Bryan Caplan  

Where Tiebout Goes Wrong

Nominations for Top Economic S... My Inflation Bet with Bob Murp...
The Tiebout model implies that local governments will be redistribution-free and waste-free.  If you find these predictions absurd - and you should - local governments must violate one or more assumptions of the Tiebout model.  What are the key violations?

1. The lowest-hanging fruit: Tiebout explicitly falsely assumes that local governments are perfect competitors.  In the real world, though, local governments face a downward-sloping demand for residency curve.  A locality does not lose all its residents simply because it offers a tax-benefit package 1% worse than the district next door.

2. Tiebout implicitly assumes that dissatisfied residents can take their real estate with them.  Virtually the opposite is true.  So as David Friedman points out, and as one of my papers in Public Choice argues at length, the main effect of unfavorable tax-benefit packages is not to induce migration, but to reduce property values.

Ultimately, though, neither (1) nor (2) get to the heart of the problem with Tiebout.  True, local governments aren't perfect competitors.  But imperfectly competitive markets normally work fine: Ponder Amazon, Apple, or the restaurants in your neighborhood.  Furthermore, (2) ignores reputation.  Market forces may fail to protect current real estate owners in a locality.  But local governments that somehow "bind their hands," that are renowned for their aversion to redistribution and waste, should be far better at attracting labor and capital in the first place.

The fundamental problem with Tiebout, rather, is:

3. Tiebout implicitly assumes that non-profit competition works the same way as for-profit competition.  It doesn't.  If a business owner figures out how to produce the same good at a lower cost, he pockets all of the savings.  If the CEO of a publicly-held corporation figures out how to produce the same good at a lower cost, he pockets a lot of the savings.  But if the mayor of a city figures out how to deliver the same government services for lower taxes, he pockets none of the savings.  That's how non-profits "work."

With non-profit incentives, neither the number of local governments nor the ease of exit lead to anything resembling perfectly competitive results.  The "competitors" simply have little incentive to do a good job, so they all tend to perform poorly. 

To use an educational analogy: Think about the difference between competition on a graded exam versus competition on an ungraded exam.  In both cases, there are many competitors.  But if the exam doesn't count, the competitors don't try very hard, so the average outcome is poor.

You could object, of course, that the desire for re-election provides the necessary incentive.  But (a) the whole point of the Tiebout model is to show that local government delivers the goods regardless of the quality of the democratic process, and (b) the desire for re-election often gives local politicians perverse incentives to pursue redistributive and/or wasteful policies.  In the Tiebout model, a mayor who tried to replace over-educated, overpriced high school gym coaches would see his locality swell with residents and investment.  In the real world, however, a mayor who tried to do this would be called a monster and thrown out of office.

If you're a populist, you should be delighted by this outcome.  If you're an economist, you should be horrified.  No matter how you evaluate the outcome, though, one thing is clear: contrary to the Tiebout model, local governments are not analogous to perfectly competitive firms.  Not even close.

COMMENTS (14 to date)
Chris writes:

The Curley Effect also runs directly contrary to the Tiebout Effect:

local governments are coercive governments, not profit-maximizing firms!

Fischel's Homevoter Hypothesis gets the housing point right, though:

Chris writes:

The Curley Effect is a bit like Obama/Pelosi immigration policy, in fact.

Justin Ross writes:

Bryan is one of my favorite bloggers in the universe, but his Tiebout series is so disorganized that I can't make sense of it. For instance, who is the target of this critique?

At first, I thought it was Tiebout (e.g. "Tiebout implicitly assumes..."), except that the only intention of the Tiebout paper is to demonstrate that mobility is a plausible mechanism for the revealing of preferences for public goods. Why Bryan seems to think otherwise is puzzling. Samuelson seemed to think that no mechanism was possible and that this made federal public expenditures necessary, so Tiebout sought to show otherwise. Tiebout's title "A Pure Theory of Local Expenditures" is a direct parody of Samuelson's "A Pure Theory of Public Expenditures" for this reason.

As a result, I thought maybe Bryan was actually targeting what might be described as "the Tiebout literature", but he only seems to compare back to assumptions of the original Tiebout model. This is a giant literature, much of which reaches different conclusions on efficiency in public good provision or other topics that Bryan hits on. In order to accomplish this, that literature adds many of the complexities that were simplified away from the Tiebout model so that it could be used to draw implications on other topics, such as "governments competing" and other voter outcomes (none of which were the subjects of Tiebout 1956). Ron Hamilton had a highly influential series of papers in the 70's (e.g. 1976 AER; 1977 JUE) that brought housing price capitalization into a Tiebout model with zoning, so Bryan is extremely late to the party here.

MG writes:

A model that (I will have to take Bryan's word for this) predicts the complete absence of waste and redistribution ("-free") would always strike me as unrealistic. But I am surprised that it is as bad as "Not even close." In the long-term there must certainly be a decent amount of responsiveness to avoidable taxation (out migration, underinvestment), some of which may be masked by unrelated advantages (climate, geography, existing cultural draws) that embold the pols to push taxation higher and lull the taxpayers into paying.

Arnold Kling writes:

I think that one needs to rely on barriers to consumer choice. Otherwise, how do really bad local governments persist when some governments are better?

The fact that we observe some particularly bad local governments is an argument against both the Tiebout model and your model. Even if the bad governments have no incentive to improve, why do those governments still have constituents?

Seth writes:

A benefit of local governments, that perhaps the Tiebout model does not identify but exists in reality, is the benefit accidental experiments.

If one works, others can adopt. Certainly, there's not quite as strong incentives to adopt as there might be in a private market where the owners pocket some of the savings or added value, but it sometimes still happens.

Bryan Willman writes:

I suggest that all of these models would be improved by accounting for reality on the ground and on the limits of human beings.

For many citizens in many places in the US, local government is a surprizingly minor part of their lives. Many (most?) of us pay far more in various taxes to the federal and state governments than to local entities. In many jurisdictions, school boards are separate entities and so the most important local issue (schools) isn't even under the purview of the mayor. I speculate that mayors and the like sometimes make daft sounding proposals they know to be hopeless, just to remind people who they are. (I do not actually know the name of the mayor of the town I actually live in. Nor do I need too.)

The Curly effect (both pro and con) is surely real, and eventually self limiting.

But the very diffuse nature of paying for government services, and the "protective" nature of some of them, will always cloud the debate.
Quick - what does a police officer in your city get paid? What does the fire chief make? What part of your property or local income taxes are consumed in this way? What did you pay in gas taxes last year and what part of that was spent on something other than roads? And does any of this come *close* to your federal income tax or your concerns about social security and retirement funds?

John Bourne writes:

"The Tiebout model" Tiebout?

Can he throw a football?

Jets may be interested.

Chris Thomas writes:

Isn't the fact that local governments are not autonomous a big factor? Don't the federal and state governments act to ensure uniformity in local governance, at least to a very large degree? Doesn't that fact alone show that competition among local governments is largely a fraud?

Michael Strong writes:

In Singapore, government employees receive a bonus based on the previous year's GDP growth,

One could imagine local communities that tied compensation of government employees to some metric of performance (relative increase in property values vis-a-vis neighboring communities, for instance), that would result in more effective Tiebout competition.

Costard writes:

MG has it.

I'd suggest that "waste" is a poor term for what happens in government. It implies a lack of intent, and an absence of beneficiaries. Incompetence may play its part, but in politics even the idiots are useful.

sourcreamus writes:

The problem with the competition anaylsis is that most localities don't want to win the competition. Burger King would like all of McDonald's customers, but Fairfax county does not want all of Washington DC's residents. Many localities only want enough new residents to keep housing prices up. More new residents means more crowding and more land use restrictions and more anti-developer laws.
Also most jobs are so location specific that competition is minimal. It does not matter how much better the government is a county 100 miles from DC is not going to compete with a county across the river.

Bill Nelson writes:

Local governments might not be profit maximizing, but they seem to be very good at revenue maximizing. Isn't that the problem? Wouldn't private profit-maximizing local governments also leave zero consumer surplus?

Floccina writes:

Most areas have slow growth polices they want fewer residents not more so they will not try to attract new residents. Doesn't that blow up the whole premise?

Comments for this entry have been closed
Return to top