John Kuran points us to a study on Stock Market Extremes and Portfolio Performance.
click for larger graphic
Graphic courtesy Towneley Market Timing Study
While one frequently hears T-Heads mentioning how performance drops if/when investors miss the best periods in the market, one rarely hears mention of missing the worst. I recall Tom Dorsey (of DWA) discussing this some years ago.
Note that same market index performance of 12% per year (discussed prior via Jeremy Siegel) requires a very long duration to assure that level of performance.
Source:
Stock Market Extremes and Portfolio Performance
Professor H. Nejat Seyhun, University of Michigan
(commissioned by Towneley Capital Management)
http://www.towneley.com/html/study.htm
Very interesting and timely posts, but what is the average investor to do?
The University of Michigan paper indicate the results are “virtually unreachable. In terms of the monthly data, for example, if a market timer is right 50% of the time, the probability of executing a perfectly timed investment strategy is 0.5 raised to the 816th power — or nearly zero.”
I know some portfolio managers who move into cash and defensive positions when particular thresholds are met in the market. I also loved the critique of Segiel’s book. I’ve heard too many people (myself included) use the rationale if you miss the best days you are in trouble, they don’t mention missing the worst days. Do you know any good sources of information regarding to actionable portfolio timing/allocation techniques?
Thanks for all you do with your blog!
Regards,
Lisa
Rather than ‘timing the market’ I see it as an exercise in judging risks and return based on historical data. While no one can precisely determine turning points in the market the job of evaluating risks and return can be facilitated by studying similar periods of the past. If you determine that the current environment is one of high risks relative to future expected return then reduce your exposure. You are not proven ‘wrong’ by the market continuing higher in my opinion nor are you proven ‘right’ by the market going down; as your task is to establish risks levels not to be clairvoyant.
In respect to the current environment I would suggest looking at: historic returns after the fed has raised rates to a similar degree, historic long term returns if one is starting at a p/e of 20X, historic correlation between CRB and financial assets, historic impact to valuations if the growth of ‘e’ is contracting, historic impact to valuations when rates are rising, historic impact to valuations if CPI is rising, historic returns in post election years, historic return from election year high to midterm election year lows, historic ‘bubble’ cycles and time required to stabilize asset class, impact if dual bubbles of financial assets and real estate should be pricked inadvertently within a few years of one another… and I’m sure there are many others that could be added.
If one looks at the data I believe an objective conclusion as to risks can be reached. Predicting human behavior on the other hand is a whole other matter.
It is possible to market time and do it effectively. To give a thorough explanation requires more space than it is possible to do so in this comment section (maybe I’ll use multiple comments.). Marty Zweig had a book out a few years ago that had a pretty good strategy (it’s been so long since I’ve read it, I’ve forgotten the title.).
[BR: Winning on Wall Street http://www.amazon.com/exec/obidos/ASIN/0446672815/thebigpictu09-20%5D
IIRC, is the close of the week exceeds 4% above the lowest, recent lows, you’d go long. If the close drops below 4% of the most recent highest highs, you’d either exit the market or if you’re agressive, go short. There are many variatons of this and many other effective strageties.
You would think that Wall Street would love it if people try to time the markets all the time (with the commisions these trades bring in as revenues.). But with commissions so low these days, the second and third reason they oppose market timing comes in. The 2nd reason they oppose market timing is performance comparision. If someone can outperform Wall Street, why would they need these ‘professionals’ for advice? You’d only use them for custody of the account (which calls in the 3rd reason).The 3rd reason is liability. They are the ones holding on to the money. What if the market timer blows up and take the firms down with them (either thru leverage, misplays or just sues them for not stopping the timer from his/her own bad plays.)or take down other accounts as well? Those are the primary reasons Wall Street oppose market timing.
To those who say that this is not possible, because the markets are random, I ask, have you look at the mathematics behind this or are you just regurgitating what is taught in the ‘business’ school? There are tons of people out there who have done it effectively (and you can’t pull Nassim Taleb’s surviourship bias into this as the math his theory is based on is flawed as well.). I’d suggest you read Benoit Mandelbroit’s “(Mis)Behavior of the Markets”.
To have a random market requires a Normal / Gaussian distribution (aka Bell shape curve). A NG (for short) requires a kurtosis of zero, skewedness of zero and a Hurst Coefficient of 0.5. Well, the markets (any liquid public market, traded on the exchanges, or currencies) do not have a kurtosis or skewedness of 0 or a Hurst Coefficient of 0.5. The only time they take on these values is when they are crossing above or below it.
So what does the market distribution conforms to? A Stable Pareto Levy distribution. What are the implications of this? Short term is ‘predictable’ while the long term is not (anyone remember Dow 36,000 and the “New Economy” predicitons during the bubble years?). Think of weather forecast. You can tell what may happen in the next 3 days but not 3 years from now (rain/snow/ sunshine/ temperature range – but not the exact amount.).
I haven’t kept up with Marty Zweig lately, however I do recall a presentation he gave here in Denver… maybe 10-15 years back. At that point he conceided that his methodology was such that he did not outperform the market (buy and hold) rather that he believed by timing entry and exit he could capture most of the market return while experiencing less volatility. That’s added alpha, nothing wrong with that.
Wall Street does have conflicts of interests. There is the (real) liabilty issue as numerous studies indicate the ‘folly’ of market timing. However you may view the studies, there appears to be an incredible amount of evidence that most market timing efforts fail to outperform a buy and hold startegy long term (if you live to see it!). In court or arbitration making recommendations too far from the crowd is a dangerous thing.
There is ‘business risks’ if a professional/firm advises unconventional investing as well. There are many instances of a money managers making this error and losing their position just as the market finally proves them right. And many conservative professional managers saw a flight of assets (to Janus and the like) in the 1999-2000 period as their own good judgement generated poor relative returns (below 50%!).
A key obstacle for market timing is the client/human. That is the human element (emotion?) is anxious to buy just as they should be selling and to sell when uncertainty sets in. For this reason alone the investment firm recommendation of buy and hold is defensible. Again studies suggest the average investor does far worse than the market – due to their timing of buys and sells!
It seems logical that market timing can only be effective for a minority. It is the inefficiency of the masses that allows more active investors to capture excess return. The fact that active management is not practical for Wall Street (and the majority of investors) is what allows for the opportunity of those that can develop their own methodology to profit. If one is responsible to no one else (managing there own capital) and thinks independently there is a built in advantage.
I would be uncomfortable with any ‘black box’ -mechanical or mathematical model that has worked based on computer models. But that is just me. It strikes me that such models are an ‘easy’ answer. In fact they are almost a broker substitute. What I mean is that a common error of investors is to want to be told what to do, what to buy, what to sell, and when. In other words they do not want to have to think.
This is a difficult thing we do. Many intelligent educated folks are unable to outperform the market professionally and they are your competition. You have to be willing to do the hard work of thinking. Otherwise most do not have sufficient conviction to capitalize on the opportunities when they present themselves nor the confidence to remove oneself when everyone else is excitedly making money.
A present for Barry Ritholtz :)
http://delong.typepad.com/sdj/2005/03/william_polley_.html
March 25, 2005
William Polley Inveighs Against the Efficient Market Hypothesis
Particularly as applied to market reactions to the change of CEO:
Stumbling and Mumbling: What do bosses do?: How much difference do chief executives really make to a business? ‘A lot,’ say shareholders in Prudential. They raised the price of the company by £580 million yesterday when they learned that Jonathan Bloomer was to be replaced as CEO by Mark Tucker. If this judgment right, there’s something very wrong about the market for chief executives; either Mr Bloomer was massively overpaid or Mr Tucker is grossly underpaid. But is it right? Of course, the Pru’s price fell sharply under Mr Bloomer’s watch. But how much of this is really his fault?…
If I were to say that CEOs made no difference to a company, and that the belief to the contrary were just an application of the fundamental attribution error or managerialist ideology, what hard evidence could you present to the contrary, except for pointing to a handful of extreme cases, such as Enron?
There’s one more problem here, though. Let’s say CEOs can make a difference. It doesn’t follow that investors can spot the bosses who are good enough to turn a company around.
Two cases will show my point. When Simon Wolfson took over at Next, many shareholders were sceptical. They thought he was too inexperienced for the job, and was the beneficiary of nepotism. Next’s price rose sharply in the following months. By contrast, when Rick Haythornthwaite took over at Invensys, shareholders welcomed his appointment. He’d done a good job, they thought, at Blue Circle. Invensys’ share price has since collapsed.
As Warren Buffett once said: ‘when a chief executive with a good reputation takes over a company with a bad one, it is the company that keeps its reputation.’
I do market timing and sell my services in the soft dollar market. I have been doing it since 1980 and have a good but not perfect record.
My approach is to focus on the PE side of the equation. If you can be right on the direction of PE moves you have something like a 80% to 90% probability of being right on the direction of the market. Using earnings as a market timing device is useless. The correlation between the change in the market and the change in earnings since WW II has been -0.005– perfectly random.
Even if you use a perfect forecast of next years earnings the correlation is still only 0.2 — you would be better off flipping a coin. Since WW II on average in bear markets the pe falls by a third and earnings rise by 5% to 10%. In the first year of a bull market the PE rebounds by a third and earnings fall some 5% to 10%.
I have developed a set of leading indicators that has almost a perfect record of leading the PE over the last 40 years and rely heavily on that.
But things also change, so that you have to keep up to date. For example, money supply use to be a great leading indicator. But the financial system underwent significant structural changes
in the early 1990 and the old M1 relationship to the market PE broke down and I had to develop a different money supply indicator.
It is also why I use a lot of signal. No one signal is perfect — everything has failed at one time or another — but if you get almost all of a set of indicators giving you buy or sell signals you know to ignore the few that are not giving the same signal.
Second, you have to develop good decision rules that involve a difficult trade off between being whipsawed by false signals or noise but still be
timely. Moreover, the lag times have contracted over the years. 15 to 20 years ago my index had a good 6 month or more lead time. Now, that lead time has fallen to about 3 or 4 months or less.
In additon, market timing does not only involve betting on the direction of the market. When the PE is rising you are in a bull market and you should be in growth stocks or high beta stocks
while in a bear market you should be in low beta
defensive stocks. For example, the relative performance of oil stocks has over a 0.8 correlation with interest rates and industrial raw material prices. This correlation is higher than then the correlation with either nominal or real oil prices. But if you think about it the very things that drive the relative earnings of oil firms are exactly the same thing that drive interest rates higher. This is also true with most basic industries. The relative performance of aluminum stocks, for example with rates is also over 0.8.
It is not easy, and even though I have a very good record I have made mistakes — for example I was way to early in the last bear market.
I could go on and on, but this is enough.
Thank you to everyone! I never expected to get such a good collection of information to get me started.
Lisa
Many of Hussman’s weekly market comments @ http://www.hussmanfunds.com/weeklyMarketComment.html , e.g. 3/14/05, discuss the relative likelihood of “missing out”.