Going South

Proponents of the North American Free Trade Agreement (NAFTA) assure nervous lawmakers that free trade with Mexico will not send high-wage American jobs south. They say Mexico is an industrial backwater limited to low-tech, low-productivity operations. Low wages alone, they insist, will not persuade manufacturers to uproot entrenched American operations. Rather, NAFTA will bring Mexican wages up to world-class levels, rendering the argument about wage differentials irrelevant in the long run and opening a lucrative market for goods manufactured in the United States.

But these arguments are rooted in the outdated misconceptions about Mexican manufacturing. Mexican export plants in industriesfrom automobiles to consumer electronics already match or exceed American puality and productivity levels-- and at artificially depressed Third World wages. Although average labor productivity in Mexican manufacturing has risen considerably since 1980, the hourly compensation of workers--wages and benefits--has actually fallen by more than 30 percent when adjusted for inflation.

This disparity calls into question the basic assumptions underlying both NAFTA and the recently reached side agreements that are supposed to protect American jobs. Without a stronger framework to ensure that highly skilled Mexican workers get fair compensation, the free trade agreement will encourage this trend toward a high-tech, low-wage economy in Mexico and render vulnerable even the $18-an-hour American jobs NAFTA proponents insist would remain safe after the pact is signed. Beyond the loss of jobs, the factors that depress wages in Mexico could drag down wages in the United States.

Tearing down these myths about the Mexican economy leaves an alarmingly shaky foundation--part wishful thinking, part political expediency--for NAFTA to stand on.


Myth 1: Mexico's export sector is limited to low-technology low-productivity operations.

Overall, labor productivity in manufacturing is considerably lower in Mexico than in the United States. This broad average, however, tells us little about Mexico's export sector and its potential, since the average lumps antiquated mom-and-pop machine shops for the domestic market with state-of-the-art color television assembly plants for export. Average productivity tells us where the Mexican manufacturing base has been, not its capability or its likely future. Moreover, firms do not locate production sites based on the average productivity of an economy but rather on the efficiency they expect in their operations. With this last point in mind, the auto and electronics sectors--together accounting for over half of Mexico's manufacturing exports last year--provide a good place to evaluate the new capabilities of the Mexican industrial base. Consider four complex manufacturing operations--automobile engines, car assembly, auto parts, and color television assembly. These plants, which arguably are more sophisticated than most that might be moved to Mexico, typically achieve high quality and productivity standards within 18 months to three years of starting up with young, motivated workers who have a basic education and little industrial experience. While these new factories are not yet typical, they indicate the future of Mexican manufacturing.

Auto Engines. The auto industry is one of the most advanced export-oriented sectors in Mexico today, and engine production is among the most complex and painstaking manufacturing operations in this or any other major industry. Compare the example of the $250 million engine plant in Mexico that went on line in 1982 with a U.S. plant producing the same engine with similar technology. In the key area of "machine yield"-- the central measure of the productivity of technology in a high-tech plant--the Mexican plant achieved 85 percent of U.S. performance in less than two years and 97 percent after eight years. The Mexican plant actually surpassed the U.S. plant on some complex machining lines in as little as 18 months from starting up.

The product quality of the Mexican plant was even more impressive. It surpassed the American plant in four of the six years for which company data is available (1986-1991) and exceeded U.S. quality by 32 percent in the last year of production. Not surprisingly, the automaker is investing over $500 million into the expansion and retooling of the Mexican plant, which will be capable of producing more than 400,000 engines a year.

Other Mexican engine plants report similar results. Nissan's engine plant in Aguascalientes has achieved the lowest number of end-of the-line defects of any of the company's plants. Plans call for the Mexican plant to produce all engines for the U.S.-assembled Altima. Likewise, General Motors's Toluca engine plant reports zero defects in 33 straight audits.

Auto Assembly. The results are equally impressive for auto assembly. The most sophisticated plant is Ford Motor Company's $500 million assembly and stamping plant in Hermosillo. About 2,000 hourly workers, more than 120 robots, and dozens of computerized systems produce up to 165,000 cars a year, all for export north. Three years after the plant went on line, according to company data, the Mercury Tracer it produced had the second-highest quality rating for small cars sold in the U.S. market, outdistancing even Nissan Sentra and Toyota Corolla.

In 1990, the plant switched over to assembling the more complex Ford Escort. By the end of 1992, according to J.D. Powers, the Mexican-assembled Escort had a higher quality rating than cars assembled in five of eight Japanese-owned factories in the United States and was virtually tied for fifth place out of 46 assembly plants in North America. In fact, quality in the Mexican plant was 8 percent higher than in a U.S. plant building the same car. The Mexican plant needed about 15 percent more workers per vehicle than the U.S. plant, but that is low considering the Mexican plant uses less automation and therefore has lower capital costs.

Small and midsize firms such as Breed Technologies, which makes electronic sensors for airbags, have been as successful in transferring production as corporate giants. After setting up a maquiladora, a border assembly plant, in the late 1980s, company president Allen Breed commented that "almost immediately, we had higher productivity in Mexico than we had in New Jersey," adding that "I also won't hesitate to build any new plants we might need for products which have significant labor content."

Consumer Electronics. Quality in consumer electronics products is also high. Sony operates color television assembly plants on both sides of the border, one in Tijuana and the other near San Diego. The Tijuana plant has won Sony worldwide quality awards for several years running, and the performance of both plants is so close that Sony does not break down warranty costs separately.

Subscribe to The American Prospect

Services. Advanced services also are being moved to Mexico. A dramatic example is now under construction on 100 acres adjacent to the Tijuana airport: a giant maintenance and repair station for state-of-the-art jet aircraft. Ultimately, the investment may total $250 million and employ as many as 3,000 people.

While these success stories are not yet characteristic of the industrial base, they represent some of the fastest growing areas in Mexican manufacturing. All five automakers operating in Mexico--GM, Ford, Chrysler, Volkswagen, and Nissan--built or expanded engine plants in the last decade. These plants exported almost 1.3 million engines in 1992, making Mexico one of the world's largest exporters of auto engines. The Mexican government predicts that annual engine exports will double from $1.2 billion in 1991 to $2.4 billion in 1994. Mexico's auto exports--more than 300,000 cars in 1992--could rise to almost 1 million passenger cars annually by the year 2000, the vast majority going to the United States, according to CIEMEX-WEFA, a consulting group.

What levels of experience and education do the workers in these new plants require? The firms tend to hire people without previous industrial experience to avoid "preconceptions" about how a plant should be run. In the auto engine plant, the most complex facility, the average age of the initial workers hired was in the early twenties. All lacked auto industry experience. Thirty percent had a junior high school education, and 63 percent had completed high school or a technical education. In the electronics plant, more than half of the workers had only elementary education.

A growing flow of technical and engineering school graduates provides a potential core for wider diffusion of high-tech production. Mexico had over 340,000 engineering students in 1990, almost as many as the United States, and graduated 200,000 vocational students in 1989-1990.

Myth 2: Low wages are no longer important in high-tech manufacturing.

Analysts are fond of repeating a statement by former U.S. Trade Representative Carla Hills: "If wages were the only factor, many developing countries would be economic superpowers." Wages, of course, are never the only factor. But when quality is high and productivity reaches a certain threshold--certainly the case on both counts in Mexico's export sector--wages can be decisive in plant location decisions.

Consider again the auto industry. The most advanced U.S. plant requires about 20 hours to assemble a car; the most advanced Mexican plant needs about 24 hours. Total compensation--wages and benefits--at the Big Three automakers averages about $40 an hour; in Mexico it is in the range of $5 an hour. This means that the direct labor cost for assembling a car in the most advanced plant in the United States is $800; in Mexico it is $120. If the comparable costs of two or three hundred salaried workers in an assembly plant are factored in, the savings would rise even more.

The auto assembly plant, however, is only the final point of the manufacturing process. As component plants and suppliers move to Mexico, the assembly plant savings could be repeated throughout the production chain. Overall, the average compensation cost in the motor vehicles and equipment sector in the United States was almost eight times the Mexican average in 1991--$24.21 in the United States compared with $3.33 in Mexico-- offering manufacturers a powerful incentive to relocate production.

The importance of low labor costs was underscored by Collectron, an Arizona-based management firm that sets up maquiladoras and has helped companies such as Xerox, General Electric, and ITT establish Mexican operations. In a promotional letter, the director of marketing extols the possible savings. "Companies that relocate to Mexico come primarily to take advantage of the low labor rates," he writes. "Our clients report that they have achieved savings between $10,000 and $20,000 per employee"--and this is on low-tech jobs. He then adds that "the work force is unspoiled and can be trained to a client's specifications quickly and efficiently and superior productivity can be achieved by proper management."

Low wages mean low training costs, allowing manufacturers to invest heavily in training. "When I factor in other non-labor costs--less heat, cheaper land, and cheaper construction--there is no question that Mexico's lower labor costs are decisive," John Pearlman, chairman of Zenith Electronic Corporation, told the New York Times earlier this year. The vice president of Sanyo's television operations said, "If your business is fairly labor-intensive and you have more than 500 workers, its probably a good move to come to Mexico."

Transportation costs are likely to be higher in Mexico, but even here Mexico may prove to have some surprising advantages. In an auto assembly plant, for example, shipping costs have two components: shipping parts to the plant and transporting the finished car to the dealer. The supplier base is likely to gravitate to Mexico under NAFTA, considerably reducing shipping costs for parts. Moreover, a plant in northern Mexico would have a significant cost advantage over a plant in Detroit for shipping cars to the huge markets of California and Texas.

Myth 3: World-class wages will quickly follow world-class productivity.

Economists presume that wages generally reflect productivity levels. "Mexican wages, like other low prices, have been low for a reason: Mexican productivity is low," according to Alan Reynolds of the Hudson Institute. While this may be indisputably true in introductory economics classes, the issue is more complex in the real world. Mexico's economic collapse in the 1980s, followed by government austerity measures, state strategies to attract investment, a large surplus labor force, and a highly controlled labor movement, have depressed wages in manufacturing despite gains in productivity.

The Mexican government now achieves its austerity and inflation goals through the Pact for Stability and Economic Growth. El pacto, as it is referred to in Mexico, is a state-dominated alliance of business, labor, and peasant organizations introduced in 1987. According to Business Week, it serves "to smother inflation and preserve Mexico's huge labor cost gap with U.S. and other producers."

The "official" labor group, the Confederacion de Trabajadores de Mexico (CTM), is an arm of the ruling party and, during the economic debacles of the last decade, served more to transmit government policies than to advocate for workers. Although Mexico has strong labor legislation in many areas, the state-dominated labor relations system, combined with truncated labor rights in practice, makes independent organization or opposition difficult and, at times, dangerous. Workers at Volkswagen saw how the system worked in the summer of 1992. A bitter strike ended after "Salinas gave VW permission to rip up the union contract," according to Business Week. VW terminated its 14,000 workers, instituted a new contract, and then rehired the work force minus 300 dissidents.

While independent worker organizations are stifled, firm representatives in maquiladora centers meet regularly to set compensation ranges. Rather than letting wages rise to combat turnover--the way an open labor market theoretically would operate--these associations fix wages and let turnover fall where it may, often as high as 10 to 15 percent a month.

As a result of the several factors, productivity and wages have diverged. Average labor productivity per hour in manufacturing rose by 41 percent between 1980 and 1992, according to the latest data available from INEGI, an agency of the Mexican government. (Banco de Mexico data, defining productivity somewhat differently, shows a gain of over 30 percent during this period.) Despite these productivity gains, the hourly compensation of Mexican production workers in manufacturing, in inflation-adjusted pesos, was only 68 percent of 1980 levels last year. Hourly wages alone fared even worse. They were stuck at 61 percent of 1980 levels. Indexing compensation to 100 for 1980, it climbed to 104.7 in 1982, slid to 57 in 1988, and inched up to 68 in 1992.

The gap in hourly compensation costs in manufacturing between Mexico and the United States, in dollars, has widened considerably since 1980. Mexican compensation was 22 percent of U.S. levels in 1980, 7 percent in 1987, and only 15 percent of U.S. levels in 1992. A widely expected devaluation of the Mexican peso in 1994 could widen this gap even further.

In the rapidly growing maquiladora industry, total compensation was 14 percent of U.S. levels in 1980 (in dollars) and 10 percent in 1992, despite an explosive growth in employment in the maquiladoras, from 120,000 to more than 500,000 workers. Given this rapid rise in the demand for workers, one would expect wages to be driven upwards. Instead, total compensation costs were lower in 1992 than they were in 1981--$1.64 compared with $1.68.

Wages in the cutting-edge export factories are pulled down by the same forces depressing Mexican manufacturing wages in general. As a result, plants in the U.S. and Mexico producing the same product, for the same market, and at similar productivity and quality levels have significantly different wage costs. As we have seen, the Mexican and U.S. auto assembly plants producing the Ford Escort have comparable productivity and quality. Hourly wages in Mexico, however, are about $2.38 an hour, compared with $17.51 in Detroit; total compensation, as noted above, is $5 an hour in Mexico and $40 an hour in Detroit. Are we to assume that differences of this magnitude will have no bearing on future plant location decisions or on long-term wage setting in the United States?

Mexican real hourly compensation in manufacturing began rising in 1988 and has risen 19 percent since then, significantly trailing productivity even during this period of recovery. This upward movement, however, is tempered by two factors. First, the drop was so steep in the 1980s that using the trough reached in 1988 as the base year is misleading. Total hourly wages and benefits for Mexican workers had plummeted to $1. Second, the institutional factors that decoupled wages and productivity in the first place remain in place and could slow or reverse future wage gains. Without significant reform in Mexico's industrial relations system, wages will remain depressed, preserving the enormous gap with U.S. wages that was exacerbated in the 1980s. As the American and Mexican economies become more interdependent, artificially depressed Mexican wage levels could exert a strong downward pressure on wages in the United States, given Mexican productivity in the export sector and the size of the current wage gap.

The movement of even a limited number of plants to Mexico is likely to have a powerful demonstration effect. A single plant that moves to Mexico in a given firm or industry has a chilling effect on wages and working conditions in the plants that are not moved, exerting downward pressure through example.

According to a 1992 Wall Street Journal poll, one-quarter of almost 500 corporate executives polled admitted that they are either "very likely" or "somewhat likely" to use NAFTA as a bargaining chip to hold down wages. About 40 percent indicated that they might move at least some production to Mexico within the next several years.


The Clinton administration's supplemental agreements on labor and the environment announced in mid-August supposedly add important new safeguards to NAFTA. But do they address the frayed relation between productivity and wages?

Consider the labor accord. The emphasis is on "the obligation of each [government] to ensure the enforcement of its domestic labor laws." In cases where this doesn't happen, the agreement sets out a complex four-stage process for handling grievances that begins with lodging a complaint in a National Administrative Office (NAO), located in each country. If the dispute is not resolved at this level, appeal to a ministerial council composed of the three labor secretaries is possible, and, if this still doesn't resolve the differences, a "friendly" outside panel can be appointed to study the situation. In exceptional cases, and with approval by two of the three labor secretaries, a case goes before a panel that can assess fines on the offending government.

The most glaring flaw in the accord itself is that industrial relations issues--violations of the right to strike, collective bargaining, or the right of association--are excluded from either the fact-finding panels or arbitration. If Mexican workers had limited rights before the agreement, they have gained few rights as a result of the agreement. That the Mexican government, despite its interest in passing NAFTA, would not even accept fact-finding panels that merely provide "sunshine" on labor rights heightens the fears of unions and critics in the United States that Mexico's failings are widespread and a key contributor to artificially depressed wages. Moreover, the possibility for arbitration or fines is nonexistent in those areas that most influence investment decisions. The labor side agreement does cover health and safety, child labor, and the minimum wage. These are important areas, but they will not be the primary drivers influencing new investment. Moreover, the grievance process is cumbersome, uncertain, and limited. The NAO, for example, is likely to have 10 or fewer full-time staffers, an indication of the limited role it is designed to play. If one thinks of the labor side agreement as a warranty on the original NAFTA, this warranty doesn't cover the engine or drive train.

Focusing too narrowly on the labor side accord neglects what may be the greatest lost opportunity of the NAFTA negotiations: the ability to use the promise of expanded trade with Mexico as an inducement for the Mexican government to reform its industrial relations system and allow wages to rise. Even effective side accords--far more comprehensive than the current versions--are not a substitute for this type of commitment with true collective bargaining. Mexico might still have retained a significant wage advantage, but not one based on governmental strategies that seek to attract investment at the expense of wages and a denial of labor rights. Mexican industry would have remained highly competitive, an engine for domestic Mexican economic growth, and allowed increasing numbers of workers to buy the products they produce.

Since NAFTA was proposed by President Salinas, the Mexican government has worked hard to ensure investors that investing in Michoacan would be as secure and trouble-free as investing in Michigan. The government has taken major steps to eliminate red tape and harmonize investment standards with the United States. Even in the environmental arena--where the side accords are equally weak--the Mexican government felt compelled to enact highly visible symbolic gestures from closing refineries to allocating money for environmental cleanup.

In the labor arena, in contrast, some of the most widely cited abuses have occurred since negotiations began. Among them were the arrest of a Matamoros labor leader on four-year-old tax evasion charges just as a work stoppage in the export sector began and the breaking of a Volkswagen strike near Mexico City. These acts hardly inspire confidence. Moreover, a cynic might say that the Mexican government has believed since it began negotiating with the Bush administration that the United States does not really care about labor rights. U.S. acquiescence to a supplemental agreement with a gaping hole in it seems to confirm this impression. No worse signal could be sent to the Mexican government and the multinationals that invest in Mexico.

The day the side accords were announced, President Salinas made a widely publicized address in Mexico promising to raise the minimum wage and link it to future growth in productivity. This action, in response to issues raised by House Majority Leader Richard Gephardt and other congressional Democrats, has given wage-productivity linkage such prominence for the first time in a trade debate. Salinas's pledge, however, included few specifics and is more cosmetic than substantive. The 1991 urban minimum wage in Mexico was 44 percent of 1980 levels in real terms, and Salinas made no mention of bridging this gap. Moreover, the minimum wage only applies to about 10 percent of the economically active population.

However deficient the side agreements may be, proponents argue that they provide at least a first step--maybe even only a half-step--for future progress. But accepting inadequate provisions today may dampen the prospect for change in the future since the Mexican government will have gained the trade agreement it wants, leaving little incentive for broader reform. Moreover, high productivity at low wages is addictive, and considerable corporate pressure will exist within Mexico to continue that relation indefinitely.

Increasingly, as economic arguments ring hollow, geopolitical arguments are being invoked to push NAFTA, no matter how flawed its labor and environmental provisions may be. The alternative, we are told, is economic collapse in Mexico, a wave of anti-Americanism, the dissolution of the Mexican political system, or all of the above. The likely impact on Mexico of a "no" vote on this agreement, significant to be sure, is hardly as apocalyptic as portrayed. President Salinas will suffer deep embarrassment, but neither he nor the Partido Revolucionario Institucional (PRI), the entrenched governing party, is on the verge of collapse. The political fallout in Mexico could wind up strengthening democracy rather than preserving autocratic control. The current political fortunes of the PRI, in any case, should not be the basis for accepting this agreement. Investor confidence in the short term would suffer, but long-term investment plans would continue.

The choices are not to accept this agreement or to halt trade with Mexico. The $76 billion in cross-border trade with the United States in 1992 is likely to grow in the long term whatever the fate of this pact. Clearly, the deep integration of the U.S. and Mexican economies requires more effective management and attention. A trade agreement is both necessary and desirable, but it must also ensure that labor rights and environmental standards move upwards rather than slide downhill.

You need to be logged in to comment.
(If there's one thing we know about comment trolls, it's that they're lazy)

, after login or registration your account will be connected.