It is often said that the economy and the stock market are not the same. Evidence for this can be seen during 2019. For most of the first quarter, economic news was underwhelming. This was confirmed by the slowdown in consumer spending reported in the 1Q GDP estimate by the Bureau of Economic Analysis (the 3.2% annualized increase in 1Q GDP was driven by increases in inventories and state and local government spending and improvements in trade).
Despite the lackluster economic data during the first quarter, dovish statements and dovish meeting minutes by the Fed reassured investors that rate increases were on hold. This was aided by seeming progress on the trade agreement with China. The result was a significant gain in the stock market from January through April.
Then, alas, just as a range of economic indicators are improving in May, Trump sends out several tweets about the China trade "deal" and the S&P 500 is down 2.5% in one week. With new tariffs in place and the Chinese threatening retaliation, the stock market is struggling to find new footing as investors seek to price in the effect of the escalating trade war, and whether a solution is likely. The expectations, fear, and/or psychology of investors drive the stock market in ways that go beyond pure economic activity.
Correlation between real GDP and the real S&P 500
Looking at the correlation between the forward-looking stock market and reported GDP demonstrates how weak this relationship is. Table 1 shows the correlation by quarter between GDP and the S&P 500, both adjusted for inflation, for the current and other most recent business cycles. The highest correlation occurred during the 1991-2001 business cycle, 0.14. The current business cycle has the lowest correlation, 0.08. As I will show momentarily, ZIRP is a factor for the current business cycle.
Finding a better correlation
In an earlier article, I showed how the stock market generally trends with the leftward and rightward shifts of the mean of coordinates (MoC) that is plotted on the BaR Analysis Grid© (scroll down to Grid 3 in this article). As shown in Grid 1, the BaR makes it clear when the economic growth is accelerating (shifts to the right) or slowing (shifts to the left). The MoC is the average of 19 economic indicators, and in this grid, it is a three-month average. The MoC captures the general state of the economy. If you are not familiar with the BaR, you can read about it here.
The long leftward shift in 1Q 2019 was due to lower sentiment and expectations by consumers, small businesses, credit managers, and purchasers. The BaR is designed to show when a recession is near, and these outlook measures are critical for doing so. Though exaggerated, the leftward shift is consistent with the preliminary 1Q 2019 GDP estimate that showed a slowdown in consumer spending. However, because the MoC is well above the baseline, it shows that there is still a lot of positive economic activity happening.
The sentiment of economic stakeholders, at least up until the breakdown in the China trade talks, has been improving. In May, the MoC has nearly shifted back to the Expansion quadrant (see the most recent BaR here). What effect the trade war will have remains to be seen.
Advantage of the BaR over GDP data
Because the BaR captures consumer, small business owner, credit manager, and purchaser sentiment and/or expectations, it has an advantage over GDP numbers. As Gallup polls consistently show, consumer, small business owner, and investor sentiments generally trend together (see here and here). Consequently, the measures on the BaR should track more closely with the stock market than GDP.
However, as pointed out, in the first quarter of this year, investor sentiment remained optimistic while the outlook of other economic stakeholders turned more pessimistic. This only confirms an earlier point - the correlation between economic activity and the stock market only explains so much. How much is next.
Use of BaR Leading Indicators
For my analysis, instead of using the MoC, I'm using the leading indicators that are tracked on the BaR. Because the leading indicators (LDs) are forward-looking, they should correlate better with stock market behavior. The BaR tracks eight leading indicators: unemployment claims, STLFSI, yield curve spread (10YR/3M), nonfarm job openings, consumer sentiment, private building permits, small business optimism, and temporary employment.
Correlation between the leading indicators and the S&P 500
Using data from the 1991-2001, 2001-2007, and current business cycles, I've matched the BaR LDs rate of change with the average of the real S&P 500 rates of change.
I need to clarify that for the 1991-2000 business cycle, only nine of the BaR's 19 indicators were reported then, and only five of these can be used as leading indicators. For this reason, I am using six-month averages for that correlation. For the other two business cycles, for which all eight leading indicators are available, I'm using quarterly averages.
As shown in Table 2, the correlation for the 1991-2001 and 2001-2007 business cycles were respectively .29 and .44. This correlation is considerably stronger that what was seen in Table 1, which demonstrates that the BaR LDs are a better tool than GDP for correlating economic activity with the stock market. However, most interestingly, the correlation for the current business cycle is negative .05. However, as I mentioned earlier, ZIRP is a factor.
Table 2 confirms that although economic activity is a factor in how the stock market behaves, it is a partial influence.
To visualize the correlation shown in Table 2, Chart 1 plots the BaR LDs and the S&P 500 quarterly values during the 2001-2007 business cycles. It is evident that the LDs moved with, and occasionally led, the S&P 500. The .44 correlation is observable. If you are not familiar with the BaR, don't be thrown off by the higher rates of change for the leading indicators. They are set to a base of 100 using each indicator's historic range of change (read more here). Dividing by 10 would result in the left axis values being the same as those on the right, but the graph would look exactly the same.
In Chart 2, the relationship is displayed for the current business cycle, which has a correlation of negative .05.
Impact of ZIRP
The vertical dotted line in Chart 2 signals when the Fed began to increase interest rates, starting in 4Q 2015. For the quarters to the left of the vertical line, the LDs/S&P correlation was 0.03. After, from 4Q 2015 until 1Q 2019, it increased to 0.29, similar to previous business cycles as shown in Table 2. More normal interest rates are resulting in a more normal correlation.
How the S&P 500 performs when the LDs are increasing and decreasing
Table 3 shows how the S&P 500 rates of change correspond with changes in the LDs rates of change. The LDs rates of change (RoC) are categorized as follows: <-10%, >-10% and <0%, >0% and <10%, >10% and < 20%, and >20%. The observations used were during the current and 2001-2007 business cycles. Because fewer leading indicators were tracked in the 1991-2001 business cycle, and because of the stock market bubble, I chose not to include anything prior to 2001.
In Table 3, the LDs RoCs are monthly rates of change averaged over a quarter; whereas, the S&P 500 RoCs are strictly monthly averages - they are not averaged for a quarter. I'm showing it this way to capture the variability the S&P 500. A quarterly average suppresses the highs and lows.
The quickest way to make sense of Table 3 is by looking at the Average column on the right (F). This is the average of the S&P 500 monthly percent change when the LDs are increasing (green outline) or decreasing (orange outline). When the LDs RoCs are increasing, the average of the S&P 500 monthly percent change across all RoC categories is 1.6%. This compares to an average of minus 0.2% when the LDs RoCs are decreasing. The confirms that, on average, an improving economy will lead to higher stock market returns.
That conclusion is not a surprise. What is interesting is what is shown in the Maximum % Change and Minimum % Change rows across all LDs RoC ranges. There is a significant range in the S&P 500 monthly rates of change. Even when the LDs RoCs are decreasing (orange), the monthly maximum percent change in the stock market can be as high as when the LDs RoCs are improving (green). However, there is clearly more downside risk and lower average returns when the LDs RoCs are decreasing.
Table 3 suggests that if investors don't factor in whether leading indicators are improving or slowing, they could underestimate or overestimate risks and returns.
A backdrop of risk and return
I've melded the BaR Analysis Grid with Table 3 to assess where we are now. Most recently, the LDs are moving in a direction in which there is, on average, less downside risk and, on average, higher returns. The direction of the LDs could well change with the trade war heating up.
However, with a correlation of 0.29, the current economic trend is only one factor. As already seen during 1Q 2019, the stock market has moved opposite of the BaR's leading indicators.
- Leading economic indicators provide a better correlation to the stock market than reported GDP (Tables 1 and 2).
- As the Fed has begun to normalize interest rates, the usual correlation between economic activity and the stock market is returning.
- Knowing when there are distinct periods in which leading indicators are improving or slowing can help investors determine levels of risk and return.
- The correlation between the leading indicators and the stock market beginning 4Q 2015 is 0.29, which is consistent with the last two business cycles.
- This confirms that there are many other factors that influence the stock market. Never ignore Trump tweets (sadly), Fed statements, and market dynamics.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.