Contact The Editor
Articles By This Author
WASHINGTON -- Is the American economy stronger than we think? Perhaps. A persisting puzzle about its recent performance is the stark contrast between growth of jobs (which has been unexpectedly robust) and the growth of the economy's output (which has been unexpectedly weak). How could a struggling economy produce so many jobs? The puzzle would disappear if the economy's output is consistently undercounted.
And it is, say critics. Government statistics miss many gains from the Internet, they argue, because some "free" services such as Facebook, Google and Twitter aren't counted directly in the economy's output, gross domestic product (GDP). This seems to defy common sense, considering the widespread use of these services. (At last count, Facebook had 1.6 billion users worldwide and Twitter had 320 million.) Other economists deny the undercounting argument.
The debate matters. Rapid job creation and slow economic growth translate into poor labor productivity. (Productivity is the buzzword for economic efficiency.) Indeed, growth in business-sector productivity has slumped. From 1995 to 2004, it averaged 3.25 percent annually; since then, the average is 1.5 percent, reports a new study. If the 1995-2004 pace had been maintained, GDP in 2015 would have been $3 trillion higher; that's a sixth of present GDP ($18 trillion).
So, debate rages among economists. What happened to the $3 trillion? Is it there but overlooked because economic statistics haven't caught up with the real world? Or has economic vitality declined? Then the $3 trillion is lost forever and future prospects for higher growth have dimmed.
Now, the new study rejects the undercounting argument. "We find no evidence [of] growing mismeasurement," write economists David Byrne and John Fernald of the Federal Reserve System and Marshall Reinsdorf of the International Monetary Fund in a paper prepared for the Brookings Institution. The debate is admittedly technical, but considering the high stakes, it's worth exploring.
What matters for GDP are final products purchased by consumers and businesses. Consider the steel used to make a car. It doesn't count directly in GDP, because the ultimate product is the car. Counting the steel (called an "intermediate good") separately would be double counting.
How does this affect the Internet?
Some Internet services (examples: downloading music and paying monthly access fees) are traditional consumer products and are so counted. Likewise, some Internet services are intermediate products for other businesses. If Google enables car companies to operate more efficiently, that will show up in the car industry's improved productivity.
The real argument concerns services (examples: most email and Facebook) that are ostensibly free to consumers. It's these free services that are wildly undervalued by GDP, say critics. The new study makes three main responses.
First, many of these Internet uses are private "home production" -- akin to doing housework, cooking a meal or playing soccer with children. These activities may be important and even pleasurable, but they're not part of the "market economy" where goods and services are bought and sold for cash. That's what GDP aims to measure; therefore, these Internet benefits, though significant, are properly excluded from GDP.
Second, though these Internet services aren't added directly to GDP, they are counted indirectly. Much of the Internet is supported by paid advertising; therefore, these Internet services are valued at the cost of this advertising. This is not new. The same logic was (and is) applied to so-called free TV -- over-the-air broadcasting. The costs of all these ads are assumed to be embedded in the prices of final products.
Third, the study concedes that the digital economy is hard to measure, but it doesn't find that errors have increased dramatically over time. Something else must have caused the productivity slump.
Maybe GDP should be revised to reflect the Internet's many daily uses, though this would raise difficult technical issues. How, for instance, would "free" services be valued, if not by advertising? But the economists were answering another question: Has the Internet's created a shadow economy that's increasingly missed by official figures? Their conclusion, no.
We know some things about the productivity slump. It started before the Great Recession; it affects many nations and industries; present productivity gains resemble those of the 1970s and '80s. These conclusions also tend to exonerate the Internet. But to the question of what caused the slump, the honest answer is: We don't know.
Comment by clicking here.
