trade creation, trade diversion, and non-members …...trade creation, trade diversion, and...

Trade Creation, Trade Diversion, and

Non-Members of Free Trade Areas

Robert WaschikLa Trobe University

Victoria, Australia

May 2005

Abstract

Many studies of Free Trade Areas (FTAs) focus on the detailed effects of suchAgreements on FTA members, but often use only aggregate measures of trade di-version to evaluate the effects of FTAs on non-members. The issue of whether FTAscan serve as stepping stones or stumbling blocks to attainment of Global Free Tradeis strongly influenced by the effects of FTAs on non-members. We use a multi-sectormulti-country Numerical General Equilibrium to endogenously determine the tar-iffs that FTA members would need to charge on non-member trade to maintainFTA-member imports from non-members at their pre-FTA level. This enables usto decompose the effects of FTAs into changes due to trade creation and trade di-version, and allows us to report these changes at the sectoral level. These detailedsectoral measures of trade creation and trade diversion give a much clearer image ofthe effects of an FTA, particularly on non-members. We then illustrate this methodof decomposing trade creation and diversion using our NGE model to offer somepredictions of the effects of the Australia-U.S. FTA, and an FTA between Australiaand China. Results of such experiments are relevant to the current debate under theWTO’s Doha Round of trade negotiations over GATT (1994) Article XXIV whichevaluates the consistency of FTAs with the WTO.

J.E.L. Classification Codes: F15, F13, C68Keywords: trade creation, trade diversion, free trade areas, Article XXIV

Correspondence to:Robert WaschikDepartment of Economics and FinanceLa Trobe UniversityVictoria 3086 AustraliaTel.: (+61 3) 9479 5701Fax: (+61 3) 9479 1654E-Mail: [email protected]

1 Introduction

While there has been and continues to be substantial liberalization of trade under the

auspices of the General Agreement on Tariffs and Trade (GATT) and the World Trade

Organization (WTO), more and more countries are pursuing policies of trade liberal-

ization through Preferential Trade Agreements (PTAs). There continues to be serious

debate over the issue of whether such PTAs promote deeper liberalization of trade than

would be possible through the GATT/WTO and are therefore beneficial, or whether such

PTAs generate welfare gains for PTA members at the expense of non-members, ultimately

forming stumbling blocks on the path to global free trade.

In evaluating whether PTAs are welfare-improving or welfare reducing, theoretical and

applied models typically use the notions of trade creation and trade diversion described in

Viner (1950).1 Detailed and disaggregrated measures of trade creation and diversion are

routinely used to evaluate whether PTA members benefit by joining a proposed PTA. But

when it comes to the effects of a PTA on non-members, many studies ignore the impact of

a PTA on non-members, and those that do consider the effects on non-members typically

present only aggregate effects. The objective of this paper is to focus particularly on

the effects of PTAs on non-members. We propose a specific and sectorally-disaggregated

measure of trade creation and trade diversion, and demonstrate how such a measure could

be implemented in a Numerical General Equilibrium (NGE) model of PTA formation.

The plan of the paper is as follows: Section 2 gives a very brief review of the theoretical

and applied literature on trade creation and trade diversion,2 to motivate the importance

of a detailed focus on the effects of PTAs on non-members, and to describe the theoretical

basis for our method of solving for trade creation and trade diversion. Section 3 describes

the dataset and NGE used to implement our measures of trade creation and trade diver-

sion. To show how these measure could be used in practice, we describe the effects of two

PTAs in Section 4: The recently-completed Australia-US Free Trade Agreement (AUS-

FTA), and a potential Free Trade Agreement between Australia and China. Concluding

1As noted in Panagariya (2000), Kowalczyk (1990) suggests decomposing the effects of PTAs intoterms-of-trade and volume-of-trade effects. Harrison et al (1993) provide an alternative welfare decom-position of the formation of CUs. Nevertheless, most of the literature evaluates PTAs in terms of tradecreation and trade diversion.

2For a more detailed literature review, the reader is directed to any of a number of surveys of variousapproaches to the study of PTAs, including Panagariya (1999) and (2000), Adams et al (2003), DeRosa(1998), Harrison, Rutherford and Tarr (2003), Robinson and Thierfelder (2002), Scollay and Gilbert(2000), and Lloyd and MacLaren (2004).

1

comments are offered in Section 5.

2 Trade Creation and Trade Diversion

The seminal treatment of PTAs is Viner (1950), which shows that the formation of a PTA

can yield gains due to the creation of trade between PTA members as trade barriers on

intra-PTA trade are reduced. But these gains may be offset due to trade diversion as some

of the increase in intra-PTA trade can come about at the expense of imports from more

efficient extra-PTA suppliers. This work has been extended to incorporate models with

increasing returns and product differentiation (see Krugman (1991), Bond and Syropoulos

(1996), among many others), political economy aspects of PTA formation (Krishna (1998)

and Grossman and Helpman (1994), for example), and strategic investment aspects of

PTA formation (Freund (2000)). While all of these studies argue that trade creation and

diversion are important in understanding which and how PTAs form, the effects of PTA

formation are typically ambiguous, since the relative size of trade creation and diversion

is in general indeterminate. The welfare effects of PTA formation and the size of trade

creation and diversion are an empirical issue.

The most common empirical tool used to estimate the effects of PTAs in ex post stud-

ies is the gravity equation.3 Adams et al (2003), Cernat (2001), Haveman, Nair-Reichert

and Thursby (2003), Kreinin and Plummer (1998), Krueger (1999), and Linnemann and

Verbruggen (1991) are just some of the studies which have used a gravity equation to

estimate trade creation and diversion in the NAFTA, MERCOSUR, the Australia-New

Zealand CER, ASEAN, various PTAs between primarily developing countries, and the

Generalized System of Preferences (GSP) of 1989 between developed and developing coun-

tries. But many gravity models suffer from mis-specification and omitted variable bias,

since relative price changes are often omitted from such regressions, and models are often

estimated using very aggregated data. As noted in Krueger (1999), it is very difficult

to control for changes in all relevant variables to isolate the effects of trade creation and

diversion. Most importantly, the ex post studies available often reach conflicting conclu-

sions, with some studies arguing that the NAFTA (for example) caused significant trade

3Recent studies by Clausing (2001) and Trefler (2004) present econometric estimates of the effects ofthe Canada-US Free Trade Agreement (CUSFTA) and do not use a gravity equation to derive results.But these papers present the effect of the CUSFTA on non-member countries only at the most aggregatelevel.

2

diversion while others argue that NAFTA did not divert trade from non-members to any

significant extent.

Numerical General Equilibrium (NGE) models have become popular tools for analyz-

ing the implications of formation or accession to PTAs (see Brown et al (2003), Harrison,

Rutherford and Tarr (2001), (1997), Rutherford, Rutstrom and Tarr (1997), and Scollay

and Gilbert (2000), among many others). The problem with such ex ante studies of PTA

formation is that they do not give reliable estimates of trade creation and trade diver-

sion. The counterfactual analysis used implies that PTA imports from non-PTA countries

are affected by changes in all variables in the model. Changes in trade volumes include

both trade creation and trade diversion, so welfare effects (for both PTA members and

non-members) must arise due to both trade creation and trade diversion. As a result,

it is problematic to argue that any increases in intra-PTA trade are due to pure trade

creation, or to changes in income within the PTA, changes in relative prices between

PTA and non-PTA members due to the formation of the PTA (trade diversion), or other

relative price or income changes in the new counterfactual equilibrium.

PTAs are necessarily discriminatory since trade barriers are reduced or removed only

on intra-PTA trade, implying that cet.par., relative prices of goods imported from outside

the PTA will rise. PTAs are therefore inconsistent with Article I (MFN) of the GATT,

which requires that nations treat all trading partners as they do their most-favoured

trading partner. Since the inception of the GATT, PTAs have been legitimate exceptions

to MFN through Article XXIV, which aims to ensure that as a result of formation of the

PTA:

• PTA members remove barriers on substantially all trade between PTA members

over a reasonable period of time, and

• duties set by PTA members on non-member trade are not higher or more restrictive

than prior to the formation of the PTA

Art.XXIV has proven to be an unreliable tool to evaluate PTAs. Many authors have

argued that Art.XXIV is too vague and imprecise to be useful in evaluating PTAs, and

Working Parties of the GATT/WTO typically fail to reach a conclusion as to whether or

not a given PTA is consistent with the GATT/WTO.4

4As an example, paragraph 98 the Report of the Working Party on the Canada-US Free Trade Agree-

3

An important criticism of Article XXIV is McMillan (1993), who argues that Article

XXIV focuses on the wrong measures of the effects of PTAs. By concentrating on the

level of duties set by PTA members on non-member trade, Article XXIV targets a vari-

able which only indirectly affects welfare through its effect on trade volumes. McMillan

proposes a superior indicator of the effects of a PTA which targets trade volumes with

non-members: A PTA would be consistent with the GATT/WTO if it did not result in

any reduction in trade between the PTA and non-members. As noted by McMillan, this

rule for admissibility of a PTA has a strong theoretical foundation in the work of Vanek

(1965) and Kemp and Wan (1976). Such a rule would ensure that PTAs result in no

trade diversion, implying that PTAs themselves could not pose a stumbling block to the

attainment of global free trade.

The objective of this paper is to demonstrate a mechanism which implements the

Kemp-Vanek admissibility criterion described in McMillan (1993), using a NGE model of

PTA formation. In implementing this admissibility criterion, we will generate measures

of trade creation and trade diversion due to formation of a PTA which are as exact as

possible. As opposed to the typical NGE models of PTAs, which report results that arise

due to both trade creation and diversion, we isolate trade creation by implementing a

model with endogenous trade taxes which adjust to keep PTA member imports from non-

members at their initial level. This allows us to present a detailed sectoral measure of

trade diversion due to a proposed PTA, namely the extent to which tariffs on non-member

imports must be reduced so a proposed PTA does not harm non-members.

2.1 Modelling Kemp-Vanek Admissibility

In order to implement the Kemp-Vanek admissibility criterion described in McMillan

(1993), we need to incorporate a mechanism into our PTA experiment which ensures that

the volume of imports by members of a PTA from non-members does not change when

the PTA is formed. We isolate the effects of trade creation and eliminate trade diversion

by introducing into our NGE model an endogenous tariff on each commodity imported

by each PTA member from each non-member. When we model the effects of a PTA using

our NGE model, this endogenous tariff will adjust to keep imports by PTA members from

ment (see General Agreement on Tariffs and Trade (1992:76)) notes that: “At the conclusion of itsexamination of the Agreement, the Working Party ... was unable to reach agreed conclusions as to theconsistency of the provisions of the Agreement with the General Agreement.”

4

non-members at their pre-PTA (benchmark) levels.5 Any welfare changes or changes in

trade volumes will be due only to trade creation. If we then run the same PTA experiment

without endogenous tariffs, we get welfare and trade volume changes which are due to both

trade creation and trade diversion. By comparing these counterfactuals with and without

endogenous tariffs, we can identify the changes in welfare and trade volumes which are

attributable solely to trade creation and those attributable solely to trade diversion.

To address the issue of Kemp-Vanek admissibility, we consider a further measure of

trade diversion: The difference between the baseline or benchmark tariff between PTA

members and non-members, and the endogenous tariff which eliminates trade diversion.

If this difference is zero (or very close to zero), then the proposed PTA should not result

in significant trade diversion. In this case, Art. XXIV as it is currently written should

be sufficient to ensure that the PTA is compatible with the WTO. But if the difference

between the initial and endogenous tariffs is large, then the tariff on non-member imports

needs to be reduced to eliminate trade diversion. Art. XXIV will be inadequate in

protecting non-members from the trade diverting effects of the PTA. As opposed to a

gross welfare measure of trade diversion, this measure of trade diversion as the difference

between baseline and endogenous tariffs satisfies the criteria raised by MacMillan (1993)

that PTA’s be evaluated at as disaggregated a level as possible.

We demonstrate the use of endogenous tariffs to isolate trade creation and eliminate

trade diversion using a NGE model to illustrate the effects of the recently-adopted AUS-

FTA, and a potential Free Trade Area between Australia and China. We then use our

measures of trade creation and trade diversion to comment on the Kemp-Vanek admissi-

bility of these FTA’s, and to quantify the result (noted in Lloyd and MacLaren (2004:448),

among others) that trade diversion may be either beneficial or harmful to PTA members,

though it is necessarily harmful to non-members. Results suggest that the scope for trade

diversion, due to the liberalization of merchandise trade under this Agreement, is limited.

We then contrast these results of the AUSFTA with a Free Trade Agreement which would,

additionally, liberalize merchandise trade between Australia and China. Results suggest

that such an Agreement between Australia and China would yield welfare gains to Aus-

tralia which are an order of magnitude greater than those which can be anticipated from

5We focus only on the case where PTA members form a Free Trade Area and do not choose a CET.This eliminates the need to motivate the CU’s choice of a CET, an issue deemed to be beyond the scopeof this paper. For more on the issue of the CU’s choice of a CET, see Melatos and Woodland (2003). SeeOhyama (2002) and Panagariya and Krishna (2002) for the extension of the Kemp-Wan theorem to thecase of FTA’s.

5

the liberalization of merchandise trade under the AUSFTA, though an Australia-China

FTA would result in trade diversion in a number of sectors.

Many authors have noted how trade creation and diversion are affected by character-

istics of such NGE models. For example, Rutherford et al (1997) show how trade creation

will be greater the more elastic is a PTA member’s import demand curve. Trade creation

will also be greater the higher is the initial tariff removed under the PTA. A higher substi-

tution elasticity between imports from different sources effectively benchmarks the model

to a higher import demand elasticity, resulting in more trade creation and more trade

diversion.6 To show how trade creation and diversion are affected by these parameters,

we conduct sensitivity analysis on the import demand elasticity. An important result

which should be relevant to negotiations over the reform of Art.XXIV is that our sectoral

measure of trade diversion is quite robust to different specifications of import demand

elasticities. Sensitivity analysis on the parameter used to control the import demand

elasticity shows that in most sectors, the reduction in the tariff on non-member imports

by PTA members needed to eliminate trade diversion is very insensitive to a quadrupling

of the import demand elasticity. But we begin with a detailed description of the NGE

model.

3 Numerical General Equilibrium Model

This section describes the Numerical General Equilibrium model and Benchmark Equi-

librium Data Set (BEDS). The dataset is an aggregated version of the beta release of

version 6 of the GTAP dataset, described in Hertel (1995), which represents a world trad-

ing equilibrium in the year 2001. The model has ten regions: Australia (AUS), the United

States (USA), China (CHI), and seven other regions, detailed in Table 1.7 In each region,

final goods are produced using four primary inputs—labour, capital, land, and natural

resources—and intermediate inputs. The primary factors land and natural resources are

specific factors, while capital and labour are perfectly mobile between production sectors.

6It is important to note that while both trade diversion and trade creation can be expected to risewhen the import demand elasticities are increased, their net effect on welfare will be ambiguous. Forexample, CIE (2001:98) note that a “high Armington elasticity will see lower allocative efficiency gains.”This will not necessarily be true, as demonstrated in our example of an Australia-China FTA in Table 7below.

7The regional aggregation in Table 1 is the same as that used in Centre for International Economics(2001), to facilitate comparison of our results with those in that study.

6

The production side of the economy is disaggregated so that 57 final goods are produced.8

As argued in McMillan (1993), higher levels of disaggregation are desirable, since more

disaggregated statistics will give more reliable measures of the impact of a PTA on trade

flows with non-members.9 Each region produces, imports and exports all goods, so there

is cross-hauling in all goods. Trade is accommodated using the so-called Armington as-

sumption, so that the same good produced in different regions are imperfect substitutes

for one another. The commodity disaggregation, along with some summary production

and trade statistics, is displayed in Table 2.10

The final consumption goods in any region are consumed by a representative consumer

in each region who owns all primary factors of production, and supplies all labour, capital,

land, and natural resources to the production sector. All primary factors are completely

immobile between regions. There are a number of distortions in the initial equilibrium

data set, including production taxes, export taxes and import tariffs.

For each sector i, finished goods yi are produced using intermediate inputs from each

sector j xij and primary inputs: Land Hi, natural resources Ri, labour Li, and capital

Ki. We assume that production technologies display constant returns to scale, and are

represented by nested CES production functions of the form:

yi =

∑

j∈GOODSδjx

γi−1

γi

ij + δV Vγi−1

γi

i

γiγi−1

where Vi =

[

αHHρi−1

ρi

i + αRRρi−1

ρi

i + αLLρi−1

ρi

i + αKKρi−1

ρi

i

]

ρiρi−1

, ∀ i

where xij is the amount of good j used in production of good i. The substitution elas-

ticity between primary inputs, ρi, is given in Table 2.11 Intermediate inputs xij and the

aggregate value-added Vi are combined using fixed-coefficients production technology, so

γi → 0 ∀ i. The structure of production is shown in Figure 1.

8This is the highest level of disaggreagation afforded by the GTAP dataset. The Monash modelproduced by the Centre of Policy Studies at Monash University uses a finer level of disaggregation, butprovides benchmarked data only for Australia.

9A series of disputes between the US and EC due to enlargement of the EC in 1973 and in 1986provides an excellent example of how measures of trade creation and diversion are affected by the levelof aggregation. See McMillan (1993:303).

10Import tariffs on service sectors in the GTAP dataset are virtually all equal to zero, and since ourcounterfactuals will only be concerned with the removal of import tariffs, adjustments in service sectorsbetween counterfactuals will be negligible. So to make the Tables more manageable, tabular data includeonly non-service sectors. Complete data and results including all service sectors are available from theauthor on request.

11These central case values for the elasticity of substitution between primary inputs ρi are consistentwith those in version 5.4 of the GTAP database, reported in Hertel (1995).

7

Table 1: Regional Aggregation

Code Title Countries/Regions

aus Australia Australiachn China Chinausa USA United StatesASE ASEAN Indonesia, Malaysia, Philippines, Singapore, Thai-

land, VietnamEU EU Austria, Belgium, Denmark, Finland, France, Ger-

many, UK, Greece, Ireland, Italy, Luxembourg,Netherlands, Portugal, Spain, Sweden

RNA Rest of North Asia Japan, KoreaRoE Rest of Europe Switzerland, Rest of EFTA, Rest of Europe, Bulgaria,

Croatia, Cyprus, Czech Repub, Hungary, Malta,Poland, Romania, Slovakia, Slovenia, Estonia, Latvia,Lithuania

RoN Rest of NAFTA Canada, MexicoSA South America Colombia, Peru, Venezuela, Rest of Andean Pact, Ar-

gentina, Brazil, Uruguay, Rest of South AmericaRoW Rest of World New Zealand, Rest of Oceania, Hong Kong, Taiwan,

Rest of East Asia, Rest of Southeast Asia, Bangladesh,India, Sri Lanka, Rest of South Asia, Rest of NorthAmerica, Chile, Central America, Rest of FTAA, Restof Caribbean, Albania, Russian Federation, Rest ofFormer Soviet Union, Turkey, Rest of Middle East,Morocco, Tunisia, Rest of North Africa, Botswana,South Africa, Rest of South African CU, Malawi,Mozambique, Tanzania, Zambia, Zimbabwe, Rest ofSADC, Madagascar, Uganda, Rest of Sub-SaharanAfrica

All markets are assumed to be perfectly competitive, with free entry and exit of firms,

so economic profits are equal to zero in all industries in equilibrium. Producers take

all output and input prices as given, and these are all normalized to unity in the initial

equilibrium.

The demand side of each economy is represented by a system of demand functions

derived from the solution to the representative consumer’s utility maximization problem

when utility for the representative consumer in region r is represented by a Cobb-Douglas

utility function in each region. Because we employ a Cobb-Douglas utility function the

elasticity of substitution in consumption is equal to one. As an example, the structure of

consumption of agriculture in Australia is described in detail in Figure 2, with a corre-

sponding description of the structure of consumption goods applying for all goods in all

8

Table 2: Summary Production and Trade Statistics from GTAP Data Set

Industry Cap/ Dom/as Share of Lab Imp Import

Value Value of Subst. Subst. IntensityDescription Added Prod’n Elast. Elast. AUS USA China

Paddy rice 0.03 0.03 0.23 2.2 1.15 2.60 0.93Wheat 0.42 0.35 0.23 2.2 0.08 3.95 19.26Cereal grains nec 0.23 0.19 0.23 2.2 0.23 1.36 16.36Vegetables, fruit, nuts 0.58 0.38 0.23 2.2 5.97 16.39 10.88Oil seeds 0.03 0.04 0.23 2.2 10.69 2.07 23.81Sugar cane, sugar beet 0.13 0.08 0.23 2.2 0.05 0.29 0.10Plant-based fibers 0.22 0.26 0.23 2.2 0.17 0.47 29.68Crops nec 0.40 0.31 0.23 2.2 11.55 32.65 23.74Cattle,sheep,goats,horses 0.66 0.52 0.23 2.8 1.95 2.34 3.96Animal products nec 0.26 0.30 0.23 2.8 1.50 4.52 5.45Raw milk 0.29 0.29 0.23 2.2 0.11 0.12 0.08Wool, silk-worm cocoons 0.36 0.29 0.23 2.2 0.24 40.99 11.91Forestry 0.19 0.18 0.20 2.8 1.41 2.52 10.71Fishing 0.18 0.24 0.20 2.8 1.94 90.50 6.74Coal 1.75 1.04 0.20 2.8 0.00 1.08 44.80Oil 0.88 0.52 0.20 2.8 66.82 108.49 49.15Gas 0.39 0.24 0.20 2.8 0.00 391.03 45.38Minerals nec 1.58 1.67 0.20 2.8 1.65 10.55 20.88Meat: cattle,sheep,goats,horse 0.33 0.66 1.12 2.2 0.38 3.02 10.86Meat products nec 0.10 0.35 1.12 2.2 1.71 2.21 12.67Vegetable oils and fats 0.04 0.08 1.12 2.2 43.21 12.49 21.64Dairy products 0.37 0.73 1.12 2.2 2.85 1.84 12.80Processed rice 0.02 0.04 1.12 2.2 9.50 17.28 3.79Sugar 0.14 0.18 1.12 2.2 0.35 15.64 11.43Food products nec 1.06 1.59 1.12 2.2 11.42 6.29 12.45Beverages and tobacco prods 0.55 0.81 1.12 3.1 5.95 5.44 9.67Textiles 0.53 0.76 1.26 2.2 37.18 19.28 29.39Wearing apparel 0.33 0.44 1.26 4.4 36.48 42.61 26.80Leather products 0.08 0.15 1.26 4.4 77.03 179.56 30.37Wood products 0.85 1.15 1.26 2.8 13.09 14.25 23.96Paper products, publishing 1.89 1.97 1.26 1.8 13.73 5.98 14.81Petroleum, coal products 0.12 0.80 1.26 1.9 9.78 5.04 21.54Chemical,rubber,plastic prods 1.73 2.51 1.26 1.9 41.96 11.99 28.48Mineral products nec 0.81 0.98 1.26 2.8 17.77 14.86 15.09Ferrous metals 0.64 1.12 1.26 2.8 11.98 16.27 19.28Metals nec 0.91 1.68 1.26 2.8 8.63 22.84 38.28Metal products 1.16 1.54 1.26 2.8 11.82 7.69 14.21Motor vehicles and parts 1.13 1.61 1.26 5.2 45.71 30.29 28.66Transport equipment nec 0.44 0.61 1.26 5.2 35.52 11.34 53.26Electronic equipment 0.33 0.44 1.26 2.8 230.52 53.18 44.06Machinery and equipment nec 1.55 1.75 1.26 2.8 102.68 24.06 46.02Manufactures nec 0.20 0.24 1.26 2.8 83.03 84.99 21.06

9

Natural Resources CapitalLabourLand

Added

Value

Inputs

Intermediate

of Good i

Output

. . .

γi = 0

ρi

Figure 1: Structure of Production of Output

regions.

Of total output of industry i in region r, some amount is exported, and the remainder

is consumed within region r. Trade must be balanced, and exports of good i by region r

must equal imports of good i by all other regions. Trade in any region can be distorted by

the presence of export taxes/subsidies and import tariffs. Import tariffs are a combination

of tariffs and tariff-equivalents of quantity restrictions on imports. Tariffs for Australia,

China and the US are reported in Table 3.

To help explain the trade creation and trade diversion results in Sections 4.2 and

4.3, it will be useful to identify the source of imports by region. Table 4 reports the

share of imports by region for Australia, China, and the US. Tables 3 and 4 will assist

in identifying sources of trade creation and diversion since we would expect that trade

creation and diversion will be higher:

• the larger the tariff being removed, and

• the larger the share of imports from the region on which tariffs are being removed.

For example, trade creation due to the AUSFTA should be larger if Australia’s benchmark

tariff on motor vehicle imports from the US is 20% rather than 10%. Even if the tariff

on motor vehicle imports is 20%, the scope for trade diversion would be very small if

Australian motor vehicle imports from the US account for only 1% of total Australian

motor vehicle imports, but would be much larger if this share were 20%.

10

Table 3: Import Tariffs (%)

Aus. Tariff on USA Tariff on China Tariff onImports From Imports From Imports From

Code Description USA China Aus. China Aus. USA

pdr Paddy rice 0.00 0.00 4.20 5.98 0.00 1.00wht Wheat 0.00 0.00 2.78 0.00 1.00 1.00gro Cereal grains nec 0.00 0.00 0.39 0.52 90.90 50.26v-f Vegetables, fruit, nuts 1.14 0.45 2.64 2.99 24.36 32.32osd Oil seeds 0.18 2.45 3.27 6.74 9.00 113.43c-b Sugar cane, sugar beet 0.00 0.00 0.00 1.48 0.00 10.00pfb Plant-based fibers 0.00 0.00 0.84 0.06 1.00 1.00ocr Crops nec 0.01 0.00 2.21 2.59 2.74 7.06ctl Cattle,sheep,goats,horses 0.00 0.00 0.00 0.00 4.73 0.50oap Animal products nec 0.00 0.00 0.48 0.32 8.92 6.88rmk Raw milk 0.00 0.00 0.00 0.00 0.00 0.00wol Wool, silk-worm cocoons 0.00 4.68 1.25 0.37 1.00 3.77frs Forestry 0.04 0.27 0.17 0.62 1.21 2.16fsh Fishing 0.10 1.48 0.45 0.49 17.23 13.79coa Coal 0.00 0.00 0.00 0.00 4.50 4.45oil Oil 5.01 5.58 0.00 0.00 0.00 0.00gas Gas 0.00 0.00 0.00 0.00 0.00 6.00omn Minerals nec 0.03 0.23 0.11 0.47 0.02 2.39cmt Meat: cattle,sheep,goats,horse 0.00 0.00 4.51 1.94 13.67 20.67omt Meat products nec 3.41 3.32 1.14 5.04 8.67 10.48vol Vegetable oils and fats 0.04 0.43 4.75 1.70 27.46 24.42mil Dairy products 0.58 3.68 10.13 2.15 22.08 16.56pcr Processed rice 0.00 0.00 7.46 9.22 0.00 1.00sgr Sugar 1.21 0.00 35.03 49.70 19.53 19.84ofd Food products nec 2.52 3.33 2.93 2.86 23.13 20.64b-t Beverages and tobacco prods 6.52 19.03 4.62 2.57 57.33 55.32tex Textiles 9.88 18.11 10.23 7.42 24.78 16.03wap Wearing apparel 16.01 22.38 10.77 9.79 23.14 23.88lea Leather products 8.02 9.39 4.57 14.90 7.33 14.75lum Wood products 4.72 4.78 0.75 0.63 14.67 7.11ppp Paper products, publishing 2.83 4.14 0.16 0.48 12.82 7.90p-c Petroleum, coal products 0.00 0.00 2.10 0.22 7.67 7.65crp Chemical,rubber,plastic prods 2.82 4.51 1.73 3.23 16.59 11.06nmm Mineral products nec 4.54 4.34 3.29 4.66 18.42 15.96i-s Ferrous metals 4.58 4.28 1.27 2.01 4.04 2.75nfm Metals nec 0.98 3.56 0.74 2.17 13.01 4.89fmp Metal products 5.71 5.68 2.43 2.92 12.01 11.32mvh Motor vehicles and parts 11.02 5.45 2.01 1.32 22.51 30.30otn Transport equipment nec 0.25 4.60 0.27 4.82 4.77 3.76ele Electronic equipment 0.35 1.34 0.20 0.43 11.60 10.17ome Machinery and equipment nec 2.93 3.81 1.29 2.32 13.82 13.17omf Manufactures nec 3.73 3.72 0.86 1.52 18.80 14.38

11

Table 4: Import Shares (%)

Aus. Share of USA Share of China Share ofImports From Imports From Imports From

Code Description USA China Aus. China Aus. USA

pdr Paddy rice 1.09 16.02 0.11 6.77 0.27 3.69wht Wheat 0.07 2.19 0.03 0.15 1.12 23.62gro Cereal grains nec 40.01 4.45 0.04 0.17 54.13 1.90v-f Vegetables, fruit, nuts 19.62 8.61 0.63 2.15 1.34 15.12osd Oil seeds 49.18 17.42 15.64 1.28 2.07 37.27c-b Sugar cane, sugar beet 3.97 11.44 0.01 10.90 0.00 0.21pfb Plant-based fibers 3.38 0.58 0.32 0.38 10.29 21.85ocr Crops nec 10.89 2.64 0.39 1.87 0.14 3.53ctl Cattle,sheep,goats,horses 18.61 0.21 1.14 0.05 56.00 7.01oap Animal products nec 15.06 22.49 1.24 21.52 17.48 45.57rmk Raw milk 0.23 1.86 0.02 1.81 0.02 0.22wol Wool, silk-worm cocoons 0.01 0.06 34.23 0.39 80.34 0.06frs Forestry 20.12 5.64 0.82 1.78 0.14 2.23fsh Fishing 1.05 6.86 1.93 2.19 1.46 4.61coa Coal 36.05 2.61 1.24 0.81 32.18 13.09oil Oil 0.61 2.18 0.30 0.67 0.89 0.33gas Gas 0.25 0.63 0.00 0.00 0.00 0.00omn Minerals nec 9.88 10.09 5.21 6.97 25.34 1.57cmt Meat: cattle,sheep,goats,horse 12.52 0.43 33.67 0.04 40.98 14.66omt Meat products nec 5.89 1.55 0.20 1.12 6.38 22.98vol Vegetable oils and fats 0.51 1.97 0.23 0.82 0.09 1.53mil Dairy products 2.14 0.37 3.37 0.03 17.00 9.50pcr Processed rice 2.43 8.65 3.88 12.76 0.02 0.12sgr Sugar 24.19 3.53 5.59 0.06 18.89 3.93ofd Food products nec 11.31 5.33 0.54 4.06 2.61 17.28b-t Beverages and tobacco prods 19.80 3.21 4.22 0.96 1.01 3.20tex Textiles 5.31 32.73 0.21 10.17 0.20 2.67wap Wearing apparel 1.85 66.50 0.09 13.30 0.71 3.56lea Leather products 1.76 54.51 0.12 56.25 1.12 4.10lum Wood products 6.19 15.13 0.10 18.35 1.20 7.75ppp Paper products, publishing 20.02 4.69 0.26 5.46 1.04 14.66p-c Petroleum, coal products 12.53 3.31 0.12 0.19 0.54 1.53crp Chemical,rubber,plastic prods 23.81 6.24 0.41 6.16 0.69 9.98nmm Mineral products nec 24.28 11.47 0.33 15.09 0.26 11.12i-s Ferrous metals 3.35 3.25 1.28 3.07 2.17 5.86nfm Metals nec 2.62 3.55 3.48 1.46 13.49 10.19fmp Metal products 11.73 18.72 0.24 19.19 0.39 10.74mvh Motor vehicles and parts 12.68 0.48 0.31 0.49 0.13 5.41otn Transport equipment nec 52.34 3.13 0.61 3.13 0.15 39.58ele Electronic equipment 18.02 9.00 0.08 12.16 0.11 11.41ome Machinery and equipment nec 25.03 5.68 0.40 9.82 0.42 12.33omf Manufactures nec 13.59 31.30 0.32 37.57 0.87 9.25

12

Agriculture

Imports from China

AgricultureAgriculture

Imports from ROWImports from USA

Agriculture

ImportedAustralian

Agriculture

Agriculture Fishing, Trade and

Transport

Services

Forestry,

and Mining

in Australia

Representative Consumer

. . .

. . .

σ = 1

νagr

τagr

. . .. . .

Figure 2: Structure of Consumption in the BEDS

The nesting structure of traded goods is illustrated in Figure 2. The final good i in

region r is a nested CES aggregate of the domestically produced good i and imported

good i, which itself is a CES aggregate of imports of good i from different regions. The

substitution elasticity between goods imported from different regions (the lowest nest in

Figure 2) is τi. The aggregate imported good is combined with the domestically produced

good with a substitution elasticity given by νi, reported in Table 2.12 In setting the

substitution elasticities between imports from different regions τi, we adopt the method

used in Jomini et al. (1994: 81) and set τi = 2 × νi.13 The size of trade creation and

trade diversion will depend upon these substitution elasticities. To show how sensitive

results are to specification of these elasticities, we conduct sensitivity analysis by repeating

experiments for low (νloi = 0.5×νi) and high (νhi

i = 2×νi) values for these independently

specified parameters.

12These central case values for the elasticity of substitution between domestic and aggregate importedgoods νi are consistent with those in version 5.4 of the GTAP database, reported in Hertel (1995).

13For a discussion of and empirical support for this Rule of Two, see Liu et al (2004).

13

4 Baseline and Counterfactual Scenarios

We now give a description of the baseline scenario to which the FTA counterfactuals are

compared, followed by a more detailed description of how the AUSFTA and the proposed

Australia-China FTA are modelled. In each FTA we use endogenous tariffs to derive

precise measures of trade creation and diversion.

4.1 Description of Baseline Scenario

The benchmark dataset we use represents a world trading equilibrium for the year 2001.

Since Australia has made changes to its trade taxes since 2001, it is important that PTA

results be compared to a baseline which incorporates these changes. We implement the

baseline adopted by the Centre for International Economics study of the effects of the

AUSFTA (see CIE (2001:63) for details). These include a reduction in Australia’s tariff

on:

• imported motor vehicles from 15% to 10%

• imported textiles to 17.5%

• imported clothing to 10%

• imported footwear to 7.5%

4.2 Australia-USA FTA

Since the primary objective of this paper is to model trade creation and trade diversion,

we keep our simulation of the AUSFTA as simple as possible and only model the effects of

liberalization of merchandise trade between Australia and the United States. To this end,

we set all import tariffs on merchandise trade between Australia and the United States

to zero: tmi,s,r = 0, s, r = aus, usa.14 In the 2001 GTAP dataset, tariffs on services

traded between Australia and the United States are equal to zero,15 so we do not model

any liberalization of trade in services.

14Trade in sugar was excluded from the Agreement, so we leave tmsugar,aus,us = 35%, its benchmarklevel in the GTAP dataset.

15In CIE (2001), liberalization of trade in services is modelled as a productivity improvement. CIE(2001) also reflects liberalization in investment as a reduction in the equity risk premium in Australia of5 basis points. Since ours is a static model where the focus is on measuring trade creation and diversion,we do not model investment liberalization. For more detail, see CIE (2001:23-60).

14

The welfare effects of an AUSFTA are reported in Table 5. For the central case

values of the Armington (trade) elasticities (middle panel of Table 5), both Australia

and the US experience small welfare gains: $US 79.6 and 105.2 million in Australia and

the US, respectively,16 with most other regions experiencing very small welfare losses. If

we consider the same AUSFTA experiment using endogenous tariffs to eliminate trade

diversion, both Australia and the US see small decreases in welfare as they reduce tariffs

on non-member imports. Thus the AUSFTA does result in a small amount of trade

diversion, and this trade diversion makes both Australia and the US better off.

16CIE (2001:83) reports a welfare gain to Australia of $AUD 92 million due to merchandise tradeliberalization. The small difference between our results and theirs are due to the different version of theGTAP dataset used in the CIE (2001) study and differences between the respective baseline scenarios.

15

Table 5: Summary Welfare Changes - Australia-US FTA

Low Armington Elasticities: νi = 0.5 ∗ νcentral casei

∆ in Welfare aus chn usa ASE EU RNA RoE RoN SA RoW

Trade Creation %∆ -0.06 0.21 -0.07 0.25 0.01 0.04 0.01 0.17 0.01 0.01Only $US million -223.5 2441.9 -6612.5 1350.9 815.9 1623.4 66.5 2283.5 57.8 473.9

Trade Creation %∆ 0.03 0.01 0.00 0.00 -0.00 -0.00 -0.00 -0.01 -0.00 -0.00and Diversion $US million 91.9 68.2 244.5 13.9 -128.0 -108.5 -0.8 -135.6 -16.0 -58.6

Central Case Armington Elasticities: νi = νcentral casei


Trade Creation %∆ -0.01 0.16 -0.04 0.18 0.01 0.03 0.00 0.06 -0.00 0.01Only $US million -34.4 1824.8 -4024.3 947.0 575.5 1280.8 43.9 819.7 -12.2 209.8

Trade Creation %∆ 0.02 0.01 0.00 0.01 -0.00 -0.00 -0.00 -0.01 -0.00 -0.00and Diversion $US million 79.6 126.3 105.2 46.6 -99.6 -56.7 -1.5 -132.3 -18.5 -62.0

High Armington Elasticities: νi = 2 ∗ νcentral casei


Trade Creation %∆ 0.03 0.17 -0.04 0.18 0.01 0.03 0.00 0.05 -0.00 0.00Only $US million 116.3 1938.0 -4062.2 970.4 557.5 1343.3 38.0 695.5 -30.6 159.7

Trade Creation %∆ 0.00 0.02 -0.00 0.02 -0.00 -0.00 -0.00 -0.01 -0.00 -0.00and Diversion $US million 3.2 225.0 -64.7 96.7 -86.2 -6.9 -5.2 -115.1 -24.5 -68.4

16

Notice that both Australia and the U.S. experience a small decrease in welfare in the

“Trade Creation Only” scenario. In general, trade creation due to accession of a small

open economy to a PTA must result in a welfare gain. However, the Armington structure

employed in our NGE model implies that even small countries will have some market

power (import demand elasticities will be finite). So while the tariff reductions implied

by the AUSFTA will lead to a positive volume of trade effect on welfare, the terms of

trade will change, and the net effect on welfare will be indeterminate, even if trade with

non-members is held constant.

As a region’s import demand becomes more elastic, so the region will have less market

power, and the terms of trade loss due to the removal of import tariffs will get smaller.

Table 5 shows how these welfare results are sensitive to changes in the import demand

elasticities, modelled through changes in the Armington elasticities which control substi-

tution between domestic goods and imports (νi) and substitution between imports from

different regions (τi = 2∗νi). The different panels in Table 5 shows how Australian welfare

due to “Trade Creation Only” rises from -225 million to -34 million to 116 million as the

Armington elasticities are increased from 0.5*νi to νi to 2*νi. Likewise in the US, welfare

due to “Trade Creation Only” is more positive (becomes less negative) for larger import

demand elasticities, though the welfare results in the US are less sensitive to changes in

νi. Interestingly, trade diversion (the difference between the “Trade Creation Only” and

the “Trade Creation and Diversion” figures) decreases as the Armington elasticities are

increased, and actually makes Australia worse off in the “High Elasticity” case. These

results suggest that the liberalization of merchandise trade due to the AUSFTA should

result in small overall welfare gains to Australia and negligible welfare gains (never larger

that one-hundredth of one percent of baseline welfare) in the US.

The difference between the “Trade Creation Only” and the “Trade Creation and Di-

version” rows in Table 5 gives us a specific welfare measure of trade diversion for non-

members. As we would expect, all non-members are worse off due to trade diversion.

Notice that this specific measure of trade diversion is different and in all cases larger

than the aggregate welfare change due to trade creation and diversion. For example, the

ASEAN region (column ASE in Table 5) experiences an overall welfare gain of just over

$46 million due to trade creation and diversion, but would have seen a welfare gain of

$947 million due to trade creation alone, implying a welfare loss due to trade diversion of

17

some $900 million.17

As noted in McMillan (1993), the best measure of the effects of a FTA on non-members

will give results at the sectoral level. We can consider the potential for Kemp-Vanek

admissibility of the AUSFTA using Table 6, which reports the baseline and endogenous

tariffs for Australian and US imports from non-members where the endogenous tariff is

at least 1% (100 basis points) lower than the baseline tariff. For example, to eliminate

trade diversion between Australia and China in the beverages and tobacco sector (b-t),

Australia must reduce its tariff on imports of Chinese beverage and tobacco products by

210 basis points, from 19.0% in the baseline to 16.9% under the AUSFTA.

17In the “trade creation only” scenario, all non-member regions experience a small increase in welfare(other than South America). While the “trade creation only” experiment holds constant Australianand US imports from non-members, non-member imports are allowed to respond in the counterfactual,resulting in very small and typically positive changes in welfare for non-members.

18

Table 6: Baseline and Endogenous Tariffs to Eliminate Trade Diversion -Australia-US FTA

aus chn ASE EU RNA RoE RoN SA RoW

base endog. base endog. base endog. base endog. base endog. base endog. base endog. base endog.

b-t 19.0 16.9 17.0 14.9 9.0 7.1 8.4 6.5 6.9 5.0 3.3 1.4 22.8 20.6 12.5 10.5tex 17.5 15.5 8.9 7.1 9.7 7.9 10.7 8.9 10.0 8.2 6.6 4.9 9.0 7.3 11.5 9.7wap 10.0 8.5 10.0 8.5 10.0 8.6 10.0 8.6 10.0 8.6 10.0 8.6 10.0 8.6 10.0 8.6lea 7.5 6.1 7.5 6.1 7.5 6.2 7.5 6.2 7.5 6.2 6.3 5.0 7.1 5.9 7.5 6.2lum 4.8 3.6 4.8 3.7 4.0 2.9 6.3 5.1 4.5 3.4 2.9 1.7 3.0 1.9 4.3 3.2ppp 4.1 3.1 3.4 2.3 3.2 2.1 3.4 2.4 2.7 1.7 1.5 0.4 1.5 0.5 2.8 1.8crp 4.5 2.9 4.4 2.8 2.0 0.6 5.5 3.9 1.6 0.1 1.6 0.1 3.5 2.0 4.0 2.5nmm 4.3 2.4 4.2 2.3 4.4 2.5 5.3 3.4 3.3 1.5 4.5 2.6 5.3 3.4 3.7 1.9i-s 4.3 3.1 5.0 3.7 4.7 3.5 4.9 3.7 2.7 1.5 3.2 2.0 3.6 2.4 4.4 3.2fmp 5.7 4.2 5.2 3.7 6.4 4.9 7.4 6.0 4.8 3.3 3.8 2.3 3.9 2.5 6.1 4.7mvh 5.4 1.8 10.0 6.2 10.0 6.1 10.0 6.1 10.0 6.2 6.4 2.7 5.9 2.2 10.0 6.1ome 3.8 1.8 4.7 2.6 3.3 1.2 4.2 2.1 1.8 −0.2 2.2 0.2 3.9 1.9 4.2 2.1omf 3.7 2.3 3.4 2.0 2.8 1.4 2.6 1.3 2.7 1.3 2.8 1.5 2.5 1.2 2.6 1.3

usa chn ASE EU RNA RoE RoN SA RoWcmt 1.9 0.6 1.0 −0.4 1.4 0.1 1.2 −0.2 1.3 0.0 0.0 −1.3 4.5 3.1 4.3 2.9

19

In Australia, those sectors where trade diversion is expected to be greatest are those

where Table 6 suggests that tariffs need to be reduced by the most: Beverages and

tobacco, textiles, motor vehicles and parts, and machinery and equipment. Inspection of

Tables 3 and 4 suggests that trade diversion should pose more of a threat in those sectors

where baseline tariffs are relatively high and where benchmark imports, whose trade is

liberalized, form a non-trivial share of total imports. An exception to this is the machinery

and equipment sector, where the baseline tariff on US imports which was removed was

less than 3%. In terms of volume of trade, the only sector where the volume of Australian

imports from non-members are expected to fall by more than 4% are beverages and

tobacco (≈ 4.9%), mineral products (≈ 4.2%), and motor vehicles and parts (≈ 8.9%).18

For the US, the only commodity where import tariffs need to be reduced by more than

1% to eliminate trade diversion is the meat sector, where tariffs need to be reduced by

just under 150 basis points. As shown in Table 4, there are only three sectors where US

imports from Australia represent more that 6% of total US imports: Oil seeds (15.64%),

wool (34.23%) and meat (33.67%). Of these three sectors, baseline tariffs on Australian

imports (reported it Table 3) are highest in the meat sector: 4.51%. In the US, the volume

of imports from non-members falls by less than 1.5% in all sectors except meat (≈ 3.4%).

Even at the sectoral level, the AUSFTA should result in very little trade diversion by the

US. This should not be surprising, since there are only three sectors (oil seeds, wool, and

meat) where imports from Australia represent more than 6% of total imports into the

US, and baseline import tariffs in these three sectors are already very low.

Finally, we note that the tariff changes necessary to eliminate trade diversion are quite

robust to changes in the Armington elasticity. Table A-1 reports these tariff changes

for the low-elasticity and high-elasticity cases. While the trade elasticities between the

low- and high-elasticity cases differ by a factor of four, the US tariff on imports from

non-members never varies by more than 30 basis points, and only in the textiles, wearing

apparel and leather sectors does the Australian endogenous tariff on non-member imports

needed to eliminate trade diversion vary by more that 50 basis points.

18Comprehensive results on changes in the volume of trade are available from the author on request.Volume of trade results reported here are for the “trade creation and diversion” scenario under centralcase elasticities. Of course, in the “trade creation only” scenario, the change in the volume of importsfrom non-members is equal to zero.

20

4.3 Australia-China FTA

Australia and China signed the Australia-China Trade and Economic Framework on 24

October 2003, which included a committment to undertake a study to investigate the

feasibility of a possible Free Trade Agreement between Australia and China. This feasi-

bility study, described in Annex II of the Framework, is to include an assessment of “the

impact of the removal and/or reduction of existing barriers to goods and services trade

and investment” (Department of Foreign Affairs and Trade (2003)). We investigate the

potential for trade creation and diversion due to the removal of barriers to Australia-

China merchandise trade by setting tmi,s,r = 0, s, r = aus, chn. Again, import tariffs on

Australia-China services trade are all equal to zero in the 2001 GTAP dataset, and to

keep the model simple, we use a static model and ignore any effects due to liberalization

of investment. Welfare results of this experiment are reported in Table 7. Note that

welfare results reported in Table 7 are relative to the baseline. That is, relative to the

baseline described in section 4.1, Australia sees an improvement in welfare of 0.30% due

to the AUSFTA and an Australia-China FTA. While the AUSFTA results in an increase

in Australian welfare of $US 79.6 million, the AUSFTA and an Australia-China FTA

together would see Australian welfare rise by a further $US 996.5 million (ie: $US 1076.1

million higher than the baseline).

21

Table 7: Summary Welfare Changes - Australia-China FTA



Trade Creation %∆ 0.18 -0.02 -0.01 0.06 0.00 0.01 0.00 0.05 0.00 0.00Only $US million 646.0 -183.6 -1431.4 310.3 223.4 373.8 30.5 648.9 27.0 138.3

Trade Creation %∆ 0.29 -0.11 0.04 -0.20 -0.01 -0.03 -0.00 -0.05 -0.00 -0.01and Diversion $US million 1028.3 -1265.5 3706.9 -1065.0 -707.7 -1493.1 -34.4 -733.3 -2.9 -475.1

Central Case Armington Elasticities: νi = νcentral casei


Trade Creation %∆ 0.24 0.04 -0.02 0.08 0.00 0.01 0.00 0.03 0.00 0.00Only $US million 857.4 428.9 -1857.7 423.0 268.9 543.9 34.6 455.4 14.7 123.2




Trade Creation %∆ 0.38 0.06 -0.02 0.08 0.00 0.01 0.00 0.03 0.00 0.00Only $US million 1363.5 649.6 -1990.0 433.1 237.4 542.3 34.5 422.1 18.4 126.8


22

For central case values of the Armington elasticities, both Australia and China see an

increase in welfare due to trade creation. But while trade diversion due to an Australia-

China FTA makes Australia better off, trade diversion leaves China worse off. This should

not be surprising, since the effects of trade diversion on FTA members is expected to be

larger the higher are the tariffs being removed, and as is clear from Table 3, China’s

benchmark tariffs on Australian imports are considerably higher than Australia’s tariffs

on imports from China. The US (still a member of the AUSFTA) sees small welfare losses

due to trade creation and somewhat larger welfare gains due to trade diversion.

Of course, these welfare changes are sensitive to specification of the import demand

elasticity. Trade creation results in a welfare gain for both Australia and China (except

under the low-elasticity case where China experiences a very small welfare loss due to

trade creation) which is always more positive the higher is the Armington elasticity. Trade

diversion always leaves Australia better off, and this welfare gain due to trade diversion

is larger the higher is the import demand elasticity. But trade diversion always leaves

China worse off: China gains by reducing tariffs on imports from non-members. As was

the case under the AUSFTA, welfare changes for the US are again relatively robust to

changes in the Armington elasticities.

Compared to the AUSFTA, an Australia-China FTA results in much larger welfare

gains for Australia, with overall welfare gains always at least an order of magnitude larger

regardless of the size of the Armington elasticities.

Welfare losses of non-members due to trade diversion (the difference between the

“Trade Creation and Diversion” and the “Trade Creation only” rows in Table 7) are never

greater than 0.07% of baseline welfare, except in the ASEAN region, where the welfare

change is -0.13-0.08 = -0.21% relative to baseline welfare. Recall that these welfare effects

are due to an Australia-China FTA with the AUSFTA already in place. So if we compare

the trade diversion results derived from Table 7 with those derived from Table 5, we see

that the Australia-China FTA results in very little additional trade diversion beyond that

impled by the AUSFTA. For example, trade diversion in the ASEAN region due to the

AUSFTA alone in the ASEAN region was 0.01-0.18 = -0.17% of baseline welfare.

Table 8 allows us to consider the potential for Kemp-Vanek admissibility of an Australia-

China FTA in the presence of the AUSFTA. This Table reports the baseline and endoge-

nous tariffs for Australian, Chinese, and US imports from non-members where the en-

23

dogenous tariff is at least 1% lower than the baseline tariff.19 Trade diversion by the US

is again very small: An Australia-China FTA should have very little impact on the effects

of the AUSFTA through the US. But the Australia-China FTA means that Australia will

have to reduce tariffs on non-member trade beyond levels indicated in Table 6 in some

sectors. To eliminate trade diversion on imports of textiles, wearing apparal and leather

products, tariffs on non-member imports need to be reduced by 150-200 basis points under

the AUSFTA. But under the Australia-China FTA, these tariffs on non-member imports

of textiles, wearing apparal and leather need to be reduced by 900, 850 and 600 basis

points, respectively. As indicated by Table 4, Australian imports of these products from

China account for 32.7%, 66.5%, and 54.5% of total Australian imports of textiles, wear-

ing apparal and leather, respectively. The baseline reductions in tariffs on these products

described in Section 4.1 are not sufficient to eliminate trade diversion in Australia due to

an Australia-China FTA.

19Note that the US is a non-member for China, and China is a non-member for the US, but China andthe US are both members of an FTA with Australia.

24

Table 8: Baseline and Endogenous Tariffs to Eliminate Trade Diversion -Australia-China FTA

aus ASE EU RNA RoE RoN SA RoW

base endog. base endog. base endog. base endog. base endog. base endog. base endog.

b-t 17.0 14.5 9.0 6.7 8.4 6.1 6.9 4.6 3.3 1.0 22.8 20.1 12.5 10.1tex 8.9 −0.5 9.7 0.3 10.7 1.2 10.0 0.5 6.6 −2.6 9.0 −0.3 11.5 1.9wap 10.0 2.6 10.0 2.6 10.0 2.7 10.0 2.6 10.0 2.6 10.0 2.6 10.0 2.6lea 7.5 1.6 7.5 1.6 7.5 1.6 7.5 1.6 6.3 0.4 7.1 1.3 7.5 1.6lum 4.8 3.0 4.0 2.3 6.3 4.5 4.5 2.7 2.9 1.1 3.0 1.2 4.3 2.5crp 4.4 2.8 2.0 0.5 5.5 3.9 1.6 0.1 1.6 0.1 3.5 2.0 4.0 2.5nmm 4.2 2.3 4.4 2.5 5.3 3.4 3.3 1.4 4.5 2.6 5.3 3.4 3.7 1.9i-s 5.0 3.9 4.7 3.6 4.9 3.8 2.7 1.7 3.2 2.2 3.6 2.5 4.4 3.4fmp 5.2 3.1 6.4 4.3 7.4 5.4 4.8 2.7 3.8 1.7 3.9 1.9 6.1 4.0mvh 10.0 6.5 10.0 6.4 10.0 6.4 10.0 6.5 6.4 2.9 5.9 2.5 10.0 6.4ome 4.7 2.5 3.3 1.1 4.2 2.0 1.8 −0.3 2.2 0.1 3.9 1.7 4.2 2.0omf 3.4 1.3 2.8 0.7 2.6 0.6 2.7 0.6 2.8 0.8 2.5 0.5 2.6 0.6

chn usa ASE EU RNA RoE RoN SA RoWgro 50.3 4.6 1.0 −29.7 90.7 32.8 1.7 −29.2 0.6 −30.0 91.0 32.9 1.1 −29.7 39.8 −2.8osd 113.4 112.3 113.9 112.8c-b 10.0 8.6 6.3 4.8 0.0 −1.3 0.0 −1.3 0.0 −1.3 0.0 −1.3 0.0 −1.3 9.3 7.9oap 6.9 5.5 11.8 10.3 13.2 11.7 10.5 9.0 11.8 10.3 7.9 6.4 11.8 10.3 9.5 8.0coa 4.5 3.1 4.1 2.7 2.2 0.9 3.5 2.1 0.0 −1.3 0.0 −1.3 0.0 −1.3 4.3 2.9cmt 20.7 14.8 12.2 6.8 21.0 15.2 23.3 17.4 7.0 1.9 11.0 5.6 30.5 24.2 16.3 10.7mil 16.6 10.3 31.0 23.9 12.2 6.1 23.4 16.7 19.8 13.4 12.8 6.8 10.1 4.1 25.7 18.9sgr 19.8 10.6 18.4 9.3 18.5 9.4 19.6 10.4 7.8 −0.5 31.5 21.4 19.9 10.7 18.4 9.3ofd 20.6 19.4 20.9 19.6 20.5 19.3 22.9 21.6 18.1 16.9 24.4 23.1 7.6 6.5 16.2 15.0b-t 55.3 52.3 47.7 44.9 38.8 36.2 46.9 44.1 24.6 22.2 4.0 2.0 41.1 38.4 53.4 50.5nfm 4.9 1.8 4.7 1.6 6.0 2.8 7.3 4.0 3.2 0.1 3.6 0.5 4.2 1.1 4.9 1.8

usa chn ASE EU RNA RoE RoN SA RoWcmt 1.9 0.7 1.0 −0.2 1.4 0.3 1.2 0.0 1.3 0.2 0.0 −1.1 4.5 3.3 4.3 3.2sgr 49.7 47.9

25

China also presents some sectors where trade diversion is expected to be a concern. As

is evident in Table 3, there are a number of sectors where benchmark tariffs imposed by

China on imports from Australia are relatively high, and many of these are sectors where

imports from Australia make up a relatively large share of total Chinese imports. Not

surprisingly, Table 8 suggests a number of sectors where Chinese tariffs on non-member

imports need to be reduced by more than 500 basis points to eliminate trade diversion:

Cereal grains, meat, dairy products, and sugar. Most notable of these is the cereal grains

sector. Chinese benchmark tariffs on imports of Australian cereal grains are just over

90%. Just over 40% of Chinese cereal grains imports are sourced from the EU and the

Rest of NAFTA (virtually all from Canada), and to eliminate trade diversion from these

regions, China would need to reduce tariffs on cereal grain imports from the EU and the

Rest of NAFTA from just over 90% in the benchmark to just over 32%.

With some exceptions, these tariff changes needed to remove trade diversion are again

quite robust to changes in the Armington elasticity. Table A-2 shows how the tariff

reductions needed to eliminate trade diversion in the Australia-China FTA respond to a

quadrupling of the Armington elasticity. In many cases, the endogenous tariffs change very

little, with endogenous tariffs typically less than 50 basis points smaller under the high-

elasticity case compared to the low-elasticity case. But in some sectors, higher Armington

elasticities require a much greater reduction in tariffs. For example, Australian tariffs on

wearing apparal imports from ASEAN need to fall from 10.0% in the benchmark to 3.7%

under the low-elasticity case and 1.5% under the high-elasticity case. Those sectors which

require a much larger tariff reduction to eliminate trade diversion in the high-elasticity case

compared to the low-elasticity case are beverages and tobacco, textiles, wearing apparal,

and leather imports by Australia, and cereal grains, meat, milk, sugar, and beverage and

tobacco imports by China.

Finally, it is worth noting how the aggregate welfare measure and the sectoral measures

of trade diversion give a very different picture of the effects of an Australia-China FTA on

non-members. Trade diversion defined as the difference between the “Trade Creation and

Diversion” and the “Trade Creation Only” welfare measures suggests that an Australia-

China FTA would result in very little trade diversion beyond the (relatively small) trade

diversion implied by the AUSFTA. But if we define trade diversion as the difference

between baseline and endogenous tariffs required to maintain imports from non-members

at their pre-FTA level, a number of sectors are identified as posing a threat of trade

26

diversion to non-members. Thus while an Australia-China FTA would yield much greater

welfare gains to Australia than the AUSFTA, an Australia-China FTA would result in

more trade diversion, and more tariff cuts on non-member imports would be necessary

for an Australia-China FTA to satisfy the Kemp-Vanek admissibility conditions.

5 Conclusion

The continued proliferation of PTAs continues to stimulate debate over whether such

Agreements can promote the attainment of multilateral trade liberalization, or whether

such Agreements present stumbling blocks on the path to free trade. Central to this

debate is the effect that PTAs have on non-member countries. The effects of PTAs

are typically measured in terms of trade creation and trade diversion. The objective of

this paper has been to present and demonstrate a method of separately measuring trade

creation and trade diversion, to provide detailed overall welfare and sectoral measures

of the effects of PTAs, particularly on non-members. We described how to implement

a test for admissibility of PTAs from McMillan (1993), whereby any PTA is set up in

such a way so that the volume of imports by PTA members from non-members remains

unchanged. Such a test ensures that trade diversion due to the PTA is exactly equal

to zero. We implement this admissibility test using a NGE model in which endogenous

tariffs adjust to keep constant PTA-member imports from non-members. To demonstrate

how this admissibility test could be used to judge PTAs, we employed the NGE model to

simulate the effects of the recently-completed Free Trade Agreement between Australia

and the U.S. (AUSFTA), and a potential Free Trade Agreement between Australia and

China.

As has been suggested by other studies, the liberalization of merchandise trade under

the AUSFTA should result in small welfare gains to Australia, negligible welfare gains to

the US, and very small welfare losses to other regions. The AUSFTA should result in a

small amount of trade diversion in only a few sectors in Australia, while in the US, trade

diversion should only be an issue in the meat sector. In any of these sectors, the small

amount of trade diversion could be eliminated with relatively small reductions in tariffs

on non-member imports.

On the other hand, while our NGE model of the liberalization of merchandise trade

through an Australia-China FTA suggested larger welfare gains for Australia and only

27

a slightly larger loss in welfare for non-members due to trade diversion compared to the

AUSFTA, our detailed sectoral measure of trade diversion pointed to potential trade

diversion in a number of sectors where Australian and Chinese tariffs would need to be

significantly reduced to eliminate any negative impact of an Australia-China FTA on

non-members. This sectoral measure of trade diversion, the reduction in PTA member

tariffs on non-member imports needed to hold constant PTA member imports from non-

members, identified trade diversion which was not apparant at the more aggregate level of

national welfare, highlighting the importance of the argument made by McMillan (1993)

that tests for admissibility of PTAs sould be made at as detailed a sectoral level as

possible. This sectoral measure of trade diversion was shown to be quite robust to different

specifications of import demand elasticities.

It would be worth extending our model to incorporate other features of NGE models

of PTAs, including those described in Section 2, and the features of “second-generation”

NGE models, described in Lloyd and MacLaren (2004), for example. If our measure

of trade diversion were also robust to such extensions, it should prove a useful tool to

complement Art.XXIV in evaluating the admissibility of PTAs under the WTO.

References

Adams, R., P. Dee, J. Gali, and G. McGuire (2003), “Trade and Investment Effects of

Preferential Trading Arrangements – Old and New Evidence” Productivity Commis-

sion Staff Working Paper, Canberra, May.

Bond, Eric and Constantinos Syropoulos (1996) “The Size of Trading Blocs, Market Power

and World Welfare Effects” Journal of International Economics 40:3-4, 411-37.

Brown, Drusilla, Alan Deardorff, and Robert Stern (June 2003), “Multilateral, Regional

and Bilateral Trade-Policy Options for the United States and Japan” World Economy

26:6, 803–28.

Centre for International Economics (2001), “Economic Impacts of an Australia-United

States Free Trade Area” Report prepared for the Department of Foreign Affairs.

Centre for International Economics, Canberra.

Cernat, Lucian (2001), “Assessing Regional Trade Arrangements: Are South-South RTAs

28

More Trade Diverting?” Policy Issues in International Trade and Commodities Study

Series No.16 UNCTAD.

Clausing, Kimberley (Aug. 2001), “Trade Creation and Trade Diversion in the Canada–

United States Free Trade Agreement” Canadian Journal of Economics 34:3, 677–96.

Department of Foreign Affairs and Trade (2003), “Australia-China Trade and Economic

Framework” http://www.dfat.gov.au/geo/china/framework/economic_framework.

html.

DeRosa, Dean A. (1998), “Regional Integration Arrangements: Static Economic Theory,

Quantitative Findings and Policy Guidelines” Policy Research Working Paper No.

2007, World Bank.

Freund, Caroline (2000), “Different Paths to Free Trade: The Gains from Regionalism”

Quarterly Journal of Economics 115:4, 1317-41.

General Agreement on Tariffs and Trade (July 1992), Basic Instruments and Selected

Documents: Supplement No. 38 Geneva: GATT.

Grossman, Gene and Elhanan Helpman (1994), “Protection for Sale” American Economic

Review 84:4, 835–50.

Harrison, Glenn, Thomas Rutherford, and David Tarr (Oct. 2003), “Rules of Thumb for

Evaluating Preferential Trading Arrangements: Evidence from Computable General

Equlibrium Assessments” World Bank Working Paper 3149.

Harrison, Glenn, Rutherford, Thomas, and David Tarr (July 2001), “Trade Policy Options

for Chile: The Importance of Market Access” World Bank Working Paper 2634.

Harrison, Glenn, Rutherford, Thomas, and David Tarr (April 1997), “Economic Implica-

tions for Turkey of a Customs Union with the European Union” European Economic

Review 41:3-5, 861–70.

Harrison, G., T. Rutherford and I. Wooton (1993), “An Alternative Welfare Decomposi-

tion for Customs Unions” Canadian Journal of Economics 26:4 961-68.

Haveman, Jon, Usha Nair-Reichert, and Jerry Thursby (May 2003), “How Effective

are Trade Barriers? An Empirical Analysis of Trade Reduction, Diversion, and

29

Compression” Review of Economics and Statistics 85:2, 480–5.

Hertel, Thomas W. (ed.) Global Trade Analysis: Modeling and Applications. Cambridge:

Cambridge University Press, 1995.

Jomini, P.; McDougall, R.; Watts, G.; and Dee, P.S. The SALTER Model of the

World Economy: Model Structure, Database and Parameters. Canberra: Industry

Commission, 1994.

Jones, Kent Albert (1994), Export Restraint and the New Protectionism: The Political

Economy of Discriminatory Trade Restrictions Ann Arbor: University of Michigan

Press.

Kemp, Murray C. and Henry J. Wan (1976) “An Elementary Proposition Concerning the

Formation of Customs Unions” Journal of International Economics 6:1 95–7.

Kowalczyk, C. (1990) “Welfare and Customs Unions” NBER Working Paper 3476.

Kreinin, Mordechai and Michael Plummer (Aug. 1998), “Ex Post Estimates of the Effects

of the European Single Market Programme on the Exports of Developing Countries”

ASEAN Economic Bulletin 15:2, 206–14.

Krishna, P. (1998), “Regionalism and Multilateralism: A Political Economy Approach”

Quarterly Journal of Economics 113:1, 227–51.

Krueger, Anne O. (Dec. 1999), “Trade Creation and Trade Diversion Under NAFTA”

NBER Working Paper 7429.

Krugman, Paul R. (1991), “Is Bilateralism Bad?” in Helpman and Razin (eds.) Interna-

tional Trade and Trade Policy Cambridge, Mass.: MIT Press.

Liu, Jing, Channing Arndt, and Thomas Hertel (Sept. 2004), “Parameter Estimation

and Measures of Fit in a Global, General Equilibrium Model” Journal of Economic

Integration 19:3, 626–49.

Lloyd, Peter J. and Donald MacLaren (Dec. 2004), “Gains and Losses from Regional

Trading Agreements: A Survey” Economic Record 80:251, 445-67.

Linnemann, Hans and Harmen Verbruggen (1991), “GSTP Tariff Reduction and its Effects

30

on South-South Trade in Manufactures” World Development 19:5, 539–51.

McMillan, J. (1993), “Does Regional Integration Foster Open Trade? Economic Theory

and GATT’s Article XXIV” in Anderson, K. and R. Blackhurst (eds.) Regional

Integration and the Global Trading System London: Harvester-Wheatsheaf.

Melatos, M. and A. Woodland (2003) “Pareto Optimal Delegation in Customs Unions”

University of Sydney Working Paper.

Ohyama, M. (2002), “The Economic Significance of the GATT/WTO Rules” In: Wood-

land, A. (ed.) Economic Theory and International Trade: Essays in Honours of Murray

C. Kemp Edward Elfar, Cheltenham.

Panagariya, A. (2000), “Preferential Trade Liberalization: The Traditional Theory and

New Developments” Journal of Economic Literature 38:2 287-331.

Panagariya, A. (1999), “The Regionalism Debate: An Overview” World Economy 22:4,

477–511.

Panagariya, A. and P. Krishna (2002), “On Necessarily Welfare-Enhancing Free Trade

Areas” Journal of International Economics 57: 353-67.

Robinson, Sherman, and Karen Thierfelder (2002), “Trade Liberalisation and Regional

Integration: The Search for Large Numbers” Australian Journal of Agricultural and

Resource Economics 46:4, 585–604.

Rutherford, Thomas, Elisabet Rutstrom and David Tarr (April 1997), “Morocco’s Free

Trade Agreement with the EU: A Quantitative Assessment” Economic Modelling 14:2,

237–69.

Scollay, Robert and John Gilbert (2000), “Measuring the Gains from APEC Trade

Liberalisation: An Overview of NGE Assessments” World Economy 23:3, 175–97.

Trefler, Daniel (Sept. 2004), “The Long and Short of the Canada-U.S. Free Trade

Agreement” American Economic Review 94:4, 870-95.

Vanek, J. (1965), General Equilibrium of International Discrimination: The Case of

Customs Unions Cambridge, MA: Carnegie Endowment for International Peace.

31

Viner, Jacob (1950) The Customs Union Issue New York: Carnegie Endowment for

International Peace.

32

Table A-1: Baseline and Endogenous Tariffs to Eliminate Trade Diversion - Australia-USFTA


aus chn ASE EU RNA RoE RoN SA RoW

base endog. base endog. base endog. base endog. base endog. base endog. base endog. base endog.

b-t 19.0 17.0 17.0 15.0 9.0 7.2 8.4 6.6 6.9 5.2 3.3 1.4 22.8 20.7 12.5 10.7tex 17.5 15.8 8.9 7.4 9.7 8.3 10.7 9.2 10.0 8.5 6.6 5.1 9.0 7.6 11.5 10.0wap 10.0 8.7 10.0 8.8 10.0 8.9 10.0 8.8 10.0 8.9 10.0 8.7 10.0 8.9 10.0 8.8lea 7.5 6.4 7.5 6.4 7.5 6.5 6.3 5.2lum 4.8 3.6 4.8 3.7 4.0 3.0 6.3 5.2 4.5 3.5 2.9 1.7 3.0 1.9 4.3 3.2ppp 4.1 3.0 3.4 2.2 3.2 2.1 3.4 2.3 2.7 1.7 1.5 0.3 1.5 0.5 2.8 1.8crp 4.5 3.0 4.4 2.9 2.0 0.7 5.5 4.0 1.6 0.2 1.6 0.1 3.5 2.1 4.0 2.6nmm 4.3 2.5 4.2 2.3 4.4 2.6 5.3 3.4 3.3 1.6 4.5 2.5 5.3 3.5 3.7 2.0i-s 4.3 3.2 5.0 3.9 4.9 3.9 3.2 2.1 4.4 3.4fmp 5.7 4.2 5.2 3.8 6.4 5.0 7.4 6.0 4.8 3.4 3.8 2.3 3.9 2.6 6.1 4.8mvh 5.4 2.0 10.0 6.4 10.0 6.4 10.0 6.3 10.0 6.5 6.4 2.8 5.9 2.5 10.0 6.4ome 3.8 1.7 4.7 2.6 3.3 1.3 4.2 2.2 1.8 −0.1 2.2 0.1 3.9 1.9 4.2 2.2omf 3.7 2.4 3.4 2.1 2.8 1.5 2.6 1.4 2.7 1.4 2.8 1.5 2.5 1.3 2.6 1.4

usa chn ASE EU RNA RoE RoN SA RoWcmt 1.9 0.3 1.0 −0.6 1.4 −0.1 1.2 −0.3 1.3 −0.2 0.0 −1.6 4.5 3.0 4.3 2.8


aus chn ASE EU RNA RoE RoN SA RoWb-t 19.0 16.6 17.0 14.6 9.0 6.8 8.4 6.2 6.9 4.7 3.3 1.1 22.8 20.2 12.5 10.2tex 17.5 14.9 8.9 6.5 9.7 7.3 10.7 8.3 10.0 7.6 6.6 4.3 9.0 6.7 11.5 9.1wap 10.0 8.0 10.0 8.0 10.0 8.1 10.0 8.2 10.0 8.0 10.0 8.0 10.0 8.1 10.0 8.0lea 7.5 5.5 7.5 5.5 7.5 5.5 7.5 5.5 7.5 5.5 6.3 4.4 7.1 5.2 7.5 5.5lum 4.8 3.5 4.8 3.5 4.0 2.7 6.3 5.0 4.5 3.2 2.9 1.6 3.0 1.7 4.3 3.0ppp 4.1 3.0 3.4 2.2 3.2 2.1 3.4 2.3 2.7 1.6 1.5 0.4 1.5 0.4 2.8 1.7crp 4.5 2.8 4.4 2.7 2.0 0.4 5.5 3.8 1.6 −0.1 1.6 0.0 3.5 1.8 4.0 2.4nmm 4.3 2.3 4.2 2.1 4.4 2.3 5.3 3.2 3.3 1.3 4.5 2.4 5.3 3.3 3.7 1.7i-s 4.3 2.8 5.0 3.5 4.7 3.2 4.9 3.5 2.7 1.3 3.2 1.8 3.6 2.2 4.4 3.0nfm 3.6 2.5 3.0 2.0 3.4 2.4 1.2 0.2fmp 5.7 4.0 5.2 3.5 6.4 4.7 7.4 5.8 4.8 3.1 3.8 2.1 3.9 2.3 6.1 4.4mvh 5.4 1.7 10.0 6.1 10.0 5.8 10.0 5.8 10.0 6.0 6.4 2.5 5.9 2.1 10.0 5.9ome 3.8 1.6 4.7 2.5 3.3 1.1 4.2 2.0 1.8 −0.3 2.2 0.1 3.9 1.7 4.2 2.0omf 3.7 2.1 3.4 1.8 2.8 1.2 2.6 1.1 2.7 1.1 2.8 1.3 2.5 0.9 2.6 1.1

usa chn ASE EU RNA RoE RoN SA RoWcmt 1.9 0.6 1.0 −0.4 1.4 0.1 1.2 −0.2 1.3 0.0 0.0 −1.3 4.5 3.1 4.3 3.0

33

Table A-2: Baseline and Endogenous Tariffs to Eliminate Trade Diversion - Australia-China FTA




b-t 17.0 15.0 9.0 7.2 8.4 6.6 6.9 5.1 3.3 1.5 22.8 20.7 12.5 10.6tex 8.9 1.0 9.7 1.8 10.7 2.6 10.0 2.0 6.6 −1.1 9.0 1.1 11.5 3.4wap 10.0 3.7 10.0 3.7 10.0 3.7 10.0 3.7 10.0 3.6 10.0 3.7 10.0 3.7lea 7.5 2.5 7.5 2.5 7.5 2.5 7.5 2.5 6.3 1.3 7.1 2.2 7.5 2.5lum 4.8 3.3 4.0 2.5 6.3 4.7 4.5 3.0 2.9 1.3 3.0 1.5 4.3 2.7crp 4.4 3.1 2.0 0.8 5.5 4.2 1.6 0.4 1.6 0.4 3.5 2.3 4.0 2.8nmm 4.2 2.5 4.4 2.7 5.3 3.6 3.3 1.7 4.5 2.8 5.3 3.6 3.7 2.1fmp 5.2 3.4 6.4 4.6 7.4 5.6 4.8 3.0 3.8 1.9 3.9 2.1 6.1 4.3mvh 10.0 6.9 10.0 6.8 10.0 6.8 10.0 6.9 6.4 3.3 5.9 2.9 10.0 6.8ome 4.7 2.7 3.3 1.3 4.2 2.2 1.8 −0.1 2.2 0.3 3.9 1.9 4.2 2.2omf 3.4 1.7 2.8 1.1 2.6 0.9 2.7 1.0 2.8 1.1 2.5 0.8 2.6 0.9

chn usa ASE EU RNA RoE RoN SA RoWgro 50.3 17.2 1.0 −21.2 90.7 48.8 1.7 −20.6 0.6 −21.5 91.0 48.9 1.1 −21.2 39.8 9.0c-b 10.0 8.7 6.3 5.0 0.0 −1.2 0.0 −1.2 0.0 −1.2 0.0 −1.2 0.0 −1.2 9.3 8.1oap 6.9 5.5 11.8 10.3 13.2 11.7 10.5 9.0 11.8 10.3 7.9 6.4 11.8 10.3 9.5 8.1coa 4.5 3.2 4.1 2.8 2.2 1.0 3.5 2.2 0.0 −1.2 0.0 −1.3 0.0 −1.2 4.3 3.0cmt 20.7 15.1 12.2 7.0 21.0 15.4 23.3 17.6 7.0 2.1 11.0 5.8 30.5 24.4 16.3 10.9mil 16.6 11.0 31.0 24.6 12.2 6.8 23.4 17.4 19.8 14.0 12.8 7.4 10.1 4.8 25.7 19.6sgr 19.8 11.3 18.4 9.9 18.5 10.0 19.6 11.0 7.8 0.1 31.5 22.0 19.9 11.3 18.4 9.9ofd 20.6 19.6 20.9 19.8 20.5 19.5 22.9 21.8 18.1 17.1 24.4 23.3 16.2 15.2b-t 55.3 53.7 47.7 46.2 38.8 37.4 46.9 45.4 24.6 23.3 4.0 2.9 41.1 39.6 53.4 51.9nfm 4.9 2.1 4.7 1.9 6.0 3.1 7.3 4.4 3.2 0.5 3.6 0.8 4.2 1.4 4.9 2.1

usa chn ASE EU RNA RoE RoN SA RoWcmt 1.9 0.9 1.0 −0.2 1.4 0.3 1.2 0.0 1.3 0.2 0.0 −1.2 4.5 3.3 4.3 3.2

34




b-t 17.0 13.5 9.0 5.8 8.4 5.1 6.9 3.7 3.3 0.2 22.8 19.1 12.5 9.1tex 8.9 −1.7 9.7 −0.9 10.7 0.0 10.0 −0.7 6.6 −3.7 9.0 −1.5 11.5 0.7wap 10.0 1.5 10.0 1.6 10.0 1.7 10.0 1.6 10.0 1.6 10.0 1.7 10.0 1.6lea 7.5 0.8 7.5 0.9 7.5 0.9 7.5 0.9 6.3 −0.2 7.1 0.6 7.5 0.9lum 4.8 2.8 4.0 2.0 6.3 4.3 4.5 2.5 2.9 0.9 3.0 1.0 4.3 2.3crp 4.4 2.6 2.0 0.3 5.5 3.7 1.6 −0.2 1.6 −0.1 3.5 1.7 4.0 2.2nmm 4.2 2.1 4.4 2.4 5.3 3.2 3.3 1.3 4.5 2.5 5.3 3.3 3.7 1.7i-s 5.0 3.8 4.7 3.5 4.9 3.7 2.7 1.5 3.2 2.0 3.6 2.4 4.4 3.2fmp 5.2 2.9 6.4 4.1 7.4 5.2 4.8 2.6 3.8 1.6 3.9 1.7 6.1 3.9mvh 10.0 6.3 10.0 6.0 10.0 6.0 10.0 6.3 6.4 2.8 5.9 2.3 10.0 6.1ome 4.7 2.3 3.3 0.9 4.2 1.8 1.8 −0.5 2.2 −0.1 3.9 1.5 4.2 1.8omf 3.4 1.1 2.8 0.5 2.6 0.4 2.7 0.4 2.8 0.6 2.5 0.2 2.6 0.3

chn usa ASE EU RNA RoE RoN SA RoWgro 50.3 −3.4 1.0 −35.1 90.7 22.8 1.7 −34.5 0.6 −35.3 91.0 22.9 1.1 −35.0 39.8 −10.1v-f 32.3 30.8 23.1 21.6 17.6 16.3 27.5 26.0 2.8 1.6 4.4 3.2 24.7 23.2 28.4 26.9osd 113.4 111.7 28.9 27.8 113.9 112.1c-b 10.0 8.2 6.3 4.5 0.0 −1.7 0.0 −1.6 0.0 −1.7 0.0 −1.7 0.0 −1.6 9.3 7.5ocr 13.2 12.1 19.2 18.1 33.7 32.5 25.6 24.5oap 6.9 5.3 11.8 10.1 13.2 11.4 10.5 8.8 11.8 10.1 7.9 6.2 11.8 10.1 9.5 7.8fsh 13.8 12.6 18.6 17.3 6.9 5.8 14.5 13.3 13.7 12.5 17.1 15.8 19.2 18.0 12.4 11.2coa 4.5 2.8 4.1 2.4 2.2 0.6 3.5 1.9 0.0 −1.6 0.0 −1.6 0.0 −1.5 4.3 2.6cmt 20.7 14.4 12.2 6.5 21.0 14.9 23.3 17.0 7.0 1.6 11.0 5.3 30.5 23.9 16.3 10.4omt 17.0 16.0 16.7 15.7 14.8 13.8 17.2 16.2 14.5 13.5 15.1 14.1mil 16.6 9.1 31.0 22.6 12.2 5.0 23.4 15.5 19.8 12.2 12.8 5.6 10.1 3.1 25.7 17.7sgr 19.8 9.5 18.4 8.2 18.5 8.3 19.6 9.3 7.8 −1.5 31.5 20.2 19.9 9.6 18.4 8.2ofd 20.6 18.7 20.9 18.9 20.5 18.6 22.9 20.9 18.1 16.2 24.4 22.4 7.6 5.9 16.2 14.3b-t 55.3 47.1 47.7 40.0 38.8 31.6 46.9 39.3 24.6 18.1 4.0 −1.4 41.1 33.7 53.4 45.4nfm 4.9 1.1 4.7 1.0 6.0 2.2 7.3 3.4 3.2 −0.5 3.6 −0.1 4.2 0.5 4.9 1.2omf 21.3 20.3 17.0 16.0 20.9 19.9 20.9 19.9 16.7 15.6 14.4 13.4

usa chn ASE EU RNA RoE RoN SA RoWcmt 1.9 0.6 1.0 −0.2 1.4 0.3 1.2 0.0 1.3 0.2 0.0 −1.1 4.5 3.3 4.3 3.2mil 2.2 0.8sgr 49.7 46.9

35