Central Bank Workshop Discussion 2023

Would Order-By-Order Auctions Be Competitive?

Thomas Ernst, Chester Spatt, Jian Sun

Discussion by Katya Malinova

DeGroote School of Business

McMaster University

18th Central Bank Conference on the Microstructure of Financial Markets

Federal Reserve Board, Washington, D.C.

October 19, 2023

Several Questions re: Retail

Q1: Should retail flow be segmented out?

Benefits to retail?
- Battalio, Jennings (2023):
  - Rule 605 vastly understate wholesaler execution quality
- Dyhrberg, Shkilko, Werner (2023):
  - wholesalers save $-billion/month to retail
Benefits/costs to the rest?
- Comerton-Forde, Malinova, Park (2018 JFE):
  - demise of Alpha (= Canadian RLP) $\to$ depth $\nearrow$

A non-random selection of the many papers on this topic

Several Questions re: Retail

Q2: If retail orders are to execute away from the "toxic flow"

$\to$ what is the best practice?

Q2A: Is the current system good enough?
- SEC: $1.5 billion annual shortfall!

https://www.sec.gov/news/sec-videos/office-hours-gary-gensler-order-competition-rule

Where would $1.5 billion come from?

"Office hours with Gary Gensler"

Several Questions re: Retail

Q2: If retail orders are to execute away from the "toxic flow"

$\to$ what is the best practice?

Q2A: Is the current system good enough?
- SEC: $1.5 billion annual shortfall!
Empirical research:
- Battalio, Jennings comment letter & Battalio, Jennigs, Saglam, Wu (2023):
  - SEC is wrong re: retail id based on BJZZ (2021) algo
  - $\to$ costs of failed auctions $\approx$ benefits of successful
- Dyhrberg, Shkilko, Werner (2023):
  - retail investors in less liquid stocks would likely be worse off under proposed!

This paper: current system vs. order-by-order

Q2: If retail orders are to execute away from the "toxic flow"

$\to$ what is the best practice?

Q2B: THIS PAPER
- (When) is the SEC proposed order-by-order auction mechanism ("order competition rule") an improvement on the current system?
- Auctions $\Rightarrow$
  - allocative efficiency
  - "winner's curse"
  - larger rents to market makers
    - ex-post bidding $\to$ heterogeneity among MMs

Auctions welfare-dominant in:

liquid stocks
lots of competition
MM don't care much about "common component" /other MM inventory but allocative (in-)effficieny matters!

Would like to know more re:

when, where, & the extent of the inefficiency?

Send "unwanted" $\to$ exchange(?)
Barardehi, Bernhardt, Da, Warachka (2023)
- internalized imbalance $\Leftarrow$ institutional liquiidity demand

Order-by-order auctions: $\to$ "unbundling"!

Current system: random allocation to wholesalers
- based on ex-ante bids (or past performance)
  - forces participation (?)
- $\to$ encourages cross-subsidization
  - improves liquidity in illiquid stocks

The paper: auctions
- $\to$ larger variation in speads: liquid vs. illiquid
- redistribution across small vs. large stocks
Empirically Foley, Liu, Malinova, Park, Shkilko (2020):
- Canadian market: cross-subsidization helps illiquid without worsening the liquid!

Comment: No "tether" to a given stock/order size.

If costly participation

$\to$ may drop small/illiquid?
more detrimental than redistribution?

Cross-subsidization + ex ante allocation!

Order-by-order auctions: who provides liquidity?

Chairman Gensler: "These auctions will include competition from possibly hundreds or thousands other market participants."

https://www.sec.gov/news/sec-videos/office-hours-gary-gensler-order-competition-rule

Do we expect to see "thousands" other participants?

What is their impact?

The paper: skeptical

Theoretically, institutions may be better informed (directional flow?)
- $\to$ increase the adverse selection in auctions
- $\to$ exacerbate the "winner's curse"?
RLP: little empirical evidence of extra bidders standing by to trade

Comments/questions:

Connect institutions & retail directly, bypassing the wholesalers?
- Barardehi, Bernhardt, Da, Warachka (2023)
- "In response to institutional liq. demand, wholesalers internalize retail trades"
Is the aggregate order (always) directional?
How good is the RLP as a predictor?
- Exhaust/"externalized" flow?
Changes in info environment?
- ex-post vs. ex ante (size?)
- info revelation through the auction?
Will new institutions enter?
- slow liq. providers? non-directional?

Order-by-order auctions: other considerations

SEC: untapped (hidden) liquidity at the midpoint!
- Canadian experience: require a min price improvement for dark
  - min one tick or at the midpoint
  - $\to$ "killed" the retail (RLP-style) dark pool
  - $\to$ "sit on the outside, catch the retail flow & earn rebates"
Failed auctions?
- Likelihood of no bidders $\to$ route to exchange?
Effect on broker commissions?
- $\to$ on retail trading and volumes?
Literature on bundling?
- E.g., Adams&Yellen (1976), Bakos&Brynjolfsson (1999)

If "break it" ... what is the OPTIMAL mechanism design?

The paper:
- status quo vs. proposed
Other auction designs?
- Second price?
Periodic auctions? Bundle retail over time? Akin Budish et al (2015)?
Decentralized finance AMMs?
- passive liq. provision = alternative to "buy&hold"
  - $\to$ liq providers want asset exposure
  - fees compensate for "incremental" adverse selection
  - use idle capital!
- pro-rata risk & fees for liq providers
  - improved risk sharing
- continuous trading, no intermediaries

Malinova and Park (2023):

propose AMM implementation for equity markets

$\to$ estimate $-billions saving in transaction costs

$\to$ in theory, should work best for the less toxic retail flow

Summing up

Very nice paper
- elegant parsimonious model!

(Theoretical) assessment of the SEC proposal
- status quo vs. proposed auction-by-auction

Would like (the SEC) to think "bigger picture" re: optimal mechanism design
- Auctions vs. bundling
- Auction design
- Alternate out-of-the-box designs?