Regulatory Protectionism, Developing Nations, and a Two-Tier World Trade System
Richard E. Baldwin
Graduate Institute of International Studies, Geneva
[Comments and Discussion]
[Figures]
Army generals, it is said, are always preparing to fight the last war they won. Is the World Trade Organization (WTO) in danger of following the same logic? The WTO-GATT (General Agreement on Tariffs and Trade) has all but won the war against tariffs. Tariffs on imports of OECD-manufactured goods--which account for more than two-thirds of world merchandise trade--are to average only 3.8 percent when the Uruguay Round cuts are fully implemented, with duty-free treatment applying to fully two-fifths of these trade flows. As in all wars, victory is never total. Tariffs on rich nations' food and clothing imports and poor nations' industrial imports are still high, as table 1 shows, but these items account for little of world trade. Moreover, the industrial tariffs of poor nations, at least, seem to be falling of their own accord. Trade is nonetheless far from unfettered.
"Regulatory protection" is but one name for the tens of thousands of cost-raising, behind-the-border measures that continue to substantially inhibit trade. Most of these measures are seemingly innocuous when viewed individually, but tangled together they are able to significantly fragment world markets. This point is not new. The last time "globalization" was in fashion--when it was called internationalization and interdependence--Robert Baldwin wrote, "[t]he lowering of tariffs has, in effect, been like draining a swamp. The lower water level has revealed all the snags and stumps of non-tariff barriers that still have to be cleared away." 1 The intervening thirty years have witnessed completion of the swamp draining, but the stumps have started to grow; three [End Page 237] [Begin Page 239] decades of ever tighter regulation of goods--most of which was adopted for purely domestic policy aims--have escalated regulatory protection.
Trade nevertheless is, overall, much freer than it was when GATT was signed. Somewhat paradoxically, this means that remaining barriers have a more important, not less important, impact on production. The magnification effect of globalization is the name for this. Falling tariff levels, teamed with lower transportation and communication costs, have cleared away many manmade and natural obstructions to trade, making the international playing field flatter than it has ever been. 2 This very flatness, however, means that even a slight tilt tends to have large effects on the location of production. In particular, seemingly minor differences in technical norms can have an outsized effect on production. While it is easy to exaggerate the international footloose character of production, one thing is clear. As the field gets flatter, regulatory protection's influence on production will continue to grow, forcing its liberalization ever higher on the negotiators' agendas.
The main thesis of this paper has five steps: technical barriers to trade (TBTs) are important; liberalization of TBTs will continue; this liberalization will involve hegemonic harmonization or mutual recognition of rules and test; such liberalization will almost surely entail preferential arrangements among rich nations, creating in essence a two-tier system of market access with developing nations in the second tier. The final step is to conclude that the WTO should be modified to address the potentially discriminatory aspects of regional TBT liberalization initiatives. The WTO, in short, should be gearing up to "fight" the battle of frictional barrier liberalization.
TBT Liberalization Is Important
TBTs result from norms (regulations and standards) that control the sale of goods in a particular market by specifying required product characteristics or production processes. 3 There are two distinct aspects of this control: content of the norm and testing procedures necessary to demonstrate that a product complies with the norm. TBTs thus come in two basic forms, content-of-norm TBTs and testing TBTs. The mind-numbing minutiae and technical complexity behind most regulations and standards make it difficult to speak of [End Page 239] TBTs as a whole. Some examples may help. We start with a content-of-norm example.
All cars sold in Sweden must have wipers on the headlights. This policy may have started as a meritorious safety regulation (Swedes keep their lights on at all times for safety's sake, and in the old days Sweden had lots of dusty roads). Today it mainly serves to raise the relative price of imported cars. From the drawing board onward, Volvos and Saabs--and their factories--are designed with headlight wipers in mind. For other carmakers--Renault for example--the Swedish market is too small to really matter, so Renaults are not optimized for headlight wiper installation. Thus, while it is expensive to put headlight wipers on both Swedish and French cars, the regulation raises the relative cost of French cars in Sweden. This gives the Swedish carmakers an edge in Sweden. Indeed, the outcome may well be that Renault will only export to Sweden its luxury models for which headlight wipers are already an option, so the regulation may act as a zero import quota for low-end cars.
Testing can also lead to TBTs. A frequent source of U.S.-Japan trade tensions in the 1990s was asymmetric testing procedures. For example, Japanese manufacturers, who were trusted by Japanese regulators, could establish compliance with regulations by having a few units tested periodically while foreign firms had to have each import shipment tested. 4
Vertical and Horizontal TBTs
It is useful to classify content-of-norm TBTs into two subtypes, horizontal and vertical. Vertical-norm TBTs involve norms that can readily be characterized as being more or less stringent. For example, the United States adopted a law (the so-called Corporate Average Fuel Economy [CAFE] standard) requiring each car company's fleet to attain certain maximum levels of emissions when averaged over all company models sold in the United States. This norm would have forced European carmakers to invest in more emission abatement than their U.S. rivals since U.S. auto makers lower their average, in part, by selling small-engine cars while European car companies sell mainly high-end cars with big motors. Europe also has much stricter rules than the United States on the permissible level of growth hormones in beef. TBTs stemming from such "vertical norms" generate much emotion and media attention since they can be portrayed as protecting local consumers from low-quality imports. [End Page 240]
Horizontal norm differences, however, are probably more common. Many TBTs arise when a nation, or subnational government, adopts the specifications of the local firm's differentiated product as its norm. For example, France and Germany regulate the nature of permissible electric plugs (both have two round pegs, but the German pegs are wider). The two plugs are equally safe and equally expensive to produce, but they are different. This raises the French firm's cost of selling in Germany (since German firms do not have to modify their product to meet the regulations while French firms do). It also raises the cost of German firms selling in France, so TBTs based on horizontal norm differences tend to create reciprocal trade barriers.
It is easy to see why horizontal TBTs arise so often. Citizen concern and industrial efficiency demand product norms, and a typical rich nation will have tens of thousands of standards and regulations. Most are highly technical, and a large fraction covers intermediate inputs--products unknown to most voters. Owing to their technical complexity and political invisibility, product norms are often written, directly or indirectly, by domestic firms to which they apply. Quite naturally, these firms write the norms in a way that favors their varieties or at least disfavors foreign varieties. The result is a horizontal content-of-norm TBT.
Of course, the mere fact that standards and regulations inhibit trade and competition is by no means an argument for their removal. Good governance requires regulation to protect the health, safety, and well-being of citizens, animals, and plants as well as to facilitate market transactions. The main problem with TBT liberalization is that it is difficult to know whether a particular norm serves the public interests or protectionists' interest, and, indeed, both motives are often combined in a single TBT.
Modeling TBTs
As the examples show, product norms and testing procedures can distort trade when they raise foreign firms' costs relative to those of domestic firms. Of course, in certain industries, the impact of TBTs is radically more complex. In industries with network externalities, like mobile telephones, standards can be manipulated to throw up barriers against nonlocal firms. In industries with patent races, like pharmaceuticals, a regulation that merely delays the introduction of foreign goods can radically alter the market outcome in favor of home firms. In industries with learning curves, product standards that apply to only a fraction of the market--government or military purchases, for example--can have large effects on the market equilibrium. Modeling TBTs in such [End Page 241] "sexy" industries will certainly be the focus of much future work, but in this chapter I focus on mundane industries where TBTs act by raising the costs of foreign firms more than the costs of local firms.
The cost-raising aspects of TBTs can affect foreign firms' fixed costs of
selling to the regulated market as well as their marginal costs of doing
so. Figure 1 shows how a
typical TBT involves up-front, one-time costs--for example, of learning
about the regulation and bringing the product into conformity--as well
as ongoing costs, such as periodic testing, or higher marginal production
costs that stem from a low scale of production.
The Economics of Regulatory Protection
Surprisingly little work has been done on the economic effects of TBTs, even at an abstract level. Most trade economists working in the subject tend to assume that TBTs affect trade in the same way as frictional barriers (that is, tariff-like measures that generate no revenue or rents). Most of the simulation studies of the European Union's (EU) 1992 program, for example, [End Page 242] modeled the liberalization effects as lowering the real cost of intra-EU trade by about 2 percent. 5 The industrial organization literature has focused on the impact of standards on a narrow range of industries, namely, industries marked by network externalities such as mobile phones, consumer electronics, computers, and the like. As the survey by Carmen Matutes and Pierre Regibeau argues, the general presumption is that harmonization of standards within a single market can be manipulated by incumbents to raise barriers to entry. 6 In such industries, regional TBT liberalization--which usually involves some form of harmonization of standards--would be presumed to act as a trade barrier to outsiders. While these network industries tend to attract a lot of attention, they cover only a fraction of the trade affected by TBTs.
A Baseline Model
TBTs can affect fixed costs as well as variable costs, so even a minimalist model requires imperfect competition and increasing returns. With these elements, however, almost anything can happen theoretically. As Elhanan Helpman and Paul Krugman show, quotas can lower prices, export subsidies can improve welfare, and small countries can have positive optimal tariffs. Clearly then, one should not be surprised that, theoretically, TBTs have an ambiguous impact on trade flows and welfare. 7
To highlight the main sources of ambiguity, a baseline model is helpful. Specifically, consider an industry marked by Cournot oligopolistic competition with firms facing constant marginal costs and two types of fixed costs, a firm-specific fixed cost of setting up production (unrelated to TBTs) and a fixed cost of selling to each market (related to TBTs). To allow discrimination, we need three nations: home (H), partner (P), and rest of world (R), assumed ex ante identical to reduce complexity. The inverse demand function of a typical nation is pj = 1 - Qj, where pj and Qj are the price and total sales in market i (i = H, P, or R). The total quantity in market j is (Sniqi j), where qi j is sales of a typical i-based firm in market j; ni is the corresponding number of i-based firms. Marginal production costs are normalized to zero. The three markets are assumed to be segmented.
Here we focus on TBTs based on horizontal norm differences and model them as simply as possible; nonlocal firms face higher variable costs or fixed costs (or both) when selling to the local market. Specifically, firms located in [End Page 243] home (called H-based firms for short) face no additional costs when selling to the home market. Partner-based (P-based) firms face a specific cost 0 < t when selling to the H-market. Rest-of-world-based (R-based) firms in the home market face per unit cost, t* , that is initially identical to t. The home and partner nations are always identical. TBTs may also impose an additional fixed cost on nonlocal firms. To capture this simply, we assume that each market has its own norm, and complying with these costs F in each market. (F is the same in each market since we are addressing horizontal norm differences.)
The equilibrium is found by solving the nine segmented market first-order conditions for the nine levels of sales. The model's linearity allows us to find solutions for prices, consumption, welfare, the numbers of firms, trade flows, and so on. In particular, the equilibrium price in market H, pH, equals the sum of two terms. The first term, 1/(1 + Sni), reflects the level of overall competition. The second term, (SniciH)/(1 + Sni), reflects the average marginal cost of firms active in the market (ciH is the market-specific marginal cost, normalized to zero for H-based firms, to t for P-based firms and to t* for R-based firms). 8 The focus of my inquiry, however, is on the distortionary impact of TBTs, so I highlight the "trade bias," defined as the difference between home's imports from wholly P-based firms and home's imports from wholly R-based firms; this roughly corresponds to a preference dummy in a gravity model equation. When TBTs among all nations are symmetric (that is, there is no discriminatory dimension favoring P-based firms over R-based firms), the bias equals zero. Specifically, 9
Trade Bias = pH(n-n*)-
where n is the number of home and partner firms (always identical
in the exercises we consider) and n* is the number of rest-of-world
firms. In some cases,
[End Page 244]
I allow the number of firms to vary with TBTs. The equilibrium expressions
for n and n* are messy, but it is easy to show that the
equilibrium "n" falls as the fixed cost F rises, and it
rises as the R-based firms' fixed cost F* rises.
10
Likewise, the equilibrium n* falls as F* rises or F
falls.
TBT liberalization can take the form of lower fixed costs or lower
marginal costs of selling to a particular market, or both. Each of these
may be "open" or "exclusionary." The trade impact of these various forms
differs radically, as we shall see.
Open versus Exclusionary Mutual Recognition:
Liberalization of Fixed-Cost TBTs
Consider first the liberalization of fixed-cost TBTs; for simplicity's
sake, we neutralize variable-cost TBT discrimination by assuming t =
t*. Initially suppose that the same fixed-cost TBT applies on an most
favored nation (MFN) basis, so R-based firms, H-based
firms, and P-based firms have to pay a total of 2F to sell
in both H and P markets. The liberalization considered
is meant to resemble one aspect of the EU members' mutual recognition
of one another's norms and testing. After the liberalization, a
product that complies with either home's or partner's norms can be sold
freely in both markets. That is, H-based firms and P-based
firms have to pay only a total of F to access both the H
and P markets.
This mutual-recognition privilege may or may not extend to R-based
firms. On one hand, mutual recognition can be "open": that is,
R-based firms can also sell to both H and P markets
after certifying their product in either. In this case, the fixed-cost
liberalization benefits all firms equally; after liberalization,
the total fixed cost of selling to the three markets is reduced from
3F to 2F. This raises profits and attracts new entrants in
all nations. Given symmetry, n and n* rise equally so the
trade bias remains at zero (t = t* and n = n*). In short:
open, fixed-cost TBT liberalization is not discriminatory--even when it
is regional.
On the other hand, mutual recognition can be "exclusionary" in the sense
that it applies only to products made in H or in P. In
practice this exclusion is enforced by rules of origin. For example,
the EU-Swiss Bilateral Accords
[End Page 245]
encompass TBT liberalization, but only Swiss-made goods circulate freely
in the EU market after having been certified in Switzerland. In our model
this means that H and P firms pay only F to access
the combined H-P market, but R-based firms must
continue to pay 2F. This results in improved profitability for
H and P firms, which will lead to a rise in n. There
is also a drop in n* as the new H-based and P-based
firms crowd out R-based firms. This creates a positive trade bias
as can be seen by inspection of (1) combined with the facts that t = t*
and n > n*.
In short, regional liberalization of fixed-cost TBTs is not discriminatory
when it is "open," but it is discriminatory when combined with rules
of origin.
TBT Liberalization That Favors Rest-of-World Firms
Open, fixed-cost TBT liberalization can actually favor rest-of-world
firms. Owing to fixed market-entry costs, not all firms will sell to all
markets. Rather, firms will sell to markets where their cost advantage
makes sinking F worthwhile. For example, an East Asian manufacturer
of electronic components might not find it worthwhile to export to
Europe if each European nation had its own product norm (each involving a
compliance cost) but might find Europe very interesting with an EU-wide
norm. In the context of our model, suppose that home and partner are
geographically close, and this gives H- and P-based firms
a distance-related marginal cost advantage (compared with R-based
firms) in the H and P markets. If this advantage is large
enough, R-based firms may chose not to sell in either the home
or the partner markets, since the operating profit they would earn on
such sales would not cover the per market fixed cost F. An open,
fixed-cost liberalization, which lowered the fixed cost from F
in each market to F for both the H and P markets
could induce new R-based firms to start selling in the H
and P markets. In terms of the trade bias, we see that a rise
in n* (number of R-based firms selling in the H
and P markets) would reduce the bias.
To make the point more formally, we start with the MFN situation where
each firm must incur a fixed cost F to sell in each of the three
markets. To simplify, suppose that TBT-related marginal costs are zero (t
= t* = 0), but H and P firms have a natural advantage since
R-firms face a trade cost of T when selling to the H
or P markets, but H- and P-based firms
face zero trade costs when selling into each others' markets. From
the first-order conditions, a typical R-based firm's sale
to the H-markets equals pH - T, where
pH is the equilibrium H-market price, while
the sales of a typical P-based firm in H,
[End Page 246]
namely, pH, will be higher than that of an
R-based firm. Noting that with linear-demand operating profit equal
to the volume of sales is squared, we see that R-based firms may
not find it profitable to sell to either the H or the P
market when T is sufficiently large. Specifically, we can have
a situation where (pH - T)
2
< F, so R-based firms sell to neither the H nor
the P markets, but (pH)
2
> F, so H-based firms and P-based firms will
sell to both markets. Now, consider the impact of an open,
fixed-cost TBT equalization between the H and P
nations. Postliberalization, R-based firms will sell to both
H and P markets as long as the sum of operating profit
earned in the H and P markets exceeds F. With
symmetry, 2(pH - T)
2
> F is the condition, where pH is the
postliberalization price. The impact of this on the trade bias is
straightforward. The liberalization has no impact on P-based
firm sales to home (since the liberalization only affected F
and all P-based firms were already selling to home), but it
induces a whole new set of R-based firms to export to home. The
trade bias, which was positive before the liberalization, now falls.
Liberalization of Variable-Cost TBTs
Regional TBT liberalization can easily affect the variable cost of
complying with product and process norms. For some products, such
harmonization may create a preference for EU products. The EU could,
for example, adopt a common norm that is systematically more costly
for non-EU firms to comply with, owing to scale economies or specific
technological advantages in the EU. This sort of case may arise when,
for example, EU firms sell the majority of their output in the EU (and so
design their products and factories with the EU norm in mind). If scale
economies are important, the per unit cost of meeting the EU norm may
be lower for EU firms than for non-EU firms for whom the EU is a minor
market. Alternatively, adoption of a common norm and testing procedure
might result in scale economies that would lower compliance costs for
all firms.
In the formal analysis, the first type of harmonization--call
it prejudiced harmonization--shows up as a change from the initial
situation of t = t* to a situation where t* exceeds t. This sort of
regional liberalization has the usual "supply switching" effect of
inducing H and P consumers to switch away from R-made
products and toward H-made and P-made goods. In particular,
taking n = n*, for simplicity, the trade bias--which reduces
to n(t* - t) in this case--is positive. In the second type
of harmonization--call it neutral harmonization--t = t* both before and
after liberalization, but the level of costs falls. Plainly, this will
have no effect on the trade bias.
[End Page 247]
Trade-Reducing Regional TBT Liberalization
TBT liberalization may affect trade bias by altering firms' ability to
price discriminate internationally, that is, by switching the nature
of competition from segmented markets to integrated markets. As the
literature on trade policy and imperfect competition has made clear,
the impact of protection can be radically different with and without
international price discrimination.
11
Only one of such implications is exposed here; see Richard Baldwin
and Anthony Venables for more on the economics of preferential
liberalization.
12
Markets are segmented when firms can assume that their customers will not
resell across markets. Markets are integrated when firms treat the two
markets as one. As it turns out, even when the two markets are symmetric,
that is, equilibrium prices are identical, market segmentation matters. In
particular, the volume of trade is higher under market segmentation
due to "reciprocal dumping" trade. The relevant link with TBTs is that
product regulations may help firms to segment markets. Consequently,
the imposition or liberalization of TBTs may have important effects that
are not captured by standard nontariff-barriers (NTB) analysis.
To make the point as cleanly as possible, consider two identical nations,
each with a single firm. The firms produce identical output--a consumer
good--at identical costs; trade in the good is subject to some natural
transport costs. Initially, the two nations impose a safety regulation
of a very special type. For safety's sake, the good and its instructions
must be sealed in tamper-resistant packaging. The instructions must
be written in the language of the country, and only the language of
the country (to avoid confusing consumers). Presume that instructions
would be included in any case and translation costs are minimal, so the
regulation has a negligible impact on marginal and fixed costs.
Using the analogy of tariff analysis, this regulation would seem to have
no trade effects. This is incorrect. The regulation prevents resale by
consumers or middlemen and thus allows each firm to treat the two markets
as segmented. When the two markets are segmented and yet there are some
natural trade costs, the two firms engage in reciprocal-dumping trade
à la Brander and Krugman.
13
In particular, firms charge a lower producer price for their exports
than they do for their local sales. If the regulation is changed to
require, for example, instructions to be printed in both languages,
the markets become integrated
[End Page 248]
in the sense that resale across markets is possible. Of course, such
arbitrage goes on until producer prices are equalized.
Interestingly, this market integration has a dramatic and
counter-intuitive trade effect. All trade halts since each firm supplies
only its local market. The reason is that due to arbitrage, firms treat
a sale in either market as if it were a sale in the local market and in
this case, there is no return to incurring the trade costs of selling
abroad. Nevertheless, prices fall since the potential competition due
to arbitrage is a more powerful source of competition than are actual
imports without arbitrage possibilities.
The economics profession does not yet have direct empirical measurements
of the trade-inhibiting effects of TBTs in general. There is hope that
this lacuna will soon be filled since there is a World Bank project on
this topic under way. At this point, indirect evidence is all that can
be marshaled.
Sapir's Evidence
André Sapir aims to gauge the impact of EU integration on
nonmember nations since 1960. To this end, he estimates year-by-year
gravity equations for the 1960-92 period on aggregate, bilateral
trade flows among a set of Western European nations.
14
The key interest lies in the sign and size of the dummies he introduces
for all the European preferential trade arrangements (PTA). Interpreting
the dummy coefficients, however, requires a bit of care. Sapir does not
include a dummy for trade flows among the original six members of the
EU; this level of integration is taken as the standard against which all
other PTAs are judged. Thus the signs of the included PTA dummies tell us
whether the particular bilateral trade flow is higher or lower than it
would have been if the two nations had joined the EU in 1958. Standard
controlling variables, such as gross domestic product (GDP) of sending
and receiving nation, distance, and common language, are included.
It is important to note that all tariffs and quotas on intra-EU trade and
all industrial-goods tariffs on intra-European Free Trade Association
(EFTA) trade had been eliminated by 1968. The signing in 1973 of the
EU-EFTA free trade agreements removed all tariffs on industrial trade
between EFTA and EU nations (industrial goods account for the vast
majority of EU-EFTA trade). As a result, all of Western Europe was part
of an industrial duty-free zone by
[End Page 249]
1973. The importance of this fact should be clear. Almost all of the
Western European integration since 1973 has involved the removal of
frictional trade barriers, with TBT liberalization being an important
component of this, especially the Single Market Programme (initiated
in 1987). It is also important to note that the EFTA nations did not
participate in the Single Market integration until the European Economic
Area agreement entered into force on January 1, 1994.
It would be interesting to directly measure the impact that this
frictional barrier liberalization had on intra-EU trade, but Sapir's
results do not permit this. We can, however, garner some evidence on
the trade volume impact of the liberalization by looking at changes in
the coefficient on the EFTA-EFTA and EC-EFTA dummies.
The rapid decline in the EFTA-EFTA coefficient, together with the
fact that EFTA was not pursuing any significant intra-EFTA integration
initiative, suggests that EU-EU trade was being promoted by the Single
Market's removal of frictional barriers. (Recall that the dummies indicate
market access relative to the market access available on trade among
the original six EU members).
A second piece of evidence comes from the evolution of the coefficients
on the EC-EFTA dummy. The jump up in market access due to the 1973 free
trade agreements (FTAs) is clear in the point estimates. Somewhat less
evident is the gradual decline of point estimates after the mid-1980s. The
modest change in the point estimates, however, hides the fact that
point estimates become significantly less than zero in 1985 and remain
significant right up to the end of the period (not shown in the figure).
Head and Mayer
Keith Head and Thierry Mayer find more conflicting evidence
using
industry-level data and the more sophisticated gravity model approach
inspired by Shang-Jin Wei. Head and Mayer use fairly disaggregated
industry-level data (three-digit NACE) to estimate the so-called
border effects, that is, the effect that an international border has
on purchasing patterns. In particular, the authors set out to answer
three questions.
16
How big were border effects in Europe before the nontariff
liberalization embodied in the Single European Act? How closely were the
industry-by-industry border effects associated with indirect measures
of the severity of trade barriers? Has the change in border effects
been associated with the removal of trade barriers under the Single
European Act?
In addressing the first question, they pool the ninety-eight industries
in their sample over the 1984-86 period to find a large border
effect. Depending upon the exact specification, they find that EU nations
buy between thirteen and twenty times more from local producers than
from producers based in other EU nations.
17
This base-line regression provides no direct evidence on the cause
of this border effect, but two facts are suggestive. First note that
there were no tariffs or quotas on intra-EU trade during this period,
so nontariff barriers
[End Page 251]
are a likely culprit. This suspicion is reinforced by the fact that the
regression results also show EU membership stimulate trade by something
like 50 percent. The point is that the only nonmember nations in their
sample--Spain and Portugal--have enjoyed duty-free industrial trade
with the EU since the 1970s, yet they did not participate in the EU's
factor-market and nontariff barrier liberalization. Consequently,
one possibility is that the impact of EU membership reflects the
trade-inhibiting effects of the EU's preferential liberalization of
nontariff barrier, and the border effect reflects the impact of the
remaining nontariff barriers.
The authors, however, fail to find any correlation between
industry-specific border effects and two measures of intra-EU NTBs. The
first measure is based on a 1980s survey of EU firms, the second on an
educated guess by Pierre Buiges.
18
This presents a puzzle. Industry-specific border effects are uncorrelated
with proxies for trade barriers, yet EU membership has an important
impact on trade flows. This suggests one of three interpretations. Deep
EU integration does matter, but the trade barrier proxies are faulty;
deep EU integration does matter, but it is due to something other than
TBTs and similar to nontariff barriers; or nontariff barriers do not
matter, and the Spanish and Portuguese trade flows are unusually low
for some reason that is unrelated to trade barriers. Unfortunately,
the authors' results do not allow us to distinguish among these.
Results pertaining to the third Head-Mayer question are also
interesting. Year-by-year regressions using their pooled data show a
marked decline in the border effect, suggesting that policy changes or
transport cost reductions have been important. However, the decline seems
to be much sharper before the Single European Act came into force. Again,
the interpretation of this finding is ambiguous. Either the Single
European Act was ineffective at liberalizing regulatory protection, or
it was effective at removing the barriers, but the barriers had little
impact on trade.
Moenius
One of the most direct attempts to measure the trade impact of TBTs is
Johannes Moenius.
19
The paper focuses on the trade impact of standards (voluntary norms)
rather than on TBTs more generally due to data limitations. In
particular, the author has constructed a unique data set on
industry-specific standards that includes information on whether a
specific standard is country
[End Page 252]
specific or common to the specific bilateral trade relationship. The
latter aspect is of the greatest interest. The point is that standards
shared bilaterally may preferentially lower marginal or fixed costs of
market and thus may have a trade-distorting effect. As argued earlier,
however, the expected sign of the distortion is not clear.
Moenius's data set is an enormous panel covering 471 industries in
twelve West European nations from 1980 to 1995. He finds that a shared
standard has a large trade-promoting effect among the nations sharing
the standard. For instance, in his central regression, which controls
for a wide range of other factors, he finds that a 1 percent increase
in the number of bilaterally shared standards results in a one-third
percent increase in bilateral trade volume. This is perhaps the only
clear empirical result in the literature on the discriminatory impact of
preferential TBT liberalization. The sharing of a common norm promotes
bilateral trade, so--by the usual logic of relative prices--we can
conclude that a shared norm inhibits trade with nations that do not
share the standard.
The impact of country-specific standards is even more intriguing. Moenius
finds that country-specific standards in importing countries promote
trade in manufacturing sectors but hinder trade in nonmanufacturing
sectors such as agriculture. Since the current comparative advantage of
many developing nations lies in nonmanufacturing sectors, this result
suggests that TBTs may create a bias against developing nations. Testing
this hypothesis directly, by including developing nation data, would be
an important contribution to the literature.
Finally, it is worth citing some important but highly indirect
evidence. Governments and firms engaged in international trade act as
if TBTs have important trade-inhibiting effects. They also act in a way
that suggests that they believe that preferential TBT liberalization
has an important discriminatory trade effect. As the discussion of
TBT liberalization below makes clear, the TBT liberalization in the
EU's Single European Act induced non-EU governments to negotiate trade
arrangements--notably EU membership, the European Economic Area agreement,
and the EU-Swiss Bilateral Accords--that redressed the discrimination thus
arising. Moreover, the rapid spread of 'mutual recognition agreements'
between the EU and its major trading partners (the United States,
Canada, Australia, and so on) shows that firms and governments take
the trade-inhibiting impact of TBTs seriously enough to invest time and
energy in lowering them.
Given this admittedly scant empirical evidence that TBTs are important,
we turn next to the real-world TBT liberalization schemes that have
been pursued
[End Page 253]
in the past. The lessons from historical attempts are crucial to
predicting the path of future TBT liberalization.
TBTs inhibit trade by raising costs faced by foreign firms more than
those faced by domestic firms. Liberalization requires a lowering of
the gap, and there are two main dimensions to this, content-of-norms
and conformity assessment. Liberalization of the first involves making
product norms more cosmopolitan and thus narrowing the cost advantage
of domestic firms. Liberalization of the second involves lowering the
excess costs that foreign firms face in demonstrating compliance of
their goods to accepted norms.
There are two ways forward along both dimensions, harmonization (that
is, convergence to a single norm or conformity assessment procedure) and
mutual recognition (that is, acceptance of foreign norms and conformity
procedures). Harmonization can be accomplished through negotiation, that
is, by narrowing the differences among various nations' norms, or through
the hegemonic route, that is, other nations adopt the norms of a hegemon.
20
We turn first to the European experience for one simple reason. The
introduction argued that the growing prominence of TBTs on the world
trade agenda is largely because of the near-total victory over the "easy"
barriers, tariffs, and quotas in OECD nations. Intra-European trade
reached this stage more than three decades ago. Subsequent liberalization
interest has focused on behind-the-border measures, with TBTs being one
of the chief attention-getters. More generally, since trade integration
among European nations is two or three decades ahead of WTO integration,
Europe's experience with thorny issues--such as labor standards and
trade, trade and competition policy, trade among rich and poor nations,
dispute settlement, and so on--provides a natural departure point for
the study of most "new" trade issues.
The EU recognized TBTs as a barrier to the Common Market in 1957
with Article 100 of the EEC Treaty, also known as the Treaty of Rome
(see box 1 for details). This requires approximation (Euro-speak for
harmonization) of national regulations for the proper "functioning of
the common market." As
[End Page 254]
in the WTO-GATT system, Europe's first liberalization efforts focused on
the "easy" barriers, tariffs, and quotas. With these eliminated by 1968,
liberalization attention turned to TBTs.
21
[End Page 255]
The Old Approach, 1969-84
The EU systematically took up the removal of technical barriers in 1969
with its General Program, which launched what came to be called the "old"
approach to TBT liberalization. The old approach was, in essence, based on
the idea that the EU would become as a unified economic area, functioning
much like a single national economy. Specifically, the General Program
was composed of four Council resolutions and a framework decision was
adopted on May 28, 1969. The resolutions concerned a detailed timetable
for a large number of directives on industrial products, the same for
foodstuffs, expression of the council's intention to institute mutual
recognition of conformity assessments, and a procedure for adapting
directives to technological advances. The framework decision prescribed
a standstill of member state measures concerning products covered by
the General Program, and a requirement that member states inform the
Commission of new provisions for products not covered; this also was
largely ignored.
22
In short, the old approach dealt with the content-of-standards issue
through negotiated harmonization. It strove for the adoption of a single
standard laid out in detailed technical regulations for single products
or groups of products implemented by unanimously agreed-on "directives"
of the European Council (the EU's main decisionmaking body).
23
Once adopted, such directives supplant national laws, regulations,
and standards. The conformity-assessment issue was tackled in a more
decentralized fashion. For products on which directives were adopted,
EU members recognized conformance assessments performed by designated
bodies in any EU nation. To prevent new TBTs, the so-called framework
decision provided for standstill and notification rules on new national
standards and regulations.
Evaluation of the Old Approach
"This approach to technical harmonization failed completely, " according
to Giandomenico Majone, although it deserves credit as the world's first
serious attempt to tackle TBTs.
24
The detailed timetable was roundly ignored, and harmonization of
standards and regulations proceeded much more slowly than the development
of new national TBTs. For example, 10 years were required to adopt a
directive on gas containers made of unalloyed steel, and the average
delay for the fifteen directives adopted en masse in 1984 was 9.5 years.
[End Page 256]
Furthermore, rules on "standstill" and notification were ignored since the
council qualified these rules as a "gentleman's agreement." In practice,
member states had full discretion to adopt new national regulations--as
long as they did not cross EU case law. The Commission received an
average of only eleven notifications annually between 1975 and 1982
while thousands of new regulations sprang up in member states.
25
The manifest failure of the old approach is confirmed by many such as
Michael Emerson (the chief Commission economist on the Cecchini Report)
who states that little progress was made since the exhaustively detailed
directives were "difficult to agree and quick to become obsolete."
26
The New Approach, 1985-Present
The EU's "new approach," formally launched in January 1985, is something
of a misnomer since it evolved over the 1960s and 1970s.
27
In particular, decisions of the European Commission and the EU Court that
are described in box 1 and box 2 introduced the main elements of the
approach by the end of the 1970s. The genius of the new approach is to
distinguish sharply between the goals ("essential safety requirements"
in Euro-ese) and means ("harmonized standards") of product and process
regulation. Two elements must be distinguished, "mutual recognition"
and "technical harmonization."
For the vast majority of European products, TBTs are eliminated by the
principle of "mutual recognition" of testing and content-of-norms. EU
case law and interpretations by the European Commission (see box 1)
create the presumption that the standards and regulations of all member
states are merely different means of implementing equivalent
regulatory goals. Thus, products made or sold in one member
state have "Single-Market access." That is, they can be sold freely and
without further testing in all member states (and in Norway, Iceland,
and Liechtenstein since the 1994 EEA agreement). Nonetheless, member
states are free to set standards and regulations on their territory,
but any resulting barriers to trade are allowed only if they can be
justified as being the least-trade-restrictive means of achieving a
legitimate purpose (for example,
[End Page 257]
health, safety, consumer protection, environmental protection). The EU
Court has taken a very narrow view of allowable justifications.
28
Of course, implicit is the mutual recognition of conformity assessments
performed by any body sanctioned by a member of the EEA.
For products where regulation-linked trade restrictions are justifiable,
the EU has turned to technical harmonization. It is here that the new
approach is truly new. With the new approach, the goals are completely
harmonized through
[End Page 258]
Council directives that supplant national law. Detailed technical
specifications and standards that show how to comply with these
requirements are left to private European standard-setting bodies,
such as the Comité Européen de Normalisation (CEN) and
Comité Européen de Normalisation Electrotechnique (CENELEC).
29
For example, the Lifts Directive (95/16/EC), which is
[End Page 259]
only fifteen pages long (not counting the annex), lays out essential
health and safety requirements relating to the design and construction
of lifts and safety components. Paragraph 1.2 in annex 1 states, "The
car must be designed and constructed to offer the space and strength
corresponding to the maximum number of persons and the rated load of the
lift set by the installer." The Council leaves the listing of technical
specifications that would fulfil this goal to CEN.
The effectiveness and political acceptability of this TBT liberalization
is bolstered by two principles. First, the essential safety requirements
must be transcribed into member states' laws. Second, while the resulting
harmonized standards must also be transposed at the national level--and
all conflicting national standards must be withdrawn--the standards are
voluntary. More specifically, goods made in accordance with harmonized
standards are presumed to conform to the essential requirements and are
thus free to circulate in the EEA. However, a firm may choose to deviate
from these, say because of an innovation. In this case, Single Market
access can be obtained by presentation of a certificate of conformity
with essential requirements issued by a designated body.
30
For products subject to technical harmonization (rather than mutual
recognition), the conformity-assessment aspect of TBT liberalization is
addressed by the 1989 "global approach."
31
Four stages (design, production, marketing, and sales-service) are
distinguished, and at each stage conformity may be assessed by one of
three bodies: the manufacturer, the local member state's regulatory
authority, or a designated third-party certification body.
32
Depending on the precise product in question, approval from the various
standards bodies must be obtained at various stages. Products for
which this procedure is clear and for which a harmonizing "essential
requirements" directive exists can earn the CE mark; this mark ensures
free access to all EEA markets.
The entire process of content-of-standards technical harmonization was
expedited by the 1986 Single European Act (that is, the 1992 program)
which replaced unanimity with qualified majority voting in the European
Council for all internal market decisions, including TBT issues. Majority
voting also now applies to CEN and CENELEC decisions on the adoption
of standards.
The new approach also made substantial progress on the prevention of new
TBTs with the 1983 Mutual Information Directive. Member states must notify
[End Page 260]
the Commission of new regulations. Either the Commission or other member
states can object to the regulation and request that it be modified to
reduce its trade-distorting effects. In either case, adoption of the
new regulation must be delayed and may be prevented altogether, if the
Council adopts a harmonization directive in the meantime. Additionally,
national standard-setting organizations are required to notify the
Commission and certain European standards bodies of proposed standards
and regulations before their adoption. A standing committee can determine
whether such standards would substantially interfere with trade and can
then refer the matter to a European standards organization or require
the national agency to collaborate with interested member states. When
the matter is referred to a European standards body, member states
cannot enact their own standard until a European standard is set, or
a one-year period elapses. Some exceptions, concerning urgent health
and safety considerations, are permitted. While this directive is not
radically different from the gentleman's agreement in the old approach,
the Commission actively pursues cases of noncompliance as an infringement
of European Community law (a Council directive has the force of a national
law in all member states and is enforceable through the EU Court). This
notification, comment, and objection procedure is credited with reducing
the emergence of new TBTs in Europe.
Given the sparse empirical evidence available, it is too early to
convincingly evaluate the new approach. In terms of output, however,
it has been a clear success. In 1975 there were 20 harmonized standards
in the EU; in 1999 there were nearly 5,500.
33
It is important to note that Europe's success with managed mutual
recognition is not a very good indicator of the route's success in the
wider world, as box 3 argues.
The Reaction of European "Outsiders": EEA and Switzerland
TBT liberalization in the EU had discriminatory effects on the West
European nations that were not EU members. In the late 1980s, EFTA firms
and governments had decided that they had to react. Several considered
applying for EU membership (Austria actually did), while others considered
bilateral negotiations. Jacques Delors forced the decision in January
1989 by proposing the European Economic Area agreement.
34
The agreement is very complex
[End Page 261]
and required the creation of extensive institutions, but for our purposes
it can be thought of as bringing the members of EFTA into
the new approach (see box 4).
While the effort to counter discrimination is easy to understand, the
outcome is not. Two aspects of the EEA are truly extraordinary. First,
the EEA is unbalanced in terms of the rights and obligations of EFTA
nations when it comes to future EU legislation. In essence, it forces
the EFTA nations to accept future council directives (the acquis
communautaire) concerning the Single
[End Page 262]
Market, without formal participation in the formation of these new laws.
35
Second, the EU insisted on a good deal of supranationality among
EFTA nations to simplify the task of keeping the Single Market
homogeneous. EFTA nations have resisted such supranational authority
since the end of World War II, so it is astounding that they said they
would accept it. As it turns out, virtually none of the EFTA governments
was willing to live with the EEA as it
[End Page 263]
was negotiated. By the end of negotiations, all but Iceland and
Liechtenstein (populations 300,000 and 30,000 respectively) had put in
EU membership applications, so for most EFTA nations the EEA was to be
merely a transitional arrangement. Swiss voters rejected the EEA in
December 1992, effectively freezing their EU application. Norwegian
voters rejected EU membership but implicitly accepted the EEA. The
EEA thus consists of the EU 15 on the one hand, and Norway, Iceland,
and Liechtenstein on the other.
In 1997 Switzerland initialed a bilateral accord with the EU after six
years of on-and-off talks; this should enter into force in 2001. This
bilateral deal essentially extends EEA rights and obligations to
Switzerland.
As far as the liberalization of TBTs is concerned, the lessons from this
episode are twofold. First it is not easy to deeply liberalize TBTs among
[End Page 264]
nations--even among rich industrialized nations that share similar
views on the subject of regulation (for example, protection of human,
animal, and plant life). Moreover, when a new nation gains access to
an existing arrangement, the likely result is a lopsided arrangement in
which the newcomer is forced to accept an immutable status quo. Second,
deep TBT liberalization requires institutions that permit surveillance,
enforcement, and adjudication.
Compared to the EU's active and ongoing efforts to reduce TBTs and
avoid new ones, TBT liberalization in the WTO is quite passive, aimed
primarily at avoiding the most obvious protectionist misuse of standards
and regulations.
[End Page 265]
[Begin Page 268]
These disciplines, which have evolved over the decades--see box 5 for details--are contained mainly in the so-called TBT agreement
that was part of the Uruguay Round. These apply to standards and
regulations governing both products and production processes. There
are five main disciplines, most favored nation for import restrictions,
national treatment once the good has entered, the sham principle, the
least-restrictive-means principle, and the transparency principle. MFN
and national treatment require no comment. The sham principle states
that standards and regulation should not be "a disguised restriction"
on international trade. The least-restrictive-means principle states
that regulations should accomplish their regulatory goal by means
that are the least restrictive to trade. As part of this, there is a
general preference for performance standards as opposed to standards
based on product characteristics. There is also a "requirement" to
use international standards, but this is largely nullified since the
requirement does not apply when international standards would be an
"ineffective or inappropriate" means of fulfilling legitimate regulatory
objectives. The transparency principle takes several forms. There is an
obligation on signatories to go through a notice-and-comment period when
introducing new measures that may affect trade. Signatories should also
maintain an information clearinghouse for standards-related inquiries.
Additionally, there are a number of hortatory elements such as the one
urging signatories to give "positive consideration" to unilaterally
recognizing foreign standards and regulations.
It is difficult to judge the liberalizing impact of the TBT agreement,
but few would claim that it has had a major impact. For one, compliance
with the TBT agreement has been quite spotty, outside of OECD nations.
36
TBTs have, however, been the subject of many disputes. Of the forty-eight
requests for consultation made to the WTO's Dispute Settlement Body
during the first two years of its existence, eleven concerned standards
or invoked the TBT agreement.
37
As of December 1999, the count had risen to twenty-five.
38
Another measure of the TBT agreement's success, or lack of it, is the
fact that much of the energy and momentum in TBT liberalization has
developed outside the WTO framework, in regional arrangements (EU,
the North American Free Trade Agreement [NAFTA], Asia Pacific Economic
Co-operation [APEC]), the U.S.-EU Mutual Recognition Agreement [MRA],
and so on) and on a sectoral
[End Page 268]
basis. According to John Wilson, there is "little evidence to indicate
that the WTO Committee on Technical Barrier to Trade has influenced
the debate."
39
There are three main international standard-setting organizations,
the International Organization for Standardization (ISO), the
International Electrotechnical Commission (IEC), and the International
Telecommunications Union (ITU). The ISO and IEC are private,
nongovernmental agencies in which each member country designates its own
representative. In the most important of the three, the ISO, most members
are the national standards bodies from member countries, and these may be
private or public bodies. Development of standards is a lengthy process
involving voluntary participation and a consensus decisionmaking rule.
40
Member nations are not obliged to transpose the adopted
standards. Moreover, compliance to an ISO standard does not bring with
it automatic market access.
According to Sherry Stephenson, European members have dominated the
international standardization process.
41
For example, European members hold two-thirds of the ISO secretariats
(roughly equivalent to a chair). The United States, by contrast,
participates in almost all of the ISO's work, but it held only 13 percent
of the secretariats. Stephenson claims that the predominance of European
members in this process derives from the fact that there are eighteen
European countries with long-established and sophisticated national
standards bodies well acquainted with international standardization.
42
Most of these coordinate their actions under the aegis of the EU.
While the ISO and brethren organizations have an important role to play
in TBT liberalization, their actions are essentially limited to areas
where
sector-by-sector negotiations can lead to Pareto improvements. In concrete
terms, the ISO has produced about 10,000 standards in its fifty-year
history. While this is a large number, it is small compared to the
100,000 or so standards that exist in the United States and EU today.
43
[End Page 269]
In June 1997, the EU and the United States signed a Mutual Recognition
Agreement (MRA) that deals with the trade-inhibiting aspects of
conformity assessment. The MRA commits each party to recognize the
results of product testing or certifications of both governments
in the six specified sectors. This should liberalize trade in these
sectors by reducing the cost of duplicative testing, inspection, and
certification. The covered sectors are telecommunications equipment,
electromagnetic compatibility, electrical safety, recreational craft,
pharmaceutical good manufacturing practice, and medical devices.
Importantly, the agreement does not mutually recognize standards and
regulations. Article 4 of the MRA states: "This Agreement shall not
be construed to entail mutual acceptance of standards or technical
regulations." It merely permits some EU standards to be assessed
by U.S.-based laboratories and some U.S. standards to be assessed by
EU-based laboratories. EU and U.S. industries, nevertheless, viewed this
as an important step to reducing trade barriers. Indeed, the initiative
came mainly for industry-to-industry talks in the TransAtlantic Business
Dialogue.
A number of other MRAs have also been signed. One of the first outside
Europe is the Joint Accreditation System--Australia and New Zealand
(JAS-ANZ). This is an open system that promotes mutual recognition
of testing in its member nations and beyond. As part of its mandate,
JAS-ANZ facilitated the negotiation of an MRA with the EU.
Having argued that TBTs are important and having examined the main liberalization efforts, the next step is to argue
that TBT liberalization will continue, and that it will almost surely
proceed along the lines of mutual recognition agreements among rich
nations augmented by some unilateral harmonization by poor nations to
the standards of rich nations.
The argument requires a bit of background. First, it is necessary to
clarify how TBTs are different from old-fashioned trade barriers like
tariffs and quotas. Second, a political economy framework is needed to
help us understand historical liberalization efforts and to predict the
path of future TBT liberalization.
[End Page 270]
It is often asserted that liberalization of behind-the-border
measures--product norms being a prime example--intrinsically engages
issues of sovereignty that did not arise in tariff-cutting talks.
44
This is misleading.
Tariffs, like many TBTs, are adopted in pursuit of governance goals--for
example, income distribution--that the community of nations views as
legitimate. However, through GATT-WTO, the same community rejects tariffs
(or more specifically the raising of bound rates) as a legitimate means
of attaining such goals, the simple reason being that tariffs are not
the least-restrictive means of fulfilling the goals. Developed nations
can achieve the same goal with income transfers, active labor market
policies, and the like.
45
The point is that tariffs are not prohibited because nations rejected
income distribution as a legitimate policy goal. Tariffs are prohibited
because equivalent but less-trade-restrictive measures clearly exist.
This brings us to the key distinction between tariffs and TBTs as far
as liberalization efforts are concerned. TBTs are harder to liberalize
because it is radically more difficult to objectively determine whether
less-trade-restrictive measures could accomplish the same legitimate
governance goals.
In short, the difference between regulatory protection and tariffs
boils down to issues of obscurity, not sovereignty. This simple but
important point has significant ramifications for the nature of future
liberalization.
Negotiated Harmonization Does Not Work
TBT liberalization can proceed in two ways, mutual recognition or
harmonization, with harmonization coming in two basic "flavors,"
negotiated harmonization and hegemonic harmonization. I argue
that negotiated harmonization does not work.
Begin with the historical evidence. The EU in the 1970s was a near-ideal
setting for such an approach. EU members had a "constitutional" obligation
to approximate product norms, made regular side payments by way of the EU
budget, had a supranational judicial system for dispute settlement, and
had similar income levels and regulatory goals. Even in this near-ideal
setting, negotiated
[End Page 271]
harmonization by unanimity proved unworkable in all but a handful of
cases. This suggests that the prospect of negotiated harmonization in
the wider world is nil for most products.
More can be learned by studying why the EU failed at harmonization
in the 1970s. As a matter of fact, the "failure" took the form of
interminable delays (in one famous example, it took eleven years to set
norms for mineral water).
46
Why the delay? International negotiations must strive for a balance
of commercial gains since each government must align a political
consensus behind the final liberalization package. Moreover, when
it comes to product norms, all governments must be convinced that
the synchronized norm meets their governance goals. Obscurity renders
both tasks enormously difficult. It is difficult and time consuming to
determine the commercial impact of each proposed norm. Further complexity
is added by the need to determine whether each proposed norm permits
an appropriate level of regulatory protection in each nation. Given
that a typical international negotiation involves many proposals, and
many modifications of each of these, leads one to the conclusion that
obscurity can quickly render such negotiations impractical. In fact,
obscurity-induced delays are important enough to make the negotiated
harmonization approach a nonstarter.
Mutual Recognition
This leaves hegemonic harmonization and mutual recognition as the
only routes forward, and, in fact, both routes are currently being
pursued. Hegemonic harmonization is the default option for many small
nations that are heavily dependent on a large trading partner such as the
EU. The Europe Economic Area agreement and the EU-Swiss Bilateral Accords,
for instance, commit nations to the EU's standards and regulations en
masse and without substantial modification.
Mutual recognition, or more precisely mutual recognition teamed with
some "new approach" harmonization in sensitive areas, is the standard
procedure inside the EU. It is also the liberalization route adopted
in the bilateral trans-Atlantic trade talks, APEC, and other regional
arrangements.
47
Interestingly, the most important MRA pursued to date--the U.S.-EU
MRA--is limited to mutual recognition of conformity assessments, not
of norms themselves.
[End Page 272]
Good governance demands product regulation, so the existence of standards
and regulations is no mystery. But why do these norms inhibit trade?
Obviously, protection does not result from random acts of nature,
and trade liberalization is not the outcome of heroic deeds by high-minded
civil servants and political leaders. Protection is endogenous. In a
price-like manner, proposed protection "clears" the political market by
equalizing the demand and supply of protection.
48
Regulatory protection, like most protection, arises from Mancur Olsen's
asymmetry. Asymmetric organization and information costs means that
industries organize more effectively than consumers. Norm-setting
agencies are thus "captured" by domestic firms in the regulated
industry. The obscurity factor greatly enhances the ease of capture
since most industrial standards and regulations are way off the radar
screens of voters and politicians. Obscurity also shapes the mode of
capture. In many nations and industries, domestic firms are the de
facto authors of their own norms. This is easy to understand.
When regulating a technical field, take elevators for example, obscurity
abounds. The government, which probably does not employ many full-time
experts on elevator manufacturing, asks the opinions of domestic
elevator-producing firms. With an eye to their foreign competitors, these
firms quite naturally push for norms and conformity practices that raise
the production cost of imported goods beyond that of locally produced
goods. For the domestic firms this is a straightforward cost-benefit
exercise. Domestic firms incur the costs of influencing the norm-setting
process when these norms create sufficient rents. This is not to call
the domestic industry greedy philistines. Most firms are convinced of
the superiority of their own products. Crafting standards to reflect
characteristics available only in local products may strike the firms
as a way of promoting the public interest. Given the obscurity factor,
it is difficult, maybe even impossible, to separate the protectionist
and public interest content of a particular norm. A good example is the
European paper industry (see box 6).
Taking it as a given that the level of protection is set in a political
market, trade liberalization poses a puzzle. Why should a nation find
it politically optimal to remove barriers that it previously found
politically optimal to impose? This is especially puzzling since the
revision-of-optimality has occurred on a
[End Page 273]
very regular basis over fifty years, and the revisions coincide with
multilateral trade negotiations.
The Juggernaut Model of Reciprocal Trade Liberalization
Elsewhere, I have proposed a puzzle solution that focuses on exporters
as the key antiprotection force and import competitors as the key
pro-protection force.
49
Announcement of a reciprocal trade talk--where better foreign market
access is bought with domestic liberalization--starts the juggernaut
rolling by multiplying the ranks of pro-trade forces. Exporters, in
search of foreign market access, cast off their normal indifference to
domestic protection and don the cloak of antiprotectionists. The same
goes on in other nations, so the political market for protection clears
at a lower level, implying some liberalization in all participating
nations. With the phasing in of the liberalization, export interests
get stronger as they expand output and employment. Import
[End Page 274]
competitors get weaker as they scale back or shut down. A few years down
the road, another round is launched, and the juggernaut rolls on. Barriers
are initially lower, but antitrade forces are weaker and pro-trade forces
stronger than they were at the conclusion of the last round, so barriers
again come down. The endpoint of this process is zero tariffs in the
markets of nations engaged in periodic reciprocal trade talks. Tariffs
remain high in countries where exporters get better market access "for
free" (for example, developing nations owing to MFN and special and
differential treatment) and in markets left off the bargaining table
(for example, food and clothes).
Application to TBT Liberalization
Consider next the application of this model to TBT liberalization. In
the political economy approach, changing the level of regulatory
protection requires a change in the forces that determined the political
equilibrium. In particular, liberalization has to offer something to
domestic firms that they find more attractive than the status quo. A
tried-and-true method is to engage nations in an exchange-of-market-access
negotiation. The political economy reasoning, however, differs somewhat
from the model explained above.
Most of the TBT liberalization seen so far has occurred among nations
and within industries where the exporting and import-competing firms are
one and the same--that is to say, in trading relationships dominated
by intraindustry trade. We must therefore modify the juggernaut model
to allow a different set of political actors. When it comes to TBT
liberalization, the key distinction will be between large efficient
firms and smaller less efficient firms. Two-way regulatory protection
results in fragmented markets where firms are dominant in their home
market while being marginal players in other markets. This market
fragmentation reduces competition, raises profit margins, and thus keeps
too many firms in business and keeps prices too high. Tearing down these
barriers defragments the markets, producing a pro-competitive effect
that puts pressure on profits, especially those of the inefficient (and
therefore small) firms. The end result is an industrial restructuring
that involves weaker firms merging, exiting, or getting bought up. In the
end, a more efficient industrial structure emerges with fewer, bigger,
more efficient firms competing more effectively with one another.
The key result is that two-way liberalization of TBTs systematically
favors large firms' profits. Intuitively, big firms gain from the
liberalization and from the exit of the small firms. Given this, the
juggernaut logic applies directly with large firms in the role of the
main pro-liberalizers and small firms in the
[End Page 275]
role of the main anti-liberalizers. Of course, a negotiating framework
based on reciprocity is again of the essence. Large and small firms
would be united in their opposition to unilateral liberalization.
With all this on the table, it is a simple matter to predict that large
multinational firms will continue to press for two-way liberalization
of TBTs, and this liberalization is most likely to take the form of
preferential arrangements among rich nations.
Most liberalization of regulatory protection in the past has been bi-
or plurilateral and based on the MRA approach. Future liberalization is
likely to continue being so for reasons just discussed. Yet the
MRA approach requires a high level of trust in a nation's governance
capacity. Since few developing nations are able to generate such trust,
the ongoing liberalization process will almost surely exclude them. A
two-tier system of market access may thus emerge, with developing
countries occupying the second tier. Note that this proposition is not
entirely based on abstract reasoning. The EU's Single Market program
liberalized TBTs on a preferential basis, which implied that non-EU
Western European nations had second-class access to EU markets. Outsiders,
such as the EFTA nations, had to scramble to redress the situation by
joining the EU or the EEA.
A couple of features make this two-tier market access more pernicious,
"escalation" and the "magnification effect." The magnification effect
of globalization emphasizes the tendency for a general lowering of
natural and manmade barriers to trade to make the remaining barriers
and discrimination more important, not less important.
"Escalation" concerns the link between product sophistication and
the cost of TBTs. Product regulation applies to all goods ranging
from apples to jet engines. Generally speaking, however, these norms
become more restrictive, complex, and costly as one moves up the
product-sophistication scale. What this means is that preferential
TBT liberalization has an "escalating" effect on the level of
discrimination facing third nations. The analogy with tariff escalation
should be clear. Even now, OECD nations tend to have higher tariffs on
industrial goods than on raw materials apart from food. In the 1970s,
when OECD tariffs were more important, this pattern of protection had
the unintentional
[End Page 276]
effect of distorting developing nations' post-tariff comparative
advantages away from industry.
What is wrong with a two-tier system?
--It is undisciplined. The WTO's TBT agreement contains no explicit
strictures on the form of MRAs or on their discriminatory effects on
third nations.
--It violates the nondiscriminatory and multilateral spirit of the WTO,
and this, in turn, tends to undermine the rule-based world trade system.
50
For both bilateral MRAs and hegemonic harmonization, the basic principle
is "might makes right."
--It harms the competitiveness of firms located in excluded nations--a
group that is likely to be dominated by developing nations.
For instance, bilateral foreign direct investment flows clearly show
that the EU's Single Market program led to investment diversion and
trade diversion.
51
By the same logic, the EU-U.S. MRAs may well divert investment and
trade from nations such as Mexico. At the moment, MRAs outside of Europe
concern only testing and so have a limited impact on competitiveness. If
the MRA process gets extended to norms as well as testing, the actual
content of standards will be affected. Public choice reasoning would
indicate that such an outcome is likely to lead to product norms that
disfavor goods made in excluded nations. A good European example of
this, described in box 6, involved an attempt by French and German
paper firms to impose a standard that would negate the resource-based
comparative advantage of Swedish and Finnish firms (by requiring that
cardboard boxes contain a high fraction of recycled paper and rags as
opposed to new wood). One could similarly imagine the United States
and the EU agreeing to a standard that would offset the comparative
advantage of developing country manufacturers.
Professors are generally better at identifying and framing policy
problems than they are at coming up with workable solutions. Yet even
admitting this absolute disadvantage, there do seem to be a couple of
basic proposals that would redress some of the problems.
[End Page 277]
Open Liberalization or Article 24 Disciplines
TBT liberalization can have radically different results when it is "open"
and when it is "exclusionary" (see above for the arguments). When it is
open--as is the case in the EU's Single Market--the TBT liberalization
may actually improve the market access of third nations.
52
For example, a Turkish-made skateboard can, in principle, be sold
in all fifteen EU markets after having been certified in any one. To
illustrate the importance of open liberalization, consider what the
impact would be if EU's mutual recognition principle were limited to
EU-made goods. Recalling that EU nations are allowed to have different
standards, third-nation discrimination in this counterfactual world could
be enormous. EU-made goods would have to comply with one norm, while
third-country goods would have to comply with fifteen different norms.
Generally, exclusion in TBT liberalization is enforced through rules of
origin. This then suggests the first proposal. Regional TBT liberalization
initiatives, including MRAs, should not in principle include rules of
origin. If they do, they are likely to be discriminatory, and thus they
should be subject to disciplines like those that fall under GATT Article
24. The point is simple. Rules of origin in a free trade agreement are
essential to prevent tariff evasion. However, rules of origin make
no sense in an MRA. A good, regardless of where it was made, either
meets the requisite norms, or it does not. Country of origin should
not be an issue. Given this, rules of origin in a TBT liberalization
arrangement create unnecessary discrimination. GATT allows for such
deviations from the MFN principle, but all such deviations are subject to
disciplines designed to ensure that the primary intent of the preferential
liberalization is liberalization rather than preference.
I would propose that these disciplines include something like the
"substantially all trade" dictate in Article 24. This would clearly be
difficult since all existing MRAs are limited to a relatively narrow
range of industrial goods. Recall that the substantially-all-trade rule
was adopted to reduce the likelihood that the main motive behind a
free trade agreement was to discriminate against third parties. A similar
discipline would be welcome for MRAs. For example, one might imagine
that the EU and Japan could craft an MRA on the norms and testing of
semiconductors that would substantially raise the costs faced by U.S. chip
makers in the EU and Japanese markets. The regulatory protection thus
provided might be very appealing to the body politic of both the EU and
Japan. Of course, the United States is big enough to counter
[End Page 278]
such moves. But what would happen if the United States and the EU signed
an MRA deal in grapes that had the effect of reducing the competitiveness
of Chilean grapes? Surely, this is exactly the sort of thing the WTO
was set up to discipline.
Additionally, something akin to the test on a custom union's common
external tariff should be applied to MRAs, although obscurity of TBTs
would make this difficult to apply. The point is that the signatories
of an MRA should be forced to examine the impact of their preferential
liberalization on third nations. There should be an obligation to pay
compensation to third parties if the MRA results in worsened market access
conditions. The details of such a stricture would require a great deal
of elaboration, but the "injury test" in antidumping and countervailing
duty cases might provide an example of how one might proceed.
It is worth noting that the issue of rules of origin in MRAs could
become important. Reportedly, the EU asked for rules of origin during
the negotiation of the EU-U.S. MRA but were rebuffed by the United States.
53
The stance of the United States, however, was based on the impracticality
of ascertaining "the" country of origin for sophisticated industrial
goods (nowadays, things are made nowhere in particular). The impact of
such rules on third nations was not, apparently, an issue for either
the EU or the United States.
GSP-like Treatment
A second proposal concerns policies that could actively seek to
offset any anti-developing-nation bias that arises from regional TBT
liberalization. The WTO TBT agreement already contains hortatory
statements about providing technical assistance to developing
nations. The WTO, or some other organization such as the World Bank
or the UN Conference on Trade and Development, should more explicitly
promote the TBT equivalent of the Generalized System of Preferences
(GSP). For instance, industrial nations might directly or indirectly
subsidize conformance assessment procedures for products made in
developing nations. Right now, the WTO's TBT agreement prevents members
from charging foreign firms certification fees that are too high. To
offset the cascading effects of discriminatory TBT liberalization, the
WTO might encourage a subsidization of fees charged to firms based in
developing nations.
[End Page 279]
This paper covered the economics of TBTs, evidence on their importance,
and the various initiatives made to liberalize them. It was found
that most liberalization of regulatory protection in the past has been
discriminatory and based on the MRA approach; future liberalization is
likely to continue to do so. The MRA approach requires a level of trust
in a nation's governance capacity that few developing nations are likely
to be able to provide. Consequently, a two-tier system of market access
is likely to emerge, with developing countries occupying the second tier.
I have pointed out that the discriminatory liberalization of regulatory
protection, unlike preferential tariff cutting, is largely undisciplined
despite violating the WTO's MFN spirit; this lack of discipline
may undermine the rules-based system as TBT liberalization becomes
increasingly important. More directly, MRAs--especially when they employ
rules of origin to exclude third nations from the liberalization--can
harm the competitiveness of firms located in excluded nations.
I also propose that the WTO be modified to discipline regional
TBT liberalization. Specifically, bilateral and plurilateral TBT
liberalization schemes should not include rules of origin. Rules of origin
exclude third nations from the liberalization without serving a legitimate
purpose (either a product meets the norms or it does not). If nations
do choose to include rules of origin, then the liberalization schemes
should be presumed to be discriminatory and should thus be subject to
Article 24-like disciplines. Finally, some sort of GSP-like policy
to offset the impact on developing nations should also be considered.
1.
Baldwin (1970, p. 2).
2.
See Baldwin and Martin (1999) for a detailed comparison of current and
pre-World War II levels of openness.
3.
According to common usage, regulations are mandatory, while standards
are voluntary.
4.
Sykes (1995, p. 25).5. Gasiorek, Smith, and Venables (1992, p. 45).
6.
Matutes and Regibeau (1996, pp. 183-209).
7.
Helpman and Krugman (1989).
8.
For a given market, say H, there are three types of firms
active; H-based firms, P-based firms and R-based
firms. Normalizing marginal production cost to zero (with linear
demand the intercept, call this a, and the marginal cost,
call this c, always enter together as a - c
in the first order conditions, and by choice of units we can
normalize a - c to unity). The first order conditions are:
pH = qHH for H-based firms,
pH = t + qPH for P-based
firms and pH = t* + qRH for
R-based firms. With ni (i = H,
P, R) firms in each nation, the sum of all the first
order conditions is (Sni)pH
= nH qHH + nP (t +
qPH ) + nR (t* + qRH
). Using the definition QH =
nH qHH + nP qPH
+ nRqRH and the demand function
pH = 1 - QH , the sum of first
order conditions becomes pH = 1 + nPt
+ nR t*, so pH = 1/
(1 + S(ni) + (nP t +
nR t*)/ (1 + Sni), which presents
that two terms discussed in the text.
9.
Using the first-order conditions pH
= t + qPH for P-based firms and
pH = t* + qRH for R-based
firms, we have nPqPH -
nRqRH = nP
(pH - t) - nR
(pH - t*). Rearranging this and employing the
convention that nP = n and nR
= nt yields the expression in the text.
10.
The simulations and full solution of the model can be found in the MAPLE
worksheet "brook.mws" available from the author upon request, and on
http://heiwww.unige.ch/~baldwin/papers.htm#policy (posted September 27,
2000).
11. Markusen and Venables (1988, pp. 299-316).
12.
Baldwin and Venables (1995).
13.
Brander and Krugman (1983, pp. 313-21).
14.
Sapir (1997).
15.
Most EFTA-EFTA trade is among adjacent nations, namely, among Nordic
nations or among Alpine nations.
16. Head and Mayer (1999); Wei (1996).
17.
Head and Mayer (1999, pp. 14, 17).
18.
Buiges and others (1990).
19.
Moenius (1999).
20.
Liberalization additionally involves surveillance, enforcement,
adjudication, and the introduction of new standards.
21. The year 1968
also saw the eradication of tariffs on intra-EFTA industrial trade by the
Stockholm Convention. The eradication of tariffs on EU-EFTA industrial
trade began with the 1973 EU-EFTA free trade agreements.
22. See Lauwaars
(1988, p. 152) for details.
23.
An exception, which was a precursor to the new approach, was the
"low-voltage" directive of 1973; see Lauwaars (1988, p. 156) for details.
24.
Majone (1994).
25.
See Pelkmans (1990, p. 109); Majone (1994, p. 166).
26.
Emerson and others (1988, p. 21).
27.
Commission communication of January 31, 1985, to the Council, "Technical
Harmonization and Standardisation, a New Approach." Elements of this
emerged earlier in 1973 with the Council's favorable experience in
framing the low-voltage directive, in 1974 with the Dassonville
case, and in 1979 with the Cassis de Dijon case. Other elements
emerged later, for example, the "global" approach to conformity assessment
in 1989. For a succinct summary, see www.NewApproach.org. [October 2000].
28.
See Sykes (1995) for legal analysis.
29. These harmonized standards are
prepared in accordance with the "General Guidelines" agreed between the
Commission and the mandate issued by the Commission after consultation
with the member states.
30.
Pelkmans (1990, p. 105).
31.
"A Global Approach to Certificate and Testing," European Commission,
July 24, 1989.
32.
Pelkmans (1990, p. 107).
33.
Moenius (1999, p. 1).
34.
The idea was first suggested at a meeting of EFTA and EU ministers in
Luxembourg in 1984. This produced the Luxembourg Declaration. EEA talks
began informally in 1989, continuing more formally in 1990 and 1991. It
was signed May 2, 1992, in Oporto, together with an agreement establishing
the EFTA Court of Justice and the EFTA Surveillance Authority.
35.
EFTA nations do participate in Euro-standards bodies, such as CEN and
CENELEC.
36.
Stephenson (1997, p. 42).
37.
Stephenson (1997, p. 17).
38.
Wilson (2000, pp. 16-17).
39.
Wilson (2000, p. 20).
40.
Stephenson (1997).
41.
Stephenson (1997).
42.
Stephenson (1997, p. 51).
43.
Stephenson (1997, p. 47).
44.
See Vogel (1997); Bagwell and Staiger (1998); Brittan (1998).
45.
Counterexamples strengthen the point. For developing nations that do not
have sophisticated systems for gathering and distributing tax revenue,
tariffs are more likely to be tolerated. Moreover, GATT allows import
prohibitions when these are the only practical means of pursing a
legitimate goal, such as reducing the consumption of illegal drugs.
46.
Stephenson (1997, p. 59 n. 31).
47.
See Stephenson (1997); Wilson (2000); and OECD (1998) for details.
48.
Mancur Olsen is the Freud of this literature. See Hillman (1989) for
the basic logic and Dixit, Grossman, and Helpman (1997) for recent
refinements.
49. Baldwin (1994, chap. 3).
50.
Note that the TBT agreement actually encourages MRAs, so the current
course of liberalization is by no means GATT illegal.
51.
Baldwin, Forslid, and Haaland (1996, pp. 635-59); Sapir (1997).
52.
The Cassis de Dijon mutual recognition principle applies to any
good that is made or marketed in an EU nation.
53. Wilson (2000,
p. 39).
Bagwell, Kyle, and Robert W. Staiger. 1998. "The Simple Economics of
Labor Standards and the GATT." Working Paper 6604. Cambridge, Mass.:
National Bureau of Economic Research (June).
------. 1999. "The Simple Economics of Labor Standards and the
GATT." In Social Dimensions of U.S. Trade Policies, edited by
Alan V. Deardoff and Robert M. Stern (University of Michigan Press).
Baldwin, Richard E. 1994. Towards An Integrated Europe. London:
Centre for Economic Policy Research.
Baldwin, Richard, Rikard Forslid, and Jan Haaland. 1996. "Investment
Creation and Diversion in Europe." World Economy 19 (6):
635-59.
Baldwin, Richard, and Philippe Martin. 1999. "Two Waves of Globalisation:
Superficial Similarities, Fundamental Differences." Working Paper
6904. Cambridge, Mass.: National Bureau of Economic Research (January).
Baldwin, Richard E., and Anthony J. Venables. 1995. "Regional Economic
Integration." In Handbook of International Economics, vol. 3,
edited by Gene Grossman and Kenneth Rogoff, 1597-1644. North Holland.
Baldwin, Robert E. 1970. Nontariff Distortions of International
Trade. Brookings.
Brander, James, and Paul Krugman. 1983. "A 'Reciprocal Dumping' Model
of International Trade." Journal of International Economics 15
(November): 313-21.
Brittan, Leon. 1998. Globalisation vs. Sovereignty? The European
Response. Cambridge University Press.
Buiges, Pierre, Fabienne Ilzkovitz, and J.-F. Lebrun. 1990. "The Impact
of the Internal Market by Industrial Sector: The Challenge for the
Member States." In European Economy: Social Europe,
special edition. The European Commission, Economic and Financial Affairs.
Dixit, Avinash, Gene M. Grossman, and Elhanan Helpman. 1997. "Common
Agency and Coordination: General Theory and Application to Government
Policy Making." Journal of Political Economy 104 (August):
752-69.
Emerson, Michael, and others. 1988. The Economics of 1992: The
E.C. Commission's Assessment of the Economic Effects of Completing the
Internal Market. Oxford University Press.
Gariorek, Michael, Alasdair Smith, and Anthony J. Venables. 1992. "Trade
and Welfare; A General Equilibrium Model." Discussion Paper 672. London:
Centre for Economic Policy Research (March).
Head, Keith, and Thierry Mayer. 1999. "Non-Europe: The Magnitude and
Causes of Market Fragmentation in the EU." University of British Columbia
(March).
Helliwell, John. 1998. How Much Do National Borders Matter?
Brookings.
Helpman, Elhanan, and Paul Krugman. 1989. Trade Policy and Market
Structure. MIT Press.
Hillman, Arye L. 1989. The Political Economy of Protection. Chur,
Switzerland: Harwood Academic Publishers.
Krueger, Anne O. 1999. "Trade Creation and Trade Diversion under NAFTA,"
Working Paper 7429. Cambridge, Mass.: National Bureau of Economic Research
(December).
Lauwaars, Richard H. 1988. "The 'Model Directive' on Technical
Harmonization." In 1992: One European Market? edited by
Roland Bieber, Renaud Dehousse, John Pinder, and Joseph Weiler,
pp. 151-73. Baden-Baden: Nomos Verlagsgesellschaft.
Majone, Giandomenico. 1994. "Comparing Strategies of Regulatory
Rapprochement." In Regulatory Cooperation for an Interdependent
World. Paris: OECD.
Markusen, James R., and Anthony J. Venables. 1988. "Trade Policy with
Increasing Returns and Imperfect Competition: Contradictory Results
from Competing Assumptions." Journal of International Economics
24 (May): 299-316.
Matutes, Carmen, and Pierre Regibeau. 1996. "A Selective Review of the
Economics of Standardisation, Entry Deterrence, Technological Progress,
and International Competition." European Journal of Political
Economy 12 (September):183-209.
Moenius, Johannes. 1999. "Information Versus Product
Adaptation: The Role of Standards in Trade." Department of
Economics, University of California at San Diego (November 24)
(http://weber.ucsd.edu/~jmoenius/hp/stantrad6.pdf [November 2000]).
Organization for Economic Cooperation and Development
(OECD). 1994. Regulatory Co-operation for an Interdependent
World. Paris.
------. 1998. Regulatory Reform in the Global Economy: Asian and
Latin American Perspectives. Paris.
Pelkmans, Jacques. 1990. "Regulation and the Single Market: An Economic
Perspective" in The Completion of the Single Market: Symposium
1989, edited by H. Siebert, pp. 91-117. Tübingen: Mohr.
Sapir, André. 1997. "Domino Effects in Western European Trade,
1960-1992." Discussion Paper 1576. London: Centre for Economic
Policy Research (February).
Stephenson, Sherry M. 1997. "Standards and Conformity Assessment
as Nontariff Barriers to Trade." Policy Research Working Paper
1826. Washington: World Bank, Development Research Group (September).
Sykes, Alan O. 1995. Product Standards for Internationally Integrated
Goods Markets. Brookings.
------. 1999. Regulatory Protectionism and the Law of International
Trade." University of Chicago Law Review 66 (Winter):
1-46.
Vogel, David. 1997. Barriers or Benefits? Regulation in
TransaltanticTrade. Brookings.
Wei, Shang-Jin. 1996. "Intranational versus International Trade:
How Stubborn Are Nations in Global Integration?" Working Paper
5531. Cambridge, Mass.: National Bureau of Economic Research (April).
Wilson, John S. 2000. "The Multilateral and Regional
Trade Policy Response to Technical Barriers." In World
Bank Research in Progress on Standards. Washington.
(http://www1.worldbank.org/wbiep/trade/papers/Hoekman-chapter-3-wb-paper.pdf [October 2000]).
Van Wincoop, Eric. 2000. "Borders and Trade." Federal Reserve
Bank of New York, International Research Function (April)
(www.ny.frb.org/rmaghome/economist/vanwincoop/trade1.pdf [November
2000]).
TBT Liberalization
Empirical Evidence
The annual point estimates are plotted in figure 2. According to the estimates,
up to 1975 EFTA-EFTA trade volumes were higher than the EU-EU flows,
controlling for the usual factors. This somewhat unexpected result
(the received wisdom is that the EU membership provides the tightest
integration in Europe) may be because of Sapir's omission of an
adjacency variable.
15
In
[End Page 250]
any case, the point estimates on the EFTA-EFTA dummies are never
significant (at the 95 percent confidence level) when they are positive
(not shown in the figure). The EFTA-EFTA dummy becomes solidly negative
in the early to mid-1980s but become statistically significantly less
than zero only in 1989.
Liberalization Experiences
EU Initiatives
TBT Liberalization in the GATT-WTO
The ISO
The EU-U.S. Mutual Recognition Agreement
Future Liberalization
Obscurity, not Sovereignty
Political Economy of TBT Liberalization
A Two-Tier World Trade System?
Some Proposals
Conclusions
Notes
References