by Mike Kimel
The Historical Relationship Between the S&P 500 and the Economy, Part 2: The Abrupt Rise of Short Term Investing
Last week I had a post looking at how long it takes for events in the economy (in the guise of real GDP) to affect the stock market (represented by the S&P 500), and vice-versa. I noted that if you go back to 1950 and played with some correlations, you’d find that the S&P 500 tends to lead real GDP with the maximum effect occurring after about 10 quarters, and real GDP tends to lead the S&P 500 with the maximum effect occurring after about 16 quarters. Put another way – both the S&P 500 and real GDP affect each other (i.e., there seems to be virtuous cycle when things are going well, and a vicious cycle when things are going badly).
But… what if you look at shorter periods of time? Do these results hold? Using the same methodology as in the previous post – i.e., comparing the the correlation between the S&P 500 and real GDP in the same quarter, the correlation between the S&P 500 and the real GDP in the next quarter, the correlation between the S&P 500 and the real GDP two quarters later, etc. – we can find the number of quarters for which the effect of the S&P 500 on real GDP is greatest. But this time, instead of looking at for the period from 1950 to the present, we also look at it for the period from 1951 to the present, from 1952 to the present, from 1953 to the present, etc. This can tell us whether the degree to which the S&P 500 leads real GDP changes over time.
And here’s what that looks like graphically.
So how do we interpret this? Well, for the 1950 to the present period, the strongest correlation between the S&P 500 and a lagged real GDP (lags checked: 0 to 24 quarters) occurred at around ten quarters. That rose to 11 quarters for the 1970 to the present period, and then in 2002 that rose again to 13 quarters. Results from the most recent sample periods may simply be a factor of there being very few years in those periods. Nevertheless, it does seem that the relationship between the S&P and lagged real GDP is fairly stable. The stock market does seem to be leading the economy, and it seems the biggest effect of the stock market on the real GDP seems to occur after 10 to 13 quarters.
But what about the other way? For the 1950 to the present period, it does appear that real GDP takes about 16 quarters, or four years, to have its greatest effect on the stock market. (That isn’t to say there aren’t shorter term effects – announcements about the latest quarter’s economic growth do tend to move the economy, but that effect may be short lived in the fickle world of modern stock market trading.) But has that relationship changed over time? Here’s a graph:
Urghhhh. How do we interpret this? The graph seems to indicate that early on, the economy led by about 4 years, and that lead-time slowly decreased over time until…. it simply dropped off a cliff in 1974. When the maximum lag is zero, that means that the highest correlation between the economy and the S&P 500 occurs with no lag at all. Note … not shown in the graph is the fact that the correlation between the real GDP and the S&P 500 in the same period began dropping at the same time as well.
What was that all about? My guess, and it is just a guess right now is that when the economy started to suffer high inflation, people lost the ability to judge whether there was “real” growth or not. When prices are rising rapidly, who really knows how the economy is performing or what they might do to company’s listed on the exchange. An alternative explanation that may be related… maybe big market players started paying a lot more attention to real GDP all of a sudden, and started becoming more interested in the short term.
The interesting thing is that once the stock market started to decouple and/or focus more on the economic short term, it stayed that way. To tie it in to something I wrote a while back, in Presimetrics, Michael Kanell and I note that the stock market performed better during the George Herbert Walker Bush administration, when the economy grew mediocrely, than it did under a number of presidencies that produced rapid growth, such as those of LBJ or Reagan.
So what about that big change at the end of the sample? Is the market starting to pay attention to the long term again, or is that spike a function of a very short sample toward the end? My bet, given how unstable it is, is the latter.
I wonder what all this means for value investing.
This post is the second in a series on the historical relationship between the economy and the stock market.
1. Data used in this post is the adjusted close of the S&P 500 going back to 1950 and quarterly nominal GDP going back to the same date. Because quarterly GDP figures measure the economy at the midpoint of the quarter, the S&P 500 for February, May, August and November are considered the analogous “quarterly” S&P 500 figures.
2. I’m not a financial advisor. I strongly suggest against making investments based on anything written above.
3. If you want my spreadsheet, drop me a line via e-mail with the name of this post. My e-mail address is my first name (mike), my last name (kimel – with one m only), and I’m at gmail.com.