algorithmic trading - attracting the buy side

8/2/2019 Algorithmic Trading - Attracting the Buy Side

1/12

A-TeAmGro

up

AnA-TEAMGROUPPublication

ElectronicTrading presents

Algorithmic Trading:

Attracting The Buy Side

June 2009

Sponsored By:

www.a-teamgroup.com

Whether they accept it or not, sell-side institutions arefnding themselves in the unamiliar role o inormationtechnology vendor. The adoption o algorithmictrading models by buy-side frms o all shapes andsizes is shiting trading strategies, and the technology

inrastructure to supply and support them, rom the realmo nice-to-have appendage to must-have service oering.

With more sell sides than ever oering both standardbenchmarks and their own takes on old avourites,competition between algorithmic trading strategies isheating up. And its not all about alpha. The buy sidedoesnt like surprises. What many und managers areseeking rom their sell-side suppliers is certainty oexecution, low market impact and some degree oaccuracy on hitting stated targets.

As a result, brokers and their technology suppliers areall working uriously to help dierentiate algorithmicoerings, launching custom and so-called adaptivealgorithms, and taking great lengths to prove that theirmodels perorm as stated on the tin.

In essence, the world o algorithmic trading is enteringa new phase as models acceptance by the widermarketplace is increasing pressure on frms to perormand, to some extent, productize their oerings. Expect

more innovation, more customization and more choice.


2/12

As algorithmic trading evolves,more rms seem to be using their

systems to trade equities rom

across the U.S., European, Asian, Middle

Eastern and Latin America markets.

Interactive Data is committed to support-

ing electronic traders by providing them

with consolidated low-latency global

data that includes extensive level 2 data,

as well as real-time and historical tick data.

Following is a look at some key oerings

rom Interactive Datas Real-Time Services

business that may interest clients.

EuropeIn Europe, because no consolidated tape

exists, many rms have had to ace the

challenge o consolidating data across

Europes growing number o execu-

tion venues and building that logic into

their algorithms.

With the launch o Interactive Datas

PlusBookTM, a new consolidated order

book service or the European nancial

industry, rms may no longer need tobuild algorithms with the ability to de-

termine the best price out o a range o

markets. PlusBook provides a consolidat-

ed order book designed to help nancial

institutions gain a more complete view

o global market liquidity by aggregat-

ing orders rom multiple venues, includ-

ing those based in the US.

Providing low latency access to

streaming level 1 and 2 data, PlusBook

has been designed to identiy each trad-

ing venue by Market Identication Code(MIC) and, subject to client entitlements,

covers the markets available on Plus-

FeedSM, Interactive Datas consolidated,

low-latency digital dataeed. Venues or

Algorithmic Trading or the Buy-Side - Interactive Data

June 2009 Issue 022 AnA-TEAMGROUPPublication

which a price is available via PlusFeed,but not part o the customer author-

ized entitlements, are identied as part

o the consolidated view. Customizable

eatures can allow end users to tailor the

coverage provided to the venues most

relevant to them.

PlusBook derives the sum o orders at a

specic venue at a specic price, designed

to provide nancial institutions with an

aggregate o oers rom dierent venues

which can help track market liquidity. Sub-

scribing companies eed the consolidatedorder book data directly into algorithmic

or automated trading applications, there-

by optimizing trading strategies.

PlusBook can be integrated quickly

and easily with PlusFeed, which delivers

a broad range o global nancial inor-

mation rom over 450 sources and ex-

changes worldwide, covering more than

six million instruments.

Emerging Markets

As rms expand their trading intoemerging markets, data coverage avail-

able through the low-latency PlusFeed

has been expanded to include more

emerging markets exchanges.

PlusFeed delivers extensive level 2 data

or developed and emerging markets,

and is designed to provide ull depth o

book and every tick o data. Interactive

Data does not conate any o its level 2

data, and oers customers the option o

receiving all available updates.

Following are some o the emerg-ing markets exchanges recently added

to PlusFeed:

Multi Commodity Exchange (India)

Dalian Commodity Exchange (China)level 1 and 2

Dubai Gold and Commodities Ex-

change

Ho Chi Minh Trading Center (Vietnam)

Hanoi Securities Trading Center (Viet-

nam) level 1 and 2

Ljubljana Stock Exchange (Slovenia)

Zagreb Stock Exchange (Croatia)

Belgrade Stock Exchange (Serbia)

Micex (Russia) level 1 and 2

Tick DataThrough its PlusTickSM services, Inter-active Data provides a range o options

rom a ully managed, onsite database

with streaming real-time tick data to

end-o-day le delivery to help with:

Back-testing algo strategies

Transaction cost analysis

Proving best execution

Compliance and regulatory require-

ments

Pre- and post-trade transparency

For rms looking or end-o-day tick

data, PlusTick FTP is a daily le that de-

livers time and sales, minute bar and

cross-reerence data rom over 60 global

exchanges. For rms that require real-

time streaming tick data, PlusTick Server

is a deployed container that provides

tick capture, storage and management

o millions o real-time and historic tick

messages at the client site.

The upcoming version o PlusTickServer, due out summer 2009, will have

enhanced tick history storage capabili-

ties and will support both level 1 and 2

equities data.

Traders Eye EmergingMarkets, EuropeanMarket TransparencyBy Don Finucane, Vice President o Product Management and OTC Data Services at Interactive Data


3/12

An A-TeAmGROUP publicationwww.a-teamgroup.com

Read the Latest Headlines & Get a FREE Issue Now!

Electronic-Trading.com

A-TeAmGROUP

Electronic Trading is the only publication to ocus exclusively on the electronictrading marketplace. It oers insight into how technology is aecting the business

o liquidity: how sell-side f rms are leveraging new, reliable and inexpensive

technologies to attract order fl ow, and how buy-side f rms are turning new

capabilities into operationally superior business processes.

Electronic Trading will help you understand major f rms activities in the areas o:

Algorithmic trading models

Matching systems for the buy- and sell-side and interdealer segments

Trade order management systems

Liquidity aggregators

Exchange gateways, ECNs, ATSs

Transaction messaging and protocols

View the latest Electronic Trading headlines, read the f rst paragraphsand get a FREE issue at Electronic-Trading.com

To fnd out more, [email protected] call +44 (0)20 8090 2055

Or visit our website:

www.Electronic-Trading.com

SPONSORS

ElectronicTradingwww.Electronic-Trading.com

30-Day Free Trial

Electronic-Trading.com/ree-trial

Access all articles and

resources or ree, no

credit card required


4/12

interesting to reect on where we might

move rom here will algo trading in-

creasingly become a high-added-valueservice, ultimately provided by relative-

ly ew brokerage rms, or might there

be ways in which it can become more

democratized?

Both the buy side and smaller brokers

have powerul incentives or this direc-

tion: the brokers will naturally want to

win back the provision o added value

that they have (arguably dangerously)

ceded to their larger competitors, while

sophisticated buy-side rms can gain

exibility, unique trading advantage and

signicant brokerage cost savings by

taking at least part o their algo destinyinto their own hands.

For both communities, the strength o

this incentive will increase as algo trad-

ing continues to become more prevalent

remembering also the natural growth

multiplier o an algo (now oten ollowed

by a smart router), which almost always

produces multiple trades rom a single

original order.

A Necessary Enabler

A necessary enabler or this broaden-ing o algo development capability is a

degree o commoditization in some o

the algos and also in the underlying de-

velopment platorms. There is an oppor-

Algorithmic Trading or the Buy-Side - SunGard


We see this beginning to hap-pen. Smart Order Routing hasalready been successully pack-aged by ISVs across many clientimplementations, though morecan be done to make it available

on an ASP/SaaS basis

U

p to the present time, outside the

community o trading-ocused

hedge unds and specialist high-requency traders who have undertaken

their own developments, algo trading

has been predominantly viewed as a

sell-side product, provided as part o the

overall range o brokerage services by

sell-side rms to the buy-side.

The growth o algo take-up over the

last several years in both the US and

Europe has been impressive, and the

eld continues to evolve rapidly. It is

worth highlighting some o the signi-

cant developments o the past year ortwo, which have been impacted in par-

ticular by the post-crisis market turbu-

lence and, in Europe, by the increasing

ragmentation o markets since the in-

troduction o MiFID:

Doubts about the universal eec-

tiveness o certain standard approach-

es, which unsurprisingly have turned

out not to t all market circumstances

equally well and now what happens

to VWAP benchmarks in an increasingly

ragmented world? The integration of smart routing into

algo trading processes an execution or

benchmark algo engine having decided

the when or a trade, the smart router

determines the where.

The overall complexity of the above,

leading many smaller brokerages to

outsource their algo provision to Tier 1

rms, some o which have been aggres-

sively marketing these capabilities to

sell-side as well as buy-side rms.

Increasingly wide availability of themajor brokers algo oerings via the net-

works and platorms o the major trad-

ing ISVs and EMS providers.

In the light o these developments it is

tunity here or ISVs and sotware service

providers to move algo development out

o the rocket-science laboratory and intothe everyday world o replicated (though

customizable) packaged sotware.

We see this beginning to happen.

Smart Order Routing has already been

successully packaged by ISVs across

many client implementations, though

more can be done to make it available

on an ASP/SaaS basis. SunGard has also

worked with numerous clients on both

the sell and buy sides to integrate algo

trading platorms into their operations.

These usually run specic implemen-tations o benchmark and execution

algorithms, developed using a work-

bench (Tactics Studio) that is specically

adapted or the purpose o trader-driven

development. Again, the next step is to

take the operational complexities out o

the process by providing all o these ca-

pabilities as ASP services.

Not All or NothingThe decision whether to work in house

or to buy rom a broker is by no meansan all or nothing scenario. As mentioned

above, the algos and smart routers o the

major brokers can also be reached via in-

tegrated ront-ends and DMA links across

the order routing networks o SunGard

and the other major providers.

We envisage a exible and open-ar-

chitecture approach in which an asset

manager or broker can optimize which

parts o its algo/smart routing strategy

it chooses to implement in-house or in

ASP, while accessing external brokerageservices or others. This seems likely to

be the best route that the industry can

adopt or delivery o cost-eective serv-

ices to investors.

Algorithmic Trading:For and By the Buy-Side

By Vincent Burzynski, Chie Product Ofcer, Global Trading, SunGard


5/12

An A-TeAmGROUP publicationwww.a-teamgroup.com

MiFID Monitor Expands toBecome Risk & Regulation IT!

A-TeAmGROUP

MiFID Monitor is nowRisk&RegulationIT, with a broader ocus on:

Regulations and their impact on financial markets infrastructure, data, technology and systems

Emerging regulations as they are discussed and brought into affect such as possible regulationo hedge unds or ratings agencies

Input from the global regulators themselves

Risk management - data and technology-based approaches to managing market, credit,operational, and liquidity risk, and the likely evolution to an enterprise risk managementramework

Governance, risk and compliance systems and suppliers

Impact of derivatives and alternatives on regulatory and risk systems

Industry initiatives and best practices, and more.

To receive your 30-day ree trialor to sign up or a subscription(495 / $745), contact:[email protected]: +44 (0)1522 850 913

Or visit our website:

www.Risk-RegulationIT.com

Risk& Regulation IT

www.Risk-RegulationIT.com

SPONSORS

30-Day Free Trial

Risk-RegulationIT.com/ree-trial

Access all articles andresources or ree, no

credit card required


6/12

Algorithmic Trading or the Buy-Side - Roundtable


The use o algorithms has

been embraced by the main-stream. To what actors doyou attribute their broadappeal, and are there par-ticular buy-side frms that arebest suited to their use?

Gozlan: I think there are several ac-tors that trigger the broad use o algos

by the buy-side community, depend-

ing on the category o the buy-side irm

and considering that algos can be clas-siied in our categories : alpha seek-

ing ; best execution ; market impact ;

and cross-asset execution. O great-

er importance, I believe, is the need

to achieve best execution in a deeply

ragmented market, and secondly, to

reduce market impact. This concerns

most o the buy-sides trading, includ-

ing non-listed activities coverng Fixed

Income and Foreign Exchange.

The aim o high-requency trading

irms is to generate direct proits romtheir alpha-seeking algorithms. For

these hi-requency groups, owning a

stake in a venue and controlling the

way the engines are built and inter-

act with their algos has been a key to

being successul.

But consider this : You could reverse

the question and ask, could we run a

buy-side business today without using

algorithms? In equities, probably

not, but in other markets like OTC in-

struments, the answer would be yes.However, in the rapidly changing For-

eign Exchange markets, the answer

would be no. As or cross-asset trad-

ing, perhaps the burden lies on the

sellside invited by buy-side irms to

provide them with multi-asset or evencross-asset trading capabilities.

Idelson: Dierent considerationsdrive the demand or algorithmic trad-

ing on the sell and buy sides, although

the sell side will try to service customer

demand by delivering packaged prod-

uct to generate customer loyalty. The

primary appeal rom the buy-side per-

spective is that or a rm with a success-

ul trading strategy the single best ROI is

oten the introduction o an appropriatealgorithmic execution strategy to gen-

erate execution price improvement.

When assessing which buy-side

irms beneit the most, clearly trading

requency is an important actor. The

direct contribution to proits o good

algorithmic execution increases with

trading requency to the point where

at very high requencies the execution

strategy can become as important as

the underlying trading strategy. How-

ever, even irms with long-term trad-ing strategies, which may have pro-

duced at most only daily baskets to

make small adjustments to their trad-

ing positions, are recognising the key

advantages o algorithmic execution.

The increases to annual ised returns

can be signiicant and provide a mar-

ketable edge over competitors.

Additional synergies are available or

those buy-side irms that also employ

algorithms in their trading strategies,

as well as as or execution and smartorder routing (SOR). Its worth noting

that a signiicant number o irms do

employ algorithmic strategies or parts

o their trading and that they cover a

Algo Trading:Attracting theBuyside

Ask the Experts

Harry Gozlan, founder and CEO,

Smart Trade Technologies

Nick Idelson, Technical Director,TraderServe


7/12

Algorithmic Trading or the Buy-Side - A-Team Group

June 2009 Issue 02 7AnA-TEAMGROUPPublication

spectrum stretching rom low-latency

non-vanilla arbitrage through to long-term statistical models.

What are the key elementsbeing used by algorithmicstrategy providers to dier-entiate their oerings or thebuy side?

Gozlan: Obviously, this is a broad-basedquestion requiring a detailed reply. It could

be latency, oering co-lo hosted acilities.Or design and modeling tools or the algos,

including simulation and back-testing o

applications. Or packaging under one UI-

accessible interace. Or integration with

OMS and post-trade systems. Or purely

providing algos that can do better than the

competition. Or, last, the cost or the classi-

cal access to some undamental research in

the case o brokers.

As weve learned at Smart Trade, our

dierentiating elements include the ex-

ibility and design o the various combina-tions o rules-based aggregation, smart

order routing and crossing eatures that

our engines enable, in an extremely

open, ast and secure way the opposite

o a xed-packaged application.

Still, no matter who the vendor may be,

they must position their oering dier-

ently or the buy-side than the sell-side,

but the underlying technology, i open

enough, could very well be the same.

Idelson:The sell-side is understandablymost concerned with demonstrating reg-ulatory compliance with best execution.

In the U.S., this can lead to a narrow ocus

on achieving the statutory minimum best

execution, while in Europe MiFID requires

periodic review and measurement o the

best execution policy.

This results in vendors needing to

provide not just straight OATS-compli-

ant reporting but drill-down tools or

examining trading by customer and

market. Ideally, the algorithmic plat-orm should also supply transaction

cost analysis (TCA) tools, which allow

ingerprinting o the cause o each

order slice and resimulation over actual

traded history to allow the algorithmic

execution strategies to be properly as-sessed and improved. Pure backtest-

ing o algo execution strategies cannot

properly assess market impact and

while necessary or development pur-

poses is not a good way to evaluate re-

al-world perormance.

All this oten distracts rom using algo-

rithmic execution as a true execution im-

provement tool. It is a challenge or a sell-

side-ocused algorithmic trading platorm

vendor to repackage its products to meet

the real requirements o the buy side, butmany are trying to embrace this.

Key actors or buy-side rms do vary

partly by size o rm and asset classes

employed, but common issues include

improved execution prices to boost per-

ormance o existing trading strategies,

and the ability to minimise inormation

leakage and disguise order ow.

For those rms with onsite rather than

collocated technology ability, a Zero

Latency mechanism can help compete

on a level playing eld. This approachmoves maintenance o multiple order

slices at multiple price levels across mul-

tiple execution venues into a trade con-

troller (providing automated order state

management and synchronisation) dis-

crete rom the execution strategy itsel.

Orders can be executed in the market

ahead o institutions with much higher-

speed market access by intelligent man-

agement o queued order slices pre-

placed in the market both at and away

rom the best bid and oer.When moving on rom conventional

algorithmic execution strategies to cus-

tomised strategies, trading rms oten

look or ease and speed o development

independent o the platorm vendor

and the means to maintain secrecy or

in-house strategies. Vendors typically

address this by allowing their platorm

to call or be called rom external proc-

esses and by providing rameworks o

primitives to cope with real-time event

handling and trade processing. Somevendors also supply graphical interac-

es, which can simpliy the algorithmic

execution strategy development proc-

ess while maintaining exibility.

Is beating the benchmark im-

portant, or are buy-side prac-titioners more interested inconsistency and hitting speci-fed targets?

Gozlan: I think the latter, actually.Beating the benchmark is their job, o

course. What eective algorithmic trad-

ing solutions should provide is the abil-

ity to embed in various ways the logic

enabling each buy-side irm to hit spe-

ciic targets, irst in terms o precisiono execution and reliability in the trad-

ing patterns overall, and then having

the cleverness to beat the benchmark.

I dont believe that trading provid-

ers can pretend to oer the buy-side the

same systems to everyone to systemati-

cally beat the benchmarks. Only custom

behaviour, based on each rms character-

istics, can enable you to achieve this goal.

Idelson: In general the buy side

wants alpha. I the return o their ex-isting strategies can be boosted by the

use o algorithmic execution to gener-

ate simple execution price improve-

ment, that is usually a more important

goal than speciied targets.

That is not to say that any additional

risk is acceptable. But well-designed

algorithmic execution platorms oer

integrated risk management to detect

and prevent this. This point is well

demonstrated by the common desire

o buy-side irms to want to move onrom the packaged algorithmic trad-

ing strategies provided by the sell side

These strategies are oten target-based,

as typically this allows demonstration

o compliance with an agreed execu-

tion policy and thereore can be a sae

option or a sell-side irm oering.

How is the growing ragmen-tation o liquidity, particu-larly in Europe, impacting the

design and use o algorithms?

Gozlan: Greater agility and speed areneeded. Typically, passive orders have to


8/12

be managed in a very dynamic manner,

dierent rom aggressive orders. This iswhere a lot o the ticks in the execution

can be saved. With the appearance o

dark pools, the search or liquidity is also

more complex.

To this extent, algos cannot manage

all o the execution optimization proc-

ess. At some stage, they need to rely pn

a very thorough liquidity management

system in order to manage the states

o these complex execution processes,

beyond the algos themselves.

Idelson:The most eective algorithmictrading platorms encompass smart order

routing unctionality, which allows algo-

rithmic execution to cope with multiple

execution venues including dark pools.

Due to the growing number o MTFs

and dark pools, the design o algorith-

mic platorms, execution strategies and

particularly their SOR components now

needs to be lexible enough to quickly

and easily add new execution venues.

In Europe, at least the sell side is likelyto drive the platorm vendors urther in

this direction, since it will become dif-

cult to periodically review best execution

policies (a MiFID requirement) without

access to multiple execution venues.

What are the new considera-tions when developing todaysmodels? Have adaptive algo-rithms proved useul to buy-side practitioners, and in what

ways?

Gozlan: Theyre using some real-time inormation, such as current

latency to all trading destinations,

whether theyre brokers, dark pools or

exchanges. Close proximity to execu-

tion venues will clearly impact the like-

lihood that traders will hit the target

price as expected .

Looking at this another way, the

measurement o hit-ratios enables youto measure the quality o each trading

destination, meaning the percentage

o missed hits on each venue.

Other post-trade statistical inor-

mation, such as the number o tick-

ets/trades per order, will also have animpact on the overall execution cost

management o your strategy.

Idelson: It is increasingly importantor modern algorithms to cope ei-

ciently with substantial and growing

volumes o order book data. This cre-

ates challenges o its own or the plat-

orms. Vendors need to produce archi-

tectures capable not only o handling

sophisticated models that can gener-

ate ongoing price improvement, butto do so using distributed processing

and ever lower latencies.

Algorithmic execution strategies can

take into account substantial tick and

order slice market impact histories to

adapt to current trading conditions.

Controlling this adaptation is vital to

prevent unintended consequences,

so parameter variation rather than

wholesale change in algorithm is a

more usual course.

A normal practice would be to usetransaction cost analysis to review

market impact and adjust execution

strategies explicitly. Review and im-

provement o algorithmic execution

strategies should be considered an im-

portant business process key to main-

taining a irms market edge in execu-

tion price improvement.

What kinds o perormanceinormation are buy-side

users o algorithmic modelsrequiring?

Gozlan: Most algorithmic tradingtechnology today comes rom either

the sell-side or third- party platorms

that oer a ully packaged (but hard-

to-change) application, oten hosted

and provided through an ASP model.

The ootprint o standalone engines,

such as a smart-order router, or in some

specic cases crossing engines, is stillrather small but this has begun to change

due to the increased demand or a higher

level o independency rom brokers and

third-party trading platorms.

In several cases, implementing your own

algo-trading solution can be benecial bybeing able to control and personalize strat-

egies involved in the trading process.

Idelson:This varies widely, oten withsize o irm, traded sector and the irms

internal risk management processes.

The most general statistic used to

assess price improvement is imple-

mentation shortall against generation

time o trade. This is a relatively clean

and simple benchmark compared to

other post-trade statistics.Transaction cost analysis and market

impact studies are increasing in

popularity.

In general, irms express interest in

proo that the algorithmic execution

strategy, whether developed in-house

or by a third party, is perorming in

line with its mandates and execution

policy and without the introduction o

any unexpected risk.

Having dierentiated throughout be-

tween sell-side and buy-side irms overtheir interest in best execution report-

ing and compliance, virtually all irms

are interested in reports suicient to

meet their regulators requirements.

How equipped are buy-sidecustomers rom a technologystandpoint to take advantageo algorithmic trading capabil-ities? What has been buy-sidefrms appetite or third-party

or outsourced trading tech-nology platorms that enablealgorithmic trading?

Gozlan: Today, its mostly outsourcedusage. However, speciic to your ques-

tion, the use o SORs to dispatch li-

quidity between brokers, or between

brokers and trading platorms even

though theyew oering dierent de-

grees o SOR capabilities to the market

is a very clever way to control its li-quidity, execution and improve returns

on execution. In truth, a SOR sitting

on top o the brokers SOR can prove

extremely eicient.




9/12

April 2008 Issue 51

Algo trading ad

A-TeAmGRO

UP

The A-Team Algorithmic Trading Directory is the industrys only reerence orproessionals active in the algorithmic and electronic trading community.

The directory provides:

An easy to use guide to help buy-side professionals understand the algorithms on

oer rom their brokers and other trading counterparties and suppliers.

A series of detailed supplier profi les that offer descriptive information on the

algorithms available.

Analysis of who should be using the algorithms, under what market conditions and

when.

Download your ree copy o the second edition o

A-Teams Algorithmic Trading Directory at

www.algotradingdirectory.com

AlgorithmicTrading

Directory2009

Sponsored by:

Make sure youdont miss out.

Download your

reecopytoday!

Second Edition

NOW AVAILABLE

The Who, What, When and Why of the Algorithmic Trading Universe

The Second Edition of A-Team Groups Algorithmic Trading Directory for 2009 is now available


10/12

Idelson: It is commonplace or buy-

side rms, particularly those that al-ready have established IT inrastructures

or trade generation, to attempt to FIX/

DMA-enable their existing sotware at

an order or order-slice level. While this

may be an economic use o in-house de-

velopers, the results are usually disap-

pointing. Without a separate execution

algorithm and platorm capable o man-

aging the real-time data and implicit

synchronisation o trade slices with the

execution venues, price improvement is

oten marginal or non-existent.Better results are usually obtained by

using third-party algorithmic trading and

smart order routing platorms. The plat-

orm selection decision depends upon

the buy-side rms current methodology,

technology, uture plans and algorithmic

expertise. In our experience, rms are

best to agree a small-scale proo o con-

cept with an algorithmic trading plat-

orm vendor with well-dened success

criteria to ensure that the platorm se-

lected meets the rms real-world needs.This can also help a busy buy-side rm to

establish cost-benet beore contracting

or a ull platorm deployment.

There is denite buy-side appetite or

the right platorm with varied support

level requirements provided it is per-

ceived to t that organisations needs.

It would be air to say that enthusiasm

or third-party platorms varies depend-

ing upon the success o other third-

party projects within the rm in ques-

tion. Outsourced algorithmic platormsappear to have gained in popularity, but

or many buy-side rms having hands

on their technology remains axiomatic.

How ar have they embracedOMS/EMS platorms and smartorder routing systems? Do youoresee use o SOR algorithmsto seek out and execute liquid-ity between brokers?

Gozlan: I believe that major progresscan be realized by oering a hosted de-

ployment o their privately owned SOR

and algorithmic platorms, connected

to their OMS, brokers and venues in a

secure, low-latency set up, controlledtotally by their strategy teams.

This is very dierent rom using an ASP

version o an algorithmic trading plat-

orm or service and would combine the

outsourcing o the heavy duty inrastruc-

ture as well as the exibility and sharp-

ness o a sel-controlled liquidity man-

agement systems. In act, Smart Trade is

currently developing such an oer.

Idelson: Penetration o OMS/EMS

across the buy side is very variable.Smaller irms oten attempt direct FIX/

DMA-enabling o their existing trade

generation technology as the irst step

rom daily baskets towards algorithmic

trading, rather than install an OMS/EMS.

Larger buy-side irms requently do in-

stall OMS/EMS as the most practical

way to operate trading desks, enorce

compliance and handle reporting.

Conventional EMS smart order routing

modules have tended to be inexible and

o limited use to the buy side this is un-derstandable when latency is taken into

consideration. I the buy-side rm is using

onsite technology, as most do, the latency

via the EMS and telecoms to execution

venues is requently relatively poor.

The zero-latency approach, which can

be so successul in levelling the playing

eld with larger rms and managing

higher physical latencies or algorithmic

trading, is harder to apply in the case o

smart order routing without risking at

least some overcommitted trades.That said, dark pool probing SORs are

virtually certain to grow in popularity

as ragmentation increases in act,

while higher latencies or typical buy-

side irms do complicate conventional

SOR between transparent execution

venues, this actor may be o more lim-

ited impact in the case o dark pools.

What are the next major de-velopments in terms o en-

couraging the buy sides useo algorithmic trading?

Idelson: Buy-side irms are gradu-

ally understanding the advantages o

algorithmic trading technology. Priceimprovement adds to the bottom line

and provides one o the best ROIs or

many irms. Deploying algorithmic ex-

ecution strategies on the buy side pro-

vides key advantages:

use of execution strategies that fit

the buy-side irms trading and can

optimise price improvement rather

than aim or arbitrary benchmarks in

an attempt to show an actual or as-

sumed minimum best execution.

Disguise of trading activity. Exe-cution algorithms typically use ran-

domisation and other techniques to

a stay below the radar. With a buy-

side algorithmic trading platorm,

the order slices remain with the orig-

inator until stealthily committed to

market, minimising the spread o in-

ormation to the market.

To level the playing field with the

sell-side and larger irms, platorms

that provide automated order state

management independent o theexecution algorithms allow or zero

latency trading where large num-

bers o order slices can be let rest-

ing at execution venues and thereby

be close to the ront o the queue as

the price moves.

The continued spread o algorithmic

trading throughout the buy side has

hitherto been more evolution than

revolution and led by buy-side irms

seeking execution price improvement,

thereby raising both returns and proitwhile generating a marketable edge.

As take-up o algorithmic execution

strategies becomes commonplace

amongst the buy side, late adopter e-

ects are likely to become key drivers.

Asset allocators already conduct ex-

tensive due diligence, which includes

trade execution methodology and

regulators are demanding ever-great-

er transparency.

As disclosure increases, most buy-

side irms will need to demonstratecompetitive trade execution mech-

anisms, which will increasingly be

synonymous algorithmic execution

strategies.




11/12

A-Ta Groups experienced analysts and journalists cut through the noise todeliver our acclaimed commentary and analysis, giving you the

NeWS & INDUSTRY INTeLLIGeNCe THAT mATTeR TO YOU.

Take your pick from these topical industry issues and view our news, commentary, archives,and in-depth research:

markt Data elctronic Trading Risk & Rgulation IT

Low Latncy Rfrnc Data

Find out how A-Ta Group can deliver for you:

Get customised daily alerts delivered direct to your inbox to make sure you dont miss athing sign up for free alerts on our website at www.a-tagroup.co/ail-alrts

Register for free access to extensive news services, archives and free research reports

Want to try our premium news and analysis services? Register for a free one month trial

at www.a-tagroup.co/fr-trial

More than just the news...

delivers

BUSINESS INTELLIGENCEFOR FINANCIAL MARKETS IT

A-Ta Groups online news and research services - an invaluable research

tool for busy executives seeking insight to help them excel in their roles.

www.a-teamgroup.com


12/12

algorithmic trading - attracting the buy side

Documents