RDP 9312: A Re-examination of the Determinants of Australia's Imports 4. Data

The data used in the estimation of import demand are listed below with a brief comment on their features. All data are quarterly and expressed as natural logarithms. Where data are in index or constant price form, 1989/90=100. Where seasonal patterns are evident, seasonally adjusted data have been used. Further details and sources are given in Appendix 1.

Import volumes are represented by real endogenous imports, excluding computers. Computers have been excluded because the rapid pace of advance in computer technology has posed problems for measurement of their constant price value (Meade 1991; McCarthy 1989). In essence, the dramatic rise in power of computers in recent years has been equated with a fall in the unit price of such power. In consequence, while the value of computer imports as a share of GDP has risen little since the mid 1980s, the volume has risen significantly. Since estimates of import demand are sensitive to changes in the reported volume of computer imports, computers have been excluded.^[13]

There are a variety of published price series for imports. That used here is an implicit price deflator for endogenous imports (excluding computers). It records import prices ‘free on board’ – that is, over the docks – and excludes tariffs. A proxy for the landed price of imports is also constructed by multiplying the implicit price deflator by the average rate of duty paid. Ideally, a landed price would also include the scarcity premium that arises from import quotas, transport costs and all other charges that account for the difference between prices over the docks and those that face purchasers.^[14] Given that such information is generally unavailable, it is common practice to assume that such changes have not been significant over the period or are, at least, less significant than the changes in tariffs (Gregory and Martin 1976; Athukorala and Menon 1988).

Real GNE is nominated as the appropriate activity variable as it represents domestic demand for goods.^[15] In practice, however, the explanatory power afforded to activity is altered little by the choice of activity variable (Horton and Wilkinson 1989; Wilkinson 1992).^[16]

Activity and price variables are also decomposed into the components that correspond to consumption, capital and intermediate imports. Given that imports are not a homogeneous bundle of goods, this facilitates an examination of the extent to which the behaviour of aggregate imports is driven by a given component class.

Finally, a measure of openness was required. A popular measure of openness is the ratio of imports to GDP.^[17] However, use of such a measure would introduce endogeneity into a function for import demand: it is the very thing that we want to explain. Consequently, we measure openness by changes in protection and use the effective rate of assistance calculated by the Industry Commission. A decline in the effective rate of assistance is interpreted as an increase in openness.

Footnotes

Bullock, Grenville and Heenan (1993). [13]

An attempt is also made to account for tariffs and the scarcity premium that arises from import quotas by using the Industry Commission's nominal rates of protection for each of the twelve major industry categories in the manufacturing division. However, as discussed by Phillips (1989), there are difficulties associated with the use of nominal rates of protection for this purpose. In compiling measures of nominal protection the Commission uses weights proportional to unassisted output values, rather than import shares. Given this limitation, landed prices derived from the nominal rate of protection are not reported in this paper. Results are, however, broadly consistent with those generated by inclusion of tariffs only and are available from the authors. [14]

GDP would be an appropriate measure of activity if imports are primarily intermediate goods, the use of which is a function of domestic output rather spending (Horton and Wilkinson 1989, p.6). Intermediate goods in 1991/92 accounted for about 50 per cent of Australia's total real endogenous imports. Given this share, it is common for either type of activity variable to be used. [15]

Although, it may affect the elasticities of the nominated relative price terms in an import demand function. [16]

See, for example, Romer (1991). With respect to discussion of Australia's increased openness, see Bullock et al. (1993). [17]