In the quest to defend more protectionist policies, Commerce Secretary Wilbur Ross is resorting to an old trick with large scale economic models. These models routinely show negligible economy-wide effects of nearly any trade policy change, such as trade liberalization, a free trade agreement, or the imposition of new tariffs.
Economists know that the sign (positive or negative) is usually reliable, and that results in particular sectors are extremely valuable in tracing the effects of a policy change.
But because the economy-wide effects are routinely tiny, the overall effects are not usually a useful tool to promote or defend a policy move, particularly one that is known to be a drag on economic growth such as import taxes (e.g. tariffs).
To use the finding of ‘negligible effects on the overall economy’ to defend new import taxes on steel and aluminum, which nearly all US manufacturers use and hence boils down to a manufacturing tax, is misleading.
Large scale computable general equilibrium (CGE) models are used widely in economic policy circles to aid the analysis of trade and other policy changes. A CGE model uses actual economic data to estimate how an economy might react to changes in policy, technology, or some other external factor. With a large scale representation of the US economy and its trading relationships, a great power of these models lies in their ability to trace the effects of a policy change industry by industry, capturing the input-output relationship across sectors. The model reports on estimated changes in employment, output, imports, exports, and other variables.
But, like any tool, their value is in how they are used.
Resorting to an old trick
At least one Department of Commerce (DOC) official has gone on record to say, “DOC modeling has not shown any substantial impact on the overall economy as a result of the proposed steel tariffs…This is consistent with the finding of the International Trade Commission that the Section 201 (in 2002) had negligible effects on the overall economy.”
Secretary Ross appears to be leaning on the ‘negligible effects’ on the overall economy as a justification to impose import taxes on steel and aluminum, but these models consistently show negligible effects on the overall economy of just about any trade policy shock.
Imagine a large bridge collapses in Boston. The impact for the local economy will be felt widely. The impact for the economy as a whole, however, will be negligible (e.g., likely small but negative).
Does that mean that we should be dismissive of the large negative impact of the bridge collapse in Boston, or that we should dismiss the overall negative economic impact just because it is small? Of course not.
Yet this is exactly the kind of argument that Ross and his team are trying to make when they defend steel tariffs by saying that the impact on the overall economy will be negligible. It’s bad enough that for the sake of a few steel producers, Secretary Ross is comfortable hurting the overall economy with an import tax, even if just a little. But it is shocking that he would dismiss the plea of workers in the downstream industries that would be dramatically and adversely affected. Don’t they count, too?
CGE models routinely show negligible effects of trade policy shocks
In 1993, the International Trade Commission (ITC) found the probable effects of the North American Free Trade Agreement (NAFTA) to be less than one percent on US employment, GDP, and wages.
In 1999, on China’s World Trade Organization accession, the ITC did not report GDP effects, but noted that at the time, total US-China trade was less than one percent of US GDP, and hence any shock by that model and that approach would get similarly small effects.
In 2016, the ITC estimated the economic effects of all U.S. trade agreements implemented since 1984. They found the economic effects of all of these agreements was an increase in real GDP by 0.2 percent.
In 2017, they found that removing the significant import restraints in the United States results in an economic benefit of less than one percent (less than 0.05 percent).
There is a long list of these reports.
A big shock to a small part of the economy will tend to have a small overall effect. Trade with China is three percent of US GDP, and steel and aluminum are just a fraction of that. So even a big tariff would be expected to have a small economy-wide impact.
A better way
But there are ways to use the models productively. The power of CGE models lies in their sectoral level results, and their ability to help trace out the economic effects across sectors of the economy. The sign (positive or negative) is usually correct and helpful. A more genuine reporting would include the full sectoral results at the most detailed level possible.
Suppose the model would show negative effects on downstream steel using sectors (as it surely does), and negative, but small, overall economy wide effects. In this case, in order to justify the steel tariffs, Ross would have to place a higher weight on factors outside the model in order to justify the import taxes. That is his prerogative, but if this is what he is doing, then he should be honest about it: he is recommending that the president forego economic growth and jobs across the economy, with most job losses in manufacturing, in order to boost profits in a particular sector.
History is not kind to those who ignore the evidence
The Bush era steel tariffs did not change China’s behavior. We are in exactly the same place today with China as we were then. Further, we now know that thousands of people in steel-consuming sectors lost their jobs after the Bush-era steel tariffs. As the price of steel rose, more people lost their jobs in steel consuming sectors than existed in the entire steel industry.
This is not a criticism of the models, but rather a statement about how administration officials are using these models to justify their positions to impose new import taxes on manufacturing.
With negligible effects, you can justify just about anything.