For those of us who trade every day, it’s easy to dismiss what goes into supporting a listing. We have humans to help with investor relations questions, systems to assist with governance, tech to surveil the market, and we also advocate to improve capital formation. In addition to a great marketing team that helps listed companies grow their brand via their listing.
But those who trade do care about the benefits of good market structure and cheaper trading. And that’s the focus on today’s data. Using stats from the recent IBKR listing switch to IEX (and back to Nasdaq on Oct. 7), we can see results have important implications for many of the open market structure debates right now.
Some facts on the rebate debate
IEX has never been shy about criticizing rebates. That’s somewhat ironic, not because they offer plenty of their services “free,” but because they often charge takers the max 30mils on a sweep, creating 33mils in revenue for themselves on liquidity-taking trades.
Maker-taker markets also charge high fees to liquidity takers, but instead of pocketing them for themselves, they offer rebates to liquidity providers for improving market quality. Our earlier research has shown that this results in consistently cheaper spreads for investors. That’s arguably a “market good” as it also improves off-exchange executions. More importantly, we showed that large takers might even benefit from the additional liquidity in the queues, acting to reducing the costs of liquidity by as much or more than the cost of the rebates.
The data from IBKR’s time listed on IEX seems to confirm this. IBKR’s spreads increased 83% on average following their switch to IEX. Over the same period spreads on S&P500 stocks increased just 4%.
There is plenty of research that shows that investors are hurt by wider spreads. We estimate that over the past year, investors may have spent over $1million more than necessary, just in trading IBKR.
Chart 1: NBBO was almost twice as wide on IEX
Source: Nasdaq Economic Research
Market makers aren’t to blame
It’s tempting to say, “maybe the market makers should have made better markets for IEX.” But our data shows that electronic market makers moved their quotes to IEX almost instantly improving IEXs quote in IBKR from being on the NBBO 12% of the time to around 96% of the time.
Interestingly, that resulted in IEX having even deeper queues at the NBBO that before—something IEX has said in the past hurts investors who aren’t fast enough to “set” price.
Although that fact is irrelevant if IEX really believes that all investors send every trade across the spread.
In reality, we all know that all investors trade with a combination of spread capture and liquidity taking orders. Either way, data show IEXs market structure wasn’t fair to investors interacting with their lit IBKR quotes. Their 38% deeper queues would hurt passive investors, but because the spreads were 83% wider the cost of liquidity for investors crossing the spread was actually much higher.
Chart 2: NBBO depth did not increase enough to offset the wider spreads on IEX
Source: Nasdaq Economic Research
Higher trading costs hurt market makers and liquidity
One big question is “why didn’t market makers make spreads the same” when IBKR switched exchange.
The answer is economics. Market makers, almost by definition, need to do a lot of quoting and trading, often holding positions for short periods.
Capturing the spread also means they have very small expected return per trade. The most that a two-sided market maker will make is the whole spread, but adverse selection and fees can quickly reduce those profits.
That means market makers are especially affected by per-trade costs, but benefit from being able to achieve economies of scale.
IEX has frequently said that lower fees for more active traders aren’t fair. To support their views, they (mostly) don’t charge for fixed costs of data and ports and don’t offer co-lo. Instead, their business model focuses on both sides of a trade paying on a “per-trade” basis. Our previous analysis has shown that they are the most expensive exchange on a per-trade basis.
That’s not very fair to market makers, who need to capture wider spreads to cover these costs, as this data seems to confirm.
Data also shows that as IBKR spreads widened, market-wide liquidity also fell 27%. That makes sense, as reduced trading spreads can be linked to increased liquidity in the market. Higher costs also mean investors can’t accumulate as large a position without eating into their expected outperformance.
Chart 3: IBKR liquidity fell significantly on its IEX listing, especially in the close
Close volatility spiked too
Reduced liquidity on IEX may have had its biggest impact on investors trading the close (MOC). Our data show that after switching to IEX, the IBKR close probably cost investors around $0.5 million:
- Volumes ahead of the close fell over 39% while volatility more than doubled.
- MOC volumes fell from almost 10.7% to just 5.1% of ADV.
- With that, came additional volatility and market impact from MOC trades. The dislocation between continuous trading and the official close more than quadrupled from 2 cents to an average of 9 cents per day.
As our own study found, prices move into the close because of supply and demand imbalances. That means natural investors likely incur these liquidity costs. We estimate the index funds and others that needed to invest in the close likely found their trading costs increased around $0.5 million on IBKR alone.
Chart 4: A less liquid close increased volatility and added costs to index funds
Source: Nasdaq Economic Research
Higher trading costs hurt issuers too
Wide spreads affect tradability in a variety of ways. Our recent study of the stock split premium found that the market discounts the valuation of companies with wider spreads and higher trading costs.
Academic research linked to that study found that wider spreads and worse tradability were reflected in increased weighted average costs of equity capital.
IEX has created a lot of negative stories about how the U.S. market operates. Their rhetoric has driven policy on rebates, speed bumps, data costs and port pricing.
Yet the results of their listings experiment seems to show that putting many of those policies in action seems to do more harm than good, at least to issuers and investors.
It’s easy for any of us to say the market isn’t fair when we can’t get what we want. It’s much harder to build a market that balances the needs and costs of quote providers, liquidity takers, issuers and investors all at once.
Perhaps the main lesson, though, is that current market economics actually do that pretty well.