سرمایه گذاری غیرقابل قرارداد و ادغام عمودی در صنعت کفش مکزیکی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|1079||2002||28 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : International Journal of Industrial Organization, Volume 20, Issue 8, October 2002, Pages 1197–1224
This paper examines patterns of integration among manufacturers and retailers, using data from a survey of footwear manufacturers in Mexico. The property rights framework, developed in papers by Grossman and Hart [1986, J. Polit. Econ. 94: 691] and Hart and Moore [1990, J. Polit. Econ. 98: 1119], is differentiated from the standard empirical transactions cost framework. In the context of this industry, the most relevant distinction between the two frameworks is that the property rights framework addresses both the benefits and costs of integration, while the transactions cost framework focuses only on variation in the benefits of integration. We show that the costs of integration are highest where the retailer’s non-contractible investment has an important effect on the overall profits from the relationship. Consistent with the property rights framework, the data suggest that integration is less likely in these circumstances.
Transactions cost economics posits that a firm’s decision to make or buy a part or service is affected by the presence of relationship-specific investments. Vertical integration (making) reduces the risk of hold-up once relationship-specific investments are sunk. The power of this idea, pioneered by Klein et al. (1978) and Williamson (1979) has been demonstrated by an extensive empirical literature.1 This empirical literature has shown that the benefits of integration increase in the presence of relationship-specific investments. Grossman and Hart (1986) and Hart and Moore (1990) provide a formal model which explicitly considers the costs and benefits of vertical integration. Integration is seen simultaneously to alleviate and to create hold-up risks. Efficient ownership depends not only on the degree of relationship specificity of investments, but also on the importance of those investments in determining the profits of the trading relationship.2 Grossman/Hart and Hart/Moore have received considerable attention in the literature, and numerous theoretical refinements have recently appeared.3 But 9 years after the formal model was published, Hart reported “Unfortunately, there has to date been no formal test of the property rights approach…” (Hart, 1995, p. 49).4 This paper attempts to address this gap by providing a test of the property rights model using data on manufacturer-retailer integration in the Mexican footwear industry. Asset ownership in the property rights framework is driven by the need to balance incentives for non-contractible investments made by two managers. If one manager’s investment has a greater impact on the gains from trade or is more specific to the relationship (as these are defined below), then that manager should be given stronger investment incentives. For contractible investments, stronger incentives can be written into the contract. But when contracts cannot be written and enforced, owning assets increases investment incentives. Hence, determining the efficient pattern of asset ownership requires an understanding of how both the specificity and the importance of investments made by both managers vary across firms in the industry. This paper considers relationships between footwear manufacturers producing made-to-order goods and the retailers to whom they sell. A manufacturer must make non-contractible investments which affect the quality of workmanship of the goods produced. Standard purchase contracts allow retailers to cancel delivery of previously ordered merchandise under conditions described below. Cancellation may come after the manufacturer has produced the goods. The possible abuse by the retailer of cancellation rights subjects the manufacturer to risk of losses resulting from the cost of finding alternative buyers for heterogeneous goods. Hence his incentive to invest in workmanship quality is compromised. Integration into retailing by a manufacturer relieves the risk for the manufacturer, but simultaneously creates a risk for the manager of the retail store. Retailer managers make effort investments in providing better service for customers and in learning the tastes of their customers, allowing a better selection of goods on the shelves. But since a non-owning retail manager can be separated from the store (fired), she faces the risk of losing her investment. Hence her incentive to invest is compromised. The efficiency loss in either case will depend on the specificity and importance of the investments made by either manager, discussed in more detail below. The importance and specificity of the investments made by the two managers vary across different types of products. We focus on how the frequency of integration changes as three product characteristics change: the heterogeneity of goods produced, the quality of materials used in production, and the rate at which fashions change. In the first two cases, the patterns of integration are consistent with the predictions of both the property rights framework and the standard empirical transactions cost framework. With respect to increases in the rate of fashion turnover, however, the transactions cost framework leads to an expectation of more integration and the property rights framework to less integration. This allows the property rights framework to be differentiated from the transactions cost framework. The data indicate that integration is less frequent when fashion turnover is more rapid, suggesting that the stronger incentives store ownership provide for the retailer are needed when her investment is more important to the relationship. Thus the property rights framework is found to do a better job of explaining the patterns of integration than is the standard transactions cost framework. The motivation of local managers through ownership has also been examined in the franchising literature.5Maness (1996), for example, argues that local ownership (franchising) provides a greater incentive for cost minimization by the managers of local outlets. Lutz (1995) examines franchising with a model which follows the property rights framework used in this paper very closely. The relevant actions are ‘day to day decisions’ made by the local manager. These decisions affect both current and future profits. Contracts, including those governing division of future profits, are incomplete, and strengthening incentives to one party necessarily weakens incentives for the other. Using her framework, Lutz concludes that franchising is more likely when the manager’s non-contractible actions have large impacts on the future profitability of the local unit. The manager’s incentives are increased by her ability to sell the unit and capture the gains from these investments in future profitability. In company-owned outlets, the incentives of managers are dampened by the possibility that they will be fired before realize the returns to these actions. Lutz model captures the essential tension in the case examined in this paper. The empirical application of property rights requires not only data but also an understanding of contractual relations among firms in the industry. Such an understanding is needed to match the property rights model to the industry and develop a set of testable predictions. For this paper, the necessary institutional detail of the Mexican footwear industry comes from some 250 interviews conducted by the author during 1992 and 1993. The interviews and resulting data are described in Appendix A. The next section of the paper provides a brief overview of the industry, focusing on the relationships between manufacturers and retailers. Section 3 relates these relationships to a simple property rights model and develops a set of predictions of patterns of integration in the industry. The data are discussed and the results reported in Section 4. Finally, Section 5 discusses the results and concludes.
نتیجه گیری انگلیسی
The transactions cost and property rights frameworks are differentiated by the predicted effect of increased fashion turnover on integration. Driven by an increase in the absolute level of specificity in the relationship, transactions cost predicts an increase in integration as fashion turnover increases. Driven by a change in the relative specificity and importance of marginal investments made by manufacturing and retailing managers, property rights predicts less integration as fashion turnover increases. Overall, the data provide support for the property rights framework: independent ownership is more likely in segments with high fashion turnover. While the negative relationship between fashion turnover is and integration is consistently negative, the finding is statistically significant only in some of the specifications. The negative association is not significant, for example, in a randomly selected subsample. The other empirical findings-increased integration among producers of heterogeneous goods and goods using higher quality materials — are consistent with either framework. These findings suggest it is important to consider how the costs of integration vary in the data. Previous empirical work in transactions cost has focused exclusively on variation in the benefits of integration. The importance of the retailer’s investment in high fashion segments of the industry drives the results. This depends upon the particular characteristics of the industry — a very large number of manufacturers, most of small size (fewer than 100 workers), and an even larger number of mostly independently owned retail stores. This structure is quite different from the footwear industry in the United States, for example, which is characterized by a much smaller number of large manufacturers. In the US market, where the manufacturer’s brand name is much better known, the retailer’s investment in building a client base is likely to be less important. Nevertheless, the conclusion that retailers need stronger incentives in situations where their investments are more important may apply to other industries, such as art galleries or high end restaurants. Where investments in building a client base are important, both vertical and horizontal integration should be less frequent. Brickley et al. (2000) provide some additional support for this conclusion, showing that the importance of local knowledge is an important determinant of bank branch ownership in Texas. There are, of course, many alternative explanations for integration. One alternative rationale for integration is examined in Shepard (1993), who looks at vertical integration in the gasoline market under the assumption that firms integrate to avoid double markup pricing. From this perspective, differences in demand elasticities across segments could generate a pattern of integration incentives. No measures of demand elasticities in the various segments are available. Greater heterogeneity might well imply less substitutability between goods (and lower elasticities) in women and dress segments, which might produce the higher integration found in those segments. However, differentiation across time, measured by fashion turnover, should also result in lower demand elasticities and an incentive to integrate. The data here suggest the opposite. Agency theory provides another explanation.25 The agency literature views ownership as the right to the residual profit stream from an asset, and generates integration from the need to provide profit incentives. Grossman and Hart assumes away agency considerations with a claim that integration does not make available any new contracts. An integrated manufacturer can pay his retail manager the profit stream of the retail store without selling the store to her. If in practice profit streams and asset ownership are correlated, then it may be that the integration measure here is proxying for lower powered incentives for the retailer. But for agency theory to explain the patterns in the data, monitoring the retailer would have to be easier in the heterogeneous segments than in the homogeneous segments. This seems unlikely, though no direct evidence is available. The industry demonstrates that the property rights framework can lead to patterns of integration different from those identified by the transactions cost framework, especially as the latter has been developed in the empirical literature. The exercise also demonstrates the difficulty in matching the specific components of the property rights model to a real world setting. Detailed information about the structure of trade in an industry is required. Some part of the required information must be gathered in a qualitative manner, making replication more difficult. Nevertheless, the attention which the Grossman/Hart and Hart/Moore papers have received, further empirical evidence, such as that provided by Baker and Hubbard (2000), is needed.