July 9, 2007

"In other fields do you see people spouting 'I just thought of this, and I'm an economist, so it must be true!'"

Well, in other fields they don't say "I'm an economist" because they're not economists, but we know exactly what you mean there.

That's a comment in a Metafilter thread about that book "More Sex Is Safer Sex." Someone else says, "Ah. I see the new trend in economics, post Freakonomics, is to say something outrageous to get people's attention and then try to back it up with numbers."

We talked about the "More Sex" book last Friday, and a really cool thing was that the author, Steven E. Landsburg, participated in the comments and got into a whole back-and-forth with people here. Here's the exchange he had with me. I said:
The phrase is "more sex is safer sex," not "more sex is more sex and assuming sex is generally good, there are more total benefits weighed against the costs of disease, depending on just how relatively bad you think it is to have HIV." I still don't see how more sex is safer sex. But it is a catchy phrase. I'll give you that. It made me blog about it.
He said:
I hope you will read the book, or at least the chapter; if you do, you'll find that there are several related points, and I am careful to distinguish among them.

The first point is that under quite reasonable assumptions drawn from standard epidemiological models, more sex probably IS safer sex; i.e. it actually reduces the spread of STDs. That conclusion relies on some modeling assumptions, but the assumptions are nowhere near as naive or silly as many of the commentators in this thread seem to assume.

The second point is that REGARDLESS of your modeling assumptions, and REGARDLESS of whether more sex actually reduces STDs, more sex (by the relatively chaste) MUST increase the excess of benefits over costs, not as a matter of idle speculation but as the conclusion of a careful logical argument.

And the third point is that following logical arguments is both illuminating and fun. Far more illuminating and far more fun than constructing and attacking straw men, in fact.
I said:
I don't see why this assertion is susceptible to logical analysis -- careful or not. Isn't it just [an] article of faith that sex with a partner produces happiness?
He said:
The assumption is that on average, sex produces the expected amount of happiness. Sometimes much more than expected, sometimes much less than expected, but not systematically one or the other. The justification for that assumption is that if people are systematically wrong in a predictable direction, they'll eventually learn and correct for that.

That turns out to be all you need to get the analysis going (and if you won't grant me this much of a start, do be aware that the same assumption underlies 90% of economic analysis, and, stereotypes to the contrary, economic analysis does yield lots and lots of accurate predictions).

Now you can reason as follows (this is highly stylized, but it gives the flavor): You take a new partner when the expected benefits to you exceed the expected costs and not otherwise. Sometimes that will cause you to turn away partners even when the total expected benefits (to you and other members of society) do exceed the expected costs (because you count only some of the benefits, not all of them). And because, on average, expected benefits equal actual benefits, it turns out that actual benefits-minus-costs would be higher if you took additional partners.

(The ways in which your taking an additional partner benefits your neighbors are, incidentally, more substantial---and more subtle---than those mentioned in the New York Times review.)
I quoted him -- "The justification for that assumption is that if people are systematically wrong in a predictable direction, they'll eventually learn and correct for that" -- and got sarcastic:
I guess I'm cynical. I stopped believing in fairy tale romance long ago.
Lots more at the Friday post. I really loved that Landsburg came over and duked it out in the comments. Let's see more of that sort of thing.

IN THE COMMENTS: Theo says:
Yes! That was one of the best threads in recent memory. It was a great topic. And it was interesting, logically challenging, and pound-on-the-table funny by turns—yet remarkably free of slash-and-gash trolldom.

You know, I've been grousing about trolls, the quality of comment writing, etc. here and elsewhere. I know at least a few longtime bloggers and commenters who seem depressed and generally bummed out by the blogosphere. But this is the kind of thread that restores your faith in the whole enterprise.

Hate to turn all Polyannaish on everybody.

31 comments:

ricpic said...

More sex; less excitement.

The Drill SGT said...

I'm a bad person. I'm not going back to read everything the author posted. It just seems obvious to me. take the most simplistic, absurd example.

1 man, 1 woman, both infected with HIV. They have sex. the net result is 1 sex event and 0 transmission events.

now add 1 additional NON-infected woman. (or man, but keep it simple). (3 players now) result 1 sex event yields a 50% likelihood of it being between a clean woman and an infected male. the number of sex events is the same, the likelihood of passing HIV is increased.

It seems that any addition of clean partners into the mix, decreases the likelihood of pairings between already infected pairs, thus MUST increase the likelihood of spreading HIV to the clean participants.

Tim said...

Regardless of the merits or lack thereof of Landsburg's idea, the collateral concern about "economists imperialism" seems laughable, as after all, economist actually know something about something.

As opposed to journalists, who rarely know anything about anything - but never hesitate to let basic incompetency in the subject matter stop them from reporting or editorializing. That is a form of "profession" imperialism about which folks should be outraged.

Theo Boehm said...

Let's see more of that sort of thing.

Yes!  That was one of the best threads in recent memory.  It was a great topic.  And it was interesting, logically challenging, and pound-on-the-table funny by turns—yet remarkably free of slash-and-gash trolldom.

You know, I've been grousing about trolls, the quality of comment writing, etc. here and elsewhere.  I know at least a few longtime bloggers and commenters who seem depressed and generally bummed out by the blogosphere.  But this is the kind of thread that restores your faith in the whole enterprise.

Hate to turn all Polyannaish on everybody.

Internet Ronin said...

1 man, 1 woman, both infected with HIV. They have sex. the net result is 1 sex event and 0 transmission events.

Unfortunately, Drill Sgt., as I understand it, that is not necessarily true. There are strains of HIV that one carrier may have that another carrier may not have, particularly if one has been taking HIV-suppressing medications. Specific drug-resistant strains can build up over time. If the partner is treatment naive, or not previously resistant to certain drugs, a transmission can occur which could limit future treament options for that partner.

(BTW, I disagree with your statement that "I'm a bad person." No, you aren't!)

Seven Machos said...

I'm sorry I missed it live.

That's really the sad thing about trolls. Threads have a certain momentum and almost a life force. Trolls come in and snuff it out somehow, even if people try to ignore them.

As an aside, I saw a Youtube clip by the Freakonomics guy -- I think -- on work he did on the economics of gang life. It was one of the most fascinating things I have ever seen.

Also, economics is a great discipline because you get to back up your theories with real facts and your real facts with theories. Hence, even Paul Krugman is occasionally interesting simply because of the accepted methodologies.

The Drill SGT said...

Internet Ronin said...

Unfortunately, Drill Sgt., as I understand it, that is not necessarily true.

yeah, yeah. you are correct. I was keeping it simple. how about

1 man, 1 woman, both infected with the exactly the same strain of HIV.


:)

The Drill SGT said...
This comment has been removed by the author.
Internet Ronin said...

That works, DST!

P. Rich said...

Uh, risk. Every sexual encounter involving one participant with a STD carries a risk of infection for the other. The risk is linearly additive - more encounters, more risk (a bad thing).

Choosing partners at random only means one will eventually have sex with the same percentage of infected individuals as the percentage in the general population from which one is choosing (another bad thing). Two bad things combined equal a really bad thing, the General Theory of Sex clearly not understood by economists.

PS
Condoms fail at a rate of 15+%. Strap that scary fact on the old mental dingus.

blake said...

Let's hear it for Det. Detriech!

Seriously, it was good of him to come and run the gauntlet here. I still think it's a dopey theory but I'm intrigued to see how he sets it up.

mythusmage said...

p. rich said...

"Condoms fail at a rate of 15+%."

That's the overall failure rate. Used properly high quality rubbers have a failure rate of around 2%. Improperly used cheap knockoffs have a failure rate closer to 50% or worse.

So spend the money and learn how to put them on right, it could save your life.

Smilin' Jack said...

I quoted him -- "The justification for that assumption is that if people are systematically wrong in a predictable direction, they'll eventually learn and correct for that" -- and got sarcastic:

I guess I'm cynical. I stopped believing in fairy tale romance long ago.


Doesn't that just confirm his point?

But when Landsburg says "if people are systematically wrong in a predictable direction, they'll eventually learn and correct for that" it's not consistent with his later statment that "it turns out that actual benefits-minus-costs would be higher if you took additional partners." If that were true, people would already have learned and corrected for it.

Fen said...

Used properly high quality rubbers have a failure rate of around 2%.

If there was a 2% chance I would die everytime I operated a motor vehicle, I would stop driving...

Steven E. Landsburg said...

Ann Althouse:

I quoted him -- "The justification for that assumption is that if people are systematically wrong in a predictable direction, they'll eventually learn and correct for that" -- and got sarcastic:

I guess I'm cynical. I stopped believing in fairy tale romance long ago.


Smilin' Jack:


Doesn't that just confirm his point?


Yes, absolutely.

Smilin' Jack again:


But when Landsburg says "if people are systematically wrong in a predictable direction, they'll eventually learn and correct for that" it's not consistent with his later statment that "it turns out that actual benefits-minus-costs would be higher if you took additional partners." If that were true, people would already have learned and corrected for it.


No no no no no. If you took additional partners, *your* benefits minus *your* costs would decrease. That's why you don't do it. But *total* benefits (including benefits to your neighbors) minus costs would increase. That's why there's a disconnect between your private incentives and the social optimum. That disconnect is the entire point.

Analogy: I say that if General Motors shuts down a particular polluting factory then benefits (value of cars produced) minus costs (costs to GM plus costs of pollution) would rise. You say: If that's true, why wouldn't GM realize it and shut down the factory on its own? I say: Because GM doesn't care about the pollution costs.

Thank you, Ann, for making me welcome. I had some trepidation about invading your blog; I'm glad you enjoyed it.

Smilin' Jack said...

Steven E. Landsburg said...

But *total* benefits (including benefits to your neighbors) minus costs would increase. That's why there's a disconnect between your private incentives and the social optimum.


Reminds me of the advice Victorian women were supposedly given re performing the wifely duty: "Just lie back and think of England."

But nowadays no one is going to have sex to benefit their neighbors--as an erotic fantasy, the "social optimum" just doesn't cut it.

Steven E. Landsburg said...

Smilin' Jack:


But nowadays no one is going to have sex to benefit their neighbors--as an erotic fantasy, the "social optimum" just doesn't cut it.


Doesn't that just confirm my point?

reader_iam said...

I understand and appreciate the theory, and the intellectual exercise, as well as the concept of personal incentive and benefits as opposed to the collective.

But then--what to do with that? I'm raising my child to recycle, and so forth. What would you have me teach him in terms of implementation of the insight you're providing?

Ann Althouse said...

"Thank you, Ann, for making me welcome. I had some trepidation about invading your blog; I'm glad you enjoyed it."

It was sublime! Thanks for hashing it out with us.

blake said...

Mr. Landsburg--

Here's where I can't track your thinking (and I am trying): In your analogy GM is polluting. They're actively putting stuff into the environment that harms all.

The chaste aren't polluting. In fact, they're removing themselves as vectors for disease (and not just among sexual partners but children who may inherit diseases).

So, isn't the correct analogy inverted? You're chiding GM because it's not consuming the pollution that others are making? Granted, that's a terrible analogy....

Steven E. Landsburg said...

blake: Thanks for trying hard to follow; I'll try to make it easier.

When GM decides to open a factory, they weigh the benefits (to themselves) against the costs (to themselves) without accounting for the costs to others. Therefore, they'll sometimes open factories where the TOTAL cost (including pollution costs) is greater than the benefit.

When people decide to take extra partners, they weigh the benefits (to themselves) against the costs (to themselves) without accounting for the benefits to others. Therefore, they'll sometimes forego encounters where the TOTAL benefit (including benefits to others) is greater than the cost.

Chastity, then, is like pollution in reverse. Polluters fail to account for the costs they impose on others, so they pollute too much. The relatively chaste fail to account for the benefits they could confer on others, so they loosen up too little.

Internet Ronin said...

I thank you as well, Steven. Not many people would take the time to respond as frequently, or as straightforwardly, as you have. Reader_Iam hits close to the point I apparently failed to clearly state the other day:

I understand and appreciate the theory, and the intellectual exercise, as well as the concept of personal incentive and benefits as opposed to the collective.

But then--what to do with that? I'm raising my child to recycle, and so forth. What would you have me teach him in terms of implementation of the insight you're providing?

Steven E. Landsburg said...

internet_ronin (and reader_iam):

There are a lot of interesting and important intellectual exercises that have no implications at all for raising your children. I expect this is one of them.

The lessons I want to teach are that cost-benefit analysis is subtle, that its implications can be surprising, that it is both fun and enlightening to be surprised. I could say exactly the same things about the fundamental theorem of calculus. I don't think you need to teach the fundamental theorem of calculus to your four year old, but I do think your life will be richer if you understand it.

Bruce Hayden said...

I also thank the author, Steven E. Landsburg, for his contribution here. I await finding his book at the local library with great anticipation. (Well, ok, I probably should buy it, but then why have libraries?)

I have always had an interest in economics, but the normal dry concentration on macro matters caused in the dismal science caused me to go in different directions.

But I found that economics can be interesting and personally advantageous when I was introduced to Freakonmics. Indeed, I follow their blog to this day.

The problem here is that the thesis is a bit counter-intuitive. But that is a result of most of us having a problem transitioning from the specific to the general or societal view. The idea here seems to be that it is obvious that the personal health of those who were chaste and give that up to have more sex and get infected with HIV would be adversely affected by this. But what the author is doing is looking at things from society's point of view, and noting that while these people suffer, the overall rate of infection drops because the newly infected are less likely to pass HIV along than if the carriers had infected more promiscuous partners. In other words, if an HIV positive carrier infects one person, infecting a non-promiscuous person is less likely to lead to additional infections, and thus, in the long run, the societal infection rate will be lower.

Dave Schuler said...

economic analysis does yield lots and lots of accurate predictions

So do astrology, phrenology, and numerology. That's not the standard for a science. The standard is something closer to “accurate most of the time” and at its present state, predicting real world events, economics isn't quite there yet.

When physicists apply the principles of physics properly, their predictions make it possible to create nuclear reactors that actually produce energy and rockets that actually make it to the moon.

If economics were at that point, there would be no stock market, for example (the stock market requires a difference of opinion).

Economics, a study of human nature, remains mostly descriptive rather than predictive. I'm not disparaging economists in saying this. It's not their fault. The study is complicated.

Anne said...

I think what Landsburg is trying to say is: suppose you have a guy who thinks casual sex would be lots of fun, but chooses not to engage in it because he's scared he'll catch a disease. From his point of view, he's made the right choice: the fun of sex is not worth the chance of getting a disease.

But looking at it from society's point of view, if he were to allow himself an occasional fling, his chance of catching something would go up, but his partner's chance and hence all her future partner's chances would go down. When you add in the benefits to all the other people to the fun he would have from the sex, it would balance out the possibility of harm to his health.

Note that none of the above applies if the guy is abstaining from casual sex for moral reasons, because he thinks it's gross, because he's married and doesn't want to hurt his wife, etc. The logical train starts after you assume that the guy does want to have sex; another way of putting it is that the argument only applies to those who abstain only for health reasons.

Steven E. Landsburg said...

Ann:

Note that none of the above applies if the guy is abstaining from casual sex for moral reasons, because he thinks it's gross, because he's married and doesn't want to hurt his wife, etc. The logical train starts after you assume that the guy does want to have sex; another way of putting it is that the argument only applies to those who abstain only for health reasons.

Not quite. What matters is that the guy has toted up costs and benefits it's a close call. Say he really really wants to have sex with his neighbor and also really really wants to stay faithful to his wife. If the second motive dwarfs the first, we should be glad he's staying faithful. But if the second motive only *slightly* dwarfs the first, he can do more good (for his neighbor) by cheating than he does (for himself) by staying faithful.

It doesn't matter whether he's motivated by health, faithfulness, morality, or whatever. What matters is that it's a close call. That won't apply to everyone, but it will always apply to some.

Anne said...

Mr. Landsburg,

I stand corrected.

IMHO, though, it would be helpful for you to stress that at some level the desire to have sex has to be there. I think a lot of people are misunderstanding you as saying, "So what if you find casual sex revolting? Take one for the team, buddy!", or believing that you think you have some logical chain which proves that everyone would like casual sex.

People will find your intellectual exercise a lot more fun if they don't get hung up on these misunderstandings.

Steven E. Landsburg said...

Ann: Thanks for your comments. Do let me point out that what you've seen on the New York Times site is Chapter One of a book that also has a substantial introduction which (I hope) puts the discussion in a larger context and clarifies what I am and am not trying to illustrate.

Ann Althouse said...

Please note that "Anne" isn't me. I'm Ann -- no "e" -- and my last name always appears here with my first.

Steven E. Landsburg said...

Please note that "Anne" isn't me

Oh! I was indeed confused about this, and it was entirely due to careless reading on my part. I stand corrected.