In the past week Seeking Alpha readers have been treated to an article and a rebuttal about a supposed "scary graph" showing suspicious similarities between the Dow Jones Industrials Chart in 1928-29, and 2013-14. For my readers' amusement the chart is shown below.
Just a brief aside: I taught college statistics from 1980 to 2002 in the US, Asia, Russia, and Europe for various academic institutions as part of their economics departments. Early in the first chapter of every Stat textbook I used (usually page four or five -- not to rub it in) was a caution about spurious correlations. Examples were legion, and included:
- correlation between cola consumption and the incidence of polio.
- correlation between the % of blacks in a community and the crime rate.
- correlation between full moons and robberies.
- and, lest I forget, correlations between hemline length and the level of the Dow Jones average.
After putting my students on guard for such statistical ignorance -- which is the proper word for it, because the damage caused by such spurious analysis is incalculable -- I cheer them up by pointing out
- you now are in possession of more statistical expertise than 90% of the analysts on Wall Street.
They would chuckle and not believe me.
They can believe me now! IS the hemline graph more, or less sophisticated than the chart at the top of my article? The answer? It is more sophisticated. Why?
- first it covers a far longer time period, which makes it less likely to be spurious.
- second, the hemline chart is logarithmic, which is the only scale to use for long term market comparisons.
- the hemline index has some theoretical justification (hemlines are shorter when society is frisky and optimistic). The earlier chart comparison is purely conjectural and curve fitting.
How does such a chart, even if posted in jest, cause the "incalculable damage" I referred to a couple paragraphs ago? Simple: it distracts investors from any true similarities between the two periods, which might be far more informative and ominous.
Is there such a similarity? Definitely.
- The US economy slips into recession. The stock market takes a tumble ... interest rates fall as the FED quickly trims the discount rate in the hope of cushioning the business cycle ... the low interest rate environment sets off a massive wave of construction and an asset bubble in real estate ... bank balance sheets and household savings become dependent on the profound mispricing of real estate and other equity holdings ... the housing bubble implodes, pushing the economy into a long, deep recession. This is the economic story of the last decade ... and the 1920s.
As Evan Soltas makes clear in this Bloomberg article last fall, this is the dreaded similarity between the 1920s and the last few years.
I might add that the FED also kept interest rates low in the 1920s to prevent draining capital from shaky European economies: exactly the complaint some analysts made recently when the Fed "ignored foreign central banks concerns" (e.g. Turkey) when they decided to continue the taper.
Since markets have come to expect the FED to continue the taper, stock prices have rallied strongly this month. It appears the FED has realized its primary concern should be domestic price stability and a strong financial system -- not unemployment, or bankers overseas (both topics for another article).
In the meantime ladies, trim those hemlines! Wall Street analysts are counting on you.