The economics profession, or well over 90 percent of it according to polls, continues to support free trade. Above all, most economists remain stuck in a cheery "win-win" fantasy of how trade works and are unable to see the brutally adversarial dynamics of trade in the real world.
The basic justification for their delusion, of course, is David Ricardo's venerable 1817 theory of comparative advantage. However, economists do not consider free trade justified today simply on the strength of the original 1817 theory alone. Ricardo's ideas have been considerably elaborated since then, and they generally use sophisticated "computable general equilibrium" (CGE) computer models, built upon his work as the foundation, to assign actual dollar amounts to the purported benefits of free trade.
As a result, it is well worth looking at problems with these models a bit in order to understand why economists remain so mistaken.
These models are called "computable" because, unlike economic models that exist purely to prove theoretical points, it is possible to feed actual numbers into them and get numbers out the other end. They are called "general equilibrium" because they are based on the fundamental idea of free market economics: that the economy consists of a huge number of separate equilibria between supply and demand and that all these markets clear, or match supply with demand, at once.
The most obvious problem with these CGE models is that they often make rather implausible assumptions.
For example, they often assume that government budget deficits and surpluses will not change due to the impact of trade, but will remain fixed at whatever they were in the starting year of the model. Hmm...
Worse, they assume that trade deficits or surpluses will be similarly stable, with exchange rates fluctuating to keep them constant. Yeah, right.
And they assume that a nation's investment rate will equal its savings rate: every dollar saved will flow neatly into some productive investment. O-kay.
These assumptions are understandable, as devices to simplify the models enough to make them workable. They are, however, both clearly untrue and serious objects of controversy in their own right.
That investment will equal savings is basically a form of Say's Law, "supply creates its own demand," named after the French economist Jean-Baptiste Say (1767-1832). This basically makes both underinvestment and unemployment theoretically impossible. (This is a recurring problem in free trade economics: ideas long discarded in other areas of economics recur with alarming regularity.)
Furthermore, these models often assume that nations enjoy magical macroeconomic stability: the business cycle has been mysteriously abolished. I wish.
And, of course, their financial systems enjoy unruffled tranquility, without booms, busts, or bubbles. I want to move to this country!
Many of these assumptions are pre-Keynesian, and are thus at least 70 years behind mainstream domestic economics. That is to say, they are innocent of the thinking of John Maynard Keynes (1883-1946), the great British economist who revolutionized economics by explaining why economies do not naturally reach an equilibrium of full employment (and thus why deficit spending can help economies climb out of recessions.)
These models also generally leave out transition costs. These sound temporary, but such transitions can take decades. Consider the pain experienced by the Midwestern manufacturing areas of the U.S. as their industries have gradually lost comparative advantage since the mid-sixties.
Given that the world economy is not static, but constantly moving into new industries, there are always new transitions being generated, which means that transition costs go on forever as an intrinsic cost of having a global economy based on shifting patterns of comparative advantage. Somebody will always be the rustbelt. This does not of itself mean that economic change is a bad thing, but it does mean that these costs must be factored in to get an accurate accounting.
Arbitrary accounting for trade in services (AKA offshoring) is another big flaw in CGE models.
The root problem here is that this trade usually isn't regulated the same way as trade in goods. Due to the fact that, prior to cheap long-distance telephony and the Internet, many services were rarely internationally traded, there are actually few outright tariffs or quotas on them. Instead, there is a crazy-quilt of hard-to-quantify barriers, ranging from licensing requirements to tacit local cartels and linguistic differences.
As a result, when these barriers come down, they rarely come down in a neatly quantifiable way like reducing a tariff on cloth from 28 to 22 percent. So economists must, to put it bluntly, guess how to quantify nonquantitative changes in order to model them. (The standard term for this is "tariff equivalent" numbers.) As a result, the conclusions generated by many CGE models of trade in services are so dependent upon arbitrary guesses as to border on arbitrary themselves.
Another caveat: because these CGE models are predictions about the future, they are of necessity somewhat speculative under the best of circumstances and are notoriously susceptible to deliberate manipulation.
It is easy, for example, to generate inflated predictions of gains from trade by extrapolating calculations intended to apply only within certain limits with back-of-the-envelope calculations that go far beyond these limits. (These are known in the trade as "hockey stick" projections due to their shape when graphed.) As a result, as Frank Ackerman of the Global Development and Environment Institute at Tufts University puts it:
The larger estimates still being reported from some studies reflect speculative extensions of standard models, and/or very simple, separate estimates of additional benefit categories, not the core results of established modeling methodologies. ( "The Shrinking Gains From Trade: A Critical Assessment of Doha Round Projections," 2005.)
Similarly, the standard way for free traders to play down the damage done to the victims of free trade is to count only workers directly displaced from jobs as its losers. Unfortunately, these workers crowd into the labor market of everyone else with similar education and skills, dragging down wages for other people, too.
Even if all statistical gamesmanship is removed and other reforms made, there is a deeper problem with CGE models: no such model can predict what choices of trade strategy a nation will make.
For example, none of the models used in the 1950s predicted Japan's ascent to economic superpower status. Quite probably, no model could have. Indeed, no model based upon purely free-market assumptions will ever readily predict the outcomes from such strategic choices, as free-market economics, with its insistence that it is always best to just do what the free market says, rules out a priori the possibility that most such deliberate economic strategies can even work.