The Frontal Cortex

Optimism Bias

I’m pretty fascinated by this chart from the McKinsey Quarterly, which is a great demonstration of the optimism bias. The chart captures the earnings estimates of equity analysts for S&P 500 companies. The downward slope of these yellow lines is what happens when our hopeful projections meet dismal reality:


Needless to say, these estimates come from highly paid professionals, with access to vast amounts of data. (They’re also making projections about the relatively near future.) Unfortunately, all that data is no match for a deep-seated bias, which leads us to accentuate the positive and downplay the prospect of potential losses. (This helps explain why earnings projections are even less accurate during economic downturns.) Interestingly, the only segment of the population that isn’t vulnerable to the optimism bias are people with major depressive disorder. Maybe Wall Street should think about hiring them.


  1. #1 Doazic
    April 16, 2010

    Are there any studies out there comparing stock picking strategies of the chronically depressed?

  2. #2 dmaas
    April 16, 2010

    Its a beautiful graph! To appreciate it, I took out the labels:

  3. #3 Jerry
    April 16, 2010


    this has nothing to do with optimism. Equity analysts are just cheerleaders for the companies they cover, so that their banks can get investment banking business. If the analysts started becoming pessimists, the banks would get no deals.

  4. #4 Mike
    April 16, 2010

    It’s regression to the mean, right?

    S&P 500 companies are those which have performed well in the past… companies that are pretty successful. The problem is, we fail to regress our projections.

    My guess, if you looked at analyst expectations for poor stocks, you’d see the opposite – you’d see rising expectations over time. People would have initially failed to regress past performance as they project it into the future.

  5. #5 royniles
    April 16, 2010

    Optimism bias is the buyer’s need that salesmen fulfill with their purposes.

  6. #6 DeborahInLA
    April 17, 2010

    All the data points look like sperm–does that mean anything?!

  7. #7 Dennis Gorelik
    April 18, 2010

    Optimism bias could be easily eliminated by discounting optimist forecasts by, say, 5%/year.
    The problem is that even after such discounting forecasts are still not accurate at all (either too optimistic or too pessimistic).

  8. #8 royniles
    April 18, 2010

    Optimism is a form of the hope that allows you to get out of bed in the morning and face again the real prospect of failure. A bias formed for an unbiased purpose.

  9. #9 Mozza
    April 18, 2010

    This is an important graph. (1) Demonstration of optimism bias in human nature; (2) demonstration of the uselessness of financial forecast (even in catastrophic years, the predictions were standard); (3) demonstrates the power of a good chart to finally drive home an idea supported by tons of data. I wish that there were more details on the years that were different from the average (words with arrows for instance).

  10. #10 Justin
    April 18, 2010

    Actually, there’s a fairly rich literary history of the “adventures” of the majorly depressed on Wall Street… Bartleby the Scrivener anyone? (,_the_Scrivener) Maybe Melville wasn’t quite the introspective neuroscientist Proust was, but his take on the perversion of human nature in the emerging industrial/financial environment is worth analyzing as much as this fascinating graph.

  11. #11 richard
    April 19, 2010

    Interesting, but not the full picture though. These analysts are not investing, merely selling their research (so called sell-side). I think when it comes to actual investing (buy side), loss aversion trumps the optimism bias, on average. What do you think?

  12. #12 Jay Ulfelder
    April 19, 2010

    I’m puzzled by the claim that this chart confirms an innate bias toward optimism. That claim seems predicated on the idea that analysts consistently overestimate a company’s earnings, but that’s not what I see in the chart. In some years I see the downward trend on which that claim seems to rest, but in some years I see the reverse (analysts underestimating eventual earnings), and in some years I see substantively little difference between analysts’ initial expectations and the earnings eventually reported. It looks like there are probably more years with downward trends than there are flatlines or upward trends, but some of the downward trends are a matter of a few cents per share or less. More important, if the downward trends are to be explained by an innate bias, then why isn’t there a downward trend every single year?

  13. #13 Mozza
    April 20, 2010


    I think you’re reading the graph the wrong way. Still, your way seems to make the case for the optimism bias: 18 out of 24 years (75%), analysts have started by overestimating the EPS, then adjusting downwards as reality set in. The years where they didn’t were exceptional in growth and proved to be irrational and lead to a major economic crisis. It was too good to be true – even for optimists.

    I think the stronger evidence for the optimism bias lies on the vertical axis: at a given date, the predictions for the longer time frame are consistently higher than for the short run. To this rule, I see only one exception across all the graph: the 1999 prediction for 2002. This is what optimism is about: it will be better in the future.

New comments have been disabled.