By proposing the phantom GDP effect, I am not denying the normal benefits of trade, as believed by most economists. One potential benefit is the ability to shift more of our workers into higher productivity activities (i.e. rather than assembling iPods, we design them). Another benefit is the potential improvement in terms of trade (our import prices fall relative to our export prices).
Phantom GDP, by comparison, is a statistical mirage. When a good or a service is offshored, the phantom GDP effect causes us to understate the magnitude of the terms of trade effect. It therefore causes us to overstate the domestic productivity gains from trade (and whatever other forces are driving productivity growth. Whatever the productivity gains from trade—and they may be very real—phantom GDP makes them look bigger than they are.
Let’s do a simple and ridiculous example. Let’s suppose that up to now, all production of flimflams were done domestically, at $10 a flimflam.
One day Libya shows up and offers to provide all the flimflams that the U.S. needs, for $5 a flimflam. In fact, by fiat, the Libyans commit to supply flimflams at that price forever.
Three things happens. First, all the domestic flimflam factories fold. Second, the former flimflam workers find new jobs, either easily or with great difficulty. This shift to other parts of the economy is part of the advantages of trade.
Third, our imports will become cheaper relatively to our exports and domestically produced goods and services. In fact, that’s the main reason why this deal is good for the U.S—the terms of trade have shifted in our favor. In the aggregate, we are clearly better off.
But how much have the terms of trade shifted in our direction? We need to know how much import prices have dropped. It should be a lot, right? After all, we are paying a lot less for flimflams.
But wait! Flimflams were not imported before. So how can the Bureau of Labor Statistics give us the right decline in imported prices?
They can’t—not with their current procedures. In fact, as I will describe below, the BLS procedures—while eminently sensible in many ways—will consistently underestimate the magnitude of the decline in import prices when a good is first offshored with a big drop in costs.
To understand why, let’s turn to the BLS procedures for calculating import prices, which are described here and in the subsequent linked sections. As the BLS says: “The goal of the International Price Program is to produce valid price indexes that track the price trends for consistent items over time.” And they work very hard at achieving this goal. Imported goods are always changing—changing models, changing names, changing importers. So the BLS goes to an importer, say, and picks out a sample of goods. Then when they go back again, they ask, for each good in that sample, how the import price has changed.
What happens if a new imported good enters the scene? The BLS clearly explains their procedure: “…the historical movement of the index is used to begin the series for the item.” Uh-huh.
So when the Libyan flimflams show up for the first time, what happens? In the largest sense, the BLS assumes that the price change for flimflams—in the period before they showed up—was the price change for all existing imports. The same is true for any subcategory that flimflams might belong to. So if overall prices for existing imports were rising at 1% per year, then by golly, so are the newly imported flimflams.
So the BLS has no way of picking up the fact that the imported flimflams were half the price of the domestic ones. That fact simply doesn’t show up on their radar screens. And since imported flimflam prices stay at $5 forever, they never contribute a price decline.
As a result, the import price drop from offshoring is undestimated, and so is the magnitude of the terms of trade effect. So the productivity gains from trade and other forces are overestimated.
Are we having a good time yet?