RDP 9401: Resource Flows to the Traded Goods Sector 2. Existing Measures of Tradeable Capacity

There are two steps involved in defining tradeable capacity: identifying the traded goods sector, and then allocating resources to it. Most challenging is the first step. In principle, the distinction between traded and non-traded goods is simple. However, in practice, this distinction is not clear cut. Goldstein and Officer (1979) describe a spectrum of traded goods. At one end of the spectrum, a narrow class of goods can be defined as traded. These are goods that enter international trade (that is, actual exports and imports) and satisfy the law of one price. At the other end of the spectrum is a much broader class of goods. These are goods that, at an appropriate relative price, could earn or save foreign exchange (that is, potential exports or import replacements).

Whilst interpretations of what constitutes a traded good are varied, in this paper, in accordance with the Australian literature, the domestic traded goods sector is defined as that which produces exports and import competing goods.[6] There are, however, a number of practical issues involved in making such a definition operational. Few industries produce solely for export or to compete with imports: part of production tends to be consumed locally. Consequently, a judgment must be made about the threshold at which the degree of export orientation or import substitutability warrants inclusion of an industry in the traded goods sector. Given the complexity of this task, it has been common to form proxies of the traded goods sector. The traded goods sector has been variously represented by:

  • the real value of exports plus imports, with the latter as a proxy for import competing goods (Pitchford 1986);
  • the output from a limited number of broad industry divisions that are nominated as traded a priori (Shann 1982; Treasury Dept 1987); and
  • the output from industries that meet particular criteria about the extent to which their production is exported or import competing (BIE 1989, 1990; Wood et al. 1990; Kent and Scott 1991; Dwyer 1992).

Having identified the traded goods sector, the second step in defining tradeable capacity is to allocate resources to it. Kent and Scott (1991) provide a review of estimates of tradeable capacity in the Australian literature focusing on the direction of private investment. From this review, two main approaches to the allocation of resources emerge: the “rule of thumb” approach and the propensities approach. These two approaches are discussed below, where it is shown that estimates of tradeable capacity are sensitive to decision rules about both the classification of traded industries and the method of allocating resources to them.

2.1 The Rule of Thumb Approach

The first, and arguably most common approach, will be described here as the “rule of thumb” approach where industries are described as traded or non-traded a priori and all resources in that industry are defined as tradeable capacity (Shann 1982; Treasury Dept 1987, 1988). The work by Treasury will be taken as representative of this approach. Treasury Dept (1987) defines traded industries as the divisions of agriculture, mining and manufacturing, based on a prior that agriculture and mining are primarily exported oriented, whilst manufacturing is primarily import competing. All other industries are defined as non-traded.

Treasury measures resource flows to the traded goods sector using data on new private fixed investment from the Capital Expenditure Survey (CAPEX). However, given that investment in agriculture is unavailable from this source, their analysis is confined to the sectoral allocation of non-farm investment. Using recent CAPEX data, their estimate of tradeable capacity of the non-farm sector is updated and reproduced below. An estimate which includes a proxy for investment in tourism is also shown, following Treasury Dept (1988).[7]

In Figure 1 it is shown that, by this measure, the share of total investment allocated to the traded goods sector was as high as 63 per cent in the early 1970s. This share fell sharply through most of the 1970s, but rose during the resources boom era of 1979–82 and again during the episode of currency depreciation in the mid-1980s. However, by the end of the 1980s, the share of investment in the traded goods sector was significantly lower than at the beginning of the decade. In recent years, the share of total investment in the traded goods sector appears to have increased.

Figure 1: Real Tradeable Investment as a Share of Total Non-Farm Investment – Rule of Thumb Approach

The rule of thumb approach is, however, impressionistic. Without actual knowledge about the extent to which the output of an industry is tradeable, industries cannot be classified accurately. Any error in judgment is compounded by the use of aggregate data. Even if results are accurate, the share of investment in the traded goods sector as a whole may belie significant developments in the exportable and importable subsectors.

2.2 The Propensity Approach

The second main approach to measuring tradeable capacity will be described here as the “propensity” approach. Industries are defined as traded according to their propensity to produce traded goods. For exportable industries, this propensity is measured by the ratio of exports to total sales or output, while for importable industries it is often measured by some sort of import penetration ratio (Wood et al. 1990; BIE 1989, 1990; Kent and Scott 1991). This approach lends itself to the use of disaggregated data and the identification of a more detailed profile of the traded goods sector and its subsectors. Nonetheless, it has only been applied to identify the tradeable capacity of the manufacturing division (for which disaggregated data are more readily available) and not the economy as a whole.

Users of the propensity approach have also measured resource flows to the traded goods sector with data on new private investment. However, they tend to allocate investment in one of two ways. A benchmark is nominated above which an industry is defined as traded and then all investment in that industry is defined as tradeable capacity (BIE 1989, 1990). Alternatively, the use of arbitrary benchmarks is avoided. Instead, a fraction of investment is apportioned to tradeable capacity with that fraction equal to the industry's propensity to export or compete with imports (Wood et al. 1990; Kent and Scott 1991). The latter method of allocating investment has been more prominent in the literature and will be described here as typical of the propensity approach. The results of Kent and Scott (1991) will be focused upon in this paper.

Applying the propensity approach to highly disaggregated unpublished data, Kent and Scott (1991) estimate the direction of manufacturing investment between 1984/85 and 1988/89.[8] Their results are reproduced in Figure 2.

Figure 2: Real Tradeable Investment as a Share of Total Manufacturing Investment – Propensity Approach

Their estimates, over the period from 1984/85 to 1988/89, indicate that 32 to 38 per cent of investment in the manufacturing division is in tradeable capacity. This share fell slightly in 1986/87, despite currency depreciation, and subsequently increased during the episode of growth in the late 1980s. This increase was shown to be the product of growth in the share of investment allocated to the production of import replacements. The share of investment in export creation, on the other hand, changed little over the period.

Others employing the propensity approach have found similar results.[9] Thus, using this approach, one might conclude that the nation's capacity to produce import replacements had increased, while that to produce manufactured exports had declined. However, the reported increase in import replacement capacity is inconsistent with the continued growth in import penetration. Similarly, the reported lack of growth in exportable capacity is inconsistent with the growth of Australia's exports (in particular manufactured exports) since the mid 1980s.

It might be expected that the increase in the capacity to produce import replacements reported by Kent and Scott (1991) is exaggerated by properties of the data available at the time of their estimation; in particular, the issue of computer prices.[10] However, a more substantive question is the extent to which the reported changes in tradeable capacity are an artefact of an inappropriate methodology.

Certainly, few industries produce solely for foreign markets so that the export orientation of an industry is usefully measured by the ratio of exports to output or final sales. Similarly, few import replacement industries produce solely to compete with imports and are best described according to their propensity to do so. The point is whether or not an import penetration ratio accurately captures this information. Two points warrant attention.

First, high import penetration ratios may exist not because a domestic industry produces import competing goods, but for the very opposite reason that such goods are not readily available to domestic consumers.[11] In such cases, use of standard import penetration ratios will overstate the capacity to produce import replacements and, thereby, will distort information about trends in the subsectoral allocation of investment.

Second, if investment is then allocated to the import competing subsector in proportion to the import penetration ratio, another distortion can arise. For instance, there may be no change in the absolute level of investment but, because of the trend rise in import penetration observed in Australia over the last decade, import creating capacity will be recorded as having increased.[12]

In the following section, an alternative approach to the estimation of tradeable capacity is presented. It draws on elements of existing measures. However, it employs a more sophisticated method of identifying traded goods, in particular import competing goods, than do other studies.

Footnotes

See Salter (1959) for an early debate. [6]

While “exports” of tourism earn foreign exchange, the treatment of tourism as a traded industry is contentious. The tourism industry cannot easily be delimited in terms of production, as are all other industries in the national accounts. Consequently, a proxy industry is often chosen. That used here is “other non-manufacturing service industries” which comprises “entertainment, recreational and personal services” and “restaurants, hotels and clubs”. [7]

They define their propensity to export (compete with imports) as the ratio of exports (imports) to total supply. The data used by Kent and Scott were unpublished ABS, Capital Expenditure Survey data and unpublished data on exports, sales and imports of manufactured goods at the four digit ASIC level. [8]

BIE (1989) report that over the same period, investment in import competing industries increased while that in manufactured export industries declined. Wood et al. (1990), who examined export industries only, reported that there was no clear increase in exportable capacity over the 1980s. [9]

The methodology employed by the Australian Bureau Statistics to measure computer prices, in effect, equates the dramatic rise in power of computers to a fall in their price. This translates into a large increase in the volume of computers that are, say, imported or form part of investment. [10]

For example, as import penetration climbs towards 100 per cent, under the criterion described above, an industry would be classified as one which produces import competing goods. And yet, continued penetration of domestic sales may exist because production of the domestic substitute is falling. [11]

In principle, since there has also been a trend rise in export propensities, the same argument applies to the apportionment of investment to the exportable subsector. However, in the manufacturing division, the propensity to import is significantly higher than it is to export (see Kent and Scott (1991, pp. 31–32)). Thus the impact of the anomaly is exaggerated in the import competing subsector. [12]