Logica Capital commentary for the month ended February 28, 2023.
Summary
Set the stage by telling the reader what the main topic is. Your main story should be at the top of the email with a CTA button that links to a corresponding web page or blog post. Don’t include the entire story in your email.
Commentary & Portfolio Return Attribution
“Google can bring you back 100,000 answers. A librarian can bring you back the right one.”
– Neil Gaiman
February was a bit of a perfect storm for our strategies. While we didn’t experience much payoff from volatility’s mild upside move, the S&P 500’s drawdown wasn’t quite enough to dig us out of the “valley” of the straddle. The “valley” of the straddle is the area below water, i.e. the cost of the options’ premiums that must be recovered by enough of a move in the price of the underlying. The straddle valley becomes the greatest source of pain when there is not enough magnitude of a move in the underlying to enable volatility (or convexity) to kick-in, but when there is still enough of a change in delta/direction of the underlying to push one side of the straddle lower than the other side. In February, this valley amounted to about a -2.6% move in the S&P, and some degree more than that within various sectors, as discussed below.
So what contributed to the weight that pushed our performance down? There were a few different contributors worthy of discussion, so we’ll jump further into each. Albeit, before diving into the more powerful contributors, it is important to note that the performance of implied volatility in February was not to blame. Implied volatility was certainly below the norm for an S&P 500 move of -2.6%, but only slightly below:
Thus, setting Volatility aside, our strategies suffered from three major areas in February, which we shall discuss in greater depth:
- Factor Rotation
- Gross Notional Exposure
- Macro Headwind
Factor Rotation: Our higher sector exposure in our Sector & Single Stock Calls module was to energy (XLE, -7.1%), which was the weakest sector in the S&P, and which greatly influenced the greater decline in the Dow (DJIA, -3.9%), as the weakest of the 4 major indices. As mentioned several times last year, it was our overweight exposure to the energy sector that enabled outperformance of this same module in 2022, with energy holding up far better than much of the S&P during last year’s extended decline in the broader market. However, the more recent factor rotation saw this relationship flip. Our momentum models do not typically react that quickly, so we’ve been caught in this flip for 1-2 months now. More specifically, we saw underperformance with our momentum sub-module given its overweight in energy. Concurrently, the anti-momentum/diversified module took hits from its positions in Utilities (XLU, -5.9%) and the Biotech Sector (IBB, -6.8%). More broadly, with technology strongly outperforming the broader market – and other major indices — in February (QQQ, -35bps), our defensive sector positioning that saved the day in 2022 has thus far hurt us in 2023.
So do we think this factor rotation is finished? Do we hypothesize that this was likely a mild correction of the previous outpacing, and healthy for a broader resetting of relationships back to acceptable levels? Or do we feel the rotation is the new order and should now make way for another corner of the market to take the lead into 2023? Fortunately, as a systematic fund, we don’t have to make these types of decisions. Having rigorously assessed decades of empirical data and market behavior, our models are built to answer such questions by optimizing our positioning to the highest probabilistic expectation. To this end, it’s nice to see that our models did call for a portfolio level shift early in March (March 8th to be exact), which has since been beneficial.
More specifically, the reduced level of Implied Volatility (“IV”) for a prescribed period (after several years of higher level IV) triggered our “fast-scalping S&P Calls” module, as if screaming “let me in” while banging on the portfolio door. And so, once fast-scalping kicked in, our portfolio construction exposure guidelines required us to reduce our Single Stock & Sector module to let the Fast-Scalping module in. Accordingly, we reduced the Single Stock & Sector module’s notional exposure by 50%, in exchange for an allocation to the Fast-Scalping S&P module with that freshly available notional exposure. Thus, whether we do or don’t believe that Energy, Utilities, and Biotech are overdone or underdone, half of that exposure is now delightfully participating in higher frequency volatility scalping, which has improved total portfolio performance since the transition. And of note, after existing without the Fast-Scalping module since approximately mid-2020, we have been eagerly waiting for it to kick back in, so were quite excited to see this happen, and are now even more excited to experience the fruits of this sub-strategy over the foreseeable future. And higher level, we are of course happy to further diversify our up-capture strategies given the last few months of underperformance from our factor exposure models.
Gross Notional Exposure: Over the past several months, our top-level portfolio construction model has had us increasing our gross notional exposure in line with the reduction in Implied Volatility levels. Quite simply, we scale up our portfolio when holding more volatility/vega becomes increasingly attractive, e.g. as it gets “cheaper.” Looked at analogously, we see volatility as a “spring” which upon being fully released is fully extended and so has no upside left. However, as one begins to push down a spring, there becomes more available upside (or bounce) upon letting it go. And at its extreme, if one were to fully compress a spring into its most compacted position, there is literally no more room for it to squeeze more together, but on the other side, a large amount of upside bounce for it to go once released. Implied Volatility shares this general characteristic in that as it gets cheaper (perhaps closer to 2017 VIX levels of 11, 12, 13), it has both less likelihood of further declining (could IV get to 8 or 9 – or half the historical average for the S&P at about 16? highly improbable), and has even more available right skew when it pops (12 to 40 is far greater of a spike, of course, than 25 to 40).
With this quantitative framework in mind, the lower IV levels get, the larger the opportunity set, and therefore, the further our model will increase our gross notional exposure. Obviously, while based on cheaper pricing and greater potential upside (an improved risk/reward ratio with a powerful lens targeting right skewness), the increase in overall notional exposure will exacerbate the ups/downs as we move forward, and already increased the magnitude of the down month we experienced in February. Simply speaking, with more exposure, our portfolio will move at higher variance. That said, we are excited at the improved risk/reward of volatility levels – and fatter right skew — and what they can potentially offer on the desirable side of the coin. For reference, below is a graphical view of our increasing notional exposure over the last few months.
Macro Headwind: Lastly, we experienced a headwind from our Macro Overlay during February, with gold having its worst month (-6.5%) since June 2021. Further, long-term Treasuries also suffered in the month (-5.1%). As is planned/expected, when those two assets within our risk-off macro module draw down, our 3rd exposure in USD typically swoops in to buoy the whole module and provide some much-needed positive PnL. And while USD did help to some degree in February (+3.3%), its gain was still not enough to offset the poor performance of both Gold and Treasuries.
“Problems are solved, not by giving new information, but by re-arranging what we have known for a long time.”
– Ludwig Wittgenstein, Philosophical Investigations
As many of you recall from our writing or have experienced directly in other parts of your portfolio, there were similar difficult months in Treasuries and Gold during 2022’s market decline. And so, the re-emergence of this phenomenon leads us to some frustration around the reliability of these risk-off assets in the face of equity market downturns. Empirically, we see many decades of high reliability (of both negative correlation to equity market stress, and more powerfully, some convexity on such occasions), alongside out of sample “misbehavior” in 2022, leading us to stand behind their long-term flight-to-quality behavior and point to the recent environment as more anomalous given inflationary pressures.
Moreover, as of mid-March, Gold and Treasuries have rallied nicely, and as a result, our macro module has recovered nicely. At the same time, given some further digging into the “2022 anomaly” and the surrounding macroeconomic issues that the underlying markets have been, and still are, facing, we observe that from 1926 to May 2022, when inflation has been greater than >3%, equity and bond returns were positively correlated approximately 90% of the time. To be clear, and while it’s still a worthwhile view, this chart speaks to bonds in general, which infuses credit spreads and duration as related risks, whereas Treasuries (TLT) are of course pure rate exposure at long duration – a major bifurcation that differentiates risk-off (flight to quality) from risk-on (credit spreads/yield curve):
Relatedly, we can see that this relationship between TLT and SPX has been positive for some time now:
The significance of this with respect to our strategy is that we can see one of the popular “pillars” of hedging/diversification has not been so for quite a while, and its future as such is seemingly in doubt. From our perspective, a lack of reliability from these risk-off assets under recent market conditions leads to our increasing conviction in the negative correlation of Volatility and its high reliability during risk asset declines. More so, and as touched upon above, as Volatility levels get cheaper (relatively speaking), we see greater opportunity and potential upside in carrying an even larger exposure to Long Volatility’s “spring” potential. The more right skew, the merrier!
Moving on from the portfolio attribution, after printing its single lowest close of the “post-Volmageddon” era in January, the CBOE VVIX Index6 once again trickled downward, closing at its low point on the month.
“Strategy is actually very straightforward. You pick a general direction and implement like hell.”
– Jack Welch
Relatedly, we saw a very interesting – almost perplexing — phenomenon among the EurekaHedge Volatility Indices in February, wherein both Long Volatility and Short Volatility Indices were concurrently down! In other words, in February, it seemed difficult to make money in either direction. To state the obvious, one would assume (and would normally see), one up when the other is down. Seeing both long volatility and short volatility indices down speaks to the overall complexity of trading volatility during certain regimes.
Finally, taking a closer look at the daily movement of our strategies for the month of February, we see what we’d generally expect, with LAR participating mildly, and LTR holding ground in the market’s gain, while remaining net short, with bundles of convexity in the waiting.
