All government in two (and a half) layers

Every layer of democracy – local, regional, national, super national – is another set of institutions to understand alongside their competencies, and another set of incumbents to evaluate. Very commonly, voters end up evaluating all layers against the primary national layer, either reacting against it or expressing support for it.

The platonic ideal would be to condense all government competence in as few layers as possible, namely two. The British Centre for Cities has provided a plan to divide England in such a way, with a single level of local councils below the national level, in the vein of Wales, Scotland, and Ireland. Only metropolitan areas get two layers: individual districts and metropolitan governments.

(I do worry that this is primarily a cognitive bias I have in favour of “cleanliness” and nice, clean categorization. But it is good for accountability for things to be very clear, and in any case I’m in the 99th percentile of being obsessed with politics, so if there’s anything that seems needlessly complex to me, it definitely is for everyone else.)

Reducing government to two levels also allows for the addition of a higher third level without much difficulty: the (con-)federation, such as Canada or the EU. Is the US a nation or federation of nations? That is the million-dollar question.

Important note: there can be politically relevant geographies other than these two layers: local councils should be divided into electoral districts, public services can be managed at smaller levels, and on the other side, multiple councils can coordinate joint ventures (although border should be drawn so that as much as possible can be managed within them).

Indeed, the 3-5 local councillors from a particular district can/should organize general town meetings or get advisory councils made up of citizens from their areas. All I’m saying is that there shouldn’t be multiple layers of councils sharing local responsibilities, because this spawns confusion about who is allocated which responsibility, which officials can take advantage of by shifting blame.

(Yes, I am from New York, how could you tell?)

To put some concrete numbers, the local councils in the Centre for Cities’ plan are each 300-800 thousand people. Using the cube-root law (which is usually applied to national politics but spits out reasonable numbers here too), that creates councils of 67-93 members. Each member represents 4,481-8,618 residents.

In urban areas, this ratio seems adequate. It’s much less than the ratio in my current district in Tokyo (15k), while Montreal has 1 councillor per 16-21k residents. But in rural areas, I suspect it probably isn’t. Rural councils should have a much smaller citizens-per-MP ratio, possibly by introducing an area term in the equation. (Incidentally, in a smaller council with 100k residents, the cube-root ratio is 2,154 residents per councillor.)

(I do wonder if the English case is particularly un-generalizable: it has relatively few areas that are completely sparse, with no small cities at all, while most other large countries do. On the other hand, recall that Scotland has a single-tier council system, and the largest council, the Highland council, has 230,000 people in an area the size of Maryland.)

Two further issues to discuss: one is whether the metro council should also be of a size that follows the cube-root law, or be supplementary in nature (and thus small). The Greater London Assembly, for instance, has 25 seats, because borough councils do much of the everyday governing.

The other is whether executives should be directly elected or rely on a council majority: the Centre for Cities wants directly elected executives, which I have a strong instinctive aversion to. But it’s probably also true that coalition-forming dynamics in national politics don’t apply to local elections at all, leading to different recommendations.

And lastly, when drawing borders, an important criterion from the Centre for Cities is to put home and workplace under the same government for as many residents as possible, which seems like a reasonable measure to avoid collective-action problems.

In Japan, replicating this plan would mean creating a single layer that’s smaller than the 82-ish former provinces but larger than the 600-ish former districts, with an ideal of 200-300 units. (I have no illusions that this will ever happen, given how married the Japanese are to the number 47 for the number of prefectures.) In France, this would mean consolidating everything to the department and while keeping metropolises united. In Italy, the provinces. Some boundary reworking would be necessary in all cases.

In the cases of France and Italy, I realize this is wading into a minefield, so I glanced through several articles about the question of abolishing regions or departments/provinces. From what I see, it’s obviously not a popular idea, but it isn’t super-bonkers, and I’m happy to defend it. (The subject, from either direction, is largely argued in terms of saving public funds, which is of secondary importance in my personal view.)

The question of abolishing municipalities, on the other hand, doesn’t seem to be even discussed; at most, people call for merging the least-populous ones. Which is understandable!