Re: Insurance ruling in Yrp

1

The car insurance issue strikes me as unfair to women, but easy to manage. Whoever is best at selecting women without looking like they are selecting women wins. The life insurance/annuity issue strikes me as potentially problematic. Women will pay quite a bit more for life insurance if the rates ignore gender. That will probably have its biggest effect on two income families with children or households where the main earner is a woman.


Posted by: Moby Hick | Link to this comment | 03- 1-11 9:02 PM
horizontal rule
2

dsquared tweets now?


Posted by: teofilo | Link to this comment | 03- 1-11 9:05 PM
horizontal rule
3

He certainly seems to.


Posted by: nosflow | Link to this comment | 03- 1-11 9:05 PM
horizontal rule
4

2: We live in diminished times.


Posted by: JP Stormcrow | Link to this comment | 03- 1-11 9:08 PM
horizontal rule
5

The obvious solution is to print auto insurance cards on Kotex products.


Posted by: Moby Hick | Link to this comment | 03- 1-11 9:10 PM
horizontal rule
6

2, 4: Funny, I was just thinking that I'd point to his as an example of someone who manages to communicate with brevity and yet remain substantive and non-cutesy. Good Lord do I hate cutesy abbreviations.


Posted by: Witt | Link to this comment | 03- 1-11 9:13 PM
horizontal rule
7

o rly?


Posted by: JP Stormcrow | Link to this comment | 03- 1-11 9:14 PM
horizontal rule
8

Most drivers don't have serious accidents, so the car insurance prior translates into not much money. Speeding tickets and moving violations are the other source of information. Balances out backing into a parking lot pylon while fooling with makeup.

Seriously, who buys annuities or even whole life? These make sense only if equity + bond markets are unavailable, maybe in Japan or Russia.


Posted by: lw | Link to this comment | 03- 1-11 9:15 PM
horizontal rule
9

8.2: I'd think the term life policies would also be affected. Possibly more than whole life, depending on the traits of those who got for it.


Posted by: Moby Hick | Link to this comment | 03- 1-11 9:19 PM
horizontal rule
10

So what European right is this violating? It seems pretty odd. I'm looking over echr.coe.int but haven't found it.


Posted by: Minivet | Link to this comment | 03- 1-11 9:33 PM
horizontal rule
11

Also, the equity market kind of sucks if you want consistency or a small measure of security.


Posted by: Moby Hick | Link to this comment | 03- 1-11 9:33 PM
horizontal rule
12

it seems like this sort of issues is weird because we ignore the significant substantive difference to focus on the secondary, formalistic legal difference.


Posted by: yoyo | Link to this comment | 03- 1-11 10:15 PM
horizontal rule
13

Has Twitter always made it difficult to click the "Back" button away from their site? I hate that shit.


Posted by: Stanley | Link to this comment | 03- 1-11 10:30 PM
horizontal rule
14

I believe it's a misfeature of the recent redesign. That shit is well and truly hateful.


Posted by: nosflow | Link to this comment | 03- 1-11 10:34 PM
horizontal rule
15

Don't know whether the ruling is a good one or not, but if it stands there'll be a lot of champagne corks popping int he various software consultancy firms, because boy will this mean a lot of adaptations to insurance software...


Posted by: Martin Wisse | Link to this comment | 03- 1-11 11:30 PM
horizontal rule
16

Uh-oh. First MayorRahm, now DsquaredDigest. One more and I'm gonna have to sign up.


Posted by: k-sky | Link to this comment | 03- 2-11 12:02 AM
horizontal rule
17

16: Punk Planet was my first magazine subscription! I was unduly excited by that news, even though I hadn't been following the dude's fake Rahm feed.


Posted by: Stanley | Link to this comment | 03- 2-11 12:06 AM
horizontal rule
18

Seriously, who buys annuities or even whole life?

Everyone with a personal pension fund buys annuities in the UK, because it is compulsory to do so. More or less nobody buys them in the rest of the EU, which is possibly how this monumentally fucked decision got through the court in the first place.

The annuity, btw, is intrinsically a very good product and much maligned by unfair comparisons between annuity returns and equity returns. When interest rates are low, it's actually difficult to provide a retirement income out of a fixed pool of savings, and ignoring this risk by pretending that you can live off a savings account doesn't make it go away.


Posted by: dsquared | Link to this comment | 03- 2-11 12:13 AM
horizontal rule
19

I'm hoping Stanley comes up with a suitable pun for tontine, because I'm too tired to think one up.


Posted by: fake accent | Link to this comment | 03- 2-11 12:28 AM
horizontal rule
20

What an incredibly stupid ruling.


Posted by: Robert Halford | Link to this comment | 03- 2-11 12:38 AM
horizontal rule
21

19: Sorry, I got nothing.

On interest rates being low: I was annoyed by the bank teller today, when I was making a deposit, who told me I had a pretty high checking account and could move it over to a new savings account at like zero-point-eleventy interest rate.

Yeah, awesome. Thanks. I'll buy a boat and look for fish instead. Seems like a higher return rate. Bastards.


Posted by: Stanley | Link to this comment | 03- 2-11 12:40 AM
horizontal rule
22

19: Sorry, I got nothing.

No worries. I was just taunting you.


Posted by: fake accent | Link to this comment | 03- 2-11 1:19 AM
horizontal rule
23

dsquared has now posted at greater length about this at CT.


Posted by: chris y | Link to this comment | 03- 2-11 1:48 AM
horizontal rule
24

||

If true, this could all go a bit funny:

http://pakobserver.net/detailnews.asp?id=78009
>


Posted by: nattarGcM ttaM | Link to this comment | 03- 2-11 2:19 AM
horizontal rule
25

If true, this could all go a bit funny

Ha! Bleedin' Ha!

Also, yesterday it was reported that the Yanks have 400 marines on a boat in the Med (presumably relocated from the Halls of Montezuma). I foresee nothing good.


Posted by: chris y | Link to this comment | 03- 2-11 2:34 AM
horizontal rule
26

25.2: to the shores of Tripoli, you mean?

Yes, there are two US landing ships in the Med - the USS Kearsarge, and the rather unfortunately named USS Ponce. The marines are on the Kearsarge.
But Obama's apparently not keen on the idea even of a no-fly zone, so we might yet manage to avoid being involved in a third concurrent Middle Eastern war.


Posted by: ajay | Link to this comment | 03- 2-11 3:06 AM
horizontal rule
27

Obama's apparently not keen on the idea even of a no-fly zone

apparently I read in my Sunday newspaper that this is a world-historical crime for which his blood-drenched name will forever be cursed in the books of infamy, alongside such tyrants as Douglas Hurd.


Posted by: dsquared | Link to this comment | 03- 2-11 3:22 AM
horizontal rule
28

There's also the Mount Whitney, which seems to be essentially a huge floating office block/data centre. Being a superpower means you get to have some really weird kit.


Posted by: Alex | Link to this comment | 03- 2-11 4:00 AM
horizontal rule
29

It's been illegal for a few years in the EU for hair dressers to have different prices for men and women, but it's never honored or enforced here, which upsets my little sister to no end.


Posted by: David | Link to this comment | 03- 2-11 5:28 AM
horizontal rule
30

27: They'll value their new government much more if they have to fight for it themselves. What we obtain too cheaply we esteem too lightly, etc.


Posted by: ajay | Link to this comment | 03- 2-11 5:37 AM
horizontal rule
31

It's been illegal for a few years in the EU for hair dressers to have different prices for men and women, but it's never honored or enforced here, which upsets my little sister to no end.

Yeah but, no but. There's no law which says they have to charge the same for a quick number 3 all over as for a perm and layer job that takes two people 90 elapsed minutes. In my limited experience the people who do the latter sort of work for men charge the same to all customers and women who want the two minute buzz cut go to men's barbers for it.

Does your little sister wear her hair in a style commonly affected by men? Because otherwise it's apples and oranges.


Posted by: chris y | Link to this comment | 03- 2-11 5:38 AM
horizontal rule
32

I used to work in auto insurance, on the pricing side. I can tell you, the primary determinant of your rate in the US is whatever their best approximation for socioeconomic status is, the holy trinity being occupation, education level, and credit score (location factors into this to some extent as well). Everything else is really small potatoes.


Posted by: fiatmoney | Link to this comment | 03- 2-11 6:30 AM
horizontal rule
33

19: The Tontine Macoute is the one you don't want to join.


Posted by: mcmc | Link to this comment | 03- 2-11 6:30 AM
horizontal rule
34

32: I thought it was all about the zip code? On the theory that people of means don't crash into each other, but people in poor neighborhoods do so all the time? That's how it's always seemed to work in Boston.


Posted by: mcmc | Link to this comment | 03- 2-11 6:32 AM
horizontal rule
35

34: Used to be, way back. But the regulatory infrastructure on auto insurance can be really hardcore as far as forcing you to justify your rate decisions (insurance commissioners are elected in a good number of states), and they really don't like geographic distinctions - except places like NY and FL where you're really talking about 2-3 different states in terms of traffic patterns. Once they accuse you of "redlining," even if it's justified by claim data, the relationship gets hostile and they might just start refusing your requests to change rates.

Boston is a special case because in MA, rates are almost entirely determined by the state. Something like 1.5 years ago, they loosened up a bit, but I believe the state does mandate a lot of location-based pricing there.

Folks in the industry are so paranoid about being accused of redlining that I was literally told not to make a map that plotted claim density or any of our pricing variables by zip code.


Posted by: fiatmoney | Link to this comment | 03- 2-11 6:39 AM
horizontal rule
36

20

What an incredibly stupid ruling.

Do you believe insurers should be allowed to base rates on race and religion also? If not, what is different about sex?


Posted by: James B. Shearer | Link to this comment | 03- 2-11 7:23 AM
horizontal rule
37

If not, what is different about sex?

When you're older, it will all make sense. (Not really.)


Posted by: Moby Hick | Link to this comment | 03- 2-11 7:25 AM
horizontal rule
38

Do you believe insurers should be allowed to base rates on race and religion also? If not, what is different about sex?

In the case of pensions, I'd be astonished if they didn't already do so. There are a bunch of diseases that different ethnic groups are differentially likely to get, and they are well documented. If those odds aren't factored into the actuarial calculations I'd be amazed.

On the car insurance front, I don't suppose there are differences that aren't already covered under gender and income.

How would an insurer benefit from discriminating on grounds of religion? Do people drive more dangerously if they've got their plastic Jesus riding on the dashboard of their car? If they do, I'm sure the data is in the algorithms.


Posted by: chris y | Link to this comment | 03- 2-11 7:32 AM
horizontal rule
39

On the theory that people of means don't crash into each other, but people in poor neighborhoods do so all the time?

As well as the redlining laws, you need to take into account that rich people sometimes crash into more expensive things than poor people (this is why people working in classical music, even in administrative roles, will sometimes find insurance companies behaving very strangely - it only takes one low-ranking employee to crash while giving a lift to someone who was carrying her Stradivarius, and all sorts of internal directives and exclusions come down!)


Posted by: dsquared | Link to this comment | 03- 2-11 7:33 AM
horizontal rule
40

There are a bunch of diseases that different ethnic groups are differentially likely to get, and they are well documented. If those odds aren't factored into the actuarial calculations I'd be amazed

Be amazed. There are such differences, but they're not really enough to be a major contributor to the variance. Something like 95% of all life policies in the UK end up being offered on standard terms.

Remember, as per the D2D post I linked at CT, "medical risk to the insured" doesn't map at all well on to "financial risk to the insurer".


Posted by: dsquared | Link to this comment | 03- 2-11 7:36 AM
horizontal rule
41

Remember, as per the D2D post I linked at CT, "medical risk to the insured" doesn't map at all well on to "financial risk to the insurer".

Maybe they need a combined medical-life policy. That way, the insurer would have trouble saying money by denying treatment.


Posted by: Moby Hick | Link to this comment | 03- 2-11 7:38 AM
horizontal rule
42

There are such differences, but they're not really enough to be a major contributor to the variance.

Then they're factored in with a zero value. If a virus suddenly evolved which allowed Welshmen, and only Welshmen to live to 150 as a matter of course, how long do you think it would take the Pru to push your premiums through the roof? If in fact there's no available reason for companies to vary policies on racial grounds, then JS's question becomes meaningless.


Posted by: chris y | Link to this comment | 03- 2-11 7:43 AM
horizontal rule
43

How would an insurer benefit from discriminating on grounds of religion? Do people drive more dangerously if they've got their plastic Jesus riding on the dashboard of their car?

Ceteris paribus, you'd expect Muslims to be safer drivers because they're less likely to be drunk at the wheel. But I suspect that so much of the variation can be explained by other factors that being able to discriminate by religion isn't really that valuable. There are Muslim car insurers, but that's more a question of Muslims wanting a halal workaround - conventional insurance is haram, because it involves gambling.


Posted by: ajay | Link to this comment | 03- 2-11 7:48 AM
horizontal rule
44

How would an insurer benefit from discriminating on grounds of religion?

In increasing order of car-crashiness:
Amish (over 18), Unitarian, Jewish (orthodox), Quaker, Jewish (other), Episcopalian (regular), Lutheran, Catholic (Not Irish, Not Hispanic), Episcopalian (Anglican kind), Methodist, LDS, Catholic (Irish), Southern Baptist, Catholic (Hispanic), Snake handler, Amish (under 18).


Posted by: Moby Hick | Link to this comment | 03- 2-11 7:53 AM
horizontal rule
45

Moby doesn't even see Muslims.


Posted by: ajay | Link to this comment | 03- 2-11 8:09 AM
horizontal rule
46

Do people drive more dangerously if they've got their plastic Jesus riding on the dashboard of their car?

I've heard anecdotal reports that they are indifferent to safe driving conditions.


Posted by: foolishmortal | Link to this comment | 03- 2-11 8:11 AM
horizontal rule
47

Nobody sees them. Which is why they are in so many car accidents.


Posted by: Moby Hick | Link to this comment | 03- 2-11 8:11 AM
horizontal rule
48

47 to 45.


Posted by: Moby Hick | Link to this comment | 03- 2-11 8:12 AM
horizontal rule
49

44.

i. Did you just make that up, or where's your source?
ii. What is the variance? Is it significant?
iii. How would it be weighted against other factors?
iv. WTF is Snake Handlers doing in there?


Posted by: chris y | Link to this comment | 03- 2-11 8:12 AM
horizontal rule
50

49.4: Crashing into bridge abutments as a result of rattlesnake bites, looks like.


Posted by: LizardBreath | Link to this comment | 03- 2-11 8:13 AM
horizontal rule
51

If a virus suddenly evolved which allowed Welshmen, and only Welshmen to live to 150 as a matter of course, how long do you think it would take the Pru to push your premiums through the roof?

In this horrifically dystopian scenario, I think dsquared's life insurance premiums would actually go through the floor, because the Pru would get 150 years worth of them before it had to pay out the same lump sum, rather than 78 years worth. The same goes for medical costs: most of your lifetime medical costs happen in your last year of life, so dsquared's medical insurers would have 149 years of relatively low expenditure before that rather than 77.

His pension provider, on the other hand, would have a screaming fit, and probably start looking into Chester bylaws and whether they could be extended to cover North London.
http://www.chesterwiki.com/Shooting_the_Welsh!


Posted by: ajay | Link to this comment | 03- 2-11 8:15 AM
horizontal rule
52

I have to go to a meeting now. No time to find a link to the source.


Posted by: Moby Hick | Link to this comment | 03- 2-11 8:15 AM
horizontal rule
53

50: and the meth and moonshine probably don't help either.


Posted by: ajay | Link to this comment | 03- 2-11 8:16 AM
horizontal rule
54

Christian Scientists are probably on the less-crashy side of things.


Posted by: JP Stormcrow | Link to this comment | 03- 2-11 8:19 AM
horizontal rule
55

54: They go between Episcopalian (regular) and Lutheran. The full list takes too much space.


Posted by: Moby Hick | Link to this comment | 03- 2-11 8:22 AM
horizontal rule
56

re: 51

Yeah, several of the newspaper articles I've read keep seeming to get things the wrong way round, vis changes in life insurance rates versus cost of annuities for men and women. But I'm always a bit wary since I'm likely to be told I'm making some Actuarial 101 error and am financially stupid.


Posted by: nattarGcM ttaM | Link to this comment | 03- 2-11 8:25 AM
horizontal rule
57

52: It's not just made up? I figured the car insurance premiums for Amish adults was a dead giveaway.


Posted by: LizardBreath | Link to this comment | 03- 2-11 8:29 AM
horizontal rule
58

Yes, it's made up.


Posted by: Moby Hick | Link to this comment | 03- 2-11 8:31 AM
horizontal rule
59

Who let the dogs out?


Posted by: Pauly Shore | Link to this comment | 03- 2-11 8:33 AM
horizontal rule
60

To 31: Right, they could have switched to pay based on hair length, but the didn't. It's still based on sex. Note that "here" for me is Sweden, not the UK, btw.


Posted by: David | Link to this comment | 03- 2-11 8:45 AM
horizontal rule
61

Then they're factored in with a zero value

Yes that's about right, but things have to be very extreme indeed before your average life actuary will assign anything a non-zero value - there are smokers' annuities but that's about it. Even the HIV-positive can't always get favourable annuity rates (this is mainly due to a lot of chancers having bought life policies off HIV sufferers in the 1980s and having absolutely done their nuts with the invention of antiretrovirals).


Posted by: dsquared | Link to this comment | 03- 2-11 8:52 AM
horizontal rule
62

I'm surprised this thread has turned to the economic and not the legal side.


Posted by: Minivet | Link to this comment | 03- 2-11 9:23 AM
horizontal rule
63

62. Easy. No British lawyers here, but everybody feels able to bs about economics (even economists).


Posted by: chris y | Link to this comment | 03- 2-11 9:32 AM
horizontal rule
64

Also there isn't really a legal point to argue. ECJ's made its mind up.


Posted by: ajay | Link to this comment | 03- 2-11 9:38 AM
horizontal rule
65

Also there isn't really a legal point to argue. ECJ's made its mind up.

As if impracticality stopped discussion before.


Posted by: nosflow | Link to this comment | 03- 2-11 9:45 AM
horizontal rule
66

40

Be amazed. There are such differences, but they're not really enough to be a major contributor to the variance. Something like 95% of all life policies in the UK end up being offered on standard terms.

In the United States the difference in life expectancy at birth between blacks and whites is approximately the same as between males and females. However it is apparently illegal to factor this into insurance rates.


Posted by: James B. Shearer | Link to this comment | 03- 2-11 6:31 PM
horizontal rule
67

James is right about life expectancy at birth. However, most people don't buy life insurance at birth (stay away from me, you creepy Gerber ads that haunted me for months), I looked at life expectancy at 65. Obviously, this is nearly as pointless for these purposes as looking at life expectancy at birth. Life expectancy at thirty could have been better, but this was starting to be effort and I quit with what I had.

Life expectancy gets more even between races as time goes on. At 65, for white people, men can expect 17.3 more years and women 19.9, for a difference of 2.6 years. For all races, the gender difference is 2.7 years. White people as a whole get 18.7 more years and black people get 17.2 years more, for a difference of only 1.5 years. Most of that difference is because of lower life expectancy for black men. Black women are only 1.2 years below white women.

Anyway, I don't doubt there is a race difference, but I don't think you can look at overall life expectancy to judge how big it is. In addition to the lower interracial difference at higher ages, I'd also think that the people most at risk of dying before paying nice premiums are those least likely to try to buy life insurance.


Posted by: | Link to this comment | 03- 2-11 7:00 PM
horizontal rule
68

That was me.


Posted by: Moby Hick | Link to this comment | 03- 2-11 7:00 PM
horizontal rule
69

38

How would an insurer benefit from discriminating on grounds of religion? ...

At one time Jews had difficulty obtaining fire insurance because of their purported propensity for arson. See this for example.


Posted by: James B. Shearer | Link to this comment | 03- 2-11 7:28 PM
horizontal rule
70

I have heard of that stereotype.


Posted by: Moby Hick | Link to this comment | 03- 2-11 7:32 PM
horizontal rule
71

67

According to table A (p. 2) in this large pdf file at age 30 life expectancy for males and females is 46.9 and 51.3 and for blacks and whites is 45.5 and 49.5. So the race difference (4.0 years) is still 90% of the sex difference (4.4 years).

Interestingly if this source is to be believed over age 80 blacks have a longer life expectancy but I suspect this is some sort of error.


Posted by: James B. Shearer | Link to this comment | 03- 2-11 7:40 PM
horizontal rule
72

71.2: That's a well known effect, not an error.


Posted by: Moby Hick | Link to this comment | 03- 2-11 7:43 PM
horizontal rule
73

69: but that wouldn't benefit the insurer, since Jews don't really have a propensity for arson. (And accidents involving unattended menorahs can't be common enough to cause a significant effect.)

72: presumably because if you've survived eighty years of being a black man in the US then there really isn't much that can kill you.


Posted by: ajay | Link to this comment | 03- 3-11 2:17 AM
horizontal rule
74

73,69: this is a classic example of a confounding factor: condtional on the industry, Jewish customes had no different propensity for arson, but historically in the USA they tended to be well represented in the garment and textile industries, where things often go on fire during difficult trading conditions (Lloyds brokers occasionally used to refer to "Scottish bankruptcy insurance" ie fire insurance)


Posted by: dsquared | Link to this comment | 03- 3-11 4:14 AM
horizontal rule
75

Ah ha. That makes a sort of sense, but it still wouldn't benefit the insurer to charge higher premiums to Jews rather than simply charging higher premiums to people in the rag trade. After all, most Jews wouldn't have been in the rag trade, and Gentile textile manufacturers would have been just as likely to suffer convenient fires.
After all, the insurer is presumably already taking the nature of the business into account. A flour mill isn't going to be charged the same fire insurance premium as an ice house.


Posted by: ajay | Link to this comment | 03- 3-11 4:29 AM
horizontal rule
76

Still not clear to me why some of you think it is fine to legally prohibit setting insurance rates based on race or religion but not on sex.


Posted by: James B. Shearer | Link to this comment | 03- 3-11 6:20 AM
horizontal rule
77

Well, as earlier discussed, there's quite a lot about economics that's not clear to you or your cohorts.


Posted by: ajay | Link to this comment | 03- 3-11 6:48 AM
horizontal rule
78

Is it in fact the case that setting insurance rates based on calculations which include factors associated with race or religion, NOT 'based on race or religion' is illegal? What has been argued is that these factors are in practice so trivial that they are generally disregarded actuarily, whereas the discrepancy by sex for life expectancy in middle age and for accident rate while driving is large enough to show up in the sums*.

I think everybody would agree that setting insurance rates based on race or religion, e.g. "I'm a Catholic and you're Orthodox so I'm going to charge you an excess premium because I hate you", would be unacceptable. That isn't the point.

*All I know about actuarial science is that I'm not interested in it, so I don't have a view as to the correctness or otherwise of people like dsquared's assertions, except that usually Daniel doesn't say stuff like that without knowing what he's talking about. Where I made a guess upthread I was apparently wrong.


Posted by: chris y | Link to this comment | 03- 3-11 6:52 AM
horizontal rule
79

78

Is it in fact the case that setting insurance rates based on calculations which include factors associated with race or religion, NOT 'based on race or religion' is illegal? What has been argued is that these factors are in practice so trivial that they are generally disregarded actuarily, whereas the discrepancy by sex for life expectancy in middle age and for accident rate while driving is large enough to show up in the sums*.

Insurance rates (in the US) are set on gender directly (not factors associated with gender). See here for example. It is my understanding that at one time companies also set rates on race directly but that this is now illegal.


Posted by: James B. Shearer | Link to this comment | 03- 3-11 7:18 AM
horizontal rule
80

78: I wouldn't say that the racial differences are trivial in the U.S. Rather, I would say that they are more likely to be spurious. On many health questions, you see a great many racial differences that disappear or weaken when you control for other factors (education, income, employment, etc.). The gender ones are less likely to go away which would seem to show that gender isn't serving as a proxy for something else, but rather is a key difference. I'm speaking generally here, as I don't do actuarial work, but I don't understand how you could sell life insurance at the same rates to men and women. You could compute the expected costs of a policy, but if you were wrong on the gender ratio of your clients, you would go broke. They (European life insurers) will likely find a work-around, which shouldn't be that hard statistically, but will be more burdensome for those buying insurance and probably more litigation.


Posted by: Moby Hick | Link to this comment | 03- 3-11 7:24 AM
horizontal rule
81

74

this is a classic example of a confounding factor: condtional on the industry, Jewish customes had no different propensity for arson, but historically in the USA they tended to be well represented in the garment and textile industries, where things often go on fire during difficult trading conditions (Lloyds brokers occasionally used to refer to "Scottish bankruptcy insurance" ie fire insurance)

I have no idea about relative propensity for arson but according to Andrew Tobias in cases like that you want to underwrite the owner not the property. It would not surprise me if different cultural groups in fact have significantly different attitudes about insurance fraud.


Posted by: James B. Shearer | Link to this comment | 03- 3-11 7:26 AM
horizontal rule
82

80

... but I don't understand how you could sell life insurance at the same rates to men and women. You could compute the expected costs of a policy, but if you were wrong on the gender ratio of your clients, you would go broke. ...

It would be simple enough, at worst you could just charge everybody the man's rate. Insurance companies have to worry about adverse selection all the time. For example people who buy annuities live longer than average. So you make a guess. Of course you can lose money if you are way off but there are a lot of ways to go broke selling insurance.


Posted by: James B. Shearer | Link to this comment | 03- 3-11 7:35 AM
horizontal rule
83

On many health questions, you see a great many racial differences that disappear or weaken when you control for other factors (education, income, employment, etc.). The gender ones are less likely to go away which would seem to show that gender isn't serving as a proxy for something else, but rather is a key difference.

Exactly. You'll never be selling insurance or whatever based solely on race; if you were then it would indeed make commercial sense to discriminate, because there are real racial differences in life expectancy. But, given all the other information that you'll get, race doesn't have much additional explanatory value. A black guy may well live (say) three years less on average than a white guy; but the difference in life expectancy between a black university-educated male accountant aged 45 who's married, doesn't do dangerous sports, has no heart problems and lives in Manchester and his white colleague ditto is likely to be minuscule. The difference between the black accountant and his female counterpart ditto, however, will be noticeable.


Posted by: ajay | Link to this comment | 03- 3-11 7:37 AM
horizontal rule
84

82: If you charged everyone the man's rate, you'd sell very few policies and go broke for not selling any policies. And insurance companies don't guess about adverse selection. Obviously, they sometimes make mistakes, but it is in the models. This is one of the reasons you can get life insurance much cheaper as part of a group (e.g. employer provided) than buying as an individual.

Also, this isn't adverse selection as that refers to an information asymmetry where the buyer knows something the seller doesn't (e.g. buyer doesn't mention they huff paint). Both parties know the gender of the buyer.


Posted by: Moby Hick | Link to this comment | 03- 3-11 7:49 AM
horizontal rule
85

but the difference in life expectancy between a black university-educated male accountant aged 45 who's married, doesn't do dangerous sports, has no heart problems and lives in Manchester and his white colleague ditto is likely to be minuscule.

I would guess that 'minuscule' is way too strong of a word for that, at least in the U.S. "Small enough to ignore in the cause of greater social justice" is probably more accurate.


Posted by: Moby Hick | Link to this comment | 03- 3-11 7:51 AM
horizontal rule
86

75: yes, you're right - this was probably the root of the prejudice but the actual pricing was motivated by simple racism.

76: I would advise against a career in underwriting at this stage then.


Posted by: dsquared | Link to this comment | 03- 3-11 7:51 AM
horizontal rule
87

86.1: I'm sure it was...

James, here's something else for you to puzzle over: why do these horrible liberals support legislation designed to prevent age discrimination, but still think that annuities should be priced depending on the age of the buyer?


Posted by: ajay | Link to this comment | 03- 3-11 8:15 AM
horizontal rule
88

87: Ooh, I know that one! Hypocrisy!


Posted by: JP Stormcrow | Link to this comment | 03- 3-11 8:29 AM
horizontal rule
89

I wonder if there was also a stereotype of recent immigrants more generally committing arson; if so, that early scene in Double Indemnity would be a reflection of it. A lot of the keywords come up in a book called "Nativism, discrimination, and images of immigrants," but its text isn't viewable.


Posted by: Minivet | Link to this comment | 03- 3-11 12:10 PM
horizontal rule
90

84

Also, this isn't adverse selection as that refers to an information asymmetry where the buyer knows something the seller doesn't (e.g. buyer doesn't mention they huff paint). Both parties know the gender of the buyer.

According to wikipedia it is:

The term adverse selection was originally used in insurance. It describes a situation where an individual's demand for insurance (either the propensity to buy insurance, or the quantity purchased, or both) is positively correlated with the individual's risk of loss (e.g. higher risks buy more insurance), and the insurer is unable to allow for this correlation in the price of insurance[1]. This may be because of private information known only to the individual (information asymmetry), or because of regulations or social norms which prevent the insurer from using certain categories of known information to set prices (e.g. the insurer may be prohibited from using information such as gender or ethnic origin or genetic test results). The latter scenario is sometimes referred to as 'regulatory adverse selection'.[2]


Posted by: James B. Shearer | Link to this comment | 03- 3-11 6:48 PM
horizontal rule