A different question about federalism

Matt Yglesias writes:

States seem to differ primarily in how they deal with some fairly trivial regulatory matters. Each state's rules governing alcoholic beverages differ somewhat from its neighbors, cigarette taxes and where (if ever) you're permitted to smoke indoors vary, but you don't see a ton of policy variation. No state, no matter how right-wing, has just voted to dismantle its public school system nor have we seen a state attempt single-payer health care. I wonder if this is parasitic on the fact that there's shockingly little institutional variation among American states.

US federalism is somewhat unusual in that the states have essentially total autonomy in terms of how they want to arrange the institutions of state government. The federal constitution only contains a vague requirement of a "Republican form of government" which seems to offer a lot of leeway. Nevertheless, 49 out of 50 states choose bicameralism. Zero states out of fifty opt for parliamentary-style governance where the state executive must maintain the confidence of the legislature. All fifty states, including tiny Rhode Island, implement a an interstitial country (or "parish") level of government between the state and towns and cities. All the states elect their legislators on the basis of single-member constituencies. You'd think that some state, at least, would try something different along some of these dimensions and see how it works out.

As a number of his commentators note (and see also Will Baude), there really is a fair amount of variation o those policy questions that haven't been taken away from the states by either Congressional preemption or federal judicial constraints. And some of the range of permissible institutional variation got removed in the 1940s-60s. (This is one of those assaults on federalism that I think was necessary to break Jim Crow but that we should still recognize was constitutionally costly-- ideally the 14th Amendment, 15th Amendment, and Voting Rights Act should not be interpreted to constrain institutional choice as tightly as they have been read in the last half-century, but otherwise southern states seemed likely to use their institutional leeway to concoct further ways to keep blacks out of political power in perpetuity.) Moreover, the different systems fo selecting judges and the different rules about initiatives, referenda, and recalls do make for very different kinds of political systems.

But still-- he's right, there's been no radically green or libertarian state government, no state government that was operationally socialist no matter how aspirationally socialist some of the old Scandianvian midwestern states were, etc. And every state has a separately-elected unitary governor, only one state has a unicameral legislature, no state has tried to get a PR or statewide STV system through VRA approval. Two observations.

One is that there was much more institutional variation during the period 1776-89. After the federal constitution was ratified, it became a focal point for how Americans thought about constitutional organization.

The second is that, as far as I can tell, this is actually very common in federations. So I doubt that "shockingly little" and that "somewhat unusual." I think the U.S. is typical here. I know a lot about a lot of federal systems (though of course not everything about all of them) and this kind of isomorphism between provincial governments and the federal government is, to the best of my knowledge, universal. To take the simplest case: I don't know of any federation with a presidential form of government that has even one province with a parliamentary one, or of any federation with a parliamentary form of governance that has even one province with an independently-elected executive.

Some federations do constrain institutional choice more than the "republican guaranty" clause does in the U.S. constitution-- India, for example-- but in general the constraint just seems to be familiarity. There might also be party issues at stake. Federations are hard on parties to begin with-- any parties that aren't explicitly provincial/ regional (e.g. the Parti Quebecois) have to juggle their national and their provincial positions, trying to appeal to the very different median voters of each province severally as well as of the country as a whole. If parties had to compete in completely different electoral environments from one province to another, I imagine that the task would become hopeless. First past the post rules send you toward the median voter; proportional representation rules and STV rules mandate very different strategies. If I were a political party, I wouldn't want my strategic calculations to be rendered impossible like that. So the dominant political parties in any system might have a strong interest in isomorphism between the central and the provincial governments.

Indeed as a constitutional designer I'd worry that different electoral systems from state to state would encourage the growth of very different party systems at the state and the federal levels. (This turns out to be bad. See Mikhail Filippov, Peter C. Ordeshook and and Olga Shvetsova, Designing federalism: A theory of self-sustainable federal institutions, one of the best political science books on comparative federalism.) But that worry can't explain the isomorphism, just justify it. I suspect the explanation lies in some combination of party self-interest and sheer familiarity.