Stock-market forecasting.

Interesting Wall St. Journal article today on why most stock-market forecasts are wrong. The whole article is a good read, but here’s the part that surprised me:

History shows that the vast majority of the time, the stock market does next to nothing. Then, when no one expects it, the market delivers a giant gain or loss — and promptly lapses back into its usual stupor. Javier Estrada, a finance professor at IESE Business School in Barcelona, Spain, has studied the daily returns of the Dow Jones Industrial Average back to 1900. I asked him to extend his research through the end of 2008. Prof. Estrada found that if you took away the 10 best days, two-thirds of the cumulative gains produced by the Dow over the past 109 years would disappear. Conversely, had you sidestepped the market’s 10 worst days, you would have tripled the actual return of the Dow.

I wonder what bearing this has on the debate over whether or not share prices follow a random walk.

Comments

  1. I guess this explains why Lenny Dykstra was once a very successful investor.

  2. I’m not sure that the article has much bearing on the debate on random walk. The final paragraph says the market’s or individual stocks’ big moves in either direction seem to come “out of a clear blue sky.”

    In my opinion, the article does make a good point about the risks of market timing. If you are or were prescient or lucky enough over the last 109 years to be out of the market on its 10 worst days and in on its 10 best, you’ve made a fortune. However, I submit that very few are so prescient and, if you are relying on luck, your odds might be better in a casino. On the other hand, if you had invested part, most or all of your long-term time horizon – money that you did not need for a long time – in the stock market and left it there through the ups and downs, it has probably earned close to the 10% per year that the article quotes.

    My final point is that most of the time when human beings predict long-term future events that will be affected (indirectly and directly) by other human beings and by factors that are out of the control of humans, we should all be skeptical about the predictions. For example, look back at the April 2008 predictions by the baseball experts on the four-letter. How many had Tampa winning the AL East and the ALCS and Philadelphia winning the NL East and the World Series? None. I guess you could say Peter Gammons’s prediction of an Indians vs. Braves World Series was akin to a stock market analyst predicting last January that the market would gain 10% in 2008.

  3. The Dow has very little relevance these days Keith.

  4. Bob Warja, while you’re right that the Dow and its 30 stocks do have little relevance any longer, the Dow does correlate highly to the S&P 500, which does have relevance. Therefore, if the article’s author had used the S&P 500 since its inception in the 1950’s instead of the Dow since 1900, his findings would not have differed greatly.

  5. So, unless you are unusually gifted in sports, the best bet is to put your head down, work, save 10%, donate 10% and spend 80%.

    That should produce a decent retirement.

  6. H: I may have misinterpreted the part about “out of a clear blue sky,” but I took it to mean only that wide swings were not predictable from previous market data. I interpreted that phrase to mean that those big days, plus or minus) were caused by exogenous events (like a terrorist attack) that were not random but weren’t tied to previous market performance.

  7. I think this article provides an interesting thought experiment, but little actual relevance. Why would you even consider taking away the 10 biggest losses/gains?

    Rational expectations says that price = value, so if those 20 days were truly dismissable outliers, the market would have quickly corrected for them. Instead, prices/valuations stayed at the subsequently lower/higher levels, indicating that price (still) equals value.

    What’s the difference between a 10% one-day drop or a cumulative 10% one-week drop? A dam can either leak or burst, but when all is said and done, the result is the same.

  8. Keith,

    I don’t see the correlation between the findings mentioned in the article and random walk. Random walk is a forward-looking theory and no studies of an individual stock’s or the enitre market’s past history can tell me right now whether GE’s stock is going to move higher from its current price ($12.54) or whether the S&P 500 is going to move higher from its current level (849.71).

  9. I think this offers credibility to the idea that picking stocks and active asset management is a useful endeavor. That said, if you don’t want to pay someone to help you invest, you should double down on the market indexes following a bloodbath.

    An important aspect of the article’s analysis that we overlook is the dividend payment. Valuation growth is sexy, but solid, reliable valuation with steady dividend payments is a smart way to invest.

  10. Good points FQ. In addition to doubling down (or at least adding to equities in some way) following a bloodbath, investors would be smar to reduce equities following a day with huge gains.

  11. What FQ said.

    Markets are unpredictable and gains may be in nominal terms because of inflation. For instance, right now the Dow is priced just under 9:1 in gold. It used to be 45:1. It may go to 1:1 like it once was.

    That will likely balance out, barring a dollar collapse. Thus a steady flow of cash via a strong dividend yield is the best bet, IMHO.

  12. Anyone have thoughts about TBT? I think Treasuries are maxed out right now, and the bubble on that is going to burst. If that happened TBT will go to 200+, but–then again–200 bucks if that happens won’t buy too much.

  13. I’m in TBT right now. I got in at $37 and I’m pretty happy. However, I’m not sure getting in at $44 is as profitable as getting into corporate bonds and some distressed assets. The driver for TBT is going to be China. The Fed could buy longer term treasuries in the near future, but I think China will do whatever it has to grow at 9% – political and demographic realities dictate this. I’m not sure the Fed has another $T on it’s balance sheet should the Chinese flood the market with Treasuries. Don’t think of TBT as an inflation bet either – a better call on that front would be commodities – but be wary because a lot of commodities ETFs have operational issues.

  14. but be wary because a lot of commodities ETFs have operational issues.

    I always say: Buy the real thing. If crap hits the fan, who knows if those ETFs are really holding what they say they are. Call me a pessimist, but I’d rather just own gold and silver in bullion form than a GLD.

  15. The “best ten days, worst ten days” while technically correct does not tell the whole story, especially if you expand it to the “best 30 days, worst 30 days’. Basically, a lot of these “best days” were withing a day or two of a lot of the “worst days” and were, a lot of times, what they call the “dead cat bounce”. So, even if you did know when these days were going to happen, the transactional and tax costs of buying and selling an entire porfolio is a huge drag. The moral of the story…time IN the market is more important than timing the market.