...Archive for February 2021

Fix the Senate III: Stochastic Gong Show

American government has an obvious problem. The only elected official held accountable for whether the government as a whole is effective is the President. But the President, under our checked-and-balanced system, cannot govern effectively without the active cooperation of Congress.

For members of Congress, how effectively we are governed during their tenure is basically a non-issue. Representatives are held accountable not for how the country is doing, but for the bills they personally sponsor and for how they individually vote. They have every incentive to pander pointlessly to key constituencies and to avoid contentious votes that might anger people, regardless of how essential the issue. Ultimately, the public interest is bound to the portfolio of legislation a Congress produces, while a legislator’s interest is bound to the popularity of individual bills and votes they become identified with. And to the unpopularity of votes they can avoid becoming identified with — much of why we have an “imperial” Presidency is because, on matters like war and peace for which Congress is Constitutionally responsible, our representatives rationally abdicate. They prefer to cede hard calls to the executive, so they cannot be blamed for whatever happens.

This helps explain a stylized fact of American politics: The public detests Congress as a whole, but loves their own representatives. Each member of Congress tailors her actions to keep the love of her own constituency, and usually succeeds. But all that preening and ducking of controversy fails to compose to an intelligent portfolio of legislation, conducive of high-quality government. Even when a single party controls Congress and the Presidency, for most representatives, the electoral cost or benefit of how well the country is governed under their party’s “brand” is modest compared to the personalized costs of a vote that might upset important constituents.

We would be better governed if Congresspeople had a stronger stake in the success of their legislative bodies as a whole than they had in their personal popularity with constituents. But we have no institutions that rewards elected officials for the collective successes, or punishes for the collective failures, of the bodies they constitute. We should create such institutions.

Suppose that at every Federal election, a yes/no question were placed on the ballot. “Do you approve of the job the US Congress is doing?” The result would be aggregated, one person one vote, into a national approval rating. Senators are ordinarily guaranteed six-year terms. But suppose (following the Constitutional amendment that would be required to enact this) each just-reelected or two-year-in Senator stood in jeopardy of losing their protracted terms, and being forced to stand again for election in just two years. For each such Senator, we’d perform a lottery in which the probability of having their terms abruptly shortened would be (1 - approval_rating). If Congress had a 100% approval rating, then senators could be as secure in their jobs as they are today. If Congress has only a 33% approval rating, however, then two-thirds of incumbent senators who might have looked forward to four or six years of job security would find themselves thrown untimely before the tribunal of the people. This would create a strong incentive for Senators to govern in ways that not only endear them to their own constituents, but also persuades the national public that Congress as a whole is discharging its duty of representation and governance well. [*]

I call this idea a “Stochastic Gong Show”, after the television variety show during which judges would cut poor performances short by banging on a gong. We would all stand in judgment of Congress, as we should. But rather than deterministically ending bad numbers with a bonk, we’d decide the risk our players face if they fail to choreograph a dance that delights the public.

An objection might be that it would only apply to senators, who would be judged for the performance of Congress as a whole. But maybe that’s okay. The Senate in practice is the more powerful house of Congress, populated by the more senior members of each party who are not without influence over their colleagues in the other chamber. If Senators really need Congress to work, they can go a long way towards making it happen.

A practical objection is that, under the only Constitutional amendment process ever thus far used, a supermajority of the Senate would have to approve adoption of this proposal, whose function and intent is to make senators uncomfortable, for the good of the nation. However, anything is possible if the public is sufficiently convinced it’s a good idea. “Stochastic Gong Show” sounds like a dorky thing attractive to people like me (and not just) who think injecting some randomness into democracy could do a lot of good. But the basic idea of a mechanism that automatically throws the bums out when, overall, we agree they’re doing a crappy job, strikes me as one that could be popular.


Notes: I first presented this idea a couple of years ago. It owes something to this proposal by Robert Merkle, who similarly proposes using an approval rating to drive political institutions, though he goes well beyond that, proposing (after Robin Hanson’s “futarchy“) that policy choices should be driven by market-predicted approval ratings conditional upon adoption.

[*] If this level of risk seems too strong, the formula could be generalized to 1-(k+approval_rating)/(k+1), for a chosen value of k. If k is zero, we have the original formula. If k were 1 and Congress’ approval rating was 33%, each senator’s risk of a shortened term would be 1-(1+1/3)/2, 1/3 rather than 2/3.


Office Hours: I’ve taken to doing Zoom office hours on Friday afternoons, 12pm Pacific / 3pm Eastern / 8pm. If you’d like to join, let me know by e-mail or Twitter DM or in the comments here and I’ll send you an invite. (If you use your real email when leaving a comment, I’ll have it but it won’t be published.)


Fix the Senate II: Integrate

One way to address the absurdly disproportionate representation of the US Senate is to take it as a challenge, rather than a problem. The broken Senate does not in fact represent some kind of “wisdom of the Founders”. They knew it was a mess. Here is Alexander Hamilton (as recorded by James Madison), opining presciently:

Another destructive ingredient in the plan, is that equality of suffrage which is so much desired by the small States. It is not in human nature that Va. & the large States should consent to it, or if they did that they shd. long abide by it. It shocks too much the ideas of Justice, and every human feeling. Bad principles in a Govt. tho slow are sure in their operation, and will gradually destroy it.

States were given equal Senate representation despite vast differences in population because the ratification of the Constitution required smaller states to sign on, and without this concession they would not. There was not much pretense of principle in the thing. It was a politically necessary compromise.

Nevertheless, in the spirit of lemons to lemonade, let’s suppose there could be some virtue in the arrangement. What might it be? Well, the disproportionality of representation is only harmful to the degree that smaller states as a group have interests at variance with larger states. The political party whose power currently depends upon the Constitution’s several disproportionalities has taken to a rhetoric of the “left behind”. If one wanted to impute wisdom to what in fact was crass compromise, one might argue that the purpose of the Senate’s “equal Suffrage” of states is to make sure that smaller political subunits, which are individually less influential, could not as a group simply be ignored or left behind by the centers of population and industry. That comes to “shock…the ideas of Justice” only when there are stark divergences in values and interest between political subunits sorted along the dimension of population size. If voters in small states (as a group) behaved similarly to voters in big states, the Senate’s disproportionate enfranchisement of small states wouldn’t matter. There shouldn’t be two Americas, one predominant in small states, another in big states, whose values and interests sharply differ. Our fortunes should rise and fall together. When the states’ equal Suffrage in the Senate is a live issue, it means that we have failed, as a country, to integrate.

As a practical matter, how could we cure this divergence between the interests of small states (as a group) and larger states? Two approaches are obvious. The simplest but not sweetest approach is to simply augment the existing group of small states with new states whose values and interests would make the full complement more closely resemble big states. Under contemporary partisanship, if we make a city-state of Washington DC, the mostly Republican politics of small states as a group would shift a bit towards Democrats, because the urban density that would characterize this new state is a key predictor of Democratic vs Republican politics. If Puerto Rico were made a state, the ethnoracial demographics of small states as a group would come to resemble larger states and the nation as a whole more closely. [*]

Diluting away the differences between smaller states and larger states is a straightforward approach to remedying the Senate’s legitimacy crisis, but it is also an ugly way. It leaves citizens of current smaller states understandably cynical that beneath all the high-minded talk of proportionate enfranchisement, the goal is simply to override their interests in a sphere where they happen to have an advantage. Statehood for DC and Puerto Rico are just causes on their own terms. Regardless of their partisan inclinations, the populations of those territories ought not be disenfranchised. But gerrymandering (or degerrymandering) the Senate for partisan advantage is zero-sum hardball politics. I’ll take it if the alternative is the other guy’s hardball bonks me on the head. But I’d prefer other ways.

A better, but slower and harder, way of addressing the Senate’s legitimacy crisis is to address the causes of the divergence of interest between small and large states that leads to divergent and extreme partisanship. Strong predictors of US partisanship now include urban versus rural, more versus less formal education, less versus more religious, nonwhite versus nwhite, young versus old, and female versus male. Controlling for these other factors, “poor” versus “rich” is not, I think, a reliable predictor of party identity in the United States. (Am I wrong?) That strikes me as extraordinary in a country where one party still gets coded “left” and the other “right”. Furthermore, at the state rather than individual level, affluence (in per-capita GDP terms) correlates with Democratic rather than Republican party identification, and smaller states are likely to be poorer. The Senate does, in this sense, overrepresent Republicans as the left-behind.

There are lots of ways that groups of states could diverge from one another, but if we are interested in addressing the Senate’s legitimacy crisis, we should want to encourage smaller and larger states to become more alike along these dimensions that become entrenched in partisan politics.

Robert Manduca has a wonderful essay on “place-conscious” policy. “Help people, not places” is an econowonk catchphrase that is finally, fortunately, finding its way to the rubbish heap. For most people, the costs of relocation are high. Humans are not like the atomic individuals economists commonly model. That academics and career-driven professionals are geographically mobile renders them an exception they too often mistake for normal, or impose as a norm. You cannot help most people independently of the places and communities to which they belong. To require geographic mobility — interneighborhood, let alone interstate — is to impose harm upon the lived communities that substitute for capital among all but the most affluent classes. If your values, like mine, are liberal, you cannot help but want to enable geographic mobility. To be sure, the space between “require” and “enable, between coerced and voluntary, is always a spectrum, a continuum, a slippery slope. Still, we should strive to create a world in which, for most people, staying where you’re from is a fine option, but it’s easy to go elsewhere too. And (contentiously when you think about it this way) we should strive to bring more people into the affluent classes, so that people need to rely less upon particular relationships for risk management, so that the freedom whose flip side is rootlessness becomes less harmful.

We should want poorer states to become richer, in per-capita terms, and the divergence of prosperity across states to decline. Just as the law in its majesty forbids rich and poor alike to sleep under bridges, liberalism in its majesty enables rich and poor alike to uproot themselves and choose their own communities and lifestyles. If Mississippi were as prosperous as California, a greater share of citizens in Mississippi would become more “liberal” in the simple sense of valuing choices that liberals value. But those choices mean little to those who cannot exercise them, because they are not rich enough, or because they are relied upon by others who cannot afford to depend solely on the market for support. There are lots of policies that could help to equalize prosperity across places, including a UBI (whose level, importantly, would not vary by local cost of living), a high Federal minimum wage; and airline reregulation that prioritizes universal service at similar rates, rather than “efficiency” and cheap fares for the most competitive markets.

Smaller states are disproportionately rural and low density, while larger states are disproportionately urban and high density, contributing to (or at least predicting) their partisan tilt. Every state of the union has cities, though. Adopting policies that encourage development and growth of smaller and midsize cities, perhaps at the population expense of contemporary superstar cities in high population states, would help bridge the partisan divide. The experience of the pandemic, and the normalization of remote work, may go some way towards this result even without new policy. Matt Yglesias points out that (absent some subsidy through policy), refugees from superstar cities are more likely to move towards places with a low cost of living and nice weather or tourist amenities than they are to the struggling post-industrial cities that policymakers often seek to “revive”. But from the perspective of integrating small and large states, that’s probably okay. Nice weather and tourist amenities are much less concentrated in high population states than superstar cities are, and smaller states tend to be cheaper to live in. Tesla and Oracle are huffily relocating their headquarters from California to Texas, and some Silicon Valley VCs are making a big show of migrating to Miami. The irony is that while principals and VCs are looking for red-state amenities — lower taxes and lighter regulation — they will carry with them large workforces of educated professionals, who are likely to shift the political cultures of their new homes blue. The result will probably be more “purple states”, which are great for national cohesion. But it would be even better if this tech exodus to red, cheap states alit on smaller, warm states like coastal Mississippi, or tourism havens like Maine, Montana, Idaho, and Utah, to develop new headquarters. Wyoming has tailored a regulatory environment friendly to blockchain fintech, but that seems to have drawn corporate domiciles more than human enterprise so far. Wyoming is beautiful! Go forth, my cryptolibertarian friends, and actually build your cryptoutopia there! Smaller, poorer states are friendlier to new development than more crowded and prosperous states, and several have stunning places not as resort-famous. The Dakotas, Arkansas, West Virginia are all great candidates to develop de novo tech hubs attractive to megacity refugees. Small-state cities are also great candidates for diversifying the geographic footprint of the Federal government. Conversely, we should think about ways of strengthening (perhaps subsidizing) rural small towns. That would both diminish the prosperity gap between smaller more rural states and large states, and it would revitalize rural areas in big states, decreasing the urbanization gap from the other direction.

Educational “attainment” tends to distinguish smaller states from large. Noah Smith has argued that non-“elite” colleges and universities could be key nuclei for revitalizing rural parts of the United States. Smith emphasizes the stimulative effect of university research rather than teaching, pointing out that undergraduates often move away after graduation. But I don’t think that’s so true of less elite institutions. The United States’ community college system is a tremendously underrated asset that could help bridge a wide variety of social divides. If we allowed community colleges to earn accreditation to confer four-year degrees, they could do a tremendous amount to overcome education polarization, working within more geographically rooted communities. A Federal program aimed at expanding the scope and reach of community colleges (which could and should be made tuition free) would be popular throughout the country. On both cultural and local-economic grounds, supporting community colleges is more broadly acceptable than supporting flagship state research universities (or elite private higher ed). Small state Senators would I think support a community college expansion, particularly if it included some targeting towards places where the share of college graduates is low. The net effect of such a program would be to reduce the polarization of educational attainment between small states and large.

Ethnoracial and religious polarization are harder to address, because the legitimacy of policy that directly targets these “identities” is contentious in the United States. Policies like those described above would, I think, also encourage demographic convergence. If we did want to tackle demographics more directly, I still think “neoliberal desegregation” might be a good idea.

This post has become a laundry list of policy suggestions. But let’s pull back again to the big picture. The design of the United States Senate means that, if there are systematic divergences of values and interest between small and large states, the nation will be subject to legitimacy crises. On the one hand, the remarkably disproportionate influence of smaller states in the Senate makes a mockery of one-person, one-vote democracy. On the other hand, the authors of our Constitutional adopted this framework eyes-wide-open, precisely because small states would not consent to joining a union in which their voices would be consistently overwhelmed. Altering the “equal Suffrage” of states in the Senate is foreclosed even by Constitutional amendment. The only way to mitigate this tendency towards corrosive crisis is to ensure that differences of interest between larger and smaller states are generally modest. When, as now, those differences become large, the stability of the nation requires that they be addressed. One way they could be addressed is by adding states, small states whose values and interests are like existing big states, or big states whose values and interests resemble existing small states. But that’s a bit ugly, as it seeks national comity by overriding the preferences of existing states, diluting them into a country where they might “democratically” be ignored. Alternatively, there are policies, including place-based economic development, support for midsize cities and small-towns, and expansion of community education, that might be welcomed in states large and small, while reducing the divergences that threaten the democratic character of our union. We should pursue such policies aggressively.


[*] However, I don’t think it’s remotely clear that Puerto Rico would favor Democrats. Two out of the last three of the island’s non-voting House delegates identified as Republican, and education polarization would place the island squarely among the red states. Statehood for Puerto Rico would be good for the US in part because Republicans would contend for it, and would be a better party for having to contend for it.


Note: I accidentally hit publish while I was still editing this piece, then reverted it to draft status. If you happened to see the version I let briefly slip, there’ve been more and more substantial changes than I’d usually allow without an “update history”. The intended “final” publication time was 9:28 EST / 6:18 PST on February 2, 2021.