The effective collapse of the Doha round of trade talks in Cancun in September has left the world trade scene in chaos, with an increase in protectionism likely. Free trade, beloved by economists, has only a modest worldwide political constituency, and severe practical defects in a world of floating exchange rates. And the world is turning toward the worst kind of protectionism. Instead of mere tariffs that are at least quantifiable, regulatable and revenue-producing, the world is adopting subsidies, quotas and non-tariff barriers, all less transparent and all more distorting of trade flows. The current world trade regulation system is thus both dangerous and unsatisfactory.

There is a trade structure that would better suit the world's needs, that would alleviate both the political and economic problems inherent in international trade, but not eliminate them (which would be impossible). Politically, the need is for an initiative similar to that of the Reagan Administration's tax reform of 1986: special interests need to give up their favored tax breaks and trade loopholes, replacing them with a modest general tariff. This would then build a system that satisfies the most important needs of all parties and, by its simplicity and efficiency, comes close to maximizing overall welfare.

Like most economic theories, the doctrine of comparative advantage rests on a number of theoretical assumptions that are often untrue in practice. In particular, the doctrine ignores the frictional costs of people losing their jobs and being forced to change their livelihoods. Thus, in cases where the advantage from free trade is modest and its disruption severe, free trade may not be wealth-maximizing, for the economy as a whole, when the time value of money is taken into account. Moreover, free trade involves inequality of sacrifice. In most cases, the benefits of free trade are hidden and accrue in small amounts to consumers all over the society. The costs, on the other hand, are highly visible, bearing particularly harshly on a relatively small group of people, generally of modest income and education. It has frequently been said that free trade is a mechanism to redistribute income from poor people in rich countries to rich people in poor countries; the history of blue-collar wage levels in the United States since 1973 tends to validate this thesis (although heavy immigration, a separate question, has exacerbated the problem by still further augmenting the supply of modestly skilled U.S. labor).

The benefits of freeing trade in an item do not in practice accrue in perpetuity; varying wage and other input costs and the inflow of new technology mean that cost differentials between production sources are quite unstable, occurring for only a few years, or at most a decade or two. Moreover, movements of exchange rates between currency blocs cause the comparative advantage between product sources to swing to and FRO, as respective currencies rise or fall. If we lived in a world of fixed parity gold-standard exchange rates and minimal technological advance, like Adam Smith's 18th century, then completely free trade would optimize the system over time--production factors would migrate to their nexus of greatest comparative advantage. But we have not lived in such a world since 1914 at the latest.

In the world of exchange rates that fluctuate against each other by substantial percentages (apparently in a rough long-term cycle), the optimization potential of comparative advantage does not work--even if you ignore technological changes. For example, from 1985-95, the deutschemark strengthened from DEM 3.17 = US$1 to DEM 1.42 = US$1, then weakened to the equivalent of DEM 2.38 = US$1 by 2002 (by which time it was of course subsumed in the euro). Inflation in both Germany and the United States was low and comparable throughout the period.

Consider manufacturers of equivalent products in Germany and the United States, whose products are primarily domestically manufactured but can be exported to the other country at a cost of 5 percent of their value. Suppose the products are closely equivalent, with the German product's manufacturing cost being DEM 200, and the U.S. product's manufacturing cost being $100. Then, at the deutschemark's nadir in 1985, the German company could export to the United States at a cost of $63.09, plus $3.15 freight, or $66.24 in the U.S. market, compared with the U.S. manufacturer's $100. The U.S. producer, facing such a huge cost differential, would be rapidly driven out of business.

Ten years later, the U.S. manufacturer could sell into Germany at a cost of DEM 142 plus DEM 7.10 freight, or DEM 149.10 total, compared with the German manufacturer's DEM 200. This time, it is the German manufacturer who would be driven out of business. By 2002, the German manufacturer again would have the advantage, selling in the United States for $84.03 plus $4.20 freight, or $88.23 total, with the U.S. manufacturer at a severe disadvantage.

Needless to say, driving manufacturers out of business in alternate countries once a decade is not optimal for anybody. In such a case, a low tariff, no more than 10 percent or so, imposed by each side on imports from the other, may help stabilize the system. Such a tariff is equivalent to raising shipping costs from 5 percent to 15 percent, thus making exporting more expensive. In the case above, the German product costs $72.86 in the U.S. market in 1985 and $97.11 in 2002, while the U.S. product costs DEM 164.01 in 1995. Since the extreme exchange rate values last for only a few months, with a reversion closer to the mean quite quickly, such a 10 percent tariff would probably be enough to prevent the bankruptcy of either manufacturer from exchange rate fluctuations alone.

The low tariff does not prevent truly substantial cost advantages from winning out. If the German company, instead of DEM 200 in manufacturing cost, had DEM 160, then its U.S. cost, even with the tariff, would be $58.29 in 1985 and $77.69 in 2002. This is enough of an advantage for it to gain dominance in the U.S. market, which it would maintain through most of the exchange rate cycle. The same analysis applies in the more complex case where there are several different manufacturers, each located in a different currency bloc.

In a completely free-trade system, temporary overvaluation of a particular currency against its trading partners causes bankruptcies and job losses that, taking into account the long term trend of exchange rates, are not justified. Experience has shown that currencies do not hover around their "purchasing power parity"; they fluctuate substantially beyond any rational estimation of their value, driven by speculative money flows and capital market transactions, far more than by trade.

Such an overvaluation of the dollar in 2002, combined with the political strength of the U.S. steel industry, caused President George W. Bush to impose high "anti-dumping" duties on many foreign steel imports. As the dollar dropped after the spring of 2002, the problems in the U.S. steel industry lessened, and the anti-dumping duties became superfluous, but they had already distorted trade by imposing a much higher tariff level than could have been justified in an open long term negotiation. At the same time, the United States faced retaliation: the EU, Japan and South Korea threatened to impose punitive tariffs of up to 30 percent on U.S. exports. Much better to have an overall low tariff on imports, openly negotiated and in place over the long term. It would be far less economically damaging to all parties than pre-emptive and unilateral impositions of high tariff rates, let alone quota systems.

When designing a free-trade agreement, one must look for areas in which the comparative advantage is large or the barriers being removed are high, because it is in those areas that the gains for free trade are greatest and most likely to outweigh the frictional costs of currency fluctuations. Reducing or removing tariffs that are already low produces only small gains (even proportional to the value of tariffs removed) and may well incur large costs.

Low tariffs may be beneficial in a world of floating exchange rates, but high tariffs, whether imposed overall as in many Third World countries, or in the form of anti-dumping duties as in the United States, are pernicious obstacles to economic optimization and should be outlawed in a well-designed tariff system. The high tariff rates between Third World countries are probably the greatest single economic obstacle that poorer countries face. Their removal by a strong World Trade Organization (WTO) would bring enormous gains--and all achieved entirely without participation by the wealthy West.

Not that tariffs are the worst form of protectionism: subsidies are more damaging because they cause huge distortions between the subsidized item and the rest of the economy, which is taxed rather than subsidized and must pay extra to provide the subsidy. Quotas are also more damaging than tariffs, as they provide a tariff that is effectively infinite at the margin, and hence allow the perpetuation of inefficient production, whatever the cost advantages of imports. Non-tariff barriers such as environmental or labor regulations are also more damaging than tariffs. They are not transparent, they impose costs often far beyond those of a simple tariff, and they are generally discriminatory in favor of Western suppliers and against less well established Third World producers.