Models have had a mixed ride in the wake of the financial crisis.
Bank regulators may have reined in capital models for operational risk, credit risk and elements of market risk but, at the same time, they have pressed banks to spend more time modelling the performance of their businesses under stress, and to look further into the future. Accounting standard-setters are doing something similar with loan-loss reserves.
This work typically starts out as a compliance project, but it has much greater potential. Which is where technology vendors such as Moody’s Analytics come in.
The firm scooped four categories in Risk’s separate Market Technology Awards – where winners were chosen by a panel of industry judges – including categories for stress testing and economic scenario generation (ESG) products. It lands our overall technology vendor award because its strength in these crystal-ball products put it at the intersection of some fascinating trends – the rise of data science, cloud computing and the growing ubiquity and power of analytics in daily management.
The proximate challenge these products address is that of coming to an enterprise-wide view of an institution’s risk position. “When challenges come, they don’t come handily in silos,” says Andy Frepp, managing director at the firm. “That banks and insurance companies have had to work across silos to properly stress test has been challenging.”
But this has also brought opportunity. As the cost of technology has dropped, and specifically the cost of running hugely complicated calculations across thousands of servers, the potential to transform regulatory compliance into actionable insight has exploded.
“We’ve taken products designed to help with regulatory compliance and turned them into analytics tools,” says Steve Tulenko, the firm’s executive director of enterprise risk solutions. He gives the example of the US Federal Reserve’s annual stress-testing process. While regulatory compliance requires the production of a static report, the tests offer the potential for deeper, ongoing insights, such as how the risk position alters under the effect of changing oil prices or interest rates, where correlations might exist, and how risk managers might respond.
The practical implications are that institutions can drive awareness of risk towards the front officeSteve Tulenko, Moody’s Analytics
This process has been enabled by the revolution in computing power, and the ability to tap the almost limitless capacity and flexibility offered by cloud computing. “It’s driving an enormous wave of fintech innovation,” says Keith Berry, who leads the emerging business unit for Moody’s Analytics.
Berry gives an example of a customer that conducted the same exercise – carrying out more than 1 million simulations, of more than 1.5 million loans, with a 10-year history – in two different ways, to test the power of the cloud. Running the simulations on-site took 300 hours; using the cloud, and the ability to run the simulation on thousands of machines in parallel, it took less than five. “This elastic computing concept, using parallel processing, can cut run times gigantically,” says Tulenko.
This allows Moody’s to offer its clients more dynamic insights into their overall exposures. “The practical implications are that institutions can drive awareness of risk towards the front office,” he says. So a loan officer or a trader could, in real time, consider how a loan or a trade would affect their bank’s capital adequacy position.
“These sorts of analytic insights are the features we’re trying to build into the product every day, to gear towards the objective of the front office, of the board and, ultimately, of the regulator, to help the banks understand what would happen ‘if’,” says Tulenko. “It’s about producing management dashboards rather than merely regulatory reports.”
This kind of power is prompting firms to look for insights in previously neglected datasets – in risk management, for example.
What’s fascinating to us is taking some of those same data sources and applying them to credit – looking at how that data affects an institution’s real-estate portfolio or, from an insurance perspective, which buildings have or have not had maintenance on themKeith Berry, Moody’s Analytics
“The cost of storing data used to be huge – now, the marginal cost is almost nothing. We have customers who are literally keeping website click data alongside their risk data, because they believe there may be correlations they could identify that could reveal fraud, for example,” says Berry.
Alternative sources of data can also offer investment insight, Berry continues. He gives the example of a company that collects data on every renovation that requires a building permit across every county in the United States: it sells that data to hedge funds trading stock in Home Depot, who have been able to use it to accurately predict the DIY retail giant’s quarterly earnings numbers.
“What’s fascinating to us is taking some of those same data sources and applying them to credit – looking at how that data affects an institution’s real-estate portfolio or, from an insurance perspective, which buildings have or have not had maintenance on them,” says Berry. “There are hundreds of companies commercialising these datasets. It’s going to change what we can do.”
The award judges were impressed with the technical capability of Moody’s Analytics, and the sense that there are few – if any – companies that can offer ESG and stress-testing technology with similar capabilities. But as important to Moody’s clients are more traditional attributes of customer service. One European insurer praised Moody’s Analytics’ “good ongoing support”, while a second described the supporting documentation as “a great strength”.
For Tulenko, flexibility in service offering will be key to Moody’s Analytics’ continued success. By allowing clients to combine elements of its product suite with other technology providers, and integrate them with the client’s own systems, the company can provide ‘micro-services’ closely tailored to the client’s needs. “We’re engineering for modularity and adaptability. That’s our primary strategic direction,” he says.