WIOD conference, Groningen, 24-26.04.2012 Recent EU Enlargement: The Evolution of Services Trade Costs between EU Members Joseph Francois, Olga Pindyuk www.wiiw.ac.at
motivation 2 Barriers to services trade Cross-border services trade accounts for >20% of global trade - Even more if trade through foreign affiliates is added Regulation driven both by efficiency and equity concerns - Affecting establishment or ongoing operations - Non-discriminatory or discriminatory - Affecting price of services or costs of service providers Bilateral heterogeneity of regulation in each pair of countries matters
motivation 3 EU and services trade EU members have quite heterogeneous services regulation (Kox and Lejour, 2007) Still the most advanced services trade liberalization among existing RTAs (Francois, Hoekman, 2010) Thus we should expect to see positive effect of the EU membership on services trade of new members
motivation 4 Openness in the EU to services
motivation 5 Openness in the EU to services 6 is maximum value of an index, indicating the highest level of regulation
motivation 6 Services exports to the EU, index, 1999=100
data description 7 Geographic structure of EU10 services exports, USD mln
data description 8 Sectoral structure of EU10 services exports, %
literature review 9 Ways to measure services trade barriers “Indirect” modeling (price-cost margins by sector across countries or gravity regressions) - Not possible to attribute price-cost margins or differences in trade volumes to specific trade policies - Data limitations „Direct“ methodology using regulatory indicators of APC, World Bank, OECD (Dee, 2005; Dihel and Shepherd, 2007; Berden, Bergstrand, et al 2009) - Assumption of sample average responsiveness of countries‘ performance to policy settings - Often lack of in-sample variation; insufficient differentiation of services regulation
modeling approach 10 Gravity-based representation of trade Assuming that import values depend on a mix of importer characteristics, exporter characteristics, and bilateral properties, we specify total trade as follows:
modeling approach 11 Estimation approach We group effects as follows: Pair-wise Exporter Importer Time varying Time invariant We only are interested in the first cell. We can isolate this by differencing over means (pairs, exporter/time, importer/time). Difference-in-difference approach (Egger and Pfaffermayr, 2004; Frazer and Biesebroeck, 2007; Hornok, 2009)
modeling approach 12 Estimating approach We estimate a polynomial pair-wise time trend for trade with the EU relative to the general baseline (similar to Francois and Woerz, 2009) This captures changes that are not explained by general exporter/ importer time varying effects, or time-constant effects, but represent pair-wise trends, which might be attributed to regulatory changes in the single market
modeling approach 13 Estimating procedure Regional regressors: - 6 regional pair-wise dummies ( old-old, new-new, old-new, new-old, third-old, third-new) are interacted with time trend, and third degree polynomials are constructed 2 stage selection model estimation procedure : - Heckman selection to account for zero flows – gives inverse Mills ratios (pdf/cdf) to use as an additional regressor in the second stage - Services imports (logs) and regressors are demeaned with respect to time, exporter and importer to isolate time varying pair-wise effects - Clustered errors to deal with remaining heterogeneity. - Separate regressions for 2 periods: 1999-2002 and 2003-2007 - Chi-squared tests for full polynomial expressions Regressions for each services sector
data 14 description Data description Dataset based on the Eurostat, OECD, UN, and IMF data Bilateral services trade flows for 244 reporting countries and 244 partners (plus World) About 20 sectors In total (1995-2009) we have more than 2 mln observations, 18% of observations are missing values, and 35% of observations are zero flows. Mode 1+2 (cross border trade)
results 15 Change in bilateral trade relative to global baseline, %, 1999 to 2002 205 268 200 Transpor 236 245 249 253 260 262 Other BOPS Total t Travel Communications Construction Insurance Financial Computer business A. old_old -2.37% -0.49% -0.49% 5.34% -0.37% -0.42% -0.43% -3.66% -0.21% B. old_new 2.35% -1.21% 7.16% 11.01% -4.51% 43.22% -10.52% 18.29% 1.24% C. new_new -9.92% 7.47% 25.11% 31.26% -58.62% -14.85% 1121.18% 43.45% D. new_old 3.79% -1.90% 4.22% -3.75% 6.35% 5.05% 0.63% 1.97% -1.11% E. third_old -0.07% 0.97% 0.65% 0.98% -0.46% -0.20% -0.37% -0.66% -1.80% F. third_new 1.59% 0.58% -9.23% -8.29% 31.57% -15.67% -46.80% -8.79% -3.01% Observations 8417 5594 4849 2577 2127 2210 1944 2031 4382
results 16 Change in bilateral trade relative to global baseline, %, 2003 to 2007 205 268 200 Transpor 236 245 249 253 260 262 Other Total t Travel Communications Construction Insurance Financial Computer business BOPS A. old_old -3.39% 2.12% -6.18% -2.63% -9.85% 3.08% -3.29% -11.75% -7.16% B. old_new -8.95% 10.08% 2.56% 3.59% 89.65% 18.53% 2.76% 2.17% 129.91% C. new_new 20.32% -14.36% 1.63% 7.45% -19.55% -5.47% -4.97% -76.31% -53.81% D. new_old 5.83% -0.67% 13.31% 5.07% 16.33% -13.50% -1.48% 8.30% 17.64% E. third_old -0.48% -0.61% -0.01% 0.68% 1.38% 1.75% 1.26% -0.12% -0.17% F. third_new 0.42% -0.41% -0.57% -1.88% -7.90% -35.76% -24.33% -7.48% 0.18% Observations 24566 12208 11261 7294 5264 5758 6333 6560 11197
results 17 Main trends in services trade 1999-2002 Accession countries seem to have started liberalizing access to their services markets prior to 2003 - Relatively faster growth in exports from old to new in communications, construction, other business services and, most prominently, financial services Exports in the opposite direction ( new->old ) did not pick up much above the global trendline Likely trade diversion in exports of third and new to new
results 18 Main trends in services trade 2003-2007 Exports from old to new grew faster than the global baseline for most of the sectors - now as well for travel and computer services - dramatic growth in exports of financial services Exports from new to old also grew faster than the global baseline for most of the sectors, apart from insurance and finance Trade diversion continues: - third-new in most sectors - new-new in computer and other business services No strong impact on third country access to the old EU members’ market
future work 19 Estimating trade cost equivalents consistent with cumulative trade volume changes Objective: translate trade volume effects into price/cost effects. These represent a mixture of possible effects: - Simple RTA dummies would miss third county benefits of non- discriminatory changes linked to accession of new Members. Method: small computational model (general equilibrium) to identify set of trade cost reductions consistent with estimated volume changes. - 3 regions (EU15, EU12, rest of world) - 12 sectors (8 services, 4 others: agrofish, mining, food, manufactures, transport, communication, construction, finance, insurance, business services, consumer services, other services)
future work 20 Basic computational approach Model: - Data (including intermediate linkages and two-way trade flows) are benchmarked to 2003 - CES production structures, final demand, and trade (consistent with sector level gravity equation) - Iceberg trade costs linked to bilateral import demand Selected services trade flows are made exogenous, in a “closure swap” with trade costs Trade flow changes are imposed on the model (for 8 estimated service sectors, for EU12 EU15, EU15 EU12, ROW EU12, ROW EU15), yielding trade cost reductions
Recommend
More recommend