Nature and Structure of the U.S. Mortgage Business, 1940–85

The U.S. mortgage business was destroyed and painfully rebuilt during the Great Depression. Across the country, banks failed. Thousands of homes were foreclosed. New lending almost ceased during 1931–33. Diminished credit availability strangled the economy.

Franklin Roosevelt’s New Dealers realized that reviving credit required a rehabilitation of mortgage lending. Major reforms were enacted with this in mind. Most of these involved either a guarantee or insurance for depositors and lenders. The Federal Deposit Insurance Corporation (FDIC) was instituted to insure customer deposits with commercial banks. The Federal Housing Administration (FHA) and the Veterans Administration (VA) undertook to insure qualifying home mortgages. Then in 1938, the Federal National Mortgage Association (FNMA or Fannie Mae) was created to buy FHA and VA guaranteed mortgages from banks. With the VHA/VA and Fannie Mae in the picture, mortgage originators would not have to hold mortgages to maturity. Government entities would now provide mortgage lenders with both protection against default and a way to sell off loans they didn’t wish to hold.

This might seem like an overabundance of protection to latter day observers. However, the mortgage business turns out to be both riskier and less lucrative than it might at first appear.

Public policy is one reason that mortgage lending is risky. Widespread homeownership is viewed as a bedrock of democratic political stability. Homeownership is seen to encourage many civic virtues; owners tend to care about communities, pay their taxes and generally promote law and order. Meanwhile, the home building industry is a huge employer and an engine of economic growth. Its supply chain reaches all the way back to oil, timber and cement. Its multiplier effects spill over into appliance manufacturing, maintenance and repair, home improvement, real estate marketing and, of course, mortgage financing.

Impressed by homeownership’s benefits, U.S. public policy has consistently supported its expansion. One policy component was of special importance—the 30 year, fixed-rate mortgage loan. This financial tool solves a big problem—how can middle class households afford to buy a home? The answer—by saving for a 20% down payment and then devoting about 30% of their monthly income to service a 30 year loan. This math worked for decades. Homeowners knew what their monthly payment was, knew it wouldn’t change, and could budget accordingly. The millions of suburban housing units that sprouted in the post-WWII era testified to the effectiveness of the “30 year fixed.”

Unfortunately, the 30 year mortgage, so good at protecting the borrower, transfers difficult risks to the lender. Mortgage lenders always bear the risk of default. However, the 30 year mortgage maximizes two other hazards—repayment and interest rate risk. The standard 30 year mortgage imposes no penalty on the borrower for prepaying the loan. When are they most likely to repay? When interest rates have fallen and refinancing is attractive. That unfortunately, is precisely when lenders are seeing revenues falling due to the same interest rate decline. Repayment flexibility for the borrower equals reinvestment risk for the lender.

As for interest rate risk, lenders found this even more dangerous. Mortgage lenders typically fund their 30 year loans with short-term borrowing. Should interest rates rise, their funding costs could go above the fixed rates of their assets. Mortgage lenders then face negative margins, losses, depleted capital and eventual insolvency.

The 30-year, fixed rate mortgage loaded these risks onto an industry whose business model involved high fixed costs and thin margins. The mortgage loan quickly became a commodity; one lender’s 30 year mortgage is essentially the same as another’s. Mortgage lenders became numerous. Competition, if not always fierce, was enough to keep prime mortgage loan spreads (lending rate – funding cost) from widening beyond 1–2%. That margin then had to support a business characterized by “bricks and mortar” and numerous employees. Mortgage lenders needed physical sites to attract depositors and borrowers; these sites needed to look like banks, both to project security and to compete with actual banks. The mortgage process was also work intensive. Applications were complex, information needed to be checked, appraisals obtained and each credit analyzed. A separate department was needed to handle time-consuming loan foreclosures. All of these costs made mortgage origination, in the words of Wall Street analyst Mike McMahon, a “negative cash-flow business.” Upfront costs ate up origination fees at the outset, and were only recovered over the life of the loan; the more lenders originated, the deeper their initial cash flow hole became.⁴

Given this business model’s risks, the web of protections around mortgage lending made more sense. Government insurance gave lenders an option to protect against default. Selling loans to Fannie Mae and Freddie Mac (Government Sponsored Entities, or GSEs) provided the flexibility to manage repayment and interest rate risk. Mortgage originators could decide how much risk to retain by deciding how many loans to sell off.

Mortgage lenders still had to make prudent decisions about how much risk they could bear. Volatile interest rates were the danger. Very high interest rates punished mortgage lenders’ funding costs. Very low rates triggered waves of loan prepayments. The mortgage industry only thrived when interest rates stayed within a predictable band. For the three decades after WWII, interest rates fit this pattern. Mortgage lending flourished. In the late-1970s, however, rates broke out of their historic range. A persistent inflation worsened, driving the Federal Reserve to raise short-term rates all the way to 20% in 1979–80.

This savaged the mortgage industry, especially the Savings and Loan Companies (S&Ls). Trapped in 30 year loan assets carrying 5–7% interest rates, their new funding costs meant painful negative margins. Losses quickly eroded S&L capital. By the end of the 1980s, the S&L industry was devastated. Congress authorized a workout—the Resolution Trust Corporation (RTC) took over many institutions, and sold off the loans and properties for what they could get. This process lasted into the 1990s.

Another mortgage business existed, overshadowed by all the drama in the prime mortgage sector. This business was known as “hard money” lending, and would later be called “subprime.” Many of the lenders were known from their sporting event commercials: Household Finance, Beneficial, and the Money Store. These lenders catered to borrowers who couldn’t qualify for prime mortgages. Their mortgage loans were inherently risky, but long experience taught them how to make money in this sector—require ample down payments, lend against both the house and other collateral, be brutally effective in collecting and foreclosing.

It was a tough business, but it made money. Growth was limited by the stringent criteria used to qualify a subprime borrower. Then, Wall Street developed a product that profoundly changed subprime lending.

Wall Street Develops Collateralized Mortgage Obligations (CMOs)

Mortgage-backed bonds were not created from scratch. Well before Salomon Brothers Lew Ranieri and First Boston’s Larry Fink developed CMOs, GSEs had sold $230 billion of mortgage “pass-through” securities.⁵ Still, investor appetite was somewhat limited. Along with “passing through” mortgage repayments, these securities passed along undiluted prepayment and interest rate risk.

This began to change in 1983. In June of that year First Boston issued the first CMO. This security was designed using a different technique—instead of just passing through repayments, First Boston tranched the repayments into 5 year, 12 year and 30 year obligations.⁶ This was done by aggregating GSE guaranteed mortgages into a pool. The pool was sized to give it a high probability of producing cash flows that could service a collection of 5-, 12- and 30-year year bonds. Prepayment risk was then aggregated into a residual tranche, which was sold off at a risk-adjusted price. This approach gave bond investors more of what they wanted: specific maturities, higher yields than conventional bonds, plus minimal default and prepayment risk.

The First Boston innovation was immediately successful, and the CMO market took off. The mortgage securities market, however, was still dominated by the GSEs. Their guaranteed mortgages were the only collateral investors would accept for CMOs. Wall Street’s investment banks made attempts to secure legislation breaking the GSEs’ stranglehold; these failed to gain any traction. As the 1980s turned into the 1990s, the CMO market seemed to settle into a groove — GSEs produced growing numbers of guaranteed mortgages for Wall Street to pool, tranche, underwrite and trade.

Ironically, it was another government agency, the Residential Trust Corporation (RTC), that destabilized the CMO market. In taking over the failed S&Ls, the RTC inherited a hodge-podge of assets—residential loans, real estate, raw land, commercial loans and junk bonds. To monetize these disparate holdings, the RTC turned to Wall Street. The banks decided to securitize the assets, using them to back new bonds; however, the banks had a problem—cash flows from the RTC assets were even more unpredictable than those from mortgage pools.

Pushed to innovate, the banks came up with a fillip on their tranching technique. Instead of just tranching to produce different maturities, the RTC’s assets would also be tranched into senior and subordinate instruments. Any cash thrown off by the RTC assets would go first to the senior ranking instruments. Only if/when additional flows materialized, would cash flow go to those holding subordinated bonds. The senior and subordinate securities would be rated appropriately by the rating agencies and priced according to risk.

This innovation allowed some senior RTC securities to be rated AAA. The eyes of Wall Street’s banks were opened. Securities collateralized by assets of uncertain quality could be engineered to obtain the highest possible rating.⁷

It did not take Wall Street long to act on this insight. The first candidate spotted was Metropolitan Mortgage of Miami. Metropolitan was a successful hard money lender whose business was starved for capital.

Enter Bear Stearns and Lehman Brothers. Each gave Metropolitan a warehouse line of credit. This financing enabled Metropolitan to expand its lending sufficiently to produce pools of subprime mortgages. The investment banks then bought these pools, applied their tranching technique, took the resulting CMOs through the rating agencies and sold the bonds to investors. Each deal freed up more capital for Metropolitan to expand. A new model was born. Metropolitan’s profits grew along with its mortgage volume. Its Wall Street creditors were soon able to also launch the firm’s IPO. Imitation quickly followed these successes. Banks like Alex Brown and Oppenheimer joined the game, lending to new subprime firms like NovaStar, ContiMortgage, FirstPlus and Greentree Financial.⁸ The subprime mortgage market took off. From 1994–1997, annual subprime originations grew from $35 billion to $125 billion.⁹

There was one dark side to this development. Securitization, the process of converting assets into securities backed by assets, weakened the connection between originator and the mortgage asset. If the originator was not going to hold the mortgage, inevitably it would care less about whether it was repaid. Meanwhile the joys of expanding volume were becoming ever more evident. Weakened commitment to credit quality combined with more intense focus on volume expansion would soon prove a toxic mix.

‘Subprime 1.0’ Temporarily Sobers the Market

Securitization spread beyond subprime mortgages. Soon credit card receivables, student loans and subprime car loans were being turned into investment grade bonds. By the end of 1996 subprime used-car loan securities alone had grown into a $60 billion market.¹⁰

This growth was aided by two other developments: computerized loan applications and credit scoring models. The former streamlined the process of gathering borrower financial data. The latter promised to short cut the credit analysis process. Using predictive metrics based upon past borrower behavior, scoring models seemed to promise a reliable default risk forecast without the hassle of checking borrower references and documents. Fair Isaac Corporation (FICO) produced a model that eventually became the industry standard. Putting automated applications and scoring models together allowed asset-backed loan originators to process greater volumes without having to add much staff.

Unfortunately, automated applications and scoring models only reinforced the trend toward weaker due diligence. Tempted by easier credit decisions and quick loan sales, originators saw less and less reason to care about the data they were feeding into their scoring models.

Bad results were not long in coming. They struck first in the subprime car loan market. By the end of 1997 the number of subprime used-car financing companies had grown from about 25 to 170. To get their bonds sold, these firms gave undertakings to buy back any loans that quickly failed. Unfortunately, the loosened underwriting standards resulted in many more defaults than the credit models predicted. Large loan losses ensued. In January 1997, Mercury Finance restated $90 million of prior year profits, and disclosed $61 million in loan defaults. Two other subprime lenders shortly followed suit and filed for bankruptcy. By the following year most subprime auto lenders were gone—merged or bankrupted out of business.¹¹

This debacle did not immediately spill over into the subprime mortgage market. At the end of 1997, 40% of new originations were of the subprime variety.¹² Many originators were showing spectacular results. Their financials were often “juiced” by the use of “gain-on-sale” accounting. Major players began bidding to acquire subprime lenders. First Union bought the Money Store for $2.1 billion early in 1998.¹³ Subprime lenders understood the game—spiffy reported results could quickly turn into a nifty acquisition price.

The subprime market hit the wall in August 1998. This began with Russia defaulting on its sovereign debts. Shortly thereafter, a crisis engulfed the giant hedge fund, Long-Term Capital Management (LTCM). Only a rescue orchestrated by the Federal Reserve staved off a bank panic. In the wake of these events, investor appetite for subprime mortgage-backed securities disappeared. Originators could no longer use new volume to mask losses on loan buybacks or their problems with overstated accounting.

Subprime lenders collapsed en masse. United Companies Financial, Southern Pacific Funding, FirstPlus, Amresco, ContiMortgage and others vanished. Investors were left with bonds backed only by suspect mortgages; the originators were no longer around to honor a buyback pledge.

As “Subprime 1.0” played out, Angelo Mozilo felt a sense of satisfaction. By pursuing a purposeful strategy in the prime mortgage sector, Countrywide had avoided disaster and still grown beyond its wildest hopes.

Countrywide’s Strategy in the 1990s

Angelo Mozilo attacked the challenges of the mortgage business with gusto. Part of his approach involved changing the industry’s cost structure. Mozilo was one of the first “quality” mortgage lenders to use independent brokers. This allowed Countrywide to limit its in-house staff. Countrywide was also early in using automated loan applications and computer scoring. And Countrywide was nimble in terms of product development. When Mozilo detected other lenders pushing some new, customer-friendly product, Countrywide was quick to respond.

By far, Mozilo’s most important strategy was to grow fast, far and wide. The company was called Countrywide for a reason. Mozilo set up offices throughout the nation. By 2003 Countrywide had 500 offices generating $400 billion in annual mortgages.¹⁴ Countrywide’s national reach gave it important economies of scale. This transformed Countrywide’s business model to one with the lowest unit costs in the industry.

Bargaining power was another benefit produced by Countrywide’s operating scale. Early in the 1990s Mozilo used his leverage to extract a concession from Fannie Mae. The GSE typically charged a .23% fee on mortgages it purchased and guaranteed. Jim Johnson, Fannie’s CEO, was eager to ramp up the GSE’s mortgage securitization and retention businesses. Mozilo convinced Johnson to cut Countrywide’s fee to .13%. In return for this 40% discount, Countrywide became Fannie Mae’s largest single supplier of mortgages.

Over and above a low-cost business model and a preferential Fannie Mae relationship, Mozilo and Countrywide were also blessed with good timing. Countrywide’s expansion coincided with a huge boom in mortgage refinancing. Countrywide helped lead the boom, widely advertising its innovative product, the ‘home equity loan.” By 1994, refinancing accounted for 75% of the mortgages originated by Countrywide.¹⁵

The last piece of Countrywide’s strategy was strong financial control. Wall Street analysts came to extol its disciplined culture. This reputation was earned less by Angelo Mozilo than by his long-time partner, David Loeb. Whereas Mozilo never saw a market he didn’t want to dominate, Loeb confined himself to assuring that Countrywide’s sprawling operation maintained good practices.

Under Loeb’s tutelage, Countrywide largely sat out the first subprime mortgage boom. Mozilo did launch a subprime operation in 1997. Dubbed Full Spectrum, it made first mortgage and home equity loans to borrowers who couldn’t qualify for a GSE-eligible mortgage. Loeb and his colleague, Stanford Kurland felt that sound practices were needed if subprime opportunities were to be distinguished from likely defaults. Loeb and Kurland issued three rules to govern Countrywide’s foray into subprime:

1. All subprime loans, including residuals, must be sold. No portion of a mortgage or mortgage pool is to be retained on Countrywide’s books.
2. Subprime borrowers had to make a 20% down payment or obtain mortgage insurance covering 20% of the loan.
3. Countrywide could not offer any subprime loan product with a higher probability of default than products offered by the FHA or VA.¹⁶

These guidelines insured that Full Spectrum got off to a careful start. Full Spectrum did only $140 million in subprime mortgage origination in 1998, its first full year of operation.¹⁷

When other subprime lenders folded in the second half of the year, Countrywide’s caution seemed well founded. However, the firm didn’t count on subprime origination rebounding so quickly.

AmeriQuest Launches a Subprime ‘Race to the Bottom’

Roland Arnall knew the subprime mortgage market from the ground up. Arnall had launched Long Beach Savings & Loan, pioneered its use of independent brokers, and became one of the first sellers of subprime mortgages to Wall Street. Eventually Long Beach was bought by Washington Mutual. By then, Arnall had launched a new firm, Ameriquest Capital Corporation (ACC).¹⁸

ACC applied many features of Countrywide’s model to the subprime market. It pursued scale, developing more than a dozen subsidiaries specializing in different subprime sectors.

ACC didn’t have to be quite so concerned about cost management because the margins on its mortgages were much higher than for prime mortgages. Whereas prime mortgages might produce margins of 3–5% of loan value, subprime margins could reach 15%.¹⁹ With margins like that ACC could afford ample sales commissions and splashy advertising.

To grow the mortgage pipeline, AmeriQuest borrowed techniques from telemarketing. AmeriQuest sales people didn’t wait for customers to walk through the door. Each month there was a quota of “cold calls” to be met. Sales staff that didn’t close enough new loans were quickly replaced. Why didn’t such aggressive sales tactics turn off ACC’s customers? This didn’t happen because ACC’s salespersons were selling products that sounded “too good to resist.” These products, however, always had a catch.

ACC’s key product, the “2/28 loan” was a good example. The 2/28 offered customers a low fixed rate for two years; then the rate adjusted to a high and variable rate. The real catch, however, was a big prepayment penalty located at the end of year 2. Seduced by the prospect of “getting home ownership at an affordable rate,” clients would wake up in year 3 finding themselves trapped in expensive financing.²⁰

Ameriquest developed many such products. There were “high loan-to-value” mortgages, zero down payment loans, no documentation loans and “interest only” loans. All came with big points, fees, and fine print catches.²¹

Perhaps the most toxic product was ACC’s “pay-option adjustable rate mort-mortgage” (ARM). With this product, customers got to pick an initial interest rate from a menu. In most cases, the option didn’t cover the real interest due on the loan. This shortfall would then be “capitalized,” increasing the loan principal. When the loan’s principal reached 115% of the initial amount, the “pay option” fell away. Borrowers then had to pay the full interest rate on higher loan principal. Monthly payments would escalate dramatically; often this set the stage for having to refinance or, more likely, sell the home.²²

Inside the ACC pressure cooker, the sales force came to see the benefits of cheating. How do you get customers to buy something that won’t be good for them in two years? Don’t tell them about the fine print. Much later, audits of ACC’s books would reveal numerous and persistent instances of fraud.²³

Why wasn’t ACC more worried about its customers defaulting? Because Wall Street banks were buying everything it could produce. In the new low interest environment, bank clients were screaming for yield and were not too particular about from where it came. The banks especially loved the pay-option ARM. Because of its inherent risk, it could be bought from AmeriQuest at a cheaper price. After packaging and securing high investment grade ratings, the resulting CMO could be sold at a price close to high quality corporate or GSE bonds. The resulting margins for Bear Stearns, Lehman, Merrill, etc., were very good. As with ACC, the banks thought default risk was something for the next buyer to worry about.

AmeriQuest’s results with this business model were spectacular. Starting in 2002, AmeriQuest’s lending grew twelve-fold in three years. Annual mortgage origination would surpass $80 billion, leading to profits of close to $3 billion. This growth would make AmeriQuest the largest subprime lender in the country.²⁴

By launching this “race to the bottom,” AmeriQuest had indeed changed the game. It was also clearly outearning Countrywide. The question now was whether Angelo Mozilo’s Countrywide would join that race.

Mozilo Reconsiders Countrywide’s Subprime Strategy

Angelo Mozilo was deeply conflicted about AmeriQuest and subprime lending in general. Periodically he told friends something along the lines of subprime lenders all being crooks who fleeced unsuspecting customers with toxic products.²⁵ Yet, he increasingly felt that Countrywide could not settle for a minor subprime role. AmeriQuest was showing that subprime was too big and too profitable to be ignored.

One piece of Countrywide’s strategy had already fallen into place. Mozilo had urged Fannie Mae to approve the purchase of subprime loans. Fannie had grudgingly moved in that direction during the 1990s. Now Fannie Mae, under CEO Franklin Raines, opened the door to subprime mortgage purchases.

Given Countrywide’s relationship with Fannie, this open door created an interesting option. Countrywide and Fannie together could try to define what constituted the “quality” subprime market. If they were discriminating, they might convince investors that the only sure way to buy a quality subprime CMO was to buy what Countrywide and Fannie Mae vetted. This would define for subprime something like the standards Fannie had imposed on prime mortgages.

There were problems with this approach, however. First, given that the rating agencies were stamping AAAs on subprime CMOs without looking at the underlying mortgages, would investors even care about Countrywide/Fannie’s due diligence? Certainly the Wall Street banks wouldn’t encourage their clients to care; they were delighted to have bypassed Fannie and would tell clients to “just look at the rating.”

A second problem was the threat posed to Countrywide’s high growth image. Investors and analysts were aware that mortgage growth was concentrating in the subprime sector. Could Countrywide afford to be small and selective there? Analysts covering Countrywide were already asking why the firm did not have a bigger subprime presence.²⁶

Mozilo sometimes thought about going on the offensive. He and Raines could take a high profile in questioning and even denouncing AmeriQuest’s predatory products, and seek new laws to curtail abuse. They could also demand strong enforcement of existing rules. Here the political lay of the land didn’t appear promising. The Wall Street and subprime lobbies were doing a good job of stymieing Federal predatory lending legislation. There was some interest in new laws at the state level. Even there, however, it would be a fight. Fannie’s enemies would accuse it of discouraging “affordable housing” in the interest of maintaining its monopoly.

What to do about growth was another problem. Was subprime just another bubble, a Subprime 2.0 in the making? Isn’t that what “races to the bottom” are all about? Should Countrywide just wait it out? Mozilo pondered this question in the light of current trends in the prime mortgage market and the possibilities for diversification. Attachment 3 provides information on new prime mortgage origination, 1999–2002.

Or, should Countrywide go all in on subprime? Mozilo thought about affordable housing and all the products AmeriQuest was hawking. Countrywide could craft those products, and maybe better ones. If necessary, Countrywide could develop 100 affordable housing products to sell through its nationwide network.

Angelo Mozilo picked up the draft speech and a pen. Countrywide was the nation’s mortgage leader. Which way should it lead?

Author’s Note

This case challenges students to consider the dilemmas posed by a “race to the bottom.” Such races are occasionally commented on but seldom analyzed in depth. Races to the bottom typically involve some form of “cheating.” The “cheating” can be illegal behavior or the exploitation of lax standards. A competitor builds its business model around this cheating advantage. Other competitors must decide whether to follow suit. Should they do so, the First Mover may restore its advantage via new forms of cheating. Competitors then face another need to respond. The result can be a spiral to ever lower standards, a “race to the bottom.”

Anthony Mozilo, CEO of Countrywide, faces such a challenge. Students must stand in Mozilo’s shoes and decide whether to follow AmeriQuest downward. The case allows Mozilo to ruminate on alternatives. It also provides history about earlier problems in the subprime market. Students should take pains not to engage in 20/20 hindsight. Instead they must work up alternative strategies, drawing on the facts available to Mozilo at the time. Mozilo’s upcoming speech affords an opportunity to outline Countrywide’s future path and comment on market trends. Students working on the case should produce a speech outline laying out conclusions on these matters.

Some of Mozilo’s problem involves the inherent challenges to sustaining margins and growth in the mortgage business. The case provides considerable background on these matters. Students should incorporate these structural fundamentals into their thinking about alternative strategies available to Countrywide in 2003.

Both All the Devils Are Here (Bethany McLean and Joe Nocera) and Reckless Endangerment (Gretchen Morgenson and Joshua Rosner) provide detailed accounts of Countrywide’s rise to prominence. An account of Mozilo’s speech is found in Reckless Endangerment, which also describes Countrywide’s relationship with Fannie Mae. All the Devils Are Here gives a good description of Subprime 1.0, and of Roland Arnall’s tactics at Long Beach Savings and AmeriQuest.

Attachment 2, Mozilo’s speech outline, is a Historical Recreation. It is also a key document for the case. No such document has been made public by Countrywide Public Affairs. This text, however, is consistent with published reports of Stanford Kurland’s approach to controls. A detailed account of Kurland’s perspectives appears in All the Devils Are Here. Attachment 2 draws on this account to imagine what Kurland might have said if asked to comment on chasing AmeriQuest with ever-more-problematic loans. The speech outline thus becomes a counterpoint to Mozilo’s well-reported appetite for making Countrywide bigger. Taken together with the case’s account of Countrywide’s business strategy, the text gives students a variety of options to consider.

This case is situated near the front of the financial crisis cases because the generation of toxic mortgage products was at the heart of the crisis. It is no coincidence that the chapter discussing these events in All the Devils Are Here is entitled “Ground Zero, Baby.” Mozilo’s decisions at this moment turned out to be highly consequential. For students, their thinking about this case’s solution will likely influence their approach to subsequent cases.