Curiouser and Curiouser

The WISE 2008 conference took place last month in Paris, and was a great experience. First of all, kind thanks are due to all the people who took the time to write with suggestions on how to have a good time in that city. Hemingway was right —  if Paris didn’t exist, we’d have to invent it. Every time I go there I’m reminded ofMFK Fisher’s wonderful observation that even though it’s an ancient town it’s always reinventing itself and offering something new, and that even those of us who have been coming for many years will never have to worry about it becoming stale. I remember when there were two classes of service on the Metro. Now the subway is highly egalitarian as is the Velib bicycle rental program, which is still used even on bitter cold winter days. It was good to be back.

The conference itself was a treat, as always. I think it’s the year’s best opportunity to catch up with both the top academic work being done on IT, and with the people doing it.  In the first session Erik Brynjolfsson and I presented our work using Gould’s Full House hypothesis to assess the effects of IT on competition in US industries. After that I got to sit back and enjoy the rest of the intellectual smorgasbord.

My favorite piece of work came from Adam Saunders, one of Erik’s doctoral students at MIT. I say ‘favorite’ not only because it was innovative, relevant, and rigorous, but also because it directly address the ways and the extent to which IT is reshaping the economy, a subject near to my heart.

In a paper entitled “Has Information Technology Leveled the Competitive Playing Field?” (email him for a copy, or to find out its status) Saunders investigated whether IT was like other forms of capital that a company can invest in, like warehouses, machine tools, and oil wells. Capital typically acts as a barrier to entry in an industry, making it harder to start up a new company because of the high initial investment required. Oil extraction, for example, is a highly capital intensive industry, and the expense of an oil well is one of the things (but not the only one) that explains why we see relatively few oil drillingstartups.

Saunders obtained from the Census Bureau data on the number of establishments each year between 1998 and 2005 for every US industry. An ‘establishment’ is simply a location where business is conducted. He also gathered data on the amount and type of of capital within each industry over the same time period. He reasoned that if basic economic theory about capital as a barrier to entry was correct, than as an industry became more capital intensive over time (in other words, as its companies acquired more warehouses, machine tools, etc.) then its rate of new establishment formation should slow down.  The greater the required capital, in short, the fewer new establishments we should see.

And this is exactly what Saunders found. As he writes, “as an industry becomes more ordinary capital-intensive, there is less entry of small firms and fewer establishment openings by large firms.”

Note that in that sentence he was careful to say ‘ordinary capital-intensive,’ a phrase he invented out of necessity. This is because he went a big step further with his analyses and took advantage of the fact that the Census Bureau breaks out IT as a separate form of capital. Saunders did the same, and found something remarkable: while other forms of capital (‘ordinary capital’ ) had the effect predicted by theory, IT behaved very differently. In fact, it had precisely the opposite impact. As industries became more IT-intensive they saw higher numbers of new establishments over time. IT capital, in other words, appears to be unique in that it lowers barriers to entry rather than raising them.

Saunders’ analyses assessed both the effects of both IT and ‘ordinary’ capital simultaneously, so he could draw conclusions about the effects of IT not in isolation, but instead in the presence of other capital. He also took the very smart step of excluding from his analyses all the industries that produce IT — hardware, software, and networking gear. This means that there’s no chance that his findings are are distorted by the peculiarities of the high tech sector. Instead, Saunders has uncovered a phenomenon taking place in the mainstream economy where most of us still work.

Saunders hypothesizes that IT has this barrier-lowering effect because it lowers many of the other costs associated with starting up a new establishment. For example, since communication is now so cheap many types of company locate themselves wherever they like. And computers can also substitute for costly human labor.

One other important aspect of Saunders’ work, and one that yields further surprises, is that he looked primarily at establishments rather than companies. An establishment can be a standalone firm (think of your local dry cleaner) or a part of a larger entity (which would be the case if your local dry cleaner was part of the Zoots chain). Saunders found that as an industry becomes more IT intensive, both kinds of establishment become more plentiful —  new small companies appear, andpre-existing larger companies open more locations (it also appears, furthermore, that companies from other industries start “spilling over” into the IT-intensive ones). So one big surprise in his work for me was the finding that IT capital actually appears to lower barriers to entry rather than raising them. And there was also another major surprise.

One of the most persistent pieces of lore around IT is that it levels the playing field between big and small companies —  that in the era of the Net and fast computers startups and small businesses have the same reach and have access to the same technology-based capabilities as large enterprises. All other things being equal, this fact should favor small companies relative to larger ones, and lead to a greater percentage of small firms. If IT removes one of the disadvantages of being small, after all, we should see more small companies in IT-intensive sectors.

But Saunders’ analyses revealed that the opposite dynamic is taking place. As industries became more IT intensive they became less hospitable environments for small companies instead of friendlier ones. it’s true that as industries computerized they saw morestartups, but it’s also true that they saw more deaths among startups , and greater expansion of establishments that are part of larger incumbent companies. Overall, greater IT intensity was associated not with a more even battle between large and small firms, but instead with greater concentration —  a trend toward domination a few large players.

Here again, the opposite pattern held for non-IT capital; greater capital intensity was associated with greater fragmentation rather than concentration. As ordinary capital increases, small companies appear to have an easier time defending their niches against big established rivals.

Saunders used different data and a different methodology than Erik, Michael Sorell, Feng Zhu, and I did in our paper “Scale Without Mass” but reached just the same conclusion: that greater IT intensity is associated with greater concentration. Evidence is mounting that technology is not the great competitive leveler. Instead, it is helping separate winners from losers and leading to the creation of a smaller number of more dominant companies within an industry.

I believe that this is because modern IT increases the scope, the precision, and the fidelity with which a business innovation can be propagated throughout a company. To put it as tersely as possible, good ideas and good execution separate winners from losers, and IT helps companies execute on their good ideas (technology also helps companies generate good ideas, but that’s a subject for other posts).

This explanation could be wrong, of course, and it could also be the case that something other than technology, something correlated with IT investment but totally separate from it, is actually driving the changes in competition we’re observing. But I don’t think so. I see a compelling story coming together about how and why IT separates winners from losers, and I see this story being backed up by a growing body of data and research like that done by Saunders.

Do you see the same thing?  Leave a comment, please, and let us know.