> I have no doubt that insuring banks benefits long term growth. So the FDIC is > helpful. It would be interesting if a Berkshire Hathaway were to start insuring > some bank deposits, say, amounts over $100K, to see whether we could get > at least limited competition in the bank insurance market.
I presume you know some of the fundamentals of both the banking and the insurance industry. In banking, the reserves need only be 10% of M1 see http://www.theshortrun.com/data/Financial/aggregates/msexplain.html for good descriptions of M1, M2 & M3). > As for the Fed, they can be helpful in a severe liquidity crunch, such as during > bank runs, but the Fed is more intrusive than it needs to be. I'd like to see > the Fed keep quiet except in the case of banking panics. I have not seen anything > to convince me that it is anything more than an grand illusion that the Fed > actually has significant control over the economy (outside of a liquidity > crisis) by varying short term interest rates. The Fed is a bit like the rooster that > thinks his crowing raises the sun each day. Hmm, as you also mention below, you are rather skeptical about correlations. What do you base that on? Let me bring up something you said earlier > Sounds like the rooster again, I think. The problem with this sort of analysis > is that there is a lot of randomness in economic growth figures. I expect you can > probably find a lot of variables that correlate with the data you mention (have you seen > the graph of global warming versus number of pirates?). You would have to do a huge > statistical study to convince me that FDR actually caused the economy to change > that drastically that quickly. Well, I should probably give you a bit of my background here. I've been working in physics for the last 30 years or so (finished my Phd Dissertation 1-82 in particle physics) and am fairly familiar with statistical analysis. I've been dealing with false and true correlations for all of that time. There are rigorous ways to look at correlations and to ask questions precisely that decrease the chance at false correlations. For example, there are a lot of things that went down as global warming increased, as well as a lot that didn't. You have a very simple model (up/down) for global warming and piracy and the chances of false correlations are good. Indeed, Monte Carlo techniques lend themselves very well to this type of analysis. Another example of false correlations is the correlation of cancer with high voltage power lines. The trick involved is to look for a 3 sigma relationship in any type of cancer. There are so many types of cancer, that one of them is bound to have a 3 sigma signal. Again, Monte Carlo techniques are invaluable in doing rigorous, straightforward, efficient modeling of these types of questions. I've been involved in similar debates to this in the past on Brin-L, and have done Monte Carlo analysis. With today's computer speeds, it is straightforward to run 1 million cases in a few seconds (maybe a few minutes for complicated analysis). Further, Monte Carlo's allow/require one to explicitly instead of implicitly state assumptions. And, furthering the value of Monte Carlos, the analytical models employed by the major houses, Freddy and Fanny did not accurately predict the recent meltdown/Black Swan, but Taleb (who tends to rely more on Monte Carlos and less on analytical approximations) did. But, before I apply the toolbox of techniques I've learned such techniques to analyzing your and my hypotheses, I'd like to know if it would mean anything to you. Are you convincible by numbers? I'm cutting the rest of your response because I think that a rigorous reply would only be meaningful if you accept what I've been taught as rigor as valid. BTW, I do realize that economics is no more than a dismal science, but that doesn't mean that techniques of science cannot be borrowed to look at some of the historical information. - Dan M. _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
