What Is Segmentation?

The new SEC chief has been talking a lot about segmentation recently, and it also took up a few pages of the recent MEME stock trading report.  

But what is segmentation? And what does it do to markets and trading? 

The theory of segmentation 

Academics have studied segmentation for years, but we start our summary of how it affects equity markets with a 1996 paper by Maureen O’Hara (see Table 1). 

In contrast to fragmentation, which just adds to the number of venues, segmentation refers to the ability of different venues to select who they trade with – thereby “segmenting” that flow away from the rest of the market. 

Academics being academics, classify traders as “informed” and “uninformed.” Rather than a sign of investing acumen, “informed traders” represent those who are most likely to cause adverse selection, where prices change permanently, causing losses for liquidity providers.   

Trading with “uninformed” traders should, therefore, make capturing spreads easier, which makes market making more profitable. In a competitive market, uninformed traders should also receive lower effective spreads, commensurate with their lower “toxicity.” 

The academic theory says that because most segmentation occurs off-exchange, on-exchange trades become, on average, more often “informed.” The academic theory then follows to say that on-exchange bid-ask spreads will adjust to reflect the costs of market making to more informed flow, and depth will be reduced as it is fragmented into other markets – as we illustrate in Chart 1 below.  

In short, segmentation should create winners and losers. In an efficient market, a narrowing of spread for some investors should be offset by an equal but opposite widening of spreads, and a reduction in depth, for all others. 

Chart 1: The theory segmentation suggests that lit quote widens and depth falls in response to off–exchange trades with lower effective spreads increasing 

But what do they find in practice? 

What does academic research say? 

We’ve summarized all the academic papers we could find on this topic below.   

Most find that off-exchange trades do, in fact, have a lower adverse selection, indicating that segmentation works (for some liquidity providers).  Some also find lower effective spreads, indicating that segmentation also helps investors who qualify for the segment. 

A few well-known studies find a U-shaped result, where some off-exchange trading is actually better than none (up to a point). One well-known paper suggests that dark trading helps when share is below 10%; another introduces the concept of a “tipping point” as dark trading increases that is at 40%, or quite a bit higher. 

Of course, at the extreme, in bilateral markets with no public quotes, even doing a study to show investors are harmed becomes hard. 

Table 1: Summary of academic papers on off-exchange trading and market segmentation 

^{1 Meaning that off–exchange trading reduces spreads or improves depth. 
2 See links to papers below}  

Source: Nasdaq Economic Research 

Interestingly, most academics also find high off-exchange market share makes prices “more efficient.” That mostly reflects how academics think about efficiency.  High price efficiency means lit prices reflect news instantly, which should happen if the chance of adverse selection increases. 

What does industry data say?  

Markets frequently advertise the fact that adverse selection is lower in their (segmented) venues.  Industry data, like 605 reports, also confirms that segmentation also usually results in lower effective spreads for segmented investors. 

Separately, there is also some evidence that market makers price spreads to include all-in-costs of trading very effectively. That suggests that higher adverse selection on exchange would lead to widening of spreads, or lower depth, suggested by the example in Chart 1. In fact, Best-Ex Research recently used adverse selection costs data to estimate that NBBO spreads could be 25% lower if off-exchange flow were executed on a centralized market. Another study suggested depth is falling in large-cap stocks, which may support the reduction of depth theorized in Chart 1 too. 

Most industry studies focus on so-called “inaccessible liquidity.” For those left in the public markets, the ADV you see isn’t the ADV you can trade with. Some studies suggest that increases trading costs for mutual funds, especially in stocks that trade heavily off-exchange. 

What does this all mean? 

Research from others seems to show that segmentation provides better spreads and profits to those in the segment. It also generally finds that the opposite happens to those outside the segment, where spreads widen, depth declines and price changes become more permanent.   

In short, the academic theory proposes that segmentation creates winners and losers.  

The big questions it raises are: Who are the losers today, and what kind of market structure would leave investors, as a whole, better off? 

Eugenio Piazza, Research Specialist for Economic and Statistical Research at Nasdaq, contributed to this article. 

APPENDIX: Links to papers in Table 1 

1. Easley, Kiefer & O’Hara (1996) 
2. Battalio (1997) 
3. Hendershott & Mendelson (2000) 
4. O’Hara & Ye (2011) 
5. Zhu (2014) 
6. Weaver (2014) 
7. Preece & Rosov (2014) 
8. Comerton-Forde & Putniņš (2015) 
9. Degryse, de Jong & Van Kervel (2015) 
10. Foley & Putniņš (2016) 
11. Buti, Rindi & Werner (2017) 
12. Hatheway, Kwan, Zhang (2017) 
13. Menkveld, Yueshen, & Zhu (2017) 
14. Farley, Kelley, & Puckett (2018) 
15. Comerton-Forde, Malinova, & Park (2018) 
16. Hasbrouck (2019) 
17. Johann, Putniņš, Sagade, &  Westheide (2019) 
18. Battalio, Hatch, Saglam (2019) 
19. Aramian & Norden (2020) 
20. Eaton, Green, Roseman & Wu (2021)