Interview by Matthew Dove
2017 was a rough year for Manoj Narang. As the quant market faltered, Narang’s fund MANA Partners had the plug pulled by its primary investor, JPMorgan, in its first year of trading. Having made an 8% loss for 2017, this determined band of quants fought back. Here, senior editor Matthew Dove asks what the forecast looks like for MANA, and the market more generally, now that the storm clouds have begun to clear…
TFT: What happened to the quant market last year and why?
MN: Quant has become a highly bifurcated space lately. Some firms are doing well, but in my opinion, most are struggling. This has been going on since the end of last year. As for why this is the case, I believe there are several reasons.
(1a) in Q4 last year, the market sold off precipitously. That is not a great climate for most equity market quants. The reasons are technical, but I will attempt to explain here. The first reason why down markets are tough for quants is that when markets go down, implied volatility goes up. When implied vol goes up, value at risk goes up. This leads to de-levering.
Equity hedge funds can often be highly levered (6x is fairly common). When they reduce leverage due to greater value at risk, they do not create any market impact in beta or other directional factors; they strictly hit the equity cross-section (i.e. alpha) because that’s what they hold in their portfolios. Thus, that adversely impacts other equity market neutral managers who did NOT yet de-lever. This can create a bit of a chain reaction.
(1b) Another reason why down markets are tough for many equity quant managers is that when volatility goes up, correlations between stocks also go up. That means stocks temporarily stop moving idiosyncratically based on their stock-specific information flow (earnings, fundamentals, etc); instead, they tend to move in lock step with the market.
Those kinds of environments can be challenging for market neutral strategies because they hedge all systematic risk factors, and instead focus on explaining stock-specific movements. When correlations are extremely high, the amount of stock-specific movement in the market is minimal, making it hard to generate enough alpha to overcome t-costs. This is exacerbated by the fact that t-costs themselves can be higher when volatility is high.
(1c) The combination of the previous two effects can often lead to drawdowns. To mitigate drawdowns, some managers may de-lever further, which can create a positive feedback loop leading to more losses across the industry.
Quant has become a highly bifurcated space lately. Some firms are doing well, but in my opinion, most are struggling.
TFT: Is the quant market slump a sign of worse to come?
MN: I don’t think so. I think mechanical investment styles have been gaining market share every year for at least three decades, and I highly doubt this trend will stop any time soon.
That said, I believe the space is full of landmines due to the ongoing threat of liquidations. For instance, this March/April, from my understanding, there were billions of dollars of liquidations of equity market neutral books at some of the biggest quant managers. I believe that definitely created temporary pressure on returns, but those pressures always historically mean-revert after the liquidation ceases. In my opinion, there will always be an ongoing continuous shake-out as new managers get funded, many of whom don’t perform well, and then they have to shut down, which creates pressure on performance of other managers. However, the flip side is there as well — as new money comes into the market, that creates tailwinds for other managers as that capital gets deployed.
TFT: What is the status of MANA Partners now?
MN: The manager has been in continuous operation since it launched in 2017. However, I was pretty unhappy with what we accomplished in the first year we launched, so we pretty much did a massive overhaul of our strategies and systems at the beginning of 2018. I’m very heartened by the strategy’s performance since the completion of the overhaul last June, especially how well it fared during the volatile fourth quarter of 2018. Now that we have four quarters under our belt since the overhaul, the strategy has begun generating considerable investor interest.
I was pretty unhappy with what we accomplished in the first year we launched, so we pretty much did a massive overhaul of our strategies and systems at the beginning of 2018.
TFT: What remedial measures were taken?
MN: If this question is about how we “fixed” our business after the rocky start in 2017, the answer is that we retained our best and most productive people and everyone worked extremely hard to turn things around. We had a very ambitious blueprint, and it took us 18 months to pull it off instead of the 12 we had hoped for initially. This is a very technology intensive business, and you can’t build something successful overnight in such a hyper-competitive space. It takes time, and it took us a bit longer than we hoped. But we’ve managed to pull it off.
If the question is about what we do to protect ourselves in markets that are challenging for other quant managers, there are several things:
(1) we seek to build economically sound, top-down strategies rather than what many in the industry seem to be working on these days (i.e. AI / ML). Investors have really miscalculated, in my view, by pouring money into AI strategies. I believe those are now extremely overcrowded and highly undifferentiated. Not to mention, I believe predicting of security prices is one of the least compelling applications of AI I can think of.
(2) we understand that simply being market neutral is not enough protection against down markets, because of the effects I described above related to de-leveraging. The best protection against that is to do unique things that are not overcrowded. The next best thing is to include a healthy dose of “risk off” signals into your alpha mix. These are strategies that we believe are explicitly designed to do well in falling markets Finally, we also run HFT strats alongside our stat arb strategies. These are very pro-volatility and hence, are typically considered protective when the market melts down.
TFT: To what extent is increasing geopolitical uncertainty causing funds to take a disproportionate number of short positions?
MN: If you’re talking about quant funds, I believe the biggest component of that space is market-neutral, so there are no “extra” shorts — EVER. For instance, the strategies seek to have approximately 50% of our capital in short positions, regardless of market conditions. We just have a different set of ways to go about managing that uncertainty from our long-only cousins. We operate several strategies that specifically benefit from this sort of geopolitical uncertainty. I’m not going to tell you how we do that, but I believe it is an advantage. In addition, we have very interesting ways to measure stock-by-stock sensitivity to effects like trade wars, Brexit risk, and other similar phenomena. Once you can measure such sensitivities, you can hedge if necessary.
For instance, the strategies seek to have approximately 50% of our capital in short positions, regardless of market conditions.
One thing I will say is that the market action has demonstrated fairly conclusively that investors are even MORE worried about missing out on upside than they are about experiencing downside. This is no doubt because of the dual effects of the “Fed Put” and the “Trump Put”. The Fed has made it clear that it has Trump’s back on the Trade wars, and that if these exert downward pressure on growth or employment, they will “act accordingly”. This has further cemented the already ingrained “buy every dip” mentality among investors. Similarly, people know that Trump’s policy actions are a bit fickle and tenuous, and can be unwound as quickly as they were initiated. That also creates upward pressure on dips.
One big risk I see lurking in the market is the election. Polls (as usual) are showing the Democrat party with a big advantage. If investors buy into this thesis, and if Biden loses his front-runner status (which I see as likely given the zeitgeist and the crowded field), then this could create a self-reinforcing downward momentum spiral. I believe if the market becomes convinced that democrats will win the presidency and unwind Trump’s corporate tax cuts, the market will fall, which could be the catalyst that precipitates a recession. A falling market will reduce Trump’s re-election chances, which will only further strengthen those fears, creating a feedback loop.
We could be in for some significant volatility later this year as a result of this.
One big risk I see lurking in the market is the election.
TFT: Could the downturn make a recession inevitable:
MN: IF the market falls back to its lows of Q4, I think that will lead to a recession. I’m skeptical of a deal with China; in my opinion, they know that if the trade wars hit US Growth and cause a recession, Trump is likely to be kicked out of office, and they’d prefer to negotiate with the next president. Even if a global recession hits China asymmetrically hard, it will be temporary and the Communist party doesn’t really need to worry about re-election. So, I think both sides are pretty much dug in, unless Trump caves.
Anyway, I think the prospect of the democrat party nominating an unabashed socialist to contest the white house is DEFINITELY enough to trigger a recession. So I believe it really depends how the debates and polls play out on the Democrat side in the next several months.
TFT: What role will AI play in the future of quant trading?
MN: AI is like anything else — there is a high degree of skill differentiation amongst the players. I’m a significant AI skeptic, but I still sit on the board of an AI fund because those guys happen to be extremely skilled and quite differentiated (which I also try to help them with).
Overall, I think the role of AI in finance today is HIGHLY overstated. Thanks to Google’s (and a couple of other firms’) advances in image recognition, search autocompletion, chess and go playing, text processing and other advances, I think investors naturally assume stock trading is next. I do not believe Stocks are the same thing. I believe a good test of what tasks you could expect an ML program to be good at; If a human can be trained to be good at something, there is a good chance that a computer can also be trained to be good at it using ML algos.
I think the role of AI in finance today is HIGHLY overstated.
Every successful application of AI thus far is something that humans can be trained to excel in. Chess, driving, understanding a sentence, etc. However, can humans be trained to be good at stock picking? I don’t think there is any evidence that this is the case; in fact, I believe there is a mountain of evidence that it is NOT the case, which is why passive investing and indexing have become all the rage. If you can’t train a human to do something well, how on earth will you train a computer to do it? The answer is that you CAN’T.
If you can’t train a human to do something well, how on earth will you train a computer to do it?
In my view, we are a long way off from AI taking over for traditional quant. In fact, I worry about something totally different — that when investors totally and irrevocably sour on AI in quant trading, which I believe to be inevitable, that this will tarnish the entire quant space for years, including strategies that have nothing to do with AI. This happened after LTCM blew up in the late 90s — that caused the entire quant industry and quant pedigree to temporarily be out of fashion.
Longer term, however, I believe quant will recover. Technology can not be stopped, nor can progress. And, you can bet that after this AI bubble bursts, there will be another one in 5-10 years after that…
