Agribusiness Review - Vol. 7 - 1999

Paper 9
ISSN 1442-6951

Intra-industry trade of processed food in APEC economies

Tina Yiping Chen and Ray Trewin ¹

Australia-Japan Research Centre
Asia Pacific School of Economics and Management
The Australian National University
ACT 0200, Canberra, Australia

• 1. Introduction
• 2. Literature survey
• 3. Measuring IIT
• 4. Processed food IIT in APEC
• 5. Determinants of IIT in processed foods
• 6. Conclusion
• References
• Footnotes

1. Introduction

A considerable part of the growth in world trade, particularly amongst developed countries, is of an intra-industry trade (IIT) nature - the simultaneous export and import of products that are very close substitute for each other in terms of factor inputs and consumption (Tharakan, 1985) . Since such trade is very difficult to explain in a neoclassical model, a substantial amount of literature has developed to explain such trade. But there is still room for studies of such trade both theoretically in general and empirically in particular. In light of this gap, this paper examines the trends and patterns of IIT in the processed food 2 sector for APEC economies.

The focus of this paper on processed food IIT in APEC economies is important for three broad reasons. First, most empirical work on IIT has focussed almost entirely on manufactured products in general. Processed food products are more or less regarded as one component of broad agricultural products whose trade is dominated by comparative advantage (inter-industry trade in nature). But being a component of manufactured industry, processed food industries tend to have imperfectly competitive market structures characterized by high seller concentration, some degree of plant level economies of scale, and product differentiation (McCorriston and Sheldon, 1991) . Also larger growth has been observed in trade, investment and other economic components of the processed foods sector than in the raw material sector, although agriculture and related agribusiness sectors have tended to decline in terms of their position in the economy. At the same time, it is believed that urbanisation has played an important role in the increased consumption of processed foods. The more consumers change their consumption patterns towards processed foods, the more likely diverse tastes will be generated and grounds provided for demand driven IIT. These suggest that IIT in processed foods may be an issue of interest.

Second, most of the current empirical work on IIT deals with European or OECD economies, and there is little empirical analysis of IIT for APEC economies. APEC economies differ considerably from European economies and amongst themselves in respect of per capita income, size of economy, level of industrialisation, exposure to external trade, commodity patterns of trade and so on (see Pacific Economic Cooperation Council 1995 and Fukasaku 1992) . Despite the recent financial crises experienced in some of APEC economies, APEC is one of the most dynamic regional groupings in the world economy. Further more, agribusiness trade and investment is a high priority issue within APEC economies as reflected by the food sector being accepted amongst nominations of sectors for consideration of Early Voluntary Sector Liberalisation (EVSL) at a recent APEC summit. Therefore, the pattern of IIT in processed food in the Asia and Pacific region will be of particular interest.

Third, study of IIT gives rise to potentially important policy results. From a welfare perspective, trade liberalisation under IIT will bring about not only the traditional gains from exchange and specialisation, but also gains from greater product variety and increased scale economies. From an adjustment perspective, it has been suggested that if trade is of an intra-industry nature, then industry adjustment to competitive forces will be easier than if it were inter-industry in nature (See Greenaway and Milner 1986 for a discussion, Chen 1998 for a more solid empirical analysis).

2. Literature survey

Theoretical explanations of IIT

The first group of models considered that explain IIT incorporates competition between a large number of firms. There are several such models within the Heckscher–Ohlin framework. These models are based on factor endowment differences and specify production as different combinations of basic factors such as capital and labour in a way that is consistent with constant returns to scale and perfect competition, demonstrating that the pattern of IIT is driven by relative endowments.

One well-known such model was developed by Falvey (1981) . Based on differences in factor endowments, this model reveals that IIT occurs along vertically differentiated products giving the reciprocal demand for both high and low qualities of a product between two countries. On the other hand, for models incorporating monopolistic competition, scale economies in production and diverse consumer tastes, horizontally differentiated ITT has been explained by Spence (1976) , Dixit and Stiglitz (1977) , Krugman (1979) and Lancaster (1980) . The main idea behind these models is that, if the number of varieties enters directly into consumers' utility function (desire for variety), the economies of scale limit the number of varieties in production, then IIT indeed may take place and, by increasing the number of varieties, have positive welfare effects.

The second group comprises oligopolistic models that focuses on the strategic interdependence between firms in an industry. A distinguishing feature of these models is the form of conjecture assumed to influence a firm's decision. Brander and Krugman (1983) , using a Cournot-type conjecture, developed a model which explains IIT in an identical commodity which is often referred to as two-way trade or ‘cross-hauling'. That this two-way trade can occur is a consequence of price being above marginal cost in both markets, both producers seeking to maximise their profit by selling to both markets, taking the sales of the other producer as given so long as transportation costs are not high.

Some general empirical studies

Several economists have estimated the degree of IIT. The results of a comparative study undertaken by Greenaway and Milner (1989) suggest three important findings. First, as expected, the level of IIT is lower when a more detailed level of industry classification is applied. Second, the level of IIT is higher for manufacturing than for other industries. Third, among different economies, IIT is dominant for all developed market economies (DMEs), especially in trade between the DMEs. It is less important but still significant for the newly industrialising economies, but it is only of relatively minor importance for less developed countries.

Greenaway and Milner (1989) surveyed the literature on the testing of hypotheses about IIT. Their survey covers a wide range of empirical studies including those of Pagoulatos and Sorenson (1975) , Finger and De Rosa (1979) , Loertscher and Wolter (1980) , Caves (1981) , Toh (1982) , Lundberg (1982) , Culem and Lundberg (1986) , Havrylyshyn and Civan (1983) , Bergstrand (1983) , Tharakan (1984, 1986) , Greenaway and Milner (1984) and Balassa (1986a, 1986b) . Since then, other important studies have been published, such as those by Lee (1989) , Lowe (1991) , Fukasaku (1992 ) and Clark (1993) . The hypotheses tested in these studies are either derived from various theories of IIT or suggested by more casual empiricism. According to Greenaway and Milner (1989) , the hypotheses can be grouped under three headings: country-specific variations in IIT intensity for any given industry dependent on the characteristics of the trading partners; industry-specific variations in IIT intensity across industries dependent on commodity/industry-specific demand and supply characteristics; and policy-based variations in IIT intensity influenced by policy/institutional factors.

The major country-specific hypotheses are that the average levels of IIT will be high: (1) in DMEs compared with LDCs because of differences in income and in economic structure; (2) in ‘large' economies compared with ‘small' ones since the scope for product diversity and economies of scale may be expected to be higher in the former; (3) when there is taste overlap between trading partners, since this may increase the scope for the exchange of differentiated commodies; and (4) when trading partners are geographically close, either because proximity means lower transport costs or because of similarities of culture and taste.

There are five industry-specific hypotheses. IIT will be higher: (1) the greater the product differentiation; (2) in commodities where there is scope for scale economies; (3) when the market structure tends towards monoplistically competitive conditions; (4) when there is potential for product cycle trade and/or technological differentiation; and (5) when there are more multinational corporations.

The two policy-based hypotheses are that IIT will be greater (1) when tariffs and non-tariff barriers are low; and (2) when economies are subject to some form of economic integration.

Existing econometric studies which test some of these hypotheses generally confirm the expected signs of the estimated coefficients. In some cases, the scale economy variables are less consistent; and tariff barriers are often an insignificant variable. A major difficulty with such studies is to obtain data which are appropriate proxies for the explanatory variables, as economic theory suggests. This is especially so for two important industry-specific explanatory variables: product differentiation and scale economies. Given these difficulties, the explanatory power of the regressions in these studies is often low. Another feature of the econometric studies in the field is that there are very few studies of vertical product differentiation and the activities of multinational corporations as they affect IIT. Greenaway, Hine and Milner (1994) use an intuitively plausible criterion to disentangle vertical and horizontal IIT in the bilateral trade of the United Kingdom, and show that in that country over two-thirds of all IIT is vertical. From this finding, it is worthwhile to distinguish between horizontal and vertical IIT and to work on their explanations separately. On the other hand, using the case of the automobile industry, Becuwe and Mathieu (1992) show that intra-firm trade is the major determinant of IIT in that industry.

In recent years, a little literature of IIT on processed food has emerged. Probably the first paper on this was published by McCorriston and Sheldon (1991) . In their paper, the authors examine trade in a sample of high-value products for the US and the EC using indices of IIT and intra-industry specialisation. The results indicate that for total trade in 1986, the EC exhibited more IIT across the sample than the US. Further, over the period 1977-86, the EC had a greater tendency towards intra-industry specialisation in its geographical pattern of trade than the US. Based on these results, Christodoulou (1992) tests formally a general and a restricted set of country and industry hypothesis for the EC meat market. Results from this study suggest that tastes overlap as reflected by countries' cultural and economic convergence, as well as the imperfectly competitive structure of the market, as supported by product differentiation, significantly explain IIT. Based on Helpman and Krugman's model, Hirschberg, Sheldon and Dayton (1994) analyse determinants of IIT in food processing for a 30-country sample over the period 1964-85. This study concludes that IIT is a positive function of a country's GDP per capita and the equality of GDP per capita between countries. Hirschberg and Dayton (1996) using the 30-country sample again test the determinants of IIT on processed food at the detailed industry level. And it is found that certain industries are significantly more prone to engage in IIT given the increasing similarity of the technology of the countries that trade than others.

3. Measuring IIT

There are two major issues in empirical studies involving IIT: defining an industry and measuring the extent of IIT within such industries.

Two main criteria have been used to define an industry. Different products are the output of a single industry either if they are relatively easy to substitute one for the other in the production process or if consumers put them to essentially the same use. The choice between these two criteria depends on the use to which the data generated by the criteria is put. Economists studying IIT often use data from published statistics on trade in various recognised ‘categories'. The most commonly used classification is the Standard International Trade Classification (SITC). There are 10 sections at the 1-digit level, 63 2-digit groups, and so on.

The level of IIT depends crucially on the level of aggregation. If IIT is measured at a very detailed classification, there might be very little IIT. On the other hand, if IIT is measured at very high level of aggregation, much trade will be IIT. Although relatively little effort has been devoted to measurement problems, they are of absolutely vital significance to the entire subject. The standard documentary work often reports IIT at the 3-digit level of aggregation. Efforts have been made to find a more constructive way to take some account of the bias due to the aggregation level or even to establish a more appropriate trade classification.

One measure of the level of IIT was proposed by Balassa in 1966. The most widely used measure is the Grubel–Lloyd index. The problem with the Grubel–Lloyd index is associated with biases created by trade imbalances at the multilateral level (Grubel and Lloyd 1975) . Some economists have attempted to correct this, but a widely acceptable method of correction has yet to be found. As argued by Helpman (1987) , attempts to modify the Grubel–Lloyd index to correct for trade imbalance bias are inappropriate since the nature of the bias is not known. In particular, we do not know whether the imbalance is caused by homogeneous or differentiated products and whether the trade structure is in equilibrium or not. This explains why in general, bilateral IIT is more interesting than overall IIT. 3

Following Grubel and Lloyd (1975) , the individual industry IIT index between countries i and j for product k in year t is given by

(1)

X X and M are exports and imports of product k in year t between two countries respectively.

The aggregate IIT index is calculated as

(2)

This is a weighted average of the individual industry indices, where the weights are the share of the industry in total trade.

4. Processed food IIT in APEC

Trends of total processed food IIT

IITs of processed food in APEC economies are significant for most of the economies in this region. Over the last nearly three decades, processed food IIT in this region experienced significant change but there is no uniform pattern amongst all the economies.

There are, of course, many methodological questions which have been raised in the literature about calculation of IIT indexes. See Chen (1998) for detail.

The IIT of processed food in this study is measured as a Grubel - Lloyd index for ISIC 31 (Manufactured food, beverages and tobacco) industry at 4-digit level through 1970 to 1996 4 . Data are obtained from International Economic DataBank (IEDB), the Australian National University. The trends of processed food IIT in APEC are shown in Figure 1.

For the 18 APEC economies, the levels of processed food IIT during the early 1970's can be classified into three categories. Say high (over 50 per cent of total processed food trade is of IIT), medium (level of IIT in total trade of processed food is between 25 to 50 per cent) and low (level of IIT in total trade of processed food is less 25 per cent). Canada and Singapore are the only two economies that have more than 50 per cent of processed food IIT in total trade, in which Canada has a little over 55 per cent and Singapore has almost 70 per cent. There are 8 economies in this region whose level of processed food IIT fall in the medium range. They are Chile, Indonesia, Japan, Korea, Malaysia, Mexico, Papua New Guinea and USA. For the rest of the 8 economies in this region, the levels of processed food IIT are well in the low range, including Australia and New Zealand. In summary, the levels of processed food IIT of APEC economies in early 1970's are widely different with 10 of its economies having significant evidence of IIT in processed food and another 8 economies no significant evidence.

Figure 1 also illustrates changes over 1980's and early 1990's in IIT in processed food in APEC economies. Several economies experienced significant change, so that their processed food IIT changes from one category to another. The levels of IIT in processed food of Japan and Papua New Guinea were within the medium range in the early 1970's. They continued to decrease and fall into the low range in the early 1990's. Some East Asian economies like Philippines, Taiwan, Thailand and China had their levels of processed food IIT within the low range in the early 1970's, but they have experienced quite significant increase over time, especially Taiwan and China. The levels of processed food IIT of the Philippines and Thailand change from low to medium range; The levels of processed food IIT of the USA are in medium range only in the early 1970's. After that, the levels of processed food IIT of USA are well above 50 per cent of its total trade and increase continuously. Australia and New Zealand's levels of IIT in processed food are low. But over last two decades, it has increased steadily for both economies: Australia changes from around 10 per cent to almost 30 per cent; New Zealand's increased from around 5 per cent to almost 20 per cent.

Though other economies didn't experience change to make their levels of processed food IIT shift from one catagory to another, there are some very interesting changes of the trend over the time period. The processed food IIT of Indonesia experienced dramatic ups and downs over the time. The first down was over 1970 to 1974, where the IIT on processed food dropped from 45 per cent in 1970 to almost 10 per cent in 1974. But from 1974 to 1988, its processed food IIT climbing back and even jumped to more than 60 per cent. But after that IIT in processed food went down again to nearly 35 per cent.

Though Korea's processed food IIT is within the medium range over the whole time period, it experienced similar continuous decrease to Japan. The development of Mexico's processed food IIT exhibit another pattern. Since 1970's it persistently had a downward trend until mid-80s, after that it rapidly returned to the original level. While both Hong Kong and Singapore's processed food IIT fluctuate over the time period, Hong Kong's remained low and Singapore's stayed high on average. Chile's processed food IIT fluctuates and keeps a downward trend for the first 20 years, then it goes upward.

The selection of a time period to cover in this study is based on the data availability but it represents a period of significant development in APEC.

Figure 1. IIT of processed food in APEC economies, (per cent) 1970-96

Industry characteristics

It was shown above that the total trade at the aggregated processed food industry level (ISIC 2-digit) exhibited a high intra-industry oriented pattern (where more than 50 per cent of its total trade is of an intra-industry type) for some of the economies in APEC and in some of the years during the 1970-96 period. If this is examined at more disaggregated industry levels, then analysis may be provided answers to questions like will the trade pattern be different to that at the more aggregated level and how IIT is concentrated at sub-industrial level of the processed food sector for each of the economies.

Table 1 Industry classifications and codes for processed food

Source: ‘Classification of Commodities by Industries Origin', Department of Economic and Social Affairs, United Nations, 1971.

The processed food IIT is measured again using the Grubel-Lloyd index but at 3 and 4-digit ISIC industry level and the information is summarised in Tables 2 and 3. In these tables, if an economy's processed food trade is intra-industry oriented for a particular year, then the value in that year is 1 otherwise the value will be 0. At the 3-digit industry level, as showed in Table 2 , industry 312 (Other food manufacturing) is overall more IIT oriented compare to other industries as there are more economies involved and number of years in the period with intra-industry oriented processed food trade.

There are exceptions from individual economies' perspective. Canada's processed food trade at the 3-digit industry level is intra-industry oriented among 311 (Food manufacturing), 312 (Other food manufacturing) and 314 (Tobacco manufactures) industries. Indonesia, Malaysia, New Zealand and Philippines have industries 312 (Other food manufacturing) and 314 (Tobacco manufactures) with intra-industry oriented trade almost equally; Taiwan and USA's processed food IIT are more concentrated on industries 311 (Food manufacturing) and 312 (Other food manufacturing). All of Singapore's processed foods at the 3-digit level are strongly intra-industry oriented.

At the 4-digit industry level, the information on APEC economies' processed food IIT is summarised in Table 3 . The overall situation here is that taking APEC as a whole, IIT is more concentrated in industries 3121 (Manufacture of food products not elsewhere classified), 3114 (Canning, preserving and processed of fish, crustaces and similar foods) and 3117 (Manufacture of bakery products) both in terms of economies involved and/or in number of years. For industry 3121 (Manufacture of food products not elsewhere classified), there are 17 out of 18 economies that have intra-industry oriented trade with some economies having more years than others. Also for that industry, most of its economies have more than half of the years in the period of concern exhibiting intra-industry oriented trade. In particular for the USA, New Zealand and Papua New Guinea, those economies have intra-industry oriented trade in industry 3121 (Manufacture of food products not elsewhere classified) for each single year in the whole period of study.

Another characteristic of intra-industry oriented trade at the 4-digit processed food level is that except for industries 3114 (Canning, preserving and processed of fish, crustaces and similar foods), 3117 (Manufacture of bakery products) and 3121 (Manufacture of food products not elsewhere classified), all other industries have intra-industry oriented trade more concentrated in several economies. For example, for industry 3112 (Manufacture of dairy products), only Canada, Korea, Singapore and USA exhibit their trade to be more intra-industry oriented.

Another interesting point here is that for certain industries, bilateral characteristics are more important than industrial characteristics in determining the presence of IIT in processed food. One example is industries 3118 (Sugar factories and refineries) and 3132 (Wine industries). The low intensity with intra-industry oriented trade does not mean the degree of processing for those industries is low. It may result from very different consumers' tastes or trade restrictions on those industries.

It is useful to compare the IIT in processed foods with that in manufacturing in general. Table 4 presents total IIT indices for manufacturing and processed foods in APEC economies in 1975, 1985 and 1995. For most of the economies the levels of IIT for processed foods are well below those for manufacturing as a whole. The growth of the levels of IIT in processed foods is faster than that for manufacturing industry as a whole. Canada, Korea, Singapore and the United States are still the economies with the highest levels of IIT for manufacturing initially and with steady growth from 1975 through to 1995, like processed foods. Thailand, the Philippines, Mexico, Malaysia and Indonesia experience the highest growth in both processed foods and manufacturing. Japan and Korea experienced significant growth in the level of IIT for manufacturing whilst experiencing declines for processed foods.

Table 2 - Intra-industry oriented processed food trade at 3-digit level

Source: Authors' calculation using ISIC data in International Economic DataBank, ANU.

Table 3 - Intra-industry oriented processed food trade at 4-digit level

Note Note: There are no processed food IIT data available for Taiwan in years 1970 and 71 for commodity 3122 and 3134; and the Korea in 1974 and Chile in 1971-75 for commodity 3134.

Source : Authors' calculation using ISIC 4-digit data in Internationa Economic DataBank, ANU.: Authors' calculation using ISIC 4-digit data in Internationa Economic DataBank, ANU.

Table 4 - Indices of total IIT for manufacturing and processed food in APEC economies, 1975, 1985 and 1995 (per cent)

Notes: 3 represents manufacturing industry; 31 represents processed food industry.

Source: Authors' calculation using ISIC Trade data from International Economic Databank, ANU.

Bilateral characteristics

The significant extent of total IIT in processed food in APEC economies suggests that the bilateral characteristics may be an issue. Therefore this part examines who are the major partners of processed food IIT for each APEC economies and how they change over time.

In order to simplify the analysis, this study divides the partners for processed IIT of APEC individual economies into two broad groups of economies namely EEC-12 and APEC at the global level, and ASEAN and NIES within the APEC region. At the same time, this study breaks the whole time period into three even periods, and looks at the average IIT in processed food in these three time periods.

At the global level, there are four economies in APEC namely Australia, Chile, Japan and Thailand that have a greater extent of processed food IIT in EEC-12 to that in APEC in the 1970 -78 time period. This means that in these economies' total processed food trade, the share of IIT is larger in trade with EEC-12 compared to that with APEC. But in time periods 1979-87 and 1988-96, all APEC individual economies have a larger share of IIT in total processed food trade with APEC compared to that with EEC-12. If we compare individual APEC economies' share in IIT in total processed food trade with APEC as a whole and with that with the world, most of the economies are or are becoming more intra-industry oriented in processed food trade with APEC as a whole. This phenomenon demonstrates that the intra-industry processed food trade for APEC economies is or is becoming more concentrated in the APEC region compared to the rest of the world.

Since there are larger shares of IIT in total processed food trade with APEC for individual APEC economies, it is of interest to further break down APEC economies' processed food IIT within APEC region. As shown in Table 5, the number of APEC individual economies' with larger shares of processed food IIT with ASEAN compared with NIES is 9 in 1970-78, 13 in1979-87 and 14 in1988-96. Clearly the data here demonstrates that APEC individual economies' processed food IIT within APEC region, are increasingly likely to be concentrated in ASEAN compared to that in NIES.

Similarly, we can compare the extent of processed food IIT of APEC economies to APEC as a whole and to ASEAN and NIES. As shown in Table 4 , for the majority of APEC economies' share processed food IIT in total processed food trade with ASEAN and/or NIES, this is lower than that with APEC as a whole. This indicates that the share of processed food IIT for APEC individual economies with the rest of the APEC economies must be bigger than that with APEC as whole.

Table 5 - Major partner of processed food IIT for APEC eocnomies, (per cent), 1970-78, 79-87 and 88-96

Notes: * Singapore is included in. ** Singapore is not included in.

Source: Authors calculation using ISIC data in International Economic DataBank, ANU.

5. Determinants of IIT in processed foods

So far this paper has examined the importance of IIT in processed food sector in APEC region. But it is important to look at the possible forces behind this development. These are examined from following aspects in conjunction with industry specific, country specific and policy specific determinants as reviewed in section 2.

Economic development

From previous empirical studies as reviewed in section 2, it can be appreciated that economic development levels are important determinants of IIT in general. This is believed to be the same for processed food IIT. This is because, from the production side, the higher the level of development amongst economies, the higher the capability to develop and produce highly differentiated processed foods. In general, processing an agricultural product provides opportunities for differentiate products. Higher levels of processing provide greater opportunities for differentiation. Economic development brought about technological advance at all stages of processed food production and therefore, highly differentiated processed foods.

From the consumption point of view, the higher the level of development, the higher differentiated demand that allows the exploitation of economies of scale in the production of a wide variety of individual commodities. Some demand analysis indicates that economic development (income) is a very important determinant in changed consumers' consumption patterns. Although, existing studies showed that consumers would spend less of their income on food as income increases. But within the group of food, consumers will spend more on processed food compare to unprocessed food. This is because, as income increases, the opportunity costs of time is increased. Processed food provides consumers with time-saving and convenience. It is believed that urbanisation has played an important role in the increased consumption of processed foods. The more consumers change their consumption patterns towards processed foods, the more likely diverse tastes will be generated and grounds be provided for demand driven IIT.

Further, the similar the level of economic development, the similar consumers' tastes will be and the more likely it is that markets for differentiated products will be created.

These economic development aspects certainly help to explain observed development of processed food IIT in APEC region. For example, as observed in Section 4, although Korea's processed food IIT is within the medium range over the whole time period, it experienced similar continuous decrease to Japan. This can be explained by the sharp decrease of its agriculture sector accompanied by a sharp increase of its manufacturing sector and therefore its trade pattern becoming increasingly dominated by inter-industry trade.

Also as demonstrated in the bilateral characteristics of processed food IIT, APEC individual economies' processed food IIT within the APEC region are increasingly likely to be concentrated in ASEAN compared to that in NIES. This empirical change may be because of the development strategies for NIES as Hong Kong is increasingly concentrating its economic development in the service sector, and Korea and Taiwan are increasingly concentrating on manufacturing industry.

Processed food IIT for APEC individual economies with the rest of APEC other than ASEAN and NIES is larger than that with APEC as whole. This may be because the rest of the APEC economies other than ASEAN and NIES contain several developed economies. For these economies, their economic developments are higher than ASEAN and NIES economies. Therefore, scale economies from the production side, and high income induced high demand for processed food from the demand side are possible forces behind the high share of processed food IIT with these economies for the APEC economies. The phenomenon demonstrated in Section 4, is that the IIT in processed food for APEC economies is or is becoming more concentrated in the APEC region compared to the rest of the world. This change may be contributed largely to diversified consumer demand as the result of the fast economic development of some of APEC economies during the later time period of this study.

By taking APEC as a whole, the most notable feature of GDP levels is their diversity. APEC comprises very large and small economies; developing and developed economies; slowly and rapidly growing economies. It is made up of economies vastly different in size and at different stages of development. This may explain why observed processed food IITs among and between APEC economies are at very different development stages.

Factor endowments

Similarities in factor endowments have been shown in previous theoretical studies to have a mixed role in explaining the level of bilateral IIT. A number of APEC economies appear to have similar factor endowments with respect to processed foods IIT, for example Australia and New Zealand, Canada and the United States, and Hong Kong and Singapore. Also from that perspective, Australia and New Zealand are natural resource based economies in this region hence their IITs in processed food sector are low as expected. Yet, there are some IIT in processed foods between economies with similar factor endowments and some between economies with different factor endowments. A large amount of APEC IIT in processed foods would not appear due solely to relative resource endowments as would appear the case for agricultural commodities. Processed foods are different from agricultural commodities. Technology advance, highly differentiated branded and convenient foods, and the market structure of many of the processed food industries, adds dimensions that are not captured by focusing solely on economies' resource endowments. The mixed outcome for APEC IIT in processed food contrasts to that for manufactured products, suggests the importance of quality differentiation in explaining IIT in processed foods.

Economic integration

It is observed that lower transaction costs among trading partners, be they from a relative decrease in prices for transport and communication services or from a removal of policy imposed-trade and investment barriers, tend to be accompanied by an increase in IIT. That is, economic integration positively influences IIT among integrated economies. This economic integration influences the development of processed food IIT within this region in several ways. Firstly, studies have shown that FDI and IITs are positively correlated. The increased processed food IIT within APEC may to some extent be explained by the increased FDI in the processed food sectors. At the same time, the side effects from FDI such as the spill-over effect and the demonstration effect from multinational firms (MNFs), may boost the development of domestic processed food sectors. Secondly, in the spirit of openness, people in APEC economies will share in the benefits of increased mobility. This increases the acceptability of different culture and hence diversified tastes.

APEC as a region has experienced significant economic integration during past decades. This is evidenced by a growing share of trade and investment among all member economies with other APEC members that is being driven by the private sector, which is seizing the opportunities created by the complementarity of the region's economies (PECC, 1995) . Some arrangements such as closer economic relations between several subgroups within APEC will assist in making these economies more important trading partners for processed foods over time. This economic integration not only reduced the barriers to trade and investment but also made these economies more open to each other and facilitated the merging of cultures and consumer demands. This may explain why observed processed IIT between economies that have closer economic relations tends to be higher than otherwise. For example, the sub-economic trade arrangements such as between Australia and New Zealand and in North America, may facilitate processed food IIT amongst these economies. These aspects explain why the share of processed food IIT for APEC individual economies with the rest of the APEC other than ASEAN and NIES is bigger than that with APEC as whole. Also this explains the development of Mexico's processed food IIT. Since 1970's Mexico persistently had a downward trend until the mid-80's, after which it rapidly retained to original level. The latter may be contributed to by the form of NAFTA which enables Mexico to access to a wider market. This economic integration may contribute to the explanation of the phenomenon as demonstrated in Section 4, the IIT in processed food for APEC economies is or is becoming more concentrated in the APEC region compared to the rest of the world.

Since IIT in processed foods is positively related to FDI and trade liberalisation, the existence of self-sufficiency and other impediments to trade and investment in this region, certainly restricted the development of more substantial IIT in processed foods.

Degree of processing

It was observed earlier that total and APEC bilateral IIT in processed foods have increased significantly but are still relatively low compared to manufacturing as a whole. There is some evidence that the degree of processing is related to the level of IIT. This could explain the just mentioned differences if the degree of processing of processed foods is less than that of manufacturing as a whole. Similarly, the increased level of IIT in processed foods over time could be explained by the degree of processing in processed foods increasing over time.

6. Conclusion

Existing studies on IIT are mainly concentrated on manufactured industry as a whole, and trade in agricultural products is often regarded as inter-industry in nature. This paper examined the processed food IIT in APEC economies. It is found in summary that there is an increasing trend in total IIT in the processed food sector for most economies in APEC. And the processed food IIT are mainly concentrated on industry 312 (Other food manufacturing) at the 3-digit ISIC industry level; industries 3121 (Manufacture of food products not elsewhere classified), 3114 (Canning, preserving and processing of fish, crustacea and similar foods) and 3117 (Manufacture of bakery products) at the 4-digit ISIC industry level. Further processed food IIT for individual APEC economies are increasingly likely to be concentrated in the APEC region as a whole compared to that with EEC-12 and the rest of the world. Within APEC, the individual APEC economies' processed food IIT is increasingly likely to be concentrated from NIES and ASEAN to the rest of APEC. But compared to manufactured industry, IIT on processed food are lower, possible due to the degree of processing being lower for processed food compared to other manufacturing industries. The overall result from the analysis of this paper is that a significant extent of processed food IIT exists in some of APEC economies for some of the years over 1970-96 period.

The results obtained from this study, not only suggest the importance of IIT on processed food in the APEC region, but also have significant policy implications. One important aspect of examining IIT, is its adjustment implications. As revealed in this study, IIT of processed food in APEC region as whole is still quite low on average. The trade pattern of processed food of Asia Pacific economies was and still is characterised by inter-industry trade between economies. Processed food sectors of Asia Pacific economies might be expected to suffer high adjustment costs in the face of trade liberalisation. However, facilitating the development of IIT in Asia Pacific economies will help ease the adjustment process in further economic integration; and industrial structural adjustment will be less painful in future trade liberalisation.

Then what will be the future perspective of processed food IIT in the region? As economic development is regarded as an important source for development of processed food IIT within this region, there is vast scope for the further development based on the initial low extent of IIT for some of the APEC economies. The rapid growth of many Asia Pacific economies was led, particularly in the early years, by a rapid expansion in exports. This was especially the case in East Asia in respect of goods whose production required intensive use of their abundant labour. China and the more populous Southeast Asian economies are still at this stage. Some more advanced economies have come to rely more heavily on expanding and maturing domestic demand. Even so, trade expansion is important in terms of improving industrial structures and incomes (PECC 1995) . Further expansion of exports and domestic demand will create new markets for differentiated products and consequently IIT.

Reconciling the diversity in development among APEC economies will also be important for the development of IIT in this region. As agreed at the Bogor summit by economic leaders, industrialised economies will provide opportunities for developing economies to increase their economic growth and the level of development. Developing economies are commited to aiming for high growth rates. If the development gap is narrowed in ways consistent with sustainable growth, equitable development and economic stability, the development of IIT is likely to be encouraged from the demand side.

The recent financial crisis in East Asian economies will undoubtedly hamper economic growth, not only in those economies experiencing the crisis. This might obscure the development of IIT in the region. Early recovery from the crisis is therefore important for sustained economic growth and smooth adjustment in regional trade.

Increased processed food IIT from economic integration results mainly from reductions in trade barriers and increased foreign direct investment among economies. The commitment to liberalising trade and investment in the region by 2010 for the forum's developed economies and by 2020 for its developing members promises good prospects for the growth of IIT in economies within this region.

Given the focus of this paper is mainly on the trend and pattern of processed food IIT in APEC region, the possible determinants for the development of such trade in this region is not discussed in a rigorous analytical manner. Further studies, for example, econometric studies to ascertain the significance of determinants of processed food IIT for both bilateral and industry characteristics, for which some possible reasons have been suggested by theory will be worthwhile. Furthermore, since the processed food IIT has increasingly become more important in the APEC region, the examination of welfare effects and policy implications would be a interesting direction for future study in this area. This aspect is particularly important in relation to liberalization on trade and investment among APEC region.

References:

Balassa , B. (1966) ‘Tariff reduction and trade in manufactures among industrial countries', American Economic Review , 56, pp. 466–73.

Balassa, B. (1986a) "The determinants of intra-industry specialization in United States trade", Oxford Economic Papers , 38, pp. 220-33.

Balassa, B. (1986b) "Intra-industry trade among exporters of manufactured goods", in D. Greenaway and P.K.M. Tharakan eds, Imperfect Competition and International Trade , Sussex:Wheat-sheaf Books.

Becuwe , S. and C. Mathieu (1992) ‘The determinants of IIT: the case of the automobile industry', Weltwirtschaftliches Archiv , 128(1), pp. 34–51.

Bergstrand , J.H. (1983) "Measurement and determinants of intra-industry international trade", in P.K.M. Tharakan ed., Intra-Industry Trade: Empirical and Methodological Aspects , Amsterdam: North Holland.

Brander , J.A. and P.R. Krugman (1983) ‘A reciprocal dumping model of international trade', Journal of International Economics , 13.

Caves , R.E. (1981) ‘IIT and market structure in the industrial countries', Oxford Economic Papers , 33, pp. 203–23.

Chen , T.Y. (1998) Trade liberalization, IIT and adjustment costs , PhD dissertation, Australian National University, Canberra.

Christodoulou , M. (1992) ‘IIT in agrofood sectors: the case of the EEC meat market', Applied Economics , 24, pp. 875-884.

Clark , D.P. (1993) ‘Recent evidence on determinants of IIT', Weltwirtschaftliches Archiv , 129(2), pp.332–44.

Culem , C. and L. Lundberg (1986) "The product pattern of intra-industry trade: stability among countries and over time", Weltwritschaftliches-Archiv, 122(1), pp.113-30.

Dixit , A.K. and J.E. Stiglitz (1977) ‘Monopolistic competiton and optimum product diversity', The American Economic Review , 67.

Falvey , R.E. (1981) ‘Commercial policy and IIT', Journal of International Economics , November 1981, pp. 495–511.

Finger , J.M. and D. De Rosa (1979) "Trade overlap, comparative advantage and protection", in H. Giersch ed., On the Economics of Intra-Industry Trade , Tubingen: Mohr.

Fukasaku , K. (1992) ‘Economic regional and IIT: Pacific-Asian perspectives', Technical Papers No 53, OECD Development Centre, OECD, Paris.

Greenaway , D. and C. Milner (1984) "A cross-section analysis of intra-industry trade in the UK", European Economic Review , 25, pp.319-44.

Greenaway, D. and C. Milner (1986) The Economics of IIT , Oxford: Basil Blackwell.

Greenaway, D. and C. Milner (1989) "Effective protection analysis and optimal trade policy with intra-industry specialisation and imperfect competition", in P.K.M. Tharakan and J. Kol eds, Intra-industry Trade: Theory, Evidence and Extension , New York: St. Martin's Press.

Greenaway, D., R. Hine and C. Milner (1994) ‘Country-specific factors and the pattern of horizontal and vertical IIT in the UK', Weltwirtschaftliches Archiv, vol. 130(1), pp. 77–100.

Grubel , H.G. and P.J. Lloyd (1975) IIT: The Theory and Measurement of International Trade in Differentiated Products , London: Macmillan.

Havrylyshyn , O. and E. Civan (1983) "Intra-industry trade and the stage of development: a regression analysis of industrial and developing countries", in P.K.M. Tharakan ed., Intra-Industry Trade: Empirical and Methodological Aspects , Amsterdam: North Holland.

Helpman , E. (1987) "Imperfect competition and international trade: evidence from fourteen industrial countries, Journal of Japanese and International Economics , 1, pp. 62-81.

Hirschberg, J.G., I.M. Sheldon and J.R. Dayton (1994) ‘An analysis of bilateral IIT in the food processing sector', Applied Economics , 26, pp.159-167.

Hirschberg , J.G. and J.R. Dayton (1996) ‘Detailed patterns of IIT in processed food', in I.M. Sheldon and P.C. Arrott eds, Industrial Organization and Trade in the Food Industries , Westview Press.

Krugman , P.R. (1979) ‘Increasing returns, monopolistic competition and international trade', Journal of International Economics , 9.

Lancaster , K.J. (1980) ‘IIT under perfect monopolistic competition', Journal of International Economics , 10.

Lee , Y.S. (1989). ‘A study of the determinants of IIT among the Pacific Basin countries', Weltwirtschaftliches Archiv , 125(2), pp. 346–58.

Loertcher , R. and F. Wolter (1980) ‘Determinants of IIT: among countries and across industries', Weltwirtschaftliches Archiv , 116.

Lowe , P. (1991) ‘Resource convergence and IIT', Research Discussion Paper , 9110, Economic Research Department, Reserve Bank of Australia.

Lundberg , L. (1982) "Intra-industry trade: the case of Sweden", Weltwirtschaftliches Archiv, 118, pp.302-16.

McCorriston , S. and I.M. Sheldon (1991) ‘IIT and specialization in processed agricultural products: the case of the US and the EC", Review of Agricultural Economics , Vol. 13, No. 2, pp. 173-184.

Pacific Economic Cooperation Council (1995) Survey of Impediments to Trade and Investment in the APEC Region .

Pagoulatos , E. and R. Sorensen (1975) "Two-way international trade: an econometric analysis"" Weltwirtschaftliches Archiv , 3, pp. 454-65.

Spence , M. (1976) "Product differentiation and welfare", American Economic Review , 66, pp. 407-14.

Tharakan , P.K.M. (1984) "Intra-industry trade between the industrial countries and the developing world", European Economic Review , 26, pp. 213-27.

Tharakan, P.K.M. (1985) "Empirical analysis of the commodity composition of trade", in D. Greenaway, ed., Current Issues in International Trade: Theory and Policy , London: Macmillan.

Tharakan, P.K. (1986) ‘The IIT of Benelux with the developing world', Weltwirtschaftliches Archiv , 122, pp. 131–49.

Toh , K. (1982) "A cross-section analysis of intra-industry trade in US manufacturing industries", Weltwirtschaftliches Archiv , 118, pp. 281-301.

Footnotes

1 - The major part of this paper was undertaken when the first author was visiting the APEC Study Centre at Columbia University. Thanks to Professor Hugh Patrick for providing the opportunity and the necessary resources provided by the Centre for this work to be undertaken. The International Economic DataBank at ANU offered a lot of assistance in making the calculation of the IIT indices much easier. Original funding for this work was provided with ACIAR's involvement through AusAID's APEC Support Program.

2 - The processed food here is defined as ISIC 31 (manufactured food, beverages and tobacco) industry.

3 - There are, of course, many methodological questions which have been raised in the literature about calculation of IIT indexes. See Chen (1998) for detail.

4 - The selection of a time period to cover in this study is based on the data availability but it represents a period of significant development in APEC.