One way to keep America’s global edge is to strengthen the infrastructure surrounding government statistics. Without accurate data, many economic policy decisions are being made in the dark—a crime in the Information Age. Specifically, there’s evidence that we are miscomputing many key figures regarding trade and productivity, among others, which doesn’t give us a clear picture of our competitiveness.
Supply chains in the 21st century make classic “wine-for-cloth,” the two-country trade patterns described by classic economists Adam Smith or David Ricardo, seem quaint. Complex goods like cars and computers are made up of commodities from many countries, internal components from others, and finally assembled in another. Unfortunately, our statistics haven’t kept up. Studies involving iPhone production found that China imported around $172 of components from Europe, Japan, South Korea and other countries, assembled them, and then exported a completed unit to the US for approximately $180, then sold at retail prices. That means the Chinese “value added” is really only about $8 of labor per unit. But official US statistics register the full $180 import from China, thereby distorting our trade picture. Moreover, some assembling countries often “double count” these imported components and exported finished goods, often making their economies look more active than they really are. We’re not capturing these important trade and labor subtleties in our official statistics.
In another modern supply chain problem, US labor productivity figures are obscured. For example, if Cadillac lowers a widget cost in its Escalade, this is captured as a statistical US labor productivity gain—but in reality, we can’t be too sure. How did the widget become cheaper? Maybe Cadillac found an Ohio company to manufacture it for less. If so, it may be a US productivity gain. Oddly, if Cadillac sourced it in Mexico versus Ohio, that also would be registered as a US productivity gain—but it’s not a true increase in American labor productivity, is it? What happens if Cadillac then switches from Mexico and finds cheaper widgets in China? That registers another US productivity gain because input costs go down again. And what if that Chinese widget company actually gets more “productive,” and can further lower its cost to Cadillac? That’s a US productivity gain, too—but in reality, it’s China’s improvement that oddly gets tangled into our official statistics. As some have suggested, we know less about American productivity than our numbers would have us believe.
Research professor Andrew Reamer of George Washington University likens the federal data system (pdf) to:
…a large black box in a dark shadow. We know a few high-profile stats shed light on how we’re doing economically, such as GDP and unemployment, but most everything else is opaque. We don’t quite understand what else the system contributes to economic policy or, to be honest, how it works.
In the US, these numbers are compiled largely by three agencies: the Census Bureau, Bureau of Economic Analysis, and Bureau of Labor Statistics. Together, we give them a measly $1.6 billion of Washington’s $3.7 trillion budget to study our complicated $15 trillion economy. So is American labor truly productive? And with which countries do we really run trade surpluses and deficits? Given our antiquated, pre-globalization methodology and the agencies’ small resources, it’s tough to say.
Reamer recommends increasing funds to government number crunchers by $300 million to better capture 21st economic activity, but we should probably spend even more. Good data is not only important to business, but also to government for crafting effective public policies to stay competitive. In a country that has produced some of the greatest icons of the Information Age—Microsoft, IBM, Apple, Google, Bloomberg—it’s a shame we can’t better capture what’s going on in the world and our own economy.