Multilateralizing 21st Century Regionalism

Date July 4, 2013
Speaker Richard E. BALDWIN(Professor of International Economics, The Graduate Institute, Geneva)
Moderator Willem THORBECKE(Senior Fellow, RIETI)


Richard E. BALDWIN's PhotoRichard E. BALDWIN

Today's presentation will cover several key topics in relation to multilateralizing regionalism. The first section will explain how 20th century and 21st century globalization are different and why this means that 20th century and 21st century trade are different. Following this, I will look at how and why 20th century and 21st century regional trade agreements (RTAs) are different. Finally, I come to the key point. Since 20th and 21st century RTAs are different, multilateralization of 20th and 21st century RTAs must be viewed differently. The main messages are that it is erroneous to consider multilateralizing 21st century regionalism (MR21) in the same light as multilateralizing 20th century regionalism (MR20). Having made this point, I admit immediately that I do not have all of the answers. More research is needed—specifically with regard to the legal content of the various 21st century RTAs around the world and the economic impact of the various provisions in them.

Traditional view of globalization

The standard way of understanding globalization at present is that a slow transition from a state of "no trade"—or autarky in economics jargon—to one of completely free trade. There are two key points in this traditional conceptualization of globalization. First, it is mostly about trade costs. Second, globalization can be viewed as a single process that has—with some important setbacks—been shifting the world from near-autarky before the Industrial Revolution to something approaching free trade. I assert that globalization is actually two distinct processes. The traditional view—with its focus on trade costs—was right up until the late 1980s or early 1990s. After that time, however, globalization should be considered as driven in part by a new process—although the traditional trade-cost process is still in effect.

Globalization as three cascading constraints

To explain this, it is useful to conceptualize globalization as governed by three constraints—not one. Before the Industrial Revolution, production and consumption were spatially bundled due to three types of costs: the cost of moving goods, ideas, and people. In the pre-globalized world, all three costs were very high, thus almost everything that was consumed in each village was made in the same village. This meant that the world was effectively broken down into village economies, each of which consumed almost all of its own production. The traditional view of globalization considers only the reduction in the cost of moving goods. The other two constraints also matter, but before turning to that, it is worth reminding you of the traditional globalization perspective.

First unbundling: Relaxing the transportation constraint

The first real shock came with the Steam Revolution. This produced a radical drop in the cost of transporting goods due to the development of steam-powered ships and trains. The cost of moving ideas became easier with telegraph and telephone, yet both remained extremely expensive. The cost of moving people was also still high and time-consuming. In a paper I wrote in 2006, I refer to this period in time as the "first unbundling," that is to say, lower costs of moving goods allowed for the spatial unbundling of production and consumption. Once shipping goods over long distance became feasible, scale economies and comparative advantage reshaped the world's economic geography. Industry concentrated in some countries and products were sold all over the world.

The separation of production and consumption, however, did not make the world flat in economic geography terms. Indeed, it was quite the opposite. Being able to sell to the world market favored firms that could achieve vast scale economies. These new industrial processes, however, were radically more complex than the old pre-globalization processes—think of the contract between a blacksmith's shop and a modern steel foundry. What this meant was that the second and third constraints became binding. In order to reduce communication and face-to-face costs, production was concentrated in factories and industrial districts. Thus, even as production dispersed globally, it clustered locally. This local clustering had very little to do with the cost of moving goods. It occurred due to the cost of moving ideas and people. In short, relaxing the transportation constraint led to a dispersion of activity internationally, but, as the communication and face-to-face constraints continued to bind, production was clustered at the sub-national level.

Second unbundling: Relaxing the communication constraint

The next big step came with a revolution in information and communications technology (ICT). This relaxed the communication constraint by radically lowering the cost of moving ideas. This, in turn, led to what I called the "second unbundling" in my 2006 paper. More directly, this meant that the factories themselves could be unbundled. Complex manufacturing processes could be broken down into several stages and organized from a distance. In other words, the ICT revolution allowed manufacturing to be offshored.

Critical differences between the first and second unbundling

The main difference between the first and second unbundling lies in the nature of the cross-border flows. In the first unbundling, it was mostly goods crossing borders. In the second unbundling, many of the flows of goods, ideas, people, training, and investment that used to happen only inside factories now occurred across borders.

Two critical changes occurred when the ICT revolution related the second constraint. First, managerial, marketing, and technical know-how that used to move only inside firms within a single nation was now moving internationally. This led to extremely rapid industrial growth in the handful of developing countries which received the offshored stages of production. This is why the impact of the second unbundling was so different than the first. The technology "transfers" that we always thought to be very important for development started to happen. But it was not a "transfer"—it was organized inside global value chains (GVCs), most of which were by multinational firms from the United States, Japan, and Germany. The recent sudden expansion of developing industry in such countries has come about due to the moving of ideas across borders, as opposed to how industry grew after the first unbundling. Trade was previously used for the purpose of selling goods. However, it is now more important as a means of producing goods. The second big change is that we have to start thinking of know-how as firm-specific, not nation-specific. Before the ICT revolution and the boom in offshoring manufacturing, it was convenient to think of a nation's technology as bundled together with its labor. Now that offshoring is easy, we see that a nation's technology is actually controlled by its firms, and these firms can now choose to combine their high technology with low-wage labor abroad.

Implications for trade

Now I'll turn to the implications of these changes for trade. In the 20th century, goods were bought and sold across borders, and trade was relatively simple. In the 21st century, trade flows became more complex, and supply chain linkages were developed. This led to a reliance on each stage of a supply chain working smoothly and efficiently, much like a widely-dispersed factory. As tangible and intangible assets are now exchanged all over the world, property rights have become a necessity for 21st century trade. 21st century trade is clearly significantly different from 20th century trade—not merely a continuation—as know-how and ideas are now a significant part of supply chains.

Another related yet significant factor involves the development of 20th century trade over time. In the late 1980s, production sharing and so-called "vertical specialization" trade within Europe, the United States, and Canada was common, but little of this occurred North-South. However, over the last 20 years, there has been a shift to North-South production sharing, including Asian nations. Yet, North-South production sharing hasn't been developing to the same degree across all sectors, and, in fact, growth in trade has mostly developed in the fields of electrical and mechanical machinery.

There are a number of factors to consider in how this understanding of 21st century trade could affect policy. Rather than considering simply one aspect of trade such as free trade agreements (FTAs), it is crucial that all trade disciplines are considered together as a package. Before starting on this line of reasoning, I should note that I believe that the case of China is quite different. A different set of rules apply to China since it can use its huge internal market as a lever to encourage offshoring even without providing legal assurances that come with deep RTAs.

Implications for trade policy

Due to its different nature—in particular its close association with production sharing—21st century trade requires different disciplines than did 20th century trade. These can be broken down into two main categories. The first is supply-chain disciplines. These are essential for connecting the geographically dispersed stages of production. They include not only liberal trade policy barriers but also excellent transportation services, guaranteed business mobility, and guaranteed communication services. Rather than thinking of supply-chain disciplines in terms of lowering tariffs in order to sell more, thinking in terms of ensuring that each part of the chain can work smoothly and efficiently together may be more beneficial.

The second category of 21st century trade disciplines are offshoring disciplines. These refer to protection of companies' tangible and intangible assets that they have placed abroad. This ranges from factors such as international investment to application protection of technology abroad and the local availability of business services. Looking at 21st century trade as being different from 20th century trade shows that these factors are also a part of modern trade which must be taken seriously as disciplines in their own right.

Trade policy as a package

Supply-chain and offshoring disciplines work best when packaged together. In relation to this, 21st century RTAs represent a convenient package. High-tech firms find such packages favorable and are willing to offshore some of the manufacturing jobs. In turn, developing nations that wish to join GVCs are generally willing to make certain concessions to be able to do so. Deep RTAs appear to be the solution in moving forward with 21st century multilateralizing of regionalism. The World Trade Organization (WTO) had issues with the Doha Development Agenda (DDA), and 21st century regionalism has developed consequently through the explosion of growth in bilateral investment treaties (BITs), deep RTAs, and unilateral liberalization in developing nations.

To introduce some basic statistics, the WTO database reveals that the number of RTAs began to increase significantly from around the mid-1990s. The number of related offshoring and supply-chain provisions in RTAs started to accelerate rapidly from around 1989, and the number of BITs started to rise rapidly since around 1988. In the mid-1990s, developing countries started unilaterally lowering tariffs. This information paints a picture of 21st century regionalism as one of deep agreements, BITs, and unilateral liberalization. Developing countries have made their economies more business-friendly, although this is difficult to document prior to 2005.

Concrete examples of 21st century disciplines

In order to discuss more specifically the meaning of 21st century regionalism, the agreements which have been signed by the United States, Japan, and the European Union (EU) should be mentioned. In looking at the frequency of each provision in the United States in the WTO database, preference is clearly given to those which are both mentioned in WTO 1.0 along with having deeper RTAs in place. Examples of this include tariff cuts, Trade Related Aspects of Intellectual Property Rights (TRIPS), Trade Related Investment Measures (TRIMS), the General Agreement on Trade in Services (GATS), and public procurement. In terms of provisions which are not included in the WTO such as visa and asylum, movement of capital, and labor market regulation, there is actually a percentage of U.S. agreements which include such provisions as well. In the United States, a high percentage of bilateral agreements include the movement of capital and are legally enforceable.

Japan's RTAs in the WTO database tells a slightly different story from those of the United States. Japan has a much higher percentage of visa and asylum-related RTAs, with a similar percentage of movement of capital and investment provisions. The United States and Japan in particular use RTAs to underpin value chains. The value chains of the EU tend to be within the EU itself, and it doesn't have to sign such deep agreements as do Japan and the United States. As a result, it may have a much less clear interest in underpinning GVCs through RTAs, as the EU represents the ultimate RTA in itself. A pattern can be seen by investigating RTAs conducted by all countries in the world. The categories of visa and asylum, movement of capital, intellectual property rights (IPRs), investment, and competition policy have the highest rate of RTAs in the world as a whole, and it is easy to imagine that they are the most strongly related to the current GVCs.

With regard to supply-chain disciplines in RTAs, examples include customs cooperation, beyond WTO GATS liberalization, FTA of industrial goods, and visa disciplines. Examples of offshoring disciplines include TRIMS, GATS, TRIPS, competition policy, IPRs, investment, movement of capital, and the approximation of laws. As GATS is mentioned as both a supply-chain discipline as well as an offshore discipline, it would be beneficial to research which aspect of GATS matches each discipline.

In summary, the traditional 20th century view of regionalism is that RTAs of FTAs are specifically related to tariff preferences, with a simple concept of political economy. However, 21st century regionalism is more closely related to disciplines which underpin second unbundling. Therefore, it could be argued that traditional analysis is insufficient and irrelevant for 21st century regionalism. It is now more appropriate to utilize regulation-economics as opposed to tax-economics in analyzing current 21st century regionalism. Furthermore, the basic political economy is different in 21st century regionalism than it was in 20th century regionalism. In 21st century regionalism, interest groups between countries are very diverse. Developing countries tend to aim to satisfy national industrialization strategies through FTAs, whereas in advanced economies, FTAs tend to satisfy the development of GVCs.

Multilaterizing 21st century regionalism

Having laid out the differences between the first and second unbundling and the associated types of trade and trade policy (what I called 20th century trade and 21st century trade), I turn now to the mulilaterizing 21st century disciplines that are embedded in 21st century RTAs. Let us start, however, by recalling what multilateralization of 20th century RTAs involved.

Multilateralizing regionalism in the 20th century focused on extending tariff preferences, rules of origin, and rules of cumulation in order to reduce discrimination. The concept of multilateralization was to go from bilateral to multilateral agreements in gradual steps, eventually leading to worldwide free trade. 21st century RTAs still include a similar concept, but they also revolve around deeper disciplines that support global supply chains.

A key distinction between shallow and deep RTA provisions is that the latter is non-discriminatory, or at least much less obviously discriminatory. And this is by their very nature—not by policy design. At the heart of this lower degree of discrimination is the difficult of ascertaining the nationality of companies, services, and intellectual property in today's interconnected world.

Nowadays, it is difficult to pinpoint specifically where the value-added originates or the nationality of companies. Of course, we all know that Sony Corporation is a Japanese company, but when it comes to the North American Free Trade Agreement (NAFTA), Sony USA is largely treated as a U.S. company as far as Mexico and Canada are concerned. This also applied to capital as banks tend to operate on a multilateral scale. As a consequence, rules of origin on deep provisions are "leaky"—it is difficult to write an RTA provision on, for example, rights of establishment that limits the benefit to companies from the signing countries.

As a consequence, RTA provisions are fundamentally non-discriminatory. When two countries sign a FTA, it is not necessarily discriminatory toward other countries. In fact, such an agreement could even benefit other countries. As an example, the adoption of the euro was originally expected to lead to an increase in trade within Europe but a reduction in trade outside of the Eurozone. However, the introduction of the euro led to an increase in trade both inside and outside of the Eurozone as it became easier for the countries outside of the Eurozone to do business with nations which all use the same currency. Rather than being discriminatory, the adoption of the euro could more accurately be described as slanted multilateral liberalization. Many 21st century RTA provisions arguably have similar features and effects.

Another aspect of multilateralizing regionalism in the 21st century is that preference margins in the real world are actually very small. In looking at the tariff preferences of the 20 largest economies in the world, a large number of them have more than half of their trade as duty free. Furthermore, when high tariffs of greater than 10% exist, they tend to be excluded from any FTA in which the country is engaged. As a result, between high tariffs being excluded from FTAs, which display no preference, and multilateralism, where no preference exists, preference margins are clearly smaller now than they were previously.

Recent estimates of trade creation and trade diversion due to RTAs have provided some interesting empirical evidence. In assessing them in relation to various trade agreements around the world, it was found that a trend of positive trade diversion exists in the majority of agreements. Of the few trade agreements which reflected negative trade diversion, the figures were very low. Only the Common Market for Eastern and Southern Africa (COMESA) and the Caribbean Community (CARICOM) agreements had high percentages for negative trade diversion, yet they also showed negative percentages for trade creation at the same time. These agreements could almost be thought of as anti-trade agreements as they reduce trade both inside and outside of the affected area.

The South African Development Community (SADC) represents an agreement which closely reflects the classic 20th century view on regionalism, where the numbers for trade creation are significantly positive and those for trade diversion are negative. This view would suggest that trade diversion is a problem issue, with trade creation being desirable. However, the state of trade diversion and trade creation in the SADC is generally not reflected in the majority of other trade agreements. This is due to tariff preferences not being high and trade diversion and trade creation not being driven by market access. This information could be used to argue against the necessity of multilateralizing. However, based on this evidence, it may be beneficial to cease viewing FTAs simply in terms of trade diversion and trade creation. It would be ideal if this evidence could be analyzed in more detail in terms of comparing trade diversion and trade creation in individual provisions of trade agreements such as competition policy, investment, and GATS.

At present, there is a general lack of discrimination technology for deep RTA provisions. Supply-chain disciplines assure the rapid movement of goods, ideas, people, and capital, and the goal of developing nations is to foster supply-chain industrialization. Discrimination in trade agreements is not useful ordinarily. In other words, multilateralism is often embedded in deep RTAs. This is also another factor which should be considered in looking at multilateralizing 21st century regionalism. Furthermore, the origin of discrimination in RTAs is also difficult to determine for services, capital, firms, and communication. Liberalization is also often embedded in host nation regulations whose justifications exclude discrimination. As an example, if the United States were to get Colombia to change its labor laws, this would have an effect not only on the former, but also on any other nation conducting business with the latter.

More research needs to be conducted in a variety of fields. Assessing the impact of various RTA provisions on trade in goods, services, and investment through using the WTO database on provisions, the World Input Output Database (WIOD) on parts and components, and the data on foreign affiliates' sales or employment to measure investment effects of various provisions would all be beneficial. Analysis of the WIOD could be used to see if trade in parts and components is affected by certain provisions, which in turn would provide for a better understanding of GVCs.

Another priority area would be to identify RTA provisions with negative spillovers for third nations. Perhaps there is a chance that certain investment provisions or capital provisions have discriminatory elements to them, and it would be very useful to pinpoint exactly where such provisions exist.

At this stage, it would also be good to start to search for network effects of RTAs. Most current research is on a bilateral level, although evidence shows that even bilateral agreements have an effect on a multilateral level. This, in turn, could lead to a better understanding of what could be gained from multilateralizing FTAs.

The goal of 20th century multilateralizing regionalism was to reduce discrimination. The goal of 21st century multilateralizing regionalism is to realize network externalities. At present, there are a few thousand BITs in existence. Japan itself has about 30 BITs. It would be simpler and more efficient to have only one BIT rather than several. Realizing such a network externality would lead to a more predictable environment, and it would also make it easier for developing countries to join such agreement. This entails a trade-off of determining an optimal level of harmonization and multilateralization. The downside of multilateralization is that diversity of situations and preferences between nations favor bilateral agreements instead, yet network externalities and sales economies favor multilateralization at higher-than-bilateral-level. It is essential to determine what elements should be multilateralized and to what extent. For example, if harmonization was low cost and low gain, it would be a non-issue. However, if it is low cost and high gain, it could lead to a unilateral adoption of regional rules. As an example, the Harmonized Commodity Description and Coding System (HS) of the past meant that every country had its own system of trading classifications. Over time, it was deemed sensible and beneficial for countries to share the same classification system, thus this was developed and adopted unilaterally. This system allows for differences between countries below the six-digit level and is an example of harmonization at a certain level, which is also low cost.

In cases of harmonization with high cost and low gain, such as the legal alcohol limit for driving, it would be sensible to maintain national rules.

In turn, high gains from having regional rules would lead to hub and spoke regionalism, and high gains from global rules would result in unilateral adoption of rules at low cost, or mega-regional or global multilateralization at high cost.

In conclusion, in order to move forward with multilateralizing regionalism, finding agreement on minimum principles as in GATS, such as investment disciplines, infrastructure service openness, and deeper IPR disciplines is important. Furthermore, legal research on how different are the deep provisions in existing RTAs, and developing investment rules along with a WTO trade facilitation package for customs cooperation would also aid this process.

Questions and Answers

Q1: Is there any hard empirical evidence showing that foreign direct investment and BITs are related?

Richard E. BALDWIN
It is surprisingly difficult to show that BITs are related to foreign direct investment. Part of the problem appears to be that foreign direct investment is measured in terms of capital flows.

Q2: In terms of multilateralization, how important do you feel the Trans-Pacific Partnership (TPP) is? Also, what do you feel should be the guiding principles of FTA negotiations in Japan?

Richard E. BALDWIN
Korea has signed some very deep trade agreements with the United States and the EU. The TPP could possibly look similar to those. If the TPP turns out to be multilateralized, it would surely be a success worldwide. Regarding FTA negotiations in Japan, the underlying commercial logic of GVCs is fundamentally asymmetric. With this in mind, the International Supply Chain Agreement (ISCA) seems to be a good institutional vehicle for multilateralizing FTAs in Japan and Asia.

Q3: You mentioned the FTA agreement between the United States and Korea as a possible template for the TPP. What is your view on the prospects of a potential Transatlantic Trade and Investment Partnership (TTIP) between the EU and United States, particularly in the area of rules and regulations?

Richard E. BALDWIN
Just to be clear, I actually don't think that the FTA between the United States and Korea should be used as a template. Rather, if I were to guess how I imagine a Japanese TPP would turn out, it would be similar to the U.S.-Korea FTA. Regarding the TTIP, it appears that it will not be easy to complete because a similar thing has been tried before. The EU and the United States have been discussing similar issues for about 60 years. There still hasn't been much of an evolution on hard issues in both the United States and Europe, unlike in Japan. I think that it is not likely that the TTIP will be completed if the TPP isn't. In order for the TTIP to be successful, it would surely need the full support of people like David Cameron and Angela Merkel. Yet, it is unclear as to how much an agreement on making the TTIP would lead to regulatory convergence. There is some hope in that the globalization of production has led to a certain degree of standardization in parts and components. However, the fact that the TTIP is referred to as an agreement, rather than a partnership, is in itself revealing about its nature.

*This summary was compiled by RIETI Editorial staff.