Saturday, July 10, 2010

Utilitarian Policy

I think the world would be a much better place if public policy debates were more focused on cost-benefit analysis. Too often, people refuse to acknowledge trade-offs or opportunity costs (health and military spending are obvious candidates here). Other times, people seem more interested in harming out-groups or "looking tough" than in actually securing better outcomes for everyone (think immigration, prisons, torture). And then there are the obvious cases of legislative capture by special interests (farm subsidies, retroactive copyright extensions). It seems like there's a lot of scope for "no brainer" policy improvements that every reasonable person should be able to agree on. But maybe I'm missing something. So let me take a stab at outlining some of the issues where the answer seems to me completely obvious -- and I hope that others will add more suggestions in the comments, and/or explain where you think I'm going wrong.

Foreign Policy and the 'War on Terror'

War is a hugely expensive negative-sum game, almost never worth getting into if you can possibly avoid it, and probably worth withdrawing from as soon as you can. (You may leave behind a mess, but surely more good could be done by reallocating the saved money to more effect forms of humanitarian aid.)

Military action also seems incredibly counterproductive to the nation's security interests: inflaming anti-American sentiment and hence aiding terrorist recruitment. Security interests would seem better served by investing in "PR" -- e.g. Fulbright scholarships for Arab students to immerse themselves in American culture before returning home to share their (hopefully positive and liberalizing) experiences. Whatever the details, the aim should be to improve America's image abroad, and especially among Muslims. (Torturing their neighbours seems unlikely to help with this.)

Also, let's stop overreacting to terrorism. We're something like twenty times more likely to be struck by lightning than to die in a terrorist attack. It's not worth clogging up airports or abandoning civil liberties over. (You could save a lot more lives by reducing the speed limit to 10 mph, but who thinks that would be worth it? Of course, we don't have front page headlines and official hand-wringing every time someone dies in a car accident.) As Brad Templeton puts it, "The goal of counter-terrorism is not to stop the terrorists from attacking and killing people, not directly. The goal of counter-terrorism is [or should be] to stop the terrorists from scaring people."

Subsidies and Corporate rent-seeking

Farm subsidies, coal and oil subsidies, etc., are obviously detrimental. Likewise retroactive copyright extensions -- which, in the absence of backwards causation, do nothing to incentivize production. IP laws more generally have swung much too far in the direction of regulation. I'm sure there's much more that could be added to this list.

It may be worthwhile, even from a "limited government" perspective, to invest in public financing of elections to help prevent such regulatory capture. (This is not such a "no brainer", admittedly, but at least highly plausible, I think.)

Medical spending

Given limited resources, we should allocate public spending so as to do the most good -- in case of medical spending, that means maximizing quality-adjusted life years. (Of course, people should be free to spend their own money on less-efficient treatments. But they shouldn't expect taxpayers to foot the bill. Personally, I'd rather die peacefully in hospice care than be plugged into life-support and fighting to the bitter end.)


... should be encouraged. However, given widespread misperceptions among conservatives that immigrants are a "burden" on the economy, it might be worth considering Will Wilkinson's proposals to end birthright citizenship and vastly increase the number of temporary worker visas instead.

(Charter cities sound like a promising idea, too. Worth exploring, at least.)


Introduce congestion pricing, price curbside parking at market-clearing rates, relax zoning regulations to allow for increased residential density, and invest in effective public transport -- basically all that sensible stuff that Matt Yglesias goes on about.

Tax and Redistribute

Increase taxes on things that are better used less: carbon / gasoline, alcohol, cigarettes, soda, etc. -- and then redistribute the proceeds.

In general, redistribution seems preferable to government spending (more efficient, respectful of individual autonomy, etc.). Implementing a full-blown universal basic income may be ideal, especially if combined with labour deregulation (abolish the minimum wage, etc.).

Criminal Justice System

End the war on drugs. Legalize (and tax) marijuana. Prison costs a lot of money and ought to be a last resort; alternative punishments should be found for non-violent offenders and others who do not pose an ongoing risk to society. The focus should be on rehabilitation rather than punishment for its own sake.

* * *

Does that all sound right? What would you add?


  1. I suppose this is just an extension of your war thesis, but the U.S. military is vastly bigger and more expensive than there is any good reason for it to be. I can't offhand think of anything else I would add, and I agree with everything you have here. You may be wrong about zombies, but that doesn't seem to have distorted your views on public policy.

  2. "I think the world would be a much better place if public policy debates were more focused on cost-benefit analysis."
    But what's the counterfactual here?

    Yes, if the median voter was a elite university graduate student in economics or analytic philosophy, the world would be radically different. However, most of the electorate isn't composed of such folk, for whom a variety of non-utilitarian appeals (going with the grain of human emotions and biases) work better. Then, the areas where the wonkish types disagree (either ideologically, or just in terms of the allocation of money, power, and patronage amonst themselves) lead to the use of appeals to 'win' at the expensive of massive policy collateral damage.

    A cross-party wonk cartel can work, and does work in some areas (sometimes for the public welfare, sometimes in bipartisan exploitation of the public), but it depends on barriers to entry to block political entrepreneurs from using the most effective rhetoric.

    Innovations to increase the intelligence and education of the electorate would seem to work in this area (Bryan Caplan's book is relevant here), and making the areas of wonk consensus clear has value, but we shouldn't forget that the marginalization of wonk ideas is endogenous.

    To your policy list I would add the systematic use of randomized trials and empiricism in public policy: rolling out new programs in randomized fashion, or locally suspending some old programs, in order to see what they're actually doing. This has been growing in recent years but is still pretty marginal.

  3. This is nothing more than a load of typical liberal/socialist ideological propaganda, full of logical argument, reasonable thinking and common sense. I find myself in complete agreement.

  4. You left out public sector rent-seeking and indiscriminate redistribution to the elderly (regardless of need or benefit, and insensitively to changes in the age distribution). Those, combined with military and medical spending, dominate rich country budgets, and cause a lot of trouble for less rich countries like Greece.

    Restrictive occupational and business licensing technically fits under your rent-seeking and corporate welfare, but its structure is quite different from cash payout, and its importance is massive. Eliminate arbitrary rent-seeking educational requirements (favoring accreditation cartels and delaying quick learners) in favor of knowledge tests.

    Shifting resources from medicine to (currently more important in driving health outcomes) non-healthcare determinants of health like diet, smoking, and lifestyle is a short-term utilitarian winner. Shifting resources from current healthcare to medical research likewise in the longer-term.

    In education, have high-stakes tests administered by outside agencies without incentive to cheat (like teachers and schools or regional governments getting funding for high scores), and making the data and analyses of it transparent.

    In criminal justice and corruption, elect or appoint internal affairs and anti-corruption officials independently of those their reports may embarrass.

    Cut massive indiscriminate housing subsidies (still ongoing in the U.S.). Like transfers to the elderly, they're broad-based, and not primarily the pet of an industry (although real estate agents are big supporters), but dwarf most industry subsidies.

    Allow government healthcare programs to pay for treatment in low-cost jurisdictions, with patients who engage in such medical tourism getting a share of the savings.

    Require that police officers be recorded at all times they may be interacting with the public, with the feed being sent to a remote location (not under the control of the police department in question, but some other agency) to prevent abuses. Protect citizens' ability to surreptitiously or publicly record all interactions with government officials.

    Permit real-money prediction markets, including policy ones.

    Collect exhaustive medical, economic, and other data on a population basis, like the Scandinavian countries, boosting public health and policy research.

    Cut tax subsidies for religion.

    Increase funding for long-term agricultural research, and reduce regulatory obstacles to agricultural biotechnology.

    Ease labor market regulations (hiring and firing, minimum wages), especially in non-Scandinavian Europe to reduce the costs of large-scale unemployment among lower social strata. Add in low-income wage subsidies if need be.

  5. The real question is how you could get this stuff enacted. Most of these policies, if you think about it, are things conservatives should embrace. (Not that they do or would.) At least in the US, we might have more success selling as many of these things as possible as right wing/libertarian/anti-big government ideas.

  6. The idea that rehabilitation can or should be the primary aim of punishment has been all but rejected in the legal community. There's no reliable way to implement it consistently. Plus, someone with an impulse to steal cars could take a lifetime to rehabilitate, while a man who commits a single murder of passion may need no rehabilitation, yet it doesn't sit well with our sense of justice to let the murderer go free and yet keep the thief locked up for life.

    Deterrence is the more realistic utilitarian policy, and it is dominant in the criminal justice system. That said, the very notion of "punishment" is shot through with retribution theory, and that will likely never change.

  7. There I more had in mind prison conditions than sentencing. For example, apparently there's research to suggest that offering drug rehab programs to inmates severely reduces the risk of recidivism, and yet relatively few places bother to implement this. We would all be safer if prisons were more geared towards enabling inmates to reintegrate with society and become productive citizens upon their eventual release.

  8. (Though, fwiw, I would much rather see the rehabilitated murderer doing something to "repay" society -- e.g. hefty fines, community service, etc. -- rather than leeching further resources by having us pay to keep them in prison.)

  9. So, here's the question as I see it: Almost anyone you ask will tell you that there are such things as "no-brainer" policy issues, i.e. issues where a policy question, while politically undecided, has a clear right answer that all reasonably intelligent, well-informed, and disinterested people should be able to agree on. We should expect, if there are such issues, that all reasonably intelligent, well-informed and disinterested people would agree on what they are--which policy questions really are no-brainers. If such people can disagree about an issue, then, more or less by definition, it's not a no-brainer. And of course the existence of disagreement on those issues, and on their no-brainer status, does not in and of itself disconfirm this expectation, because some people are stupid, poorly informed, or have a vested interest in a particular issue.

  10. The trouble, then, is two-fold: First, it's tough to determine who should count as intelligent, well-informed, etc on a given issue (for convenience, let's abbreviate call such people "qualified" wrt the issue). We all tend to see those who disagree with us as unqualified in various ways, because (a) we each tend to be aware of a distinct body of facts which support our side of an issue, and which those on the other side are likely to be ignorant of, (b) we tend to view people who say things that strike us as absurd as less intelligent than we are, and (c) on issues that break down along lines of class, race, etc, a large percentage of the disputants on either side have some vested interest in the issue.

    Second, by any reasonably expansive, impartial set of standards for qualification on policy issues, I would expect that we will almost never find *universal* agreement. Whatever we might like to believe, there is at least one (and probably quite a few) smart, well-educated climatologist out there, not in the pocket of any oil company, who still believes that global warming cannot be traced in any significant part to human activity. And ditto for almost any other question that we take to be a no-brainer. So, if an issue is truly a no-brainer, how can there be any disagreement at all concerning it among qualified persons? And if we are to continue believing in no-brainers, what level of disagreement can we accept before we are forced to the conclusion that our own views are not as obvious as they seem to us?

  11. To put the point concretely: I agree with most of the items on the above list of no-brainer policy issues. But there are one or two where, while I agree with the policy position being propounded, I don't think that the balance of evidence is decisive enough to qualify the issue as a no-brainers. And on at least one issue, I actively disagree with the putatively obvious conclusion.

    To take the second case: As a libertarian, I don't support redistributive taxes on behaviors that don't create externalized costs. So, while a redistributive gas tax does seem like a good idea (because the consumption of fossil fuels has externalized costs), a redistributive tax on cigarettes, for instance, does not (one could argue for (very light) tobacco and alcohol taxes on libertarian grounds based on externalities like second-hand smoke and drunk driving, but I assume that that's not the argument being made and it doesn't seem at all like a no-brainer). Even thought I see the undesirability of cigarette smoking, for instance, as a no-brainer, I don't think it immediately follows from that stance that government policies aimed at penalizing smokers will have net beneficial outcomes.

    Of course, it could be legitimately suggested that I'm not unqualified on the issue of cigarette taxes. Although I have no obvious vested interest (I don't smoke or own tobacco company stock), I'm also not an economist or a behavioral psychologist, so I may not be adequately well-informed to assess the potential effects of such a tax. And my general intelligence, of course, is entirely open to question. But it seems to me, at least, that by any reasonable standard which would count me (and everyone else who shares my view, i.e. at minimum almost every other libertarian) as unqualified would very severely restrict the total pool of qualified persons.

  12. That might be a good thing. Really, we're all vastly underinformed about most of the social and political issues on which we have opinions. Human behavior, especially in large groups, is extremely complex, and even those with multiple advanced degrees in the relevant sciences have a very hard time explaining and predicting it. For those of us without so much as an undergraduate social science education, our main claim to certainty comes mostly from what we believe to be our sound knowledge of expert opinion. But that knowledge is not always sound--expert opinion can be misrepresented, and can change quickly without the lay population being immediately aware of it.

    At the same time, though, unless we are happy to turn human affairs over to the wise management of the wonkish few, we want to regard a reasonable portion of the general population as qualified on the majority of issues--qualified enough, at least, to get the no-brainers right.

  13. To wrap this comment up before it crashes any servers: It seems like there are at least a few sufficient conditions which we might reasonably endorse for viewing a policy issue as a no-brainer: If the contrary position to our own is advocated exclusively, or almost exclusively, by a small minority with a clear and direct financial stake in the issue, then we may reasonably discount their views. This seems to cover issues like farm, coal and oil subsidies. Similarly, if there is some clear-cut and objective threshold of education or intelligence above which agreement on an issue becomes virtually universal (say, above 95 or 99 percent), then we can also regard it as a no-brainer except where the issue concerns the particular interests of those in that sub-group (for instance, I would guess that 95 percent of university professors agree that university professors ought to be better paid, but on this question their opinion is not authoritative).

    How many issues will truly come out as no-brainers by one of these tests, though, is open to question. If the answer turns out to be "few if any," then the takeaway might be further evidence of the proclivity we each have towards overconfidence in our own beliefs. Still, I would hope that at least a few of the issues which seem to be no-brainers truly are.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)