5-31-2013 Close Comments

[Mike]

The market looks like it is setting up to roll over. It has looked this way before and the bears didn't take it down. The Log-Periodic Power Law analysis from Didier Sornette however pointed to last week as the top and those prior times as just pull backs so possibly this is not a false alarm. Jerry plans to do some investigation over the weekend to analyze the closed end bond funds which are most sensitive to flagging liquidity problems. A preliminary look using a two year old group of CEF bond funds does point to liquidity issues building up. He will have to download a new set of CEF bond issues. Tom McClellan is saying the same thing about liquidity. If we are having liquidity problems it could be coming from margin calls in Japan.

Today looks like distribution to me on all indices. I usually wait until I see final volume figures so this call may be premature. If we have distribution on the NASDAQ it will produce an S13 distribution cluster sell rule. These come when 4 out of the last 8 trading days has distribution or stalling. Clusters of distribution are rare in the middle of a rally and are usually found in market topping structures. Yesterday was a stall day on the S&P500 producing an S13 on that index.
So there you have my rationale for a possible major top in process.

[Mike]

I was asked in a private mail about the Log-Periodic Power Law analysis. This is analysis that I have done and previously reported on. It follows the work of Didier Sornette in his book Why Markets Crash. This work is particularly difficult to access because the book is written at a very high mathematical level however the conceptual framework can be understood by anyone. Essentially the market goes about its random price fluctuations until the market self organizes through the herding nature of man when we all start to move in the same direction. When this happens markets become unstable. Recently we traders have all been marching upward on the long side producing a super exponential growth. This is unstable behavior and all good things will come to an end. Didier Sornette found a signature in market prices that precedes market crashes that he calls a log periodic power law.

Below I update the excel chart of the NASDAQ. The arrow points to a place where a singularity occurs in a LPPL equation described in the book. I have reversed a term in the LPPL right at the singularity to allow following the market on the down side, so you don't see the effect of the blow up in the red line here. Essentially when markets move up super exponentially they become unstable with a predictable instability point. The market tends to enter a new regime when it reaches the singularity. This could be an outright market crash like in 1987 or 2010 flash crash or something less spectacular. The major market indices have been behaving this way and last week reached its singularity in the LPPL equation. I placed a URL at the bottom if you want to research further. The essential features of the analysis are to first detect that the market is acting super exponentially (increasing in value an a faster than exponential growth). Then fit the LPPL equation to the daily close prices of an index or individual stock. This involves fixing seven parameters in the equation to best fit the data. The red curve is the result after optimizing the seven parameters in a least mean squares sense. Inside the equation is a cosine (log(t-tc)) function that blows up when t (calendar date) approaches the critical time tc as logarithms of negative numbers are not allowed. This log-periodic power law growth is typical of market tops such as 1929, 2000, 2007, etc and amazingly the critical time can often be estimated six months in advance and refined as the zero hour is approached. The October 1987 crash could be predicted in May of that year using this approach.

18628

Here is a URL leading to a Zero Hedge discussion and a paper on the subject:
http://www.zerohedge.com/article/didier-sornette-critical-market-crashes

[adam ali]

Mike, I really appreciate all the work you've undertaken to produce this. I've always kept Sornette's work in the back of my mind and to see it demonstrated in real time here is just awesome. Thanks!