Deutsche Bank Calculates The Level Of The New "Fed Put"
Back in February, in the immediate aftermath of the Feb. 5 volocaust, we quoted a troubling - for stock bulls - statement from outgoing NY Fed president Bil Dudley, who revealed that not only was the recent market correction not "a big drop", but that the Fed's "Powell Put" was far lower.
Specifically, commenting on the February rout, Dudley said an equity rout like the one that occurred in recent days "has virtually no consequence for the economic outlook" and added that, if the market continued to go down sharply, “that would affect my view,” he said at a New York event, but "this wasn't that big of a bump in the stock market" and " is not a big story for central bankers yet."
In other words, the Fed's "Powell put" is well lower from the current S&P level... but how much lower?
That is the question Deutsche Bank's derivatives strategist Aleksandar Kocic, who had been discussing where one can find the new "Fed put" for weeks, believes he has answered in a note released overnight.
Following a lengthy introduction in which he admits that the Fed's new role has basically transformed to be the market's Chief Risk Officer, although using far more words to get to the same point, to wit...
... Kocic then lays out his empirical approach of quantifying the new Fed put, which is another way of stating how high will the Fed raise rates before it relents amid rising market volatility and instability.
Since the way Kocic's thought experiment is structured has just one variable, it is easy to plot various manifestations of short rates vs the S&P. Seen in this light, "stock prices show alignment with the short rate across different cycles (at least, that was the case in the past)."
Specifically, to demonstrate his point, Kocic shows a chart of of 2Y swaps vs. the Log(S&P) levels "to keep things in scale." Here, it becomes obvious that where there "collinearity between the two has been rather compelling in the past", the Fed's QE broke the correlation starting with the financial crisis, and continuing roughly through 2014:
Breaking up the chart above into three distinct time frames, roughly defined as pre-QE and after, reveals that historically short rates vs the S&P scaled appropriately for two Fed cycles: 1999-2000 and 2004-2007. Specifically, Kocic calculates In both cases the beta is about 10 – for a 1% rise in rates, there is a 10% appreciation in S&P.
In striking contrast, the next chart focusing on the post-2014 period, shows a dramatic rise in S&P to 2s beta, roughly 30x. The chart also shows something else: a sharp, post-January divergence between the two. To Kocic this is "the market’s restriking of the Fed put."
Which brings us to the crux of the argument: calculating the actual strike price of the Powell put.
According to Kocic, "one can estimate this strike in two different ways." First, delta neutral, in which a rise in vol is compensated with lowering of the strike in such a way that the underlying is unchanged. Using this approach we find that the S&P drop started at the point when S&P was at 2800 which places the new strike of the Fed put somewhere in the 2300 — 2400 range.
Here, as Kocic explains, the "Fed put is embedded in the beta, which represents response of stocks to rates rise. Higher strike, i.e. lower deductible, implies a more protective Fed; lower beta corresponds to a higher deductible or a lower strike of the put." And as shown above, where previously the beta was 10, it has since jumped to 30, meaning that the Fed remains highly protective (as Bank of America repeatedly showed before). According to Kocic, "this is a residual of the Fed's awareness of the distortions caused by QE as well as the market's vulnerability to stimulus withdrawal."
In quantitative terms, this would sugges that using the old beta of 10, the SPX should have been around 2300. To the Deutsche analyst this means that the current restriking of the Fed put is not very far from what the "old" striking would imply.
In simple English, what all of the above means stated simply is that the S&P has no less than roughly 300 points of downside from here before the Fed even bothers to intervene; it also means that if stocks demand Fed intervention as they often have in the past, the S&P will have to drop to the revised Fed put level of 2300-2400 before the requested response is triggered.
* * *
But wait there's more, because as Kocic further explains, it is very unlikely that the market will sell off in a calm, cool and collected manner from its current level to 2,300; in fact, the drop would likely be far more stormy as a result of unwinding convexity flows, which push investors out of equities and into bonds. His explanation below:
Restriking of the Fed put is a withdrawal of convexity from equities. It is effectively a removal of a put spread from the market. However, in the environment where everything is bound to sell off (a market mode that is a mirror image of QE), volatility is one of the key decision variables. More volatile equities are less desirable than less volatile duration. In that environment, convexity withdrawal creates a reinforcing loop where more turbulence in risk assets tends to cause stability in fixed income. The figure shows the convexity flows across the two markets.
The implication of these cascading convexity flows is that as equities tumble, there would be an outsized bid for all fixed income instruments:
However, even here there is a peculiar paradox, because as flows flood into fixed income, they create the kinds of curve vol distortion and migration we discussed last week in "The Key Story For The Bond Market Is About To Play Out", a process which as Kocic explains once again, makes rate hikes less effective at a parallel shift in the curve, instead assuring curve flattening and subsequent inversion.
Send this article to a friend: