Over at this post, someone was asking about the actuarial profession. The main action is in the comment thread, and you can see that there is a huge amount of silliness going on, because I told some people at the Actuarial Outpost about it and we brought our in-jokes over. But they went away and I stayed to answer questions seriously.

One of the last questions I was asked was this:

Just one more questions actually, to what extent do you think being an actuary can help people?

So, here's the thing.

Actuaries tend to be very far removed from the beneficiaries of their work. We sit in cubicles, staring at computer screens. We wrangle with crap data. We have legacy spreadsheets we are trying to tame into something that gives us sane answers. We have proprietary database systems that create new bugs after 20 years of operation.... yay.

It can be a very isolating job. Many people find it very boring. The barrier to entry of the exams is rather high -- I've taught exam seminars and tried to help people with their frustrations of having to sit the same exams multiple times (and even tried helping people on exams I didn't teach). Then many people see that the math and concepts on the exams are way beyond what they have to implement at work, and don't see the intellectual challenge.

That's one perspective.

Here's another.

We've seen in a major way that risk exists. We're all going to die; it's just a matter of when. If you own stuff, that stuff can get destroyed or damaged and might need to be replaced. If you do stuff, something bad could happen, and you could get sued. It would be nice if there were ways that we could prepare for these risks, reduce the costs for preparing for these risks, and be able to count on that protection.

I work in the field of "Life", which means, of course, I'm concerned with death.

The types of products I look at are life insurance and annuities. Now modern versions of these products also have investment risk protections, but I'm going to focus solely on mortality/longevity risks. And I will look at these two products separately.

For life insurance, the main risk is that you'll die early, and people who depend on you will be bereft not only of your presence but also the financial support you had been giving. If things go to expectation, one could save money up while working, and then when one dies, there's a nice inheritance for these people. But what if, like my father, you die when you're 38? You didn't have enough time to save up anything.

So you need to protect against such early death. I'm not going to get into all the nasty details, but in general, the cheap life insurance coverage people get in term insurance (which is what I recommend for most people -- get term insurance, plus save up money. Past a certain age, you should have enough money saved to provide for beneficiaries and you will no longer need life insurance coverage. There are special situations, though, that may require something more complicated.)

So actuaries come in on two parts of this: setting the price (or premiums) and setting the reserves (and checking on capital).

Lots of things go into pricing, but the main thing is that the insurance company needs to be charging enough for the risk...otherwise, there's a pretty good chance the insurance company will go bust. And they did go bust pretty often early on in the history of insurance.

But then, we could just set really high prices, and that would be sufficient... but then the protection for your risk coverage wouldn't be worth it. This has also been in evolution over the ages. There are pricing swings (and again, I'm not going into the details here. But those interested in becoming actuaries, I can explain another time).

On the other side is reserving. This is setting aside money to cover the promises made. If prices are good, then adequate reserves can be set up (and adequate capital, which is the "extra" to cover adverse situations...like, say, a catastrophic earthquake/tsunami). Again, without reserves, chances can be good that the promises being made by the life insurance policy won't be fulfilled.

Obviously, insurance is a highly regulated industry. So actuaries are involved on the regulatory side. And also as auditors.

Actuaries are also involved in the design of these risk protection products.

So the whole point is that we are trying to help people reduce the financial impact of very real risks, by transferring these risks into affordable, regular premiums -- and setting it up so that you can depend on that protection being there. It took lots of disasters and failures, and a long time of developing the tools of actuarial science, for us to have a dependable insurance system. I've gone back and researched a lot of the 19th century insurance biz.... and it's ugly.

It can help to keep this big picture in your mind while you do your work. This is a very details-oriented profession (those details are very important); but this does not mean you cannot see how it contributes to helping people deal with risk better.

Yes, I've greatly simplified a lot of the issues, but that's what happens when you take the view from Mt. Everest. There is actuarial work other than life insurance, there are other risks and actuaries haven't always been that great at dealing with some of those risks in recent history (cf the failure of Equitable Life in the UK which resulted in a review of the profession there (hmm no wikipedia article on that), and their self-governance was taken away. The closest thing to that in the U.S., I think, is what's going on in public pensions.)

And someone else wrote the following:

reminds me of about schmidt where he sees all his work thrown in the garbage; yours will end up there also. 90% of actuarial work is meaningless.

I do the minimal amount of work and only extra if I find it interesting as a puzzle.

My response:

100 years from now no one will remember you either. So what?

The vast majority of people's work and lives have no lasting effect. Doesn't mean being a slackass is a great idea. If you want to be a slackass, then hey - go for it. But you're going to have a particular kind of career arc.

The stuff I talked about I found interesting. It was an intellectual challenge to me, I learned a lot, and I've gotten to reuse what I learned in a variety of places...leveraging it to all sorts of interesting things. I'm sticking my nose into other people's business all the time, which I know is annoying to these other people sometimes, but it has gotten me some of my most interesting projects and connections.

I find that opportunity rarely walks up to a person and shouts howdy... usually I have to hunt it down.

A lot of my work =has= been thrown out, in front of my face, while I was still there. I knew why it was thrown out, too. A lot of what we do is not for the ages.... but guess what? That's true of all professions. That's true of all human life. Very little lives on past us. This does not make it meaningless.

The point is to think of what is important to other people. How your small piece contributes to the wider world.

Here's an example from my working life: the last major project I worked on at TIAA-CREF was a portfolio segmentation. I'm not going to get into the details, but the point was that it was becoming difficult to "price" the retirement annuities so that annuity payments wouldn't get reduced. They hadn't reduced people's payments since the early 1990s. They knew the impact of people in retirement of lowering the amount being paid.... but you can't just increase or hold payments level if the money can't support it. TIAA has been around since 1918, and has always fulfilled its promises. They've got people who have had accounts with them for decades... people who get retirement income from them for over 40 years! We needed to figure out a way to be able to keep the payments up in a very challenging interest rate environment.

My piece of the project was but one of many. I cracked jokes that if I screwed up, no one would know for twenty years. But the way it would really work is that my part was just the starting point, and that this would be continually adjusted every year. There will be dozens of people who touch this to keep it going while this strategy is around. And who knows? Perhaps in 5-10 years, they'll do something different and what I did will no longer be pertinent. But maybe its effect will still be around 50 years from now, rolled up in the efforts of all these other people.

Finally, another way actuaries can help people is to explain these technical concepts of risk to help people make better decisions about it. To that end, I co-wrote an article with Benjamin Goodman on annuities, because not enough people are protecting themselves from outliving their savings.

So no, we're not saving lives. The people we help don't know our names (and we don't necessarily know theirs....but a big part of my job is checking over claims files. And that can get depressing.) And we won't be remembered long after we retire. But we do have some impact on people.

Meep

Meep is a member of the Irish Catholic mafia, having a suspiciously high number of green-eyed, red-haired friends. While she doesn’t have red hair herself [except when she goes into the sun (rare for any vampire)], she does have green eyes. She’s a raving Papist and is a life actuary on the side [i.e., she counts dead people]. An amateur pain-in-the-ass [willing to go pro!], she likes covering retirement, mortality, math, and education issues.

Factors that come into play in making these recommendations: rate of false positives, rate of incidence of whatever is being screened for, and how the benefits are measured.

Before I get into my stuff, let me link other coverage of this issue, and you can ignore all the stuff I have to say.

NPR's coverage on cancer screening guidelines

Coverage from a Florida newspaper

Coverage from ABC News

Let me run a very simplified model. To simplify matters, I'm going to pretend that the rate of false negatives [i.e., you have the condition screened for but the test doesn't find it] is zero; this is not going to appreciably change the analysis below. If we have a test with a high false negative rate, then who is going to want to use the test? So let's assume that the test we use will catch all actual incidences of the condition.

However, some conditions are more rare than others, and there is always some level of false positives - some growth found that isn't cancer, but you still will have to followup and do a biopsy, say.

Let d = the percentage of people in a given population who have a disease, and p = the probability of false positive [i.e., probability that the test gives a positive result, given you don't actually have the condition]. We are going to test the population: in percentage terms, how many false positives, true positives, and true negatives are we going to get? [remember, we're saying false negatives are not occurring].

true positives = d
false positives = (1 - d) p
true negatives = (1 - d) (1 - p)

Probability you actually have the disease, given you test positive: d / [d + (1 - d) d]

So let's try different combos of d and p, and see results

So when the disease has low incidence in the population, even for a relatively accurate test, most of the positive results will be false. This is just one piece in the decision of recommending which groups [by age, family history, race] should get screenings. On the cost side, you can figure in the costs of the screening tests themselves, and then the costs of followup [for false positives and true positives alike.]. Sometimes the anguish of false positives is added into the cost equation, but I think that's usually for the press releases and not necessarily for an official cost-benefit analysis. The costs are the followups. For those who are true positives, you need to wrap in the cost of treatment for the condition.

But what about the benefit side? While the costs are very easy to define, these benefits are harder to define. Forget about improved quality of life [which obviously has value to the individual] for treatment of a condition, let's say we use are measuring stick as added years to life expectancy if true positive is treated. But what dollar value are you going to put on those extra years of life?

That's where it gets murky and political.

I'm going to swerve into the personal for a moment. My father died of a heart attack at a rather young age, and my family sued a medical center and the head doctor for misdiagnosis and mistreatment [our lawyer? John Edwards. This was his pre-senatorial days, of course]. I was a party to the lawsuit, but only 19 when it was settled, so I don't know all the details, but there is one thing I know -- part of the number in the requested restitution was based on the fact that my dad was an engineer working for IBM, and there was a calculation for lost wages due to early death. This is quite common in medical malpractice cases [esp. when the person who died or is disabled had been in a high salary career before the medical incident].

Can you imagine if the official government screening advice were based on your salary? Yeah, that's not going to fly.

So what usually happens is some dollar amount is put on extra years of life, not really paying attention to which years they are, and not treating the value of these years as infinite. Think about it - you do something similar in your own life, though probably more implicitly than the explicit calculation made here.

People aren't perfectly rational, to be sure, but we're making these cost/benefit tradeoffs in terms of risk all the time. We know that traveling outside the house is risky compared to staying at home, but most of us think the benefits outweigh the risks we take on. But we don't all value benefits the same -- some of us don't think the benefit of sky diving, say, outweighs the extra mortality risk.

Back to governmental recommendations. They've got to use some number to determine the cutoff, and depending on how they value those lives. This is where it becomes political. The higher the dollar value you put on those extra years, the more screening you're going to recommend even into those areas where there are lots of false positives; the lower the dollar value, the less screening will be recommended.

I have nothing against these recommendations being set. The problem is how they will be used, and how they will be adjusted over time. In a competitive market, those who put a higher dollar value on those extra years will pay more for more screenings and more treatment; they will see more false positives and costly followups, but will also catch more true positives than a group that puts a lower dollar value on those extra years of life.

In the private insurance marketplace, the insurers can and do make different decisions on screenings supported, and they go with two things: what their customers want and what is required by state insurance regulation [if it pertains. I'm not getting into insurance regulation right now.]

They do do these sorts of analyses, too, though they will look in terms of costs in a different way -- if one doesn't do the screening, what will treatment costs be if the condition is detected later on? It's not so much cost/benefit analysis as cost/cost analysis - what will the costs be under two different screening set ups? Sometimes the insurance companies offer screenings as part of coverage, even without being asked by customers, because they know the ultimate costs will be less if problems are found early on. [Similarly, some employers will do stuff like free flu shots on site, because that cost is much less than the cost of sick days due to flu].

You cannot escape tradeoffs in costs and benefits, no matter who is doing the choosing. People are going to have an emotional reaction, because nobody wants to put an explicit dollar amount on their own life, and when these recommendations change, that calculus becomes more evident to all. The question becomes who gets to decide, and whether there will be one decision for all.

As it is now, there are many ultimate "deciders". Will that continue under "health care reform" as envisioned by the current crop of politicians?

Meep

Meep is a member of the Irish Catholic mafia, having a suspiciously high number of green-eyed, red-haired friends. While she doesn’t have red hair herself [except when she goes into the sun (rare for any vampire)], she does have green eyes. She’s a raving Papist and is a life actuary on the side [i.e., she counts dead people]. An amateur pain-in-the-ass [willing to go pro!], she likes covering retirement, mortality, math, and education issues.

Here is something that has not brought up in many of these health care debates: women use health care a lot more than men do.

First, there are the obvious costs of pregnancy and childbirth. I've done that three times, and it's not cheap, especially if you get a C-section. C-sections are real surgery with all the costs that go with that [and anyone who goes for elective C-sections are freaking nuts! Ladies, the vagina is built to let kids out, it does bounce back a lot better than do your abdominal muscles being sliced through. Okay, no more graphic mommy talk. In this post.]

Another thing is that women are more likely to go to the doctor regularly, and part of that are those annual Pap smears [at younger ages] and mammograms [at older ages]. These aren't free, though much cheaper than surgery.

All of this health care "reform" is about money. Who is paying for what and how much, and who is getting paid and how much. The federal and state governments already have a very heavy hand in this stuff. Only a few health-related industries escape much of the onerous government funding and regulation: LASIK surgery, plastic surgery, and abortion to name a few. There are already a bunch of price distortions due to Medicare, Medicaid, S-CHIP, and state insurance regulations; but the octopus has decided that level of power is not enough. The people who are going to be most impacted by federal government control of health care are those who use the most health care. The biggest groups: the elderly and women.

But wait, women make up half the elderly, so they're being double-counted, right? No, it's worse than that. Women make up a lot more than half the elderly. Here's where life expectancy comes into play. I'm going to use the 2000 calendar year Social Security tables [without mortality improvement] to calculate differences. And I'm going to measure life expectancy from different points, not birth. I'm going to check from age 21, and from age 65 [when Medicare coverage begins].

[A quick aside: absent abortion or other artificial conception methods, more males are born than females. Males in the U.S. have higher mortality rates at all ages than do women. I will explain this another time.]

Here are the age at death stats starting from age 21:

And here they are starting from age 65

Let's pretend there are a whole bunch of actuarial caveats here. The point is not the exact numbers, though, but the relative numbers, the gaps between men and women.

Let's start with the from age 21. I'm not using life expectancy, because I want to look at actual probabilities - the 50th percentile [the median] is a good starting point. Using the mortality table numbers, half will die before this age and half after this age -- note there is a 5-year gap here. If we want to get more specific, at the median death age for women is close to the 70th percentile for men. If there are the same amount of men and women are alive at age 21, then at age 83, you will have half the women left, but only 30% of the men [this is not exact, but it's close] -- that means there will be 5/3 - 1 = 66% more women than men at that age. Of the age 83 people, 63% will be women. It gets worse at higher ages.

This is going to come up again when I look at public pensions - especially when the men are married to women who are already younger than they are.

You're going to find, even without the bias for women to go to the doctor more often than men, women simply live longer to require more medical care. If you drop dead of a heart attack at age 75, that's a heck of a lot cheaper than living past age 90 in a nursing home. If we look at the stats starting at age 65 [thus ignoring all the people who died during middle age], you will see using this mortality table almost 30 percent of women will make it to age 90.

But let's go back to the mortality table I'm using: a calendar year table with no mortality improvement. That table is based on actual deaths at certain ages in calendar year 2000; so you count up how many people are age 21 at begin of year 2000, see how many make it to the end of the year alive, and put down the mortality rate [it's a little more complicated than that, but not by much]. The group of people you're looking at for mortality at 21 are different from the group you're looking at for mortality at 65.

A better table to use would be a cohort table, meaning you follow a group of people who are born in a particular year, from age to age. But there's a problem with this - you can't get a full table until they all die. And not even all the people from 1900 cohort are dead yet. But actuaries don't wait until everyone is dead; all sorts of projection techniques are used, including mortality improvement from earlier mortality tables. I don't have a cohort table loaded up right now [I'll have to get back to you on that], but the qualitative results I discuss above aren't going to be much different: women outlive men, and you'll see this especially at advanced ages where senility and general systemic failure mean there's a heavy usage of health care at this point.

When Sarah Palin said "death panels", she was on point... but what she really should point out is that women really will be hardest hit by these death panels. She has written on the recent cancer screening recommendations, but I don't see anything in there that it was inevitable that this sort of thing would happen. It's not just a matter of screening and prevention; the actual healthcare of women will be cut more than that of men under government-run health care, just because there's more to cut.

But hey, perhaps under the new regime, life expectancies and survival probabilities will be equalized between the sexes. But I don't think that's what feminists were shooting for.

More stuff on childbirth health care costs: This is from an actuarial meeting, but the speakers are a nurse and a doctor. Births of a Nation [pdf of Powerpoint slides]. I want to point out the following on page 39 - the very last page:

Though hospital and physician fees
are still the main obstetric costs, lab
tests make up a growing portion of
costs.

These enhancements to care over the
last 20 years have clearly added to
costs but have debatable health
benefits.

So expect the pregnancy and childbirth lab tests and OB visits to be the next on the chopping block from government advisory panels.

RELATED POSTS: Instapundit linking to Althouse.

Discussion at the Actuarial Outpost on mammograms.

Meep

Meep is a member of the Irish Catholic mafia, having a suspiciously high number of green-eyed, red-haired friends. While she doesn’t have red hair herself [except when she goes into the sun (rare for any vampire)], she does have green eyes. She’s a raving Papist and is a life actuary on the side [i.e., she counts dead people]. An amateur pain-in-the-ass [willing to go pro!], she likes covering retirement, mortality, math, and education issues.