2  Economic Integration

During the last decade, several authors have explored how the degree of regional economic integration depends on whether different regions are in the same country or in different countries. This research has been undertaken in response to several different factors including Europe’s commitment to the Maastricht Treaty, increased academic interest in economic geography, large falls in transport and communication costs, and the availability of new data sets. The research has yet to reach definitive conclusions, but most of it suggests that national borders are considerably greater impediments to economic integration than had been previously imagined.

“Economic integration” does not have a single, straightforward definition, and two variants should be distinguished. The simplest case concerns commodities such as wheat which are sold in international auction markets and for which sales transactions are largely unaffected by the identity or characteristics of the producing firm. For such commodities, markets are integrated if people in different regions have access to the same goods at similar prices. A notable feature of these commodity markets is that they can be integrated without inter-regional trade being extensive, for a producer who usually sells into one region can substitute sales into another region very rapidly if necessary. The second case concerns producers who have limited ability to substitute sales between regions because sales are affected by such characteristics of the firm as its marketing strategy and its distribution chain. In this case, a second component needs to be added to the concept of spatial integration: in addition to prices being similar, products produced in one region must be routinely sold and used in other regions. Because most products and services are not sold anonymously through auctions, the second case is most important. For this reason, it is not generally sufficient to consider that regions are integrated because they have similar products at similar prices; rather, trade in these products is also necessary.

This section begins with a discussion of several issues pertaining to the integration of goods markets, before examining financial market integration. It ends with a short discussion of how economic integration can affect the location of output across space.

(a)  Goods Market Integration

(i)  Trade Volumes, Prices, and Exchange Rate Volatility.

Trade Volumes

Until recently it has not been possible to examine how economic trade between different regions is affected by political borders because data on trade volumes within countries have not been collected. However, Statistics Canada now collects data measuring trade flows between each of the Canadian provinces, and also between the Canadian provinces and the different U.S. states. This data was first analysed by McCallum (1995), and subsequently by Helliwell (1996, 1998). Using 1988 data, McCallum found inter-provincial trade between any two Canadian provinces was approximately twenty times as great as trade between Canadian provinces and U.S. states, once distance and economic size are taken into account.[2] [3] The analysis of provincial “exports” shows that Canadian firms do not find U.S. destinations to be close substitutes for Canadian destinations; similar analysis for “import” flows shows that U.S. products are not close substitutes for Canadian products in Canadian provinces.

Helliwell (1998) confirmed McCallum’s result for 1988, 1989 and 1990. He also extended the analysis in three directions. First, he analysed the impact of the North American Free Trade Agreement on Canada-U.S. trade by estimating how the home bias changed between 1990 and 1996. After the implementation of the agreement, the home bias fell steadily from a factor of twenty to a factor of twelve, where it remained from 1993 to 1996. While the decline shows the extent to which customs barriers deterred trade prior to 1990, the fact that it remains at such high levels shows that other aspects of the “border” - separate national institutions - are important.

Secondly, he analysed the border effects by industry. While the estimates were considerably less accurate, home bias was evident for all sectors, including the transport sector, although smaller for natural resource sectors than for others. The extent of the home bias in the transport sector is strange, for the 1965 Auto Pact between the U.S. and Canada instituted free trade in automobiles and automobile parts and all evidence suggests that trade in automobile parts between Michigan, Ontario and Quebec (the main producing regions) is now unaffected by the border. Further analysis showed that the home bias is only a factor in the other nine provinces, suggesting that the bias is a consequence of retail activity in the transport sector, perhaps because the distribution chains are organised along national lines.

Thirdly, Helliwell calculated the border effects by province. They were significantly higher in Eastern provinces than Western provinces; or as Helliwell (1998; p27) noted,

“ The ranking in border effects follows the ranking in terms of resource dependence, with the lowest border effects in the three western provinces with the largest concentration in the production and export of natural resource commodities.”

In conjunction with the analysis of sectors, Helliwell uses this data to suggest that home bias is more pronounced for regions where intra-industry trade, not inter-industry trade, is important.

This literature is still in its formative stages, and it is important not to generalise from the results of one country as to the relative size of intra-national and international trade. As Bayoumi (1997; pp104 – 108) notes, intra-provincial trade in Canada is only the same fraction of GDP as inter-European Community trade. Nonetheless, other evidence is broadly consistent with this result. Wei (1996) estimated a gravity model for OECD nations using a similar methodology as McCallum, and found that on average a country’s internal goods trade is ten times as large as its external trade, taking distance and economic size into account. A more sophisticated estimation procedure suggested that this ten-fold home-bias factor reduced to 2.5 once additional allowance was made for the fact that countries typically have much greater trade with countries which share a common land border and which speak the same language. Helliwell (1998) used similar methodology and found similar results, although his estimates of home bias taking language and land border effects into account are higher. He also explored the sensitivity of these estimates, and showed that they were sensitive to how authors estimated the average distance separating trading agents within countries. This raises some questions as to the reliability of Wei’s results as these distances are much less precisely known than the distances between major North American cities. Consequently, while the results of Wei, McCallum, and Helliwell point to the same phenomena - that there is a large home bias which means that trade within a country is several times as large as trade between countries - until there is more data it is premature to conclude exactly how large this bias is.

Prices

A substantial body of evidence shows there are large price wedges between the prices of the same good in different countries, and that these price wedges are greater than those within countries. This evidence has taken two forms. First, there is a vast empirical literature examining the behaviour of real exchange rates between countries which shows that real exchange rate movements tend to be large and persistent. A recent consensus suggests that the average half life of deviations from relative purchasing price parity between OECD countries is 4 - 5 years (see the reviews by Froot and Rogoff (1995) and Rogoff (1996)). The literature also shows that these slow rates of convergence are not just an aggregate phenomena but that prices for individuals goods and services vary widely between countries, and are slow to respond to each other.

Secondly, several recent papers analysing disaggregated price data in the United States and Canada show that price differences between these two countries tend to be larger than price differences within them (Engel (1993), Rogers and Jenkins (1995), Engel and Rogers (1995), Engel and Rogers (1996); and Parsley and Wei (1996)[4]). The paper by Engel and Rogers (1996) is a good example. They calculated “disaggregated real exchange rate indices” using price index data disaggregated into seven sub-indices for fourteen US and eight Canadian cities. They found that the quarterly volatility of these relative price indices was higher on average for a pairing of one Canadian and one US city than for city pairs which were either both Canadian or both from the United States. They also found that the quarterly volatility of the relative price indices increased with the distance between the cities. The distance between Canadian and United States cities could not explain this additional volatility, however; rather, the additional volatility caused by the border was equivalent to a distance of at least 1800 miles.[5] In line with their more general study (Engel and Rogers (1995)), they argue that the price volatility was in part due to exchange rate volatility, in part due to distance, and in part due to different distribution networks in different countries.

Exchange Rates

Since most countries have their own currency, it is natural to estimate the extent to which separate currency zones hamper economic integration. Two mechanisms are usually postulated: lower trade can occur because of the higher costs associated with the need to make currency transactions or because of the greater uncertainty associated with the possibility of future exchange rate changes. Despite business surveys which frequently suggest that exchange rate uncertainty has an adverse effect on trade and investment, it has proved difficult to quantify the importance of exchange rate uncertainty because there are few data sets available which allow a direct estimate of its effects. These difficulties have prevented the European Commission from reliably estimating the effect of exchange rate uncertainty on trade[6]. Nonetheless, they have convincingly estimated that the elimination of currency transactions costs would save approximately 0.4% of European GDP (Emerson et al 1992).

There is a small, intermittent literature examining the effect of exchange rate volatility on trade volumes, but to date it has been inconclusive. Part of the problem has been that different authors have used different measurements of exchange rate volatility, with a consensus as to the best method yet to emerge.[7] In particular, it has proved difficult to test whether long term exchange rate uncertainty has an adverse effect on trade.[8]

Two studies have addressed this issue in a manner that is directly relevant to this paper. First, de Grauwe (1988) related the change in trade growth rates between the 1960s and the 1970s in a cross section of countries to a variety of variables including exchange rate volatility.[9] He found that higher long term exchange rate volatility associated with flexible exchange rate regimes was a cause of lower trade growth in the latter period, adjusting for other factors. Secondly, Pozo (1992) found that trade volumes between the United Kingdom and the United States between 1900 and 1940 were lower in flexible exchange rate regimes than in fixed exchange rate regimes, and that these declines could be linked statistically to the greater exchange rate volatility in flexible exchange rate regimes. While both of these papers find that exchange rate uncertainty has a small negative effect on trade, this evidence is still far from conclusive. It may be the case that evidence of this kind is not capable of answering the qualitatively different question ‘what is the effect of freeing economic agents from the need to worry about the exchange rate at all?’ The adoption of the Euro will provide more evidence in due course. Until then, however, it seems unlikely that economists will be able to make convincing estimates of the extent to which currency unification enhances goods market integration.