Previous empirical research has assumed that goods trade responds to goods trade preferentialism only, while other forms of preferentialism — such as services trade or investment preferentialism — are irrelevant for goods trade. This article provides novel evidence for the gains from a broader scope of preferentialism (in terms of the number of dimensions covered: goods, services, and investment) at the intensive and extensive country margins of bilateral goods trade.
We would like to thank the editor and three anonymous reviewers for numerous helpful comments on an earlier version of the manuscript. Egger acknowledges funding from GA ČR through grant number P402/12/0982.
Arcand, Jean-Louis, MarceloOlarreaga, and LauraZoratto. 2010. “Weak Governments and Trade Agreements,” unpublished manuscript, University of Geneva.Search in Google Scholar
Baier, Scott L., and Jeffrey H.Bergstrand. 2007. “Do Free Trade Agreements Actually Increase Members’ International Trade?”Journal of International Economics71(1):72–95.10.1016/j.jinteco.2006.02.005Search in Google Scholar
Baier, Scott L., and Jeffrey H.Bergstrand. 2009. “Estimating the Effects of Free Trade Agreements on International Trade Flows using Matching Econometrics.” Journal of International Economics77(1):63–76.10.1016/j.jinteco.2008.09.006Search in Google Scholar
Baldwin, Richard E.1995. “A Domino Theory of Regionalism.” In Expanding Membership in the European Union, edited by R. E.Baldwin, P.Haaparanta, and L.Kiander, 25–53. Cambridge: Cambridge University Press.10.1017/CBO9780511983368.003Search in Google Scholar
Baldwin, Richard E.2011. 21st Century Regionalism: Filling the Gap Between 21st Century Trade and 20th Century Trade Rules, CEPR Policy Insight No. 56, Centre for Economic Policy Research, London.10.2139/ssrn.1869845Search in Google Scholar
Bergstrand, Jeffrey H., and Peter H.Egger. 2011. “The Determinants of BITs,” unpublished manuscript.Search in Google Scholar
Bergstrand, Jeffrey H., Peter H.Egger, and MarioLarch. 2010. “The Timing of Preferential Trade Agreement Membership,” unpublished manuscript.Search in Google Scholar
Blonigen, Bruce, and Ronald B.Davies. 2004. “The Effects of Bilateral Tax Treaties on U.S. FDI Activity.” International Tax and Public Finance11(5):601–22.10.1023/B:ITAX.0000036693.32618.00Search in Google Scholar
Bond, Eric W., and ConstantinosSyropoulos. 1996. “The Size of Trading Blocs. Market Power and World Welfare Effects.”Journal of International Economics40(3–4):411–37.10.1016/0022-1996(95)01412-8Search in Google Scholar
Davies, Ronald B.2004. “Tax Treaties and Foreign Direct Investment: Potential versus Performance.” International Tax and Public Finance11(6):775–802.10.1023/B:ITAX.0000045331.76700.40Search in Google Scholar
Davies, Ronald B., HartmutEgger, and Peter H.Egger. 2010. “Profit Taxation and the Mode of Foreign Market Entry.”Canadian Journal of Economics43(2):704–27.10.1111/j.1540-5982.2010.01590.xSearch in Google Scholar
Davies, Ronald B., Pehr-JohanNorback, and AycaTekin-Koru. 2009. “The Effect of Tax Treaties on Multinational Firms: New Evidence from Microdata.”The World Economy32(1):77–110.10.1111/j.1467-9701.2009.01158.xSearch in Google Scholar
Dorn, Sabrina, and Peter H.Egger. 2011. “On the Distribution of Exchange Rate Regime Treatment Effects on International Trade,” CEPR Discussion Paper no. 8654.Search in Google Scholar
Egger, Peter H., and MarioLarch. 2008. “Interdependent Preferential Trade Agreement Memberships: An Empirical Analysis.” Journal of International Economics76(2):384–99.10.1016/j.jinteco.2008.08.003Search in Google Scholar
Egger, Peter H., MarioLarch, and MichaelPfaffermayr. 2004. “Multilateral Trade and Investment Liberalization: Effects on Welfare and GDP Per Capita Convergence.” Economics Letters84(1):133–40.10.1016/j.econlet.2003.12.020Search in Google Scholar
Egger, Peter H., MarioLarch, and MichaelPfaffermayr. 2007a. On the Welfare Effects of Trade and Investment Liberalization.” European Economic Review51(3):669–94.10.1016/j.euroecorev.2006.03.007Search in Google Scholar
Egger, Peter H., MarioLarch, and MichaelPfaffermayr. 2007b. Bilateral versus Multilateral Trade and Investment Liberalization.” The World Economy30(4):567–96.10.1111/j.1467-9701.2007.01005.xSearch in Google Scholar
Egger, Peter H., MarioLarch, MichaelPfaffermayr, and HannesWinner. 2006. “The Impact of Endogenous Tax Treaties on Foreign Direct Investment: Theory and Empirical Evidence.” Canadian Journal of Economics39(3):901–31.10.1111/j.1540-5982.2006.00375.xSearch in Google Scholar
Egger, Peter H., MarioLarch, and KevinStaub. 2012. “Trade Preferences and Bilateral Trade in Goods and Services: A Structural Approach,” CEPR Discussion Paper no. 9051.Search in Google Scholar
Egger, Peter H., MarioLarch, KevinStaub, and RainerWinkelmann. 2011. “The Trade Effects of Endogenous Preferential Trade Agreements.” American Economic Journal: Economic Policy3(3):113–43.10.1257/pol.3.3.113Search in Google Scholar
Egger, Peter H., and ValeriaMerlo. 2012. “BITs Bite: An Anatomy of the Impact of Bilateral Investment Treaties on Multinational Firms.” Scandinavian Journal of Economics114(4):1240–66.10.1111/j.1467-9442.2012.01729.xSearch in Google Scholar
Ekholm, Karolina, RikardForslid, and James R.Markusen. 2007. “Export-Platform Foreign Direct Investment.”Journal of the European Economic Association5(4):776–95.10.1162/JEEA.2007.5.4.776Search in Google Scholar
Frankel, Jeffrey A., ErnestoStein, and Shang-JinWei. 1995. “Trading Blocs and the Americas: The Natural, the Unnatural, and the Super-Natural.”Journal of Development Economics47(1):61–95.10.1016/0304-3878(95)00005-4Search in Google Scholar
Horn, Henrik, Petros C.Mavroidis, and AndreSapir. 2010. “Beyond the WTO? An Anatomy of EU and US Preferential Trade Agreements.” The World Economy33(11):1565–88.10.1111/j.1467-9701.2010.01273.xSearch in Google Scholar
Huang, Hui, JohnWhalley, and ShunmingZhang. 2009. “Exploring Policy Options in Joint Intertemporalspatial Trade Models using an Incomplete Markets Approach.” Economic Theory41(1):131–45.10.1007/s00199-008-0414-0Search in Google Scholar
Karacaovali, Baybars, and NunoLimao. 2008. “The Clash of Liberalizations: Preferential vs. Multilateral Trade Liberalization in the European Union.” Journal of International Economics74(2):299–372.10.1016/j.jinteco.2007.07.003Search in Google Scholar
Limao, Nuno.2007. “Are Preferential Trade Agreements with Non-Trade Objectives a Block for Multilateral Liberalization?”Review of Economic Studies74(3):821–55.10.1111/j.1467-937X.2007.00456.xSearch in Google Scholar
Limao, Nuno, and Patricia TovarRodriguez. 2011. “Policy Choice: Theory and Evidence from Commitment via International Trade Agreements.”Journal of International Economics85(2):186–205.10.1016/j.jinteco.2011.06.002Search in Google Scholar
Rosenbaum, Paul R., and Donald B.Rubin, 1983. “The Central Role of the Propensity Score in Observational Studies for Causal Effects.” Biometrika70(1):41–55.10.1093/biomet/70.1.41Search in Google Scholar
Sauvant, Karl P., and Lisa E.Sachs. 2009. The Effect of Treaties on Foreign Direct Investment: Bilateral Investment Treaties, Double Taxation Treaties and Investment Flows. New York: Oxford University Press.10.1093/acprof:oso/9780195388534.001.0001Search in Google Scholar
Wooldridge, Jeffrey M.1995. “Selection Corrections for Panel Data Models Under Conditional Mean Independence Assumptions.” Journal of Econometrics68(1):115–32.10.1016/0304-4076(94)01645-GSearch in Google Scholar
Wooldridge, Jeffrey M.2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.Search in Google Scholar
Wooldridge, Jeffrey M.2005. “Simple Solutions to the Initial Conditions Problem in Dynamic, Nonlinear Panel Data Models with Unobserved Heterogeneity.” Journal of Applied Econometrics20(1):39–54.10.1002/jae.770Search in Google Scholar
In general, there is heterogeneity about any type of agreements regarding the extent of market access and the so-called preference margins granted across different activities (goods with goods trade, services with services trade, and investment-related activities with foreign direct investment). We account for such heterogeneity to the extent that we distinguish between agreements that contain goods trade provisions only, ones that capture goods and services trade provisions, ones that capture goods and investment provisions, and ones that capture goods, services, and investment provisions. However, even GTAs with goods provisions only differ with regard to the tariff lines (products) covered and the preference margins granted (i.e. the difference between the applied tariff outside the agreement and the one charged within the agreement). Accounting for the latter type of heterogeneity lies beyond the scope of this article (see Baier and Bergstrand, 2004, 2007, and 2009, for a similar approach in that regard).
Many GTAs include services trade and investment provisions. For instance, this is obvious from a new data-set compiled by the World Trade Organization. In accordance with that data-set, we classify country-pairs to have services trade provisions whenever they appear in a GTA or a separate STA. Similarly, we classify country-pairs to have investment provisions no matter whether they appear in a trade agreement or a bilateral investment agreement.
In the following, we provide a highly selective list of examples of work in economics which focused on the determinants or consequences of just one dimension of preferential market access. Work on the causes of agreements. GTAs-theory: Baldwin (1995, 1997); Bond and Syropoulos (1996); Limao and Tovar Rodriguez (2011); Arcand, Olarreaga, and Zoratto (2010); GTAs-empirics: Magee (2003); Baier and Bergstrand (2004); Bergstrand, Egger, and Larch (2010); STAs-theory: Huang, Whalley, and Zhang (2009); STAs-empirics: Egger and Lanz (2008); Francois and Hoekman (2010); Egger and Wamser (2013); BITs-theory: Egger, Larch, and Pfaffermayr (2007a, b); BITs-empirics: Bergstrand and Egger (2011); DTTs-theory: Davies (2003, 2004); DTTs-empirics: Egger, Larch, Pfaffermayr, and Winner (2006). Work on the consequences of agreements. GTAs-theory: Frankel, Stein, and Wei (1995); Freund (2000); Ornelas (2005a–c); Limao (2007); Karacaovali and Limao (2008); GTAs-empirics: Baier and Bergstrand (2007, 2009); Egger et al. (2011); STAs-theory: Huang, Whalley, and Zhang (2009); STAs-empirics: Egger, Larch, and Staub (2012); BITs-theory: Egger, Larch, and Pfaffermayr (2004, 2007a, b); BITs-empirics: Egger and Merlo (2012); Sauvant and Sachs (2009); DTTs-theory: Davies, Egger, and Egger (2010); DTTs-empirics: Blonigen and Davies (2004); Egger et al. (2006); Davies, Norbäck, and Tekin-Koru (2009).
In principal, all time-invariant variables and their parameters in Table 4 could be thought of as being part of the vector of country-pair fixed effects. However, we report the corresponding parameters for convenience. Moreover, one could allow for continent-specific or even country-specific year effects rather than pooled year effects. With the data at hand, this leads to an extremely flat likelihood function which is difficult to optimize for parameter values. Therefore, we resort to the more parsimonious specification.
The results in Egger, Larch, Staub, and Winkelmann (2011) suggest that these two margins can be analyzed as two separate parts of an integrated model.
In general, endogenous treatment effects’ problems are ones of missing data. In our context, say, a country-pair ij with a scope of preferentialism of unity in period t is only observed with that treatment level. Ideally, we would like to compare this country-pair to itself with another treatment level (say, one with a scope of preferentialism of four) in the same period. However, such a data point does not exist. We can only impute (or estimate) it by finding other country-pairs in the same year which are very similar (ideally, they are identical) to country-pair ij.
Similar to linear regression models, matching based on the propensity score requires that all relevant variables are included in the model (so that the estimates – with matching, including the propensity score – are consistent). Notice that Assumption Al is not testable.
For observations to be comparable, country-pairs with a specific treatment and control level of the scope of preferentialism in a given year t should have overlapping probabilities of having the same scope as the treated. Otherwise, the units would be too dissimilar to infer causal treatment effects from.
Suppose units of observation with specific treatment and control levels of the scope of preferentialism had similar, overlapping probabilities of exhibiting the level of treatment of the treated but the respective determinants of those treatment probabilities would be very dissimilar for units with treatment s versus . Then, similarity in propensity scores would be an artifact and not a compact measure of similarity in the underlying , which they are supposed to be.
In principal, one could allow even for country-specific rather than continent-specific year effects. However, with the data at hand, the model’s objective function is difficult to optimize for parameter values in that case. Therefore, we resort to the more parsimonious specification.
Suppose that the propensity of having any scope other than 0 is zero in some initial period. Then, the level of propensity of having a given scope greater than 0 in a later period does represent not only the propensity of having that state but also the change in propensity relative to the initial period.
Notice that, following Lechner (2001), one may compute average treatment effects — say, and – based on the results reported in Tables 7 and 8, which can be compared to the scope-specific average treatment effects in Tables 5 and 6. In Table 9, we do so for the intensive margin of trade . Besides these scope-dyad-specific ATEs, we report scope-specific average treatment effects in the same table. The latter can be compared with the results on the scope-specific average treatment effects on the intensive margin of exports in Tables 5 and 6. The findings suggest that the average treatment effect of broadening the scope depends on the initial state of liberalization, which is not accounted for in the estimates in Tables 5 and 6.
©2013 by Walter de Gruyter Berlin / Boston