Domestic Reform, Trade and Investment Liberalization, Financial Crisis, and Foreign Direct Investment in Mexico

Introduction

Foreign direct investment, which serves as the primary vehicle by which the global activities of multinational corporations are created, is today eagerly sought by the vast majority of the world's developing nations. Indeed, the perceived desirability of FDI has been enhanced by the Asian financial crisis of 1997, wherein it was (apparently) revealed that, by a number of measures, FDI flows into the affected nations were more stable than other forms of international financial flows (World Bank 1998). Worldwide, policy-makers in developing nations seek a policy mix that will lure inward FDI into their economies. However, the love affair of developing nations with FDI is of rather recent origin. As little as twenty years ago, multinational corporations were viewed with suspicion and outright hostility by policymakers in many developing nations, and policies of most such nations towards direct investment were restrictive.

Among the most restrictive of these policies twenty years ago were those of Mexico. However, for a variety of reasons, Mexico began to reevaluate its position during the 1980s such that, by the late 1980s, these policies were undergoing substantial liberalization. This liberalization of policy towards FDI was not, however, carried out in a vacuum. Rather, it took in place in conjunction with other policy reforms, to which we shall turn shortly. The liberalization was capped by Mexico's entry into the North American Free Trade Area (NAFTA) in 1995, which inter alia binds Mexico to a number of obligations pertaining to direct investment policy as per chapter 11 of the NAFTA agreement.

Mexico thus has served as something of a "laboratory" in which a number of research issues pertaining to direct investment can be tested. These include:

• Does policy liberalization serve effectively to induce FDI flows?
• If so, were the effects of entering into international obligations to reinforce this inductive effect (or, otherwise expressed, did the international obligations serve, at the margin, to increase further the inward flow of FDI)?
• Did the expected benefits of FDI actually materialize? And
• Were FDI flows more stable in times of financial crisis than flows of other forms of international capital.

On the last item, Mexico was "lucky" enough to have its own financial crisis in 1995.

The value of Mexico as a "laboratory" is arguably enhanced because the majority of FDI to enter Mexico has been from the United States. Table 1 indicates total our own calculation of the total stock of FDI in Mexico by year since 1980, along with the stock of FDI in Mexico of US origin. The figures in this table are, for the years 1981 onward, simply calculated as cumulative flows added to UN estimates of the 1980 stocks, where the flow data are from Mexican sources. Mexico calculates FDI flow on a different basis than does the United States, and thus Mexican figures for FDI inflows from the United States do not equal US figures for FDI outflows to Mexico. (The two series are, however, closely correlated.) We use the Mexican figures in order to calculate US-origin FDI as a percentage of the total in Mexico. As can be seen, this percentage has dropped steadily since 1980, but US-origin FDI nonetheless remains over 60% of the total.

While it might have been desirable, for some research purposes, for this FDI to have more diverse origins, one advantage of a dominant role of US-origin FDI for research is that the US Government provides more detailed data for US outward direct investment than does any other government. Thus, the focus of this article is, to some large extent, what is revealed by these data.

Alas, even in the context of a "laboratory" such as Mexico for which unusually detailed data are available, it is difficult to answer these questions unequivocally. The reason is that, as already stated, policy reform in Mexico has taken the form of an ensemble of actions, and it is almost impossible to determine what specific actions led to what specific results. In the case of FDI, what is clear is that the ensemble produced an overall climate that foreign direct investors found highly favorable, such that this investment into Mexico began to accelerate following the onset of reform in the late 1980s (well in advance of NAFTA) and has grown at pretty much unabated rates ever since. The financial crisis of 1995 did cause a temporary attenuation of this growth, but the effect now appears to be a "blip" and little more. Nonetheless, there is some evidence that foreign direct investors did in the aggregate withdraw liquid funds from Mexico at the time of the crisis. For technical reasons, this withdrawal does not in fact appear in the FDI data per se. However, after the period of crisis ended, these funds appear to have moved back into Mexico. Overall, the evidence points to significant economic benefit to Mexico from this investment, albeit that some commentators worry about changes in the Mexico's culture resulting from an element of "Americanization" that has doubtlessly resulted at least in part from so much involvement of foreign firms in Mexico's domestic economy.

Mexican Reform and the Response of US Direct Investment in the 1980s

Figure 1 indicates the annual flow (broken line) and year-end stock (solid line) of US direct investment in Mexico 1966-1998. Because direct investment is lumpy (i.e., individual undertakings can represent substantial percentages of the total flow in any one year), flow figures tend to exhibit a volatility that is exaggerated by this lumpiness. Trends are probably thus better observed via stock figures than flow figures. In this matter, it should also be kept in mind that foreign direct investment is important above all else because it creates economic activity and that the stock of direct investment is a better indicator of this activity than is annual flow.

Cursory examination of Figure 1 suggests that the stock of US FDI in Mexico grew steadily from 1966 through the late 1980s, and that thereafter the rate of growth accelerated sharply. (An apparent drop in the stock in 1982 is the result of recalibration of the data by the US Commerce Department and not of an actual pullout of US investors.) More careful analysis (Figure 2) reveals that indeed there was an identifiable trend break in 1989.1 In fact, during the period 1966-88, the stock of US FDI grew at an average annual compound rate of about 3.1%, while from 1989-1998, this rate jumped to almost 5.6%. This latter rate was maintained in spite of a lapse in the growth at the time of the Mexican financial crisis in 1995.

Three things stand out about this trend break. First, it occurred well before the onset of NAFTA. The governments of Mexico and the United States indicated that they would seek to negotiate a free trade agreement only in June, 1990, and negotiations began one year later in 1991. The earliest indications that NAFTA might be in the offing seem to be in the form of "leaks" from the Mexican Government that appeared in the press in the early spring of 1990. Thus, the trend break cannot be attributed to NAFTA nor even to expectations that NAFTA would occur.

Second, the trend break occurred right after the Mexican election of 1988, wherein the incumbent Institutional Revolutionary Party (PRI) was returned to office.2 This followed a serious challenge from populist rivals that, had they been elected, might have slowed or dismantled the economic reforms that were initiated by the PRI under the previous administration of President Miguel de la Madrid. It is tempting in light of this second fact to assert that the return to office of the reformist PRI in 1988 was the principle cause of the upswing in direct investment that subsequently flowed into Mexico. However, to assert this would be to ignore that direct investment typically entails rather long lead times, such that actual realization of a direct investment can lag the decision to make this investment by two years or more.3 Thus, the trend break that was actually observed in 1989 can be attributed to decisions that were made in 1987 or even earlier. And these decisions would have thus been made prior to the 1988 election.

The third thing that stands out is that the trend break occurred in the same year that the Mexican Government announced sweeping (and liberalizing) changes in the highly restrictive Law to Promote Mexican Investment and to Regulate Foreign Investment of 1973 (often abbreviated to Law on Foreign Investment or LFI).4 However, the temptation to ascribe the trend break to this announcement runs into the same objection as just above: this announcement came too late (May, 1979) to have affected investment decisions that created the trend break. Again, these decisions had been taken well in advance of the changes in the LFI. It cannot be ruled out, of course, that investors held expectations of changes in the LFI prior to these changes being announced. But this gets us slightly ahead of our story.

One further thing can be said: the trend break was not accounted by what has become known as the "maquiladora phenomenon" in Mexico. Maquiladora are plants in Mexico, mostly foreign-owned, that engage in in-bond processing or secondary assembly of semifinished goods or components imported into Mexico that are then re-exported. Most maquiladora in Mexico both import from and export to the United States. Such operations have been authorized under Mexico's Border Industrialization Program since 1965. Originally confined to the US border area, a 1972 law enabled maquiladora to be located almost anywhere in Mexico. These operations are exempt from most provisions of the LFI, and the imported inputs into these plants are exempted from Mexican import duties5. Products that are reimported into the United States have been free of duties on the component of value-added that is of US-origin and, in addition, have benefited from special provisions of the Generalized System of Preferences (GSP). Until 1983, the output of a maquiladora had to be reexported 100%. In that year, a decree authorized sales of up to 20% of output in the domestic market, and a 1989 decree effectively increased permissible sales in the domestic market of 33% of output.

The maquiladora program in Mexico was a success virtually from inception and the number of such plants has grown steadily since the late 1960s. In fact, during both the decades of the 1970s and the 1980s, the number of maquiladoras (and the number of employees in them) more than quadrupled. But no discernible trend break in the creation of maquiladora operations is evident in 1988.

What then did account for the trend break? The two probable causes are (1) a series of bilateral trade agreements struck between the United States and Mexico during the period 1985-89 and (2) policy reforms implemented by Mexico unilaterally. With respect to the first of these, there were three such agreements6: an Understanding on Subsidies and Countervailing Duties (1985), the Framework of Principles and Procedures for Consultation Regarding Trade and Investment Relations (1987), and the Understanding Regarding Trade and Investment Facilitation Talks (1989). However, with the exception of the first of these, these three agreements would seem to have been struck too late to have affected investment decisions taken in 1987 (they could, however, of course have figured in later investment decisions). The 1985 agreement largely brought Mexico into compliance with the GATT Agreement on Subsidies and Countervailing Measures and adjusted US procedures to account for this fact. This unlikely would have had major impact on US investment decisions regarding Mexico. The 1987 Understanding resulted in an "immediate action agenda" for the steel and textile sectors; subsequent US direct investment in Mexico, however, did not seem to originate from these sectors.

Thus, in a search for an explanation for the trend break, the first potential explanation, bilateral agreements between Mexico and the United States seems to yield little fruit. Attention thus turns to the second potential explanation, unilateral policy reform undertaken by Mexico itself. This reform was begun in the middle 1980s, under duress, following a twenty year period where the Mexican economy might be best likened to a roller-coaster ride.

In fact, the reforms of the 1980s were in many ways a continuation of reforms that had been implemented during the late 1960s. These earlier reforms, inter alia, reduced import protection and acted to reverse still earlier policies of import substitution under which Mexican economic growth, while quite high, fell behind that of a number of other middle income nations, especially in East Asia. For a time, during the late 1960s and early 1970s, these reforms seemed to be working: growth was up and prices were stable. However, the oil price shock of 1974 sent Mexico's current account into deficit, and one response was for the Mexican Government largely to reverse the reforms and to increase import protection. At the same time, to maintain growth, public expenditure was increased, but with no offsetting increase in public revenue. The result was to cause a further deterioration in the current account. However, external borrowing was increased such there was upward pressure on the exchange rate. This policy mix created a bias against exports, and one consequence of Mexico's actions was loss of market share in export markets during a time when the export markets themselves were in recession. The slowdown in these sectors in turn drove Mexico's rate of increase of total factor productivity essentially to zero (World Bank 1993). One consequence was that there was a severe recession in Mexico in 1976, and a devaluation of the peso.

However, substantial new oil reserves were discovered in Mexico in 1977. Thus, in contrast to five years earlier, a second round of oil price increases in 1979 benefited Mexico. And, following on the declines in real per capita GDP that had been registered in 1976-77, per capita GDP grew sharply in 1979-1980. Alas, the response by the nation was to step up even more its international borrowing spree and to use the revenue to finance populist programs offering little or no long run return. This set the stage for near-disaster when, as soon happened, international events turned against Mexico. After peaking in 1981, oil prices began to fall in subsequent years, while international interest rates rose. Concurrently, primary fiscal deficits (i.e., public sector deficits before interest) rose, reaching 8% of GDP in 1981. Real per capita GDP declined, and continued to do so until 1988.

In 1982, the Mexican Government announced that it could not service its international debt obligations, triggering the Latin American sovereign debt crisis that ultimately was resolved through international debt relief. Faced with crisis, the Mexican Government responded with fiscal tightening, mostly in the form of reduction of the bloated expenditures.7 The results were that in 1983 there was recorded a primary fiscal surplus of 4% of GDP. 1985 witnessed further adversity, however, in the form of collapsing oil prices, and this surplus was reduced to 1.5% of GDP.

It was at this time that the Mexican Government began to introduce structural reforms. These were motivated by observations by economic analysts that rates of productivity gains in specific Mexican sectors had tended to slow following imposition of additional import protection in the early 1970s relative to the earlier period of reform and, indeed, in some sectors there had been negative growth in productivity (Samaniego 1985, cited in Trigueros 1989). Also, as noted by Trigueros (1989), Mexican policy had provided "fertile ground for implementation of an increasing number of regulations that interfere with market activity". Accordingly, the goal became to reduce levels of protection not simply by reducing tariffs but by deregulating sectors where imports had been subject to licensing and other regulatory requirements. Again according to Trigueros (1989), such licensing requirements applied to over 90% of traded goods in 1985 but by 1988 to only about 25% of traded goods (based on value of domestic output). At the same time, tariffs were lowered from an average of more than 28% to an average of just under 12%. Trigueros indicated that, by 1988, the domestic process had reached an "impasse" with respect to further liberalization of import protection, and he argued that further such liberalization would require that Mexico enter into international negotiations whereby access to Mexican markets could be traded for greater access by Mexican producers to foreign markets - meaning, of course, mostly the US market.8

As already noted, the reforms of 1985-1988 in trade and regulatory policy were followed by a major relaxation of the restrictive Law on Foreign Investment in 1989. The point to be established here is that this last reform was part of a whole string of reforms designed to open the Mexican economy to greater foreign participation and, indeed, the reforms in foreign direct investment policy lagged behind trade and regulatory reforms. It is thus likely that the surge in direct investment that followed was the product of the whole ensemble of reform and not simply the reform in direct investment policy per se (and, indeed, the trend break in FDI that was registered in 1989 almost surely could not have been created solely by the changes in the LFI that were announced that same year).

Likewise, the continuance of the rapid rate of growth of FDI stock arguably required more than reform of investment policy per se. The trend in Mexican policy continued in the direction of liberalization in the 1990s. In particular, a large number of state-owned enterprises were privatized in the period 1989-92. The 1989 changes in the LFI enabled substantial foreign participation in these privatizations, including in the telecommunications, insurance, banking, petrochemicals, and mining sectors. For example, a large minority block of the stock of the formerly state-owned telecoms firm Telmex was acquired by a consortium of foreign firms, giving this consortium effective control (the remainder of the stock was widely held).

Impact of the NAFTA

The case is strong, as indicated in the last section, against the NAFTA being the main causal agent behind the favorable trend break in US direct investment into Mexico. However, it is plausible, even if not provable, that the NAFTA ensured that the flow of FDI into Mexico did not abate. However, even so, embedded in the issue of the role of the NAFTA is the issue of exactly how important to this flow are the provisions of the agreement pertaining to direct investment per se and how important are other provisions. The NAFTA in fact contains the most extensive set of rules yet agreed upon in an international context to pertain to direct investment and, for the most part, these rules act in the direction of liberalizing government policy.9 However, these same rules are conditioned by exceptions, and the exceptions that apply to Mexico are large in number, much larger than for the other two nations that participate in the NAFTA, the United States and Canada (Rugman and Gestrin 1994). The rules are not hollow. Individual investors who believe that their interests have been hurt by a NAFTA government's violation of these rules can turn to an investor-to-state dispute resolution mechanism established under NAFTA chapter 11 part B. If the finding of the tribunal established under this mechanism goes against the government, the investor can recover monetary damages. The first case to be brought under these provisions was that of The Ethyl Corporation against the Government of Canada, and the case has had wide ramifications (Soloway 1999). Arguably, the chapter B provisions make the investment provisions of the NAFTA more credible than they otherwise would be.

Indeed, much the same statement would apply to Mexican participation in NAFTA in general: liberalization, including but not limited to investment liberalization, undertaken to meet NAFTA obligations by Mexico was more credible than unilateral liberalization. This is for two reasons. First, the obligations of NAFTA were enforceable, or at least so to the extent that violations could be subject to sanctions. Second, the NAFTA arguably created something of a lock-in: Once Mexico committed to meeting a NAFTA obligation, it would have been difficult to back out.

However, of course, NAFTA is much more than a set of provisions pertaining to investment. It creates a free trade area among Canada, the United States, and Mexico, albeit one with some rather stringent conditions attached to its functioning. Thus, the question raises itself, is the "NAFTA effect" that foreign investment flowed into Mexico because, under NAFTA, Mexico further significantly liberalized its policies towards FDI, or is it because Mexico, as part of a free trade area, now could be even more effectively used as a production base for goods ultimately to be sold elsewhere in North America?

The answer to this question is, of course, likely to be "both reasons apply"? But, even if this is correct, to what extent do they apply?

Figure 3 attempts to get at this issue. It indicates, indexed so that 1991=100, the volume of sales of affiliates of US firms in Mexico in the domestic Mexican market, the volume of sales to the United States (which equals approximately the volume of products exported to the United States), and the volume of sales to other markets. 1991 is chosen as the base year on grounds that firms from that year forward could reasonably have held expectations that a North American free trade area would come into being, even if this was not realized until 1995. As can be seen, the volume of sales to other markets was by far the fastest growing of these three variables; this variable expanded particularly quickly after 1995, which likely can be explained by the devaluation of the peso in that year, about which more will be said in a later section. Sales to the United States grew the next fastest and, again, there is some acceleration after 1995. Sales to the domestic Mexican market actually declined in 1995, reflecting the crisis in that year, and overall grew more slowly than the other two variables.

These comparisons suggest that the greater motivation for direct investors to locate operations in Mexico was to take advantage of a Mexican location in order to serve the whole North American market than to serve the domestic market of Mexico itself. This in fact is quite consistent with anecdotal evidence regarding foreign firm operations in Mexico (see, e.g., Goad 1999). Thus, we believe that the data are consistent with the following statement: that the continued expansion of FDI in Mexico was as much or more driven by the expectation of the free trade area created by NAFTA than by the investment provisions of the NAFTA per se or the liberalization associated with them. However, to nuance this somewhat, let us note that the latter might or might not have been a necessary condition for the investment to flow in the quantities it has, but it almost surely was not a sufficient condition.10

Has Mexico Received Benefit from Recent Inward Direct Investment

A truly honest answer to the question "has Mexico benefited from the recent inflows of foreign direct investment" would be "it is really too early to know for sure." Without question, this investment is creating large transformations in the economy and the society of Mexico. Some of these transformations have manifested themselves in ways that most observers would rate as positive, e.g., higher incomes for large numbers of Mexicans (as we shall see shortly, the data simply do not support the assertions of anti-globalists who maintain that direct investment in Mexico has actually impoverished workers there). However, some negatives clearly are also present, including environmental despoliation in areas heavily affected by new investment and rising income disparities.

The latter, it should be noted, seem more related to trade liberalization than to direct investment in Mexico, in that the trade reforms of the middle 1980s caused the relative price of unskilled labor-intensive goods to fall relative to that of skill-intensive goods, possibly with the effect of depressing wages of unskilled workers relative to wages of skilled workers (Hanson and Harrison 1999). To the extent that FDI figures in this picture, it probably is in the direction of increasing the wages of skilled workers (see discussion below), thus further increasing the ratio between unskilled and skilled workers. However, this latter effect is due to rises in the real wages of skilled workers, not to declines in the real wages of unskilled workers. The former effect, in contrast, might have cause real wage declines among the low skilled workers of Mexico.

In fact, the economic performance of Mexico in recent years has been erratic. Figure 4 indicates an index of real GDP in Mexico since 1981. As can be seen, the growth rate of GDP has been rather volatile. Also shown in the figure is an index of the price of oil, and it is observable that there was a rough correlation between oil prices and GDP growth during the 1980s but that this correlation seems to have ended during the 1990s. The financial crisis of 1995 is associated with a very sharp decline of real GDP growth into negative figures that occurred when oil prices were actually rising. Overall, Mexican growth during the 1990s has been positive but not impressive.

Has FDI contributed to recent Mexican growth? Almost surely the answer is "yes" but, again, the short time horizon and the multiplicity of other factors that affect growth make it almost impossible to assess how great is the overall effect of FDI.11 Also virtually impossible to ascertain is whether FDI has contributed, either positively or negatively, to the volatility of this growth. However, what is possible to do is to examine some of the microeconomic effects of FDI. We proceed to review briefly some of the results that have been reported thus far, which, as we shall see, are overwhelmingly positive.

To return to the issue of wages: empirical work reported to date suggests unequivocally that foreign direct investment in Mexico is associated with higher, and not lower, wages in the domestic Mexican economy. This is an important point to stress because certain antiglobalists consistently maintain that foreign direct investment in developing nations - and Mexico's experience under NAFTA is often cited - tends to impoverish workers or to create "slave labor conditions".12

These assertions simply fly in the face of contrary evidence. For example, Aitken, Harrison and Lipsey (1996) report the existence of a large wage premium associated with foreign ownership of economic activity in both Venezuela and Mexico. (It should be noted, however, that two of the same authors, as noted above, do find that trade liberalization in Mexico might have had negative effects on the wages of unskilled Mexican workers.) In a related work, Feenstra and Hanson (1997) show that increases in demand for skilled labor is positively associated with growth of FDI, and that the areas in Mexico where FDI is concentrated are exactly those where wages for skilled labor have risen the most.

This effect has doubtlessly been boosted by Mexico's participation in NAFTA. This participation seems to have increased demand for Mexican exports embodying high skilled labor, where much of this demand has been met by foreign-controlled enterprise locating in Mexico (Espinosa and Noyola 1997). This finding would suggest that the main source of income inequality in Mexico is in fact rising wages of skilled workers rather than decline in real wages of unskilled workers, in some contrast to the Hanson and Harrison (1999) finding reported earlier.13

Completely consistent with the finding that FDI or, more precisely, foreign-owned economic activity is positively associated with rising wages in Mexico is a consistent finding that foreign-owned enterprises are associated with higher productive efficiency there than are domestic enterprises (Blomstrom 1985, Blomstrom and Kokko 1997 op. cit.).14 Also, spillovers from foreign-owned enterprises in Mexico have long been evident (Blomstrom and Persons 1983), and recent evidence (Aitken, Hanson, and Harrison 1997) finds the existence of spillovers associated with foreign-owned enterprises with respect to export activity (including activity by domestically-owned firms; the specific finding is that domestically-owned Mexican plants located in close proximity to foreign-owned activities have a higher than normal propensity to export). We note that this last statement goes beyond the assertion that foreign-owned activities in Mexico have a high propensity to export. Such a high propensity is, of course, an expected finding, given that so many of these activities are maquiladoras created largely or solely for the purposes of exporting (see previous discussion). What is implied by the Aitken et al. finding is that the propensity of any firm locating in a region where foreign-owned activities are clustered to export is higher than the propensity of a firm to locate in another region, even after controlling for overall exports from those regions. There is "something" about foreign ownership that induces exports that goes beyond the simple clustering of export activities. Moran (1998), observing this effect, postulates that it has something to do with both networking effects and creations of external scale economies.

Overall, then, has FDI been good for the Mexican economy? The empirical results reported to date suggest that the answer is "yes" but, as noted above, the jury is nonetheless still out on this issue. Some disturbing effects of FDI have been reported, e.g,. social dislocation and environmental despoliation, for which there is not available quantitative analysis to indicate how damaging these are in terms of economic effects.15 Those who are concerned about these effects need to keep in mind, however, that if foreign investment in Mexico does have the effect of raising incomes (and the evidence is strongly supportive that it does), higher incomes can be utilized to offset the harm that the transformations cause, e.g., the environment can be cleaned up and effective social welfare programs can be instituted. But, whether these will be effectively implemented in Mexico, only time will tell.

The Crisis of 1995 and the Response of Foreign Direct Investment and Foreign Controlled Enterprises

As noted in the introduction, foreign direct investment has received something of a blessing from the World Bank by virtue of its stability as an international financial flow during times of financial crisis. In the case of Mexico and the crisis of 1995, it is clear that FDI from the United States was indeed stable. During the crisis year, there was a slowdown in the rate new entry of FDI into Mexico, but there was no exit of this form of investment.

However, FDI is largely a measure of owner's equity in on-going business enterprises, where the owners are the foreign investors (who largely in fact are "legal persons", i.e., multinational corporations, rather than "natural persons", i.e., individual human beings) and the on-going business enterprises are largely the affiliates of the multinational corporations. In order for there to be exit of FDI from a country, these affiliates must either be sold or shut down or report dividends in excess of earnings such that retained earnings are drawn down. The latter would likely occur only in times of severe recession, and the former only if the foreign investors intended to exit the nation permanently.16

However, equity flows are not the same as cash flows, and what is most characteristic of a financial crisis is that the affected nation becomes illiquid, not necessarily that its overall international investment position deteriorates. It is argued that Mexico's overall problem in the early 1990s was exchange rate appreciation, which combined with the trade liberalization of the late 1980s, led to a deteriorating current account balance (Williamson 1997). Ultimately, this policy mix a drain of official reserves as investors began to anticipate that the peso would have to devalue and, accordingly, moved funds offshore. Of relevance is the question, were foreign investors among those who did so?

The answer, unsurprisingly, is that they probably did. To see this, it is necessary to examine not data on FDI per se, but rather balance sheet information for foreign-controlled affiliates. Figure 5 presents some such information for affiliates of US firms in Mexico. Two items are shown, owners' equity (which corresponds roughly to the stock of FDI) and current assets of these affiliates. The key thing is that the latter of these quantities, reflecting the liquid component of assets of these affiliates, dipped in 1995, whereas the former did not. Both quantities continued to grow in years following the crisis.

The dip in current assets of US affiliates is consistent with their withdrawing of funds from Mexico during that year. Alas, one cannot say for sure from these data what the magnitude of the withdrawal was or, indeed, even for certain that it occurred at all. (This latter is true because some of the dip might be accounted by currency translation losses reported by affiliates due to the devaluation of the Mexican peso that occurred in that year; how much of the dip, however, is the result of such asset revaluation cannot be determined from the published data.) All that can be said is that the dip is suggestive of some pullout of funds. It can be further said, and this is of comfort to those who have faith in the stabilizing effects of FDI, if liquid funds were withdrawn from Mexico by foreign controlled enterprises operating in Mexico in 1995, these funds seem to have been restored by 1996.17

What is important for analysts to realize is that flow of funds within a network affiliates of multinational enterprises is not the same as flow of FDI as reported in balance of payments statistics. The evidence suggests in fact that there might have been a net outflow of funds from affiliates of US firms in Mexico in 1995 even though there was a small net inflow of FDI. This should neither be surprising or alarming. After all, if investors do behave rationally in their own self interests in the face of policy errors by national governments, it should be expected that foreign direct investors would behave similarly to other investors. It is quite clear that domestic investors and foreign holders of portfolio investment moved money out of Mexico in anticipation of the 1995 financial crisis and thus it should be no surprise that foreign direct investors might have done likewise.

Table 1: Share of US FDI in Total FDI in Mexico

References

1966 Aharoni, Yair, The Foreign Investment Decision Process (Boston: Harvard Business School Division of Research).

1997 Aitken, Brian, Gordon H. Hanson, and Ann Harrison, "Spillovers, Foreign Investment, and Export Behavior", Journal of International Economics 43.

1996 Aitken, Brian, Ann Harrison, and Robert E. Lipsey, "Wages and Foreign Ownership: A Comparative Study of Venezuela, Mexico, and the United States", Journal of International Economics 40.

1997 Belausteguigoitia, Juan Carlos, and Luis F. Guadarrama, "United States-Mexico Relations: Environmental Issues", in Bosworth, Barry P., Susan M. Collins, and Nora Claudia Lustig, editors, Coming Together? Mexico-US Relations (Washington, DC: The Brookings Institution).

1986 Blomstrom, Magnus, "Foreign Investment and Productive Efficiency: the Case of Mexico", Journal of Industrial Economics 35.

1997 Blomstrom, Magnus, and Ari Kokko, "How Foreign Investment Affects Host Countries", World Bank International Economics Department Policy Research Working Paper No. 1745.

1983 Blomstrom, Magnus, and H. Person, "Foreign Investment and Spillover Efficiency in an Underdeveloped Economy: The Case of Mexico", World Development 11.

1997 Borzenstein, Eduardo, José de Gregorio, and Jong What Lee, "How Does Foreign Investment Affect Growth?", Journal of International Economics 45.

1991 Cline, William R., "Mexico: Economic Reform and Development Strategy", EXIM Review, special issue (Tokyo: Export-import Bank of Japan).

1958 Dunning, John H., American Investment in British Manufacturing Industry (London: George Allen and Unwin).

1997 Espinosa, J. Enrique and Pedro Noyola, "Emerging Patterns in Mexico-US Trade", in Bosworth, et. al., editors, Coming Together? Mexico-US Relations.

1999 Goad, G. Pierre, "Mexican Wave", Far Eastern Economic Review, issue of November 11.

1994 Graham, Edward M., and Christopher Wilkie, "Multinationals and the Investment Provisions of the NAFTA", The International Trade Journal, 8.

1999 Hanson, Gordon H., and Ann E. Harrison, "Trade Liberalization and Wage Inequality in Mexico, Industrial and Labor Relations Review 52.

1992 Hufbauer, Gary C., and Jeffrey J. Schott, North American Free Trade: Issues and Recommendations (Washington, DC: The Institute for International Economics).

1998 Moran, Thedore H., Foreign Direct Investment and Development (Washington, DC: Institute for International Economics).

1994 Rugman, Alan M., and Michael Gestrin, "NAFTA's Treatment of Foreign Investment", in Alan M. Rugman, editor, Foreign Investment and NAFTA (Columbia, SC: University of South Carolina Press).

1985 Samaniego, Ricardo, "The Evolution of Total Factor Productivity in the Manufacturing Sector of Mexico, 1972-1982, Documentos de Investigacion (Mexico City: El Colegio de Mexico).

1999 Soloway, Julie A., "Environmental Trade Barriers in NAFTA: The MMT Fuel Additives Controversy" Minnesota Journal of Global Trade, 8.

1989 Trigueros, Ignacio, "A Free Trade Agreement Between the United States and Mexico?", in Jeffrey J. Schott, editor, Free Trade Areas and US Trade Policy (Washington, DC: Institute for International Economics). 1984 Wientraub, Sidney, Free Trade Between the United States and Mexico (Washington, DC: The Brookings Insitution).

1997 Williamson, John, "Mexican Policy toward Foreign Borrowing", in Bosworth, et. al., editors, Coming Together? Mexico-US Relations (Washington, DC: The Brookings Institution).

1998 World Bank, Global Development Finance (Washington, DC: The World Bank).

Notes

1. This is revealed because a piecewise linear relationship, with cusp at t=1989, fits ln y(t) = rt + A, where y(t) = the stock of US FDI in Mexico in year t, 1966 £ t £ 1998, and r and A are arbitrary constants, better than either a simple linear relationship or a piecewise linear relationship with a cusp in any other year. We did not test for the possibility of two or more trend breaks because the visual data suggests only one such break, excluding the region around t = 1982, where a revision in the data series occurred.

2. The PRI in fact has continuously held office since the conclusion of the Mexican Revolution of 1916.

3. That realization of a direct investment undertaking lags by several years managerial decisions to make the investment has been a consistent finding of case studies of direct investment dating to Aharoni (1966).

4. Both the 1973 law and the 1979 reforms are described in some detail in Hufbauer and Schott (1992).

5. Certain other preferences were accorded these operations by Mexico, e.g., the machinery used in the plants was generally exempt from import duties. See Hufbauer and Schott 1992, op. cit., for details.

6. Details can be found in Hufbauer and Schott (1992), op. cit.

7. See Cline 1991 for a detailed exposition of Mexican policies during this period.

8. That Mexico needed to liberalize its import protection regime but that considerations of domestic Mexican politics suggested that this was best done in the context of bilateral negotiations with the United States was a conclusion that had been reached earlier by Weintraub (1984).

9. For an analysis of NAFTA chapter 11 provisions, see Graham and Wilkie (1994).

10. Thus, the issue is, had there been created a NAFTA without the chapter 11 provisions, would the investment flowed at levels equivalent to those actually recorded? To answer this requires information about the counterfactual which of course is not available. In the end, probably the best we can do is to say that the chapter 11 provisions certainly did not hurt, but they might not have been necessary.

11. We tend to be very suspicious of efforts to quantify this effect, e.g., via use of models that assume that FDI adds a measurable increment to a nation's capital stock that would not have been added in the absence of FDI. As is well established in the literature, FDI does much more than to augment a nation's capital stock, e.g., it brings new technologies into an economy, affects the level of competition within key sectors, usually, but not always, increasing the amount of effective competition, and alters the mix of output. See, e.g., Blomstrom and Kokko (1997); see also Borzenstein, de Gregorio, and Lee (1997), who show using cross sectional analysis that FDI seems to contribute more to economic growth in developing nations than domestic investment. We simply do not believe that the data available are sufficient even to begin to capture the overall effect of recent FDI into Mexico.

12. See, e.g., report on internet by Public Citizen Global Trade Watch located at HTTP:www.citizen.org/pctrade/nafta/reports/5years.htm#wage.

13. In this matter, it should be noted that Harrison and Hanson (1999) is not based on more recent data than Espinosa and Noyola (1997); rather, "publication lag" accounts for the more recent date on the former.

14. Findings specific to Mexico are consistent with similar findings in other locations, including both developing and developed nations. A long literature on this subject exists, dating to Dunning 1958.

15. Belausteguigoitia and Guadarrama (1997) discuss aspects of the environmental issue.

16. Costs of exit and reentry are quite high in most sectors and, both because of these costs and for other reasons well explained in the industrial organization literature, incumbent firms intending to retain a presence in a market over the long run tend not to exit during bad times and reenter when times improve.

17. It should be noted that if funds were withdrawn sometime in 1995 but replaced before the end of the year, this movement would not show up at all in these data.