Last Thursday, when the S&P was once again surging to within a fraction of taking out its all time high, we warned readers to "Brace For "VaR Shock" - How The Bank Of Japan May Be About To Unleash A Global Selloff." Of particular interest was whether 10Y JGB yields would soar, now that the market was questioning the BOJ's resolve to keep longer rates under control, leading to a risk-contagion scenario like the one seen in 2003, when Japanese bond yields exploded. We then said that "what that selloff - in a time of soaring cross-asset correlations, record quant leverage and virtually non-existent market liquidity - would mean for equities, we don't know, - but thanks to Haruhiko "Peter Pan" Kuroda, we will soon find out."
Well, the timing of the post could not have been better, because we got our answer the very next day when - just as predicted here on Thursday - as a result of crashing bond MTMs - the very definition of a VaR shock - global stocks suffered their first aggressive global selloff in months, the biggest one in fact since Brexit, as trader attention, which has been ignoring pretty much every development and key news update over the past two months assuming instead that central banks "have it covered", finally focused on the sharp selloff in long-dated global bonds leading to cross-asset liquidation and sharp quant deleveraging, as we predicted just hours earlier.
Friday's selloff, incidentally, also took place just two days after JPM's head quant, Marko "Gandalf" Kolanovic issued a new stark warning: "Volatility Is About To Surge", due to catalysts which - as we explained in the post - included this month's central bank (ECB, BOJ, Fed) meetings, seasonals pushing market volatility higher, and leverage in systematic strategies and option positioning provide fuel for volatility. We bring this up just to silence all those peanut gallery complainers that Kolanovic "has no idea what he is talking about."
But while Kolanovic's arguments confirm that our prediction of an imminent VaR shock is correct, another influential JPM analyst, Nick Panigirtzoglou, author of JPM's Flows and Liquidity weekly snapshot, is just fractionally more optimistic and says that while a VaR shock is a distinct possibility, he notes that "not all conditions underpinning VaR shocks are currently in place in either equity or bond markets." We disagree, but that's what makes a market. As such, we present his thoughts on the matter, and let readers decide just where in the imminent VaR shock continuum we find ourselves.
Here is Panigirtzoglou's attempt to mitigate some of the long, long overdue fears of a risk off response to another round of central bank mistakes, and his take on whether September 2016 is set to be this year's major "tantrum" event which first sweep Japanese and European bonds, and then quickly spreads across the global impacting all asset classes.
* * *
Market volatility declined further over the past month raising more concerns about complacency. Are investors too complacent currently? Are current market conditions inducive to a VaR shock? And VaR shocks do not necessarily need a fundamental trigger such as policy changes or political events. Examples of fundamental or policy triggered VaR shocks were the taper tantrum of May 2013, the Chinese devaluation of August 2015 and more recently the Brexit vote. VaR shocks can simply occur if, for example, investor positions normalize from previous extreme levels. Examples of such non-fundamental VaR episodes were those that took place in April 2013 in the JGB market or in April 2015 in the Bund market.
So, according to the Greek JPM analyst (not to be confused with his nemesis on this issue, the Croatian one), how likely are such shocks in the current conjuncture? According to JPM, there are certain conditions underpinning the emergence of VaR shocks:
1) Low volatility. Low market volatility induces investors, most of whom employ some type of volatility-based risk management framework, to increase the notional size of their positions as volatility collapses. The same investors are forced to cut their positions when hit by a shock, triggering self- reinforcing volatility-induced position shedding. Examples of VaR sensitive investors are hedge funds such as risk parity funds, asset managers, banks that set limits against potential losses in their trading operations by calculating Value-at-Risk metrics. Historical return distributions and historical market volatility measures are typically used in VaR calculations given the difficulty in forecasting volatility. This in turn induces these investors to raise the size of their trading positions in a low realized volatility environment, making them vulnerable to a subsequent volatility shock.
2) The market to be “trading long”. VaR shocks tend to materialize following an overhang of extreme long equity or duration exposures. How elevated are investor positions currently? The equity betas we regularly update in Charts A20 and A21 show a rather mixed picture. While systematic hedge funds such as risk parity funds and CTAs have high equity exposures currently, discretionary managers such as Discretionary Macro and Equity Long/Short hedge funds appear to have reduced their equity exposures recently. So not all hedge fund sectors are high in terms of their equity exposures.
The picture is also mixed in the bond space. For example in our European client survey, while Euro area single currency investors have still pretty high duration exposure at 0.36 years, multi-currency investors are close to neutral. Both of these position indicators were very elevated at the beginning of 2015, just before the April 2015 Bund VaR shock. By taking into account both the blue and the black line in Figure 2 the picture we are getting is of less extreme positioning in Euro area bonds than in April 2015
We mentioned banks as typical VaR investors. Are US banks too long duration? The Fed's H.8 release provides some guidance on this for US banks. Each week, the Fed reports the net unrealized gain on banks’ available for sale securities. This is by no means a complete measure of banks’ duration exposure. It does not include held to maturity portfolios (albeit these are much smaller than AFS portfolios), and importantly does not include the impact of any swap hedges. That said, we can infer banks’ duration exposure by relating week-to-week changes in unrealized gains to week-to week changes in the yield of our UST 10y yield. Figure 4 shows this beta, estimated over a rolling 3-month window to smooth through the noise in the estimate. It suggests that banks would make MTM losses on their available for sale portfolios of around $15bn for each 1% increase in the 10y yield. Abstracting from the volatility post last August’s episode, the current reading is rather negative i.e. US banks are rather long duration, albeit not far from the middle of its historical range.
An alternative way of gauging positions is to look at the response of equity and bond markets to economic news. This is shown in Figure 5 and Figure 6 which look at the behavior of equity prices and bond yields to economic surprises. An environment where long equity positions are heavy should see equities responding by more to negative economic news than to positive news. That is, the beta of equity prices to negative news should be higher than their beta to positive news. An environment where long bond duration positions are heavy should see bond yields responding by more to positive economic news than to negative news. That is, the beta of bond yields to positive news should be higher than their beta to negative news. The bond market is currently trading close to “neutral” according to this indicator, in contrast to the overhang of long positions seen in April/May 2015 during the “Bund VaR shock”. Similarly, the equity market is trading close to “neutral”, in contrast to the overhang of long positions seen in May this year and December last year.
Taking together, the above indicators do not currently point to overstretched equity or duration positions.
3) Stretched valuations. Admittedly this is a harder to assess metric especially on an absolute basis. On a relative basis valuations for JGBs or Bunds looked very expensive during the April 2013 and April 2015 VaR episodes, respectively. One simple way of illustrating this is via a bivariate regression of the 5y5y forward swap rate in Japan or the Euro area on its US counterpart and on the 2y spot rate differential vs. the US. This is shown in Figure 7. The residuals were very negative i.e. valuations were very expensive, for Japanese rates in April 2013 and for Euro area rates in April 2015. According to Figure 7 these residuals look a lot more “normal” i.e. close to zero, currently. In the equity space, a simple visual inspection of 12m forward PE multiples shown in Figure 8 suggests that the previous relative expensiveness of European and Japanese equity markets, the equity markets that corrected the most since August 2015, has more than unwound. In all, not all conditions appear currently in place for a VaR shock as low market volatilities are combined with rather mixed investors positions and valuation signals
Regarding market volatilities it is also important to remember that while vol is low, it is not cheap. In fact the opposite is true. Volatility is rather expensive as we highlighted in this month's Cross Asset Volatility Strategy publication (Sep 8th). Implied volatilities have indeed declined to levels last seen n August 2015, before the equity market correction. However, realized volatilities have declined even more sharply. This is also shown in Figure 9 which constructs a cross asset 1- month Realized vol metric based on the same indices and weights as the ones used for our Implied volatility metric. This Realized vol metric currently stands at the lowest level since October 2014! The greater fall in realized vs. implied volatilities means that vol risk premia increased sharply over the past month making vol even more expensive as an asset class. The ratio between our cross-asset Implied and Realized vol metrics stands at well above one, at 1.3x currently. In other words there are hefty risk premia embedded in option markets that are more indicative of skepticism rather than complacency. And this particularly true with retail investors who continue to pour money into VIX ETFs (Figure 10), despite the very negative carry. As a result of the extreme steepness in the VIX futures curve, popular VIX ETFs are currently losing around 10% per month due to negative roll. If no shocks materialize over the coming months, it is likely that negative-carry long- VIX positions are taken off, exerting downward pressure on the steepness of the VIX curve, from currently “bubble” like levels.
* * *
And there you have it: on one hand JPMorgan - via Kolanovic - expects a major selloff, potentially as much as 20% should central banks not come to stocks' bailout this time; on the other hand you have JPM - via Panigirtzoglou - explaining that the "other" JPM guy is wrong, and that the conditions for a VaR shock are only modest embedded right now. Which, of course, is good news if only for the bears out there: after all if everyone was expecting an aggressive continuation of Friday's plunge, it would be assured that that would not happen. However, now that JPM's Greek fund flow expert has inject a trace of doubt that the VaR shock may happen, the contrarian selloff may well continue.
Luckily, there will be a quick and easy way to find out of Panigirtzoglou is wrong: keep an eye on JGB trading overnight. If we the banchmark bond's yield surges, and if the Nikkei is getting whacked, the answer will be obvious: major quant and CTA funds are now deleveraging and until they are finished, there will be no rebound for the market for the time being. In fact, if central banks really think they can extricate themselves from micro managing the market and the economy (with only failure to show for it), then the next leg lower in the S&P500 could make on dizzy.
In the end, it will all be about what Kuroda - central bank governor of the country that has almost all negative yielding paper in the world - with does and says in the next few hours because if the BOJ reveals yet one more disappointing surprise, all bets may just be off. As to where the 10Y JGB ends up trading, we don't know, however we do know that a lot of investors will be crushed if the "frontrun the BOJ" trade no longer works; in that case the TINA option, or there is no alterative to holding stocks, will promptly be replaced with a "lack of alternative" to holding cash. At least, until, such time as the Ken Rogoff proposal to ban cash is fully implemented and the next and final leg of the monetary policy cycle arrives, the one which ends in hyperinflation and the collapose of the global reserve-based system.