Read on Twitter
1/ Here is a long(ish) thread on why the possibility of repeated $GME-like short squeezes is bad for certain aspects of market structure.
2/ This is a complicated question, with many aspects, but I will focus on one aspect that may be unfamiliar to market participants, and that you will hopefully find interesting.
3/ You might ask, "who is being hurt by the GME squeeze?" And your natural response might be "only hedge funds that are short sellers, and short sellers are BAD!"
4/ The answer is much more complicated than this. I won't address short sellers specifically, but will address the question of how GME affects market participants that are primarily providing liquidity. First, a few assumptions that I make.
5/ 1) GME isn't actually worth $300 / share, and is priced here due to herding and some amount of a short / gamma squeeze. Given behavior of GME, AMC, BBBY, etc, all at once, I think this is a fair assumption.
6/ 2) Market makers and other liquidity providers are an undeniably GOOD thing for healthy markets.
7/ This question could be a thread unto itself, but market participants frequently want to trade close to instantly, and MM'ers bridge the time gap between when a seller and a buyer show up at the market.
8/ Bid / ask spreads are fractions (pun intended) of what they were years ago before decimalization and before markets transitioned to being fully electronic. Gone are the years where NYSE specialists made millions because of their unique position.
9/ You only need to look at the fact that GS bought IMC (a designated market maker) for $6.5 billion in 2000, and sold it for $30 million in 2014. Easy profits in the market making business are gone.
10/ So back to the original question of why the GME phenomenon could have dangerous side effects... I ran a stat arb hedge fund for a number of years, and continue to trade some stat arb strategies.
11/ Stat arb strategies essentially act as market makers, and provide liquidity to the market. We sometimes place two-sided orders (on both the bid and the ask), and sometimes, if we have a directional view, place orders on just one side of the market.
12/ We make money two ways: off the bid / ask spread, and off our directional prediction of an overpriced / underpriced security. It varies by strategy, but we make roughly 20% of our profit from the bid / ask spread, and 80% from being directionally correct.
13/ Our directional predictions are always based on mean reversion. As one grossly simplified example, we might pair trade INTC and AMD. We might assume that the two stocks should track each other very closely.
14/ If INTC is up 1% intraday, and AMD is up 0.5%, we might assume that there is a temporary imbalance, and that there are temporarily more buyers for INTC than AMD. Maybe a large mutual fund showed up to buy INTC today.
15/ So we assume that there is a temporary imbalance and that the relative price gap is statistically likely to close. We will thus offer to sell INTC at the ask and offer to buy AMD at the bid. We will provide liquidity to both the buyers of INTC and to the sellers of AMD.
16/ INTC and AMD are large, liquid securities, and I use them as an example that everyone is familiar with. But we, and other stat arb players, provide liquidity in many names that are far smaller.
17/ There are days where I am responsible for 5% of the volume in a given name, often largely taking one side of the trade.
18/ I am providing significant liquidity to the buyers / sellers of that security, assuming that in the next 1-10 days, there will be a liquidity flow in the other direction, and I will trade out of my position.
19/ Without someone like us willing to provide liquidity, it is far harder to find natural counterparties in the market, and bid / ask spreads will widen, and trades more than a few $K will move prices much more.
20/ You may be able to guess where this is headed. Even before the last week, a stock like GME is tricky to fit into the world of stat arb. I'm sure that some stat arb funds trade GME, but not us.
21/ However, the GME phenomenon is extending far beyond GME. One of the major baskets that we trade is a basket of REITs. We maintain market neutrality. So for every dollar of a REIT that we are long, we have a similar dollar short.
22/ There have been a handful of REITs that were picked up by the WSB community and have shot up far beyond what I'll call their fair value. My model wants to short that REIT and be long a similar REIT.
23/ When a particular REIT is 50% above it's fair value, I'll be very short that REIT. But I'm now looking at a new world where a REIT that is up 50% for no good reason may only attract more WSB buyers, and may be more likely than not to go HIGHER.
24/ That is a very scary world for someone who is trading stat arb. Stat arb margins are small, and are dependent on making millions or billions of trades that have a slightly greater than 50% chance of making a small amount of money.
25/ If WSB ushers in a new regime that greatly increases the number of large losses, I may pack up shop and shut down. Or I may completely rebuild my models in a way that greatly curtails how much risk I will take in any one name.
26/ Some of you will read this, and say "Haha. BOO HOO!!!! Market maker is leaving the market!" However, I hope that most of you understand that market makers and stat arb strategies truly do add significant liquidity to the markets and make trading much cheaper for everyone.
27/ It is a tricky concept that a market maker can regularly make money, and also make markets more efficient for everyone involved. And I won't go into that concept too much here.
28/ But it's 100% the case, and it will be a negative for market structure if stat arb funds and market makers have to significantly change how they operate or shut down altogether.
addenda 1/ There are liquidity providers who function at many different time scales. HFTs live at the shortest time scales, and are typically looking to hold positions for no more than a few seconds, maybe a minute at the longest.
addenda 2/ Stat arb strategies are the next step out. We'll hold positions for anywhere from a few seconds to a few weeks. The HFTs probably have done terrifically with $GME. The bid / ask spread is fairly large, and volume has been very high.
addenda 3/ I may be blinded by self interest, but I would argue that it is the stat arb players that provide liquidity that is generally more useful in providing liquidity. And the stat arb players operate on the time scale that is most hurt by the new WSB dynamics.
