dissecting that scary 1929 market chart

Scary 1929 market chart gains traction

{ disclosure : borrowed heading from Mark Hulbert from the Scary 1929 market chart gains traction }

here is that Scary 1929 market chart , for your eye-candy purposes , that is making headlines and doing rounds across the trading desks

Scary 1929 market chart

Below is what we understood by reading the Scary 1929 market chart gains traction 

Exercise 1 ) 

“The chart was first publicized in late November of last year” -> so we took 29-Nov-2013 weekly closing value of DJIA as the base

“this chart of the DJIA (DIA) of the last 18 months superimposed on a chart of the index during the same time frame in 1928-1929” from ( Scary chart draws parallels with 1929) -> so we are slicing DJIA closing values into 18 month non-overlapping periods ( 78 weekly closing value percent changes, so we will get about 75 observations )

Step 1) download DJIA closing weekly values  from http://research.stlouisfed.org/fred2/series/DJIA/downloaddata?cid=32255

Step 2) Calculate the correlation between a) 78 weekly percent changes of DJIA with the period from 29-Nov-13 to 08-Jun-12 and b) the previous 78 weeks starting from 01-Jun-12 to 10-Dec-10

Step 3) repeat this process for 75 times , till Jan 1900 , with each 18 month period to that of the DeMark  Eureka/Benzene  moment 🙂 i.e late November 2013 , 18 month period .

below the table we got , for your convenience we have also given the next 18 month returns ..  and ranked the correlation to that of current 78 weeks period

Start End Correl Rank DJIA Fwd 18 months DJIA returns %
29-Nov-13 08-Jun-12 1.00 15536.18 ??
10-Dec-10 01-Jun-12 -0.05 52 12331.87 30.41
12-Jun-09 03-Dec-10 0.00 37 11374.84 8.41
14-Dec-07 05-Jun-09 0.09 20 8767.35 29.74
16-Jun-06 07-Dec-07 0.09 19 13498.3 -35.05
17-Dec-04 09-Jun-06 -0.08 57 10869.08 24.19
20-Jun-03 10-Dec-04 -0.13 63 10672.36 1.84
21-Dec-01 13-Jun-03 -0.05 51 9263.21 15.21
23-Jun-00 14-Dec-01 0.14 7 9996.27 -7.33
25-Dec-98 16-Jun-00 -0.06 54 10454.33 -4.38
27-Jun-97 18-Dec-98 -0.02 43 9113.33 14.71
29-Dec-95 20-Jun-97 -0.15 66 7678.85 18.68
01-Jul-94 22-Dec-95 0.03 28 5105.41 50.41
01-Jan-93 24-Jun-94 -0.01 41 3658.76 39.54
05-Jul-91 25-Dec-92 0.12 10 3316.58 10.32
05-Jan-90 28-Jun-91 0.01 31 2949.58 12.44
08-Jul-88 29-Dec-89 0.09 18 2797.3 5.44
09-Jan-87 01-Jul-88 -0.02 44 2129.4 31.37
12-Jul-85 02-Jan-87 0.06 24 1989.65 7.02
13-Jan-84 05-Jul-85 0.00 35 1331.9 49.38
16-Jul-82 06-Jan-84 -0.03 48 1278.29 4.19
16-Jan-81 09-Jul-82 -0.19 70 826.69 54.63
20-Jul-79 09-Jan-81 -0.12 61 968.72 -14.66
20-Jan-78 13-Jul-79 -0.05 50 829.47 16.79
23-Jul-76 13-Jan-78 -0.06 55 778.53 6.54
24-Jan-75 16-Jul-76 0.10 14 990.11 -21.37
27-Jul-73 17-Jan-75 -0.08 56 653.07 51.61
28-Jan-72 20-Jul-73 0.01 32 927.23 -29.57
31-Jul-70 21-Jan-72 0.08 22 897.38 3.33
31-Jan-69 24-Jul-70 0.10 13 733.19 22.39
04-Aug-67 24-Jan-69 -0.12 60 940.43 -22.04
04-Feb-66 28-Jul-67 -0.01 40 917.05 2.55
07-Aug-64 28-Jan-66 -0.03 46 981.85 -6.60
08-Feb-63 31-Jul-64 0.28 3 831.75 18.05
11-Aug-61 01-Feb-63 0.10 15 680.97 22.14
12-Feb-60 04-Aug-61 0.14 8 720.09 -5.43
15-Aug-58 05-Feb-60 -0.26 74 622.41 15.69
15-Feb-57 08-Aug-58 -0.21 71 509.25 22.22
19-Aug-55 08-Feb-57 -0.18 68 460.81 10.51
19-Feb-54 12-Aug-55 0.04 26 453.66 1.58
22-Aug-52 12-Feb-54 0.00 36 290.97 55.91
23-Feb-51 15-Aug-52 0.02 29 274.34 6.06
26-Aug-49 16-Feb-51 -0.03 47 252.39 8.70
27-Feb-48 19-Aug-49 0.09 21 179.21 40.83
30-Aug-46 20-Feb-48 -0.14 64 167.63 6.91
02-Mar-45 23-Aug-46 0.10 16 191.54 -12.48
03-Sep-43 23-Feb-45 0.13 9 159.58 20.03
06-Mar-42 27-Aug-43 -0.11 59 136.59 16.83
06-Sep-40 27-Feb-42 -0.02 45 105.36 29.64
10-Mar-39 30-Aug-40 0.28 4 131.71 -20.01
10-Sep-37 03-Mar-39 0.15 6 150.46 -12.46
13-Mar-36 03-Sep-37 0.01 34 164.93 -8.77
14-Sep-34 06-Mar-36 0.10 17 154.52 6.74
17-Mar-33 07-Sep-34 -0.17 67 89.21 73.21
11-Sep-31 03-Mar-33 -0.10 58 61.93 44.05
14-Mar-30 04-Sep-31 0.02 30 128.29 -51.73
14-Sep-28 07-Mar-30 0.00 39 274.25 -53.22
18-Mar-27 07-Sep-28 -0.21 72 240.38 14.09
18-Sep-25 11-Mar-27 -0.14 65 161.04 49.27
21-Mar-24 11-Sep-25 -0.01 42 145.99 10.31
22-Sep-22 14-Mar-24 -0.12 62 96.66 51.03
25-Mar-21 15-Sep-22 0.07 23 99.38 -2.74
26-Sep-19 18-Mar-21 0.23 5 76.87 29.28
29-Mar-18 19-Sep-19 -0.06 53 106.83 -28.04
29-Sep-16 22-Mar-18 0.12 11 76.46 39.72
02-Apr-15 22-Sep-16 0.00 38 102.05 -25.08
23-May-13 26-Mar-15 0.29 2 60.58 68.45
24-Nov-11 16-May-13 0.05 25 79.09 -23.40
27-May-10 17-Nov-11 0.01 33 81.35 -2.78
27-Nov-08 20-May-10 0.11 12 88.69 -8.28
31-May-07 20-Nov-08 0.33 1 86.95 2.00
01-Dec-05 24-May-07 -0.03 49 78 11.47
03-Jun-04 24-Nov-05 -0.22 73 89.13 -12.49
05-Dec-02 27-May-04 -0.19 69 48.18 84.99
07-Jun-01 28-Nov-02 0.04 27 62.08 -22.39

points to note are

1) highest correlation value is o.33 ( highlighted in green ) with that of period starting from 31st May 1907 to 20th Nov 1908 , when DJIA returned 2% over the next 18 months.

2) the correlation values to those subsequent 50% crashes ( highlighted in red) over the 18 months period are mere 0.00 and 0.02

3) rule of thumb -> if n*abs(correl)>10 , that can be considered something to serious , investigate further.

in this case, n being 75 , (sample size ), DeMark moment’s, correl being 0.02 , and 74*0.04 being 1.5 , that analogue statistically doesn’t make sense to us 

Exercise 2 ) 

assuming this chart is gaining more traction now , as the original purpose was fun , but now it is serious,  as quoted here ->

“Originally, I drew [the chart] for entertainment purposes only … Now it’s evolved into something more serious.” from Scary chart draws parallels with 1929 

doing the same steps mentioned above but calculating the 78 weekly change % correlations with the period starting from 07th Feb, 2014 to 17th Aug, 2012 , we got the below table

Start End Correl Rank DJIA Fwd 18 month DJIA returns %
07-Feb-14 17-Aug-12 1.00 16082.27 ??
18-Feb-11 10-Aug-12 -0.11 59 13206.33 17.64
21-Aug-09 11-Feb-11 0.06 21 12298.48 7.38
22-Feb-08 14-Aug-09 0.03 28 9297.69 32.27
25-Aug-06 15-Feb-08 0.06 20 12357.45 -24.76
25-Feb-05 18-Aug-06 -0.11 58 11314.26 9.22
29-Aug-03 18-Feb-05 -0.19 69 10718.85 5.55
01-Mar-02 22-Aug-03 -0.08 53 9356.38 14.56
01-Sep-00 22-Feb-02 0.08 18 10172.71 -8.02
05-Mar-99 25-Aug-00 -0.08 52 11204.97 -9.21
05-Sep-97 26-Feb-99 -0.05 44 9420.35 18.94
08-Mar-96 29-Aug-97 -0.03 42 7817.29 20.51
09-Sep-94 01-Mar-96 -0.02 41 5596.89 39.67
12-Mar-93 02-Sep-94 -0.07 49 3890.76 43.85
13-Sep-91 05-Mar-93 0.05 23 3460.94 12.42
16-Mar-90 06-Sep-91 0.02 33 2994.05 15.59
16-Sep-88 09-Mar-90 -0.10 54 2697.21 11.01
20-Mar-87 09-Sep-88 -0.07 50 2089.3 29.10
20-Sep-85 13-Mar-87 0.04 26 2290.65 -8.79
23-Mar-84 13-Sep-85 0.08 17 1302.49 75.87
24-Sep-82 16-Mar-84 0.01 38 1165.74 11.73
27-Mar-81 17-Sep-82 -0.31 75 924.8 26.05
28-Sep-79 20-Mar-81 -0.19 68 1003.22 -7.82
31-Mar-78 21-Sep-79 0.01 35 884.88 13.37
01-Oct-76 24-Mar-78 -0.16 63 758.16 16.71
04-Apr-75 24-Sep-76 -0.01 40 993.87 -23.72
05-Oct-73 28-Mar-75 -0.10 57 757.95 31.13
07-Apr-72 28-Sep-73 -0.17 64 959.47 -21.00
09-Oct-70 31-Mar-72 0.10 16 952.18 0.77
11-Apr-69 02-Oct-70 0.10 14 777.71 22.43
13-Oct-67 04-Apr-69 -0.19 66 927.65 -16.16
15-Apr-66 06-Oct-67 -0.08 51 922.31 0.58
16-Oct-64 08-Apr-66 -0.13 61 942.25 -2.12
19-Apr-63 09-Oct-64 0.23 5 874.19 7.79
20-Oct-61 12-Apr-63 0.00 39 710.48 23.04
22-Apr-60 13-Oct-61 0.16 9 703.96 0.93
24-Oct-58 15-Apr-60 -0.26 73 622.27 13.13
26-Apr-57 17-Oct-58 -0.26 74 542.09 14.79
28-Oct-55 19-Apr-57 -0.18 65 491.62 10.27
30-Apr-54 21-Oct-55 0.14 10 456.71 7.64
31-Oct-52 23-Apr-54 0.03 31 315.87 44.59
04-May-51 24-Oct-52 0.02 32 266.41 18.57
04-Nov-49 27-Apr-51 -0.06 46 261.02 2.06
07-May-48 28-Oct-49 0.03 30 191.26 36.47
08-Nov-46 30-Apr-48 -0.06 47 181.28 5.51
11-May-45 01-Nov-46 0.17 8 171.24 5.86
12-Nov-43 04-May-45 0.03 29 165.2 3.66
15-May-42 05-Nov-43 0.05 24 132.72 24.47
15-Nov-40 08-May-42 0.01 34 98.13 35.25
19-May-39 08-Nov-40 0.28 3 136.94 -28.34
19-Nov-37 12-May-39 0.11 13 130.64 4.82
22-May-36 12-Nov-37 0.04 27 126.9 2.95
23-Nov-34 15-May-36 0.11 12 149.43 -15.08
26-May-33 16-Nov-34 -0.16 62 100.04 49.37
20-Nov-31 19-May-33 -0.20 70 82.94 20.62
23-May-30 13-Nov-31 -0.10 56 103.03 -19.50
23-Nov-28 16-May-30 0.04 25 267.97 -61.55
27-May-27 16-Nov-28 -0.24 72 283.21 -5.38
27-Nov-25 20-May-27 -0.12 60 171.63 65.01
30-May-24 20-Nov-25 0.10 15 150.32 14.18
01-Dec-22 23-May-24 -0.04 43 90.24 66.58
03-Jun-21 24-Nov-22 0.28 2 93.81 -3.81
05-Dec-19 27-May-21 0.19 7 73.1 28.33
07-Jun-18 28-Nov-19 0.05 22 105.53 -30.73
08-Dec-16 31-May-18 0.12 11 78.85 33.84
11-Jun-15 01-Dec-16 -0.06 48 106.52 -25.98
01-Aug-13 04-Jun-15 0.23 4 68.92 54.56
02-Feb-12 25-Jul-13 -0.10 55 78.62 -12.34
05-Aug-10 26-Jan-12 0.08 19 80.84 -2.75
05-Feb-09 29-Jul-10 0.21 6 77.01 4.97
09-Aug-07 29-Jan-09 0.31 1 84.63 -9.00
09-Feb-06 02-Aug-07 -0.05 45 76.23 11.02
12-Aug-04 02-Feb-06 -0.20 71 99.87 -23.67
13-Feb-03 05-Aug-04 -0.19 67 53.08 88.15
16-Aug-01 06-Feb-03 0.01 36 66.98 -20.75
16-Feb-00 09-Aug-01 0.01 37 70.6 -5.13

points to note are

1) highest correlation value is o.31 ( highlighted in green ) with that of period starting from 9 Aug 1907 to 29th Jan 1909 , when DJIA fallen by 9% , over the next 18 months.

2) the correlation values with current 18 months period and to to that of DeMark Benzene-Eureka movement ( highlighted in red where DJIA crashed 60% over subsequent 18 months period ) is a mere 0.04.

3) re-read rule of thumb , written above pls ..

charts are from DeMark , Correlations are calculated on my excel sheets , wallet is yours !!! 

you might want to consider subscribing to our daily quant rants email  !!

check out Quant-Ideas  to aid the short term trader to find high probability winning trades

have you bought the Anatomy of $SPY on First Trading Day of the Month  , written the co-founder of this site, on Amazon yet ?

3 thoughts on “dissecting that scary 1929 market chart

  1. Adam Reply

    Fun post. Was thinking you should consider flipping the return sign for any observation where observed correlation is negative. After all, it seems these periods saw an inverse relationship… Thanks

  2. Pingback: Daily Quant Ideas from Paststat » Blog Archive $SPY Returns around Presidents Day

  3. Pingback: Daily Quant Ideas from Paststat » Blog Archive does it look like 2002 ?

Leave a Reply