Over the past several weeks, central banks globally have been keen to emphasize the extent to which many of the emergency actions taken are necessary to ensure the smooth functioning of their respective government bond markets.
In mid-March, the US Treasury market effectively “broke”, leading to all manner of RV blowups and various dislocations, discussed here at length on too many occasions to count.
Treasury market liquidity was severely impaired, and the situation was exacerbated by a fire sale, as foreign central banks liquidated more than $100 billion in USTs as part of March’s market zeitgeist which can be loosely summarized as follows: “Sell anything that’s not tied down to raised USD cash”.
The Fed has done quite a bit to ameliorate the situation. Just last week, for example, we got the foreign repo facility and an interim rule exempting Treasurys and deposits at Federal Reserve banks from the supplementary leverage ratio, a move some argue is fraught with peril, but which is necessary to ensure that some of the steps taken to support the system don’t end up tripping over themselves by accidentally curtailing banks’ willingness to lend.
Some of these measures were met with predictable lines of criticism from an equally predictable list of critics. The charges vary, but generally involve some combination of generalized shouting about moral hazard and more specific accusations tied to purported “bailouts” of hedge funds.
To be sure, there’s always a moral hazard argument. And that goes for markets and life in general. Just pick a vexing scenario where some higher power (usually governments or officials) is compelled to intervene and you can conjure a moral hazard counter-narrative.
That doesn’t mean moral hazard arguments aren’t good or that they don’t often prove prescient when something goes wrong down the road. It’s just to say that you should be wary of blindly assuming that everyone who screams “moral hazard” is intelligent. Moral hazard arguments are a dime a dozen – and not just in markets.
One particularly touchy subject was the extent to which Fed action came on the heels of blown-up basis trades, illustrated vaguely in the following visual (this is just an updated version of the chart Bloomberg used in their discussion last month):
This sounds complicated, but like many ostensibly “esoteric” dynamics, isn’t actually that complex at a conceptual level. In case you missed it, here’s JPMorgan’s Josh Younger to explain:
Basis trading strategies are designed to police the relationship between Treasury futures contracts and the bonds that meet their criteria for delivery. At its most basic, one can monitor the price differential between a given futures contract and the cheapest-to-deliver (CTD) bond. Because the short futures leg determines when and what bond to deliver the long, as long as they hold the position to expiry they can be assured of a sale at a known price and a time of their choosing (at least within the deliverable month). When the cash CTD bonds are cheap, for example, buying them (often levered using overnight or term repo) and selling the futures offers a near-arbitrage payoff—albeit for a relatively small P/L of a few 32nds of a percent.
Again, we’re not building deep space probes here – at heart, it’s just capturing pennies via a small divergence and turning those pennies into dollars with leverage.
“So far this likely sounds like yet another wonky analysis of derivative market structure and small pricing discrepancies. Which it is.”, JPMorgan’s Younger wrote, in his explainer on the subject dated March 13 (his piece was widely cited both in these pages and, for example, by Bloomberg in the linked post above).
But he went on to explain that as “wonky” as is sounds, it had knock-on effects for funding markets which were already exhibiting signs of stress. The following is a bit tedious for everyday folks, but the point is simply to explain how things spilled over, and thereby underscore one of the rationales for Fed intervention. Here’s Younger again (keep in mind, this is from a March 13 note):
We should all care about what is going on in Treasury futures basis. These positions, in addition to being a relative value trading strategy, can also be used for short-term cash management. Futures pricing implies a repo rate that can diverge from the cash market, and when that happens buying the basis unlevered (i.e., not using repo to fund the CTD leg) can be seen as a risk-free competitor to 1- to 3-month investment for real money.
When the implied repo is high enough, these positions start to look more attractive than other products. This does not occur often, since relative value trading and dealer activity will tend to enforce implied repo close to cash repo, and both much lower than those with credit or FX risk. Recent moves have, however, turned that relationship somewhat upside down. Futures basis in TY, for example, is now quite a bit more attractive than 3-month bank CP and close to carry on 3-month EUR forwards (Exhibit 3). It is also important to note that even better levels were possible intraday [on March 11]. At these levels, the incentives are quite strong for real money to reallocate away from these markets and into nearly risk-free basis positions at a higher yield. In that sense, there is a rather direct channel between stress in the futures basis market and a wide range of other asset classes, including but not limited to unsecured bank funding—and by extension, Libor—as well as FX forwards.
There you go. It’s “wonky” (as Younger put it) but as I was keen to point out a couple of weeks ago, you don’t necessarily need to understand the mechanics, just the concept.
Well, fast forward a few weeks from the worst of the turmoil, and the BIS (the “central bank for central banks”, as it’s colloquially known) is out with a pretty good summary of what happened in the Treasury market last month. It touches on almost all of the topics discussed in these pages, and because it amounts to a relatively dispassionate take on things, it’s worth highlighting a few excerpts.
Here’s the BIS describing what happened to hedge funds’ RV strategies (and this is couched in even more straightforward terms than I’ve used):
Hedge funds that employ so-called relative value strategies fund large positions in treasury securities using leverage through repos, while, at the same time, selling the corresponding futures contract. Hedge funds employing such strategies profit from small differences in the yield between cash treasuries and the corresponding futures. Moreover, since the futures yield and cash yield move together closely, the hedge fund can pocket the small yield difference regardless of which way the market moves. This is an example of a “long-short strategy”, as made famous by Long Term Capital Management (LTCM), a hedge fund that failed in 1998, where a long position in one asset is hedged by a short position in a closely related asset.
Relative value investors can achieve high levels of leverage because the collateral value of Treasury securities is normally high. For instance, if an investor can borrow $99 by pledging $100 worth of treasuries, the investor need have only $1 of own funds to hold treasuries worth $100, achieving 100-fold leverage. In this way, even a small yield difference can be magnified by leverage. Before the Covid-19 crisis, the cash futures relative value strategy delivered a steady stream of returns. One indication of the popularity of such trades was the growing short positions in futures by leveraged funds (Graph 2, centre panel).
However, as volatility picked up and margins surged, particularly at the long-end (Graph 2, right-hand panel), liquidity in futures markets deteriorated sharply (Graph 3, left-hand panel). Futures-implied yields dropped more rapidly than bond yields, imposing mark-to-market losses on relative value investors who had sold futures and bought cash bonds.
One facet of the ensuing price dislocations between bonds and futures was the fact that the implied repo rates of the cheapest-to-deliver bonds in the futures contracts rose markedly above market term repo rates (Graph 3, centre panel). In frictionless markets, the implied repo rate that an investor would earn by buying the cash bond and delivering it in fulfilment of the futures contract should theoretically be equal to the actual term repo rate. The breakdown of the relationship suggests severe strains at the time in relative value trades. Once the funds were no longer able to meet variation margins, their positions were unwound by dealers/futures exchanges, pushing prices lower. This in turn gave rise to a classic “margin spiral”, whereby the cycle of illiquidity, price dislocations and tighter margin requirements fed on itself.
Over there in the right panel of Graph 3 is a simple (and familiar) chart that shows bloated dealer balance sheets. After two years of absorbing massive supply (i.e., issuance), dealers weren’t really in a position to cushion these unwinds.
Well, as things escalated, CTA trend piled on across various assets as key trigger levels were breached. “While the locus of initial market turbulence largely involved relative value investors, de-risking quickly spread to systematic funds”, the BIS goes on to say, reminding you that “such funds follow a rules-based investment strategy with limited, if any, scope for discretionary or tactical adjustment”.
And then came vol.-control and risk parity. The BIS largely parrots the existing market commentary on the epic risk parity unwind that played out late last month. To wit:
They tend to have longer look-back periods and are less responsive to short-lived episodes of financial market volatility. However, during the COVID-19 crisis, the dislocation in the treasury market meant that bonds no longer performed as a hedge or delivered desired exposure in either type of strategy. As the funds scaled down leverage, this led to a dash for cash, exacerbating declines in both equity and bond prices. As the equity market sell-off accelerated and treasury volatility also rose, the assets and number of risk parity funds began to fall, indicating that the de-risking by systematic funds was widespread.
By most accounts, things started to unravel for risk parity in earnest during the week of March 9. By March 13, it was readily apparent that RP deleveraging was underway. By March 19, exposure had been slashed dramatically.
“As the ‘bonds as your risk-asset hedge’ correlation runs into the ‘liquidate to cash at all costs’ buzzsaw, we see an unprecedented forced-deleveraging from risk parity, from 555% gross exposure (100th %ile) to 240% (0.1%ile)”, Nomura’s Charlie McElligott wrote, marveling at the rapidity and scope.
“The simultaneous downward pressure on equities, bonds and commodities placed immense pressure on risk parity funds and other multi asset investors which rely on low correlation between asset classes to contain risk and drawdowns”, JPMorgan’s Nikolaos Panigirtzoglou said, in a note dated March 20. “Instead, over the past week risk parity funds were hit on all of their asset holdings at the same time and were thus forced to de-lever”.
“This is a liquidity crisis, so people sell everything; it is a classic ‘correlation = 1’ event”, Harley Bassman said last month, in the e-mailed color accompanying his latest commentary.
China’s $941 billion wealth fund reportedly cut its RP exposure in half just before it all fell apart. That was good timing.
Towards the end of their exposition, the BIS offers some helpful tips for policymakers and market participants going forward. Regular readers will recognize the general thrust of their suggestions, as they echo familiar warnings about the self-feeding, “doom loop” dynamic which we now know is not merely a “ghost story”, so to speak.
“The recent episode in Treasury markets illustrates yet again that high leverage through cheap funding of securities positions can lead to the accumulation of negative convexity positions”, the BIS says.
“This directly raises the risk of an endogenous feedback loop, when such positions unwind rapidly and in a non-linear way as volatility picks up”, they go on to warn, adding that “market monitoring should look beyond current conditions and ask the ‘what if’ questions that are relevant for potential market stresses”.