In an old joke, a shipwrecked economist is asked for his counsel on how the stranded group can be rescued. “Assume we have a boat,” he begins.
Robert Lawrence and Tyler Moran, writing for the Peterson Institute for International Economics, seem to have missed the joke in their recent repeat of the same flawed assumptions of their colleagues’ hugely optimistic assessment of the Trans-Pacific Partnership (TPP) Agreement which prompted our own paper, “Trading Down: Unemployment, Inequality, and Other Risks of the Trans-Pacific Partnership.”
Claiming to address contrarian findings that the TPP may well cause job losses and increase income inequality, Lawrence and Moran assume away the causes – downward pressure on wages and employment due to the consequent “race to the bottom” – which have made free trade agreements so controversial.
Assume we create jobs
To recap, in January, the Peterson Institute published new TPP estimates, updates by Peter Petri and Michael Plummer of an earlier 2012 paper. The update reiterated their claim of significant income gains from the agreement, 0.5% for the United States after fifteen years, with minimal job displacement, and with new jobs in growing industries absorbing displaced workers in declining activities.
In “Trading Down”, we pointed out that the study was flawed because it assumed full employment and unchanged national trade and fiscal balances, among other things. We applied the United Nations macroeconomic Global Policy Model to their estimated trade impacts from the TPP dropping the full employment assumption.
Even without adjusting for the assumption of fixed trade balances, we found that if one does not assume away job losses, there will be some permanent job loss, there will be downward pressure on wages, and economic growth will be slowed by the consequent decline in aggregate demand.
Congressman Sander Levin (D-MI) highlighted the problems with the kind of modeling the Peterson Institute offered, calling on the International Trade Commission, in its TPP assessment for the U.S. government due in May, to stop using models that assumed away the problems. As Inside U.S. Trade reported, the new paper is the Institute’s attempt to respond to that criticism:
“Levin in February at a U.S. International Trade Commission (ITC) hearing on the economic impact of TPP argued that its analysis must include an examination of how TPP will affect wages and income inequality; a review of whether the ITC’s economic model should assume full employment; and an analysis of who will experience gains or losses as a result of TPP and other factors. Lawrence said that his and Moran’s paper aimed to answer Levin’s demands for a more holistic analysis of TPP.”
Holistic analysis? Or filled with holes?
It does nothing of the sort, offering a misleading analysis instead. Consider:
- The new study is based on the earlier Petri-Plummer model, claiming to take those results to estimate the “adjustment costs” for workers displaced by the agreement. But the same assumption, that the TPP causes no long-term job loss, underlies the analysis. So permanent job loss is excluded by assumption, with all displacement assumed to be temporary.
- Nor do the new findings allow for trade deficits. The authors assume that TPP does not cause long-term trade surpluses or deficits, in fact, that trade itself is not a major determinant of current account balances. This, of course, flies in the face of large and persistent U.S. trade deficits, including with partners such as Korea, with whom the U.S. has seen its bilateral trade deficit nearly double since the Korea-U.S. Trade Agreement took effect four years ago. Again, the Peterson modeling assumes away the possibility of trade deficits and associated job losses.2
- With no trade-deficit-related job losses, Lawrence and Moran only estimate “adjustment costs” for the remaining few displaced workers awaiting new jobs assumed for them, offering three scenarios, each smaller than the previous.
- The first mischaracterizes our paper, suggesting that we assume that no displaced workers get new jobs. We simply do not assume that they are fully absorbed into growing industries. They estimate 1.69 million U.S. workers could be displaced over ten years.
- The second drastically reduces that total to 278,000, by invoking the full-employment assumption that rising demand will generate new jobs and limit job loss. They acknowledge, however, that the displaced workers are nearly all in manufacturing.
- The third reduces this to 238,000 workers who voluntarily leave manufacturing jobs, so the TPP can’t be blamed for that.
- They then apply a formula to estimate the temporary adjustment costs (essentially lost wages) from those “displaced”. They compare these to Petri and Plummer’s reported TPP gains for the United States of $131 billion. The resulting cost-benefit calculation does not report the costs, just the ratios, for the three scenarios. The authors report that for their “most realistic” scenario (#3), with the least displaced jobs, the benefits are 18 times the costs over the 10-year “adjustment period” (2017-26).
- Then, remarkably, when they add in the “post-adjustment years” 2027-2030, the ratio skyrockets to 115:1. Why? Presumably because with the full-employment assumption all displaced workers are, by then, happily employed in their new post-TPP jobs.
- Finally, Lawrence and Moran claim that the TPP will be mildly progressive for U.S. income distribution. Basically, they argue that the assumed income gains will be very much the same for each quintile of U.S. income distribution, with the bottom quintile seeing a percentage increase 0.007 of a percentage point higher than the top quintile. Technically, that is mildly progressive.
- But it certainly does not look that way when one looks as the absolute gains. The bottom 40% sees just $8 billion in income gains, while the top quintile would get $48 billion. That is more in absolute terms than the bottom 80% combined.
- The authors also make the unfounded assumption that U.S. wages will increase at the same rate as productivity, though that has not happened for decades. This misleadingly raises most workers’ incomes in their analysis.
Full-employment models? Abandon ship!
It is not surprising that Lawrence and Moran find that the benefits of the TPP far exceed the adjustment costs. They employed the same study with the same flawed assumptions of full employment and fixed trade balances. With such assumptions, wage and employment losses are written off as temporary adjustment costs on the path back to full employment. These are significantly understated if the TPP results in large and persistent trade deficits, an outcome they assume away.
The resulting cost-benefit calculations are misleading. First, the costs are minimized as outlined above. Second, the benefits are overstated, taking Petri and Plummer’s estimates at face value, with all their flawed growth-boosting assumptions (surge in foreign investment, most growth gains from non-trade measures).
Finally, the gains are simply asserted to be large, when even the recent Petri-Plummer estimates of gains are incredibly small, just 0.5% of GDP for the United States in 2030, i.e. a paltry 0.029% per year on average over 15 years. How small is that? For the bottom 40% of the U.S. income distribution, the gains amount to just $62 per person, in 15 years.
Those concerned that TPP modeling needs to take better account of the real implications of such agreements should not be satisfied with the Peterson Institute’s latest offering. It does little more than reiterate flawed assumptions, which understate costs and overstate benefits, besides misrepresenting them as serious cost-benefit analysis.
Before the U.S. Congress approves the TPP, the public deserves the kind of robust economic analysis that Rep. Levin has called for, that does not assume away employment losses or trade deficits and offers realistic estimates of the TPP’s impacts on wages, employment, and inequality.
 For the United States, we estimated that in 2025 the TPP would generate a 0.5% slowing in economic growth, 448,000 job losses, and rising inequality, as measured by a 1.31% decline in labor’s share of national income.
 It is worth quoting the paper’s own acknowledgment of these assumptions (from p. 3): “For analyzing the long-run impact of the TPP, it is reasonable for Petri and Plummer to assume that the agreement is unlikely to permanently affect the level of employment or the trade balance[…] Assuming normal employment levels is justified not because changes in imports and exports have no impact on employment in the short run-obviously import growth can cause job loss and exports can generate job growth-but rather because the size of the annual impact of the TPP will be smaller than the many other shocks that will occur every year[…] Moreover, over a longer period macroeconomic policies and wage and price adjustments are likely to restore the economy to the same employment level as the baseline.”