Wages, Productivity, and Foreign Direct Investment Flows
Larudoo, Mehrene, Koechlin, Tim, Journal of Economic Issues
Will an increase in economic openness cause relocation of capital and jobs from high-wage countries to low-wage countries? The standard textbook response is that it will not: low wages reflect low levels of labor productivity and thus provide no incentive for mobile firms to relocate. Wage rates vary widely by country; unit labor costs (ULCs) do not. "The nation as a whole," William Baumol and Alan Blinder [1998, 396] conclude in their popular introductory text, "need not fear competition from cheap labor . . . [L]abor will be cheap only where it is not very productive." This sentiment is widely echoed [Hufbauer and Schott 1993, 171-172; Krugman and Obstfeld 1997, 25-26; Lovely 1994; Golub 1995a, 1995b].
Here we question this commonplace conclusion. We argue that wage differentials between countries can have an independent influence on flows of foreign direct investment (FDI). Even when two countries have the same average reported unit labor cost in manufacturing, there are still reasons why a positive net annual flow of FDI from the high-wage to the low-wage country will likely occur.
In a controversy often muddied by vague claims, we begin with some clarification. The claim that production relocation is likely from high-wage to low-wage countries is often called the "sweatshop labor argument." This argument is often caricatured by its critics, who reduce it to the assertion that wages are the only determinant of FDI. For example, Stephen Golub [1995a, 1] comments that if the sweatshop labor argument were correct, "countries with rock-bottom labor costs, such as Bangladesh and Botswana, would rule world trade." Former U.S. Trade Representative Carla Hills used the same argument against NAFTA opponents [Hufbauer and Schott 1993, 12]. We agree that wages are not the only determinant of FDI flows; but it simply does not follow that concerns about low wages are unfounded. Our claim is that the host country average wage has some independent influence on FDI flows to that country. Before we offer support for this claim and reply in detail to the critics, however, we state their case more fully.
The Critique of the "Sweatshop Labor Argument"
The critique of the "sweatshop labor argument," based in Ricardian trade theory, holds that a country's wage level tends to align itself with its level of labor productivity. This may occur through adjustment in wages, exchange rates, or product mix.
To illustrate this, Golub [1995b] presents measures of average labor compensation (wages plus benefits), labor productivity, and ULCs in manufacturing for six newly industrializing countries, or NICs (India, South Korea, Malaysia, Mexico, the Philippines, and Thailand), for 1970-1992. For each of these countries, both wages and labor productivity were much lower than in the United States. Further, the ranking of these countries by productivity matched their ranking by labor compensation. Hence each country's ULC differential with the United States was dramatically smaller than its wage differential. In fact, for four of them (India, Malaysia, Mexico, and the Philippines), average ULCs rose above those of the United States for a few years. Golub [1995a, 3] concludes that "low wages are a symptom of low productivity, not an independent source of international competitiveness . . . fears about unfair competition of low wage NICs are greatly exaggerated and should not stand in the way of free international trade, which benefits both the US and developing countries."
But this familiar conclusion does not follow from Golub's analysis. Golub's findings undermine only a crude caricature of the sweatshop labor argument - that wage differentials provide an unambiguous measure of ULC differentials and thus inevitably motivate capital flight. We acknowledge that national average wages bear some relationship to national average labor productivity, but this does not constitute proof that wage differentials have no independent effect on FDI. Our disagreement with critiques like Golub's rests on one minor and two major points. First the minor point. Golub offers no theory or evidence about how sensitive FDI flows are to ULC differences (or to anything else, for that matter), and so he provides no basis for assessing whether or not a particular ULC differential is "small." Moreover, in some of the data he examines, ULC differentials are actually rather large. From 1982 to 1992, for instance, his data show that the average manufacturing ULC in Mexico (a country that in 1992 was host to a larger stock of U.S. FDI than the other five countries combined [OECD 1996]) was about half that in the United States, surely providing an inducement for FDI flows.
Second and more important, an individual multinational enterprise (MNE) will choose a production site based on the costs it will incur there, including the ULC it is likely to achieve - and there is strong reason to think its ULC will be less than the host country's average ULC. This implies that FDI may well flow between a high-wage and low-wage country with the same average ULCs. We pursue this point shortly.
Third, multinational firms that produce in low-wage countries for export to their parent firms often use transfer pricing in intra-firm transactions in part in order to minimize tax liabilities. As a result, their reported valued added per worker in many low-wage countries is likely to be less than their true value added per worker. Thus, the true productivity gap for these countries is likely smaller, and the true ULC gap therefore larger, than official figures show, creating an incentive for FDI flows.
MNEs' Firm-Specific Advantages
Trade models typically achieve sweeping generality at the cost of drastically simplifying assumptions. Among these is often the assumption that all producers use the same production function, and so there are no gaps in knowledge or in access to technology. Richard Nelson  pointed out the inconsistency of this assumption with available cross-country data and offered an initial analysis of the process of international technological diffusion. Today a growing literature recognizes that technically advanced countries and firms have persistent advantages and that this has implications for trade theory [see, e.g., Dosi, Pavitt, and Soete 1990]. In fact, economic researchers increasingly recognize heterogeneity among firms with respect to labor productivity and ULCs even within the same industry and within the same country. This heterogeneity is due to different skill levels and organizational routines that embody different tacit and explicit knowledge. Technological diffusion is costly and happens at varying rates [Teece 1976].
In addition, the theory of MNEs developed by Stephen Hymer , John Dunning , James Markusen [e.g., 1984], and others asserts heterogeneity specifically between MNE affiliates and host country domestic firms. An MNE establishes operations in a host country only when it brings with it some firm-specific advantage that more than offsets the advantages that host country firms have in their knowledge of local language, culture, marketing networks, and so forth. The MNE's advantage is typically something it owns: a patented or copyrighted process or product, a trademark it has widely advertised, its technical expertise, or a superior set of management principles. These ownership advantages are difficult to copy either because it is illegal to do so or because of their tacit nature. Some of these firm-specific advantages allow the firms to enjoy higher labor productivity than domestic firms. Of course, along with ownership advantages, MNE labor productivity is jointly influenced by location-specific factors, including infrastructure and the supply of skilled technical and managerial personnel. Our contention here, in short, is that the determinants of a MNE affiliate's productivity reside partly in the firm itself and so its ULC is likely to differ from the host country average.
MNE Productivity and Wages
Rigorous statistical studies of MNE affiliates in developing countries have generally found that labor productivity is higher than in host country firms in the same sector [Blomstrom 1988, 1989; Fairchild and Sosin 1986, 699-700; Willmore 1986; Harrison 1996]. Interview evidence corroborates this; for example, Harley Shaiken  found that five U.S.-owned plants he studied in Mexico making auto engines, computers, and electronic products were achieving "comparable or higher productivity and quality than similar U.S. plants operated by the same parent company," as reported by executives of these plants [Shaiken 1990, x]. A number of studies have found that these differences are explained by the larger average size, higher capital intensity, or higher skill intensity of MINE affiliates relative to host country firms [e.g., Willmore 1986; Caves 1996]. For our purposes, however, it is irrelevant that the MNE productivity advantage appears to flow from size, capital intensity, or skill intensity, rather than from higher technical efficiency in operating the same production technique. If domestic firms could achieve the same productivity and profitability advantage by operating larger, more capital- and skill-intensive firms, surely they would do so. So the MNE advantage is expressed through scale and factor proportions, but is nevertheless a distinct advantage.
Moreover, the only econometric studies using firm-level data, summarized by Ann Harrison [1996, 164], not only found that Moroccan and Venezuelan firms with foreign equity participation had "significantly higher levels of productivity than their domestic counterparts," but also that "the presence of foreign firms has no impact or a strong negative impact on the productivity of domestic firms." If this finding is representative of developing countries, then MNEs' productivity advantage over domestic firms is like to be enduring.
With higher labor productivity than the host country average, individual MNEs' unit labor costs must be lower than the host country ULC if the wage premium they pay is less than their productivity advantage. But is this the case? While many studies have found that MNE affiliates pay higher wages than domestic firms, the differential shrinks once the different industry composition of foreign and domestic firms is accounted for [Caves 1996; Willmore 1986; Aitken, Harrison, and Lipsey 1996]. Brian Aitken et al.  found that for firms in Mexico and Venezuela, geographic location within the host country also accounts for a substantial part of the wage differential; larger firm size and a higher proportion of skilled workers employed than in domestic firms are also contributing factors [Caves 1996, 228]. These studies find wage premia paid by MNEs ranging from zero to about 30 percent before controlling for any of the causal factors just listed. The studies cited above, together with the arguments and evidence presented in our next section on transfer pricing, appear to indicate that MNEs' productivity advantage is larger than the wage premium they pay.
A well-specified, careful econometric study of the determinants of FDI to developing countries should capture the independent influence of the host country wage. Such a study would also include ULC or productivity as a regressor. Existing studies reach no uniform conclusion: with FDI as the dependent variable, significant coefficients are sometimes found on the average host country wage [Schneider and Frey 1985], average manufacturing productivity, or ULC, but sometimes not; in some cases the wage even has a positive sign [Caves 1996, 218]. These inconsistent results seem due partly to the omission in some regressions of significant determinants of FDI such as infrastructure, market size, political and exchange rate risk, and partly to samples that include or emphasize developed countries. In addition, we think that separating out export-oriented FDI - the category we would expect to respond most strongly to ULC and wages [Caves 1996, 218] - would more likely produce a significant negative coefficient on the host country wage. David Wheeler and Ashoka Mody , in a sample including developing countries, found that FDI in the electronics sector was more responsive to host country average hourly manufacturing wage than was total FDI in manufacturing; this was particularly true for low-income countries. Of course, underreporting of value added due to transfer pricing (reporting value added as a small markup over cost) would also affect the outcome, since reported host country average productivity would be more collinear with its average wage than is truly the case, and ULC would be overestimated. We turn next to the issue of transfer pricing.
Transfer Pricing and Distortion of Reported Value Added
Labor productivity (or value added per worker) in MNE affiliates in developing countries is likely to be substantially underreported due to transfer pricing. Roger Y. W. Tang  observes that an MNE can often reduce its tax bill (and thus enhance after-tax profits) by shifting profits from one country to another, typically by over-invoicing imported inputs and under-invoicing exported output. Tang cites four reasons that MNEs often choose to shift profits out of a host country, based on the example of China:
1. To recover the cost of investment quickly, to avoid political and currency exchange risks;
2. To shift profits to a parent firm before sharing them with joint venture partners in the host country;
3. To inflate the costs of technology and equipment transferred to the MNE affiliate to obtain a greater share of equity investment and to increase the costs of fixed and intangible assets for tax purposes; and
4. To reduce tax liabilities by shifting profits to low tax jurisdictions [Tang 1997, 197, citing Li 1995].
In addition, under the production-sharing provisions of the U.S. tariff code - Harmonized Tariff System (HTS) 9802 - when U.S. firms export inputs to an assembly plant abroad and re-import the finished product, a tariff is levied only on value added abroad. Hence U.S.-based MNEs have an incentive to underreport value added in foreign assembly operations. A high proportion of the imports to and exports from such plants are intra-firm transactions, and so the manipulation of internal prices may lead to a substantial difference between actual and reported value added. Many other developed countries have similar provisions, so this distortion is not only an issue for U.S.-based firms.
Is the size of such underreporting large enough to worry about? The available data suggest that it is, at least for some countries. For example, in 1991, as part of this effort, the Chinese tax authorities investigated the extent of transfer pricing for 1,500 commodities and concluded that the transfer of funds out of China through false invoicing added up to $2.7 billion, or at that time 27.1 percent of the total gross value of imports plus exports [Tang 1997, 200]. This implies that the distortion as a percentage of true value added was quite likely greater than 27 percent.
Gary Hufbauer and Jeffrey Schott, in an effort to dismiss concerns of NAFTA's critics, argued that U.S. firms had little incentive to close U.S. plants and move to Mexico [1993, 171-173] because the wage differential between U.S. manufacturing workers and maquiladora workers was matched almost exactly by the differential in value added per worker. It is notable, for starters, that Hufbauer and Schott cite data for the maquiladoras, rather than for Mexican manufacturing as a whole, which shows a substantial ULC gap with the United States in the aggregate, as we have seen above [Golub 1995a, 1995b]. But more to the point, these data on the export-oriented maquiladoras understate value added. In fact, "Maquiladoras normally sell their finished products for cost plus a small profit (between 1 to 5% of total costs)" [Tang 1997, 155]. In other words, typically a subsidiary of a U.S. MNE would export its product back to its parent firm, reporting value added at just over cost; and the bulk of the cost is typically labor cost. In effect, such pricing practices would guarantee that reported MNE productivity would closely track MNE wages, while true MNE productivity would be much higher. Thus, host and source country ULCs might be reported as equal, but the MNE's true ULC could be substantially lower. FDI would then flow even where average ULCs were reportedly equal.
The textbook critique of the "sweatshop labor argument" relies implicitly on the simplistic trade theory assumption that all firms in a country have access to the same technology. But both MNE theory and abundant evidence assert the contrary: an MNE affiliate derives its labor productivity in significant measure from the firm-specific advantages it brings with it. The differential between source and host country in average manufacturing wage should therefore be an independent determinant of FDI flows.
In addition, value added per worker is likely substantially underreported by export-oriented MNEs in many low-wage countries through transfer pricing, so that the productivity gap is likely smaller and the ULC gap larger than reported. For these reasons even where a low-wage country's reported ULC is approximately equal to that of a high-wage country, we would expect significant FDI flows from the high-wage into the low-wage country. Removing any existing barriers to such flows would then increase these flows and typically result in some job loss in the source country. Thus, the "sweatshop labor" argument is better grounded in theory and evidence than its critics would have us believe.
Aitken, Brian, Ann Harrison, and Robert E. Lipsey. "Wages and Foreign Ownership: A Comparative Study of Mexico, Venezuela, and the United States." Journal of International Economics 40 (1996): 345-371.
Baumol, William J., and Alan S. Blinder. Economics: Principles and Policy. 7th ed., 1998 update. Fort Worth, Tex.: Dryden Press, 1998.
Blomstrom, Magnus. "Labor Productivity Differences Between Foreign and Domestic Firms in Mexico." World Development 16, no. 11 (November 1988): 1295-98.
-----. Foreign Investment and Spillovers. New York: Routledge, 1989.
Caves, Richard E. Multinational Enterprise and Economic Analysis. 2d ed. Cambridge: Cambridge University Press, 1996.
Dosi, Giovanni, Keith Pavitt, and Luc Soete. The Economics of Technical Change and International Trade. New York: New York University Press, 1990.
Dunning, John H. "Explaining the International Direct Investment Position of Countries: Towards a Dynamic or Developmental Approach." Weltwirtschaftliches Archiv 117 (1981): 30-64.
Fairchild, Loretta, and Kim Sosin. "Evaluating Differences in Technological Activity between Transnational and Domestic Firms in Latin America." Journal of Development Studies 22 (July 1986): 697708.
Golub, Stephen. "Productivity and Labor Costs in Newly Industrializing Countries." Pacific Basin Notes no. 95-27 (August 18, 1995). Federal Reserve Bank of San Francisco, Center for Pacific Basin Monetary and Economic Studies, 1995a.
-----. "Comparative and Absolute Advantage in the Asia-Pacific Region." Federal Reserve Bank of San Francisco, Economic Research Department, Center for Pacific Basin Monetary and Economic Studies, Pacific Basin Working Paper Series, Working Paper no. PB95-09, October 1995b.
Harrison, Ann. "Determinants and Effects of Direct Foreign Investment in Cote d'Ivoire, Morocco, and Venezuela." In Industrial Evolution in Developing Countries, edited by Mark J. Roberts and James R. Tybout. New York: Oxford University Press and World Bank, 1996.
Hufbauer, Gary Clyde, and Jeffrey J. Schott. NAFTA: An Assessment. Revised ed. Washington, D.C.: Institute for International Economics, 1993.
Hymer, Stephen. The International Operations of National Firms. Cambridge, Mass.: MIT Press, 1976.
Krugman, Paul, and Maurice Obstfeld. International Economics. 4th ed. Reading, Mass.: Addison-Wesley, 1997.
Li, Jinyan. "Transfer Pricing Law and Policy in China." Tax Notes International (October 2, 1995): 931-944.
Lovely, Mary. "Thinking Locally, Acting Globally: Congressman Jim Walsh and the NAFTA Vote." Pew Case Studies in International Affairs, Instructor Copy 160. Washington, D.C.: Georgetown University, Institute for the Study of Diplomacy Publications, School of Foreign Service, 1994.
Markusen, James R. "Multinationals, Multi-plant Economies, and the Gains from Trade." Journal of International Economics 14 (1984): 205-226. Mundell, Robert A. "International Trade and Factor Mobility." American Economic Review 47 (June 1957): 321-335.
Nelson, Richard R. "A 'Diffusion' Model of International Productivity Differences in Manufacturing Industry." American Economic Review 58 (1968): 1219-48.
OECD (Organization for Economic Cooperation and Development). International Direct Investment Yearbook, 1996. Paris: OECD, 1996.
Schneider, Friedrich, and Bruno S. Frey. "Economic and Political Determinants of Foreign Direct Investment." World Development 13, no. 2 (February 1985): 161-75.
Shaiken, Harley. Mexico in the Global Economy: High Technology and Work Organization in Export Industries. San Diego: University of California, San Diego, Center for U.S.-Mexican Studies, Monograph Series, 33, 1990.
Tang, Roger Y. W. Intrafirm Trade and Global Transfer Pricing Regulations. Westport, Conn.: Quorum Books, 1997.
Teece, David J. The Multinational Corporation and the Resource Cost of International Technology Transfer. Cambridge, Mass.: Ballinger (J. P. Lippincott), 1976.
Wheeler, David, and Ashoka Mody. "International Investment Location Decisions: The Case of U.S. Firms." Journal of International Economics 33, nos. 1-2 (August 1992): 57-76.
Willmore, Larry N. "The Comparative Performance of Foreign and Domestic Firms in Brazil." Worm Development 14, no. 4 (1986): 489-502.
Mehrene Larudee and Tim Koechlin are Assistant Professor of Economics, the University of Kansas, and Associate Professor of Economics, Skidmore College, respectively. This paper was presented at the annual meeting of the Association for Evolutionary Economics, New York City, New York, January 3-5, 1999. Helpful discussions with Steve Burks are gratefully acknowledged.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Wages, Productivity, and Foreign Direct Investment Flows. Contributors: Larudoo, Mehrene - Author, Koechlin, Tim - Author. Journal title: Journal of Economic Issues. Volume: 33. Issue: 2 Publication date: June 1999. Page number: 419+. © 1999 Association for Evolutionary Economics. COPYRIGHT 1999 Gale Group.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.