When is it good to be a bad forecaster?

I’ve been engaged in a very interesting discussion with Kash at Angry Bear on economic forecasting which will appear in the Wall St. Journal’s Econoblog later this week. Kash has a couple of posts [1], [2] based on his contributions over at Angry Bear. Here are the remarks that I used to open the discussion:

People often complain about the inaccuracy of economic forecasts. But for many variables, the inability to forecast the variable can be an indication that our economic understanding of the market is right on target.

The classic example of this is the stock market. According to one theory of stock prices, in a properly functioning market it should be completely impossible to forecast the change in stock prices. The argument is that if somebody knows that the price of the stock will be higher at the end of the week than it is now, they should buy right now while the price is still cheap. But such purchases drive the price of the stock up immediately. This theory thus maintains that the current stock price should reflect all available information, so that no predictability remains in the change between the current price of the stock and that at some future date. This is sometimes described as the “random walk” theory of stock prices, or, more accurately, the view that the stock price should follow a martingale.

There is both theoretical and empirical evidence against this martingale hypothesis. When investors care about avoiding risk, for example, there could be higher expected returns in times when the stock is more risky to hold. But there is no question that the martingale description is a pretty good first approximation to the behavior of most stock prices.

Actually economic theory suggests that not just stock prices but a great number of other economic magnitudes of interest should exhibit near-martingale behavior. Predictable price changes in any commodity that can be stored, for example, can be arbitraged away by increasing or decreasing current inventories. Such inventory adjustment again would result in the current price jumping to the value that best reflects the expected future price. Although costs of storage, changing inventories, and other benefits from holding inventories can introduce the possibility of more complicated dynamics, a martingale proves to be a good approximation to exchange rates and many commodity prices.

Indeed, it turns out to be a standard feature of any dynamic optimization problem in the face of uncertainty that a certain function of the magnitudes you choose should prove to be impossible to predict ahead of time– if it were ex ante predictable, you’re not using all the currently available information in the optimal way.

I’m sympathetic to the popular perception that if you really understand something, you should be able to predict what’s going to happen next. But there are at least some examples in economics where our theory suggests that if we really understand the phenomenon, it should be impossible to predict the variable. When someone tells me they can predict with a great deal of accuracy what the stock market is going to do next, my first reaction is that the person may well be a charlatan. Whereas when someone tells me they don’t have a clue, I think, now there’s a real expert!

38 thoughts on “When is it good to be a bad forecaster?

  1. Jack

    Would you be willing to postulate that long term trends may be known with more certainty than short term trends?
    Or, would you say that long term fluctuations carry equal uncertainty with short term fluctuations.
    If long term fluctuations are possible to predict, then at what timescale would you expect that to hold?

  2. pgl

    Consider the 1983 paper by Meese & Rogoff on forecasting exchange rate movements as one example of your opening remark. This Angrybear looks forward to your discussion with Kash.

  3. T.R. Elliott

    I find it more accurate to say that markets are complex phenomena that are not easily predicted. All the points about reflecting current information and the likes sounds good but saying as such is convenient, in a way: it elevates economics to a higher pedestal: “See everyone, our theories said we couldn’t predict it and we didn’t” instead of saying “the darn things so complex, we’ve barely a clue what it will do next.”
    See what I mean? I claim that the statement “the market contains or reflects all available information,” in many ways, contains little or no information. Just noise.

  4. JDH

    Jack, certainly long-term trends are relatively straightforward to recognize statistically in the sense that, if you try to forecast the level of a variable that is dominated by a trend, a summary of the forecast in terms of an R^2 or a correlation should look very good.
    On the other hand, the errors you make in identifying the exact parameters of the trend will accumulate as you try to project that trend farther and farther into the future, so that the absolute size of the difference between your forecast and the variable will get bigger and bigger as the horizon of your forecast grows. In that sense, even long-run trends are objects that we can make big mistakes about.

  5. JDH

    T.R.Elliott, the statement that the “price reveals all available information” is an extremely precise and highly testable proposition. It implies that if you regress the change in the price on any variable or group of variables that was known at time t or earlier, you should find all the regression coefficients are zero. This idea has served as the basis for many, many hundreds of articles in economics and finance that are devoted to testing such hypotheses to see whether or not they are true.

  6. Tracy W

    Long-run trends are unpredictable. The future depends on new knowledge and the discovery of new knowledge cannot be calculated ahead of time.
    To explain further, let us say you want to predict if an Aids vaccine will ever be invented. So you study the Aids virus, current medical knowledge, etc. Now to predict if it will be invented you need to figure out ways the Aids vaccine could work. At some point as you gather knowledge refining your prediction about whether an Aids vaccine could ever be invented, you have gathered so much information about how it could work that you have in fact determined that it does work (or doesn’t). Thus you can only perfectly predict the invention of an Aids vaccine by inventing it or by proving somehow that an Aids vaccine cannot be invented. At which point it is not part of future knowledge but part of current knowledge. And the new current knowledge will then be used by tomorrow’s researchers to invent new things that may well not have been invented had you not first invented the Aids vaccine.

  7. T.R. Elliott

    JDH: I don’t doubt the precision and testability of the statement, though I’ve only considered from the perspective of writings such as “Random Walk Down Wallstreet” and its ilks. I would expect there to be very low to no correlation between various measures of the market from one time to another or they would be discovered and pursued for financial gain–and therefore disappear (or decorrelate).
    But from my non-specialists perspective, I’m still not sure that statement of efficient markets tells me much more than the fact that I can’t predict. And when I combine that thought with the idea that market sentiments can change, that the “information” the prices incorporate include the willingness of someone and herding that goes along with it–with attendend prices bubbles and bursts–I’m just am not enamoured with the statement that the price reveals all information.
    At some point, I suppose this sort of argument because too metaphysical. For me, the saying the “price reveals all available information” still is tantamount to saying “we don’t have a clue what it will do next.”
    And I’m not complaining. It’s complex stuff.

  8. Jack

    I guess I’m wondering that if you state that you can’t predict what the market will do next and you apply that to any timeline, then how do you have any assurance that the market will rise at all over the long term. And, if you can’t say that the market will rise in the long term with any more than 50% accuracy, then what is anyone doing investing in the market at all?

  9. JDH

    Well, we almost always include a constant in these regressions to pick up the trend growth or the average required return. So, the more complete statement would be about deviations from those averages. Certainly stocks on average go up over time, no quarrel there! But the question is, can you predict ahead of time whether this year is going to have a higher-than-average return?

  10. JDH

    Or, to be more accurate yet, what is usually tested is the return on the stock over the period minus the return you would have earned on a Treasury bill over the same period minus the average difference between those two. For weekly stock returns that I talked about in the original post, these adjustments are very minor.

  11. anon

    You know, it is random and all. But somehow some people end up 19-1 and some end up 10-10 on their predictions. Random I guess.
    Work experience may help. Just as professors might be able to forecast student behavior better than non-professors, experienced workers may be in a position to forecast better than people with no work experience. Markets may appear efficient from the outside. However, for people with work experience, clues are everywhere.

  12. clarence darrow

    Who is “Kash”, and why would you pay attention to anyone who blogs anonymously? Does he have a ph.d. from a reputable school, or is he another Den Beste (college dropout who pretends to expertise)??

  13. Joo Carlos

    So, are you saying that Economy ISN’T falseable?
    Are you sure that Economy is a Science? I am sorry to say it, but if Economy’s predictions aren’t falseable as you are saying they aren’t, Economy ISN’T a SCIENCE. Maybe economy (with a little “e”) be a religion…. or some kind of pseudo-science as astrology.
    Maybe be better the newspapers put the economists forecasts at the same place they put the astrology’s forecasts if they have the same pseudo-scientific irrelevance.
    Joo Carlos
    Sorry my bad english, my native language is portuguese.
    My formation is Biology that is a TRUE Science. And articles as this one make me think that economy is a pseudo-science.

  14. JDH

    Joo Carlos, as I explained to T.R.Elliott above, the statement that “the price reveals all available information” is highly falsifiable, and there are many hundreds of studies investigating whether or not it is true in different markets.
    But, I agree with you that economics is not as pure a science as biology.

  15. AB

    “Clarence Darrow” — Kash isn’t anonymous, and he has a PhD from Princeton. And even were his real name not known, the correct term would be “pseudonymous,” not “anonymous.” In any case, that was a rather uninfomred comment, though I agree with the sideways criticism of den Beste

  16. eric bloodaxe

    There is an algorithm, which will generate an equation, which will pass thru’ all the points, on a set of random dots, but it cannot be used for extapolation.

  17. T.R. Elliott

    Adding a couple thoughts:
    1. Ormerod in Butterfly Economics argues similar to JDH above–and rightly in my opinion–that predictability is near impossible (my layman speak). He goes on to say that controllability is also near impossible–other than letting the market do whatever it’s going to do. Though he brings up important caveats when controlling the market may be possible and should be attempted. I think some in the economics community, depending on their ideological persuasion, would like to ignore these possibilities.
    2. The observation that markets are not predictable does provide us with some important information. E.g. many of the experts, even those consistently successful, may in fact be skilled, but they are also probably lucky. Caveat emptor.
    3. The non-predictability of markets, similar to the Omerod point made above, is often used by free market advocates to argue against intervention. And rightly so. But the argument that the market contains all available information is, in my opinion, sometimes pushed beyond reason, used for ideological purposes, making the “market” some sort of god mammon, an oracle, that knows all. Hence the pabulum on television that often passes for business and financial news, when we hear that the market has spoken on topic X or, vice-versa, topic X has affected the market. All bow down before said mighty market. 🙂

  18. anon

    steve jobs was a college dropout. in recent history, steve jobs probably got it right and eisner was probably wrong.
    so college is not everything.

  19. T.R. Elliott

    I agree with anon that college isn’t everything. That is not an argument against college. But the fact that someone drops out of college (as did Gates if I’m not mistaken) is not a mark against them.
    Steve Jobs was exactly the example I was going to use for someone who has hit several home runs in his life. This does not make him a genius. He has also been extremely lucky. But I think he’s demonstrated the ability to pursue a line of thought to success when most around him scoffed and were detractors.
    On the topic of colleges, I read a study a while back that indicated intelligence, not college, is the best indicator of success. It just so happens that intelligent people struggle and compete to get into the top colleges, so one would expect a lot more successful people who originated from these top colleges–but the correlation is with intelligence, not the college.
    Granted, there is a networking factor that should be taken into account, e.g. if you go to Harvard or Yale you’re more likely to befriend a future senator, president, or the likes.

  20. Tracy W

    Anon – there is a scam that sometimes gets carried out (and I don’t know which country you live in, but it’s generally illegal, so I advise against carrying it out).
    Write a letter to 16,000 people, telling them that your company has developed a new, high power, predictive model, and are sending the receipient a free prediction to build your company’s reputation. For 50% of those letters predict that the sharemarket will go up next week. For the other 50% predict it will go down.
    Next week, write a letter to those 8,000 people who got a correct prediction last week. Same terms, 50% of the letters predict the sharemarket will go up, the others that it will go down.
    Next week, write a letter to those 4,000 people who got a correct prediction last week. Same deal.
    Next week write a letter to those 2,000 people …
    Next week write a lettter to those 1,000 people….
    The next week write a letter to those 500 people who got the right prediction last week. This letters says “Our predictions have been right 5 tiems in a row. If you are interested in purchasing our services, please pay us $500 for our next prediction.”
    That some people manage to pick which way the market is going for a long time does not tell us much in itself. The next question is how many people failed to predict the market, and was the percentage of successes greater than you would get by pure chance?

  21. John

    Random Walk Theory is invalid — for example: Price range in the first 10 minutes sets the daily high or low about 20% of the time whereas a “random walk” in the first 10 minutes would set the daily high only about 6.25% of the time. There are trading strategies that exploit such reversions to the mean. Timeframe is paramount.
    Two efficient market guys are walking down the street and notice a green slip of paper that says “$20” and has a picture of Andrew Jackson on it. One considers picking it up but the other stops him, saying, “No, if that were really a $20, someone would have picked it up by now.”

  22. Anonymous

    Joo Carlos – Economic theories are falsifiable, at least the interesting ones.
    Some terminology:
    The economy is not falsifiable, any more than a rock is. The economy exists. Resources are produced and consumed, for example I had toast for breakfast this morning.
    Economics is not falsifiable. Economics is often defined as the study of the allocation of scarce resources. Even if every single current economic theory was falsified tomorrow, on the day after tomorrow people could still sit down and study the allocation of scarce resources. Equally, if every current biological theory was falsified tomorrow, biology itself would not be falsified. Whether we should be studying biology or economics is another question of course.
    Economic theories are sometimes falsifiable. At least the interesting ones are.
    One theory that has been thoroughly falsified is that there is a stable trade-off between inflation and the unemployment. The 1970s, by providing stagflation, did that.
    Another theory that has been thoroughly falsified is communism – a central planner cannot do a much better job of managing the economy than any market.
    Sometimes people come up with economic theories that are not conceivably falsifiable. You can believe them or not as you prefer. These are generally very boring theories as they make no difference to what sort of policy advice you would offer.
    And there are some basic economics tools that are really facts like rocks are. E.g. marginal utility. What this says is that you get different value from a good depending on how much you already have of it. Anyone who has experienced in their life both being hungry and being so full you have to struggle to swallow another bite knows that the marginal utility of another bite of food can change depending on how much food you have already had. I can’t see marginal utility ever being falsified. It would be like biologists announcing there are no, and never were, any birds.

  23. pgl

    Adding onto what AB said – our blog lists Kash as Kash Mansori as in Dr. Kashif S. Mansori who teaches economics at Colby.

  24. JDH

    John, bid-ask bounce is well-documented, but there’s not so much you can do to profit from it.
    Do you know where the Dow is going to be in March? That’s where the real money is. And if you don’t know, I wouldn’t be quite so quick to dismiss the theory as entirely invalid.

  25. John

    JDH: Some people make ‘real’ money by fading intraday highs or lows (not bounce). At this time frame, they don’t care where the Dow is going to be in March. (Volatility would be nice.) APPB was a nice play today.

  26. T.R. Elliott

    John: Other than anecdotal evidence, can you point us to research indicating long-term consistent performance?
    For others: An amusing read a while back: “Fooled by Randomness” by Nassim Taleb.

  27. T.R. Elliott

    John: What are you pointing to? As far as I can tell, you’ve referenced the hypothetical performance of some trading rules. Back tested results tell us nothing about future performance.

  28. John

    T.R: If you look further into the threads, you should find abundant evidence that those strategies have been traded successfully, consistently, in real time, not just back tested. I’ve emailed you a clip from my trading account showing you my trade of APPB this morning, which traded through its high and low within 15 minutes of the open. As you can see, I went long from about 0633 to 0639 for a nice profit for 5 min work. The point is not to brag but to show that, yes, it is a statistical fact that there are transients that can be exploited over certain time frames, thus invalidating Random Walk Theory.
    I’m not sure JDH and I are positing the same time frame. If RWT applies only for longer time frames, okay, goodie for the academics. If you want to make $, you need merely learn how to exploit the high probability price action (admittedly easier said than done).
    To reiterate, I don’t know what the Dow will be in March, and I don’t care. I certainly don’t know what APPB will be, and I don’t care. But I do know, through statistics, decent knowledge of market microstructure, good intraday scanning software, years of experience and just plain gut instinct, a high probability trade when I see one. I urge everyone to stare at today’s chart for APPB and entertain the notion that there was ORDER in that particular squiggle if only briefly and for the swift.

  29. T.R. Elliott

    This appears to be the most complete study on the topic so far:
    I’ve not assessed it entirely yet, in particular I need to better understand the long-term performance of the 20% who profit over a six month period. 80% of day traders lose money over that time period.
    Also note that five years, the extend of the study, is not a long time. Many day traders finally blow up. I’m going to bet, my intuition tells me at least, that over a good investment horizon–say thirty years (from a retirement perspective)–day traders will not outperform the market and will in fact underperform it. And that those who do survive: their success can just as easily be attribute to luck. Flipping coins, throwing darts, etc.

  30. a

    T.R. Elliott
    Good points on college vs. no-college. College is probably correlated with success as measured by money. However, there are noteworthy outliers – Bill Gates, Steve Jobs, Michael Dell, and others – who have no college degree.

  31. John

    “This appears to be the most complete study on the topic so far:”
    Oh, boy… another academic study. Why study losers if you want to learn about success?
    “80% of day traders lose money over that time period… that those who do survive: their success can just as easily be attribute to luck. Flipping coins, throwing darts, etc.”
    And most small businesses fail within 2 years. The ones that survive are sufficiently capitalized and have an edge. Same for trading, which should be approached no differently. Just b/c something appeals to the “get rich quick” mentality and gambler’s anonymous rejects does not make it equivalent. And driving a Buick like Warren Buffet won’t make you a billionaire. (Although driving Warren Buffet’s Buick could land you in jail.)
    NATI was a good play this morning and not chosen at random.
    Another simple method that works:
    When price has reached 90% of daily range, enter on a reversal with decreasing volume.

  32. T.R. Elliott

    John: Wow. You’re not John. You’re Micheal Parness. That guy on TV late at night. Trend Trading. (http://www.trendtradingtowin.com/) It’s an honor to exchange messages with you Mr Parness.
    Seriously though, my opinion is that everything comes down to luck and effort. Intellectual capabilities, for example, are a combination of luck (genes, upbringing, schooling) and effort (taking advantage of them). Success in life is a combination of luck and effort. One can break out the luck component to consider issues such as strategy, planning, execution, etc. But I claim it still comes down to luck and effort.
    So then let’s look at those small business owners. You are right. Most fail. This does not argue against starting a small business. But one should take into account the failure rate. Many who succeed are those well capitalized and with good business plans. Unfortunately, many of those that are well capatalized with good business plans fail. The only difference often enough: The circumstances were supportive of the one, not supportive of the other. Luck. This does not argue against personal or organizational merit. It does argue that even if you do everything right, you may fail.
    My own personal experience, having worked at QUALCOMM, which grew significantly while I was there, bears this out. And my knowledge of other successful companies tells me the same. A simple decision–Sprint deciding to go GSM–could have buried the QUALCOMM business plans. Maybe not. It’s hard to tell. QUALCOMM was successful, but as the saying goes, one should not get cocky about it.
    So when I consider day-trading, I look for analysis that demonstrates it is worth the effort. I know some people will do well day trading. But the odds are against them. And if they are betting large positions in order to make real money, they are likely to blow up, destroying everything they’ve accumulated.
    I see no evidence to indicate otherwise.

  33. T.R. Elliott

    One addendum: I should have said one can break out Luck-plus-Effort to consider issues such as strategy, planning, etc etc. It is erroneous to say stragety is just luck, for example. It’s a component of luck and effort.
    But my point is tha tI’ve come to believe in luck plus effort as the fundamental cornerstones of life.

  34. T.R. Elliott

    John: I completely agree. Luck favors those who prepare. Unprepared people won’t even recognize lady luck when she passes their way. And I’m not saying one cannot make money as a day trader, perhaps even long term. Just not me.
    May luck shine on you and your well-prepared trading strategy.

Comments are closed.