April 23, 2015
Glenn Kessler, the Washington Post fact checker, gave four Pinocchios this morning to Ohio Senator Sherrod Brown for for mis-attributing a claim on lost jobs from the trade deficit to George W. Bush. Since I may have played a role in the Pinocchio warranting comments, let me try to clear up some possible confusion on the issue.
At the most basic level there are two different ways to view trade based on two different views of the overall economy. The conventional view is that trade affects the allocation of output (i.e. we produce more of some goods and services and less of others) but has little impact on the overall level of output. This is because the economy is assumed to be at or near full employment. The other view is that trade can have a large impact on employment and output because the economy is often not near full employment. In this case, the size of the trade deficit can make a big difference.
Until the recent downturn, the vast majority of economists held the first view. There were a small number of progressive economists who believed that the economy was often far from full employment. However most economists, including liberals like Alan Blinder and my friend Paul Krugman, believed that the economy was generally near its full employment level of output.
From this perspective, trade does not directly create jobs. Trade can increase efficiency, which means that wages may rise on average. This can lead more workers to decide to work. However there are three important qualifications on this point.
First, the effect is likely to be small especially with the trade deals now being considered. For example, the analysis of the Trans-Atlantic Trade and Investment Pact (TTIP) by the Centre for Economic Policy Research in London (no connection to my CEPR) shows the deal will increase GDP by 0.4 percentage points when its effects are fully felt a dozen years in the future. If the gains are felt evenly, this means that a worker who would otherwise earn $20 an hour in 2027, will instead be able to earn $20.08 an hour. As a result, we might expect some increase in the number of people willing to work, given that they can make 8 cents (0.4 percent) more for each hour of work. However this effect is likely to be small.
The second point is that we are not really reducing unemployment in this story. The worker who decides to work for $20.08 an hour always had the option to work for $20.00 an hour, they just decided that for $20.00 an hour they would rather sit home and watch TV. In other words, we are seeing a reduction in voluntary unemployment, not involuntary unemployment, which is what concerns most of us.
The third point is that gains surely will not be felt evenly. There is now considerable research showing that many workers have seen large reductions in their wages as a result of the trade openings of the last two decades. Given the unevenness of the gains and losses, it is not possible to say apriori whether a trade agreement will be a net job loser or gainer in this full employment story. (It is also worth noting that the projections produced by the London CEPR do not factor in any losses that would be associated with higher prices for prescription drugs and other products that will gain stronger patent or copyright protection. It is entirely possible that the distortions from the stronger protections that will be part of the TTIP will cause this deal, as well as the TPP, to be net losers even before considering the distributional effects.)
Anyhow, the first view of trade is not consistent with any story of large job gains. Kessler rightly whipped out the Pinocchios on the Obama administration for using a study from this first perspective to claim large job gains from the TPP.
The second view of trade does mean that there can in principle be large job gains or losses from trade, but this stems from the change in the trade balance. The story here is that the economy is typically below full employment, meaning that if we had more demand in the economy, we would have more employment. The basic equation for aggregate demand in the economy is:
GDP = C + I +G +(X-M)
This means that total demand is equal to consumption demand, plus investment demand, plus government spending, plus net exports, with the last term being exports minus imports. It is very important to have this “minus” in the picture. If we increase exports, but see an even larger increase in imports, then we have not increased demand in the economy through trade. This is especially clear when the exports are directly linked to the imports, for example when we export car parts to Mexico to be assembled into a car that is imported back into the United States.
There has been a tendency for many advocates of trade deals to do accounting where they only consider the export side of the equation. For example, 13 former Democratic governors recently signed a letter in support of fast-track trade authority that referred to exports:
“adding $760 billion to our economy between 2009 and 2014.”
It associated this increase in exports with the creation of 1.8 million jobs. The letter ignored the fact that imports increased by $890 billion over the same period. This means the change in trade over this five year period on net reduced demand in the economy. There was no remotely honest way the governors could get their 1.8 million jobs story. It was a pure fabrication. (My comment on the governors’ letter is here.)
This issue of net exports gets to the heart of the disagreement between Kessler and Brown. Brown used a figure from the Commerce Department on the number of jobs created per billion dollars of exports and applied it to the trade deficit. Kessler rightly points out that Brown has been using a 1993 number, which is not adjusted for inflation and wage growth over the last two decades. In other words, $1 billion of exports would imply far fewer workers today than in 1993.
However on the more important point of applying the Commerce Department number to the trade balance, Brown is very much on the mark. The Commerce Department’s figure was always a rough number. Not all exports will create same number of jobs. For example, a $1 billion increase in exports of car parts or airplanes will lead to many more jobs than a $1 billion increase in Microsoft or Merck’s licensing fees. The Commerce Department was presenting a number that attempted to average over all types of exports.
If we apply the Commerce Department number to imports, it will also not be exactly right, but it should be a ballpark figure. In fact, because wages tend to somewhat higher in export industries than in import competing industries, there would likely be more jobs lost per billion dollars of imports than are gained per billion dollars of exports.
But that would be very much a second order effect. On the basic story, if we assume that the economy is below full employment, then the sort of calculation that Brown did using the change in net exports was entirely kosher. Furthermore, from the perspective of an economy that is below full employment, the impact of any increase in the trade deficit that results from a trade deal is likely to swamp any gains from the greater efficiency that might result from reductions in tariffs or other trade barriers.
For this reason, Brown and others who have raised the issue of currency rules in the TPP are very much on the right track. If we continue to run large trade deficits (currently around $500 billion a year), it will be extremely difficult to make up this demand from domestic sources. As a result the labor market would remain weak, and wages would likely continue to stagnate.
If we believe that the economy is below full employment, then we should be asking about net exports. Exports taken in isolation tell us nothing.
Correction: It has been called to my attention that Kessler’s point was that Brown was wrong to attribute his claim to Bush, since Bush never talked about trade deficits, only exports.
Comments