SocGen: ‘We Cringe’ Whenever We See That Chart Of The VIX And The Yield Curve

Over the past six or so months, a handful of bloggers and some members of the finance Twitterati (and those are not mutually exclusive) have periodically rolled out a lagged chart of the VIX plotted with the 2s10s in an effort to suggest something or other about an imminent spike in volatility “caused” by inversions or near inversions in the yield curve.

It’s a testament to something (nothing good) that such a chart crime was able to get any traction whatsoever, but it did. In fact, it was cited by at least two analysts that I know of after being initially passed around by the bearish crowd and before long, otherwise reputable Twitter accounts were tweeting it out themselves, thus perpetuating things.

The problem, as we gently pointed out in December, is that you can’t really take a 3-year lead slope (which is itself an unreliable and inconsistent recession predictor), slap it on top of the VIX, (a 1-month option), and then claim to have discovered something profound about conflicting narratives. Even if you neglect the obvious fact that the two series have nothing to do with each other (R-squared cannot possibly be materially different from zero), there is no logical connection there – at all.

Of course that kind of reasoning doesn’t matter in many circles and the general investing public is more than happy to dance around like a bunch of circus monkeys if you wave something ostensibly scary in front of their faces.

Well, five months after we initially set about trying to dispense with this silliness, SocGen is out with a note that finds the bank literally “cringing” at the same chart we described above.

“In recent months we have seen different versions of the chart below doing the rounds [and] it is being used to argue that equity volatility will rise in the future because the yield curve is now flattening”, the bank writes, adding that “we cringe every time we hear this ‘flattening curve’ rationale for higher volatility.”

(SocGen)

SocGen – a bank which, by the way, isn’t generally averse to making bearish calls or otherwise suggesting that volatility could spike – goes on to call that purported correlation “spurious”.

“The yield curve historically starts flattening as the economy reaches the middle (uneventful) phase of the business cycle”, the bank writes, adding that because that “boring phase” is “when investors are unable to enjoy capital gains on risk assets or sharp moves on safe assets”, it’s generally the time when folks start to hunt for yield.

Now, think about what happens when the hunt for yield gets going – it compresses volatility and risk premia. “Selling volatility generates yield, therefore, in our view, a flattening yield curve is consistent with lower, not higher, volatility parameters”, SocGen goes on to say, before driving the point home even further by noting that if you “play around with the two time series” you can, if you like, “demonstrate that the VIX both lags and leads the US yield curve, and that a flatter yield curve can lead to either higher or lower volatility.”

So, what does matter when it comes to forward-looking equity vol. (because, as we said in December, and very much contrary to what somebody selling a narrative and/or gold coins might have told you, it is absolutely futile to correlate the VIX with the yield curve)?

Why, real rates, of course. “The key takeaway from our work over the past few years analyzing the impact of macro factors on equity volatility is that it is the real central bank policy rate that drives the (subsequent) volatility in equities”, SocGen goes on to say, adding that “increases in real rates take a few years to feed through to the economy”. Jerome Powell himself has said just that recently (on November 28, for instance).

The bank continues, noting that “while the QE period has distorted the tight fit these two parameters had in the 1990s, the inflection points have overall been forecast quite accurately by our real rates model.”

(SocGen)

Additionally, tweaking lead/lag times and inverting the axes doesn’t disprove the model as it apparently does with the the purported VIX/yield curve relationship. Indeed, except for oil shocks, SocGen’s model generally holds going back nearly five decades.

So, if you’re looking to predict S&P volatility, do yourself a favor and focus on what’s important – not what someone cooked up by playing around with lead/lags in between writing about apocalyptic super volcanoes and the relative merits of buying beanie weenies and distilled drinking water in bulk to prepare for the moment when the system finally collapses in a glorious once-and-for-all purge of misallocated capital.