RDP 9110: Resource Convergence and Intra-Industry Trade 3. Intra-Industry Trade

(a) Models

There are two principal sets of models of intra-industry trade, both resting on the assumption of imperfect competition. The first, developed by Brander and Krugman (1983) focusses on oligopolistic firms' incentives to price discriminate between countries. It assumes that the same good is produced in two countries and that transportation costs are positive. The producers in each country have an incentive to export to the other country (provided transportation costs are not excessive) at a price below the current price. This incentive arises from the fact that sales in the foreign market yield a higher marginal revenue than domestic sales, even though price is lower, as the exporting firm does not suffer a decline in price on the infra-marginal units. The result is two-way trade in identical commodities.

The second and richer set of models rest on consumer preferences which exhibit a desire for variety and on economies of scale in production. These models were developed by Krugman (1979, 1980), Lancaster (1980) and Ethier (1982). The following is a sketch of the simplest model. It closely follows the exposition in Helpman and Krugman (1985). There are two countries (home and foreign), two industries (X and Y), and two factors (capital and labour). One of the industries (Y) uses a constant returns to scale technology to produce a homogeneous good. The other (X) produces a differentiated good. The production of each variety of the differentiated good requires both a fixed and variable cost. It is assumed that the fixed cost is small enough to allow a monopolistically competitive market structure. Technologies are assumed to be the same in both countries and consumers have identical and homothetic preferences represented by the following utility function:

Production of the differentiated good is relatively capital intensive and the home country is assumed to be relatively capital abundant. The home country therefore, imports the homogeneous good and is a net exporter of the differentiated good. Predictions concerning net trade are thus the same as in the conventional H-O-S factor abundance model. However, given the desire for variety in consumption, the home country will both import and export varieties of the differentiated good. The result is intra-industry trade.

Denote the share of home income in world income by s, and home production of the two commodities by x and y. Foreign variables are denoted by a star (*). The price of X in terms of Y is the same in both countries and is denoted by P. Assuming balanced trade the volume of trade equals twice the exports of the home country. Given the structure of preferences, home's exports equal foreign's share of world income (s*) multiplied by the total value of home production of the differentiated good (Px). The volume of trade is thus given by:

The total volume of intra-industry trade equals twice home's imports of the differentiated good:

The share of intra-industry trade in total trade is thus given by:

Given the distribution of income (s and s^* held fixed) intra-industry trade will be greater the closer is x^* to x, i.e. the closer are the outputs of the two country's differentiated goods sectors. These outputs will be closer, the closer are the two country's capital/labour ratios. Thus, the more similar are the resource endowments of two countries the more important should be intra-industry trade. Similarly, the smaller the size of the capital rich country (i.e. the lower is x) the more important should be such trade. In the limit if two countries are identical (s=s^* and x=x^*) all trade will be in differentiated goods.

When we move from a 2 × 2 × 2 world to a multi-country, multi-industry and multi-factor world, the exact pattern of trade becomes indeterminate. However, just as in the H-O-S model, the factor content of country i's imports from j will be higher in those factors which in autarky were more expensive in i than in j. While it is not possible to specify exactly which goods i will trade with j, the analysis in Helpman and Krugman (1985) suggests two testable propositions concerning the importance of intra-industry trade in a world of more than two dimensions.

Proposition 1: As a group of countries' resource endowments become less disparate over time, the share of intra-industry trade in the within group volume of trade should increase.
Proposition 2: The closer are two countries' factor compositions, the more important should be intra-industry trade in their bilateral trade.

In addition to resource dispersion, the simple model suggested that the smaller the size of the capital abundant country the more important should be intra-industry trade (provided the monopolistic competition assumption remains valid). This prediction does not generalize straightforwardly to a multi-dimensional world. The model, assuming monopolistic competition in the increasing returns to scale sector, is only a parable for reality. In small countries fixed costs may prevent the establishment of any firm. As Caves (1981) argues, large fixed costs may result in only a single world producer in which case there will be no intra-industry trade. Such cases are, however, relatively rare. Relaxing the monopolistic competition assumption allows a more important role for country size than suggested by the simple model. Two propositions regarding size are:

Proposition 3(a): As the economic size of a group of nations increases, the exploitation of economies of scale is likely to become more pervasive. Providing this does not lead to a further increase in monopolistic production, intra-industry trade should increase.
Proposition 3(b): The larger the size of the smaller country in bilateral trade the more important should be intra-industry trade in bilateral trade.

These propositions are tested in Section 4.

(b) Measurement

Numerous authors have proposed ways of measuring the importance of intra-industry trade. The simplest and most widely used measure is that proposed by Grubel and Lloyd (1975). Intra-industry trade between countries i and j as a share of their bilateral trade in industry k in year t is calculated by:

where represents exports (imports) of good k by country i to country j in year t. To calculate the proportion of intra-industry trade in total trade, we take a weighted average of the individual industry indices where the weights are the share of the industry in total trade. This paper uses 3 digit SITC data so that k = 238. If there is complete specialization across countries the index takes a value of zero. Alternatively, if exports equal imports in all industries the index takes a value of one and all trade is said to be intra-industry trade.

As Aquino (1978) pointed out this measure is biased when aggregate trade is not balanced. Numerous corrections have been applied in an attempt to correct for the bias; however, as Helpman (1987) argues, these corrections are inadequate as the nature of the bias depends on whether the imbalance is generated by homogeneous or differentiated goods. Consequently, we focus our attention on the above measure of intra-industry trade.

(c) Growth and Importance

The shares of intra-industry trade in intra-OECD trade for each country and for the OECD as a whole are shown in Table 3 for selected years. Overall the importance of intra-industry trade shows a steady increase over the period 1965 to 1979, interrupted only in 1974 by the effects of the oil crisis. In contrast, the period between 1979 and 1985 saw no growth in the importance of intra-industry trade, its share in total trade falling by one percentage point over the period. The latest available data suggest an increase over 1986 and 1987. There are significant differences in both the level and growth rate of intra-industry trade across nations. Such trade is highest in the western European nations (Austria, Belgium, France, Germany, Netherlands, Switzerland and the United Kingdom) and Canada where it accounts for approximately 50 per cent of total trade. In a number of these countries the share of intra-industry trade has, however, shown little increase over the last 10 years.

Of all the OECD nations Australia has the most highly specialized intra-OECD trade pattern with intra-industry trade accounting for just 12 per cent of total trade in 1987. Australia is closely followed by Turkey (15 per cent in 1987) and New Zealand and Greece (16 per cent in 1987). The degree of specialization in Japanese trade is also low compared to that of most other OECD countries. The other striking observation concerning Japanese intra-industry trade is its failure to increase in the 1970s and first half of the 1980s. In 1969, intra-industry trade's share of total trade was just 19 per cent. Seventeen years later the share was the same. In Section 4 we return to the reasons for this low and static share.

In the 1980s, it has been the less wealthy nations that have experienced the most rapid increases in intra-industry trade. Turkey increased its share by 11 percentage points between 1980 and 1987, Spain and Portugal by 8 percentage points, New Zealand by 6 and Greece by 5 percentage points. These increases compare with a 2 percentage point increase for the OECD as a whole.

Table 4 presents the weighted average share of intra-industry trade for each of the 1 digit SITC categories for each country for 1987. The table shows that there is significant variation across SITC classes. As expected intra-industry trade is least important in those industries which are thought of as producing relatively homogeneous goods – i.e. mineral fuels and crude materials. Intra-industry trade is most important in chemicals (SITC 5) where in 1987 it accounted for 55 per cent of total trade. The importance of intra-industry trade in the other three manufacturing sectors (SITC 6–8) is just slightly lower at around 50 per cent.

The last two rows of the table show the increase in intra-industry trade over time in the different sectors. Over the entire period from 1965 to 1987 chemicals (SITC 5), manufactured goods classified chiefly by materials (SITC 6) and food and live animals chiefly for food (SITC 0) achieved the greatest increases. In the ten years to 1987, over which the share of intra-industry trade increased only slightly, chemicals and food products were the two industry groups which achieved the largest increases. Food products was the only one digit category where all countries experienced an increase. The increase in the chemicals category was largely concentrated in those countries whose trade patterns have traditionally been relatively specialized.