One year ago, in the aftermath of last August's ETFlash Crash, we wrote about the risk of "violent, sharp and unexpected market moves" tied to the delevering of "risk parity funds" (see our post entitled "This Is What The Historic "Risk Parity" Blow Up Looked Like"). For those unfamiliar with the term, we previously described a risk parity fund as "a cross-asset portfolio allocation model that assigns weights inversely proportional to volatility and typically prescribes being overweight fixed income assets and underweight equities" - the best, and biggest, example of a risk parity fund is Bridgewater's $70+ billion "All Weather" fund.
For equity markets, the risk associated with risk parity funds is their programmatic nature and the sheer amount of levered capital allocated to the strategy which can result in substantial selling pressure when risk measures exceed pre-defined parameters. As BofA pointed out about a year ago, volatility swings in August 2015 likely resulted in $28 - $48 billion in equity selling pressure from risk parity funds alone resulting in a massive equity selloff that wiped out the YTD performance of a couple of prominent hedge funds.
A recently published report estimated the size of assets tracking risk parity at $400bn. Based on typical characteristics in many popular risk parity funds, we can assume vol control overlays with target vols between 6% and 10% vol along with leverage caps from 1.5x to 3.0x. Under these assumptions, anywhere between 40% and 120% of unlevered assets under management could have been deleveraged last week.
If we assume $200bn of the purported $400bn risk parity AUM uses vol control, then the above 40% to 120% range suggests that between $80bn and $240bn in levered risk parity notional could have been unwound over the prior week. With a 35/65 split between equities and fixed income, this would have translated to $28bn to $84bn of potential selling pressure in equities and $52bn to $156bn in fixed income.
Which brings us to last week's sharp sell-off in JGBs (chart 7), which in turn raised renewed concerns of forced selling by risk parity funds. Showing just how fragile these funds can be, BofA finds that a mere 2% daily decline in the S&P coupled with a 0.6% decline in the 10Y Treasury could trigger 25% deleveraging by risk parity funds. "Last Friday 10-Year US Treasury futures declined about 60bps. Had the S&P 500 declined 2.0%, we would have expected about 25% of the unlevered notional of a model 8% vol-targeted, 2.0x max leverage risk parity portfolio to deleverage. The S&P 500 was in fact up 86bps on a total return basis which according to the tool falls in the region of no deleveraging."
Luckily for the central bank's "bull thesis", BofA also found that little to no risk parity deleveraging took place:
It is intuitive to think that rising volatility corresponds to an increase in model-driven selling pressure from risk parity strategies. However, what’s less appreciated in our view is the impact on risk parity allocations as a result of the relative dynamics between component volatilities and correlation. For example, through the close on the Monday post-Brexit (27-Jun-16), S&P 500 volatility rose from 9.6% two days prior to 17.9% (increase of 1.9x) while 10-Year US Treasury Futures return volatility rose 4.3% to 6.6% (increase of 1.5x). Despite these outsized vol moves, in a recent report we showed that owing to the diversification (increasingly negative correlation) between equities and bonds, unlevered risk parity portfolio volatility remained stable and hence, target vol overlays were less likely to be subject to model-driven selling.
What changed this time to prevent a potentially market-moving risk parity event? BofA believes that equity/bond moves through the Aug-15 risk flare would not cause a deleveraging today. The reason is bonds have increased in allocation since last August (78% vs. 66%), and hence the portfolio is more resilient towards equity market declines. The flipside, however, is that while risk-parity is now more insulated to equity market crashes, it is far more sensitive to sharp, sudden selloffs in Treasurys.
The chart below is a useful tool to help assess deleveraging risk associated with daily S&P and 10Y Treasury movements. As BofA points out, the most risk of deleveraging from vol controlled risk parity funds comes when both volatility and correlation of the underlying components rise together (i.e. quadrants 2 and 3 in the chart below).
What the schematic above shows is how much more sensitive to the Treasury market risk-parity funds have become, especially in the case of a combined melt down... or up.
For example, the first and third quadrants (upper right and lower left sections) are dominated by scenarios of greater than 50% deleveraging. On the other hand, the second and fourth quadrants have episodes of more benign model-driven deleveraging. In each quadrant exists examples of simultaneously increasing equity and bond volatility (that is, high absolute equity and bond daily returns). However, in the first and third quadrant, equity and bond moves are in the same direction, which would likely be an example of increasing correlation. On the other hand, in the second and fourth quadrant equity and bond moves are in opposite directions and hence correlation is subject to a decrease. The main takeaway here is the most risk of model driven deleveraging from vol controlled risk parity funds comes when both volatility and correlation of the underlying components rise together.
So does the fact that the market avoided a major risk-parity freak out two weeks ago suggest that the system is more stable? Here is BofA's take:
Latent risk remains worth monitoring, as (i) leverage is still near max levels across a variety of risk parity parametrizations, (ii) bond allocations are historically elevated, and (iii) markets continue to be sceptical of a 2016 Fed hike.
If BofA is correct, it would mean that a day which sees a -4% SPX drop and +1% bond rally (good diversification) would generate no selling pressure, "underscoring the critical role played by bond-equity correlation in governing the severity of risk parity unwinds." However, a troubling scenario is one where even a relatively benign 2% selloff of the S&P coupled with just a 1% selloff of the 10Y could result in up to 50% deleveraging, which in turn would accelerate further liquidations by other comparable funds, and lead to a self-fulfilling crash across asset classes.
Which incidentally sounds like precisely the scenario that could happen when the Fed tries to raise rates, and is also why asset classes continue to move without fear of any rate hike, as they now realize - very well - just how trapped the Fed truly is. That said, in 4 months we will see if the Fed, for once, has the intestinal fortitude to actually raise rates in the face of the extreme volatility awaiting equities in the event they do... we doubt it.