The future of trading

I recently chaired a breakfast meeting on the
future of trading with three key figures from the industry: John Serocold,
Director, London Investment Banking Association; Wolfgang Eholzer, Head of
Trading Systems Design, Eurex; and Eli Lederman, Managing Director, Electronic
and Agency Trading Services, Morgan Stanley.  I thought you might be
interested in their views, so here's the opening salvo from each speaker.
A complete transcript is available if anyone wants one.  Just mail me.

CHRIS SKINNER:     
The debate is on the future of trading and will there be any traders around in
15 years, or 10 years, or 5 years?  A report came out last year about
algorithmic trading and latency, which said that nine out of ten dealers would
be out of their jobs by 2015.  I have readily asked people that question,
“Do you believe that nine out of ten dealers will be redundant within seven or
eight years?” and everyone in the market tells me, “No, no, more like four out
of ten, three out of ten, five out of ten” depending on what sort of company
you work for so.  We will get into some of that discussion this morning by
asking John to pick up on the theme.

JOHN SEROCOLD:  Thanks very much,
Chris.  I should preface my remarks by reminding everybody that making
predictions is dangerous, especially about the future.  It seems to me
that the future of trading is going to be bound up with a number of existing
themes some of which Chris touched on in his introduction and most of which
will play out in either a major or a minor key over the next few years.
The laundry list of themes includes security, end to end authentication and all
of the issues around that.  I was interested to read this morning that the
people who are trying to crack the system from the outside are now turning
their attention to Facebook and Bebo as a way in to identify theft.

          
My second theme is constant refreshment of the infrastructure and you only have
to look out of the window of this room to see the physical manifestations of
that.  That green glass building was built as the London Stock Exchange in
the 1960s and it is just being refreshed as another landmark for the city of
London.  All those cranes are contributing to the constant refreshment of
the infrastructure just to accommodate the human beings.  What you can’t
see from here that is equally important is the investment that is going in -
invisibly, silently - all the time into hardware, software, communications and
capacity.

          
When buildings in London started to get taller the market leaders at the time
thought they were going to be running with hydraulically powered lifts so an
enormous network of relatively large tunnels was quickly built under the city
of London to provide high pressure water to power the hydraulically powered
lifts.  This was at a time when electricity was in its infancy and thought
to present a fire risk.  Those hydraulic tunnels quickly fell out of use
but then had a new lease of life in the 1990s when it was a very convenient
place through which to throw large quantities of high capacity fibre optic
cable to take advantage of the next turn in the technological progress.

          
That leads me to my next theme.  Constant refreshment of the
infrastructure costs money and the end to end cost of trading still needs to
fill.  This is a major theme of LIBA’s work over the last few years and
for now, in my introductory remarks, I will just divide those costs into two
buckets.  There is the cost that the investment banks pay away externally
to infrastructure providers which continue to be too high, and the second set
of costs are costs about which the investment banks would dearly love to do
more but they cannot, which is the cost of complexity - the cost of a broken
Europe, the cost of the absence of a single approach to regulation, market
microstructure issues, conduct of businesses issues.  The potential for a
significant win in that area as a result of the Financial Services action plan
is very much in the yet to be delivered box as far as the industry is
concerned.

          
Fourthly, I should mention attention to detail.  I think attention to
detail as a theme will continue to be very important over the next few years
and attention to detail not just in capturing business requirements in order to
code them but attention to detail in terms of the details of the application of
the laws of physics will be important.  We will come later on to the
services that some of the major exchanges are offering called co location which
is an attempt to work around the fact that there is a thing called the speed of
light by allowing people to put very resource hungry applications very, very
close to the exchange’s trading engines rather than have the tiny extra latency
of sending large quantities of messages from the bank’s trading engine to the
exchange’s trading engine.

          
Along with attention to detail goes constant attention to the changing needs of
the client.  The automation story, more or less all my life and for some
time before that, has been a story which is fundamentally about
simplification.  If you go back to Ford’s insight into the production
line, if you look at robotisation of production, if you look at the prefabrication
of buildings - and again you can see that out of the window - deskilling jobs,
making them easier, making them easily replicable, making them easily
exportable doesn’t necessarily fit very well with the changing needs of the
client.  Clients’ needs have always been and will continue to be very,
very heterogeneous.  I am sure that with 20 of you in the room we will
have at least 25 opinions about the future direction of the financial
market.  You are all at different stages in your lives; you all have
individually different financial needs - scale that up across the population of
Europe, scale that up across the population of the world.  The clients are
always going to be heterogeneous and the challenge for bankers and investment
bankers is to find ways of meeting those needs efficiently.

          
So I think that is probably enough from me but that’s just to set the scene and
give you an idea of some of themes which are going to influence the future of
trading over the next few years.

CHRIS SKINNER:    
Wolfgang from Eurex, give us a little bit more insight into where the markets
are going and how technology in the markets and global integration of markets
are coming together.

WOLFGANG EHOLZER: Yes, from my side
we have heard a lot of, I would say, general principles which for sure will
lead the way into the future.  From my side, looking a little bit more on
the technology side of things, I can see for any place where trading really
occurs, whether it is an exchange or trading venue, ECN or whatever, at least
five criteria which will determine whether they will be successful.

          
These days, obviously latency is one of those; that is an issue.  I think
one has to be careful in talking about latency because there are different
definitions and I think up front it is one of the important things, talking
about things to try to define them.  So at least, in my high level and
naïve point of view, when somebody enters his order to buy or sell whatever it
might be until he gets the response, what really has happened to that
order?  You send, “Buy me 100” and I get back, “Bought 50, 50 in the
book”.  That is one part of latency which I think is important.

          
More importantly, I think that achieving low latency is definitely not enough
because I think what we’re seeing in the market situation at the end of
February, beginning of March and also in August, is that a very high
consistency is extremely important for the market.  So even in situations
when there is a lot happening, all the people who trade whatever it is want to
see the same behaviour of all the IT infrastructure; whether it’s from outside
providers, exchanges, ECNs, they behave exactly in the same way which is a
challenge for all the guys who are doing the technology side of things because
volumes are up by a factor of three, four, five or ten maybe, and in a peak
situation that is a challenge.  So I think consistency is extremely
important.

          
I think the third thing on my list is obviously high speed market data, so it’s
not only knowing what your order has been doing - whether it has been matched
or you’re in the book - but it’s also to know where is the market and that, of
course, also consistently with low latency.

          
I think the next two things, and one of them is probably mine -- I think there
is a general trend that trading venues don’t want to have a large footprint on
members or participant’s side, whether it is exchanges or ECNs, to bring down
the cost so that there can be consolidation on the service side of things or
technical infrastructure side of things.  But there is also one, I find
extremely important, point and I’m pretty sure we will hear more of that in the
future and that is really state of the art risk management.  If we see
these days investment decisions being taken by machines with extremely high
speed and we see order rates coming in at hundreds per second then the value at
risk there is extremely high on very low time scales.  From my perspective
it is quite clear that there will be drastic changes happening on this side
because it can’t be only the trading side of things which speeds things up; it
has to be the clearing and the risk management as well.  I think we will
see a lot of that happening in the future.

          
Now, from a derivatives exchange point of view, of course they are different
investment strategies.  It’s not everybody who needs this ultra low
latency and all the other attributes.  There are strategic investors who
want to buy 5% or 3% of a company, maybe have a certain influence on certain
decisions at that company, and they are thinking about a much longer time
scale; they are not doing statistical arbitrage.  So there are very many
different groups of investors and I think this is another generalistic thing we
will see in the future, that there is more diversification also with respect to
technology.  There are some people who need this high speed game, which in
the end will also be at some cost to them because keeping up in this race for
speed will and already leads to costs on the side of the people playing this
game, so we will have a strong segmentation of the market.

          
All of that said, obviously we see the electronic pits forming these
days.  You used to have the pits with the traders; these days it is all
the service moving close to the exchanges wherever it is around the globe and I
am pretty sure that we will see this not only happening on the trading side but
also happening on the clearing side.  Risk management has to be almost
real time soon because the risk of all these guys sending their orders in at
very high frequency is going to add up.  There is no point in knowing the
risk two hours later or at the end of the day or whatever.  I think those
people who take the risk, typically in the exchange business side, these are
the clearers who have to know that in time.

          
Now, long term, coming back to your question, I strongly believe in traders and
human beings so I am pretty sure that we would see in the area of, say, more
straight forward, plain vanilla, high liquid products less human interaction,
so less people getting a job there - more machines or more people who know how
the machines work and can be configured and tested and whatever.  But I
think there will be plenty of other complex products like complex derivatives
pricing which you don’t want to hand over completely to a machine and where
innovation is extremely important.  Maybe it’s different kinds of people
sometimes but personally I think there will be a lot of humans still involved
on the trading side.

          
I also think that we will see more of a move that more products will come on
exchange or on large trading venues which are OTC at the moment to essentially
in the end bring down the total cost of the process by means of standardisation.
I am pretty sure derivatives will continue to grow at enormous rates in the
future and maybe long term we will eventually also see that we get some global
geographical clustering of exchanges.  If latency really is very important
then that might lead to the fact that all the services of the exchange or
trading venues -- look, for example, at the States’ equity markets; there are
lots of trading venues there.  There is a natural centre for equities
trading in the States.  New York.  For futures -- obviously that is
Chicago.  That will be the next step.  If that concentration process
goes on will there be only one financial centre in the States and everything
moves to Chicago maybe, or here to London or to Frankfurt or to Milan or to
Paris?  But you can’t rid of the laws of physics; there is the speed of
light and on cables it’s not even the speed of light in a vacuum - it’s a
little bit less.  So maybe that is driving things together so that in the
end we have only one or two places per time zone on the globe where they at
least say the bulk of the trading is really going to be conducted.

CHRIS SKINNER:    
Let me just ask one question.  Looking at your background, you’re a
physicist and when you look at the skill set in trading these days, the skill
set is rocket scientist.  There are people who have mathematical degrees,
there are people who understand quantitative analytics whereas 10, 15 or 20
years ago those skill sets didn’t have that value.  So how are the skill
sets changing and is it going to be a future which is dominated by all these
rocket scientists?

WOLFGANG EHOLZER:  Yes and no, of
course.  I would say on this plain vanilla, high liquid stuff which will
be highly automatic in the future and already is in parts, you will need people
with a very solid IT background and understanding of the laws of physics, I am
pretty sure.  On the other hand there will be a lot of complex products
and I think the skill set that was good 10 years ago or 15 years ago to enter a
job in the financial industry will still be good in this area, I am pretty
sure.  So I think we will have more people with a natural science
background moving in on one side of the trading product range but there is lots
of space for, I would say, traditional people there, I am pretty sure.

ELI LEDERMAN:       
I completely agree with that.  I think the demographics of who gets
employed by investment banks has changed dramatically and much more so on the
trading side of things probably than on the pure investment banking side of
things.  There is certainly every reason to believe that that
continues.  It probably goes some way in beginning to address your
question about exactly how many traders there still are employed, first of all
after banks report earnings this week and then looking a little bit further out
5 or 10 years or whatever.

          
The question I think people on the inside would probably argue is maybe not
completely relevant or answerable or other things without clarification.
The reality is that the very definition of what a trader is is changing and you
might more fairly ask how many people are going to be employed by operations
that do trading because it really is all about building systems, designing
systems, using systems and these are things fundamentally that humans do and
humans who we call, and I think other people call, trading scientists as much
as anything else.  It is certainly remarkable.  I usually don’t tell
this to anyone although I think it may be in whatever kind of bio is in your folder,
but I also have a PhD in physics.  I didn’t realise you did so there is at
least something about the fact that four panellists ... we have 50% physicists
up here.

          
I think looking at the next, say, five years I see a few major trends playing
out and they are in chronological order: proliferation, consolidation and
rationalisation.  Looking at these I realise that if I put them in a
different order it would be C, R, P and if I had thought of something that
began with an A, I would know I was on the wrong track but in any event --
automation, thanks, very helpful.

          
So, proliferation.  We are clearly on the cusp of some very dramatic
changes in the trading landscape, the proliferation of trading venues.  We
have now the Chi X platform from Instinet which is really very rapidly
attracting liquidity away from the primary exchanges in Europe.  We have
Project Turquoise which my firm and others in the city are involved in which
will create a Pan-European MTF, or multilateral trading facility, which of
course is promoted -- or that sort of competition is promoted by MiFID although
it very likely would have happened with MiFID or without.

          
When there are multiple liquidity buckets you inevitably get fragmentation.
You have liquidity in bucket one, two or three and it becomes someone’s job to
segregate that liquidity.  That, of course, is called fragmentation.
The thing that happens immediately upon the creation of these alternative liquidity
sources is fragmentation but the fact that the orders get smaller and smaller
and people spend more time trying to hide what they’re doing in the various
markets which itself leads to the proliferation of more data.  So
something that used to be one order for 100,000 shares becomes 1,000 orders for
100 shares; as we’ve seen the trend inexorably lead in the US that will happen
in Europe.  So you really have an exponential growth in data -and it’s
entirely appropriate that COLT is buying the croissants today because they’re
probably the only real beneficiaries of this - as people develop this sort of
insatiable requirement for things like bandwidth and low latency, and of course
right around the corner we will have digital news and so on as someone alluded
to earlier.  Digital news driving investment decisions will lead to a
whole new type of investment strategy and there will be more black boxes and
they’ll do more insidious things when the markets get rocky.  So, anyway,
as I say that’s the proliferation side of things.  It’s all about
incredible complexity and the requirement, of course, for more people to design
and build systems to deal with it.

          
The next phase I would predict is consolidation.  People are going to look
at this market place - our competitors; I hope not us and I don’t expect it
will be us - and say this is an extraordinarily difficult, expensive business
to be in.  Indeed, not just on the sell side but a lot of what I’m talking
about is the buy side, who have invested very heavily in their trading
operations over the last few years.  I think maybe it will be a market
turn down or something but people on the buy side are, I think, inevitably
going to look at the trading operations they have developed and ask does it
make sense that there are one million trading desks in the world?  What is
the right number?  Maybe there should only be 1,000; maybe there should
only be 5.  I think it’s entirely reasonable to expect that there will be
something a little bit - I realise there is some irony here - like the old
world where people on the buy side looked to brokers to provide the science and
the technology behind trading and we get into an environment where people are
competing on the objective quality of what they’re doing because these things,
after all, are measurable, are getting more measurable.  With all of the
dark pools and all of the different things that you have to navigate through to
do a best execution type of job, it’s entirely reasonable to imagine a world
where there are only five or ten or whatever providers of such a service.
It will be, in any event, extremely complex and people will have obligations to
demonstrate that they have done as well as Morgan Stanley or Goldman Sachs or
whoever the pre-eminent providers of these services end up being.

          
So, anyway, that brings me to the last of these things and that is
rationalisation.  I think it’s very, very important to pause, to look
around and to see what’s happening and to remember what exchanges were built to
do in the first place.  There could be some debate about this but in
general terms they were designed to be places where people who cared about an
asset and wanted to buy it were able to meet someone who cared about an asset
and wanted to sell it.  What has happened, I think, beyond any possible
argument after what happened in August -- what has evolved is actually a system
of interconnected pools of liquidity that are effectively a playground for
arbitrageurs and people who are not really fundamentally about investing in
assets that they care about.  If that is the way things have been and if
that continues to be the way things go, it makes it difficult to understand
exactly why we all have to invest so much.  So there’s not some evil
machine out there that’s lurking, waiting to gain the system or pick up the
footprint of an institutional order to trade against.  I think in this
last phase, it’s imminently reasonable to expect that market participants,
either in concert or not, look at the market and say, “We’ve got to rationalise
this” and look at crossing tools or whatever to reintroduce some order to the
trading world.

          
Finally, I would just say as we all talk about black boxes, even in August when
they were in every headline when things started to go bad, people noticed and
people turned off the black boxes and people decided when to turn them back on
again so it’s not all as robotic as you might believe.