Bryan Caplan  

Insurance, Reputation, and Caricature

PRINT
Insurance, Reputation, and Kri... Various Links...
Here's a view Krugman attributes to me:
Of course, there's also an alternative universe in which insurance companies would never, never treat their clients badly, because that would hurt their reputations.
Here's what I explicitly said in my original piece on insurance and reputation:
Am I saying that health insurance companies never play dirty tricks on their customers?  Of course not.  It's a big world, lots of bad stuff happens.  What I'm saying, rather, is that reputation works well even in industries where firms have big, lumpy liabilities.
If the .5% rescission number that Krugman cites were true and correctly interpreted as fraudulent-in-spirit, I'd admit that reputation in the health insurance industry isn't working too well.  But I suspect that the real story is usually that (a) insurers are well within the letter and spirit of their contracts, and (b) are fighting genuine customer abuse. 

In any case, no matter how you interpret the .5% number, there's good reason to be skeptical of it.  A recent non-Krugman editorial in the NYT suggests a figure closer to what I'd expect:
But Congressional investigators found that three big insurers canceled about 20,000 individual policies over a five-year period -- allowing them to avoid paying more than $300 million in medical claims.
In 2008, U.S. health spending was about $2.4 trillion; 2004-2008 spending would be roughly $10 trillion.  So even if these three big insurers held just 1% of the market each and their behavior were typical, we're talking we're about .1% of total spending.  If companies were really targeting big claims, the fraction of affected individuals would be much smaller. 
 
Question: If Krugman really wanted to fight dishonest health insurance companies, wouldn't it be more productive to produce a table of companies' rescission rates, so customers would know where take their business?  Can anyone find a link to such a table?

* Note: Even Krugman hedges his bet with an "if that's really true."


Comments and Sharing





COMMENTS (26 to date)
Marcus writes:

"If Krugman really wanted to fight dishonest health insurance companies, wouldn't it be more productive to produce a table of companies' rescission rates, so customers would know where take their business?" -- Bryan Caplan

Precisely.

This morning I was reading a NYT article which claimed that most people who get laid off never get back to their prior income levels. I wondered the same thing as you, shouldn't the article look into what successful people do and inform people of it? Instead, it was a pointless article rooted in fear.

The same tactic Krugman uses.

Skyler Collins writes:

A thought I had was how often do complaints about a health insurer get through a policy holder's employer to the insurance company? I've never had a complaint with mine, but if I did, it's the only option my employer provides. (So how many find it worthwhile to complain?) Of course there's private insurance, but since canceling insurance with my employer is not likely to increase my take-home pay, it's the cheaper option for me.

Dan Weber writes:

I think you underestimate the fear people have of being denied by a profit-seeking enterprise.

Still, I look forward to the public option, because it will deny people stuff they think they should have gotten, and then we will have some fun cognitive dissonance to work through.

NadavT writes:

If, as you suspect, insurers are fighting genuine customer abuse, rather than figuring out new ways to deceive their customers, why would they wait until a claim is submitted to search for errors in people's applications? If genuine customer abuse were the real problem, one would expect that they would develop screening indicators of what kinds of people were likely to abuse the system and target their investigations on those people. Instead, what they actually do is wait until a customer files an expensive claim and only then look for some excuse to deny coverage. That way, the insurance company collects premiums for years and then never has to provide services to the former customer. Is that really the behavior you would expect to see from an entity concerned about its reputation?

Dan Weber writes:

NadavT, I think life insurance has laws in which they must pursue fraudulent applications within 12 or 24 months, or else everything is assumed to be true. Maybe health insurance needs this.

... Come to think of it, I can't imagine any insurance company advertising itself as "we don't cut people off."

I can guarantee you that this is a very real concern among insurance shoppers; yet none of them seem to market themselves this way.

I'd love to see the numbers on rescission, whether by this blog or by Krugman.

Rich writes:

I have seen plenty of opinions on the issue of recission, but no hard data on the actual incidence of it. State insurance regulators may collect it.

Many policies provide that two years after a contract is issued, no false statements that might have been in the insured's application can be used to void the contract or deny claims.

Marite writes:

I advise you to read a very good post reccomended by Angry Bear about health insurance recisions and Monte Hall:

http://tauntermedia.com/2009/07/28/unconscionable-math/

Robert writes:

The .5% figure is from the CEO of Wellpoint, America's largest insurance company.

Also:

"WellPoint's Blue Cross of California subsidiary and two other insurers saved more than $300 million in medical claims by canceling more than 20,000 sick policyholders over a five-year period, the House committee said."

So 3 insurers saved $300 million, not the insurance industry as a whole.

It's really surprising that you would defend insurance companies. When 1% of the insured account for 22% of all costs, rescission makes a great deal of sense. Screw over a small number of people and it doesn't matter -- they can't tell everyone.

Dan Weber writes:

Marite, that post was blood-chilling to read.

As a customer, I wouldn't mind rescission as much if it was explicit. That is, they told me that they would cut me off after $X million, instead of trying to see how they can get out of paying their expensive claims after they come up.

Arnold Kling said one of the best things about insurance companies here a year or two ago: "I trust government very little, but I trust insurance companies even less."

Chris writes:

NadavT, insurance companies don't contract with individuals, it is mostly with employers so there isn't a way to screen up front.

Robert, Bryan did the full extrapolation. IF the top 3 only had 1% of the total market and $300 million was typical for healthcare as a whole then the total rate would be under .1%. He wasn't assuming that was the total picture.

Rich writes:

Unfortunately, the tauntermedia analysis seems wholly predicated on an ill-defined "less than 0.5%" comment by some insurance company executive, not on actual recission data. It would be helpful if someone could offer actual data.

Troy Camplin writes:

What do Medicare and Medicaid refuse to pay? What percentage?

Dan Weber writes:

NadavT, insurance companies don't contract with individuals, it is mostly with employers so there isn't a way to screen up front.

In the employer-provided group insurance model, there isn't any fraud. You state your age and gender and that's about it. Maybe in rare cases you make a mistake with your SSN.

In the individual market, there are detailed applications. And they suck. People who are interested in saving the private insurance market ought to campaign for a uniform application, with a set series of rules on how prior conditions are treated, and just what kind of "errors" on the application can cause coverage to be yanked.

matt writes:

I think someone has there math wrong.

Assumption: Company X dropped .5% of there policy holders.
Assumption: Company X cannot drop people for lying on their application if they get their insurance through their work. (law says they can't ask your medical history for group policies.)
Assumption: 90% of company Xs sales are through the group market
Result 5% of policy holders are subject to rescission.
That's a huge number.

Actually, if reputation-based mechanisms are working at all, it should be extremely easy to find the data Bryan wants. Insurance companies with relatively low rescission rates ought to be loudly trumpting the fact that unlike our competitors, we don't toss our sickest policies to the wolves.

Gary writes:

In the event of fraud, rescission and claim denial are a result of the insurance fund administrators looking out for the honest portion of their pool. Your premiums go up when your insurance company pays claims it didn't agree to in its contracts. Nobody wants to be part of an insurance pool that gets taken advantage of by fraudsters.

Ari Timonen writes:

Funerals will always be the most cost-effective treatment and insurance companies have every incencitive to deny claims. Besides its always better to risk a court case with army of lawyers than pay for an expensive treatment. Many people won't need expensive treatment, and those who do are in margin, and those who get denied are in a smaller margin. You just hope you don't get that bullet.

Even if a person fooled by insurance company wants to tell about his experience, it will get lost in the noise of information. And as long as insurance companies can make enough satisfied customers (with those 50% who only need little care) there will be enough resistance for change. And human memory is limited.

Ultimately I think health care is a distribution problem. If total free markets were allowed, insurance companies wouldn't ever cover people with high risk of sickness, or with extremely high premiums. Basically if you were 90% likely to have a certain kind of illness (prediction could be possible with gene tecnology), you would be paying 90% out of pocket of the costs of any medical operation. Given that demand curve is highly inelastic compared to other goods, the prices will always be high.

Some families just can't afford to pay such sums because they don't have the skillset to make the ends meet. Sadly, I read from Mankiw that only 28% of american people support taxing so everyone can have health care. I guess you have kind of every man or woman for him or herself attitude there. Luckily we don't have it here. We have a bit of social responsibility. I really feel sorry for the 40+ million uninsured people there.

The Cupboard Is Bare writes:

Bryan,

Krugman is no dummy. He twisted your words by taking the quote out of context, and he knows it.

He also knows that rather than read your blog, liberals will take what is the equivalent of a soundbite and run with it.

Anyone who is worth his salt, however, will know exactly what you meant.

:)

Dan Weber writes:

In the event of fraud

Have you ever filled out an individual insurance application?

They ask things like "please tell us about every visit to the doctor in the past five years." Did you forget about your visit to the dermatologist for acne? Then you committed fraud.

It is directly in the insurance companies' interest to make this stuff hard to fill out accurately. I looked around a bit when shopping and they all seemed equally onerous. There might be some company that checks things up-front and takes responsibility to make sure things are accurate, but wouldn't such a company market itself that way and be easy to find?

Gary writes:

They ask things like "please tell us about every visit to the doctor in the past five years." Did you forget about your visit to the dermatologist for acne? Then you committed fraud.

Perhaps you shouldn't rely on your memory when entering into an obligation as serious as a medical insurance contract. Would it be so hard to contact your GP and get a record of your doctor visits? If the insurance company can investigate your medical history, you can too!

Gary writes:

Just thinking this through from an insurance company's perspective, how hard would it be to attract honest people to your risk pool if you had a policy like this:

"Oh, sure, just fill out the forms based on your best recollection. If you forget some doctor visit or malady and end up paying lower premiums as a result, no worries. I'm sure the others in the pool won't mind making up for what you didn't pay as a result of your omissions. Plus, it's not a problem for us that we might not get a complete picture of your health risks, because we never even think about trying to balance our portfolio of insureds to prevent too many of one kind of payout falling due at the same time. Finally, we wouldn't dream of optimizing the investments from our policy holders' premiums according to the schedule of expected payoffs we'll have to make based on who we've insured, even though that might help us offer competitive prices for our policies. We're cool like that."

Seriously, when did people get the idea that they have no responsibilities when entering into contracts with insurance companies?

Dan Weber writes:

Would it be so hard to contact your GP and get a record of your doctor visits?

I haven't seen a GP in five years (although I keep on meaning to). Even if I had one, the GP wouldn't necessarily have records of all of my outside contacts. I could just see a podiatrist or a psychiatrist on my own. Hopefully electronic medical records would help with this without totally destroying privacy.

(And having filled out life insurance applications a few times, I made a text file of all the health visits of me and my family, but I only update it when I need it, so I might forget to include, say, a visit to the Minute Clinic for an ear infection.)

If the insurance company can investigate your medical history, you can too!

Then they should, at the time of application. As Taunter pointed out, it is in their interests to not investigate.

I had to pee in a cup for my life insurance application. Life insurance companies apparently care enough to check.

I think it's okay to refuse to cover a problem if I left off things that are directly related to the problem. If I forget to mention my irregular heartbeat and I later have a heart attack, then I think it would be fine to say that I wasn't covered. That's a pretty significant thing.

But if I forget to mention my visit for acne, my heart attack should still be covered.

Bob Calder writes:

Dan Weber's remarks are the only ones I can agree with here. Most of the commenters really don't understand how the insurance business makes money or what it does.

An insurance company really doesn't care about its reputation. Reputations have VERY little effect on the bottom line unless we're talking about complete piracy. The fact is that all of you would take an AIG car insurance policy tomorrow if it were offered at an attractive price.

As to recission, nobody knows what the trending numbers are in recission from company to company so the discussion is useless. Recission is basically the discovery of fraud on the part of the customer by the insuror. The problem is that a weasel can find a lie anywhere because as Weber said, the app. asks for things that are in the mists of personal history, or vaguely worded. Sometimes deliberately so. Recision is caused more by incompetent salesmen than any other thing. Why? Because they have little education and no medical experience in most cases. They want to feed their families, so there is little incentive to accept a truthful application.

That brings up the erosion of trust. This is a HUGE issue for me. I have seen the insurance business undergo a change from a small network of trusted agents to a large network of untrusted agents. Although this has meant increasing revenues overall, the subsequent increase in fraud and questionable acts has been substantial.

Probably the best place to track this is in New York. That's because almost every other state lets Insurance problems pass. Indeed, other states actively encourage things that are bad for the citizens in general due to lax and incompetent oversight.

OMG I forgot to say: Russ, I love the podcast!

Gary writes:

Then they should [investigate], at the time of application. As Taunter pointed out, it is in their interests to not investigate.

How do you get "should" from "can"? Do you apply that principle to most areas of life?

Also, if you know it's not in an insurance company's interest to investigate, but it is in your interest to investigate (because the terms of the contract usually say you will provide an honest and thorough account of your medical history, and you agree to those terms), then why would you be so resistant to investigating yourself? Hire a PI if you have to.

If you were buying a house, would you be willing to hire an inspector to check every house at which you looked for rot and bad piping between the walls and under the floors? Or would you expect the sellers to hire inspectors and show you the results? Or would you just say, "just tell me about the floors and walls based on memory - don't bother looking up the inspection history."?

Finally, you didn't respond to my points about how omitting some "irrelevant" information harms the others in the risk pool. If you and I were in a risk pool together and you needed help for a heart attack but you omitted information about your acne visits in your application, I would be pretty upset! I'm glad my insurance company protects me by not putting up with people "forgetting" minor and supposedly irrelevant details.

Dan Weber writes:

These days health insurance companies are very unpopular. The media loves stories about people losing their coverage, whether legitimately or not. There's a big desire among the executives to Keep Your Head Down.

So it's really surprising to see someone suggest "Hire a PI if you have to" when dealing with an insurance company application. Is this answer designed to make people as angry as possible with insurance companies? If I were determined to get the public option to pass, I would go out on the Internet, pretend to be a health care lawyer and laugh it up at people who lose their insurance with a "you should have known better."

Or would you just say, "just tell me about the floors and walls based on memory - don't bother looking up the inspection history."?
You mock this approach, but this is exactly what the health insurance companies do. They have the chance to inspect up-front, but they don't. Instead, they buy the house, and if they don't like it 18 months later they go back through the paperwork to find some way to weasel out.

Finally, you didn't respond to my points about how omitting some "irrelevant" information harms the others in the risk pool.
I did. If I find out that other guy in my risk pool forgot to declare his one visit for acne and gets his heart attack treated, I will be happy, because it means I am dealing with a reasonable party instead of a bunch of thugs.

He also hasn't driven up my costs. If one in a thousand people who had a single dermatology visit for acne end up having a $50,000 acne attack, and there are a thousand such people in my pool who didn't list their single acne treatment, that means there will be one person with a $50,000 claim. But I said I'm fine with them denying coverage for things directly related to any errors or omissions, and this would count, so the pool doesn't pay out anything, and so I am just at the same place where I started.

A business model based on "we hope the other party makes a mistake" is corrupt and deserves to die. I can only hope, for the sake of the industry, that you have been hired to portray the worst caricature of the industry that the radical left could possibly imagine.

Gary writes:

I can only hope, for the sake of the industry, that you have been hired to portray the worst caricature of the industry that the radical left could possibly imagine.

Sadly, I'm not being paid.

You mock this approach, but this is exactly what the health insurance companies do.

Are you sure? Insurers usually advise you to divulge your complete medical history (in the metaphor, they are house buyers who ask whether there are problems with the house) - and they warn of the consequences if you do not. Is hiring someone to help you provide exactly what you are agreeing to in the contract so onerous? I have nothing against insurance companies offering in the contract to take responsibility for the investigation, but you shouldn't expect that they will if both of you agree to a contract where it is stipulated that you are taking responsibility for providing a complete medical history. I'm utterly baffled that you think you should get extra services that aren't stipulated in the contract you agree to.

He also hasn't driven up my costs. If one in a thousand people who had a single dermatology visit for acne end up having a $50,000 acne attack, and there are a thousand such people in my pool who didn't list their single acne treatment...

That last assumption is a doozy, but let's go with it. The less information the insurer has about its risk pool -- even if everyone forgets the same thing, as you've assumed -- the harder it is to understand the risks it faces. As a result, everyone's premiums would be higher than they would be otherwise. It gets worse if different people omit different things. Everyone's premiums will be higher than they need to be, and those who are thorough will pay the highest premiums, subsidizing the dishonest or forgetful ones.

But I said I'm fine with them denying coverage for things directly related to any errors or omissions...

Why would it be up to you to decide what is or isn't directly related? Perhaps those with acne are more likely to have other health problems, and the insurer would have charged higher premiums if it had known about it. Since those higher premiums weren't collected, the rest of the risk pool winds up paying for them. I have no way of knowing if that's true for acne, but it could be true of other things that you might not easily recognize as being directly related.

Comments for this entry have been closed
Return to top