Real Impact of No Round Lots

By Phil Mackintosh, Chief Economist, Nasdaq

There has been a lot of discussion about the odd lot problem in U.S. markets, especially as odd lots have been increasing. The latest range of solutions focuses on making the problem smaller.

But what if we just made the odd lot problem go away?

What if we eliminated round lots altogether?

From NBBO to nBBO

We wouldn’t be the first country to do this. Most developed markets already trade in integer shares, as do many emerging markets. 

The biggest fear is that moving from a round lot (NBBO) allows smaller odd lots to set the best price (let’s call it the nBBO), which would also reset benchmark prices used to measure retail price improvement and institutional execution quality

That would result in tighter spreads, which would be good for investors. But there is also a fear that depth would fall, perhaps so much that the nBBO would not be large enough to complete most trades.

And if you can’t execute a whole trade at the nBBO, we would see trade-throughs and effective spreads that are wider than the nBBO, making it harder to assess true routing efficiency.

However, our first pass at the data suggests the real impact on trading might be smaller than you might think.

70% of volume would be unaffected if all orders were treated equally

For this study, we break all stocks into the same equal quartiles we used in our recent report on market share. That means our “large-cap” group includes around 1,400 stocks.

Although we’ve seen before that many stocks trade wider than their one-cent tick, those that do also tend to be relatively illiquid stocks. When looking at the stock universe by volume (shares traded), we see that almost 70% of all volume is traded in tick-constrained names.

That’s important because, in tick-constrained stocks, odd lots and round lots all join the BBO at the same price. 

In short, treating all orders as the same and allowing any quantity to set NBBO wouldn’t change the spread or depth in the tick-constrained stocks much at all.

Chart 1: 70% of volume traded occurs in tick-constrained stocks, with almost all widespread volume in large-cap names

Trading distribution across market cap

NBBO depth currently varies across market cap and stock price

The fear is that for stocks with multiple ticks between the NBBO, odd lots might “micro-penny” each other resulting in a very small order setting the benchmark prices.

But how much depth do we need to fill an “average” trade?

We know from other studies we’ve done that the average trade sizes for retail and algos are both just under $10,000. That’s likely why the SEC, in their NMS-II proposal, reset their variable lot sizes to be at least $10,000.

When we look at the average depth across the spectrum of stocks, we see two potential issues with treating all lots equally:

  1. Large-cap stocks over $100 often have wider spreads, where odd lots could easily reduce the size and spread of the nBBO.
  2. Microcap stocks already have low depth even with a 100 share NBBO.

Chart 2: Depth is not constant across market cap or stock price

Average nominal size at the 100-Share NBBO

Most stocks are near a good trading price

The thickness of the lines in Chart 2 represents the number of stocks in each category. From this, we can see that: 

  • Large-cap stocks have a U-shape for value on the NBBO, but that shape highlights two separate trading problems we discuss below. However, both tails are thin, which means the problems with higher- and lower-priced stocks affect a small proportion of all tickers.
  • Small and microcap stocks are mostly lower-priced, although even with an NBBO, their quotes have the least depth. However, many of those stocks trade as much as 80% off-exchange.

Looking closer at the large caps U-shape problem

On the left side of the U-shape are we see very high NBBO depth in low-priced large-cap stocks. That actually reflects the fact that these are already mostly tick constrained (see 59% of their ADV in Chart 3), which causes wide spreads that are expensive for investors to cross.

For those stocks in this group, an odd lot joins the same long queue as round lots already – rarely fragmenting (or improving) the quote. So it would make no difference to these stocks if odd lots were treated the same as round lots.

Chart 3: Large-cap stocks under $100 are also mostly tick-constrained

Trading distribution for large cap stocks

In contrast, the right-side of the U-shape are high-priced large-cap stocks. As we see from darker blue in Chart 3, these see the opposite, with almost 60% of their ADV in stocks with wider spread tickers (over 10 cents). 

However, these stocks also have artificially high NBBO depth because of their higher prices. The round lot times their higher price means significant additional capital is required to form an NBBO in their names. That results in significant use of odd lots inside the NBBO and also makes their NBBO artificially wider.

In fact, these are the majority of stocks contributing to the odd lot problem. Consequently, allowing odd lots into their queue should make these stocks trade more efficiently.

Looking at the Nasdaq order book to confirm these theories

We can look at the “best odd lot depth on Nasdaq” (let’s call it the qBBO) to see what impact equally treating odd lots might have.

Without even consolidating odd lots from other exchanges, the data corroborates much of our hypothesis above. In short, odd lot depth:

  1. Isn’t as fragmented for high priced large-cap stocks as many fear
  2. Isn’t as shallow for microcap stocks as many fear

Even though data also shows that one share orders are the “most common” size, most of the time, the quote is much larger. Based on our analysis, the average notional “best odd lot” on Nasdaq (qBBO) is still over $10,000 (Chart 5).

Focusing on high-priced large-cap stocks, data shows that the qBBO still has plenty of depth to complete an “average trade.” The notional value of the qBBO actually increases as prices rise (chart 4), showing that “micro-pennying” wider spreads matters less than the rising price of each share. Remember, it only takes three shares of a stock like AMZN to fill an “average” sized trade.

Chart 4: Nasdaq odd lot depth for high priced large-cap stocks remains above average trade sizes

Median nominal size

Of course, averages don’t tell the whole story. But the grey line shows that the qBBO is above $10,000 over 70% of the time across all these stocks too. And remember, that’s before we consolidate other markets.

Less happens to depth than you might think

In fact, the current qBBO is deep enough to fill an average-sized trade most of the time, even for microcap stocks. So treating all lots the same makes less difference than you might think to spreads and tradable depth.

Chart 5: Nasdaq depth at our best odd-lot quote across market cap and price

Median Nominal Size

Should all lots be treated the same?

Many fear that treating all sized orders equally will lead to micro-pennying and reduced depth at the BBO.

Looking at the data, the real impact seems relatively small. It might be that round lots are really the bigger problem.