There is the meme going round that somehow a "bear" market is different. That bear markets are the opposite of the hot hand. That because for multiyear periods in 29-43 and in 68-82 the S&P index gave zero returns, that this is "proof" that bear markets are not random. That Bear markets are different, they have such strong momentum we must throw out the zero correlation results that apply to "normal times".

That such a couple of strings of bad luck is "proof" that we should throw out the random walk theory. In fact a certain contributor to this web site has written a book on these "secular range bound bear markets"

If I understand his summary correctly, he is basically suggesting that we abandon randomness and predict a long secular bear, reverting to economic arguments, because 4 "secular range bound bear" existed for more than 5 years and 1 "depression" lasting from 1929-33, and then returning by 1943 in about the last 120 years of the Dow. His suggestion is that we ignore the dividend returns and go with the P/E. This of course, despite no correlation to change in P/E and change in Index. Hence forget the numbers, go with the economics. Of course I am not suggesting that the P/E or economics over these "secular bears" where not bad. Sure, due to the outlook of Labor Unions, Japan eating our lunch in manufacturing, high capital gains and tax on the rich that drove the P/E down in 1968-1982, to very low levels… But this is all in hindsight.
This of course has the media in a frenzy, those without the numerical literacy buy the story hook-line and sinker.   Even many, that should know better, have bought into the argument calling for an "inevitable depression".

If anyone remembers the first e-mail I sent to this web site was just about such "hot-hands" in basketball. My response to the coaches was that randomness is much more clumpy than we think.

It would appear to me that the burden of proof goes to those claiming these streaks are proof of inevitable structural failings continuing (having "momentum") i.e. to the "secular bears" camp, not those claiming "bear markets don't exist except in hindsight."

But what does a simulation say simply using randomness: how abnormal are these "secular bears". My simple simulation using 9% drift per year and 20% volatility with log normal distribution using 500 separate 120 year periods, suggests that it is pretty close to average. Slightly above average, but nothing close to "proof" of a "secular bear" that is anything but random fluctuation. (The 9% drift assumes 2% dividend each year which over the last 100 years was probably too generous to the bears camp. I did use Excel random generator, which I understand has some problems. So someone may want to use code with a "real" r.n. generator to verify the results).

My results average of 3 "secular bears" per 120 year period, but  many 6, 7, 8 separate (separate by at least 1 year with new high) periods. Average max length without new high 11 years, but many  15/16 and 17 year periods.   Of course if you allow 1 year new high before going range bound ( like the 2000-? secular bear)   this  extends such "streaks".

So  it would seem to me that what we have seen is only slightly above average.   Which I would suggest the meme has more to do with  psychology and "bear baiting" than economic inevitability.

Now this is not to suggest that understanding the new new meme on why USA is doomed is not worth  while short term as I suggest this will catch many on the wrong foot short term. Just expect it: when the random news goes the bear way the short term swings will be big. Nor is this to suggest that some poor fool with a 5-10 investment time horizon and low tolerance for risk, whose advisers don't understand the risk, should be 100% + in equity, thinking 10% drift means risk free 5-10 years out.

Indeed: Et tu counters?





Speak your mind


Resources & Links