Real stock market returns reveal the true frequency of ‘once in a century' crashes,
says Paul Kaplan

When former Federal Reserve chairman Alan Greenspan characterised the financial crisis of 2008 as a "once in a century credit tsunami", I was stunned. Being familiar with long term data on the US capital markets, I thought a more apropos statement was the one made in the CFA Institute magazine by Leslie Rahl (founder of Capital Market Risk Advisors) more than a year before the crisis: "We seem to have a once in a lifetime crisis every three or four years."

In a 2009 article for Morningstar, ‘Déjà Vu All Over Again', I illustrated the frequency and severity of the major drawdown for various countries using time series of stock market total returns. The results clearly demonstrated that Greenspan was in need of a history lesson. Later I expanded the analysis at the request of Laurence Siegel, director of research at the CFA Institute, as a contribution to his book, Insights into the Global Financial Crisis. Larry asked me to use monthly real total returns - that is, returns that include the reinvestment of dividends and are adjusted for inflation - and to go back into history as far as it was possible with reasonably reliable data.

To complete the study, I needed to find monthly data from before 1925 on both stock returns and inflation, and calculate real returns. Since there was no such return series in existence, I would have to create one out of readable available data. Professor Robert Shiller of Yale posts a monthly history of US stock market returns and inflation on his website that goes back to 1871.

Unfortunately, Shiller's stock data is based on monthly average prices rather than month- end prices. So I could use his inflation data, but not his stock market data. Separately, Roger Ibbotson and some colleagues created, for the December 2000 Journal of Financial Markets, an annual price and total return series for the NYSE that goes back to 1825. However, annual returns are at too low a frequency to measure the largest drawdowns of the period, such as the large drop in the stock market during the panic of 1907.

Fortunately, the Phyllis Pierce-edited Dow Jones Averages 1885-1980 (published in 1982) provides daily price data on that index; Larry Siegel advised me to estimate the monthly price returns in the broader NYSE price index from the monthly price returns on the Dow Jones Averages and then interpolate the total returns by assuming that the level dividends remained constant during each year.

Following Larry's advice, and soliciting the help of Morningstar intern Kailin Liu, I produced a time series of real total returns for the US stock markets that runs from 1871 through the present. While for the first 15 years we only have annual returns, we now have more than 123 years of monthly total real returns. This data now appears in the Ibbotson SBBI Yearbooks.

Over the entire 138-and-a-half-year period, the Real US Stock Market index grew from $1 to $5,179 in 1869 dollars. This is a compound annual real total of just under 6.4%, almost the same as the post 1925 period. However, figure 1 - which shows the growth of $1 invested in the US stock market at the end of 1869 through June 2009 in real terms, along with a line that shows the highest level that the index had achieved as of that date - reveals that it was a very bumpy ride with a number of major drawdowns, some of which can be linked with specific economic and political events.

Figure 2 lists all of the drawdowns that exceeded 20%. In total, there were 17 such declines, including the present one, from which we have yet to recover. Not surprisingly, the granddaddy of all market declines started just before the crash of 1929 and did not recover until towards the end of 1936. The US stock market lost 79% of its real value in less than three years, and it took more than five years to recover. What may be more sobering, however, is that not only are we currently in the second greatest decline, but it started nine years ago. The combined effect of the crash of the internet bubble in 2000 and the financial crisis of 2008 caused the US stock market to lose 54% of its real value from August 2000 to February 2009.

Who knows how long it will take to recover from that and when our next crisis will occur? The history of stock market drawdowns presented here shows that investing in stocks can be very risky business, indeed, and that the current crisis is hardly a ‘once in a century' event. But to more than just state the obvious, we should use this data to better gauge the potential risks and long term rewards of investing in risky assets such as stocks. Specifically, we should supplement our traditional measures of risk, such as standard deviation, which relies on a normal distribution, by measures that better capture the fat tailed nature of the historical returns and drawdowns as presented here.

Paul Kaplan is quantitative research director at Morningstar Europe