About Us  |  Search  | FAQ  | Contact Us
Capital Markets
Risk Management
Human Resources
Smarts Cards
Optimise CRM
Data Warehousing
Disaster Recovery
Swift Messaging
BPM & Workflow
Capital Markets
Global Custody

Page last updated
February 16, 2003




Capital Markets:
Implementing an appropriate transactional architecture


The capital markets sector is an industry whose restructuring will come through a delicate combination of new concepts, new technologies, while respecting "ancestral" rules and habits. This industry has seen historical huge growth rates in the past 30 years, with regular conquests of new territories. Its overall profitability increased more by the rise of the revenues than by the reduction of its costs and the streamlining of its processes. At the forefront of the challenges this industry now faces, lies the necessity to re-arrange the transaction processes beyond what the first generation online dealing systems have offered until today.


I.1. The key reasons of capital markets expansion
The key components of the capital markets expansion have been, and still are :

n An always-larger community of participants

Correlated to the developments of the Telco capabilities around the globe, the number of communities of participants in the capital market industry has increased year after year: from the large international institutions down to the small local company and the inclusion of private individuals, access to financial products has seen a tremendous "democratisation".

In parallel, the adoption by new countries of the techniques and regulations of the world's capital markets has generated a strong geographical expansion.

n l An harmonisation and liberalisation, across the main OECD countries, of the rules of circulation of the capital.

From the Bretton Woods Agreement to the launch of the Euro, many complicated hurdles have been passed, freeing a bit more each time the movements of capital around the planet, and increasing its speed of circulation.

n A stronger sophistication of the financial products

Thanks to the progress of the computing power and capabilities, the capital market industry has regularly been able to invent or adapt new financial instruments to cope with always more demanding requirements from the clients. By improving the measurement of all types of market risks, new frontiers have been crossed, and new financial risks have become "hedgeable".

As a result overall, year after year, crisis after crisis, bubble after bubble, the revenues increased strongly. Yet, due to this fast rate of expansion, the consolidation of the operating processes, the organisation of the "production lines" have remained somehow erratic; each generation of financial products has introduced in a bank its own array of systems and processes. This means non-optimised production costs at the end.


1. 2. The first answer to cost cutting: merging with another institution
Regularly, with a clear acceleration in the late 90's, banks found a first good answer to the need of controlling and reducing their operating costs, while preserving their revenues growth; to merge or to acquire another financial institution.

Behind the merging logic lie the assumptions that economies of scale can be achieved, principally by streamlining the "factory" processes and production lines. Among which:

- decreases the use of interbank assets

- decreases the number of branches pricing and booking the same instruments

- lessens the number of back offices locations & systems

- increases the local sales presence for the customers

- decreases the number of front and back office systems doing the same task, and consequently reducing the population of the support teams needed to make these heterogeneous systems work together

- increases the liquidity of all instruments within its own group; that translates into better prices for the clients and more deal flows

- streamlines the sales and distribution process

- knows better the clients behaviours and needs

- increases the transparency of information inside its own group, particularly the data related to the clients' activity

- decreases the brokerage and exchange fees

The reality and practical cases have shown that achieving these goals (streamlining the operating processes from banks which hadn't done so before merging) can become mission impossible. Very few mergers have enabled to properly leverage the existing business while implementing true economies of scale


Since a bank has to rethink its role of intermediary around a few core principles; serving various clients needs while controlling its risks and capital allocation -, it must now concentrate its efforts in the transaction processes, and the integration of the new transaction processes into various existing and heterogeneous legacy architectures.

II.1. New Technologies & Internet promises
Internet and the new technologies promised to bring the answer. However, by excess of precipitation, and probably by naive belief that things could be changed rapidly, and effortlessly, most new Internet projects have failed to deliver the awaited promises. On the operating front, they often resulted in adding another layer to the existing systems in place, instead of eliminating one of several heavy processes. It is clear that some good results have been obtained in areas external to the bank; we see that now most of bank-to-bank and some of the bank-to-client executions are done electronically, whether it is for FX, Bonds, or equities, as long as the credit issues are taken care of in a simplified manner. Still, the benefits are limited to the pure execution processes, and lots of efforts are to be done in the clearing, STP, integration of the deals within the banks' internal processes, homogenisation of the global transaction processes, irrespectively of the nature of the market considered. This is where the real challenge lies in.

II.2. Elements to think of before re-designing a new transaction architecture

What needs to be re-considered to "milk" the benefits offered by the new technologies.

n Start from the abstraction of the transaction processes; think transversally

Transaction processes must be reconsidered generically before looking at the market's products segmentation. A bank is an intermediary: various trading communities exchange various instruments with or through the bank, respecting precise sets of constraints and rules. This principle is valid for all markets. It must be the cornerstone of a new transversal approach.

n Think transactions, interests & quotes in terms of orders

A transaction can always be viewed as the matching of two opposite orders. Whether we consider specific or benchmarked instruments, all transaction protocols (RFQs (Requests For Quotes), click-and-trade, Branch Dealing (i.e. dealing in the name of someone else), back-to-back trades, etc), can be viewed through the prism of order matching. A transaction architecture based on order-matching principles can furthermore help build an exhaustive database on clients and traders' behaviours and so implement CRM tools.

n Split cleanly the steps composing a transaction

Create appropriate APIs (Application Programming Interfaces) with existing legacy systems, in order to interact, at each step of the transaction process, and in an appropriate manner, with the pricing engines, the credit engines, the margining engines, the instruments reference databases, the users databases, the distribution rules etc

n Separate the matching process from the rest

When orders have to be sent to external ECNs, Marketplaces, Exchanges, the pure matching is done outside the bank. However, the price discovery, the credit checking, the input of the order must remain the same whether the matching is done in-house with one of the bank's traders or even clients, or if it is done outside the bank. In this case, the transaction server must act as a router. The router becomes intelligent when it is capable of arbitraging between an internal and an external matching. For example, banks should consider all direct and indirect costs (price, risk, capital cost, traders and sales time, brokerage & exchange fees) before deciding to match an order internally in its own books or externally. This is valid whether we deal with equities, FX, deposits, bonds, cash or derivatives, commodities, etc...

n Accept various; possibly all; types of front-ends. Respect the ecology of existing GUIs

One of the drawbacks of 1st generation online dealing systems is that they focused primarily on the front-end. In reality, the stickiness of existing front-ends is strong, and imposing a new front-end to a trader, a sales person, or a client, is a tough ambition. Moreover, a front-end is different for a client, for a salesperson, and for a trader. In some case, the front-end is replaced by a host-server, or a pricing robot, or another ECN.

This constraint translates into the obligation to access to the server through clean and documented APIs so that all cases are treated.

n Support various known and unknown messaging protocols

One must not be locked into today's standards messaging formats (FIXs, FpML, Twist, GSTPA, ), as well as on the communication protocols (JMS, TIB, MQseries, X25, Http, IIOP), and be prepared to deal with various unknown protocols. For orders inputs, as well as for STP, the formats of messages have to be managed by an independent module, and mustn't interfere with the core transaction process.

n Propose clean and fully documented APIs in most standard technologies

By proposing external APIs in Java, Microsoft .Net, and COM, the transaction server and its modules will technically guarantee an easy access from any type of front-end, and an interaction with almost all legacy systems, without disturbance.

In trading rooms, and at the clients level, the tools manipulated are most of the time based on Microsoft technologies, very often out of Microsoft Excel spreadsheets. In this case, offering a .Net GUI that is fully fungible in the existing environment is absolutely key. If it has to be rebuilt from scratch, then a .Net external API is needed.

n Consider outsourcing some of the elements, without losing control on the global design of the architecture

It takes a long time to build and develop its own transaction server, respecting criteria of flexibility, power, scalability, security, etc. It is however important to control the design of the global transaction processes within a bank. Outsourcing completely this task creates the risk to be entangled in one single vendor's goodwill. No perfect offer exists anyway in the market. On the other hand, as long as the proposed external components have a full capacity to be integrated natively through their own APIs into any global architecture, it makes sense to assemble such key components together, and to concentrate its efforts on the global architecture and the design of some specific elements (the GUIs for example, the build up of global reference databases). With such conditions, time and money is saved.

Today's new technologies enable to reach the goals presented above. But not so much by optimising the sales-to-client relationships, also from the "inside", by drastically reviewing the process of a transaction.


The smartTrade suite of trading application are all based on an architecture and a development that respects the above-mentioned principles. The chart shows an overview of the functional architecture, where we see a precise split between the role of each component. More information to be found on www.smart-trade.net.


One important direction leading to costs reductions in the capital markets sector now encompasses the implementation of solid principles and processes in the way transactions are conducted. This transformation is achievable today thanks to the possibilities offered by the "new technologies". More than other industries, the capital market sector has piled-up many operating systems and processes over the last 30 years. The global and risk free homogenisation into a "one unique system in one unique centre" becomes almost impossible today.

By respecting the existing environment, keeping alive the current trading and sales methods and habits, but by refocusing intelligently its efforts on the transaction process, a bank can rapidly open its door to a new virtuous circle bringing a better service to its clients, a better allocation and control of its capital among its activities, and a sharp reduction of its process and operating costs. This is what the new economy must deliver to be really new.

Harry Gozlan
C.E.O & Founder
Smart Trade Technologies


Six Pillars
Funds of Funds
Capital Markets




Home  |  About Us  |  Search  | FAQ  | Contact Us