The Fed put is real, and it's spectacular

It's payroll day today, and while the number of course is very important, somehow it feels a bit less so given that the Fed is almost certainly out of play this month.  Yesterday's ISM was worse than expected and better than feared, judging by the published consensus and the word on the strasse.  Markets seemed uncertain of how to digest things, even though the annualized vehicle sales numbers were a blowout, registering their highest reading in 15 years.

Also released yesterday was a very interesting paper authored by Boston Fed president Eric Rosengren and two others, arguing that the Fed acts as if it had a treble mandate, with financial stability joining inflation and (un)employment as explanatory variables in determining Fed monetary policy.  The authors identified a number of buzzwords that reflect concern over financial stability in Fed meeting transcripts, and then included their frequency of use in a regression to explain changes in Fed policy.

If Macro Man were to summarize the findings of the paper in one sentence, it would be this:  The Fed put is real, and it's spectacular.

Cynics will be unsurprised to see that the Fed's use of "downside" financial stability words like "volatility" and  "bust" occur much more frequently than "bubble" during the 1987-2009 sample period.   Perhaps not coincidentally, the FOMC's use of the buzzwords has trended noticeably higher since the 1990's.

                                                         source: https://www.bostonfed.org/macroprudential2015/papers/Rosengren-Peek-Tootell.pdf

What is somewhat surprising, however, is how much of an impact the buzzwords (or, more accurately, the concerns conveyed by them) have had in explaining Fed policy.  Over the sample period, each 100 buzzwords included in the transcripts equated to a 45 bp move in the Fed funds rate.  Believe it or not, that's a stronger impact than a 1% change in the forecast for inflation or the unemployment rate had over the same period.

Perhaps most telling is the skew in the Fed's reaction function.  The paper broke the sample down into quintiles using credit spreads, with the widest spreads representing a "bust" and the tightest spreads a "boom."  During the "bust periods", each 100 words of moaning about financial stability explained nearly 67 bps of easing, while a similar volume of concern during booms (which we've already established doesn't happen anyways) would only explain 36 bps of tightening.  In fact, the explanatory power of the financial stability buzzwords during "busts" is stronger than any reading for inflation or unemployment, the Fed's ostensible policy targets.  As many have suggested, the Fed not only doesn't take away the punch bowl these days, it spikes it with Everclear.   A simplified summary of the findings is set out below.

Based on these findings, perhaps the appropriate question is not "does the Fed make policy based on financial stability", but "does the Fed make policy based on anything else"?   

Elsewhere, the Atlanta Fed slashed its GDPNow Q3 forecast to just 0.9% yesterday, which raised eyebrows (and more doubts that the FOMC will ever pull the trigger.)  The basis of the downgrade was preliminary goods trade data for August, which showed sharp declines in both exports and imports and a widening of the deficit.   Forecasting the impact of trade on GDP is sketchy at the best of times, using preliminary data even more so.

One counterpoint to the notion that trade collapsed in August is the Long Beach Port data, which showed quite a substantial surge in traffic in that month.  Obviously the port data can be quite sloppy, but as a sense check it certainly doesn't confirm the basis of the GDPNow downgrade or suggest undue cause for concern.  After all, if the economy's going into the tank, why are Americans buying so many cars?

Finally, today's number.  The consensus forecast is rather unimaginative at 201k for the headline number, but Macro Man is hard pressed to argue against it; his own model looks for a rise of 209k.  Ordinarily, one might say that the unemployment would be of interest, given that it's approaching the NAIRU identified by the FOMC in the latest SEP last month.   The again, that NAIRU projection itself has come down by 0.3% since the end of last year, so it's pretty clear that the Fed is willing to lower it as much as necessary based on the spot unemployment rate reading.

Moreover, the table posted above suggests that it doesn't really matter anyways; during the good times, the Fed doesn't give a monkey's about the unemployment rate.   Indeed, based on all available evidence, they don't give a monkey's about anything but keeping the market supplied with (virtual) puts, thus blowing up the next (unmentioned) bubble.  The problem is that when the bubble bursts, that tends to be pretty spectacular, too.