OWe recently reviewed the LULD rules and found that exchange-traded funds (ETFs) have triggered only a fraction of all LULDs over the past two years. This is partly due to the diversification benefits of ETFs, which represent a portfolio of underlying stocks. We have shown that the volatility of ETF tickers is usually a fraction of the volatility of one of the underlying stocks.
We then compared ETFs in Tier 1 and Tier 2, and we find that:
- Volatility is virtually identical (Graph 1, left). Typical all-day price ranges for most tickers in both groups are below 2%.
- Liquidity is lower in Tier 2 ETFs. This is by definition since liquidity is how ETFs are allocated level.
- The spreads of some Tier 2 ETFs are widerbut despite the lower liquidity, the majority of level 2 spreads are below 50 basis points, which is still a fraction of the smaller level 1 band (chart 1, right), meaning that a narrower band should not be triggered by wide gaps.
This got us wondering if it would make sense for Tier 2 ETFs to be combined with Tier 1 ETFs?
Chart 1: Tier 1 and Tier 2 ETFs have almost the same volatility and spreads that are well within the Tier 1 bands
What would happen if the LULD ETF groups were merged?
Currently, Level 2 ETFs have bands twice as wide as Level 1 ETFs (10% instead of 5%) for most of the day. Although after 3:35 p.m. Level 1 stocks switch to “double-width” bands, which means that Levels 1 and 2 stocks already have identical width bands at the end of the day.
Because we are modeling tighter LULD bands, we are able to use actual trading to see how often actual Level 2 ETFs would have reached Level 1 bands. Our analysis covered all dates in 2020 and 2021 (except MWCB days). The results suggest that moving Tier 2 ETFs into Tier 1 would increase their LULD stops tenfold, from just 56 to 565 last year.
Chart 2: Placing Tier 2 ETFs in Tier 1 would significantly increase LULD stops (excluding MWCB days)
While that may seem like a lot, that’s less than two a day in 2021. It also follows that Tier 2 ETFs (with Tier 1 bands) have a similar probability of a breakout as a Tier 2 stock ( with level bands 1) .
Why is the increase in LULD so important?
Since Tier 2 ETFs were no more volatile than Tier 1 ETFs, we wanted to understand how LULD stops rose so much.
It turns out the answer is in LULD math. Remember that the LULD bands are based on the average trade prices of the previous 5 minutes. For Level 2 ETFs, this turns out to be very important because Level 2 ETFs are the least liquid and risk-free stocks. any trades, LULD bands are not updated.
Instead, what happens to many Level 2 ETFs can be seen by the example of Chart 3 below, which plots the De-SPAC ETF (Nasdaq:DSPC). Note that DSPC is the first ETF to offer portfolio exposure to completed SPACs, so it only has US-listed stocks in its portfolio, which trade at exactly the same time as the ETF. This means that market makers can calculate an accurate “real-time” net asset value, allowing them to quote with tight spreads and frequent price updates (what we see in Chart 3 from the lines green which move a lot but remain very close to each other):
- Light green and dark green lines indicate bid and offer. The fact that they are touching shows that the gap is very tight. The fact that they move shows that market maker prices are very frequently updated to match NAV.
- But there are no yellow dots (no round lot trades) until the end of the day.
- This lack of trades also meant that there weren’t even any arbitrage opportunities. Thus, the ETF’s net asset value rose steadily throughout the day, rising from around $21 the night before to close to $22.50 at the close.
- However, the LULD bands have not been updated (dark blue represents the current level 2 bands calculated by the SIP).
- Due to increased supply during the day, quotes eventually hit our “virtual Level 1 LULD” (light blue lines), which would have caused the limit to stop around 1pm under our new proposed scheme (red quotations).
Graph 3: Modeling the operation of half-bands for level 2 actions
Importantly, LULD’s additional stop auction helps reset bands by:
- Reopening a trade will reset the LULD reference price to the trade price.
- If there is no trade and the LULD Stop Auction reopens, the LULD Reference Price is based on the midpoint of the primary listing exchange’s quote.
- In the unusual case of a LULD stop auction without a trade or bilateral quote, the LULD benchmark price will reset to the price range that was in a boundary state before the trading pause (the red prices in Chart 3).
Ironically, since some of these ETFs are less liquid and have wide spreads, and may NOT be able to calculate a live NAV (like we can for DSPC), the third option above might actually work better than the second choice (using the midpoint) even in an auction without trade.
In fact, the LULD plan has already changed its rules to not use “quotation midpoints” in the official opening, as it sometimes takes the bands away from fair market levels, resulting in more LULD when the first trade real took place.
Most of the new LULD stops aren’t bad at all
Although halting trades more often sounds bad, as it limits investors’ liquidity, our analysis shows that it actually affects very few trades and would in fact affect very few investors.
In fact, the data shows that for more than 91% of additional “virtual” stops, no exchange took place. These stops where exchanges actually took place represented only 0.035% of all Level 2 liquidity, which, it should be remembered, only includes the least liquid ETFs.
Our results also suggest that half (49%) of these additional stops occur within the first 15 minutes of the day, as NAV calculations and ETF quotes are updated for the day’s market news at next day.
Chart 4: Even with more stops, very few trades would be affected
It therefore seems that the impact on liquidity and investors would be minimal.
But more importantly, by resetting the LULD bands to stricter Level 1 bands for all ETFs, the guardrails should prevent more trades that fall far short of the ETF’s fair market valuation, which is good for investors. .
What does this really mean?
ETFs are both stocks and stock portfolios. For this reason, the liquidity of ETFs does not necessarily mean that ETFs will be more volatile or harder to trade in size.
LULD bands create guardrails that protect investors from bad trades and runaway markets. It follows that investors could benefit from better protections if all ETFs used level 1 bands.
However, we estimate that placing all ETFs in Tier 1 would trigger around 10 times more LULD than we currently see. Turns out it’s because Tier 2 ETFs are often thinly traded, and without trades, LULD tapes may become obsolete.
However, since Level 2 ETFs are often thinly traded, these additional stops would affect trading very little. Instead, the additional stop represents an opportunity to reset the band near market prices. And that, overall, should be good for investors.