This week was chaotic in markets.
From multiple curve inversions (2s5s, 3s5s) to Tuesday’s lunchtime plunge in the S&P (which coincided with a monumental rally in the long-end as tactical steepeners were stopped out en masse), to Wednesday night’s dramatic crash in U.S. equity futures to Thursday’s manic round-trip in stocks (and concurrent aggressive repricing of the Fed path via the short-end and interest rate futures) to Friday’s selloff, it was a week to remember.
Given the manic and often inexplicable price action and considering that all of this unfolded in a week that by almost all accounts “should” have been defined by a post-G20 trade truce risk rally, it’s no surprise that traders, investors and the mainstream financial media went looking for answers from some of the folks who are usually willing to weigh in when ostensibly strange things are happening.
The elephant in the room this week was obviously Nomura’s Charlie McElligott, the living embodiment of “more cowbell” when it comes to real-time updates on systematic flows. If you’ve “got a fever” and the “only prescription” is more rapid-fire analysis of programmatic de-leveraging/re-leveraging, McElligott is your Gene Frenkle.
In addition to (and in some cases, in response to) McElligott, we heard from a bevy of other names who often find their way into the discussion at times like this. We heard, for instance, from Nikolaos Panigirtzoglou, Wells Fargo’s Pravit Chintawongvanich and, of course, Marko Kolanovic, the boss of it all.
While everyone and their brother was obsessively staring at every E-mini tick in a largely fruitless effort to sort out exactly what was going on, Deutsche Bank’s Aleksandar Kocic released a 28-page note on Tuesday called “Through the looking glass: Central bank adventures in convexity management.” Here is the opening paragraph:
During the last few decades, we have learned how to handle booms and busts, but, at the same time, we have gradually internalized the message that we cannot prevent these cycles from happening. If stability is indeed destabilizing, the main challenge then lies not in how to avoid the mistakes, but instead in how to minimize the costs. We see convexity as the essential ingredient of leveraging and deleveraging dynamics and its management as an integral part of effective monetary policy. This convexity management role for central banks provides a new perspective on monetary policymaking in the post-crisis world.
In other words, while the rest of us were fumbling around for something novel to say about what curve inversions portend for the economy while simultaneously speculating about whether Skynet had become self-aware on the way to deciding, on its own, that the G20 needed to be faded, Kocic was detailing the role of convexity in crises management on the way to, quite literally, providing the Fed with a strategy blueprint based on the myriad concepts and frameworks that together comprise his wholly unique approach to analyzing the evolution of monetary policy.
Regular readers are familiar with Kocic. His name doesn’t ring the same bells in the mainstream financial media as some of the Street’s more widely-followed analysts, but that’s probably because he’s incomparable. His notes transcend financial analysis. The sheer number of references to history, literature and other disciplines in his work and the way in which he manages to weave them all together in his weekly missives is almost impossible to fathom. Readers know we like to bring in pop culture references. Kocic seems like the modest type, so we’re pretty sure he doesn’t think about it this way, but one is reminded of the classic Jay-Z cameo on Outkast’s “Flip Flop Rock”: “I’m on a whole different plane/ a whole different lane/ a whole different game that I’m playin’/ understand what I’m sayin’?”
Given that, it comes as no surprise that “Through the looking glass: Central bank adventures in convexity management” is a veritable masterpiece, the profundity of which is impossible to convey in a post here, but we’re going to give it a shot.
Kocic kicks things off by framing the 2008 crisis as an en masse “short covering of negative convexity exposure in a market with acute shortage of convexity supply.”
“In the lead up to the crisis, years of risk premium compression and the search for yield forced investors into positions that were negatively convex through conventional callables, mortgages, credit, CDS, and structured products”, he recounts, emphasizing that “everyone was short optionality in some form: issuers, investors, bond holders, asset managers, insurance companies, pension funds, banks, dealers, hedge funds, and so on.”
“No one”, Kocic says, “was exempt.” That means everyone was short vol. one way or another. That is of course only a tenable proposition as long as benign market conditions persist.
“Like ordinary high deductible insurance policies, they were not reactive to local market moves, but, when spreads exploded following initial default shocks in the subprime market, volatility spiked and rates moved lower in a risk-off mode, all of these options quickly and simultaneously converged in the neighborhood of the strike (where an option’s convexity is maximal), raising the market’s negative convexity exposure practically exponentially and, thus, triggering subsequent hedging which further reinforced the trend”, Kocic goes on to write, describing the mechanics of the panic while simultaneously capturing how a veneer of stability can be shattered virtually overnight (recall “the avalanche“).
He goes on to recount the subsequent panic as everyone suddenly realized that the illusion of stability was just that – an illusion. “As everyone ran for the door at the same time, there was an inadequate supply of convexity — no one in the private sector was willing or capable to offer sufficient downside protection to cover all of the shorts”, Kocic recalls, before setting up his characterization of the Fed as a convexity manager by noting that “the imbalance between the surge in demand and simultaneous disappearance in supply evolved into a spike in risk aversion and volatility levels at which the markets could no longer function properly.”
Enter the Fed and a new monetary policy regime.
While convexity considerations used to be the purview of investors and capital allocators, post-crisis, the management of convexity flows became a defining feature of the central bank-market nexus. The Fed’s response to the GFC was, Kocic contends, “essentially as a program of convexity supply in order to provide stability and restore the market’s normal functioning.” He crystalizes this by couching the three aspects of the Fed’s policy response in terms of convexity management.
On rates, Kocic notes that the date-and-state dependent nature of forward guidance “assured investors that short-term interest rates would remain low and, more importantly, would not be a source of uncertainty or volatility.”
Kocic then characterizes MBS purchases as the “nationalization of the mortgage negative convexity” – an effective supply of convexity to the key market at the heart of the problem (for more on this point, see here).
As far as QE goes, it’s straightforward. Between forward guidance and Treasury purchases, the Fed effectively installed rate caps and floors, thereby ensuring that dramatic moves in the curve would not manifest themselves in destabilizing spikes in rates vol.
Inherent in any emergency response is an element of moral hazard. In this case, the Fed’s crisis response engendered the global hunt for yield and, ultimately, created a scenario where free choice was replaced with a “free selection” among a series of options which all amounted to the same binary proposition: participate in what amounts to a global carry trade or else attempt to rebel against the state of exception and risk going out of business as you burn premium waiting on something to happen (there are multiple references there to Kocic’s earlier work).
If you read that last paragraph and come away wondering whether the policy response to the crisis has now put us right back into the same conditions that caused the crisis in the first place, the answer is “yes.” The global hunt for yield pushed everyone out the risk curve and capital misallocation is now a defining feature of markets.
“In a perverse way, after almost a decade of accommodation, as markets calmed and the economy recovered, we are where we started before the crisis: Everyone is once again in a short convexity position”, Kocic writes. That means (and this is intuitive), that “the unwind of monetary stimulus is thus a mirror image of the QE trade.”
The natural question, then, is how can the Fed mitigate things as the state of exception is lifted and markets are re-emancipated? If you’re familiar with our coverage of Kocic’s work, you already know part of the answer. The key is careful management of convexity flows.
“Convexity cannot be withdrawn simultaneously from all market sectors; it has to be taken out slowly and selectively”, Kocic calmly explains, in what amounts to a policy recommendation. Given the renewed proliferation of negative convexity positions, volatility becomes the paramount variable. The Fed, Kocic writes, must then deftly navigate the precarious waters around marking the transition from convexity supplier to convexity manager.
Next, Kocic outlines the three steps the Fed has taken so far in the course of normalizing policy. The first step didn’t go well. What was most worrisome about the taper tantrum, Kocic says, wasn’t so much the steepening, but rather the rapidity of the spike in rates vol.
“While it took the curve three months to realize about 150bp of steepening, short-dated (implied) rates volatility doubled and credit spreads widened by 50% in just the first six weeks”, he writes.
That episode underscored the need for policy normalization to be carefully telegraphed – there could be no surprises. Even if only implicit, there had to be an agreement between the market and policymakers about the schedule for the removal of accommodation.
The second step was making it clear that the removal of accommodation would be more gradual than previous cycles. The third step Kocic describes as the “iterative removal of the training wheels”, involving, first, the the restriking of the Fed put and the implicit withdrawal of convexity as tighter financial conditions are allowed to take hold with no accompanying nod to a change in the Fed’s conviction in the rate path. Second, forward guidance became less transparent as strict data dependence superseded the previous arrangement wherein the Fed and markets chose the path of least resistance when interpreting incoming information.
“As a greater distance between zero and the fed funds rate opened up, and as policy began to approach a neutral setting which itself was uncertain, the future path for policy became more data dependent and thus murkier”, Kocic writes, underscoring a point we’ve tried to drive home since Powell first took the reins. Paradoxically, “plain English” presaged less transparency, not more.
Kocic notes that Powell’s recent efforts to walk back the “long way from neutral” misstep that helped catalyze the October rout means “this process is not monotonic, it is iterative [and] requires the Fed to pull back at certain points if the tightening of financial conditions becomes too abrupt and risks the attainability of the Fed’s objectives on an ongoing basis.”
Now the “game plan” (if you will) becomes clear. One thing Kocic has emphasized continually is that a volatile repricing in the long-end is a non-starter. This is the key to his contention that the Fed is effectively “daisy chaining” the two ends of the curve via the equity market.
The concepts here are familiar and speak, in part, to why the market is concerned about investment grade credit and, in the same vein, the buildup of duration on balance sheets in an environment where the post-crisis regulatory regime makes the Street unwilling (and indeed, unable) to lend its balance sheet in a pinch.
“Tail risk resides in credit and the long end of the curve because that is where risk premia have been compressed the most and capital misallocation has been concentrated”, Kocic writes, underscoring the point and adding that “this sector, to which allocation has grown multiple times during QE, is associated with lower liquidity than the front end [and] in the light of regulatory changes and general scarcity of the balance sheet, its unwind would be impossible to absorb by the street.”
And so, the Fed must avoid a volatile back-end repricing at all costs by ensuring the supply of convexity to that sector remains some semblance of constant. Here’s how that will/should work, according to Kocic (regular readers will note that this is a crystallization of the concepts he’s been building on for years):
The plan is to daisy chain the three market sectors — the curve front end, equities and the long end — and recycle the convexity withdrawn from the front end of the curve and from the equity market into the back end of the curve. In other words, the Fed must be willing to inject greater volatility in the front-end and accept greater volatility in equities, while simultaneously working to limit volatility and downside for long rates. This will be done with the help of four dials: Renormalization of the curve, prevention of disorderly bear steepening, managing volatility, and allowance for tighter financial conditions along the path of removing some of the positive externalities behind equity performance by restriking of the Fed put.
Ok, we’ve been over “dial one” here before. In “normal” times, shocks arrive to the front end of the curve (i.e., bull steepening or bear flattening). But in the QE era, that has changed. With the front end glued to the lower bound, shocks arrive at the back end (i.e., bear steepening and bull flattening). The following chart shows the regime shift at the beginning of the QE era, when suddenly, the typically negative correlation between the 2Y rate and the 2s10s flipped:
Is the post-crisis mode an issue? Well, it wasn’t – but it might be now.
“Curve correlations are an implicit expression of the distribution of volatility along the curve”, Kocic notes, on the way to warning that while “the explosive process does not present a problem as long as the front end is in a ‘sleeper’ mode, as soon as it starts moving — when rate hikes commence — the risk of the long end getting unhinged becomes a problem.” So, as convexity supply is withdrawn from the front-end, it has to be matched by a supply to the long-end.
Here’s where this gets particularly interesting. Kocic goes on to explain why this is so critical in the grand scheme of things. He puts it in the context of the Street and its ability to step in as a middleman.
“The capacity of dealers and other market players to extend liquidity hinges on the ability to replicate any given asset with a portfolio of other securities”, he writes, emphasizing how paramount it is that “liquidity providers can always hedge their exposure in other market sectors [so that] market functioning can remain fluid.”
Well, as it turns out, instability in yield curve correlations screws that up (to speak far more colloquially than Kocic puts it). Note how he says dealers need to replicate the exposure inherent in what they’re holding. That replication process, Kocic notes, usually depends on PCA analysis. Here he is bringing this all together:
Figure 7 shows the long history of the yield curve according to a PCA in terms of their explanatory power in a rolling sample. The post-2008 period corresponds to major — in terms of both size and duration — instability in yield curve correlations, as the second factor has gained significant explanatory power relative to the level factor. This instability represents a compelling argument for a need for curve normalization — in order for liquidity providers to be willing to serve this role, they must be able to reliably predict curve correlations moving forward.
So there’s that. Moving on to “dial two”, the process of normalizing the mode of the curve (and now you know why that’s critical) is of course made immeasurably more complicated by the fact that the Fed is trying to unwind the balance sheet against a backdrop where the Trump administration is running a fairly anomalous experiment in debt-funded, late-cycle stimulus.
Not only does that experiment further distort the supply/demand picture in the Treasury market (supply is increasing while the Fed is pulling its support for the market), the effects of the stimulus risks pushing up inflation and reawakening the Phillips curve. “These forces work to inject volatility into the back end of the curve and thus work against the normalization laid out in the previous section”, Kocic writes.
This gets back to Kocic’s previous work on the “breather mode” of the curve. The bottom line here is that given the conditions outlined in the above paragraph, the Fed needs to guard against any vicious bear steepening episodes. The way they’ve achieved this is by effectively promoting long periods of bear flattening following brief episodes of bear steepening.
If you’ve been following along, you can see this as yet another manifestation of convexity management. Post-2016 election, for instance, this entailed sticking to the rate path despite weakness in core inflation and, more recently, in the face of turmoil in other markets. As Kocic puts it, “they are reducing the likelihood of facing overheating and inflation risks that could come if they fell behind the curve and instead pushing rates higher themselves — instead of risking that markets disruptively raise rates, they are taking control of that process and doing it on their own terms.”
Next comes “dial three” and if you were perusing these pages on November 17, you are well-versed in the dynamics that dial entails.
Why has rates vol. remained the outlier recently despite an uptick in equity vol. and widening credit spreads? This discussion goes back to the “shrinking playground”. The policy gap (the “playground”) defines the scope for rates volatility. As Kocic is fond of putting it, “anything that can happen, happens inside that gap.”
Obviously, it’s an oversimplification to say that the policy gap is the only thing that matters for vol. That is, if the gap disappears entirely, volatility isn’t going to just flatline altogether. “There is always residual uncertainty in rates determined by liquidity and other factors [so] volatility never collapses to zero”, Kocic wrote last month.
Here is realized rates vol. plotted with the playground:
Think about where the policy gap was when the current cycle started, versus previous cycles (as shown in the chart). During the 2004 experience, the gap was roughly 500bps. This time around (i.e., in the current cycle), the gap was just 200bp to start. There’s some math involved here, but the bottom line is that the tighter the gap, the less scope there is for rates vol. to move and if you look at the annotation there, you’ll note that “this time is different”, if you will. As Kocic put it last month, “with each subsequent hike, it appears as if the short rate is pushing the long rate from below. In this way, rate hikes continue without allowing volatility to take off.”
Finally, we get to “dial four”, which is the restriking of the Fed put as the withdrawal of convexity from the equity market by allowing financial conditions to tighten even as stocks sell off in the interest of ensuring that an economic overheat doesn’t end up manifesting itself in a “rogue” inflation print (i.e., some kind of scorching-hot AHE number or, potentially more concerning, a way above-consensus ECI print).
“By implicitly or explicitly communicating a greater acceptance of tighter financial conditions, the Fed is removing convexity from risk assets like the equity market, and supplying convexity to the long end of the curve by signaling the Fed is willing to act to prevent overheating risks that could translate into a more meaningful inflation overshoot and potential disorderly bear steepening events”, Kocic writes, in the course of delivering the following visual.
This is a dynamic we’ve discussed here on too many occasions to count. You can think about the restriking of the Fed put as the normalization of beta and the equity selloff as stocks’ reaction to the FOMC’s attempt to re-emancipate markets — to lift the state of exception.
Kocic reiterates and expands on his previous “Fed put” analysis, but the bottom line is the following (importantly, he does go out of his way to note that the co-movement of the short rate with equities is sometimes conducive to the promotion of the Fed’s mandate – i.e., it’s not always just a manifestation of a desire to protect the stock market):
In the current hiking cycle until earlier this year, the beta between equities and yields was unusually high, around 30, suggesting that monetary policy has been protective of risk — i.e., the Fed put between early 2014 and early 2018 had been struck very close to ATM. This relationship broke down earlier this year in February when Fed officials made it clear that they were willing to accept tighter financial conditions without adjusting their expectations for the appropriate future path of monetary policy. At that time, equities corrected sharply but expectations for short rates continued to march higher as the Fed did not relent (Figure 14). This disconnect re-emerged with the equity sell-off in October. This is the re-striking of the Fed put in practice. It is a reduction in the sensitivity of short rate expectations to financial conditions.
Having thus rolled up a number of the frameworks developed and expounded on over the years into a comprehensive assessment of the Fed as convexity manager during the re-emancipation of markets, Kocic then proceeds to deliver a four page “status check” which serves as the “outro” to this crowning achievement in analytical brilliance.
For months, analysts, investors and commentators have been tying themselves in knots trying to answer and otherwise address the only question that really matters right now which, roughly, is this: Why have virtually no assets (with the exceptions of USD “cash” and leveraged loans) managed to perform for investors in 2018? To be sure, there are some simple explanations, including, obviously, Fed tightening and the fact that “peak QE” is now in the rearview mirror.
What has perplexed many an astute market observer, though, is the apparent disconnect between the still strong U.S. economy (and really, the global economy hasn’t entirely rolled over yet either, despite the impression you’d get from looking at global equity benchmarks) and performance of risk assets.
Kocic answers this question, and the most amusing thing about his answer is that it follows so naturally from the analysis summarized above, that he delivers it at the tail-end of the note. In other words, the answer to what the rest of us consider to be an impossibly vexing quandary, is so obvious to Kocic that it’s basically just a footnote to him. Here it is:
However puzzling at first sight this disconnect may be, the preceding discussion suggests that it can be resolved, at least in part, through the dialectics of the unwind of monetary stimulus of the past decade. This dispersion has coincided with the turning of the fourth dial noted above — the Fed allowing tighter financial conditions — and is the natural consequence of the Fed stepping back more materially from its role as a supplier of convexity. In this context, the divergence between financial markets and the economy emerges as the logical consequence of the recovery process, a mirror image of the early stages of the crisis when monetary policy engineered a stabilization and a rebound in financial markets as a precursor of the subsequent economic recovery.
Now let’s hope Kocic’s analysis somehow got passed along to a Fed official or two, because it certainly sounds like he understands their plan better than they do.