1. HFT - Brussels Exchange Forum

High Frequency Trading:
To be Nurtured or Banned?
Hans Degryse
KU Leuven and CEPR
Brussels Exchange Forum, April 25, 2014
Hot topic
• Michael Lewis book “Flash Boys”
• Investigation by FBI
Let me aim to summarize academic evidence…
1.High Frequency Trading (HFT):
– Definition
– Identification
– HFT in the US and Europe: some stylized facts
2.Impact of HFT: what does theory suggest?
3.HFT and market quality
4.HFT and market stability
5.HFT, social welfare and regulatory responses
1. High-Frequency Trading: Definition
Difficult animal to define. SEC (2014) defines them as:
1. Use of extraordinarily high speed and sophisticated programs for
generating, routing, and executing orders.
2. Use of co-location services and individual data feeds offered by
exchanges and others to minimize network and other latencies.
3. Very short time-frames for establishing and liquidating positions.
4. Submission of numerous orders that are cancelled shortly after
5. Ending the trading day in as close to a flat position as possible (that
is, not carrying significant, unhedged positions overnight).
1. HFT: Identification
Different methods
1. Direct classification based upon trader IDs, i.e. HFT Flag
– E.g. NASDAQ dataset (used by e.g. Brogaard, Hendershott and Riordan (2013),
Hirshey (2013), Zhang (2013)); ESMA dataset (Degryse, De Winne, Gresse and
Payne (in progress))
– Pure HFT firms
– Typically all HFT flags have co-location
2. Quantitative method employing “order-to trade ratios”, “intraday
inventory management”, or “order modification and cancellation
Apply to all IDs, apply to specific trades
E.g. E-mini datasets (e.g. Kirilenko et al. (2011)); Canadian dataset (e.g.
Malinova, Park and Riordan (2013), Euronext (e.g. Verschelden (2014)).
1. HFT: stylized facts (US)
1. HFT: stylized facts (Europe)
1. HFT: stylized facts (Europe (2))
HFT more important on Multilateral Trading Facilities than Regulated Markets
1. HFT: stylized facts (Europe (3))
Order-to-trade ratios of HFT much larger than other participants
1. HFT: stylized facts (Europe (4))
HFT more active in stocks that are more fragmented across trading venues ->
HFT “klit” together different trading venues (see Menkveld (2014))
1. HFT: stylized facts – general
• HFT is not a monolithic phenomenon but encompasses a
diverse range of trading strategies
• Not all HFT trading is passive
• NASDAQ dataset (Brogaard, Hendershott and Riordan (2013)):
more than half of trading is attributable to liquidity taking (market)
orders, even more so in small cap stocks
• UK dataset (Benos and Sagade (2012)): less than half of trading is
liquidity taking
• HFT much less active in small stocks in the US; seems less the case in
Euronext (Verschelden (2014))
• Quite some variation across countries within Europe (ESMA (2014))
2. HFT – Theory
• Theory essentially models two forces : speed and information
• Implications for HFT, other traders, social welfare
1) Speed
• Hoffmann (JFE forth)
• fast traders reduce exposure to picking off risk -> beneficial for HFT and
social welfare
• Presence of fast traders changes strategies of slow traders that submit limit
orders with a lower execution probability such that trading rate declines
• Speed endogenized: speed is market power and allows to extract rents
from slower traders => arm’s race leading to overinvestment from a social
welfare perspective
• Calls for randomized “speed bumps” as now inplemented in some FX
• Menkveld and Jovanovic (2012):
• HFT may lead to more competition reducing spreads
2. HFT – Theory (2)
2) Information – advantage of machines over humans is ability to process vast amounts of
information at superhuman speed
Jovanovic and Menkveld (2012):
• HFT process hard information faster which lowers adverse selection for them
• But they throw an adverse selection problem on others
• Calibration exercise shows that social welfare improves
Biais, Foucault and Moinas (2013):
• HFT have a higher likelihood of finding trading opportunities which induces a higher trading rate which is
good for welfare
• but HFT expose in this way adverse selection on others reducing trading rates and thus welfare.
3) Combination of these two:
Bernales and Daoud (2013)
• HFT benefit in two ways: (1) picking off “stale” orders (2) react faster on information => slow traders
modify their strategies and trade more through market orders
• Benefit or cost to slow traders depends on their relative presence: if many slow traders, picking off risk
dominates and they are worse off; if few slow traders, HFT are beneficial.
• HFT with informational advantage is good for welfare; HFT with only speed advantage is bad for welfare;
having both is better for welfare => suggests 70% of HFT is optimal
Bongaerts and Van Achter (2013):
• endogenize number of HFT and slow traders => HFT drive some slow traders out of market.
• With substantial asymmetric information, HFT shun the market inducing slow traders also to leave =>
endogenous market freezes and small crashes.
3. HFT and Market Quality
• Fragmentation in “lit markets” (proxy for HFT) improved market quality but
dark trading decreased it (see e.g. Degryse, de Jong and van Kervel (2013))
• HFT effect on market quality:
• Passive HFT strategies have beneficial effects: lower spreads and intraday volat
• Jovanovic and Menkveld (2012) find that the entry of a large, primarily passive HFT
reduces spreads by 15% in Dutch markets
• Malinova, Park and Riordan (2013) exploit chock in exchange fees that affect HFT
traders and find that bid-ask spreads increase in Canadian markets
• Liquidity consuming activities of HFT less beneficial:
• More price impact (Zhang and Riordan (2011) for large stocks, Zhang (2013),
Brogaard, Riordan and Hendershott (2013))
• Competition between HFT harmful?
• Breckenfelder (2013) finds that when HFT compete there are more liquidity consuming
trades by HFT
• HFT may lead to “ghost liquidity” for slow traders
• van Kervel (2013) shows that trades executed in one venue lead to substantial
cancellations on other venues
4. HFT and Market Stability
Do HFT increase financial instability and systemic risk in financial markets?
Example: the flash crash, May 6, 2010
Role of HFT in Flash Crash (see e.g. Kirilenko (2011)): not triggering the
shock but maybe deepening the volatility
Hagstromer and Norden (2013): more passive HFT activity reduces intraday
Systemic risk concern:
• HFT have little capital
• HFT trade a lot with each other -> contagion?
5. HFT, Social Welfare and Regulatory Responses
• Evidence suggests that market quality has improved and markets may
have become more informational efficient
• Should be beneficial to firms and real economy
• However,
• Slow traders may need to change trading strategy and may ultimately be
worse off => switch from limit orders to market orders
• Is gain in informational efficiency socially productive? Information
would have been incorporated at slightly lower speed anyhow
• HFT may improve slow traders outcomes when they act as market
makers, but reduce slow trader welfare when picking-off risk
• Should HFT be allowed to buy preferential treatment? Probably not!
• Earlier access to information releases is disturbing fair level playing field
• Co-location for every one -> should look into business model of RM and MTF
to see how this is allocated
5. HFT, Social Welfare and Regulatory Responses (2)
• Regulatory responses:
• Fair level playing field in terms of access to information
• If HFT by quote stuffing or excessive order flow delay other
participants -> introduce order withdrawal fees; but difficult to separate
“good” from “bad” order flow
• MiFID II: HFT strategies subject to regulatory authorization
• Change market structure: batch auctions every “xxx” milliseconds
• Some concluding thoughts
• HFTs have allowed new trading platforms to arrive and have induced
competition in trading and post-trading fees. Both are important as
“cum-fee liquidity” has improved substantially (see e.g. Colliard and Foucault
(RFS2012) and Degryse, Van Achter and Wuyts (2013))
• In sum, HFTs have been beneficial but regulators need to take care of
potential externalities
• Too much dark trading might be a more important concern (MiFID II)

similar documents