Daily Speculations The Web Site of Victor Niederhoffer and Laurel Kenner

 

The Chairman
Victor Niederhoffer

Home

 
Write to us at: (address is not clickable)

4/26/05
Working Hypotheses, by Victor Niederhoffer

Often the screen will go down, and to great consternation the floor will keep me updated as prices vary wildly from what's on the screen, as happened today beginning at 10:52 a.m. in the S&P E-mini. Invariably, the screen will come on a while later and the actual price will be within a gnat's eyelash of the broken price on the screen. 1166 was the price this time. [Editors' note: Always the unexpected: the price was 1161 when electronic trading resumed.]

Holbrook Working developed some statistics in the 1950s based on the distribution of expected move assuming independence of, say, 100 changes versus 1 change, and found that prices tended to move inordinately less in the big period than would be expected if the moves were random.. Such a test should be made vis-a-vis the daily changes in a week versus weekly moves and monthly moves, and the hourly moves in a day versus the day to day change. Admittedly this is a Big League query, but there would seem to be many extensions of same that might reduce the number of big cons in the market.

Jim Sogi adds:

Hedgers pay speculators for the avoidance of risk according to the textbooks. When CME shut down yesterday afternoon at the strangely coincidental 1166, many speculators became instant hedgers. Those holding SP minis would have turned to alternate markets to hedge their positions, and accordingly the market oozed down for the afternoon, leaving an odd midday gap. Some testing on overnight gaps shows a tendency to fill over a few day period, but how to test such a rare shutdown and midday gap is unclear. The "trend' and economic numbers led to an overnight gap down, but speculators stepped in, and the hedgers paid them for doing so.

This has been the third 20 point elephant stampede Lobagola in a week. What's the 'story' behind the trade? Picture a school of fish rushing away, then all turning on a dime and rushing back. How about the personal stories behind the action; panicky sellers unloading inventory at any price, or hard nosed short hedgies with their sleeves rolled and their army of executioners driving prices down, and looking for their profits at the very bottom. The best story is right there in the mirror. Where are you in relation to the others? Its like watching tourists in New York, Hawaii, or anywhere. Tourists are looking up and saying, "Skyscrapers n' everthang." You 'see' them and what they are doing, but they don't 'see' you or what is going on around them. They are locked in their own little world, basically oblivious to what is going on right around them and can become marks.

Dr Philip McDonnell replies:

In A Non-Random Walk Down Wall Street, Andy Lo looks at variance ratios for individual stocks and portfolios for various weekly periods. The idea behind the variance ratio is that dispersion is linear in the variance (not the standard deviation). So the ratios should remain constant at about 1.0 if the market moves are independent and random. What he found was very much in line with the Chair's suggestion.

Using weekly data for 625 individual stocks he found ratios for 2 and 16 week periods as follows:

              2 wk        16 wk
625 stocks    0.97         0.89
equal weight  1.21         1.76
cap weighted  1.04         1.12

The results for the 625 individual stocks supports the idea that the variance ratio decreases with time for individual stocks. In contrast, for equal weighted portfolios the ratio increases with time, with a lesser effect for cap weighted portfolios.

In a subsequent elaborate analysis he attributes the above discrepancy to the tendency for stocks to follow other stocks in successive periods. The idea is that if Exxon Mobil goes up one day on news, the other oil stocks may benefit in successive days. Thus the cross correlations with lags cause portfolios to show increasing variance ratios over time while individual stocks do not.

For more of Victor Niederhoffer's writings, click here>>>