,
15 tweets,
4 min read
Read on Twitter
The Financial-Entertainment Complex these days seems fixated with the idea of a looming recession (in 2020, apparently). Does this make sense? NO! Here is what MODERN MACRO has to say...
\begin{rant}
\begin{rant}
First, much of the empirical evidence points to economic activity being essentially unpredictable beyond 1-2 quarters ahead. Nowcasting and ST forecasting has actually improved a lot. Beyond that, the best we can hope for is a good density forecast around the right LR mean (2/N)
Financial-Entertainers' view that one could predict business cycles 2 yrs ahead seems to me a variant of their obsession with well defined, alternating, business cycle "phases" that inevitably follow each other: this was state-of the-art macro... in the 1940's! (3/N) 
The modern view of fluctuations is all about the transmission & propagation of unpredictable SHOCKS. The 2020-recession narrative would imply a shock which has happened already, with a propagation mechanism so slow that its effects will only be seen in 2 year's time... (4/N)
Moreover, despite the shock being common knowledge (at least among sell-side luminaries) & despite there being enough time for the usual operational lags of monetary & fiscal policy, no one (gov't, hh'ds, firms) will act or change behaviour to stop it from here to 2020...🤔(5/N)
Compare that narrative to the "Lehman shock", widely believed to be important for the 2008-09 recession. Lehman collapsed mid September. Credit markets froze within days. GDP growth fell down a cliff on the fourth quarter. That's a transmission lag of less than 90 days! (6/N)
Okay, even if you think shocks are unpredictable, you might think certain conditions make us more vulnerable to a yet unknown shock, so p(recession) goes up. A popular story is the length of the expansion itself: "the cycle had a good run, we are due a recession soon"... (7/N)
The literature calls this idea "duration dependence". Key papers by Diebold and Rudebusch find no evidence that expansions die of old age in the Post-WWII period. See also this recent @sffed note: frbsf.org/economic-resea…
Interestingly, there appears to be some evidence of duration dependence in PREWAR business cycles. Another manifestation of the intriguing link between present day market commentary and outdated theories from the 1930s!
Wait, what about the slope of the yield curve? Doesn't that predict a recession in 2020? (10/N)
I am a bit skeptical of this. First the sample is really small (N=11, and some studies don't even use the full postwar period) so conclusions are bound to be non robust. E.g. Two different Fed notes come to different conclusions: frbsf.org/economic-resea… vs federalreserve.gov/econres/notes/…
Second, not all recessions are alike and the economy has changed a great deal from the beggining of the postwar period to now so in fact it is likely N<11. When your theory is based on three data points, you should really not be making bold predictions! (12/N)
Finally, the weak evidence that does exist points to *inversion* of the yield curve being the signal. A positive but declining slope might not necessarily mean anything. Again, the evidence is VERY WEAK! (13/N)
So why is the Financial Entertainment Complex spinning the 2020-Recession narrative? In one word: INCENTIVES. First, it is entertaining and compelling, so they will get more clicks from Short-Attention-Span-Traders. Second, it's their shot at becoming famous if it happens...
...but no-one will remember it if it doesn't! When you don't care about being an honest scientist, that's a one way bet. SAD!
\end{rant}
\end{rant}
