If businesses were to grow faster than GDP, they’d eventually represent virtually the entire economy, at which point the two would be one and the same. If the opposite were true, businesses would shrink until they became nonexistent.
But what does the data say? The results are, actually, quite interesting.
U.S. GDP and S&P 500
Comparing the quarterly values for the S&P 500 to the nominal US GDP from 1950 to today reveals a quite strong relationship between the two. This is not something unexpected. What is surprising, however, is the effect that financial crises seem to have on the relationship between the two.
Looking at the chart above, GDP almost seem to function as a moving average (or support even) for the market. In the long run, the two tend to be at the same level.
To reinforce this point, it is also useful looking at the trend lines for the two series. It is straightforward to see how they’re virtually parallel.
However, more and more in recent times the S&P 500 has detached itself from GDP, rising to incredibly high levels. Visibly, this only happened four times:
In the 1950s, when the world economy was boosted by the reconstruction effort following WWII. This 15-year boom was then followed by 15 years of underperformance compared to GDP, following the oil crises of the 70s;
In the late 90s, when the dotcom bubble was inflating. This was followed by a sharp correction in 2001, which brought the S&P 500 back to the level of GDP;
In the 2004-2007 period, when the subprime mortgage bubble was inflating. Again, stocks were promptly brought down to the level of GDP in 2008 with a loud crash;
From 2012 to today.
This doesn’t bode well for the stock market. Nonetheless, looking at the relative of the two shows that this deviation is not as bad as the dotcom bubble. Still, this shouldn’t be a cause for celebrations, as the 2007 financial crisis too occurred after a comparatively smaller deviation from GDP.
Given this bleak information, we should try to understand whether there is any other trend at play that could explain this deviation from GDP. After all, stocks are representative of only one of the contributing factors to GDP: businesses.
Business contribution to GDP
To achieve this, I looked at the data presented by FRED St. Louis on the Gross Value Added (GVA) by Business, Households and Government.
Whereas the chart above does show a slight increase in the contribution of business to GDP since 2012 (from 75.3% to 76.2%), this isn’t, in my opinion, sufficient to explain away the detachment of the S&P 500 from GDP. In actual fact, in the past the percentage contribution of businesses to GDP was much higher (82.2% in 1950) and the relationship between the two never faded.
In addition, the data seems to suggest that the contribution of business to GDP varies significantly with the business cycle. To highlight this, I’ve detrended and scaled both the S&P 500 and the business GVA. From the chart, it is clear that a relationship between the two does exist.
Further to this, the correlation between the two series has been quite high since 2007, at around 74%. It is unclear, however, whether a causal relationship exists and, if so, which way it goes.
What does seem to happen is that the business contribution to GDP is much more volatile compared to the other components. This means that it moves around the GDP trend depending on the business cycle. Keeping this in mind, the idea that the stock market might currently be overpriced is reinforced.
So, are we heading for a market correction? And if so, how large?
Whereas it is almost impossible to correctly predict the timing of such a correction, macroeconomic data suggests that it should indeed occur. In the last 7 years there’s been a growing disconnection between GDP and the stock market. Every time this occurred in the past, a correction followed.
As for the second question, scaling GDP and the S&P 500, the former is at a level of 1,362 whereas the latter is at around 2,800. Using data as of 01/07/2019 (the end of Q2, that is), a correction of around 54% is expected.
We need to apply some caution when interpreting this data. Using 1950 as the starting point for GDP was my somewhat arbitrary choice. Using other starting points, the results might differ. Nonetheless, the fact that the S&P 500 repeatedly returned to the level of this GDP series is a good indicator that my choice was at least sensible. It is also difficult to dispute with the fact that the S&P 500 has been trending above its long run average for the past 7 years.
Combining this information with other indicators, like the inversion of the yield curve and the current political climate, makes a significant correction all the more likely. It might be time to buckle up.
Written by Davide Buccheri