Automation and trade
By Scott Sumner
I’ve been arguing that the problems often attributed to trade are more likely due to automation. Indeed in a sense both trends are two sides of the same coin — creative destruction. It might help to make this point with a simple example.
Suppose you have an air conditioner factory in Indiana, which employs 2000 workers. Now consider two possible corporate decisions:
1. Move 1000 of the 2000 jobs to Mexico, and import much of the resulting production.
2. Automate the factory so that the same output can be produced with 1000 workers instead of 2000.
In practice, both of those hypotheticals occur all the time, although the automation example is far more important—indeed it’s the only one happening at the global level, as obviously Earth doesn’t trade with other planets.
Now let’s think of secondary effects, starting with the export of jobs and import of AC units. An increase in American imports tend to reduce the value of the dollar, which boosts exports. Will the rise in exports replace the jobs lost to imports? Probably not. Partly because trade may not be “balanced” (although if measured properly the US actually does have roughly balanced trade.) And partly because a given value of exports may require fewer workers than that same value of imports.
If 1000 jobs are shipped to Mexico, then we might end up with an extra 200 jobs in exporting high tech goods like Apple phones, Facebook ads and Boeing airplanes, and another 800 jobs in areas like home building, restaurants and health care. Critics of trade will point out that the 1000 new jobs might not go to the very same 1000 workers who lost their jobs at the AC plant, and that concern is legitimate. But it holds equally well for automation.
If 1000 jobs are automated, there will be some jobs created in building robots, and other means of automation. But almost certainly not 1000 jobs, otherwise there would be no cost saving for automation. What would be the point? More likely, you might get another 200 jobs building robots and other forms of automation, and another 800 jobs in areas like homebuilding, restaurants, and health care. The effects are essentially identical to trade.
You might argue that the extra 800 jobs in the non-traded sector assumes proper monetary policy. I think it’s more sensible to think in terms of assuming non-perverse monetary policy, but in any case, the assumption applies equal well for trade and automation.
If you are opposed to free trade but not opposed to automation, you probably wrongly think trade results in a net loss of jobs and automation does not. It doesn’t. And almost no one seems to oppose automation. I conclude that arguments against trade are bogus, based on ignorance. Trump and his apologists in the media have simply not thought through the issues.
Trade and automation are largely responsible in the rising living standards for average Americans over the past 50 years. (And yes they are rising; another myth is that they are not.) As I drive down the highway, 95% of the cars I pass are luxury cars, or more precisely cars that would have been viewed as luxury cars in 1966. Automation has made that possible. In another 50 years, lower middle class Americans will drive cars that are better than BMW 700 series cars, or perhaps I should say “riding in” cars that good–I have no idea if they’ll be driving them.