Reply to Bryan Caplan on abortion nudges

UPDATE: Caplan has responded, and makes an interesting point: it looks like the turnaway study shows reduced depression and anxiety for women who are turned away from abortions, relative to women with similar physical health and financial situations. There may be problems with drawing that inference from the study design – I can’t make a clear determination on that because it looks like the most relevant publications from the study are still in press. Based on the NYT writeup, it looks like the Turnaway researchers were trying to show that mental health outcomes in people who had abortions were no worse than in people who carried the pregnancy to term – I don’t know if there are problems with using non-inferiority studies to show actual equivalence, so it seems possible that the women who had abortions actually did have better mental health and the study just wasn’t constructed/powered to detect an improvement. (If that line of reasoning is nonsense, please let me know!)
All that said, it sounds like a relevant pro-natalist point, and one that ought to be included in the shared decision making that would ideally exist around pregnancy/termination decisions. Maybe in the idealized environment I described below, it would be sufficient to tip the balance in favor of nudging women away from abortion, but it’s certainly not a strong enough independent finding to have substantial policy implications yet.

Bryan Caplan had an interesting query for Richard Thaler yesterday, which started on Twitter and then Caplan expanded in his blog post on Nudges and Abortion. In it, Caplan – who, it’s worth noting, is a strong natalist with good arguments for having kids – proposes several policies that would, in a vaguely non-coercive way, reduce access to abortions, and so (perhaps) make people better off.

The post amounts to a challenge primarily because it at least appears to be the case that most would-be “nudgers” are pro-choice, so the issue has at least the potential to test whether nudgers are serious about making people better off, or are simply using nudges as a stealth tool to push their preferred policies.

Obviously, the current legal regime around abortion complicates the discussion. I would bet that, encountering them in a modern context, most pro-choice would-be nudgers find Caplan’s suggested anti-abortion nudges revolting and even infuriating. That’s not because of anything inherent in the suggestions – they are, at least in principle, well-motivated and fair-minded. But since the modern abortion debate boils down to transparently disingenuous claims that we need to make abortion “safer for women,” and proposals for doing so that serve only to limit or eliminate abortion access so society can use pregnancy to punish women for having sex, nobody who seriously proposed the changes Caplan suggests “to make women better off” would (or should) be taken at his word.

That in itself illustrates the first problem: nudges are often best directed at decisions people make casually, and that are driven by unnoticed cognitive errors. That risk is already low when we as a society have made it deliberately difficult to pursue a particular decision. It would hardly be reasonable to support nudges to reduce chocolate consumption if you had to drive across three states and wait 24 hours to buy a Snickers bar; it’s when you’re faced with a half-dozen candy displays on an average Tuesday that slightly reducing access looks reasonable.

But imagine, for the duration of this argument, that abortion was accessed as if it was any other aspect of the medical system, with no onerous regulations beyond those deemed generally appropriate for regulating procedures with a commensurate level of risk. In such a system, it would not be inherently repugnant to suggest that decisions about abortion are subject to some of the same cognitive biases that plague decisions about saving for retirement, becoming an organ donor, or eating more fruits & vegetables.

Even in such a world, I think Caplan’s basic assumptions are misguided. From the post:

“Parents very rarely regret having children – even initially “unwanted” children.  This is not mere status quo bias: Most childless adultseventually regret not having children.  As I’ve said about parenthood before, ‘Buyer’s remorse is rare; non-buyer’s remorse is common.’  Implication: Most women who want to terminate their pregnancies would probably change their minds after their babies are born.  Most won’t go through the next eighteen years thinking, ‘I wish I’d gotten that abortion.'”

Even if all of that is accurate, it’s not especially relevant to this specific discussion. In most cases, choosing to have an abortion now is not the same as choosing never to have kids, so the comparison to childless adults is nearly irrelevant. Rather, the relevant choice is generally between “kid now” and “kid later,” bearing in mind that there’s at least some (probably predictable) chance that whatever conditions make having kids now unappealing (unstable family circumstances, poverty, being in high school/college/grad school/&c.) will later be alleviated.*

Second, the data saying that women are generally happy with their children, even after unplanned pregnancies, are unlikely to be representative of the population we’re trying to nudge. More relevant evidence comes from the recent study of women who were just barely denied abortions (vs. women who just barely got them) presents a far less rosy picture of life outcomes and mothers’ relationships with their children.

Finally, most of Caplan’s proposed “nudges” aren’t very good examples of the concept. Here they are:

1. Waiting periods: Abortions must be scheduled at least a week in advance.  This gives women time to reconsider their decision, so they don’t abort rashly.

This one is hard, because waiting periods are already part of the policy space, and so are steeped in existing messaging that women can’t be trusted to make the right decision. Moving away from that, though, this still strikes me as more coercive than a nudge ought to be. The nudge version of this would be for the default scheduling to be a week after the first appointment – patients should still be able to choose an earlier appointment if they want one. Also, it’s worth noting that abortions (particularly safe ones) can be time-sensitive, especially for people who find out they’re pregnant close to fetal viability.

2. An opt-out rule for counseling.  The libertarian paternalist could schedule all women who want an abortion for a pre-procedure session with a psychologist – or maybe just volunteer mothers who previously considered abortion.  Women who don’t want counseling would have to explicitly refuse to participate.

This one actually qualifies as a nudge, and well-implemented, could be really good policy: it’s low-cost to opt out, and offers a service that some substantial number of women would be happy to use. The tripping point is, of course, the content of the counseling (I’m skeptical of specifically involving mothers who previously considered abortion without the other side), but I’ll assume that it’s aimed at offering emotional support during a difficult decision and helping women understand and consider all of the available options, not at shaming women into not having abortions. Shared decision making about pregnancy and abortion is a great idea, and I wish it was more commonly available. Props to Caplan for suggesting it.

3. Inconvenient locations: Abortions have to be performed in remote rural hospitals.  Women who definitely want abortions will make the extra effort, but more ambivalent women will decide to keep their babies.

This illustrates another general principle about nudges: they have to be well-targeted. Restricting abortions to facilities in rural areas will obviously limit abortion access for the urban poor more than for other groups – so why is the policy targeting them? Are the urban poor more susceptible than the rural poor or suburban middle class to abortion-related cognitive biases? If not, it doesn’t make sense. (Never mind that this would be a substantial infringement on the rights of OB/GYNs to practice where they wish, and place an unconscionable burden on women whose wanted children are inviable, and who would have to schlep out to the middle of nowhere, away from their emotional support systems, to have medically necessary abortions.)

4. Deny government funding for abortion.  If the government thinks that a procedure is generally ill-advised, the first step is to refrain from encouraging it.  If people want to pay for it out of their own pocket, they’re still free to do so.

Another general principle: a good nudge is not a financial incentive (or barrier). Again, financial barriers inherently pose targeting problems, but perhaps more important is the violation of the “low-cost avoidance” rule. Limiting public funding for abortions, or soda, or coal-burning power plants is perfectly reasonable public policy, but it doesn’t really qualify as minimally invasive.

Finally, here are a few anti-abortion (or at least pro-natalist) nudges that might be closer to what Caplan was looking for:

  1. Posters at clinics with messages along the lines of: “Most couples keep their unexpected pregnancies,” “Kids can be a lot of fun,” or “You’re more ready for a baby than you think.”
  2. PSA campaign for young people considering delaying childbearing: “Your baby may be healthier if you have her now.”
  3. Condom wrappers with baby faces on them and slogans like “Don’t you want your own?” (I can imagine this doing double-duty, encouraging some young men to complain less about using condoms, and perhaps encouraging some couples to decide not to.)
  4. Placing condoms right in the front of the drugstore, where you can’t hide when looking at them, and limiting their sale to between 8 AM and 6 PM.

*You can argue that in some cases, conditions are unlikely to improve, but in that case, using Karl’s model, the abortion isn’t the rational decision in the first place. Under those circumstances, we’d more likely be talking about nudges to encourage abortion – or, preferably, nudges to reduce unwanted pregnancy.


Observations on gun culture and “self defense”

In the wake of the Newtown shooting, it looks like gun control is on the political agenda for the immediate future (at least, more than it has been in the recent past). Following each of the recent mass shootings, I’ve tended to spend some time in the days after looking at gun-related websites–everything from forums discussing the best home-defense calibers, to Guns & Ammo magazine, to the webpage for Colt’s .22 rimfire M-16 lookalike. In that time, I’ve noticed some things about the gun owners who comment on those sites–things that I thought would be useful for other people like myself to understand during the coming policy debate. This will not be a heavily data-driven post, other than my own observations, but I hope it will be useful.*

I feel some need to open with a quick statement on some of my biases.

I think it’s fair to say that my family has little use for guns. Not just my immediate family–I’m not aware of any gun owners in my entire (large) extended family. I may have missed one or two, but I doubt it. I also live in DC, and lived before that in New York City–not among the most gun-happy areas of the country. I have had little exposure to the culture of gun ownership.

But even with that background, I’m not totally ignorant about guns. I learned to shoot in Boy Scouts, and wasn’t god-awful at it–I shot .22s, shotguns, and once a percussion(?) muzzleloader. I have never shot a handgun, although I’m intrigued by the possibility. And I’ve had a lifelong fascination with weapons of all sorts, so I occasionally kill a few hours reading about guns on Wikipedia.

Finally, as a health policy guy, I need to admit that I see guns as an environmental health hazard. The evidence is pretty ambivalent on whether having more armed citizens around has much of an effect on crime, but with respect to suicide and accidental death, it’s pretty obvious that having more guns around leads to having more people dead from accidents and suicides. With that noted, my observations.

First, several beliefs that appear common among gun owners:

  1. Responsible gun owners don’t pose any threat to other people. Most of the comments I’ve seen attribute gun risk to individuals, not to guns–there are “responsible, law-abiding gun owners” and irresponsible or non-law-abiding gun owners, and the responsible ones don’t cause injuries or death. Rather, gun deaths are caused by irresponsible gun owners, the mentally ill, and criminals.That belief is true, to a certain extent–in any given year, or even over a lifetime, any given gun owner is unlikely to commit a crime or kill someone with a gun. But moving from that fact to the idea that legal guns pose no risk is obviously in conflict with the data and what I just got done saying about guns as a public health hazard. That conflict points out the deception inherent in the phrase, “Guns don’t kill–people do.” Because while it’s true that guns are not independent actors who choose to kill people with no human involvement, the fact remains that the presence of guns means more people end up dead. Because while responsible, law-abiding gun owners are unlikely to commit crimes, their guns can still be involved in accidents and suicides–and both of those risks remain, simply based on the presence of a gun.
  2. Violent crime is a common (or at least likely) occurrence. Gun forums are the only place I’ve ever seen someone use the term “gangbanger.” It’s common there, though, as a generic term for (as nearly as I can tell) people who actively are out to commit violent crimes against ordinary people. “Gangbangers” seem to be the boogeyman that will rob your house, or assault your wife or elderly grandparent in the street–if you’re not armed. Editorializing again, this seems contrary to reality. Violent crime (and particularly random street crime) is, by the standards of history and for most people in almost all areas of the US, very rare. Clearly there are areas where it’s more common, and people in particular occupations may be more likely to be targeted, but a belief in common violent crime seems necessary to make a general case that “I need to protect my family.”
  3. The police will not save you. When the gangbangers come, the police will either be (depending on which site you’re looking at) uninterested in helping you, or just not around when you need them. That means protecting yourself (and your family–these sites always invoke protecting the family, in what seems to me like a shoutout to hypertraditional family-first thinking) cannot be effectively outsourced to the power of the state.This belief is, in a certain respect, probably right. I’m not familiar with typical police response times, but there are at least some situations where criminals escape before the police show up. But I’m also not sure how much they would accomplish if they did frequently show up to, for example, a home invasion in progress. The idea that the police are supposed to provide armed protection to the citizens (but they just don’t) seems like a misunderstanding of the role of a modern police department: to deter, investigate, and prevent crime, promote public order, and catch criminals after the fact.
  4. Owning and using a gun is an effective way to protect yourself against violent crime. There are two ways that guns protect you against violent crime–by deterrence (both group and personal), and by responding to a threat. There are a couple ways this logic works. The deterrence case for widespread concealed carry is that criminals will be unlikely to attack people if they believe many of their potential victims or bystanders are armed, because they don’t want to get shot. Obviously, in this worldview, the more common concealed carry is, the better off everyone is–it’s a classic collective action problem. (This is also a relatively testable proposition–if criminals fear the general concept of an armed citizenry, there should be pretty solid evidence that overall crime goes down when more people carry concealed. The deterrence case for open carry is more direct, and harder to test–if you are openly carrying a weapon, criminals are presumably less likely to attack you, but that may simply mean they pick a softer target. In that case, there is no positive externality to more common open carry.)Finally, there’s the response case, i.e. the presumptive point of carrying a weapon (or at least a loaded one–deterrence could perhaps be accomplished without ammunition): a person who is armed can attempt to use their gun to remove the threat, either by killing the assailant, scaring them off, or maybe making a citizen’s arrest. (I may just have looked in the wrong places, but I remember seeing very little writing on what to do once you have a gangbanger/robber/rapist/whoever in your sights.)

    The data seem to suggest that homicides are not substantially affected either way by concealed carry. My recollection is, however, that there are pretty strong data saying that guns kept in the home for protection are substantially more likely to injure a member of the household than be successfully used against an intruder. That’s plausible on its face, simply because intruders are rare (as noted above), family members are frequently home, and late-night accidents and miscues are all too common. My guess, though, is that gun owners are as aware of that data as the rest of us–they simply fall prey to the Lake Wobegon Effect, and assume that their superior equipment/skills/training/awareness in the middle of the night will make them more likely to use their gun effectively.

  5. Given 2, 3, and 4, owning a gun is the responsible thing to do, to protect yourself, your home, and your family. I haven’t and can’t address the deeper philosophical and psychological differences between people who own guns for protection and those who don’t. But if these beliefs above are true, the case seems pretty obvious. I see most of these as either inaccurate, misunderstandings of the evidence, or just not terribly relevant to whether or not owning a gun makes sense, but it’s at least possible to tell a rational story about why one might choose to do so.

For more on gun owners’ beliefs, read the excellent “Tactical Reality” at TPM and this Atlantic piece. (The latter is interesting for “Registration, confiscation, genocide,” but I don’t see any reason to believe that’s a common view–it’s got more in common with Stormfront than Guns & Ammo.)

Two final observations:

  • Gun owners are not a homogeneous bunch. (Duh.) Obviously, goldbug doomsday-preppers don’t necessarily have a lot in common with antique gun collectors, who may have nothing in common with deer hunters, who perhaps don’t agree with the home-defense types on much.  Writing this has made me realize that the “beliefs” I listed above are probably most reflective of  one particular group–the people who believe in owning guns for home defense. I would not be surprised, however, if these beliefs and the arguments that come from them (“Criminals will get guns anyway, so it doesn’t make sense to take guns from law-abiding citizens who aren’t a threat anyway”) will form the core of the opposition to additional gun control.
  • Last, gun & ammunition manufacturers are having a field day with zombies. I’m not going to read into this too much, because if I owned a gun I’m sure I would be all about blowing holes in some zombies. But in the end, tactics and equipment for killing zombies are pretty similar to tactics and equipment for killing people. That makes the extent of the fantasy mildly unsettling–it goes beyond targets and forum discussions to full-scale articles on Guns & Ammo on the best anti-zombie guns, “zombie ammunition” and specially-marketed “zombie guns.” And amid the rest, I can’t help but find this target’s resemblance to a caricature of  President Obama disturbing.

Please, if I’ve misrepresented anything, or gotten some important fact wrong, point it out. As I’ve said, these are merely observations, but I do hope they’re informative.

*Given my lack of extensive data, I will endeavor not to make unreasonable generalizations. However, the whole point of this post is to generalize some of what I’ve seen. Thus, this is my disclaimer that the points I make here may not be representative of all or even most gun owners, or even of participants in gun forums. They are simply my recollections from  looking around the web.

A simple fix(?) for the carried interest question

Dylan Matthews explains carried interest.

Here’s a simple criterion to determine whether managers are earning income or investment returns: If you own an asset (and are eligible to pay cap gains taxes on it), you must have symmetrical exposure to the downside and the upside. If hedge fundies want to take that risk they can do so–they are, in that circumstance, providing the socially valuable service of risk-taking and investment, and we encourage that with a lower tax rate. If your management contract only says you get X% of investment gains, though, that’s ordinary income and you get to pay the full tax rate on it.

It seems like the issue should be more complicated than this. Am I missing something, or is this a reasonable framework?

Note: If you insure against losses, I’m inclined to think that insurance payment ought to be taxed as ordinary income. That might end up being incredibly hard to do in practice, though, and I’m really not certain how to deal with risk mitigation in general.

Machine Ethics Isn’t Actually a Problem

At least, it’s not in any way a new one.

A while ago, The Economist ran an article posing the question of how we should deal with drones and other artificial intelligences that have to make ethical decisions. Alex Tabarrok pointed out that the question is not entirely hypothetical, with the emergence of Google’s self-driving car: a truly self-driving vehicle has to decide what to do if, in a variation of the classic “trolley problem,” the car was hurtling toward a crowd of pedestrians and faced the choice to kill them or turn and kill just one pedestrian.

The issue is interesting on its face, but I don’t think the question really stands up to much additional thought. It rests on a basic misunderstanding of the nature of ethics and the nature of computers, at least as computers currently exist.

It’s useful to think about the disconnect by analogy to regulatory systems. There are two basic ways to construct, for example, financial regulations: they can be rules-based or principles-based.

A rules-based regime might say “Systemically important institutions must maintain an additional cash capital buffer equal to 3.1% of total demand deposits plus 0.8% of all debts with maturities less than 180 days. A bank or other regulated entity is considered ‘systemically important’ and subject to additional scrutiny if it has total liabilities greater than $38 billion, or is a counterparty to more than 10% of the notional value of all contracts in any of the derivative classes named in subsections c-f, below.” The rule is algorithmic–if you follow the rules precisely, you are in compliance, and there’s no uncertainty involved.

A principles-based regulatory regime, by contrast, might say “Systemically important institutions (those institutions whose default would pose a substantial risk of financial sector contagion) must maintain an additional capital buffer sufficient to maintain their solvency and liquidity during times of heightened financial stress.”* In this case, there aren’t clear lines between “OK” and “bad.” What one person honestly believes constitutes compliance may seem to another person to be in clear violation. (Obviously, there’s more room to fudge the rules to your own advantage, too, but leave that aside for the moment.)

Ethics is clearly a system of principles-based regulation. There are no readily applied algorithms that spit out answers–even once the principles are established, individuals have to use judgement to apply abstractions to specific scenarios.

That leads us to the central problem: computers do not (to my knowledge) exercise judgement. They are sets of algorithms. While we can create a computer that appears to parse the meaning of words, it does so based on a search mechanism that finds information, not based on a logical application of abstractions. (No, Toronto is still not a city in the United States.) As long as that is the case, computers don’t have any real independence–and therefore, do not make decisions that could give them moral culpability. We would not hold a misprogramed calculator responsible for poor performance on a math test–that would be the fault of the calculator’s programmer (and the unprepared math student). Similarly, we can’t hold a drone responsible for shooting what it registered as a threat, even if that threat was actually a bouncy castle.

To the extent that the emergence of more autonomously-operated technology poses any ethical problems, those problems just require us to come to some agreement on existing moral questions. The trolley problem is still the trolley problem; the decision of whether to fire a missile at a suspected insurgent based on limited information remains the same; we still have to decide how much collateral damage is acceptable. But any sort of widespread use of technologies programmed with an answer set for those situations would seem to require a set of answers that an overwhelming majority of people would find acceptable, to avoid public uproar. I suspect that widely acceptable answers sets would include a bias toward inaction, which poses its own ethical problems.**

The next question is who bears ultimate responsibility for the decisions of programmers, and what kind of accountability mechanisms are appropriate.  Translating principles-based systems into algorithms is imprecise and unpredictable–we saw that with the financial crisis. If algorithms are more often making decisions of moral relevance, we need to decide whether someone can be sent to jail for a program whose unpredicted outcomes we find objectionable. Presumably the programmer’s responsibility for a wrongful death lies somewhere between that of totally uninvolved bystander and that of an actual murderer. That seems like the new and interesting question here.

*(Yes, I made both of those passages up. Yes, I had fun doing so. I hope they illustrate the differences.)

**A commenter on Tabarrok’s blog post also notes that Google’s engineers and programmers will be more focused on avoiding the situation altogether than deciding whether to kill three grannies or one infant. That is as it should be. This probably deserves a post in itself, but in general I propose that one’s ethical duty in a real-life trolley scenario is to break the rules of the ethical dilemma (to beat the Kobayashi Maru), rather than to accept the deaths of the five workers, or of the fat man.

Karen the Collective Action Problem

I’m a little ‏up in the air about what this means, in an economic sense. Ultimately, I suppose it’s just another illustration of a collective action problem.

When you do a survey, lots of people will say “Yes, I would like to give $2 to prevent this beautiful waterfall from being strip-mined.” You get the same results if you ask if people want to spend $.50 a year to prevent pandas from going extinct (despite the pandas’ best efforts), or $8 to bring 100 people in Africa out of poverty, or whatever. But when you ask people to put their money on the table, they vote for lower taxes and lower levels of public goods like protected waterfalls & living pandas.

This scenario isn’t entirely parallel to that–after all, some people are stepping up and offering money to this bus monitor. But the campaign doesn’t acknowledge the fact that there are about 60,000 elementary schools in the US, many of which have bus monitors, many of whose lives are identically shitty. It doesn’t make any effort to push for better treatment or compensation for equally deserving people who happen to be less salient (because they haven’t had a Youtube video posted of their torments). Maybe in the end, this is just a post noting that people respond to what’s salient rather than what’s important–I’m not entirely sure . Would be interested in your thoughts.

Sentences worth reiterating ad nauseam

Meant to post this ages ago, don’t have time to comment at the moment. Hopefully blogging more this weekend.

From The Economist‘s Free Exchange blog:

Having lots of well-paid manufacturing workers isn’t the way one grows rich; replacing lots of those workers with massively productivity enhancing machines is.

A new candidate for the pantheon of crap jurisprudence

Today, in Florence v. Board of Chosen Freeholders, the Supreme Court held that detention officials may strip-search any person arrested for any reason, if that person is being held in the same room as other people. They offered as rationale only the notion that judges do not know how to run a jail, and shouldn’t underestimate the difficulty of doing so. Given that, it would apparently be irresponsible to require any level of reasonable suspicion before a strip search is allowed.

I’d been waiting for this decision for a while, and when I heard the ruling I really hoped there was more to this case. I respect most of the conservative Justices (those in the majority here), particularly Roberts and Scalia, and I really wanted there to be some obscure legal justification that would validate that respect.

There isn’t. If there was, SCOTUSBlog’s incomparable Lyle Denniston would have found it.The majority abdicated its responsibility to enforce (or even, it seems, to read) the Fourth Amendment. There is no other word for this decision than “disgrace.” Justices Thomas, Alito, Scalia, Roberts, and most of all Kennedy, you should be ashamed.

The one bright spot I see in this particular case is that the ruling is amenable to legislative remedy–it articulates merely a constitutional permission, rather than a constitutional duty. Right-minded states may enact statutory prohibitions on unreasonable searches (i.e., searches that are not founded on any reasonable suspicion). The federal government, it seems, already follows a more rational set of rules. Florence may, therefore, end up being less destructive than other recent misapplications of the law (I refer to Citizens United, which appears to require remedy by constitutional amendment). It remains, however, a gross miscarriage of justice, and may be expected to take its place in the pantheon of the Court’s most absurd and embarrassing decisions.