How IT can help save the global financial system

ZDNet: Let's start simply: What is The New Data Imperative about?

Baskan: Managing the information use in financial institutions. The degree of readiness to react to financial markets can be improved. The ability to respond can be much more improved if information is managed as a strategic asset.

Nathan: Historically we've focused on transactional changes. The measurement of risk in real-time is just as important. Not centered around transactional processing, but surrounded by analytics. Smaller amounts of latency.

Khan: It’s really about what got us here, and what we need to be aware of to make adjustments. So that all employees, top to bottom, have an understanding of where they fit in.

Baskan: The visibility to the firm's balance sheet got lost in the data. Bits and pieces were stuck in one system, and if you knew you needed it, you had to go get it. If you missed something – thought it was less important, or didn't know – you left it there. It's about bringing relevant information together in a relatively short period of time and making it visible to the people who look at the institution's viability.

There were actual limitations of physical systems. The impact of one nasty thing happening is very big.

Khan: In an investment bank, you’ve got four to five thousand developers making thousands of applications. How do you manage that data? Over the last 15 to 20 years, people were driven by the creation of products. Applications are created but siloed, not having an awareness of how they fit within the institution. You have a significant amount of legacy, because people wont change things until they break. There is no overarching way to manage information. You have to grapple with it and tie it all together.

ZDNet: Where are we seeing underinvestment? Overinvestment?

Khan: Underinvestment isn't really there. Technology is in abundance, there's no shortage of technology, but in terms of using that technology more efficiently...

There is a need for more methodology and a framework. You need somebody who's actually a domain expert in terms of jerry-rigging this information together for what you're dealing with and what you see on the horizon. There is a need for a much more framework-based approached. Getting off your existing siloed applications and toward a framework.

Baskan: A lot of investment went into securitization, trading – constantly acquiring a lot of assets and portfolios. All of that has seen a tremendous amount of investment. That put the pressure on the middle office – post-transactional, all the data has to be gathered. That's where the latency begins. The custodians of the firm's assets weren't working with the most up-to-date data, and didn't have the visibility to the depth of data. That's strain on the back end of the technology.

Nathan: The managers aren't yelling as much as the front-end traders are, and there's no framework. There's a natural resistance there. We believe there's underinvestment in that space.

Nathan: What level of integration produces the most benefit? If you have a framework, then you have an ability to do it incrementally and get benefit from it and progressively go forward.

Khan: It's an early rendition of self-governance, in a way. By creating a framework, people feel the need to conform to something. Before, with a lack of framework, you have mayhem.

Nathan: There's got to be a hybrid model. You actually can have a hybrid model. In some cases it's OK for the query to come back in 15 seconds, in other cases, 15 milliseconds is too slow. The framework establishes where your needs are and where they aren't. There's no one size fits all in terms of what kind of database you use and things like that. If you have a framework, you can get there.

Khan: The siloed approach has bred a level of infrastructure to cope with the inadequacies of the original programs. A foundation satisfies those inadequacies.

Baskan: There wasn't incentive [to improve transparency across silos].

Nathan: It's hard for anybody to say they're not a part of the broad ecosystem. China and India say their banks aren't as involved in it, but the fact is, they're big economies participating in the system. After having seen what has happened [in the latest meltdown], they're putting in regulations to ensure it won't happen.

Khan: The degree of transparency. Counter-party risk – the ability of anybody to do business with you will be compromised. The climate is completely different now.

ZDNet: One oft-used phrase in the book is “one version of the truth.” Explain the importance of this.

Khan: The challenge in the past is that everything was so batch-driven that there was no ability to view anything during the day. So first, the ability to deal in near-real-time; and second, the ability of intraday trading.

Baskan: [Understanding] exposure to risk all the way up to the balance sheet. That probably isn't happening in a lot of institutions today. A couple hours a day you get a complete view. Look, nobody went out and bought new technology to satisfy Tim Geithner. It shows you how severe the limitations are; how long it took. It's all the other logistics before you get to the number crunching. And that I think is revealing.

Nathan: Compliance is not a bounded time frame. The time to prove that you're compliant has reduced drastically. The granularity that you must prove compliance is getting much less high-level. Compliance is built-into the activity itself. These points [that we've been making] are another reason why you've got to look at how you’re managing your data.

ZDNet: In the book you talk about traders gaming the system with rosy data. How can we avoid this again?

Nathan: It is trying to measure the impact of an activity while you're doing that activity. In the front office, you focus on how they’re doing the activity and the impact of it. You mitigate the ability to game the system.

Baskan: If you have the core data available, you can anticipate. There are examples of models in which everybody believed would work -- but in fact, failed. One part of governance is to detect those instances. Once you establish a norm...you can't hide that information if you designed the system to distribute the signals. Somebody will see it.

Khan: Most investment banks have migrated toward black box trading. If you look at data in a short period of time...for example, one institution had 150 Ph.D's on staff, but they had older data, and it turned out that 35 percent of their predictions were wrong.

The methodology you have to use must look at data over good and bad times. Alerts must sound for internal data and external sources.

ZDNet: Why did you write this book?

Baskan: There's an underlying element about capability. Institutional capability to anticipate, identify, respond to problems and prevent deterioration. Without that capability being there...if you want a more stable, transparent, healthy system, you need the capability to design, it, run it, control it. There's a cause and effect relationship between what happened and what happened on the technology side.

Khan: We wanted to provide an instructional guide that tries to be comprehensive and concise. We wanted to try to cut through the superfluous information and get to the "guide."

Nathan: We could have done more to help the firms through this process. Management of information is not a desirable thing to do but a very necessary thing to do. Managing it in a tactical way doesn't meet the firm's interests, just strategic interests. Let's do it now before we hit the next crisis.

IT can certainly help save the financial system. Managing the firms through those times.

Baskan: The financial system is healthier when data is managed properly and reflects what's going on in the time frame decisions are made. All the other behavior, policies, is beyond the realm of technology. Good science is never going to lead to improved morality.

And you know what? Some of Sybase's clients have already seen the benefits: 22 have signed up; one firm has moved to measuring risk in hours rather than days. That's the best proof point.

The New Data Imperative: Managing Risk Real-Time Risk in Capital Markets will be available June 16 for $21.95.