ساختار انتخاب محل برای سرمایه گذاری های تولیدی جدید آمریکا در آسیا و اقیانوس آرام

Despite an extensive literature on the determinants of the foreign location choices by multinational firms, researchers have only recently begun to systematically examine how these firms form their location consideration sets. When considering new foreign locations, do firms evaluate the attributes of the alternatives at the national level, the sub-national regional level, at some other level of geographical aggregation, or using some combination of these? This paper employs discrete choicemodels to examine how U.S. multinational firms form their consideration sets over locations in the Asia-Pacific area and to identify some of the relevant location attributes. The results are consistent with a sequential, or hierarchical, decision-making process in which firms first select a host country based on one set of attributes and then select a region within that country is selected based on another set of attributes. Most of the location attributes that are found to be significant are related to productivity-enhancing features, such as worker skills, industrial agglomeration, and extensive transportation infrastructure, rather than features related to factor prices, such as low-wage labor.

Multinational firms seeking locations for new investments abroad are faced with a choice over at least 150 alternative host countries. Many studies of the location of foreign direct investment assume that the representative firm making these choices is a rational decision maker that considers all relevant attributes of all of these alternatives. This strong assumption is not easily reconciled with case study literature and with research in human decision making. In a number of case studies of U.S. firms investing abroad, Aharoni (1966) (p. 82) notes that ‘‘the decision to look abroad is made in most cases in terms of a specific project in a specific country’’ rather than an all-encompassing evaluation of a menu of alternatives. When firms do evaluate more than one alternative, it is doubtful that the number of choices considered even begins to approach the number of all possible host countries in the world. Research on human cognition and on marketing suggests that people are adept at analyzing groups of four or fewer items. Gigerenzer (2007) (p. 155) notes that, when the typical person encounters a group of people, he or she is able to instinctively count up to four people. The author also cites the Ancient Roman custom of giving ordinary names to the first four children but giving a numerical name (e.g. Quintus, Sixtus, etc.) to each subsequent child. In the marketing literature, Wright (1975) found that individuals are very effective at deliberating over a few alternatives but that the effectiveness of their decisions declines rapidly as the number of alternatives increases. This study examines U.S. multinationals’ choices of new manufacturing sites in four major host countries in the Asia- Pacific area (Australia, China, Japan, and South Korea), and the subnational regions within those countries. It is concerned with both how multinationals evaluate potential sites and the location attributes considered. The first question deals with the ability or willingness of multinationals to deal with a large number of location alternatives. Do firms engage in a computationallydemanding simultaneous choice over all sub-national regions in the Asia-Pacific area, or do they settle for a less-demanding sequential choice process consisting of a choice of country followed by a choice of region within that country. The second question deals with the location attributes that U.S. multinationals consider. The location attributes that will be considered include market size, labor characteristics, industrial agglomeration, and whether or not the investor had a prior investment in the host country.

The central question of this paper is how to best approximate the location choice process of multinational firms in econometric studies. In particular, can the choice process be modeled as a simultaneous choice process in which the decision maker considers all relevant attributes of all alternatives, or are the alternatives and their attributes so numerous that multinationals must employ heuristics to bring the problem down to a manageable size? The heuristic considered here is a sequential choice process in which a country is first selected based on national attributes and then a region within that country based on regional attributes. Three econometric models are considered: two conditional logit models that depict simultaneous choice processes, over either the four candidate host countries (‘‘national choice model’’) or the 36 sub-national regions in these countries (‘‘regional choice model’’), and a nested logit model that depicts a sequential choice process beginning with the selection of country, followed by the selection of a region within that country (‘‘nested choice model’’). In addition to the structure of the location choice process, the results also shed light on the location attributes that are important to multinational firms.