The Economist - July 24-30, 1999
LEADERS How real is the new economy?
THE recent performance of the American economy has been so remarkable that it seems only proper to call forth a very big idea to explain it. The theory of the "new economy" meets this need. A large idea in more ways than one, it accommodates many followers. A lot of them these days are sober, sensible types, making uncontroversial claims about the growth of the high-tech industries, about the spread of the Internet, the expansion of e-business and so on. At the outer edge still stand the techno-visionaries, declaiming prophecies that are one part heroic forecast and three parts science fiction. These true believers argue that the information-technology revolution goes not just wide but unfathomably deep-that thanks to IT, nothing humans do will ever be the same. Next to this revolution, the visionaries sometimes seem to think, the invention of the internal-combustion engine was merely noteworthy. As for electricity, its real significance is only now coming to be understood (you need it for IT). Certainly, they would say, the new economy explains America's record-breaking expansion, the recent surge in productivity, the vanishing of unemployment and the death of inflation. Abolishing the sad old laws of economics was but an afternoon's work for the new paradigm.
Who knows, the visionaries may be right about the distant future. Advancing technology has made clowns of forecasters for centuries. That makes it rash to believe the fundamentalists, but no wiser to say they are wrong: there is simply too little evidence to allow an intelligent discussion. However, what can and should be discussed, intelligently if possible, is the light that a more level-headed version of the new-economy thesis casts on the American economy right now.
Careful of that shifting paradigm
A respectable position-indeed it has lately become the consensus position-goes roughly as follows. Information technology is transforming the way America does business. This is revealing new opportunities for growth and helping companies to lower their costs. That is why Wall Street, recognising and confirming this, has soared to great heights. Also, compared with a few years ago, the economy can henceforth grow faster without risk of inflation. The proof lies in the current combination of rapid growth, very low unemployment and barely visible inflation.
It all seems to fit. Given America's marvellous performance, it seems curmudgeonly, unAmerican even, to question this thinking. Question it anyway. You will find that the evidence is not after all quite so clear-cut.
Especially puzzling are the productivity figures. Contrary to a widespread impression, the recent uptick in America's productivity raises more questions than it answers. Robert Gordon of Northwestern University, one of the country's top authorities on the subject, has found that more than 100% of the acceleration in productivity since 1995 happened not across the economy as a whole, nor even across IT at large, but in computer manufacturing, barely 1% of the economy. Elsewhere, growth in productivity has stalled or fallen. Might these strange figures be wrong? Yes, of course: the difficulty of measuring productivity in services, now most of the economy, is well-known. The trouble is, the more reliable figures for non-computer manufacturing show a worsening too-and that takes some explaining if the IT miracle is real.
This is not to deny the scale and intensity of the IT revolution. Firms across America are embracing it; jobs have been lost, new ones created; ways of doing business are changing all over (see article). If the productivity figures are true, how can they be reconciled with this gale of creative destruction? One possibility is that firms are having to go wired in order to preserve or extend their share of the market, but are not in the end producing all that much extra output from given inputs. In other words, radical change in a competitive economy-even though it may be a matter of life or death for the companies involved-need not yield commensurate improvements in economy-wide productivity. Maybe, as Mr Gordon speculates, the highest-productivity applications of computers are already in place, so that the economy is now quite far along the path of diminishing returns. Maybe (to stray for a moment into long-range prognostication) the IT revolution will never come close to matching the significance of the internal-combustion engine and electrical power, which both gave rise to vast new supplies of entirely new final goods-as opposed to, say, taking advertising away from newspapers and giving it to the Internet, or reducing lead-times in supply chains.
What about stock prices? What about the death of inflation? Wall Street is a fickle analyst, liable to change its mind at any minute. Put it this way: its Internet valuations call its judgment into question. As for the heavenly combination of low inflation and very low unemployment, there is no need yet for a big new idea to explain it. A combination of small old ones will do: the strong dollar, cheap commodities, the strength of Wall Street, cyclical weakness outside the United States, downward pressure on non-wage labour costs, and so on. To say this is not to belittle the success: America's performance has been awesome. And the new rise in productivity is no less real for apparently being concentrated in one sub-sector: it still boosts growth. The only question is whether new economic forces are needed to account for these blessings.
With luck the IT revolution will be everything its most zealous prophets claim, and mankind is on the verge of a leap forward as great as the one it made between the late 1800s and the middle of this century. Short of that is a more plausible wish: that today's far-reaching changes in the way business is done will soon lift the trend of growth in productivity across the economy as a whole. Even this, a bit less than a miracle, would be a wonderful prize. But the puzzling fact is, it does not appear to have happened yet. As John Maynard Keynes might have said: "When the facts are unclear, I keep an open mind. What do you do, sir?"
====
SPECIAL Work in progress W A S H I N G T O N , D C , A N D I N D I A N A P O L I S
On the surface, America's economy is changing dramatically-that much is plain. But just how deep the changes go, and what they imply for the country's growth in the long term, remains an open question
IT WAS Robert Solow, Nobel laureate in economics, who memorably observed a few years ago that "you can see the computer age everywhere these days except in the productivity statistics." He pointed out that the arrival of computer technology had affected people and firms in visible and obvious ways-the desktop PC was already ubiquitous-yet the slowdown in productivity growth that had been afflicting the American economy since the 1970s had apparently not been reversed. Lately, however, things have changed. Undaunted by Mr Solow's amused scepticism, the economy has moved on, and how. Growth is up, productivity is up, unemployment has all but disappeared and inflation appears to have vanished. Is it therefore time to declare the "Solow paradox" resolved, to conclude that it was just a matter of waiting for the miraculous effects of the computer revolution to come through?
Certainly, the grounds for that view are firmer now than before. Inspiring anecdotes and visionary predictions have been in ample supply throughout. Many executives (and even some economists) have been claiming for years that information technology, the communications revolution and globalisation were radically altering the way that American companies did business. Computer technology coupled with a global marketplace, they argued, would allow America's economy to grow much faster without overheating. Men of true faith went further: inflation was dead, they claimed, the business cycle abolished and the old economic rules repealed.
As long as the productivity figures suggested that nothing very profound was going on, this glut of optimism was easy to ignore. Recently, though, the productivity figures have been saying something else. In the past year or two, America has been increasing the output it squeezes from a given amount of input faster than before-surprisingly fast for this stage of the economic cycle. Add this to the fact that Wall Street seems convinced that an economic miracle is under way (nothing less could justify current valuations) and that unemployment continues to fall with no bad effects on inflation, and the "new economy" view looks more plausible than it did. Yet what is plausible is not always true. If the debate is more interesting today than it was before, that is partly because, even now, it is far from settled.
In one narrow sense, of course, the "new economy" school was right all along: if "new economy" means no more than "lots of high-tech firms and workers", all controversy on the point can cease. In real terms, American companies increased their annual investment in computers fourteenfold in the 1990s, while other investment hardly rose at all. As a result, the high-tech industry grew at a startling rate. Between 1995 and 1998 the IT sector, despite accounting for only about 8% of America's GDP, contributed, on average, 35% of the country's economic growth.
By 2006, according to a new report by the Commerce Department ("The Emerging Digital Economy II"), almost half the American workforce will be employed in industries that are either big producers or intensive users of information technology. And not only are "new economy" sectors growing fast, their productivity is shooting up too. Value-added per worker in IT-producing industries grew at an annual average of 10.4% in the 1990s, far higher than in the rest of the economy.
The web ahead
Attention is now shifting to the Internet and its attendant software, the newest and most publicised aspects of the IT revolution. This goes well beyond the hype of eBay and other stockmarket wonders. According to the Business Software Alliance, software companies now employ more than 800,000 people; employment in the industry is growing by 13% a year, compared with growth of 2.5% in the rest of the private economy. Although e-commerce firms such as Amazon hog the headlines, most of the business on the Internet is firm-to-firm (rather than firm-to-consumer). Numbers on Internet-based business-to-business commerce are sketchy, but they are undoubtedly big. According to Forrester Research, a leading Internet consultancy, American firms did e-transactions worth $43 billion in 1998. By 2003 the consultants estimate the total could rise to $1.3 trillion.
As the "new economy" (thus defined) mushrooms, it affects ever more aspects of American business and is transforming ever more parts of the country. Even three years ago, you could delineate the new economy geographically. Silicon Valley (obviously), Seattle, Boston, the Research Triangle Park in North Carolina and Austin, Texas (home of Dell Computer) were its homes. Now every state is, or wants to be, a high-tech state. California is still by far the biggest (employing 784,000 high-technology workers in 1997), according to a study by the American Electronics Association, the country's largest high-technology trade group. But other states are showing remarkable increases in high-tech activity. With 82 technology workers per 1,000 private-sector workers, New Hampshire has the highest concentration of new-economy employees. And the fastest-growing technology industries (starting, admittedly, from a low base) are in South Dakota, Utah and Idaho.
Throughout America, joining the new economy has become an aspiration. Take Indianapolis. Deep in the manufacturing heartland of the mid-west, this city is reinventing itself as a high-tech hub. Under the slogan "Wired. Inspired", Indianapolis wants to become a high-tech leader by 2005. The city's three high-technology clusters (advanced manufacturing, health technology and software) are booming. The number of software companies has quadrupled from 375 to 1,500 in five years. But the competition is fierce: from Minneapolis to Kansas City, heartland America is waking up to high-tech.
Fact and figment
If this is what "new economy" means, it is already an established fact. But this is not, or not only, what most new-economy prophets are talking about: their claims are much bigger. The real question is whether the emergence and expansion of the new high-tech industries is making the economy as a whole more productive. The answer will crucially affect America's long-term prospects for growth and higher living standards.
At first sight, the figures are encouraging. Since 1996 America's non-farm productivity (the benchmark measure) has improved on average by 2.2% a year-up from a 25-year average of barely 1% a year. If you estimate trends through the fluctuating quarterly figures over this whole period, there seems to be a point of inflection at about the middle of this decade (see chart), as though the macro-impact of the new economy kicked in about then. During just the past year or so, the rise in productivity has been especially impressive. In the second half of 1998 non-farm productivity grew by an annual average of 3.4%. During the first three months of 1999 it grew by 3.5%. If this improvement can be sustained, many (if not quite all) of the stronger claims of the new-economy enthusiasts will be shown to have been justified. But there is a problem. Impressive as they are, these recent productivity numbers do not settle the matter. A first point is that such increases are not unheard of. There have been two other spurts of high productivity growth since the trend of America's growth in output per worker first slowed sharply in the early 1970s. Between 1975 and 1978 labour productivity grew by an annual average of 2.3%; between 1983 and 1986 it grew by an annual average of 2.4%. Neither episode heralded a new economy. Admittedly, the situation was different then. In both the mid-1970s and mid-1980s, productivity rose temporarily as the economy recovered after a steep recession. Idle capacity was being put back to work. Today's circumstances are in sharp contrast. Big increases in productivity have occurred well into the expansion phase of the business cycle, after a period of heavy investment in new capital equipment. Still, the point stands: more data will be required to establish where the new trend, if any, lies.
In addition, some of the improvement is probably due to distortions in the numbers. The Bureau of Labour Statistics has changed its methods of measuring inflation (to try to correct the well-known upward bias in the consumer-price index): these revisions have increased measured productivity, perhaps by as much as 0.3 to 0.4 percentage points a year. Second, the surge could also be partly the result of unexpected growth in demand late in the business cycle. If companies were caught on the hop by the unexpected, and continuing, strength of the American economy, one way to keep pace with demand (especially in a tight labour market) would be to work existing employees harder. For as long as the expansion continues, this thesis cannot be disproved. Not until improvements in productivity growth have survived the next downtown will it be clear that they are permanent ("structural") rather than temporary ("cyclical").
But a third point is perhaps the most upsetting to new-economy optimists. The improvement in measured productivity since the mid-1990s is surprisingly, extraordinarily, concentrated in one small sector of the high-tech economy: computer manufacturing. Robert Gordon, a professor of economics at Northwestern University and one of America's leading authorities on productivity, has carefully broken down the aggregate numbers. He finds that, as prices collapsed, productivity growth in computer-manufacturing improved at a staggering 42% a year between the fourth quarter of 1995 and the first quarter of 1999 (see table above). Even though computer manufacturing is just 1.2% of America's output, that improvement was big enough to move the figures for the whole of the private non-farm economy. Indeed, allowing for other factors as well, productivity in durable-goods manufacturing apart from computers, and in the manufacturing of non-durable goods, actually grew more slowly in the most recent period than during the "slowdown" years of 1972-95.
Mr Gordon sums it up this way: "the productivity performance of the manufacturing sector of the United States economy since 1995 has been abysmal rather than admirable. Not only has productivity growth in non-durable manufacturing decelerated in 1995-99 compared to 1972-95, but productivity growth in durable manufacturing stripped of computers has decelerated even more." The new productivity numbers, far from settling the debate in favour of the new-economy optimists, seem thus to point the other way. And if the productivity miracle is so narrowly confined, its sustainability must be in doubt. To date, the IT revolution would appear to boil down to this: computer technology has proved unbelievably effective at reproducing itself; beyond that, its apparent influence on productivity (in manufacturing, at any rate) has so far been somewhere between imperceptible and adverse.
Lies, damned lies
Such a conclusion will strike many as frankly incredible-and not just new-economy zealots. For instance, Alan Greenspan, head of the Federal Reserve, is best understood as a cautious new-economy optimist. For several years he has been pointing out that American companies' high rates of investment, especially in capital that embodies high technology, were probably resulting in productivity improvements. In recent speeches and congressional testimony, he has explained in detail how "innovations in information technology...have begun to alter the manner in which we do business and create value, often in ways that were not readily foreseeable even five years ago."
The crux of Mr Greenspan's argument is that the easy accessibility of up-to-date information offered by information technology has allowed substantial improvements in corporate efficiency. Production planning is made easier; inventories can be reduced; delivery lead-times fall; the nature of distribution is altered. All of these factors increase the flexibility of capital goods, making capital investment more attractive and productive than it used to be, and encouraging firms to substitute capital for (scarce) labour. Although Mr Greenspan is always careful to point out that productivity growth cannot rise indefinitely, he appears to believe that the process is not yet over.
His case seems impeccable. No one would quarrel with the idea that this expansion has seen extremely high investment, and hardly anybody would question that this is likely to yield (soon if not now) productivity improvements. In part this capital investment was driven by prospect of high returns on technology; in part by the scarcity of labour and the need to substitute capital for labour.
Anecdotal evidence (none the worse for that: anecdotes are data) can be found to back this up all across corporate America. Navistar, an engine producer whose Indianapolis plant was named one of America's top ten manufacturing sites in 1998, is typical. Since 1995 it has poured more than $285m into capital improvements at the plant. In 1994, 900 people produced 175 engines a day. Today 1,900 workers produce 1,400 engines a day. The same story can be told again and again: enormous investment in new equipment, often heavily computer-based, helps to explain why (despite high growth) America's rate of capacity utilisation is relatively low. That, in turn, has helped to keep inflation subdued despite extremely tight labour markets.
Inner wiring
In addition, also difficult to deny, IT is fundamentally transforming the way companies are run. This is not just a matter of accumulating extra capital. The new economy is about the specific potential of information technology to change the way businesses work and thereby yield a quantum shift in productivity. The computer, and especially now the Internet, can change how companies deal with suppliers and customers; they can revolutionise firms' management, as well as create a whole set of new firms to capture Internet-based gains in efficiency. So even cautious new-economy advocates will say that Mr Gordon is talking bunk. They will claim that, although the trend may still be hard to isolate statistically, evidence of the broad impact of information technology-and more specifically of the Internet-is accumulating fast. The best-known examples are the big computer manufacturers. Dell Computers, for instance, says it handles 30% of its computer orders on the web; its supply chain and customer base are both "web integrated". Cisco Systems, the company that sells most of the gear that powers the Internet, handles nearly 80% of its orders through the web. It, too, has pioneered upstream and downstream web-based integration. Mr Gordon, of course, would acknowledge that productivity is surging in computer manufacturing. But if computer manufacturers like Dell and Cisco can make a success of these methods, why not other firms as well?
In fact, signs of a much wider revolution throughout American business are all around. Many of America's biggest companies are well ahead in developing and implementing their Internet strategies. Among smaller firms, in the words of one wired executive, the process "is still at the evangelical stage." Meanwhile, thousands of software companies are springing up to exploit and advance this revolution. Even Indianapolis has its share. One instance is Made2manage Systems, a software company that provides business-management systems for small manufacturing firms (those with less than 1,000 employees). It will run a company's entire operations, from inventory control to customer service and payroll. Its web-based e-business also allows suppliers and customers to be brought together online. Currently 1,200 firms use it. With a market of more than 50,000 firms in the United States and Canada alone, the potential for growth is huge.
Or take Attune, another Indianapolis start-up that designs web-based marketing software. Its product claims immediate gains in efficiency (the planning cycle can be cut by a third) and, more important, provides a way to connect corporate marketing people with a vast marketing-supply chain (advertisers, market-research companies, and so forth) online. Given that companies around the world spend in the region of $800 billion a year on marketing services, here, too, the potential is enormous.
It seems to be a sad case of the irresistible story meeting the immovable statistic. Maddeningly, the Solow paradox cannot be retired quite yet. In fact, the paradox has a new lease on life: it now looks even deeper and more puzzling than it did at the start. The figures for manufacturing productivity say that little if anything has happened except in the 1.2% of the economy that makes computers. Every other sort of evidence says that the world has been turned upside down.
Still waiting
Perhaps it is still a question of waiting for the statistics to catch up with reality. Productivity in services is especially hard to measure. The qualitative improvements that computers have ushered in (the now-classical example is the automated teller machine) are pervasive, but uncountable. Growth in output and productivity alike tends to be understated as a result.
Even if, to take the worst case, the increase in the growth of productivity is largely confined to computer-manufacturing, it is not to be sneezed at. The increase in quality-adjusted output of PCs is so vast in relation to the inputs used that it has perceptibly lifted the whole-economy figures for productivity and GDP. Mr Gordon happily concedes that the underlying rate of growth in American GDP has improved from a rate of between 2% and 2.5% to a rate of between 2.5% and 3%. If that could be sustained, such is the power of compounding, it would transform America's prospects. To take just one example, it would maintain the "solvency" of America's Social Security system, on unchanged policies, for many more decades yet and maybe even into the 22nd century.
The question, however, is whether it can be sustained. This seems likely only if big productivity improvements begin to migrate from the centre of the high-tech industry to the rest of the economy. In services, this may be happening already and the figures may simply fail to show it. In manufacturing as a whole, contrary to the evidence about particular firms, it does not yet appear to be happening. If pressed, one can think of reasons for this. Maybe the first 15 years of PC development were a bit of a let-down in productivity terms: firms spent a fortune on them and, for one reason or another, reaped comparatively few gains in efficiency. In future, things look brighter: the technology is maturing, spinning off new industries faster than during its adolescence (and thereby repeating a pattern that is familiar in the history of technology). And now there is the Internet, which as a strong force is much less than 15 years old.
All these are grounds for optimism, to be sure. The fact remains, the issue is not yet settled.