We recently studied how trades moved around the market – usually starts with a broker order in secus, traveling at the exchanges around the market at the speed of fiber to take orders, then causes a reaction function at the microwave speed.
Today, we are seeing how quotes updates usually flow around the market.
What we find is that primary listing exchanges mostly set the new national best bid and proposal (NBBO). Then, most of the venue see orders reaching a fairly consistent rate over time, but in some places there is a rapid increase in quotes in new NBBOs within the first milliseconds.
Whether it is good for market structure and router or bad is an interesting question.
Set listing venue most NBBOS
We see who sets new NBBOs using the site timestamp in all exchanges. This removes any delay that reports new trades back to SIP.
We see that primary exchanges set the new NBBO by a majority of time. This should not be surprising, as primary exchanges also need to compete for company listing, and companies compare them based on things that reduce their capital costs, such as frequent quotes and tight spread.
For example, NASDAQ market manufacturers set quotes in NASDAQ-listed shares close to 58% time. In contrast, all other exchanges jointly improve NBBOs for less than 43% NASDAQ shares.
Chart 1: The new NBBO price of more than half of the time is determined by the Primary Listing Exchange
Other places include NBBOs at different rates and speeds
It is likely to be more frequent that other locations are involved in NBBO prices (or copy) that are set on primary.
Importantly, we do not see all places running to copy NBBO quotes at the speed of light. Instead, most exchange time (X-Xis) looks at the frequent arrival of new orders as progression.
But some interesting exceptions are:
- Rapid jump Show rapid copying of new quotes. Ltse, and later IEX, stand out, both were well quoted within 1 millisecond with a sharp jump in copy.
- Line height Shows that IEX stands out, especially for more liquid shares in NASDAQ-100 (Chart 2A), where it has more quotes that mimic NBBO than any other exchange-despite other exchanges from one significant differences, actually offers more liquidity (ARCA, Edgx and bats) despite other exchanges.
- line thickness The universe reflects the width of shares where quotes are copied. IEX and LTSE are standing out again, far more with copy-elements on stocks, providing investors the presence of wider liquidity, especially their trading market share (Chart 3).
Chart 2A: Iex liquid (NASDAQ-100) send orders to copy NBBO quotes for stock
Interestingly, the patterns change very little when we look at less liquid stock.
Given that there are about 3,300 companies in the chart below (just 100 in the versus chart), the increase in total quotes is relatively low. It is probably true that, with these small companies, they are less profitable for quotes. But these are also ticks that require liquidity support. Even SIP Revenue Allocation Sutras are tilted in favor of quotes in these shares.
What the data shows that for these low liquid stocks, IEX and LTSE experienced a single sharp jump in the number of copied quotes. However, as you see in Chart 3, these large numbers do not eventually go towards trades.
Chart 2B: Low liquid, small-cap, less interest in mimicking quotes for stock
Is this good or bad for the market structure?
It matters differently for different participants.
For brokers and traders, it reduces the special requirement of speed at the top of the queue. This is because you can be at the top of their queue until you can find any order at that place. In turn, those exchanges helps to increase the occupation and profitability spread to traders.
But fission connects other costs for brokers, including more connections and more complex routing. It also adds cost to investors through high opportunity costs.
There are other costs to reduce queue priority.
System does not reward competitive quotes that lead to trades
Even more importantly, the fragmentation of quotation is bad for market makers to really establish quotation in the first place. Their business is to benefit from spread capture, but copy quotes are less likely.
Perhaps worse than this, the way data economics in the US are regulated adds to disabilities. The SIP Revenue Allocation Sutra was designed to “equally” quotes and reward trades, even if those prices were set more often. Before exchanging more than a dozen, with some protected quotes using speed bumps, it was set to reward all quotes equally.
Given the recent SIP revenue, it has been shown that some exchanges really earn a lot of quotes, without doing any business. In some exchanges, there are clear costs for exchanges, such as discounts, to reward market manufacturers providing those quotes, but in other cases the cost of exchanges and benefits to market manufacturers is less transparent.
Chart 3: Send a large number of orders copying LTSE and IEX NBBO quotes
SIP quotes revenue may add to millions of dollars for some exchanges. It is an artificial incentive that supports fragmentation, actually without making the market more competitive and cheaper to investors.
We need to ensure that NBBO is good in protecting investors and issuers
Most of us agree that NBBO is important for investors and issuers. Educational research also suggests that tight proliferation reduces the cost of capital and increases liquidity. In turn, it helps to make the American market more attractive than many markets around the world.
We spend a lot of time to debate about the benefits of a “public” and inexpensive NBBO. Perhaps we also need to ensure that economics also greatly rewards places and traders who establish those quotes. Even can help in making markets more efficient.
