As we promised in our recent Updated Intern’s Guide to the Market Structure Galaxy, today we follow up with an Intern’s Guide to Trading.
Where do stocks trade?
In our guide to market structure, we talked about all the different “trading centers” in the market – from exchanges to broker dark pools to brokers filling orders directly – and how the total trading in the market is allocated between those venues.
But to understand how trading works, we need to look at how customer orders (retail and mutual funds) make their way to market.
All customer orders need to be handled by brokers.
Retail orders are often sent by a retail broker to a wholesaler for execution and fills often come back at a sub-decimal price (known as “Price Improvement”). In order to protect retail investors from bad prices, a rule known as NMS rule 605 requires retail brokers and wholesalers to compute how many orders get price improved and by how much.
Data seems to show that retail investors mostly trade very infrequently and usually for a very small amount with 92% of their orders for less than $20,000. That’s important when we talk about depth later.
Mutual funds, in contrast, usually have much larger trades to do, often for millions of shares. Adding all that demand to the market at once would increase trading costs, so mutual fund orders are “worked” by institutional brokers. That usually means a computer (algo) slices the order up into smaller pieces to execute over time, often resting in a dark pool (or pools) before being added to exchange flow. Sometimes two investors with opposite trades can agree to do a large trade, known as a “block trade,” with each other simultaneously.
Dark pools, block trades and principal fills (like wholesaling, where a broker takes on the other side of a trade) all result in off exchange trades that occur and without “advertising” an order publicly (by posting an order on an exchange, that order size and price is publicly shared with the whole market).
The latest data shows that over 43% of all shares are traded off exchange. That means just over 50% of shares end up trading on the 16 exchanges in the U.S. now.
Chart 1: Where stocks trade
One side note, for those interested in the economics of trading, is that each of the trading centers has different ways to incentivize customers to their venues.
- PFOF and PI: Retail brokers often get direct payments for their order flow (PFOF) as well as beating the prices available elsewhere (price improvement).
- Tiers: Dark pools typically sort customers into “tiers” based on their profitability to liquidity providers. Traders are attracted to tiers that allow them to capture more spreads (profits) with less adverse selection (which we talk about below).
- Rebates: Many exchanges offer “rebates” to anyone doing trades on an exchange.
- Speed bumps: Being able to avoid trading with incoming orders which in hindsight moved the market also help a market maker capture more spread.
Where do prices come from?
Orders that make it to exchanges are important to the whole system for another reason though.
Exchanges are the only market centers with “equal access” and “public quotes.” That means everyone can trade with anyone, and the prices we all see on our TVs and terminals come from exchanges too.
Special processors, commonly called the “SIPs,” collect and compile quotes from all 16 exchanges. As we show in Chart 2, the national best bid (NBB) and national best offer (NBO) might actually come from different exchanges.
The SIPs create a consolidated national best bid and offer (NBBO), which is then shared with all participants. The NBBO is also an important benchmark price for off-exchange trades, as other NMS Rules require all trades to be executed no worse than the NBBO.
Chart 2: Where prices come from
Source: Nasdaq Economic Research
Where do the best prices come from?
Some exchanges operate what is called a “maker-taker” structure. These markets offer rebates to anyone providing liquidity at the best quote. That includes retail, mutual funds and market maker orders that rest displayed limit orders.
In studies we’ve done in the past, the data shows that it’s hard to attract competitive bids and offers to both sides of all 8000-plus stocks in the U.S. market. However, data clearly shows that offering rebates of a fraction of a cent (around 0.3 cents typically) makes a big difference to quoting activity. In fact, the tight spreads of many stocks are most likely to come from maker-taker markets.
Chart 3: Rebate markets have, by far, the most competitive quotes
Source: Nasdaq Economic Research
Are you ready to trade yet?
Sometimes things on a trading desk happen really fast. Quotes change, prices disappear and opportunities are missed. Misunderstandings can also cost a lot of real money. This has led traders to develop jargon to pass on a lot of information in a consistent and unambiguous way.
For instance, you bid to buy and offer to sell. And each separate trade is called a fill.
The difference between the bid and offer is called the spread (Chart 2 and Chart 4). It creates one of the first trade-offs that a trader needs to make in order to fill their order.
- Market order: A trader can buy all the stock on the offer (or far-touch) immediately with a market order. Market orders don’t care what price they pay; they are happy to accept the prices sellers are offering.
- Limit order: Setting a lower (limit) price allows you to try to buy the stock cheaper, say by joining the bid.
Marketable limits are limit prices that, like market orders, should also cross the spread, and therefore, should execute (they just reduce the risk of paying a much higher price if something happens to the offer after you hit “send” on your order).
Hide or seek?
There are other price levels you might hear used as well.
Mid-point orders are also often used to rest in between the bid and offer. Although this seems like a compromise between paying the whole spread or none of it, mid-point orders are also not “displayed.” That means other investors don’t know they are there.
There are benefits and costs of using hidden orders. A benefit is that you don’t visually add to the supply or demand for a stock – so you have less chance of moving the price before you even get to trade. However, because other traders don’t know your order is there, it’s also more likely that a hidden order won’t trade.
What happens next?
If you use a market order, your order costs more (you paid most of the spread) but should also be filled within microseconds (green line in chart 4). For a mutual fund, that frees you up to trade again and complete your larger “parent” trade faster.
In contrast, when a buyer joins the bid, they almost always start at the back of the queue. That introduces uncertainty and delays. At a minimum you need to wait for enough sellers to cross the spread (“hit the bid”) to get you to the top of the queue so the next trade you get a fill too (blue line in chart 4).
That doesn’t usually take long, but there is a risk while you wait that the market instead ticks higher (yellow and red lines in Chart 4). When that happens, it’s known as “opportunity cost”. It can mean you’ll end up paying more to complete your trade, sometimes even more than using a market order in the first place.
Chart 4: Traders’ choices and consequences
Source: Nasdaq Economic Research
Of course, the opposite can happen too, and prices can fall (black line in chart 4).
Although that means you get your fill and save the spread, we know in hindsight that we could have got even better prices if we instead waited for prices to fall. That’s called “adverse selection,” and for a market maker, it’s bad.
That’s because market makers provide two-sided markets all the time, so a fill as a price-level changes most likely represents an unrealized loss to trade-off. Here is why:
- A buy order will fill as the price falls, leaving the market maker long and having to sell to a lower new bid to close their position.
- A sell order will fill as the price rises, leaving the market maker short and having to buy back at a higher new offer to close their position.
However, it’s not so clear that adverse selection is “bad” for an investor. For a start, it is a fill, allowing more of the order to trade faster, with the next fill potentially at that better price. It also reduces remaining exposure to the market and, therefore, opportunity costs on the total order.
Interestingly, we found that the “benefits” of waiting increase roughly in line with the risks of not getting a fill. That seems to indicate the market is especially efficient at pricing liquidity.
That said, that study also showed that markets are pretty random, and the probability of getting a fill at the near touch (bid for a buyer) was only just over 50%. That means the chance of seeing opportunity cost on a missed fill is about the same as the chance of getting a fill that captures the spread.
Chart 5: Realizing spread capture
Source: Nasdaq Economic Research
To learn more on routing decisions and how algorithms work orders, see Routing 101: Identifying the Cost of Routing Decisions and Routing 201: Some of the Choices an Algo Makes in the Life of an Order.
How fast should you trade?
One important implication from Chart 5 is that the faster you trade, the more it will cost you. It’s also true that the larger the trade you want to do, the more it will cost.
That’s consistent with a lot of academic and industry research. It also makes economic sense. If you increase demand more, prices should rise.
The problem for most traders is going slower has costs too.
If other investors have the same information and trade idea, trading slower allows others to buy the stock first, and their impact on the stock’s price results in higher opportunity costs and worse trade fills for you.
This is the conundrum most traders face:
- Trade faster, and it costs more (market impact).
- Trade slower, and it costs more (opportunity cost or “alpha decay”).
However, there is a mathematical way to optimize this problem which we discussed in How Fast Should You Trade? This shows that you can minimize your trading costs by balancing your trading impacts against how fast other traders are impacting the same stock. The problem is you need to know a lot about each order on your blotter, as each trade will have different alpha, liquidity and market impact.
Chart 6: Optimal speed to trade-off impact and opportunity cost can be mathematically determined
Wait! Not all stocks trade the same way
Although U.S. trading rules are pretty consistent – with almost all stocks trading in all markets, using 100-share round lots and one-cent ticks – not all stocks trade the same way.
In fact, the differences between stocks can be significant.
For example, the size of companies is very different. In fact, the smallest 50% of companies add to a total market cap of less than $800 billion. That’s smaller than our single largest company (and a few other Nasdaq-listed companies below that)!
Market capitalization also tends to affect liquidity. Larger stocks trade more, roughly in proportion to their size. Stocks that trade more also tend to have more depth, meaning a market order can trade for more value, helping to finish large orders faster and with less impact (circle size in chart 7).
In addition, stocks come in all sorts of prices—from sub-$1 to over $1000. That means the economic value of a one-cent tick also changes materially. You can see from the diagonal nature of the dots in Chart 7 that lower priced stocks have wider spreads. However, the sharp diagonal line shows where the one-cent tick stops spreads from getting any smaller. Where that happens, you can see the dots are bigger, representing longer queues.
The colors also show that small cap stocks typically have lower prices and more expensive spreads (yellow dots are top and left) while very large stocks are the opposite (blue dots).
Chart 7: Liquidity (in $) is a key driver of spreads, but so is the one-cent tick
Source: Nasdaq Economic Research
As algorithms adapt to these different trading costs, we see other problems created. For example, there are more odd lots as prices rise, often at better prices. However, odd lots are not protected, nor do they contribute to the NBBO, which is a problem that regulators and the industry are currently trying to solve.
If you’re interest, there are other important trends we’ve seen in our research too:
- Asset managers seem to focus on larger-capand more liquid stocks, which makes them trade more competitively, which gives them tighter spreads.
- That leaves retail investorsdominating microcap stock trading, especially in 2021. That also means they’re adding to off-exchange market share while trading mostly stocks with much wider spreads.
- We’ve also found that spreads get consistently better as liquidity(measured as value traded) increases. So, more liquidity reduces trading costs and spreads.
- But liquidity itself is mostly driven by the size (market cap) of a company (with a little boost from periods of volatility, which also tends to widen spreads and increase trading costs).
- Finally, we estimate that “real investors” actually make up just a fraction of overall trading in the market. The prevalence of futures and options arbitrage as well as market-making strategies makes the market very fast but also very efficient. That’s also likely why U.S. markets are considered the most liquid and cheapest to trade in the world.
Don’t stress — computers do most of the trading for investors
Although this all sounds complicated, the reality is that computers (trading algorithms and market maker models) do most of the trading these days and they can be optimized with data and programmed to fix much of the complexity that human traders face.
From deciding what price level to use or when to advertise an order – to which venues to use – most of the market is also interconnected and automated. The SIP and NMS rules require it.
The biggest input required from investors is to tell the algorithm how fast they need to trade.