Healthcare and Insurance (2)

Insurance Model
Insurance is very effective for risks with low likelihood and high impact, but it stops working if the risk changes from being low likelihood to extremely widespread. That is why insurance companies have exemptions for extreme events like war and so-called acts of God. If the city is bombed during a war, nearly every house might be burned down. In that situation, sharing the cost does not help. Paying for a thousandth of the cost of rebuilding all the houses is no cheaper than the cost of rebuilding your own house.

Pooling the risk of an event that will affect everyone similarly makes no sense. Insurance cannot deal with a widespread risk because there is no benefit in pooling the costs.

An insurance model does not work for healthcare, because the conditions necessary for pooling risk do not apply. Health claims are not rare, because almost everyone makes claims against their health insurance. If everyone is making regular claims, the insurance model is inefficient, because the bureaucratic costs of handling claims are high, and costs are not really shared. Insurance companies tend to “gold plate” treatments, because that allows them to increase their margins.

With health care, there is very little uncertainty about where the risk will strike. Most expenses will be incurred by the elderly and by people with existing health issues. The probability of a claim is skewed towards these groups.

Forced Sharing of Risk
The problem with medical insurance is that young people have a very low probability of making a claim, while those who are older or unhealthy have a higher likelihood of making a claim. If they pool the risk, younger people will end up subsidizing the old and sick. That makes no sense, so young people would usually be better off not having health insurance.

If only sick or elderly people were insured, they would not be pooling risk, because everyone would be just making claims from each other. If everyone is likely to make a claim, that is a situation of high likelihood and medium impact. This kind of risk is hard to pool, so there is not much point in having insurance. It just adds a costly administrative layer that transfers money between people. They would just be paying administrators to manage expenses they could manage themselves.

To be effective, health insurance needs young and healthy people to subsidise the old and the infirm. People were willing to do that once, but now they have woken up and are opting out. Consequently, free-market medical insurance does not work very well.

Obamacare gets around this problem by making insurance compulsory and preventing insurance companies from refusing to insure the elderly and people with pre-existing conditions. This forces the young and healthy to subsidize the old and sick. This is fine if voters accept that this is what is happening, but they usually do not.

A single-payer system has the same effect. Everyone pays tax at the same rate (more or less) and everyone gets the health care that they need. Again, this means that the young and healthy subsidize the healthcare of the old and infirm.

Single-payer is a misnomer. There is not a single-payer, as everyone pays through their income taxes. Universal payer would be a more accurate name.

The disadvantage of a single payer is that people lose control over how much health care they will receive. Governments rarely budget enough money, so there is never sufficient health services for everyone (Of course, when services are free, the demands expand significantly). Bureaucratic processes have to be put in place to ration scarce health care resources. People who are sick often struggle to persuade bureaucratic gatekeepers that they are sick enough to need care.

The situation is made worse by technology. The range of practical surgical interventions has grown immensely. People are living longer, so everyone needs their knees and hips replaced and their shoulders reconstructed. Some need these several times during their lifetime.
New drugs for cancer and other diseases are extremely expensive. A treatment can sometimes cost several hundred thousand dollars a year, for uncertain outcomes and marginal benefits. This places enormous stress on single-payer systems, because decisions about treatment often become political.

Politicians often get involved in deciding on the value of a life, sometimes on an extra three months of life. In an insurance model, the pressure of technology falls on premiums. In a single payer system, it falls on politicians.

Insurance (4) - Banks and Risk

The banks and other financial institutions that bought CDOs also misunderstood the nature of risk. They assumed that insurance against default had eliminated the risk, so they purchased these financial instruments as if they were risk free. The reality was that they had swapped one form of risk for another. The risk of many homeowners defaulting had been replaced with the risk of a large monoline insurance company failing and defaulting on its obligation.

The risk of AIG defaulting could not be estimated on the basis of previous history, because this type of history does not repeat frequently enough. The default risk for AIG was probably very, very small, but the impact would be absolutely huge. The banks and institutions who trusted in AIG did not take this risk into account. Once the risk of mortgage default became widespread, the risk of AIG defaulting changed quite dramatically, but by then it was too late for anyone make change to mitigate the risk.

The government has now come to the rescue of AIG, but that does not eliminate the risk. The risk has just been transferred to all the taxpayers of the United States. This is not pooling of the risk, but pooling of the costs of default.

Note: the AIG bailout money flowed on to the banks with debt and securities insured by AIG. According to the Wall Street Journal, these include Goldman Sachs ($6 billion), Deutsche Bank ($6 billion), Merrill Lynch, Société Générale, Calyon, Barclays, Rabobank, Danske, HSBC, Royal Bank of Scotland, Banco Santander, Morgan Stanley, Wachovia, Bank of America, and Lloyds Banking Group.

Insurance (3) - Finance and AIG

The huge AIG insurance conglomerate has just received a second bailout from the US Government. The principles outlined in my previous posts explain why they got into trouble. AIG was insuring various kinds of debt instruments (CDOs) against the risk of default. Most of the CDOs were based on residential mortgages in the United States.

Insurance made sense while the risk of mortgage default was low likelihood and high impact. In normal times, the risk of a homeowner defaulting on their mortgage is quite low, even though the loss could be quite large, on the rare situations when they did default. Pooling the risk of mortgage default due to the vagaries of life was sensible.

Everything changes in a housing boom. When credit is flowing freely and every man and his dog is flipping houses, the nature of the risk changes. The boom eventually collapses and house prices fall. The risk of default ceases to be low likelihood, as default becomes widespread. The insurance companies set their fees as, if the risk was low, but now the risk is widespread and high impact, they are unable to compensate all those who face losses.

Insurance (2) - Risk Remains

Insurance pools risk, but it does not eliminate it. When all the people in a city take fire insurance, the risk is transformed, but it is now concentrated in the insurance company. If dishonest people gain control of the insurance company, and embezzle the premiums and reserved held in trust, the company might default on its promise to rebuild houses that are burnt down. The risk of the insurance company defaulting is even lower likelihood than the risk of a fire, but the impact is even greater. The people of the city can spread this risk by dealing with several insurance companies, but the risk can never be totally eliminated.

Insurance (1) - Pooling Risk

Understanding and managing risk has become a bit of science. Risk has two parts.

1. Liklihood
2. Impact

Liklihood tells us the probablity of an event happening. Impact tells us how much harm the event will do. The two factors are combined for an assessment of risk. An event that is highly likely is not a problem, if it has no impact. A very high impact may not be a worry, if it only happens once in thousand years.

Insurance is a method of pooling risk to minimise the cost of rare events to any one person.

Insurance works well for situations where the risk has low likelihood, but high impact. For example, the likelihood of a person’s house burning down is quite low. However, the impact is enormous, if it does happen. House insurance is method for pooling the risk and the costs.

In a city with a thousand homes, only one homeowner might experience a catastrophic fire in their house in any year. If everyone who lives in the city pays into an insurance fund, each home owner will only have to pay one thousandth of the cost of rebuilding a home to give the insurance fund enough money to pay for the cost of replacing the one house that is burnt down. Homeowners have several good reasons for paying into an insurance fund.

• No one knows in advance who will have a fire. It might be me.
• The cost of a fire would be devastating for the family whose house burns.
• The cost of the insurance is relatively cheap.
• The probability of a house going on fire can be estimated by looking at the history of house fires.

Insurance deals effectively with risk like a house fire, because it is relatively rare. The insurance company is able to estimate the probability of fires occurring and calculate an appropriate level for premiums. Everyone benefits from sharing the costs of the fire, because they know that next time they could be the one facing a tragedy.

Insurance is very effective for risks with low likelihood and high impact, but it stops working if the risk changes from being low likelihood to extremely widespread. That is why insurance companies have exemptions for extreme events like war and acts of God. If the city is bombed during a war, nearly every house might be burned down. In that situation, sharing the cost does not help. Paying for a thousandth of the cost of rebuilding all the houses is no cheaper than the cost of rebuilding your own house. Pooling the risk of an event that will affect everyone the same makes no sense. Insurance cannot deal with a widespread risk, because there is no benefit in pooling the costs.