Think the Internet Leads to Growth? Think Again

Photograph by Sara Cwynar for Bloomberg Businessweek

Remember the year 2000 in the months after the Y2K bug had been crushed, when all appeared smooth sailing in the global economy? When the miracle of finding information online was so novel that The Onion ran an article, “Area Man Consults Internet Whenever Possible?” It was a time of confident predictions of an ongoing economic and political renaissance powered by information technology. Jack Welch—then the lauded chief executive officer of General Electric—had suggested the Internet was “the single most important event in the U.S. economy since the Industrial Revolution.” The Group of Eight highly industrialized nations—at that point still relevant—met in Okinawa in 2000 and declared, “IT is fast becoming a vital engine of growth for the world economy. … Enormous opportunities are there to be seized by us all.” In a 2000 report, then-President Bill Clinton’s Council of Economic Advisers suggested (PDF), “Many economists now posit that we are entering a new, digital economy that could inaugurate an unprecedented period of sustainable, rapid growth.”

It hasn’t quite worked out that way. Indeed, if the last 10 years have demonstrated anything, it’s that for all the impact of a technology like the Internet, thinking that any new innovation will set us on a course of high growth is almost certainly wrong.

That’s in part because many of the studies purporting to show a relationship between the Internet and economic growth relied on shoddy data and dubious assumptions. In 1999 the Federal Reserve Bank of Cleveland released a study that concluded (PDF), “… the fraction of a country’s population that has access to the Internet is, at least, correlated with factors that help to explain average growth performance.” It did so by demonstrating a positive relationship between the number of Internet users in a country in 1999 with gross domestic product growth from 1974 to 1992. Usually we expect the thing being caused (growth in the 1980s) to happen after the things causing it (1999 Internet users).

In defense of the Fed, researchers at the World Bank recently tried to repeat the same trick. They estimated that a 10 percent increase in broadband penetration in a country was associated with a 1.4 percentage point increase in growth rate. This was based on growth rates and broadband penetration from 1980 to 2006. Given that most deployment of broadband occurred well after the turn of the millennium, the only plausible interpretation of the results is that countries that grew faster from 1980 to 2006 could afford more rapid rollouts of broadband. Yet the study is widely cited by broadband boosters. Many are in denial about the failure of the IT revolution to spark considerable growth.

Innovation in information technology has hardly dried up since 2000. YouTube was founded in 2005, and Facebook is only a year older. Customer-relations manager, the first cloud-based solution for business, only just predates the turn of the millennium. And there are now 130 million smartphones in the U.S., each with about the same computing power as a 2005 vintage desktop. Meanwhile, according to the U.S. Department of Commerce (PDF), retail e-commerce as a percentage of total retail sales has continued to climb—e-commerce sales were more than 6 percent of the total by the fourth quarter of 2012, up from less than 2 percent in 2003.

Yet despite continuing IT innovation, we’ve seen few signs that predictions of “an unprecedented period of sustainable, rapid growth” are coming true. U.S. GDP expansion in the 1990s was a little faster than the 1980s—it climbed from an annual average of 3 percent to 3.2 percent. But GDP growth collapsed to 1.7 percent from 2000 to 2009. Northwestern University economist Robert Gordon notes that U.S. labor productivity growth spiked briefly—rising from 1.38 percent from 1972 to 1996 to 2.46 percent from 1996 to 2004—but fell to 1.33 percent from 2004 and 2012.

Part of the labor productivity spike around the turn of the century was because of the rapidly increasing efficiency of IT production (you get a lot more computer for the same cost nowadays). Another part was because of considerable investments in computers and networks across the economy—what economists call “capital deepening.” But even during the boom years it was near-impossible to see an economywide impact of IT on “total factor productivity”—or the amount of output we were getting for a given input of capital and labor combined.

Within the U.S., investment in the uses of the Internet for business applications had an impact on wage and employment growth in only 6 percent of counties—those that already had high incomes, large populations, high skills, and concentrated IT use before 1995, according to a recent analysis (PDF) by Chris Forman and colleagues in the American Economic Review. Investments in computers and software did yield a return for most companies—but the return wasn’t anything special.

So what happened to the promised Internet miracle? While the technology has had a dramatic impact on our lives, it hasn’t had a huge impact on traditional economic measures. Perhaps that shouldn’t come as a surprise.

To understand why, think about television in the 1970s. Broadcast to the home for free, and all we had to pay for was the set and the electricity to run it. Despite that small expenditure, we spent hours a day watching TV. Fast-forward four decades: 209 million Americans spent an average 29 hours online in January, according to Nielsen; 145 million U.S. Facebook visitors spent an average six hours in January on that site alone. And each month, YouTube users spend 6 billion hours watching videos—or 684,000 times as long as it took to paint the Sistine Chapel. They pay for the PC and Internet connection, but per hour, surfing the Web is a cheap form of entertainment.

And all that time online has knock-on effects. Andirana Bellou at the University of Montreal suggests that broadband adoption has led to increasing marriage rates among 21- to 30-year-olds as they meet online in chatrooms and dating sites, but has “significantly reduced the time young people spend on socializing and physically communicating.” This may be why the link between Internet usage and measures of contentment are pretty weak. According to the World Database of Happiness, answers to a poll that asked if you were not happy (a score of 1), somewhat happy (2), or very happy (3) averaged 2.28 in the U.S. in 1990. That fell to 2.11 in 2010. Studies of Internet use tend to suggest that people who spend more time online are less happy (PDF) than the rest of us—although it may be that less happy people are surfing more rather than surfing being the cause of their misery. Regardless, the Internet has been behind a massive shift in our use of time during the past two decades, and not necessarily one that has generated a huge amount of positive feelings.

Of course, we also use the Internet to sell stuff on e-Bay and look for jobs. While Betsy Stevenson has suggested (PDF) in a National Bureau of Economic Research paper that the widespread use of the Internet in job searches may be one factor behind employed people switching jobs more often, she concludes there’s little evidence that the Web helps the unemployed find jobs faster.

Perhaps one reason we haven’t seen a huge impact on productivity because of access to the Internet is because, once we find a job, we spend quite a lot of time surfing the Web at the office. (Some of that time is used to look for a different job, apparently.) Ninety percent of workers with a PC also say they surf recreational sites at work. Almost the same number say they send personal e-mails, and more than half report they cybershop. The reality may be worse: Tracking software suggests that 70 percent of employees visit retail sites, and more than one-third check out X-rated sites. Perhaps we’re using the Internet to do more in less time at work. Yet we’re using the extra hours to check out pictures of Kate Upton or cats playing the piano rather than producing more widgets for the boss.

More broadly, Gordon argues that the big productivity wins from IT occurred long ago. He notes (PDF) that telephone operators disappeared in the 1960s as the first robots were arriving in factories. Reservations systems, electronic calculators, and barcode scanners spread in the 1970s and 1980s. Business-to-business electronic invoicing through electronic data interchange has been around since 1965—when Holland-America Steamship Lines started sending shipping manifests through telex messages that were automatically converted into computer data. Online retail shopping may be a comparatively recent phenomenon, but it isn’t a very important part of the overall productivity story.

The Internet has changed the economy and will continue to change it. Some industries—not least print media, booksellers, and broadcast TV—will continue to see dramatic upheaval. But the biggest impact of the technology has been as a more addictive form of entertainment than watching Friends reruns or talking to real friends in real life. If we’ve learned anything over the past 10 years, it’s that there’s no simple Web-based solution to an economy in the productivity doldrums.

That may be why Jack Welch has moved on to natural gas. Last year he announced that “the gas that we have found is in the first inning—it’s like the Internet in 1990. This is the first inning of a great American century.” Let’s hope this time he’s right.