Doing good by doing better – Fraud and defect management as a way of life.

The notion of equal treatment for all is an egalitarian ethic that is ingrained into American culture.   And in most things it works fairly well. In most of American business everyone gets charged the same price, for example in retailing,  a high end boutique and a dollar store have different product mixes and serve different customers but they charge essentially the same price for the same product to everyone whether it be a ten pack of kid’s tube socks or the latest Givenchy original. As the song goes: “if you’ve got the money, they’ve got the time” for you. Entities that undertake risk like Insurance Companies don’t have this luxury:  they need to price their policies to reflect the risk of the policy holder which varies substantially.  Nevertheless, mutual insurance companies do their best to treat all members of the group the same. And it’s certainly true that similar customers who live in the middle of an inner city slum whether they be fabulously rich or dirt poor gets the same  rate. And likewise all the low income residents of Beverly Hills get the same demographic and property adjusted deal as their Movie Star neighbors.

But wait: rich people don’t tend to live in poor, high risk neighborhoods. This means the poor and the rich neighborhood’s loss experience are very different so their rates in fact vary widely. People living in poorer, rougher neighborhoods pay more for their insurance because the losses that premiums need to cover are much higher. This state of affairs exists not because mutual and other insurers are bad or discriminatory but because anyone with property in those areas tends to be exposed to higher risk. But there are three kinds of risk:  individual specific risk, general loss risk rooted in their location and fraud and error driven risk. And each form of risk has a different characteristic.  Individual risk reflects the attributes of the individual, including their age, gender, and driving experience.  General loss risk reflects the risks that are associated with the general environment that the property is established in.  There there is the risk of Fraud and Errors which historically we haven’t been able to directly measure so it gets bundled in with the overall loss experience.  But fraud and errors are a big deal. In the Auto Insurance market fraud alone is estimated at 15 to 20 percent of premiums paid. And fraud and error risk isn’t like general risk because there are many very honest and organized poor people who are being charged higher premiums because they live in an area or belong to a demographic where up front fraud and errors are rife.

Historically fraud and errors have not been identifiable up front which meant that they were simply reflected in the overall loss experience for the neighborhood or segment.  This is unfortunate because up front fraud and errors vary from neighborhood to neighborhood with affluent neighborhoods likely experiencing quite a bit less than average and lower income, more disorganized communities experiencing far more. For example it may be that in places where everyone upper middle class the losses due to fraud and errors could be quite a bit lower – say 10 percent of premium. While in poorer, more disorganized neighborhoods the cost of fraud and errors could add 30 percent to the price of a premium. Thus in my example, the honest in the poor neighborhood are carrying a much greater burden than the more affluent.

are not like general risk to property:  they are specific choices and capabilities possessed by individual risks.  A person may be working class and live in a rough neighborhood and still be very honest and organized.

Yet historically the honest have had to pay for the dishonest’s fraud and the disorganized’s errors. And crucially, the cost of fraud and errors in the auto insurance market are likely to be much higher in lower income areas. This isn’t necessarily because the poor are more or less honest it simply reflects the fact that cost of a homeowner or auto policy is a far greater share of their income than to the more affluent. And their education,  command of English, and understanding of bureaucratic institutions like insurance is often substantially less.  So the likelihood that they’ll be tempted by their circumstances to shade the truth or that they’ll make mistakes of fact or simple errors on their applications is probably much higher. This makes fraud and error incidence very ‘lumpy’ across an entire book of business:  high in places where people are living closer to the economic ‘edge’ with less organization and help and lower in places that are more advantaged. 

 For a committed mutual insurer this creates a moral conundrum:  if you recognize the issue and subsidize lower income working member’s premiums you run the risk of pricing yourself out of the market for higher income members.  On the other hand if you price based upon loss experience, you are charging a lot of otherwise honest lower income members a premium not because they’re dishonest but because of who they live near.  It’s a real head scratcher.

But it doesn’t need to be.  I work for a company called VeracityID:  our focus is on helping financial and other risk bearing institutions identify and screen out fraud and errors up front during the application process.  We can demonstrate to Mutual Insurers that by using our solution they can screen out deceptive and mistaken application information and by doing so enable them to recognize and reward the honest people in lower income neighborhoods and help the disorganized get to the right information and product:  simply by evaluating their applications up front using our idFusion fraud identification and management platform.  By doing so Mutual insurers will be able to serve more of the working class and working poor at better rates and accept

able loss ratios.

And since one of the key purposes of Mutuality has always been to expand access in their target population without regard to their circumstances, VeracityID and idFusion can help them keep their promises. Helping insurers do well by doing good – just one more way that VeracityID analytical solutions deliver.

A bedrock American value is that all Americans should be treated the same regardless of wealth, education or upbringing. And that works in a lot of business settings:  a high end boutique doesn’t have the same clientele as a dollar store but they certainly will sell you the same products at the same price at both regardless of your demographic details.  This doesn’t work in financial services, including insurance for the obvious reason that when an institution makes a loan, issues a credit card or binds an insurance policy they are undertaking risk and that risk is specific to a customer.  Everybody gets their own price based upon their personal characteristics and financial reputation as well as based upon where they live. They can’t be treated ‘the same’.

The goal in pricing products is to comprehend the specific risk profile of each customer so that the appropriate price can be charged to cover the expected value of the risk and leave enough left over for operations and to reward the contributors of capital. In practice risk is broken into two buckets:  customer specific risk:  they’re FICO score, their demographics, their earning power, their ability to pay and broader ‘community’ risk:  what is the risk profile of their community? Companies must combine both customer specific risk and overall community risk because not all losses can be attributable to specific individual attributes.  That is to say that all the incurred losses
include factors that can’t be identified and tied to specific individuals therefore those extra, undifferentiated losses need to be apportioned to everyone in a category or community’s risk. And the reason for this:  fraud and errors.  In auto insurance fraud constitutes an estimated 15 to 20 percent of premiums.  But certain segments or communities experience much higher level of loss that likely include far more fraud, while other segments experience much less.  So there must be parts of an insurer’s book that experience very little fraud and some that experience large amounts say 30 or even 40 percent of premium.

It stands to reason that the places with the highest fraud burden are from lower income communities with low socioeconomic status, sometimes limited facility in Englih and fragmented family structures.  These people live closer to the edge so the cost of a financial product or policy is necessarily far more burdensome to them than to the affluent.  This is particularly true of automobile insurance because it’s a requirement to have mobility in our society.

But how does that play out in insurance where everyone has a different risk
This is particularly true for Mutual insurers whose mission is to ensure equal access
In one respect we treat everyone the same but in the real world this doesn’t work.
This is because rating is based upon total loss experience which includes both legitimate losses and losses deriving from fraud and errors.
It is reasonable to believe that up front underwriting fraud and errors are greater in lower income, lower education, less english speaking areas than in affluent areas.
This doesn’t mean that they’re less honest, it just means they face a much more acute tradeoff with this product.
The result is that fraud isn’t 15 to 20 percent everywhere it’s 10 percent of premium in some places and 30 in others.
And the problem in all these places is the current rating methods don’t differentiate between customers who are telling the truth and not imposing a heavy fraud cost and those who aren’t.
So assume that one third of today’s fraud can be identified and eliminated – in a rich neighborhood that might constitute 3% of premium.  In a poor neighborhood that could exceed 10% of a substantially higher premium.  The cost of not addressing addressable fraud in a poor place could easily be five, six or seven times that in a rich place. Cost differentials that are so significant that they result in denial of access to insurance or to sufficient comprehensive insurance to many very honest, trustworthy customers.
And so long as we couldn’t identify fraud this had to be acceptble.
But we can identify and stop quite a bit of it now.
And if by focusing on fraud an errors you can expand access to your product by marginalized groups with the same or higher profitability, shouldn’t both enterprise profitability and social conscience dictate that you do so?
Particularly if you’re a Mutual Insurer.