Stats › Bear markets and drawdowns › Correlation in crashes
Correct crash correlations for volatility and the contagion disappears in 15 of 16 cases
| Market, 1997 Asian crisis | Unadjusted correlation | t | Adjusted | t |
|---|---|---|---|---|
| Germany | 0.902 | 4.60 | 0.642 | 0.97 |
| France | 0.886 | 5.19 | 0.596 | 1.35 |
| Australia | 0.865 | 3.59 | 0.561 | 0.57 |
| Italy | 0.896 | 6.05 | 0.618 | 1.79 |
| Indonesia | 0.749 | 1.75 | 0.399 | -0.10 |
| Netherlands | 0.742 | 2.08 | 0.397 | 0.18 |
| Philippines | 0.705 | 1.83 | 0.388 | 0.22 |
| Korea | 0.683 | 2.30 | 0.380 | 0.69 |
Correlation with Hong Kong during the turmoil window of 17 October to 16 November 1997, tested against the full period correlation at a 5 per cent critical value of 1.65. Stability window 1 January 1996 to 16 October 1997. Italy is the only market whose correlation increase survives the volatility adjustment, and it survives at 1.79 against a critical value of 1.65.
Measured correlations rise in crises, but higher volatility in the source market biases the measurement upward. Adjust for that bias and crisis correlations stop being statistically distinguishable from normal period correlations in almost every market tested across three crashes.
Members only
The rest of this page is for members
Below this point there are 5 sections, 1 table and 3 named sources, roughly 700 words of it. Every figure carries the source it came from and the date the data is from.
You can keep browsing every statistic in the library for free. The intro and the summary are always open.