Banks are working to refine the stress-testing processes that regulators began requiring in the wake of the financial crisis, with a goal of making the expense pay off in the form of better business decisions.
Stress tests can uncover untapped opportunities, as well as flag unrealized risks, the thinking goes. So rather than viewing these regulatory mandates only as a burden, bank executives are starting to recast them as opportunities to get valuable insight into their banks.
"It's really a business intelligence opportunity," said Aite Group analyst David O'Connell, voicing the consensus of speakers at "Stress Testing 2015," a summit hosted by American Banker.
Even smaller institutions should run stress tests on their loan portfolios using calamitous assumptions and "Black Swan" hypotheticals that may never happen, because of the valuable knowledge analyses of these scenarios produce, several speakers said.
"One of the benefits of stress testing is it forces you to keep track of the data that you need to make better decisions," said Andy Spero, head of model development at the $122-billion asset Regions Financial Corp. in Birmingham, Ala. "If the data set is good enough for modeling, then it should be good enough for everything else. It's something every bank should do as a normal course of business."
Stress testing uses mathematical models fed by loan data to forecast how the values of a bank's assets will react to changing economic circumstances. It's a practice that naturally provides a more intimate understanding of the bank portfolio, because managers must organize and analyze the loans at a granular level. And the better the portfolio is understood, the better informed bank executives are when they try to decide how to safely grow assets and, thus, the bottom line.
"We can use the results from the stress tests to actually make credit risk management decisions, whether it's avoiding some areas or concentrating in others," said Paul Spotts, senior vice president for credit underwriting and administration at the $1.2 billion-asset Orrstown Bank in Shippensburg, Pa. "Our challenge is to make stress testing a part of the credit question."
Large banks (those with $50 billion of assets or more) and midsize banks (between $10 billion and $50 billion of assets) are required to apply annual company-wide stress tests to their loans to reassure regulators they have enough capital to survive various economic and market shocks. Large banks also undergo stress tests administered by the Federal Reserve.
Smaller institutions (under $10 billion of assets) have no proscriptive requirements. But in practice, regulators often have community banks do stress tests for specific loan types, particularly if risk management practices are found wanting during safety and soundness exams, speakers said.
What concerns regulators the most lately are commercial loans. "There are some red flags that are out there with respect to leveraged lending all the way down to the community bank level," said Matthew White, a stress test analyst for midsize banks at the Office of the Comptroller of the Currency. "There are some signs that the market is frothy. We're not seeing that bear out in actual defaults at this point." But regulators are scrutinizing whether banks' methodologies reflect changes in portfolio quality and higher loss estimations for these loans, he said.
In seven recent formal agreements with smaller banks, the OCC required the banks to do stress testing, and in five of those agreements, the agency cited insufficient oversight of commercial loans.
Forcing bank staff to tell the story of what can happen to loans in multiple adverse scenarios regulations require banks to report stress test findings narratively, as well as numerically lends itself to discovery of business solutions.
"Narrative reporting is not only for the regulators; it's an actionable and valuable document for the bank," said Peter L. Cherpack, a director at the risk management consultant Ardmore Banking Advisors. "It says what we learned about our risk and how it may affect our capital, and this is what we're going to do about it."
Banks also must demonstrate how they've reached their conclusions. "One of the most critical pieces is the documentation portion," said Taylor Pool, stress testing manager at the $12.6 billion-asset Hilltop Holdings in Dallas. "Making sure every step along the way has been well documented, every conversation that's been had with line of business experts. Having those notes in place every day, we can see every position that's been made and all the model functionality, version one and version two. Obviously, the goal is not just to comply, but to be able to make these models usable for your business."
Coming up with plausible forecasts requires collecting and sorting the right data into the mathematical models. "It starts with improving data capture and origination," said Kevin Kirksey, managing director at ALM First Financial Advisors. "What we're saying is not for you to keep in your core every single social security number and borrower name and start creating correlation coefficients on the social security numbers. But if you can put data stewards in place that have real ownership of data at their various business lines, being very clear about who owns the data fields, who owns the responsibility for updating and refreshing the tools over the lifecycle of those assets, you'll be at the point where you feel very good about your modeling."
So "big data" is a misnomer: Finding nuance amid the morass is where the value's at; and gleaning real knowledge can only occur if the data is first finely organized. Specifically for banks, this entails grouping loans with similar characteristics together, so apples-to-apples comparisons can be made. This is not always easy, particularly with commercial and industrial loans, which can have very different features. Standardized data sets on floor plan lending exposures, for instance, can be difficult to assemble.
"C&I is all over the board," said Christian Albela, associate director at Protiviti, a Princeton, N.J.-based consulting firm. "Airplane lending is one product class that hits the balance sheet from three lines of business. You technically have multiple variables for each one of those areas, because the underlying demographic supporting that credit is slightly different one's a lease; one's large corporate customers; and another's in private wealth management."
Given that, it's unsurprising Aite Group's O'Connell finds negative perceptions of "stress test hell" persist among senior management. It takes headache-inducing hard work to cull what matters from the minutiae. But enlightenment resides in those details, along with the devil.