Monthly Archives: February 2012

Immoral Altruism

Eighty-five per cent of them said it would be morally wrong to push one person off [a bridge] to save five [from a trolley], whether these people are brothers or strangers, confirming the idea that there is a rule against killing. However, despite thinking it wrong, 28 per cent said they would still push a stranger off to save five, while 47 per cent said they would push a brother off to save five brothers. (more)

One of the study’s authors offers an explanation:

Social cohesion demands we have rules, regardless of what they are, to help resolve disputes quickly and peacefully. DeScioli says our rule-making system is arbitrary, producing the belief that masturbation is “bad”, for instance.

But why resort to randomness when other good explanations remain? We naturally want simple clear social norms against murder. While simple rules create unfortunate incentives in specific cases, they are overall easier to monitor and enforce. This trolley problem seems to be one of those specific cases where many of us think that our simple rule against murder goes wrong – while we agree that killing in this case violates our murder norms, even so many of us are willing to violate such norms in order to help associates, especially if we care a lot about them.

While morality may be in general pro-social, it is not in every specific case. So there are times when you must choose between being moral, and being helpful.

GD Star Rating
Tagged as: ,

How Social Are Signals?

We are aware that do many things for show, and I often suggest that we do such “signaling” more often than we realize. But while I’m eager to see writings on signaling theories and their empirical support, I’ve come to suspect that most tend to be unrealistically asocial. Let me explain.

In the iconic signaling story, one person has a hidden feature, which they choose to show to one other person, via some visible action. For example, on Valentine’s day a man traditionally buys a gift, writes a poem, etc. to show a women the strength of his feelings for her. The bigger the gift, the bigger his feelings, supposedly.

In this iconic situation, only these two parties matters. And this allows for simple sharp predictions. For example, if the person watching can’t see the signal, or already knows about the feature, there is no point in signaling. And there is no point in taking an action A to show feature F if that feature is unrelated to willingness to do A.

In realistic signaling, however, third parties typically matter a lot more. For example, the man might want to signal that other women want him, or that he knows that other men want her. The woman might care less about what she infers from his signal, and more about being able to let slip details to her friends, to show them the kind of man she has. This inclusion of a wider social circle makes it harder to find simple sharp tests.

I’ve talked about how schooling could be such a more social signal, and how that could complicate empirical testing:

Firms want to impress customers, suppliers, investors, etc. with the quality of their employees, and hiring graduates from prestigious schools helps them signal such quality. Hiring such graduates can also help a manager to impress his bosses, potential employees, and sister divisions about the quality of his employees. … The fact that attending school seem to cause changes in students that employers are willing to pay for does not show that school isn’t all about signaling. (more)

Similarly, people often respond to my suggestion that medical care functions in large part to “show that you care” with the example of people buying medicine for themselves. “Surely that can’t be signaling,” they suggest. But consider that unattached women often buy themselves flowers or chocolates on Valentines day. As signals become more social, and involve wider circles, it gets harder to isolate situations where no signaling should happen.

By the way, one way to think about “status” is as the limit of very social signals. The more that an action or sign is generally seen as positive, without being very specific about what good features it indicates or who exactly cares about such features, the more that this action or sign looks like a signal of general social status.

GD Star Rating
Tagged as: , ,

Don’t Torture Mom & Dad

A doc’s eloquent plea:

It’s typically the son or daughter who has been physically closest to an elderly parent’s pain who is the most willing to let go. Sometimes an estranged family member is “flying in next week to get all this straightened out.” This is usually the person who knows the least about her struggling parent’s health. … With unrealistic expectations of our ability to prolong life, with death as an unfamiliar and unnatural event, and without a realistic, tactile sense of how much a worn-out elderly patient is suffering, it’s easy for patients and families to keep insisting on more tests, more medications, more procedures. … When their loved one does die, family members can tell themselves, “We did everything we could for Mom.” … At a certain stage of life, aggressive medical treatment can become sanctioned torture. When a case such as this comes along, nurses, physicians and therapists sometimes feel conflicted and immoral. … A retired nurse once wrote to me: “I am so glad I don’t have to hurt old people any more.” (more; HT Amanda Budny)

Our urge to use medicine to show that we care costs more than just spending more for mostly useless treatment. It often literally tortures our loved ones.

GD Star Rating
Tagged as: , ,

Me On Marketplace

I’m on today’s edition of the NPR radio show Marketplace Money (transcript; audio ~25:30 to 29:00):

The average dog owner spends $655 a year on health care, that’s up 50 percent from a decade ago. Cat owners are in for $644, up nearly 75 percent, close to how much our health care costs have risen by. And that’s a puzzle to economists, like Robin Hanson at George Mason University.

Robin Hanson: Everyone’s got a favorite villain or bugaboo about why human health care costs are increasing; it’s too much regulation, too much government involvement, too much third-party payment.

Too many malpractice lawsuits. None of these factors apply to pets. You can’t blame insurers for pushing up costs either. Pet insurance is rare; only 1 percent of pet owners in this country have it. The 99 percent are paying full freight.

Hanson: But in pet medicine, people put their money on the barrel head. And yet pet expenses are increasing nearly as fast as human expenses.

What gives? Hanson and other economists give two explanations. Explanation one: Love. We treat our pets like family. They eat our food, they sleep in our beds, they relax at the spa, they have Facebook accounts. Of course we’re going to pay for their health care. Take dogs.

Hanson: So we want to show loyalty to these dogs who are showing loyalty to us. One way to do that is to spend more on medicine for them.

Explanation two for the rising cost has nothing to do with your pets; it’s how we see ourselves.

Hanson: We compare ourselves to people around us. And we ask the doctor and they say well, lots of people do this, most people do this, and the bar has been raised on how much you need to spend on your pets to show you’re a caring pet owner.

In the interview I tried to pose the choice as supply vs. demand explanations, as I’ve done in my last two posts, but I guess they didn’t find as engaging.

GD Star Rating
Tagged as: ,

Dog vs. Cat Medicine

Yesterday I said that med spending increased faster for pets, vs. farm animals, suggests that med spending increases are due mainly to demand, not supply, effects. We spend more on pet medicine now more because we care more about pets now, or want to show we care, and less because doctors have invented new useful treatments.

Now consider dog vs. cat medicine. A 2007 source said that at one point annual med spending was $200 per dog and $81 per cat. (It was $92 per horse, $9 per bird. Today we spend $655 per dog; other current figures available here for only $3000. Sigh.) So we spent 2.5 times as much on dog med, vs. cat med. Yet dogs and cats have about the same lifespan (dogs, cats), and similar rates of medical problems:

50% of today’s cat owners never take their cats to a veterinarian for health care. … Because cats tend to keep their problems to themselves, … cats, on an average, are much sicker than dogs by the time they are brought to your veterinarian for treatment. (more)

I doubt we should blame this on cats. It seems more likely that cat owners pay less attention to cats, because they care less:

74 percent of the test sample like dogs a lot, while only 41 percent like cats a lot. … 15 percent of the adults questioned said they disliked cats a lot while the number who said they disliked dogs a lot was only 2 percent. … Dog people were 11 percent more conscientious than cat people. … Cat people were generally about 12 percent more neurotic. (more)

Yet there are more cats than dogs. Note also that both WebMD and wikipedia have pages devoted to dog lifespan; neither have such a page for cats. Dogs are famously more loyal than cats, and it seems plausible that dog owners thus feel more loyal to dogs, and more obligated to help when sick.

I tentatively conclude that we spend 2.5 times as much on dog vs. cat pet medicine mainly because we care more about dogs. This shows a huge demand effect on med spending.

Now consider that in our society many consider men more expendable than women. We send men to war, expect men to put themselves in harms way to protect women, and try to save “women and children first.” Women also go to the doctor a lot more often than men, even though men are on average sicker (they die faster). For 2008 US doctor office visits, here is the ratio of women to men by age:

All,  1.43; <15,  0.93; 15–24, 2.24; 25–44, 2.26; 45–64, 1.39; 65–74, 1.11; >75,  0.95. (more)

This also seems likely to be a demand effect – we spend more on female medicine mainly because we care more about women, or care more to show that we care about them.

Added 7p: That Marketplace show quotes similar numbers for dog and cat spending:

The average dog owner spends $655 a year on health care, that’s up 50 percent from a decade ago. Cat owners are in for $644, up nearly 75 percent.

So did we once to care more about dogs, and now care about the same?

GD Star Rating
Tagged as: ,

Farm vs Pet Medicine

We now spend a huge fraction of income on medicine. Today the US spends ~18% of GDP on medicine, while in 1940 we spent ~4%. Why the huge increase?

A supply explanation is that doctors have invented lots of new useful treatments. A demand explanation, in contrast, is that we want more medicine as we get richer, either because we care more about health, or about showing that we care.

One way to distinguish supply vs. demand explanations is to look at farm vs. pet animal medicine. Both kinds of animal medicine are treated similarly by most supply changes – new medical treatments help both kinds of animals. But most demand changes treat them differently – farm animals today aren’t that much more valuable than they were long ago, but we treat our pets as if they were far more valuable.

While I can’t find good historical data, what I do find suggests we’ve seen a huge switch in animal medicine, from a focus on food animals to a focus on pets. On recent pet med spending increases:

The average household in the U.S. spent $655 on routine doctor and surgical visits for dogs last year, up 47% from a decade ago, according to the American Pet Products Association. Expenditures for cats soared 73% over the same time frame—on pace with human health-care cost increases. Expenditures for people in the U.S. were up 76.7% between 1999 and 2009, according to the U. S. Centers for Medicare and Medicaid Services. (more)

On vets long ago:

Very early veterinarians were mainly concerned with the care of livestock and horses and mules. … Prior to World War II, very few people would consider paying more than a token amount for the medical care of their pets any more than the average person today would consider taking an injured chipmunk to the vet. (more)

On the focus of US vets in 2011:

Food animal exclusive 1.8%; Food animal predominant 6.0%; Mixed animal 6.8%; Companion animal predominant 9.7%; Companion animal exclusive 67.2%; Equine 6.0%. (more)

Thus much, perhaps most, of the rise in animal med spending is a demand effect. More careful data analysis might give a more precise estimate.

Now pets probably live to be older than farm animals, so a supply shock mainly relevant for older animals might explain an increase of pet med relative to farm animal med. But that seems pretty unlikely to be the main thing going on here.

GD Star Rating
Tagged as: , ,

On Being Self-Aware

Diane Rehm yesterday, interviewing an expensive matchmaker:

REHM: And, Janis, you said that your clients are men. So you don’t take women who may have lots of money looking for a male?
SPINDEL: No, thank you.
REHM: Tell me why.
SPINDEL: Been there, done that.
REHM: Well, tell me why.
SPINDEL: To be honest with you, when I first started in business I had lots and lots and lots of fabulous women clients, really great women. And they seem to be needy and very high maintenance and you can never satisfy them.
REHM: Interesting.
SPINDEL: We would introduce them to amazing men. They’re not available, which is one of the biggest problems that I hear about women. See, I own the minds of men. I know what they want and I know what women do wrong. I could literally do this in my sleep. Men are very simple. You deliver exactly what they’re asking for and you leave the rest up to chemistry and the universe.

I have to admit this is somewhat at odds with my suggesting:

We should expect men to be more self-aware, transparent, and simple regarding their feelings about short-term sexual attractions. … In contrast, women should be more more self-aware, transparent, and simple regarding their feelings about long-term pair-bonding.

GD Star Rating
Tagged as: ,

Rah Power Laws

The latest Science has an article by Michael Stumpf and Mason Porter, complaining that people aren’t careful enough about fitting power laws. It mentions that a sum of heavy-tail-distributed things generically becomes has a power law tail in the sum limit. And it claims:

Although power laws have been reported in areas ranging from finance and molecular
biology to geophysics and the Internet, the data are typically insufficient and the mechanistic insights are almost always too limited for the identification of power-law behavior to be scientifically useful … Examination (15) of the statistical support for numerous reported power laws has revealed that the overwhelming majority of them failed statistical testing (sometimes rather epically).

Yet in reference 15, where Aaron Clauset, Cosma Rohilla Shalizi, and M. E. J. Newman looked carefully at 25 data sets that others had claimed fit power laws, only for 3 did they find less than moderate support for a power law fit, and in none of those cases was any other specific model significantly favored over a power law! It this is the best criticism they’ve got, this seems to me resounding support for power laws.

Here are the phenomena where the power is less than one, meaning the few biggest items get most of the weight:

intensity of wars 0.7(2); solar flare intensity 0.79(2); religious followers 0.8(1); count of word use 0.95(2)

The number is the power and the digit in parens is the uncertainty of the last digit shown. Here are the phenomena where the power is greater than one, meaning most weight goes to many small items:

telephone calls received 1.09(1); bird species sightings 1.1(2); Internet degree 1.12(9); blackouts 1.3(3); population of cities 1.37(8); terrorist attack severity 1.4(2); species per genus 1.4(2); freq. of surnames 1.5(2); protein interaction degree 2.1(3); citations to papers 2.16(6); email address books size 2.5(6); sales of books 2.7(3); papers authored 3.3(1)

For quake intensity they give power 0.64(4), but say a better fit is a different power (unspecified) and a cutoff. For net worth (of the US richest 400) they give power 1.3(1), but say a power-law doesn’t fit, though no other model tried fits better.

On catastrophic risk, I wrote in ’07:

We should worry more about disasters with lower powers, such as forest fires (area power of 0.66), hurricanes (dollar loss power of 0.98, death power of 0.58), earthquakes (energy power of 1, dollar loss and death powers of 0.41), wars (death power of 0.41), and plagues (death power of 0.26 for Whooping Cough and Measles).

So the above study suggests we worry most about wars, quakes, religions, and solar flares. I hadn’t been worried about solar flares so much before; now I am. On city inequality, I think I trust that other paper more.

Added 4p: Cosma Shalizi says:

In ten of the twelve cases we looked at, the only way to save the idea of a power-law at all is to include this exponential cut-off. But that exponentially-shrinking factor is precisely what squelches the WTF, X IS ELEVENTY TIMES LARGER THAN EVER! THE BIG ONE IS IN OUR BASE KILLING OUR DOODZ!!!!1!! mega-events.

I’m happy to admit that worse case fears are reduced by the fact that <1 power law data tend to be better fit by a tail cutoff. Good news! I don’t want to believe in disaster, but I do think we must consider that possibility.

GD Star Rating
Tagged as: ,

Paying For Med Quality

Some hope to cut US med spending, and raise med quality, by paying more for higher quality outcomes. But this doesn’t work so well if the people you pay are also in charge of telling you the outcomes. Or if you must prove that their care caused bad outcomes, instead of being due to especially weak/sick patients. I have high hopes for a system that paid for med outcomes determined by independent third parties, where price competition for specific patients could deal with patient selection issues. But I’m pretty skeptical that the US govt will allow that:

Medicare has begun publishing the rates of complications as a step toward using them to set payment rates for thousands of hospitals. But leaders of a number of the nation’s prestigious teaching hospitals are objecting …

A central tenet of the 2010 federal health-care law will tie Medicare reimbursement to a variety of measures, including how patients rate their stays, readmission, mortality rates and how closely hospitals adhere to basic guidelines for care. … Officials at many of the hospitals listed as having high rates of complications say the measures are fundamentally skewed in ways that exaggerate problems at hospitals that treat many complicated cases or very sick patients. …

Hospital officials examined the cases that led Medicare to rate her hospital as having a high rate of accidental cuts and lacerations. They found most of those cuts had been intended by the surgeon, but erroneously billed to Medicare under the code for an accidental cut. … “These patient safety indicators, they’re not real­ly well risk-adjusted.” …

Medicare identified 190 of 3,330 hospitals as having very high levels. Of those, 82 were major teaching hospitals, … Cleveland Clinic, said the clinic’s high rates of accidental tears and lacerations and serious blood clots were because “people are careful at documenting, almost to a fault, things that are incidental to the case.” … Gregg Meyer … predicted that many hospitals will react to the publication of the patient safety data by instructing those who fill out the billing records to change what they include, or by lavishing staff attention on the areas flagged by Medicare even if they aren’t a real problem. (more)

GD Star Rating
Tagged as:

Info Market Failure

Unless project gains can be very clearly proven to analysts, or perhaps so small and numerous to allow averaging over them, public firms are basically incapable of taking a loss on earnings this quarter in order to make gains several years later. … CEOs are strongly tempted to instead please analysts by grabbing higher short-term quarterly earnings. …

Private firms are 3.5 times more responsive to changes in investment opportunities than are public firms. … IPO firms are significantly more sensitive to investment opportunities in the five years before they go public than after. (more)

A month ago I said that these results imply that we need wealth inequality, to ensure we make the discretionary investments on which all our future wealth depends.

Today I want to admit that these results also imply that even thick speculative markets, full of lots of people trading lots of money, often have big info failures. While I am a big fan of using speculative markets to aggregate info, I must admit that they quite often fail to aggregate all relevant info, even when a lot of money can be won there.

CEOs at private firms choose investments based on private info on likely rates of return. If the same firm were to be made public, however, the above evidence suggests that CEOs would make less than 25% of those investments. In the other 75+% of cases, the CEO would estimate that market speculators would not credit the stock price for the value of those promising investments, but would instead punish the firm for lower short term earnings. It seems that market speculators cannot distinguish these investments from other less promising ones that CEOs would undertake if speculators were to credit these. CEOs typically know crucial investment details not available to speculators.

Now I can see ways to improve existing stock markets, so that they could aggregate more investment info. We could allow and even encourage “insider” stock trading by firm insiders like the CEO. And we could create decision markets, trading the stock value conditional on specific investment decisions. But while these changes should raise that <25% figure, i.e., the fraction of investments by private firms that would also be made by a public firm, they might not raise it by much.

Speculative markets can work info aggregation wonders, at least compared to common methods like surveys or committee meetings. But if you really want as much info as possible on big investments, we still know of nothing better than rich private investors with a lot on the line.

GD Star Rating
Tagged as: ,