(initially April 2022; revised August 2023)
I’m planning for this to be a running list of prescriptions about what sort of democratic institutions are optimal, hopefully adding citations as I go through the literature, and hopefully keeping a record of where I’ve changed my mind and why.
1. Legislative supremacy. Growing up in the American education system, I of course learned of the virtues of the three branches of government, co-equal and balancing each other out. As I’ll mention below, there is some merit to this, but at this point I value the balancing game between parties that is inherent to coalition government over the balancing game between branches. If multiple branches are elected, then you either have 1) unified government, where checks and balances are few, or 2) gridlocked government, where the opposition has no incentive to cooperate with the government.
Instead, the legislature should be supreme, because only it can be proportional. And proportional party politics is the most difficult way to form a majority (compared to directly-elected positions), and it has its own set of feedback loops. Radical parties will decide they want to be taken seriously as governing partners and moderate. Populist parties will need to play some role in governing if they get too popular, at which point they have to start having an actual agenda. When policy areas get neglected, parties emphasizing them will gain votes and play larger roles in negotiations. When newcomers spring up, incumbents will take note of the reasons for their support, and shift their own positions.
It’s safer to trust that those feedback loops will fulfill their functions, than it is to place any significant amount of power in more centralized branches (elected presidencies) or branches with uncertain accountability (many judiciaries).
Some combination of a majoritarian institution and a proportional institution (say, a Presidency + a proportional legislature) wouldn’t really work either, since by observation, there is only ever one first-order election that takes up most of the attention, and it usually ends up being the one with directly-elected candidates. Directly-elected candidates usually end up being the ones held accountable for outcomes, regardless of the actual institutional setup.
(There’s a fuller case for proportional representation that I’ve been meaning to write out for a while now: it lowers the stakes and reduces the risk; it allows different types of parties to make different types of appeals; it doesn’t distinguish between safe seats and swing seats (a dynamic that privileges suburbs over cities); and it gives more value to the process of deliberating, rather than submitting pre-made manifestos for an up-or-down vote.)
2. A Senate that delays. To quote the old adage, ”A passive upper house is useless, an active upper house is meddlesome.” An upper house is pointless if it replicates the representation structure of the lower house (which should be broadly proportional); but allowing any other representation structure to gain equal weight as the lower house would be undemocratic.
The upper house’s principal raison d’être is as ”sober second thought”, and as such it should have the power to delay legislation for a year or so, but not to overturn something the more democratic house was elected to do. An appropriate membership for this role might be subject-matter experts (as in Canada, or France’s Conseil économique, social, et environmental), appointed by Parliament.
Alternately: a second chamber of random citizens (added in 2023). I increasingly buy into the notion of deliberation as a source of legitimacy. A lot of partisanship and ideological animosity is the result of our primal tribal instincts, to identify ourselves as belonging to a group in opposition to other groups; these tribal attitudes harden as we watch representatives of “our side” attack and by attacked by those of “the other side”. A political arena that is dominated by these dynamics has negative effects – it means that the most divisive issues with the least space for accommodation consistently get the most attention (often the exact same issues across decades), and the strength of partisan loyalty gives elite ideologues space to pursue policies that are out of sync with the population.
Getting randomly-selected average people to deliberate – hear experts argue both sides and weigh the trade-offs, and also watch oft-derided “career politicians” grapple with the same decisions – might result in meaningfully better decision-making and stronger respect for democratic procedure. In any case, I’d be happy to try it and see.
3. A notwithstanding clause? While we’re discussing how at the core of democracy is a tension between expertise and representativeness: the judiciary! Obviously the stakes here – the degree to which individuals have rights – are much higher, which is why the norm here has been to defer to expertise (lawyers/judges) much more. But the judicialization of politics never sits right with me, leaving us spectators of arbitrary decisions from up high, with no recourse in case we find something to be wrongly decided. (And in America, where I grew up, things were wrongly decided all the time, often purposefully and brazenly.)
There’s also the angle of realism: judiciaries aren’t actually detached from the political system, and if a judiciary (whose word cannot be contravened) stays proudly and dramatically out of sync with the populace, this usually invites a backlash. Populist governments campaign on taking it down entirely, and then it cannot fulfill any part of its responsibilities. This has already happened in countries such as Hungary, Poland, Israel, and the US.
So we need a compromise. The Notwithstanding Clause allows elected legislatures to overturn court decisions, but only by expending tremendous political capital every few years. With consensus-based coalition government instead of Canadian-style majoritarianism, the barrier to implementing “notwithstanding” legislation would be much higher. Further restrictions would also be good: a delay of one year before any “notwithstanding” legislation takes effect, for instance.
I know, however, that a great deal of social progress in the US was achieved only through the judiciary, with Brown v. Board of Education particularly striking me as something that no other branch of government could have forced. There are theoretically elegant institutional solutions to the American South – racial power-sharing in particular – so the point is, I suppose, that America in particular needs various changes before contemplating an override clause. (I will note, though, that the American judiciary has since been overtaken by partisanship, and is now reversing rights-affirming decisions like Roe.)
I will note lastly that I come from three countries whose political cultures steer clear of constitutional revision for various reasons (Japan, the US, and Canada). It might be that in more functional systems, responding to a judicial decision with a constitutional amendment is actually a workable check and balance.
4. No referenda. An extension to my aversion to single-member districts is an aversion to all ballot papers with only two options. This has proven itself to be a path to an oversimplified and antagonistic tribalism. Democracy should be a constant tug-of-war between various issue positions and various priorities.
Now, I am aware that several countries conduct a steady stream of frequent referenda on banal issues without tearing apart their citizenries. I will have to do more research on when referendums take on antagonistic characteristics. (I do not consider California’s endless stream of local and state ballot questions to be a model to emulate.)
Maybe a condition where constitutional reform requires a majority in a referendum is justifiable. Requiring majorities (supermajorities?) in two consecutive parliaments is also an option, but even leaving the possibility open for defections from rebellious MPs seems wrong for such an important matter.
5. 1-2 local governments, 1 national government, and 1 federal/supranational government. I describe the reason behind simplifying layers of government here. In the months since, I still have trouble imagining this to be feasible, but I still believe simplicity is desirable, and the process in England of individually splitting up large counties and gathering together metropolitan areas is probably the model to emulate.
(added in 2023) As for federal/supranational government: I am sceptical of forms of federalism where competencies are constitutionally allocated to one government or the other; average voters have demonstrated much difficulty keeping up with these. The government that people regard as the “primary” government representing “the people” (whatever the dominant interpretation of this is) should be the one with most of the power. I’ve long believed that much of America’s dysfunction stems from the mismatch between its self-conception as a nation-state and its confederal institutions.
Federalism makes sense where a set of voters feel attachment and solidarity to different national communities. Quebecers (who general feel equal attachment to both Quebec and Canada, with the latter not overwhelming the former) are a prime example, as are the constituent nations of the European Union. In these cases though, the distinction between the governments should be less a distinction in fields of competency, and more that the higher government should have a supermajority requirement, requiring consensus across the various regions. Quebecers are attached to Canada but are still wary of Canadian decisions made over Quebecers’ objections.
Note that this can be differentially true within the same federation. In these cases, asymmetric federalism should be normalized, like Britain’s devolution of power to Scotland, Wales, and Northern Ireland (except constitutionally entrenched). I realize, though, that citizenries absolutely hate even the appearance of preferential treatment.
6. Shorter parliamentary terms. Broadly speaking, shorter terms (of three years, or perhaps even two years) allow for closer monitoring and accountability between representatives and populations. It minimizes the time that an unpopular government, or an unnatural post-election coalition with limited support, can cling to power. This should be balanced against the fact that some programmes require time to properly legislate, enact, and for their effects to be felt. I’m not convinced there are many programmes that require more than three years. And in any case, even if policy effects take so long that the next election is voted with “incomplete” information, adopting proportional representation (with its smaller seat swings between elections) should minimize the effects.
Also, there are countries like Japan where elections happen every three years anyway, but with the Prime Minister deciding the timing to their own benefit. New Zealand has had the chance to extend terms from three years to four at two referenda, and both failed. And I would not say that governance in New Zealand is particularly more short-sighted, or less capable of long-term thinking, than the UK!
Koji's Posts 