As a practicing software developer, I regret that I must own up to the role software has played in facilitating concentration of risk in our society. By “concentration of risk,” I mean that people (or classes of people) who suffer misfortune are often required to pay more for services, which increases their vulnerability.

Paradoxically, this situation arises due to the desire of those that have wealth to minimize risk and maximize return from passive investments such as lending and insurance. That supports a class of investment and financial advisers who seek to segregate populations into “high-risk” and “low-risk” communities. In the health care industry, that “high-risk” obviously includes someone with a pre-existing condition. “Low risk” would be someone that exercises regularly and moderately and does not smoke.

The advent of computerized information gathering and processing means that identifying and marketing to “low-risk” populations is possible today in ways that were not possible before. Now that might seem to be a good thing – we obviously want to reward responsible behavior such as moderate exercise, and discourage irresponsible behavior such as smoking. Charging people more money is one way of sending those signals.

The difficulty comes when behaviors previously thought to be acceptable are discovered through statistical analysis to be correlated with high cost. Smoking is the obvious example. Many doctors smoked prior to the publication of the cancer studies that resulted in the warning labels on cigarettes. It was a rational choice: smoking helped them to manage stress, and by restricting blood flow in the extremities, helped them to think more clearly.

Even more difficult is when we actually have no control over our risk. Let’s say that we learn that our genes themselves are risk indicators. What are we supposed to do about that? Go back and tell our parents not to have sex?

But this is why it’s called “insurance.” Life is full of circumstances beyond our control – just think of the victim disabled in a head-on collision with a drunk driver.

So far, though, we’ve been considering situations that involve meaningful learning. That’s a desireable application of statistical analysis. But that’s not the kind of analysis that created the two great financial disasters of the modern era: America’s grossly inefficient health care market and the mortgage industry meltdowns. Both of these were driven by risk analysis unrelated to personal conduct.

In the health insurance market, the problem began with the formation of companies that sought to isolate and insure only those that were healthy. They offered tempting premiums to those in traditional full-service health plans, which caused many of them to switch carriers. Unfortunately, this meant that the traditional plans were starved of the premiums that financed care for sick people. To stay in business, the traditional health plans raised premiums, which eventually began to force the sickest people (often disadvantaged as income earners) out of the plan.

Unable to afford insurance due to their pre-existing condition, the chronically ill either went without care or applied for coverage that did not include their preexisting condition. Discovering this trend, the low-cost insurers hired claims agents to vet insurance applications. Then the real catch-22 came in: when the insured became sick with another illness, they were denied coverage because they did not report their pre-existing condition. They paid for insurance, and were denied coverage. Eventually, the profitability of this practice became such that profit-conscious insurers would routinely deny coverage for expensive treatments, forcing patients into lengthy and obscure claims adjustment procedures that they lacked the understanding to navigate.

Let’s be certain that we understand clearly: people who enrolled when not sick and led normal lives became ill, and were denied the benefits of their life-long participation in health insurance because people not so misfortunate were poached away by insurers that offered them lower premiums. Some among those insurers chose to maximize their profits by using complex statements of coverage and simple intimidation to avoid paying expensive claims. In conclusion: the application of sophisticated data analysis techniques distorted the health care coverage system by increasing the number of insurers, and therefore the total cost of its administration, while isolating the sick and poor from health care.

In the mortgage industry, the process was more subtle, and more directly reflected the divorce between financial management and service provision. Historically, banks made money on mortgage interest payments. They provided the money for the home purchase, and carried the risk of default. As the housing market became less and less stable, the large money market banks sought methods to distribute this risk. Sensible enough. They created “mortgage-backed securities”: essentially stocks that pooled mortgages, allowing investors to buy mortgages in bulk without having to administer loans. Particularly for overseas investors, American interest rates represented an attractive premium over those available in their relatively impoverished markets.

There were two twists in the implementation of the program. I’ll focus on the first, because mortgage security risk pools is too arcane for casual discussion.

First, how were banks to make money for placing the loans? They were giving up the long-term revenue of interest payments. There was another source of profit in the mortgage process, however: the closing costs paid on the transaction itself. This was baked into the system however, and so not particularly easy to increase.

So another strategy was chosen: the adjustable-rate mortgage, or ARM. This was structured to enable underqualified buyers to get into a home with low fixed interest rates, with a switch to much higher floating rates after five years. While many home buyers may have thought that improved earnings would allow them to manage the higher payments at five years, downward pressure on wages actually meant that most of them were forced to refinance their mortgage at five years with another ARM. Now this might seem unfair to the mortgage holder, who was losing out on the high interests rates expected after five years. But the holder didn’t have access to the customer – the banks did. And the banks profited because the refinancing allowed them to collect closing costs again.

Eventually, this system went completely out of control. In the aftermath of the 2008 financial meltdown, it was discovered that many of the largest and most aggressive mortgage aggregators (such as CountryWide) routinely falsified loan applications to make the loans appear less risky than they were. Effectively, they were defrauding those buying the loans as securities, and those (such as Freddie Mac and Fannie Mae) that insured them.

Of course, when the system collapsed, it was the homeowners that were hurt the most. A mortgage default is an incredibly abusive process: the homeowner loses all of their equity. Let’s be specific: if you’ve paid off 80% of your mortgage and fall behind on payments on the remainder, ownership of the property is returned to the mortgage holder in full.

For this reason, many states have laws that protect homeowners in the event of default on their primary mortgage. Home-equity loans, however, violate that protection, as does (you guessed it) refinancing.

How did information technology contribute to this mess? By enabling the creation and marketing of mortgage-backed securities.

But my point here is that in both situations, it was the desire to avoid risk and maximize profit that created dysfunctional systems focused solely on profit creation to the detriment of those actually paying for the service – either the patient or the home owner. These are average members of the public who of necessity must trust the expertise of those providing the service, just as the insurance agent or realtor must trust the plumber that comes to unclog their toilet.

Prior to the modern era, one of the fastest ways to wealth was to sell “death insurance” to the poor. This was often a fraud, with the “insurer” skipping town when people began to die. To limit this public nuisance, regulations were established. In the ’90s, however, information technology drove evolution in these industries that did a complete end-run around the regulatory restrictions. It behooves the public to be conscious of that, and to hold their representatives in government responsible for any failure to anticipate and moderate the excesses the ensued.