This content is from: Home

Plastic meltdown

Providian mastered credit card marketing, then got hammered by the slowing economy - and by its own overconfidence.

Providian mastered credit card marketing, then got hammered by the slowing economy - and by its own overconfidence.

By Gregory J. Millman
December 2001
Institutional Investor Magazine

Executives of Providian Financial Corp. could at times act more like devil-may-care dot-commers than risk-savvy bankers. They swaggered through the halls of their San Francisco headquarters with an aura of invincibility worthy of the most self-assured Silicon Valley entrepreneurs.

And they made money, lots of it, feasting on loans to borrowers whom other banks regarded as marginal. Built and run by longtime CEO Shailesh Mehta, a fastidious engineer who prided himself on his ability to design extremely accurate and highly predictive mathematical models, Providian thumbed its nose at traditional notions of business cycles and credit quality, approving credit cards for people who otherwise couldn't get them.

Even when the economy was booming, the company suffered outsize loan losses, but blessed with a seemingly magical combination of computerized marketing and modeling techniques, Providian's executives shrugged off the losses, and so did investors. After all, profits had mushroomed since Providian came out of nowhere in the 1990s to become one of the ten biggest credit card lenders in the U.S. - and one of the rising stars of the financial services industry. Net earnings jumped by 45 percent annually between 1997 and 2000; card loans outstanding soared from $18.8 billion to $32.2 billion today. In 2000 Providian posted a 39.21 percent return on common equity, 17 points better than Citigroup, on net income of $652 million, or $2.23 per share.

"People at the low end may be a higher risk, but not necessarily a bad risk," said a confident Mehta in an interview with Institutional Investor earlier this year. "You can design a product that meets the customer's needs and still compensates you for the risk."

It sounded good - in theory. There may indeed be no such thing as a bad credit risk, only poorly designed products and pricing. But in the real world, Providian may not be around to prove that axiom.

Providian started 2001 on its usual roll. Despite a charge-off rate of 9.34 percent of managed loans in the first quarter - by far the highest among large credit card banks - Providian reported a 32 percent year-over-year rise in net income, to $231 million. The second quarter brought a 24 percent increase, to $232 million, even as the loss rate jumped to 10.29 percent (which was later revised upward to 10.63 percent). But even before the economic chill that set in after the September 11 terrorist attacks, Providian's glossy image had begun to chip and crack.

In a September 4 conference call with analysts, Mehta warned of higher charge-offs into 2002, causing shares to plunge 22 percent that day, to $30.36. He also forecast that 2001 earnings would still be 17 to 19 percent higher than those of 2000, but the market wasn't impressed - and a far worse reaction was to come. Over the next month Mehta's relatively rosy prediction vaporized. On October 18, in its third-quarter earnings statement, Providian revealed that charge-offs stood at 10.33 percent, double those of its nearest rivals, Capital One Financial Corp. and MBNA Corp. Net income had plummeted 71 percent from 2000's third quarter, to $57.2 million, or 20 cents per share, as the company was forced to double its loan-loss provision, to $996 million. Providian expected per-share income of 10 to 15 cents in the fourth quarter, for a full-year total of $1.87 to $1.92, down by about 14 percent.

There was one other piece of news. Describing himself as "frustrated" and conceding that "we have not delivered the earnings we promised," Mehta stepped aside as chairman of Providian's board and said he would stay on as CEO only until a replacement could be found. The share price, which hit a peak this year of $60.91 on May 22, shed 58 percent on October 19, falling to $5.15. It continued to decline steadily until more bad news hit on November 14, after the market closed: Providian withdrew its previous earnings guidance and announced a plan to sell or securitize $3.9 billion of loans. The stock dropped 22 percent the next day, to $2.87.

Mehta departed on November 26; Providian announced that Joseph Saunders, the 56-year-old head of FleetBoston Financial Corp.'s credit card subsidiary, would replace him. The stock price jumped 16 percent that day, to $3.80, but the company's market capitalization was still an anemic $1 billion, down from $14 billion only five months earlier.

Why Mehta's business model went bad - and why it went bad so quickly - remains something of a mystery. Providian's closest competitor, Falls Church, Virginia-based Capital One, shrugged off the post-September 11 economic slump, reporting relatively low charge-off rates and a 35 percent jump in its third-quarter net income, to $165 million, or 75 cents a share (see box, page 109). The most likely scenario, accepted by most Wall Street analysts, is that an overconfident - or overly greedy - Providian erred by continuing to market heavily to risky customers even as the economy soured and other card companies were turning more cautious. Never doubting its proficiency at designing products for any risk profile, Providian grew its loan portfolio by 31 percent, or $7.7 billion, in the 12 months ended September 30. In November federal regulators ordered the company to halt the growth of its subprime lending.

"The most shocking statistic is that their average balance in the subprime market was $1,500. That is double to almost triple Capital One's," says Caren Mayer, a specialty finance company analyst at Banc of America Securities.

Says one insider, speaking on condition of anonymity, "The company basically shot for too much growth at the expense of some other things it should have been assigning more priority to."

Perhaps, in the end, Mehta was just another genius who pushed his insights a bit too far. He has avoided interviews since October 18, when he turned the chairmanship over to veteran Providian board member and Mayfair Capital and Glenview Trust Co. chairman J. David Grissom. Two weeks later Providian engaged Salomon Smith Barney and Goldman, Sachs & Co. to consider "a broad array of financial and strategic alternatives." As often as not, of course, that's synonymous with a sale, but when new CEO Saunders took the helm, he vowed to keep Providian independent.

Some research analysts continue to believe that Providian's competitive advantages in credit analysis and financial mass marketing were genuine. Most bankers were unwilling to take Providian's risks, but few could match the size of its consumer database or the sophistication of its analytical models. "I think Providian really did have a superior platform," says J.P. Morgan Securities credit industry analyst Michael Freudenstein. "I still think they do relative to most of the industry."

More widespread among even formerly bullish analysts, blessed with the benefit of hindsight, is the sentiment that they may have given Mehta a little too much credit. "Their return on assets was double that of the better players in the industry," says disenchanted ABN Amro researcher Robert Napoli, who in January had Providian on a short list of favorite stocks. "You could justify it by the segments they were going after, but generally, when somebody is that much better than everybody else, it's a key risk indicator."

The irony, of course, is that risk is precisely what Mehta and Providian claimed to know better than anyone else in their industry.

To revolutionize the U.S. credit card business, Shailesh Mehta traveled far.

Born in 1949 in Bombay, Mehta earned a bachelor's degree in mechanical engineering from the Indian Institute of Technology, then went to Cleveland in the early 1970s to earn his master's and Ph.D. in operations research and computer science at Case Western Reserve University.

In 1973, while still a student, Mehta took on a consulting assignment for a local bank, AmeriTrust Co., which was trying to find a way to save gasoline expenses during the Arab oil-producing countries' embargo on exports. The bank's management wanted to reduce the mileage on trucks that picked up checks from branches and delivered them to the Federal Reserve Bank of Cleveland. Mehta crunched the numbers and came to the counterintuitive conclusion that economizing on transportation would be a mistake because of the money that would be lost by delayed check-clearing.

Pleased, the bank offered Mehta a full-time job. He stayed 13 years, running back-office systems and rising to executive vice president. In 1986 he left to become chief operations officer at Capital Holding Corp., an insurance company in Louisville, Kentucky, that was seeking to diversify its financial services offerings. Capital, which would rename itself Providian Corp. and sell its insurance lines to Aegon of the Netherlands, owned a small bank that would become the nucleus of today's Providian Financial.

The credit card industry had been dominated by big commercial banks like Citibank and Bank of America that had controlled the business since it went mass-market in the 1960s. The banks mainly marketed their cards through advertising and direct mail aimed at their own customers and others living in regions served by their branch networks. Banks typically invited consumers to send in applications that were evaluated according to a standard scoring system that used age, income, homeownership or rental status and other factors to determine creditworthiness.

In the late 1970s Citi broke the old marketing pattern by buying mailing lists and flooding the country with solicitations. Although Citibank went overboard - some of its offers famously went to children and pets - the strategy made Citi the biggest card issuer in the world; to this day cards are one of its biggest revenue generators, bringing in more than $3 billion out of Citigroup's $20 billion in the third quarter.

Nearly every major commercial bank stepped up its direct mailings in response to Citibank's success, but there was little difference in their offerings. Except in states where usury laws limited interest rates, bank credit card interest generally ranged between 18 and 20 percent. When the record interest rates of 1979-'80 erased their margins, banks raised credit card rates above 20 percent wherever they could and began to impose annual fees.

By the mid-1980s the rigidity of bank pricing attracted new competitors. Sears, Roebuck and Co. introduced its Discover card - now part of Morgan Stanley - with an annual rebate, something no bank had ever tried. And American Express Co., then a marketer of travel and entertainment cards without a revolving credit line, rolled out the Optima card, undercutting most banks with a 15 percent interest rate.

Even then, competition among the big banks remained relatively restrained. "When we started to look at the credit card business in 1987, we didn't know anything about it," says Capital One co-founder, chairman and CEO Richard Fairbank. "But we did know that every one of the top ten players were charging the same interest rate - 19.8 percent. Something was wrong with that. We set out to lower the price and democratize credit."

Capital One, Providian, MBNA and First USA emerged as a new breed of companies - called monolines - that focused only on the credit card business. (Some of these companies were in fact spin-offs of established banks.) The monolines sold most of their cards by direct mail, but they took a more aggressive, targeted approach than the banks. Rather than wait for customers to come to them, the monolines relied heavily on demographics and credit scoring systems to seek out and fashion pitches to every potentially profitable applicant. These new companies pioneered the use of low interest rates as teasers to attract customers and their balances away from established competitors. And they scrutinized account activity, watching for credit problems and identifying opportunities to pile more credit on good customers.

The innovative Mehta and his team at Providian quickly began to discover profitable nuances to the credit card business. "Most people underwrote credit based [only] on demographics," he recalled earlier this year. "We found that the credit behavior of two people with the same demographics could be very different. The challenge was to discover the variables and build models."

Mehta came up with models that placed every U.S. consumer in one of 17 risk categories. If the old, blanket approach to mass mailings was equivalent to carpet bombing, then Providian engaged in surgical strikes, tailoring credit card offers to different kinds of customers and in ways that would maximize its own profits. The ability to customize has become especially valuable in an industry that generated an estimated 4 billion mail solicitations in 2001, up from 1.1 billion in 1990, while response rates have fallen from 1.5 percent to 0.5 percent.

The monoline model worked. In 1990 the ten biggest card issuers accounted for 51 percent of the $118 billion in MasterCard and Visa account debt outstanding, and only one of them was a monoline: MBNA, which ranked third behind Citicorp and Chase Manhattan Corp. By 2000 the top ten controlled 82 percent of a $471 billion market, according to "The Nilson Report," an industry newsletter based in Oxnard, California. Citigroup headed the list, and J.P. Morgan Chase & Co. came in fourth, but the rest of the top six were monolines: MBNA, First USA, Providian and Capital One. (First USA was acquired by Bank One Corp. in 1997 and is just recovering from credit losses in 1999 that forced longtime Bank One CEO John McCoy into retirement.)

Capital One and Providian epitomized the aggressive and intelligent use of technology and computerized modeling. UBS Warburg card industry analyst John McDonald says that the two set themselves apart "by using information about customers to be more creative on the marketing end or smarter on the pricing end than everyone else."

Providian set itself apart in another way - it aggressively courted riskier customers. After a couple of years in the 1980s when Providian primarily went after the higher end of the credit spectrum, the company decided to move downscale, focusing on people it labeled "unbanked" or "underserved." Today about 70 percent of Providian's card loans and losses are concentrated in the riskier subprime segments.

"Around 1989-'90 we found we were declining prospects who were not bad credits - we just didn't have enough information in our files to make an informed decision," David Petrini, Providian's vice chairman overseeing finance, technology and administration, told Institutional Investor earlier this year. "This was a large market that included people new to credit, people with a life event like a divorce who hadn't developed credit on their own and people whose credit history was less than stellar."

Providian's enthusiastic pursuit of the subprime market became its hallmark. In 1999 the company's tactics prompted investigations by the City of San Francisco and the State of Connecticut into some allegedly questionable sales and collection practices. Related customer and shareholder class-action suits were filed; Providian agreed to settlements and restitutions as high as $300 million.

Analysts and investors excused those incidents because they didn't hurt Providian's bottom line and because the company responded contructively, increasing its charitable contributions and appointing former banking regulator Konrad Alt as chief public policy officer. But the legal problems were an indication of what UBS Warburg's McDonald calls "overaggressive marketing practices" that would come back to haunt Providian. "They overexposed themselves to the riskiest segments of the subprime market," says McDonald.

Providian's delinquency and loss numbers came to reflect that. In the first and second quarters of this year, when an American Bankers Association survey showed that 4.13 percent of all credit card loans were at least 30 days past due, MBNA and Capital One both hovered between 4 percent and 5 percent. But Providian hit 8.04 percent in the second quarter - up from 7.64 percent in the first quarter and a portent of higher charge-offs to come. A high delinquency rate was fine as long as the company continued to make big profits. And Providian's competitive analysis seemed to confirm its strategy. As recently as May its investor relations people were brandishing charts illustrating how the profit per account, at $56, far outstripped those of MBNA ($33), Citi ($32) and Capital One ($16).

The attitude that produced those profits was that "losses are never too high," as chief risk management officer James Redmond explained in an interview earlier this year. "There is no such thing as a level beyond which you won't lend. The point isn't how high losses are but are they predictable and can you price for them."

But by September of this year, with the stock market bubble long since collapsed and the U.S. economy flirting with recession, Providian's prices weren't covering losses like they used to. Mehta could no longer seem to get his earnings and charge-off projections straight, and investors took out their frustrations on the stock. The price-earnings multiple, in the high 20s earlier this year, fell well below 2 in November, while MBNA's was at 18 and Capital One's at 19.

The postmortems about what went wrong at Providian should continue for some time, but there are damning clues in some of the details that the company disclosed about third-quarter loan performance. The company divides its card loans into three broad categories: platinum, the most creditworthy and highest-balance accounts; and two predominantly subprime segments, known as middle and standard.

For the third quarter, the platinum category had an average balance of $4,331, total loans of $9.6 billion and a loss rate of 7.8 percent. In the middle portfolio the average balance was $1,950, total loans were $12.7 billion, and the loss rate was 9.9 percent.

The average standard account balance was $1,496 - far higher than the $500 to $1,000 in competitors' subprime card portfolios. And on Providian's $9.4 billion in standard loans, the third-quarter loss rate was 13.9 percent - far worse than the company's 10.33 percent blended rate, which itself had raised analysts' alarms.

J.P. Morgan Securities analyst Freudenstein, who believes standard card losses could climb next year to a whopping 24.5 percent, says the charge-offs were simply too large. With the economy weakening, Providian didn't pull back from the subprime market in time, he says. The final straw may have been worse-than-expected losses in the platinum accounts, which Freudenstein attributes to a spike in bankruptcy filings. He says that proposed federal legislation designed to make such filings more difficult motivated many of those heavily indebted cardholders to seek bankruptcy protection while it was still an easy option.

Providian not only offered the riskiest customers more credit than they could get elsewhere, but also provided it at a lower price. "They capped out their APR [annual percentage rate] to the highest-risk customers at 23.9 percent. If you look at subprime cards from other issuers, you see rates closer to 30 percent," UBS Warburg's McDonald says.

Higher-than-expected losses. Pricing that was a little too low. Shailesh Mehta now has plenty of time on his hands to do the math.

Capital's ideas

Not so long ago Providian Financial Corp. and Capital One Financial Corp. were twin stars of technology-intensive credit card marketing. As Providian's then-CEO, Shailesh Mehta, said in an interview in the spring, "If anyone comes close to our skill set, it is Capital One."

At about the same time, Capital One's investor relations director, Paul Paquin, complimented his rival but presciently pointed to a difference: "Providian is a good, information-based strategy player. They approach the process from a technology point of view much the way we do. But we are much more risk-averse."

That risk aversity helped to propel earnings growth at the Falls Church, Virginia-based company 35 percent in the third quarter, to $165.3 million, and 36 percent for the first nine months of the year, to $464.3 million. Capital One entered 2001 with 33.8 million consumer loan accounts; by September 30 that figure had jumped 19 percent, to 40.1 million. Loan balances climbed $9 billion, or 31 percent, to $38.5 billion, but the charge-off rate remained low at 3.92 percent, "the best credit performance of any of the major credit card institutions," boasted Capital One chairman and CEO Richard Fairbank.

Capital One got its start when Fairbank was a consultant with New York-based Strategic Planning Associates in the 1980s. He and another consultant at the firm - Nigel Morris, who is now president and COO of Capital One - wrote a business plan incorporating their ideas about computerized credit card marketing. They met with executives from 20 of the largest U.S. banks in hopes of raising money to back their vision, but only one, Signet Bank of Richmond, Virginia, agreed to back them. They launched Capital One in 1987, and seven years later the company went public. (Signet became part of what is now Wachovia Corp.)

Capital One builds its business through trial and error on a massive scale. It designs and tests customized card offers by the thousands - last year it tested 45,000, an increase from 36,000 in 1999 and 1,130 in 1992, when the strategy was in its infancy. If consumers respond positively to a test, Capital One directs a marketing campaign toward more people who it believes are likely to do the same.

Says Fairbank, "We built an information-based company that can create an infinite number of products and put them through massive scientific testing to get down to a market of one, or mass customization."

The company has occasionally been tempted to apply its marketing savvy to other sectors. In 1997 it made a foray into the cellular phone business, reasoning that its methods of identifying card customers would carry over. But in March Capital One pulled the plug on a joint wireless services venture with Sprint PCS Group, because the intense competition in that business had eroded profit margins to "significantly lower than our hurdle rates," says Paquin.

In 1998 Capital One moved in a more banklike direction, entering the automobile lending business. It underscored its seriousness in September when it agreed to acquire online auto lender PeopleFirst Finance for $148 million in stock. "Simple and transaction-based financial products are growing faster than credit cards. The model is very extendable," Fairbank asserts.

That may be, but the economy is worsening, Capital One is growing at rates reminiscent of Providian before its crash, and a stock market that tends to value specialized financial companies more highly than diversified ones might frown on Capital One's expansion. Still, the company is not backing away from its 20 percent earnings-growth projection for 2002, and Fairbank and Morris are optimistic enough to be taking all their compensation in options, a majority of them performance-based, for the next two years.

Says UBS Warburg card industry analyst John McDonald: "The $100,000 question is how much of Providian's problems were Providian-specific and how much is reflective of the weakening economy, with Providian feeling the effects first. That question is going to be answered over the next six to 12 months." McDonald is betting on Capital One with a buy recommendation, and he is not alone.