RDP 2006-01: Modelling Manufactured Exports: Evidence from Australian States 6. Domestic Influences on Export Outcomes

The implicit assumption in the modelling thus far has been that the estimated equations represent export demand curves, with the Australian export supply curve taken to be perfectly elastic (so that prices can be assumed to be exogenous). This is an approach taken in much of the previous literature, and accords with the notion that Australian firms satisfy all foreign demand for their goods at a given price. However, it is likely that domestic conditions (domestic demand and capacity utilisation) influence manufacturers' desire or ability to supply exports. For example, for a firm that is a price-taker on world markets, an increase in domestic demand will in theory cause exports to be reduced one for one to satisfy that demand.

Alternative models have also been developed to examine firms that have some degree of pricing power in world markets.[12] For example, Ball (1961) presented a model in which firms set marginal revenue from exports equal to marginal revenue from domestic sales and marginal costs, implying that changes in domestic conditions influence its desired export sales. Alternatively, if firms respond to changes in demand with a lag, perhaps reflecting delays in expanding production, then increases in domestic demand may also cause such firms to divert production from export to domestic sales, if domestic sales are more profitable (because of transport costs, for example). Similarly, Artus (1970) suggested that changes in domestic demand may cause firms to alter the effort (such as marketing) they exert to sell products overseas, thus influencing their non-price competitiveness and affecting export sales. These considerations have resulted in a large literature that allows for the possibility that exports also depend on domestic demand conditions, a possibility which is allowed for in the following section.

One way to capture these domestic influences is to estimate a simultaneous equation model, with both export volumes and prices modelled as functions of explanatory variables (including relative prices, world demand and domestic demand). While this is the most common approach in the literature, the results for Australian states suggest the price equation has little role to play in modelling exports,[13] consistent with Australian manufacturers being price-takers on world markets.

A more appropriate way to measure the influence of domestic demand pressure on Australian manufactured exports is to include some proxy for such pressures in our earlier DOLS model. The augmented model thus becomes:

where Kt is a proxy for the strength of domestic demand (relative to supply) – so that β is expected to be negative – and y = {z, K}.

A natural proxy for such pressure is domestic final demand (DFD). Alternatively, a measure of capacity utilisation in the manufacturing industry may be an effective proxy of the strength of domestic demand relative to supply. Two alternative measures of capacity utilisation are used; one calculated by the Australian Chamber of Commerce and Industry and Westpac (the ACCI-Westpac survey) and the second by the National Australia Bank (NAB).[14] The profiles of these series are shown in Figure 4.

Figure 4: Domestic Demand Proxies
Figure 4: Domestic Demand Proxies

The properties of these proxies are as follows. Domestic final demand is an I(1) process, according to unit root tests, and is cointegrated with exports, trading partner GDP and the real exchange rate. Both capacity utilisation series are also I(1), indicating that the spare manufacturing capacity created by the recession in the early 1990s has been gradually utilised. A cointegrating relationship is also found when these capacity utilisation series are included in place of domestic final demand.

The domestic demand elasticity of manufactured exports is shown in the left-hand column of Table 9. This elasticity is significant and of the expected (negative) sign for NSW, Victoria and Queensland, resulting in a significant negative mean-group estimate of the national elasticity. Similarly, the fixed-effects estimate of the domestic demand elasticity is also negative and significant. These results support the hypothesis that domestic demand pressure influences exports independently of competitiveness, consistent with the theory presented earlier. However, interpreting the magnitude of this result is made difficult by the instability of the income elasticity when domestic final demand is included; for those states in which domestic final demand is found to reduce manufactured exports, the income elasticity rises to implausibly large levels of between 5 and 6 (this also occurs in the fixed-effects estimate). This instability in the income elasticity estimate is likely to stem from the collinearity between trading partner GDP and domestic final demand.

Table 9: Elasticity of Manufactured Exports to Domestic Demand Pressure
  Domestic
final demand
Capacity utilisation
ACCI-Westpac measure NAB measure
Vic −2.18** −2.21** −0.87
NSW −3.45** −1.57 0.97
SA 1.03 −3.92** −4.10**
WA 0.20 −1.24 −0.97
Q1d −3.05** −1.92 −0.46
Australia
Mean-group estimate −1.88** −2.18** −0.83
Fixed-effects estimate −1.70** −2.23** −1.08*
Notes: * and ** represent significance at the 10 and 5 per cent levels respectively.

Using the ACCI-Westpac measure of manufacturing capacity utilisation as a proxy for domestic demand pressure provides more stable results that also suggest a role for domestic demand pressure in determining manufactured exports. Increased capacity utilisation is found to constrain manufactured exports in Victoria and South Australia, with both the mean-group and fixed-effects estimates of the elasticity negative and significant (middle column, Table 9). The price and income elasticities are also largely unchanged from the baseline model. Increases in the NAB measure of capacity utilisation are found to have an insignificant effect on national manufactured exports using the mean-group estimate, with only the South Australian elasticity significant (right-hand column, Table 9). Nonetheless, the sign of these elasticities are as expected for most states, and the price and income elasticities are stable. Furthermore, the fixed-effects estimator finds a negative coefficient on the NAB measure of capacity utilisation that is significant at the 10 per cent level.[15] While these results are not completely satisfactory, they are highly suggestive that domestic demand pressure has some role to play in determining manufactured exports.

Footnotes

Dwyer, Kent and Pease (1993) found that Australian manufacturers are price-takers on world markets. For contrary evidence, see Swift (1998). [12]

Specifically, the supply price elasticities are very large, and the volumes equation is relatively unchanged from results shown earlier in the paper. [13]

Domestic final demand and both measures of capacity utilisation are national, rather than state-specific, measures. This is done because national measures of demand pressure should better capture the incentive to divert production to domestic (local or interstate) sale. [14]

An alternative method would be to estimate a non-linear relationship between exports and capacity utilisation (such as squared capacity utilisation), or to estimate separate elasticities for periods of low and high capacity utilisation (using interactive dummies). The use of interactive dummies produces qualitatively similar results, although the high-capacity utilisation elasticities are significant for more states than when a single elasticity is used. Results using squared capacity utilisation are largely unchanged from those using its level. [15]