As European policymakers wrestle with plans to curb high frequency trading, Colin Lambert takes a
look at a paper that suggests taxation, not speed limits, is the answer.
F-or all the opprobrium poured upon the
heads of hedge fund managers and then
bankers over the last decade, neither
comes close to the emotions that have been
stirred by high frequency traders – even though
the majority of the wider populace has no idea
how that strategy works.
Concerns about HFT activities were bubbling
under the surface through the second half of the
last decade but were brought very much into
the public eye in the aftermath of the “flash
crash” in May 2010 – even though subsequent
evidence proved HFT was not to blame for
events. Since then however, HFT has been on
the defensive as a strategy even though some
larger firms continue to contribute to market
liquidity without causing any undue disruption.
Nowhere has the assault on HFT been fiercer than in Europe where suggestions have covered much of the
spectrum from outright banning of the strategy to mandatory
registration of firms and their strategies. The European
authorities, along with those in the US, continue to work to find a
solution to the “problem” of HFT, especially in equity and futures
Into this mix has be thrown a working paper, A Dynamic Limit
Order Market with Fast and Slow Traders by Peter Hoffman of
the European Central Bank, which studies the role of high
frequency trading in limit order markets. Although the paper
looks at equity markets there are, of course, some parallels that
can be drawn in FX markets. Equally, should the paper’s
recommendations be taken up by the European authorities, there
would be implications for HFT firms in all financial markets.
The paper seeks to throw more light upon the debate which has
polarised much of the financial world, academics and even the
authorities: Does HFT contribute to market instability or
The author builds two trading models that operate in the same
way, in the same market, but with two distinctly different speeds.
The intention is to highlight the advantage fast traders have over
slow, especially when it comes to being “picked off” by other
traders as new information comes into the market.
The paper argues that fast traders do not risk being picked off
by slow traders when market conditions change through the
arrival of news or other information, it also points out that fast
traders can alter orders in the market quicker than slow traders.
This is true of course; however, it should be noted that the vast
majority of fast traders are governed by algorithms which react to
market data, and not news. This means that a slow – or human –
trader, seeing the news has the (brief) opportunity to exploit that
information to perhaps trigger the move that the fast traders’
algorithms are waiting for. This is one example of where human
can beat machine in financial markets.
The model cannot take this example into account due to its
parameters, instead it merely seeks to demonstrate the advantage
that high speed traders have over their slower brethren when it
comes to resting orders. The author also argues that slow traders
have a difficult choice when it comes to posting orders, “They
can either keep their chances of execution constant by posting
more aggressive quotes that are attractive to both [slow traders] and [fast traders], or instead accept a decrease in the execution
probability of their limit orders by only targeting [slow traders].”
One could reasonably ask, if the slow traders were posting bids or offers at a level then it should be assumed they want to buy/sell
there? In this the analysis reflects more of a bygone era, when
market making was a largely manual task. Now, unless a trader
has a specific interest, the machines have driven the manual two-
way market maker out of the business – at least in liquid markets.
The research goes on to note the model used “predicts that [fast
traders] are more likely to be [market] makers than takers…and
display a higher ratio of limit to market orders than [slow
The paper claims that slow traders, upon observing the entry of
HFT to the market, react by submitting orders with a lower
probability of execution, a claim it is hard to argue with when
high frequency traders have proven over the years to quote in a
much more granular fashion than manual traders.
Tax Rather than Speed Bumps
The big thing to come from the paper, however, is a suggestion
that curbing HFT activity will merely build inefficiencies in
markets. Of course, for some in the financial markets industry,
this inefficiency is the price customers pay for accessing
guaranteed liquidity and, in some cases, credit.
Rather than suggest speed bumps, such as those put in place by
some foreign exchange platforms, the paper declares that taxation
is a better option. “Currently, a number of different measures are
being discussed among regulators that aim to curb HFT activity,
including minimum order resting times and limits on the number
of messages that individual traders can send,” the paper states.
“While these measures could eventually improve on the market
outcome by eliminating HFT activity, the present analysis
suggests that they would also deprive the market of potential
efficiency gains because it is precisely fast traders’ ability to
revise their limit orders quickly that allows more gains from trade
to be realised.
“It may be more effective to tax HFT activities directly in order
to ensure that at least some of the potential benefits of speed are
reaped,” it adds, going on to note that fast traders create market
efficiency by being quickly able to revise orders after news arrives. That said, “While this dampens an existing inefficiency, it
constitutes a loss in bargaining power for slow market
participants and leads them to submit limit orders with a lower
The paper cites the practice on the part of some trading venues
to fine users for “excessive” system usage – and highlights what
it terms a “sensible” solution implemented by Intercontinental
Exchange to only penalise traffic beyond the best quote.
The idea of fining “excessive” traffic has one obvious flaw –
how can you define what is “excessive”? Equally, something that
even a committee of professionals under the auspices of the
Commodity Futures Trading Commission, took an eternity to
decide – at what point does one cross over to be a high frequency
When deciding the threshold of “excessive” traffic, a venue
risks being too conservative – adding to market inefficiencies in
the eyes of the author – or not being strict enough so that it
maintains what the author (and most of the authorities) seem to
believe is an unfair advantage for fast traders.
Indeed it can be argued that by using the term “fast traders”,
the paper is highlighting the conundrum. In the foreign exchange
market, for example, several banks can be termed “fast” traders –
would trading venues wish to curb that type of activity?
If an effective definition of high frequency can be reached, then
the European authorities could follow the idea from the paper, to
tax these traders. Of course, other traders could be caught up in
the definition, therefore the overall level of market efficiency may
increase, but it could be a thinner market.
If the authorities get it right, then this could be another blow
against high frequency trading which is already reeling from
assaults on several fronts – both geographic and market.
It has been noted in these pages before, that one problem that
faces HFT firms is unique, when compared to previous targets
(hedge funds and banks), because unlike the latter, the wider
populace cannot access or use the services of an HFT in a
quantifiable manner. This leaves the strategy vulnerable to
regulation or even worse taxation, and the probability is that the
only potential source of support – other traders – is unlikely to be
realised because as one trader recently put it, “High frequency
was great, it made the banks raise the bar – but it would be good
if it went away now.”