The aim of this report is to examine the link between trade and total factor productivity (TFP) in two manners. First, we study the impact of tariffs on the plant-level TFP, and then, we separately analyze the impact of TFP and other plant characteristics on the survival and export decision of plants. Using Olley and Pakes (1996) method, we calculate TFP of Spanish and Turkish manufacturing firms. In the period analysed, the reallocation of market shares to more productive plants contributes positively to productivity growth while, within plant productivity improvements are significant only in some sectors.
We find that productivity improvements resulting from declining protection levels are statistically significant and economically important in both cases. For Turkey, this is especially true in import competing sectors. For Spain, we also show that TFP also increases with the presence of foreign products in the domestic market. Importing firms benefit from additional gains. In both countries, we find evidence of positive effect of agglomeration at the province level for productivity. As a consequence, industrial policy should favor external scale economies by encouraging clustering. Our analysis also suggests that there is a huge degree of heterogeneity among plants. In particular, results from Spain and Turkey differ concerning the reaction of small firms’ productivity, in response to changes in protection rates. In Spain, small firms react more positively than others to tariff reduction while in Turkey, large plants would obtained larger gains.
Concerning the entry decision of firms on export markets, estimation results are strongly in favor of the hypothesis that more productive firms self-select into export markets. Secondly, there is a positive and highly significant effect of previous experience to export, suggesting the presence of considerable sunk costs at exporting. Concerning the effect of protection on the participation in exports markets, the results are not so homogeneous among the two countries. For Spain, our study shows that a decrease in import tariffs and the growth of imports from developed countries increase the probability of exporting. In the Turkish case, the study shows that firms operating in protected non-trading industries are more likely to export. As for the impact of trade reforms on the survival patterns of firms, results provide evidence that exporters are more likely to survive (after controlling for productivity and size), and trade protection worsens the survival prospects of exporters.
Keywords: Total factor productivity, Spain, Turkey, Exports behavior, Survival, tariffs, heterogeneity of firms. JEL codes: F12
Non technical synthesis
The switch from an inward-oriented growth process to a much more outward-oriented one would bring about, as was expected, a more efficient allocation of resources which would then transmit into a sustainable growth path with a steady increase in per capita income. Several studies, surveyed in Rodrik (1995), distinguish between two effects of trade liberalization on growth in DCs: on the one hand, static effects which entail intersectoral resource transfers due to the modification of the relative price structure and, on the other hand, dynamic effects arising from productivity growth due to increased exposure of local firms to competition on foreign and domestic markets, to a increase in technology imports embodied in capital and intermediate goods and to the transfer of other kind of technical knowledge through informal means. These dynamic effects are to be materialized through sustained productivity growth in firms and a switch to export activities. These proposals can be verified thanks to firm-level data.
This report aims at contributing to the empirical literature about the link between trade, total factor productivity (TFP) and export behaviour, at the micro level. We study the impact of protection levels and other important variables like foreign ownership, agglomeration effects, competition of the markets, export and import status on productivity and the impact of firm’s and market’s characteristics on export status. We include two case studies: Spain and Turkey, two countries with a bilateral trade agreement with the EU.
A special effort has been made to make the two studies comparable by estimating similar empirical models, unless we were unable to avoid some restrictions.
First, the Turkish survey is larger and totally exhaustive while the Spanish survey is not. All the Turkish firms have to answer the survey, so we have information about all firms after 1982, in particular firms created after 1982 and those who exited the market after 1982. This is not the case for the Spanish survey which is not exhaustive for small firms. Some firms could have disappeared from the survey, even if they are still active. In particular, this reduces the number of firms for which we are able to affirm that they exit and the possibility to estimate a survival model in the case of Spain. For this reason, we were able to run estimations for some subsets of sectors in the Turkish case but not in the Spanish one. In the Turkish case, we use an unbalanced panel of more than 150,000 observations per year for the 1983-2001 period. Data for exports at the plant level are only available after 1990. In this huge data set, we have sufficient observations to distinguish among three types of sectors based on their trade orientation: Export-oriented, import-competing, and the non-traded sectors. In the Spanish case, the sample is much smaller: 3,107 firms (20,882 observations for the period 1991-2002). Though, we have information about imports and exports of the firms, all over the period.
Second, in both cases, data are available for very large periods (1982-2000 for Turkey and 1991-2002 for Spain) but do not cover the period of trade liberalisation of Spain since Spain entered the CEE in 1986. Spanish tariff was already lowered and harmonized with the Common trade policy for third countries product and set to zero for EU products. Though, Spanish imports increased all over the period under study and the import penetration rate (IPR) captures accurately the effect of trade liberalisation. IPR and tariffs drove different information in this case. IPR reflects the importance of EU and non-EU products in the Spanish markets, while tariffs reflect protection towards third countries. IPR was not taken into account in the Turkish case since it engages in bilateral free trade with the EU later. Actually, the Customs Union agreement entered into force in 1996.
Nevertheless, we follow similar estimations strategies in both studies, as far as we are able to. Total Factor Productivity (TFP) is measured in both studies following the procedure of Olley and Pakes so as to eliminate potential simultaneity and selection biases. In both cases, TFP gains have been large in the studied period. In the majority of sectors, in Spain, the aggregate productivity has increased between 1,5% to 3,3% per year except for the Food sector from 1991 to 2002. TFP growth is explained both by the increase of intra-firm productivity, reallocation among firms, exit and entry. The Turkish case is different. Firms have experienced on average 1.77% of TFP growth per year (1982-2000) but 2.96% in Import competing sectors. Intra-firm productivity negatively contributes to TFP growth at the sector level. TFP growth is explained by reallocation among firms, exit and entry.
For each country, we focus on two sets of questions. The first one concerns the impact of trade on TFP. The second one concerns the export behaviour of firms. The studies are original contributions to the literature and lead to interesting conclusions for policy makers.
Concerning the entry decision of firms on export markets, our results provide strong support for the hypothesis that sunk costs in entering foreign markets are important. Once a firm commits itself into foreign markets by covering the sunk costs at the time of entry, the firm tends to stay in foreign markets. Therefore, sunk costs create persistence in export behavior, i.e., exporters tend to remain as exporters.
Secondly, the most productive firms and the large firms are more willing to export. Estimation results are strongly in favor of the hypothesis that more productive firms self-select into foreign markets. Moreover, even after controlling for productivity, large firms are found to have higher probability of participating in foreign markets.
Concerning the effect of protection on the participation of exports markets, the results are not so homogeneous among the two countries. For Spain, our study shows that a decrease in import tariffs increases the probability of exporting. Consistent with this fact, the growth of imports from developed countries have a positive impact on export status, that is the presence of foreign products in the domestic markets is an important incentive for exporting. The growth of exports to developed countries of the industry is a negative incentive for exports since it may increase the costs for new entrants to increase their share of the market. Firms with medium size, importers and firms with low foreign participation in their capital are especially sensitive to the growth of trade with developed countries while large firms or firms with an important participation of FDI are sensitive to tariffs.
In the Turkish case, the study tends to show that firms operating in protected non-trading industries are less likely to export. But in general, for Turkish firms, the decision of exporting ins not influenced by import tariffs. As for the impact of trade reforms on the survival patterns of firms, results provide strong evidence for the hypothesis that exporters are more likely to survive even after controlling for their productivity and size. However, high tariff barriers are detrimental for the survival of exporters. This result may be explained by the fact that foreign trade protection increases the relative competitiveness of non-exporters in relation to the one of exporters. Firms operating in industries that experience a surge in exports to developed countries have better survival prospects. Export status and size are the main determinants of survival. Consistent with previous results, we find that small firms are more willing to exit the market.
As pointed by Tybout (2001), the effects of trade policies on the exports and, on the structure of the domestic market, depend widely on the initial conditions, the importance of sunk costs and the importance of the heterogeneity of firms. Our results are very much in this line. For small firms, firms that do not import intermediate goods or capital goods, the previous experience is more important than the productivity level. For these firms, the competition of the other exporters is a more important barrier for exporting than the growing presence of foreign products or the lowering of tariffs. Though, large firms also face large costs to enter export markets but they additionally have to be very productive and care about protection. The presence of sunk costs at exporting is associated with costs of gathering market information, establishing networks of distribution and adjusts to national standards and legislation. Although trade policy has traditionally concentrated in trade costs, a policy that aims at reducing fixed costs of exporting could have considerable effects on exporting. To this end, public policies should, not only increase the capacity of the firms to assimilate innovations in order to increase their productivity, but also facilitate the knowledge and access to foreign markets and favor the permanence in the export market. Export promotion agencies and international harmonization standards may have an important role to play, especially for small firms.
Impact of trade on productivity
We investigate the effects of protection rates on plant/firm level productivities using fixed-effect estimations for Turkey and System-GMM for Spain. In both cases, we include the lagged level of productivity since we believe that TFP determinants are highly persistent. In these regressions, we control for endogeneity of protection rates as well as other potential important determinants of productivity improvements during the period under consideration, such as intensity of domestic competition.
We find that productivity improvements resulting from declining protection levels are statistically significant and economically important in both cases. For Turkey, tariff cuts lead to more important productivity gains in import competing sectors. For Spain, w also find that a removal of European tariffs would translate into improvements of TFP. Moreover, the presence of foreign products leads to additional gains. Another important finding is that, firms that import intermediate and capital goods from abroad benefit from additional productivity improvements. We conclude that, even in a European country, with relatively low levels of protection such as Spain, there are additional gains to expect from trade liberalisation process. However, a large part of the positive effect comes from the presence of foreign products and more indirect effects of openness rather than from tariff reduction.
In both countries, we find evidence of positive effect of agglomeration at the province level for productivity. As a consequence, industrial policy should favor external scale economies by encouraging clustering.
All firms do not react in the same way to trade liberalisation and some firms are more sensitive than others to tariffs and to the presence of foreign products in the domestic market. In Spain, small firms and firms that do not participate in foreign market via exports, imports or ownership react more positively than others to tariff reduction. In turn, large firms, importers and exporters, take more advantage of foreign competition in terms of TFP gains than the other firms. In Turkey, the opposite results holds: the larger the plant size, the larger will be the productivity improvement as a result of a tariff cut. In the case of Turkey, the small firms are more likely to face credit constraints than the large ones. Consequently, their response to trade liberalization is constrained by their limited access to credit due to high interest rates. This could have been the case of Spanish small firms at the beginning of the eighties but not during the period under study, marked by a stable macroeconomic context and low interest rates. Spanish small firms could rather be in the second step of the trade liberalisation where the context allows them to benefit from trade liberalisation while in a first step, only robust firms seem to be able to do so.