On the 24th, at National Gypsum Company in Ohio, Obama said (Remarks for Senator Barack Obama: National Gypsum in Lorain, Ohio),
One million jobs have been lost because of NAFTA, including nearly 50,000 jobs here in Ohio.
Where did these numbers come from, and how significant are they as a proportion of jobs in Ohio?
Obama doesn't explain whether he was talking about gross job loss (which would have been offset by job gains elsewhere in the economy, but would still have involved temporary unemployment and income loss for those who lost jobs) or whether this is meant to be an estimate of net job loss - we have a million fewer jobs (even after accounting for job gains) than we would have had without NAFTA. From the context I think he meant net job loss.
These numbers appear to be estimates produced by the Economic Policy Institute (EPI). This EPI leaflet reports the figures: NAFTA & the U.S. The leaflet itself is based on Revisiting NAFTA. Still not working for North America's workers (Robert E. Scott, Carlos Salas, and Bruce Campbell; Introduction by Jeff Faux, EPI Briefing Paper #173, September 28, 2006).
Its clear from the EPI materials that the estimate is an estimate for net job loss. The totals include direct and indirect job losses (so the factory job is lost, as well as the jobs that depended indirectly on that factory job). The EPI method accounts for goods but not services producing industries.
The approach used by the EPI is described in an appendix to the paper and in other EPI reports. Basically, the EPI draws on Bureau of Labor Statistics estimates of the employment required to produce a dollar of goods in different industries. The EPI then looks at net value of imports in each industry (in this case with respect to Mexico or Canada) and multiplies the net value for each industry by the estimate of employment per dollar of output from the BLS.
Call this the "jobs embodied in net imports" approach following the lead of three economists from the New York FRB, who use a very similar approach: U.S. Jobs Gained and Lost through Trade: A Net Measure (Erica L. Groshen, Bart Hobijn, and Margaret M. McConnell , August 2005).
The EPI uses the method to estimate the total U.S. job loss, and then allocates that job loss among the states using BLS data on industry employment by state.
The 50,000 jobs lost in Ohio include direct and indirect job losses. The estimate was made for 2004. The BLS estimate of total non-farm seasonally adjusted employment for Ohio in December 2004 was about 5,400,000. So if NAFTA has actually cost 50,000 jobs in Ohio, then that is a loss of about nine-tenths of a percent.
The EPI approach assumes that the entire value of Ohio's trade deficit with Canada and Mexico is due to NAFTA. Ohio's trade deficit with Canada and Mexico is a component of the U.S. trade deficit, which is driven by many things. This approach ignores everything else (low savings rates in this country and high savings rates abroad for example) and attributes the whole thing to the trade agreement. Thus, on its own terms, this approach overstates the job loss that would be due to NAFTA.
Here's Clinton using numbers from the same source. In a speech last fall, she said,
The Economic Policy Institute estimates that we've lost 1.8 million jobs to China...
Rebuilding the Middle Class: Hillary Clinton's Economic Blueprint for the 21st Century (October 8, 2007). The numbers come from Costly Trade With China. Millions of U.S. jobs displaced with net job loss in every state. (Robert E. Scott, originally released on May 2, 2007 and revised on October 9, 2007).
I'm not the only one who's been skeptical about the EPI's NAFTA job loss estimates. Gary Hufbauer and Jeffrey Schott of the Peterson Institute criticized an earlier version of the EPI estimates in their NAFTA Revisited (Peterson Institute, 2005 - see chapter 2 on Labor). Here's the EPI piece they're critiquing: NAFTA's Hidden Costs. Trade agreement results in job losses, growing inequality, and wage suppression for the United States (Robert Scott, 2001).
Here's what they say:
The most extreme estimate of job losses from NAFTA is 879,280 actual and potential jobs lost between 1994 and 2000, according to Robert E. Scott (2001) of the Economic Policy Institute. Scott's estimate is based on his calculations of how many more jobs there would be if the US trade deficit with Canada and Mexico were the same in 2002 as it was in 1993, adjusting for inflation.
Blaming NAFTA for 100 percent of the growth in the US trade deficit with Canada and Mexico ignores the macroeconomic determinants of these two bilateral trade deficits. The growth in the US trade deficit with Canada and Mexico is in line with, but slightly lower in percentage terms than, the growth in the total US trade deficit.
Even assuming Scott's estimates were plausible, over half of the alleged job loss comes from the growth in the US trade deficit with Canada, which competes with high-valued US products. Scott concedes that the US economy created 20.7 million jobs between 1992 and 1999, or 27 times the number of jobs allegedly lost due to NAFTA. The estimated 879,280 jobs lost over seven years due to NAFTA is less than 15.2 million US workers displaced during seven years.
Minor edit (2004 BLS jobs est. not preliminary) Fed 29; addition of text from Hufbauer and Schott on March 1; Clinton text on March 3.
There are more posts on the 2008 election here: Election 2008 or over here at the International Economic Law and Policy Blog.
David Leonhardt noted in his New York Times column (The Politics of Trade in Ohio , Feb 27), "Ohio’s troubles haven’t really been caused by trade agreements..."
Comments