Capital management picks up speed

RMA Journal, The, June, 2003

Remember when stress-testing was given a lot of airtime but few institutions had it in practice? Just about any advance in risk management undergoes a similar rollout of a few highly technical articles that sound like they probably make sense--at least to someone trained to be highly technical, lots of discussion based on sound bites, lots of counter-claims about insurmountable obstacles and unproven performance, lofty claims by technology companies that they've made it easy, and so forth. Finally that elusive concept is captured, everyone's comfort level has moved up a couple notches, efficacy is visible, and the practice is widely implemented. Then something new pops up.

For the past several years, we've been hearing about economic capital and the various means of arriving at this more realistic method of allotting and using capital. VaR, EVA, and a few other acronyms slowly found their way into our hearts and minds. But in 2003, how many banks are really making the most of their capital? Truth be told, relatively few banks have bragging rights in capital management.

Attendance at RMA's 2003 Capital Management Conference showed that regionals and many community banks are now ready to step up to the plate, because the plate appears to be closer and more accessible and the rewards are very promising. The fact that Basel II is supplying a gentle nudge toward better risk techniques in capital management doesn't hurt.

The 2003 conference brought together many of those professionals now in the forefront of capital management and dozens of others who are ready to move past airtime and into reality. The conference opened with an address by Roger W Ferguson, vice chairman of the Federal Reserve (1) (below), and also included an address by Suzanne Labarge, vice chairman and CR0, RBC Financial Group.

Three Views of Risk

Maurice H. Hartigan II, RMA's president and CEO (left), moderated panelists who addressed capital management from the perspectives of credit risk, operational risk, and market risk. Kevin Blakely (second from left), EVP and CR0 for KeyCorp, began by admitting his institution is relatively new at using economic capital to set limits.

"Institutions have always been engaged to some degree in setting limits," he said, "but it was historically done by picking a number out of a hat and telling everyone to remain with the limit. It really didn't address risk within that limit. For example, an institution might set a $10 billion limit on commercial real estate, but that leaves a lot of room for error. The line of business could then invest $10 billion in mezzanine-level commercial real estate or an equivalent amount of very poor quality senior loans. KeyCorp's efforts to factor in the risk element led us to use economic capital. Instead of assigning a dollar risk limit to a line of business, we allocate a block of economic capital and tell the line-of-business managers that it's up to them to determine how to spend it. Let's say that we give a line of business $500 million in economic capital. They could make a lot of high-quality senior loans or make fewer mezzanine-type investments, since those consume more economic capital. It's a risk-adjust ed way to set a concentration limit; we meet our objective by capping the risk level, but the line of business has much more flexibility in determining where to spend that capital and how to distribute the portfolio--whether it be equity, mezzanine, or senior lending. Each of those consume different amounts of that economic capital.

"Let's say that we allocate $100 million worth of economic capital to a line of business. At KeyCorp, we have a 20-grade rating system for our commercial loans. If a loan is at a grade one, it hardly consumes any capital at all. If it's a grade 17, it consumes a lot of capital. A line of business can go through its portfolio and easily calculate, almost on a daily basis, how much capital it is consuming from a credit perspective.

"We use a model that distributes our credits according to our 20 grades. To change a grade from what the model tells us requires permission from Credit Administration, who owns grading system. While we tend to go with the grade that comes from the model, we know that grade isn't always right. If someone can convince the Credit Administration officer that the grade seems inappropriate, he or she will then change the grade. If that happens, however, the Credit Administration officer is required to write up an exception report, which goes back to our quant group, who will use it to refine the automated grading system. We've found a number of areas where the grading system does need to be changed; it's a continuous feedback loop.

"There can be a lot of interesting activity as a line of business closes in on its limit. They may try to cut out their bottom 10% to make room for others, or they may find other ways of generating revenue- for example, fee-driven business. We hit a cap limit several years ago, and the first thing our line of business said was, "You're going to put us out of business." We said "No, let's figure out how to manage better within the allowed limit of capital." And they did. They came up with some very creative ideas that they might not have thought about had they not had these types of limits in place."