Statistical Shocker: S&P 500 Performs Best When Economy is Shrinking

Impossible, right? As a money manager I spend a decent amount of time explaining to clients, readers, family, and friends that the stock market does not mirror the economy in real time. Just because the unemployment rate is 9.5% and GDP growth is decelerating does not mean that the stock market is a poor investment option. Stock market returns and GDP growth simply do not track each other, and as a result, reading economic reports will not help you figure out where stock prices are headed.

As always, I try to present numbers to people so they do not simply have to take my word for it. In today’s world of media sound bytes and political maneuvering Americans all too often repeat something they heard from one of their favorite media or political pundits as if it was fact, even when a tiny bit of research can disprove the claim.

In order to show that stock market movements do not mimic the economy, I decided to compile data from 1958 (the first full year the S&P 500 index was published) through 2009. While I had no idea what the actual numbers would be, I was confident they would show that stocks and the economy shared a very low correlation. Sure enough, the results were even surprising to me. It turns out that the S&P 500 has performed best when GDP growth is actually negative (i.e. when the economy is in a recession). Since 1958 there have been 7 years when U.S. GDP shrank and the S&P 500 gained an average of 24% per year during those periods. Pretty interesting, right?

Here is the full data set. I divided economic growth into 4 subsets (negative, zero to 3%, 3 to 5%, and above 5%).

As you can see, there is very little correlation between the economy and the stock market. Not only that, investors choosing to own stocks only in years with negative GDP growth would have earned nearly 4 times as much than investors choosing to invest only when GDP was growing at 5% or better. So the next time someone tells you the market is going to drop because the economy is bad or unemployment is high, send them a link to this blog post.

Enjoy this post? Subscribe and never miss another one: RSS | Email | Twitter

8 Thoughts on “Statistical Shocker: S&P 500 Performs Best When Economy is Shrinking

  1. Interesting. Results probably don’t change, but it would be interesting to see if there’s a calendar effect, and if these results hold up if you do it on a trailing 12 month basis.

  2. Tariq on July 27, 2010 at 4:27 PM said:

    Good post. I agree with the comment above in that the next step would be to see if there’s a lagged correlation – between the stock market returns of given year and the GDP growth of the following year (since presumably markets are anticipating GDP trends by at least 12 months, perhaps more).

    Under that scenario, the outsized returns in years of negative GDP growth would relate to the market already having priced in the GDP drop and starting to anticipate the return to growth in the following year. Another factor added on top could be the potential for overshooting in either direction also (i.e., arguably last year’s rally from March on was not only the anticipation of a return to growth following government stimulus etc, but also a realization that the markets had overshot on the downside and that the policymakers would prevent a wholesale financial sector collapse).

  3. Chad Brand on July 27, 2010 at 4:58 PM said:

    Indeed, market returns do seem to predict GDP strength the following year, which makes sense:

    S&P Return / GDP Growth Following Year

    Negative / +1.1%
    Less than 10% / +2.8%
    10-20% / +3.8%
    More than 20% / +4.6%

  4. Corey McGuire, CFA on July 28, 2010 at 7:22 PM said:

    I would have to look at the specific data, but I think this speaks to two concepts 1) Equity markets anticipate the direction of earnings, so the positive equity market performance coincident with negative GDP growth simply reflects investor’s buying future earnings recovery. 2) Earnings can grow through productivity despite a weak economic environment.

  5. John Hussman (Hussman Funds) says the same thing in his December 1, 2003 – Rationale Expectations, weekly market newletter. Nevertheless, nice to be reminded of this information, as most long term trends seem temporal but it’s handy to know it hasn’t changed.


  6. Chad, are you sure this isn’t data mining? Why do the stats only go back to 1958 when GDP data is available at the St. Louis Fed website back to 1929? I suspect the answer is because it may change the conclusion.

    Another test would be to use 1- or 2-quarter GDP periods instead of full years, especially since we’ve had a mix of short, mild recessions and big, long ones during the measurement period.

  7. Chad Brand on August 22, 2010 at 8:42 PM said:

    @Hondo – I would never misleadingly exclude data. The S&P 500 did not exist back in the 1920’s. I purposely included that tidbit in the article (reread sentence one of paragraph three).

  8. Andrew on August 30, 2010 at 12:40 PM said:

    What people fail to realize that a recession is needed to weed out the malinvestment and pave way for the competent companies to take over.

    Therefore, we should get the Federal Reserve out of the way and let the market truly function the way it is.

    It wouldn’t be just S&P 500 performing well but a lot of other sectors. It’s good to feel the short-term pain for the long-term gain (hey that rhymes).

Post Navigation