WPS #50 POLICY RESEARCH WORKING PAPER 1850 Determinants of There is a good deal ofintr industry trade between Intra-Industry Trade naonsin Central andstern between East and West Europe and the European Union, Most of it is vertical Europe Ithe exchange of similar goods of different qualtity. Chonira Aturupane Simeon Djankov Bernard Hoektman The World Bank Development Research Group November 1997 | POLICY RESEARCH WORKING PAPER 1850 Summary fincings Intra-industry trade as a share of total trade between Controlling for country effects, they find a statistically Central and Eastern European nations and the European significant positive association between horizontal intra- Union (EU) is among the highest of all the EU's bilateral industry trade (the exchange of close substitutes of trade flows. similar quality) and foreign direct investment, product Aturupane, Djankov, and Hoekman break down data differentiation, and industry concentration. They find a on these trade flows into horizontal and vertical significant negative relationship for economies of scale components and investigate the determinants of each. and labor intensity. They find that vertical intra-industry trade (the These results do not hold if they do not control for exchange of similar goods of different quality) accounts country effects, suggesting that country-specific factors for 80 to 90 percent of total intra-industry trade. It is are key determinants of horizontal intra-industry trade. positively associated. with product differentiation, labor intensity of production, economies of scale, and foreign direct investment. This paper - a product of the Development Research Group - is part of a larger effort in the group to analyze the role of trade and foreign investment in the process of transition in Eastern Europe. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Jennifer Ngaine, room N5-056, telephone 202-473-7947, fax 202-522-1159, Internet address trade@worldbank.org. November 1997. (32 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The I papers carry the names of the authors and should be cited accordingiy. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center Comments Welcome Determinants of Intra-Industry Trade between East and West Europe Chonira Aturupane* Simeon Djankov** Bernard Hoekman*** JEL code: F13; Keywords: Intra:industry trade, foreign direct investment, Eastern Europe * Stanford University; ** World Bank and University of Michigan; *** World Bank and CEPR. Conesponding author: bhoekman @worldbank.org; tel: 202-473-1185. Address: World Bank, IS1H St. NW, Washington DC 20433. We are grateful to Jim Tybout and Alan Winters for helpful suggestions. Summary There is a large empirical literature that investigates the determinants of intra-industry trade (IIT). Most studies find strong support for country effects such as size, distance, and relative income levels, but less evidence for the effects of the various industry- specific variables that theory suggests should be important. This is puzzling as most attempts to test theories that give rise to IIT focus on exchanges between highly developed economies. Given the relatively high degree of similarity of industrialized countries, one would expect to see strong support for the theory as regards industry- specific determinants of IIT. The existing literature focuses on trade flows that occur in the context of a relatively stable environment, with little change occurring in explanatory variables such as market structure or the size of technology and capital flows. In this paper we analyze the determinants of IIT between the European Union (EU) and eight Central and Eastem European countries (CEECs) during the 1990-95 period. These countries provide an interesting opportunity to improve our understanding of IIT. All CEECs are relatively industrialized and most have significant stocks of human capital. In conjunction with their geographic proximity to the EU and significantly lower real wages, there should have been significant scope for rapid growth in IIT after the collapse of central planning, driven by the opening of the economies and associated changes in managerial incentives, market structure and flows of technology. IIT between the EU and the CEECs has been growing rapidly. As of 1995, three countries (Czech Republic, Hungary, and Slovenia) were among the top ten countries in terms of the share of IIT in total trade with the EU. We follow the recent literature in distinguishing between horizontal and vertical IIT. Loosely defined, the latter consists of exchange of similar goods of different quality; the former comprises exchange of similar goods that are differentiated by characteristics rather than quality. In the CEEC context the distinction is particularly relevant because the level and growth in horizontal IIT is a good indicator of the extent to which the CEECs are "similar" to the EU. This in turn is an important consideration in terms of "convergence" and the prospects for accession to the EU. Our findings suggest that most IIT is vertical in nature: between 80 to 90 percent of total IIT with the EU is vertical. Horizontal IIT levels are less than half of those of countries such as Austria, Spain, or Switzerland and has been static over the 1990-95 period for the majority of countries. However, for the Czech Republic and Slovenia it has been growing rapidly and has attained levels that exceed those reported for countries such as Greece, Finland and Israel. After controlling for country-specific factors, vertical IIT is found to be positively associated with product differentiation, economies of scale, labor intensity of production, and foreign direct investment (FDI). A statistically significant positive association is also found between horizontal IIT and FDI, product differentiation and industry concentration, while a significant negative relationship is found for scale and the labor intensity of production. Only two of the coefficients (on FDI and scale economies) are significant if country dummies are not included in the regression. Overall, industry- specific factors explain less than 15% of horizontal IIT. From this we conclude that country-specific effects dominate industry-specific determinants of horizontal IIT. The empirical literature on IIT has generally found more support for the importance of country as opposed to industry factors. Given that vertical IIT accounts for most of the observed IIT between the EU and the sample of CEECs, one would have expected country factors to be particularly important determinants of vertical IIT. This is not the case: about 85% of the systemic variation in vertical IIT can be explained by industry- specific factors. The estimation results are quite robust when compared to existing studies on the determinants of IIT and its components. It can be hypothesized that this is due to the specifics of the initial post-reform period in the CEECs which were associated with a very significant opening of the economy to international competition, high levels of FDI (in 1995 the FDI-to-GDP ratio in the Czech Republic and Hungary was 17% and 15%, respectively); and substantial increases in the incentives to pursue product differentiation strategies following demonopolization and the break-up of the old conglomerates. The high share of vertical IIT that is observed is not surprising given the differences in relative real wages for comparable skill levels that existed between the EU and the CEECs and the geographic proximity of the CEECs to the EU. It is precisely these characteristics that make the CEECs particularly interesting in terms of investigating the effects of different industry-specific variables on IIT. I. Introduction There is a large empirical literature that investigates the determinants of intra-industry trade (IIT). Most studies find strong support for country effects, but little evidence for the effects of the various industry-specific variables that theory suggests should be important (Greenaway, Hine and Milner, 1995). This is puzzling as most attempts to test theories that give rise to IIT focus on exchanges between highly developed economies. Trade between such countries should be driven less by differences in endowments or technologies than North-South trade. Given the relatively high degree of integration of high income countries, past diffusion of know how, cross-hauling of foreign direct investment (FDI), movement of people, and so forth, one would expect to see strong support for the theory as regards industry-specific determinants of IIT. The existing literature focuses on trade flows that occur in the context of a relatively stable environment, with little change occurring in independent variables. There are no large shocks that affect managerial incentives, changes in market structure or the size of technology or capital flows. In this paper we analyze the determinants of IIT between the European Union (EU) and eight Central and Eastern European countries (CEECs) during the 1990-95 period. These countries provide a unique opportunity to improve our understanding of IIT. All CEECs are relatively industrialized and most have significant stocks of human capital. The scope for rapid growth in IIT after the collapse of central planning can be expected to have been substantial, driven by the opening of the economies and associated changes in managerial incentives, market structure and flows of technology. In conjunction with their geographic proximity and significantly lower 1 real wages, CEECs are a particularly appropriate set of countries for which to explore the effect of different industry-specific variables on IIT. Previous research has found that IIT has indeed been growing rapidly in the region. Much of the IIT that is observed at relatively high levels of aggregation comprises a pattern of trade where CEECs import intermediate inputs which are used to produce goods for export that are classified in the same industry.1 Studies that calculate IIT indices at more appropriate levels of disaggregation also find, however, that IIT has been rising rapidly.2 As of 1995, most CEECs had levels of IIT comparable (or higher) to those of Portugal, Greece, and Israel. Three countries (Czech Republic, Hungary, and Slovenia) were among the top ten countries in terms of the share of IIT in total trade with the EU. Existing studies of IIT between the EU and the CEECs do not distinguish between horizontal and vertical IIT. Loosely defined, the latter consists of exchange of similar goods of different quality and the former comprises exchange of similar goods that are differentia'ted by characteristics rather than quality. As argued by Abd-el-Rahman (1991) and Greenaway, Hine and Milner (1995) making such a distinction is important as the determinants of each type of LIT differs. In the CEEC context the distinction is particularly relevant because the level and growth in horizontal IIT is a good indicator of the extent to which the CEECs are "similar" to the EU. This in turn is an important Such trade reflects ongoing efforts by CEEC firms to upgrade production facilities and improve quality. See Hoekman and Djankov (1997) for an analysis of the importance of sourcing of inputs from the EU in changing the export structure of the CEECs. 2 See, e.g., Neven (1994). 2 consideration in terms of "convergence" and the prospects for accession to the EU. More generally, given that the empirical literature has come to ambiguous conclusions regarding the determinants of horizontal IIT, additional evidence from a data source that has not yet been explored is informative. The dataset that exists for the CEECs is of high quality and includes industry-specific variables that are of interest. Our findings suggest that vertical IIT accounts for 80 to 90 percent of total IIT with the EU, and that it is positively associated with product differentiation, economies of scale, labor intensity of production, and FDI. A statistically significant positive association is also found between horizontal IIT--the exchange of close substitutes of similar quality--and FDI, product differentiation and industry concentration, while a significant negative relationship is found for scale and the labor intensity of production. Only two of the coefficients (on FDI and scale economies) are significant if country dummies are not included in the regression. Overall, industry-specific factors explain less than 15% of horizontal IIT. From this we conclude that country-specific effects dominate industry-specific determinants of horizontal IIT. Conversely, about 85% of the systemic variation in vertical IIT can be explained by industry-specific factors. The paper is structured as follows. Section I1 briefly summarizes the literature on IIT. Section III describes the dataset and discusses summary descriptive statistics. Section IV turns to an econometric analysis of the determinants of IIT, using the explanatory variables that are commonly used in the literature. Section V concludes. 3 II. Literature Review Horizontal IIT arises when there is two-way trade in products of similar quality, but different characteristics or attributes. The theoretical basis for such trade was developed by Lancaster (1980), Krugman (1981), Helpman (1981, 1987) and Bergstrand (1990). These models suggest that the more similar countries are in terms of their endowments (incomes), the greater the share of horizontal IIT, which is driven by product differentiation and scale economies; the smaller the minimum efficient scale of production, the greater the number of firms in an industry, the greater the number of varieties supported by the market and the greater the magnitude of IIT. Vertical IIT involves simultaneous export and import of similar goods of varying qualities. The theoretical basis for this type of IIT was first developed by Falvey (1981), who showed that vertical IIT may arise in situations where large numbers of firms produce varieties of different qualities but there are no increasing retums in production. The pattem of vertical IIT follows traditional endowment-based models, with the relatively capital abundant country exporting higher quality products and the relatively labor-abundant country exporting lower quality goods. Shaked and Sutton (1984) showed that vertical IIT may also arise in market structures with small numbers of firms and increasing returns. No clear predictions therefore arise regarding the impact of scale or concentration as a determinant of vertical IIT. However, as in the case of horizontal IIT, the greater the number of varieties supported by the market, the more vertical IIT is observed in equilibrium. 4 Although the general presumption in the literature is that multinational activity and IIT are positively correlated, the relationship between FDI and IIT is ambiguous. Vertical IIT is likely to be associated with the presence of inward FDI, as foreign firms can be expected to combine their technological knowledge with local endowments to produce goods of varying qualities that are then shipped to export markets. In the case of horizontally differentiated products, FDI may substitute for exports of the goods that were previously produced in the investor's home country (Markusen and Venables, 1996). Whether this would reduce IIT depends on the export structure of the industry in the foreign country prior to entry by the multinational. If the industry did not produce similar goods or if the foreign entrants have positive net exports, horizontal IIT may increase. Helpman and Krugman (1985) conclude that multinational activity will be positively correlated with horizontal IIT once country-specific effects are controlled for. The empirical literature has focused on "testing" all or a subset of the industry- specific and country-specific determinants of IIT predicted by theory. These studies have generally found more empirical support for country-specific (i.e., endowments; income levels, distance) than industry-specific hypotheses (market structure, scale, product differentiation). Estimated coefficients on proxies for product differentiation and scale economies have often been insignificant or of the wrong sign, and the explanatory power of estimated equations is frequently very low. Greenaway, Hine and Milner (1994, 1995) argue that this may be the result of mis-specification, in particular the failure to distinguish horizontal from vertical IIT.3 3They conclude that the determinants of vertical and horizontal IIT differ, but not always in the expected manner. For the UK, vertical IIT appears to be better supported by models with large numbers of 5 Ethier (1982), Harrigan (1995) and Tybout (1993) all note that the appropriateness of regressing IIT indices on measures of scale or product differentiation is questionable, as the Grubel-Lloyd index is invariant to changes in these variables in the standard trade model with monopolistic competition. Moreover, Deardorff (1995) has demonstrated that reduced form equations where bilateral trade is regressed on income and distance can be consistent with a wide range of theoretical models, including neoclassical ones where there is no role for scale economies or imperfect competition. The implication of this is that regression analyses of the type commonly found in the literature cannot be regarded as tests of specific hypotheses or theories, and that no strong priors can be maintained as regards the signs of coefficient estimates that emerge from such exercises. Notwithstanding these methodological criticisms, we follow the recent literature in focusing on the industry-specific determinants of vertical IIT and horizontal IIT, while controlling for country-specific factors. This approach is motivated in large part by our interest in investigating the role of IIT in the process of transition and exploring where the CEECs stand in relation to the EU and the EU's other trading partners. It also makes it easier to compare with the results of previous studies on IIT, based on North-North country data. The use of country dummies is motivated by the absence of reliable data on incomes (GDP) and endowments for the CEECs. More generally, it allows us to distinguish country from industry-specific effects. As noted by Hummels and Levinsohn firms, but this is not the case for horizontal IIT. Scale economies were found to be significant only for horizontal IIT, while FDI was not a significant detenninant of either type of IIT. In a more recent analysis of intra-EU IIT, Fontagne, Freudenberg and P6ridy (1997) find that FDI and scale are positively associated with both horizontal and vertical IIT, while product differentiation is positive for vertical and negative for horizontal IIT. 6 (1995), the former include more than the incomes and distance variables commonly used in empirical work. Indeed, they conclude that country-pair dummies do more to explain bilateral IIT than differences in relative factor endowments. We use this insight by proxying for the multitude of country-pair factors that determine IIT with a fixed country- pair effect (one of the trading partners always being the EU). III. Data and Measurement Levels of IIT between eight CEECs and the EU(9)4 is calculated for the 1990-95 period at the 6-digit level of disaggregation of the EU's Combined Nomenclature (equivalent to the Harmonized System). Data was obtained from COMEXT, Eurostat's trade database, using the EU as the reporter for both import and export flows. There are 5,019 six-digit product categories, which were concorded to the 3-digit NACE industry classification as provided in the EUROSTAT COMEXT software. The full sample covers 109 NACE industries5 across the 8 CEECs, giving us a cross-section of 872 observations. 4 Bulgaria, Czech Republic, Hungary, Moldova, Poland, Romania, Slovak Republic and Slovenia. In order to be able to compare CEEC data with those of other European countries we have excluded Austria, Finland, Greece, Portugal and Spain from the EU. The resulting EU (9) includes Belgium, Luxembourg, Germany, France, the United Kingdom, Italy, the Netherlands, Denmark, and Ireland. Belgium-Luxembourg is reported as one aggregate. The original sample consisted of 194 3-digit NACE industries. We exclude all agriculture-related and service sectors. Industries that correspond to the CN categories 460000 (wickerwork and basketwork), 910000 (clocks, watches and parts thereof), 920000 (musical instruments, parts and accessories), 930000 (arms and ammunition), 960000 (miscellaneous manufactured articles), 970000 (works of art, antiques), 980000 (power production) and 990000 (other products) are excluded from the sample due to data limitations and reporting problems. 7 We use the adjusted Grubel-Lloyd (1975) index: E -Xj mXJkl 1] (1) IlTik = 1 E k)J00 E(Xdjk + Mok)J where i refers to the 6-digit product categories that make up each 3-digit "industry"j and k identifies countries. The index of IIT varies between 0 (complete inter-industry trade) and 100 (complete intra-industry trade). Following Greenaway, Hine and Milner (1995), horizontal IIT is defined to exist for trade in product i in industry j that satisfies the criterion: 1- a  exportUV- 1+ importUV#k Vertical IIT comprises trade where: exportUV >+a importUVijk importUVijk Relative unit values of exports and imports are utilized to disentangle horizontal from vertical IIT. The underlying assumption is that relative prices tend to reflect differences in qualities. Thus, vertical IIT is defined as two-way trade in a 6-digit product whose per kilogram unit value of exports (measured f.o.b.) relative to its per kilogram unit value of imnports (measured c.i.f.) falls outside a specified range of ħa Trade in products whose relative unit values fall within the range ħa is defined as horizontal IIT. Once IIT has been separated into the two types at the 6-digit level, trade flows are aggregated over the 6-digit categories to compute vertical and horizontal IIT at the 3-digit industry level. As 8 in Abd-el-Rahman (1991) and Greenaway, Hine and Mvilner (1995), we use a unit value dispersion of 15 percent (i.e., a=0.15) for the analysis, as well as a=0.25 as a robustness check. Descriptive statistics on unit values for the eight CEECs during 1993-1995 are reported in Table 1A. They illustrate that the significant variance in unit values across countries. The values for the Czech Republic, Hungary, Poland and Slovenia are somewhat lower than what is observed for comparator countries such as Greece, Portugal or Spain, but the difference is not very large. Tables 2A-4A report summary statistics for total IIT, horizontal IIT and vertical IIT (a-+15%) between the 8 CEECs and the EU(9) as well as between 31 comparator countries and the EU(9) for the years 1990-1995. In addition, Tables 5A and 6A present data at the a=ħ25% level for the eight CEECs. The numbers reported are not absolute levels of IIT but shares in gross industry trade, i.e. zEXj mj)z(|,kmj| _______M_k (~ Uk - Ill IIT(z)jk= + 1-F[ Z + ) xl00, l Xijk +M#jk) ( Xijk + M#k) where i refers to the 6-digit CN products in each 3-digit industry, j is a subscript for the 3- digit industry, and z varies over horizontal and vertical IIT. All the CEECs display relatively high levels of IIT. The Czech Republic has the highest share of IIT in the sample (42.5%), the fifth largest of any EU trading partner for 1995. Hungary and Slovenia are also among the "top ten" countries in terms of the share of IIT in total trade with the EU. In general, average IIT indices for the 8 CEECs are similar and show little variation over the reported period. The exceptions are the Czech 9 Republic, which has an above average mean level of IIT of approximately 43% during 1993-1995, and Moldova, where the mean IIT varies significantly from year to year.6 The extent to which vertical IIT dominates horizontal IIT for all eight CEECs is striking. Vertical IIT accounts for 80 to 90 percent of total IIT. The horizontal IIT levels are similar to those observed for Finland, Greece, Israel, Portugal and Tunisia, and are less than half the level of countries such as Austria, Spain, or Switzerland. Noteworthy is also that horizontal IIT has been static over the 1990-95 period for most countries, the only exceptions being the Czech Republic and Slovenia. Similar conclusions obtain if a is set at 25%. Industry specific variables are calculated using firm-level data from a comprehensive enterprise dataset on CEECs. A detailed description of the dataset can be found in Pohl et al. (1997). The data contain balance sheets and profit and loss statements for 1992-95 for the eight CEECs in our sample, obtained from private firms (Czech Republic and Hungary) or central statistical offices (Bulgaria, Moldova, Poland, Romania, Slovak Republic, and Slovenia). Typically, the data are annual observations at the plant level and cover the majority of plants in manufacturing industries. Two types of selection bias are present: "informal" enterprises are excluded and small firms are under- represented. The sample primarily covers medium and large enterprises in the formal sector. We also used the 4-digit CN disaggregation to calculate IIT, HIIT, and VIIT (not reported). The results proved to insensitive, to the initial level of disaggregation - the IIT shares derived from the 6- and 4- digit levels are similar. The only measurable difference is that the numbers for Moldova do not display significant variation anymore. 10 In an attempt to use similar data across all eight countries, we have restricted the samples to firms that have more than twenty-five workers. The exclusion of small firms undoubtedly presents a possible problem in terms of capturing the true extent of, say, FDI flows to the eight CEECs. Since foreign investors are, however, likely to be attracted by firms with significant market power, the results are probably not affected significantly.7 We exclude all firms which have missing observations in 1995. The majority of the excluded firms reported in 1992-94, which suggests that they may not be liquidated, but simply failed to report. This could give rise to a selection bias if smaller firms (or firms without FDI, etc.) are more likely to exit (or not turn in their reports), leading to an overestimate of all our variables, but particularly the industry concentration variable. This will, however, be the case for all countries -- a priori we cannot sign the selection bias that results from this data cleaning. The data include detailed information on firm revenues and expenditures, as well as its ownership status and equity stakes of strategic investors. A firm is regarded as "foreign" when more than a third of its shares are foreign-owned. This choice was made based on the existing corporate laws in the Central and Eastern European countries. In all eight countries, major strategic and inr estment decisions at the finns' Board of Directors can be taken with only two-thirds majority. Thus if more than one-third of shares are owned by foreign nationals they can block decisions of the Board. 7 In four countries (Bulgaria, Romania, Slovak Republic, Slovenia) we have the complete industrial census, including firms with less than twenty-five workers. We construct the four explanatory variables using the whole population of firms, and then the truncated sample with finns which employ more than twenty-five workers. Since the resulting variables (not reported) do not differ significantly, we proceed with truncated samples in all eight countries. 11 IV. Estimation Consistent with the literature on the determinants of IIT, we estimate a regression model of the following form: IITk(Z) = 0o + I3lLABik + P2CONCjk + I33FDIjk + I4MESJk + ,SPDjk + P6BGR + 7CZE + J8HUN + PgMDA + PIOROM + PI jSVK + U2SVN+cjk where IIT(z-total): 3-digit industry j IIT between country k and the EU(9) IIT(z--H): 3-digit industry j HIIT (ħ15% ) between country k and the EU(9). IIT(zV): 3-digit industry j VIIT (ħ15% ) between country k and the EU(9). LAB: The inverse of the share of energy in total costs CONC: Four fimn sales concentration ratio FDI: FDI output as a share of industry total MES: Minimum efficient scale: ratio of output of top 4 firms to rest of industry PD: Number of 8-digit categories in a 3-digit industry BGR: Bulgaria country dummy CZE: Czech Republic country dummy HUN: Hungary country dummy MDA: Moldova country dummy ROM: Romania country dummy SVK: Slovak Republic country dummy SVN: Slovenia country dummy The four firm sales concentration ratio (CONC) is a proxy for the influence of market structure on IIT. Existing theory suggests markets with a large number of firms are more likely to generate horizontal IIT than markets with a small number of firms.8 Therefore, For example, Lancaster (1980) demonstrates that a market structure of perfect monopolistic competition will necessarily lead to a high degree of horizontal IIT. Models have been developed where horizontal IIT occurs in a smail numbers setting, but large number models "form the dominant paradigm" (Greenaway, Hine and Milner, 1995, p. 1507). 12 the expected sign on 12 is negative for horizontal IIT. Theory is more ambivalent regarding the effect of market structure on vertical IIT. Thus, 12 may be greater or less than zero depending on whether a small or a large number model applies. The minimum efficient scale of production (MES) is measured as the ratio of gross value-added per employee in the largest four firms to gross value-added per employee in the remaining firms. The expected sign for this variable on horizontal IIT is negative, as low scale economies will lead to easier entry, a greater number of monopolistically competitive firms and thus more varieties and increased IIT. The predicted effect of scale on vertical IIT depends on market structure and may therefore be positive or negative. The product differentiation variable (PD) is defined as the number of 8-digit CN product categories in each 3-digit NACE sector.9 The expected signs are 135>0 for horizontal IIT since this type of IIT is directly related to the existence of differentiated products. Conversely, we expect P35<0 for vertical IIT. In addition to the foregoing variables, we also investigate the relevance of foreign direct investment (FDI) and labor intensity (LAB) for IIT. FDI is generally hypothesized to be positively associated with the level of IIT, as multinationals are often multi-product firms. One result of FDI is greater specialization in production by plants located in different countries, giving rise to more IIT, both horizontal and vertical. We therefore expect the sign on FDI (A3) to be positive for both types of IIT. Given the absence of reliable data on labor utilization in the CEEC context, the inverse of the share of energy in total costs is used as an indicator of labor intensity. The higher the energy intensity of 9 There are a total of 11,257 8-digit categories in the EUROSTAT database. 13 an activity, the lower will be the share of labor in total value added. This suggests that there will be less scope for vertical IIT, as variations in quality will generally be associated with activities that allow variations in inputs of skilled and unskilled labor. More specifically, in the CEEC context industries with high energy use (fertilizers, basic metals, plastics/rubber) were confronted with large increases in input costs as energy subsidies were eliminated. Some also became subject to greater pricing scrutiny in export markets (through antidumping and related policies). Such factors implied greater pressures to (a) "price to market" and (b) differentiate output to compete with foreign producers. This in turn could be reflected in an observed rise in horizontal IIT. We therefore expect the sign on PI to be negative for vertical and positive for horizontal IIT. Descriptive statistics for all five independent variables are reported in Table 1. Table 1: Descriptive Statistics of the Explanatory Variables, 1995 (3-digit NACE) CONG Bulgaria Czech Hungary Moldova Poland Romania Slovakia Slovenia Mean 0.443 0.346 0.354 0.553 0.287 0.312 0.366 0.441 Median 0.396 0.252 0.325 0.458 0.272 0.296 0.287 0.362 St. Dev 0.287 0.224 0.245 0.264 0.174 0.189 0.229 0.234 LAB I Xl l l l Mean 8.623 9.615 9.258 7.756 8.843 7.813 9.005 9.345 Median 10.981 11.364 9.806 8.064 8.621 8.064 9.176 9.432 Std. Dev 16.358 25.643 24.392 20.833 29.415 22.241 24.395 31.254 FDI Mean 0.032 0.174 0.198 0.021 0.089 0.042 0.052 0.096 Median 0.000 0.113 0.143 0.000 0.010 0.000 0.000 0.035 Std. Dev 0.058 0.207 0.224 0.126 0.149 0.096 0.112 0.146 MES Mean 1.023 1.042 1.028 0.997 1.016 0.967 1.047 1.073 Median 1.002 1.016 1.012 0.978 0.987 0.961 1.026 1.031 Std. Dev 0.317 0.172 0.178 0.315 0.254 0.274 0.172 0.198 PD Mean 117.968 117.968 117.968 117.968 117.968 117.968 117.968 117.968 Median 63.000 63.000 63.000 63.000 63.000 63.000 63.000 63.000 Std. Dev. 162.591 162.591 162.591 162.591 162.591 162.591 162.591 162.591 14 Learner (1994) has argued that it is important to look at the simple correlation matrix between dependent and independent variables as it can be quite difficult to interpret the partial correlations that emerge from the regression analysis. Table 2 reports the simple correlations between all variables used in the analysis. There is a relatively strong negative relationship between labor intensity and IIT, and a strong positive correlation between FDI and IIT. Correlations between the explanatory variables and horizontal IIT are quite low, although it can be noted that the highest (negative) correlations are with MES and CONC. Correlations with vertical IIT are very similar to those with total IIT: there is a high positive correlation with FDI, a substantial positive correlation with LAB, and no correlation with CONC and MES. FDI is positively associated with LAB, suggesting that FDI has been going into relatively labor-intensive sectors. This is consistent with the high correlation between FDI and vertical IIT, as the latter will involve activities where there is scope for quality differentiation through employment of more labor intensive techniques that build on lower labor costs in the CEECs. Note also that the correlation between CONC and MES is not very high. Table 2: Correlation Matrix, 1995. IIT HIIT(15%) VIIT(15%) LAB CONC FDI MES PD IIT 1.000 HIIT(15%) 0.416 1.000 VIIT(15%) 0.895 -0.038 1.000 LAB 0.358 0.072 0.351 1.000 CONC -0.089 -0.152 -0.033 -0.178 1.000 FDI 0.618 0.091 0.637 0.245 0.072 1.000 MES -0.027 -0.128 0.021 -0.154 0.145 -0.001 1.000 PD 0.156 0.038 0.152 -0.016 0.206 0.051 0.001 1.000 15 As mentioned previously, country dummies are used to capture the country- specific determinants of VIIT and HIIT which are generally assumed to include factors such as incomes, distance, and differences in endowments. As our primary interest is to explore the significance of industry-specific variables as determinants of IIT, the use of country dummies is an effective way of controlling for country-specific effects. Given the widespread presence of zero observations on trade flows at the 6-digit level, we follow Balassa and Bauwens (1987) in using nonlinear least squares to estimate the following logistic function: 1 IIT(z)uk= 1 + exp(-b' Xijk) + where b 'is the regression coefficients vector, x the explanatory variables vector and S is the random disturbance term. In order to correct for possible heteroscedasticity in the disturbances, all regressions were estimated with heteroscedastic consistent standard errors. V. Regression Results The results of the estimation for IIT, VIIT and HIIT at the a=ħl 5% with (1) and without (2) country dummies are reported in Table 3. For total IIT, labor intensity, FDI and the product differentiation proxy are statistically significant, the first two variables having by far the largest coefficient estimates. The adjusted R2 is 0.599, which is quite high for cross-section regressions of this type. Most of the country dummies are not significant. If the regression is run without the dummies, the goodness of fit does not decline very 16 much, and the concentration and scale variables become significant. This suggests country-specific variables are not very important determinants of IIT. The relative unimportance of the country dummies is somewhat surprising in light of the literature, which concludes that these are generally more robust explanatory factors than industry variables. The fit of the estimation for horizontal IIT is less good than that for IIT as a whole: the adjusted R2 falls to 0.372. Compared to earlier work this is nonetheless relatively high.10 The sign of the coefficient estimate on LAB is negative as expected. The MES coefficient also has the expected negative sign and is significant, while the coefficients on FDI, CONC and PD are positive and again significant. Without the country dummies the explanatory power of the equation drops to 0.059, and only FDI and MES remain significant. It therefore appears that horizontal IIT is driven primarily by country-specific effects. If account is taken of the wide differences in distances from the eight CEECs in the sample to the EU, as well as in per capita incomes--Moldova, the poorest country has an estimated per capita income level that is one-tenth that of Slovenia, the richest country--this result is not that surprising. The VIIT results are much closer to those obtained for total IIT (R2 of 0.556), reflecting the fact that VIIT accounts for 80 to 90 percent of total IIT (see the Appendix Tables). FDI has the predicted positive sign and is highly significant. The product differentiation and scale variables are also positive and significant but the concentration 10 Greenaway, Hine and Milner (1995) obtain an R2 of only 0.06, while Greenaway, Milner and Elliot (1996) obtain an R2 of 0.12 in a regression that adds country-specific explanatory variables such as income levels and distance. Fontagne et al. (1997) obtain an adjusted R2 of 0.46 in a panel setting for intra-EU IIT that includes country, industry, and policy variables. 17 variable is not significant. As is the case for total IIT, the fit of the equation is not very sensitive to the inclusion of country dummies. Table 3: Nonlinear Least Squares Estimation Results (109 3-digit NACE sectors at aħ1 5%) Independent IIT(1) IIT(2) HIIT(1) HIIT(2) VIIT(1) VIIT(2) Variables Constant -0.099 -0.207 2.055 -2.863 -0.318 -0.445 (-0.461) (-1.185) (1.711) (-5.335) (-1.787) (-2.008) LAB 0.118 0.053 -0.233 -0.138 0.125 0.083 (9.536) (5.362) (-2.638) (-1.322) (8.372) (3.187) CONC -0.023 -0.284 4.169 -0.137 -0.055 -0.218 (-0.157) (-2.234) (8.241) (-0.394) (-0.435) (-1.536) FDI 3.132 3.314 1.269 2.069 2.633 2.805 (13.852) (17.536) (4.761) (7.316) (15.253) (10.078) MES 0.171 0.371 -10.079 -1.141 0.469 0.643 (1.233) (3.054) (-7.769) (-2.659) (4.108) (4.548) PD 0.001 0.001 0.002 0.004 0.001 0.001 (8.432) (8.393) (4.663) (1.458) (7.559) (7.578) BGR -0.112 -3.578 -0.052 (-1.387) (-3.834) (-0.566) CZE 0.354 2.504 0.268 (5.206) (3.254) (3.305) HUN -0.213 0.271 -0.186 (-2.774) (0.586) (-2.161) MDA -0.931 -3.374 -0.806 (-6.263) (-5.486) (-6.641) ROM 0.062 -4.032 -0.006 (0.742) (5.024) (-0.062) SVK 0.182 0.793 0.106 (2.361) (1.587) (1.255) SVN 0.145 -- 1.716 -- -0.145 -- (1.795) (3.735) (-1.654) Number of obs: 872 872 872 872 872 872 Adjusted R2 0.599 0.501 0.373 0.059 0.556 0.489 Notes: t-statistics are in parentheses. CONC and MES may be highly correlated resulting in multicollinearity problems in the estimation. However, examination of the correlation matrix gave no indication of strong collinearity between the two variables. 18 To test the robustness of our results to the definition of horizontal and vertical IIT, we re-run the regressions using the data for HIIT and VIIT using a=+25% as the criterion (Table 4). No significant differences with the results reported in Table 3 emerge if country dummies are included. Table 4: Nonlinear Least Squares Estimation Results (109 3-digit NACE sectors at a=ħ25%) Independent HIIT(1) HIIT(2) VIIT(1) VIIT(2) Variables Constant -0.314 -1.448 -0.894 -0.856 (0.789) (-3.857) (-4.759) (-3.547) LAB -0.064 -0.082 0.106 0.064 (-1.875) (-3.182) (6.528) (3.498) CONC 0.614 -0.682 0.133 -0.046 (1.945) (-2.347) (1.008) (-0.297) FDI 2.534 2.224 2.168 2.297 (10.428) (9.487) (12.984) (9.1458) MES -1.592 -0.824 0.659 0.831 (-4.767) (-2.774) (5.513) (5.402) PD 0.001 0.001 0.001 0.001 (4.164) (2.627) (7.738) (7.563) BGR -1.687 0.169 (-3.245) (1.748) CZE 0.514 0.458 (3.104) (5.287) HUN 1.185 0.014 (6.304) (0.164) MDA -1.128 -0.662 (-3.498) (-5.014) ROM -0.194 0.081 (1.144) (0.834) SVK -0.384 0.327 (-2.067) (3.628) SVN 0.542 -0.143 (4.325) (-1.465) Number of obs: 872 872 872 872 Adjusted R2 0.274 0.115 0.504 0.424 Note: t-statistics are in parentheses. 19 If country-dummies are excluded, however, the CONC and PD coefficients in the HIIT specification become significant and the overall fit of the regression increases to 0.115. Overall, the results are not very sensitive to the choice of a. VI. Concluding Remarks The magnitude of IIT is relatively high in bilateral trade between the CEECs and the EU. Levels of total IIT are comparable to those observed for countries such as Canada, Israel, Korea or Portugal. Most of the IIT is vertical in nature. Horizontal IIT levels are less than half of those of countries such as Austria, Spain, or Switzerland. Horizontal IIT has also been static over the 1990-95 period for the majority of countries. However, for some countries such as the Czech Republic and Slovenia it has been growing rapidly and has attained levels that exceed those reported for countries such as Greece, Finland and Israel. After controlling for country-specific factors, we find a positive and significant relationship between FDI and product differentiation and both vertical and horizontal IIT. Scale is negatively (positively) associated with horizontal (vertical) IIT, while concentration is positive and significant for horizontal IIT, but is insignificant for vertical IIT. Horizontal IIT is highly dependent on conditioning on country specific variables. If country dummies are not included in the estimation. the explanatory power of the industry-specific variables declines substantially. The empirical literature on IIT has generally found more support for the importance of country as opposed to industry factors (Balassa and Bauwens, 1987; Greenaway et al. 1995). Given that vertical IIT accounts for most of the observed IIT between the EU and the sample of CEECs, one 20 would have expected country factors to be particularly important determinants of vertical IIT. This is not the case for vertical IIT between the EU and the CEECs. The estimation results are quite robust when compared to existing studies on the determinants of IIT and its components. It can be hypothesized that this is due to the specifics of the initial post-reform period in the CEECs which were associated with a very significant opening of the economy to international competition, high levels of FDI (in 1995 the FDI-to-GDP ratio in the Czech Republic and Hungary was 17% and 15%, respectively); and substantial increases in the incentives to pursue product differentiation strategies following demonopolization and the break-up of the old conglomerates. The high share of vertical IIT that is observed is not surprising given the differences in relative real wages for comparable skill levels that existed between the EU and the CEECs and the geographic proximity of the CEECs to the EU. It is precisely these characteristics that make the CEECs particularly interesting in terms of investigating the effects of different industry-specific variables on IIT. 21 References Abd-el-Rahman, Kamal, (1991) "Firms' Competitive and National Comparative Advantage as Joint Determinants of Trade Composition," Weltwirtschaftliches Archiv, 127:83-97. Balassa, Bela and Luc Bauwens, (1987) "Intra-Industry Specialization in a Multi-Country and Multi-Industry Framework," Economic Journal, 97:923-939. Bergstrand, Jeffrey H., (1990) "The Hecksher-Ohlin-Samuelson Model, the Linder Hypothesis and the Determinants of Bilateral Intra-Industry Trade," Economic Journal, 100:1216-1229. Deardorff, Alan. 1995. "Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World? NBER Working Paper 5377. Ethier, W. 1982. "National and International Returns to Scale in the Modem Theory of International Trade," American Economic Review, 72:389-405. Falvey, Rodney E., (1981) "Commercial Policy and Intra-Industry Trade," Journal of International Economics, 11:495-511. Falvey, Rodney E. and Henryk Kierzkowski, (1985) "Product Quality, Intra-Industry Trade and (Im)perfect Competition," in edited by H. Kierzkowski Protection and Competition in International Trade, Blackwell: Oxford. Fontagne, Lionel, Michael Freudenberg and Nicolas Peridy. 1997. "Trade Patterns Inside the Single Market," Discussion Paper 97-07, CEPII, Paris. Greenaway, David, Chris Milner and Robert Elliott, (1996) "UK Intra-Industry Trade with EU North and South: A Multi-Country and Multi-Industry Analysis," Centre for Research in Economic Development and International Trade, University of Nottingham, Paper No. 96/9, May. Greenaway, David, Robert Hine and Chris Milner, (1995) "Vertical and Horizontal Intra- Industry Trade: A Cross Industry Analysis for the United Kingdom, " Economic Journal, 105:1505-1518, November. (1994) "Country-Specific Factors and the Pattern of Horizontal and Vertical Intra- Industry Trade in the UK," Weltwirtschaftliches Archiv, 130:77-100. Grubel, Herbert G., and P. J. Lloyd, (1975) Intra-Industry Trade: The Theory and Measurement of International Trade in Differentiated Products, London. 22 Harrigan, J. 1995. "The Volume of Trade in Differentiated Intermediate Goods: Theory and Evidence," Review of Economics and Statistics, 77:283-93. Helpman, Elhanan, (1981) International Trade in the Presence of Product Differentiation, Economies of Scale and Monopolistic Competition: A Chamberlin-Hecksher- Ohlin Approach," Journal of International Economics, 11:305-340. _(1987) "Imperfect Competition and International Trade: Evidence from Fourteen Industrial Countries," Journal of the Japanese and International Economies, 1:62-81. Helpman, E. and P. Krugman. 1985. Market Structure and International Trade. Cambridge: MIT Press. Hoekman, Bernard and Simeon Djankov. 1997. "Determinants of the Export Structure of East European Countries," World Bank Economic Review, September Hummels, David and Jamnes Levinsohn, (1995) "Monopolistic Competition and International Trade: Reconsidering the Evidence," Quarterly Journal of Economics, 110:799-836. Krugman, Paul, (1981) "Intra-Industry Specialization and the Gains from Trade," Journal of Political Economy, 89:959-973. Lancaster, K., (1980) "Intra-Industry Trade Under Perfect Monopolistic Competition," Journal of International Economics, 10: 151-175. Leamer, E. 1994. "Testing Trade Theories," in D. Greenaway and L. Alan Winters (ed.), Surveys in International Trade. Oxford: Blackwell. Markusen, James R., and Anthony J. Venables, (1996) "The Theory of Endowment, Intra-Industry and Multinational Trade," Centre for Economic Policy Research, London, Discussion Paper No. 1341, February. Neven, Damien. 1995. "Trade Liberalization with Eastern Nations: Why So Sensitive?," in R. Faini and R. Portes (eds.), European Union Trade with Eastern Europe. London: CEPR. Shaked, A and J. Sutton, (1984) "Natural Oligopolies and International Trade," in edited by H. Kierzkowski, Monopolistic Competition and International Trade, Oxford University Press: Oxford. Tybout, Jim. 1993. "Internal Returns to Scale as a Source of Comparative Advantage: The Evidence," American Economic Review, Papers and Proceedings, 83:440-44. 23 Table IA: Average ratio of export to import unit values, 1993-95 (109 3-digit NACE sectors) 1993 Bulgaria Czech Hungary Moldova Poland Romania Slovakia Slovenia Mean 0.711 1.119 1.052 0.826 1.008 1.034 0.896 1.153 Median 0.659 0.795 0.836 0.658 0.742 0.742 0.736 0.956 Std.dev 0.425 1.199 0.716 0.732 0.937 1.012 0.613 1.041 1994 1 Mean 0.766 1.058 1.074 0.754 0.972 0.950 0.911 1.182 Median 0.648 0.784 0.826 0.587 0.786 0.687 0.705 0.923 Std.dev 0.443 0.915 0.749 0.674 0.626 0.903 0.883 1.231 1995 _ _ _ Mean 0.908 1.002 1.265 0.644 1.126 0.837 0.869 1.221 Median 0.694 0.834 0.962 0.508 0.885 0.660 0.714 1.046 Std.dev 0.721 0.611 1.221 0.520 1.125 0.594 0.585 1.272 1993 Albania Estonia Greece Latvia Lithuania Portugal Spain Austria Mean 1.148 0.806 1.361 0.633 0.851 1.444 1.235 1.432 Median 0.865 0.442 0.948 0.483 OA72 1.140 1.034 1.315 Std.dev 0.884 1.170 1.874 0.591 0.985 0.982 1.874 0.711 1994 = Mean 1.308 0.653 1.128 0.570 0.871 2.138 1.175 1.442 Median 1.000 0.447 0.935 0.530 0.524 1.092 0.996 1.337 Std.dev 1.673 0.466 1.261 0.313 1.005 6.991 0.826 0.734 1995 Mean 1.614 0.925 1.284 1.033 1.040 1.922 1.416 1.517 Median 1.016 0.512 0.927 0.685 0.615 1.097 0.935 1.351 Std.dev 2.642 0.482 1.174 1.467 1.637 3.244 2.040 0.788 Note: Since unit values may not be dependable for minimal amounts of trade flows, unit values were calculated for only those flows which exceeded 5,000 ECU. 24 Table 2A: Total Intra-Industry Trade with EU(9) (109 3-digit NACE sectors) COUNTrRY I 1.9 1 1991 1 1992 1 1993 1 1994 1 199S CEEC: . Albania Mean 48.47 33.86 35.82 41.22 38.31 34.49 Median 58.06 30.51 25.11 34.94 35.42 28.97 ._____________ Std. Dev 30.66 27.15 31.19 31.04 30.49 29.89 Bulgaria Mean 27.95 30.87 28.81 33.32 31.04 24.57 Median 24.46 23.64 23.13 29.91 27.37 18.80 Std. Dev 21.08 20.22 19.33 24.16 24.19 18.57 Czechoslovakia Mean 31.56 36.82 41.04 -. Median 31.13 36.30 44.14 .. Std. Dev 20.56 22.34 21.95 - -- - Czech Republic Mean -- - - 43.87 43.43 43.68 Median _ , _ 42.38 41.89 42.52 Std. Dev _ . 22.32 23.03 22.41 Slovak Republic Mean . . 34.98 33.46 29.41 Median _ _ _ 30.28 31.01 21.53 Std. Dev .. .. .. 23.50 23.46 22.66 Hungary Mean 36.43 35.84 34.43 34.65 35.53 33.09 Median 34.39 33.70 33.96 32.20 35.38 32.53 Std. Dev 22.53 20.03 19.25 20.57 19.20 19.01 Poland Mean 29.12 28.87 28.03 28.98 29.01 29.69 Median 25.74 20.48 24.23 22.20 22.54 21.62 Std. Dev 19.99 20.87 18.32 22.29 22.42 22.12 Romania Mean 25.88 27.55 27.19 28.47 27.22 25.74 Median 18.26 21.86 24.17 22.68 20.49 17.96 Std. Dev 24.40 22.53 19.85 22.51 22.42 22.96 Slovenia Mean 37.52 33.66 36.87 37.35 Median _ - 36.20 32.75 37.82 37.15 Std. Dev _ _ 19.54 20.48 21.02 23.04 EU/EFTA* Austria Mean 50.37 50.23 50.56 50.45 51.04 47.03 Median 51.36 52.78 51.77 51.97 53.92 50.29 Std. Dev 20.09 19.58 20.61 21.20 20.05 20.49 Finland Mean 33.08 34.89 35.86 35.34 34.64 31.72 Median 28.64 32.45 32.24 33.10 32.83 33.60 Std. Dev 22.57 22.36 22.91 20.80 20.93 17.89 Greece Mean 25.69 25.78 24.02 20.59 22.99 22.50 Median 20.21 20.60 19.10 15.34 19.05 15.51 Std. Dev 21.60 19.16 19.66 17.02 18.09 18.75 Portugal Mean 30.28 32.39 31.49 30.12 30.15 30.46 Median 22.04 26.49 25.23 27.02 25.05 25.52 Std. Dev 23.20 24.76 23.09 20.75 23.46 22.19 Spain Mean 50.48 47.92 47.70 48.93 50.53 51.29 Median 49.24 48.31 46.94 49.03 52.73 53.52 Std. Dev 20.24 18.64 19.65 18.07 19.47 21.56 Switzerland Mean 51.33 51.98 52.88 53.22 52.68 52.31 Median 52.49 52.70 53.57 55.11 52.31 53.81 Std. Dcv 20.94 20.06 19.14 18.61 18.99 19.31 ME,NA: 1990 1991 _1922__ 1S93 1994 1995 Egypt Mean 21.11 20.09 19.36 22.71 22.05 24.50 Median 13.83 13.78 15.56 13.73 12.51 15.17 Std. Dev 18.97 18.44 16.81 22.32 24.69 25.68 Israel Mean 34.07 30.28 33.61 31.56 29.68 33.14 Median 29.73 24.95 26.30 27.30 24.68 28.15 Std. Dev 24.52 22.90 23.82 23.30 20.57 22.68 Morocco Mean 21.17 19.86 17.30 18.38 18.85 18.71 Median 11.60 9.73 10.53 10.34 8.86 11.13 Std. Dev 23.49 22.00 18.92 20.01 20.83 20.39 Tunisia Mean 28.29 26.88 28.85 25.91 21.13 24.86 Median 19.63 18.27 20.12 15.62 16.56 16.21 Std. Dev 24.36 23.26 25.20 23.39 17.80 22.36 Turkey Mean 28.75 26.61 26.39 25.20 28.20 29.05 Median 21.42 21.51 20.85 18.73 25.40 21.15 Std. Dev 24.77 23.14 20.19 22.59 19.88 24.26 25 NlCs Indonesia Mean 19.55 15.75 17.80 17.43 22.78 17.96 Median 11.87 8.73 12.02 8.89 14.81 8.75 Std. Dev 20.20 15.79 17.60 20.01 23.43 19.07 Korea Mean 26.35 23.35 25.74 26.31 28.38 30.85 Median 25.96 22.47 21.46 21.18 27.07 29.86 Std. Dev 19.01 13.70 17.87 19.19 18.50 19.38 Malaysia Mean 25.27 27.21 26.94 28.17 27.41 26.89 Median 18.19 16.29 19.33 19.62 23.22 23.38 _Std. Dev 22.75 25.72 22.66 24.62 23.31 23.04 Taiwan Mean 25.86 23.17 24.64 25.86 27.10 29.80 Median 22.20 21.89 19.76 19.83 25.08 27.38 ___ut_America iStd. Dev 21.97 19.72 19.03 21.84 20.44 21.23 South America: _________ Argentina Mean 28.90 26.41 24.26 23.68 20.88 21.14 Median 23.62 24.58 21.14 14.90 12.68 13.75 Std.dev 21.67 18.45 19.34 24.80 19.70 20.20 Brazil Mean 31.75 28.77 26.90 27.98 31.15 22.79 Median 29.89 23.83 23.34 22.78 26.29 20.80 Std.dev 24.46 23.36 20.91 21.29 22.92 16.41 Chile Mean 26.00 32.39 26.92 23.19 26.11 20.45 Median 15.27 16.89 12.97 11.40 10.87 9.83 Std.dev 27.62 30.71 28.18 24.41 26.97 25.09 South Asia: Bangladesh Mean 34.73 17.74 32.53 21.57 24.84 23.96 Median 29.09 7.82 19.47 17.11 17.89 20.45 Std. Dev 29.86 22.71 31.02 24.25 24.60 22.55 India Mean 27.09 24.44 25.69 27.13 24.06 29.09 Median 22.02 17.92 21.45 24.27 21.33 21.45 Std. Dev 23.61 20.81 20.66 23.58 18.78 23.59 Pakistan Mean 21.71 18.75 17.26 19.53 23.17 20.08 Median 11.62 10.08 8.09 7.88 12.34 8.40 Std. Dev 22.33 19.91 19.83 21.23 25.55 22.77 Cs: 1990 1991 1992 1993 1994 1995 Australia Mean 18.09 16.77 17.14 17.62 20.86 17.59 Median 9.64 11.49 9.99 12.46 13.57 12.63 Std.dev 21.24 19.19 19.51 16.86 21.48 16.62 Canada Mean 28.43 31.62 33.58 34.27 32.82 32.61 Median 22.31 27.56 31.04 30.67 29.36 28.12 Std. Dev 21.34 21.23 21.58 23.68 21.09 28.12 Japan Mean 34.68 34.82 36.23 35.40 36.03 34.51 Median 30.99 30.94 31.03 31.94 34.63 32.53 Std.dev 22.41 22.62 24.03 21.09 21.16 22.32 New Zealand Mean 23.55 25.51 23.64 27.17 25.97 25.89 Median 15.02 19.30 16.82 22.18 17.60 19.30 Std.dev 24.42 23.37 22.64 22.59 24.14 27.34 United States Mean 45.86 47.72 48.30 46.63 49.19 48.10 Median 49.92 51.63 53.32 48.57 50.52 51.70 _Std. Dev 21.81 21.06 21.11 22.02 21.29 21.75 Former USSR: Estonia Mean -- -- 45.81 29.07 30.88 28.79 Median 41.43 26.43 27.31 27.32 Std.dev 23.47 20.97 21.10 20.58 Latvia Mean 37.65- 34.59 35.24 29.04 Median 43.15 31.50 24.61 18.08 Std.dev 28.18 24.81 29.07 27.95 Lithuania Mean 40.62 25.56 29.88 28.08 Median 34.58 20.43 22.43 21.62 Std.dev 29.45 22.57 26.83 23.39 Moldova Mean -- -- 60.30 48.05 1 22.31 37.69 Median 69.64 43.45 13.08 29.45 _____ __ Std.dev -- 34.18 25.04 21.29 26.65 26 Table 3A: Horizontal Intra-Industry Trade (ħ15% range) with EU(9) (109 3-digit NACE sectors) COUNTRY 1990 1991 91 1992 1993 1994 1995 Albania Mean 11.43 3.65 3.66 5.59 5.29 6.34 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std. Dev 26.53 11.33 13.19 17.37 16.37 14.95 Bulgaria Mean 1.12 1.21 4.13 4.50 2.15 1.88 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std. Dev 4.54 3.24 11.55 13.82 6.75 6.19 Czechoslovakia Mean 4.33 4.28 5.33 - -- -- Median 0.00 0.00 0.34 Std. Dev 12.84 11.70 10.44 -- -- Czech Republic Mean -- -- -- 4.59 6.83 7.63 Median 0.18 0.85 1.59 Std. Dev 11.47 13.00 11.87 Slovak Republic Mean . 4.96 4.30 4.91 Median 0.00 0.00 0.00 Std. Dev 12.40 12.61 13.03 Hungary Mean 3.54 4.00 2.93 4.74 5.73 4.79 Median 0.00 0.11 0.49 1.08 1.45 1.00 Std. Dev 11.31 10.72 5.54 10.77 9.93 8.21 Poland Mean 3.27 3.08 3.42 2.70 6.01 3.35 Median 0.15 0.05 0.07 0.54 0.18 0.33 Std. Dev 7.12 8.00 8.11 5.93 15.17 8.27 Romania Mean 2.33 4.61 1.19 3.51 3.29 3.74 Median 0.00 0.00 0.00 0.00 0.04 0.00 _____________ Std. Dev 8.14 16.18 4.68 10.99 9.65 14.27 Slovenia Mean _ 4.86 6.28 6.37 8.65 Median 0.55 0.53 0.47 0.81 _Std.dev -- -- 12.58 12.33 12.00 16.94 IEU/EFTA: Austria Mean 18.80 18.29 17.99 16.00 17.75 16.94 Median 11.80 10.01 10.00 10.85 11.87 11.07 Std. Dev 20.63 19.80 20.76 16.45 18.19 18.72 Finland Mean 6.70 6.11 6.32 7.47 10.00 5.75 Median 2.16 1.13 1.26 2.07 1.29 0.97 Std. Dev 13.32 10.56 9.45 12.37 17.77 9.47 Greece Mean 5.00 5.51 4.25 3.11 4.31 4.45 Median 0.93 1.32 0.10 0.04 0.37 0.40 Std. Dev 9.85 10.39 8.38 7.74 10.52 9.63 Portugal Mean 5.55 5.31 8.13 7.83 9.68 6.18 Median 0.80 1.01 2.29 2.10 1.50 1.29 Std. Dev 10.37 10.18 14.82 13.41 19.00 11.58 Spain Mean 13.16 13.30 13.90 15.05 14.05 11.65 Median 5.11 9.30 8.52 9.80 8.91 4.88 Std. Dev 16.31 14.97 15.63 16.12 14.55 15.77 Switzerland Mean 13.41 13.10 11.93 11.36 12.36 13.14 Median 8.92 7.70 7.19 5.29 4.48 7.52 Std. Dev 16.83 16.86 15.44 15.13 16.05 16.60 MENA: 1990 1991 1992 1993 1994 1995 Egypt Mean 1.24 1.56 2.80 2.03 4.82 0.81 Median 0.00 0.00 0.00 0.00 0.01 0.00 Std.dev 4.34 4.07 11.48 5.01 16.36 2.61 Israel Mean 5.12 5.82 6.19 4.86 3.84 5.71 Median 0.14 0.30 0.72 0.41 0.49 1.09 Std. Dev 13.46 14.29 14.17 9.72 6.55 9.42 Morocco Mean 1.90 0.87 1.33 1.15 0.91 2.28 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std. Dev 6.42 2.34 3.33 2.34 2.23 9.31 Tunisia Mean 2.41 4.66 2.32 3.39 2.93 4.58 Median 0.00 0.31 0.09 0.00 0.00 0.09 Std. Dev 8.48 12.27 7.78 8.24 8.09 11.67 Turkey Mean 3.97 1.93 3.04 2.42 3.09 1.80 Median 0.01 0.04 0.28 0.00 0.36 0.01 Std. Dev 9.58 3.80 5.33 9.87 6.14 4.19 27 NICs: Indonesia Mean 2.50 0.63 0.91 1.46 1.26 2.36 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std. Dev 9.88 1.80 2.78 4.81 3.86 10.20 Korea Mean 1.75 1.47 1.98 1.19 3.73 3.56 Median 0.00 0.04 0.00 0.00 0.00 0.19 Std. Dev 4.79 3.38 5.77 2.72 8.84 8.58 Malaysia Mean 3.02 2.41 2.92 1.41 0.71 1.37 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std. Dev 9.79 13.00 9.88 4.19 1.73 4.58 Taiwan Mean 2.29 1.79 1.42 1.57 2.53 3.04 Median 0.00 0.00 0.05 0.00 0.00 0.00 Std. Dev 7.96 5.94 5.40 4.71 6.48 7.56 South America: Argentina Mean 2.29 2.53 2.44 4.63 1.81 2.13 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std. Dev 7.59 7.29 6.75 11.62 6.11 5.90 Brazil Mean 1.51 2.40 1.82 2.52 2.22 2.67 Median 0.00 0.00 0.00 0.00 0.06 0.02 Std. Dev 4.23 7.31 4.16 5.80 5.00 5.85 Chile Mean 3.94 5.24 0.46 1.29 2.42 2.26 Median 0.00 0.00 0.00 0.00 0.00 0.00 I Std. Dev 15.44 17.64 1.64 5.31 8.84 11.20 South Asia: Bangladesh Mean 0.03 0.77 8.50 0.21 3.80 2.69 Median 0.00 0.00 0.00 0.00 0.00 0.00 _Std.dev 0.09 2.56 22.43 0.61 18.11 11.36 India Mean 2.82 4.02 3.18 5.07 2.21 4.83 Median 0.00 0.00 0.00 0.13 0.19 0.09 ______________ Std.dev 10.03 13.62 9.18 13.75 4.42 14.52 Pakistan Mean 1.32 0.57 1.49 2.01 3.61 2.85 Median 0.00 0.00 0.00 0.00 0.00 0.00 Std.dev 5.73 1.45 4.73 7.78 8.86 10.69 IDCSe I I 1990 1 1991 12 1993 11294 1295 Australia Mean 3.98 2.42 3.23 2.32 1.71 1.58 Median 0.17 0.05 0.01 0.06 0.12 0.08 lstd.dev 10.48 4.99 10.92 7.40 4.88 3.19 Canada Mean 5.47 4.96 6.29 6.10 7.92 5.12 Median 0.77 0.74 0.85 1.61 1.06 1.27 |____________ IStd.dev 10.36 8.20 15.41 12.08 14.55 1.27 Japan Mean 5.74 3.93 3.96 5.76 5.03 6.07 Median 1.42 0.45 0.68 0.83 0.97 0.60 I___________ IStd.dev 11.42 7.52 7.74 9.77 8.47 10.61 New Zealand Mean 4.69 3.31 2.71 2.29 2.83 1.76 Median 0.02 0.00 0.00 0.00 0.00 0.00 I____________ lStd.dev 14.83 8.05 7.41 5.86 6.72 4.69 United States Mean 8.12 9.65 13.12 9.35 10.30 11.19 Median 2.76 3.64 8.20 3.02 4.18 4.99 I____________ IStd.dev 13.27 14.54 14.80 14.42 13.71 14.37 F~ormer USSR: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Estonia Mean -- -- 9.43 2.12 2.06 5.70 Median -- - 0.00 0.00 0.00 0.00 Std.dev -- -- 22.97 6.69 7.02 14.57 Latvia Mean -- -- 0.72 4.94 3.77 3.02 Median |0.00 0.00 0.00 0.00 Std.dev -- -- 2.51 13.97 13.60 7.58 Lithuania Mean -- -- 3.47 3.76 1.63 2.27 Median -- -- 0.00 0.00 0.00 0.00 I____________ IStd.dev -- -- 14.32 14.69 5.84 6.87 Moldova Mean - - 7.17 0.13 0.86 1.06 Median -- - 0.00 0.00 0.00 0.00 I____________ gStd.dev -- - 24.84 0.52 2.42 3.99 28 Table 4A: Vertical Intra-Industry Trade (ħ15% range) with EU(9) (109 3-digit NACE sectors) COUNTRY I 1990 I 191I 1992 1 1993 1 1994 1 1 WS CEEC: ___ Albania Mean 37.04 30.21 32.16 35.63 33.03 28.15 Median 33.06 28.16 21.76 26.30 25.58 16.81 Std. Dev 32.79 28.08 32.25 30.88 30.55 28.50 Bulgaria Mean 26.83 29.66 24.67 28.82 28.89 22.70 Median 22.33 22.83 20.51 20.00 24.61 18.06 Std. Dev 21.54 20.72 18.77 23.85 23.67 18.55 Czechoslovakia Mean 27.24 32.54 35.72 -- - Median 24.91 31.53 32.49 Std. Dev 19.88 23.12 23.28 -- - Czech Republic Mean -- -- -- 39.28 36.60 36.05 Median _ 39.09 35.35 35.22 Std. Dev _ 23.02 23.67 22.30 Slovak Republic Mean 30.02 29.16 24.50 Median _ _ 24.14 27.00 19.33 Std. Dev -- - 23.42 21.47 19.74 Hungary Mean 32.89 31.84 31.50 29.91 29.80 28.31 Median 28.09 28.02 32.45 28.78 29.23 23.93 Std. Dev 22.03 20.24 19.37 19.76 17.55 19.48 Poland Mean 25.86 25.79 24.61 26.28 23.00 26.33 Median 23.37 20.02 22.61 20.59 18.02 18.49 Std. Dev 19.58 20.75 17.37 21.65 19.06 21.12 Romania Mean 23.55 22.94 26.01 24.96 23.93 22.00 Median 14.47 16.64 21.29 19.31 17.10 15.89 _Std. Dev 24.29 20.20 19.85 22.63 22.36 20.95 Slovenia Mean -- 32.67 27.38 30.50 28.70 Median -- 30.24 25.11 28.33 24.01 _Std.dev -- 19.52 18.68 21.60 19.65 EU/EFTA: Austria Mean 31.57 31.94 32.57 34.46 33.30 30.08 Median 30.32 29.05 31.55 34.23 28.99 26.50 Std. Dev 19.28 19.47 19.70 20.61 21.24 18.24 Finland Mean 26.38 28.78 29.54 27.88 24.64 25.96 Median 21.28 22.80 24.36 24.72 21.10 26.10 Std. Dev 21.06 21.27 22.35 19.28 18.68 17.77 Greece Mean 20.69 20.26 19.77 17.49 18.68 18.05 Median 12.43 13.60 13.88 11.41 13.30 11.03 Std. Dev 21.83 18.89 19.80 14.97 17.10 18.04 Portugal Mean 24.73 27.08 23.36 22.29 20.48 24.29 Median 17.02 20.86 20.66 17.85 15.78 21.20 Std. Dev 21.48 24.19 20.51 19.62 19.22 20.83 Spain Mean 37.32 34.62 33.80 33.88 36.48 39.65 Median 35.13 35.35 30.44 32.01 35.35 41.91 Std. Dev 21.62 19.32 18.41 19.20 21.39 21.37 Switzerland Mean 37.92 38.88 40.96 41.86 40.33 39.17 Median 34.62 36.64 41.28 40.77 38.23 38.96 Std. Dev 20.75 20.82 20.09 20.37 19.66 18.97 MENA: 1990 1991 1992 1993 1994 1995 Egypt Mean 19.87 18.53 16.56 20.67 17.23 23.69 Median 13.50 11.87 12.07 10.91 9.03 14.67 Std.dev 18.97 19.08 15.17 22.63 20.59 25.96 Israel Mean 28.95 24.46 27.42 26.71 25.8.4 27.44 Median 22.76 16.88 22.49 21.90 21.77 24.45 Std. Dev 23.74 21.58 _ 22.18 22.72 21.26 21.87 Morocco Mean 19.27 18.99 15.97 17.23 17.94 16.43 Median 9.54 8.23 8.14 7.85 8.24 10.56 Std. Dev 23.16 22.10 19.42 20.05 20.64 19.21 Tunisia Mean 25.89 22.22 26.53 22.52 18.20 20.29 Median 17.81 12.77 18.75 12.89 11.99 11.81 Std. Dev 24.41 23.18 25.63 23.27 18.10 21.19 Turkey Mean 24.79 24.68 23.36 22.78 25.11 27.25 Median 15.16 19.03 17.55 18.24 21.79 17.83 Std. Dev 24.55 23.35 20.30 21.72 19.89 24.07 29 NlCs:__ _ _ l__ _ _ __ _ _ _ Indonesia Mean 17.05 15.12 16.88 15.97 21.52 15.60 Median 8.95 8.31 9.98 8.68 14.11 8.36 Std. Dev 18.85 15.82 18.00 19.74 23.53 16.86 Korea Mean 24.59 21.88 23.76 25.12 24.65 27.30 Median 21.37 21.98 19.25 19.98 21.11 24.35 Std. Dev 19.47 13.85 18.05 19.20 18.50 19.97 Malaysia Mean 22.25 24.80 24.02 26.76 26.71 25.52 Median 15.15 15.52 15.02 18.12 19.79 22.68 Std. Dev 22.89 24.48 23.03 25.10 23.39 23.13 Taiwan Mean 23.57 21.38 23.22 24.29 24.56 26.76 Median 19.06 15.97 18.52 19.83 23.32 23.41 Std. Dev 21.64 19.35 18.59 20.90 20.52 21.02 South America: Argentina Mean 26.61 23.87 21.83 19.04 19.07 19.01 Median 21.04 20.02 18.02 10.75 8.95 9.36 Std. Dev 22.24 18.72 19.15 22.81 19.78 20.21 Brazil Mean 30.25 26.37 25.07 25.45 28.93 20.11 Median 28.20 23.51 20.78 20.59 23.76 17.06 Std. Dev 25.00 21.95 20.78 21.63 22.69 16.17 Chile Mean 22.06 27.15 26.45 21.91 23.69 18.19 Median 8.75 15.73 12.97 9.02 9.30 7.14 Std. Dev 26.41 28.53 28.36 24.97 26.94 23.90 South Asia: Bangladesh Mean 34.70 16.97 24.03 21.35 21.04 20.41 Median 29.09 6.21 16.52 17.11 15.56 15.88 Std.dev 29.89 23.14 27.39 24.40 20.90 21.83 India Mean 24.27 20.43 22.52 22.06 21.85 24.26 Median 18.86 15.11 17.72 18.03 17.71 18.93 Std. Dev 22.70 18.99 20.68 23.31 19.34 22.87 Pakistan Mean 20.39 18.19 15.77 17.51 19.56 17.23 Median 8.28 9.46 6.69 6.34 7.50 7.81 Std.dev 22.30 20.24 20.15 21.21 25.70 21.89 DCs;ffi: 1990 1299 1992 1993 1994 1995 Australia Mean 14.11 14.36 13.91 15.30 19.15 16.01 Median 7.24 8.73 9.28 10.83 12.26 11.66 Std. Dev 19.97 19.15 16.33 15.85 21.15 16.03 Canada Mean 22.96 26.66 27.30 28.18 24.90 27.49 Median 17.94 20.54 22.53 21.57 22.19 20.53 Std.dev 19.46 21.24 20.82 22.69 18.96 20.53 Japan Mean 28.95 30.88 32.27 29.64 31.00 28.44 Median 23.64 24.87 26.86 24.38 27.81 27.17 Std. Dev 22.40 22.72 24.48 19.82 21.28 20.69 New Zealand Mean 18.86 22.21 20.93 24.88 23.14 24.13 Median 12.42 13.78 12.20 16.73 13.65 11.97 Std. Dev 22.01 22.65 22.23 22.80 23.80 26.48 United States Mean 37.74 38.07 35.18 37.28 38.89 36.91 Median 36.83 38.66 35.86 38.25 42.01 36.37 _Std.dev 21.67 20.70 19.68 19.40 19.99 19.70 Former USSR: Estonia Mean -- - 36.38 26.95 28.82 23.09 Median 36.19 20.74 26.13 21.73 Std. Dev _ 24.55 21.54 22.22 18.45 Latvia Mean 36.93 29.65 31.47 26.02 Median 42.75 24.72 19.58 15.64 Std. Dev _ 28.19 23.36 28.85 28.00 Lithuania Mean _ _ 37.15 21.80 28.25 25.81 Median _ _ 28.72 19.60 21.87 18.64 Std. Dev _ 29.59 19.52 25.96 23.81 Moldova Mean _- 53.13 47.92 21.44 36.64 Median - 58.45 43.45 12.18 24.48 Std.dev - _ 37.18 25.25 21.92 27.01 30 Table 5A: Horizontal Intra-Industry Trade (ħ25% range) with EU(9) (109 3-digit NACE sectors) COUNTRY 1990 1991 1992 1993 1994 1995 (%/.) (%) ) ( (%) (%/) (%/0) CEEC: Bulgaria Mean 2.90 3.51 5.95 7.07 3.27 2.83 Median 0.00 0.00 0.82 0.73 0.00 0.03 Std. Dev 7.77 6.87 11.93 16.34 7.99 6.62 Czechoslovakia Mean 5.00 6.99 8.57 -- -- -- Median 0.49 1.56 2.10 Std. Dev 12.35 12.48 12.69 -- -- -- Czech Republic Mean -- - -- 9.35 11.64 11.50 Median 2.35 4.11 2.80 Std. Dev 14.40 17.19 17.16 Slovak Republic Mean 7.02 7.00 6.30 Median 0.70 0.40 0.83 Std. Dev 13.99 15.48 13.08 Hungary Mean 4.84 6.77 8.86 8.01 8.67 7.61 Median 0.88 2.04 2.97 2.05 3.09 3.17 Std. Dev 11.54 12.53 13.28 12.92 12.56 9.51 Moldova Mean -- -- 12.01 0.13 1.47 5.38 Median 0.00 0.00 0.00 0.00 Std.dev -- -- 28.68 0.52 2.85 16.36 Poland Mean 4.61 6.03 6.29 6.99 8.29 8.00 Median 0.73 1.00 1.16 1.53 1.61 0.87 Std. Dev 8.09 11.39 10.01 15.24 15.16 14.66 Romania Mean 2.,4 4.90 2.72 8.37 4.04 5.84 Median 0.00 0.00 0.00 0.91 0.20 0.23 Std. Dev 8.07 15.73 9.32 18.98 9.36 15.09 Slovenia Mean -- -- 8.13 9.50 11.48 13.36 Median 2.30 2.86 4.37 5.46 Std.dev 14.81 14.69 15.59 18.83 31 Table 6A: Vertical Intra-Industry Trade (ħ25% range) with EU(9) (109 3-digit NACE sectors) COUNTRY 1990 1991 1992 1993 1994 l.995 (°/°) (%) (%) (%) (%) (%/0) CEEC: Bulgaria Mean 26.29 27.76 23.57 28.11 29.19 21.75 Median 22.03 21.93 17.13 19.92 23.81 18.00 Std. Dev 22.58 21.31 18.39 25.46 25.16 18.25 Czechoslovakia Mean 27.08 30.43 33.41 -- -- - Median 25.47 30.02 31.07 Std. Dev 20.61 23.77 22.69 -- Czech Republic Mean -- -- 34.90 32.52 32.76 Median 34.69 32.53 31.63 Std. Dev 23.78 22.84 23.21 Slovak Republic Mean 28.31 27.02 23.22 Median 23.36 24.66 15.95 Std. Dev 23.98 21.18 19.86 Hungary Mean 31.87 29.69 26.57 27.41 27.63 25.29 Median 26.62 26.97 24.53 26.38 26.17 21.65 Std. Dev 22.17 20.23 18.03 18.26 17.48 17.92 Moldova Mean -- -- 48.29 47.92 20.84 32.31 Median 47.06 43.45 11.80 21.40 ________ Std.dev - -- 35.37 25.25 22.40 27.47 Poland Mean 24.31 23.56 22.18 22.97 21.89 21.79 Median 23.41 18.20 19.65 18.64 16.49 14.77 Std. Dev 18.98 19.92 17.82 20.32 19.98 19.56 Romania Mean 22.81 22.73 27.29 22.39 23.53 19.48 Median 14.47 13.48 21.29 16.71 17.98 12.77 Std. Dev 24.48 22.24 22.46 22.20 22.52 20.08 Slovenia Mean -- -- 29.39 24.16 25.39 23.99 Median 26.00 22.04 24.20 19.91 Std.dev -- 19.45 17.80 19.13 18.06 32 Policy Research Working Paper Series Contact Title Author Date for paper WPS1828 The Determinants of Banking Crises: Asli DemirgUO-Kunt September 1997 P. Sintim-Aboagye Evidence from Developed and Enrica Detragiache 38526 Developing Countries WPS1829 Economic Reform and progress in Norman Loayza September 1997 E. Khine Latin America and the Caribbean Luisa Palaclos 37471 WPS1830 Private Ownership and Corporate Roman Frydman September 1997 B. Moore Performance: Some Lessons frcm Cheryl W. Gray 38526 Transition Economies Marek Hessel Andrzej Rapaczvnski WPS1831 How Trade Patterns and Technology Wolfgang Keller September 1997 J. Ngaine Fiows Affect Productivity Growth 37947 WPS1 832 Pension Reform in Bolivia: Innovative Hermann von Gersdorff September 1997 C. Pavlak Solutions to Common Problems 82099 WPS1 833 Cost Recovery and Pricing of David B. Humphrey October 1997 T. ishibe Payment Services Robert H. Keppler 38968 Fernando Montes-Negret WPS1834 The Comparative Advantage of Pedro Belli October 1997 L. Schunk Government: A Review 31779 WPS1835 Cost-Benefit Analysis of the Global Aehyung Kim October 1997 A. Kim Dracunrculiasis Eradication Campaign Ajay Tandon 35029 Ernesto Ruiz-Tiben WPSl 836 Health and Labor Productivity: The Aehyung Kim October 1997 A. Kim Economic Impact of Onchocercal Ajay Tandon 35029 Skin Disease Asrat Hairu and Others WPS1837 How Estonia's Economic Transition Rivo Noorkoiv October 1997 S. Fallon Affected Employment and Wages Peter F. Orazem 38009 (19&9-95) Allan Puur Milan Vodopivec WPS1838 The Limits and Merits of Participation Paulo Vieira da Cunha October 1997 E. Thomas Maria Valeria Junho Pena 31019 WPS1839 The Effects of Financial Liberalization Cevdet Denizer November 1997 E. Khine and New Bank Entry on Market 37471 Structure and Competition in Turkey WPS1840 Efficient Public Sector Downsizing Martin Rama November 1997 S. Fallon 38009 WPS1841 Patterns of Metropolitan Gregory K. Ingram November 1997 j. Ponchamni Development: What Have We Learned? 31052 Policy Research Working Paper Series Contact Title Author Date for paper WPS1 842 Motorization and the Provision of Gregory K. Ingram November 1997 J. Ponchamni Roads in Countries and Cities Zhi Liu 31052 WPS1843 Externalities and Bailouts: Hard and David E. Wildasin November 1997 C. Bernardo Soft Budget Constraints in 37699 Intergovernmental Fiscal Relations WPS1844 Child Labor and Schooling in Ghana Sudharshan Canagarajah November 1997 B. CaseJy-Hayford Harold Coulombe 34672 WPS1845 Employment, Labor Markets, and Sudharshan Canagarajah November 1997 B. Casely-Hayford Poverty in Ghana: A Study of Dipak Mazumdar 34672 Changes cluring Economic Decline and Recovery WPS1846 Africa's Role in Multilateral Trade Zhen Kun Wang November 1997 J. Ngaine Negotiations L. Alan Winters 37947 WPS1 847 Outsiders and Regional Trade Anju Gupta November 1997 J. Ngaine Agreements among Small Countries: Maurice Schiff 37947 The Case of Regional Markets WPS1848 Regional Integration and Commodity Valeria De Bonis November 1997 J. Ngaine Tax Harmonization 37947 WPS1849 Regional Integration and Factor Valeria De Bonis November 1997 J. Ngaine Income Taxation 37947