Join our community of smart investors

When competition doesn't work

Profit margins have, on average, been trending upwards around the world for years. Might new technologies reverse this?

When competition doesn't work

One of the first things students of economics learn is that competition tends to reduce and to equalise profits. In the real world, however, this is not true. A new paper by Jan De Loecker and Jan Eeckhout shows that profit margins around the world have risen since the 1980s, and that the variance among margins is huge.

There are countless reasons why this might be. Nick Bloom and John Van Reenen point out that there are big variations in managerial ability, and that competition does not quickly drive badly-managed companies out of business. Perhaps relatedly, Boston University’s James Bessen shows that differences in companies’ ability to use IT contributes to industry concentration and to variations in profit margins. And then there are many obstacles to competition. These include the 'economic moats' that Warren Buffett speaks of such as high capital requirements or brand power; patents and intellectual property rights; or consumers’ reluctance or inability to shop around. And on top of this the declining power of organised labour in many countries has allowed profits to rise at the expense of wages.

In this context, Google’s chief economist Hal Varian asks an important question: might new technologies reverse this trend?

Of course, technical change always destroys some big profits. Remember how digital cameras hurt Kodak and Polaroid or how smartphones did for Nokia. It’s possible, though, that it might also change the overall distribution of profits.

The key thing here, says Mr Varian, is how it affects the relationship between fixed costs and variable costs. High fixed costs encourage large companies with big mark-ups, because they act as a barrier to entry, and because companies need to be big to cover those fixed costs. But, he says, it’s possible that new technologies are cutting fixed costs because new companies can cheaply outsource what used to be costly business processes. They can use Quickbooks for accounting, Nolo for legal documents and so on. Further into the future perhaps, machine learning – which itself is cheap – can be used to reverse engineer some businesses and thus better allow rivals to compete with them. Such developments might reverse the upward trend in overall margins.

But there are countervailing tendencies. One is that machine learning will better enable companies to use price discrimination, charging more to customers that don’t shop around and less to more price-sensitive ones. That will tend to increase margins.

Also, the decline of mass manufacturing does not mean an end to high fixed costs. As David Evans and Richard Schmalensee point out, many businesses are platforms connecting buyers and sellers: think of Facebook, cable TV providers or Uber. These typically incur large losses before they acquire enough users to become profitable: there’s a reason why Trivago advertises so much. That’s a high fixed cost.

And then the facts pointed out by Bloom, Van Reenen and Bessen won’t disappear: firms companies differ greatly in their ability to manage technical change. That will contribute to winners and losers.

Which poses the question: can we identify these winners and losers in advance? Maybe not. Sussex University’s Alex Coad has shown that firm growth is very largely random. This is consistent with the fact that there’s often massive variation in returns to private equity funds: one or two big successes make all the difference. William Goldman’s famous maxim applies at least in part: “nobody knows anything”.

Except we do know one thing – that the elementary textbook idea of competition reducing profits is deeply questionable.