Data Management Special Report trading technologies for financial-market professionals

Sponsored by:

waterstechnology.com June 2014 FREE attendance for qualified end users

June 25 Grand Hyatt Singapore

The most comprehensive trading technology conference for end-users

Speakers include:

KEYNOTE ADDRESS: AFTERNOON ADDRESS: Neal Cross, Endre Markos, Chief Innovation Officer, Director / Asia Pacific Head of DBS Execution to Custody, Citi (Hong Kong)

Angelo Roxas, Ian Kay, COO APAC Technology, Regional Head of Global BARCLAYS CAPITAL Execution Services, Asia Pacific, ABN AMRO CLEARING

Michel Dinh, Stuart Gurr, Director, Regional Head APAC Head of Global Trading Prime Services IT, Technology & Risk Change CREDIT SUISSE Delivery, Asia Pacific, ROYAL BANK OF SCOTLAND

For sponsorship opportunities please contact: Joseph Donato, Business Development Executive, WatersTechnology [email protected] or call (+1) 646-736-1895

Panel sponsors: Hosted by:

waterstechnology.com/aptas Editor-in-Chief Victor Anderson [email protected] tel: +44 (0) 20 7316 9090 US Editor Anthony Malakian [email protected] Deputy Editor, Sell Side James Rundle [email protected] US Staff Writer Jake Thomases [email protected] US Staff Writer Tim Bourgaize Murray [email protected] European Staff Writer Marina Daras [email protected] Head of Editorial Operations Elina Patler [email protected] Contributors Max Bowie, Editor, Inside Market Data Michael Shashoua, Editor, Inside Reference Data Global Commercial Director Jo Webb [email protected] tel: +44 (0) 20 7316 9474 US Commercial Manager Bene Archbold A Product of Participation [email protected] US Business Development Executive Joseph Donato [email protected] Business Development Executive Shayna Hersh t’s diffi cult to fully articulate and appreciate the critical role data plays in the [email protected] fi nancial services industry. It’s been written before in the pages of Waters, Account Manager Tom Riley [email protected] Ibut it’s worth noting again that data is analogous to blood fl owing through Senior Marketing Manager Claire Light the human body: Remove it and the body dies. Additional illustrations of data’s [email protected] Marketing Manager Natasha Gordon-Douglas pivotal role in the capital markets can be gleaned from Waters’ sibling publica- [email protected] Design Lisa Ling tions, Inside Market Data and Inside Reference Data, which, over the years, have [email protected] helped inform, shape, and educate their respective communities. Multiple Subscriptions/Site Licences Continuing the data-as-blood analogy, like the body and its myriad complexi- Monazer Rashid tel: +44 (0)20 7316 9606 (UK) ties that rely on blood for nourishment and the most fundamental requirement Subscription Sales to support life—oxygen—every conceivable business process across the buy Hussein Shirwa tel: +44 (0)20 7004 7477 (UK) Dennis Bond tel: +1 646 755 7374 (US) side and the sell side is contingent on data to some degree. And, like blood, [email protected] the “purer” the data, the better. Technology also plays an important role when it Group Publishing Director Lee Hartt comes to fi rms’ data management practices, although its part is more that of a Chief Executive Tim Weller Managing Director John Barnes supporting actor than the starring role. But make no mistake, without technol- Incisive Media Head Offi ce ogy, automation of every aspect of the capital markets would be impossible. Haymarket House Data management in our industry is a multi-faceted phenomenon, illustrated 28–29 Haymarket SW1Y 4RX by the variety of responses from our seven sources who took part in the virtual tel: +44 (0)20 7316 9000 fax: +44 (0)20 7930 2238 roundtable on page fi ve. One issue that did yield something of a consensus, however, is the notion that effi cient and effective data management disciplines Incisive Media US 55 Broad Street, 22nd Floor are not driven by technology projects. In this respect, data management is more New York, NY 10004 tel: +1 646 736 1888 of a process, a continuum for which no discernible start or end points exist, and Incisive Media Asia which, in an ideal scenario, combines fi rm-wide objectives, clear data govern- 14th Floor (Unit 1401-3) ance policies, a chief data offi cer responsible for driving the entire initiative, and Devon House, Taikoo Place 979 King’s Road Quarry Bay of course, the technology underpinning the fi rm-wide program. Hong Kong tel: +852 3411 4888 Data management schemes, especially those multi-year, all-encompassing data warehouse projects, have come under scrutiny by executive committees Incisive Media Customer Services tel (UK): +44 0870 240 8859 (UK) over recent years. And rightly so: Far too often, those initiatives, conceived at tel: +44 (0)1858 438421 (International) the outset to be the panacea for a fi rm’s data-related maladies, failed to deliver anything resembling a half-decent return on the substantial time and fi nancial investments for a variety of reasons ranging from constantly changing sponsors To receive Waters magazine and non-accountability, to scope creep, poor coordination and loosely defi ned every month you must subscribe to Buy-Side Technology online, objectives. Now, however, such initiatives tend to be more focused and better Sell-Side Technology online or one disciplined, demanding buy-in and genuine participation from right across the of our multi-brand subscription options. For more information business. Q and subscription details, visit waterstechnology.com/subscribe Victor Anderson Editor-in-Chief Waters (ISSN 1068-5863) is published monthly (12 times a year) by Incisive Media. Printed in the UK by Wyndeham Grange, Southwick, West Sussex.

©Incisive Media Investments Limited, 2014. All rights reserved. No part of this publication may be reproduced, Sell Side stored in or introduced into any retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the copyright owners. Special Report Data Management

Stress Tests, Data Aggregation Get Bigger Parts in Risk Management

Risk data aggregation efforts are pulling in more sources of data and more functions and tasks, such as stress testing. By Michael Shashoua

Stress testing will play an increased role in or industry classifi cation level is driven by the industry’s eff orts to meet the principles diff erent regional classifi cations as well as the set out in the Basel Committee on Banking classifi cations used in certain industries.” Supervision’s BCBS 239 recommendations, Kanungo detailed numerous codes and aggregating the risk data necessary to and ratings that comprise an intricate follow those principles will be a three-layered web of data that is subject to aggrega- challenge that will have to connect multiple tion, including ones such as the North sources and types of data, according to American Industry Classifi cation System, industry experts from RBC Capital Markets, which denotes the insurance industry, the UBS and S&P Capital IQ, speaking during a Global Industry Classifi cation Standard for recently held Inside Reference Data webcast. sector allocations, credit rating indicators Sanjay Sharma “The emphasis is now more on stress test- particularly relevant for credit default RBC Capital Markets ing, which is more deterministic,” said Sanjay swaps, as well as International Securities Sharma, chief risk offi cer for global arbitrage Identifi cation Number (ISIN) codes, entities coming from diff erent vendors. The and trading at RBC Capital Markets. “When Bloomberg IDs and Sedols. LEI simplifi ed the risk data aggregation doing deterministic analysis, such as for landscape, Feddo said, and now UBS had US Comprehensive Capital Analysis “a very clear, single, golden-source view of Review (CCAR) stress testing, the qual- the data we bring into our risk aggregation ity of the data you are using is paramount. process.” “It’s a very challenging process If your data is not trackable, uniform or UBS is still working on aligning transac- consistent, you will see errors or results to combine all this data into one tion data through common feeds, according that are unspecifi ed or plain wrong.” container. That’s even when there to Feddo. That starts with charting market Institution-wide stress tests have to are multiple providers feeding into risk in its investment bank. “We’re trying collect data that is likely to be dispersed it.” Sanjay Sharma, RBC Capital to agree to principles across the control among diff erent silos, according to functions for risk and fi nance, bringing these Sharma. “It’s a very challenging process to Markets control functions together and aligning combine all this data into one container,” them, then aligning the data sources across he said. “That’s even when there are the bank to help pull everything together,” multiple providers feeding into it.” “All the underpinnings of data frame- he said. works are being established to leverage Client-onboarding processes, as well as Three Layers within the organization and to answer simple the creation, maintenance and management The three layers involved in cross- questions, such as: What’s my exposure to of partner data, have changed at UBS in referencing data to aggregate it for risk ‘X,’ which could be a certain entity that response to regulations and the overall management and to fulfi ll the demands of made the headlines for good or bad reasons,” principles refl ected in BCBS 239, said stress tests are the entity layer, the instru- he said. “At the push of a button, you would Feddo. “We found we must have linkages ment layer, and the sector or industry layer, be able to aggregate all the entities affi liated between our client data, account data and explained Rick Kanungo, senior director of with that, mapping an entity back to an legal trading agreement data very clearly enterprise solutions at S&P Capital IQ. In organization’s holdings and counterparties, in place for us to understand that the client dealing with entities, legal entity identifi ers and be able to say, ‘This is exactly what my has a regulation applicable, based on the (LEIs) need to be mapped to vendor IDs, he exposure is.’” products in the combination of entities noted. Without an LEI, vendor IDs may end they are trading,” he said. “It’s become an up being adopted as the primary identifi er, Connecting and Coordinating open process with much more know-your- he said. Refl ecting the silos that Kanungo described, customer due diligence.” Q “At the instrument level, many entities— Simon Feddo, director and head of change but not all—issue securities, and that could for legal entity, client and account at UBS, Michael Shashoua is editor of ,QVLGH5HIHUHQFH be across asset classes,” he said. “The sector said his fi rm had a “splintered view” of 'DWD, part of the WatersTechnology stable.

2 June 2014 waterstechnology.com News

SunGard Builds Risk Reporting Service for Hedge Funds

Financial industry operations software and risk model database to high- can characterize the services provider SunGard has launched a light exceptions or queries risk on derivatives new risk reporting service for hedge funds, quickly before proceeding to products with much providing independently validated data and the risk engine for overnight lower operational analytics. The Hedge360 Risk Reporting report generation. overhead than if this Service provides detailed risk attribution “In addition to an were done in-house. and stress testing, as required by regula- effi cient, automated service, Most importantly, tions such as the Alternative Investment we off er data enrichment, our Hedge360 Risk Fund Managers Directive in Europe. Risk based on engineered market Reporting Service reports can also be customized for com- data such as credit curves and database contains a modity strategies, long–short equity, and volatility surfaces, which in Laurence Wormald wealth of historical specialized credit funds. many cases are of better qual- shock scenarios for Users of the service upload position ity than those easily available to a hedge stress testing and proxies for illiquid assets information via a secure SunGard network. fund,” says Laurence Wormald, London- which have been created for institutional The data is validated by SunGard’s man- based COO and head of research of clients and are very hard to source,” aged service team against SunGard APT’s SunGard’s APT business. “In this way, we Wormald adds. Q

Nasdaq OMX to Launch Business Intelligence Tool

Nasdaq OMX has announced the launch discovery solutions, allows users to tap market technology at Nasdaq OMX. of MiQ, a business intelligence solution into their data and derive valuable insight “Enabling organizations to be more that provides actionable data for exchanges, from it, according to Nasdaq OMX. informed through better data intel- clearinghouses, central security deposi- “The launch of MiQ will signifi - ligence taken across the entire trade tories (CSDs), and other fi nancial market cantly improve how market operators life cycle will help drive liquidity and operators across the globe. eff ectively and strategically manage market share, grow revenues, enhance The product, developed in partnership and analyze their data,” says Lars client services, and stay ahead of the with Datawatch, a provider of visual data Ottersgård, senior vice president of competition.” Q

Markit EDM Signs CLAMC as First Chinese Client

China Life Asset Management Company a selection process that included interna- Markit EDM has also introduced double- (CLAMC), a -based provider of tional and local EDM vendors. byteware software. “That means we can life insurance and annuity products with According to Stuart Plane, -based display the data and software in any language approximately $300 billion under man- managing director at Markit, the vendor the client requires,” says Plane. agement, has become Markit Enterprise was chosen because, in addition to satisfying According to Plane, Markit EDM began Data Management’s (Markit EDM’s) fi rst the typical requirements of a large fi nancial talking to CLAMC a couple of years ago, Chinese client, a move that is expected institution, Markit EDM has local, Mandarin- but only embarked on “serious negotiations” to lead to more business in the country speaking support and implementation teams with the fi rm over the past year. He describes for the London-based vendor, which has in Hong Kong and has introduced a number CLAMC as a “marquee client,” which he created adaptors for domestic Chinese of enhancements to its software to suit the believes will lead to other client wins in data feeds and introduced multilingual Chinese market. China for Markit EDM. software. “China has some domestic vendors for “What we have found when pushing CLAMC, the largest institutional data and software that are unique,” says into any new market is, once you have one investor in China’s capital markets, chose Plane. “The data feeds come in the Chinese client and have proved you can get it up and to deploy Markit EDM’s software locally language, including the documentation. So running very quickly, and we understand to create a central hub for all of its securi- we built adaptors to the domestic vendors, as the local requirements, then you will have ties, pricing, positions and transaction part of the proof of concept.” He adds these a number of additional clients signing up data. Markit EDM was chosen following adaptors will now be available to other clients. afterwards,” Plane says. Q

waterstechnology.com June 2014 3 Special Report Data Management

Mizuho Taps GoldenSource for Data Quality, Compliance

Mizuho International has chosen “As a broker-dealer, Mizuho recognizes European Market Infrastructure Regulation, GoldenSource’s enterprise data manage- the importance of establishing links between Basel III, and the Dodd–Frank Act. ment (EDM) platform to help its investment securities and issuers. Customer due diligence “GoldenSource will facilitate compli- banking business comply with new and onboarding, and generally the de- ance with these regulations by signifi cantly regulations by improving data quality and duplication and centralization of securities improving data quality and transparency,” transparency across its London operations. and counterparty data, have been key drivers says Vanlint. “With an ever-increasing GoldenSource’s platform will replace for this project,” explains Neill Vanlint, emphasis on the importance of data transpar- in-house systems at Mizuho International GoldenSource’s London-based managing ency, data governance-enabling technologies and is part of a strategic approach to director for EMEA. like GoldenSource that prove your data improve data quality and operational Vanlint says GoldenSource’s 360 EDM lineage are now essential, both for internal effi ciency, while facilitating regulatory platform will help Mizuho International auditing and for external regulatory compli- compliance. comply with requirements such as the ance.” Q NYSE Plans Corporate Actions Data Changes

NYSE Group is planning service and NYSE’s Arca equity platform, which include dividend declaration fre- pricing changes for users of its corporate covers more than 1,300 exchange-traded quency and comprehensive dividend actions data, to take eff ect on September products. Ex-dividend information for breakdowns. The exchange also will 1 of this year, according to offi cials at the these Arca exchange-traded products will require approval requests from users exchange operator. be added to its ex-date reports. distributing its corporate actions data Ticker notices reports will include In addition, daily corporate actions to third parties that are re-distributing full corporate actions information from and ex-date reports will now also that data. Q

BBVA Selects Markit EDM for Interactive Data Hires Spanish, Mexican Data Hubs Lauterbach

Banco Bilbao Vizcaya also intended to enable business Market data vendor Interactive Argentaria (BBVA) Asset users to access and use the data Data has hired Joachim Management, the asset man- as part of their daily activities. Lauterbach as president, global agement arm of the Bilbao, “The fi rm was looking for a head and chairman of the board Spain-headquartered Banco truly enterprise-wide solution of Interactive Data Managed Bilbao Vizcaya Argentaria that could trace back and source Solutions (IDMS). Lauterbach, banking group, is set to roll where the data came from and who replaces former Interactive out the Markit Enterprise Data who touched it,” says Daniel Data Managed Solutions Management (Markit EDM) Simpson, Markit’s London- managing director Matthias Joachim Lauterbach platform to create hubs for the based managing director and Paul, joins Interactive Data management of its securities, head of enterprise software. from consulting fi rm CSC, where he served as managing portfolio, fund, issuer and posi- “After several workshops and director and senior managing partner, initially responsible tion data in Spain and Mexico. a successful proof-of-concept, for its fi nancial services consulting business in Central and BBVA opted for a locally BBVA was confi dent that Markit Eastern Europe, and then of its consulting business across installed version of the platform, could successfully transition all industries in the region. which will be implemented in and manage its pricing, securi- Prior to that, Lauterbach was chief customer offi cer several phases during 2014. ties, portfolio, fund, issuer and at German market data provider VWD, and spent the BBVA will use Markit position data.” previous 15 years at Reuters, Capco, and Thomson EDM as part of an initiative to Simpson says BBVA’s use of Financial. streamline its operations, reduce the platform may be extended to Based in Frankfurt, Lauterbach reports to Interactive manual processes and mitigate other Spanish-speaking coun- Data COO Jay Nadler. Q operational risk. The platform is tries in the future. Q

4 June 2014 waterstechnology.com Roundtable A Process, Not A Project

Data management challenges proliferate in an industry where success, irrespective of how one might defi ne it, is contingent on clean, reliable and timely data. And, as this virtual roundtable illustrates, effective and effi cient data management practices are more a process than a project.

waterstechnology.com June 2014 5 Special Report Data Management

client reporting. Fundamentally, as organizations have become more complex, they’ve ended up with more and more back-end systems (databases), and therefore the report at the front end, whoever it is destined for—internally, or for clients and regulators—is depend- ent on the processes that bring that data in from new data sources alongside all the existing sources. As each link in that process is implemented, the process becomes ever more complicated. Brian Sentance CEO Brian Sentance, CEO, Xenomorph: Risk management is one Xenomorph of our key focus areas for data management, and we are seeing that Tel: +1 212 401 7894 Web: www.xenomorph.com the current bout of regulation is aff ecting risk management and the data it uses much more than in any other area within institutions. Regulators have, over the years, always pressed for institutions to improve data quality and control in the data feeding risk models, but Q What business processes across the buy side and sell both direct data management regulation and indirect consequences side are currently most reliant on clean, consistent, and of regulation are intertwining to drive big changes in how data is up-to-date data? processed and managed in this key area. John Bottega, senior advisor, EDM Council (previously chief data offi cer, Bank of America): When responding to a regula- Paul McInnis, data management product manager, Eagle tory report, for example, you really need to have precise data for Investment Systems: Due to a host of regulatory changes measuring liquidity or capital or risk. On the marketing side—and over the last decade, coupled with an ever-increasing amount this is the big data side of of information investment things—there’s an argument managers need to process and in the industry that says for consolidate, there has been that type of analysis, maybe “Data integration is a huge challenge for larger increasing pressure on invest- the data doesn’t have to be institutions and I think that this integration has ment fi rms to be able to provide pristine because a lot of the transparency throughout new technologies and ways of probably received less technological focus than their investment portfolio. trolling through big data can other areas such as database and data management In order to maintain a com- be very informative, without technology. For example, many of the current ‘big petitive advantage, improve the data being perfect. It’s a data’ technologies are designed more for ‘green the decision-making process, little bit risky, because the field’ application scenarios than for trying to unravel reduce trade errors and meet value of big data evaluation is regulatory standards, it is vital always going to be contingent the spaghetti that is the legacy data architecture to have clean, consistent and on the quality of that data. As of many larger financial institutions. That said, you up-to-date reference data across fi rms start to explore those can’t ignore the people aspects and putting together the enterprise. This focus on capabilities and what data can data governance that delivers in practice is still a big accurate and reliable data has tell you, you probably have led to the development of the a little room to extract value challenge.” Brian Sentance, Xenomorph concept of an investment book from the data without it being of record (IBOR), or as we pristine, but when it comes to operational functions—fi nancial, risk, frequently refer to it, “a single version of the truth.” compliance—you need precise and accurate data. There isn’t room However, the reality is that many investment managers today for inconsistencies and inaccuracies; it’s got to be clean. are working within legacy systems that were implemented in silos 10 to 15 years ago and are no longer fi t-for-purpose, given today’s Ralph Baxter, CEO, ClusterSeven: One can answer that ques- demands for greater transparency and risk management. As a result, tion by saying that every business process requires clean, up-to-date the need for a single version of the truth has emerged as a top data, but I think that the one area that has proven most challenging business priority and investment managers across the board are col- in recent times is wherever you need to aggregate data from multiple lectively stepping back to examine their data management practices back-end sources, and that is particularly true for the ever-increasing to ensure they are utilizing the appropriate technologies and have waves of internal and external reporting, including regulatory and the right governance in place to be competitive.

6 June 2014 waterstechnology.com Roundtable

John Robertshaw, principal consultant, Investit: Investit’s experience is predominantly on the buy side, where we typically fi nd that all pre- and post-investment decision processes are reliant on clean, consistent and up-to-date data. This is because the core buy-side process is the creation of an investment valuation, which shows the current holdings and values of a client/product portfolio. This means that all upstream data processes must be performed in a timely and accurate manner: from the processing and transfer of Ralph Baxter all client monies to corporate action updates, trade processing, and CEO margin call management. It follows that all downstream data processes ClusterSeven will then also be able to be correct. This means that, ultimately, data Tel: +44 (0) 20 7148 6270 Web: www.clusterseven.com accuracy is a process, not a project—a process in which all parts of the organization are necessarily engaged, and accountable for the part they play and produce. Data management is necessarily a decentralized transparency oriented; and the Federal Reserve’s SR Letter 11-7 on activity in most buy-side organizations—it is the management and Supervisory Guidance on Model Risk Management published in 2011. orchestration of these activities that is the key challenge still facing All of them focus on data, data quality, and data processes as being most practitioners today. the key element to satisfactory supervisory or quality-type activity. Q How has the I think the regulators have played introduction of regula- a huge part in this, but the other tions in the US and “Solutions are always a combination of technology aspect that is coming into play is the pan-European and human components. The biggest challenges that in the wake of the 2008 global marketplaces spurred are about responding to constant change. This fi nancial crisis, margins are much capital market fi rms means establishing processes that recognize new more under pressure. Whereas before, to manage their business needs—often satisfied in the short term it was perhaps a lot easier to ignore data better? Is this some of the irregularities in your data data management by spreadsheets—and putting in place migration because the margins in the business drive an internal one processes, which can prioritize these new requirements were big enough to overcome them, (company-driven), for more robust solutions. But this is not a one-off now, operational effi ciency and the or an external one requirement. Constant business change creates a contribution that data errors can (regulatory-driven)? make to your margins is much greater Enrique Smith, CIO, continuous conveyor belt of new demands. The most on a percentage basis. So there is a CIFC: If regulatory mature organizations recognize this.” Ralph Baxter, strong internal drive to focus on data requirements are the ClusterSeven quality as well. If you compare the main driver for your data philosophy around data between UK management, you may have other issues. Those trying to stay ahead grocer Tesco, for example, with its Clubcard business, and a large of their peers will likely encounter the need for better data and better fi nancial services institution, Tesco had to sort out its data manage- data management before it is required by an agency. ment issues many years ago because its margins are so much smaller.

Baxter: The straight answer is absolutely yes. If you look through Mike Atkin, managing director, EDM Council: Are regulations the last three or four years of regulation on both sides of the driving data management? The answer without question is yes, abso- Atlantic, there are specifi c examples of new regulations that focus lutely, no doubt about it. It is the reality of linked-risk analysis, where on data quality, data processes, and manual data manipulation. you have to connect instruments and their obligations and entities Those include: the Public Company Accounting Oversight Board and rules, and you have to connect indexes and underlyings, and you (PCAOB) with its staff alert on system-derived data and reports in have to link relationships across your instruments, entities, obligations, October 2013; the Basel Committee on Banking Supervision’s Principles and holdings. That requirement is the baseline for all of the systemic for eff ective risk data aggregation and risk reporting; the 2013 Committee risk analysis under way; all the stress testing that’s under way; all the of Sponsoring Organizations of the Treadway Commission transparency objectives under way. All of these things are built off (COSO), where they produced revised guidelines last year around of a baseline of data, and if you don’t get it right, the regulators can’t manual data manipulation; Solvency II, which is data-quality and perform their function.

waterstechnology.com June 2014 7 Special Report Data Management

It’s the “You ain’t got no choice” business case, and what it really does is make data management mandatory—mandatory to the top of the house, with the full authority of the process behind it. So it has changed the equation within fi nancial institutions, because now they have an obligation that they have to meet. You can look specifi cally at the evolution of the risk data aggregation (RDA) principles as a prime example.

McInnis: The global fi nancial crisis of 2008 unfortunately high- Paul McInnis lighted the fact that capital market fi rms lacked the proper ability Data Management Product Manager to provide an accurate and reliable picture of their risk exposure. Eagle Investment Systems Tel: +1 781 943 2200 This naturally spurred the introduction of regulations in the US and Web: www.eagleinvsys.com Europe focused specifi cally on transparency and proper risk manage- ment. These regulations have increased the amount of transactional and counterparty data investment fi rms are required to manage and Indeed, regulations have proved to be a distraction, and have report on a regular basis. It therefore follows that a large part of the possibly led to a misallocation of resources away from investment in focus on better data management right now is perhaps unsurprisingly creating accurate client-facing, performance, risk and management driven by the introduction of and response to these regulations. information data toward holistic institutional data which has been of The silver lining is that in responding to these regulatory changes little direct use or bearing. and thereby enhancing data governance policies and technologies, The data management drive has generally been an internally investment fi rms have also been able to reduce their operational risk. driven one—companies need good data upon which they can The benefi ts of clean, accurate make and execute investment data extend far beyond decisions in a timely manner complying with regulations, and without error. and investment managers have “Data governance is a business decision. Before The increasing complexity of been able to take advantage embarking on a costly data management overhaul, undertaking portfolio valuations of those benefi ts to maintain has been due to: a competitive edge and create firms need to understand and evaluate their • Complexity of instruments operational effi ciencies. current environment, data sources and operational (derivatives) workflows in order to identify the expected benefits • Volumes and market Robertshaw: To the extent of their future state environment. One of the most conventions that the introduction of commonly overlooked aspects of a data management • Unreliable/partial pricing/ regulations in the US and corporate action sources pan-European market places project is having the right people involved. Senior • Inherent errors in the data has provided senior sponsoring management and IT must work together and clearly sources data management champions articulate the challenges and benefits of such a • Inability of core platforms to with the external leverage to large-scale project upfront.” Paul McInnis, cope/scale enable them to build a case for Furthermore, the desire for Eagle Investment Systems a more coherent and unifi ed clients to see their data accu- data management approach rately in real time (because of internally, they have been useful. internet) has driven the need for these data processes to work well. However, the data gathered at behest of regulators has generally In this context, regulations have been an unnecessary distraction only been designed to be of benefi t to the regulator, and not used by that has had no direct bearing on improving the actual perform- the fund manager and/or their clients—the submission of European ance/risk profi les of funds under management, nor on improving Market Infrastructure Regulation (EMIR) derivatives-related data client focus and delivery. being a recent case in point. Current regulatory data requests are focused on the fi rm, not the client. From this perspective we would Sentance: On the sell side, for example, you have data management say the answer is no—regulations have increased the demand for practices directly being mandated with documents such as BCBS more data more often but this is not necessarily useful, pertinent, or 239 but, as I suggested earlier, other indirect eff ects of regulation relevant data. are having big eff ects on data management. For example, capital

8 June 2014 waterstechnology.com Roundtable

calculations such as credit valuation adjustment (CVA) require near- and potentially leverage the data sets being created for use in real-time integration and aggregation of counterparty exposures. The risk management. Certainly, enterprise-wide data management motivation to optimize capital costs will probably do more to move projects have proven diffi cult and costly historically, which again to real-time data management and break down data silos at large biases me to starting small in one department and spreading sell-side organizations than any direct data management mandates. outwards incrementally. With this in mind, regulation, capital Other regulatory initiatives, such as the legal entity identifi er (LEI), costs, and operational cost control now provide drivers and were intended to assist in counterparty exposure aggregation but, shared motivations that help to overcome some of the political without parent/ultimate parent data being available, other methods diffi culties that have hindered larger data management projects will need to be used. As ever, the eff ects of regulation are sometimes in the past. as intended, sometimes more than intended, and sometimes not what was intended at all. McInnis: Whether a single, enterprise-wide data warehouse model or discrete tools are best to achieve a sound data manage- Bottega: Clearly, the regulatory require- ment strategy depends on the source, owner and consumer ment has spurred markets to respond. It’s of the data. If data is a commonly used asset, such as security forced attention, but the good thing is reference data, then a centralized approach is in line with a data that in any fi rm, “to satisfy a regulatory governance framework. If it is a specialty without the need for requirement” is usually the justifi ed use across the enterprise—i.e., quantitative research—then a business case—this is not something that’s discrete application makes sense. up for debate; it’s something that fi rms However, when dealing with legacy systems especially, it can have to do. be tempting to approach data governance by simply focusing on The good news, though, is that connecting the pipes to get immediate benefi ts, especially when what I have seen is that are now budgets are tight and resources are spread thin. It is vital for looking at the requirements for data investment managers to remember that sound data management John Bottega management not as a check-the-box begins with sound data governance—businesses need to own EDM Council exercise—they are focused on imple- the data, and the process in which it is sourced, and institute an menting the best practices within their fi rms. organization-wide data governance vision, which may require There are lots of regulations around this—perhaps the most prom- both organizational change and a change in IT. Organizations inent one is risk data aggregation (RDA). The principals of RDA need to recognize good data as an asset and treat it as a resource direct fi rms to improve their information infrastructure and their risk with value. The quality of that asset is obtained by applying the practice. What’s unique about that regulation is that it’s not overly appropriate business practices, methods and tools to safeguard prescriptive; instead, it’s based on a series of principles. This regula- that data. tion has given fi rms the business case to move dollars to improve their infrastructures and to shore up their risk operations. There will Robertshaw: The key initiative every fi rm can launch is to be a lot of benefi ts to come from that because improving information implement appropriate organization-wide data governance infrastructure doesn’t only benefi t risk, but all operations. culture and structures so that every element of core data that makes up a portfolio’s holdings/valuation/performance/risk Q Typically, what initiatives or approaches can capital should have a clear and relevant data owner; it should also have markets fi rms take with respect to improving their data appropriate process/procedure/accountability underlying its management disciplines? Is the single, enterprise-wide production; and should have workfl ow, escalation and service data warehouse model still valid, or are fi rms looking to level agreements (SLAs) underlying its production. deploy smaller, discrete tools that can provide them and Only once these data governance structures are in place can their business users with more immediate benefi ts? advances in technology and tooling be deployed with eff ect— Sentance: Taking a traditional “Well, where to start?” this is not a technical problem. Technology alone cannot solve approach to data management, then I believe that risk manage- data quality. Data quality is a process, not a project—one in ment is a good place to begin an initiative. It covers all asset which the whole fi rm is and must be interdependently engaged. classes/exposures and as such is a good place to implement a data Data quality is the output of these individually owned data management solution that can incrementally be extended to the management processes. departments being covered by the risk function. Risk is also Neither the enterprise-wide data warehouse nor the smaller highly relevant to the front offi ce, so there is the chance of more discreet tools approach is wholly suffi cient to achieve data involvement and engagement by the front offi ce as they can see quality. For example, data warehouse models are only as good

waterstechnology.com June 2014 9 Special Report Data Management

as the data that fl ows into them. The main problem with data is establishing an adaptable environment—one where they’re not warehouses has been and continues to be the tragedy of the trying to predict the solution that they build today is going to commons—unless the data is owned, it cannot be trusted. As solve all their problems in the future. such, people will always then fi nd it easier to create their own trusted data source. Smith: Partnership with the business is Establishing data quality requires much more than simply still the single most important ingredi- throwing technology at the problem. In the fi rst instance, ent in data management. The tools organizations require a proven data governance structure that technology can provide are probably facilitates all core data processes being owned and actioned to the simplest part of the puzzle. Having an agreed model. Only then does it become benefi cial to ensure everyone aware and agreed to the proc- that each group, team and individual has the tools they need to ess those tools support is critical. own, manage and deploy the data under their control. In this The enterprise data warehouse still context we would expect an increasing use of end-user driven plays an important role in keeping workfl ow platforms and escalation/incident practices to coor- sanitized “master” data. However, a dinate and orchestrate data management processes. Crucially, good master data management (MDM) implementing such a program requires major and across-the- strategy will also take into considera- Enrique Smith board organizational sponsorship and executive support—it must tion the business appetite for intra-day CIFC come from the top. and good-enough/estimated data that can be more cost- eff ectively serviced through operational data stores or business Bottega: Most important is that if you look at information specifi c data marts. These all have to be integrated, however, management from the end-user’s perspective, the objective is to and not considered separate solutions. enable your infrastructure to provide accurate, timely, and con- sistent data to critical functions. It depends on where a fi rm is in Q What are the biggest challenges facing buy-side and its technology journey. If its solution is an enterprise warehouse, sell-side fi rms in terms of their data management practices? or if it’s satellite “data marts,” or a federated data environment Are these challenges mostly technology related, or are they with a semantic infrastructure that enables discovery of the data a mix of technology and operational/governance issues? without moving it around—yes, there are evolving technolo- McInnis: The biggest challenges facing investment fi rms are due gies, but the focus has to be on the objective and the end-user to a lack of proper data governance across the enterprise. Data, and shouldn’t really care how the technology is implemented, as long data governance at large, continues to be viewed as a “technology as it achieves the goal of providing immediate access to critical problem,” when clearly its value and impact stretches across the fi rm. data. In order to leverage the data appropriately and make it most advanta- Now, technologists are smart at this and there are technolo- geous for the fi rm, the business needs to be an equal partner with the gies that are coming out that are far in advance of our ability technology. Good data governance is not a purely technical initiative to manage this stuff . If a fi rm’s technology roadmap points to a and should be shared between the managers of the business and the warehouse environment versus a semantic environment, as long stewards of the IT. as they get the job done, that’s all that matters. Technologists might have diff erent opinions as to the best way to achieve that, Baxter: Solutions are always a combination of technology and but my focus is on the end game. human components. The biggest challenges are about respond- ing to constant change. This means establishing processes that Baxter: In terms of the initiatives and approaches, I think it’s recognize new business needs—often satisfi ed in the short term by about taking specifi c business processes and looking at them spreadsheets—and putting in place migration processes, which can at a more detailed level. It’s very easy to put down visually prioritize these new requirements for more robust solutions. But attractive architectural plans that are far too distant from the this is not a one-off requirement. Constant business change creates day-to-day complexity of what’s taking place in the business. a continuous conveyor belt of new demands. The most mature However, when you try and solve them at a lower level, that’s organizations recognize this and know that business systems will when you encounter problems. While there is defi nitely a role always contain some tactical spreadsheet systems that will last a few for enterprise data warehouses, the idea that a single warehouse years before replacement. As long as you have transparency and is a long-term cure doesn’t pay suffi cient attention to the real control of the end-to-end data management process, including the complexity of data aggregation and manipulation in the face of spreadsheets, then you have the understanding you need to prioritize continuing business changes. What people need to think about and implement continuous improvement.

10 June 2014 waterstechnology.com Roundtable

Atkin: Look at this as a troika of activities. First, there are the data Sentance: Data integration is a huge challenge for larger institu- management issues that are not an IT problem but rather a govern- tions and I think that this integration has probably received less ance problem. Then there are the IT problems—acquire, store, technological focus than other areas such as database and data integrate, distribute. And data harmonization—controlled and management technology. For example, many of the current “big defi ned—is a data problem and a governance problem. And then you data” technologies are designed more for “green fi eld” application have the operations staff , and they have to be able to facilitate the scenarios than for trying to unravel the spaghetti that is the legacy daily business functions and understand requirements. It all becomes data architecture of many larger fi nancial institutions. That said, you a little bit more challenging because it’s collaborative. can’t ignore the people aspects and putting together data governance that delivers in practice is still a big challenge. Bottega: Another important thing that has come out of these new regulations is that it’s not just about technology. What a lot of these Q How do buy-side and sell-side fi rms make the business regulations are doing—and what fi rms are realizing—is there’s a lot of case for embarking on expensive and often onerous data focus on context, as well as technology. The technology is the enabler, management projects, and what challenges do they most where it comes down to how you manage the content, which means often underestimate/overlook? precise, consistent shared meaning of data; understanding how the Baxter: In our world of analyzing end-user computing, the best data fl ows; understanding the lineage; and having proper data govern- contribution to the business case is the provision of visual evidence ance in place, and having a data management program that is driven and analysis of what’s actually going on, such as how data is fl owing by policy and standards. Overall, I think that the regulations have from systems through spreadsheets, back to systems and into other had a positive impact on the industry in terms of mobilizing fi rms to spreadsheets. Very often we fi nd that individual process owners respond to the regulation, which helps improve the infrastructure. do know what is happening but it is only when that information is presented visually do others also appreciate the real complexity. Robertshaw: There are several major challenges facing buy-side Once stakeholders are armed with this demonstrable visual evidence fi rms in terms of their data management practices. Here are some over the extent and complexity of manual spreadsheet activity it comments in priority order: becomes clear that doing nothing is not an option. But equally the • Establishing eff ective organizational data governance structures/ scope of any transformation project needs to be carefully defi ned as a accountabilities—all parts of the organization have their part to play series of smaller steps to avoid being over ambitious. in this respect. A large degree of operational interdependence is required to achieve the necessary levels of data quality. Robertshaw: Prior to making a business • Proven methodologies via which data governance can be imple- case, the greatest challenge is to identify mented across an organization are just emerging. Many fi rms a senior data champion who can get understand the importance of data governance, but don’t necessarily overall organizational buy-in at all levels know where to begin. of the organization, including executive • It is diffi cult to make tangible return on investment (ROI)-based level senior sponsorship, and also elicit business cases to implement improvements in data management and nominate operational management practices. accountability. Unless that level of buy-in • The data market is inherently lacking in quality, hence the need to is there and shared, a data management clean market data. Exchange-based data in particular needs to be project should not start. cleaned. This is a persistent and ongoing problem, even before you A collaborative and authoritative look at data generated within the organization. implementation approach is required, John Robertshaw • Misdirected and superfl uous regulatory interventions and data which recognizes that data management Investit requirements and a lack of clear requirements/specifi cations, which is the new business as usual, and will need to build the organiza- are left to the last minute. There is also the failure on the part of the tional structures, accountabilities and escalation processes to ensure regulators to understand the time and cost it takes to eff ect change data is comprehensively managed—e.g., that the operational, mana- in large fi rms with daily operations and billions under management. gerial and strategic lines of defense are in place to ensure ongoing • Many investment instruments/products are increasingly complex, data integrity and quality. arcane and diffi cult to understand, administer and manage—tech- Many make the case for data management projects by focusing nology/testing processes can’t keep up. There is a similar problem on client needs and/or the integrity/improvement of the investment when it comes to operational/reconciliation processes. process—these are diffi cult to substantiate in tangible ROI terms. • Expensive skill sets are required to manage data at the exotic end of The regulator’s new hard line approach to fi ning fi rms is creating the spectrum. some more tangible ROI drivers.

waterstechnology.com June 2014 11 Special Report Data Management

Many expect the major challenges will be about technology health and well-being. The benefi t that I’m seeing is that those implementation. However, our experience is that most diffi cult implementations are going to have a benefi t across other func- hurdle is to implement appropriate data governance, processes tions. As we build around this regulatory demand, we’ll create and accountability regimes. Many overlook the testing challenges other capabilities. involved and underestimate the complexity, time and eff ort involved—especially managing extended end-to-end data processes Smith: Very few business leaders like to hear the words “no” or and fl ows. “getting that data will take a few days.” Without tackling data management, the spreadsheet-intensive manual processes required to Atkin: There are two things to look at: bridge disparate data—when even possible—cause folks to say these What’s the total size of the bad data tax, things to business leaders. As a CIO, the tough part is getting the and what’s the diffi culty of getting deeper leadership to understand that a successful project starts before there is into this because it’s simply hard to do? a problem. Speaking the language of the business is really important The data tax is signifi cant. You need to to get people to understand why this problem needs to be solved have systems consolidation. You also need before it hurts. to reduce the reconciliation processes In terms of the challenges buy-side and sell-side fi rms most often while also reducing unproductive underestimate, partnership with the business is the most important headcount by putting those people into success factor for these (and many) projects. Just getting the check better functions. for a project won’t get the data stewardship and business processes You also have model precision. A in place. The technology we put into place is only to support and Mike Atkin lot of fi rms are moving to model-based simplify these business processes. EDM Council investment and valuation, and these models require data as one of their input factors. As a result, McInnis: Data governance is a business decision. Before embarking accurate, trusted data allows them to do what they’re supposed to on a costly data management overhaul, fi rms need to understand do. So the business case is that it serves the needs of the analytics and evaluate their current environment, data sources and operational teams even better—they can report and analyze opportunities, and workfl ows in order to identify the expected benefi ts of their future up-sell better. state environment. One of the most commonly overlooked aspects No one should minimize the diffi culty of unraveling 30 years of of a data management project is having the right people involved. acquisition. There are time-to-market and growth pressures. And Senior management and IT must work together and clearly then there’s business silo management: How do you take tens of articulate the challenges and benefi ts of such a large-scale project thousands of applications and unravel them to produce better data? upfront. Ultimately, senior management sponsorship is essential for We don’t want to misrepresent that—it’s hard, expensive stuff , but a process that could be long and expensive, but the potential upside it also costs a lot to maintain that disjointed, operational environ- of increased transparency and operational effi ciency coupled with ment. Sometimes you swallow hard and invest up front, but the better risk management will likely appeal to key stakeholders and benefi t is there at the end or even at the mid-range. outweigh any potential downsides.

Bottega: It’s always challenging in this industry to make the Sentance: Business stakeholders typically highlight the obvious, business case for data infrastructure because of the simple fact direct benefi ts of data management such as reduced data duplica- that data hygiene has not been traditionally viewed as a clear and tion, and hence reduced vendor costs; increased effi ciency, through present danger. Information security is something that everybody automation; improved reliability; and higher data quality. However, recognizes as vital to protecting the fi rm, so those business cases some of the benefi ts such as reduced time reconciling data are harder are usually easier to get agreement on. Data hygiene is a challenge to quantify, particularly when they cut across departments. I remem- because it’s often not something that has an immediate impact, ber recently some industry fi gures that for every $1 of direct costs even though there are security concerns if you don’t manage your on data, there is at least another $5 to $10 spent on processing it, so information correctly. the cost savings are there to be made. In terms of the operational The fi nancial crisis pointed out a lot of weaknesses in our challenges associated with these kinds of projects, they typically industry, and one thing was the net-eff ect of having inconsistent involve getting all the stakeholders involved and on-message. In our data, and not having clarity around unique identifi cation of experience, it’s key to ensure that all parties—front, middle and back entities, and understanding lineage—this all put the Street at a offi ce—are involved and that the work is undertaken with systems disadvantage because data wasn’t accessible immediately in order people who know the business well and are good communicators. to respond to the crisis. That awareness has helped to make the Communication and clear understanding of requirements are key in business case a little easier because data quality does help our my book. Q

12 June 2014 waterstechnology.com $OOWKHðQDQFLDOPDUNHW WHFKQRORJ\LQWHOOLJHQFH\RX QHHGòDQ\WLPHDQ\ZKHUH

:DWHUV7HFKQRORJ\SURYLGHVDOOWKH ðQDQFLDOPDUNHWWHFKQRORJ\QHZV

DQGDQDO\VLV\RXQHHGLQRQHSODFH 3G 4:08 PM

+VU»[MVYNL[`V\JHUHJJLZZ[OPZ PU]HS\HISLJVU[LU[UVTH[[LY^OLYL `V\HYLHZ[OLZP[LPZM\SS`JVTWH[PISL

^P[OHSSTVIPSLKL]PJLZ 3G 4:08 PM

^H[LYZ[LJOUVSVN`JVT   FRUSVXEV#ZDWHUVWHFKQRORJ\FRP waterstechnology.com June 2014