A free-trade pact between the United States and Mexico is being portrayed as little more than an extension of the U.S.-Canada free trade deal of two years ago.

It is no such thing.

Free trade between the United States and Mexico would affect the distribution of income within the United States in ways that need careful consideration before such a system is implemented. The U.S.-Canada pact did not produce these effects, so this problem did not arise two years ago.

Although free trade between the United States and Mexico would increase total incomes in this country, it would also redistribute income away from unskilled and semi-skilled labor and toward professional and technical labor and capital. Because the "winners" would be people whose incomes are already above average, while the "losers" would start with below average incomes, this arrangement would make the distribution of U.S. incomes more unequal.

This is because of the types of products each country would export to the other, and because of the resulting expansion of some industries and contraction of others. It is an argument first developed 70 years ago by two Swedish economists, Eli Hecksher and Bertril Ohlin.

They noted that when there is free trade between two countries when one has an abundance of labor and the other an abundance of capital, each will export products that use a great deal of its abundant input.

Labor-abundant countries export textiles because that industry uses a great deal of labor. Countries with large amounts of capital export products such as chemicals that require a lot of capital.

In the U.S.-Mexico case, this means that the United States will import labor-intensive goods such as garments and shoes from Mexico and will export capital and professional and technical labor-intensive goods such as computers and machinery. With an underemployed population of about 90 million people, Mexico could produce a huge volume of garments and shoes.

This pattern of trade would mean that labor-intensive industries in the United States would shrink, while capital and technical labor-using industries would expand. The U.S. demand for unskilled and semi-skilled labor would fall, while the demand for capital and for highly educated labor would grow. As a result, U.S. wage rates for unskilled and semi-skilled labor would fall, while returns to capital and to professional and technical labor would rise.

The U.S.-Canada free-trade agreement did not produce this effect because these two countries have very similar economies. U.S. labor is not threatened by competition from Canadians whose wages are similar to those prevailing here, but competition with Mexican labor is a very different matter.

Retraining laid-off workers, with the goal of making them high-income skilled workers, is often seen as the answer, but experience with such programs has been very disappointing. Most of the affected workers have limited educational backgrounds, and many are not young. Despite retraining efforts, they generally have ended up with lower incomes than in the jobs they lost.

There is a solution to this income redistribution problem, but it is difficult to implement. Because total U.S. incomes would undoubtedly rise as a result of free trade with Mexico, the winners would gain more dollars than the losers would lose. This makes it possible for the winners to compensate the losers and still gain. If, for example, half of the population gains $100 each from free trade, while the other half loses $50 per person, the winners could pay the losers $50 each, thus restoring their original incomes, while still having net gains of $50. If those benefiting from free trade with Mexico paid part of their gains as additional taxes, and if the revenues were used to compensate those whose incomes would decline, this would be an arrangement in which nobody loses.

The compensation approach is theoretically simple, but politically and administratively difficult. It implies a more active income redistribution policy from Washington, which is not a politically popular idea. It is also far from easy to measure the gains and losses with precision, so the compensation would be approximate at best. In any event, this approach has never been seriously considered by the Congress as part of U.S. trade policy.

Until it is clear that compensation will be provided, the AFL-CIO and other representatives of labor are correct in opposing the U.S.-Mexico free trade proposals. Most Americans are not unskilled or semi-skilled and would clearly gain from this arrangement, but the losers would be irrational if they did not oppose it.

There is, however, one way in which U.S. labor and the AFL-CIO would gain from free trade with Mexico, but it only partially offsets these income distribution effects.

Because Mexico is on the opposite side of the Hecksher-Ohlin process from the United States, it would export labor-intensive goods, and experience an increase in wage rates. Higher Mexican wages would reduce pressures to emigrate, thus cutting the number of illegal immigrants coming to the United States to compete with U.S. workers.

For years, the AFL-CIO has been looking for ways to reduce competition from foreign workers within our economy, and now it could have one, although it is a very expensive solution to the problem of illegal immigration. It requires trade flows that would reduce U.S. wage rates, which defeats the original reason for trying to keep foreign workers out.

Free trade with Mexico or the free international mobility of labor would produce the same income distribution effects. U.S. wages fall, while those in Mexico rise. The rational AFL-CIO goal is to both avoid free trade and keep foreign workers out of this country.

The United States faces a difficult dilemma in designing a policy for its trade with developing countries. On one hand, this country wants the gains in total income that result from free trade and would also like to encourage the growth of these economies and reduce pressures for their workers to try to come here. On the other side, there is already strong evidence that the distribution of U.S. incomes has become more unequal in the last two decades, and free trade with countries such as Mexico would make that problem worse.

If the United States is going to pursue free-trade discussions with Mexico as well as other Latin American countries, as suggested recently by the White House, serious thought must be given to providing compensation.

Robert Dunn is a professor of economics at George Washington University.