When a major international bank bought a large North American finance company, there were concerns the finance company would take its new parent into risky lending where it had little experience. Investor fears were calmed by the finance company's use of software from Santa Fe-based Strategic Analytics, which predicted bad debt provisions accurately and explained recent performance.
“We have developed software that quantifies the key performance components of consumer portfolios,” said Joe Breeden, president of Strategic Analytics. “With that as a basis, our software does loss and revenue forecasting, stress testing, and calculates economic capital. We enable our clients to gain new insight into the dynamics of their consumer portfolios, and use that knowledge to do scenario-based forecasting.”
One large financial services company, which has been using Strategic Analytics for more than a year on cards, has now extended the analytical tool to attribution rates on vintage pools and to payment rates, said a credit executive.
“We didn’t have a good forecasting platform in the past; we used Excel charts. This is the first rigorous analytical tool and it has allowed us to do something amazing – we can separate the impacts of the economy from our internal actions, such as tightening credit score, and higher or lower credit lines – the multiple things we do to control credit,” said an executive at the company.
In regression tests, the tool from Strategic Analytics works well, he added.
“It gives pretty good results. It has more noise on the edges, but I think that is a data issue. It seems to hold up very well and we use it in everyday forecasting,” he said.
Strategic Analytics creates analytics to drive decision-making. Its tools run on Microsoft Windows XP, and its analytical tool easily fits on a notebook computer.
“We also use Microsoft Excel as a standard file format and editor of scenario files,” explained Breeden. “This allows one user to easily send their forecast assumptions to another user, recreating the forecast exactly. Our software provides a number of scenario-design tools, but when the user wishes to create something custom, they have access to all of Excel's mathematical flexibility.”
Strategic Analytics greatly speeds the generation of forecasts, he added. “We are doing analytics for strategic decision making,” said Breeden. “We help senior executives make the right decisions about managing their portfolios. Typically, in the past, they would have teams cranking out forecasts over weeks or months. Then they’d look at them and agree or disagree, but not have time to change them. We can turn around a forecast in under an hour; if management objects, we ask them which of their assumptions to change, revise the forecasts and we can then present an hour later. Until you get that sort of fast turnaround, the analytics are not part of the decision making. Now the decision makers can use the analytics in real time, while before it was based on what feels good, and we’ll check back in a year.”
The company, which has focused on retail lending portfolios, counts six of the top ten consumer finance firms in North America as customers, including a major credit card firm and HSBC. In fact, the company’s software currently analyzes over $1 trillion in consumer loans in North America and is beginning to expand overseas. Its LookAhead application can analyze all types of consumer loans including mortgages, credit cards, and auto loans.
“We can pull a portfolio as large as any of the major card players, run that through a laptop and have the forecast in the afternoon,” Breeden said.
Strategic Analytics will take data from a variety of systems, such as SAS or Cognos or relational databases. It uses monthly vintages in retail lending. That way, 12 years of data fill an Excel spreadsheet.
“If we touched every account we would need a mainframe, but by distilling the data into vintages we can do it on a laptop. At the vintage level, we see structures and information that can be compared across vintages, information you would miss if you were looking at accounts, where you can miss the forest for the trees.”
The software offers decision makers guidance on their future direction.
“Loan portfolio managers want to understand what to do next,” said Breeden. “They want to understand the current trajectory in areas such as attrition and charge-offs. They also want to know what could happen under a range of different scenarios. Is the economy going to get better or worse? As you pursue new marketing efforts do you need to steer your marketing to pursue one particular class of customer or concentrate on a particular geography. They want to create a diversified portfolio while not leaving money on the table.”
With the software, a lending company can examine results across products and measure the originations quality, vintage maturation process, and the impact of outside economic factors.
For many lending institutions, the housing boom – or bubble – is an area of current concern. Customers are taking regional economic data to determine if they want to shift strategies, such as slowing down in California and increasing loans in North Carolina.
Because it is highly dynamic, Strategic Analytics can help banks avoid the classic rearview mirror view of the economy, which often drives them to respond to market turns 12-18 months late.
“During the last recession, a financial institution was using our analysis across all their portfolios. The recession ended well before the worst impact on consumers, which kept coming for another 12 months,” said Breeden.
The CEO didn’t believe his marketing staff, so his staff placed a call to Strategic Analytics for help.
“We showed that the economy had gotten 20 percent worse, while their portfolio had declined only 10 percent because of their countermeasures,” Breeden said. “We persuaded them to ride it out for 12 months and then get aggressive in booking more accounts as the economy changed. They had been talking about paring back to a fraction of their normal books, taking themselves out of the market with a huge loss of revenue.”
Instead, banks can take other measures, such as changing their credit score cutoff or revising their pricing and product mix, Breeden noted.
“Often, at the point of maximum pain they dial everything back, which is expensive when the economy improves,” he said. “When retail lenders overreact to the downside and upside, they double their losses going down, and cut their revenue in half on the upside through reactionary management.”
The complexity in the Strategic Analytics tool is in the development, not in the use.
“You no longer need an in-house Ph.D. to concoct a hand-coded model,” explained Breeden. “Instead, you can take any data asset from any retail portfolio anywhere in the world and drop it in. Within minutes it will train up a model on your laptop and give you a forecast. We take all the witchcraft out of the process. You get a result that is reproducible and explainable. We brought concepts from engineering and science into banking to accomplish this.”
www.strategicanalytics.com