Establishing why some countries trade considerably less than others is one of the most important items on the agenda of international economists. A deeper understanding of the factors that impede international trade is important because it would enable a better evaluation of their welfare costs. Those are suspected to be large – on their own, policy-related trade barriers may be worth more than 10% of national income (Anderson and van Wincoop 2002).
In that context, the case of the EU is particularly appealing, since trade integration is expected to be high among its member states due to several factors. First and foremost, the process of economic integration was launched in the 1960s with the creation of customs unions, abolishing internal tariffs and trade quotas. This process was revived by the Single European Act of 1986, which aimed at the completion of the Single European Market by the end of 1992. The Single Market was motivated by the observation that in the 1980s, many barriers to trade (especially non-tariff barriers such as technical barriers to trade) continued to impede intra-European trade. More recently, the introduction of the single European currency, the euro, aimed at accelerating the process of trade integration by eliminating exchange rate uncertainty and increasing transparency and competition across markets.
But assessing trade integration is not an easy task. The main difficulty lies in the measurement of overall trade impediments as “direct measures are remarkably sparse and inaccurate” (Anderson and van Wincoop, 2004, p.692). Data coverage is often limited to a few countries and years, and it can be hard to gather disaggregated trade barriers data at the industry or product level. As a result, researchers often need to rely on indirect proxies that are inferred from data on trade flows.
The problem is that the trade integration measures generally used in the literature - the most common being trade-to-output ratios or openness indices - are ad hoc proxies that lack a clear theoretical foundation. They are therefore inadequate to study international trade integration. For instance, trade-to-output ratios are scale-dependent and vary with country size, which clearly biases the measurement of trade integration. Smaller countries will be characterised by larger trade-to-output ratios regardless of their degree of trade integration. Trade-to-output ratios are also inappropriate to measure bilateral trade integration because they not only capture bilateral barriers but also multilateral trade barriers with other countries in the world. These are referred to as “multilateral resistance” barriers by Anderson and van Wincoop (2003). In their recent Vox column, Hiau Looi Kee, Alessandro Nicita and Marcelo Olarreaga also criticise the use of such proxies of trade impediments and stress the necessity of deriving alternative measures that have strong ties with trade theory.
A micro-founded measure of trade integration
To overcome the shortcomings inherent to crude measures of trade integration such as trade-to-output ratios, we propose in a recent paper (Chen and Novy, 2008) an alternative measure of bilateral trade integration that can be applied to disaggregated data at the level of industries or products. We derive this measure by modelling trade flows at the industry level in the gravity framework pioneered by Anderson and van Wincoop (2003, 2004). Anderson and van Wincoop (2003) show that trade flows are determined not only by bilateral trade costs between two countries but also by average trade barriers with other countries (“multilateral resistance”). We derive an analytical solution for multilateral resistance variables that vary across industries and over time (Novy, 2008). In turn, this enables us to derive a micro-founded measure of bilateral trade integration that has an in-built control for time-varying multilateral resistance and can therefore be applied to panel data. We are thus able to focus on bilateral trade integration because our micro-founded measure is not distorted by multilateral resistance effects. In contrast to trade-to-output ratios, our measure is also scale-independent.
Arguably, the Anderson and van Wincoop model is one of the most parsimonious trade models of recent years. It rests on the Armington assumption that countries produce differentiated goods and trade is driven by consumers’ love of variety, yielding a gravity equation. However, we show that isomorphic trade integration measures can be derived from other leading trade models – such as the Ricardian trade model by Jonathan Eaton and Samuel Kortum (2002), the trade model with heterogeneous firms by Thomas Chaney (2008), and the trade model with a linear demand structure by Marc Melitz and Gianmarco Ottaviano (2008) – thus giving our measure a broad theoretical underpinning.
Intra-European trade integration
Our measure of trade integration enables us to document and explain the variation of trade barriers across 166 manufacturing industries in 11 EU countries over the period 1999-2003. Consistent with the standard gravity literature, the variation of trade integration across country pairs can to a large extent be captured by typical gravity variables such as distance, adjacency, and language but also by policy-related variables such as membership in the Eurozone or participation in the Schengen Agreement.
Most importantly, we investigate the role of several sectoral characteristics in explaining trade integration across industries, with a particular emphasis on policy-related variables such as the extent of technical barriers to trade. Technical barriers to trade result from norms (regulations and standards) that affect the sale of goods in some markets by requiring specific product characteristics or production processes. Richard Baldwin (2000) stresses the importance of technical barriers to trade in shaping trade flows between countries and industries. He argues that in the case of Europe, such barriers have become more and more visible over time, especially since tariff barriers were completely eliminated by 1968. In addition, such barriers are a predominant concern in today's global trade negotiations, and for the WTO in particular as it precisely seeks to ensure that “technical regulations and standards, including packaging, marking and labelling requirements [...] do not create unnecessary obstacles to international trade.”1
We find that trade integration is indeed lower in countries and industries where technical barriers to trade are high. We also show that trade integration tends to be high for industries characterised by low transportation costs (captured by bilateral distance but also by the weight-to-value ratio of exports and the ratio between “cost, insurance, freight” and “free on board” trade values) and a high degree of transparency in public procurement. From a dynamic perspective, average trade integration has improved for most countries over the period 1999-2003, as well as individually for a large number of industries in our sample.
Table 1. Decomposing the variance of trade integration
|Industry fixed effects
|This table reports the contribution of various categories of factors in explaining the variation of trade integration across 166 manufacturing industries in 11 EU countries between 1999-2003. Source: Chen and Novy (2008).
In addition, we perform a variance decomposition to illustrate the quantitative contribution of each factor in explaining the variation in trade integration (see the above table). Interestingly, we find that geography and transportation costs explain 25% of the variance in trade integration, the most important factor being weight-to-value (17%), followed by bilateral distance (5%). Policy factors explain 7% of the variation in trade integration, which is far from negligible. Technical barriers to trade are the most important factor (5%), while public procurement, Schengen and the euro only play very minor roles.
The policy implications of these results are clear. While the barriers related to geography and transportation costs exist due to the very nature of spatial separation between markets, policy barriers such as technical barriers to trade are in principle removable, suggesting there is room left for policy action and that further gains are possible through the reduction of trade barriers in Europe.
Anderson J.E. and E. van Wincoop, 2002, Borders, Trade, and Welfare, Brookings Trade Forum. In: Collins and Rodrik, Eds., Washington: Brookings Institution, 207-244.
Anderson J.E. and E. van Wincoop, 2003, Gravity with Gravitas: A Solution to the Border Puzzle, American Economic Review 93(1), 170-192.
Anderson J.E. and E. van Wincoop, 2004, Trade Costs, Journal of Economic Literature 42(3), 691-751.
Baldwin R.E., 2000, Regulatory Protectionism, Developing Nations, and a Two-Tier World Trade System, Brookings Trade Forum. In: Collins and Rodrik, Eds., Washington: Brookings Institution, 237-293.
Chaney T., 2008, Distorted Gravity: The Intensive and Extensive Margins of International Trade, American Economic Review 98(4), 1707-1721.
Chen N. and D. Novy, 2008, International Trade Integration: A Disaggregated Approach. CEPR Discussion Paper 7103.
Eaton J. and S. Kortum, 2002, Technology, Geography and Trade, Econometrica 70(5), 1741-1779.
Melitz M.J. and G.I.P. Ottaviano, 2008, Market Size, Trade, and Productivity, Review of Economic Studies 75(1), 295-316.
Novy D., 2008, Gravity Redux: Measuring International Trade Costs with Panel Data, University of Warwick, mimeo.
1 Agreement on Technical Barriers to Trade (p.117). This Agreement, negotiated during the Uruguay Round, is an integral part of the WTO Agreement.