94 EICHENGREEN and FRANKEL TABLE 3.Intra-regional Trade Shares in the Interwar Years Intra-regional trade Ratio 1 Total trade of the region Intra-regional trade /Total trade of the region Ratio 2 Total trade of the region/World trade 1928 1935 1938 Franc Zone Ratio 1 0.247 0.439 0.423 Ratio 2 0.318 0.568 0.729 British Commonwealth Ratio 1 0.403 0.464 0.484 Ratio 2 0.141 0.163 0.172 Western Hemisphere Ratio 1 0.212 0.203 0.197 Ratio 2 0.091 0.107 0.101 German Bloc Ratio 1 0.255 0.234 0.235 Ratio 2 0.178 0.159 0.153 Guilder Zone Ratio 1 0.124 0.110 0.139 Ratio 2 0.287 0.288 0.330 Yen Zone Ratio 1 0.289 0.383 0.478 Ratio 2 0.533 0.535 0.563 Source: League of Nations (1941). countries according to both measures.In contrast,there is no evidence by either measure of the German bloc relying more on intra-bloc trade.The evidence for the other blocs is mixed.There is no evidence of a shift in intra-dollar area trade according to Ratio 1 but some rise according to Ratio 2.(This difference may be attributable to the fact that U.S. trade collapsed dramatically after 1928.)Ratio I suggests an increase in the share of intra- group trade on the part of the yen-bloc countries,but not so Ratio 2,while the converse is true for the members of the guilder zone. Table 4 reports gravity equations for 1928,1935,and 1938.We include only the leading trade blocs of the period:the Commonwealth,the German bloc,and the dollar area.3 The coefficients on the product of national incomes are strikingly close in 1928 to their postwar values of 0.7,suggesting that the tendency for larger economies to trade more than smaller ones was equally pronounced in the 1920s.4 These coefficients decline in magnitude in 1935 and 1938,however,as if the tendency for tariff increases to discourage trade reduced the trade dependence of larger countries disproportionately.The coefficients on the product of per capita incomes is smaller in 1928 than after World War II,suggesting a somewhat smaller tendency for rich countries to trade more than poor ones between the wars.(This coefficient is sometimes interpreted in terms of the prevalence of intra-industry trade,since this is the type of trade in which pairs of high-income countries are thought to engage in disproportionately.Thus,the rise in its magnitude between the 1920s and the 1970s plau- sibly reflects the growing importance of this form of trade.)The coefficients on distance and adjacency are also smaller in 1928 than in the 1970s and 1980s,as if other factors,such as a history of colonial ties,mattered more than geography for the pattern of trade.The magnitude of the coefficients declines between 1928 and 1935-1938,as if factors like polit- ical and imperial ties increasingly overrode geography as a determinant of trade
94 EICHENGREEN and FRANKEL TABLE 3. Intra-regional Trade Shares in the lnterwar Years Ratio 1 = Intra-regional trade Total trade of the region Ratio 2 = Intra-regional trade / Total trade of the region Total trade of the region /World trade 1928 1935 1938 Franc Zone Ratio 1 0.247 0.439 0.423 Ratio 2 0.318 0.568 0.729 British Commonwealth Ratio 1 0.403 0.464 0.484 Ratio 2 0.141 0.163 0.172 Western Hemisphere Ratio 1 0.212 0.203 0.197 Ratio 2 0.091 0.107 0.101 German Bloc Ratio 1 0.255 0.234 0.235 Ratio 2 0.178 0.159 0.153 Guilder Zone Ratio 1 0.124 0.110 0.139 Ratio 2 0.287 0.288 0.330 Yen Zone Ratio 1 0.289 0.383 0.478 Ratio 2 0.533 0.535 0.563 Source: League of Nations (1941). countries according to both measures. In contrast, there is no evidence by either measure of the German bloc relying more on intra-bloc trade. The evidence for the other blocs is mixed. There is no evidence of a shift in intra-dollar area trade according to Ratio 1 but some rise according to Ratio 2. (This difference may be attributable to the fact that U.S. trade collapsed dramatically after 1928.) Ratio 1 suggests an increase in the share of intragroup trade on the part of the yen-bloc countries, but not so Ratio 2, while the converse is true for the members of the guilder zone. Table 4 reports gravity equations for 1928, 1935, and 1938. We include only the leading trade blocs of the period: the Commonwealth, the German bloc, and the dollar area.3 The coefficients on the product of national incomes are strikingly close in 1928 to their postwar values of 0.7, suggesting that the tendency for larger economies to trade more than smaller ones was equally pronounced in the 1920~.~ These coefficients decline in magnitude in 1935 and 1938, however, as if the tendency for tariff increases to discourage trade reduced the trade dependence of larger countries disproportionately. The coefficients on the product of per capita incomes is smaller in 1928 than after World War II, suggesting a somewhat smaller tendency for rich countries to trade more than poor ones between the wars. (This coefficient is sometimes interpreted in terms of the prevalence of intra-industry trade, since this is the type of trade in which pairs of high-income countries are thought to engage in disproportionately. Thus, the rise in its magnitude between the 1920s and the 1970s plausibly reflects the growing importance of this form of trade.) The coefficients on distance and adjacency are also smaller in 1928 than in the 1970s and 198Os, as if other factors, such as a history of colonial ties, mattered more than geography for the pattern of trade. The magnitude of the coefficients declines between 1928 and 19351938, as if factors like political and imperial ties increasingly overrode geography as a determinant of trade
Economic Regionalism:Two 20th-Century Episodes 95 We again estimate pairs of dummy variables,with the subscripts"2"and"N,"for the Commonwealth,the German sphere of influence,and the dollar area.An advantage of the interwar data is that it may be possible to distinguish the effects of explicit trade prefer- ences from other factors shaping international transactions.The data for 1928 essentially predate Britain's imperial preferences,Germany's bilateral agreements,and other discrim- inatory trade-policy initiatives that are sometimes thought to have had such a powerful impact on the direction of trade in the 1930s.Thus,the coefficients for 1928 capture other bloc-specific determinants of the pattern of trade such as the legacy of colonial and impe- rial ties.The change in these coefficients between 1928 and 1935-1938 then provides a measure of the effects of trade-policy initiatives. The coefficient for 1928 shows that members of the British Commonwealth already traded twice as much [exp(0.75)=2.11]as would have been predicted by their other char- acteristics and by the average behavior of countries in the sample.They showed no ten- dency to trade less with other parts of the world.The members of what became the German sphere of influence in the 1930s displayed a similar tendency to trade disproportionately TABLE 4.Regression Estimates of the Gravity Model,1928-1938 Independent Variable 1928 1935 1938 Constant -4.02 -3.53 4.29 (0.59) (6.13) (0.58) GNP 0.69 0.56 0.60 (0.03) (0.03) (0.03) GNP/Capita 0.19 0.08 0.10 (0.05) (0.04) (0.04) Distance 0.48 -0.37 -0.34 (0.06) (0.06) (0.05) Adjacency 0.34 0.19 0.18 (0.19) (0.18) (0.18) Commonwealth2 0.75 1.15 0.92 (0.23) (0.21) (0.20) CommonwealthN 0.04 0.04 -0.07 (0.10) (0.10) (0.10) German2 0.84 0.88 0.58 (0.22) (0.21) (0.19) GermanN 0.17 0.21 0.22 (0.10) (0.31) (0.09) Dollar2 0.33 0.57 0.53 (0.35) (0.31) (0.32) DollarN 0.25 0.06 0.20 (0.13) (0.12) (0.11) Number of observations 419 418 426 Adjusted R2 0.66 0.62 0.62 Notes: Standard errors in parentheses. All variables except the dummies are in logarithms
Economic Regionalism: Two ZOth-Century Episodes 95 We again estimate pairs of dummy variables, with the subscripts “2” and “N,” for the Commonwealth, the German sphere of influence, and the dollar area. An advantage of the interwar data is that it may be possible to distinguish the effects of explicit trade preferences from other factors shaping international transactions. The data for 1928 essentially predate Britain’s imperial preferences, Germany’s bilateral agreements, and other discriminatory trade-policy initiatives that are sometimes thought to have had such a powerful impact on the direction of trade in the 1930s. Thus, the coefficients for 1928 capture other bloc-specific determinants of the pattern of trade such as the legacy of colonial and imperial ties. The change in these coefficients between 1928 and 19351938 then provides a measure of the effects of trade-policy initiatives. The coefficient for 1928 shows that members of the British Commonwealth already traded twice as much [exp(0.75) = 2.1 l] as would have been predicted by their other characteristics and by the average behavior of countries in the sample. They showed no tendency to trade less with other parts of the world. The members of what became the German sphere of influence in the 1930s displayed a similar tendency to trade disproportionately TABLE 4. Regression Estimates of the Gravity Model, 1928-1938 Independent Variable 1928 1935 1938 Constant GNP GNP/Capita Distance Adjacency Commonwealthz CommonwealthN German2 GermanN Dollar2 DollarN Number of observations 419 418 426 Adjusted R* 0.66 0.62 0.62 -4.02 -3.53 (0.59) (6.13) 0.69 0.56 (0.03) (0.03) 0.19 0.08 (0.05) (0.04) -0.48 -0.37 (0.06) (0.06) 0.34 0.19 (0.19) (0.18) 0.75 1.15 (0.23) (0.21) 0.04 -0.04 (0.10) (0.10) 0.84 0.88 (0.22) (0.21) -0.17 -0.21 (0.10) (0.3 1) 0.33 0.57 (0.35) (0.3 1) -0.25 0.06 (0.13) (0.12) -4.29 (0.58) 0.60 (0.03) 0.10 (0.04) -0.34 (0.05) 0.18 (0.18) 0.92 (0.20) -0.07 (0.10) 0.58 (0.19) -0.22 (0.09) 0.53 (0.32) -0.20 (0.11) Notes: Standard errors in parentheses. All variables except the dummies are in logarithms