10431.pdf
Media
- extracted text
-
TWN Trade & Development Series
The WTO Negotiations on Industrial
Tariffs: What is at Stake for
Developing Countries
Yilmaz Akyuz
Third World Network
24
The WTO Negotiations on Industrial Tariffs: What
is at Stake for Developing Countries
Yilmaz Akyuz
TWN
Third World Network
The WTO Negotiations on Industrial Tariffs: What is at Stake
for Developing Countries
is published by
Third World Network
121-S, Jalan Utama
10450 Penang, Malaysia.
© Third World Network, 2005
2
Printed by Jutaprint
Solok Sungei Pinang 3. Sg. Pinang
11600 Penang, Malaysia.
ISBN: 983-2729-39-4
1U *31
CONTENTS
1
Introduction
I
2
The Key Elements of the NAMA Framework
4
3
A Brief History of Industrial Protectionism: Good for
the Goose, but not for the Gander
8
4
Industrial Development and Tariffs
19
5
Assessing Costs and Benefits: Are the Gains Worth the Pains?
33
6
Reciprocity or Unequal Exchange?
45
7
Conclusions: The Way Forward
49
References
53
NOTE:
This paper was prepared for the Third World Network. I am grateful to Ha-Joon
Chang, Bhagirath Das, Martin Khor, Richard Kozul-Wright, Kamal Malhotra,
Jorg Mayer, Chakravarthi Raghavan and Irfan U1 Haque for helpful comments
and suggestions. They are not responsible for remaining errors.
1
Introduction
A key item on the agenda of the Doha Round of trade negotiations is liberaliza
tion of trade in industrial products or, in the terminology of the WTO, “nonagricultural market access” - NAMA. Despite its significance for industrial
ization and development, and the difficulties encountered in negotiations, this
issue has not attracted much public attention in large part because the recent
discussion has focussed primarily on agriculture.
The framework adopted for modalities for negotiations for NAMA, as con
tained in Annex B of the so-called July Package (WTO 2004b), and based pri
marily on the proposals made by developed countries, stipulates the reduction
of industrial tariffs in both developed and developing countries according to a
formula yet to be agreed.
There are several proposals on the table, including linear formulas wherein
tariffs would be cut by a certain rate regardless of their levels, and non-linear
formulas which would reduce higher tariffs by greater rates, thereby bringing
harmonization both across countries and tariff lines. Almost all the formulas so
far proposed would entail deep cuts in bound and/or applied industrial tariffs of
developing countries. But this is much more so in formulas proposed by devel
oped countries.
In the debate on the implications of cuts in industrial tariffs for developing
countries, attention has focussed on two issues. First, their impact on imports,
exports, production and employment in the sectors affected by tariff cuts and
increased market access. Second, their implications for government revenues
from trade taxes, particularly where such taxes account for an important part of
the budget.
Less attention has been paid to the implication of tariff cuts for industrialization
in developing countries and their participation in the international division of
labour. While it is generally agreed that there may be temporary costs, there is
also a widespread belief, in accordance with the prevailing orthodoxy, that pro
posed tariff reductions would be beneficial to developing countries when ad
justment to a more liberal trade regime is completed and existing resources are
fully redeployed and utilized according to new incentives.
However, for developing countries what matters is not one-off welfare gains or
losses resulting from reallocation of existing resources, but the longer-term
implications of proposed tariff cuts for capital accumulation, technical progress
and productivity growth which hold the key to narrowing income gaps and
catching up with richer countries. Even if there could be an instantaneous, costless
adjustment to a new set of incentives allowing developing countries to fully
realize the benefits of their comparative advantages as determined by their ex
isting endowments and capabilities, an irreversible commitment to low tariffs
across a whole range of sectors would carry the risk of locking them into the
prevailing international division of labour.
It is true that tariff protection is not always the only or even the best way to
promote technologically advanced and dynamic industries. However, many of
the more effective and first-best policy options successfully used in the past for
industrial upgrading by today’s mature and newly-industrialized countries are
no longer available to developing countries because of their multilateral com
mitments in the WTO, notably in agreements on subsidies, trade-related invest
ment measures (TRIMs) and trade-related aspects of intellectual property rights
(TRIPS). The loss of freedom to use policy tools in these areas increases the
risks entailed by narrowing policy autonomy further through irreversible com
mitments for deep cuts in industrial tariffs.
2
This booklet focuses on the implications of the negotiations on industrial tariffs
for longer-term industrialization in developing countries.
Chapter 2 gives a brief overview of the NAMA framework without getting into
technical details of various proposals with which the trade negotiators in Geneva
grapple on a daily basis and which have been examined in various documents
and papers prepared in the WTO, the UN Conference on Trade and Develop
ment (UNCTAD) and elsewhere.1
This is followed by a brief review of the historical experience of today’s ad
vanced countries regarding the use of tariffs in the course of their industrializa
tion, which is compared and contrasted with the actual situation prevailing in
developing countries today and the proposals put forward.
Chapter 4 discusses the sectoral pattern and evolution of tariffs that may be
needed in the course of industrial development in comparison with the con
straints that would result from the proposals made by developed countries, and
advances a simple alternative formula that can help reconcile policy flexibility
with multilateral discipline.
This is followed by a chapter on an evaluation of various estimates of benefits
of tariff cuts to developing countries.
Chapter 6 turns to the question of reciprocity from a broad developmental per
spective.
The booklet concludes with a brief summary of the key points on how the nego
tiations could accommodate both the immediate needs and longer-term inter
ests of developing countries.
1 For a discussion of issues related to NAMA, modalities for negotiations, and various formu
las and their implications for tariff cuts, sec UNCTAD (2003); Laird et al. (2003); Fernandez de
Cordoba et al. (2004b); WTO (2003a. b and c); Khor and Goh (2004); and Das (2005).
3
The Key Elements of the NAMA
2 Framework
There appear to be four interrelated objectives pursued by developed countries
in the WTO negotiations on industrial tariffs, which underlie the framework in
Annex B of the July Package:
(a)
Full Binding Coverage
The first objective is that with some minor exceptions all tariffs should ulti
mately be bound. While most developed countries have almost full binding
coverage, this is not the case for the majority of developing countries, particu
larly outside Latin America. For some 30 countries binding coverage is less
than 35 per cent, and about a third of these are non-LDCs from Africa.
The proposal for these countries is to bind all their non-agricuitural tariffs at or
below the average level of bound tariffs of developing countries taken together.
These countries would be exempt from making tariff reductions through the
formula. For others, unbound tariffs would be fully bound after applying the
formula for reduction from twice the applied rate. While for some tariff lines
the newly bound rates would continue to be above the current applied rates, in
many cases there could be a reduction in applied rates as the newly bound rates
would fall below the current applied rates.
The proposed increase in binding coverage in developing countries, if adopted,
would lead to a considerable reduction in the scope to use trade policy for in
dustrialization. Commitments are not time-bound, to be renegotiated after a
pre-specified period according to the outcome obtained, but are permanent.2
It is true that GATT rules allow countries to resort to measures such as anti
dumping duties or safeguards when imports cause “injury to domestic produc
tion”, or even to renegotiate their tariffs. However, these are exceptional and
temporary provisions, or require agreements among contracting parties and in
volve compensation. They are not designed to allow developing countries to
pursue effective industrial policies in order to promote competitive firms in
more dynamic, high value-added sectors by providing them appropriate infant
industry support against mature firms from more advanced economies.
(b) Rapid and Continued Liberalization
Another objective is that, whatever their initial positions, countries should lower
their tariffs over time in successive rounds. Accordingly, a successful conclu
sion of the Doha Round is expected to include lower tariffs for industrial prod
ucts, coming on top of large reductions already committed by developing coun
tries during the Uruguay Round. Indeed, an overarching objective pursued by
some of the most advanced countries is a rapid convergence to free trade.
The formula put forward by the United States for cuts in industrial tariffs pro
poses successive reduction of tariffs in two phases of five-year duration each,
culminating in free trade after the second phase.
The objective of progressive liberalization is also implicit in the proposals put
forward by some developing countries, although there is a difference in the
speed of liberalization. Furthermore, liberalization is pursued on a line-by-line
basis; that is, tariff cuts would be applied to all product categories, with some
2 This stands in sharp contrast with free trade agreements of the 19th century, discussed in the
next chapter, which all had fixed lifespans.
5
minor exceptions for what countries may consider as sensitive products. This
stands in sharp contrast with the approach adopted during the Uruguay Round
when commitments by developing countries were for an average level of tariffs
without any obligation to apply reductions to all tariff lines.
(c)
Harmonization Across Countries
The third objective is reduction in tariff dispersion across countries. Currently
the average weighted bound tariffs are close to 14 per cent in developing coun
tries and 3 per cent in industrial countries. Under the EU proposal, the differ
ence would be cut to 4 percentage points while in the proposal by the United
States it would altogether disappear after the second phase. Even the proposals
by China, India and Korea would imply sizeable reduction in average tariff
differences between developed and developing countries (Laird et al. 2003,
table 3).
Again, there would be a considerable compression of tariff differences among
developing countries. For instance, a sharp cut according to the Swiss formula
would reduce the dispersion of industrial tariffs among developing countries,
excluding least developed countries (LDCs), as measured by standard devia
tion of average bound tariffs, from more than 20 percentage points to less than
3 percentage points. Even a more moderate application of the non-linear for
mula would reduce the intra-developing country dispersion of bound tariffs to 6
percentage points?
(d) Greater Uniformity of Tariffs Across Product Lines
Since the proposed tariff cuts would be applied on a line-by-line basis, the re
sult would be a considerable decline in tariff dispersion across products. This is*
’ For bound tariffs of individual developing countries before and after the application of vari
ous formulas, see Fernandez de Cordoba el al. (2004b, Appendix table Al).
6
explicitly stated in the EU proposal where tariffs would be compressed into a
range with an overall cap of 15 per cent. Again the Indian proposal that tariffs
on any single product should not exceed the average tariff by more than a factor
of three effectively implies smaller dispersion.
Similarly, non-linear formulas would compress tariff dispersion by imposing
sharper cuts on higher tariffs. The application of a non-linear formula could
reduce the dispersion of tariffs among industrial sectors, as measured by stan
dard deviation of average weighted bound tariffs, by more than two-thirds.4
Greater uniformity of industrial tariffs would no doubt imply a reduction in
tariff peaks. In developed countries cuts in tariffs on products of export interest
to developing countries such as textiles, clothing and footwear would be deeper,
since tariffs applied to these products exceed by a factor of two the tariffs on
those imported mainly from other developed countries. However, the move
towards uniform tariffs would be much more rapid in developing countries where
tariff dispersion is larger. This would also mean reduced ability of these coun
tries to differentiate between imports of basic necessities and luxury
consumables; among intermediate, capital and final goods; and between high
and low value-added manufactures in their treatment of tariffs.
4
These estimates arc based on Fernandez de Cordoba et al. (2004b, table 8).
7
3
A Brief History of Industrial
Protectionism: Good for the Goose,
but not for the Gander
(a) The Goose
The principles espoused by developed countries in current negotiations on
NAM A do not conform to their historical experience regarding the use of tariffs
for industrial development. As documented in the literature on the economic
history of Western Europe and its offshoots, protectionism was the rule, free
trade the exception during the industrialization of today’s mature economies
(Bairoch 1993, p. 16). Follower countries used all kinds of policy tools to
support and protect their infant industries in order to catch up with the more
advanced economies:
‘"Almost all now-developed countries went through stages of industrial as
sistance policy before capacities of their firms reached the point where a
policy of (more or less) free trade was declared to be in the national inter
est. Britain was protectionist when it was trying to catch up with Holland.
Germany was protectionist when trying to catch up with Britain. The United
States was protectionist when trying to catch up with Britain and Germany,
right up to the end of the World War II. Japan was protectionist for most of
the twentieth century up to the 1970s, Korea and Taiwan to the 1990s.”
(Wade 2003, p. xv)
In the Western European core, following the widespread mercantilism that per
vaded the earlier centuries, there was a brief period of free trade beginning in
the 1840s.5 This coincided with the emergence of Britain as the industrial
hegemon, achieved under high barriers to imports, and started with the repeal
of Com Laws. Tariffs on manufactured imports were brought down to zero by
the 1860s from levels as high as 50 per cent in the 1820s.
Liberal trade policy spread to Europe with the Anglo-French trade agreement
of 1860 which replaced all restrictions on imports in France with ad valorem
tariffs not exceeding 30 per cent. This was followed by a series of similar
treaties among Western European countries. However, this episode of liberal
trade policy was followed by a protectionist backlash in the late 1870s and
early 1880s, leading to increases in industrial tariffs in several follower coun
tries including Germany and France.
In the period until the First World War, tariffs remained at relatively high levels
outside Britain, the Netherlands and a few smaller European countries such as
Belgium and Switzerland. In the interwar period, there was a proliferation of
tariffs and non-tariff barriers, including in Britain which started to feel the com
petitive pressures from the newly emerging industrializes, notably Germany
and the United States, and eventually resorted to high tariff barriers on the eve
of the Second World War.
The postwar era witnessed another wave of liberalization with gradual declines
in industrial tariffs in the developed world, this time driven by the new indus
trial hegemon, the United States. However, as argued by UNCTAD (1984, p.
75), “one of the elements which permitted this trend towards tariff liberaliza
tion was an incipient tendency to apply measures of a flexible nature (non-tariff
measures) on an increasing scale, a tendency to manage trade in certain sensi
5 For the historical evolution of industrial tariffs in the developed world the classic reference is
Bairoch (1993). See also O’Rourke and Williamson (2000, chaps. 3 and 6); and Chang (2002,
chap. 2). For the evolution of tariffs in general see Williamson (2003).
9
tive sectors, and a tendency to apply such restrictive measures on a discrimina
tory basis.”6
It is also notable that throughout its industrial development the United States
was more protectionist than other early industrializers. It was indeed described
as “the mother country and bastion of modem protectionism” (Bairoch 1993, p.
30). From the beginning of the 19th century until the 1840s its average tariffs
varied between 20 and 50 percent (Irwin 2003), while its industrial tariffs were
as high as 40 pcr cent in 1820, a level which was generally maintained until the
1840s.
The United States also entered a period of more liberal trade policy in the late
1840s, but its tariffs were still kept at much higher levels than in the Western
European core. Moreover, this liberal episode lasted even shorter, with average
tariffs returning to 40-50 per cent levels in the 1870s when custom duties ac
counted for more than 50 per cent of the United States government revenue
(Irwin 2002a, figure 3, p. 15). Until the First World War, tariffs were also
higher in all other European offshoots including Australia, Canada, New Zealand,
and Argentina than in the core countries (Irwin 2002b, p.5 and figure 1).
In the United States there was a brief easing of tariffs around the First World
War, before they were raised again to exceed 35 per cent in the mid-1920s, and
48 per cent with the onset of the Great Depression. It was only after the Second
World War that the United States started to move to sustained trade liberaliza
tion, having successfully established its industrial dominance behind protec
tionist barriers. Even then, as noted by Chang (2002, p. 29), “the USA never
practised free trade to the same degree as Britain did during its free-trade pe
riod (1860-1932).”
The historical evolution of industrial tariffs in the United States is described in
Figure 1. Its average industrial tariffs were relatively low at its early stages of
6 There was also a general upward drift in the importance of subsidies after the early 1950s
(Hufbauer 1983). Shutt (1985) argues that liberalization was an illusion as tariff cuts were
offset by a greater resort to market-distorting state intervention.
10
industrial development, rising rapidly in intermediate stages and falling with
maturity. Figure 1 excludes two periods where tariffs temporarily diverged
from their long-term path; that is, the liberal episode of 1846-1861, and the
Smoot-Hawley increase during the Great Depression. However, not only were
these extreme episodes temporary but, as noted above, declines in the former
period and increases in the latter were quite moderate compared to levels pre
vailing previously.
FIGURE 1: United States Industrial Tariffs
Evidence shows that there was a strong correlation between protectionism and
economic growth in the United States throughout the 19th century until the
Second World War (Bairoch 1993, pp. 52-53; O’Rourke 2000; Clemens and
Williamson 2001). Indeed during that period, not only did the United States
have the highest tariffs, but it was also the fastest-growing economy.
Although it is true that the correlation between high tariffs and economic growth
does not imply causality and there are many other factors than infant industry
protection that contribute to rapid growth (e.g. Irwin 2000 and 2002b), it is
notable that not only was the correlation valid for several Western European
11
countries, but it was also robust after taking into account other factors affecting
growth in cross-country regressions (O'Rourke 2000).
While the United States was protectionist across a wide range of industries,
Japan (like Germany and Sweden) was more selective (Chang 2002, p. 48),
closer to the sectoral pattern described above. Its experience was described by
the World Bank in its study on The East Asian Miracle in the following terms:
“As late as 1968. effective rates of protection (ERPs) in Japan were quite
high and exhibited the cascaded pattern from raw materials (low) to con
sumer products (high) that is typical of most developing countries. Unlike
many import-substituting economies, however, there was surprisingly high
protection of machinery (final producer goods) ... [which was] reduced
during the 1970s only after it was evident from export performance that the
sector had become internationally competitive. Quite high levels of protec
tion were afforded sectors such as iron, steel and non-ferrous metals as late
as 1970. Protection in capital-intensive sectors such as pulp, paper and
chemicals also remained high, to say nothing of the remarkably high levels
in textiles.” (World Bank 1993, p. 295)
Korea followed in the footsteps of Japan, except that it was less willing to move
to free trade: “even by 1983, when Korea’s success had become an established
fact, most sectors were still protected by some combination of tariffs and non
tariffbarriers. While Korea utilized a variety of instruments, especially export
targets and rebates, to ensure that exporters faced international prices for their
tradable inputs, there was considerable protection of goods sold on the domes
tic market” (World Bank 1993, p. 297).
(b) And the Gander
Compared with the historical experience of mature and newly industrialized
countries, trade policy in developing countries today appears to be unduly lib
eral. Table 1 makes a comparison of the historical experience of the three core
12
Western European economies and the United States with the current situation
in developing countries and LDCs, and three large developing economies:
•
At the end of the 19th century when per capita income (measured in pur
chasing power parity) in the United States was at a similar level as that in devel
oping countries today (that is, some $3,000 in 1990 dollars), its weighted aver
age applied tariffs on manufactured imports was close to 50 per cent, compared
to 8.1 per cent in developing countries and 13.6 per cent in LDCs today.
•
In 1950 when the United States was already an undisputed industrial he
gemon with a per capita income of almost three times the per capita income of
developing countries today, its average applied industrial tariff rate was higher
not only than the average rate applied by developing countries but also by LDCs
today. This is also true, to varying degrees, for Germany, France and the United
Kingdom.
•
When the United States had the same levels of per capita income as Brazil
or China today, its applied tariff rates were four times higher. When its per
capita income was similar to India today (that is, around the mid-19th century),
its average tariff was twice as high. Again, all Western European core econo
mies had higher industrial protection than Brazil, China and India today when
they had similar per capita income levels.
The figures in Table 1, however, do not fully reflect the extent of protectionism
in industrial countries in the past in comparison with developing countries to
day. High tariffs in the earlier period came on top of much higher transportation
and information costs than today, which provided natural protection from im
ports, particularly for the European offshoots (Clemens and Williamson 2001,
pp. 19-20). More important, the underlying rationale for tariffs in followers is
the productivity gap with the more advanced economies.
As argued by Chang (2002, p. 67), “the productivity gap between today’s devel
oped countries and developing countries is much greater than that which used
13
TABLE 1: Industrial Tariffs: Historical Comparison between Developed and
Developing Countries
Country/Year
Per Capita Income (at I990S)
Average Applied Tariffs (%)
1820
1257
35-45
1875
2445
40-50
1913
5301
44.0
1950
9561
14.0
1980
18577
7.0
1913
3648
13.0
1950
3881
26.0
1980
14113
8.3
1913
3485
20.0
1950
5270
18.0
1980
15103
8.3
1913
5150
0
1950
6907
23.0
1980
Developing
Countries 2001
12928
8.3
3260
8.1(2.!)*
LDCs
2001
898
13.6(13.6)*
Brazil
2001
5508
10.4 (4.0)*
China
2001
3728
12.3 (1.2)*
India
2001
1945
24.3 (9.0)*
US
Germany
France
UK
Source: Per capita income from Maddison (2001) at 1990 dollars based on multilateral
PPP. The latest available figures (1998/99) are adjusted for subsequent growth to arrive
at estimates for 2001.
Tariffs for developed countries from Bairoch (1993, table 3.3, p. 40) and for developing
countries from Fernandez de Cordoba et al. (2004b, Appendix).
’ Average tariffs that would result from the application of a Swiss formula according to
what Fernandez de Cordoba ef al. (2004b) call “hard scenario".
14
to exist between more developed and less developed NDCs [now-developed
countries] in earlier times. This means that today’s developing countries need
to impose much higher rates of tariff than those used by NDCs in the past, if
they are to provide the same degree of actual protection to their industries as
that once accorded to the NDC industries.” For instance, even though the United
States in 1913 had the same per capita income as Brazil today, it was already
one of the most developed economies in the world with effectively no produc
tivity gap with the industrial leader of the time, the United Kingdom.
Since Brazil now faces a larger income gap with industrial leaders, the same
level of tariffs would provide much less protection to its industry today than it
did for the United States in the earlier period.
Although major industrial countries have had continuous tariff liberalization in
the postwar era, the tariff levels reached by 1980 are not very much different
from the average rate of applied tariffs in developing countries today. Further
more, there was widespread resort to non-tariff measures which were “applied
to approximately one-quarter of Switzerland’s total imports and more than onetenth of the imports of Japan, Norway and the European Economic Commu
nity”, even without the inclusion of voluntary export restraints in the inventory
of non-tariff measures (UNCTAD 1984, p. 66, and table 14).
Some of the formulas currently proposed by developed countries imply that the
weighted average applied tariffs of developing countries would be reduced by
more than two-thirds and weighted average bound tariffs by more than threequarters of their current levels (Fernandez de Cordoba et al. 2004b). As can be
seen in Table 1, these would constitute much deeper cuts than those made by
most major developed countries in the 30 years after the Second World War.
An objective pursued by developed countries in the WTO negotiations on NAMA
is greater harmonization of industrial tariffs across countries, which calls for
deeper cuts in developing countries since their tariffs are typically higher. This,
again, stands in sharp contrast with their historical experience. Even though
the leading industrial nations used their political, military and economic lever
15
age throughout the 19th and much of the 20th centuries to push for liberalization-cum-harmonization in weaker countries, cross-country dispersion of tar
iffs was much greater than is the case today.
Lack of trade policy autonomy in colonies was a factor favouring greater har
monization of tariffs. There was indeed a high degree of correlation among the
tariff levels of imperial powers and colonies throughout the 19th and 20th cen
turies. This was particularly true for British colonies in Asia, including Burma,
Ceylon and India which operated under imperial tariff policies and mimicked
the trade policies of their masters. Similarly the Norwegian tariff policy was
subject to Swedish rule until the early 20th century, and the Indonesian policy
to Dutch rule until its independence.
It has been convincingly argued that this imperial dominance goes a long way
in explaining why colonies in Asia kept much lower tariffs than independent
Latin American countries through the 19th century until the Second World War.
Tariffs in the former region were aligned to relatively liberal regimes in Britain
and the Netherlands, while Latin American countries kept high tariffs in view
of excessive protectionism in their main trading partner, the United States
(Clemens and Williamson 2002b; and Williamson 2003, pp. 5-7).
Political and military power also exerted a strong influence over the tariff poli
cies of independent but weak states through the so-called unequal treaties. This
includes the gunboat diplomacy in Asia forcing Japan and China to open up
their markets to the United States and Britain respectively, as well as the 1860
trade agreement imposed on the Ottoman Empire by Britain when the country
defaulted on its external debt, and public finances were effectively taken over
by the creditors.7 Since the main objective in such instances was to make the
weaker states open their markets, such treaties did not always promote harmo
nization, particularly when the imposing countries were protectionist. This was
certainly the case when the United States forced Japan to open its markets while
protecting vigorously its own.
’ On the gunboat diplomacy see Clemens and Williamson (2002b, pp. 6-7), and on the Otto
man predicament see Kiray (1990, pp. 258-60).
16
An important factor that favours greater harmonization is that once countries
establish industrial dominance behind protectionist walls, they tend to advo
cate free trade in order to kick away the ladder from the followers and consoli
date their dominance, as lucidly shown by Chang (2002), reviving the term
originally introduced by the German economist Friedrich List. As noted above
this was certainly the case for Britain in the mid-19th century which led the
liberalization drive in Europe. The United States followed a similar path a
hundred years later.
However, despite these tendencies making for greater liberalization-cum-harmonization, there was still considerable diversity among the contemporaneous
industrializers in the core both during the 19th century and the first half of the
20th century. For instance, in the interwar period the cross-country dispersion
of industrial tariffs was quite wide, ranging from zero in the United Kingdom to
6-10 per cent in the Netherlands, 20 per cent in Germany, 30 per cent in France,
46-50 per cent in Italy and the United States.
As shown in Table 2, the dispersion of tariffs among developed countries, as
measured by standard deviation, was close to 11 percentage points even in 1875
when trade was relatively free. This figure more than doubled by 1913 after the
globalization backlash, before coming down to 7 percentage points in 1950.
Thus, cross-country dispersion of industrial tariffs was quite high during the
industrialization of today’s developed countries, not only in absolute terms but
also relative to average tariffs (as demonstrated by the coefficient of variation
in the last column of Table 2), coming down only after followers narrowed the
development gap with the leaders.
For a broader range of countries, both in the core and the periphery, the disper
sion was even larger. It was also in the wrong direction as tariffs in many poor
countries in the periphery were lower than those in more advanced economies.
For instance, a study contrasting the behaviour of tariffs over 1865 and 1938 in
six regions (Asia, European core, European periphery, Latin America, United
States and other offshoots) found “enormous variance in levels of protection
between the regional averages” (Williamson 2003, p. 4).
17
It is even more striking that, as shown in Table 2, the dispersion of average
tariffs applied by developed and developing countries was much higher, both in
absolute and relative terms, during the 19th century than at present, despite
harmonizing influences associated with imperial rule and gunboat diplomacy
in the earlier period. This is true whether one takes the figures during the
liberalization episode (1870) or during the tariff backlash (1890). Table 2 also
shows that the application of the non-linear Swiss formula in current WTO
negotiations could take “harmonization’’ between developed and developing
countries much further than was ever achieved under imperial rule or gunboat
diplomacy.
TABLE 2: Average Applied Tariffs and their Dispersion
Sample
Year
Average applied
tariff (per cent)
Developed countries'
Developed countries1
Developed countries'
1875
1913
1950
12.1
23.1
15.8
21.8
7.1
0.9
0.9
0.5
1870
13.8
10.7
0.8
1890
15.2
12.6
0.8
2001
9.3
5.7
0.6
After
4.1
harmor lization
2.9
0.7
Developed and
developing countries2
Developed and
developing countries2
Developed and
developing countries3
Developed and
developing countries3
Dispersion
Coeeficientof
(percentage)
variation
10.8
1. Industrial tariffs for 14 developed countries for 1875 and 1913, and 10 developed
countries for 1950, calculated from Bairoch (1993, p. 40, table 3.3.).
2. All tariffs for 27 countries, calculated from Irwin (2002b, p. 27, Appendix table 1).
3. Industrial tariffs for 84 countries (excluding LDCs), calculated from Fernandez de
Cordoba et al. (2004b, Appendix table A1). Harmonization refers to tariffs that would
result from what the authors call “hard scenario".
18
4 Industrial Development and Tariffs
(a)
Stages of Industrial Development
The key question raised by the NAMA negotiations is the extent to which the
proposals put forward by developed countries would affect longer-term indus
trialization prospects of developing countries. In examining this issue, it is im
portant to bear in mind that successful industrialization is a cumulative process
involving movements from one stage to another through the establishment of
new industries with higher value-added and technology contents.
In the earliest stages of economic development, production and exports consist
largely of primary commodities while imports comprise mainly manufactures,
both capital- and labour-intensive products. Exporting at such a stage provides
a vent for surplus; that is, it allows production to increase by making use of
formerly unemployed resources because of lack of domestic demand. As these
sectors enjoy natural resource-based comparative advantages of the kind em
phasized by the Ricardian theory of trade, their mobilization does not call for
specific support and protection.8 It does, however, raise other policy issues
linked to distribution of rents, particularly when foreign firms are involved
(Prebisch 1950, Singer 1950).
How long a country can rely on the exploitation of natural resources before
moving to industry depends, inter alia, on the relative size of its resource en
8 Fora discussion of (lie distinction between natural and nurtured comparative advantages see
Gomory and Baumol (2000, chapter I). See also Akyiiz (2005, Part II).
dowments. However, evidence strongly suggests that rich natural resources,
even when combined with a well-developed human resource base, do not auto
matically lead to processing and diversification. Without active policies de
signed to promote and support such activities, being rich in natural resources
can be detrimental to diversification away from unprocessed commodities.
On the other hand, even though commodity processing provides early industri
alization opportunities, the possibilities of maintaining rapid development
through deepening and diversifiation in the primary sector are limited.
Manufactures offer better growth prospects not only because they allow for a
more rapid productivity growth and expansion of production, but also because
they avoid the declining terms of trade that have frustrated the growth prospects
of many commodity-dependent economies. Countries rich in natural resources
can delay industrialization, but in general they cannot reach high income levels
without a strong industrial base.9
The early stages of industrialization are characterized by sectoral specialization
in exploiting endowments of natural resources and unskilled labour. This is
followed by diversification into a wide spectrum of technologically more ad
vanced activities, accompanied by increased internal integration through a dense
set of linkages among sectors.10 With industrial maturity there is again a move
towards sectoral specialization, this time at the top end of the technology lad
der. This pattern is also confirmed by empirical evidence on the evolution of
sectoral allocation of labour in the course of industrial development.
A study using data from a variety of sources covering a wide cross-section of
countries found “robust evidence that economies grow through two stages of
’ Countries such as Finland and Sweden diversified based on their natural resources, but their
success in industrialization depended on moving to technology-intensive manufacturing. For a
discussion of processing and diversification in timber and iron-related industries see UNCTAD
(1996, Annex to chap. JI). See also UNCTAD (2003, pp. 92-93).
10 For a discussion of the importance of internal integration in economic development and the
trade-offs and complementarities involved between internal and external integration see Wade
(2003, pp. xlviii-li).
20
diversification. At first, labour is allocated increasingly equally between sec
tors, but there exists a level of per capita income beyond which the sectoral
distribution of labour inputs starts concentrating again. In other words, the
sectoral concentration of labour follows a U-shaped pattern in relation to per
capita income.... The non-linearity holds above and beyond the well-known
shifts of factors of production from agriculture to manufacturing and on to ser
vices” (Imbs and Wacziarg 2000, pp. 1-2). The turnaround from sectoral diver
sification to specialization occurs quite late in the development process, around
a per capita income of $9,000.
During the initial expansion in resource-based and labour-intensive manufac
tures, the support and protection provided to industry will likely be phased out
after a relatively short period of learning and expansion in world markets, since
such sectors tend to be technologically less demanding. As traditional indus
tries mature and become competitive, a new generation of infant industries would
need to emerge and establish themselves. Indeed, an effective industrialization
strategy should recognize that currently successful industries may, over time,
confront difficulties in competing in international markets as domestic wages
rise, low-cost competitors emerge, and the limits of learning and productivity
growth are reached. Hence, more dynamic and skill- and technology-intensive
industries would need to be promoted simultaneously as resource-based and
labour-intensive manufacturing successfully carries the economy forward.
Such an approach underpinned successful modem industrializers such as Ko
rea which started to build up from an early date scale- and technology-inten
sive industries, including shipbuilding, steel and automotive industries. Rather
than seeking to maintain competitiveness by keeping down wage costs or pro
tecting traditional industries with high tariffs, they chose to upgrade rapidly as
a way of raising productivity, exports and incomes.
Eventually these scale- and technology-intensive industries will have to com
pete with firms in more mature economies which enjoy the advantage of having
begun sooner and progressed further on the technology ladder. But, as argued
by Gomory and Baumol (2000, p. 6) “entry into one of these industries, against
an entrenched competitor, is slow, expensive, and very much an uphill battle if
left entirely to free market forces.” They would thus need to be supported,
including with industrial tariffs and various forms of subsidies, of the kind widely
used in both mature and newly industrialized countries in the past. Such sup
port would likely be higher and maintained for longer periods compared to less
demanding, resource-based and labour-intensive manufacturing.
In this process, as new and more dynamic industries emerge, the traditional
ones are phased out and may even be left entirely to countries at earlier stages
of development. This pattern of modem industrialization, dubbed “the flying
geese paradigm”, was originally formulated in Japan in the 1930s when it was
still a comparatively poor economy (UNCTAD 1996, Part Two, chapter 1). It
provides a description of the life-cycles of various industries in the course of
economic development and their relocation from one country to another through
trade and foreign direct investment (FD1) in response to shifts in competitive
ness.
In this process, imports from more advanced economies allow new goods and
technology to be introduced in less advanced economies. The next stage is to
promote indigenous industries to replace imports in meeting domestic demand,
to be followed by exports. When a country loses competitiveness in a particu
lar product, its domestic production is phased out and replaced by imports from
the followers."
While the flying geese paradigm assumes an outward-oriented strategy, it is not
a market-driven process. Success in industrial upgrading would require policy
intervention in the form of infant industry support and export promotion, in
order to even the playing field with firms from more advanced economies. Ini
tially there would be no need for tariffs on products for which the economy
relies entirely on imports. Subsequently, as indigenous industry is established,
tariffs are introduced for infant-industry protection. And eventually protection
" Vernon’s (1966) product-cycle theory also gives a similar description of shift of production
across countries. However, it focuses on the behaviour of transnational corporations (TNCs)
and sees trade and FDI as successive stages in production for foreign markets.
22
and support would be removed as the industry matures. In this process, the
economy goes through a series of overlapping industries according to their life
cycles, constantly raising productivity as it moves up the technology ladder.
Similarly, building on the work of Young (1928), Kaldor (1966) described in
dustrialization as a cumulative process going through four stages, based on a
distinction between consumer and capital goods. In the first stage a local con
sumer goods industry emerges, substituting imported manufactured consumer
goods. As competitiveness is established, the economy moves to the second
stage, exporting consumables but still dependent on imported capital goods.
The third stage is characterized by mass production and export of consumables
combined with the emergence of a local industry to replace imported capital
goods, to be followed by capital goods and technology exports. In this process,
scale economies and learning play a crucial role. While industrialization fol
lows a clear trajectory of progress, it does not converge to a predefined point.
Rather, selection is involved across a whole range of industries and products in
each stage of development, influenced by policy including import restrictions.12
(b)
Pattern and Evolution of Optimum Industrial Tariffs
These considerations suggest a pattern of optimal tariffs in the course of indus
trial development as described in Figure 2. Four different categories of prod
ucts (industries) are selected according to a broad classification developed in
UNCTAD (1996, pp. 116-117; and 2002, Annex to chapter III): resource-based
and labour-intensive manufactures (RL), and low (LT), medium (MT) and high
(FIT) technology- and skill-intensive products.13 Tariffs are introduced once a
particular line of industry is entered, and kept at their initial (maximum) levels
12 See Argyrous (1996) who distinguishes between low-end and high-end capital goods, with
the former referring to standard, off-the-shelf equipment, the latter to custom-made machinery
built for special purposes.
13 This classification based on products does not capture all aspects of manufacturing produc
tion. Many technology-intensive products involve labour-intensive processes, such as the as
sembly of imported electronic parts and components in developing countries participating in
international production networks (Il’Ns). These should appear in the RL category; see Akyuz
(2003, chapter 1; and 2005).
23
for a certain period before being brought down at a constant rate as the industry
matures.14
For the reasons already noted, technology-intensive industries have higher ini
tial levels of protection and support than resource-based and labour-intensive
manufacturing.15 As technological capacities are built successfully, subsequent
shifts to more advanced sectors become relatively easier than the earlier move
from labour-intensive to technology-intensive activities. Accordingly, in Fig
ure 2, peak tariff rates are assumed to follow a non-linear path, rising initially
during the shift from RL towards LT and MT industries, and falling afterwards.
In the early stages of development, there would be no need for infant industry
protection against imports of MT and HT products since industries producing
these goods are not yet in existence. By the time the economy moves to MT
products, protection for RL products is assumed to have been fully phased out.
Clearly the process of sequencing industries can differ from country to country
depending on factors such as geography, size and endowments. In accordance
with the evidence noted above, industrial specialization in Figure 2 follows a
non-linear path, with greater sectoral concentration at the early and late stages
of industrial development, and diversification in between. In each stage there is
a diverse set of industries while different stages are characterized by different
levels of selection. Selection made in different countries in each stage of indus
trial development can show considerable variations depending on a host of fac
tors including institutional arrangements and endowments.
14 It is also possible to have non-linear paths for tariffs, falling at an accelerated or decelerated
rate after their introduction.
15 Since subsidies are substitutes for tariffs in maintaining domestic production above the level
that would be possible under free trade, the vertical axis may be conceived as including also the
tariff equivalent of subsidies. The horizontal axis could be defined in terms of per capita income
rather than time, but the latter is preferred here in order to emphasize the sequence of industries.
24
FIGURE 2: Tariff Profile of Sequenced Industries for Infant Industry Protection
MT
Tariffs
LT
\
HT
RL
\
\
RL: Resource-based and labour-intensive products
LT: Low technology-intensive products
MT: Medium technology-intensive products
HT: High technology-intensive products
Figure 2 describes the pattern and evolution of optimum tariffs that would be
needed for infant industry protection in late industrializes in order to overcome
their technology and skill gaps with the more advanced economies at each stage
of industrial development. For industrial leaders where technological advance
depends on innovation rather than adaptation of foreign technology, industrial
ization would not call for the kind of infant industry protection described in
Figure 2, but a host of other policies that help promote innovation and internal
ize its benefits.
In reality tariffs tend to be set and evolve in quite different ways from the pat
tern depicted in Figure 2 since they are imposed, inter alia, for balance-ofpayments and government revenue reasons. Furthermore, pressures by interest
groups or distributional considerations could push tariffs from the levels that
25
would maximize their economic benefits.16 Even though it would not generally
be efficient to have all four types of industry' operating simultaneously, labour
intensive sectors are often maintained behind barriers in economies which have
attained technological maturity. Indeed, in many developed countries textiles,
clothing, and footwear and leather goods receive far greater protection than
technology-intensive sectors against competition from low-cost developing
country producers.
Again, in some developing countries there have been attempts to establish MT
or even HT industries under strong protection at relatively early stages of de
velopment, before achieving efficiency in labour-intensive manufacturing. More
generally, under import-substitution regimes in many developing countries tar
iffs were often levied on an ad hoc basis with the consequence that “a great
hodgepodge of rates appeared, with virtually no evidence of any consideration
of costs or efficiency” (Bruton 1998, p. 912). More recently the tendency has
been towards indiscriminate liberalization adopted in many low-income coun
tries as a result of conditionality attached to adjustment lending by the Bretton
Woods institutions, and in some middle-income developing countries seeking
access to markets in the North through bilateral or regional trade agreements.
The use of tariffs in the course of technological upgrading along the lines de
scribed in Figure 2 implies that countries at the intermediate stages of develop
ment would have relatively low tariffs for products both at the top and bottom
ends of the technology spectrum, and higher tariffs on middle-range products.
By contrast, industrially advanced countries would have higher tariffs at the top
end. Available evidence from a recent study distinguishing tariffs on low, inter
mediate and high value-added products suggests that in reality this happens
only to a very limited extent (Fernandez de Cordoba and Vanzetti 2005, p. 11,
figure V). Both developed and developing countries have lower applied tariffs
on low value-added industrial products. Developing countries have higher av
erage applied tariffs for intermediate products than for high value-added prod
16 Several ad hoc theoretical and empirical models have been developed to account for politi
cal-economy influences on tariff policy. For a survey sec Gawande and Krishna (2003).
26
ucts, while in developed countries average tariffs are higher on high valueadded products. Nevertheless, in developed countries tariffs on intermediate
products are only marginally lower than those on high value-added products,
reflecting in large part tariff peaks on products of export interest to developing
countries. On the other hand, both developing countries and LDCs keep rela
tively high tariffs for high value-added products even though most of them have
not yet advanced beyond labour-intensive and resource-based industries (Akyiiz,
Kozul-Wright, and Mayer 2004).
(c)
Implications for NAMA Negotiations
The pattern of tariffs needed to support overlapping generations of industries in
developing countries conflicts with the objectives pursued by developed coun
tries in the current WTO negotiations on NAMA in several respects:
•
At any point in time, effective use of tariffs for industrialization would
require the coexistence of very low and very high tariffs. In Figure 2 at initial
stages of industrialization, tariffs are zero for MT and HT products, high for LT
products and moderate for RL products. Similarly during industrial maturity,
tariffs are zero for RL and LT products, moderate for MT products and high for
HT products. In the intermediate stages tariffs are concentrated on LT and MT
products, and there is no need for tariff protection for the RL industries because
they are competitive, and for the HT industries because they are not yet in
existence. Briefly, since at any point in time different industries would require
different degrees of infant industry protection, dispersion across tariff lines can
be very wide.
•
Over time tariff dispersion may be rising or falling according to the stage
of industrial development reached. In Figure 2 it is initially increasing as the
economy moves towards more demanding industries, but subsequently decreas
ing with industrial maturity.
27
•
In the course of industrialization tariffs are raised on some products but
lowered for others; that is, there is no continuous liberalization on a line-byline basis.
•
The behaviour of average tariffs in Figure 2 in the course of industrial
development is quite similar to the evolution of industrial tariffs in the United
States shown in Figure 1; they rise in the intermediate stages of industrialization
as the economy diversifies away from resource-based and labour-intensive manu
factures and then start falling with industrial maturity.
•
Since countries at different stages of industrial development can coexist,
there would be little harmonization across countries. Mature industrial coun
tries would have relatively low average applied tariffs compared to those at the
intermediate stages of industrialization, but not necessarily lower than those at
the early stages of industrialization. In other words, countries at the intermedi
ate stages of development need higher average applied tariffs than both mature
industrial countries and LDCs. However, while mature industrial countries are
expected to dismantle tariffs over time, LDCs would need to move towards
higher tariffs as they enter technology-intensive industries.
Although Figure 2 is a highly simplified picture of the possible evolution of
optimum infant-industry support and protection that may be needed at different
stages of industrial development, the results above do not depend on a particu
lar description of this process. A key conclusion is that in a process of sequen
tial build-up of competitive industries under temporary infant industry protec
tion, the optimal level and structure of tariffs would change over time. Conse
quently, focussing on the needs of existing industries or taking current levels of
tariffs as the basis for commitments in the WTO could subsequently present
serious setbacks to technological upgrading.
A country at an earlier stage of industrialization might be inclined to have very
low bound tariffs for high-tech products because it has not yet entered into such
sectors, aiming, instead, at retaining high bound tariffs for labour-intensive and
low-technology manufactures in order to protect its existing industries. But
28
emphasizing short-term benefits to the neglect of longer-term industrialization
could lock it into the current pattern of industrial specialization, making it dif
ficult to move up on the technology ladder. Similarly an LDC can remain de
pendent on primary production and exports if it is denied the space to develop
industries for labour-intensive manufactures.
It has been pointed out that the proposed tariff cuts in the WTO negotiations
would erode flexibility of developing countries in using trade policy for indus
trial development by bringing the new bound rates below the currently applied
rates (Laird et al. 2003, p. 16). The analysis here shows that binding at a low
rate could also present difficulties in entering an industry even if the bound rate
remains above the currently applied rate. Whether or not bound tariffs would
allow sufficient policy space for industrialization should be assessed not in terms
of a comparison with the currently applied rates, but the rates that may be
needed when the time comes to enter higher value-added, more dynamic indus
tries.
The key issue here is how to reconcile multilateral discipline with policy flex
ibility needed for industrial development. As shown above, developing coun
tries do not need high tariffs for all sectors and all the time. But they should
have the option of using tariffs on a selective basis as and when needed for
progress in industrialization. They should not be expected to keep moving
tariffs downward from one trade round to another, but be able to move them in
both directions in different sectors in the course of industrial development.
The analysis above suggests that this kind of flexibility is best accommodated
by binding the average tariff without any line-by-line commitment; that is, to
leave tariffs for individual products unbound, subject to an overall constraint
that the average applied tariffs should not exceed the bound average tariff.
Clearly, the average bound tariff should be high enough to accommodate the
needs of different sectors at different stages of industrial maturity. This would
not necessarily lead to high bound average tariffs. On the contrary, it could
result in lower average tariffs than would be the case under line-by-line com
mitments.
29
This is illustrated by the example given in Table 3 which draws on the pattern
described in Figure 2 wherein countries go through various phases of industrial
development, moving through a sequence of overlapping industries in RL, LT,
MT and HT products. The bold numbers are the maximum tariffs that would
initially be needed for each product for infant industry protection. Numbers in
the last column give the average applied tariffs in each phase of industrializa
tion. As in Figure 2, over time tariffs are lowered in some sectors but increased
in others. In this example, on a line-by-line commitment, a country at the first
phase of industrial development would want to bind its sectoral tariffs at the
maximum rates needed, ending up with an average bound tariff of 37.5 per cent
[(20+40+50+40)/4j. For an economy in phase II, the resulting average bound
rate would be 35 per cent, and in phase Ill 30 per cent. But for these economies
an average maximum bound tariff of 25 per cent would be sufficient to provide
infant industry support to all sectors in all phases of industrial development
provided that they are free to set applied tariffs for different sectors as needed.
Average applied tariffs would remain below 25 per cent in early stages of in
dustrialization, rising gradually in the intermediate stages and eventually fall
ing with industrial maturity. A country with an average bound rate of more than
25 per cent would be willing to cut it provided that it retains the freedom to set
sectoral rates as needed. Furthermore, once phase IV is reached, it would be
possible to reduce the average bound tariff gradually while maintaining higher
applied tariffs for more advanced sectors.
As countries are free to choose their applied tariffs for individual industries/
products subject to the overall constraint of an average bound tariff, such an
approach would balance multilateral discipline with policy flexibility. It would
also have the additional advantage of encouraging countries to view tariffs as
temporary instruments and to make an effort to ensure that infant-industry pro
tection succeeds in establishing competitive industries. This is because in or
der to stay within the overall limit of the bound average tariff, they would need
to cut tariffs in industries at the lower end before moving up towards the higher
end. For instance, a country cannot effectively move into phase 111 and estab
lish MT industries under tariff protection without first lowering its tariffs on RL
30
TABLE 3: Tariffs at Different Phases of Industrialization (per cent)
Phase
RL
LT
MT
HT
Average
I
20
0
0
0
5.0
11
10
40
0
0
12.5
III
0
30
50
0
20.0
IV
0
20
40
40
25.0
V
0
10
30
40
20.0
VI
0
0
15
25
10.0
VII
0
0
5
15
5.0
VIII
0
0
0
0
0.0
Notes: Bold numbers are the maximum tariffs initially needed for infant industry protection for
each product.
RL:
Resource-based and labour-intensive products.
LT:
Low technology-intensive products.
MT:
Medium technology-intensive products.
HT:
High technology-intensive products.
and LT industries from their initial levels, since this would result in an average
tariff of higher than 25 per cent. Finally, it would encourage developing coun
try trade negotiators to take a long view in making multilateral commitments,
rather than focussing on the immediate needs of their industries.
In a recent paper Das (2005, p. 5) argued that “developing countries should
only agree to cut in the average tariff” and should not undertake “obligation for
line-by-line reduction.” This would certainly be a significant improvement
over the current proposals by industrial countries. But the approach advocated
here goes further and calls for unbinding sectoral tariffs and negotiation of av
erage tariff’s only. Furthermore, because of different initial conditions, it is
unlikely to be compatible with any formula-based procedure of tariff reduction
Ct-
in the current negotiations even if such formulas were to be applied to average
tariffs only. Indeed, it may call for an increase in average bound tariffs for some
countries, particularly those which have recently acceded to the WTO on highly
unfavourable terms.
32
5
(a)
Assessing Costs and Benefits: Are
the Gains Worth the Pains?
Doubling the Effort
Fanaticism, according to the Spanish philosopher George Santayana, means re
doubling your effort when you have forgotten your aim. Economics is no ex
ception to such thinking. Thus, developing countries are once again facing
intense pressure to liberalize trade in industrial products even though the ex
travagant benefits claimed from the Uruguay Round have been belied by subse
quent experience.17
According to one estimate, the Uruguay Round’s combined liberalization in
creased global economic welfare by S75 billion, of which almost S70 billion
went to developed countries, $5 billion to Newly Industrialized Economies
(NIEs; Korea, Singapore and Taiwan), and none to developing countries taken
together.18
Despite this, recent years have seen a proliferation of similar exercises, claim
ing large benefits from further trade liberalization for the world economy in
general and for developing countries in particular. The gains estimated by vari
ous studies from full liberalization of goods and services trade, or trade in goods
alone, range from a couple of hundred billion dollars to more than S2,000 bil
17 For predictions of potential benefits of the Uruguay Round see Martin and Winters (1996).
18 Brown et al. (2001, p. 32, table 1). These estimates do not take into account the kind of
adjustment costs discussed in the following section. They thus overestimate the benefits of
liberalization. See Dorman (2001) for a critical assessment.
lion.19 On some accounts, the incidence of gains to developing countries reaches
as much as 65 per cent of the total.
It has also been argued that while a successful outcome of the Doha Round
would greatly improve the growth prospects of developing countries, these ben
efits would come primarily from liberalization in these countries themselves.
According to a scenario designed by the World Bank (2004, pp. 48-54) liberal
ization of both agricultural and manufacturing trade by both developed and
developing countries would generate some S290 billion in global economic
gains, of which $ 160 billion would go to developing countries and $132 billion
to developed countries.20
Manufacturing liberalization in developed countries would cany a small loss to
themselves while benefiting developing countries. However, developing coun
tries would gain significantly more from their own reforms than from increased
access to markets in developed countries. In commenting on these results the
WTO secretariat thus concludes that: “This lesson, that a large part of the
economic gains from trade liberalization accrue domestically, should not be
overlooked in the context of reciprocal bargaining for market access” (WTO
2004a, p. 2).
Other estimates are less sanguine about the potential benefits to developing
countries from liberalization of trade in industrial products. According to a study
on the incidence of benefits of liberalization of trade in mining and manufac
tures during the Uruguay Round, global gain was in the order of S90 billion, of
which $65 billion went to mature industrial countries, $6 billion to NIEs, and
less than $20 billion to developing countries (Brown et al. 2001, p. 31, Table 1).
” For a survey of these studies see Anderson (2004)
20 This scenario assumes, inter alia, that in developed countries average tariffs in manufactur
ing are reduced to one per cent with a cap of 5 per cent while in developing countries they are
reduced to 5 per cent with a cap of 10 per cent.
34
The same study estimates that a one-third reduction in post-Uruguay Round
manufacturing tariffs would lead to global gains of some $210 billion. Of this
around $ 160 billion would go to industrial countries ($58 billion to Japan, S31
billion to the United States and $63 billion to the EU plus EFTA), $16 billion to
NIEs and around $30 billion to the developing world (p. 37, Table 5).
Only a few attempts have been made to estimate the costs and benefits of the
various formulas proposed for reducing industrial tariffs in the current negotia
tions on NAMA. One such study simulates various scenarios, including one on
universal free trade drawing on the United States proposal, a scenario of ambi
tious liberalization, and a simple formula designed primarily to reduce tariff
peaks and escalation (Fernandez de Cordoba et al 2004a).
In all cases there are modest welfare gains for the world economy as a whole,
ranging from $28 billion (simple formula) to $42 billion (free trade). Under
free trade, winners are concentrated in Asia. In the scenario based on the simple
formula developing countries would obtain less than a third of the total gains.
Their exports increase as much as in developed countries, partly due to increase
in exports to other developing countries, while their imports increase faster. As
expected, total tariff revenues would fall less under moderate liberalization than
under free trade.
(b) Modelling the Benefits, Assuming Away the Costs
The estimates above are one-off static gains expected to result from realloca
tion of resources after trade liberalization.21 They are derived from computable
general equilibrium (CGE) models based on the neoclassical paradigm of com
petitive equilibrium where markets always clear and resources are fully em
21 For a discussion of static versus dynamic effects see Akyiiz (2005, section 11). Some of the
exercises noted above try to incorporate “dynamic" effects. For instance in the World Bank
(2004) scenario, trade is assumed to induce productivity growth so that the "dynamic” gains
from trade liberalization would be a multiple of static gains.
35
ployed?2 These estimates do not provide a reliable guide to what might happen
in reality because of two interrelated shortcomings of the CGE models.
First, the structure of a particular model determines the range of results that can
be obtained, and it is often the underlying theory and its assumptions that deter
mine what the numbers would show. Thus, a CGE model founded on conven
tional trade theory will generally behave and yield results in the manner deter
mined by its underlying assumptions, but shed no light on the validity of the
theory itself (Stanford, 1993. p. 100). For this reason it is almost impossible to
find any CGE model fashioned on the traditional trade theory which does not
predict gains from trade liberalization:
“Empirical comparative static studies of the economic welfare gains from
trade liberalization typically generate positive gains for the world and for
most participating countries. (Exceptions are when a country’s welfare is
reduced more by a terms of trade change or reduced rents from preferential
market access quotas than it is boosted by improvements due to reallocat
ing its resources away from protected industries).” Anderson (2004, p. 11)
As one commentator remarked, the “best definition of a CGE model is: ‘theory
with numbers’” (O’Rourke 1995, p. 1), rather than an empirical test of a theory.
But it is a theory that assumes away various imperfections and rigidities that
pervade developing economies and lead to market failures, including externali
ties, incomplete markets, imperfect and asymmetric information, monopolies
or imperfect competition. The incorporation of any of these could lead to prob
lems for the stability and even the existence of “equilibrium”. They could also
yield multiple equilibria, leading to uncertainty about the outcome of trade lib
eralization.
22 Not all CGE models generate full employment equilibrium. This depends on the assumption
about what is called “macro-closure”; that is, how the equilibrium in the product market is
attained. While the neoclassical closure gives full employment, under Keynesian closure aggre
gate demand and supply (or savings and investment) can be brought into equilibrium at less than
full employment. For a classification of empirical CGE modelling see Thissen (1998). The
estimates surveyed here generally use the GTAP (Global Trade Analysis Project) model devel
oped at the Purdue University, assuming neoclassical competitive equilibrium; see Anderson
(2004, Appendix).
36
The second problem is that although these are comparative statics exercises,
the difference between two equilibrium states is often presented as a change
from one to another. This tendency has long been noted:
“Following Joan Robinson’s strictures that it is most important not to apply
theorems obtained from the analysis of differences to situations of change
(or, at least, to be aware of the act of faith involved in doing this), modem
writers usually have been most careful to stress that their analysis is essen
tially the comparison of different equilibrium situations one with another
and that they are not analysing the actual processes. Nevertheless, in their
asides, they sometimes speak as if their results were applicable to a world
of change and as if‘back-of-an-envelope’ excursions into the statistics can
provide ‘realistic’ orders of magnitude to try out their theorems.” (Harcourt
1972, pp. 122-123)
Presenting comparative statics exercises as analyses of change can be mislead
ing. For instance, in competitive equilibrium an economy with a bigger labour
force will have a higher level of employment than another one similar to it in all
aspects except the size of the labour force, but it does not follow that an in
crease in population will lead to an increase in employment. Likewise, the
CGE models compare two equilibrium states of an economy differing in trade
policy but similar in all structural aspects. From the way the economies are
assumed to work follows the prediction that the equilibrium state with lower
tariffs will have higher income and welfare. But even then, this does not mean
that a reduction in tariffs would necessarily lead to another equilibrium state
with a higher level of income. This depends on how the economy will react to
disequilibria generated by liberalization. The kinds of problems involved were
explained by John Flicks some forty years ago:
“Even if the equilibrium exists, it has still to be shown that there is a ten
dency towards it. ... Even in the single market, under perfect competition.
and such that the existence of equilibrium is indubitable, there may be no
tendency to equilibrium, if speeds of reaction to price change are perverse.
Something has to be specified about reactions to disequilibrium before the
existence of a tendency to equilibrium can be asserted.” (Hicks 1965, p.
18)
37
The CGE economists do not specify reactions to disequilibrium brought about
by policy shocks, but assume that a new equilibrium is always reached. Never
theless, they also hedge by taking refuge in the concept of “adjustment costs”,
acknowledging that in the transition from one equilibrium state to another, re
sources may be temporarily unemployed, skills may be eroded, equipment may
become obsolete, government revenues may fall, trade imbalances may emerge,
or there may be all kinds of costs in learning to live with the new set of incen
tives (Akyiiz 2005, section II). Thus, for instance, the World Bank qualifies its
estimates on the benefits of unilateral trade liberalization in developing coun
tries in the following terms:
“The positive impact on overall growth, accompanied by a sharp boost in
trade and a poverty outlook improvement leaving all regions better off in
aggregate, does not signify that the reforms are without adjustment costs,
even over the long term.” (World Bank 2004, p. 53)
However, these “adjustment costs” are almost never quantified and incorpo
rated in the estimated benefits from trade liberalization simply because the CGE
economists are engaged in comparative statics, making comparisons between
two equilibrium states (i.e. two solutions, for different tariff rates, to a set of
simultaneous equations) without specifying how the economy moves from one
state to another. There are attempts to incorporate some of the deviations from
neoclassical assumptions into these models, such as labour market inflexibility,
as causes of adjustment costs. While this would alter the comparative statics
effects of liberalization, as measured by welfare differences between two equi
librium states, it would not account for what happens in between, which is what
the adjustment cost is about. More importantly, the very same deviations from
the conventional assumptions of the CGE models that lead to adjustment costs
could also prevent the economy from moving towards a new equilibrium with
an improved allocation of resources, or could actually lead to an equilibrium
with lower income and employment levels, making temporary adjustment costs
permanent. But this question is rarely asked, let alone answered.
38
When adjustment costs out-of-equilibrium are relatively large, there may be no
net benefits from liberalization even if comparative statics show positive re
sults, “since in discounting streams of costs and benefits for welfare calcula
tions, the near-present counts more heavily than ‘the long run’” (Baldwin et al.
1980, p. 407). But there is hardly any work on developing countries in the CGE
tradition specifying the time path of adjustment to a new equilibrium, assessing
the adjustment costs throughout, and measuring them against comparative stat
ics benefits. There is a tendency to underplay adjustment costs on grounds that
they are relatively small compared to (potential) benefits from reallocation of
resources, as maintained in a recent WTO paper: “Although the economy may
be worse off in the short-run, the gains from trade will outweigh short-run ad
justment costs in the medium to long-term ... Existing studies find that the
benefits from trade exceed adjustment costs not only in the long-run where the
cost to benefit ratio is estimated to be lower than 4 per cent, but even during the
adjustment period” (Bacchetta and Jansen 2003, p. 16). However, as the au
thors themselves recognize “measures of adjustment costs in existing empirical
work are crude and imprecise” and “the empirical evidence ... is restricted to
industrialized countries”, and “may not be representative for the case of other
countries” since “the institutional settings and the functioning of domestic mar
kets will affect the size of adjustment costs.”23
Until adjustment paths are properly defined and out-of-equilibrium reactions
and costs are better specified, the assumed one-off benefits of trade liberaliza
tion in developing countries would remain an act of faith, and estimates based
on CGE models would provide little guidance to the impact of liberalization.
But the real question is whether a sound analysis of trade liberalization can be
undertaken in models premised on the neoclassical tradition. As long argued
by Keynesian economists, “comparisons of equilibrium positions one with an
other are not the appropriate tools for the analysis of out-of-equilibrium pro
23 Bacchctta and Jansen (2003, p. 18). According to one of the few models, again designed for
developed countries, that specifies the adjustment path between steady states and accounts for
training costs of unemployed labour, aggregate adjustment costs could reach 90 percent of the
comparative statics gains from freer trade even under modest assumptions regarding training
costs and the discount factor; sec Davidson and Matusz (2001).
39
cesses or changes, and ... that the neoclassical procedure is singularly ill-equipped
to cope with the problem of time” (Harcourt 1972, p. 5). This is also reflected
by a sloppy use of the concepts of short-term and long-term in the CGE litera
ture. In economic analysis, short-term typically refers to a situation where,
inter alia, resources are fixed while long-term implies capital accumulation,
technical progress and economic growth, none of which really happen in the
CGE models.
(c)
Beyond Adjustment Costs: Industrialization
The key question for developing countries is not what they can gain or lose
from trade liberalization as a result of its one-off effects on the allocation of
existing resources, or even the temporary adjustment costs generated when pass
ing from one resource allocation to another. Rather, it is the implications of
leaving industrial progress, technological upgrading and economic growth to
global market forces dominated by large and mature firms from advanced in
dustrial countries. Even if developing countries could avoid adjustment costs
and instantly benefit from improved allocation of resources and increased ac
cess to markets in industrial countries, these one-off benefits may be quite in
significant compared to longer-term losses that may be incurred as a result of
losing policy space for rapid industrialization.
According to conventional trade theory, under free trade developing countries
with abundant unskilled and semi-skilled labour should specialize in labourintensive activities while industrial countries specialize in skill- and technol
ogy-intensive activities. Thus, with a significant move towards free trade in
industrial products, developing countries would be expected to exit partly or
wholly from skill- and technology-intensive and potentially high value-added
sectors maintained behind tariff and non-tariIT barriers. The same goes for in
dustrial countries for labour-intensive products.
Detailed studies of the sectoral impact of trade liberalization on employment
and production are hard to come by. Nevertheless, as expected, available evi
40
dence from a series of CGE model simulations is quite consistent with the tra
ditional trade theory. According to one such study, the impact of the combined
Uruguay Round liberalization in agriculture, industry and services in both Ja
pan and the United States is to lower output and employment in low valueadded, labour-intensive sectors including textiles, wearing apparel and leather
products, but to raise them in almost all other manufacturing sectors including
transportation equipment, metal products, and machinery and equipment (Brown
et al. 2001, Tables 2-3). The same result is obtained in a simulation of a post
Uruguay Round liberalization scenario of one-third cuts in industrial tariffs as
well as similar reductions in barriers to trade in agriculture and services (tables
7 and 8). A simulation of a free trade agreement among ASEAN plus 3 (China,
Japan and Korea) shows employment losses for Japan in textiles, wearing ap
parel and leather products, and gains in all other manufacturing sectors. For
China, there are sizeable losses in chemical industry, metal products, transpor
tation equipment, machinery and equipment, but gains in almost all the sectors
in which Japan experiences losses (pp. 24-25). These are also broadly con
firmed by the results obtained from a CGE model simulating the sectoral im
pact of various formulas proposed for cuts in industrial tariffs in the negotia
tions for NAMA (Fernandez de Cordoba et al. 2004a).
Thus, the mainstream theory and the CGE models tell us that at the current
levels of technological capability, firms in developing countries cannot com
pete with those in advanced industrial countries in skill- and technology-inten
sive products so that any rapid move towards free trade implies that developing
countries would withdraw their resources from these sectors and redeploy them
to low value-added, resource-based and labour-intensive industries. In other
words, it would have the consequences of establishing an international division
of labour based entirely on static competitive advantages derived from existing
endowments and capabilities. Still, the pace of industrialization and growth in
developing countries would depend on how fast they move away from such a
division of labour by improving their technical and skill endowments and es
tablishing more dynamic and challenging industries that promise higher pro
ductivity and per capita income. In other words, the return on industries that
developing countries would be exiting is potentially greater than industries in
41
which they are expanding. Therefore, the key question is: can developing coun
tries re-enter and successfully establish such high value-added, technologically
dynamic activities over time without being able to provide them infant industry
protection and support because of their commitments in the WTO?
In answering these questions the conventional trade theory and the CGE mod
els are even less useful since "they can't unlock the secrets of economic growth”
(O'Rourke 1995, p. 4). This issue is addressed by another class of ad hoc
models linking trade to growth. However, there is no generally accepted theory
that economies that are more open grow faster. Furthermore, several cross
country studies which show a positive relation between growth and some mea
sure of openness have come under strong criticism because of their method
ological and conceptual weaknesses, as well as their failure to account for cau
sality - that is, whether greater openness causes faster growth or faster growth
allows greater openness. The more recent experience also shows not only that
import liberalization does not guarantee a strong export performance, but also
that improved export performance is not always mirrored by acceleration of
industrialization and growth.24
While evidence on the link between growth and tariffs appears to be mixed for
advanced industrial countries, the relation between the two is generally found
to be positive in developing countries. As already noted, O'Rourke (2000)
finds a positive correlation between tariffs and growth among Western Euro
pean countries and their offshoots during 1875-1914 while Clemens and
Williamson (2001) contend that this was reversed after 1950.25 Edwards (1992)
finds a negative relation between growth and tariffs for a sample of 20 develop
ing countries, but the relationship is statistically insignificant. By contrast
Yenikkaya (2003) provides cross-country evidence for 100 countries that initial
tariffs were positively correlated with subsequent growth during 1970-1997,
24 For a discussion of these issues see Akyiiz (2005).
25 However, as noted above, while tariffs lost their relative importance, other forms of interven
tion, including subsidies, non-tariff barriers and voluntary export restraints, have proliferated in
the post-war era. For an attempt to test possible explanations for this reversal, including wide
spread use of non-tariff barriers, see Clemens and Williamson (2002a).
42
particularly in developing countries. According to a more recent study, the
relationship between tariffs and growth is negative and significant among de
veloped countries but positive among developing countries (DeJong and Ripoll
2005).
These studies all focus on average tariffs while one of the conclusions of the
analysis here is that success in the use of tariffs for industrial development
depends not so much on their average level as their pattern and evolution over
time. Consequently, two countries at the same level of development and with
the same average level of tariffs may obtain different results in industrial devel
opment and economic growth depending, inter alia, on the sectoral profile of
their tariffs. This is all the more important since, as noted, in reality tariffs
often diverge from the pattern maximizing their dynamic economic benefits. In
general success comes where they are designed to protect learning in dynamic
sectors rather than deep-seated inefficiencies or vested interests in sunset in
dustries.26 Indeed, a rational tariff structure based on selective and temporary
protection appears to be one of the factors distinguishing East Asian economies
such as Taiwan and Korea from less successful countries which had similar or
even lower average tariff protection and price ‘distortions.’27 In other words,
given the strong evidence against the orthodox idea of technological leapfrog
ging through big-bang liberalization, “some form of protection for learning is
26 According to a study of 63 countries including several LDCs as well as advanced industrial
countries, the average tariff is uncorrelated with growth, but countries that focus protection in
skilled manufacturing exhibit faster growth than those with higher protection for unskilled manu
facturing (defined to include several industries classified as RL and LT here) (Nunn and Trefler
2004). Higher protection for unskilled industries is associated with rent-seeking behaviour.
While this is highly plausible for advanced and even middle-income countries, higher tariffs on
RL and LT products in low-income countries and LDCs dependent on primary' commodities
cannot always be said to reflect the inability of governments to prevent rent-seeking behaviour,
rather than a rational choice. For the same reason it is doubtful if higher protection for skilled
industries in Bolivia, Ghana and Haiti is more beneficial (or less harmful) to growth.
27 For tariffs and price distortions in East Asia compared to other ‘interventionist’ developing
countries see Bruton (1998, pp. 912 and 924-25), and Wade (2003, p. xix).
43
necessary.... The major policy issue then is to design protection measures that
induce learning rather than the easy life” (Bruton 1998, pp. 930-31).28
To sum, while infant industry protection is no guarantee for successful industri
alization and growth, there is no example of modem industrialization based on
laissez-faire. Consequently, there is little economic rationale for developing
countries to narrow their options to use tariffs for industrialization by agreeing
to bind and lower them significantly in the current negotiations on NAMA.
First, the one-off comparative static benefits that may be brought by the reallo
cation of their existing resources are likely to be small, and there may even be
net losses. Second, they can face high "‘adjustment costs” resulting from dislo
cation and disequilibria generated by trade liberalization, and these costs can in
fact wipe out static benefits. Finally, over the longer term such a commitment
can create serious difficulties in industrial progress and development. Even
when net benefits in the short term are positive, they are unlikely to be large
enough to justify losing policy space and jeopardizing development prospects.
28 It is sometimes argued that because of increased participation of developing countries tn
IPNs organized by TNCs, tariffs are no longer needed for industrial progress. However, work
ing with TNCs promises no more leapfrogging than liberalizing trade or even faster learning
(Bruton 1998, p. 930; Akyiiz 2005, section 6). For instance tariffs on imported skill- and tech
nology-intensive parts and components can help encourage TNCs to contribute to indigenous
learning of national firms by making greater use of local products. This may indeed be an
effective way of increasing the domestic content of production in sectors dominated by TNCs in
view of restrictions placed on domestic content requirements through the TRIMs Agreement.
For a discussion of the impact of the rise of TNCs on the ability of developing countries to
conduct strategic industrial policy see Chang (2003, chap. 7).
44
6 Reciprocity or Unequal Exchange?
This takes us to one of the most contentious issues in the current WTO negotia
tions: the question of reciprocity. The General Agreement did not define or
require reciprocity to be a necessary condition for trade negotiations, as reci
procity was generally seen as a threat to the unconditional and non-discriminatory most-favoured-nation (MFN) principle of the GATT system. It came to be
emphasized with growing economic difficulties during the 1970s and 1980s,
linked to the free ride argument promoted by the opponents of the MFN prin
ciple (UNCTAD 1984, part II, chap. Ill; and Cline 1983). The principle of non
reciprocity for developing countries was first accepted in GATT Article XXXVI.8
which stated that “The developed contracting parties do not expect reciprocity
for commitments made by them in trade negotiations to reduce or remove tar
iffs and other barriers to the trade of less-developed contracting parties.” Both
the Doha Ministerial Declaration and Annex B of the July Package recognized
that in negotiations on NAMA, special needs and interests of developing coun
tries would be fully taken into account in accordance with Article XXVIII bis
of GATT, and explicitly referred to “less than full reciprocity in reduction com
mitments” by developing countries.
As is generally the case with special and differential treatment, there is consid
erable confusion and debate over what “less than full reciprocity” really means.
As discussed in some detail by Khor and Goh (2004, pp. 20-26), much of this
debate has revolved around respective cuts in tariffs by developed and develop
ing countries, obfuscated by various proposals for linear and non-linear formu
las. Since tariffs in developing countries are generally higher than in industrial
countries, an equal percentage reduction would mean that they would be cut-
ting a lot more in percentage points, and this could be so even when they cut by
a smaller rate. If agreement is reached on a Swiss-type “harmonization” or an
EU-United States-type “compression” formula, then developing countries’ tar
iffs would be slashed considerably even when in percentage terms they cut by a
smaller rate than the industrial countries. There are suggestions that such deep
cuts could be balanced by allowing developing countries longer time to imple
ment their commitments, thereby slowing the speed of liberalization.
From an economic point of view, it should be evident that the question of reci
procity should be addressed in terms of distribution of costs and benefits of
trade liberalization among countries, rather than respective cuts in tariffs by
developing and developed countries. This is because the same rate of tariff cut
can generate different costs and benefits according to the circumstances under
which it is implemented. One procedure would be to compare changes in bilat
eral trade balances that may result from the concessions exchanged, an ap
proach advocated by the opponents of the MFN principle in the 1970s and 1980s
(Cline 1983). Less-than-full reciprocity in this context would require that re
spective tariff cuts should result, ceteris paribus, in an improvement in trade
balances of developing countries vis-a-vis industrial countries. Das (1998, pp.
22-23) explains that in previous GATT rounds reciprocity was dealt with on an
offer-request basis through a comparison of tariff revenues lost as a result of
concessions. Yet a more comprehensive assessment could be to go beyond
what may happen to imports and exports and compare and contrast economic
gains resulting from different formulas. In this respect estimates based on CGE
models are of little use for the reasons already explained.
A correct characterization of the nature of ongoing negotiations is essential for
making a sound judgement on what constitutes reciprocity. In industrial coun
tries high tariffs (tariff peaks) are concentrated in labour-intensive manufac
tures of export interest to developing countries (Fernandez de Cordoba et al.
2004b, pp. 4-6). Even though these sectors have long lost their viability in the
face of the emergence of low-cost producers in the developing world, they have
been maintained behind tariffs and non-tariff barriers in large part because of
the failure to upgrade the skill profile of labour and to deal effectively with
46
rising unemployment. Such protection is not temporary since these sectors
have no chance of regaining competitiveness vis-a-vis cheaper producers in the
South, and efficient utilization of global resources would have called for them
to be phased out long ago. Developed countries are in effect offering to cut
such tariffs in return for across-the-board cuts in tariffs by developing coun
tries, including those which have protection for learning in skill- and technol
ogy-intensive industries that should eventually be removed with maturity. In
any case, as recent developments in trade in textile and clothing clearly demon
strate, the value of these offers in practice is much less than what they appear to
be because of the tendency and capacity of major industrial countries to restrict
exports of developing countries through anti-dumping duties, safeguard mea
sures and ‘voluntary’ export restraints.
Such a deal cannot be considered as an equitable exchange regardless of the
percentage cuts involved. In this respect, the proposal made by the Chairman
of the Negotiating Group on Market Access (WTO 2003c, para 9) for zeroing
tariffs in a number of labour-intensive products including textiles and clothing,
footwear and leather goods, as well as in technologically more demanding prod
ucts such as electronics and electrical goods, motor vehicles, parts and compo
nents, is perhaps the most blatant push for an unequal exchange between tariffs
used to protect unviable industries in developed countries and those needed for
infant industry promotion in developing countries.
It is true that not all tariffs in developing countries are forward-looking in the
above sense. In fact, much like the major industrial countries, many middle
income countries have been unable to maintain competitiveness in traditional
labour-intensive sectors in the face of the emergence of cheaper producers.
They have, however, persisted in such sectors behind barriers, in part because
their producers have found it difficult to upgrade and diversify, not only run
ning the risk of being squeezed between the bottom and top ends of the markets
for manufactured exports, but also contributing to a global glut in labour-inten
sive manufactures and the deterioration of manufacturing terms of trade of de
veloping countries (Akyiiz 2003, chap. 2). There is no question that this
situation owes a great deal to shortcomings in policies and institutions in devel
47
oping countries themselves, including failed adjustment programmes designed
to address such shortcomings (UNCTAD 2003. part 2). However, this is no
reason for throwing the baby out with the bathwater and denying these coun
tries the policy tools used widely and with varying degrees of success in the
past by both mature and newly industrialized countries.
48
7 Conclusions: The Way Forward
According to the traditional theory, while opening up to trade is mutually ben
eficial, the distribution of its benefits among trading partners is indeterminate,
susceptible to being influenced, inter alia, by power, intra- or extra-market.
Certainly there is considerable power play in current negotiations in the WTO.
But if these negotiations are to live up to their characterization as a Develop
ment Round, industrial tariff cuts should be so designed as to provide maxi
mum benefits to developing countries. This is not the case with the proposals
put forward by major industrial countries and a different approach would be
needed.
First of all, developed countries should not use tariffs on industrial products of
export interest to developing countries in labour-intensive sectors as bargain
ing chips. As part of unfinished business, such tariffs should be cut to the
maximum degree and as rapidly as possible irrespective of the commitments to
be undertaken by developing countries. UNCTAD (1999, pp. 137-144) esti
mated that rolling back protectionism in this area could create additional export
earnings of up to $700 billion for developing countries, to be realizable over a
10-year period. This is less than 5 per cent of the combined GDP of industrial
countries, but could absorb an important part of unemployed labour in the South
and generate a vent for surplus.
Abolishing tariffs on products of export interest to developing countries would
no doubt lead to preference erosion for several poor countries. In many of these
countries labour-intensive industries are not competitive vis-a-vis low-cost pro
ducers in countries such as China in large part because of lack of skill, technol
ogy and services infrastructure needed to support export activities. This is,
however, no reason to maintain a differentiated system of market access for
such products among developing countries. One way of dealing with this prob
lem is to allow poorer countries to subsidize export industries in these products
on a temporary' basis, until the necessary experience is gained and competitive
ness attained. Given that many of these countries lack resources for such a
subsidy, consideration should be given to establishing a trust fund for the provi
sion of grants to them. The amounts involved are likely to be small and man
ageable.
As for the trading regime for industrial products in developing countries, the
crucial issue is how to reconcile policy flexibility with multilateral discipline.
The proposals on the table leave little room not only because they stipulate
deep cuts in industrial tariffs in developing countries, but also because they
require tariff binding and reduction on a line-by-line basis. Such an approach
could pose a number of problems not only for developing countries but also for
the trading system as a whole.
There are more than 5000 tariff lines in the Harmonized System for customs
tariff classification, organized in 21 sections. Even if commitments were to be
made for broad categories of products rather than for individual tariff lines, an
attempt to shape national commercial policy through multilateral commitments
at the lowest layers of public intervention could pose problems of practicality
and potential conflicts that may surpass even those caused by the proliferation
of conditionalities attached to lending by the Bretton Woods institutions. Such
problems can be particularly serious for developing countries which need much
greater policy flexibility because of their structural rigidities and vulnerability
to external shocks. Greater binding coverage would provide increased predict
ability of market access and greater multilateral discipline only if countries are
able to cope with its consequences. If countries are forced to make commit
ments which they cannot fulfil without suffering from serious disruptions, they
would be inclined to resort to other, less transparent means of import restric
tions thereby creating trade frictions and weakening multilateral discipline.
Recent history of international development policy is replete with examples
50
showing an inverse correlation between proliferation of rules and conditions on
the one hand, and the degree of compliance, on the other, particularly when
rules are set without a full understanding of their consequences.
One way of addressing these difficulties would be to have a fixed life span for
the agreements so that they can be automatically re-negotiatcd after a certain
period. This was the case with almost all free trade agreements of the 19th
century, including the Anglo-French agreement which had a life-time of twenty
years, and was not renewed after its expiration. Granting this option to devel
oping countries would be an appropriate application of the principle of lessthan-full reciprocity, and would represent an important advance over the cur
rent procedures for the re-negotiation of tariffs. Such an approach should be
worth considering on its own merits regardless of how tariffs are bound.
Setting maximum (bound) line-by-line tariffs at sufficiently high levels so as to
accommodate all contingencies would provide considerable flexibility to de
veloping countries, but it would also render multilateral commitments super
fluous. The proposal developed in this paper, binding the average industrial
tariff without line-by-line commitments, does not only have the advantage of
simplicity compared to a complex system of tariff commitments. It would also
reconcile multilateral discipline with policy flexibility since countries would
be subject to an overall average ceiling in setting tariffs for individual products.
Furthermore, for most countries in the early and intermediate stages of indus
trial development, it would result in lower average tariffs than would be the
case under line-by-line commitments. In practice it would have the effect of
balancing tariff increases with reductions; a country would need to lower its
applied tariffs on certain products in order to be able to raise them elsewhere.
This would encourage governments to view tariffs as temporary instruments,
and to make an effort to ensure that they effectively serve the purpose they are
designed for; that is, to provide a breathing space for infant industries before
they mature and catch up with those in more advanced economies.
Developing countries would need to make an assessment of the appropriate
rate and pattern of tariffs that would be compatible with their industrial devel
•51
, o
,
) .
opment, and should not enter into any commitment which is not consistent with
such an assessment. Certainly this would involve technical difficulties and
uncertainties. But it is likely to promise a much better outcome than that im
plied by various ad hoc formulas currently proposed by industrial countries
because, inter alia, it would allow them considerable flexibility in using tariffs
for industrial development. In the final analysis, it is successful industrializa
tion in the developing world that should be expected to lead them to free trade,
not the other way round.
52
REFERENCES
Akyiiz, Yilmaz. (ed.) (2003). Developing Countries and World Trade. Perfor
mance and Prospects. London: Zed Books.
Akyiiz, Yilmaz. (2005). “Trade, Growth and Industrialization: Issues, Experience
and Policy Challenges”. Geneva: TWN.
Akyiiz, Yilmaz, Richard Kozul-Wright, and Joerg Mayer (2004). “Trade and In
dustrial Upgrading: Catching Up and Falling Behind”. In Will Milberg (ed.), Labour
and the Globalisation of Production: Causes and Consecptences of Industrial Up
grading. New York: Palgrave Macmillan.
Anderson, Kym. (2004). “Subsidies and Trade Barriers”. Copenhagen Consensus
Challenge Paper. Centre for International Economic Studies, University ofAdelaide.
Argyrous, George. (1996). “Cumulative Causation and Industrial Evolution: Kaldor’s
Four Stages of Industrialization as an Evolutionary Model”. Journal of Economic
Issues, 30(1): 97-119.
Bacchetta, Marc, and Marion Jansen. (2003). “Adjusting to Trade Liberalization.
The Role of Policy Institutions and WTO Disciplines”. Special Study 7. Geneva:
WTO.
Bairoch, Paul. (1993). Economics & World History. Myths and Paradoxes. Chi
cago: The University of Chicago Press.
Baldwin, Robert, E. John Mutti and J. David Richardson. (1980). “Welfare Effects
on the United States of a Significant Tariff Reduction”. Journal of International
Economics, 10:405-23.
Brown, Drusilla, Alan Deardorff, and Robert Stem. (2001). “CGE Modelling and
Analysis of Multilateral and Regional Negotiating Options”. University of Michi
gan School of Public Policy Research Seminar in International Economics. Discus
sion Paper 468.
Bruton, Henry J. (1998). “A Reconsideration of Import Substitution”. Journal of
Economic Literature, 36:903-36.
53
Chang, I la-Joon. (2002). Kicking Away the Ladder. Development Strategy in His
torical Perspective. London: Anthem Press.
Chang, Ha-Joon. (2003). Globalization. Economic Development and the Role of
the State. London: Anthem Press.
Clemens, Michael A., and Jeffrey G. Williamson. (2001). “A Tariff-Growth Para
dox? Protection's Impact the World Around 1875-1997". NBER Working Paper
8459.
Clemens, Michael A., and Jeffrey G. Williamson. (2002a). “Why Did the TariffGrowth Correlation Reverse after 1950?” NBER Working Paper 9181.
Clemens, Michael A., and Jeffrey G. Williamson. (2002b). “Closed Jaguar, Open
Dragon: Comparing Tariffs in Latin America and Asia Before World War II”. NBER
Working Paper 9401.
Cline, William R. (1983). “Reciprocity: A New Approach to World Trade Policy”.
In W. R. Cline (ed.), Trade Policy in the 1980s. Washington D.C.: Institute for
International Economics.
Das, Bhagirath L. (1998). An Introduction to the WTO Agreements. Penang, Ma
laysia: Third World Network.
Das, Bhagirath L. (2005). “NAM A Negotiations in the WTO: Binding of Tariff and
Tariff Reduction Process”. Geneva: Third World Network.
Davidson, C., and S. Matusz. (2001). “On Adjustment Costs”. GEP Research Pa
per 24. University of Nottingham.
DeJong, David N., and Marla Ripoil. (2005). “Tariffs and Growth: An Empirical
Exploration of Contingent Relationships”. University of Pittsburgh.
Dorman, Peter. (2001). “The Free Trade Magic”. Briefing Paper. Economy Policy
Institute.
Edwards, Sebastian. (1992). “Trade Orientation, Distortions, and Growth in Devel
oping Countries”. Journal of Development Economics, 39: 31-57.
54
Fernandez de Cordoba, S. F., and David Vanzetti. (2005). “Now What? Searching
for a Solution to the WTO Industrial Tariff Negotiations”. Geneva: UNCTAD.
Fernandez de Cordoba, S. E, Sam Laird and David Vanzetti. (2004a). “Trick or
Threat? Development opportunities and challenges in the WTO negotiations on
industrial tariffs”. Geneva: UNCTAD.
Fernandez de Cordoba, S. F, Sam Laird and David Vanzetti. (2004b). “Blend it like
Beckham - Trying to read the ball in the WTO negotiations on industrial tariffs”.
Geneva: UNCTAD.
Gawande, Kishore, and Pravin Krishna. (2003). “The Political Economy of Trade
Policy: Empirical Approaches”. In James Harrigan and E. Kwan Choi (eds.), Hand
book of International Trade. Basil Blackwell.
Gomory, Ralph E., and William Baumol. (2000). Global Trade and Conflicting
National Interests. Cambridge, MA: MIT Press.
Harcourt, Geoff C. (1972). Some Cambridge Controversies in the Theory ofCapi
tal. Cambridge: Cambridge University Press.
Hicks, John. (1965). Capital and Growth. Oxford: Clarendon Press.
Hufbauer, Gary C. (1983). “Subsidy Issues after the Tokyo Round”. In W. R. Cline
(ed.), Trade Policy in the 1980s. Washington D.C.: Institute for International Eco
nomics.
Imbs, Jean, and Romain Wacziarg. (2000). “Stages of Diversification”. CEPR Dis
cussion Paper 2642.
Irwin, Douglas A. (2000). “Tariffsand Growth in Late Nineteenth Century America”.
NBER Working Paper 7639.
Irwin, Douglas A. (2002a). “Interpreting the Tariff-Growth Correlation of the Late
Nineteenth Century”. NBER Working Paper 8739.
Irwin, Douglas A. (2002b). “Did Import Substitution Policies Promote Growth in
the Late Nineteenth Century?” NBER Working Paper 8751.
55
Irwin. Douglas A. (2003). “New Estimates of the Average Tariff of the United
States, 1790 -1820”. NBER Working Paper 96\6.
Kaldor. Nicholas. (1966). Causes ofSlow Rate ofEconomic Growth of the United
Kingdom. Cambridge: Cambridge University Press.
Khor, Martin, and Goh Chien Yen. (2004). “The WTO Negotiations on Non-Agricultural Market Access: A Development Perspective”. Geneva: TWN.
Kiray, Emine. (1990). “Turkish Debt and Conditionality in Historical Perspective:
A Comparison of the 1980s with the 1860s”. In T. Aricanli and D. Rodrik (eds.),
The Political Economy of Turkey. Debt, Adjustment and Sustainability. Hong Kong:
Macmillan.
Laird, S., S. F. Fernandez de Cordoba, and D. Vanzetti. (2003). “Market Access
Proposals for Non-Agricultural Products". Geneva: UNCTAD.
Maddison, Angus. (2001). The World Economy. A Millennial Perspective. Paris:
OECD.
Martin, W., and A. L. Winters (eds.) (1996). The Uruguay Round and the Develop
ing Countries. Cambridge: Cambridge University Press.
Nunn, Nathan, and Daniel Trefler. (2004). “The Political Economy of Tariffs and
Growth”. University of Toronto.
O’Rourke, Kevin. (1995). “Computable General Equilibrium Models and Economic
History”. Dublin: University College, www.gams.com/solvers/mspge.
O’Rourke, Kevin. (2000). “Tariffs and Growth in the Late 19th Century”. Eco
nomic Journal, 10(4): 456-83.
O’Rourke, Kevin, and Jeffrey G. Williamson. (2000). Globalization and Economic
History. The Evolution ofNineteenth Century Atlantic Economy. Cambridge, MA:
The MIT Press.
Prebisch, Raul. (1950). “The Economic Development of Latin America and Its
Principal Problems”. Economic Bulletin for Latin America, 7. New York: UN.
Shutt, Harry D. (1985). The Myth of Free Trade. Oxford: Basil Blackwell.
56
Singer, Hans W. (1950). “The Distribution of Gains between Investing and Bor
rowing Countries”. American Economic Review, 15: 473-85.
Stanford, James. (1993). “Continental Economic Integration: Modelling the Im
pact on Labor”. Annals of the Academy of Political and Social Science. 526: 92110.
Thissen, Mark. (1998). “A Classification of Empirical CGE Modelling”. SOM
Research Report 99C01. University of Groningen.
UNCTAD (various issues). Trade and Development Report. Geneva: United Na
tions.
UNCTAD (2003). Back to Basics: Market Access Issues in the Doha Agenda.
UNCTAD/ DITC/TAB/Misc.9. Geneva.
Vernon, Raymond. (1966). “International Investment and International Trade in
the Product Cycle”. The Quarterly Journal ofEconomics, 80: 190-207.
Wade, Robert. (2003). Governing the Market. Princeton University Press. Second
Edition.
Williamson, Jeffrey. (2003). “Was It Stolper-Samuelson, Infant Industry or Some
thing Else? World Tariffs 1789-1938”. NBER Working Paper 9656.
World Bank (1993). The East Asian Miracle. Economic Growth and Public Policy.
New York: Oxford University Press.
World Bank (2004). Global Economic Prospects. Washington, D.C.
WTO (2002). “WTO Members’ Tariff Profiles - Note by the Secretariat - Revi
sion”. TN/MA/S/4/Rev. 1. Geneva: WTO.
WTO (2003a). “Negotiation Group on Market Access - Overview of Proposals
Submitted”. N/MA/6/Rev.l. Geneva: WTO.
WTO (2003b). “Formula Approaches to Tariff Negotiations - Note by the Secre
tariat - Revision”. TN/MA/S/3/Rev.2. Geneva: WTO.
57
WTO (2003c). “Draft Elements of Modalities for Negotiations on Non-Agricultural Products'. TN/MA/W/35 and TN/MA/W/35/rev. 1. Geneva: WTO.
WTO (2004a). “Coherence in Global Economic Policymaking and Cooperation
between the WTO, the IMF and the World Bank". WT/TF/COH/S/7. Geneva: WTO.
WTO (2004b). “Doha Work Programme”. Draft General Council Decision of 31
July 2004. WT/GC/W/535. Geneva: WTO.
Yenikkaya, Halit. (2003). “Trade Openness and Economic Growth: A Cross-Coun
try Empirical Investigation”. Journal of Development Economics, 72: 57-89.
Young, Allyn. (1928). “Increasing Returns and Economic Progress”. Economic
Journal, 38(152): 527-42.
58
Titles in the TWN Trade & Development Series
No. 1
From Marrakesh to Singapore: The WTO and Developing Countries
by Magda Shahin
(48 pages
US$6.00)
No. 2
The WTO and the Proposed Multilateral Investment Agreement: Implications for
Developing Countries and Proposed Positions by Martin Khor
(40 pages
USS 6.00)
No. 3
Some Key Issues Relating to the WTO by Bhagirath Lal Das
(40 pages
US$6.00)
No. 4
The New Issues and Developing Countries by Chakravarthi Raghavan
(48 pages
US$6.00)
No. 5
Trade and Environment in the WTO: A Review of its Initial Work and Future
Prospects by Magda Shahin
(68 pages
US$6.00)
No. 6
Globalisation: The Past in our Present by Deepak Nayyar
(40 pages
US$6.00)
No. 7
The Implementation of the WTO Multilateral Trade Agreements, the ‘Built-In’
Agenda, New Issues, and the Developing Countries byXiaobing Tang
(68 pages
US$6.00)
No. 8
Strengthening Developing Countries in the WTO by Bhagirath Lal Das
(48 pages
US$6.00)
No. 9
The World Trade Organization and its Dispute Settlement System: Tilting the
Balance Against the South by Chakravarthi Raghavan
(48 pages
US$6.00)
No. 10 Negotiations on Agriculture and Services in the WTO: Suggestions for
Modalities/Guidelines by Bhagirath Lal Das
(24 pages
US$6.00)
No. 11
The Implications of the New Issues in the WTO by Bhagirath Lal Das
(20 pages
US$6.00)
59
No. 12
Developing Countries, the WTO and a New Round: A Perspective
by Ransford Smith
(40pages
USS6.00)
No. 13
Review of the TRIPS Agreement: Fostering the Transfer of Technology to
Developing Countries by Carlos Correa
(48 pages
US$6.00)
No. 14 The Proposed New Issues in the WTO and the Interests of Developing
Countries by Martin Khor
(32 pages
US$6.00)
No. 15 WTO: Challenges for Developing Countries in the Near Future
by Bhagirath Lal Das
(24 pages
US$6.00)
No. 16 Dangers of Negotiating Investment and Competition Rules in the WTO
by Bhagirath Lal Das
(32 pages
US$6.00)
No. 17 WTO Agreement on Agriculture: Deficiencies and Proposals for Change
by Bhagirath Lal Das
(28 pages
US$6.00)
No. 18 Some Suggestions for Modalities in Agriculture Negotiations
by Bhagirath Lal Das
(24 pages
US$6.00)
No. 19 The WTO Agriculture Agreement: Features, Effects, Negotiations, and
Suggested Changes by Martin Khor
(48 pages
US$6.00)
No. 20 Market Access for Non-Agricultural Products: A Development View of the
Principles and Modalities by Martin Khor & Goh Chien Yen
(32 pages
US$6.00)
No. 21
60
Financial Effects of Foreign Direct Investment in the Context of a Possible
WTO Agreement on Investment by David Woodward
(28 pages
US$6.00)
No. 22
Implementation Issues Again Off WTO Radar Screens?
by Chakravarthi Raghavan
(28 pages
USS6.00)
No. 23
Effects of Agricultural Liberalisation: Experiences of Rural Producers in
Developing Countries by Meenakshi Raman
(40 pages
US$6.00)
No. 24 The WTO Negotiations on Industrial Tariffs: What is at Stake for Developing
Countries by Yilmaz Akyiiz
(68 pages
US$6.00)
61
The WTO Negotiations on Industrial Tariffs:
What is at Stake for Developing Countries?
This booklet focuses on the implications of the negotiations on industrial tariffs
taking place at the World Trade Organization for longer term industrialization in
developing countries. It gives a brief overview of the framework on non-agricultural market access (NAMA) and provides a review of the historical experience of
today's advanced countries regarding the use of tariffs in the course of their indus
trialization. and compares and contrasts it with the actual situation prevailing in
developing countries today and the proposals put forward.
The booklet also discusses the sectoral pattern and evolution of tariffs that may be
needed in the course of industrial development in comparison with the constraints
that would result from the proposals made by developed countries, and advances a
simple alternative formula that can help reconcile policy flexibility with multilat
eral discipline.
It makes an evaluation of various estimates of benefits of tariff cuts to developing
countries and in another chapter deals with the question of reciprocity from a broad
developmental perspective.
The booklet concludes with a brief summary of the key points on how the negotia
tions could accommodate both the immediate needs and longer-term interests of
developing countries.
DR. YILMAZ AKYUZ was the Director of the Division on Globalization and
Development Strategies and Chief Economist at UNCTAD until his retirement in
August 2003. He was the principal author and head of the team preparing the
UNCTAD Trade and Development Report, and UNCTAD coordinator of research
support to developing countries (the Group of 24) in the IMF and the World Bank
on international monetary and financial issues. Before joining UNCTAD in 1984
he taught at various universities in Turkey, England and elsewhere in Europe. He
has published extensively in macroeconomics, finance, growth and development.
Since retiring from UNCTAD, he has held the Tun Ismail Ali Chair in Monetary
and Financial Economics, University of Malaya. His current activities include policy
research for international organizations, and advising governments on development
policy issues, and TWN on research in trade, finance and development.
TWN TRADE & DEVELOPMENT SERIES
is a scries of papers published by Third World Network on trade and de
velopment issues that are of public concern, particularly in the South. The
series is aimed at generating discussion and contributing to the advance
ment of appropriate development policies oriented towards fulfilling human
needs, social equity and environmental sustainability.
Position: 2614 (2 views)