In the last six months, output has barely grown, but hours worked have fallen sharply. As a result, labor productivity -- output per hour worked -- has grown at a healthy pace.

Productivity growth of 3 percent or more a year is generally considered good. In the last quarter of 2001, productivity grew at an annual rate of 5.5 percent. In the first quarter of 2002, it grew at a phenomenal annual rate of 8.4 percent, the fastest in 19 years.

Quarterly estimates are notoriously volatile. But regardless of possible future revisions, recent productivity growth seems strong, and that is good news for the economy.

But is it good news for business? Not necessarily. The relationship between productivity and profitability is looser than one might think.

If productivity grows in highly competitive industries, the benefits can quickly be competed away. The telecommunications industry is a good example.

After the AT&T breakup in 1984, labor productivity in telecommunications services grew 8 percent a year. But the competitive environment in this (partly) deregulated industry has meant that most gains have shown up in the form of lower prices. This is great for consumers but does not do much for the bottom line.

That the telecommunications industry is in such terrible shape now is in part because productivity growth has been so strong: the increased output of services created a glut of capacity, which in turn led to increased competition and lower prices.

Conversely, profits can increase without productivity growth. If one company grows at the expense of others in its industry, it may become more profitable. But total industry output divided by total labor hours in that industry may not change much.

Everything hinges on what it is that made the company grow. If its growth resulted from labor-saving technology, which reduced costs and made it more competitive, one would tend to see a productivity increase at the industry level, because a larger fraction of industry output would be produced using more efficient technology.

But if the increased sales were caused by a clever marketing campaign, with no fundamental improvement in the way labor was used, aggregate productivity would not be affected.

If the Gap sells a lot more T-shirts because it guesses correctly that fuchsia is the season's cool color, that's great for its profitability. But these profits will not show up in industry figures, because additional sales of fuchsia T-shirts tend to reduce sales of other retailers' T-shirts.

Government economists predict that productivity will continue to grow for the next 5 to 10 years at around 2 percent a year. This is not as good as the heady growth of 2.5 percent during 1995-2000, but it is a lot better than the dismal 1.4 percent from 1974-1995.

Most economists think our recent productivity growth results from some combination of a better-educated and more experienced work force, adoption of better ways of doing business and information technology, though opinions are mixed about the relative importance of each factor.

What about the Internet? Lots of companies have put up Web pages, but these have generally been oriented toward revenue enhancement. Online catalogs, electronic shopping and after-market support are pretty easy to put in place and seem to help attract and retain customers.

Attracting customers is great for profits, but because those customers tend to come from other companies, such gains have a relatively low effect on overall productivity.

Productivity gains, for the most part, come from cost savings. It is applications like supply chain management and work-flow management that really change productivity. These are more difficult to carry out, because they change the way things are done. But changing the way things are done is precisely what affects aggregate productivity.

When Henry Ford and his executive team were down on the factory floor 90 years ago, fine-tuning the assembly line, they were revolutionizing the way goods were produced. Today Amazon constantly tinkers with its Web page -- offering new services, testing features, experimenting with ways to make online shopping more efficient. Ultimately, it is such tinkering that drives productivity growth.

It's not just communication with customers and suppliers; companies are also using intranets to improve internal communication. Last year a Silicon Valley executive told me downsizing was now much easier than it used to be, because companies could use intranets to communicate with employees more effectively. This allowed them to squelch rumors quickly and to minimize, to some extent, the productivity hit that inevitably accompanies layoffs.

It may seem odd that a significant contribution of the Internet is in making it more efficient to lay people off, but on reflection, perhaps it is not too surprising.

Companies do not like to lay people off. One reason productivity declines during recessions is ''labor hoarding.'' Companies hold onto workers, even when demand drops off, leading to hours worked declining more slowly than output produced.

But that did not happen this time. One possible explanation is that when the economy was booming, companies could have used fewer workers to produce the same output, but because they were not forced to lay them off, they just assigned them to lower-priority tasks. When the squeeze came, companies were forced to downsize, but production did not really suffer, as these workers were not critical to daily operations.

Now these laid-off workers can move on to jobs where their skills can be put to more efficient use. This is not easy for anyone, but it is that reallocation of labor to more productive uses that leads to a healthier economy in the long run.