Who is Trading on U.S. Markets?

By Phil Mackintosh, Chief Economist, Nasdaq

The U.S. equity market is the largest and most liquid stock market in the world (Chart 1).

As of year-end 2019, the market cap of publicly traded companies listed in the U.S. totaled almost $38 trillion, representing nearly 40% of total global market capitalization. On an average day in 2019, trading in company stocks (excluding ETPs) adds to $233 billion. That results in turnover, multiplied by a company’s market cap trades each year, which is more than double what we see in Asia or Europe.

Better liquidity for all companies should result in lower trading costs for investors and lower costs of capital for issuers.

Chart 1: The U.S. equity market is the most liquid in the world (2019 data shown)

Which raises a question: “Who is doing all this U.S. stock trading?” 

The answer isn’t obvious, as very little trading is attributed to counterparties in public data sources. But using a number of sources, the estimates we create below show that all the participants in the U.S. markets ecosystem have different and important roles to play. These different players keep trading spreads low and prices efficient, and transfer liquidity between markets rapidly and cheaply.

Chart 2: The U.S. market ecosystem shows lots of participants contribute to liquidity

Our estimates are shown below in Table 1. But there are a few things to highlight before we get going:

• Liquidity is relevant to both buyers and And risk is hedged in dollars, not shares. So we look at the value of gross liquidity (trades x 2). So trading of $233 billion per day actually represents liquidity of double that, or $466 billion per day.
• Because most of the data from 2020 isn’t available yet, we’re basing these estimates on 2019 trading and assets. Total liquidity in 2020 actually increased 50% to almost $700 billion per day, thanks to an increase in retail trading and volatility caused by COVID-19.
• Public data becomes harder and harder to find the more we progress down the list in Table 1. Accordingly, our estimates would increasingly adjust once more data becomes available.
• We don’t include ETF trading, even though they are NMS stocks, which would add another $88 billion per day, or $176 billion in liquidity.

We also try to separate our estimates by strategies available to investors (“natural” investors) and those deployed by liquidity providers (intermediators).

Table 1: Estimating who contributes to market liquidity

So, let’s walk through how we arrived at these estimates.

How much do investors trade?

As Chart 3 shows, investors represent a small section of the market ecosystem, although data shows that they own the majority of stock market capital.

Around two-thirds of corporate shares are owned by U.S. households via mutual funds (22%), pension funds (11%) or retail brokerage (34%) accounts. And an additional 16% are owned by foreign investors, likely mostly through professionally managed investments too.

Chart 3: Ownership of U.S. stocks

How much do mutual funds trade?

Let’s start with professionally managed investments: mutual funds (domestic and international), ETFs and pensions. We know their assets are around $21.2 trillion (Chart 3).

ICI tracks the average turnover of all mutual funds. Their data shows that average turnover has been steadily declining for most of the past 35 years and now sits below 30% per annum each side.

Some of the decline is due to the rise in index funds. As we discuss below, they have a much lower turnover. Even when we account for that, we estimate active trading has still fallen to around 47% each side, equating to longer average hold times even in active institutional accounts.

Chart 4: Mutual fund turnover has declined to around 30%

Knowing that mutual funds tend to stay close to fully invested, we know that their buys are likely offset by sells, so turnover will result in a two-sided trade. Doing the math:

Mutual fund trading = $21.1 trillion in assets x 28% turnover x 2 sides
= $47.2 billion per day, or less than 10% of liquidity.

This total is somewhat understated as ICI doesn’t include cash inflows and outflows. If gross cash flows are (say) half the size again of trades, that would lead to around $70 billion in liquidity each day, or just 15% of all liquidity.

Index funds trade much (much) less

One of the common myths of liquidity is that because index funds have such large assets and consistent inflows, they also do a lot of trading, which also distorts prices.

The truth is very different. In fact, BlackRock estimated in 2017 that for every $1 in index trading, there was more than $20 in active trading.

That’s because index funds are designed to do very little trading. Once an index portfolio buys an index weight holding of a stock, in principle, they don’t have to trade it ever again. As the price of the stock rises, so does its weight in the index and in the portfolio, and the portfolio tracks the performance of the index perfectly. That should make index funds some of the longest-term holders of stocks in the market.

In the real world, some trading is required. As companies raise more capital or buy back stock, index funds also need to adjust holdings. IPOs, mergers and price changes can also cause stocks to be added and deleted from indexes, also requiring the index portfolio to trade. But index providers tend to do that infrequently. In fact, S&P 500 turnover is estimated at around 4.4%, a statistic consistent with Credit Suisse’s estimates for the annual Russell 1000 reconstitution of just 3% over recent years.

Although small-cap indexes see higher turnover as their largest holdings are often promoted to large-cap indexes during the year. For example, Credit Suisse also estimates turnover in the small-cap Russell 2000 is much higher at around 20%.

In addition, because index funds are focused on replicating their index, not trying to beat it, they tend to trade exactly when the index trades. That means almost all their trading happens at the close.

Interestingly, despite their sizable assets, Credit Suisse’s index team estimated that all U.S. index funds traded just over $300 billion in 2019, including adds and deletes (both buys and sells). The BlackRock study of 2017 estimated that index funds traded $460 billion annually. Even this higher turnover estimate means index funds make up less than 0.5% of all liquidity in the stock market.

Because indexes themselves tend to minimize their rebalances, we also see that there are nine dates each year when index funds do most of their trading. Data suggests that index funds are around 40% of all trading on index rebalance dates, but cash flows that occur on other days add to less than 5% of close volume, hardly enough to impact price discovery.

Chart 5: Trading in the close higher on index rebalance dates

That leaves active funds providing almost all of the $70 trillion in mutual fund liquidity, and basically to 100% of intraday trading and price discovery done by mutual funds.

There are two important points to note about this. Even though index funds are large:

• Their turnover is low, making them even longer-term holders than active investors.
• Their contribution to price discovery (or price distortions, as some claim) is muted.

Chart 6: Index funds have grown to around 40% of all mutual fund assets

Hedge funds have fewer assets but higher turnover

Based on estimates, hedge funds hold around 3% of total corporate equities. This translates into roughly $1 trillion in assets under management (AUM). However, most hedge funds are both levered and operating higher turnover (lower hold time) strategies, which increases the liquidity they supply to the market.

Hedge funds are notoriously secretive. They also have a wide range of strategies from long term activists to high turnover stat-arb funds. So finding good average trading data is hard.

• This paper suggests that equity hedge fund leverage could be around two times. That puts gross assets are closer to $2 trillion, although short interest data suggests that a lot of hedge funds actually short with ETFs (reducing stock trading). Other data suggests some hedge funds lever up during the day.
• Other sources suggest hedge fund turnover is between 120% and 200% per annum, which implies an average hold time for each stock of less than one year.
• Some funds likely trade much more intraday and minimize overnight exposures (increasing stock trading), but we classify them as intermediation liquidity in Table 1, rather than as a strategy used by natural investors.

Using these conservative estimates:

Hedge fund trading = $1 trillion in assets x (2 leverage x 2 sides) x 200% turnover = $9 trillion in liquidity each year
= $36 billion/day or 8% of all liquidity.

What about ETFs?

With ETF liquidity of $176 billion per day (both sides) in 2019, it is easy to say they cause a lot of stock trading, but that’s not true. There are three types of ETF liquidity:

1. ETFs as mutual funds: Just like other mutual funds, their portfolio will turnover as indexes change. However, most ETF assets are in index-style portfolios where turnover by the portfolio manager is limited. Consequently, we have included these in the mutual fund turnover metrics in the mutual funds section above.
2. ETF trading: ETFs are their own ticker. An investor trading an ETF usually has nothing to do with underlying company stocks. In many QQQ trades, both the buyer and seller usually represent traders choosing ETFs because they are a cheaper way to gain portfolio exposure.
3. ETF to stock arbitrage: Only authorized participants have the ability to arbitrage between the ETF and the stocks. We classify them as intermediaries (not naturals) and discuss their impact on stock liquidity later.

What proportion of professional investors trade in dark pools?

All of the categories above could be considered “professional investors.” They are likely to use brokers’ algorithms to work orders in the market. As a whole, these professional investors (mutual, international, pension and ETF funds) represent almost 60% of all investor assets, but just 23% of all liquidity.

We also know that brokers route their orders through dark pools—and we know that dark pools represent around 12% of all liquidity.

Even with market makers in dark pools, the math shows that professional investors execute a material proportion of their trades in dark pools, possibly 30% or more.

What about retail?

Retail trading grew significantly in 2020 to around 20% of the market. However, even in 2019, with 15% of all trades being retail, they represented a meaningful amount of “investor” liquidity.

However, the way people refer to retail market share overstates their contribution to liquidity for two reasons:

1. Each retail trade represents a buyer or seller, so they are on one side of 15% of shares
2. Retail tends to trade lower-priced stocks (see Chart 7). Using non-ATS TRF data by ticker, we estimate their contribution to value traded is actually 15% lower.

That adds to closer to $30 billion each day in 2019, or just over 6% of all liquidity.

Chart 7: Retail tends to trade lower-priced stocks more, reducing their contribution to notional liquidity

Where do Naturals trade?

So far, we have accounted for around 93% of assets but less than 30% of all liquidity.

That’s less than all the liquidity that trades off-exchange. With dark pools representing 12% of all trades and other off exchange adding to 29% of all trades.

That seems to indicate that a material proportion of naturals execute off-exchange – even though they rely on the exchanges for efficient prices and tight spreads. Ironically that makes the role of intermediaries, and competitive quotes on exchanges, even more important to the whole ecosystem. (It’s also a reason to reward the public market for sharing price data – but that’s a topic for another day)

Chart 8: These calculations indicate natural investors rely on exchanges for efficient prices but do a large proportion of trading off exchanges

What about the rest?

The breakdown for the other 70% of liquidity is much harder to support with public data, but our estimates show this is where specialist liquidity providers step in to keep markets and spreads efficient.

It includes a lot of different strategies, from two-sided liquidity provision to cross market arbitrage to statistical arbitrage strategies, all targeting short-term mispricing caused by other traders.

Their size in the U.S. market is a testament to low transaction costs, which allow for risk to be transferred via one or more additional trades. It is also a result of fragmentation and segmentation that increases facilitated trades.

Here our estimates are based on less hard data but inferred from other levels of activity. To see how intermediaries could add to 70% of liquidity, consider:

• Wholesaler facilitation: Each retail trade facilitated off-exchange has a wholesaler on the other side. So the intermediation doubles the liquidity recorded (totaling 13% of all liquidity). However, TRF data shows the major wholesalers actually print trades representing closer to 25% of all liquidity, or about $117 billion per day. That additional 12% of all liquidity represents $57 billion, and likely trades with institutional investors or even other brokers.
• Equity futures traded around $620 billion each day in 2019, even more liquidity than the whole underlying U.S. stock market (Chart 9). S&P 500 Futures spreads are tight, at 0.8 basis points (bps), and there are arbitrage opportunities with SPY ETFs at 0.4bps and underlying portfolio at 4bps. Given stocks are the most expensive, it’s likely that most futures trades do NOT cause stocks to trade. But if just 5% of futures trading led to stock arbitrage, that would account for around $31bn/day or 7% of all stock liquidity, almost all in S&P500 stocks.
• ETF arbitrage: As discussed above, some “authorized participants” are able to convert rich stocks into cheaper ETFs and vice versa, a critical market function that keeps ETFs fairly priced for investors. However, ETF spreads are often so cheap that there are no arbitrage opportunities. Industry estimates peg the amount of trading that is ETF arbitrage at between 4% and 10% of ETF trading. If that’s the case, then ETF arbitrage contributes around $11 billion per day in liquidity, of just 2% of all liquidity. That’s consistent with the average creations plus redemptions per day (Chart 9), although that also includes many non-U.S. equity ETFs.

Chart 9: ETF creations are a fraction of ETF liquidity, which is a fraction of stock liquidity, which itself is a fraction of equity futures liquidity

• Options markets require initial hedges and constant delta hedging. Notional trading of stock options totals $221 billion per day. If just 30% of notional leads to stock hedges and arbitrage, that translates to $66 billion per day, or 14% of liquidity.
• Market makers’ purpose is to provide two-sided liquidity, not hold stocks long or short. They make money by accurately pricing the spread, and investors benefit from instantaneous liquidity on market orders. Given how few natural investors actually trade on-exchange, it’s more likely than many expect that the quotes on the screen are in fact market makers intermediating trades. If market makers provide liquidity on one side of every second trade on-exchange, that would add to around 16% in total liquidity.
• Opportunistic traders: Some traders run arbitrage strategies that correct mispricing, often by taking liquidity. We note that this paper suggests hedge funds made up as much as 30% of all trades, almost double our estimate above. That’s likely because traders with short-term mispricing strategies and small overnight balance, like statistical arbitrage, are excluded from our estimates above. These traders provide important risk transfers, even though there are little overnight positions or leverage. Given liquidity providers make up 16%, it’s possible these taker strategies add to 8% of all liquidity, or $37 billion per day.

That still leaves us with a pretty large “other” category, at least from the perspective of natural investors. However, there are many other trading and hedging strategies we don’t know much about. For example, hedge funds also buy swaps and OTC options contracts from banks who still need to hedge their risk. Without data on the size of those markets, it’s difficult to estimate their impact on stock trading, but their gamma could even explain some of the growth in market-on-close trading.

It’s also possible some of our intermediation estimates are off by at least this amount (given the lack of real data to reconcile to).

Chart 10: Naturals hold most assets, but intermediation makes up most of the trading

What does this teach us?

We would love to have more data to refine our results. But even as they stand, it helps us understand the role of each participant, bringing liquidity and price efficiency to the ecosystem. The data also busts some common myths:

• Index funds don’t trade much at all, which means they can’t permanently distort prices either.
• Price discovery comes from active funds and hedge funds. Informed trading overwhelms index fund inflows.
• Lots of other participants specialize in risk transfer, from futures arbitrage to statistical arbitrage to classic market making. By process of elimination, these participants may trade even more than all the natural investors combined. But having all this competition for lit trading is also critical to the rest of the market who rely on tight spreads and instant liquidity.

The key takeaway though is that very low costs to trade in the U.S. allow so many specialists to play a role in keeping markets efficient. Rather than degrading market quality, all the competition for marginally profitable trades translates into more liquidity and tighter spreads for investors, which in turn leads to more attractive valuations for issuers.