Reinsurance and Modeling

Right now, I'm currently typing between classes. Finished with Dr. N's Probability I class and Dr. S's class.

As for Dr. N., I get the suspicion that a lot of people are floundering in that class. Unless he talked about variance a lot on Monday (the day I wasn't in class), a lot of people are going to be completely lost if he gives a variance question.

However, he did give an interesting lecture on how a lot of elementary probability problems can be solved by multiple methods: it almost implies that there's a way to do each problem by either a) combinations, b) permutations, or c) the multiplication principle. One of the few lectures where I felt I learned something new.

Dr. S's class was interesting. First, a quiz on how to transform random numbers (or "random variables of U(0,1) distribution") into random variables of other types, say, X=Binomial(3,0.5) or Z=N(0,1). I felt like I nailed the quiz.

Second: the beginnings of hedging and mathematical modeling. FINALLY!! I've been waiting for this part of the course like a starving man waits for a steak dinner.

But third was very interesting: Dr. S. and I chatted for about ten minutes or so. I had a question about bird flu, believe it or not. It seems that SOA is going to call for research regarding a bird flu pandemic. Dr. S. talked in class about how for catastrophic events -- say, terrorist attack, lightning strike, and other "acts of God" -- insurers build clauses in insurance contracts, stating basically "this is not covered".

I asked him about how insurance companies typically handle things like bird flu.
His answer: they don't, for the most part. Premiums go up to cover the costs of the catastrophe in the following year and take decades to come back down to something approaching expected loss. We then discussed things like reinsurance, where insurance companies purchase insurance (!!) to move the risk of such catastrophes off onto someone else. We also talked about the mortgage and bond markets, and I listened with particular interest to Dr. S.'s comments about mortgage insurance ("there should be no way they make money, but they do...the numbers don't add up") and bond markets.

There was also a bit of a discussion about the role of the free market versus the role of the state insurance commissioner. For example, they say that in New Jersey you can't get auto theft insurance -- insurance fraud is so prevalent that a true premium for auto theft would be sky high. But in cases like that, state insurance commissions sign a law saying "no insurance premium can be above x dollars", obstensibly in order to keep the little guy from getting screwed.

What happens then is that an insurer sets up a LLC -- a limited liability corporation. If Joe's Insurance is forbidden to charge above x for auto theft, then Joe's Insurance sets up a "Joe's Insurance of New Jersey" as a LLC. Now, JINJ can only charge x dollars, and by the theory of pricing by expected loss, the probability of JINJ going bankrupt approaches 1.0. When that happens, the federal government steps in and covers the cash. So we all end up paying for the auto theft in NJ sooner or later, and Joe's Insurance isn't stuck with a ton of liability.

Very interesting stuff. This "reinsurance" business intrigues me. I also got to hear Dr. S. do his Arnold Schwarzenegger impersonation, which was definitely worth the trouble of coming to class today.

BTW, our good friend Ben didn't show up for class today. We'll see if he shows up for Intro Stats.