The following is raw data on the performance of the Dow Jones Industrial Average (DJIA) versus the National Bureau of Economic Research (NBER) call of a recession from the peak to trough from 1900 to 2018.
Our aim is the determine if there is any coincidence or correlation between the two. We’d also like to emphasize that it would be especially ideal if there is confirmation of the idea that the stock market is a leading indicator for the economy.
Simple Coincidence
Below is the simple coincidence of the DJIA and NBER. This takes the date when the NBER calls a recession or an expansion and registers the level of the DJIA for the first trading day of month that an NBER call takes place and registers the level of the DJIA when the next NBER call begins.
The times when an recession was called but the DJIA was instead higher is indicated in red. As an example, On August 1, 1918, the NBER indicated that there was a recession until March 1, 1919. In that same period, the Dow increased from 80.71 to 85.58.
There was no instances of an expansion period being called by the NBER that was followed by a lower level in DJIA.
Date |
NBER |
DJIA |
December 1, 1900 |
expansion |
66.43 |
September 1, 1902 |
recession |
66.55 |
August 1, 1904 |
expansion |
52.73 |
May 1, 1907 |
recession |
83.87 |
June 1, 1908 |
expansion |
74.38 |
January 1, 1910 |
recession |
98.34 |
January 1, 1912 |
expansion |
82.36 |
January 1, 1913 |
recession |
88.42 |
December 1, 1914 |
expansion |
56.76 |
August 1, 1918 |
recession |
80.71 |
March 1, 1919 |
expansion |
85.58 |
January 1, 1920 |
recession |
108.76 |
July 1, 1921 |
expansion |
91.26 |
May 1, 1923 |
recession |
97.40 |
July 1, 1924 |
expansion |
96.45 |
October 1, 1926 |
recession |
159.69 |
November 1, 1927 |
expansion |
181.65 |
August 1, 1929 |
recession |
350.56 |
March 1, 1933 |
expansion |
52.54 |
May 1, 1937 |
recession |
174.59 |
June 1, 1938 |
expansion |
110.61 |
February 1, 1945 |
recession |
153.79 |
October 1, 1945 |
expansion |
183.37 |
November 1, 1948 |
recession |
189.76 |
October 1, 1949 |
expansion |
182.67 |
July 1, 1953 |
recession |
269.39 |
May 1, 1954 |
expansion |
319.35 |
August 1, 1957 |
recession |
506.21 |
April 1, 1958 |
expansion |
445.47 |
April 1, 1960 |
recession |
615.98 |
February 1, 1961 |
expansion |
649.39 |
December 1, 1969 |
recession |
805.04 |
November 1, 1970 |
expansion |
758.01 |
November 1, 1973 |
recession |
948.83 |
March 1, 1975 |
expansion |
753.13 |
January 1, 1980 |
recession |
824.57 |
July 1, 1980 |
expansion |
872.27 |
July 1, 1981 |
recession |
967.66 |
November 1, 1982 |
expansion |
1,005.70 |
July 1, 1990 |
recession |
2,899.26 |
March 1, 1991 |
expansion |
2,909.90 |
March 1, 2001 |
recession |
10,450.14 |
November 1, 2001 |
expansion |
9,263.90 |
December 1, 2007 |
recession |
13,314.57 |
June 1, 2009 |
expansion |
8,721.44 |
December 1, 2018 |
recession |
25,826.43 |
There were 8 instances (17%) where there was no coincidence with the call of a recession or expansion and a commensurate decline or increase in the DJIA.
The above coincidence data is graphically represented below. The areas in red includes the the divergence of the NBER call for a recession and the DJIA along with the period that immediately followed. This is basically showing that any recession indication that is followed by an increased in the DJIA, and the subsequent expansion calls, are not considered to be coincidence until after the last expansion and the next coincidence of a recession call.
![image image](https://i0.wp.com/www.newlowobserver.com/wp-content/uploads/2018/12/image_thumb-188.png?resize=500%2C363)
Naturally, this puts the coincidence level at 65% instead of the previous 83%. This is a literal take on whether there is a coincidence between the direction of the DJIA and when the NBER actually calls a recession or an expansion. It could be said that the DJIA follows the directions of the NBER except that the call made for the economy usually takes place at least a year after the fact.
Recession Length and Coincidence
Another method for measuring the coincidence of the DJIA and the official NBER definition of a recession is to rank the recession by length. In this case, we take the beginning of a recession and end of a recession and use the first trading day of that month and measure to the first trading day of the month when the recession is considered to have ended. Below is the ranking of the length recessions from shortest to the longest compared to the DJIA.
NBER peak |
NBER trough |
previous expansion (months) |
DJIA |
January 1, 1920 |
July 1, 1921 |
10 |
-37.16% |
January 1, 1913 |
December 1, 1914 |
12 |
-35.81% |
July 1, 1981 |
November 1, 1982 |
12 |
3.93% |
January 1, 1910 |
January 1, 1912 |
19 |
-16.25% |
September 1, 1902 |
August 1, 1904 |
21 |
-20.77% |
August 1, 1929 |
March 1, 1933 |
21 |
-85.01% |
May 1, 1923 |
July 1, 1924 |
22 |
-0.98% |
April 1, 1960 |
February 1, 1961 |
24 |
5.42% |
October 1, 1926 |
November 1, 1927 |
27 |
13.75% |
May 1, 1907 |
June 1, 1908 |
33 |
-11.32% |
November 1, 1973 |
March 1, 1975 |
36 |
-20.63% |
November 1, 1948 |
October 1, 1949 |
37 |
-2.97% |
August 1, 1957 |
April 1, 1958 |
39 |
-12.00% |
August 1, 1918 |
March 1, 1919 |
44 |
6.03% |
July 1, 1953 |
May 1, 1954 |
45 |
18.55% |
May 1, 1937 |
June 1, 1938 |
50 |
-36.65% |
January 1, 1980 |
July 1, 1980 |
58 |
5.78% |
December 1, 2007 |
June 1, 2009 |
73 |
-34.50% |
February 1, 1945 |
October 1, 1945 |
80 |
19.15% |
July 1, 1990 |
March 1, 1991 |
92 |
0.37% |
December 1, 1969 |
November 1, 1970 |
106 |
-5.84% |
March 1, 2001 |
November 1, 2001 |
120 |
-11.35% |
In this perspective on the coincidence between recessions and the performance of the DJIA, we can plainly see that there is a 63% coincidence. Overall, not a bad amount of coincidence. However, we think that we can generate an outcome that is closer to 100% coincidence if we twist the data to fit our agenda.
There is a saying that “the stock market is a leading indicator for the economy.” We promise we didn’t make this up. Furthermore, we have quoted the venerable Richard Russell of Dow Theory Letters fame to prove our point.
"The stock market is an indicator for the economy, a leading indicator (Russell, Richard. Dow Theory Letters. October 4, 1967. page 2.).”
"Just as [Elliot] Janeway senses new leading indicator of the business-market condition, I too, often sense an index which I feel should be accorded great authority. Right now I would say it is world stock exchange averages (see last Letter). The leading stock markets of the world are now heading down in earnest (statistics in Barron's each week), and this has an ominous ring to it (Russell, Richard. Dow Theory Letters. August 29, 1973. page 6.)."
"...if you believe that the market is its own best leading indicator then you have to believe what this market is saying (Russell, Richard. Dow Theory Letters. September 19, 1984. page 2.)."
"I continue to remind my subscribers that the crucial issue here is NOT whether the CPI turns up or down next month, it’s NOT whether the leading indicators blip up or down in July. No, the critical issue here is the direction of the primary trend of the stock market (Russell, Richard. Dow Theory Letters. June 8, 1994. page 2.)."
For nearly 6 decades, Richard Russell impressed upon his readers that the market leads the way when it came to understanding the direction of the economy. Naturally, William Peter Hamilton, fourth editor of the Wall Street Journal, had the following to say about the insights of the stock market:
“The market is not saying what the condition of business is to-day. It is saying what that condition will be months ahead (Hamilton, William Peter. The Stock Market Barometer. Harper & Brothers. 1922. page 42.).”
Not to be outdone, Charles H. Dow, co-founder of the Wall Street Journal, has the following to say about the stock market as a leading indicator:
“The stock market discounts tendencies. Stocks went up before the improvement in business became pronounced. Stocks will discount depression before depression actually exists, but this discounting quality in stocks make them run to extremes. They discount shadows as well as substances and often anticipate that which does not occur (Dow, Charles H. Review and Outlook. Wall Street Journal. May 10, 1900.).”
We have spanned over 100 years of claims that the stock market is a leading indicator for the economy. If this is true then we can then surmise that any of the years where the NBER called for a recession, the stock market had already embarked on a meaningful decline and if the data somehow shows a gain in stocks from a peak to trough period then it is because the decline and subsequent recovery was already in place.
Let’s see if the years when the DJIA registered a gain in the period from peak to trough of a recession was already preceded by a decline in the Dow Jones Industrial Average.
Evidence of Market Coincidence preceding Economic Reality
NBER peak |
NBER trough |
DJIA date |
DJIA peak |
DJIA date |
DJIA trough |
% change |
July 1, 1981 |
November 1, 1982 |
4/27/1981 |
1,024.05 |
8/12/1982 |
776.92 |
-24.13% |
April 1, 1960 |
February 1, 1961 |
1/5/1960 |
685.47 |
10/26/1960 |
566.05 |
-17.42% |
October 1, 1926 |
November 1, 1927 |
2/11/1926 |
162.31 |
3/30/1926 |
135.20 |
-16.70% |
August 1, 1918 |
March 1, 1919 |
6/8/1917 |
98.58 |
12/19/1917 |
65.95 |
-33.10% |
July 1, 1953 |
May 1, 1954 |
1/5/1953 |
293.79 |
9/14/1953 |
255.49 |
-13.04% |
January 1, 1980 |
July 1, 1980 |
2/13/1980 |
903.84 |
4/21/1980 |
759.13 |
-16.01% |
February 1, 1945 |
October 1, 1945 |
3/7/1945 |
161.52 |
3/26/1945 |
152.27 |
-5.73% |
July 1, 1990 |
March 1, 1991 |
7/19/1990 |
2,993.81 |
10/11/1990 |
2,387.87 |
-20.24% |
Of the eight periods when there was a positive change in the DJIA within the defined NBER recession, five of them had already experienced a decline and recovery which explains why there was a positive result in our initial review.
The remaining three periods declined after the NBER recession had already started. However, each of the three DJIA troughs occurred before the end of the NBER trough. In this respect, even in failure, the stock market managed to fulfill half of the market bromide. This means that 93% of the dates provided by the NBER since 1900 for both recessions and expansions were led by stock market changes in conformity with the later call in the economy.
Conclusion
In our simple coincidence evaluation, we found that only 17% of the periods did not conform to the idea that stock markets coincide with recessions and expansions. Somehow, all available data suggests that expansions in the economy are perfectly aligned with stock market increases.
When ranked by the length of the recessions, there is a clear majority of recessions that align with declines in the DJIA. However, the minority of recessions that show DJIA gains is somewhat confounding.
However, when we recognize that the stock market is a leading indicator for the economy, we find that the remaining 17% that don’t conform to the theory that the stock market is a leading indicator for the economy shrinks to 6.52% when accounting for the fact that market gains during a recession result from the market having recovered in advance of the recession low.