به سوی یک مدل موازنه برای انتخاب بازار بین المللی

The literature on International Market Selection (IMS) contains many proposed models which make significant contributions but do not effectively address the IMS problems. A new tradeoff model is proposed that uses two key constructs, demand potential and trade barriers, as well as firm strategy as a contingency construct. Each key construct is measured by only four variables, resulting in simplicity and low application cost, and strategy is used to develop weights for the variables. The model is tested in real market conditions for 17 target countries and three diverse products over a six-year period. The weighting solution that favors demand potential is shown to have strong predictive power for the short to medium term. In addition to being tested and validated, the model introduces a number of advances including the weighting of indicators, an approach to quantifying nontrade barrieers, and validation for two different exporting countries.

International market selection (IMS) is the first and most important step in export strategy (Root, 1994), making it a critical success factor for both smaller exporters and mature multinational firms. Systematic IMS contributes to export success while wrong choices can put the firm in an unfavorable strategic position ( Papadopoulos & Denis, 1988). Given its importance, IMS attracted significant research attention from the 1960s to the mid-1980s. However, this interest waned in later years, mainly because of the difficulty in developing IMS models that would be generalizeable to various industries while also having adequate predictive power for business ( Douglas & Craig, 1992). As a result, the literature to date is limited to either general qualitative frameworks or operational models that have not been tested sufficiently, offer little or no evidence that they can in fact predict market attractiveness, and/or are too complex to apply in practice. On first review, a promising exception appeared to be the shift-share model ( Green & Allaway, 1985), which Douglas and Craig (1992) rightly called “the only new approach” to have been proposed in recent years. The model’s core strength is that it appears to be able to predict the industry-specific attractiveness of target markets using two simple and readily available measures: their import size and import growth rate. The model’s major apparent weakness is its exclusive reliance on import measures. Since exporters compete not only among themselves but also against domestic producers in the target countries, the initial intent of this research was to explore ways for extending the model to include total demand measures, in order to reflect the true attractiveness of potential target markets. However, an in-depth review indicated that the base model itself in fact lacks predictive power and is not well grounded theoretically. As a result, this study was revised to develop a new model that would build on the core logic of shift-share but also attempt to meet Douglas and Craig’s (1992: 301) call for “more sophisticated, parsimonious [market selection] techniques in a multi-country context”.

Research of this type is limited primarily by the deficiencies of secondary data. Key challenges include the lack of direct conversion schemes between the trade coding systems (the SITC codes alone have been revised three times since 1965 and some countries still report using earlier versions), varying data definitions, and the unavailability, unreliability, or aging of data for some countries, particularly less developed ones (e.g., see Nachum, 1994). Greater product specificity would also be preferable to manufacturers of product sub-categories (e.g., lawn vs. upholstered furniture), for whom data aggregated at the SITC-ISIC levels used here may be misleading. Lastly, the nature of some of the selected variables may also limit model applications. For example, apparent consumption measures would be of little interest to firms seeking first mover advantages in countries where the market for their product is presently nonexistent, and the country-level variables reflect factors that affect some industries more than others (e.g., geographic distance has a greater effect on the transport of bulky goods, and nontariff barriers would ideally be measured at the industry level). In this study we attempted to address as many of these limitations as possible, mainly by following various approaches proposed in earlier studies (e.g., use of the UN Statistical Office guidelines for code matching); applying a variety of theoretically-grounded ad hoc solutions where necessary; using variables that past research has confirmed as the most appropriate and important, given the screening nature of the model, and as being relevant to a wide a range of industries; and using a very parsimonious set of indicators. This, coupled with the model validation in real market conditions, makes us reasonably confident that potential negative effects were minimized. Larger multinational firms, which would need to apply the model separately for each of the industries in which they compete, also have more resources at their disposal and may in fact be able to obtain more product-specific information. As well, modern information technology should further help to address the deficiencies of secondary data. The situation is improving even in the case of less developed countries where, more than 10 years ago, Day et al. (1988: 17) were able to find “enough data [18 indicators] on enough countries [96] to permit a meaningful analysis.” Carefully selected proxy variables remain a promising alternative for other cases (e.g., in another phase of this research we found that the imports-to-GDP ratio is a good substitute for import penetration, for which data is often unavailable). Therefore, notwithstanding its limitations, the model provides a significant improvement over earlier ones since (a) it captures total rather than import-only demand; (b) it is industry-specific and efficient, unlike most multiple criteria models; (c) it was tested using three different products and, unlike any previous model, using two very different exporting countries; (d) unlike econometric methods, it is generalizeable across industries, and (e) it was externally validated. The performance of the total score model was good except for a few cases, suggesting that careful selection of constructs, measures, and weights can pay off. With the proper weighting scheme, which in this case reflects opportunity-oriented business behavior, the model’s ability to predict future imports and market shares in selected target markets was high. Excluding product exports at the growth stage, the lack of significant differences in predictive power between the two exporting countries suggests that the model may be used by both developed and developing countries. Long-term shifts in the structure of a country’s GDP or the ratio of domestic production to imports, and/or major events causing such shifts in the near-term, cannot of course be predicted and would limit any model’s applicability. However, this model showed no sign of sharply declining predictive power over the six-year period, suggesting the absence of major lag effects and, therefore, that it may be used with reasonable confidence for short- to medium-term market planning. The consistency of findings between the two-dimensional and total score approaches, and the predictive power of Wdp, point to several practical and theoretical benefits from this research. For business, statistics can only tell a small part of the story. Detailed sales potential analysis would be needed, and firm and product characteristics would need to be taken into account, before a decision to expand. Therefore, what a model such as this can provide is a systematic, relatively inexpensive, and effective method for scanning a large number of markets, whose potential usefulness is underscored by the high cost of wrong IMS decisions. The individual indicators can also be used to help develop different strategies even for target countries which appear to offer similar overall opportunities. In addition to having more resources, and thus likely being in a better position to obtain more product-specific data, large multinational firms may find the model particularly useful since they typically export several related products in each sector, making the SITC-ISIC data level less of a concern. For public policy makers and export promotion agencies, the finding that firms in different industries from the same country should explore different export markets indicates that generic export assistance programs, which are common in most countries (Seringhaus & Rosson, 1991), will likely be less effective than programs tailored to the needs of particular sectors. From the theory standpoint, the in-depth review of the shift-share model may lead to further research in marketing and/or regional economics aimed at rechecking its assumptions and perhaps revising it. In addition, the study provides a theoretical framework for IMS through a tradeoff model that draws from and improves upon the tradition of multi-criteria and demand estimation approaches. Instead of providing a “universal list” of market opportunities, which can be highly misleading for different industries, the model can identify unique industry-specific opportunities. Given that the more specific the segmentation basis, the less stable the segment (Wind, 1978), the model uses a reasonable analytical level (SITC 2- or 3-digit) while also representing a step forward compared to general country rankings. The predictive power of the Wdp weighting system is, in itself, an interesting comment on the export behavior of firms, as well as being the first validated IMS model that accounts for firm strategy. Lastly, the research suggests a potentially useful approach for further work on IMS model development and validation. Future studies may focus on at least three areas that call for a renewed research effort. First, replications of the tradeoff model to test its validity using different importing country sets, exporting countries, products, and time periods. Second, efforts to expand the model beyond the exporting domain, since with suitable adjustments the underlying logic should be applicable to market selection for other entry modes, including higher control ones (such as direct investment) which would be of greater interest to larger MNEs. Finally, future studies may improve the model by examining its various individual aspects and limitations in more detail. This may include, for example, testing of different variable sets, confirmatory factor analysis using a larger country sample to validate the two dimensions, refinements that allow for the differential weighting of critical factors such as geographic distance, depending on their relevance at the firm level (Hoffman, 1997), or a search for proxy variables to further simplify the procedure or to enable its application to countries for which the kind of data needed for this model may not be available.