Competition Policy and FDI: A Solution in Search of a Problem?

I would like to thank Li-Gang Liu for excellent research assistance, and Dave Richardson for helpful comments on an earlier version of this paper. Forthcoming in the journal, Japan Economic Studies.

Abstract

This paper reviews a variety of documentary evidence to find evidence of the impact of private practices in discouraging inward foreign direct investment (FDI). Outside of a few countries, there is little documentary evidence that this is indeed a problem. Sectoral data on inward FDI from the United States and Japan are then analyzed econometrically. The regression results reinforce the impression gleaned from the document review: barriers to FDI are more likely to take the form of general economic conditions or specific policies facilitated by private practices than systematic aspects of industrial structure.

Introduction

It is difficult to identify a country that does not restrict inward foreign direct investment (FDI) in some way. This is usually through official prohibitions, restrictions, or official approvals processes, in which the government intervenes directly to affect the level, composition, or form of foreign investment. However, the behavior and practices of private parties, sometimes facilitated by government policies not directly aimed at foreign investment, may also affect the level, composition, and form of FDI. Often this takes the form of privately supported but publicly sanctioned barriers to entry. In some cases these may affect domestic and foreign potential entrants equally, though in other cases, foreign firms may face greater impediments. There are yet another set of circumstances in which government policies and private behavior may affect foreign firms' decisions to service the host market via exports or via local production, again affecting the level, composition, and form of FDI.

At the same time, there are many circumstances or conditions (such as large sunk costs of production or the existence of brand loyalty to name two) which may intrinsically discourage new entrants-foreign or domestic. So, for example, a Japanese government survey of foreign firms operating in Japan found that private impediments were relatively low down on the list of difficulties respondents cited in operating in Japan, reinforcing the findings of an earlier survey of American firms conducted for the American Chamber of Commerce in Japan.¹

Obviously there is a need then to narrow the discussion of from a broad definition of impediments which could encompass general economic conditions (such as high land prices) or nearly any practice that differs from home country practices and could thus be said to discourage FDI. This paper focuses on two principal channels through which private practices, government competition policies, and FDI may interact.

In product markets, restrictive business practices can impede FDI associated with the distribution, serving, development and production of goods. In service markets barriers to entry can discourage investment which is essential to service local markets. Firms interacting horizontally (that is with their notional competitors or rivals) can behave in such ways to affect potential entrants' investment decisions. Horizontal agreements that could affect FDI include price fixing, cartels or market allocation schemes, bid rigging, and refusals to deal and other abuses of incumbent position. Vertical restraints on trade involving buyers and sellers of intermediate input markets and the organization of distribution of final products can also affect FDI. Conventional vertical restraints include such practices as refusals to deal and boycotts, retail price maintenance, exclusive-dealing arrangements, and tie-ins. It should be recognized from the outset that these practices do not necessarily discriminate exclusively against foreign firms (indeed in some cases incumbent foreign firms participate in these activities). Rather, these practices, which can in principle discriminate against domestic new entrants, will be examined from the perspective of their impact on foreign firms that through this on FDI.

Factor markets may also provide a channel through which incumbent firms may deter entry. Capital market imperfections provide a second channel through which private arrangements in host countries can affect FDI. Specifically incumbent firms can exploit capital market imperfections and impede the development of a market for corporate control. This can frustrate merger and acquisition activity which has been the primary mechanism for FDI in developed countries, thereby affecting FDI throughout the economy.² In the labor market, lifetime employment practices can discourage FDI by impeding firms' access to potential staff.

The issue of government economic policy is salient for two reasons. First, government policy (apart from competition policy per se) can affect the ability of private firms to engage in anticompetitive private behavior and impede FDI. So, for example, cartels are unlikely to be able to raise prices and exclude new entrants to markets unless there is some mechanism (such as government licensing) which impedes the ability of new firms to enter the market and bid down prices. Economic policy in a broad sense is critical in this regard.

More narrowly, if the incumbent firms are able to use the regulatory environment to deter entry, their ability to exploit their positions may be constrained by domestic competition policies. There is no single international standard, however, and what constitutes unacceptable or illegal behavior varies enormously across countries.³ Some countries quite literally have no competition policy rules. To cite an example, according to the WTO, Uganda has no competition law or other legislation banning or sanctioning collusive or restrictive business practices, such as price cartels, price maintenance, bid-rigging, and boycotts-precisely the practices discussed in this paper.⁴ Private monopolies are not subject to restrictions or controls.⁵ Other countries, such as the Dominican Republic and Cameroon, have laws oriented towards consumer safety or protection, not toward the industrial organization issues under discussion here.

Even where laws do exist, there may be significant differences in content, sectoral scope, and entities covered.⁶ Many, though not all, countries for example, prohibit horizontal cartels, but not vertical exclusive dealerships.⁷ These differences even extend to core competition policy rules in major industrialized countries.⁸

These cross-national differences in legal standards are compounded by differences in enforcement procedures and the severity of penalties for illegal behavior. In the case of Japan, for instance, alleged lax enforcement of competition policies have been a source of trade and investment friction with both the United States and the European Union (Bergsten and Noland 1993; MITI 1996, 372-73). Likewise, Swiss authorities have exercised "considerable restraint" in the enforcement of competition policy (WTO 1991, 107), contributing to trade and investment friction (USTR 1996, 288), (though again, recent legislative change holds forth the prospect of more vigorous enforcement).

The approach of this paper is to review a variety of documentary evidence to find evidence of the impact of private practices in discouraging inward FDI.⁹ To preview the conclusion, outside of a few countries, there is little documentary evidence that this is indeed a problem.

Sectoral data on inward FDI from the United States and Japan are then analyzed econometrically. The regression results reinforce the impression gleaned from the document review: barriers to FDI are more likely to take the form of general economic conditions or specific policies facilitated by private practices than systematic aspects of industrial structure.

Horizontal Agreements

Horizontal agreements among competitors, including price fixing, cartels or market allocation schemes, and bid rigging are all mechanisms for exploiting market power.¹⁰ Their impact on foreign direct investment is less obvious, however. If prices are artificially raised, then new entrants (foreign or domestic) will enter the market attracted by supernormal profits, undermining the arrangement. In the obverse case of predation, prices may be lowered to drive competitors out of the market. However, when the incumbent firm raises prices, new entrants will re-enter the market. Similarly, successful bid rigging requires some mechanism for restraining entry to sustain the arrangement, and without such a device, cartels may not be sustainable, even if legal.¹¹

It is not surprising then, that successful horizontal arrangements with significant implications for FDI are relatively rare.¹² Probably the best documented example is the existence of bid rigging in Japanese government procurement, especially in construction.¹³ (In the case of service industries like construction or civil aviation, restrictions on entry are tantamount to restrictions on investment since the local market cannot be serviced without a local presence.) Restrictions on entry into the construction business (especially for foreign firms) have contributed to cartelization, and the associated rents are allocated through the practice of dango, a form of bid-rigging in which firms negotiate with each other as to which firms will participate in bidding on a given project and at what prices. McMillan (1991) estimates that excess profits from collusion in public works projects typically amount to 16 percent to 33 percent of the price. Evidence from the few cases that have been prosecuted supports this range of estimates, the most famous case being the 1989 episode in which the US Department of Justice reached an out-of-court settlement with 99 Japanese construction firms for bid-rigging at the Yokosuka Naval Base; fines levied in that case amounted to $32.4 million, or 24 percent of the billed costs.¹⁴ Ultimately, it is the licensing restrictions to entry which facilitate these practices, and these regulations, which impede entry, especially the entry of foreign firms, have become the object of comment by the United States, the European Union, Canada, and Australia (MITI 1996, 365-66 and 372).

Another type of market access barrier can arise if industry associations are given regulatory power or status, in the establishment of product standards or testing and certification procedures. (Incumbents can use their quasi-regulatory authority to impede new entrants.) Examples, involving both international trade and foreign investment in the service sector, have been cited in Japan (Balassa and Noland 1988), and the US government has requested that foreign firms' access to industry groups be increased (MITI 1996, 371). In the case of Korea, in 1994 Korean authorities investigated 68 industry associations and found that 48 of the associations had been engaging in anticompetitive or unfair practices. the following year Korean authorities investigated 218 industry associations, and ordered them to revise 369 anticompetitive or unfair measures or practices. In 1996 the Korean government announced that it would begin regulating not only collusion among rival firms, but trade associations as well (USTR 1997, 249). Such anticompetitive practices may discourage trade and or investment, depending on the specific circumstances.

The issue of an even more subtle impediment to investment has been identified by the EU in connection to airline travel in the US (EU 1996, 41). American airline companies control electronic reservations systems, thereby giving preference to same line connections over interline connections. This puts foreign carriers without extensive US domestic networks at a disadvantage.¹⁵ One solution in this case would be for a non-American carrier to buy a US airline for its domestic routes, but US law imposes prohibitions on foreign ownership of airlines. Another possibility would be "code-sharing," but this kind of arrangement may put foreign carriers at a commercial disadvantage to US domestic carriers, by forcing them to enter into agreements with incumbent US carriers, encumbering the ability of foreign carriers to implement their preferred commercial strategies. Similarly, the EU alleges that establishment of integrated radio telecommunications systems in the US is impeded by the greater obstacles getting relevant licenses (EU 1996, 46). In both cases, government intervention acts as a facilitating device to impede foreign access.

Vertical Restraints

Typical vertical restraints include resale price maintenance, exclusive dealing or distribution, tied sales, reciprocity agreements, territorial restrictions on dealers or distributors, refusals to deal, or vertical mergers. Starting with Bork (1966, 1978), one line of argument held that the traditional competition policy approach of discouraging these practices was misguided in that these vertical arrangements were efficiency-enhancing, in the sense of reducing transactions and search costs, removing downstream (retail) price distortions, and encouraging optimal level of investment in production and distribution. However, subsequent research pointed to potential divergences between private and social welfare associated with these practices. Salop and Scheffman (1983, 1987) explored the possibility that dominant firms might be able to induce rivals to exit or deter their entry or simply put rivals at a disadvantage by raising their costs through the application of these techniques. Hart and Tirole (1990) subsequently identified situations in which vertical integration has the effect of foreclosing the market and is the optimal strategy for the firms involved.¹⁶ From a policy perspective these opposing effects present a conundrum. Given the apparent difficulty of vertical foreclosure and the possible efficiency enhancements associated with vertical integration, regulatory authorities typically have taken a less decisive stance with regard to vertical agreements in comparison to the horizontal agreements previously discussed.¹⁷

The prime example of alleged vertical restraints affecting international commerce are the kieretsu or affiliated firm groups in Japan.¹⁸ These networks of affiliated firms typically have long-standing financial, managerial, and product market interlinkages. A kieretsu might consist of a group of large core firms (including financial firms) linked across markets, together with their vertically linked input suppliers, and possibly a captive distribution network. kieretsu are a multifaceted phenomenon but two aspects of this form of industrial organization can be distinguished in this context: so-called vertical kieretsu involving vertical supply relationships in product markets, and distribution kieretsu involving distribution networks in intermediate or final products.¹⁹

According to the WTO (1995), vertical supplier kieretsu (seisan-keiretsu) "may create an entry barrier of outsiders. Firstly, a kieretsu relationship may induce apparent anti-competitive practices such as vertical boycotting and exclusive dealing . . . long-term relationships cultivated through a kieretsu may push up 'switching costs' and make it more costly for kieretsu participants to change their business partners" (p. 92).²⁰ Repeat transactions relationships (such as kieretsu ) may work to the disadvantage of new entrants, foreign or domestic.²¹ This may be of particular relevance to firms from LDCs which are almost by definition new entrants to these markets.

A second potential impediment to inward FDI is vertical restraints within the distribution system.²² This occurs both in consumer and in capital goods. Vertically integrated firms refuse to carry the products of competitors, and product return and rebate systems are used to tilt retailer incentives toward domestically produced products. According to the WTO, vertical distribution affiliation kieretsu (ryutsu-keiretsu) "usually set up by manufacturers, tie wholesalers and retailer together through various links including special agent contracts, extension of technical and financial assistance, exchange of personnel and cross-shareholding. Some Japanese business practices, such as the suggested retail and wholesale price system (tatene), a complex rebate system and provisions for return of unsold goods are important aspects of such relationships. Distribution kieretsu are not found equally in all sectors; they are concentrated in specific sectors, such as cosmetics, electrical appliances, and automobiles, all of which are characterized by product differentiation, specialty goods, or occasional purchases, and important market segmentation. Outside the automobile sector where the affiliation ratio is exceptionally high, kieretsu distributors cover at most about half of the total market. (Fifty percent in the cosmetic sector and 40 percent in the electrical appliance sector.) While manufacturers cannot legally prevent affiliated dealers from dealing in competing products, such deals are likely to be discouraged by additional burdens on after-sale servicing and sales-promoting activities, as well as by other advantages extended or pressures exerted by kieretsu manufacturers" (pp. 92-93).²³ This is confirmed by the sophisticated econometric research of Ariga et al. (1991) which points to administered prices in sectors where there are strong vertical relationships or kieretsu, suggesting that control of the distribution system acts as an effective barrier to entry.

Japan is not alone in confronting these issues which have arisen in other countries as well. According to the WTO, "wholesale traders appear to have played a pivotal role in organizing and implementing many cartels" (WTO 1991, 113), abetted by the legality of exclusive supply contracts in the Swiss context (WTO 1991, 114). Competition in distribution is further impeded by government regulations that effectively discourage competition at the retail level, and have contributed to concentration in retailing (Ibid.). Indeed, auto parts appears to be one sector where exclusive supply networks, warranty requirements that "genuine parts" be used in repairs, and homologation procedures have discouraged new entry into the market (WTO 1996, 106).

Competition authorities have faced similar issues in the UK. According to Hay (1997), in the case of auto parts, in 1981-82 when allegations similar to the Japanese and Swiss cases were made regarding the UK auto parts sector, the Monopolies and Mergers Commission (MMC) went after auto parts suppliers who required that distributors only stock replacement parts supplied by them. In 1992 the MMC went after similar manufacturer-dealer arrangements. With respect to vertical restraints on the distribution of final products, the MMC recommended that leading brewers be required to divest themselves of a large number of their outlets on the grounds that control of the outlets represented a barrier to entry to brewing and to the retail market for other suppliers. It has issued similar judgements in other cases, and Hay concludes that there "is no equivalent of the vertical kieretsu " or informal vertical obstacles to foreign direct investment in the UK (Hay 1997, 17).

These restrictive practices may have an impact on both goods trade and investment, as well as both potential foreign and domestic entrants. The relevance for FDI is that for many products (especially sophisticated manufactures) a local distribution and service network is essential for sales. Arms-length exporting, without a local presence, is not really an option for capital goods or high end consumer durables such as autos. Rather, exports, local sales, distribution, and service, and local product development and production are all of a single cloth. The inability of new entrants to penetrate a exclusive distribution system (or establish a parallel system) may effectively foreclose entry into the market. At the same time, as survey data from Japan has shown, such obstacles may not be the most important impediments that potential foreign investors face.

Capital Market Imperfections

Capital market imperfections may inhibit the development of a well-functioning market for corporate control and thereby provide another possible mechanism for private agreements to affect the level of inward FDI. Specifically, informal obstacles to merger and acquisition activity including cross-holding and tactical impediments to acquisition may impede the primary channel of FDI in developed country markets.

To begin with, stock markets (which can be taken as a crude proxy for the market for corporate control and hence the opportunity for FDI through merger and acquisition activity) vary considerably in size and depth across countries. Table 1 reports some recent data on stock market capitalization as a share of GDP and the number of listed shares for selected countries. As can be seen in table 1, among the major developed countries there appear to be two groups: continental European countries where stock markets are relatively small, and the other developed countries where they are larger. Among the developing country emerging markets, stock market capitalizations in Taiwan, South Africa, Malaysia, Korea, and Brazil exceed that of Italy in both relative and absolute size, and have more listed firms than the bourses of either Italy or France.

Market capitalization alone, while providing some indication of the extent of the potential for merger and acquisition activity and therefore FDI through this channel, can be misleading. Due to cross-holdings, for example, in Japan only 30 to 40 percent of shares are actually traded, effectively precluding a market for corporate control.²⁴ Similar tales emerge in continental Europe. According to Rao and Ahmad (1996) in Germany banks and insurance companies together own a sizable share of non-financial firms. Moreover, concentrated patterns of ownership are reinforced by laws which in effect require acquisition of a 75 percent share of voting stock to assure control, creating a major obstacle to corporate takeovers. The WTO concluded that in Germany "access barriers resulting from intensive cross-shareholdings between domestic banks and industrial conglomerates" could be an impediment to FDI (WTO 1993, 57). In France over half of the 200 largest companies are family-controlled, and family control of the corporate sector is if anything tighter in Italy (Rao and Ahmad 1996).

This relative dearth of merger and acquisition opportunities is exacerbated by tactical barriers to takeover. Rao and Ahmad (1996) report that German and Canadian law provides for liberal ability to issue non-voting shares. (Certain French companies can do this as well.) In Germany shareholder often deposit shares with a bank which then exercises their proxy vote, and German banks have been able to control as much as 98 percent of votes at some annual shareholder meetings. Also in Germany and France, restrictions on the voting rights of individual shareholders, and schemes to weight votes by the duration of time holding shares have been used to frustrate takeover bids.

It has been argued that stable shareholding or large institutional shareholders provide better monitoring of management than more diffuse patterns of ownership, contributing to superior economic performance. In the case of Japan, Nakatani (1984) found that bank-affiliated firms had lower, though more stable, profits. He interprets bank or kieretsu affiliation as acting as an insurance scheme. This conclusion is reinforced by recent research by Kawai, Hashimoto, and Izumida (1996), which finds that distressed firms with main bank connections pay a significantly lower interest rate premia than do firms without main banks. So from this perspective, bank or kieretsu affiliation could be seen as a stabilizing force on the economy. However, Lichtenberg and Pushner (1992) found that high degree of intercorporate stock ownership in Japan was associated with lower productivity, profitability, and growth. They interpreted this as indicating managerial shirking facilitated by protection from takeover. Similarly, in the case of Germany, Chirinko and Elston (1996) found that bank finance does not lower the cost of capital or result in higher profitability. Weinstein and Yafeh (1995) argue that the corporate finance incentives faced by kieretsu member firms encourages overexpansion, and this is the explanation for their apparently low profitability, and the difficulties new entrants face in breaking into markets dominated by kieretsu. This provides the link between these alternative modes of finance and FDI, though the quantitative importance of this is unclear.

Quantitative Analysis

The analysis thus far has pointed to mechanisms by which private practices may affect foreign direct investment, and provided some examples in which this has occurred. This section extends the previous analysis with some simple statistical evidence on industrial structure and inward foreign direct investment.

Data on inward foreign direct investment (as a share of industry sales) was assembled for the US and Japan, the world's largest economies.²⁵ Herfindahl indexes of concentration were then regressed against the FDI data and this is reported in table 2. As can be seen in regression (2.1), the degree of industry concentration is negatively correlated with inward FDI, suggesting that oligopolists may indeed be able to impede foreign entry using the techniques described in the preceding section. In the next three specifications (2.2)-(2.4), kieretsu variables were added to the regression. (Because kieretsu is a vague concept, different definitions from three alternative sources were used as a check on robustness.)²⁶ The regression results are not wholly consistent. Using the data derived from Dodwell Marketing Consultants, horizontal kieretsu (keiretsu affiliations across industries) are negatively associated with inward FDI, while the result for vertical kieretsu (keiretsu affiliations vertically within industries) is statistically insignificant. For the variables derived from Toyo Keizai and Keizai-Chousa-Kyoukai, horizontal kieretsu are positively associated with inward FDI, while the vertical kieretsu variables have a negative sign. In the final three specifications (2.5) to (2.7), research and development expenditures are added to the specifications. In all three regressions, the coefficient on the research and development variable is positive and statistically significant. At the same time, the positive coefficient on the horizontal kieretsu variables in (2.6) and (2.7) become insignificant.

In specifications 2.8 to 2.13, additional variables are added to the regression, beginning with industry shipments. If anything, it appears that FDI is negatively correlated with industry size, though this result is not robust. Likewise when the labor, capital, and human capital shares in value-added are included they are almost always insignificant. The addition of these variables also tends to reduce the statistical significance of the coefficients on the other variables due to multicollinearity, with the exception of the coefficients on research and development which are relatively robust to specification.

Taken together, it appears that research and development expenditures are positively associated with inward FDI in the US and Japan. Industry concentration is negatively, though not particularly robustly, associated with FDI.²⁷ The role of the kieretsu in Japan is less clear, though the evidence in specifications (2.5)-(2.7) suggest that both horizontal and vertical kieretsu may be negatively associated with inward FDI. This would be consistent with the result of Weinstein (1996) who obtained weak evidence that the presence of kieretsu was negatively correlated with inward FDI in a panel data set. There is only one problem with these regressions: they are probably misspecified. The regressions reported in table 2 impose the restriction that the intercept terms of the Japanese and US subsamples be equal. If this restriction is violated, then the coefficients reported in table 2 will be biased and inconsistent.

Table 3 reports results obtained using the simplest test and correction for this problem, that is a fixed-effects estimator, constructed by adding a Japan dummy variable to the regression. As can be seen in table 3, the null hypothesis that the subsample intercepts are equal is rejected in most cases (i.e., the Japan dummy is significantly different from zero).²⁸ If one combines the regressions from table 2 where the null is not rejected in table 3 (2.2, 2.5, 2.8, 2.11, 2.12, and 2.13) and the table 3 results for the remaining specifications a very different set of inferences emerge: the Herfindahl index of concentration, the existence of vertical kieretsu, industry size, physical capital intensity, and human capital intensity are never statistically significant. The variable measuring the prevalence of horizontal kieretsu is significant (and negative) when measured one way, but not when measured another. Industry research and development intensity and labor intensity are occasionally statistically significant, always with a positive sign.

The bottom line from these regressions is that Japan has lower levels of inward FDI than the US, but that industrial organization variables have almost no explanatory power in this regard. Rather, an uninformative Japan dummy variable explains much of sample variance.

Conclusions

Incumbent firms often attempt to exclude new entrants by denying them effective access to goods or factor markets. This can be done by firms acting in concert horizontally across an industry (as in the case of a cartel) or vertically (such as in the case of a network of exclusive dealerships). The ability to actually foreclose entry is typically facilitated by some kind of government regulation (licensing requirements, restrictions on international trade etc.) which impedes the entry of destabilizing new entrants. Sometimes this may occur through incumbent firms' exploitation of capital market imperfections which impede merger and acquisition, the dominant mechanism for FDI in developed country markets. The forms of these restrictions may create greater impediments for potential foreign entrants than for domestic entrants, but this is not necessarily the case. As government policy often provides the facilitating devices that sustain anticompetitive behavior, government competition policies and their enforcement can constrain incumbents' abilities to implement anticompetitive strategies.

These practices can affect both goods trade and investment. FDI, in particular, can be affected in three ways. First, product market impediments may deter complementary investment in distribution, service, product development, and production. Second, service industries intrinsically require a local presence, and impediments in service markets retard this associated investment. Third, impediments to merger and acquisition activity in capital markets can discourage FDI in all sectors.

This paper has surveyed these issues with respect to the current global economic environment. In general, informal barriers do not appear to be a major impediment to FDI. Although as formal restrictions are liberalized there may be some attempt by incumbents to use anticompetitive practices to impede new entry, the global trend (in both developed and developing countries) is towards stricter competition rules and more vigorous enforcement. Perhaps even more important that competition policies per se may be the general trend toward deregulation and liberalization in both goods and factor markets with the consequent elimination of anticompetitive practice facilitating devices. This can be expected to have a salutary effect on both goods trade and investment (which are complements) as well as entry by both domestic and foreign firms.

The econometric evidence would appear to support this skepticism. Little evidence from the US and Japan can be derived to support the notion that industrial organization or private practices impede FDI. Rather than an outcome of industrial structure, it is more likely that Japan's low levels of inward FDI are due to the general economic conditions identified in the investor survey data, and anticompetitive practices facilitated by government policies particular industries identified in the literature review.

References

Ariga, Kenn, Yasushi Ohkusa, and Hisashi Namikawa. 1991. The Japanese Distribution System. Ricerche Economiche 45, no. 2-3: 167-84.

Balassa, Bela, and Marcus Noland. 1988. Japan in the World Economy. Washington: Institute for International Economics.

Bork, Robert H. 1966. The Rule of Reason and the Per Se Concept. Yale Law Journal 75, no. 3: 373-475.

Bork, Robert H. 1978. The Antitrust Paradox: A Policy at War with Itself. New York: Basic Books.

Bergsten, C. Fred, and Marcus Noland. 1993. Reconcilable Differences? United States-Japan Economic Conflict. Washington: Institute for International Economics.

Cairns, Robert D., and John W. Galbraith. 1990. Artificial Compatibility, Barriers to Entry, and Frequent-Flyer Programs. Canadian Journal of Economics 23, no. 4: 807-16.

Caves, Richard E. 1996. Multinational Enterprise and Economic Analysis, 2d edition. Cambridge: Cambridge University Press.

Chirinko, Robert S., and Julie Ann Elston. 1996. Finance, Control, and Profitability: An Evaluation of German Bank Influence. Paper presented at the Johns Hopkins University Institute of Contemporary German Studies, 10 October.

Department of Foreign Affairs and International Trade. 1996. Register of United States Barriers to Trade. Ottawa: Department of Foreign Affairs and International Trade.

Dobson, Paul W., and Michael Waterson. 1996. Vertical Restraints and Competition Policy. Office of Fair Trading Research Paper 12. East Molesey (December).

Dodwell Marketing Consultants. 1997. Industrial Groupings in Japan. Tokyo: Dodwell.

EU. 1996. Report on United States Barriers to Trade and Investment. Brussels: European Commission.

Flath, David. 1989. Vertical Restraints in Japan. Japan in the World Economy 1, no. 2: 187-203.

Graham, Edward M., and J. David Richardson, eds. 1997. Global Competition Policy. Washington: Institute for International Economics.

Hart, Oliver, and Jean Tirole. 1990. Vertical Integration and Market Foreclosure. Brookings Paper on Economic Activity: 205-76.

Hay, Donald. 1997. United Kingdom. In Global Competition Policy, ed. by Edward M. Graham and J. David Richardson. Washington: Institute for International Economics.

Kawai, Masahiro, Juro Hashimoto, and Shigemi Izumida. 1996. Japanese Firms in Financial Distress and Main Banks. Japan and the World Economy 8: 175-94.

Lichtenberg, F., and G. Pushner. 1992. Ownership Structure and Corporate Performance in Japan. Working Paper No. 4092. Cambridge: National Bureau of Economic Research.

Matsushita, Mitsuo. 1997. The Antimonopoly Law of Japan. In Global Competition Policy, ed. by Edward M. Graham and J. David Richardson. Washington: Institute for International Economics.

McMillan, John. 1991. Dango: Japan's Price Fixing Conspiracies. Economics and Politics 3, no. 3: 201-18.

MITI. 1996. 1996 Report on the WTO Consistency of Trade Policies by Major Trading Partners, Tokyo: Ministry of International Trade and Industry.

Nakatani, Iwao. 1984. The Economic Role of Financial Corporate Grouping. In The Economic Analysis of the Japanese Firm, ed. by Masahiko Aoki. Amsterdam: North-Holland.

Noland, Marcus. 1997. Public Policies, Private Preferences, and the Japanese Pattern of Trade. Review of Economics and Statistics 79, no. 2: 259-66.

OECD. 1996. Antitrust and Market Access. Paris: Organisation for Economic Cooperation and Development.

Rao, Someshwar, and Ash Ahmad. 1996. Formal and Informal Investment Barriers in G-7 Countries. In Investment Rules for the Global Economy, ed. by Pierre Sauve and Daniel Schwanen. Ottawa: CD Howe Institute.

Salop, Steven C., and David T. Scheffman. 1983. Raising Rivals' Costs. American Economic Review 73, no. 2: 267-71.

Salop, Steven C., and David T. Scheffman. 1987. Cost Raising Strategies. Journal of Industrial Economics 36: 19-34.

Sheard, Paul. 1997. kieretsu, Competition, and Market Access. In Global Competition Policy, ed. by Edward M. Graham and J. David Richardson. Washington: Institute for International Economics.

Suzumura, Kotaro. 1997. Formal and Informal Measures for Controlling Competition in Japan. In Global Competition Policy, ed. by Edward M. Graham and J. David Richardson. Washington: Institute for International Economics.

USTR. 1996. National Trade Estimate Report on Foreign Trade Barriers. Washington: GPO.

USTR. 1997. National Trade Estimate Report on Foreign Trade Barriers. Washington: GPO.

Warner, Mark. 1996. Private and Public Impendiments to Market Presence: Exploring the Investment/Competition Nexus. In Investment Rules for the Global Economy, ed. by Pierre Sauve and Daniel Schwanen. Ottawa: CD Howe Institute.

Weinstein, David. 1996. Structural Impediments to Investment in Japan. In Foreign Direct Investment in Japan, ed. by Masaru Yoshitomi and Edward M. Graham. Cheltenham: Edward Elgar.

Weinstein, David, and Yishay Yafeh. 1995. Japan's Corporate Groups: Collusive or Competitive? Journal of Industrial Economics, 18.

WTO. 1996. Trade Policy Review (various issues). Geneva: World Trade Organisation.

Notes

1. A survey of 2,656 foreign firms conducted by MITI suggested that the five major difficulties encountered by such companies in Japan were: competition of Japanese firms; difficulty recruiting skilled labor; consumer preference (for Japanese goods); differences in business practices; and high corporate taxes. The consulting firm AT Kearney conducted a similar survey of American firms in Japan at the behest of the American Chamber of Commerce in Japan. The impediments to investment as ranked by the respondents were the high cost of doing business; difficulties in locating and hiring qualified personnel; general complexities; multitiered distribution system-including anticompetitive practices, distribution kieretsu ; interlocking business ownership-keiretsu; and lack of transparency of ministry guidelines. In both surveys, private practices, while identified as obstacles to investment, were not at the top of the list.

2. According to Rao and Ahmad (1996) acquisitions accounted for almost 85 percent of foreign investment outlays in US in the 1980s and similarly almost 60 percent of EU cross-border investment in the late 1980s and early 1990s was through acquisitions.

3. See OECD (1996) and Graham and Richardson (1997) for excellent overviews of industrial country competition policies.

4. For the sake of brevity, references to World Trade Organization Trade Policy Reviews of individual countries are simply cited as "WTO" and should be understood to refer to the 1996 report unless otherwise noted.

5. According to the WTO, the Dominican Republic, Morocco, Indonesia, and Zimbabwe do not currently have competition laws, though laws are in draft or under discussion.

6. See OECD (1996) for elaboration of these issues.

7. There is also a separate issue of the situation of a small country whose local producers are de facto monopolists. For example, the Israeli Ministry of Industry and Trade considers that a monopoly exists when a single enterprise supplies more than 50 percent of a given product or service's output in the domestic market. Because Israel is a small country this is quite common, and over the years, monopolies have been declared in almost all sectors in Israel. Monopolies are not prohibited provided they do not make abusive use of monopolistic powers. If a monopoly acts in such a way that it harms competition, a specific legal action can be initiated and penalties applied. For example, in 1991, the beverage company Tempo was warned against conditioning the sale of other beverages to major supermarkets on their purchase of Pepsi Cola. In addition, import competition may restrain monopolistic behavior.

8. Under the Japanese Antimonopoly Law (AML), for example, a private monopolization or cartel is illegal if and only if it restricts competition "contrary to the public interest" a phrase which is subject to multiple (and contradictory) interpretations (Matsushita, 1997). Similarly, cartels (and their associated restrictive practices) not are not prohibited under Swiss law (though the revision of the law coming into force in 1996 introduces new constraints).

9. Given this obvious lack of a universal standard, this discussion will proceed to examine practices that observers have indicated may affect the competitive environment and foreign direct investment with the explicit caveat that these practices may or may not be illegal in particular countries. That is to say this paper examines these issues primarily from the positive standpoint of their potential impact on FDI-not the normative standpoint of whether they are desirable or justifiable, and not the legal perspective of whether or not they are permissible under local law.

10. See Warner (1996) for a discussion of these issues.

11. For example, the WTO reports that even though there are no laws against concerted bidding to establish price floors in government procurement, Moroccan authorities report few instances of bid rigging.

12. The EU has taken action against cross-national cartels in cement, cartonboard, and steel beams, but these agreements were apparently aimed at restricting goods trade, not investment per se. Likewise, in the case of Switzerland, many sectors of the economy have historically been cartelized, though recent changes in competition laws have lead to a reduction of the number of formal cartels and their replacement by "tacit collusion or informal arrangements among oligopolies" (WTO 1991, 116), facilitated by government policies relating to standards, health and safety, and international trade, to cite some examples. However, given the relatively small size of the Swiss economy and the products under consideration (books, pharmaceuticals, metal fittings, autos, car parts, and razor blades to cite some examples), it is unclear to what extent local consumption would be served by inward FDI in these sectors were restrictions to be removed.

13. It should be noted, however, that the number of legally sanctioned cartels in Japan has fallen significantly over time (Suzumura 1997, tables 1 and 2).

14. McMillan (1991) discusses this and three additional cases: the case of soil brought in to build the Kansai International Airport, in which excess profits due to collusion may have amounted to 9 percent; the 1982 case of Matsuyama City kitchen equipment purchases, in which the bids incorporated a profit margin of 31 percent; and a 1979 river-dredging case in which excess profits due to collusion were 27 percent of the contract price.

The US Justice Department subsequently settled another case against 11 Japanese electronics firms for rigging bids at the US Air Force base at Yokota. The firms agreed to pay $36.7 million in fines, or nearly 36 percent of the value of the $103 million in contracts.

15. Similarly, Cairns and Galbraith (1990) argue that "frequent flyer programs may offer a device whereby major carriers can discourage entry (retain market share)" (p. 815). This has relevance to FDI in countries with large internal civil aviation networks inasmuch as the existence of these programs tilts consumer preference towards carriers with large domestic networks and away from foreign carriers which do not. Predictably, foreign firms with low degree of overlap in their networks will tend to form alliances to counteract this competitive advantage on the part of incumbent domestic carriers.

16. As in the case of horizontal agreements, foreclosure is difficult without some regulatory mechanism to facilitate blocking competitors' counterstrategies of parallel vertical integration or establishment of distribution networks.

17. See Dobson and Waterson (1996) for an overview of these issues.

18. The term kieretsu in Japanese is typically preceded by a modifier such as kin'yu kieretsu (financial groups), kigyo kieretsu (corporate groups), kigyo guruupu (affiliated firm groups), or ryutsu kieretsu (distribution groups), specifying the type of relationship among the affiliated firms. However, as Matsushita (1997) observes, the term is rather vague and does not necessarily involve contractual relationships between the affiliated firms. Instead, these may be de facto relationships based on repeated transactions.

19. Vertical kieretsu may involve considerable cross-ownership as well as product market linkages. Sheard (1997) notes that Toyota owns 25 percent of the stock shares of its top suppliers which in turn sell 43 percent of their output to Toyota; likewise Nissan owns 30 percent of its suppliers' stock and they sell 55 percent of their output to Nissan. (Dodwell Marketing Consultants (1997), however, reports that the vehicle assemblers are encouraging their suppliers to diversify their customer bases.) In the electronics sector, Matsushita owns 54 percent of the stock of its primary suppliers which sell 51 percent of their output to Matsushita, and Hitachi owns 50 percent of its principal suppliers' shares, and they sell 30 percent of their output to Hitachi.

20. It is unclear whether these practices should be considered illegal or not. According to the WTO, group boycotts and refusal to deal violate the AML; however, Matsushita (1997) argues that "as long as there is strong interbrand competition, including price competition, between kieretsu systems, such systems are considered lawful under the AML" (p. 79).

21. An example is provided by the auto parts sector which has been the subject of recent dispute between the US and Japan. The US alleges that "close and durable Japanese intercorporate relations make it difficult for foreign automotive parts suppliers to compete with Japanese-owned suppliers for business from Japanese motor vehicle manufacturers. This situation is compounded in the Japanese market by complex and expensive Government of Japan regulations that encourage consumers to have their vehicles repaired at automobile dealership and other certified garages that utilize only original equipment (OE) parts" (USTR 1996, 192).

22. See Flath (1989) for a description of vertical restraints in Japan.

23. In the case of autos, USTR (1996) alleges that "access to the motor vehicle market is restricted primarily through Japanese vehicle manufacturers' control over the dealer networks. The Japanese automakers maintain control over their dealers through financial ties (equity, loans, rebates, transfers of personnel to dealerships), as well as through practices generically found in the automobile industry, such as allocation of the most desirable models, and other forms of cooperation and technical support. In many important respects, dealers function as captive distributors of the vehicle manufacturers, rather than independent businesses. This system in Japan tends to keep dealers marginally profitable and highly dependent on a single manufacturer. Dealers affiliated with one Japanese automaker strongly resist handling a competitor's products (domestically produced as well as foreign), unless enlisted to do so by their affiliated manufacturer. Some dealers have privately expressed interest in selling foreign vehicles, but have declined to do so, fearing retribution from their affiliated Japanese vehicle manufacturer. These market practices have severely limited the opportunities for foreign vehicle manufacturers to establish direct franchise agreements with Japanese dealers" (p. 191).

While these practices may indeed restrict foreign firms' sales through existing dealer networks, it is unclear (beyond high sunk costs) how these practices deter new entrants from establishing their own dealer networks. Indeed, the Japanese competition authorities have stated that exclusive purchase agreements would violate the AML, but that JFTC investigations have failed to reveal any such violations.

24. There is some evidence that the pattern of cross-shareholding is loosening, and receptiveness to foreign merger and acquisition activity increasing. According recent news reports, the value of foreign merger and acquisition in Japan increased to Y304 billion ($2.6 billion) in 1996, more than quadrupling 1995's Y72 billion, and 66 percent higher than the previous record (Y183 billion in 1994) (Financial Times, 23 January 1997).

25. The data are from 1994. There are nine Japanese and eighteen US industries in the sample, the finest level of disaggregation for which a complete data set could be assembled.

26. See Noland (1997) for details.

27. This result can be compared with previous research which has tended to find a positive relationship between concentration and outward FDI or an insignificant relationship between industry concentration and inward FDI, once other structural barriers to entry are taken into account. For a survey of the empirical literature see Caves (1996).

28. One possible reason for this outcome is omitted variable bias due to the omission of variables on capital market imperfections. Also, it would be interesting to test for inequality of the slope coefficients, in addition to the intercepts. Unfortunately a lack of degrees of freedom effectively precludes this.