Over the past 3 months, the name Marko Kolanovic, head of JPM's Quant Team, has become one of the most loved, or feared (depending on which way he is leaning) and respected on all of Wall Street for one simple reason: think Dennis Gartman, only correct every time.
The reason, as we have profiled before, is simple: he has somehow succeeded in calling every single market inflection point since the August 24 flash crash, and we have documented them all:
- August 21, just before the Black Monday flash crash: "Why The Market Is Crashing Into The Close: JPM Explains"
- August 27: "JPM Head Quant Warns Second Market Crash May Be Imminent: Violent Selling Could Return On Thursday"
- September 3, before the next leg lower in stocks: "Home JPM Head Quant Is Back With New Warning: "Only Half The Selling Is Done; Expect More Downside""
But his most prodigious call came on September 24 when we wrote "Bears Beware, JPM's Head Quant Just Flipped To Bullish: "The Technical Buying Begins." So it did, leading to the biggest market ramp in history, and biggest monthly point gain ever.
So moments ago we were excited to find that after a one month silence, Kolanovic released an update of what he sees in the arcana of technical trading, risk parity flows, convexity, gamma, and options hedging.
Here is the summary:
In our reports in August, we forecasted the selling pressure from option hedging and pointed to the role various systematic strategies had in the selloff. In our note from September 24th, we predicted a reversion of these technical flows and their potential to lift the market. We believe that most of these equity inflows played out over the past month and were a significant driver of the October market rally.
Just add a historic short squeeze and buybacks greater than even during last year's record, and you get the three catalyst that led to the massive rally since September. More importantly, according to Kolanovic the "technical inflows" that led to the rally are now over.
After the September option expiry, investors rolled protection lower, and as the market moved higher, convexity in index option products declined. We estimate that hedging of index options during the week of September 28th contributed to ~$20bn of equity inflows. During the month of October, the gamma exposure of put options declined significantly (and gamma of call options increased), such that the net effect of option hedging has been muted since (and likely contributed to lower realized volatility given the imbalance of option gamma towards calls).
Where did the inflows come from? Why the momos of course, as the best performing stocks were once again those which were going up because they were going up:
Perhaps the most significant inflows came from trend following strategies, i.e. CTAs. As discussed in our previous reports, all of the equity momentum signals (short, medium and long term momentum) turned negative in late August. As a result, CTA Equity exposure (as measured by its equity beta) reached record short levels in mid-September. On October 2nd, short term momentum turned positive (e.g. 1M momentum) and shortly afterwards long term momentum turned positive as well (e.g. 12M momentum). This implied a significant re-levering of CTAs during the week of October 5th (which we observe from the sharp increase in CTA beta – as shown in Figure 1).
At the end of October/early November intermediate equity momentum (e.g. 3M-6M) also turned positive, and this resulted in another large equity inflow from CTA strategies. Currently, all of the equity momentum terms are positive, which suggests that CTA equity exposure should be at the high levels observed in early summer (also confirmed with the short term beta of a CTA index to the S&P 500 in Figure 1). In short, trend followers made a full circle of equity investing from record long, to record short and then long again over the past quarter. Our estimate of equity inflows from trend following strategies over the past month is ~$70-90bn.
Then it was the constant vol traders, who too were caught in a feedback loop of buying stocks as vol dropped:
Given the sharp decline in realized volatility, strategies that target constant volatility also had to re-lever. Figure 2 shows estimated equity exposure of Volatility Targeting strategies (our asset/signal assumptions about Volatility Targeting (VT) strategies are unchanged: ~$300bn in assets, with an average 8-9% target volatility).
VT strategies likely started buying equities in late September and through October, at a pace of ~$5-8bn per week (or a total of ~$30-50bn). Given that levels of volatility are still below those observed in early summer, in theory, VT strategies could continue buying equities if volatility were to decline further. However, our view is that realized volatility is unlikely to drift much lower (e.g. to the summer lows), so any residual buying from VT strategies may not be sufficient to push the market much higher.
And then, the infamous risk parity funds:
Finally, we want to address Risk Parity strategies. Our estimate of assets following various versions of Risk Parity is ~$500bn. However, Risk Parity strategies employed by Hedge Funds may be substantially different from those employed by e.g. Pension funds (using risk parity in house as a longer term asset allocation method). For this reason, we use different models for ~$150bn in ‘HF-like’ risk parity assets (leverage >1, higher rebalance frequencies, and typically using volatility target overlays) and ~$350 in Risk Parity pension asset allocations (leverage < 1, slower rebalance frequencies and signals, and typically not using volatility targeting overlays). Figure 3 shows the equity exposure of a prototype ‘HF-like’ Risk Parity allocation, which indicates that these funds de-levered in August and September, but re-levered in October to finish at their pre-crash equity allocations.
Equity inflows from these funds may have amounted to ~$20bn in October. Risk parity strategies that use slower signals (e.g. 12M covariances) did not materially change their equity exposures.
His conclusion: the catalysts behind the furious, technically-driven October rally are now gone, but unlike mid-August, at least the likelihood of another flash crash is lower...
Summarizing technical flows from option hedges, volatility targeting, CTA and Risk Parity funds, we believe that these strategies largely re-levered to pre August crash levels. This was a significant driver of the S&P 500 performance in October and hence poses some downside risk. Additionally, given the tight trading range over the past year, CTA signals have risk of changing on relatively small market moves (i.e. there is elevated ‘CTA gamma’). On the other hand, given the lack of a large put option gamma imbalance, and perhaps some residual buying from VT funds, near term the market is likely more resilient to the risk of another technically driven flash crash.
... unless the Fed surprises: according to Kolanovic one person can overturn the cart, and that person is Janet Yellen if she once again confuses the market:
Over the past year, macro momentum trades increased exposure to various liquid assets in anticipation of a rise in US rates. Example trades include going long USD and Developed Markets, and short Commodities and Emerging Markets. These macro trends have also percolated into equity long-short momentum trades which are currently short Energy, Materials and Industrials, and Long Health Care and Consumer Discretionary sectors. Several of these macro and stock trends are relying on an anticipated Fed tightening that would boost the USD and further weaken commodities and EM assets. The risk of this increasingly one dimensional positioning across CTAs, Macro and some of Equity Long-Short managers is that these trends don’t materialize and trades become too crowded. The result could be a sharp reversion as positions are exited.
The only question then is: does the Fed want to risk such a "sharp reversion as positions are exited." The answer is revealed on December 16.