The Devil is in the (Debt) Details
One of the more overrated developments of the digital revolution is the concept of “big data” or “data mining.” In the last decade, computing power and easy access to inconceivable amounts of data have combined to create the ability to find meaning in vast amounts of correlations, interpolations, extrapolations, and so forth.
According to economist and scholar Dr. Horace Brock of SED, Inc., the tendency toward induction, or using data to project future trends, is a major flaw in today’s research. An example of induction (using oversimplified variables and hypothetical values) might be as follows:
“In the past 100 years, we never went into recession within 12 months of having 5% unemployment, therefore, because unemployment was just 5 percent in December, our chances of recession for 2016 are extremely low.”
I might point out here that the current Fed Chairman, Janet Yellen, is a big fan of “data.” I also would like to point out that pure data analysis is backward-looking and often makes assumptions about correlation and causality that may not apply “the next time.”
Dr Brock insists that “deduction” leads to more robust conclusions: if you can establish the relationships between the variables, you can best predict the outcome. I will give another example using hypothetical variables and values, though the key here is to note how the process differs:
“Because the risk of a downward economic shock rises and falls directly with the business and financial leverage that was used to produce the economic output, and because the growth in China, in emerging markets, and in the US energy industry was produced with combined operating and financial leverage of 80%, the risk of an economic shock stemming from those sectors is 80% higher than it would have been had they been funded with no debt.”
[pullquote align=”full” cite=”” link=”” color=”” class=”” size=””]Having seen how debt acted as the wild card in exaggerating the integrity of the economic growth and later magnifying the effects of the last two recessions, I am more likely to dismiss any prognosis for the future that does not include these lessons from the past.[/pullquote]
I believe that the preceding deductive example is much more useful in trying to forecast future global economic events than “data mining” analysis. While I do not have access (yet) to the numbers I would need in order to assess the amounts and kinds of leverage (and thus the risk), I can remember a powerful story I heard about 2001—after the tech markets had imploded. A former Lehman credit analyst (if I remember that part correctly) saw that the telecom sector had been borrowing insane amounts of money at increasingly high rates of interest; in the last round of borrowing, many paid close to 20%. They were hoping to build cellular and data networks, then sell them to the larger operators. But this borrowing, possibly as much as 500 billion back then, created the effect of a powerful new secular economic expansion. This created a rush for other firms to expand capacity, and it led to growth, jobs and a massive bubble in technology shares. When the funding ended and the demand petered out, firms were “all dressed up with nowhere to go.” By 2002, twenty-three telecommunications companies had bankrupted, the largest being Worldcom. Many “dotcoms” went out of business. The ensuing capital-spending recession (almost no investment spending by companies) paved the way for next round of super-stimulus and the housing bubble.
Of course, the example of the 2008 housing bust is similar to the tech bubble in that the debt created an “apparent” and unsustainable growth rate, but no one would have been able to assess the probability of either of those outcomes using pure data—or at least, none that I know of who did not also formulate their forecast with deductive methods.
My point and premise regarding the economic and stockmarket risk today is that most of the analysis today does not consider the heightened risk that comes with leverage—both operating leverage (firms with high fixed costs relative to profits have higher operating leverage) and financial leverage (firms who fund their operations with much debt have high financial leverage, i.e. they have high debt/equity ratios).
I submit that while we may not enter recession this year, the data-happy analysis of the times does not properly consider the effects of debt (leverage) on both the chances or the magnitude of a recession. Having seen how debt acted as the wild card in exaggerating the integrity of the economic growth and later magnifying the effects of the last two recessions, I am more likely to dismiss any prognosis for the future that does not include these lessons from the past.
As Warren Buffet famously quipped, “You never know who was swimming naked until the tide goes out.” His use of “swimming naked” is a good metaphor for “using too much leverage.”