Friday, June 12, 2009

Macro and Market Timers

I am sure that you have sensed a bias that I bring to the table about market and macro timers but I think I should make it explicit. I believe that there are ways in which you can beat the market in the long term, but very few of those ways involve market timing or calls about the macro economy. I know that there are stories in every market about great market timers, i.e., investors who made exactly the right call at exactly the right time about a market: Japan in 1989, dot-com stocks in 2000, housing in 2007 and financial assets in late 2008. Here is why I remain a skeptic:

a. Small sample
: Unlike stock pickers who have to put out recommendations on hundreds of firms, the nature of market timing and macro calls (about exchange rates or the economy) is such that even the most long-standing forecasters have made only about 15-20 calls in their entire lifetime. As a result, rejecting the null hypothesis that successful calls are due to luck becomes much more difficult. For instance, a stock picker who gets 60 out of 100 calls right, is statistically beating randomness but we cannot reject the hypotheis that a market timer who is right 10 times out of 15 is just lucky.

b. Fuzzy recommendations
: One aspect of market timing and macro forecasting that is frustrating and makes testing difficult is the fact that market timers often do not make specific recommendations. Again, the contrast with a stock picker is stark: when a stock picker tells you a stock is cheap, you can buy the stock and test out the recommendation. Market timers and macro forecasters often make recommendations that are not just difficult to convert into action but also impossible to put to the test.

c. Timing is everything: Anyone who makes a market call and sticks with it for a long period will eventually be right. However, the call itself becomes a bad one for investors who followed it, since they often would have lost far more money in the period where the call was wrong than they made back at the time the call turns out to be right. Thus, calling the dot-com bubble in 1997, the housing crisis in 2004 and the Japanese stock market bubble in 1986 should all be classified as mistakes rather than the right calls.

Historically, there have been far more investors who have been successful, over long periods, picking stocks than timing markets, but the allure of market timing remains strong. Here are three tests that I would suggest you put any market timer or strategy to:

1. Has the market timer been right often enough to reject the hypothesis that his or her success is entirely random?
t statistic for success = Proportion of calls that are right/ (0.5/ Square root of the number of calls)
Thus, a market timer who has been right 15 times out of 25, will have the following t statistic:
t = (15/25)/ (.5/5) = 1.00
Statistically, this does not beat randomness?

2. Is the market timer right at the right time or is he or she a Cassandra?
Market timers who are consistently bullish or bearish are dangerous. There is more of a personal psychological component to their recommendation than an analytical component.

3. Does the market timer provide fuzzy stories or make specific recommendations?
I have more respect for market timers who are categorical about what investors should do - buy or sell short a specific market - than those who tell meandering stories (that may actually read well) but leave investors confused at the end.

Finally, ask the question that needs to be asked of any successful investor or strategy? Why does the investor or strategy succeed? Every successful strategy needs an edge. Since that edge cannot be better information with market timing (whereas it can for individual companies), what is it?

3 comments:

vsh said...

1) There are regimes where market risk has overwhelmingly large impact on individual stock returns. Stock picking was a pretty ineffective enterprise last year and market timing was unnecessary from 2003 to 2007. The two need not be mutually exclusive as you suggest - awareness of which regime we are in helps followers of both disciplines. The regimes can be pretty easily detected using market barometers like VIX, credit market index spreads, economic indicators, etc.

Alternatively, investors can rely on stock pickers that factor out market risk by shorting weak stocks.

2) Your article overweights the importance of being >50% right. One can outperform by making < 50% correct picks if the winners outperform the losers by some distance. This applies to both stock picking (an Apple over the last decade will make most stock portfolios look good) and market timing (someone who got 2003 to 2007 wrong but last year right by a larger margin would have still done well). On the flip side, someone who is 60% correct with his stock picks may not outperform if he has a C/GM in his portfolio among his 40% losers. He would be a better than average stock picker but so what?

Portfolio performance, correlations, max drawdowns, etc are additional measures to look at.

Similarly on market calls, one doesn't need to be gloriously correct on timing and magnitude to outperform. But, awareness of market risk is key to atleast be able to take a prudent and cautious stance. There is too much anxiety of being left out of market moves and too much yield chasing that makes this difficult.. thanks to "Bubbles" Ben.

3) "great market timers, i.e., investors" - we should distinguish between market timers who would be better categorized as traders than investors.

4) "Market timers and macro forecasters often make recommendations that are not just difficult to convert into action but also impossible to put to the test" - Right, because many times it requires trading skills - risk/money management to allow for failures until the trader gets the timing right. Many a times, it also requires going against government actions. Hence, my previous point about clearly distinguishing between trading and investing. As we all know the goals/desires of trading are quite different from those of investing. Hedge Funds/Actively managed funds address this difference for the common investor. One doesn't have to act on the market calls himself.

Instead of your 3 point checklist, one should just look at the performance of the person in question - and if they are not actively managing money (like say your colleague Roubini) then it would be prudent to listen to their arguments but not necessarily to their market timing calls. There is a difference between the economy and the markets and then there are various distortions like due to government intervention.

Aswath Damodaran said...

There is nothing here that I would contest. Of course, you can make more money with good market timing than you could with good stock selection. You can also make more money with the right lottery ticket than you could working at a regular job. The question is both one of probability and the facts. The facts, I am afraid, are against you. Notwithstanding all the anecdotal evidence you can come up with for great market timing, the reality is that market timing when practiced has been a miserable failure. From asset allocation funds to market strategists, the evidence is overwhelming.

vsh said...

Sorry for being verbose.

1) But, my point is that it doesn't have to be stock selection vs market timing. Attention to market risk is an important aspect of investing when having a stock selection strategy as well. In fact, stock selection would many a times implicitly factor in market/macro calls - for eg; rotating into consumer staples from consumer discretionary when suspecting a recession - would you call that a stock selection strategy or a market/macro call?

I'm contesting the point of pitting one against the other.

2) Note also that I'm not equating market timing to only cases of being spectacularly right with catching tops and bottoms which as you say is much more equivalent to buying lottery tickets. I'm arguing for recognition by investors of regimes when market risk is high and tweaking their investments accordingly. That is the definition of market timing/calls that I am using. You would not expect stock selection success to enable you to catch tops and bottoms of individual stocks so why define market timing success by whether someone exactly catches the market top/bottom or not?

2) The anecdotes were to discount the importance of being > 50% right in the number of calls (whether its judging stock selection or market calls). They were not to suggest that market timing is a better strategy.

3) For evidence I would not look at asset allocation funds and market strategists where the agents do not have skin in the game and are paid by fees only - that would exclude mutual funds and equity research analysts where the incentive to deviate from relative performance to the market is absent.

4) Saying that stock selection is a better/easier strategy than market timing is more a statement on market properties than on the strategies themselves - that markets trend upwards over the long term and market risk is low for a majority of the time so that in the long term differentiation is the only game to play.

Well, if the market properties change over the long term - if I was in Japan and the Nikkei recently made 20 year lows it is not obvious to me that stock selection was the better and easier strategy during that fairly long term period. Again, I'm not saying that it was easy to catch the tops/bottoms in the Nikkei but it may have been easy to dial down/up portfolio risk based on market cycles.