The Problem of Structure

What would the constitution of a free nation look like? In trying to answer that question we immediately think in terms of a Bill of Rights, restrictions on governmental power, and so forth. And any constitution worth having would certainly include those things. But if a constitution is to be more than a wish list, it must also specify the political structure necessary to ensure that these freedoms are not eroded or ignored. Consider the old Soviet Constitution, which guaranteed all sorts of fine-sounding freedoms for its citizens — but which in practice proved only an empty promise, since its interpretation and enforcement lay in the hands of an unfettered monolithic centralized state.

Framing a constitution is an exercise in public-choice economics; politicians react to incentives, and so the political incentive structure must be designed in such a way that those in authority cannot profit by the aggrandizement of state power.

Such was the intent of the Framers of the U.S. Constitution when they established the Federal system. Each branch of government was expected to be jealous of the others, and so to be motivated to serve as a check on their expansion. At the same time, the broad base of representation was expected to ensure that no special interest could succeed in manipulating the government.

As we have since learned all too well, the experiment eventually proved to be a failure. Madison and his colleagues could not foresee the logrolling process whereby government branches and special interests (“factions”) that were intended to hold one another in perpetual check instead made concessions to one another’s ambitions in exchange for like concessions to their own. Yet some foresaw the danger; one Anti-Federalist poet, bemoaning the recent ratification of the Constitution, wrote:

In five short years, of freedom weary grown,
We quit our plain republics for a throne;
Congress and President full proof shall bring
A mere disguise for Parliament and King.

Thus did the streamlined confederation become the swollen imperial Leviathan.

Decentralize!

Thomas Jefferson wrote many stirring passages about the natural rights of man. But when asked to summarize his political philosophy, he replied that he could state it in one sentence: “Divide the counties into wards.” In other words: decentralize, decentralize, decentralize!

There are many advantages to political decentralization as a structural limitation on government power. Imagine a country the size of the United States, but consisting of only five states. Now imagine the same region containing 500 states. All other things being equal, the second situation is likely to be much more hospitable to freedom than the first. The smaller the political unit, the greater the influence an individual citizen can have in politics, thus decreasing the lobbying advantage that concentrated special interests have over the diffuse general public. Further, as the number of available alternative political jurisdictions increases, the citizen’s exit option becomes more powerful. The freedom to leave one state is small comfort if there are only a handful of others nearby to go to; but with many states, the odds of finding a satisfactory destination are much better.

In addition, competition between states can serve as a check on state power, since if any state becomes too oppressive its citizens can vote with their feet. Also, decentralization softens the impact of government mistakes. If a single centralized government decides to implement some ill-conceived plan, everybody has to suffer. But with many states implementing different policies, a bad policy can be escaped, while a good policy can be imitated. (Here too, competition can serve as a discovery process.) The Federal structure of the United States, imperfect though it is, may well explain why this country has not plunged as rapidly into socialism as its European counterparts — individual states having free entry while most European countries do not. (Fifty states is certainly better than one — though still a far cry from Jefferson’s notion that six square miles was the optimum size for a basic political unit within a republic.) And Switzerland’s even more decentralized canto! n sy stem has doubtless played a similar role in the preservation of Swiss freedom. (Frances Kendall and Leon Louw’s book After Apartheid has helped to bring to the libertarian community’s attention the usefulness of a canton-style system to countries torn by ethnic strife; but its appeal is not limited to such cases.)

Housetrailers and Empty Landscapes

The constitution of a free nation, then, should most likely be characterized by a radically decentralized power structure, along the lines of some sort of canton system. But can the Swiss model be improved on? I think so.

The effectiveness of competition among political jurisdictions is inversely proportional to the costs of changing one’s jurisdiction. Massachusetts faces serious competition from New Hampshire, but little from Alaska, since the costs of voting with one’s feet are so much higher in the second case. (The same is true at the international level; there’s no mystery as to why Cuban and Haitian refugees are trying to get into Miami rather than into Geneva.) And even when the alternative jurisdiction is nearby, the costs of switching are not exactly low. Uprooting oneself and perhaps one’s family in order to move to another state can be costly, both financially and emotionally.

The high cost of switching results from the fact that political jurisdictions correspond to geographical regions, and geographical relocation is not always feasible. Yet a decentralised system serves as a more effective check on the growth of state power to the extent that switching costs are low. Therefore, it seems desirable to decouple political jurisdiction from geographical location.

David Friedman offers an example: “Consider our world as it would be if the costs of moving from one country to another were zero. Everyone lives in a housetrailer and speaks the same language. One day, the president of France announces that because of troubles with neighboring countries, new military taxes are being levied and conscription will begin shortly. The next morning the president of France finds himself ruling a peaceful but empty landscape, the population having been reduced to himself, three generals, and twenty-seven war correspondents.” (Machinery of Freedom, 2nd ed., p. 123.)

If people could switch political jurisdictions without switching location, we would have the functional equivalent of the situation Friedman envisions. Competition among jurisdictions would be higher, and the amount of state interference that people would tolerate without switching would be lower, than in a political system where jurisdiction and geographic location are linked.

The Case of Iceland

There are a number of historical precedents for this idea. To pick one famous example, the Icelandic Free Commonwealth (930-1262) operated on the Thing system. A “Thing” was a court or assembly. (The English word “thing” originally had this meaning also; when Hamlet says “The play’s the thing wherein I’ll catch the conscience of the King,” the pun is successful because in Shakespeare’s time the word “thing” had begun to have its modern meaning, but also still retained the earlier flavor of a judicial proceeding to establish guilt or innocence.) The national legislative assembly, with its attendant national judiciary, was called the All-Thing; beneath it were four Quarter-Things, corresponding to the four geographical regions of Iceland. But here the tie between geography and jurisdiction ended.

Under each Quarter-Thing were three or four Varthings, and assigned to each of these were three Things. Residents of a Quarter were free to choose membership in any of the nine or twelve Things attached (through the Varthings) to their Quarter-Thing. Membership in a Thing determined who your Godhi or chieftain was; a Godhi protected his Thingmen against local threats, appointed judges from his Thing to serve on the judiciary, and represented his Thingmen in the national legislature. In return, a Godhi’s Thingmen paid him fees and did him various favors. One could officially switch one’s membership from one Thing to another simply by making the appropriate announcement in front of witnesses. Since the cost of transferring one’s allegiance to another Godhi was far smaller than it would have been if the Things had been purely territorial entities, competition put a brake on the ability of any Godhi to oppress his Thingmen too severely or to demand excessive favors or tribute. ! This decentralized system appears to have been quite effective. The Icelandic Free Commonwealth did eventually succumb to centralization, but it took three hundred years; the United States took much less.

(For more information on the Icelandic system, see Jesse Byock’s Medieval Iceland, William Miller’s Bloodtaking and Peacemaking, and David Friedman’s “Private Creation and Enforcement of Law: A Historical Case” (Journal of Legal Studies 8, 1979). For a historical survey of similar systems, see Bruce Benson’s Enterprise of Law, and bibliographic essays by Tom Bell and Albert Loan in Humane Studies Review 7, no. 1 (1991/92).)

Virtual Cantons

The Icelandic case has been a popular model among libertarian defenders of market anarchism. But it’s important to see that it offers valuable lessons to framers of governments as well. Within the framework of a state, the divorce of jurisdiction from geography is not an option at the national level; but it remains a very live option at the local level. Just as a nation may be divided into many small geographically distinct cantons for purposes of local government and national representation, so it might also be divided into analogous political units that had no territorial significance. These might be called “virtual cantons.”

Two Functions of Virtual Cantons

Like the Icelandic Things, the virtual cantons would have two functions: representation at the national level, and government at the local level (with “local” now serving as a structural rather than a geographical concept).

In the first case, each virtual canton would send a representative to the national legislature. Citizens would be free to switch their allegiance to another canton whenever they chose, without having to change their residence; so it would be like being able to live in New York and yet choose a Senator from Arizona as one’s representative. Also advisable would be a constitutional provision allowing any group of citizens above a certain size to form a new canton. (The lack of this crucial feature proved to be a fatal flaw in the Icelandic system: since Godhordh, or seats on the legislature, were marketable commodities, it eventually became possible for a small number of families (who had acquired their wealth by being recipients of tithe taxes, which, unlike Godhi fees, divorced income from accountability — a result of Iceland’s forcible conversion to Christianity) to buy up these Godhordh and monopolize the legislature. The Icelandic constitution had no provision for the creation of new Godhordh to counteract this threat.)

On the “local” level, each virtual canton would pass its own laws and provide its own enforcement. Citizens would be subject to the laws of the national legislature and to those of their own canton, but not to those of other cantons. A principal job of the national government would be to regulate relationships among cantons, laying down guidelines for the adjudication of disputes among members of different cantons, resolving conflicts between laws of different cantons, and so forth. But within that nationally-determined framework, there would be free competition among virtual cantons.

Such competition would have many benefits. The threat of losing “customers” would push taxes and wasteful spending far below their current monopoly levels. The presence of alternatives would also lower the incidence of government oppression by linking revenue with accountability. (Imagine, for example, how quickly the LAPD would have lost paying customers after the Rodney King beating if rival law enforcement agencies in the same area had been competing for public support.)

A virtual-canton system is also more fair than a purely majoritarian system. Under majority rule, if 51% of the population favors law X and 49% favors law Y, then law X is imposed on everyone, including the dissenting minority. In short, majority rule creates negative externalities for minorities. A system of virtual cantons would help to internalize these externalities: the minority opposed to law X need not be subjected to it, but may instead join a virtual canton offering law Y. Those in the majority cannot conscript the minority into supporting their projects (or vice versa), but must bear the full costs themselves.

Virtual cantons also provide a better check against local tyranny than do other decentralized systems. In a territorial system, those in a given geographical region may find voting with their feet prohibitively costly, and so must suffer whatever the local government decides to impose; the option to change cantons without changing residency offers the functional equivalent of voting with one’s feet, at a far cheaper rate. In general, virtual cantons would provide far more effective checks and balances than those among the three branches of government in the Federal system, because of the competitive potentialities (to which the Federal system has no analogue) of switching between cantons or creating new ones.

The virtual canton system also solves information problems: would cantons tend to divide at all along geographical boundaries, or not? would they vary greatly in size? how many would there be? Competition would allow the market to determine the optimal answers to meet citizens’ needs.

National Government

The national government has a vital role to play in providing coordination among the policies of the various cantons. Yet it must be severely restricted in its powers, or the whole purpose of decentralization will be defeated. If the national government, rather than the cantons, becomes the chief locus of decision-making, then the competition among jurisdictions will become otiose, and the cantons will degenerate into special interest blocs vying for centralized power. Hence national powers must be more severely constrained (not just by a written wish list, but structurally) than canton powers, in order to force most political disputes down to the canton level (and thus onto the competitive market).

There are a number of ways of doing this: severe term limits, supermajority requirements, a plural executive, etc. One promising suggestion, offered by Robert Heinlein in his novel The Moon is A Harsh Mistress, would be a bicameral legislature: one house requiring a two-thirds vote to pass laws, the other a one-third vote to repeal laws. It might be especially useful to compose the former of canton representatives (thus ensuring maximum participation of various interest groups in the legislative process) and the latter of popularly elected representatives (thus accommodating Isabel Paterson’s principle, in God of the Machine, that any stable political regime must provide an official conduit for the masses to exercise a veto power). A weak central power and thriving virtual cantons could combine the best of anarchy and limited government.