Spoofing in Equilibrium
Basil Williams & Andrzej Skrzypacz
Discussion by Katya Malinova
DeGroote School of Business
McMaster University
WFA 2022
June 2022, Portland OR
Spoofing: why care now? new? big(ger) problem?
- Not new
- But: was it easier to detect "bluffs" & bad actors on the floor/in the pit?
- SEC filed charges against spoofing in Nov 2001
- violating Securities Act of 1933 and Securities Exchanges Act of 1934
- 6 individuals, collectively pay $43,860.63
- Dodd-Frank: defined spoofing & explicitly illegal in futures in 2010
- enforcement ramped up (?)
BUT: not that many cases made the news!
Little spoofing? Threat of prosecution?
Difficult to catch and/or to prosecute?
As an aside: "Cryptocurrency Pump and Dump Schemes" Li, Shin, Wang (2021)
"bidding or offering with the intent to cancel the bid or offer before execution"
Spoofing: remains difficult to identify
- How to distinguish intent from legitimate cancels?
- Change of mind, new info, ...
- 2014: Michael Coscia
- pre-programmed algorithms
-
2021: Edward Bases and John Pacilio
- online chat
- “I know how to game this stuff.”
- “One can definitely manipulate if you’re aggressive.”
- “If you spoof this it really moves.”
Ideally: need a model/method to address this problem
High legal bar!
This paper
- Cool tractable model to conceptualize spoofing
- Spoofing is an equilibrium outcome
- most prevalent in "moderately" liquid markets
- Predictions on the impact
- slower price discovery
- lower liquidity
- higher volatility
Model: modified Glosten-Milgrom
- Traders: informed or noise, use market orders against MM, trade one unit
MM posts T1 quotes
MM posts T2 quotes
fundamental revealed
T1 Short-term trader: market order
T2 Short-term trader: market order
T1
T2
T1 market order may be cancelled: by the T1 trader or exogenously
Model: modified Glosten-Milgrom
MM posts T1 quotes
MM posts T2 quotes
fundamental revealed
T1
T2
Long term trader (buyer)
market buy
market sell
cancel
Model: modified Glosten-Milgrom
T1
T2
long-term buyer \(\Rightarrow\) E[V| history] = Pr (informed) = \(\alpha\)
\(p=\alpha\)
\(0<p<\alpha\)
market buy
market sell
cancel
market buy
long-term buyer \(\Rightarrow\) E[V| history] =\(\alpha\)
short-term seller + short-term buyer \(\Rightarrow\) E[V| history] =0
- Cool approach & gets the job done!
- tractable model of price manipulation
- But:
- interpretational challenges
- high price impact of spoofing
- What about a stylized limit order book model (with "simple" pricing rules)?
- Kaniel and Liu (2006)
- Brolley and Malinova (2021)
Comment 1: market order cancellations?
- Single parameter describes both the mean and the variance of the fundamental
- Spoofing by a buyer ("fake sells"):
- lower expected value
- \(\to\) mechanically higher variance \(\to\) higher bid-ask spread
- How robust are the liquidity and volatility predictions to richer info structures?
Comment 2: two states of the world
- A single order enough to move the needle
- Sell + cancel = always "bad news"
- what if "cancel" is due to new positive info?
- Long-term trader is a temporal "monopolist"
- only exogenous costs to spoofing
- no competition, no prices moving against you
Comment 3: too "easy" to spoof in the model?
Difficult to draw strong policy/practical implications
E.g., authors: HFTs facilitate spoofing
True: easier to "hide" the cancels
But is it easier to spoof with HFTs?
- Very nice, tractable, parsimonious model to conceptualize spoofing
- Spoofing is an equilibrium outcome
- can't "educate it away"
- Main takeaways for policy? Regulators? Prosecutors?
- Key insights on differentiating price manipulation from legitimate behavior?
To sum up
@katyamalinova
malinovk@mcmaster.ca
slides.com/kmalinova
https://sites.google.com/site/katyamalinova/
https://www.justice.gov/opa/pr/eight-individuals-charged-deceptive-trading-practices-executed-us-commodities-markets
