Bryan Caplan  

Taxation Under Tiebout

Too Many Bad Banks to Fail?... Why Don't the Chinese Learn fr...
The economist Charles Tiebout is famous for analogizing local government to perfectly competitive firms.  His model inspired this question on my last public finance midterm:

If the Tiebout model were correct, how would you expect local governments to raise revenue?  Carefully explain your answer.

Who wants to take a stab at this?  I'll post my preferred answer as a followup.

Comments and Sharing

COMMENTS (24 to date)
Eric Crampton writes:

Straight fee for service on all services, but it doesn't really much matter. If moving costs really are zero and if there really exist Arrow-Debreu worlds in towns, then anything else invites moving until folks are in the town that's charging them no more than they're consuming in services. Fee for service would do this most cleanly, but I suppose you could get the same outcome with just about any tax regime as folks would optimally sort across towns until any tax regime effectively replicated fee for service. So everyone in the town with an income tax is paying no more in income tax than they're consuming in town services (or else they'd be attracted away by some other town); similarly for the town with the property tax, or the town with the poll tax...

henry writes:

Interesting question. I assume the point is that you don't necessarily expect them to tax in the most efficient manner - i.e. with the least dead weight loss to society as a whole.

Given that my initial response was that you would expect them to raise revenue entirely from non-residents in some manner. So some type of tax that they might impose entirely on outsiders doing business in the jurisdiction. But any such tax would ultimately be borne by residents.

So my second shot is that you would expect local governments to raise revenues entirely through user fees. That would seem to be in equilibrium, because if any other jurisdiction tried some other scheme where taxes were imposed on those who didn't use the services, they would move to another jurisdiction where they got their money's worth of the services.

henry writes:

Looks like Eric beat me to it.

Provide services which prospective residents find valuable, thus increasing demand for living in the town. This can increase revenue either by allowing the town to charge a higher price (raise taxes) or by expanding the tax base at current rates.

Or adjust tax rates towards the revenue-maximizing tax rate, which should be just enough less than the market price to overcome switching costs. And in a perfectly competetive market, switching costs are assumed to be zero. So cut your taxes to once cent lower than the next-cheapest town.

El Presidente writes:

User fees for all/most services.

Using sophisticated price discrimination, a local government could, theoretically, maximize 'producer surplus'. Of course, this would largely undermine our most familiar and perhaps romantic notions of government actively producing positive externalities for the benefit of all market participants who divide them up as they see fit through our system of capitalism. It would reduce jurisdictional burden-shifting and back-door subsidies which account for a considerable amount of confusion amongst taxpayers who feed their money into the black box and aren't quite certain how their taxes benefit them.

John writes:

I'm not an expert on this, but here's my take.

My first guess is why people won't be charged for the services they use, just like businesses. Thinking further, if local governments are perfectly competitive that means that the price they are charging for their services are at marginal cost. If people choose to move places and are sorted into communities based on preferences for services, why can't they also sort themselves into places where they are taxed differently. It is possible that some communities might tax income and some might tax just based on services received. Or within each one there might be a combination of them. I don't know enough about the theory to be able to say, straight up, that one type of taxation would/could emerge.

Drew writes:

C'mon guys, the answer is obvious and if you disagree with it you are selfish and ignorant and you hate polar bears: CARBON TAXES! It would allow local government to drastically expand in order to measure and account for the tax itself until government was the only carbon producer left. And we are all saved! Wait, what was the point of this question?

John Thacker writes:

Given that my initial response was that you would expect them to raise revenue entirely from non-residents in some manner. So some type of tax that they might impose entirely on outsiders doing business in the jurisdiction. But any such tax would ultimately be borne by residents.

Well, you do see such "taxes" as aggressive ticketing for speeding of people driving out-of-state plates (or out-of-county, when the license plates indicate that, as in Florida) in some areas with popular roads and lots of non-resident traffic passing through. In some cases the advantage of the route passing through the jurisdiction may be high enough that the "highway robbery" rents can be extracted without ultimately placing too much of a burden on the residents.

Residents would have small or no tax rates, primarily geared toward services offered ONLY to residents, while anyone and everyone making use of services and facilities in the region would pay for the use of those services and facilities. Toll roads would be common. Sales tax would be high. Government services, e.g. vehicle registration and licensing, would be more expensive.

That's my take, anyway, as someone passingly-interested but not formally-trained in economics.

Alex J. writes:

User fees won't cut it for everything. Tieboutville is just the right size to contain the benefits from the public goods it provides. People want their local government to provide certain public goods. Tieboutville gives people what they want, ergo Tieboutville provides public goods. Public goods are non-excludable, so they can't be funded by user fees.

Consider that even though the residents are equally charitable, that doesn't make them 100% charitable. They are equally uncharitable as well. so you can't rule out free riders just because everybody wants the same services and is equally willing to pay for them.

All of the residents are the same, so they get the same benefit from city services and they are willing to pay the same price. Ergo: head tax. It's your presence or absence in Tieboutville that determines if you get the public services, therefore you have to pay to be present.

It's possible that the residents of Tieboutville also have a taste for government provided private goods. These services, unlike the other services which the town is best suited to provide, could be funded by user fees.

The Tiebout model sheds light on municipal annexation. To my knowledge, townships commonly can annex nearby unincorporated land with some ease. I don't know of any circumstance where it is easy for a part of a local government to spin off into a smaller unit. And so, most townships are too big. Some townships form for the sole reason of resisting external annexation, since even small local governments better able to protect themselves from annexation compared to unincorporated areas. Here, you would expect the townships to be too small.

Russell Bertrand writes:

Consumers demand some mixture of high quality and low price. There is an indifference curve.

Municipal governments will seek both to actually raise quality and lower price but also to appear to be doing so.

Governments will first have to figure out the relevant competitors for the "business" of attracting new residents. They might think of the relevant competitors as nearby localities, or all cities in the region, or far-flung metropolitan areas of comparable size, or even "world-class destinations." With local competition, there will be "market discipline." Any attempt to be an outlier (extremely high or low taxes; lavish or stingy services) is subject to failure in the marketplace. On the other hand, policy innovation can also yield above-average Return on Investment. (Perhaps a new local service like a town pool or better parks system will attract more citizens; perhaps a new revenue source like a lottery can keep down unpopular taxes [e.g., property])

In short, an effect closely related to Tiebout sorting should be that the policies of nearby localities should serve as moderate constraints on the policies of any given municipality.

EconNewbie writes:

My guess would be to specialize, or cater to a specific audience.

Say there is a community in Florida that sees benefits from retired persons living in their town. They should offer services that the retired persons would see as most beneficial to them, and therefore be willing to pay for through tax dollars.

Or perhaps there is a supply of young families with children in a certain area. It might be good to focus on parks, little league sports (hopefully ones where they keep score), and active involvement in the local schools in order to justify a certain tax rate (or user fee) that can raise revenues.

Otherwise, I know my hometown has seen some revenues come in from community festivals, but I don't know how much, or how effective they have been.

Arnold Kling writes:

I would think that a pure land tax (not a tax on improvements to the land). Think of the city as a landlord providing services and charging rent. The land tax is like the rent you pay to live somewhere.

If I use other taxes, then other towns can compete with me by offering lower tax rates--they may not lose revenue, because they will lure high earners (if I charge an income tax) or big spenders (if I charge a sales tax). But if other towns reduce the tax on land, all they get is less revenue. (I'm assuming that land is 100 percent owned everywhere)

Troy Camplin writes:

I don't think it would matter what the tax structure would be, because you would reach some sort of equilibrium among communities once a community found a way of collecting taxes. Once a community found a way of taxing, other communities would emulate them (that being easier than coming up with other ideas), and little real choice would then be involved for people.

Local communities, I would think, would look for taxes that would be "hidden," such as a sales or property tax. The taxes we pay are included in our mortgage payments, so we don't think too much about them, and renters have that tax further hidden from them. I personally don't like property taxes, because it amounts to rent being paid to the government, meaning they are the ones who actually own the land, and can take it away from you (evict you) if you don't pay the rent. So long as the government pretends you have property rights and ownership, you do -- but when they decide to stop pretending, you won't.

Another hidden tax is the sales tax. Most people don't really pay attention to how much they pay in sales taxes, even though it's on our receipts. So it's not literally hidden, though for as much attention as anyone pays to it, it is virtually hidden. Local governments could have different kinds of sales taxes, which would affect locations of shopping more than anything, I would think.

A lottery might work, but as my brother once observed, the lottery is more of a poor tax. Still, that might raise money from people not only in the community, but from people outside it as well.

In any case, if his theory is right, it will actually end up in the creation of the same tax structures across communities, as they reach an equilibrium that keeps people in with services and does not drive them off with taxes. And isn't that essentially what we see across the U.S.?

B.H. writes:

Levis Kochin showed a few years ago that if capital and labor use local public goods that are proportional to their size (more structures cause more need for fire protection; more people cause more need for policemen), then it is not efficient to use pure site rents. Doing so would encourage excessive accumulation. Tax whatever consumes the services.

That means that the value of structures is taxed also.

We ought to charge tuition at all schools.

Dan Weber writes:

Does Tiebout have the ability to keep people out? If they give everyone in town free health care, do all the sick people in the world get to move there?

8 writes:

Rent seeking. State and federal aid offer a "free lunch".

Property tax makes the most sense, either as currently constituted or the pure land tax described by Dr. Kling. People can leave, but the land cannot.

Massachusetts residents leave MA for NH, where higher property taxes are much higher, but there's no income or sales tax.

Every year the Tax Foudnation or someone publishes highest and lowest lists of municipalities with respect to taxes. The highest taxes are where everyone lives (e.g. NY, LA, DC) and the lowest (e.g., Bismark). So to be competitive with those hughly desireable locales, wouldn’t you implement high personal, property and sales taxes? Or would you admit that taxes are so low in importance that you shouldn’t let that tail wag the dog?

(And please don’t bring up Texas, it has high property taxes, etc so it can keep pimping the “No Income Taxes” brand)

El Presidente writes:


It would have to be a static competition. If there was any room for comparative advantage or exclusive innovation the concept would fail. Your land tax actually offers less opportunity for price discrimination because you have taxed something everybody needs but not in proportion to how badly they need it, or how much they are willing to pay for it. You are suggesting price competition, which would settle at a stable equilibrium where price equals marginal cost, but not value competition. If you allow cities to market different baskets to different consumers, then the user fees would raise more revenue and more profit (whatever that is in the public sector). Differentiation (non-price competition) is the only way to gain an advantage in a perfectly competitive market. Cities would still have an incentive, I'm guessing, to compete. Isn't that the logic behind free markets, or am I missing something?

I suggest issuing municipal bonds like there's no tomorrow.

After all, the people on the city council don't hold any equity in the city. Unlike the people sitting on a corporate board, they don't have to worry about losing capital and future dividends. Sure, this strategy will eventually bankrupt the city, causing living conditions go down. But, as Tiebout assumes, there are many other cities to move to, and no costs associated with moving. Assuming that our council members have the foresight to rent rather than own property, there's no downside.

Bond buyers and landlords will eventually wise up to this strategy. In the short-run, however, there's some great opportunity for legal, risk-free fraud here.

Politically, the scheme is most viable in cities with many renters and few property owners. Conversely, this fraud can be prevented by weighting each person's vote in proportion to the value of the real estate they own with city limits. (In analogy with the "one share, one vote" arrangement typical to jointly owned businesses.)

mgroves writes:

In marketing classes, the simplified answer was that there are two options:

1. Convince the existing market to use more of your product/service
2. Expand the market

So, I guess you would either try to convince more people to move to your city (perhaps lower taxes a bit, pass certain ordinances, offer other incentives), or make the city bigger in size?

bill fischel writes:

A student pointed this question out to me, so let me give it a try. There is some literature on this that suggests that a land tax is the ideal method. It is incentive compatible (you don't want to trash your neighbor because it would lower taxable values) and it has no deadweight loss. Henry George had a point. I have argued that a system of local zoning with ordinary property taxation (land and buildings) is actually a pretty good approximation of a land tax in a Tiebout model. The citation is (please excuse my pedantic aside) Fischel, William A. 1998. “The Ethics of Land Value Taxation Revisited: Has the Millennium Arrived Without Anyone Noticing?” In Land Value Taxation: Can It and Will It Work Today? ed. Dick Netzer. Cambridge, Mass.: Lincoln Institute of Land Policy. The reason zoning is essential is that it keeps supply of capital inelastic (in fixed proportion to land), so the tax on both is like a tax on land. A related argument is Glaeser, 1996. “The Incentive Effects of Property Taxes on Local Governments.” Public Choice 89 (October): 93-111. I was a little surprised to learn that Bryan had asked this question, since I gather that he does not think too highly of voters' ability to make rational decisions, one of which would be choice of tax base. I would demur on that on the basis of median voter evidence and Condorcet's jury theorem, but now I've really veered from the topic.

AO writes:

One effect, I imagine, would be towns would merge to form larger and larger local governments, as many public goods, like national defense, have increasing returns to scale (the thirteen colonies certainly thought so anyway). The increasing returns to national defense would suggest a natural monopoly government might arise, which would be able to erect military barriers to entry.

Miracle Max writes:

Fees and charges evades the question, which is really how to finance local public goods for which charging prices is prohibitively costly. A tax on site value as AK suggests is part of the answer. If the towns in question are as built up as the voters prefer, a tax on structures on top of site value works too.

People forget that in Tieboutville, people sort by preferences, not just to run away from taxes. Some communities like Takomaparkville could decide to have income taxes for redistributive purposes and control in-migration with zoning. Others could choose the libertarian ideal of private affluence and public squalor.

I would probably not be one of BC's favorite students.

Comments for this entry have been closed
Return to top