The main idea: the Tiebout model of government competition can be said to truly fail to the degree that it ignores the perverse incentives created by the modern versions of feudalism, federalization and devolution.
1.2 In short, that model prescribes that local government will be efficient, economically speaking, as it operates under the threat of exit from those constituents who finds the services it provides insufficient to justify the taxes it asks for. Now, surely the assumptions of the model are clearly impossible to ever fully satisfy even in the market for non-territorial services (where exit costs very, very little), let alone in the market for local government.
1.3 But after one strips such artificial elements designed to allow such models to smoothly translate mathematically, one can easily see that the core logic of the Tiebout model is what I myself have relied upon when charting my Preemptive History of the Free Society: the smaller states will become, the cheaper the exit, the more efficient the law. Trade my complication from nuclear blackmail for Tiebout ‘s “perfect competition” artifices, and you get the same general feeling.
1.4 Though the main participant in the debate, Bryan Caplan, has so far only discussed how that model seems to fail in practice, leaving the much juicier discussion of how it should fail in theory for another post, I will do my best to assess whether the criticisms levied against that model are of an accessory or central nature. Nothing less than the future of the Free Society is at stake.
2. Accessory issues
2.1 I will try to separate the issues of the model in two categories: central and accessory.
2.2 As for the accessory issues, the main criticism levied again the model are that its predictions do not even come near to being observable in practice. As Prof. Caplan states, while the model indicates that there should be neither redistribution nor waste at the local government level, practice speaks otherwise.
2.3 Any libertarian who attaches any but the most perfunctory meaning to this term will readily agree with Prof. Caplan’s points. Yet these contradictions emerge only if one takes the interests of the Tiebout model as in describing the exact, actual behavior of every last Mayor. If that be the case, Prof. Tiebout should indeed have availed himself of far better assumptions.
2.4 Still, our interest in the model rests in trends it predicts, not the full-on results. Thus, though the assumptions are artificial, they will become gradually less so with secession. In this sense, issues such as those raised by Prof. Caplan can be categorized as accessory, as they rightly state why the world as it is now is not as good as it could, should and will be, but leave the main trend untouched. Unless, of course, the coming criticism of the theory itself makes one rethink the value of that trend.
2.5 Another issue with the model, this time raised by Arnold Kling, is that the alleged competition between local governments cannot work that well since citizens get to choose between bundles of services, and not every service separately. Thus, though my city may insist on confiscatory marginal rates of a steeply progressive property tax, it also provides some very good roads, schools, parks, libraries and so on. Or else, some very good services along with some mediocre ones.
2.6 To the degree that these services cannot be unbundled, citizens will have to select the whole package, and the pressure to get rid of the unsatisfactory parts of deal will subsist. Imagine the issues with City-States, who provide many, many more services than local government does.
2.8 I cannot see much of a problem here, in two distinct senses. First, to the degree that the services really can be unbundled, i.e. provided independently, competitive pressure will incentivize their independent improvement. For example, if Singapore provides a free market, low taxes but a repressive police state and that last bit is not a direct result of the first two, than Hong Kong could provide a free market, low taxes and a free civil society. After all, both the ‘ordinary’ free market and biological evolution work fine as competitive systems even if the competing agencies engage in bundling as far as they can.
2.9 But, and this is my second point, these services could be impossible to unbundle. Imagine, for example, an American-style liberal decrying the fact that the future independent San Francisco provides a very extensive portfolio of public property, yet only few jobs. We may suspect that these “services”, far from being arbitrary bundled, are causally linked. Jobs are scarce precisely because the government insists on withholding scarce resources from the private sector.
2.10 Unbundling, in such cases, can be seen as little more than wishful thinking. San Francisco will be unable to attract leftish citizens form Hong Kong, not because it cannot unbundle its services, but because the system it has chosen to sell does not, and indeed cannot deliver what it is supposed to.
2.11 Finally, an interesting point has been raised by David Friedman, who comments that the actions of local governments are readily reflected in the value of the land in their jurisdictions. If said governments can provide efficient services but then tax the resulting land revaluation away, they will gain and citizens will be left with no surplus. Hence, competition will be, form the point of view of citizens, pointless.
2.12 But how would this, I ask, differ from the similar effect of prices on the market (leaving aside the issue of morality)? If my mobile carrier offers outstanding service and sees itself flooded by new requests, it will surely drive up its prices until the demand for its services equal the supply. Would we than say that competition between mobile carriers – or any other businesses, for that matter – is pointless as those who can provide good service can scoop the surplus away by charging higher prices?
2.13 Yet, Prof. Friedman’s point is more complex. If, after all, Singapore institutes silly tax rates, the value of land in Singapore will fall. Those who wish to leave will have to sell their land at the new, low price and purchase it elsewhere at much higher prices. Thus, exit costs will be very high. High enough, Prof. Friedman holds, that there will be almost no backlash from inefficient policies. This does not happen with my mobile carrier: the quality of service does not make it cheaper or dearer for me to leave, though it makes it cheaper or dearer for others to join.
2.14 But again, the actual market may make us take such a mechanism with a grain of salt. And not any ordinary market, mind you, but the very heart of capitalism (at least according to Ludwig von Mises): the stock exchange. Indeed, the buyer of any stock finds himself in the very same predicament as the landholder in a City-State: the company/City-state has enough power to lower the value of his stock/land, and once his property has lost value it shall be very costly to sell it and buy some better performing stock/land. But still, can we hold that competition in the stock exchange is thus inexistent or even muted? Many further considerations (expectations of future policies, dividends, etc.) will ensure that Prof. Friedman’s mechanism is kept from breaking the market, even if competition between City-states shall never be as strong, or exits costs as low, as what we can witness in the weakest stock exchange. The main idea, though, will be present.
2.15 Indeed, far from seeing the point Prof. Friedman raises as an issue, I actually count on this mechanism to drive City-States toward efficient law: to the degree that they can keep their citizens happy and attract others, they will see their coffers abound with land tax money. Land prices in Hong Kong are, after all, among the highest on earth, yet people still seek to move there.
3. Central Issues
3.1 The serious, central and theoretical issue I see with the Tiebout model is the wholly lacking role of central government. To update a point I made in my The State, our false enemy post, the idea that federalization can be as good as actual secession is negated by the fact that stacking two (or more) layers of government on the same area will create perverse incentives for all players.
3.2 Indeed, a few years ago the Enlightened Saints staffing the central government of my own country decided to slash taxes on certain businesses, only to see local governments (conveniently, mostly presided over by opposition parties) raise the local equivalent of those very taxes to the same degree. The lesson was clear: it pays no one to slash taxes, and pays everyone to boost them.
3.3 Thus, to the degree that local governments are not the sole recipients of the boon resulting from providing good services, their competition will break down far more easily. Hamburg is, after all, supposed to be a City-State just as much as Singapore, yet it faces no comparable pressure to provide decent services. Indeed, if it did it would see much, if not most, of the resulting proceeds going to the Federal Government using it to prop up the failing welfare state of the day.
3.4 Still, this issue with the Tiebout model goes some way to push my idea of the evolution toward a Free Society. The road I prescribed does not include aggregation up to a World Government which would then adopt a federalist constitution, with the federal units being City-States. That would move us not an inch close to freedom (well, perhaps a few inches closer at most).
3.5 My point was that there was no alternative to full-blown secession. And to the degree that federalization or devolution stall full secession, they are to be seen as inimical to libertarian goals.
3.6 All in all, the criticism of the Tiebout model thus far levied can be seen as rightly attacking its artificial stack of assumptions, but leaving its theoretical core unscathed. The model only truly fails to the degree that it assumes competition between free City-States and not local satrapies of large empires.