.

Figure 5.1: Change in TCF Share of Manufacturing Employment by Region: Run 2

» Actual Size Version of This Graphic (943 x 521 pixels)

5.3 Historical Tariff Changes

To put the results of Runs 2 and 3 in perspective, we can estimate the effects of tariff changes to date. Ideally this should be done by basing the model in a period prior to the start of the decline in tariffs which has taken place over the last decade and a half, and then modelling that decline. This sort of exercise is beyond the scope of the current study, but as a proxy we run the model for 2010 as if there had been no tariff reductions. The results can then be compared with the BAU scenario.

We should stress that this is by no means a plausible projection of what the economy might look like in 2010. It is merely an artifice, a means by which the effect of past tariff reduction can be examined.

For our starting level of protection we have used the rates of tariff which prevailed in 1986/1987. This year is not the high point in New Zealand's import protection regime, but by this time most border protection was in the form of tariffs, with import licences starting to be sold in tender rounds (thereby providing an estimate of their tariff equivalent). It is the earliest year for which we have reasonably reliable data. The approximate tariff profile in 1986/1987 is shown in Table 5.4. (This data was estimated from the 1986/1987 inter-industry study and thus may not match the data used in Section 2.)

We look at three scenarios, all relative to the BAU:

Run 4 - 1986/1987 tariffs; wage rates adjust to ensure the BAU level of employment, no change in real government consumption, personal income tax rates adjust to maintain fiscal balance, the real exchange rate adjusts to maintain the BAU current account balance.

Run 5 - As in Run 4 with New Zealand exports as price takers.

Run 6 - As in Run 4 with real wage rates set at the BAU level so that employment is determined endogenously by the model.

Table 5.4: Approximate Mean Tariff Rates 1986/1987

Run 4 shows a reduction in GDP of about 0.3%. In other words had the tariff regime which prevailed in 1986/1987 still existed, the gain in 2009/10 from reducing tariffs from their 1986/1987 levels to their actual current rates is 0.3% of GDP. This equates to around $414m in 1995/1996 prices, or about $300 per household. As noted earlier though, the use of the 1995/1996 inter-industry table probably biases this result downwards.

It will be seen that private consumption is actually somewhat higher than in the BAU. This is a common result in general equilibrium modelling of tariff changes because of the inclusion on downward sloping export demand curves.¹⁵ Higher tariffs raise industry cost structures and this leads to a marked decline in the volume of exports, but those exports are sold at a higher price; the terms of trade rise by 1.96% in Run 4. Higher terms of trade translate into higher Effective GDP and thus higher domestic absorption.

Table 5.5: Macroeconomic Results for 1986/1987 Tariffs

In Run 5 this effect has been nullified by forcing New Zealand exporters to sell at the world price. The result is drop in private consumption which is twice the size (in absolute terms) of the gain seen in Run 4. This is reflected in imports which decline by 5% compared to 2.9% in Run 4.

Another key assumption in Run 4 is that the real wage adjusts to maintain employment at the BAU level. In other words, no compensation is sought in gross wage rates for the rise in consumer prices brought about by the increase in tariffs. This requires a drop in real wage rates of 1.6% (and 2% in Run 5).

Run 6 looks at the situation when real wage rates are held at BAU level. (Note that the fixed terms of trade assumption in Run 5 does not apply.) Mirroring the changes between the two free trade runs (2 and 3), holding real wage rates fixed leads to a drop in competitiveness. This feeds through into lower levels of economic activity with exports down by 5.1%, employment down by nearly 23,000 FTEs, private consumption down by 0.4% and GDP down by 0.9%.

5.3.1 Dynamic Effects

A criticism of this type of modelling is that it fails to take more subtle and dynamic changes into account. For example with respect to the labour market closure assumptions, is it possible to comment on which of the two extremes (Runs 4 and 6) is the more realistic? A common characteristic of protection regimes is that they encourage rent seeking behaviour, not just in this particular sphere, but more generally as economic agents (exporters, manufacturers, labour groups etc) realise that with enough lobbying it may be possible to secure economic privileges and advantages over competitors favours. This is likely to extend to wage bargaining as more protected economies are susceptible to a cost-plus mentality.

One of the rationales for removing protection is that governments can signal that they are not (or are no longer) in the business of granting politically expedient favours. Non-competitive behaviour in one area of the economy is more difficult if other areas of the economy are becoming progressively more competitive. Accordingly the power of labour organisations to artificially maintain or enhance real wages is less in an open economy, implying that the behaviour of the labour market is not invariant to the level of protection, fitting more into the "fixed real wage" behaviour in a situation of high tariffs.

Another facet of protected economies is that they are thought to be less efficient - not just in terms of allocative efficiency (the standard framework for studying the effects of a tariff), but also in terms of productive and dynamic efficiency. If this is the case then a reduction in tariffs produces not only a level shift in the long run rate of economic growth, but also raises the growth rate itself. Evidence for this is scarce, but some recent research for the Australian Productivity Commission¹⁶ produced estimates for the elasticity of the percentage change in value added in response to a given change in industry assistance. These are shown in Table 5.6

Table 5.6: Effect of a 1% Change in Assistance on Value Added

Table 5.7 shows the macroeconomic results of testing these findings in the context of Run 6. The relative changes between Runs 6 and 7 could be applied to Run 5 to estimate the effects of lower productivity in the fixed employment context.

Table 5.7: Macroeconomic results for 1986/1987 - Tariffs and Lower Total Factor Productivity

The estimated productivity effect increases the loss in GDP attributable to the higher tariffs by about 25%, although in absolute terms the effect is small. The total change in GDP (in Run 7) amounts to $1500m in 1995/1996 prices. This is roughly $1100 per household per annum, or about $21 per household per week.¹⁷

5.4 Conclusion

Runs 2 and 3 illustrate the effects of the largest possible reduction in tariffs, that is moving to zero tariffs. Thus it is reasonable to assume that partial reductions in tariffs would yield even smaller effects than those observed for Runs 2 and 3, although there is a possibility that uneven reductions could lead to some unexpected industry or regional effects.

Nevertheless we would have to conclude that the macroeconomic effects of a partial reduction in tariffs from their current levels are too small to be identified with a (neoclassical) general equilibrium model. Even on the assumption that there would be some dynamic efficiency gains, the macroeconomic gains are unlikely to exceed 0.1%.

Accordingly arguments in support of further reductions in protection really need to be sought elsewhere. For example:

• administrative costs of maintaining and enforcing tariff schedules,
• compliance costs to importers (who are often also manufacturers),
• further reducing the perception that some industries are favoured by government for no obvious reason.

Lest this be interpreted as a case for raising tariffs, the model results make it clear that increasing tariffs does not enhance welfare. The simple point is that New Zealand has secured most of the gains from lower import protection already; firstly by the removal of import licensing, secondly by bringing down the "tall poppies" and thirdly by lowering the entire tariff profile to the point where the mean price of goods imports is about 1% higher than what it would be without tariffs.¹⁸ The additional gains from a possible fourth stage - free trade - are small by comparison.

5.5 Appendix: The Business as Usual Scenario

Policy analysis with a general equilibrium model requires a baseline against which policy shocks may be compared, much like a scientific experiment requires a control population. The Business as Usual (BAU) scenario fulfils this purpose. It is intended to represent a plausible picture of the economy in 2010 without any major policy changes or major international events. It is not intended to be a forecast of 2010.

The model consists of a set of unknown variables (prices and quantities etc) and a set of behavioural equations which determine those variables. To solve the model the number of variables should equal the number of equations. However, some parts of the economy are not well enough understood to be represented by an equation. The processes by which wages are set is an example. In addition, other parts are "exogenous" in the sense that they are (largely) determined outside the system. Examples include government policy such as the tax regime, and international factors such as oil prices.

Hence to produce a BAU scenario a number of assumptions are required. The most important of these relate to total factor productivity and export demand. The latter is modelled as shifts in demand curves with the model choosing the price-quantity suppy mix for each commodity.

Table 5.8: Factor Productivity - Business as Usual

The macroeconomic results are shown in Table 5.10. Gross domestic product and private consumption both rise at about 3% pa, with somewhat faster growth in exports and investment. Employment (in FTE jobs) expands by 1.6% pa, which lowers the unemployment rate to 5.5%. On a per capita basis, GDP rises by 2.5% pa.

Table 5.9: Export Growth - Business as Usual

Table 5.10: Macroeconomic Results ($m 1995/1996)

¹⁴Although these results are shown to two decimal places, this is spurious accuracy.

¹⁵See for an example Productivity Commission 2000, "Modelling the Effects of Removing General Tariffs," Supplement to Inquiry Report, Review of Australia's General Tariff Arrangements, Productivity Commission, Canberra, July, p19.

¹⁶op cit pp 63-65.

¹⁷The NZIER estimated a similar figure of $22/week in household spending power. See Consumer Benefits from Import Liberalisation: A New Zealand Case Study, NZIER, for Ministry of Foreign Affairs and Trade, June 1999.

¹⁸The Swiss formula approach delivered most of the gains under these last two stages.