The Tiebout model holds that competition for residents among local governments forces each to provide the optimal level of local public goods. One limitation to the argument is that although an individual can take his labor and his capital out of a locality whose government charges him more than the value it provides, his land remains behind.
Model a local government as a dictatorship in order to explore the non-political constraints it faces. Assume all citizens are identical. The optimal level of local public goods costs $10,000 per capita to produce. An adjacent polity currently produces that level of services at that price in taxes, as per Tibout.
The dictator sets taxes at $11,000 per capita, spends $10,000 producing the optimal level of local public goods, pockets the rest. Citizens start to leave. As they leave, the value of fixed resources, land and houses, goes down. They continue leaving until the cost of selling their house and land at a low price and replacing them at a higher price in the adjacent polity just balances the advantage of paying a thousand dollars less a year in taxes. There is then a new equilibrium with a smaller population and higher taxes. A local dictator who is selfish and rational will adjust the tax rate to maximize his revenue while still providing the optimal level of local public goods, since providing more or less would, once population had adjusted, leave him poorer.
The implication of this model is that the greater the value of living in the territory of a local government, taxes and services aside, the higher the level of exploitative taxation we should expect. If living in California instead of Nevada is worth $10,000 a year, California can afford to charge $10,000 more in taxes net of the value of services before citizens start to leave for Nevada. This looks like an explanation for why California, with a notoriously attractive climate, also has notoriously high taxes.
California, of course, is not a dictatorship. To fit the model more closely to reality, assume special interests within the polity, such as organized public employees, have effectively captured control of revenue. The threat of teacher strikes or police strikes or sanitation strikes can be used to push wages and pensions above the market level, transferring the excess tax revenue to the public employees. From an accounting standpoint there is no exploitative taxation, since the cost of the services provided, including the cost of those wages and pensions, absorbs all of the tax revenue.
This line of argument was suggested to me by an online discussion of what the disadvantages were of living in a low tax state where one participant wrote:
Whenever I've fisked the stats on such things all I've been able to see is that it is much worse to be a public employee in states without income tax. Schools don't seem to be worse, but being a teacher is worse, as an example.
One implication of the model is that there should be a correlation between the level of taxes in a state or local government and the natural advantages of its location. To test that implication, find or calculate an estimate of the natural advantages of different locations, taxes and services aside, and see if it correlates with level of taxes.
[Complaints about the formatting of this post should go to blogger.com not to me. All of the body of the post was set to the same font and font size.]