VoxEU Column International trade

What do gravity models tell us about PTAs impact on trade flows: More creation or more diversion?

Global tariff-cutting over the past decade was dominated by preferential trade agreements. These deals stimulate trade among members, but by creating tariff discrimination, they also divert trade from non-members. Recent research suggests that the majority of the world’s preferential deals created more trade than they diverted and so constitute a constructive force in the world economy.

Rapidly expanding trade is one of the pillars of the global rise and spread of economic prosperity in the post-war period. Billions of people would still be in poverty today were it not for the world trading system. A key element in this trade expansion has been the steady reduction of tariffs and other trade barriers. Much of this liberalisation was accomplished by global trade negotiations – like the ongoing ‘Doha round’ talks, and eight other GATT ‘rounds’ that have been completed since WW II. Non-discrimination is a key principle of these global tariff-cutting initiatives; any tariff cutting by, say the EU or US, is automatically extended in a non-discriminatory manner to all WTO members.

More recently, however, a great deal of tariff cutting has taken place within exclusive trade agreements, so called preferential trade agreements. For example, under a bilateral preferential trade agreement, the two nations lower their tariffs to zero but only for imports from each other. Third-nation exporters don’t benefit from the tariff-cutting, so preferential trade agreements create discrimination. As such, preferential liberalisation intrinsically has ambiguous effects on trade; the ‘liberalisation’ part of preferential liberalisation tends to stimulate global trade, but the ‘preferential’ part tends to distort trade. When thinking about the impact of preferential trade agreements, it is clearly important to quantify the amount of trade created and diverted.

Measuring trade creation and diversion

Recent empirical studies, using variants of the gravity model of trade, have concluded that trade creation under preferential trade agreements generally outweighs trade diversion. Contrary to these findings is a recent Australian Productivity Commission (APC) working paper (Adams et al., 2003). The APC study uses a standard gravity model augmented by dummy variables for a number of individual preferential trade agreements. The study reports that a majority of preferential trade agreements in the sample period (1970-1997) were, on balance, trade-diverting. Specifically, based on the sum of the estimated coefficients of the intra-bloc and extra-bloc indicators specified for each PTA, 12 of the 16 preferential trade agreements examined were found to have diverted more trade from non-members than they have created among members. Importantly, the 12 include the larger preferential trade agreements (EU, ASEAN, and NAFTA).

If correct, these findings would suggest that trade negotiators the world over are arduously crafting agreements only to the detriment of the world economy.

More recent work using similar methods but more extensive data – both in terms of country coverage and years covered – overturns the APC finding, concluding that most preferential trade agreements are, on balance, trade-creating.

In an initial pass on the important question, we conducted our own gravity model analysis closely patterned after the APC approach. Due to data limitations we could not replicate the APC econometric method exactly; we therefore used our own data set and employed econometric techniques that, in effect, encompass the APC method. However, our methodology lacks one important element in the APC approach: an index that differentiates preferential trade agreements according to their comprehensiveness. On the other hand, our coverage of merchandise trade flows and preferential trade agreements is more extensive than the APC data set. The DeRosa (2007) data set covers 30 years (1970-1999), 156 countries, and 46 preferential trade agreements, compared to the APC data set which covered 28 years (1970-1997), 116 countries and 16 preferential trade agreements. More recent work by Hufbauer and Schott (2007), which draws on additional econometric work by DeRosa, analyses nine PTA groupings covering approximately 560 individual preferential trade agreements. The Hufbauer and Schott study extends the panel data to a more recent time period (1976-2005) and covers 179 countries.

In Hufbauer and Schott (2007), the usual gravity model analysis is extended to estimate the dollar and percentage impacts of current and potential preferential trade agreements on specific regions and countries in the Asia-Pacific. The cited paper finds that five of the nine analysed PTA-groups are associated with trade diversion in agriculture. However, only the EU and EFTA were found to cause trade diversion in total merchandise trade.

Specifics

DeRosa (2007) specified PTA dummy variables following the standard approach of having the dummy ‘on’ only for those years and bilateral trade flows directly affected by the preferential agreement in question. Like the APC study, his approach additionally includes, for each PTA considered, dummy variables that directly estimate the trade-diversion impact on trade with third nations.

This structure allows the econometric estimates to be translated into percentage terms. Table 1 displays a sample of interesting cases from DeRosa (2007). The results differ sharply from the APC study. The DeRosa estimates show that the world’s major preferential trade agreements (EU, NAFTA, ASEAN, Mercosur and EFTA) are trade-creating, not only between “insiders” but in nearly every case for “outsiders” as well. On the other hand, minor preferential trade agreements are often trade-diverting, even between “insiders”.

Hufbauer and Schott (2007) rely on the Rieder techniques applied to a data set that is updated and expanded. A calculation scheme was devised to translate the gravity model results into estimates of both percentage and dollar impacts, for potential as well as current preferential trade agreements in the Asia-Pacific.

Applying this approach to disaggregated trade flows, we find that trade diversion has been dominant in some goods, but not in others. Trade diversion is pervasive against “outsiders” in agricultural trade. The following preferential trade agreements are estimated to reduce agriculture imports by members from non-members: European Union, down 15%; the EFTA, down 47%; EU FTAs, down 5%; NAFTA, down 10%; and SAFTA, down 26%. Trade diversion is not surprising given the high degree of protection prevalent in agriculture.

To save space, we don’t present the disaggregated results here, but rather focus on the overall results of the disaggregated calculations in Hufbauer and Schott (2007). Table 2 provides estimated impacts of current and potential preferential trade agreements on intra-bloc or “insider” trade in selected countries and regions in the Asia-Pacific.

One of the visionary proposals for world trade is the so-called Free Trade Agreement of the Asia Pacific (FTAAP). According to these results, the FTAAP would augment trade for most Asia-Pacific countries by roughly 50%. This translates into an estimated increase in two-way merchandise trade for the United States of nearly $1.2 trillion, an increase for China of nearly $600 billion, and for Japan nearly $900 billion. The FTAAP, if ever realised, might increase two-way merchandise trade in the region as a whole by $4.8 trillion. Among the major countries, the impact on China would be the smallest in percentage terms and the impact on Japan would be the largest.

Even without the FTAAP, Korea stands to gain significantly from bilateral FTAs. If the Korea-US FTA is ratified, it will increase total Korean two-way trade by an estimated $70 billion (16%). Korea would gain a similar amount ($63 billion) if the Japan-Korea FTA and the Korea-ASEAN FTA are both agreed.

Impact on outsiders

Table 3 is the counterpart to Table 2, but for outsiders; it provides estimates of the impacts of existing and contemplated FTAs on outsiders in the Asia-Pacific region. Table 3 is structured like Table 2, with an important difference. For any country or region, the diversion effects displayed are those caused by all agreements under a column heading (e.g. “in force”) to which the country or region is not a party.

Table 3 indicates that, in the aggregate, there is no decline in any non-member country’s two-way merchandise trade on account of all the preferential trade agreements for which it is an “outsider”. Negative trade diversion no doubt exists for individual commodities (particularly in agriculture, and textiles and clothing), and diversion surely characterises selected preferential trade agreements, but instances of trade diversion are substantially outweighed by episodes of trade creation for “outsiders”. Some of the notable estimates include a projected $18 billion increase in Chinese two-way trade if a US-Japan FTA is enacted, and a projected $260 billion increase in US two-way trade if ASEAN+3 is enacted.
We do not exaggerate the importance of the foregoing estimates, but we believe that the available evidence refutes a common assumption that every PTA inevitably decreases the commerce of “outsiders”. The overall pattern is trade creation vis-à-vis “outsiders”, rather than trade diversion. Richard Pomfret, in a recent column in this blog, presents a qualitative analysis of the likely impact of potential Asian-Pacific preferential trade agreements on “outsiders” and arrives at a similar prediction (Pomfret, 2007). To be sure, when trade diversion occurs, it should be compensated under WTO principles – a practice seldom observed.

Conclusions

The contrarian APC evidence that current preferential trade agreements are predominantly trade-diverting poses a fundamental challenge to the rapid spread of bilateral FTAs and RTAs. However, using up-to-date data and applying similar econometric techniques, we conclude that the majority of preferential trade agreements in force today are, on balance, trade-creating rather than trade-diverting. This finding applies to “outsiders” as well as “insiders”.

We concur with nearly all economists that multilateral liberalisation, under the auspices of the WTO, represents by far the first-best path for the world trading system. But if the first-best cannot be achieved, or can be achieved only slowly, our empirical results lead us to conclude that, on balance, preferential trade agreements represent a constructive force in the world economy.

References

Adams, Richard, Phillipa Dee, Jyothi Goli, and Greg McGuire. 2003. 'The Trade and Investment Effects of Preferential Trading Arrangements—Old and New Evidence'. Staff Working Paper. Canberra: Australia Productivity Commission.
DeRosa, Dean A. 2007. 'The Trade Effects of Preferential Arrangements: New Evidence from the Australia Productivity Commission'. Working Paper Series no: WP 07-1. Washington: Peterson Institute for International Economics.
Greenaway, David, and Chris Milner. 2002. Regionalism and Gravity. Scottish Journal of Political Economy 49: 574-85.
Hufbauer, Gary Clyde, and Jeffrey J. Schott. 2007. Fitting Asia-Pacific Agreements into the WTO system. HEI, SECO, NCCR Conference. Geneva, September 10-12.
Pomfret, Richard. 2007. 'Asian Regionalism: threat to the WTO-based trading system of paper tiger?' VoxEU.org.
Rieder, Roland. 2006. 'Playing Dominoes in Europe: An Empirical Analysis of the Domino Theory for the EU', 1962-2004. HEI Working Paper no: 11/2006. Geneva: Graduate Institute of International Studies.
Soloaga, Isidro, and L. Alan Winters. 2001. Regionalism in the Nineties: What Effect on Trade? North American Journal of Economics and Finance 12: 1-29.
Viner, Jacob. 1950. The Customs Union Issue. New York: Carnegie Endowment for International Peace.

 

Table 1. Sample of Estimated Percentage Effects of PTAs from DeRosa (2007)

(percent change in base level of trade)

PTA Method Intra-Bloc Two-Way Trade Extra-Bloc Imports Extra-Bloc Exports Total Change On Balance                      Creating or Diverting
Major Agreements
European Union Soloaga-Winters 245.6 8.4 32.6 137.2 Creating
European Union Rieder 136.3 8.4 0.0 63.8 Creating
NAFTA Soloaga-Winters 97.4 26.1 -7.3 42.7 Creating
NAFTA Rieder 103.4 32.1 0.0 57.8 Creating
ASEAN Soloaga-Winters 107.5 31.5 49.7 93.4 Creating
ASEAN Rieder 89.6 29.2 0.0 39.3 Creating
Mercosur Soloaga-Winters 76.8 51.0 -12.2 47.8 Creating
Mercosur Rieder 127.0 58.8 0.0 73.7 Creating
EFTA Soloaga-Winters 122.6 7.6 30.2 44.3 Creating
EFTA Rieder 47.7 2.8 0.0 5.6 Creating
Minor Agreements
Canada-Israel Soloaga-Winters -52.8 -18.4 -13.0 -31.4 Diverting
Canada-Israel Rieder 0.0 -17.3 0.0 -17.3 Diverting
EU-Norway Soloaga-Winters -35.6 0.0 -12.0 -26.4 Diverting
EU-Norway Rieder 0.0 0.0 0.0 0.0 Inconclusive
EU-Romania Soloaga-Winters -9.5 0.0 -7.0 -11.1 Diverting
EU-Romania Rieder 0.0 3.6 0.0 3.6 Creating
US-Israel Soloaga-Winters -36.2 -4.8 -13.1 -18.2 Diverting
US-Israel Rieder 0.0 0.0 0.0 0.0 Inconclusive

Source: DeRosa (2007). Calculations based on gravity model estimates of the impacts of major customs unions and free trade agreements on intra-bloc and extra-bloc trade at the 1-digit SITC level.
Notes: For details on the approaches see Soloaga and Winters (2001) and Rieder (2006).

 

Table 2. Average Annual Impacts of Selected PTAs on Intra-Bloc Trade in the Asia-Pacific Region on Total Merchandise Trade (SITC 0 through 9) by Region and Selected Countries

    Selected Asia-Pacific FTAs Combined Major FTAs under Consideration Possible
Region   In Force Signed Under Negotiation US-ASEAN US-Japan ASEAN+3 FTAAP
    Total Exports + Imports
    $ % $ % $ % $            % $ % $ %            $ %            
Asia-Pacific US 492 21.5 83 3.6 61 2.7 151 6.6 301 13.2 0 0.0 1,190 52.1
  China 147 13.2 0 0.0 19 1.7 0 0.0 0 0.0 174 15.7 598 53.9
  Japan 31 3.1 36 3.6 142 14.3 0 0.0 301 30.3 247 24.8 883 88.9
  Korea 2 0.5 71 16.3 63 14.6 0 0.0 0 0.0 94 21.7 245 56.7
  APEC 1,482 18.7 228 2.9 695 8.8 456 5.7 601 7.6 714 9.0 4,839 60.9
Asia ASEAN 248 25.1 35 3.5 379 38.4 311 31.5 0 0.0 203 20.5 644 65.2
  CER 31 14.2 0 0.0 40 18.5 0 0.0 0 0.0 0 0.0 143 66.5
  SAFTA 0 0.0 0 0.0 5 2.4 0 0.0 0 0.0 0 0.0 0 0.0
  Other Asia 113 12.5 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 597 66.1
Pacific NAFTA 906 27.9 83 2.6 61 1.9 151 4.6 301 9.3 0 0.0 1,698 52.4
  Other America 11 2.4 14 2.9 0 0.0 0 0.0 0 0.0 0 0.0 30 6.5
     Exports
    $ % $ % $ % $ % $ % $ % $ %
Asia-Pacific US 229 27.2 45 5.3 25 3.0 64 7.6 112 13.3 0 0.0 496 58.9
  China 121 17.6 0 0.0 13 1.9 0 0.0 0 0.0 95 13.9 377 54.8
  Japan 21 3.7 19 3.3 90 15.8 0 0.0 188 33.2 136 24.0 513 90.4
  Korea 1 0.4 36 15.5 30 12.7 0 0.0 0 0.0 45 19.3 132 57.1
  APEC 741 18.8 117 3.0 347 8.8 227 5.8 301 7.6 357 9.1 2,420 61.5
Asia ASEAN 116 21.7 16 3.0 176 32.9 167 31.2 0 0.0 83 15.5 327 61.1
  CER 11 10.3 0 0.0 18 17.2 0 0.0 0 0.0 0 0.0 68 64.4
  SAFTA 0 0.0 0 0.0 3 2.5 0 0.0 0 0.0 0 0.0 0 0.0
  Other Asia 10 2.4 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 225 52.0
Pacific NAFTA 458 34.2 45 3.3 25 1.9 64 4.8 112 8.4 0 0.0 763 56.9
  Other America 5 2.0 3 1.2 0 0.0 0 0.0 0 0.0 0 0.0 15 5.5
     Imports
    $ % $ % $ % $ % $ % $ % $ %
Asia-Pacific US 262 18.2 38 2.6 36 2.5 87 6.0 188 13.1 0 0.0 694 48.1
  China 26 6.2 0 0.0 6 1.5 0 0.0 0 0.0 79 18.7 222 52.4
  Japan 10 2.3 17 4.0 52 12.2 0 0.0 112 26.3 111 26.0 371 87.0
  Korea 1 0.5 35 17.3 33 16.7 0 0.0 0 0.0 49 24.5 113 56.3
  APEC 742 18.5 111 2.8 349 8.7 229 5.7 301 7.5 357 8.9 2,420 60.3
Asia ASEAN 132 29.1 19 4.1 203 44.9 144 31.9 0 0.0 120 26.4 317 70.0
  CER 20 18.0 0 0.0 22 19.7 0 0.0 0 0.0 0 0.0 75 68.5
  SAFTA 0 0.0 0 0.0 2 2.4 0 0.0 0 0.0 0 0.0 0 0.0
  Other Asia 102 21.7 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 372 79.0
Pacific NAFTA 448 23.5 38 2.0 36 1.9 87 4.6 188 9.9 0 0.0 935 49.2
  Other America 6 2.9 10 5.2 0 0.0 0 0.0 0 0.0 0 0.0 15 7.

Source: Hufbauer and Schott (2007). Calculations based on gravity model estimates of the impacts of major customs unions and free trade agreements on intra-bloc trade by 1-digit SITC categories.

 

Table 3. Trade Diversion - Average Annual Impacts of Selected FTAs in the Asia-Pacific Region on Total Merchandise Trade (SITC 0 through 9) of Non-Member Countries by Region and Selected Countries

    Selected Asia-Pacific FTAs Combined   Major FTAs under Consideration
Region   In Force Signed      Under Negotiation     US-ASEAN US-Japan ASEAN+3
    Non-Member Exports + Imports
    Neg Pos Total Neg Pos Total Neg Pos Total   Neg Pos Total Neg Pos Total Neg Pos Total
Asia-Pacific US 0 287 287 0 286 286 0 756 756   0 0 0 0 0 0 0 263 263
  China 0 105 105 0 146 146 0 255 255   -6 17 11 -10 28 18 0 0 0
  Japan 0 288 288 -49 146 97 0 303 303   -19 52 33 0 0 0 0 0 0
  Korea 0 83 83 0 62 62 0 105 105   -4 10 6 -8 11 3 0 0 0
  APEC -30 1,258 1,227 -97 1,101 1,004 0 2,250 2,250   -62 145 83 -73 141 67 0 501 501
Asia ASEAN -10 116 106 0 124 124 0 174 174   0 0 0 -16 29 13 0 0 0
  CER 0 34 34 0 27 27 0 61 61   -3 4 1 -4 7 3 0 26 26
  SAPTA 0 26 26 0 14 14 0 40 40   -2 3 1 -1 3 2 0 13 13
  Other Asia 0 207 207 0 133 132 0 445 445   -9 19 9 -14 19 5 0 172 172
Pacific NAFTA -19 419 400 -47 459 412 0 898 898   -19 42 23 -21 45 24 0 299 299
  Other America -21 56 34 -19 51 32 -2 50 48   -7 12 4 -8 13 5 0 14 14
    Non-Member Exports
    Neg Pos Total Neg Pos Total Neg Pos Total   Neg Pos Total Neg Pos Total Neg Pos Total
Asia-Pacific US 0 76 76 0 37 37 0 101 101   0 0 0 0 0 0 0 33 33
  China 0 84 84 0 84 84 0 91 91   0 17 17 0 28 28 0 0 0
  Japan 0 217 217 0 146 146 0 143 143   0 52 52 0 0 0 0 0 0
  Korea 0 51 51 0 30 30 0 37 37   0 10 10 0 11 11 0 0 0
  APEC 0 773 773 0 641 641 0 630 630   0 145 145 0 141 141 0 64 64
Asia ASEAN 0 93 93 0 80 80 0 57 57   0 0 0 0 29 29 0 0 0
  CER 0 15 15 0 19 19 0 22 22   0 4 4 0 7 7 0 6 6
  SAPTA 0 14 14 0 12 12 0 10 10   0 3 3 0 3 3 0 2 2
  Other Asia 0 100 100 0 68 68 0 79 79   0 19 19 0 19 19 0 19 19
Pacific NAFTA 0 208 208 0 210 210 0 197 197   0 42 42 0 45 45 0 38 38
  Other America 0 56 56 0 51 51 0 32 32   0 12 12 0 13 13 0 3 3
    Non-Member Imports
    Neg Pos Total Neg Pos Total Neg Pos Total   Neg Pos Total Neg Pos Total Neg Pos Total
Asia-Pacific US 0 211 211 0 249 249 0 655 655   0 0 0 0 0 0 0 230 230
  China 0 22 22 0 62 62 0 164 164   -6 0 -6 -10 0 -10 0 0 0
  Japan 0 71 71 -49 0 -49 0 160 160   -19 0 -19 0 0 0 0 0 0
  Korea 0 32 32 0 32 32 0 68 68   -4 0 -4 -8 0 -8 0 0 0
  APEC -30 485 455 -97 460 363 0 1,620 1,620   -62 0 -62 -73 0 -73 0 436 436
Asia ASEAN -10 24 13 0 45 45 0 117 117   0 0 0 -16 0 -16 0 0 0
  CER 0 19 19 0 8 8 0 39 39   -3 0 -3 -4 0 -4 0 20 20
  SAPTA 0 12 12 0 2 2 0 30 30   -2 0 -2 -1 0 -1 0 12 12
  Other Asia 0 108 108 0 65 65 0 366 366   -9 0 -9 -14 0 -14 0 153 153
Pacific NAFTA -19 211 192 -47 249 202 0 701 701   -19 0 -19 -21 0 -21 0 261 261
  Other America -21 0 -21 -19 0 -19 -2 18 16   -7 0 -7 -8 0 -8 0 10 10

Source: Hufbauer and Schott (2007). Calculations based on gravity model estimates of the impacts of major customs unions and free trade agreements on extra-bloc trade by 1-digit SITC categories.

 

2,625 Reads