With last week's passage of three trade agreements (Colombia, Panama, and South Korea), and spurred by Suzy Khimm's article on the Korea trade deal, I was reminded of the storm of numbers proponents of NAFTA threw out there in the run-up to its approval.
Gary Hufbauer and Jeffrey Schott, described by the New York Times (2/22/93; no link but available on Lexis-Nexis) as "the two most influential academic experts on the North American Free Trade Agreement," wrote a book entitled NAFTA: An Assessment in 1992, with a second edition in 1993 (all references below are to this edition). They predicted that the U.S. would gain 170,000 jobs by 1995.
This work struck me as flawed for a number of reasons. In the first place, it was odd that two economists would write about the winners and losers from NAFTA without making any reference to the economic theory on the subject, such as the Stolper-Samuelson Theorem (see below). Second, the authors did computer simulations of the effect of the agreement that said in their long-run scenario (Table 2.1, p. 16) that we would gain low-skill jobs and lose high-skill jobs, pretty much the opposite of what you would expect based on economic theory. I questioned what assumptions would be necessary to get that result. Finally, the book assumed that the U.S. would have a merchandise trade surplus with Mexico, "$7 billion to $9 billion annually through the 1990s, and perhaps $9 billion to $12 billion annually in the following decade" (p. 15) They held this view even though the authors showed on on page 4 that the Mexican peso was overvalued. If the value of the peso were to fall, Mexican goods would become cheaper in the U.S., while American products would be more expensive in Mexico, completely upending the likelihood of a U.S. trade surplus. As we know, the peso fell sharply in the December 1994 "tequila crisis," and the U.S. has had a trade deficit with Mexico ever since. See the table for details.
U.S. Goods Trade Balance with Mexico ($billions)
Jan-Aug 2011 -44.8
Source: Census Bureau
So much for that prediction. According to the then-fashionable claim that $1 billion in net trade surpluses created 19,600 jobs, the fact that the trade deficit worsened from $1.9 billion in 1990 to $74.8 billion in 2007 ($72.9 billion deterioration) implies a loss of 1.4 million jobs at its worst point. This number should be taken with a grain of salt, but it is in one sense what we expect. Still, it's only about 1% of the labor force, though we could certainly use the jobs now. But the effect on labor may have been worse, because the expanded option for companies to relocate to Mexico made it possible for them to threaten their workers with job loss and thereby hold down their wages. Kate Bronfenbrenner of Cornell's Institute for Labor Relations found that shutdowns became much more frequent after NAFTA (15% vs. 5% in the late 1980s) during labor organizing and contract campaigns.
At the same time, a contradictory prediction also failed to hold up. Wolfgang Stolper and Paul Samuelson argued that the expansion of trade was good for a country's abundant factors of production (they are land, labor, and capital) because of added markets abroad, but bad for scarce factors of production because of competition from abroad. Here, "bad" means reduction in real (inflation-adjusted) income. The U.S., which is not densely populated by world standards, is thus labor scarce but abundant in land and capital. My expectation was that NAFTA would therefore lead to lower real income for labor. In fact, however, weekly earnings for nonsupervisory workers in the private sector, which fell from a peak of $341.83 (constant 1982-84 dollars) in 1972 to a low of $266.46 in 1992, have risen ever since, despite the NAFTA and WTO agreements, hitting $284.79 in 2000 and $297.31 in 2010, according to the Economic Report of the President 2011, Table B-47. Of course, we are still $44.52 1982-84 dollars below the peak, or about $95.25 per week in 2010 dollars.
While obviously we see plenty of economic misery today, I'm curious why the trend on this measure, which Bill Clinton brilliantly politicized with "It's the Economy, Stupid," has reversed course, and how it relates to other measures of labor income. Perhaps there is a better measure of wages out there, or perhaps inflation is currently underestimated (there was a huge campaign in the mid-1990s claiming it was overestimated; maybe the Bureau of Labor Statistics paid too much attention to critics), which would overstate real income. Let me know what you think in the comments.
As for the three new trade deals, they are being sold with squishy job projections, just like NAFTA was. While Khimm's interviewees are probably right that the deals' effect on labor will be small, I still think we should go with the Stolper-Samuelson expectation that the effect will be negative.