[ A L G O R I T H M I C
simply don’t trade enough with
them to make serving them worth-
while. While this situation was
allowed to exist in the past, a more
challenging business climate for
investment banks means many now
need to rationalise their client lists.
Among the smaller funds in
particular, many now only rely on
a single broker to provide their
algorithms. While this may be a
tolerable situation for the time be-
ing, when MiFID II is introduced
early next year, it could cause prob-
lems with meeting best execution
requirements when you only have
access to a very limited pool of
algorithmic trading options.
The same trend can be clearly
seen when looking at the pro-
portion of firms using different
numbers of providers seen in Fig
4. While the percentage using
five or more providers increased
considerably from 40.5% in 2015
to 54.5% last year, but this has now
fallen back even harder in 2017 to
34.9%. The particularly large shift
seen here is likely reflective of the
mix of buy-siders reducing their
broker relationships voluntarily
and those with whom the sell-side
is forcibly cutting ties. It has not
been unknown for asset managers
to use as many as 10 different bro-
kers (sometimes even more than
that) but realistically, with such
a large broker pool, the amount
of business some brokers receive
must be minimal and in many cases
the relationship may only exist to
give the asset manager an addition-
al pool of analyst research to draw
from. Both best execution and
unbundling rules that will be part
of MiFID II clearly make this situa-
tion unsustainable for both sides of
the Street. However, using five or
more brokers still makes up more
than a third of firms, and more
than any other category, indicating
firms remain relatively diversified.
While there are less firms using
5+ brokers, there has of course also
been an increase in the proportion
using fewer brokers. Significant-
ly, the percentage of firms using
only one or two algo providers has
reached 34%, only just short of
the proportion using five or more,
“Regulators are now
attempting to push
other asset classes
such as fixed income
into the electronic and
automated space.”
showing that many more firms are
now making do with only a very
limited selection. As alluded to
above, this could pose problems
for the future if firms narrow their
broker selection too much and
risk missing out on best execution
opportunities.
CROSSHEAD
While provider numbers may be on
a downward trend, one area which
never ceases to head upwards is
the proportion of trades handled
by algorithms. The number of firms
trading 40% or more of their value
through algorithms has rocketed in
recent years and 2017 is no excep-
tion. While back in 2015 only 24%
of firms traded such a significant
proportion of their securities via
T R A D I N G
S U R V E Y ]
algo, this has steadily grown to
34.3% in 2016 and 43.3% this year.
It seems almost inevitable that
the continued march of technol-
ogy would push ever increasing
volumes of trades to be execut-
ed through algorithms. Traders
are becoming more comfortable
with the technology now and are
gaining more experience, meaning
they are now more able to make
the most of the algo tools on offer.
It’s also true that both asset man-
agers and brokers have seen their
resources strained and, in these
circumstances, it makes sense to
let an algorithm deal with much of
the day-to-day trading activity so
that a human trader’s time can be
spent on more complex or sensitive
orders. The real question is, have
we peaked? It certainly doesn’t
seem like the growth in electronic
and algorithmic trading is set to
slow down any time soon and reg-
ulators are now attempting to push
other asset classes such as fixed
income into the electronic and
automated space as equities was in
the past. Given the circumstances,
The TRADE expects algo trading
number to continue to grow in
2018. The only potential stumbling
block is a greater expectation by
regulators that buy-side traders
have a comprehensive under-
standing of how the algos they use
actually work, but this is unlikely
to put anyone off using algos, it
will simply increase the onus on
brokers and the buy-side to make
sure traders are well trained.
Interestingly, the area which has
shrunk the fastest is those firms
trading from 30-40% of their vol-
ume via algo. This has gone from
Issue 51
TheTradeNews.com
71