Sensitivity to a stock market crash
This is the outcome of recent DNB research into the development of stock prices during a stock market crash. Which stocks are hit hardest by such a crash? That is a question to which asset managers would obviously like to know the answer, given their aim to realise the best return at the lowest risk. But is it possible to predict which stocks will have a relatively stable price when a stock market crash occurs?
To answer that question, DNB researchers used price data covering a 40-year period to award a monthly score to more than 2,000 stocks listed in the United States. The score was awarded on the basis of a stock's price movements on days when stock markets suffered severe losses. As the price held up better, a higher score was awarded. This approach was taken on the hypothesis that the better the score, the smaller a stock's price loss in the event of a future stock market crash.
Further analysis revealed that the awarded scores have some predictive value. Our research shows that the prices of the 20% worst scoring stocks in the preceding month on average lost two to three times more in value during a stock market crash than the 20% best scoring stocks. The price losses that occurred in October 2008 are an example. During this month, shortly after the collapse of investment bank Lehman Brothers, with the financial market turmoil at a peak level, nearly 20% of the U.S. stock market value evaporated. Stocks with the lowest scores at the end of September on average lost more than 30% of their value, while stocks with the highest scores on average saw a 10-15% fall.
These scores also show that companies whose stocks performed worse during a stock market crash on average were smaller in size. Stocks with lower scores on average also had higher trading volumes and relatively volatile stock prices.
To calculate the scores, the researchers developed a methodology based on Extreme Value Theory. This methodology focuses on measurement under extremely adverse market conditions. Extremely adverse in the sense that the likelihood of such large stock market losses is relatively small. The aim of the score is to identify the stocks whose prices co-move most with the general market index when a stock market crash occurs.
Our research shows that the scores do provide an indication of stocks that will be hit relatively hard by a future market crash, albeit no more than an indication. A high score does not offer any guarantee that an investment is safe. For example, the stocks with the best scores also dropped significantly in value during a stock market crash. In addition, the volatility of stocks is often on account of other causes, including company-specific aspects. Another complicating factor is that extreme events and the unexpected always go hand in hand. For example, historical relationships that always appeared to hold true in financial markets can be lost on days when the stock exchanges go deep in the red. All this does does not detract from the research conclusion that analysis of major shocks of the past can always be useful in assessing financial risks. The outcome of such an assessment can provide asset managers with an extra compass to set out a course when another crisis occurs.