Arnold Kling writes

The prices of imported intermediate inputs are being handled incorrectly. When you underestimate the rate of price decline, you also under-estimate the real amount being supplied. And because these are imported intermediate goods, the error is not neutral with respect to U.S. output.

Here is the in a simple numerical example.

Suppose in 2007 the US imported \$100 of steel and made \$200 of cars with 4 workers. The US added \$100 worth of value with 4 workers. So each workers productivity is \$25

In 2009 the US imported \$50 of steel and made \$150 of cars with 3 workers. Now the US is tripling the value of its imports. The US has a value-add of \$100 with 3 workers. So each workers productivity is \$33.

A productivity boom of 50%. GDP stays the same with fewer workers. This is part of our current problems.

But wait Mandel says, we are doing it wrong. The actual amount of steel didn’t drop in half. Its just that the price of steel fell. If price had remained constant we would have imported \$75 worth of steel, had \$75 of value add and worker productivity would be stuck at \$25.

Moreover, US GDP value-add would have declined to \$75 so our GDP would have shrank.

This is the wrong way to look at it. It doesn’t matter what the actual raw steel coming in was, it matters what we paid for the steel. Why?

Well because

1. You are still getting more car per worker which makes the workers more valuable and hence more desirable to hire
2. The GDP of the US is still the difference between what we create minus what the foreign imports cost.

Now you might argue that these price drops are temporary and then productivity will decline. That is a fine case. However, to the extent they are permanent – as is probably the case in electronics components – it’s a real gain.

Suppose that instead of this being a price drop in foreign steel it was a price drop in domestic steel from some fancy new steel making process. What difference would it make to US autoworkers? They are still getting the same good at a lower price. They are still able to produce more value per hour of work.

In other words improvements in the terms of trade act like a boost in productivity. Indeed, that’s the whole point of trade.

Now, Noah Smith pointed me to a paper that disputed this, which I haven’t read and may make some important points but this is the standard view and I didn’t pick up on anything in Mandel that disputed this.