طراحی سیستم های مونتاژ بین المللی و زنجیره تامین آنها تحت پیمان نفتا

The purpose of this paper is to provide a decision support aid for the strategic design of an assembly system in the international business environment created by NAFTA. The strategic design problem is to prescribe a set of facilities, including their locations, technologies, and capacities, as well as strategic aspects of the supply chain, selecting suppliers; locating distribution centers; planning transportation modes; and allocating target levels (i.e., amounts) for production, assembly, and distribution. The objective is to maximize after-tax profits. This paper presents a mixed integer programming model that represents the complexities of the international design problem as well as relevant enterprise-wide decisions in the US–Mexico business environment under NAFTA. It deals with a broad set of design issues (e.g., bill-of-material restrictions, international financial considerations, and material flow through the entire supply chain) using effective modeling devices (e.g., linearizing non-linearities that arise in modeling transfer prices and allocating transportation charges). Examples demonstrate how managers might use the model as a decision support aid.

Global competition is increasing the need for enterprises to internationalize, using the production sharing strategy to locate operations in countries that offer comparative advantages. Trade allows countries to allocate natural, labor, and capital resources more efficiently, resulting in productivity increases and economic gains that improve income and living standards. According to a Wall Street Journal editorial ( Editorial, 2004) “The point of free trade isn’t to create jobs per se but to allow resources to find their most efficient use and re-deploy workers to better paying jobs. Manufacturing networks incorporating the comparative advantages of all three NAFTA members have made North America an attractive investment for global capital”. In particular, the North American Free Trade Agreement (NAFTA) (The NAFTA Secretariat) has encouraged US-based companies to locate assembly operations in Mexico. No widely accepted tools provide decision support for the strategic design of an assembly system in this international environment. The purpose of this paper is to address this need. The strategic design problem is to prescribe a set of facilities, including their locations, technologies, and capacities, as well as strategic aspects of the supply chain, selecting suppliers; locating distribution centers; planning transportation modes; and allocating target levels (i.e., amounts) for production, assembly, and distribution. Strategic decisions design the system and its capacity for a relatively long horizon (e.g., 2–10 years) with the objective of maximizing after-tax profits. The objectives of this paper are a model that represents the complexities of the international design problem, integrating relevant enterprise-wide decisions in the US–Mexico business environment under NAFTA, and examples that demonstrate how managers might use the model as a decision support aid. Initiated in 1993, NAFTA (The NAFTA Secretariat) furthered relationships between the US and Mexico, the world’s 1st and 10th largest economies. Trade between the US and Mexico jumped 74% from 1994 to 1999 to over $41 billion (Brezosky, 2003). The 1250-mile Texas–Mexico border fosters close socio-economic ties. The Texas Border Region (TBR) comprises four counties (Cisneros, 2001) with six cities, each of which has a “sister city” across the border. Both Texas and Mexico BRs participate extensively in production sharing under NAFTA. With low levels of education and worker skill, and high levels of unemployment and immigration, the TBR has historically been economically challenged (Yücel, 2001); “There are few places in the United States as desperate for economic development as the impoverished communities of the Rio Grande valley” (Pinkerton, 2001). Trade with Mexico promises new economic opportunities in the TBR. Before NAFTA, much of the trade between the US and Mexico was conducted under the Maquiladora Program (MP), which began in 1965. Comprised of low-cost, labor-intensive assembly plants that employed only unskilled labor (Vargas, 1998a, Vargas, 1998b and Vargas, 2000), it has allowed US companies to assemble products that are priced competitively in the world marketplace. Maquiladora import parts from the US, assemble them, and return end products to the US where distribution centers (DCs) typically offer better insurance coverage, protection against pilferage, and hedging against changes in exchange rates. Shipments into Mexico incur no tariffs, and the US shipper pays duties only on the value added in Mexico, stimulating job growth in the Maquiladora but discouraging use of suppliers in Mexico. Second-generation operations, which involve higher levels of technology using modern management techniques (e.g., just-in-time (JIT)), began in the 1980s. More recently, third-generation operations (Buitelaar and Perez, 2000) have involved capital-intensive, state-of-the-art manufacturing techniques and more highly skilled labor. Maquiladoras continue to thrive under NAFTA, accounting in 2000 for 29% of the manufacturing jobs in Mexico and exports valued at over $79 billion, 47.7% of Mexico’s total exports (Vargas, 2000). One in five Texans live in the TBR (Brezosky, 2003). The population of one pair of sister cities, Laredo and Nuevo Laredo, has grown 48% over the last decade. Daily northbound vehicle crossings at Laredo increased from 2.7 million in 1994 to 4.3 million in 2001. Each day, some 15,000 trucks cross the border. In 1999, $30 billion of US exports and $35 billion of US imports came through Laredo, stimulating transportation and distribution industries (Phillips and Manzanares, 2001). Typically, a carrier transports parts from a plant in the US to a warehouse in Laredo where they are cross-docked to a drayage carrier that takes them through customs and across the border where they are cross-docked to a Mexican carrier for transport to an assembly plant. Finished products retrace this route to a DC in Laredo for subsequent distribution. The Texas–Mexico business environment will change markedly over the next decade, presenting new needs that this paper addresses. The recent recession combined with competition—especially from China, which offers low labor rates—led to some downsizing in the Maquiladora. Various countries around the world (Canas and Coronado, 2002) now compete in production sharing. The future holds more challenges to deal with increasing labor costs, educating workers, and evolving Mexican fiscal policies. Nevertheless, Mexico offers low cost and is expected to maintain a competitive edge, especially in assembly (Buitelaar and Perez, 2000). Environmental concerns must be addressed, NAFTA will allow duty-free trade by 2010 (Buitelaar and Perez, 2000), labor rates will increase with the education of the workforce, and the Mexican market will grow substantially. Supply chain design will change markedly (Alarcon and Sepeda, 1998). Only 3% of inputs are produced in Mexico while 80–85% are produced in the US and 12–17% are from global sources (Alarcon and Sepeda, 1998). Maquiladoras operating with the JIT system are encouraging—if not requiring—suppliers to locate nearby. Over the next decade, it appears that more operations involving higher levels of technology and more suppliers will be located in Mexico and more manufacturing facilities and DCs will be located in the TBR to exploit proximity, enhancing JIT capabilities. Mexican (US) trucks may deliver directly to destinations in the US (Mexico), perhaps affecting the distribution industry in Laredo. As a positive example of growth in the border region, Toyota will build a new $800 million plant in San Antonio, TX, to assemble Tundra trucks (Houston Chronicle, 2003) and a $140 million plant in Tijuana, MX, to make Tacoma trucks and truck beds, which will be transported to their plant in Fremont, CA, for assembly into trucks. Operations of these magnitudes will attract suppliers to locate in proximity on both sides of the border. The Texas–Mexico business environment (Yücel, 2001) involves these unique issues (e.g., NAFTA terms, supplier location, proximity, transportation, warehousing) as well as issues that are common to all international operations (O’Connor, 1997) (e.g., local-content rules, border-crossing costs, transfer prices, income taxes, exchange rates) but with parameter values that depend upon NAFTA terms and country-specific laws. Products must meet NAFTA-imposed rules of origin (The NAFTA Secretariat, Chapter 4) to qualify for preferential tariffs as they move across the border between member countries. These local-content rules require that a minimum content (i.e., proportion of constituent material and value added) originate within the NAFTA countries. The proportion of “originating material” can be measured relative to either the transaction value or the net product cost. Article 405 (The NAFTA Secretariat, Ch 4), known as the De Minimis rule, relaxes local-content rules, allowing most products to comprise up to 7% of non-originating material. Transports from one country to another may incur “border-crossing” costs, including tariffs, duties, and monetary exchange losses. In addition, cross docking expenses must be incurred if rules allow only carriers that are based in a country to transport in that country. If an enterprise ships material from one of its facilities in one country to another of its facilities in a second country, the receiving facility must pay a transfer price for the material. Tariffs, duties, and income taxes are based on transfer price (e.g., Brezosky, 2003, Canas and Coronado, 2002, Phillips and Manzanares, 2001, Vargas, 1998a, Vargas, 1998b and Vidal and Goetschalckx, 2001), which can be used to allocate a disproportionate share of profits to facilities in countries with lower income tax rates to increase the after-tax profit of the parent company. To prevent fiscal malpractice, countries enact laws to restrict transfer prices. A principle commonly adopted—the “arm’s-length” standard—invokes a market-based valuation for setting transfer prices, requiring that a transfer price between related facilities (i.e., those belonging to one enterprise) must be comparable to market prices used in transactions between unrelated facilities. The US Internal Revenue Code recognizes several methods for determining transfer prices, including comparable uncontrolled price (CUP), resale price, cost plus, comparable profits, and profit split. Transfer prices may be determined based on any acceptable method—or combination of them—as long as the arm’s-length standard is met. Our model incorporates the CUP method by restricting each transfer price to a range that represents market prices charged to unrelated facilities. Transfer prices generally do not include transportation charges but a firm can also manipulate them to reduce taxes, so our model also allocates transportation charges according to applicable laws. A special case arises if one facility transfers the same component to a number of related facilities. Vidal and Goetschalckx (2001) required all transfer prices in such a case to be equal, but this led to a non-linear model that they solved with a heuristic. In contrast, we follow Barfield et al. (2002): “Multinational companies may use one transfer price when a product is sent to or received from a company and a totally different transfer price for the same product when it is sent to or received from another.” We adopt this view to craft a linear, instead of a non-linear, formulation to facilitate solution. Exchange rate fluctuations present financial risks in international transactions. Disparities between buying and selling prices caused by exchange rates can cause significant changes to the revenue streams. Such risks may be dealt with by hedging policies and arbitrage transactions (O’Connor, 1997). Finally, some design issues are common to both international and domestic enterprises. For example, dealing with bill-of-material (BOM) flows poses special problems in assembly systems. At a node representing an assembly operation, incoming arcs carry flows of different components; and the outgoing arc, flow of the assembly. Such multi-commodity flows do not conform to the classical flow balance associated with minimum-cost network-flow problems. Related issues arise if a component is assembled into several end products, if one component can be incorporated in several assemblies that comprise one end product, and if facilities are flexible so that one component may be manufactured in several facilities or one facility can manufacture several components. The body of this paper comprises five sections. The first section reviews relevant literature, the second presents our model, and the third section discusses our model in detail. The fourth section exemplifies how managers might use our model as a decision support aid in a hypothetical setting that involves an enterprise that manufactures two types of laptop computers. “What if” cases explore centralized design, decentralized design, vertical integration, outsourcing, sub-contracting all manufacturing operations, economies of scale, sensitivity to alternative bounds on transfer prices, sensitivity to exchange rates, and the effects of governmental inducements. The fifth section relates conclusions and recommendations for future research.

This paper achieves its objectives, contributing a model that represents the complexities of the international design problem, integrating relevant enterprise-wide decisions in the US–Mexico business environment under NAFTA, and examples that demonstrate how managers might use the model as a decision support aid. An assembly system is designed by prescribing facilities, including the location, technologies to be employed, and capacity of each. Strategic aspects of the supporting supply chain are designed by selecting suppliers; locating distribution centers; planning transportation modes; and allocating target levels (i.e., amounts) for production, assembly, and distribution. The objective is to maximize after-tax profit. Our model differs from other MIP models in that it deals with typical international issues (e.g., local-content rules (4d) and (4e) border-crossing costs (2a) and (2b), transfer prices (4a) and (4b), income taxes (3a), (3b) and (3c), exchange rates (1)) as well as features that are unique to the US–Mexico business environment (e.g., NAFTA terms, supplier location, proximity, transportation, warehousing). It deals effectively with design issues such as BOM restrictions (e.g., assembly operations, one component used in several end products (see discussion related to (7)), and one component used in several sub-assemblies on one end product (see discussion describing (6)), component usage, and facility flexibility (e.g., one component may be manufactured in several facilities (see discussion related to (9)) and one facility can manufacture several components (see discussion regarding (8)). It linearizes non-linearities that arise in modeling transfer prices and allocating transportation charges (see the discussion immediately following the model) and addresses strategic aspects of transportation and distribution. It addresses relevant financial considerations, prescribing transfer price and transportation-cost allocations (2a), (2b), (2c), (2d) and (4f), invoking safe harbor rules (4c), modeling graduated income tax rates (3a), (3b) and (3c), and incorporating exchange rates (1). It allows inventory and backorders to be incurred at each stage in the production/distribution process and it integrates material flow through the entire supply chain (i.e., suppliers, production, assembly, distribution, transportation, customers). Examples demonstrate that our model can be applied to analyze a number of trade-offs, for example, centralization versus decentralization, make versus buy, outsourcing versus in-house assembly, flexible versus dedicated technologies, and economies of scale versus economies of process. It can also be used to evaluate a variety of factors such as limitations on transfer prices, facility locations, tariff impacts, exchange rate impacts, tax impacts, dollar valuation, local-content rules, and the costs of transportation and distribution. Future research could expand the set of considerations represented by the model, for example, to model prices as a function of quantity produced or transportation cost as a function of the quantity shipped. Future research can also exploit the structures embedded in our model to devise a specialized solution method for application to large-scale instances. In addition, future research can devise an effective approach to optimize stochastic versions of the problem, for example, in cases for which it is important to treat demand-and/or-exchange-rate uncertainty explicitly. Our research is continuing along these lines.