Victor Niederhoffer: A Good Research Project
IBM flirted with 72.50 the last two days and Intel flirted with 25.00. Both reminded me (especially the apparent erroneous market order to sell Intel on Thursday that dropped it from 25.00 to 24.01 in a minute with a yo-yo type reversal a few minutes later) of the convulsive and disruptive behavior of an animal trying to evade a predator, which in this case was to evade the all too strong grasp of the remaining bulls. I propose an index of Intel to IBM to quickly monitor new tech to old tech and wonder what the leading indicator properties of this are, and whether the Nasdaq up 0.6% percent as of 2:40 pm on Friday, with the old S&P 500 down 0.9% percent, a spread of 1.5%, is predictive of anything.
Bruno Ombreux responds:
I don't have anything predictive, but I have explanations. I think I would make a good journalist.
I don't know if it is sensible to apply classical statistics to something as clearly nonstationary as the Nasdaq. However, I looked at the correlation between USD/EUR and the ratio Nasdaq100/DJIA, a proxy for growth vs. value.
Using daily log returns since 1992, I find a -0.04 correlation. This is highly significant given that we have more than 3300 observations. This means part of N100 overperformance could be explained by currency effects. That is a stronger dollar.
Another thing is the flattening of the yield curve. In Damodaran's Investment Fables, it is mentioned that a flattening to inverting yield curve and growth relative performance are correlated. This must be a good book, because you wrote on the cover that it is a good book. So I'll trust the author on this one.
A third thing is ever changing cycles. The chart below shows the cumulative returns of Stoxx style indices. These are euro indices, but the US is correlated and should show the same picture. Large cap growth has been underperforming. badly. It only makes sense that it starts overperforming. At least for a while.