Mon Aug 19, 2002

Firms, Markets and Information

In this post to his weblog, Brink Lindsey, author of Against the Dead Hand, tries to put the recent corporate scandals in the context of a wider debate about markets and hierarchies. He wants to answer three questions, in descending order of generality.

First, why are markets so great? Lindsey’s starting point is a standard one. Free markets are the best way to produce and allocate goods, and thereby promote growth and maximize welfare. The reason markets are better at this than (say) the state, was best expressed by Hayek. The distributed and decentralized nature of the market mechanism gathers and transmits a vast amount of dispersed information about demand and supply. The flow of information is reflected in changes in prices. Markets quickly and efficiently allow people to make more or less the right decisions about of how to allocate scarce resources with alternative uses.

Market ideologues generally stop there. Lindsey is more sensible, and asks the second question: If markets are so great at doing what Hayek claims, then why are there so many firms? Let alone, we might add, positively enormous corporate organizations? Lindsey gives the standard response from economics, which originates with Ronald Coase. Firms exist

because markets entail transaction costs (finding parties with whom to exchange, ensuring that exchanges will be performed appropriately, etc.), and administrative hierarchies can sometimes outperform markets by reducing those costs.

Well, maybe. We can note right away that Coase’s explanation for the existence of firms is at odds with Hayek’s reasoning about the superiority of markets. Hayekian markets are spontaneous, costless information processors. The lack of cost-based friction is what allows them to work their magic. When exchanges are costly, then what Lindsey calls “clarity of market feedback” will be severely impaired. A good deal of modern economics has focused on the problems that arise once information is no longer assumed to be cost-free.

Of course, this doesn’t mean Lindsey’s position is automatically undermined. Instead, he suggests that we can combine the two views:

Putting Coase and Hayek together, the organizational structure of a modern market economy reflects the interplay between transaction costs on the one hand, and what might be called “hierarchy costs” on the other the costs of ignoring dispersed information not available to the decision-makers in the organizational hierarchy. Firms grow in size and scope to the extent that reductions in transaction costs outweigh the loss of access to outside information. To look at the matter from the perspective of creating value rather than containing costs, the boundaries between firms and markets are set according to the relative value of applying specific, available information versus openness to unknown information.

I think Lindsey’s argument begins to break down here. At least, it runs into some unacknowledged issues. First, can we measure the “value of applying specific, available information versus openness to unknown information”? Transaction costs are difficult enough to estimate, though their existence is intuitively compelling. But how does a firm begin to figure the opportunity cost of forgoing some “openness to unknown information”?

Second, given the importance of good information to the smooth functioning of markets, what effect should we expect the various information revolutions of the past 200 years to have on the process of market/firm partitioning? One of the main effects of information technologies, according to many free-marketers, is to reduce friction between buyers and sellers. In other words, I.T. is supposed to lower transaction costs. Wouldn’t we expect more markets and fewer firms as information transmission became cheaper and cheaper, rather than bigger and bigger firms? Yet these revolutions—- telegraph, telephone, radio, TV and the computer/Internet—- seem to have gone hand-in-hand with the growth of the modern corporation. Shouldn’t someone adopting Lindsey’s theory be surprised at this?

That’s an honest question, by the way: I haven’t thought hard about what the expected effects should be, and I don’t know if any economists have. Alternative stories are easy to tell. On one view, information technology makes social organization and control possible on an unprecedented scale. (Recent books by Chick Perrow and Bill Roy give accounts consistent with such a view.) Or you could also claim that cheap information technology doesn’t affect the ratio of transaction costs to hierarchy costs. But if it doesn’t, what does? And, more importantly, how do we measure it? The Hayek/Coase rationale for the make-or-buy decision ought to be more than a rationalization.

Lindsey’s third question is this: If markets are so great and firms come into being only for sensible reasons to do with transaction costs, then why is there a horrible wave of corporate scandals right now? He answers that Enron happened because firms are not markets. Instead, they are “pockets of corporate central planning” which , alas, the market is forced to rely on for Coasian reasons. Corporate malfeasance can be traced to agency problems between shareholders and managers. And

Just as we can’t trust politicians and bureaucrats to pursue the public interest rather than their own selfish interests in power and perks, in the same way there is an ever-present risk that corporate managers will feather their own nests at the expense of shareholders.

Lindsey’s position is appealing—- it is a tragic market-liberalism rather than a triumphalist one. The tragedy is that, despite being so great for all those Hayekian reasons, markets can’t live without the state to support them on the one hand, or without creating islands of hierarchy inside them on the other. Lindsey takes a much more interesting and sophisticated position than many of his intellectual relatives. Note, though, that his irony-tinged neoliberalism leaves the market mechanism itself off the hook for Enron altogether. This makes me sceptical. For one thing, I think Enron shareholders were quite happy with how their stock was doing, to the point where they weren’t inclined to question their agents about how they could be doing so well. The traditional principal-agent problem is “Can’t Ask, Won’t Tell”. With Enron, it was “Don’t Ask, Don’t Tell.” It happened because of how markets (or particular types of real markets) work, not in spite of it. What Lindsey construes as a failure of bureaucracy inside the firm can just as plausibly be seen as a problem created by market incentives to cheat and lie about profits.

Lindsey sees the institutional machinery that markets need to function, as writers like Adam Smith did and so many contemporary free marketeers do not. He also knows that the market cannot, by itself, be the only check on corporate mismanagement. The difficulty with his position, it seems to me, is that he wants to absolve the market mechanism of any share of blame for Enron et al. So he blames “bureaucracies” instead. Hence his claim that firms are little “pockets of central planning” inside a wider market system. This makes it sound as if WorldCom was managed like North Korea. The distinction is much too sharply drawn. (Read some of the vast literature in organizational sociology on network forms of organization, relational contracting, internal markets or small-firm networks, to see a very different picture of firms.) Further, while Lindsey examines the problems of bureaucratic organization, he lets potential difficulties internal to the Hayekian market mechanism go by without scrutiny. If someone criticized the empirical pathologies of markets in detail and offered an idealized picture of a perfectly rational Weberian bureaucracy as an alternative, Lindsey would rightly cry foul. But he seems unwilling to ask whether an actually-existing market might differ from the idealized Hayekian mechanism, not because of external interference, but for reasons internal to itself.