学习出口：摩洛哥制造业的情况

POLICY RESEARCH WORKING PAPER 2827 Learning to Export Evidence from Moroccan Manufacturing Marcel Fafchamps Said El Hamine Albert Zeufack The World Bank Development Research Group Macroeconomics and Growth April 2002  PoLicY RESEARCH WORKING PAPER 2827 Abstract Fafchamps, Hamine, and Zeufack test two alternative firms that export. Most do so immediately after creation. models of learning to export: productivity learning, The authors also find that, among exporters, new whereby firms learn to reduce production costs, and products are exported very rapidly after production has market learning, whereby firms learn to design products begun. The share of exported output nevertheless that appeal to foreign consumers. increases for 2-3 years after a new product is introduced. Using panel and cross-section data on Moroccan Old firms are unlikely to switch to exports, even in manufacturers, the authors uncover evidence of market response to changes in macroeconomic incentives. The learning but little evidence of productivity learning. authors find a positive relationship between exports and These findings are consistent with the concentration of productivity and conclude that it is the resuilt of self- Mloroccan manufacturing exports in consumer items- selection: it is the more productive firms that move into the garment, textile, and leather sectors. It is the young exports. Policy implications are discussed. This paper-a product of Macroeconomics and Growth, Development Research Group-is part of a larger effort in the group to investigate the microeconomic foundations of export and growth performance using plant-level data. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Emily Khine, room MC3-347, telephone 202-473-7471, fax 202-522-3518, email address kkhine@worldbank.org. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at marcel.fafchamps(@)economics.ox.ac.uk or azeufack@worldbank.org. April 2002. (44 pages) The Policy Research Working Paper Series disseminates the fincings of work in progress to encourage the exchange of ideas about I development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordtizgly. 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 Research Advisory Staff  Learning to Export: Evidence from Moroccan Manufacturing Marcel Fafchamps Said El Hamine Albert Zeufack Oxford University* MCIt The World Bankt 'Department of Economics, University of Oxford, Manor Road, Oxford OX1 3UQ (United Kingdom). Email: marcel.fafchampsOeconomics.ox.ac.uk. tMinistere du Commerce, de l'Industrie, et des Mines, Rabat, Morocco. IDECRG, The World Bank, 1818 H Street NW, Washington DC 20433 (USA). Email: azeufackOworldbank.org.   1. Introduction For many poor countries, manufacturing exports are seen as the royal path to growth. Albeit a handful of countries - most of them in East and South-East Asia - have achieved unprecedented rates of growth through exports, many other countries have tried but failed to follow the same route (The World Bank 1993). One often heard explanation for the contrasted experiences of manufacturing export strategies is that exporting is subject to learning-by-doing (e.g. Bernard and Jensen 1 999b, Tybout 2000, Harrison and Hanson 1999, Bigsten, Collier, Dercon, Fafchamps, Gauthier, Gunning, Oduro, Oostendorp, Patillo, Soderbom, Teal and Zeufack 1999, Bigsten, Collier, Dercon, Fafchamps, Gauthier, Gunning, Oduro, Oostendorp, Patillo, Soderbom, Teal and Zeufack 2000). Countries or enterprises that lack experience find it difficult to compete in international markets. For this reason, countries and firm first need to learn how to produce before attempting to penetrate export markets. In this view, the domestic market is essential as learning ground for exports. The purpose of this paper is to investigate this idea using survey data from Morocco. Tt has long been recognized that exporting is not a straightforward process. This is particularly true when poor countries seek to export manufactured products to rich nations. Products sold in developed economies are typically of higher quality. Competition tends to be fierce and consumers are sensitive to minor variations in design and style. Learning how to satisfy fussy consumers may take some time for firms accustomed to less sophisticated domestic consumers in poor countries. Tn some cases, consumers in rich countries learn to appreciate simple products from poor countries, but this also takes time and (marketing) effort. While few dispute the above observations, there is considerable disagreement concerning how long learning ought to take. Some argue that learning is a protracted process that extends over years if not decades. The argument is used as a basis for the infant industry argument to justify lengthy protection (Prebisch 1963).' Others point to micro evidence that learning by doing takes at most a couple years (e.g. Alchian 1963, Searle 1945, Griliches and Lichtenberger 1984). 'See also the references cited in Tybout (2000).  It is also unclear what kind of learning is required to export manufactures. One approach focuses on the general productivity of the firm. According to this view, firms must reach a high level of productivity before they can export. As a result, exporters are expected to be found among the most productive firms (e.g. Bernard and Jensen 1 999b, Bernard and Jensen 1999a, Bigsten et al. 2000). Regarding Morocco, this view was presented most clearly in Clerides, Lach and Tybout (1998). Learning to be productive is seen as a prerequisite for exporting. If this view is correct, one would expect exporters to be firms that accumulated years of experience in their domestic market before launching into exports. A protected domestic market might then be necessary to allow firms to test their products locally and learn from their mistakes. Another approach focuses on market familiarity. To export, a firm must find a product that consumers abroad want to purchase. Low production costs are helpful but not critical because suitably designed products can fetch a high price in a differentiated market. If this view is correct, producing for the domestic market is not a prerequisite for exporting. Tt might even be a drawback if it leads to firm to learn 'the wrong thing' from focusing on its domestic market, e.g., that consumers are easily fooled, that quality variation is not an impediment to sales, or that health hazards go unnoticed. In this case, a protected domestic market might actually hinder exports. Tf foreign goods were allowed in, domestic consumers would become more demanding, that is, more like foreign consumers. This would force firms to learn how to upgrade their products which, in turn, would make their products more acceptable to foreign consumers. Because these two views have such dramatically different policy implications, it is important to disen- tangle them. Our main contribution to this debate is threefold: an original test; an different methodology; and a new insight. First, by using firm-level product-specific data, we are capable of testing market learn- ing vs. productivity learning. To our knowledge, this kind of test has never been used before. Second, our methodology differs from the existing literature in several respects - most notably the use of duration analysis and the use of fixed effect coefficients to predict pre-export productivity. This reinforces existing results by demonstrating that they obtain irrespective of methodology. We also combine 15 years of census data with a detailed firm-level survey. This enables us to gain a deeper understanding of the 2  issues as the two data sets are complementary to each other. Last but not least, we show that much of the exporting behavior of firms can be understood in terms of individual products and market learning. Using detailed firm-level and product-level information, we show that much of the evidence weighs in favor of the market learning model. Manufacturers who export usually do so shortly after the firm is created; 42% export within a year of initiating production; 75% export within three years of their creation. A similar pattern is observed for individual products: new products put on the market by existing firms are either exported right away or permanently confined to the domestic market. For products that end up being exported, 80% are exported within a year of production. These effects are robust in the sense that they obtain even if we control for sector, region, year of production, and experience. The need to familiarize oneself with export markets is a sunk cost that probably restricts entry, as suggested by Roberts and Tybout (1997). Consequently, exporting firms often specialize in exports, especially if they are small. Familiarity is not enough, however: firms also need to be more productive to break into export markets and remain an exporter. Tn this respect, we find little evidence that general firm experience raises exports. Export experience, however, significantly reduces the time firms take to export new products. These results are consistent with market learning. They also agree with those reported by Aw, Chung and Roberts (2000) for Taiwan and Korea. We also find that, for individual products, market learning or market adaptation is rapid: one or two years. During this learning period, the domestic market absorbs a slightly larger share of the firm's production. But this effect is temporary.2 As in the rest of the literature, we also find a strong relationship between exports and firm productivity. The causality seems to run from productivity to exports: firms that export were more productive before doing so. This finding is consistent with the result of Aw et al. (2000) that much of the relationship between productivity and exports is due to self-selection. Looking at Colombia, Roberts and Tybout (1997) also find that efficient firms tend to self select into the export market. They find no evidence for learning-by-exporting in explaining why exporters are more productive than non-exporting plants. 2Preliminary analysis also show no relationship between exporting and total factor productivity. If anything, exporters tenxd to display lower levels of labor productivity. This is consistent with the fact that, in Morocco, it is the least capitalistic sectors that exports. It also matches the claim made by surveyed foreign firms that they invested in Morocco because of its low labor cost. A detailed investigation of this issue is the object of future research. 3  Like these authors, we also find that, among exporting firms, export experience does not significantly raise productivity. Using US manufacturing data, Bernard and Jensen (1999a) also find a causality from productivity to exporting and not the reverse. While exporting plants have substantially higher productivity levels, there is no evidence that exporting increases plant productivity growth rates. Looking at African firms, Soderbom and Teal (2001) find that workers' schooling and experience - i.e., learning by workers - is not what drives exports. The underlying efficiency with which the firm operates is a stronger determinant of exports. All these results are consistent with Liu and Tybout (1996)'s claim that, in Chile and Colombia, productivity growth takes place largely through entry and exit of firms, not through (productivity) learning. Similar results are reported by Aw, Chen and Roberts (2001) for Taiwan.3 The paper is organized as follows. Section 2 presents a brief conceptual framework. The data sets are introduced in Section 3. Results on the relationship between firm age and the propensity to export appear in Section 4. Duration analysis is presented in Section 5. Productivity is investigated in Section 6. 2. The Conceptual Framework There is an abundant literature on the relationship between productivity, learning, and manufacturing exports. Looking at Colombia, Roberts and Tybout (1997) for instance find that efficient firms tend to self select into the export market. They attribute their finding to the presence of sunk costs in entering the export market: only productive firms choose to incur the costs and enter foreign markets. The usefulness of the sunk cost approach is further illustrated by Das, Roberts and Tybout (2001) and Aw, Roberts and Winston (2001). Tn their comparison of Moroccan, Colombian, and Mexican manufacturers, Clerides et al. (1998) find that Moroccan exporting firms do better than non-exporters. Their main hypothesis is that there are 3According to Bernard and Jensen (1999a), exporting in the U.S. is associated with the reallocation of resources from less efficient to more efficient plants. These reallocation make up more than 40% of total factor productivity growth in the manufacturing sector. Half of this reallocation occurs within industry and the direction of the reallocation is towards exporting plants. Bernard and Jensen (1999b) simulate the effect of globalization in the form of a 5% drop in all geographic barriers between countries and find that nearly 9% of US plants would die. However, among surviving firms, one in seven that had previously sold only to the domestic market starts exporting. Since globalization provides lager markets to survivors and since the survivors were larger to begin, the decline in manufacturing employment is less than 3%. 4  fixed costs associated with exporting; producers of large batches are better able to spread these costs. The authors speculate that "most of the impetus to become exporter in Morocco came from firm specific demand sides shocks. Many Moroccan exporters are young plants that were founded with the exclusive purpose of selling particular apparel and textile products abroad." Taking these insights as starting point, this paper examines more in detail one particular type of sunk cost, market learning. To demonstrate the relationship between learning and individual products, we construct a simple export model of the firm with both productivity and market learning. Firms are assumed to have one or several product lines. Why they have multiple lines of production is not modeled explicitly, but it could be because each product line is subject to decreasing returns to scale beyond a given threshold, or because product lines benefit from economies of scope. The output of product j by firm i is denoted Qi,. The total number of products is J. Each output can either be sold domestically or exported. For simplicity, we ignore the possibility of multiple export destinations and focus on a single one. For Morocco, this is a reasonable assumption given that most manufacturing exports go to a small group of European countries. Exports are denoted Xij; domestic sales are written Dij. The export and domestic prices are written pf- and pdj, respectively. Prices are net of transport and marketing costs. Firms take prices as exogenously given. To obtain a model in which producers need not fully specialize in either market, we assume an Armington function of the form: Q,j = (X,: