Yesterday the Washington Post’s WonkBlog engaged in some undeserved albeit tepid NAFTA cheerleading by profiling a new NBER study estimating that the deal has brought a tiny increase in real wages. And I mean tiny—the authors use a mathematical model to project that the deal led to a 0.17% increase in real U.S. wages over a twelve-year period (1993 to 2005). That means that the average U.S. worker, who was earning $22,000 in 1993, can thank NAFTA’s first twelve years for a real wage increase amounting to a whopping $37. That’s right—according to the NAFTA-friendly study’s own theorized calculations, each worker should consider the annual worth of NAFTA to be close to that of a box of Twinkies: about three dollars.
Does such a conclusion merit a blog post entitled “Study: NAFTA Raised Pay Here and Abroad?” Probably not. But that’s not the most concerning element of the WonkBlog piece. The bigger elephant in the post is that “raised pay” from NAFTA is simply not the reality for the majority. Neither the study nor the blog post address the real story of NAFTA and wages: that the deal has fueled a regressive redistribution of income from average workers to the wealthy.
Trade economists widely acknowledge that any U.S. income increases resulting from NAFTA-style trade deals will tend to disproportionately favor the wealthy while real wage reductions are likely to be the result for the rest of us. In standard trade theory, the Stolper-Samuelson effect predicts that open trade will create increased demand for U.S. capital-intensive goods and reduced demand for U.S. labor-intensive goods, thereby increasing income for capital owners (i.e. the wealthy) while reducing wages for workers. Under NAFTA, this regressive impact has moved from theory to reality. One study by the Economic Policy Institute estimated that even after taking into account consumer savings from cheaper imports, a U.S. household with two median wage earners was losing $1,000 of earnings each year to NAFTA-style trade by 1995, increasing to a $2,000 annual loss by 2006. The median U.S. worker probably finds little comfort in a new study that averages out gains at the top to transform her real $1,000 trade-related loss into a hypothetical $3 gain.
The regressive influence of NAFTA-style trade on income distribution may help explain why income inequality in the NAFTA era has reached historical heights, now apparently surpassing even the inequality levels of 1774. Median real incomes in the U.S. have been falling for the last decade, while the income of the richest 1% has been continually climbing. Workers’ productivity has been steadily rising while labor’s share of income has been steadily falling. Why are workers getting paid less while doing more? Economists from institutions ranging from the Economic Policy Institute to the Federal Reserve have named NAFTA-style trade as a key answer (“increased globalization and trade openness,” in the words of Federal Reserve economists). So the income-related takeaway from NAFTA is not the deal’s miniscule impact on aggregate wages (whether negative or positive), but its large impact on income inequality.
But even if we were principally concerned with the miniscule aggregate impact, the study featured in the WonkBlog omits several key factors, calling into question the small positive impact it cites. First, the study only aims to estimate the impacts of NAFTA’s tariff reductions, which were a focus of only 6 of the deal’s 22 chapters. The authors make explicit the limitations of disregarding the brunt of the agreement’s content, stating, “Unquestionably, NAFTA had more provisions than only reducing tariff between members and by no means our results should be interpreted as the trade and welfare effects of the entire agreement” (pg. 27). Non-tariff NAFTA provisions that could impact real wages include those found in the deal’s intellectual property chapter, as Dean Baker notes over at the Center for Economic and Policy Research. The chapter grants pharmaceutical firms and other corporations anti-competitive patent extensions, which tend to inflate the cost of medicines and other patented products, thereby eroding consumers’ real wages.
In addition, the wage increase cited by the WonkBlog ignores another side of the tariff reduction coin: loss of tariff revenue for all three NAFTA governments. The study itself notes that this loss in fiscal revenue actually outweighed the purported real wage gains for Mexico and Canada. Taking into account reduced tariff income, the authors conclude that the net income effect of NAFTA’s first twelve years is a 0.1% loss for both Canada and Mexico (pg. 25). In the United States, the lost tariff revenue reduces the Lilliputian wage impact even further, yielding a purported 0.1% net increase to aggregate income. At that rate, the hypothetical average U.S. worker did not see a gain of $37 under a dozen years of NAFTA, but just $22—less than an average tank of gas. Meanwhile the actual median worker continues to lose at least $1,000 each year under NAFTA-style trade. One of these facts seems worthy of a blog post. One does not.