Meme Stocks: Inaccessible Trading Share, Trading Cost, and Risk
Some say that retail may never have it better, but we can’t say the same for institutions. The analysis shows that Payment for Order Flow (PFOF) has taken valuable liquidity out of the marketplace. The result is significantly higher costs and volatility in “retail” names for mutual funds and hedge funds. With retail share at over one-third of market volume, the problem is too big to ignore.
Over the past year, we saw fundamental-defying trading activity in Hertz and JCPenney. There was a day in January when a handful of penny-stocks made up a fifth of all trading volume. Regulators to talking heads to politicians are still sifting through the wreckage of Gamestop and will be analyzing what it means for months to come. What is clear, however, is that the collective retail investor, especially when supported by market making firms like Citadel and Virtu, has more power than ever and significant influence over market direction.
While retail investors may never have it better in terms of myriad game-like apps that claim to democratize access to markets and payment for order flow arrangements that lock-in price improvement, the same cannot be said for the institutions that manage retirement funds for everyday Americans. Over this year, we have seen a new normal of heightened volatility where almost 40% of the shares traded in household names like Boeing, Roku, and American Airlines are inaccessible to traditional money managers and pension funds. Traditional alternative trading havens, such as dark pools and IEX, have seen their market share become compressed and traders have had to access lit markets to capture liquidity. As you would expect, this disruption comes at a higher cost.
We have found that impact costs for stocks with a high retail market share were three times as expensive as stocks with low retail involvement (Figure 1.) We also found that as retail market share increases, costs increase. Note that the range of results also becomes wider as the retail share increases, indicating that stocks with a greater retail share are also more volatile to trade. While it is possible that this difficulty can be due to unstable market conditions or trading tools that are incompatible with retail-heavy market conditions, these are both factors that stem from significant retail activity.
Figure 1: Execution price vs Arrival (bps) by Retail Market Share Percentage
Note that this data is from buy side firms from July to December 2020 and only represents orders with a one-day length. Multiple strategies are represented. Very small orders (<1% 5 day ADV) and larger orders (>10% 5 day ADV) were excluded.
It is important to keep in mind that commonly traded stocks have a high retail share, so this isn’t a niche problem. Apple currently has a 38% retail share and Zoom has a 32% retail share. Even Wells Fargo has a 22% retail share. As retail trading is not only not going to dissipate any time soon, it is likely going to increase, these significant findings should be a wakeup call to the entire industry. If traders and portfolio managers do not adapt quickly, the consequence can be costly.
Inaccessible Liquidity
Prior to April 2020, the overall retail market share was under 20% Due to a variety of factors, including people being subject to pandemic lockdowns and finding themselves with more disposable income as well as the heavy promotion of Robinhood, retail market share quickly rose to 25%, then 28%, and now is almost 35%. Evident in Figure 2[2], NASDAQ and NYSE are no longer the top two sources of liquidity in the US as they stood prior to the start of 2020. Now, Citadel consistently occupies the top spot—and it is important to note that a large percentage of the activity that is traded on NASDAQ has Citadel as a counterparty (note that only trades executed principally by Citadel would be in the Citadel bucket). While their retail market making business stands at approximately 15% of the total market, Citadel could represent up to 40% of total activity with the exhaust traded through other exchanges and dark pools. Consider that Virtu retail market making share has grown to almost 12% and there are many other retail shops. This drastic increase has left market share in other segments compressed, making it difficult for institutional traders to use their broker provided algorithms to source liquidity. There are no dark pools in the top 10 market participants and IEX is ninth. Instead of using dark pools and IEX, liquidity was more often sourced from lit markets, which tends to be more costly.
Data sourced from FINRA, Exchange data and third party sources via GTA Babelfish GPS Pre-trade
Trading algorithms are tuned to balance the contrasting objectives of liquidity sourcing and implicit cost. If urgency is desired, a trader will select an algo that is biased toward liquidity sourcing and vice versa if cost is the primary concern and patience is acceptable. However, even algos with a liquidity seeking mandate will still consider implicit costs. Although algorithms employ sophisticated techniques to minimize implicit cost, venue selection is almost always part of this strategy. “Lower cost” venues, typically dark pools and IEX are preferenced, however with certain stocks, especially those that have an inaccessible share that is above 25%, there is a greater chance that orders can be identified when trading in these lower volume pools, resulting in information leakage. Parameters like participation rate become distorted and the potential for signaling increased.
For example, assume that you have an order that you would like to execute at a 20% participation rate. If there is no inaccessible volume, you will execute at 20% of volume, and if volumes are normal, it will be done in the expected timeframe. As the inaccessible portion rises (Figure 3), you have two choices – either become more aggressive or take longer to execute the trade.
Figure 3: Inaccessible liquidity vs. Participation and Duration
If the stock has 40% inaccessible volume, then you will be 33% percent of the accessible volume (20%/60%) in order to achieve a participation rate of 20%. This will cause a much higher impact cost than traditional models would project. Alternatively, you can choose to be 20% of the accessible volume and the trade will take almost twice (1.66x) as long to execute, exposing you to time risk.
“But I am Not Trading In Retail Names”
Retail volume at the end of 2020 was approximately 35% of total market volume. Inaccessible liquidity is an issue that every institutional trader needs to understand. It is also an important facet of liquidity that Portfolio Managers need to factor in during portfolio construction and risk monitoring. Inaccessible liquidity exposes your portfolio to time risk and excess impact.
The impact is clear when a mutual fund manager owns a meme name. While there is a potential for improved returns, the plane can also crash. Additionally, rapid increases in market capitalization can require reducing a position and depending on a fund’s compliance mandate, timing might not be favorable. We have already discussed the difficulty of trading these names, which can be compounded by the volatility of news. There is clearly a risk of owning these meme names, but the question is how does a manager predict which names are vulnerable?
We would suggest looking toward inaccessible liquidity, and in particular retail market share. It is no coincidence that the two stocks that rose the most during the short squeeze last week, Gamestop and AMC, were the two heavily shorted stocks that had the highest retail participation before the short squeeze started. Stocks with high retail percentages could be particularly vulnerable to crowdsourced price action, increasing portfolio risk.
It is important to note that portfolio risk in retail heavy names is not restricted to short sellers. Short sellers merely highlighted the risk. Every time a portfolio manager sends an order to a trading desk to “participate at 20%” of volume, retail participation impacts the execution quality of that order. If the stock is 40% retail, like Apple, Gamestop, or Boeing, that 20% participation rate requires the trader to participate at 33% of accessible volume, thereby pushing the price of the security.
When portfolio managers assess the risk of their portfolios using metrics like days to liquidate, they make equally dangerous decisions. Take the same example as in Figure 3. The PM estimates that it should take 10 days to liquidate their position using a 20% participation rate. Due to the inaccessible share, it will take more than another week to complete that order, with the associated price risk. This also was highlighted during the Gamestop meme-ing. As ancillary stocks were liquidated to raise funds, the names with high retail participation declined more and cost more to liquidate than names with lower retail participation. Last week, S&P 500 names with more than 20% retail participation underperformed those with under 20% by over a full percentage point, led by names such as American Airlines, Tesla, and Enphase Energy.
GPS Pre-Trade Tool
GTA Babelfish has been particularly focused on inaccessible liquidity because of the relationship with higher trading costs. We have been working with our clients to help adapt to the increased retail market share by better understanding inaccessible liquidity and selecting more appropriate trading strategies. We have developed a tool, Pre-trade GPS, that can provide a portfolio-level view of the total and position-level time to unwind, emphasizing only accessible volume. This is useful to both portfolio managers, CIOs and compliance/risk officers. Please contact us at gtabsales@gtanalytics.com for more information.
