Whether it is in policy-making circles, high-level G-(pick your number) meetings or the academic literature, everyone talks and writes about the impact – good or bad – of more or less restrictive trade regimes on economic and social outcomes. In his recent Vox column, for instance, Paul Krugman asserts that US trade is significantly more open that it was just ten years ago and that this implies that US wages and income inequality are much more influenced by trade. Analysts of the differences observed in the growth experiences of China, India, Brazil and Russia usually give a central role to the differing trade restrictiveness.
In short, trade openness matters. This means that the measurement of trade restrictiveness also matters. Yet despite the concept’s importance, few participants in the policy debate provide a precise definition of what they mean by trade restrictiveness. When they go beyond vague concepts, their definition is unlikely to have strong ties with trade theory. A wide range of such trade restrictiveness indices are in use, but the basic problems can be understood by looking at one of the most common indicators of openness – the import-GDP ratio.
As a measure of trade restrictiveness, the import-GDP ratio is seriously flawed. To start with, this ratio is systematically higher for small nations, so one cannot use it to compare trade restrictiveness across countries. Moreover as manufacturing processes have become increasingly unbundled (geographically), an increasing share of world trade consists of parts and components. Since this trend is driven by a complex set of changes in communications costs and firm-level organisational changes, it cannot be linked uniquely to trade restrictiveness. What this means is that import-GDP ratios are not even useful for tracking the trade restrictiveness of a given country over time.
This is arguably a boring topic, but a very important one. Indicators of trade restrictiveness are crucial inputs to any study on the effects of trade policy on growth, poverty, or firm productivity, or when trying to understand the institutional and political determinants of trade protection. They will be needed by trade negotiators to understand what they can offer and get from their counterparts over the negotiating table, or to decide the allocation of aid across developing countries when trade reforms are part of multilateral or bilateral agencies’ conditionality. Without a well defined and theoretically sound measure of trade restrictiveness, any analysis of its impact on economic and social outcomes is likely to be inaccurate and lead to misinformed and costly policy recommendations.
Using theory to inform the measurement
The shortcomings of crude measures such as the import-GDP ratio have been known for a very long time. For example, the important conceptual work of James Anderson and Peter Neary has, over the last 15 years, lead the way to the development of theory-based measures of trade restrictiveness indications.1 There are really two major problems when talking about a nation’s overall trade restrictiveness – how to aggregate different forms of trade policies, and how to aggregate across different goods.
The first aggregation problem arises because trade policy can take many different forms: tariffs, quotas, non-automatic licensing, antidumping duties, technical regulations, monopolistic measures, subsidies, etc. How can one summarise in a single measure the trade restrictiveness of a 10 percent tariff, a 1000 tons quota, a complex non-automatic licensing procedure and a $1 million subsidy? Often the literature relies on outcome measures, e.g., import shares. The rationale is that import shares summarise the impact of all these trade policy instruments. The problem is that they also measure differences in tastes, macroeconomic shocks, and other factors which should not be attributed to trade policy. Another approach that is often followed is to simply rely on tariff data or collected customs duties, and hope that all other instruments are positively (and perfectly) correlated with tariffs. These are obviously unsatisfactory solutions.
A more adequate approach is to bring all types of trade policy instruments into a common metric. In our recent research on detailed data on trade barriers (Kee, Nicita and Olarreaga, 2006), we estimate ad-valorem equivalents of non-tariff barriers for each country at the tariff-line level.2 One result of our research is that after controlling for country and tariff line specific effects, tariffs and non tariff barriers tend to be negatively correlated. Thus the standard approach of assuming that measured barriers (tariffs) are correlated with un-measured barriers is just wrong.
The second aggregation problem arises because trade policy is set at the tariff line level and there are often more than 5000 tariff lines in a typical tariff schedule. How can one summarise all this information in one aggregate and economically meaningful measure? The two most common aggregation procedures involving taking a simple, un-weighted average of all tariffs, or taking the import-share weighted average of tariffs. Neither of these has a sound theoretical basis and both have serious drawbacks.
For example, using import-weights will typically underestimate the restrictiveness of the tariffs for Laffer-curve-like reasons. Goods that are subject to very high tariffs – say rice in Japan, or milk in Switzerland – have extremely low import shares exactly because of the high tariff. The alternative of using the un-weighted average has other problems. Since nations often have very low tariffs on goods that they do not import, the simple average tends to underestimate the restrictiveness of a nation’s tariff schedule.
James Anderson and Peter Neary tackle these problems using a theoretically sound aggregation procedure that answers very clear and precise questions regarding the trade distortions imposed by each country's trade policies on itself and on its trading partners. One important conclusion emerging from their work is that one single indicator cannot provide a measure of the trade distortions a country imposes on itself while simultaneously capturing the trade distortions imposed on its trading partners. When interested in the welfare distortions that the country imposes on itself, the aggregation procedure answers the following question: What is the uniform tariff that, if applied to imports instead of the current structure of protection, would leave home welfare at its current level? This corresponds to Anderson and Neary's Trade Restrictiveness Index (TRI).3 When interested in the extent to which trade distortions limit imports from the rest of the world, the aggregation procedure should answer the following question: What is the uniform tariff that, if imposed on home imports instead of the existing structure of protection, would leave aggregate imports at their current level? This second indicator is Anderson and Neary's (2003) MTRI, and it is here labelled Overall Trade Restrictiveness Index (OTRI) to account for differences in methodologies.4 Finally, if one is interested in the barriers faced by each country exporters when selling in the rest of the world, the relevant question is: What is the uniform tariff that, if imposed by all trading partners on exports of country C instead of their current structure of protection, would leave exports of country C at their current level?
A sample of our findings
Anderson and Neary's main contribution is conceptual. The contribution of our recent research is empirical – we apply the concepts developed by Anderson and Neary to a vast multi-country data set on tariffs and non-tariff barriers. We follow an econometric intensive approach within a simpler and empirically tractable trade model that allows us to capture the restrictiveness of trade policy at the most disaggregated level. The result is two databases. The first includes estimates of ad-valorem equivalents of NTBs and agricultural subsidies at the tariff line level. The second database provides estimates of trade restrictiveness indices as described above.5
It is impossible to summarise the databases. Here we focus on two findings – the relative importance of tariffs as a trade barrier, and the nature of rich nation’s barriers against the exports of developing nations.
Our results also show that tariffs are far from being a good measure of trade restrictiveness. Non-tariff barriers contribute a large share to countries’ trade restrictiveness. On average, they add an additional 87% to the restrictiveness imposed by tariffs. The contribution to the overall restrictiveness, however, is not uniform across nations. For instance, the relative importance of non-tariff barriers is more important for rich nations – a fact that can be seen in the Figure 1.
Figure 1: Ad-valorem equivalents (AVE) of non-tariff measures (NTM) and economic development
Note: NTM include core non-tariff barriers and agricultural domestic support.
While Figure 1 focuses on the barriers of importing nations, Figure 2 organises the data according to barriers faced by exporters based in developing nations. Figure 2 decomposes the trade restrictiveness of the OECD towards developing countries into its tariff, NTB and agricultural subsidy components for both manufacturing and agricultural goods. The message is quite clear. The restrictiveness of agricultural trade policy in high income OECD countries towards developing countries is much larger than the restrictiveness of their manufacturing trade policy. But agricultural subsidies are not the culprit, as their contribution to the overall level of agricultural trade restrictiveness is below 5%. The problems faced by developing countries’ agricultural exporters to the OECD are to be found in the OECD’s tariff and non-tariff barriers. Agricultural subsidies are a huge problem for European tax payers, not for developing country agricultural exporters.
Figure 2: Decomposition of the OECD trade restrictiveness vis à vis developing countries
Note: These were computed using bilateral OTRI indices that include tariff preferences
Interestingly, these results suggest that for the average developing country the negotiating priorities in the current round of multilateral trade negotiations may be misplaced. An important share of its negotiating capital is being spent in reducing agricultural subsidies. Some of their negotiating capital is spent on market access and none on non-tariff barriers. Our results suggest that the emphasis should be reversed.
Results show that while poor countries have more restrictive trade regimes, they also face higher barriers on their exports. This may be explained by reciprocity forces in multilateral and bilateral trade agreements – what you get in terms of market access depends on what you are ready to give up in terms of protection at home.
1 For a summary of Anderson and Neary’s work in this area, see James Anderson and Peter Neary, 2005, Measuring the Restrictiveness of Trade Policy, MIT Press, Boston.
2 Hiau Looi Kee, Alessandro Nicita and Marcelo Olarreaga (2006), Estimating trade restrictiveness indices, Discussion Paper No. 5576, CEPR, London, UK.
3 See James Anderson and Peter Neary, 1996), .A new approach to evaluating trade policy., Review of Economic Studies, 63(1), 107-125.
4 James Anderson and Peter Neary (2003), .The Mercantilist index of trade policy, International Economic Review 44(2), 627-649.
5 These databases are available at: www.worldbank.org/trade, then click on “Data&Statistics”, and then on “Trade restrictiveness indices”.