Friday, July 13, 2012

What We Have Here Is a Failure to Regulate

. Friday, July 13, 2012

If you follow the same blogs as me you will have noticed an ongoing debate about workplace regulation. I'm not going to link to all the posts -- which by now number in the dozens -- but the key players have been the Bleeding Heart Libertarians, Crooked Timber, Marginal Revolution, Matt Yglesias, and Modeled Behavior. Two of those blogs are written by economists, one libertarian and the other... well I'm not quite sure how to categorize Modeled Behavior; one blog that is decidedly of the academic left, although from various factions of it; one blog of the academic non-economist right; while Yglesias seems to take his positions somewhat a la carte these days.

Perhaps predictably, given the roster, these groups are talking about different things. Moreover, they are talking about different things in different ways. As it happens I've spent a decent amount of time reading, thinking, and writing about regulation so maybe I can help clarify.

When economists think about regulation, they are coming from at least one of two schools of thought: those emphasizing the welfare-enhancing potential of regulation, and those emphasizing the welfare-destroying potential of regulation. Let's call the first school the "Utilitarians". Let's call the second second school "Public Choice". The first tradition has its roots in Smith, Hume, and (JS) Mill, but really takes off with Vilfredo Pareto, Arthur Pigou, and other 20th century economists. The second tradition gets a lot from Rousseau and Marx, but in its contemporary form has been best expressed by folks like Stigler and Peltzman.

The Utilitarians believe that regulation can be used to enhance social welfare by correcting market externalities. This is a situation in which the cost of a market actor's actions are not borne by her, but are passed on to someone else. (Alex Tabarrok made reference to this idea specifically at one point.) In this case, society has moved off of the Pareto frontier; total utility is below potential. Regulation is a tool by which we can move the equilibrium back up to its optimum.

A classic example concerns a firm that pollutes a river as a byproduct of its manufacturing; those who live downstream are the ones who suffer the most from the pollution, but they receive no profit from the production. Arthur Pigou suggested one mechanism for dealing with this -- taxing it at a sufficiently high rate as to bring private and social costs back in line, and redistributing the proceeds -- but regulating (i.e. limiting or outright prohibiting) the activity will often serve the same purpose.

The Public Choicers acknowledge that in this specific set of cases regulation can be welfare enhancing for society as a whole. But they posit another set of cases in which it will not be. That is because regulations, being restrictions, have a tendency to reward market incumbents and punish new entrants. The net result is an uncompetitive market prone to inefficiencies and other bad things.

Suppose the firm in the previous example acquires a market advantage by polluting. Once it has driven all of its competitors out of business, it then acquiesces to regulation restricting its activities. But all other firms must also adhere to this regulation. Because the incumbent firm is already capturing economies of scale and has a dominant market share, it will be very difficult for new entrants to be successful. The dominant firm may even initiate the regulation to prevent, e.g., a new firm from coming in, moving further upstream, and polluting there. The resulting lack of competition is also Pareto suboptimal. So the Public Choicers argue that eliminating the regulation -- and with it the concomitant barriers to entry -- can actually enhance aggregate societal welfare by fostering new competition, thus moving society back up towards the Pareto optimum*.

Both of these stories can be true in certain cases. It seems to me that Matt Yglesias and the Marginal Revolutionaries have been trying to tease out which is more true in the real world, in which cases, and whether the best thing to do is to increase regulation or decrease it. To do this they do the only thing they can do: refer to stylized models, simple intellectual exercises (would you trade being searched for contraband in exchange for an extra dollar an hour?), and what empirical evidence exists.

When the academic left thinks about regulation they are not thinking about to reach Pareto frontiers. They are not using stylized models**, as they reject the core assumptions of those models***. They are skeptical of econometric studies which, as Henry Farrell rightly notes, are susceptible to selection bias and questionable causality. In general, they have little use for the positive theories of regulation. What they care about are normative properties of outcomes as they exist in the actual world.

Instead of thinking about how to maximize Pareto optimality, the left is thinking about how power inequalities condition outcomes. There are two ways to translate this back into the economists' language: the first is to say "I care every bit as much about where along the Pareto frontier we fall as the fact that we're on it... the distribution of the pie is as important as the size of it"; the second is to say "I don't give a good goddamn if we're on the Pareto fronter at all... I want to improve the lot of labor, and particularly the poorest and weakest among the laborers".

Both of these reject the positive political theories of the economists. Or, actually, they (implicitly) accept them as intermediate steps in a necessary chain, but not as the culmination. If what matters to you is where along the Pareto frontier the actual distribution of utility between capital and labor lies, then you (implicitly) accept the positive Utilitarian theory of regulation as being welfare-enhancing in aggregate; you just think there's one more step to be taken which can be justified on normative grounds. If what matters to you is to improve the lot of labor full stop, then you accept the positive Public Choice account of regulation as being inherently redistributionary, adding a normative preference for redistributing from capital to labor.

So now we have four end points:

1. Utilitarians: The point of regulation should be to maximize social utility, so the biggest concern is reaching the Pareto frontier. This means regulating in the case of market failure but not otherwise, and the instances of actual market failure -- defined as deviation from the Pareto frontier -- are pretty rare. (The Bleeding Heart Libertarians' position)

2. Public Choice: Most of the time regulation will maximize private utility, not social utility; at minimum that will often be an unintended consequence. So the biggest concern is to limit regulation in order to reach the Pareto fronter. (The Marginal Revolutionaries's position)

3. Social Democracy: The point of regulation should be to maximize social utility as a preliminary step towards shifting the ex post distribution of aggregate utility along the Pareto frontier more towards labor. (The Matt Yglesias position)

4. Marxist/Post-Marxist: The point of regulation should be to maximize the well-being of labor, not capital, so who cares where the Pareto frontier is, even assuming I believe such a thing exists? (The Crooked Timber position, at least among some of them)

Despite what those involved in this argument might have you believe, these are all coherent theoretical positions. But as each of them have noted of the others at one point or another, they have all been somewhat detached from empirical reality. For example, are workplace regulations trying to correct a market externality? This question has been danced around but not really answered. For positions #1-3 the answer to this matters quite a lot. For position #4 not so much. Nor do any of them really get at the implications of power, although the CTers have come closest. Power matters quite a lot for #4 and somewhat for #2, but not at all for #1 and only a bit (and in different ways) for #3. 

There is a lot more to be said on these points, but this is long enough so I'll have to save it for another post. 

*This doesn't actually deal with the pollution problem of course.

**I'm not actually sure about this. I believe they have some model(s) in mind -- they must -- but they haven't clearly articulated what it is. In response to a Twitter query, Henry Farrell listed as some influences Jack Knight (a Tar Heel), Sam Bowles, and "no-bullshit Marxism which happens to work". Fair enough, although I'd prefer a clearer exposition on what that looks like in reference to this discussion. Chris Bertram, one of Farrell's co-bloggers at CT, has previously espoused (at least elements of) bullshit Marxism which does not happen to work, so perhaps that's his model. I imagine the left are coming at this from a variety of perspectives, which is fine.

***At least they think they do. In some cases I think they believe these assumptions are more severe than they actually are, so they're really rejecting a straw man. E.g., Farrell's repeated reference to "perfect competition", but labor models have moved well beyond perfect competition without markedly changing many of the relevant results. (Both Alex Tabarrok and Tyler Cowen have noted this, too, I believe.) Anyway, at least some of the left's interlocutors seems to be working from the even simpler, and even more unrealistic, model of monopsony. This obviously requires major assumptions too. At times the way they describe labor markets is in quasi-feudal terms; I doubt even Marx would have truck with that.


LFC said...

I've read this post quickly -- deserves a more careful reading -- but your first footnote is very important, I think, and probably should not be buried in a footnote.

The little story about how regulation of pollution creates barriers to entry and therefore removing the regulation leads to more competition: this in fact, as your first footnote points out, does not deal with the pollution problem. Instead of one company polluting, you have two (or more) companies polluting. It seems bizarre to call this result positive, unless one's theoretical perspective includes the principle that pollution does not matter (or that externalities in general don't matter).

Anonymous said...

Is this related to the debate about "workplace coercion" going on between Crooked Timber, BH Libertarian, and Julian Sanchez?

Kindred Winecoff said...


It is important. I chose that example intentionally, although the lesson from it is not necessarily generalizable to regulations more broadly. Eg zoning regulations are often not intended to correct externalities except under the broadest definition of the term, under which nearly everything generates externalities. What about the sort of regulations that Matt Yglesias is constantly harping about, eg hairdresser licenses? No externalities there, nor for most licensing requirements.

Perhaps the distinction will be made plainer if and when I get to the future posts I hinted at.

Anon -

It's intended to be, yes.

What We Have Here Is a Failure to Regulate




Add to Technorati Favorites