Print this article | Return to Article | Return to CFO.com
Companies are starting to figure out how their banks make money. Banks actually like that.
Tim Reason, CFO Magazine
October 15, 2005
Some banks just can't take a hint. When a large apparel maker renewed its multibank credit facility a few months ago, the company's treasurer also saw it as an easy way to trim its outsized group of 13 banks. Only three years earlier, several bankers had complained bitterly that the apparel maker's business wasn't meeting their hurdle rates. Given today's tight spreads, thought the treasurer, surely those getting next to none of the company's other business would simply walk away rather than continue lending.
But, alas, even the banks getting bupkis wouldn't budge. With nary a word about losing money, they all signed on.
Welcome to the loose end of the credit cycle. But beware. Bankers obviously expect something in return for the favorable five-year backup credit lines they've been handing out lately. And when the credit cycle turns tight, complaints about hurdle rates will again make life awkward for companies that try to maintain access to credit through a diverse bank group. Such discussions have strained personal banking relationships in the past, yet they've also increased corporate awareness about the way banks calculate their returns. Going forward, companies that adopt this sort of quantitative analysis for themselves may find that it not only supplements traditional relationships, but strengthens them. Indeed, banks are surprisingly willing to help CFOs build models that can distinguish between legitimate gripes and sales pitches.
RAROC and a Hard Place
Some 20 years ago, Bankers Trust Co. first began using a risk-adjusted return on capital — or RAROC — model to evaluate the profitability of a transaction given the risk profiles of its commercial borrowers and the resulting return on the bank's capital. Today, such return models have become all but ubiquitous in the banking industry. So accepted are they that regulators are in the process of allowing banks to determine for themselves how much regulatory capital they must set aside for each transaction (see "Basel Faulty?").
For many corporate customers, however, RAROC has a bad reputation. Most first heard the term used as a cudgel, with bankers telling them that they weren't profitable customers and then pressing for other business in exchange for credit.
"Most banks began to internalize the idea that large corporate lending is unprofitable about five years ago," says Nick Studer, head of the North American corporate and institutional banking practice at Mercer Oliver Wyman. The 1999 repeal of the Glass-Steagall Act, which had formerly separated investment and commercial banking, was quickly followed by flame-outs in the M&A market, the IPO market, and the last generous credit cycle. "It's these last five years that made bankers realize that cross-selling is not a luxury, it's a necessity," says Studer.
But handled clumsily, particularly by the investment banking arms of newly formed universal banks, cross-selling irked many corporate customers. Indeed, in the past few years, the Association for Financial Professionals has been surveying its members about such practices, strongly hinting that they might constitute illegal tying. "Cross-selling is what retail bankers do well," notes Studer. "On the investment banking side, honestly, it's not something they do well. It's boring, and not as sexy as chasing deals."
Nor was the reputation of RAROC helped when some banks simply stopped lending to certain companies — even whole industries — based on their risk profiles. "We actually had a lead bank drop us in 2000," recalls Jeff Burchill, CFO of commercial property insurer FM Global. "They decided to drop all insurance companies because they didn't like the risk profile."
"Up until this time, banks and companies have had a combative relationship," observes Studer. Indeed, banks have had reason to complain as well. Corporate customers, after all, didn't hesitate to tie their banking business to demands for credit. And they rarely showed any understanding of their bankers' needs. "Clients used to haul out a spreadsheet listing all their banks in descending order of credit commitment and noncredit fees," says Bradley A. Hardy, senior vice president of corporate banking at Wells Fargo Bank N.A. "But they were only looking at the revenue line; they were not looking at profit at all. You can imagine profitability in an M&A transaction is vastly different than a cash-management business."
Show Me the Model
Increasingly, however, RAROC-style return models are contributing to improved corporate banking relationships. "Corporates are getting more sophisticated," says Studer. Bankers agree. "Clients are developing a better understanding and appreciation for the return models that we use," notes George Calfo, head of Citigroup's national corporate bank. "We think that's positive."
Of course, the degree to which companies actually understand their bankers' returns still varies widely. "Some clients are very analytical, some are more feel-oriented," says Calfo. Most probably fall into the latter category, relying on conversations with their bankers and a basic understanding of how business is allocated. Indeed, 31 percent of respondents to a survey by CFO (see "Last Banks Standing") said they rely on their banker to tell them how the bank values their relationship.
"We don't have a model that tells us how profitable each business is [for our bankers]," says Jack Wagner, assistant treasurer at Cabot Corp. "But each bank has a different appetite for each type of business, so we do try to accommodate them."
Sealed Air Corp. takes a similar approach. "We do not maintain a score card that shows how revenues and perceived profitability are getting allocated among the bank group," says CFO David Kelsey. "We do make an effort to treat these institutions fairly and have them feel that when a transaction comes up, they will have a seat at the table because of their participation in our global credit facility."
A small but growing number of companies, however, are seeking ways to quantify just how effectively they allocate that business. "We do a lot of measuring of fees," notes FM Global's Burchill, echoing 11 percent of survey respondents who said they calculate the revenue that each bank earns on all transactions with the company.
Almost as many — 10 percent — claim they estimate the margin that banks earn on transactions with the company. "We do have clients asking us about our model and how it works," says Wells Fargo's Hardy, who has made multiple presentations at treasury conferences over the past two years to explain the basic RAROC model. "And we have some clients who are trying to put together their own model."
"We have seen a handful of companies that have taken a highly rigorous approach to this kind of quantitative analysis," says Susan Skerritt, a partner at Treasury Strategies, a consulting firm specializing in treasury and financial management. By that she means actually trying to reverse-engineer the RAROC model used by each bank in the group. But the models vary, and several assumptions and variables are available only from one source: the bank itself.
"The companies that are doing this in the most rigorous fashion are actually requesting that their banks explain to them how that individual bank's model is calculated," says Skerritt. According to CFO's survey, 8 percent of respondents said they ask their bankers for that information. And while companies don't always succeed in getting it, in many cases they do. "The banks understand it is in their interest for clients to understand their profitability," she says.
Indeed, while just a few clients have asked for a look at Citigroup's model, says Calfo, "we get as detailed as folks want to get. The greater the transparency, the better our relationship." Former Bank of America CFO Marc Oken told CFO earlier this year that the bank holds "real detailed" conversations with companies that aren't meeting its hurdle rates. "We will either give a customer our model or tell them why their model is wrong," he said.
Of course, with less than 10 percent of companies actually asking for the models, banks are not under much pressure to be completely forthcoming. Wells Fargo's Hardy, a strong proponent of improved corporate understanding of RAROC models, notes candidly that banks are unlikely to give away the store. For example, he says, banks are highly unlikely to tell corporate customers how their models assign costs to noncredit products.
For most companies that estimate the profitability of their banking business, however, it's enough to have a single, simpler return model that estimates a common margin for services from members of the bank group (see "A Simple Return," at the end of this article). "The highly rigorous approach where the treasurer is doing it on an individual bank basis is exhaustive but may be more than your organization needs," says Skerritt. "But it's absolutely a best practice to have at least one common standard benchmark."
Doing the Math
Ken Schutte, treasurer at brokerage firm Edward Jones, uses just such a benchmark to manage his company's 11 banks. A commercial banker for 26 years before joining Edward Jones in 2000, Schutte explains, "I was used to the concept of banks trying to evaluate the profitability of client relationships. So I just reversed it."
Not every service can be measured. For example, Edward Jones offers a company-branded credit card to its clients through MBNA Corp. "I frankly don't know how to assess the profitability of that," he admits. But with a few such exceptions, Schutte attempts to capture all of the business done with each bank — right down to the fees Edward Jones pays for the three ATM machines on its campus.
In most cases, he says, the information isn't hard to collect. "For short-term loans," he says, "we know the outstandings we've had with each bank, we can estimate how many days so we can get an average outstanding per borrowed day, and we know the spread they've been charging." That's enough to estimate the banks' net interest income. Similar information is available for other types of loans, as well as for short-term investments, he says.
Likewise, Schutte tracks treasury fees such as cash management and fees for banks that act as collateral agents or trustees on bond issues. "We capture as many of those types of fees as we can," he says. To calculate each bank's pretax net income from fees, he says, "we then apply the [overhead] efficiency ratio that is published in their financial statements." (Schutte makes an exception for banks with excessively high ratios, instead applying the average ratio for Edward Jones's full bank group. "We don't reward any bank for being inefficient," he explains.)
Finally, Schutte totals up fee and interest income and subtracts taxes — based on the tax rate that the banks report in their financial statements. "That gives us our best guess at their after-tax net income figure for Edward Jones," he says.
Large one-time capital-market fees for services such as private placements — which Edward Jones does every few years — are spread over a similar interval. "The bank that does the private placement gets a pretty significant fee," he explains. "We allocate that over the interval so they don't have such a volatility in their earnings and return calculations."
To calculate the bank's return on assets (ROA) for his firm, Schutte divides the total net income for each bank by the average assets Edward Jones has placed with that bank. He also calculates a return on equity (ROE) using an assumed capital requirement of 8 percent. That's the amount of capital banks must set aside under international regulations known as Basel I. "We don't try to determine what their individual equity ratio is," he says.
Schutte ran early results of his model past some of his bankers several years ago and made only minor tweaks based on their feedback. "For the most part, we were in the ballpark," he says. "It is an art on our side, but it's still an art on their side, too."
Schutte compares the ROA and ROE figures to the ratios published in each bank's annual report to get a sense of how valuable Edward Jones's relationship is to each bank. "If we are significantly below where they are, then we conclude that we are deemed not sufficiently profitable. And if we are significantly above, we conclude we must be significantly profitable." He says he has used the model only once to move business from one bank to another, but he uses it regularly to award incremental business. "We're not trying to wring the last nickel out of our bank relationships; we are trying to ensure that our relationships are profitable to them so that we can rely on them being credit providers," Schutte says.
The Chance to Bid
Schutte's attitude is one bankers would dearly like to see more of, particularly as commercial banks reach for more investment-banking business and investment banks increasingly are forced to participate in low-return business such as credit facilities.
Wells Fargo's Hardy argues that companies can start improving their bank relationships simply by taking an inventory of all financial services purchased throughout the company and making sure that their bank group gets a shot at the business. Many potential bank services, he says, are purchased by the human-resources department, purchasing departments, operating divisions, and the legal department without consulting anyone in finance. "Banks want the opportunity to compete for that business," he notes.
Fidelity has been using its centralized Bank Services Division to evaluate every bank service purchased throughout the company for more than seven years, says director of bank services Brenda Walsh. "Our responsibility is to make sure management is aware of the full scope of all of Fidelity's banking arrangements," says Walsh. "The mere fact that there is a role like mine at Fidelity shows that we take it very seriously."
But not every company is lucky enough to have Fidelity's resources, and there are other challenges. "Corralling our foreign subsidiaries to use one of our corporate banks can be a challenge," says Cabot's Wagner. "I think the challenge for a lot of multinationals is just the limited amount of bank business that you can allocate back."
Indeed, 12 percent of CFO survey respondents said they have difficulty giving their banks enough additional business to justify their credit needs. "To be brutally candid, that may be a situation where we are never going to have a profitable relationship," former Bank of America CFO Oken told CFO earlier this year. "We tend to exit those relationships by mutual agreement."
But it's not always the company that's the problem: After several waves of consolidation in recent years, banks themselves don't always recognize every stream of revenue in their return models. In CFO's survey, only 29 percent of respondents said their bankers were aware of every piece of business from their companies, while almost as many — 27 percent — said bankers were unaware of all of the business their institutions received from the companies. "I have seen numerous occasions where the corporates have a better idea [of the business relationship] than the banks," notes Mercer Oliver Wyman's Studer. "The state of data in many of these banks is appalling."
Wagner says he relies on internal reporting to determine how much Cabot spends with each of the six banks that participate in its revolving credit facility and adds that only one bank in his group, Citigroup, tracks the amount of business Cabot does worldwide.
Likewise, the bank's return model — or the banker's incentive plan — may not include or value certain services. Wells Fargo's Hardy says part of managing a bank group is figuring out which banks value which services. And experts say that after chasing investment-banking fees, many banks are now seeking a more balanced mix after realizing that annuity businesses such as foreign exchange and cash management are also desirable.
Of course, there will always be limits on how far companies will go to maximize their banks' return. Several finance executives told CFO they wouldn't have their banks administer their 401(k) because their fiduciary responsibility required them to assess that business with the interests of employees rather than shareholders uppermost in their minds. Others simply don't believe their banks can handle all the services they now offer. "Our commercial bankers tried to sell us M&A," remarks one CFO of a transportation company with more than $1 billion in revenues, who asked not to be named. "There's no way I would ever use them for M&A. Ever. I'm going to go with a big investment house."
Nonetheless, the relationship between banks and their corporate customers does seem to be changing, and the better understanding of return models may be simply a first step. "RAROC serves as a foundation for a broader economic partnership between the bank and its client," observes Mercer Oliver Wyman's Charles Bralver, head of the firm's strategic finance practice. Going forward, Mercer's Studer predicts corporations will treat their banks more like industrial suppliers. "You haven't really seen the shared design and just-in-time relationships that you see in manufacturing," he says. "It's only really now that we start to see that kind of relationship." Indeed, while 22 percent of CFO survey respondents said they manage their banks no differently than any other commodity vendor, almost half — 49 percent — said they view their banks as business partners and involve them closely in business decisions.
"Our corporate clients have a good understanding and appreciation for risk-return models, and it's hard for me to imagine that it will not continue to evolve," agrees Citigroup's Calfo.
To be sure, that's happening slowly now. Despite a flattening yield curve that may yet rein in bank lending, it's hard for companies to devote resources to a more rigorous analysis of their banking relationships when credit is still cheap and easy to come by. But Skerritt notes that many companies have used today's rates to lock up five-year backup credit facilities. "If you don't have to renegotiate in 364 days or three years, perhaps you can use the resources now [to build a return model]. Then when the credit cycle changes, you have this model in place and are prepared." Since being short-staffed is one of the most common complaints of treasury departments, she notes, "managing resources to anticipate a change in the credit cycle makes a lot of sense."
That way, when the credit cycle turns, CFOs and treasurers can be confident that the members of their bank group really want to be a part of it.
Tim Reason is a senior editor at CFO. Randal Rombeiro, a former Fortune 1,000 corporate treasurer, contributed to this article.
A Simple Return
Corporate borrowers can be easily frustrated in their efforts to understand the proprietary — and often volatile — RAROC models for each bank in their group. But for most companies, it is enough to create a simpler common return on regulatory capital (CRORC) model that standardizes many of the assumptions. Such a model can be used to quantify the non-risk-adjusted return on regulatory capital for each bank (and can be easily maintained on a spreadsheet). The result can help borrowers identify any imbalances in the distribution of credit and fee-based business within their bank group. Moreover, the model allows borrowers to analyze the impact of awarding (or moving) transactions and services — a helpful tool when credit gets tight. Building it involves some simple steps:
A typical CRORC model contains a separate tab for each bank along with a summary that compiles key statistics for comparison. Each bank's tab should itemize the individual products and services along with the associated revenue, margin, and regulatory capital allocation. Tabs structured in "trailing 12 month" and "forward 12 month" formats can help CFOs and treasurers analyze potential changes. —Randal Rombeiro
Services to Think About
Beyond credit facilities, cash management, and foreign-exchange transactions, what should be included in a return model? A different way to ask this question might be, "What am I currently buying from a nonbank provider that could be sourced within my bank group?" Below are some additional items for consideration. —R.R.
Accounts-receivable or -payable outsourcing
401(k) / pension-plan administration
Commodity and energy hedging
Merger and acquisition fees
When building a common return model:
Be realistic: Borrowers with larger bank groups are more likely to benefit from a common return model, as are those from troubled industries.
Be open: Many bankers are still hesitant to disclose product and service margins but are more likely to oblige if they understand the reason. Borrowers, in turn, should avoid using this information simply to squeeze a percent or two out of disbursement costs.
Change your style: Consultants say that using a return model is part of building a mutually beneficial vendor/supplier relationship with banks. This may mean mending fences and a change in style if past relationships have been combative or adversarial.
Look for more business: Companies should examine their operations for additional business that could go to the bank group (see "Services to Think About," above). Centralizing bank product and service procurement — a tactic used by companies such as Fidelity — may help.
Look for future business: Evaluate the company's strategic plans for future product and service needs (such as M&A, share repurchase, or new hedging programs).
Ask banks for ideas: Odd as it seems to ask for sales pitches, banks are likely to view interest in other products as a sign of good faith.
Consider foreign banks: Varying tax rates and other differences may make it easier for overseas banks to make their margins on certain types of business. —R.R.
Adjusting Credit Ratings
Default probabilities vary for companies with the same credit ratings. Below is the range in basis points (bps) that Moody's KMV, a division of Moody's Corp., uses to calculate default probability, given a firm's rating.
AA/Aaa = 2 to 3 bps
AA/Aa = 3 to 10 bps
A = 10 to 24 bps
BBB/Baa = 24 to 58 bps
BB/Ba = 58 to 119 bps
Source: Moody's KMV
How RAROC Works
A RAROC model is simply a return on equity (ROE) model in which the numerator (net income) and denominator (in this case, capital) have been adjusted for risk — a simplified example of which follows:
Bradley A. Hardy, senior vice president of corporate banking at Wells Fargo Bank N.A., explains that banks risk-adjust net income by deducting expected loss, calculated as loss given default (LGD) multiplied by default probability and by exposure amount. It's the last of these that's tough for corporate borrowers to determine: a simple term loan exposes the bank to a 100 percent loss, but what is a bank's exposure to an undrawn line of credit? "Since each bank may calculate this differently, it's worth trying to get a sense of how 'friendly' each bank's model is to unfunded commitments," Hardy suggests.
Likewise, while banks may be unlikely to share overhead costs with their clients, asking about them may give a company some sense of the profitability of a particular service.
The most complicated part of the RAROC equation, however, is the denominator. While the amount outstanding is usually obvious, and capital multipliers are easily understood (for example, a AA credit has a capital multiplier of about 7X), the unexpected loss percentage represents the real "black box" calculation that determines the amount of economic capital required. Measured as the standard deviation of expected losses, it is calculated as follows:
Several of these variables are impossible for corporate clients to figure out. For a term loan, for example, the variance in exposure amount equals one, but for just about any other financial product, it can be a tricky number for banks — let alone their clients — to calculate. Likewise, companies are unlikely to be able to deduce the variance of default probability — banks use a figure based on historical experience.
A company's default probability, by contrast, is easier to determine, and can be a useful number to know. However, it can vary widely by bank, depending on whether the banks use credit ratings, Moody's KMV ratings (see "Adjusting Credit Ratings," above), internal models, or some combination of all three. "People are often surprised to see the difference between their public debt rating and their KMV score, and there are often very wide discrepancies," observes Hardy. "Understanding how the banker views your credit can give some insight into the profitability decision the lender has to make." —T.R.