November 2008
Sun Mon Tue Wed Thu Fri Sat
            1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30            

Authors' Committee

Chair:

Matt Blackwell (Gov)

Members:

Martin Andersen (HealthPol)
Kevin Bartz (Stats)
Deirdre Bloome (Social Policy)
Andy Eggers (Gov)
John Graves (HealthPol)
Rich Nielsen (Gov)
Maya Sen (Gov)
Gary King (Gov)

Weekly Research Workshop Sponsors

Alberto Abadie, Lee Fleming, Adam Glynn, Guido Imbens, Gary King, Arthur Spirling, Jamie Robins, Don Rubin, Chris Winship

Recent Comments

Recent Entries

Categories

Blogroll

Brad DeLong
Cognitive Daily
Complexity & Social Networks
Developing Intelligence
EconLog
The Education Wonks
Empirical Legal Studies
Free Exchange
Freakonomics
Health Care Economist
Junk Charts
Language Log
Law & Econ Prof Blog
Machine Learning (Theory)
Marginal Revolution
Mixing Memory
Mystery Pollster
New Economist
Political Arithmetik
Political Science Methods
Pure Pedantry
Science & Law Blog
Simon Jackman
Social Science++
Statistical modeling, causal inference, and social science

Archives

Notification

Powered by
Movable Type 4.24-en


« November 21, 2008 | Main | December 1, 2008 »

23 November 2008

Which Crash Was Most Like This?

Which past market environment was most like today's? This post shows that in terms of correlation to returns, the closest of all major markets to the S&P 500 in the last year was the Nikkei in 1991.

The S&P 500 peaked in mid-October of 2007, and has been down, down, down since then. Common comparisons include the Nikkei in 1990, the Dow Jones in 1929, and the NASDAQ crash of 2000.

But did those downturns really feel like today's? The plot below shows weekly returns overlaid for these four bear markets. All the plots start at the (then) all-time week-ending high of the market index in question.

They're not really all that similar. The 1929 and 2000 crashes both fell much more dramatically from their all-time highs initially. In contrast, this crash started slowly, gradually accelerating its decline until those dark days of August and September 2008. One depressing fact is that we're actually worse now than we were 58 weeks after the crash of '29!

Are there any precedents for this kind of decline? To answer that I scraped all week-by-week returns for all major indices, U.S. and international, from Yahoo! Finance. Then I programmatically scanned each index's past for a 58-week period where the returns have high correlation with the S&P 500's returns over the past 58 weeks since its October high. The winners, in order of correlation:

  • The Nikkei 225 from March 1991 is the closest, and the big winner. This is interesting because it is the cusp of the Nikkei's so-called ``second wave'' of declines, after it had stabilized following its better-known 1990 crash.
  • Germany's DAX 30 starting September 2000. The European markets all had a more pronounced downturn in the wake of the tech bubble crash than did the U.S. indices besides the NASDAQ.
  • The S&P 500 in May 1969 and in July 1973. Both of these were the crest of major bear markets, the latter owing to an oil crash.

In particular, the ebbs and flows of the Nikkei in 1991 are eerily close to those of the S&P 500 in the last year!

Although no one knows what our future holds, all but one of these indices went on to rally in the following two years. Perhaps it is some consolation that none declined much further, and most gained. The S&P 500 in 1969 recouped all of its losses by two years later. The worst of the bunch was the German DAX 30, which was about flat at its trough after two years.

Posted by Kevin Bartz at 11:22 AM