The financial media are working themselves into a lather these days about a phenomenon known as “inversion of the yield curve” and its storied ability to forecast “recessions”. So much so that you’d be forgiven for thinking it was actually important.
You may suspect we’re joking. It’s not that the yield curve is uninformative or that the state of the economy is unimportant. It’s that both of these terms are used to reduce a rich spectrum of colors and shades of gray into a simplistic, binary, black or white.
What exactly is a “recession”? The conventionally accepted meaning is whenever the National Bureau of Economic Research says there is one. How’s that for scientific and objective? Some take it to mean two consecutive quarters of negative economic growth, as if there is some fundamental significance to dates like January 1, April 1, July 1, and October 1. Not to mention that it leaves open the question of just what exactly economic “growth” is, or if it’s even measured accurately.
But just for the sake of argument, let’s overlook these questions and assume that we have a reliable, objective definition of economic growth, a reliable, objective way to measure it, and that “recession” is happening whenever it’s below zero. In which case the mere use of the term obliterates any information whatsoever as to how far below zero it is. Or how far above zero it is not. A barely noticeable a -0.1% rate of growth is classified the same as an all out crash, but deemed different in kind than +0.1%.
A barely distinguishable difference between -0.1% and +0.1% determines whether or not a recession is in effect. Yet the worlds-apart difference between -0.1% and -10% are assigned the same name. And so are +0.1% and +10%. The term draws a completely arbitrary line between very similar states while grouping together very different ones. No such thing would be tolerated in any area of real science.
We have a similar complaint with yield curve “inversion”. A seemingly innocuous statement like “the yield curve is inverted” gives the impression its utterer has said something intelligent or profound. Yet the statement by itself says nothing about which two arbitrarily selected points between thirty days and thirty years are under the microscope or by how much the yield of the former falls short of the latter.
As if it makes no difference. And it still suffers from the same defect as the concept of “recession” … reducing a whole spectrum of information to a simplistic binary variable.
Why does it matter? Because these words are tossed around as if something fundamental about the financial system is changed when the yield curve passes one magic threshold or the rate of growth sits just one side or the other of another. Investors are supposed to invest differently and policymakers are supposed to make policy differently. There is no more difference between -0.1 and +0.1 than there is between +1.9 and +2.1, but one is worthy of headlines and the other is not?
If so, is it not possible that it might lead investors and policymakers to faulty decisions?
Sloppy economics has consequences.
But you may still be wondering if there is anything to this notion, at least broadly, that an area of the yield curve in which rates decrease as the term increases has any concrete implications. Yes it does. It means the bond market expects a decrease in short term interest rates in that time frame. The yield curve represents a kind of running accumulation of the short term rates expected at each point along the curve.
Yet we still haven’t crossed any magic lines. But it does suggest that stock returns will likely be unusually low some months ahead. Say, maybe around the end of this year? But if you’re a Financology reader, you’re already aware of that. Synthetic Systems told us that back at the beginning of January, the October before, and before that. Now the yield curve is just coming around to its point of view.