This weekend I was watching Bloomberg and some obviously sophisticated money manager or private equity fund manager was on there selling the merits of his fund (or as the media puts it “providing us insight into the global economy”) and was asked about Treasury Yields and what they tell us about oncoming recession. He said, well throughout history when the yield curve has inverted, the market has done quite well. Oh yeah throughout history? What did the yield curve do when Michaelangelo was not a Ninja Turtle? Or how about when Marie Antoinette’s head was still attached to her neck? The fact is, modern economic history is quite new, and when people refer to it they are essentially referring to a period of time less than a century. Since business cycles tend to last 8 years or so, this gives us about 12 to choose from when referencing – not a population you want to be forming statistics on with any degree of confidence. Add into that the evolution of Monetary System, Bretton Woods II, technology, globalization, hedge funds, derivatives, and Soros’ theory of reflexivity, and historical accounts become meaningless – and the only bet worth making is that this recession will be anything LIKE the last one – or that it will even occur.
Back to my buddy on Bloomberg. According to an Article by Richard Shaw, An 80 Year Yield Curve History and its Implications,
“There were only 8 years (10% of the 80 years) when the average weekly yield curve status was inverted for the year: 1927, 1928, 1929, 1966, 1969, 1980, and 1981.
There were 17 years (21% of the 80 years) when there was at least one week of inverted yield curve: 1927, 1928, 1929 1930, 1959, 1966, 1967, 1968, 1969, 1970, 1973, 1974, 1979, 1980, 1981, 2000, and 2006 (but 9 of those 17 years had a positive average for the year).
Most of those inverted years were associated with tough times in the economy. It is generally held that a prolonged inverted yield curve produces a subsequent recession.”
There were only 8 years where a yield curve has been inverted for the year, and 17 when it was inverted for a portion of the year (This article is from January 2007). These are both very small sample sizes and about half (9 of 17) of the occurrences resulted in an up year and half in a down year – essentially telling us very little . So yes the my Bloombergian friend was write, in years that they invert, the market USUALLY goes up, but if I told you that 9 out of 17 people who use Brand X deodorant don’t drop dead after using it, would you go ahead and apply it liberally? Probably not, you would probably go around smelling the armpits of the corpses piling up on the street until you had enough of a sample. The comment interpretation that still conveys useful information should have been, yield curves can invert for a period of years where the market may go up or go down, but usually comes during great inflection points for the market. (but even that is up for debate on its statistical usefulness).
This “used car salesman with a corner office” suffers from what I call economic recollection. This is the act of data mining through the small sample of historical economic data we actually have (and the much larger pool of explanations for changes in that data) in attempt to find a norm. The unstable norm or mean is usually skewed further by recollection of the most recent examples – rendering the conclusion even more useless - to the point of being dangerous.
Recession recollections are accompanied with large amounts of bullish*t, and is uttered by the same person who said, “this is a new global economy” last week on the show. An example of other reminiscing lately have been, “P/E’s are at historical lows (or highs)”, “Corporate Spreads are historically low”, “Economies are usually strong for a second term President”, and many others I am sure. The fact is, these are usually meaningless statistics, and “meaningless” soon becomes “dangerous” due to the bias that people put on the more recent times – those that they were alive for.
The danger is that these “recollections” are appealing to most people, and invoked during times of great uncertainty (a recession recollection fallacy myself – See how easy it is!!) when people are looking for something to hold on to. Think of a bunch of Groomsmen sitting around before a Wedding talking about the good old days, when the groom used to “sleep with a new girl every weekend”. Chances are it was more like every couple months on average, refers to a 5 – 10 year period of time and reflects a universe of about 10 -15. Even if it did include a month where the guy had sex with a new girl each weekend, this was not the norm. However, his fear and trepidation at that moment will cause him to believe that he did that, or at least could have, and possibly be quoted in an post-honeymoon argument down the line – resulting in him not having sex with anyone for some period of time.
Bottom line, if presented with a small samples of economic data tied to recessions or market cycles, look at the hard data. The sample size, distribution, the subset of recent observations, and most importantly the relevance of the data. Most likely, this data is just an economic recollection that we choose to hold on to because we are uncertain or uncomfortable about what may be next.
Comments