Dec

14

One definition of volatility is stdev of daily returns over a period. Using DJIA daily closes since 1929, partitioned the series into non-overlapping 10 trading-day periods. At the end of each 10D, calculated stdev of that period.

The attached graph plots 10D stdev over the whole series, with the familiar spikes corresponding to other volatility measures such as VIX and VXO. Note historically there are relatively few spikes with SD10 0.03, but when in the past they occurred (late 1920s/early 30s, late 30s, and early 2000s), there were other large spikes which followed. The notable exception is 1987, and of course the present because:

1. The future hasn't happened yet
2. The late Dr. Samuelson won the debate with the late Dr. Friedman, at least within our current Keynesian government

In attempt to quantify waits to next volatility spike after large spikes, defined a large spike as: (a maximum at the center of 17 - 10D periods) AND (>0.03). (a major spike which was the highest for 80 days in the past and 80 days in the future). The recent large volatility spike which peaked Oct 2008 was 0.057, 29 10D periods ago.

Once spikes >0.03 were identified, checked wait time until another spike >0.02 occurred. Here are the instances and the waits:

Date >0.03 nxt spike
10/17/08 0.057 ??
09/24/01 0.031 21
09/14/98 0.034 38
10/28/97 0.031 22
10/27/87 0.089 50
09/08/39 0.030 18
10/22/37 0.034 11
07/28/33 0.045 25
03/21/33 0.060 9
08/16/32 0.051 14
02/25/32 0.047 12
10/14/31 0.071 9
06/23/30 0.036 13
11/14/29 0.075 15

mean 19.8

Note the current wait of 29 10D periods is third longest, after 38 (1998) and 50 (1987).

So here we are, waiting for Godot.


Comments

WordPress database error: [Table './dailyspeculations_com_@002d_dailywordpress/wp_comments' is marked as crashed and last (automatic?) repair failed]
SELECT * FROM wp_comments WHERE comment_post_ID = '4219' AND comment_approved = '1' ORDER BY comment_date

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search