Author Archives: Robin Hanson

Hidden Motives of Gratitude

I recently got 568 Twitter folk to say which of these 4 things they are most thankful for:

Assume first that one is most thankful for the process which increased your gains by the highest ratio, relative to prior expectations. If we are just talking about such a degree of selectivity into a better position, then it seems to me that the first answer is obviously best: there are vastly more possible than actual creatures. If you get zero value from not existing, but a positive value from existing, and your prior odds of existing were tiny, then your gain ratio is astronomical.

However, some insist that they can only compare scenarios where they exist, and so reject the counterfactual possibility that they might not have existed. For them, the second option seems obviously the most selective among the remaining three. Humans are clearly enormously special compared to the vast number of species who will ever exist. For example, far more special than are humans in any one place and time compared to humans in others.

What if the goal of a poll respondent isn’t to identify the option from which they benefited more, but instead to use this poll as an opportunity to signal something about themselves. (Other than literal understanding and honesty.) The apparently most useful thing to signal then would be loyalty to one’s immediate associates, and confidence in one’s local roles. In which case the last poll option seems obviously best.

But oddly, 48% of poll respondents picked the third option! What goal could explain that choice? One possibility is that they sought to signal their “patriotism” toward their place and time in human history. People in different places often feel rivalrous toward each other. Different nations are in different places, and even within nations different culture, ethnic, and political “factions” tend to be in different places.

What about your time in history? Well first, there is less rivalry of feeling between different times, and less to gain by signaling loyalty to your time. Furthermore, if you think you’ve seen a trend in history toward times getting better, to make your time the best so far, you should expect the future to get even better, making your time not so great compared to all the times, past and future. So I suspect that many overt time affiliations are actually covert political affiliations.

Thus I’m led to conclude that the strongest motive in choosing what to be most thankful for is signaling loyalty to one’s region. Nationalism, racism, political alignment, and cultural rivalry. If you encourage people to be more grateful, that’s what you will be encouraging. Not what I would have guessed, but I guess not so crazy either.

GD Star Rating
a WordPress rating system
Tagged as: ,

The Pandemic Monkey Trap

Back in 2007 I said “cut medicine in half”, as its marginal value is too low. (Since then, US spending is up 40%!) But prestigious health economists said yes on average marginal value is low, but don’t I agree that some identified treatments have high value, and as there must be more like that, we should wait and not cut until we can identify which are high vs low. I say cut now, and only add back good things once you can identify them. Those hidden good treatments are the nut in a medical monkey trap, which prevents us from letting go of a larger gourd of wasteful spending.

My colleague Bryan Caplan says we should cut school spending, as its social value is on average low. Many critics say yes value may be low now, but it must be possible to create high value school programs, and so instead of cutting we should work on figuring how to increase school value. Caplan says to cut now, and only add back spending when you can actually identify high value programs.

Regarding pandemic prevention spending, both Caplan and I say that we seem to be spending way too much, and so we should cut back.

Me in October:

I’m comfortable with … an estimate of 3.0 [for the ratio of life-year value over annual income]. But … [that gives] $1.08T and $0.64T for covid death and impairment harms, which gives a prevention to harm cost ratio of 5.31. (Which is close to the 5.2 median estimate from my most recent poll.) And its crazy to think that on average we are getting a 5.31% cut in covid harm for each 1% increase in prevention cost we pay! But that’s what it would take to justify this level of prevention spending.

Bryan Caplan on Tuesday:

total loss comes to about 37 million years of life. That’s about 15 times the reported estimate of the direct cost of COVID. … If normal life had continued unabated since March, how many additional life-years would have been lost? … fifteen times? No way. Upshot: The total cost of all COVID prevention has very likely exceeded the total benefit of all COVID prevention.

Tyler Cowen today:

I don’t agree with Bryan’s numbers, but the more important point is one of logic. The higher the costs of reaction to Covid, the stronger the case for subsidizing vaccines, therapeutics, and other corrective measures. Would you accept this Bryan? You have numerous posts about risk overreaction, but not one (if I recall correctly) calling for such subsidies. …

A second question is whether moral suasion — “don’t overreact to Covid!” — is likely to prove effective. … Sweden didn’t do any better on the gdp front, and the country had pretty typical adverse mobility reactions. … Brazil … have a denialist president, a weak overall response, and a population used to a high degree of risk. … about overreaction. What kinds of reaction are you expecting or viewing as feasible and attainable? If overreacting is indeed a public bad, why think you can talk people down out of it? … they don’t and indeed can’t tell you how most of those [overreaction] costs were to be avoided, given how the public reacts to risk. …

If we instead look to the relevant changes in relative prices, that means subsidies for vaccines and tests, most of all through advance market commitments, but not only. And a full-scale commitment to implementing testing and masks and therapeutics. The more you push home points about overreaction, the more you ought to favor these subsidies. Libertarians out there, do you? This chicken has come home to roost, so please fess up and give the right answer here. Do you favor these subsidies?

Cowen seems to divide pandemic prevention into two categories, the first of which is ineffective but simply cannot be avoided, while the second is highly effective and can in fact be changed, if only people like Bryan would speak up. In this case, the more there is of unavoidable ineffective prevention, the more valuable it is to spend more on effective prevention.

I question Cowen’s arbitrary claim that we intellectuals can only influence spending on very effective kinds of treatment, but not on others. We see variations in both kinds of policy across space and time, due both to private and government choices, all of which seem modestly influenceable by intellectuals like Caplan, Cowen, and I. There are people out there arguing to cut ineffective prevention, as well as people arguing to expand effective prevention, and both groups deserve our support. (“Spending” includes all choices that induce opportunity costs.)

But we should also consider the very real possibility that the political and policy worlds aren’t very capable of listening to our advice about which particular policies are more effective than others. They may well mostly just hear us say “more” or “less”, such as seems to happen in medical and education spending debates. In this case, we should consider the value of more or less prevention spending overall, holding constant the relative proportions of different kinds of spending. And in this case the clear answer seems to be: less; we should do less. Let go the nut of effective treatment in the pandemic-money-trap gourd of over-prevention. Don’t you agree Tyler?

Added 8pm: Though Tyler criticizes Caplan and my posts which are directly on the topic of overall covid over-prevention, he refuses to say if we are spending too much overall; he simply rejects the “framing” of this question. Seems a question he’d rather not talk about.

GD Star Rating
a WordPress rating system
Tagged as: ,

Five Important Topics

Some topics are especially important. You might think that on such topics we’d try extra hard to be more accurate, both individually and institutionally. Alas, we are also more likely to self-deceive on them, both individually and institutionally.

So on such topics you should less trust both your intuition and the usual institutions. Be more skeptical of all sources, and more eager to find especially reliable sources. At least if you were trying to be accurate. Which you probably aren’t.

What are these most important topics? Simple biology predicts (at least) the following five, ordered from the top down by the fraction of species for which they are important:

Death – All individuals of all species die, which is a terribly important event. As La Rochefoucauld said “Neither the sun nor death can be looked at with a steady eye.” For poor humans food, water, shelter, and many other topics are about death.

Sex/Fertility – All species reproduce, and most do this via sex. For sexual species, sex is as important as death, and easier to deceive associates about.

Status – Many animal species maintain status hierarchies, and in such species it is important both to rise in status, and to judge status well. We self-deceive on our own suitability for status, and on how high status folks differ.

Politics – Primate social groups often split into conflicting coalitions. In that case it is important to be an attractive coalition partner, and to judge partners well. We self-deceive on how our coalitions differ, and on other topics to show loyalty to our faction.

Honesty – Only a species that can communicate can care about honesty (I.e., accuracy), but humans communicate by far the most. And if we are dishonest on important topics, then we must also be dishonest on the topic of our honesty. Yet our reputations for honesty are important to us.

Note that these are also topics on which many seem quite emotionally sensitive, and where norms, laws, and retribution often limit who can say what. And even when people claim to not care about such things, those meta opinions seems to be very important to them.

Note also that many other important topics can be seem as combinations of the above. For example, class is inherited (i.e., fertility driven) status, medicine is death managed by status, and males are the sex designed to die more easily.

I’m painfully aware of how sensitive is the topic of sex/fertility, as I’ve been partially cancelled on the basis of a few brief remarks on that. But if this is a topic on which accuracy is harder, shouldn’t I be more reluctant to have opinions on it?

Yes, but the main thing I said that bothered people was to suggest that sex/fertility might be important to big groups of people. (To husbands who raise kids not theirs, and to people who get zero sex.) And one thing it seems we can be especially confident about these topics is that they are probably important to many people. We are evolved talking social creatures, and these topics are important to the survival and selection of such creatures. Thus I defended a claim about which we can all be especially sure, and doing that apparently induces an unusually strong retribution.

GD Star Rating
a WordPress rating system
Tagged as:

Who Wants Social Insurance?

During the Biden administration, we will hear many argue that we should hand out more benefits to more people. Now when we economists argue for policy, we usually make economic efficiency arguments. So it is worth noting that for many of these policies, the main economic efficiency rationale for such handouts is “social insurance”. We are already seeing related arguments regarding pandemic relief and school loan forgiveness.

The “social insurance” argument posits a scenario where many people would have wanted to buy private insurance against big risks that they (or their descendants) face, but private insurance markets failed to offer such insurance. And thus the government should step in and produce the effect that such insurance would have produced, which is to pay certain people in certain situations, and tax everyone else to pay for it.

Now it is certainly true that insurers don’t offer all possible kinds of insurance. This can be due to legal restrictions, transaction costs, and information asymmetries. But it can also be due to limited demand. What if most people don’t actually want the insurance that “social insurance” would provide?

We already see many puzzling patterns in common insurance choices. Insurance was long illegal most everywhere, but then in the 1800s the first big retail success of insurance was life insurance sold to husbands as a way to signal devotion to their wives. Today, people often insure small risks like a new piece of electronics breaking, but fail to insure many of the largest risks in their lives, like failures at school, career, or marriage. And when I’ve asked students if they want to insure against such big risks, most usually say no, they don’t.

To explore this further, I did some Twitter polls on willingness to pay for 15 kinds of risks. Here are those risks, sorted by the fraction of respondents who says they would find value in fairly-priced insurance:

Note that only a majority favors private insurance for the top six items, and private insurance is in fact available for all of these today. Note also that these results are from the 3rd version of these polls that I tried. I found smaller fractions wanting insurance in the 2nd and 1st sets.

Of course people aren’t always honest in polls; maybe they really do want to insure far more risks than they say. And the fact that people often push political systems for “social insurance” policies is supporting evidence. But equally plausible, I think, is the theory that many really just want to use government to induce transfers, but when those folks are economists they try to justify such plans using econ efficiency lingo.

A clean test would be for the government to offer fairly-priced insurance against many risks, to ensure that no market failures prevents the availability of such insurance. Or even to subsidize such insurance. If there were actually a market failure preventing such insurance, that seems the most direct way to fix the problem. Yet we almost never see proposals like this; people almost always just push for more handouts. I wonder why.

Added 9a: The usual insurance “market failures” are:

  1. Moral hazard – which government can only fix if it can see more private acts than can private insurers.
  2. Adverse selection – which can be solved by requiring purchase of private insurance, and whose existence requires the opposite correlation between risk level and insurance quantity than the one we usually see,
  3. Scale economics – which private insurers might also achieve if not forbidden by antitrust rules.

To get private long term insurance, we could let kids sign insurance contracts when young, or empower their parents or grandparents to agree on their behalf. Note that today parents could, but usually do not, implement partial poverty insurance by insisting that their richer kids transfer to their poorer kids. As more than half of of national income variance is of this within-family form, this could achieve more than half of the gains possible from poverty insurance within a nation. And parents are much better placed than government to adjust for moral hazard of varying child efforts.

GD Star Rating
a WordPress rating system
Tagged as: ,

Easy Conspiracy Tests

The most obvious kind of conspiracies to expect in the world are ones between your local organized crime, police, and political powers. These powers should each see each other as rivals to their dominance. So to reduce such threats they should should seek to either weaken each other,  or to ally with each other. There are rumors that such alliances are common around the world, and there is clear data that they have often happened in the past. 

So this is no mere theoretical conspiracy theory, to be dismissed by claiming that too many people would know to keep it a secret. This sort of conspiracy is not only verifiably common, it also has a quite credible threat of punishing those who too publicly expose it.

Given that this seems to usually be a real possibility, what sort of evidence might speak to it? Here are some indicators that, if true about your area, are at least weak evidence against such a local alliance:

  1. You see very little profitable crime in your area, such as gambling, drugs, or prostitution.
  2. You see profitable crime, but also conflicts over its control. E.g., wars over drug-selling territory.
  3. You see profitable crime, and little fighting for its control, but its consumer prices are quite near average costs.
  4. You see profitable crime, and see that those who enter such industries fear only police and competitors, not organized crime.
  5. You see a big conflicts in your area between police and organized crime. 
  6. Police in your area are bounty hunters, who regularly gain bounties from catching each other, and judges are clearly not corrupted.
  7. You do elections in a way that lets organized crime steal elections, yet they are clearly not being stolen.
  8. (what else?)

Now if you don’t see any of these signs in your area, you should estimate a higher than average chance that you have a crime-police-politics conspiracy in your area. And as the base rate is already substantial, your estimate should be even higher than that. 

If people were concerned about such conspiracies, they’d pay to hear from folks who collected and published stats on these indicators. And pay even more for stats that made it easier for skeptical observers to check on how such stats were collected and constructed. And then those with the worse indicators, suggesting local conspiracies, could learn about that fact, and perhaps coordinate to change it. 

So what does it tell you if few seem to care enough to even know if such stats are published?

GD Star Rating
a WordPress rating system
Tagged as: , ,

Centrists Outside The Overton Window

Overton window – the spectrum of ideas on public policy and social issues considered acceptable by the general public at a given time.

The main thing that sets the size of the world economy today is the amount of innovation that has happened before today. In general we don’t induce enough innovation, as innovators tend to benefit society more than themselves. Thus better innovation institutions, which offer stronger incentives to innovate well, are probably the main way to make a better wealthier more prosperous tomorrow.

Innovation is also terribly important in politics and governance. Most of today’s official policies today were once only proposals, and before that quite unpopular proposals. That is, most current policies were once outside the Overton window of policies seen as acceptable to at least consider. 

The usual political and governance processes are about choosing among policies within the Overton window. Almost everything you see in government, politics, and the media is part of this process. As well as a great many things you don’t see. But just as important are the less visible processes that happen outside the Overton window, processes which decide which of the vast space of logically possible policies are moved closer to the edge of that window, where random fluctuations might tip them in. 

These processes by which outside proposals are honed and selected seem especially neglected today. The main groups who attend to them seem to be political extremists, whose ideal policies are far outside the current Overton window. Such extremists seek ways to reframe political alliances or issues to move the Overton window, and thus gain more favorable consideration for their extreme ideal points. 

Now given how important is innovation, and how neglected is the vast space of policies outside the tiny Overton window, it is good that at least some are thinking outside this “box”. But it also seems a shame that these are mainly extremists. 

Our usual business and tech innovation would be worse if most of this was done by extremists pushing unusual agendas regarding how we should all want to live. For example, imagine if cell phones were only likely to be developed by religious cults with a deep hatred of wires. It is good that instead most business and tech innovation is instead driven by profit incentives, as that drives them to explore a much wider space of possibilities. 

This common diagram describing the Overton window is seriously misleading, as it presents policy as a one-dimensional spectrum. In fact, the real space of policies is hugely multi-dimensional, with a “Overton hypercube” that takes up only tiny fraction of this vast space. It is thus perfectly rational to want to explore that vast space, even by those who are “centrists” with respect to the usually small number of ideological dimensions. There are surely a great many good policies out there in that vast space which are attractive to centrists, if only they can find and hone them.

Thus we want many such centrists to help explore this space, and not just a few extremists. Perhaps in future posts I’ll make some new more concrete proposals for how to better recruit centrists, and most everyone, to explore the vast space of possible policies that lie outside the tiny Overton hypercube. To help hone and select candidates, moving them closer to the edge of this Overton “window.” After which they may later drift inside that window, to become live possibilities for adoption. The better we explore the total space of policy options, the better will be the policies that we adopt. 

GD Star Rating
a WordPress rating system
Tagged as:

The Real Problem

Some think that if only we could teach more history to screenwriters, we’d get more realistic historical movies. Or if only we could teach rationality to journalists, we’d get less fake news. Or if only we could teach ad writers more about products features and prices, we’d get more informative ads. Or teach charity workers about effectiveness. Or teach statistics to scientists. 

But in these cases, and many more, expert incentives are usually not sufficiently aligned with such objectives for these experts to listen well to such teaching. They often have long been well aware of such issues. 

Some see the fix in teaching about these incentive problems to the leaders of movie studios, newspapers, and ad firms, or to the patrons of charity and science. As then they’d know to create better incentives for the experts they lead. But their incentives are also usually not sufficiently aligned with these objectives for them to change their ways. They also have often long been well aware of such issues. 

Every expert has ultimate customers, though the chain between them is longer in some cases, with many intermediaries. You can’t change expert behavior without convincing some intermediary, or the ultimate customer, that it is in their interest to promote such change. Either they have misjudged their interests and incentives, or you can somehow change their environment to change their incentives. 

For example, imagine that doctors pretend to be effective at improving health, but mainly give customers the affiliation with prestige that they crave, to let them show they care about each other via paying much for prestige. But perhaps customers don’t want to admit this fact, and instead want to pretend to be buying health improvements. Then there’s a chance you could create data showing that their choices are not effective, and also create widespread common knowledge about which other choices would be more effective, at least at that moment.

In this case you might shame customers into demanding that intermediaries (e.g., insurers, regulators, NHS, Medicare) demand that doctors actually do the health-effective things. But you will likely be opposed by inertia and by doctors, intermediaries, and customers who are not eager to expose their prior incompetence and hypocrisy. They might even relent now and do what your data suggest, but then coordinate vigorously to crush you and any who might try later to emulate your approach, to make sure this never happens again.

This is the world you live in, and the problem we face.

GD Star Rating
a WordPress rating system
Tagged as: ,

ALL Big Punishment Is “Cruel”

Cruel – willfully causing pain or suffering to others, or feeling no concern about it.

Cruelty is pleasure in inflicting suffering or inaction towards another’s suffering when a clear remedy is readily available. …affirmative violence is not necessary for an act to be cruel. … are four distinct conceptions of cruelty. … first … above in degree and beyond in type the [suffering] allowed by applicable norms. … second … fault of character consisting in deriving personal delight from causing and witnessing suffering … punishment or other violence is a means to restore the offset in the cosmic order of the universe caused by a wrongdoing. Anything that goes beyond what is necessary for this restoration, then, is cruel. … third … the pain or the sense of degradation and humiliation experienced … fourth … accumulation of all the prior conceptions. (More)

A great many things seem quite wrong with the U.S. legal system, especially in criminal law. I’ve tried to work out comprehensive solutions, but I should also identify more modest changes, more likely to be adopted. And one big way our criminal law seems broken is our huge prison population, which is near a world and historical peak of residents per capita.

Many people say we define too many acts as crimes, that we make it too easy for authorities to prosecute people, and that we punish many crimes too severely. And while those seem like fine issues to explore, I see an even clearer case that jail is usually the wrong way to punish crime. Let me explain.

In principle jail can serve many functions, such as education, reform, isolation, and punishment. But prison is now more expensive than college; few see it as a cost effective way to learn. And few believe that US jails actually reform many convicts. Yes, convicts do tend to commit fewer crimes over time, but that’s mainly due to age, not reform efforts. Jail cuts residents off from their prior social connections, such as jobs and family, and connects them instead to other criminals. Which seems bad for reform. 

Jail does isolate convicts, making it harder for them to commit many crimes. But we can isolate most convicts nearly as well and far more cheaply with curfews, travel limits, and ankle bracelets. Whole isolated towns might be set up for convicts. And if isolation were the main issue setting who we sent to jail and for how long, then for each person we put into jail we’d keep them there until we saw a substantial decline in our estimate of the harm they might do if released. 

But in fact the median time served in state prison is 1.3 years, way too short a time to usually expect to see much change. And even if there is a substantial decline soon after a peak crime age, we don’t vary sentence lengths in this way with age.

Furthermore, exile offers a much cheaper way to isolate. Let convicts leave the nation for a specified period if any other nation will take them. Not every one would be taken, but each one who is taken represents a big savings. Worried about them sneaking back unseen? Just make severe punishments for that. Maybe even make them post a bond on it.

So if education, reform, and isolation are poor explanations of jail, that leaves punishment. Ancient societies used fines more often, which they often took from family members if the convict couldn’t pay, and they often enslaved convicts to make them pay. But as we aren’t willing to do these things, we can’t get much money out of most convicts, which is why we need other punishments. 

Note that I’m not saying that jail could not in principle achieve other ends, nor even that jails do not to some degree achieve other ends. I’m saying instead that the widespread popular support for using jails today mainly comes from a widespread perception that jails achieve punishment, which most see as a desirable end. 

The classic logic of criminal punishment is that most people are less likely to commit a crime if they anticipate a substantially higher chance that doing so will result in their experiencing a “punishment” that they will dislike. (Relative to the chance if they don’t commit the crime.) Yes, this effect may be weak, but most people aren’t convinced of other approaches, and they aren’t willing to give up on this approach. 

But a big problem with using jail to punish is that our jails are terribly expensive, relative to feasible alternatives. For example, most of our jails are relatively “nice” and comfortable, with nice food, beds, climate control, entertainment, etc. At least compared to other jails in history. But typically X years in a nice jail gives the same expected punishment (i.e., anticipated dislike) as Y years in a mean jail, for Y < X. So if a mean jail costs no more per year than a nice jail, this is a cost savings.

Our history and the world today clearly demonstrate that it is possible to create jails that are less nice than ours. Furthermore, corporal punishment (often called “torture”) is even cheaper than mean jails. This was quite common in ancient societies, and is still used in some places today. For any sentence of X years in jail, there is some amount of corporal punishment, e.g., N lashings, that gives the same expected punishment at a far lower cost. 

Some say that torture and mean jails are more “cruel” than nice jails, and thus immoral, and thus forbidden. But as you can see from the above definitions, when the amounts of these things are adjusted to produce the same amount of anticipated dislike for each, then some of them cannot be more “cruel” than others in the sense of the dislike convicts expect to experience. The only grounds then offered for saying that some are more “cruel” is that some might induce more inappropriate “delight from causing and witnessing suffering”. 

Yet I can find no evidence suggesting that observers achieve more inappropriate delight from criminal punishments in the form of mean jails or corporal punishment. Yes, we can see many people taking delight today in the suffering of convicts in the jails that we now have. And those people would likely switch their design to focus on other forms of suffering, if those happened instead. But I see little reason to think that those who today do not delight in seeing convicts suffer from existing prisons would start to so delight after a switch to other punishments. 

An interesting way to vividly show everyone that our jails today are in fact just as “cruel” ways to punish as these alternatives would be to give convicts a choice. A convict who is sentenced to X years in ordinary jail might be offered the choice to switch to Y years in mean jail, or to N lashings (or other corporal punishment). Or perhaps even to E years in exile. A simple standard mapping function between X and Y,N,E could be used, one that is adjusted continuously to get pre-determined fractions of convicts to choose each option.

(With enough data, these mappings might depend on age, gender, etc. Some small fraction of convicts might be forced into each option, to induce reliable data on option effects on convicts. Over time, the pre-determined fractions could be adjusted toward the cheaper options if their outcomes continue to seem acceptable relative to costs.) 

Under this system, it seems harder to complain that it is more cruel to give convicts the option to choose something other than ordinary jail, relative to just forcing convicts into ordinary jail. And the fact that many convicts do choose other options should yell a big loud lesson to all: convicts suffer a lot in ordinary jail. They lose big chucks of their lives, including their careers, friends, and families. If you see a person suffering under torture, and you realize that this person chose torture over ordinary jail, that tells you just how much they hate and dislike ordinary jail. It tells you that you should not on empathy-for-them grounds feel good about yourself for instead forcing them into the option from which they ran. 

Yes, when each convict picks their favorite punishment option from a set, that will on average reduce their expected punishment. But not usually by a lot, and the X in sentences of X years of ordinary jail could be adjusted up a bit to compensate. In this situation, one reason to exclude an option is if we are much more uncertain about how much each person dislikes that option, relative to other options. It is better to know how much we are punishing a convict. 

But I see no reason now to think that we are now much more certain about dislike for ordinary jail, relative to mean jail, corporate punishment, or exile. When the judges who sentence convicts do know something about the option preferences about a particular convict, they might on that basis exclude some of those options for that particular convict. For example, exile might be a weaker punishment for someone who recently lived abroad. 

Yes, there’d be a tendency by those who embrace a criminal culture to take the “toughest” punishment option available, to signal toughness to associates. But this would on average hurt them, and isn’t that different from many other things they do to show toughness. Doesn’t seem a big problem to me.

So that’s my pitch. Let’s stop wasting so much on expensive jails, when we could instead produce the same punishments at a lower cost by giving convicts a choice between types of punishment. This would also show everyone just how cruel we have been by putting convicts in our current jails. A majority of those who answer my Twitter polls approve; what about you?

So why do people see the other options are more cruel? My guess is the representativeness heuristic is at fault. People imagine a random moment from within the punishment, and neglect how many minutes there are in each punishment.

GD Star Rating
a WordPress rating system
Tagged as: ,

Universal Basic Dorms

Poor people have too little money. So why don’t we just give them more money, instead of giving them many specific things? The main theory in the literature is paternalism:

The traditional justification for in-kind transfers has been “paternalism.” … If members of society care about the consumption patterns of the poorest rather than their utility, then the unconstrained choices of the poor may create negative externalities for those who care about them. … while income inequality per se may be acceptable, all individuals should receive adequate food, medical service, and housing. … Parents may not take full account of the utility of their children when making decisions, or they may neglect to factor in externalities. … attempt to redistribute from parents to children within the family. … a sense that the poor cannot be trusted with cash. … it is hard to escape the conclusion that paternalism remains a fundamental underlying rationale for in-kind transfers. (more)

You may not be surprised to hear I have an elaboration of this account based on signaling and status. Let’s start with this result:

[A] 2010 paper … makes a strong case that in fact the outcome of life satisfaction depends on the incomes of others only via income rank. (Two followup papers find the same result for outcomes of psychological distress and nine measures of health.) They looked at 87,000 Brits, and found that while income rank strongly predicted outcomes, neither individual (log) income nor an average (log) income of their reference group predicted outcomes, after controlling for rank (and also for age, gender, education, marital status, children, housing ownership, labor-force status, and disabilities). (more)

So if satisfaction and preference go simply as income rank, then there is simply no way to to use income redistribution to help with such things. Whatever you give to some comes at the expense of others, for no net gain. You can still want to help the poor, but that requires that your concern about their lives not be mediated directly by their satisfaction or preferences. That is, what you want regarding them can’t equal what they want or what makes them happy.

For example, compared to the poor, you might put less weight on their desires to signal, and to rise in status via positional goods (like winning a sporting contest). Because you can see the negative externalities associated with such things. That is, you might put more weight than they do on the remainder of their preferences, which we might call their “direct welfare”.

If there are diminishing returns to direct welfare, then you can want to reduce the overall variance of direct welfare, by giving more of it to the poor, at the expense of the non-poor. But you can reasonably fear that if you just give them cash they will spend too much of that on the kinds of non-direct welfare that have negative externalities. So you can want to constrain their choices, to better ensure that it is direct welfare that you are giving them.

The big problem here is: how to distinguish the goods, i.e., products and services, that put more weight on signaling and status, relative to those that put more weight on direct welfare. The usual political equilibrium is to give the poor the sorts of goods that people tend to praise and admire, at least for others. Like jobs, medicine, education, libraries, art, churches, and fresh vegetables. And to withhold from the poor goods that people tend to criticize and dislike, at least for others. Like parties, drugs, sex, fast food, social media, movies, and video games.

But alas, typical admired goods don’t obviously have smaller positional components, nor do they obviously contribute less to signaling. For example, both education and medicine, widely given to the poor, have huge signaling and status components, plausibly even larger than for most goods. If we cannot in practice distinguish the goods that do more to promote direct welfare, we should give up on in-kind transfers and just give the poor cash. And then only to the extent that we think direct welfare has strongly diminishing returns in terms of cash.

I have a (perhaps not original) idea. We have good reasons to think that in general most product and service variety emphasizes signaling and status, relative to standard goods that can achieve large scale economies. So if we can make especially cheap cars, homes, clothes, food, etc. via mass production with a small range of variety, then we should prefer to spend our budget on helping the poor via such standard goods.

That is, assume that we the non-poor have a budget that we are willing to spend on helping the poor. We have two reasons to prefer to spend this budget on standard goods. First, standard goods can be provided much more cheaply, allowing us to give more to each person, or to help more people. Second, because these goods have a lower signaling and status components, the poor who consume them hurt the rest of us less via making us look worse by comparison, and rising in status relative to us. We should thus be wiling to increase the budget that we spend on the poor, in compensation.

Of course the poor may resent this policy, even if it results in larger budgets. Exactly because we choose these standard goods to have lower signaling and positional components, the poor will know that others who see that they are using such standard goods designed for the poor may think less of them as a result. That may not be quite the same as a “stigma” assigned to such goods, but it may have a similar effect. Even so, this looks like an efficient arrangment.

So how exactly can we give standard goods to the poor? There’s an obvious risk that the government, or a charity, managing such poverty assistance might overly micromanage the specific standard goods offered, inducing their design, variations, production, and maintenance. Resulting in greatly increased costs and lower quality, as we’ve often seen with public housing.
So it seems better to just offer the poor credits that they could spend at any competing private supplier of standard goods packages.

That is, let us offer to pay qualifying private “dorms” a fixed budget per resident served per day. Each dorm would give its residents its standard package of a bed, clothes, food, and services for transport, entertainment, schooling, and medicine. Then the qualified poor could choose the dorms they see as offering the best combination of location, other residents, and standard services. If there were many competing dorms, and if residents could change dorms frequently (say at least once a year), competition should force dorms to offer efficient and relatively attractive packages. (We may also need to change zoning and regulation to allow such dorms to be offered.)

History seems to suggest that direct welfare can often be provided more cheaply and reliably via dorm-like living. Such as we’ve seen in colleges, the military, orphanages, hospitals, prisons, and retirement homes. So it doesn’t seem a crazy idea to help the poor live cheaply via dorms as well. But I do see a few complications.

Limiting Variety – If not constrained by some rules, competing private dorms won’t necessarily offer standard goods, instead of more varied goods of a lower quality. After all, that is what the poor seem to choose in the market today. So there would have to be some rules enforcing a degree of standardization of the goods offered. Thus there would be limits on the variety that dorms offer in their rooms, clothes, food etc. I don’t have specific proposals regarding these rules to offer now.

Unusual Variety – Some of the poor will actually be unusually different from the others, and thus need unusually different services. Such as some disabled folks, or parents with young children. These might be offered something different than the standard options proposed here.

Added Variety – Even if we push for less variety and more standardization in what we give the poor, there’s no need to go to extremes on this. Internet connections would offer the usual immense variety of outside sites to which one could connect. And some fraction of the budget paid to dorms might be paid directly to the poor as “allowances,” from which residents could pay for unusual expenses and added variety, such as for some food or clothing occasionally bought outside the dorm. Perhaps the poor who insist on more variety than this dorm system offers could be offered a cash budget to spend on their own, a substantially lower amount than the budget offered to qualified dorms to host them.

Transitions – It is probably more expensive for dorms to deal with residents who come and go, relative to residents who stay. So it would probably be better if dorms were paid some extra amount for resident changes. Residents might have a limited budget of such changes, or maybe they’d have to pay for change fees out of an allowance.

Universal Offer – I see no reason not to allow dorms to let the non-poor to pay to be residents. In addition, some may argue for paying dorms the same budgets to let anyone stay there, not just the qualified poor. Then anyone could choose to live in these dorms, and save on living expenses. This would be “Universal Basic Dorms” instead of a “Universal Basic Income”. It would be much cheaper than U.B.I., as fewer ordinary people would be willing to stay in such dorms.

Valuing Neighbors – If these dorms only housed the qualified poor, then such poor may have a worse pool of associates and role models. Thus it might make sense to pay dorms some added amount per non-poor residents who mingle with their poor residents. This makes sense not only in terms of benefits to the poor, but also in terms of compensating such non-poor residents for the status hit they take by moving to such dorms.

Crime – My understanding is that crime is the other big thing that tended to go very wrong with public housing, in addition to mismanagement. So things might go very wrong if dorms were not allowed to reject residents they deem too likely to cause problems. In addition, things would probably go even better with a crime voucher system. Offer each resident a budget to pay a voucher to cover all the crime they may commit as a resident. If they can’t get a voucher to agree at that price, no matter what the other contract terms, such a resident does not qualify for dorm living. This might be a good test environment for a crime voucher system.

And that’s my proposal. Offer to pay dorms per resident who stays there, with rules to encourage less-varied more-standard dorm goods which achieve scale economies. Residents would then get more direct welfare, even if they’d gain less status and send less attractive signals. That lower status should make the rest of us more willing to increase dorm budgets well over the cash budget we might offer the poor to live outside dorms.

GD Star Rating
a WordPress rating system
Tagged as: ,

We Will Never Learn More On Consciousness

Some complained that I didn’t include a question on consciousness in my list of big questions. My reason is that I can’t see how we will ever know more than we do now. There’s nothing to learn:

Zombies are supposedly just like real people in having the same physical brains, which arose the through the same causal history. The only difference is that while real people really “feel”, zombies do not. But since this state of “feeling” is presumed to have zero causal influence on behavior, zombies act exactly like real people, including being passionate and articulate about claiming they are not zombies. People who think they can conceive of such zombies see a “hard question” regarding which physical systems that claim to feel and otherwise act as if they feel actually do feel. (And which other systems feel as well.)

The one point I want to make is: if zombies are conceivable, then none of us will ever have any more relevant info than we do now about which systems actually feel. Which is pretty much zero info! You will never have any info about whether you ever really felt in the past, or will ever feel in the future. No one part of your brain ever gets any info from any other part of your brain about whether it really feels.

These claims all follow from our very standard and well-established info theory. We get info about things by interacting with them, so that our states become correlated with the states of those things. But by assumption this hypothesized extra “feeling” state never interacts with anything. The actual reason why you feel compelled to assert very confidently that you really do feel has no causal connection with whether you actually do really feel. (More)

Your brain is made out of quite ordinary physical materials, driven by ordinary physical processes that we understand very well at near-atomic levels of organization. It is only processes at higher levels of organization that we haven’t traced out in detail. We will eventually be able to trace in great detail and at all levels the causes of what makes you, or an em, or any variation on either, inclined to passionately claim, and believe, that you really do feel. And that will let us predict well what changes to you, or anything, might induce you, or it, to claim or believe something different.

But if you insist that none of that can possibly verify that you, or an em, actually do feel, then it can’t add any info on that issue. Yes, maybe you have intuitions inside you that often tell you if you think something that you see in front of you really feels. But such intuitions are already available to you now. Just imagine various things you might see, note your intuitions about each, compare those to others’ intuitions, and then draw your conclusions about consciousness. After all, we already have a pretty good idea of all the things we will eventually be able to see.

Okay, yes, you are probably in denial about how much the intuitions of others would influence yours, and about how strong would be the social pressures on your intuitions to accept that ems feel, if in fact you lived in a would where ems dominated. I predict that in such a situation most would accept that ems feel. Not because new info has been offered, but because of familiar social pressures. And yes, we can learn more about how our intuitions respond to such pressures. But that won’t give us any more info on the truth of what “really” feels.

GD Star Rating
a WordPress rating system
Tagged as: