Via Morgan Stanley:
Retail Squeezing Institutions – but Retail Demand May be Slowing
Another short squeeze, a new worst day for Growth vs Value (MSZZGRVL) after just setting that record a week ago, the 16th worst day for Momentum (MSZZMOMO), etc. and investors are asking “who’s buying (and does it continue)” and “does the rotation have legs”. Evidence points to retail being a big contributor to both the recent equity rally and to the pro-cyclical rotation. And both the rally and rotation likely continue over the medium-term – but that likely requires a handoff from retail to broader institutional sponsorship that could mean in the near-term there is some reversion.
The rational for a continuation of these moves is simple – HF and other institutions are still running light positioning, still heavily tilted towards Tech/Growth over Cyclicals/Value, and still concentrated in the same names (QDS’s crowding metric is at all-time highs), meaning the pain trade is a cyclical driven rally. But that move may not come all in a straight line and there are signs that retail demand is slowing, systematic strategies likely won’t relever aggressively with volatility in the high 20% range, and discretionary institutional investors still appear relatively hesitant to chase. QDS suggests investors hedge near-term downside risk with SPX July put spread collars, while using IWM November bullish risk reversals to position for medium-term upside and a cyclical rotation.
While it’s probably an oversimplification to suggest retail has been the only buyer, many metrics suggest they are at least a big contributor. Total stock positions at online brokerage Robinhood (sum of users across all Russell 3000 stocks, 1st chart below) spiked higher starting in March, Daily Average Trade volume at other brokerages is elevated as well (2nd chart below), and premiums spent on new option trades are heavily tilted both towards small orders and towards call options (3rd chart below). Retail is also heavily adding to Cyclicals and likely squeezing consensus shorts – using Robinhood as a proxy (4th chart below) Cyclical (MSXXHBC) total positions are up 220% since the end of February, Defensive (MSMSDEFS) total positions are up 85%, and Tech (NDX) positions are up 55%.
While these flows have been supportive of stocks over the last two months, the growth rate in the online platforms is slowing (1st chart below), and if demand doesn’t accelerate elsewhere the rally could fade. In addition the retail demand appears to be more focused on dip-buying – stocks that underperform on any given day are seeing higher user growth on Robinhood than stocks that outperform. While that dynamic may prevent big downside moves, with fewer and fewer dips to buy as markets move higher it may result in less new money coming in unless behavior changes.
Looking at retail flows another way, retail buying appears to be playing a role in intraday price action with US equities – more stocks being bought generally coincides with intraday (open to close) rallies (1st chart below). That clearly doesn’t explain the strength in overnight returns (2nd chart below), so retail can’t be blamed for everything here, but any slowdown in retail buying does raise some risks to stocks.
On the other side of the market, institutional investor positioning still appears very light. Systematic strategies have only been modest buyers and leverage remains near the lows – and likely remains there in the near-term as volatility remains high. Futures positioning remains very light and most shorts have not been covered. And US L/S HF net leverage per the MS PB Content Team remains low at 45% (15th %ile since 2010) even though that is nearly the highest exposure since late March. And it would be hard to suggest all of retail have been big buyers too – the larger ‘slower money’ retail community that doesn’t trade as actively online has $1.6tr in MMF balances and equity mutual funds continue to see withdrawals (albeit modest).
The retail vs institutional dichotomy extends to the sector level as well. While retail is buying the dip in cyclicals, ETFs continue to see inflows into Tech more than Cyclicals (1st chart below) while HF positioning per the MS PB Content team (2nd chart below) continues to tilted away from Cyclicals.
Extreme sector positioning is also compounded by a high level of crowding amongst HFs in the same stocks. QDS’s crowding metric reflects the degree of overlap in L/S hedge fund portfolios based on public 13F filings – more funds with very similar portfolios means crowding is high, because it increases the chance that if there is a shock to one fund and they derisk their book, investors with similar portfolios suffer, derisk their books, and funds chase each other lower selling the same stocks at the same time (see Crowding Dec 3rd 2019 for details). While somewhat dated in a high volatility market, the level of crowding at the end of 1Q based on this methodology was the highest on record. Since then, performance for the crowded longs has been very strong as measured by the MSXXCRWD basket, which was rebalanced last week to incorporate filings from 1Q.
Lastly the signaling from the big factor moves that are playing out over the last few weeks is negative – for the consensus trades. Growth versus Value had its two worst days on record (since at least 2000) over the last week, and while prior large selloffs were largely concentrated in 2000/1, 2008/9, and over the last 6 months (i.e. the datapoints are overlapping) they do suggest further pressure on Growth versus Value in the future.