عمده فروشان در تجارت بین المللی

Recent empirical research in international trade has revealed overwhelming evidence that, in all countries, a remarkably small proportion of firms report exports in Customs statistics. However, a large share of these are wholesalers. This suggests that the number of producers selling their products abroad might be much greater than that suggested by a simple count of the firms directly reporting their exports. This paper sheds light on the role of wholesalers in international trade. Our model uses very general assumptions to show that intermediated exporters may contribute significantly to the extension of countries' export opportunities. The model predicts a twofold role in international trade. First, wholesalers alleviate the difficulty of reaching less-accessible markets. Second, they help less-efficient firms to supply foreign markets, thus increasing the number of exported varieties at the aggregate level. We use French firm-level export data to provide empirical support for these two predictions.

There is now well-accepted empirical evidence that internationalization is only for the few (Mayer and Ottaviano, 2007). Only a handful of producers report exports in Customs statistics. In France, about 17% of manufacturing firms export (Eaton et al., 2004), and Bernard et al. (2007) find an analogous figure of 15% in US data. Economists are thus inclined to think of the process whereby firms engage in international trade as being extremely selective. In the seminal model of Melitz (2003), selectivity results from the requirement that prospective exporters build their own distribution network abroad. But recent evidence on the role of intermediary firms in trade may in fact suggest that selectivity has to date been overestimated. Indeed, a considerable proportion of the firms filling in export declarations are not producers of goods. This indicates that many more firms than those appearing in official Customs statistics supply their products to foreign consumers. Bernard et al. (2010) calculate that wholesale and retail firms account for approximately 10% of US exports. The share is 22% of total exports from China (Ahn et al., 2011), 11% of total exports from Italy (Bernard et al., 2011) and 35% of Chilean imports (Blum et al., 2010). In France, the country we focus on here, intermediaries account for 20% of total exports. Two main stylized facts emerge from the examination of these particular exporters. First trade intermediaries are more prevalent in difficult markets. For instance, Ahn et al. (2011) find that in China the share of exports via intermediates is greater in countries with smaller market size and higher variable trade costs. Bernard et al. (2011) further note in Italian data that intermediary exports are less sensitive to proxies for market-entry costs (import barriers and a governance indicator) than are direct exports. Second, the few existing datasets providing information on firms that export indirectly reveal that these are on average less efficient than direct exporters. McCann (forthcoming) uses survey data for a large number of Eastern European countries1 to show that firms which export directly perform better than either those using an intermediary or domestic firms. His finding holds for many different proxy measures of firm performance.2 It has been confirmed by Lu et al. (2011)3 who find for 29 developing economies that the most productive firms export directly while the least productive resort to intermediaries for exporting. Abel-Koch (2011) also observes a negative correlation between firm size and the relative importance of intermediated exportsin Turkey.4 We propose and empirically test a very simple model that predicts the two stylized facts discussed above: (1) intermediaries are relatively more present in markets that are more difficult to penetrate (i.e. smaller, further away, more protected, etc.); and (2) intermediaries channel exports from less-efficient firms which otherwise would not be able to pay the fixed cost of exporting directly. Recent models extending Melitz (2003) to account for intermediary activity5 assume an intermediation technology, which allows wholesalers to exploit some kind of advantage (such as economies of scope or better knowledge) in exporting over small exporting producers (Ahn et al., 2011, Åkerman, 2010, Blum et al., 2011, Bernard et al., 2011 and Felbermayr and Jung, 2011). Domestic producers are modeled as facing a choice of how to export: by exporting directly to foreign markets, and incurring the fixed costs of exports and trade costs, or via a specialized firm (the trade intermediary). Intermediaries are typically supposed to charge their clients (indirect exporters) an intermediary fixed cost, which is lower than the fixed cost of direct exports, and an additional marginal cost. The intermediation technology thus provides a mechanism by which firms can access export markets even if they are not productive enough to establish their own direct export network.6 All these models predict an efficiency-ordering of firms into three categories (non-exporters, intermediated exporters and direct exporters): non-exporters are less efficient than those using an intermediary, while the latter are less efficient than firms which export directly. This assumption of wholesalers acting as a trade vehicle for less-efficient firms has so far remained largely untested. Along the same lines as the theoretical work noted above, our model assumes an intermediation technology that reduces the fixed cost of exporting in exchange for a higher marginal cost. Our model differs from the existing literature in two ways however. First, the prediction (1) relating the prevalence of intermediary exports to the accessibility of the foreign market is obtained in a very simplistic and intuitive framework, and does not involve any restrictive hypotheses regarding the intermediary technology. We simply introduce a specific (i.e. non-ad-valorem) trade cost into a very standard trade model. This fairly reasonable assumption suffices to show that the share of bilateral exports handled by wholesalers is negatively correlated with all possible determinants of foreign-market accessibility. Åkerman (2010), Felbermayr and Jung (2011) and Ahn et al. (2011) have proposed theoretical contributions leading to similar conclusions. In Felbermayr and Jung (2011) and Åkerman (2010), however, not all of the determinants of market accessibility influence the export share. Only the severity of contractual problems and the fixed cost of exporting play a role in these models; foreign-market size and variable trade cost have similar effects on both direct and indirect exports. Ahn et al. (2011) do show that all of the determinants of market accessibility influence the extent of indirect exports, but at the cost of a somewhat controversial hypothesis: they assume that once a firm pays to use an intermediary, it will supply all export markets. In our model, the fixed cost of intermediation is destination-specific so that firms are not systematically present on all foreign markets. Our second theoretical contribution is related to the prediction (2). While the existing literature supposes that the firms' decision to export directly or indirectly is only driven by productivity, we also introduce an efficiency-ordering of firms based on the quality of their variety. We consider two polar cases. In the first, the sorting of firms into export markets depends upon individual productivity (or marginal cost) draws; in the second, firms with higher marginal costs produce higher quality. This yields a novel prediction regarding the price of wholesalers' exports relative to that of direct exporters. In the productivity-sorting setting, we predict that intermediates will export more expensive varieties, as they export on the behalf of relatively higher-cost manufacturers. In the quality-sorting setting, by way of contrast, they export the least expensive varieties corresponding to lower-quality manufacturers. These contrasting predictions will be exploited to test the hypothesis that direct exporters are more efficient on average than indirect exporters. We use French firm-level data to test the two main predictions of the model. Our estimates confirm prediction (1), i.e. wholesalers channel a greater share of total exports in more difficult markets. The analysis of prediction (2) – that intermediaries channel exports from less-efficient firms – is conducted using export unit prices. Only two pieces of work have provided empirical evidence on the price difference between direct and indirect exporters, and they disagree. Ahn et al. (2011) find for China that export prices by intermediaries are about 6% higher than those of direct exporters, while Bernard et al. (2010) conclude that wholesalers in the US have lower (by 14%) unit values than do pure producers. Our model and our empirical results, reconcile these apparently contrasting findings. Our model predicts that the average price charged by wholesalers is higher than the average price of direct exports when firms' export performance is mainly driven by their productivity. But when quality sorting prevails, our model shows that this price premium is smaller and even negative for some parameter values. This prediction is supported by our empirical analysis. The remainder of the paper is structured as follows. In Section 2, we present our model and the two propositions (in terms of the prevalence and the prices of intermediaries) that we will test empirically. Section 3 describes the data, and Section 4 presents our empirical results. We conclude in Section 5.