Quite a bit, it turns out. Business investment as a share of gross domestic product typically hits its low at the beginning of a recovery--right where we are now. The latest official data show that business investment in IT now stands at 3.3% of GDP. That figure--including spending on computers, communications equipment, and software--is down from a high of 4.3% at the end of 2000. However, it is a full percentage point above the 2.3% of GDP recorded right after the recession of 1990-91.
In effect, tech spending has held on to about half its gains of the 1990s, or a percentage point of GDP--nothing to sneeze at. With GDP running at more than $10 trillion, that extra point of output is worth about $100 billion.
Moreover, if there was an investment bubble in the 1990s, it may have been more on the nontech side. Business spending on non-IT equipment, such as trucks, industrial machinery, and construction equipment, rose from 4.8% of GDP in 1991 to a peak of 6% in 1999. Since then, nontech investment has given back virtually all of its increase, falling to 5% of GDP in the first quarter of 2002. There is little evidence of a sustainable rise in nontech investment.
Here's another way to assess tech spending. Despite the meltdown, IT spending in 2001 was still high enough to make up for the depreciation of existing hardware and software, BusinessWeek estimates. As a result, the amount of IT equipment per worker, measured using historical costs, continued to rise in 2001. That helps explain why productivity growth has stayed strong.
Of course, not all of tech has held up equally well. Spending on communications equipment, as a share of GDP, is no higher than it was in 1991, showing the effects of the telecom crash. By contrast, business spending on software has doubled as a share of GDP, from 0.9% in 1991 to 1.8% today, barely off its 2000 high. Indeed, American businesses now spend more on software than they do on aircraft, trucks, and other vehicles--something that is not likely to change anytime soon. Historically, a handful of music superstars generate a disproportionate amount of sales. But the Internet may be helping to level the playing field among artists by giving buyers easy and cheap access to music before they buy it, according to a study by Ram D. Gopal and Sudip Bhattacharjee at the University of Connecticut and G. Lawrence Sanders of the State University of New York at Buffalo.
The authors counted the number of different artists who made it onto Billboard's Top 200 list, which comes out weekly. They found that number jumped from 498 in 1991 to 655 in 2000, showing that sales were being spread around among many more performers.
It's possible that this could reflect changing musical tastes or simply that some superstars didn't release albums toward the end of the decade. But based on their data, the authors argue that a key cause was the jump in the number of Internet users, from 3 million to 120 million in the U.S. By allowing buyers to hear a broader selection of artists before they shop, the Net reduces the benefit of a superstar reputation when it comes to record sales.
Thus, the turmoil in the music industry caused by the advent of music sharing over the Net may in part be the result of a switch from a superstar model to one that opens up avenues for new artists. "The online music-sharing networks are star nurseries," says Gopal. "Instead of creating and marketing stars themselves, the recording industry should outsource these activities to the online networks." Record companies are still reluctant to embrace that idea. Each quarter, the Bureau of Labor Statistics reports on overall productivity trends. But it's less well known that the BLS issues data annually on productivity growth in individual industries.
The latest such report, recently released, goes through 2000. It shows which industries were the productivity leaders and the laggards over the last decade (table). And it offers some useful insights into productivity gains in service industries such as retailing and transportation. For one, the spread in productivity growth is quite wide within services. Among those at the top of the heap are nonstore retailers, including online and catalog sales. Their productivity grew at a 9% annual rate in the 1990s. By contrast, productivity fell in several service sector industries, including grocery stores and cable TV.
The data also show that higher output per hour does not assure good profits. Productivity soared in the telecom industry at a 5.9% rate. But these gains came in part from overinvesting in new equipment, which lowered profits. And it wasn't necessary to be high-tech to boost output per hour. Productivity at department stores rose sharply, and apparel stores got big gains by cutting worker hours.
Perhaps the big surprise was the strong performance of the much-maligned railroad industry, which generated 5.1% annual productivity gains, the result of mergers that cut employment and boosted efficiency. By comparison, air transportation checked in with productivity growth of under 2% per year.