i r 68407 September 1993 Preferential Trade Arrangements: Estimating the Effects on South Asia Countries T. N. Srinivasan "- Gustavo Canonero I. Introduction Measuring the impact of potential trading blocks involving South Asian I countries is the aim of this paper. R. Safadi and A. Yeats (1993) have already analyzed the likely impact of the formation of the North American Free Trade Area on South Asia. We broaden this line of research to analyze the impacts of other potential preferential trading arrangements. In particular, we evaluate the effects of South Asia Regional Integration (among Bangladesh, India, Nepal, Pakistan and Sri Lanka), as well as preferential trade arrangements (PTA's) of individual South Asian countries with some non- regional partners like USA, NAFTAas a whole, EEC or an Asian Group. Our tool for empirical analysis is an extended version of the simple and well-known Gravity Model of Bilateral Trade, such as the one used by J. Frankel et al. (1993). We estimate bilateral trade using the standard approach with explanatory variables like GNP, GNPper capita, and distance, but we also attempt to quantify the impact of restrictions to trade, basically represented by tariff barriers. The coefficients on these tariff barriers are then used to quantify the outcome of the various PTA's mentioned before. l"'~ fr-e..pfr"o~ Z II. Model Specification and Estimation Method Based on the empirically well-behaved simple versions of the Gravity Model of Bilateral Trade, our working equation is the following: Log BT~,j,t=ag+a~ Log (GNPi,t*GNPj,t)+a~ Log (GNPPCi,t*GNPPCj,t) (1) +a~Di,j+a~ Log T~,j+a~Log Tj,i+a~ Log REXRi,j,t+U~,j,t where BTi,j,t is bilateral trade between countries--or group of countries--i and j in the year t and the superscript c indicates the commodity group. GNP and GNPPC are the respective Gross National Product and per capita National Product of both partners in year t, Di,j is the distance in kilometers between the relevant cities, TCi,j is (one plus) the tariff imposed by country i in its trade on commodity c with j and REXR is a measure of real exchange rate changes in any of the two trading partners. BT, GNP, GNPPC are measured in thousand of 1987 US dollars, using the GDP deflator for the US to deflate nominal values. The expected results are that trade between two equal sized countries should be greater than between one large and one small country, hence al is likely to be positivel. Moreover, it is reasonable to think that GNP per capita would also have a positive effect on trade (az > 0) since it is a proxy l\See Frankel et a1. (1993) for a brief discussion and references about this specification. Deardorff (1984) points out that the early formulations of the gravity model by Tinbergen (1962) and Poyhonen (1963) were based on ad hoc but intuitive theorizing, and in his further elaboration of the model Linnemann (1966) derived an analogue of the gravity equation from a quasi-Wa1rasian general equilibrium model under the rather implausible assumption of separate demand functions for imports from each trading partner. Anderson (1979) derived an extremely simple version of the model by assuming that each country is specialized in its own good and all countries have the same Cobb-Douglas preferences. Bergstrand (1985, 1989) derives a generalized gravity model from a general equilibrium model with products differentiated by place of production and place of sale. See Wang and Winters (1991) for a brief discussion of the theory of gravity models. 3 of degree of development and therefore of specialization. On the other hand, the distance between two countries, like tariff barriers, should restrict trade, one naturally (through transport cost) and the other artificially (through policy). Thus we would expect a3' a4 and as to be all negative. Finally, the effect of real exchange rate on total trade is ambiguous, but it is included in equation (1) to serve as a "catch-all" proxy variable for other fundamental changes not incorporated in the other variables, such as a maxi- devaluation, for instance. The utilization of the panel features of the data set required special treatment of the error term in (1). A general expression for Uci.j,t could be the following: C C /}C C C (2) Ui.j,t= fi + Uj + JJt + 'Ii,j,t where, f and 0 represent the individual effects, JJ a temporal effect and ~ a purely random effect. Leaving aside the temporal disturbance for the moment, we can always avoid consistency problems arising from individual effects (though not necessarily achieve efficiency), by estimating of equation (1) with its variables defined as deviations from their individual means. However, this option is not attractive since for evaluating the effects of trade liberalization we need to know values of the parameters associated with some individual variables, viz. Ti and Tj' which are eliminated from the estimating equation once the variables are expressed as deviations from their means. Having to estimate all parameters in equation (1) leads us to the standard variance-components model of (2). If f and 0 were fixed effects, then by using country dummies the problem is easily solved. On the other hand, if these effects were random, the best way to estimate equation (1) would be by 4 using Generalized Least Squares, taking into account the contribution of the variance of country effects to the variance of U~jt. In this case, a common problem that arises is the possible correlation between the individual country effects and the explanatory variables. This possibility should be tested through an appropriate test for the exogeneity of explanatory variables in order to see if instrumental variables are needed2. A preliminary estimation of (1) using country dummies demonstrated that the assumption of fixed effects is inappropriate: the residuals from this estimation still showed the presence of individual effects (we tested for homoscedasticity, see below). Therefore we reestimated equation (1) assuming random effects, used the lags of GNP and GNPPC as instruments and tested for their exogeneity. The results of our estimation and associated tests are presented in section IV, where we also report on our attempts to take into account the possible presence of the temporal terms. We review the sources and characteristics of our data in Section III and report the simulated effects of various preferential trade arrangements in Section V. III. The Data The raw data on trade are from the United Nations' COMTRADEdatabase (at the International Computing Center in Geneva), where imports and exports are l\Actually, instrumental variables should anyhow be used in this particular case since variables like GNP or GNPPC are probably correlated with the error term, since the dependent variable, TB, is the sum of imports and exports. In spite of this, it is indeed striking that the ruling practice is to use simple OLS in estimating these equations! 5 reported in current USA dollars for each country and all its trading partners. The data are available annually and by commodity using SITC classification (Rev. 1 and 2). With this basic information we calculate the bilateral trade (BT- imports + exports) on 10 commodities for 21 trading partners. Both commodity and country aggregation intend to enhance the most important features of the actual trade pattern of South Asia Countries3. The ten composite commodities are: Coffee and Tea, Textile Fibres, Leather and Dressed Fur., Textile Yarn and Clothing, where South Asia concentrates its exports, and other aggregates such as Non-Fuel Primaries, Fuels, Machinery and Transport Equipment, Other Manufactures and Total Trade. On the other hand, 13 individual countries were selected (the five from South Asia and their major partners: USA, Canada, Mexico, Japan, Hong Kong, Korea, , Singapore and China) and 8 groups (EEC 6, EFTA, Africa, Asia, Middle East, Latin America, OECDand European countries not considered elsewhere). In order to calculate the relevant distance we use the major city for the individual countries (mostly the capitals) and an average distance for the groups of countries (usually taking two cities in the group or a city located in the geographic center of the group). In particular, we use the 3 major ports in India to calculate its distance with other partners since the size (in area) of a country could greatly affect transport cost of internal trade relative to international trade with its neighbors, and therefore it could be an important factor in our simulations (see Srinivasan and Canonero (1993) in l\A summary of actual trade in the South Asia Region is presented in Tables B.l to B.3 in Appendix B. 6 this context). A complete description of country groups and associated cities is presented in Appendix A, Table A.I. The information on tariffs used in the estimation is shown in Appendix A, Table A.2. The figures for South Asian countries are the implicit effective tariffs calculated from tariff revenues. These are average tariffs and include import and export duties. The data on tariffs imposed by developed countries and Mexico come from the UNCTAD-World Bank Software for Market Analysis and Restrictions on Trade (SMART), where they are available desegregated by commodity. The tariff structure in Table A.2 is the reported one for Japan and Mexico (again effective tariff) but not for Canada, EEC and USA. Although the actual tariffs imposed by the latter are not completely similar, these probably are compensating differences in their NTB coverage. We therefore decided their tariff structure to be the same as the average of their reported tariffs4. One final remark should be made about the series Ti and Tj. Although we consider that tariffs affect traders uniformly, we nevertheless treat trade not involving South Asian countries as unaffected by tariffs. In other words, Ti and Tj are set at one whenever neither i nor j are Bangladesh, India, Nepal, Pakistan or Sri Lanka. Basically this procedure implies that the magnitudes of such tariffs, as compared to those associated with South Asia countries, to be small enough to be negligible. Finally the variable REXR aims to account for significant changes in real exchange rate in the South Asia countries. REXR is an index with base 1965=1 for the real exchange rate of the domestic currencies of South Asian ~\. Information on NTB imposed by Canada, EEC, Japan, Mexico and USA on South Asia Countries in 1988 is revealed in Appendix C. Like tariffs data, the source for NTB is SMART. 7 countries and the USA dollar and it is set at 1 for any other country. REXR is calculated using the nominal exchange rate and GDP deflator from IMF's International Financial Statistics (exchange rate is market value, period average -series rf-). It is the source also for the GNP and Population figures. In those observations where both the trading partners are from South Asia an average of each REXR is used. IV. The Results Table 1 to 10 report the estimation results for equation (1), for the ten commodities in our sample. We use annual data from 1968 to 1991, having 218 pairs of trading partners and 5175 observations.s Each Table has 3 different parts, which besides showing the final results, also reveal the importance of handling the individual effects. The outputs are: 1) The results of the estimation using Instrumental Variables but ignoring the panel feature of the data set, i.e. without variance decomposition. 2) The results of standard procedures for testing for the presence of individual effects. 3) The results of using Instrumental Variables, but also accounting for the variance decomposition implied by the test in 2 above. In this final output a WALD test for the significance of the whole equation is also included (the null hypothesis is: Ql-QZ-Q3-a4-aS=Q6-0) , since the reported 2\The 218 pairs of trading partners arise from the following. By pairing each country (or groups) from our total of 21 (13 countries and 8 groups) with a different country (or group), we get 210 pairs (i.e. (20*21)/2). In addition, since the countries within each of the 8 groups trade with each other, there are 8 within group trade pairs, i.e. each group trading with itself. Although data relate to a period of 24 years (1968-1991), because some data are missing, the number of observations is 5175 instead of 5232 (i.e. 218 x 24). 8 R Square becomes uninformative whenever instrumental variables are used. These final results are used in subsequent simulations. Before turning to the estimation results, two comments are in order. First, the instruments chosen were the second and third lags of GNP and GNPPC. In order to check their merits as instruments, we ran an overidentified restriction test on each of them. One way to run these tests is by regressing ~ instrument on the residuals from an equation using all but ~ instrument. The results from this round of tests were satisfactory, all instruments showing no correlation with the residuals. However, as is well known, these tests are valid only when there is at least one "true" instrument, which is seldom the case. Anyway, the lack of ideal instruments left us with few alternatives. However our simulations are not likely to depend heavily upon the quality of instruments used. We believe that the consistency of al and az is more likely to be harmed by a bad instrument than those that are critical to our simulations, viz. the parameters on tariffs (a4 and as). The second comment, is about the temporal effect or the possibility of serial correlation in the error term of equation (1). The D-W statistics in our estimates are usually below the critical value of 1.7 (approximately the lower bound at 0.05 level of significance, given the degree of freedoms we have). Unfortunately we do not have time series data long enough to estimate an appropriate structure of serial correlations. Economising on degrees of freedom by not allowing each pair of trading partners in our panel to respond differently to the same time disturbance led to implausible results.6 Q\For instance, we estimated equation (1) by Maximum Likelihood, correcting for autocorrelation of order one, assuming the same time series model for all (continued. ..) ,~; 9 The above discussion should make the reading of Tables 1 to 10 simpler. We will avoid, therefore, repetition and only comment on the final output, the one estimated by instrumental variables and accounting for the individual random effects in the variance composition (the estimation method used is: Generalized Least Squares-Instrumental Variables). Each Table presents the result for each composite commodity, starting with Total Trade in Table 1 . The remarkable explanatory power of this equation is the basic conclusion of our exercises. This is no surprise since the gravity model has been found by many earlier authors to explain bilateral trade flows exceptionally well in spite of its weak grounding in theory. All variables have the expected signs and are significant at 5% level for almost any composite commodity. The bigger the trade partners, the more significant is the bilateral trade, the greater the distance between the trading countries and the higher the tariff barriers the smaller is the value of bilateral trade too. Only the coefficients of product of GNP per capita have ambiguous signs (i.e. positive in some and negative in other equations). They are statistically significant in some but not all cases. For example, GNPPC is significantly positively correlated with trade in coffee, tea and spices, machines and transport equipment and clothing, while being negatively or not significantly correlated with trade on other composite commodities. However, two of the three negative coefficients, i.e. those relating to trade in textile fibres and non-fuel primary products, are statistically significant. These findings are not inconsistent with received theory under which a higher .§.\(...continued) individuals (i.e. the same value for Pl), and every coefficient a turned to be non significant. 10 GNP per capita indicates a higher degree of development, and therefore more specialization, which in tern is more important for manufactured goods. That both significant negative coefficients relate to trade in primary goods is consistent with this hypothesis. Since the dependent variable is the sum of imports and exports, the effect of the real exchange rate on it is ambiguous in theory since its effect is expected to be positive on exports and negative on imports (actually these are the expected signs in the long run, while in an annual basis trade- exchange rate relationship for some commodities may not always follow the long-run direction, principally because of contractual rigidities). However, in our estimations the real exchange rate shows positive correlation with bilateral trade in every commodity, although this statistical relation is not significant for the primary goods (i.e. coffee, tea, cocoa and spices, textile fibres and fuels). Obviously, the explanations for the signs and sizes of the estimated coefficients have to be found in the likely influence of the real exchange rate on the demand and production of the relevant commodities. Be that as it may, the significance of real exchange rate on total trade for non- primary goods indicates that empirically this effect is positive. Equipped with this set of results we simulate the effect of preferential arrangements involving South Asian countries. In our framework, a simple way to get an u~~er bound for this effect is to consider the total elimination of tariff barriers between those countries or group of countries for which such an arrangement is contemplated. The results of these simulations are presented in the following section. 11 V. Numerical Simulations In interpreting our simulation results, it is important to keep in mind that the simulations are driven by the estimated coefficients a4 and as for the tariff variables Tl and Tz respectively. We clarify their role while stressing some of the limitations of our calculations. Basically we measure the impact of preferential trade arrangements (PTA) by the proportionate change in the USA dollar value of trade they create. The higher the initial tariff level on trade between partners, the greater the final effect of such arrangements since, by definition, they eliminate the tariffs. We should, therefore, expect the impact to be increasing in the estimated values of a4 and as. However, the tariff is only one among many elements that determine the impact of PTA on trade. In assessing the impact of preferential tariff reductions, two other features of our model have to be kept in mind. First, the series Tl mainly represent tariffs imposed by Canada, EEC, Japan, Mexico and USA on imports from South Asian countries, while series Tz are tariffs imposed by the latter on trade with the former group. Since tariffs Tz are initially higher than Tl, the higher the coefficient of Tz, in absolute values, the greater the impact of preferential arrangement. Second, since a4 and as in equation (1) are elasticities indicating the proportionate response of bilateral trade to changes in tariffs, the initial tariff level as well as the initial trade level are relevant for determining the absolute changes in trade following a PTA. Once the forces driving our simulations are understood, their limitations also become apparent. First, our approach does not take into account possible terms of trade effects associated with the creation of trade. 12 As such, the simulated results almost certainly overestimate the true impact. On the other hand, the static framework of our exercise does not consider many" important dynamic aspects of trade liberalization and these could reinforce the short-run trade creation thus underestimating the true long-run impact. The long-run effect is very difficult to forecast without a more general dynamic setup that generates terms of trade effects and allows for scale economies, investment, and spillovers of technology etc. The basic information to quantify all these is not available: for example, some price elasticities could be approximated but information on scale economies simply does not exist. However, experience with the few studies that tried to estimate dynamic effects in trade, indicate that short-run effects are likely to overestimate the true long-run effects. Second, our measure of trade barriers, viz. an indicator for average tariffs, does not fully capture the effect of many non-tariff barriers to trade. Third, even if our simulations correctly measure the impact on trade creation, it should be realized that this impact is not the only factor to take into account in evaluating PTA's. For instance, the negative effect on bilateral trade with countries not entering in the simulated arrangement is not assessed at all in our simulations. Moreover, none of the indicators from the simulations could be viewed as a welfare measure, thus making the comparison of different scenarios rather incomplete. In other words, the results presented below serve the limited purpose providing an estimate of the potential effects on bilateral trade between each South Asian country and its partner in the simulated PTA. The expected outcomes from different preferential trade arrangement are presented in Tables 11 to 19. Each table represents the outcomes for each composite commodity, except coffee, tea, cocoa and spices. The coefficients 13 for tariffs (as and a6) on the latter were not significantly different from zero so that a tariff reduction does not affect trade flows. The parameters in the estimated equation give us the expected ~ro~ortionate increase in bilateral trade. Using data for 1990 and 1991 for all the countries (except Nepal, for which we had data only for 1989-1990), we translate the magnitude of such effect into USA dollars. As mentioned earlier, the exercise simply associates preferential arrangement with total elimination of existing tariffs between the member countries. Each column in each table corresponds to a South Asian country and each row, its potential partners in preferential trading. They are: USA, Canada, Mexico, EEC, Japan and SAS (i.e. other South Asian countries). The sum of the effects on bilateral trade with USA, Canada and Mexico in each column would correspond to the effect of the South Asian country represented by that column joining NAFTA. The data in the cells are the basic indicators of trade effects when the preferential arrangement is between a South Asian country (column) and the partner (row). The cells in the last rows indicate the effects for each South Asian country of its entering into a preferential trade arrangement with the other South Asian countries. The results are presented in tables 11 to 19 in two categories, those contained in part A of each table show the expected value of the estimated effects, and those in part B of each table refer to confidence intervals for these expected values at 95% level of significance. The latter reveal the relative precision of the simulated effects7. More specifically, in each cell of each table's part A: 1\ For a technical description of the estimated expected value and its confidence interval, see Appendix D. 14 1) The first row indicates the proportional increase in bilateral trade that would happen if the countries corresponding to the row and column of the cell joined a preferential arrangement. It is calculated as: exp(~4ln(Ti)+~5ln(Tj)+~ tl2(a4 In(Ti)+a5 In(Tj» » -1 (see Appendix D), where Ti-(l+percentage tariffi). For instance, for i being USA and j being India and the commodity being total trade, this value would be exp(3.9 In(1.13) + 4.65 In(1.42) + ~ 0.2325) -1- 8.2. Clearly, in order to express this in percentage terms, this value should be multiplied by 100. 2) The value of such potential increase in bilateral trade denoted in millions of USA $ (1991 prices) is given in the second row. This is the value in 1) times the initial bilateral trade between i and j. 3) The value of potential increase in bilateral trade given in 2) as a percentage of the initial value of total trade of the South Asian country represented by the column is given in the third row. 4) The value of the potential increase in bilateral trade given in 2) as a percentage of the initial value of GNP for the South Asian country represented by the column is given in the fourth row. The information on each cell of each table's part B is similar to the one detailed above, but expressing the estimated effects by their estimated upper and lower bound at 95% level of significance. Thus, row 1 in each table's part B has the potential range of increase in bilateral trade and the 15 other rows just translate these magnitudes in terms of millions of USA $ (1991 prices), share of total trade and GNP, as it is done in part A of each table. In short, in both parts of each table, the first two rows give the proportionate and USA dollar value magnitudes of the expected increase in bilateral trade when both partners reduced theirs tariffs to zero. The third and fourth rows achieve comparability across countries by relating the absolute magnitude to the value of total trade and of GNP respectively of the relevant South Asian country. Tables 11 to 19 reveal that the expected trade effect of preferential arrangements is impressive. For all the countries in the South Asian region, the most important potential effect is associated with textile fibres and manufactures, specially textile yarn and clothing. This is not surprising since in this case the initial trade is substantial, tariff barriers are high and the estimated values of a4 and as are also high. For the larger economies, like India and Pakistan, the principal gains seem to come from preferential arrangements with the European Community and USA. This is to be expected since a large share of their trade is with these countries. On the other hand, regional integration leads to greater gains in the value of trade for the small economies like Bangladesh and Nepal. A summary indicator of all the results obtained are the effects on total trade presented in Table 11.A. (Table 11.B shows the associated confidence intervals). It is clear that the greatest ~ro~ortional increase expected in bilateral trade would come from regional integration. The countries in the region have all higher tariffs than other countries and therefore a larger impact from tariff reduction in their tariff should be expected. For example the trade between Bangladesh and the other South Asian countries is expected "" 16 to increase by 9.5 times and almost the same is the case of Pakistan (8.9). For Sri Lanka the expected proportional increase is about 10.3 times, for India 12.8 times, while for Nepal this increase is even higher, at 17.2 times. However, given the initial trade pattern of these countries, regional integration leads to a greater increase in trade for Bangladesh and Nepal than for the other South Asian countries. For the two small economies of Bangladesh and Nepal, regional integration seems to be the most powerful preferential arrangement for trade creation. The values of expected increases in trade are incredible. Bangladesh's new trade with the region would account for US $ 4.6 billion, exceeding its actual total trade of $ 3.8 billion by 17%, and accounting for a whopping 21.1% of its GNP!. In the case of Nepal the trade expected to be created in the region is around US $ 1.7 billion. Though this is smaller than that for Bangladesh, it is not less impressive considering the (economic) size of Nepal. The new trade would be almost three times the actual total trade of Nepal and 58.5% of its GNP!. The effects of regional integration on the large economies in the region, viz. India or Pakistan, are naturally very different from those of Bangladesh and Nepal. For India and Pakistan a regional integration is important, but their much larger trade with the European Communities and USA, makes integration with the latter more attractive. Both countries would achieve the greatest impact on their trade by integrating their economies with EEC, for India bilateral trade would increase by 2 times its actual total trade, while for Pakistan the corresponding figure is around 0.95. Translating these effects into US dollars, new bilateral trade between India and EEC would 17 amount to US $ 85 billion, representing 30% of India's GNP!. For Pakistan these figures would be US $ 13 billion and 30%. The differential effects of the above scenarios on particular commodities are interesting. For example, for all South Asian countries, regional integration would clearly create more trade on textile fibres than any other preferential trade arrangement. In this commodity the tariff reductions would be significant and the actual trade within the region is strong enough to foster likely future trade. On the other hand, the actual trade pattern indicates that trade in clothing could receive an extraordinary impulse if South Asian countries integrate with EEC or USA, with the increased trade being similar for the large and small economies in the region. In conclusion, one would have to recognize that the estimated effects are sometimes too large to be believed, as it is the case of Nepal. However, it is worth noting that the results of Tables A are estimated average effects. Actually, the likelihood of the final outcome in any possible scenario is better illustrated by the more comprehensive measure presented in tables B, viz. the confidence intervals. Indeed, these tables show the wide range of magnitudes that are possible as a final output in any of the exercises presented above. Furthermore, as we emphasized earlier, some important factors affecting trade have not been taken into account in our simulations. Almost surely they could restrain significantly the numerical magnitudes of the effects reported. 18 Table 1. TOTAL TRADE Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES -NO PANEL FEATURES Variable Coeff Std Error T-Stat Signif GNP 0.72735026 0.01715866 42.38968 0.00000000 GNPPC 0.33445465 0.02252389 14.84888 0.00000000 D -0.80508512 0.04660275 -17.27549 0.00000000 T1 -5.74306686 0.41535882 -13.82676 0.00000000 T2 -5.45346591 0.37926731 -14.37895 0.00000000 REXR 0.78663776 0.09355628 8.40818 0.00000000 Constant -21.99658331 0.82354954 -26.70948 0.00000000 R**2- 0.54 R**2 Adj.- 0.54 DW-0.74 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 15259.147996383 217 70.318654361 24.399 0.00000000 ERROR 14286.355410242 4957 2.882056770 TOTAL 29545.503406625 5174 INSTRUMENTAL VARIABLES -RANDOM EFFECTS Variable Coeff Std Error T-Stat Signif == GNP 0.82954569 0.05252544 15.79322 0.00000000 GNPPC 0.01079169 0.09054588 0.11918 0.90513363 D -1.08263825 0.16999122 -6.36879 0.00000000 T1 -3.90443785 1.55071295 -2.51783 0.01183788 T2 -4.65837217 1.19281557 -3.90536 0.00009529 REXR 0.87676670 0.07449218 11.76992 0.00000000 Constant -19.15797828 2.49129123 -7.68998 0.00000000 R**2- 0.007 R**2 Adj.- 0.007 DW-1.23 WALD TEST F(7)- 2159 Significance Level 0.000 19 AND SPICES Table 2. TRADEON COFFEE, TEA, COCOA Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif = GNP 0.79439542 0.02194642 36.19704 0.00000000 GNPPC 0.01387946 0.03054852 0.45434 0.64960219 D -0.90546503 0.06031447 -15.01240 0.00000000 T1 -6.19172416 0.64326633 -9.62544 0.00000000 T2 -5.08781948 0.50324885 -10.10995 0.00000000 REXR 0.56535654 0.12045507 4.69351 0.00000275 Constant -24.82526493 1.09440921 -22.68371 0.00000000 R**2- 0.37 R**2 Adj.- 0.37 DW~0.35 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 37280.118100899 217 171.797779267 74.564 0.00000000 ERROR 11421.114748417 4957 2.304037674 TOTAL 48701.232849315 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif = -= GNP 0.62997118 0.08673594 7.26309 0.00000000 GNPPC 0.43820868 0.16527136 2.65145 0.00803917 D -1.17429747 0.30465940 -3.85446 0.00011739 T1 5.81319206 3.38555079 1.71706 0.08602821 T2 0.43351890 2.02020145 0.21459 0.83009396 REXR 0.09776266 0.08004428 1.22136 0.22200652 Constant -20.58091847 3.83574070 -5.36557 0.00000008 R**2~ -0.37 R**2 Adj.- -0.37 DW=1.32 WALD TEST F(7)- 344 Significance Level 0.000 L-- 20 Table 3. TRADEON TEXTILE FIBRES Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif -- GNP 1.03078944 0.01973066 52.24302 0.00000000 GNPPC -0.37324564 0.02754701 -13.54941 0.00000000 D -0.72120630 0.05431581 -13.27802 0.00000000 T1 -10.96516064 0.60888777 -18.00851 0.00000000 T2 -5.29645881 0.45778897 -11.56965 0.00000000 REXR 0.12815142 0.10872116 1.17872 0.23856536 Constant -32.13512929 0.98904687 -32.49101 0.00000000 R**2- 0.50 R**2 Adj.- 0.50 DW-0.47 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 22575.836911799 217 104.036114801 30.957 0.00000000 ERROR 16658.894047694 4957 3.360680663 TOTAL 39234.730959493 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif GNP --====== 0.87541551 0.06257944 =- 13.98887 0.00000000 GNPPC -0.67710245 0.11648153 -5.81296 0.00000001 D -0.61608358 0.20840770 -2.95615 0.00312918 T1 -15.38539900 2.44383255 -6.29560 0.00000000 T2 -7.78261940 1.46380592 -5.31670 0.00000011 REXR 0.05893799 0.08126182 0.72529 0.46830997 Constant -19.25591102 3.00374227 -6.41064 0.00000000 R**2= 0.06 R**2 Adj.- 0.06 DW=0.90 W'ALD TEST F(7)- 605 Significance Level 0.000I 21 Table 4 TRADE ON FUELS Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif --========= =--======- -- GNP 1.00869377 0.02652873 38.02270 0.00000000 GNPPC 0.00519856 0.03761178 0.13822 0.89007490 D -1.90069876 0.07634009 -24.89778 0.00000000 T1 -10.43882651 0.90322259 -11.55731 0.00000000 T2 -10.59727703 0.61946772 -17.10707 0.00000000 REXR -0.04427047 0.14706739 -0.30102 0.76341008 Constant -25.90536596 1.36917193 -18.92046 0.00000000 R**2= 0.50 R**2 Adj.- 0.50 DW=0.48 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Level INDIV 43512.572213762 217 200.518765962 36.390 0.00000000 ERROR 27314.322686216 4957 5.510252711 TOTAL 70826.894899978 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif -====--- =--- -- GNP 1.27355476 0.09668747 13.17187 0.00000000 GNPPC 0.15503331 0.18240413 0.84994 0.39539556 D -2.15608191 0.34185714 -6.30697 0.00000000 T1 -3.83417765 4.14337454 -0.92538 0.35481377 T2 -8.15291824 2.24106571 -3.63796 0.00027748 REXR 0.10805976 0.11785978 0.91685 0.35926392 Constant -39.74018187 4.69803194 -8.45890 0.00000000 R**2= -0.10 R**2 Adj.= -0.10 DW=1.24 W'ALD TEST F(7)- 346 Significance Level 0.000 22 Table 5. TRADE ON NON-FUEL PRIMARIES (EXCEPT COFFEE, TEA, COCOAAND SPICES, and TEXTILE FIBRES) Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif =~ --= GNP 0.80944983 0.01788552 45.25727 0.00000000 GNPPC 0.11300622 0.02468786 4.57740 0.00000482 D -0.86513988 0.04942992 -17.50235 0.00000000 T1 -7.34079731 0.55179244 -13.30355 0.00000000 T2 -6.49250838 0.41169213 -15.77030 0.00000000 REXR 0.61235380 0.09865960 6.20673 0.00000000 Constant -24.01167627 0.88664480 -27.08151 0.00000000 R**2- 0.54 R**2 Adj.- 0.54 DW-0.59 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 19169.195158996 217 88.337304880 33.357 0.00000000 ERROR 13127.385858414 4957 2.648252140 TOTAL 32296.581017410 5174 RANDOM EFFECTS-INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif - GNP 0.83519564 0.05637340 14.81542 0.00000000 GNPPC -0.20790273 0.10417567 -1.99569 0.04601967 D -0.94289393 0.18906570 -4.98712 0.00000063 T1 -4.75928644 2.20855807 -2.15493 0.03121342 T2 -5.99544834 1.31178519 -4.57045 0.00000498 REXR 0.64692815 0.07100743 9.11071 0.00000000 Constant -18.88194872 2.67321171 -7.06339 0.00000000 R**2- 0.03 R**2 Adj.- 0.03 DW-1.17 WALD TEST F(7)- 1409 Significance Level 0.000 23 FUR., ETC. Table 6. TRADEONLEATHER, DRESSED Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif === ---== , - GNP 0.97330282 0.01968766 49.43720 0.00000000 GNPPC 0.31471532 0.02588048 12.16033 0.00000000 D -1.30342039 0.05387989 -24.19122 0.00000000 T1 -3.74834935 0.50827462 -7.37465 0.00000000 T2 -1.54298974 0.43018674 -3.58679 0.00033787 REXR 0.76301862 0.10720215 7.11757 0.00000000 Constant -37.18794385 0.94085898 -39.52552 0.00000000 R**2- 0.51 R**2 Adj.- 0.51 DW~0.66 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 20385.415793523 217 93.942008265 25.141 0.00000000 ERROR 18522.458912268 4957 3.736626773 TOTAL 38907.874705791 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif -==== = GNP 1.18283676 0.06219875 19.01705 0.00000000 GNPPC 0.04365512 0.10702681 0.40789 0.68337166 D -1.46312117 0.20259018 -7.22207 0.00000000 T1 -2.07962859 1.95688795 -1.06272 0.28795752 T2 -2.72919576 1.39512913 -1.95623 0.05049166 REXR 1.11948357 0.08698916 12.86923 0.00000000 Constant -42.52512924 2.93405894 -14.49362 0.00000000 R**2= -0.06 R**2 Adj.- -0.06 DW=1.27 WALD TEST F(7)= 522 Significance Level 0.000 24 Table 7. TRADEON TEXTILE YARN, FABRIC, ETC. Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif --- GNP 0.69660662 0.01830384 38.05795 0.00000000 GNPPC 0.26216879 0.02370124 11.06139 0.00000000 D -1.21202376 0.04954256 -24.46430 0.00000000 T1 -5.98489475 0.40310636 -14.84694 0.00000000 T2 -3.66590250 0.39760927 -9.21986 0.00000000 REXR 0.59555482 0.09922755 6.00191 0.00000000 Constant -19.05216664 0.87527066 -21.76717 0.00000000 R**2= 0.45 R**2 Adj.- 0.45 DW=0.54 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 20576.733375134 217 94.823656107 36.561 0.00000000 ERROR 12856.419492376 4957 2.593588762 TOTAL 33433.152867511 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif GNP --===== 0.75474111 -====- 0.05764655 13.09256 0.00000000 GNPPC -0.11040279 0.10228867 -1.07933 0.28049295 D -1.39342456 0.19354345 -7.19954 0.00000000 T1 -5.16942084 1.60014735 -3.23059 0.00124309 T2 -4.30509003 1.32253598 -3.25518 0.00114051 REXR 0.69107729 0.07049558 9.80313 0.00000000 Constant -13.85127538 2.67484090 -5.17835 0.00000023 R**2= 0.02 R**2 Adj.= 0.02 DW=1.06 WALD TEST F(7)- 1032 Significance Level 0.000 25 Table 8. TRADE ON MACHINES AND TRANSPORT EQUIPMENT Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif ==== = GNP 0.80122481 0.01844487 43.43889 0.00000000 GNPPC 0.62056312 0.02517390 24.65105 0.00000000 D -1.02250865 0.05035561 -20.30576 0.00000000 T1 -6.03810687 0.51443211 -11.73742 0.00000000 T2 -4.94322478 0.42088777 -11.74476 0.00000000 REXR 0.79925875 0.10123718 7.89491 0.00000000 Constant -30.70944066 0.90564854 -33.90878 0.00000000 R**2- 0.61 R**2 Adj.- 0.61 DW=0.82 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 17170.379379505 217 79.126172256 22.840 0.00000000 ERROR 17172.638388365 4957 3.464320837 TOTAL 34343.017767870 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif --======= = GNP 0.95659553 0.05957198 16.05781 0.00000000 GNPPC 0.48736441 0.10603317 4.59634 0.00000440 D -1.19126592 0.19183219 -6.20994 0.00000000 T1 -3.45338134 2.04805416 -1.68618 0.09182210 T2 -3.13807621 1.37597021 -2.28063 0.02261100 REXR 0.98496967 0.08675565 11.35338 0.00000000 Constant -35.05969695 2.88039703 -12.17183 0.00000000 R**2- -0.22 R**2 Adj.- -0.22 DW-1.50 WALD TEST F(7)- 1270 Significance Level 0.000 26 Table 9. TRADEON CLOTHING Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif == GNP 0.71565614 0.02055326 34.81959 0.00000000 GNPPC 0.67238891 0.02636546 25.50265 0.00000000 D -0.97841158 0.05563760 -17.58544 0.00000000 T1 -3.16580261 0.37280910 -8.49175 0.00000000 T2 -5.83774195 0.43033358 -13.56562 0.00000000 REXR 0.98731198 0.11001981 8.97395 0.00000000 Constant -29.94160383 0.98663132 -30.34731 0.00000000 R**2= 0.52 R**2 Adj.- 0.52 DW=0.63 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 23848.610944764 217 109.901432925 29.810 0.00000000 ERROR 18275.173819457 4957 3.686740734 TOTAL 42123.784764222 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif -=- --=== -= GNP 0.75732199 0.05940591 12.74826 0.00000000 GNPPC 0.30018131 0.10207488 2.94080 0.00328823 D -0.92958086 0.19712703 -4.71564 0.00000247 T1 -2.43126291 1.31764686 -1.84516 0.06507215 T2 -10.33075859 1.32132024 -7.81851 0.00000000 REXR 1.84745132 0.07834761 23.58019 0.00000000 Constant -26.41665640 2.79471519 -9.45236 0.00000000 R**2= 0.04 R**2 Adj.= 0.04 DW=1.18 WALD TEST F(7)- 828 Significance Level 0.000 27 Table 10. TRADE ON OTHER MANUFACTURES Panel of Annual Data 1968-1991 N.Obs. 5175 INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif - GNP 0.77671380 0.01726196 44.99568 0.00000000 GNPPC 0.42108599 0.02296738 18.33409 0.00000000 D -1.09974739 0.04707855 -23.35984 0.00000000 T1 -6.17043301 0.45593362 -13.53362 0.00000000 T2 -5.58108246 0.38790671 -14.38769 0.00000000 REXR 0.93119898 0.09462534 9.84090 0.00000000 Constant -25.04135442 0.83441027 -30.01084 0.00000000 R**2- 0.58 R**2 Adj.- 0.58 DY-0.74 ANALYSIS OF VARIANCE: Source Sum of Squares Degrees Mean Square F-Statistic Sig.Leve1 INDIV 15783.765862590 217 72.736248215 25.443 0.00000000 ERROR 14171.178294314 4957 2.858821524 TOTAL 29954.944156904 5174 RANDOM EFFECTS=INSTRUMENTAL VARIABLES Variable Coeff Std Error T-Stat Signif ---=======--- --= GNP 0.97688932 0.05416433 18.03566 0.00000000 GNPPC 0.10911773 0.09517700 1.14647 0.25165312 D -1.36255730 0.17617493 -7.73412 0.00000000 T1 -4.66126145 1.76426536 -2.64204 0.00826564 T2 -4.47063117 1.24344620 -3.59536 0.00032697 REXR 1.08976835 0.07567980 14.39973 0.00000000 Constant -27.82908011 2.57947196 -10.78867 0.00000000 R**2= -0.06 R**2 Adj.- -0.06 DY=1.35 WALD TEST F(7)= 1651 Significance Level 0.000 28 Table 11.A. SIMULATED EFFECTS OF PTA -Expected Values - Total Trade Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 3.3 8.2 6.1 4.4 3.8 2,496.8 46,140.9 436.2 7,619.3 2,855.6 65.4% 116.8% 69.3% 54.8% 51.9% 11.4% 16.4% 15.0% 17.8% 34.2% Canada 3.3 8.2 6.1 4.4 3.8 352.4 3,803.6 20.1 714.8 194.6 9.2% 9.6% 3.2% 5.1% 3.5% 1.6% 1.3% 0.7% 1.7% 2.3% Mexico 3.3 8.2 6.1 4.4 3.8 6.1 569.5 1.5 26.2 79.1 0.2% 1.4% 0.2% 0.2% 1.4% 0.0% 0.2% 0.1% 0.1% 0.9% EEC 3.3 8.2 6.1 4.4 3.8 2,973.8 85,682.5 755.4 13,109.1 3,532.7 77.9% 216.8% 120.1% 94.2% 64.2% 13.5% 30.4% 26.0% 30.7% 42.3% Japan 1.8 5.0 3.7 2.6 2.2 682.1 18,869.1 293.9 4,548.5 943.8 17.9% 47.7% 46.7% 32.7% 17.1% 3.1% 6.7% 10.1% 10.6% 11.3% SAS 9.5 12.8 17.2 8.9 10.3 4,642.0 8,528.9 1,696.6 3,002.8 2,998.1 121.7% 21.6% 269.7% 21.6% 54.5% 21.1% 3.0% 58.5% 7.0% 35.9% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bi1atera1 Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 29 Table 11.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Total Trade Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 7 -1 20 -2 14 -2 10 -1 8 -1 5,065 -880 112,924 -12,301 999 -129 16,325 -2,492 5,951 -971 132.8%- 23.1% 285.7%- 31.1% 158.8%-20.5% 117.4%- 17.9% 108.1%- 17.6% 23.1%- 4.0% 40.0%- 4.4% 34.4%- 4.5% 38.2%- 5.8% 71.3%- 11.6% Canada 7 -1 20 -2 14 -2 10 -1 8 -1 715 -124 9,309 -1,014 46 -6 1,532 -234 406 -66 18.7%- 3.3% 23.6%- 2.6% 7.3%- 0.9% 11.0%- 1.7% 7.4%- 1.2% 3.3%- 0.6% 3.3%- 0.4% 1.6%- 0.2% 3.6%- 0.5% 4.9%- 0.8% Mexico 7 -1 20 -2 14 -2 10 -1 8 -1 12 -2 1,394 -152 4 -0 56 -9 165 -27 0.3%- 0.1% 3.5%- 0.4% 0.6%- 0.1% 0.4%- 0.1% 3.0%- 0.5% 0.1%- 0.0% 0.5%- 0.1% 0.1%- 0.0% 0.1%- 0.0% 2.0%- 0.3% EEC 7 -1 20 -2 14 -2 10 -1 8 -1 6,033 -1,048 209,698 -22,842 1,730 -224 28,088 -4,288 7,362 -1,201 158.1%- 27.5% 530.6%- 57.8% 275.0%-35.6% 201.9%- 30.8% 133.7%- 21.8% 27.5%- 4.8% 74.3%- 8.1% 59.7%- 7.7% 65.7%- 10.0% 88.2%- 14.4% Japan 3 -1 12 -1 8 -1 5 -1 4 -1 1,280 -261 43,678 -5,267 634 -91 9,113 -1,581 1,830 -344 33.6%- 6.8% 110.5%- 13.3% 100.8%-14.5% 65.5%- 11.4% 33.2%- 6.2% 5.8%- 1.2% 15.5%- 1.9% 21.9%- 3.2% 21.3%- 3.7% 21.9%- 4.1% SAS 27 -2 39 -2 54 -3 24 -2 29 -2 13,208 -887 25,692 -1,462 5,273 -285 8,213 -640 8,543 -584 346.2%- 23.3% 65.0%- 3.7% 838.3%-45.2% 59.0%- 4.6% 155.1%- 10.6% 60.1%- 4.0% 9.1%- 0.5% 181.8%- 9.8% 19.2%- 1.5% 102.3%- 7.0% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 30 Table 12.A. SIMULATED EFFECTS OF PTA -Expected Values - Textile Fibres Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 6.8 27.0 17.2 10.6 8.6 160.8 579.6 0.0 173.6 19.5 4.2% 1.5% 0.0% 1.2% 0.4% 0.7% 0.2% 0.0% 0.4% 0.2% Canada 6.8 27.0 17.2 10.6 8.6 7.4 22.3 2.5 10.0 0.0 0.2% 0.1% 0.4% 0.1% 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% Mexico 21.3 79.7 51.3 32.3 26.3 0.0 1.0 0.0 59.1 0.0 0.0% 0.0% 0.0% 0.4% 0.0% 0.0% 0.0% 0.0% 0.1% 0.0% EEC 6.8 27.0 17.2 10.6 8.6 82.2 1,887.9 4.4 988.7 82.0 2.2% 4.8% 0.7% 7.1% 1.5% 0.4% 0.7% 0.2% 2.3% 1.0% Japan 3.9 16.6 10.4 6.3 5.0 18.3 857.3 11.3 512.2 40.5 0.5% 2.2% 1.8% 3.7% 0.7% 0.1% 0.3% 0.4% 1.2% 0.5% SAS 483.4 1,867.3 3,994.9 304.2 930.9 38,221.2 33,012.4 2,109.3 23,198.1 11,886.3 1001.9% 83.5% 335.3% 166.8% 215.9% 174.1% 11.7% 72.7% 54.3% 142.4% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bi1atera1 Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. ~!J}~ 31 Table l2.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Textile Fibres Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 13 -3 68 -7 40 -6 22 -4 17 -4 303 -71 1,468 -161 0 -0 359 -66 39 -8 7.9%- 1.8% 3.7%- 0.4% 0.0%- 0.0% 2.6%- 0.5% 0.7%- 0.1% 1.4%- 0.3% 0.5%- 0.1% 0.0%- 0.0% 0.8%- 0.2% 0.5%- 0.1% Canada 13 -3 68 -7 40 -6 22 -4 17 -4 14 -3 56 -6 6 -1 21 -4 0 -0 0.4%- 0.1% 0.1%- 0.0% 0.9%- 0.1% 0.1%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.2%- 0.0% 0.0%- 0.0% 0.0%- 0.0% Mexico 45 -8 222 -19 130 -14 75 -11 58 -9 0 -0 3 -0 0 -0 137 -20 0 -0 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 1.0%- 0.1% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.3%- 0.0% 0.0%- 0.0% EEC 13 -3 68 -7 40 -6 22 -4 17 -4 155 -36 4,780 -523 10 -1 2,047 -378 162 -34 4.1%- 0.9% 12.1%- 1.3% 1.6%- 0.2% 14.7%- 2.7% 2.9%- 0.6% 0.7%- 0.2% 1.7%- 0.2% 0.3%- 0.0% 4.8%- 0.9% 1.9%- 0.4% Japan 7 -2 41 -5 24 -3 13 -2 10 -2 34 -8 2,127 -240 25 -4 1,047 -193 79 -16 0.9%- 0.2% 5.4%- 0.6% 4.0%- 0.6% 7.5%- 1.4% 1.4%- 0.3% 0.2%- 0.0% 0.8%- 0.1% 0.9%- 0.1% 2.5%- 0.5% 0.9%- 0.2% SAS 1,802 -65 7,931 -166 18,151 -287 1,014 -51 3,766 -99 142,465 -5,112 140,209 -2,934 9,583 -151 77,305 -3,923 48,083 -1,265 3734.3%-134.0% 354.8%- 7.4% 1523.6%-24.1% 555.8%- 28.2% 873.3%- 23.0% 648.8%- 23.3% 49.7%- 1.0% 330.4%- 5.2% 180.9%- 9.2% 576.1%- 15.2% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 32 Table l3.A. SIMULATED EFFECTS OF PTA -Expected Values - Fuels Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 4.6 22.7 13.5 7.7 6.0 0.4 9,243.9 43.8 341.1 0.6 0.0% 23.4% 7.0% 2.5% 0.0% 0.0% 3.3% 1.5% 0.8% 0.0% Canada 4.6 22.7 13.5 7.7 6.0 0.0 83.7 0.0 102.9 0.0 0.0% 0.2% 0.0% 0.7% 0.0% 0.0% 0.0% 0.0% 0.2% 0.0% Mexico 4.6 22.7 13.5 7.7 6.0 0.0 55.6 0.0 0.0 0.0 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% EEC 4.6 22.7 13.5 7.7 6.0 10.2 5,218.7 1.9 298.5 45.7 0.3% 13.2% 0.3% 2.1% 0.8% 0.0% 1.8% 0.1% 0.7% 0.5% Japan 4.6 22.7 13.5 7.7 6.0 14.7 982.2 2.5 219.7 36.0 0.4% 2.5% 0.4% 1.6% 0.7% 0.1% 0.3% 0.1% 0.5% 0.4% SAS 4.6 7.7 13.5 7.7 6.8 16.8 174.3 26.5 130.9 0.2 0.4% 0.4% 4.2% 0.9% 0.0% 0.1% 0.1% 0.9% 0.3% 0.0% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 33 Table l3.B. SIMULATEDEFFECTSOF PTA -Confidence Interval at 95% Fuels Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 11 -1 80 -3 42 -2 21 -2 15 -1 1 -0 32,659 -1,112 137 -7 934 -68 2 -0 0.0%- 0.0% 82.6%- 2.8% 21.8%- 1.1% 6.7%- 0.5% 0.0%- 0.0% 0.0%- 0.0% 11.6%- 0.4% 4.7%- 0.2% 2.2%- 0.2% 0.0%- 0.0% Canada 11 -1 80 -3 42 -2 21 -2 15 -1 0 -0 296 -10 0 -0 282 -20 0 -0 0.0%- 0.0% 0.7%- 0.0% 0.0%- 0.0% 2.0%- 0.1% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.7%- 0.0% 0.0%- 0.0% Mexico 11 -1 80 -3 42 -2 21 -2 15 -1 0 -0 196 -7 0 -0 0 -0 0 -0 0.0%- 0.0% 0.5%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% EEC 11 -1 80 -3 42 -2 21 -2 15 -1 25 -2 18,438 -628 6 -0 817 -59 118 -10 0.6%- 0.1% 46.7%- 1.6% 0.9%- 0.0% 5.9%- 0.4% 2.1%- 0.2% 0.1%- 0.0% 6.5%- 0.2% 0.2%- 0.0% 1.9%- 0.1% 1.4%- 0.1% Japan 11 -1 80 -3 42 -2 21 -2 15 -1 36 -4 3,470 -118 8 -0 601 -44 93 -8 0.9%- 0.1% 8.8%- 0.3% 1.2%- 0.1% 4.3%- 0.3% 1.7%- 0.1% 0.2%- 0.0% 1.2%- 0.0% 0.3%- 0.0% 1.4%- 0.1% 1.1%- 0.1% SAS 11 -1 21 -2 42 -2 21 -2 18 -1 41 -4 482 -34 83 -4 358 -26 0 -0 1.1%- 0.1% 1.2%- 0.1% 13.2%- 0.7% 2.6%- 0.2% 0.0%- 0.0% 0.2%- 0.0% 0.2%- 0.0% 2.9%- 0.1% 0.8%- 0.1% 0.0%- 0.0% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. . 34 Table 14.A. SIMULATED EFFECTS OF PTA -Expected Values - Non-Fuel Primaries Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 3.1 10.1 6.9 4.6 3.8 419.4 7,463.7 27.9 1,604.8 344.4 11.0% 18.9% 4.4% 11.5% 6.3% 1.9% 2.6% 1.0% 3.8% 4.1% Canada 3.1 10.1 6.9 4.6 3.8 216.2 1,297.6 1.8 94.2 37.7 5.7% 3.3% 0.3% 0.7% 0.7% 1.0% 0.5% 0.1% 0.2% 0.5% Mexico 4.3 13.1 9.1 6.2 5.1 0.2 109.2 0.7 8.9 6.5 0.0% 0.3% 0.1% 0.1% 0.1% 0.0% 0.0% 0.0% 0.0% 0.1% EEC 3.1 10.1 6.9 4.6 3.8 382.7 8,167.8 24.0 1,102.9 575.7 10.0% 20.7% 3.8% 7.9% 10.5% 1.7% 2.9% 0.8% 2.6% 6.9% Japan 3.1 10.1 6.9 4.6 3.8 119.3 11,316.7 4.4 267.4 145.1 3.1% 28.6% 0.7% 1.9% 2.6% 0.5% 4.0% 0.2% 0.6% 1.7% SAS 21.0 31.9 40.1 23.4 20.1 1,069.1 4,491.1 1,269.6 2,170.6 1,793.7 28.0% 11.4% 201.8% 15.6% 32.6% 4.9% 1.6% 43.8% 5.1% 21.5% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bi1atera1 Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 35 Table l4.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Non-Fuel Primaries Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 6 -1 25 -3 16 -2 10 -2 8 -1 807 -165 18,191 -2,063 63 -9 3,321 -569 687 -129 21.2%- 4.3% 46.0%- 5.2% 10.0%- 1.4% 23.9%- 4.1% 12.5%- 2.3% 3.7%- 0.7% 6.4%- 0.7% 2.2%- 0.3% 7.8%- 1.3% 8.2%- 1.5% Canada 6 -1 25 -3 16 -2 10 -2 8 -1 416 -85 3,163 -359 4 -1 195 -33 75 -14 10.9%- 2.2% 8.0%- 0.9% 0.7%- 0.1% 1.4%- 0.2% 1.4%- 0.3% 1.9%- 0.4% 1.1%- 0.1% 0.1%- 0.0% 0.5%- 0.1% 0.9%- 0.2% Mexico 9 -1 34 -3 22 -3 14 -2 11 -2 0 -0 283 -27 2 -0 20 -3 14 -2 0.0%- 0.0% 0.7%- 0.1% 0.3%- 0.0% 0.1%- 0.0% 0.3%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.2%- 0.0% EEC 6 -1 25 -3 16 -2 10 -2 8 -1 737 -150 19,907 -2,258 54 -8 2,283 -391 1,149 -215 19.3%- 3.9% 50.4%- 5.7% 8.6%- 1.2% 16.4%- 2.8% 20.9%- 3.9% 3.4%- 0.7% 7.1%- 0.8% 1.9%- 0.3% 5.3%- 0.9% 13.8%- 2.6% Japan 6 -1 25 -3 16 -2 10 -2 8 -1 230 -47 27,582 -3,128 10 -1 553 -95 290 -54 6.0%- 1.2% 69.8%- 7.9% 1.6%- 0.2% 4.0%- 0.7% 5.3%- 1.0% Y.,!"'.- 1.0%- 0.2% 9.8%- 1.1% 0.3%- 0.0% 1.3%- 0.2% 3.5%- 0.6% c~ "" SAS 79 -2 128 -3 161 -4 87 -3 72 -2 4,039 -104 17,965 -384 5,102 -116 8,073 -236 6,451 -211 105.9%- 2.7% 45.5%- 1.0% 811.2%-18.4% 58.0%- 1.7% 117.2%- 3.8% 18.4%- 0.5% 6.4%- 0.1% 175.9%- 4.0% 18.9%- 0.6% 77.3%- 2.5% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 36 Table IS.A. SIMULATED EFFECTSOF PTA -Expected Values - Leather, Dressed Fur., etc Bangladesh India Nepal Pakistan Sri Lanka POTENTIALPARTNERS USA 0.8 1.9 1.5 1.1 0.9 0.8 112.8 0.0 16.9 0.1 0.0% 0.3% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Canada 0.8 1.9 1.5 1.1 0.9 0.2 8.7 0.0 1.1 0.0 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Mexico 0.8 1.9 1.5 1.1 0.9 0.0 0.3 0.2 0.2 0.0 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% EEC 0.8 1.9 1.5 1.1 0.9 69.1 684.6 6.6 135.0 1.4 1.8% 1.7% 1.1% 1.0% 0.0% 0.3% 0.2% 0.2% 0.3% 0.0% Japan 0.8 1.9 1.5 1.1 0.9 9.6 37.3 0.1 15.0 0.3 0.3% 0.1% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% SAS 0.8 1.1 1.5 1.1 0.9 0.8 10.7 0.5 10.6 2.0 0.0% 0.0% 0.1% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA$. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 37 Table ls.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Leather, Dressed Fur., etc Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 2 -0 6 -0 4 -0 3 -0 2 -0 2 -0 338 -0 0 -0 45 -0 0 -0 0.1%- 0.0% 0.9%- 0.0% 0.0%- 0.0% 0.3%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% Canada 2 -0 6 -0 4 -0 3 -0 2 -0 1 -0 26 -0 0 -0 3 -0 0 -0 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% Mexico 2 -0 6 -0 4 -0 3 -0 2 -0 0- 0 1- 0 1- 0 1- 0 0- 0 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% EEC 2 -0 6 -0 4 -0 3 -0 2 -0 172 -0 2,049 -0 19 -0 356 -0 4 -0 4.5%- 0.0% 5.2%- 0.0% 3.0%- 0.0% 2.6%- 0.0% 0.1%- 0.0% 0.8%- 0.0% 0.7%- 0.0% 0.6%- 0.0% 0.8%- 0.0% 0.0%- 0.0% Japan 2 -0 6 -0 4 -0 3 -0 2 -0 24 -0 112 -0 0 -0 40 -0 1 -0 0.6%- 0.0% 0.3%- 0.0% 0.0%- 0.0% 0.3%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% SAS 2 -0 3 -0 4 -0 3 -0 2 -0 2 -0 28 -0 1 -0 28 -0 5 -0 0.1%- 0.0% 0.1%- 0.0% 0.2%- 0.0% 0.2%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.1%- 0.0% 0.1%- 0.0% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 38 Table l6.A. SIMULATED EFFECTS OF PTA -Expected Values - Textile Yarn Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 4.7 10.8 8.2 6.1 5.4 217.3 3,877.4 26.0 1,415.5 100.1 5.7% 9.8% 4.1% 10.2% 1.8% 1.0% 1.4% 0.9% 3.3% 1.2% Canada 4.7 10.8 8.2 6.1 5.4 20.1 384.0 1.0 290.9 15.1 0.5% 1.0% 0.2% 2.1% 0.3% 0.1% 0.1% 0.0% 0.7% 0.2% Mexico 3.5 8.3 6.2 4.6 4.0 3.0 17.1 0.2 5.4 0.0 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% EEC 4.7 10.8 8.2 6.1 5.4 396.6 7,924.9 540.8 3,106.4 199.8 10.4% 20.1% 86.0% 22.3% 3.6% 1.8% 2.8% 18.6% 7.3% 2.4% Japan 1.9 4.9 3.6 2.6 2.2 34.1 511.8 3.1 1,214.2 85.4 0.9% 1.3% 0.5% 8.7% 1.6% 0.2% 0.2% 0.1% 2.8% 1.0% SAS 14.2 17.2 27.0 9.9 13.1 2,682.7 2,981.4 140.6 844.8 737.1 70.3% 7.5% 22.4% 6.1% 13.4% 12.2% 1.1% 4.8% 2.0% 8.8% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA$. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 39 Table l6.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% Textile Yarn Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 10 -1 29 -2 20 -2 14 -2 12 -2 471 -70 10,316 -886 64 -7 3,260 -414 224 -31 12.3%- 1.8% 26.1%- 2.2% 10.2%- 1.1% 23.4%- 3.0% 4.1%- 0.6% 2.1%- 0.3% 3.7%- 0.3% 2.2%- 0.2% 7.6%- 1.0% 2.7%- 0.4% Canada 10 -1 29 -2 20 -2 14 -2 12 -2 44 -6 1,022 -88 2 -0 670 -85 34 -5 1.1%- 0.2% 2.6%- 0.2% 0.4%- 0.0% 4.8%- 0.6% 0.6%- 0.1% 0.2%- 0.0% 0.4%- 0.0% 0.1%- 0.0% 1.6%- 0.2% 0.4%- 0.1% Mexico 7 -1 21 -2 15 -2 10 -1 9 -1 6 -1 44 -4 0 -0 12 -2 0 -0 0.2%- 0.0% 0.1%- 0.0% 0.1%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% EEC 10 -1 29 -2 20 -2 14 -2 12 -2 860 -127 21,084 -1,811 1,339 -140 7,154 -909 446 -61 22.5%- 3.3% 53.4%- 4.6% 212.9%-22.3% 51.4%- 6.5% 8.1%- 1.1% 3.9%- 0.6% 7.5%- 0.6% 46.2%- 4.8% 16.7%- 2.1% 5.3%- 0.7% Japan 4 -1 12 -1 8 -1 6 -1 5 -1 68 -11 1,282 -115 7 -1 2,598 -357 176 -27 1.8%- 0.3% 3.2%- 0.3% 1.1%- 0.1% 18.7%- 2.6% 3.2%- 0.5% 0.3%- 0.1% 0.5%- 0.0% 0.2%- 0.0% 6.1%- 0.8% 2.1%- 0.3% SAS 41 -3 52 -3 87 -4 26 -2 37 -3 7,785 -525 8,954 -542 454 -23 2,214 -201 2,082 -154 204.1%- 13.8% 22.7%- 1.4% 72.2%- 3.6% 15.9%- 1.4% 37.8%- 2.8% 35.5%- 2.4% 3.2%- 0.2% 15.7%- 0.8% 5.2%- 0.5% 24.9%- 1.8% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 40 Table l7.A. SIMULATED EFFECTS OF PTA -Expected Values - Machines and Transport Equipment Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 0.9 2.4 1.8 1.3 1.1 22.3 1,936.2 6.3 354.0 23.5 0.6% 4.9% 1.0% 2.5% 0.4% 0.1% 0.7% 0.2% 0.8% 0.3% Canada 0.9 2.4 1.8 1.3 1.1 3.1 181.5 0.8 18.8 2.0 0.1% 0.5% 0.1% 0.1% 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% Mexico 0.9 2.4 1.8 1.3 1.1 0.0 26.2 0.0 0.1 0.1 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% EEC 0.9 2.4 1.8 1.3 1.1 153.4 5,556.9 41.7 1,037.9 156.8 4.0% 14.1% 6.6% 7.5% 2.8% 0.7% 2.0% 1.4% 2.4% 1.9% Japan 0.9 2.4 1.8 1.3 1.1 162.7 2,129.4 67.1 1,096.3 201.3 4.3% 5.4% 10.7% 7.9% 3.7% 0.7% 0.8% 2.3% 2.6% 2.4% SAS 0.9 1.1 1.8 1.1 1.1 54.9 144.3 29.5 7.6 59.2 1.4% 0.4% 4.7% 0.1% 1.1% 0.3% 0.1% 1.0% 0.0% 0.7% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989- 90 values were used. . . 41 Table 17.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Machines and Transport Equipment Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 2 -0 7 -0 5 -0 3 -0 3 -0 52 -2 5,494 -134 17 -0 879 -31 57 -2 1.4%- 0.1% 13.9%- 0.3% 2.6%- 0.1% 6.3%- 0.2% 1.0%- 0.0% 0.2%- 0.0% 1.9%- 0.0% 0.6%- 0.0% 2.1%- 0.1% 0.7%- 0.0% Canada 2 -0 7 -0 5 -0 3 -0 3 -0 7 -0 515 -13 2 -0 47 -2 5 -0 0.2%- 0.0% 1.3%- 0.0% 0.3%- 0.0% 0.3%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.2%- 0.0% 0.1%- 0.0% 0.1%- 0.0% 0.1%- 0.0% Mexico 2 -0 7 -0 5 -0 3 -0 3 -0 0 -0 74 -2 0 -0 0 -0 0 -0 0.0%- 0.0% 0.2%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% EEC 2 -0 7 -0 5 -0 3 -0 3 -0 359 -15 15,767 -385 111 -3 2,578 -91 378 -14 9.4%- 0.4% 39.9%- 1.0% 17.6%- 0.5% 18.5%- 0.7% 6.9%- 0.3% 1.6%- 0.1% 5.6%- 0.1% 3.8%- 0.1% 6.0%- 0.2% 4.5%- 0.2% Japan 2 -0 7 -0 5 -0 3 -0 3 -0 381 -16 6,042 -148 178 -5 2,723 -96 485 -18 10.0%- 0.4% 15.3%- 0.4% 28.4%- 0.8% 19.6%- 0.7% 8.8%- 0.3% 1.7%- 0.1% 2.1%- 0.1% 6.2%- 0.2% 6.4%- 0.2% 5.8%- 0.2% SAS 2 -0 3 -0 5 -0 3 -0 3 -0 128 -5 352 -13 79 -2 19 -1 143 -5 3.4%- 0.1% 0.9%- 0.0% 12.5%- 0.4% 0.1%- 0.0% 2.6%- 0.1% 0.6%- 0.0% 0.1%- 0.0% 2.7%- 0.1% 0.0%- 0.0% 1.7%- 0.1% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bi1atera1 Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 42 Table 18.A. SIMULATED EFFECTS OF PTA -Expected Values - Clothing Bangladesh India Nepal Pakistan Sri Lanka POTENTIALPARTNERS USA 7.1 40.7 23.0 12.5 9.5 3,385.3 29,699.0 1,227.3 3,685.0 4,823.8 88.7% 75.1% 195.1% 26.5% 87.6% 15.4% 10.5% 42.3% 8.6% 57.8% Canada 7.1 40.7 23.0 12.5 9.5 155.8 3,272.0 32.8 458.4 211.5 4.1% 8.3% 5.2% 3.3% 3.8% 0.7% 1.2% 1.1% 1.1% 2.5% Mexico 7.1 40.7 23.0 12.5 9.5 4.7 110.5 0.0 3.7 3.4 0.1% 0.3% 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% EEC 7.1 40.7 23.0 12.5 9.5 2,231.7 43,306.5 27.2 5,525.1 1,775.0 58.5% 109.6% 4.3% 39.7% 32.2% 10.2% 15.3% 0.9% 12.9% 21.3% Japan 7.1 40.7 23.0 12.5 9.5 6.2 3,798.8 5.7 134.4 112.8 0.2% 9.6% 0.9% 1.0% 2.0% 0.0% 1.3% 0.2% 0.3% 1.4% SAS 7.1 9.1 22.8 8.7 9.3 10.0 18.7 3.0 5.8 11.4 0.3% 0.0% 0.5% 0.0% 0.2% 0.0% 0.0% 0.1% 0.0% 0.1% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Totel Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989- 90 values were used. 43 Table 18.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Clothing Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 12 -4 92 -14 47 -9 23 -6 17 -5 5,769 -1,752 67,044 -10,292 2,522 -491 6,857 -1,697 8,591 -2,356 151.2%- 45.9% 169.6%- 26.0% 400.9%-78.0% 49.3%- 12.2% 156.0%- 42.8% 26.3%- 8.0% 23.8%- 3.6% 87.0%-16.9% 16.0%- 4.0% 102.9%- 28.2% Canada 12 -4 92 -14 47 -9 23 -6 17 -5 265 -81 7,386 -1,134 67 -13 853 -211 377 -103 7.0%- 2.1% 18.7%- 2.9% 10.7%- 2.1% 6.1%- 1.5% 6.8%- 1.9% 1.2%- 0.4% 2.6%- 0.4% 2.3%- 0.5% 2.0%- 0.5% 4.5%- 1.2% Mexico 12 -4 92 -14 47 -9 23 -6 17 -5 8 -2 249 -38 0 -0 7 -2 6 -2 0.2%- 0.1% 0.6%- 0.1% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.1%- 0.0% EEC 12 -4 92 -14 47 -9 23 -6 17 -5 3,803 -1,155 97,762 -15,008 56 -11 10,281 -2,544 3,161 -867 99.7%- 30.3% 247.4%- 38.0% 8.9%- 1.7% 73.9%- 18.3% 57.4%- 15.7% 17.3%- 5.3% 34.6%- 5.3% 1.9%- 0.4% 24.1%- 6.0% 37.9%- 10.4% Japan 12 -4 92 -14 47 -9 23 -6 17 -5 11 -3 8,576 -1,317 12 -2 250 -62 201 -55 0.3%- 0.1% 21.7%- 3.3% 1.9%- 0.4% 1.8%- 0.4% 3.6%- 1.0% 0.0%- 0.0% 3.0%- 0.5% 0.4%- 0.1% 0.6%- 0.1% 2.4%- 0.7% SAS 12 -4 16 -4 47 -9 15 -4 17 -5 17 -5 33 -9 6 -1 10 -3 20 -6 0.4%- 0.1% 0.1%- 0.0% 1.0%- 0.2% 0.1%- 0.0% 0.4%- 0.1% 0.1%- 0.0% 0.0%- 0.0% 0.2%- 0.0% 0.0%- 0.0% 0.2%- 0.1% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bi1atera1 Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 44 Table 19.A. SIMULATED EFFECTS OF PTA -Expected Values - Other Manufactures Bangladesh India Nepal Pakistan Sri Lanka POTENTIAL PARTNERS USA 3.0 7.5 5.6 4.1 3.5 161.7 18,162.2 21.8 1,972.4 318.1 4.2% 46.0% 3.5% 14.2% 5.8% 0.7% 6.4% 0.8% 4.6% 3.8% Canada 3.0 7.5 5.6 4.1 3.5 22.2 919.2 4.9 106.9 34.0 0.6% 2.3% 0.8% 0.8% 0.6% 0.1% 0.3% 0.2% 0.3% 0.4% Mexico 3.0 7.5 5.6 4.1 3.5 0.8 315.2 0.0 3.5 0.4 0.0% 0.8% 0.0% 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% EEC 3.0 7.5 5.6 4.1 3.5 348.5 34,950.3 136.9 2,845.3 1,148.8 9.1% 88.4% 21.8% 20.5% 20.9% 1.6% 12.4% 4.7% 6.7% 13.8% Japan 1.5 4.3 3.1 2.2 1.8 192.2 6,017.5 124.5 580.5 241.3 5.0% 15.2% 19.8% 4.2% 4.4% 0.9% 2.1% 4.3% 1.4% 2.9% SAS 13.2 17.7 21.7 12.6 14.9 1,085.2 2,440.6 801.0 254.9 765.2 28.4% 6.2% 127.3% 1.8% 13.9% 4.9% 0.9% 27.6% 0.6% 9.2% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA$. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bi1atera1 Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 45 Table 19.B. SIMULATED EFFECTS OF PTA -Confidence Interval at 95% - Other Manufactures Bangladesh India Nepal Pakistan Sri Lanka PARTNERS USA 6 -1 19 -2 13 -2 9 -1 7 -1 331 -55 45,545 -4,498 51 -6 4,291 -613 671 -103 8.7%- 1.4% 115.2%- 11.4% 8.1%- 1.0% 30.9%- 4.4% 12.2%- 1.9% 1.5%- 0.3% 16.1%- 1.6% 1.8%- 0.2% 10.0%- 1.4% 8.0%- 1.2% Canada 6 -1 19 -2 13 -2 9 -1 7 -1 46 -8 2,305 -228 11 -1 233 -33 72 -11 1.2%- 0.2% 5.8%- 0.6% 1.8%- 0.2% 1.7%- 0.2% 1.3%- 0.2% 0.2%- 0.0% 0.8%- 0.1% 0.4%- 0.0% 0.5%- 0.1% 0.9%- 0.1% Mexico 6 -1 19 -2 13 -2 9 -1 7 -1 2 -0 790 -78 0 -0 8 -1 1 -0 0.0%- 0.0% 2.0%- 0.2% 0.0%- 0.0% 0.1%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.3%- 0.0% 0.0%- 0.0% 0.0%- 0.0% 0.0%- 0.0% EEC 6 -1 19 -2 13 -2 9 -1 7 -1 714 -118 87,645 -8,656 320 -38 6,191 -884 2,424 -374 18.7%- 3.1% 221.8%- 21.9% 50.9%- 6.1% 44.5%- 6.4% 44.0%- 6.8% 3.2%- 0.5% 31.1%- 3.1% 11.0%- 1.3% 14.5%- 2.1% 29.0%- 4.5% Japan 3 -1 10 -1 7 -1 5 -1 4 -1 375 -64 14,442 -1,476 278 -34 1,208 -177 486 -77 9.8%- 1.7% 36.5%- 3.7% 44.3%- 5.4% 8.7%- 1.3% 8.8%- 1.4% 1.7%- 0.3% 5.1%- 0.5% 9.6%- 1.2% 2.8%- 0.4% 5.8%- 0.9% SAS 42 -2 59 -2 74 -3 39 -2 48 -2 3,442 -164 8,139 -334 2,737 -107 787 -42 2,468 -113 90.2%- 4.3% 20.6%- 0.8% 435.1%-17.1% 5.7%- 0.3% 44.8%- 2.1% 15.7%- 0.7% 2.9%- 0.1% 94.4%- 3.7% 1.8%- 0.1% 29.6%- 1.4% NOTE: The information on each cell is the following: row 1: Proportional Increase of Bilateral Trade. row 2: Value of Such Increase in Millions USA $. row 3: Ratio of Row 2 over Total Trade. row 4: Ratio of Row 2 over GNP. *Bilateral Trade, Total Trade and GNP are average of 1990-91 values, except for Nepal where 1989-90 values were used. 46 APPENDIX A: VARIABLES IN THE GRAVITY MODEL Table A.I. COUNTRYGROUPSAND CITIES USED IN THE GRAVITY MODEL'S ESTIMATION I GROUP OF COUNTRIES I CITIES I ~ Bangladesh Dacca India Bombay-Calcutta-Madras Nepal Kathmandu Pakistan Karachi Sri-Lanka Colombo ~ Canada Ottawa EEC 6 Paris-Rome EFTA Vienna-Stockholm Japan Tokyo Usa Chicago Rest OECD Sydney Other Groups AFRICA Algiers ASIA Manila Rest EUROPE Budapest-Alkara LATIN AMERICA Caracas-Sao Paulo MIDDLE EAST Riyadh Other China Beijing Countries Hong Kong Hong Kong Korea Korea Mexico Mexico City Singapore Singapore AGGREGATION: Analogies COUNTRY with IMF International Financial Statistics AFRICA Africa (IFS) ASIA Asia (IFS) excluding: SAS's Countries, China, Hong Kong, Korea and Singapore MIDDLE EAST Middle East (IFS) LATIN AMERICA Western Hemisphere (IFS) excluding Mexico Rest EUROPE Europe (IFS) Rest OECD Industrial Countries (IFS) excluding: Canada, EEC6, EFTA, Japan and USA 47 Table A.2. GROUPS TARIFFS ON COMMODITY USED IN THE ESTIMATION SITC Rev. 1 TOTAL 070 260 300 TNFP 610 650 700 840 TM Countries Bangladesh 22% 22% 22% 22% 22% 22% 22% 22% 22% 22% India 42% 42% 42% 42% 42% 42% 42% 42% 42% 42% Nepal 35% 35% 35% 35% 35% 35% 35% 35% 35% 35% Pakistan 28% 28% 28% 28% 28% 28% 28% 28% 28% 28% Sri-Lanka 25% 25% 25% 25% 25% 25% 25% 25% 25% 25% Canada 13% 2% 3% 0% 4% 10% 17% 6% 24% 10% EEC 6 13% 2% 3% 0% 4% 10% 17% 6% 24% 10% EFTA 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Japan 2% 7% 0% 2% 4% 14% 3% 0% 5% 0% Usa 13% 2% 3% 0% 4% 10% 17% 6% 24% 10% R. OECD 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% AFRICA 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% ASIA 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% R. EUROPE 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% L. AMERICA 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% MID. EAST 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% China 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% H. Kong 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Korea 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Mexico 13% 14% 10% 0% 9% 10% 12% 14% 20% 10% Singapore 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% OF DESCRIPTION COMMODITIES TOTAL Total Trade 070 Coffee, Tea, Cocoa and Spices 260 Textile Fibres 300 Fuels TNFP Non-Fuel Primaries (except coffee, tea, cocoa and spices, and Textile Fibres) 610 Leather, Dressed Fur., etc. 650 Textile Yarn, Fabric, etc. 700 Machines and Transport Equipment 840 Clothing TM Other Manufactures 48 APPENDIX B. SUMMARY OF SOUTH ASIA ACTUAL TRADE PATTERN Table B.l. BILATERAL TRADEIN SOUTHASIA Bangladesh India Nepal Pakistan Sri Lanka PARTNERS Total Trade 3,815 39,520 629 13,910 5,506 USA 762 5,600 71 1,717 746 20.0% 14.2% 11.3% 12.3% 13.5% Canada 107 462 3 161 51 2.8% 1.2% 0.5% 1.2% 0.9% Mexico 2 69 0 6 21 0.0% 0.2% 0.0% 0.0% 0.4% EEC 907 10,399 123 2,954 923 23.8% 26.3% 19.6% 21.2% 16.8% Japan 378 3,748 80 1,776 436 9.9% 9.5% 12.8% 12.8% 7.9% SAS 489 664 99 337 290 12.8% 1.7% 15.7% 2.4% 5.3% NOTE: The information in each cell is the following: row 1: Bilateral Trade, average 1990-91 in millions of dol1ars*. row 2: Share of Bilateral Trade on Total Trade. *Nepa1 data is 1989-90. 49 Table B.2. SOUTH ASIA IMPORTS BY COMMODITY Bangladesh India Nepal Pakistan Sri Lanka COMMODITY Total Trade 2,274 21,654 465 7,916 3,247 Coffee, Tea, etc. 17 12 8 194 19 0.7% 0.1% 1.6% 2.5% 0.6% Textile Fibres 200 230 23 195 38 8.8% 1.1% 5.2% 2.5% 1.1% Fuels 517 6,168 49 1,529 383 22.7% 28.6% 10.7% 19.4% 11.9% Non-Fuel Primaries 395 2,274 45 1,248 606 17.4% 10.4% 9.5% 15.9% 18.9% Leather, Dressed Fur. 3 77 0 8 7 0.1% 0.4% 0.0% 0.1% 0.2% Textile Yarn 195 186 29 128 592 8.6% 0.8% 6.1% 1.6% 17.8% Machines and T. Equip. 538 3,414 110 2,400 619 23.7% 15.6% 23.7% 30.0% 19.0% Clothing 43 2 7 0 14 1.9% 0.0% 1.4% 0.0% 0.4% Other Manufactures 365 9,293 194 2,215 968 16.1% 43.1% 41.7% 28.0% 30.0% NOTE: The information in each cell is the following: row 1: Imports for the specific commodity, average 1990-91 in millions of do11ars*. row 2: Ratio of row lover total imports. *Nepa1 data is 1989-90. 50 Table B.3. SOUTH ASIA EXPORTS BY COMMODITY Bangladesh India Nepal Pakistan Sri Lanka COMMODITY Total Trade 1,541 17,866 155 5,993 2,259 Coffee, Tea, etc. 37 794 2 19 581 2.4% 4.4% 0.9% 0.3% 26.1% Textile Fibres 116 329 1 506 19 7.5% 1.8% 0.9% 8.5% 0.9% Fuels 11 472 0 83 13 0.7% 2.6% 0.0% 1.4% 0.6% Non-Fuel Primaries 228 3,134 16 646 255 15.0% 17.5% 9.9% 10.8% 11.5% Leather, Dressed Fur. 100 742 8 292 13 6.5% 4.2% 5.4% 4.9% 0.5% Textile Yarn 310 2,167 72 2,865 37 20.2% 12.1% 46.7% 47.8% 1.6% Machines and T. Equip. 1 1,318 0 19 47 0.1% 7.4% 0.0% 0.3% 2.0% Clothing 688 2,720 47 1,190 860 44.4% 15.2% 30.1% 19.9% 37.5% Other Manufactures 50 6,190 10 372 435 3.2% 34.6% 6.1% 6.2% 19.4% NOTE: The information in each cell is the following: row 1: Exports for the specific commodity, average 1990-91 in millions of do11ars*. row 2: Ratio of row lover total exports. *Nepa1 data is 1989-90. 51 APPENDIX C. SUMMARY OF NTB ON SOUTH ASIA TRADE -MOST AFFECTED COMMODITIES- TOTAL TRADE I % coverage I USA I CAN I MEX I EEC I JPN I l%freq.-1 I I I I I BGD large 97.2 0.0 62.0 89.9 large 88.6 0.0 51.5 47.5 IND large 61.7 20.1 52.8 26.8 large 46.1 9.8 29.6 39.4 NPL large 41.2 0.0 5.1 51.1 large 38.5 0.0 9.1 7.1 PAK large 76.4 3.4 71.2 13.9 large 64.7 1.5 48.5 33.3 LKA large 80.4 86.9 42.6 43.5 large 53.5 25.0 41.8 35.8 AND SPICES COFFEE, TEA, COCOA I % coverage I USA I CAN I MEX I EEC I JPN I I%freq.~ I I I I I I BGD 0.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 0.0 100.0 IND 0.0 98.4 100.0 0.0 100.0 0.0 95.1 100.0 5.6 100.0 NPL 0.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 0.0 100.0 PAK 0.0 98.9 0.0 1.5 100.0 0.0 94.1 0.0 9.1 100.0 LKA 0.0 100.0 100.0 0.0 100.0 0.0 100.0 100.0 3.6 100.0 52 TEXTILE FIBRES I % coverage I USA I CAN I MEX I EEC I JPN I I%freq.-' I I I I I BGD 0.0 100.0 0.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 IND 0.0 12.1 0.0 1.5 0.0 0.0 25.0 0.0 7.5 0.0 NPL 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 PAK 0.0 0.0 0.0 0.0 1.2 0.0 0.0 0.0 9.1 8.3 LKA 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 AND SPICES) NON-FUELPRIMARY (EXCEPTCOFFEE,TEA, COCOA I % coverage I USA I CAN I MEX I EEC I JPN I l%freq.-1 I I I I I BGD 0.0 99.7 0.0 12.2 100.0 0.0 66.7 0.0 36.2 91.7 IND 0.0 62.1 99.9 17.7 39.1 0.0 48.4 93.3 33.5 66.9 NPL 0.0 100.0 0.0 1.0 98.7 0.0 100.0 0.0 14.3 33.3 PAK 0.0 91.7 100.0 56.5 92.0 0.0 50.0 100.0 39.0 78.6 LKA 0.0 32.2 0.0 10.7 79.0 0.0 19.4 0.0 36.4 69.1 53 LEATHER, DRESSEDFUR., ETC I % coverage I USA I CAN I MEX I EEC I JPN I l%freq.-1 I I I I I BGD 0.0 100.0 0.0 99.8 68.7 0.0 100.0 0.0 90.9 63.6 IND 0.0 71.5 0.0 92.0 87.0 0.0 71.4 0.0 82.7 68.0 NPL 0.0 0.0 0.0 100.0 33.3 0.0 0.0 0.0 100.0 33.3 PAK 0.0 97.6 0.0 89.4 98.4 0.0 70.0 0.0 85.4 76.9 LKA 0.0 0.0 0.0 42.1 0.0 0.0 0.0 0.0 85.7 0.0 TEXTILE YARN, FABRIC, ETC I % coverage I USA I CAN I MEX I EEC I JPN I l%freq.-1 I I I I I BGD 0.0 91.1 0.0 4.6 63.4 0.0 54.5 0.0 56.9 26.3 IND large 52.8 0.0 82.9 33.3 large 79.1 0.0 86.5 49.0 NPL 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 PAK large 62.9 0.0 99.9 2.5 large 68.9 0.0 94.0 24.2 LKA 0.0 0.0 0.0 83.4 10.0 0.0 0.0 0.0 82.5 42.9 54 CLOTHING , % coverage' USA I CAN I ME}{ I EEC I JPN I I%freq.-' I I I I I BGD large 100.0 0.0 99.6 0.0 large 100.0 0.0 87.8 0.0 IND large 93.6 0.0 94.5 0.0 large 89.7 0.0 84.7 0.0 NPL large 100.0 0.0 0.1 0.0 large 100.0 0.0 1.9 0.0 PAK large 96.4 0.0 91.2 0.0 large 88.4 0.0 84.9 0.0 LKA large 90.8 0.0 96.3 0.0 large 91.7 0.0 89.0 0.0 . . 55 APPENDIX D: Expected Value and Confidence Intervals of the Estimated Bilateral Trade Increase From equation (1) it follows (denoting the situation under a PTA with an asterisk and omitting all subscripts for clarity) that Log(BT)-Log(BT.) =Log(E-) =a4Log(T1) +a sLog (T2) +u-u' (A) BT. If we assume u and u* to be independently normally distributed with means zero and variance u2, then BT/BT* is lognormally distributed and hence, using E to denote the expectation operator, E(E-) =exp(a4Log(Tl) +asLog(T2)+u2)-rJ (B) BT. since the variance for u-u* is 2u2. It is natural to estimate the expectation in B by ~ -exp(a4Log(Tl) + asLog(T2) + u2) where the A over a parameter is its estimate from regression (1). If the regression is complete (i.e. the right-hand side includes all the relevant variables) and its error term is uncorrelated with the explanatory variables, the estimates of the parameters would be consistent and unbiased. As such, the estimate ~ would be a consistent though biased estimator of rJ. In fact, if the sample size is large enough for a normal approximation to the distribution of parameter estimates to be appropriate, E(~) = 11exp[iVariance (a4Log(Tl) + asLog(T2) + u2)]. Thus ~ would overestimate rJ. Our regression is certainly not complete, and we have already instrumented some of the explanatory variables to take into account their possible correlations between the error term. As such our parameter 56 estimates, while consistent, are certainly not unbiased; and the above theory is not quite applicable to the ij computed from the parameter estimates of our regression. We therefore took a pragmatic approach and used (C) below to calculate row 2 in each cell of Tables 11 to 19 in part A. (-) BT ..1-2 =exp(Ct4Log(T1) + Ct5Log(T2) +"7\"U1) (C) BT. L -2 2 -2 2 .2 -.. where u1=Log(T1) ua4 + Log(T2) ua5 + 2 Log(T1) Log(T2) ua4ua5Pa4a5 We are using this estimator since its bias appears to be less than that of the consistent estimator. Since Log(BT/BT*) is an increasing function of BT/BT*, if [I,m] represents an interval that includes, say, 95% percent of the distribution of 10g(BT/BT*) , then [exp(l),exp(m)] will also include 95% of the distribution of BT/BT*. Now, Log(BT/BT*) is normally distributed with mean Ct4 Log(T1) +a5 Log(T2) and variance u2. However, following the comment above, we have only consistent estimates for Ct4, Ct5 and u1 and with them we can calculate a "confidence" interval for mean prediction8. These limits are as follows: 1=a4Log(T 1) +a5Log(T2)-1. 96 (,1 (D) m=a4Log(T1) +a5Log(T2) +1. 96 (,1 (E) so that (exp t ,exp ill ) yields the required interval and defines the ranges in tables 11 to 19 part B. ~\Strictly speaking, we need to construct tolerance intervals (see T.N. Srinivasan (1979», but our large sample allows us to use confidence intervals. 57 REFERENCES Anderson, J. E. (1979), "A Theoretical Foundation for the Gravity Equation," American Economic Review, Vol. 69, P106-116. Bergstrand, J. H. (1989), "The Genera1ised Gravity Equation, Monopolistic Competition, and the FAcotr-Proportions Theory in International Trade," Review of Economics and Statistics, Vol. 71, pp. 143-153. Bergstrand, J. H. (1985), "The Gravity Equation in International Tocoomic Foundations and Empirical Evidence," Review of Economics and Statistics, Vol. 67i, pp. 474-481. Deardorff, A. (1984), "Testing Trade Theories and Predicting Trade Flows," Chapter 8 in Jones, R. and Kenen, P. (ed.), The Handbook of International Economics, Vol. I, Amsterdam: North-Holland. Frankel, J., E. Stein and S. Wei, "Trading Blocs: The Natural, the Unnatural, and the Super-Natural", Presented at the Sixth Inter American Seminar in Economics (NBER) in Caracas, Venezuela, May 1993. Linnemann, H. (1966), An Economic Study of International Trade Flows, North- Holland Publishing Company, Amsterdam. Poyhonen, Pentti (1963), "A Tentative Model for the Volume of Trade Between Countries," Welwirtschaftliches Archiv, 90 (1): 93-99. Tinbergen, Jan (1962), Sha~ing the World Economy: Suggestions for an International Economic Policy, New York. Safadi, R. and A. Yeats, "The North American Free Trade Agreement. It Effects on South Asia", WPS 1119, The World Bank, 1993. Srinivasan, T. N. and Gustavo Canonero (1993), "Liberalization of Trade Among Neighbors: Two Illustrative Models and Simulations," processed. Srinivasan, T. N. (1979), "Tolerance Intervals," processed. Wang, Zhen Kun and L. Alan Winters (1991), "The Trading Potential of Eastern 58 Europe," Centre for Economic Policy Research, Disucssion Paper No. 61.~