A reader has asked several important questions about liar’s loans that are critical to understanding the causes of the ongoing U.S. crisis. By 2006, half of all loans called “subprime” were also liar’s loans. Roughly one-third of all home loans made in 2006 were liar’s loans. The crisis was originally called a “subprime” crisis, but it was always a liar’s loan crisis. The reader is correct to inquire about causation and moral culpability.
“Dr. Black, are liar’s loans the same as stated income loans? In either case, how do we know whether buyers or loaners put the income for the loan? If most of these reported incomes were entered by borrowers, I would think most of the blame falls on them.”
Yes, “liar’s” loans are what the industry called “stated income” and “alt-a” loans when they were talking among themselves. Income was the primary category that was “stated” – i.e., listed without any verification as to accuracy – in a liar’s loans. Some liar’s loans, however, also “stated” employment, assets, and liabilities. “Stated income” is a euphemism for a liar’s loans, but it is at least honest about its insanity. Readers get it right immediately – they understand that no honest mortgage lender would make loans on this basis. (I expand on this point below.)
“Alt-a” is a bright shining lie. “Alt” is short for “alternative,” where the lie is that the loans are “underwritten” through an “alternative” methodology. True, if not underwriting can be considered an “alternative” to underwriting. Relying on a credit score is not underwriting, particularly in the home lending context. The borrower’s credit score does not tell the lender whether the borrower has the capacity to repay a $600,000 home loan. “A” is an even more blatant lie, it claims that the loan is “A” quality, i.e., “prime.”
Two bright shining lies were used to support the ludicrous claim that liar’s loans were really high credit quality. One, “alt-a” loans were made to entrepreneurs who could not document their income. Nonsense, there is a standard IRS form (4506t) that such a borrower can sign that allows the lender to check the income that the borrower reported to the IRS. (Borrowers have strong incentives not to inflate the income they report to the IRS.) Two, “alt-a” apologists claimed that borrowers really had the income they “stated” but were unwilling to document that income because they were hiding their income from their former spouse and children and/or the IRS. Anyone who has done honest lending will recall that one of the “C’s” an honest lender would insist upon is “character.” A borrower who is fraudulently hiding income from his children and the government is an exceptionally bad credit risk even if his income is real. There was never any evidence that “alt-a” borrowers really had the incomes stated on the loan applications. Because the lenders carefully did not seek to verify the stated income they could not have known that the wealthy deadbeat dads of the U.S. really had hundreds of billions of dollars hidden from their children. As I show below, liar’s loans were so massive that wealthy deadbeat dads and tax evaders could not have been more than a tiny percentage of the recipients of liar’s loans.
The fraud “recipe” for lenders
The reason that accounting control frauds characteristically engage in lending behavior that no honest lender would exhibit was that these perverse practices maximized reported short-term income and the executives’ compensation. There is a four-ingredient fraud “recipe” for lenders.
- Extreme growth through making
- Exceptionally bad loans at a premium yield (very high interest rate) while
- Employing extreme leverage (the lender has vastly more debt than equity), and
- Providing grossly inadequate allowances for future losses inherent in making bad loans
The same recipe maximizes three things: (fictional) short-term reported income, the senior executives’ compensation, and actual losses. Such a recipe can only come with the blessing of the officials that control the bank. When a significant number of frauds following the same fraud “recipe” use the same “ammunition” (liar’s loans) to feed their accounting fraud “weapon,” those frauds will continue to lend into the teeth of a glut of residential real estate. Epidemics of accounting control fraud can hyper-inflate financial bubbles.
We can now see why mortgage lenders that were accounting control frauds made massive amounts of liar’s loans – and greatly increased the number of liar’s loans they made as the FBI and mortgage fraud experts warned that there was an “epidemic” of mortgage fraud and that liar’s loans would cause catastrophic losses. Liar’s loans were the best available “ammunition” for accounting control fraud. They allowed fraudulent lenders to make vast amounts of loans to the uncreditworthy at premium yields. They also allowed the lender to make the loans without a paper trail demonstrating that the lender knew that the borrowers’ incomes were endemically, and severely, overstated. That paper trail would have made it much easier to prosecute the fraudulent lenders.
Liar’s loans were a terrible deal from the borrower’s perspective – they had to pay a premium yield to borrow. Borrowers who took liar’s loans were not typically poor, but they were typically had relatively low levels of financial sophistication. Before we begin a finer level of analysis we should focus on the larger picture.
Liar’s loans not only drove the crisis, they also serve as a superb “natural experiment” that allows us to test many of the most important hypotheses about the causes of the crisis. Liar’s loans are useful in this regard because never mandated that any the lenders make liar’s loans. Instead, the government repeatedly criticized liar’s loans. As I develop below, lenders:
- Knew from the beginning that the liar’s loans would cause (net) catastrophic losses
- Soon learned that the loans were endemically fraudulent
- Could have stopped the endemic fraud, “adverse selection,” and “negative expected value” at any time by engaging in prudent underwriting, but instead
- Rapidly expanded their issuance of liar’s loans after they knew the loans were endemically fraudulent
- Sold the fraudulent liar’s loans through fraudulent “reps and warranties.”
Liar’s loans allow us to test, for example, whether Fannie and Freddie purchased nonprime loans because they were mandated to do so by government mandates or because Fannie and Freddie were accounting control frauds purchasing the nonprime assets because their superior (nominal) yield maximized the controlling officers’ compensation. (Hint: Fannie and Freddie were control frauds.)
More generally, we can ask which of two overall views make sense. Under one view, the unsophisticated but fraudulent borrowers were able to defraud, for the better part of a decade, the sophisticated financial institutions. Indeed, the most sophisticated entities went “all in” after they were repeatedly warned that they were being defrauded. The lenders making the liar’s loans then sold the fraudulent loans, by making fraudulent “reps and warranties” to the (purportedly) most sophisticated financial entities in the world – the major investment banks. Under this view, the lenders are helpless victims of fraudulent borrowers, but become fraudulent sellers of the fraudulent loans to helpless investment banks. I submit that the story is facially absurd.
The alternative view is that the lenders that made liar’s loans and investment banks that purchased them were accounting control frauds. I submit that this view is logically coherent and fits the facts.
If the lenders that specialized in making liar’s loans were innocent victims of rapacious, but financially unsophisticated borrowers, then we would expect to see several observable characteristics. First, we would expect that honest lenders would never make liar’s loans because doing so must cause them to fail. Second, we would see very little, and decreasing, appraisal fraud. Third, we would see sharply declining amounts of liar’s loans after the FBI and MARI warnings of fraud epidemics in 2004 and 2006. Fourth, we would see sharp reductions in the liar’s loans made to borrowers with poor credit histories (i.e., the percentage of subprime loans that are also liar’s loans should quickly reach zero). Fifth, we should see a sharp reduction in liar’s loans that exhibit “layered risk” – other loan characteristics that add to risk such as negative amortization and reduced down payments. Sixth, we should see sharp rises in capital and allowances for loan and lease losses (ALLL) so that the honest lenders could be prepared for the massive losses inherent in making liar’s loans. Seventh, we would see the end of sales of liar’s loans to the secondary market because honest lenders would not sell fraudulent liar’s loans to others and because honest investment banks would not purchase them. Eighth, we would see the investment banks bring aggressive suits against the lenders that sold them liar’s loans under false “reps and warranties.” Each of these characteristics went in the direction that falsifies the “honest lender” theory.
Alternatively, we could look at case studies and investigations of the lenders that made liar’s loans. Honest lenders would support employees who insisted on prudent underwriting and discipline employees who made bad loans. They would monitor lending operations and adjust them to prevent any developing problems. The case studies and investigations of lenders specializing in making large numbers of liar’s loan show a consistent pattern inconsistent with the honest lender theory.
We could also investigate incentive structures. CEOs’ primary function is the creation of incentive structures. Honest CEOs would create virtuous incentive structures. Dishonest lenders would create perverse incentive structures. Again, we observe the creation – and maintenance despite warnings of endemic fraud – of intense, perverse incentives at lenders and investment banks that issued and purchased large amounts of liar’s loans.
Focusing on the perverse incentive structures of mortgage bankers offers a clear example for most people of how to test the rival hypotheses. Because the overwhelming majority of liar’s loans were sold to the secondary market the means to optimize the sale of fraudulent loans was to combine high (nominal) yield with apparent reduced risk. (These two characteristics are supposed to be antagonistic under the efficient market hypothesis, but traders have long realized that the efficient market hypothesis is false.) The way to reduce apparent risk was to lower the reported loan-to-value (LTV) and debt-to-income ratios. The lender could reduce the reported LTV by inflating the appraisal and reduce the debt-to-income ratio by inflating the borrowers’ income. Fraudulent lenders created compensation systems for loan brokers that paid them very large fees if they simultaneously charged a premium yield and low reported LTV and debt-to-income ratios. Ask yourself whether loan brokers would typically ignore their financial interests and leave it to the borrower to game the ratios sufficiently to maximize the broker’s fees. Only the lender and its agents (including the loan broker) can inflate large numbers of appraisals, so we know that the endemic appraisal fraud was generated by the lenders and their agents. Why would they not take the lead in inflating the borrower’s stated income?
Recall that the reader’s position on morality was:
“If most of these reported incomes were entered by borrowers, I would think most of the blame falls on them.”
I believe the moral issue is important and complex. I do not agree that if the lender was an accounting fraud whose controlling officers created perverse incentives to produce endemic fraud those officers should escape “most of the blame” if they are successful in inducing the borrowers to “enter” the false information on income on the loan application. I agree that the borrowers who sign a loan application they know contains inflated income bear some moral culpability. I agree that that a loan applicant who seeks guidance from a loan broker as to how much to inflate his stated income bears even more culpability.
There is a continuum. Many fraudulent lenders and brokers falsified the loan application directly. This was sometimes called “arts and crafts” weekends. Those borrowers had little or no culpability. Other borrowers were induced by fraudulent representations by the lenders and their agents. Note how pernicious it is when the lender inflates the appraisal. The loan broker or office may tell the victim that the house is worth far more than the purchase price and that it is therefore safe to purchase the home because the borrower can always sell the home for a profit even if he has trouble making the payments. The borrowers in this situation were often victims of predation and have little if any moral culpability.
Another category with some, but limited culpability are the financially unsophisticated (and/or those not fully literate in English) who are confronted by complex loan forms and told to put in a particular income by the loan broker or officer who explains that the paperwork is meaningless and that putting in the income figure the broker or officer suggests will speed up the loan processing time. As the famous Yale experiments on obedience to authority demonstrated, most people will defer to the “experts” and comply with their requests.
Overall, moral culpability must take into account differential power. The CEOs created the perverse incentives that produced the “echo” fraud epidemics among their loan officers, loan brokers, appraisers, and some of their loan customers. It would be bizarre to place primary blame on those they manipulated for giving in to the perverse incentives.
Accounting control fraud drove the U.S. crisis
The nonprime lenders followed the classic recipe for maximizing accounting control fraud
Nonprime mortgage lenders followed the classic accounting control fraud recipe in the current crisis. Growth was extreme.
In summary, the bank in our analysis pursued an aggressive expansion strategy relying heavily on broker originations and low-documentation loans in particular. The strategy allowed the bank to grow at an annualized rate of over 50% from 2004 to 2006. Such a business model is typical among the major players that enjoyed the fastest growth during the housing market boom and incurred the heaviest losses during the downturn (Jiang, Aiko & Vylacil 2009: 9).
Loan standards collapsed. Cutter (2009), a managing partner of Warburg Pincus, explains:
In fact, by 2006 and early 2007 everyone thought we were headed to a cliff, but no one knew when or what the triggering mechanism would be. The capital market experts I was listening to all thought the banks were going crazy, and that the terms of major loans being offered by the banks were nuttiness of epic proportions.
Leverage was exceptional. Unregulated nonprime lenders had no meaningful capital rules. Indeed, they had no capital – they were insolvent on any real economic basis because of the large, inherent losses on the fraudulent loans that were always in their “pipeline.”
Honest mortgage lenders would not make liar’s loans because such loans maximize adverse selection and create a negative expected value for the lender. Assuming away these facts solely for the purpose of this discussion, an honest mortgage lender making liar’s loans would establish record high allowances for loan and lease losses (ALLL) pursuant to the requirements of generally accepted accounting principles (GAAP). As these liar’s loans became far riskier (due to “layered” risk) GAAP required the ALLL provisions to grow substantially. The nonprime lenders routinely violated GAAP and did the opposite. “The industry’s reserves-to-loan ratio has been setting new record lows for the past four years” (A.M. Best 2006: 3). The ratio fell to 1.21 percent as of September 30, 2005 (Id.: 4-5). Later, “loan loss reserves are down to levels not seen since 1985” (roughly one percent) (A.M. Best 2007: 1). A.M. Best noted that these inadequate loss reserves in 1985 led to banking and S&L crises. In 2009, IMF estimated losses on U.S. originated assets of $2.7 trillion (IMF 2009: 35 Table 1.3) (roughly 30 times larger than bank loss reserves). U.S. securities registrants must file financial statements that comport with GAAP. The intentional failure to do so, as to any accounting matter that is “material,” constitutes federal securities fraud – which is a felony.
Liar’s loans were endemically fraudulent
Normal underwriting easily detects and prevents this fraud – which is why credit losses on traditional residential mortgages were minimal for nearly 50 years. Fraudulent lenders designed liar’s loans to remove these underwriting protections against fraud. Their fraud-friendly design was so successful that their own industry anti-fraud experts (MARI) denounced their product as “an open invitation to fraudsters” and lived down to the term the industry used behind closed doors to describe them – “liar’s loans” because they were pervasively fraudulent. MARI reported a fraud incidence in liar’s loans of 90 percent.
Lenders and their agents put the lies in liar’s loans
The officers controlling the lying lenders designed and implemented the perverse incentives that produced the intended “echo” fraud epidemics among loan brokers, loan officers, appraisers – and some borrowers. The combination of liar’s loans and the echo epidemics helped the controlling officers produce the first two ingredients of the lender fraud recipe – rapid growth at premium yields. The officers that controlled the lying lenders wanted to be able to make loans to the uncreditworthy – as long as they could do so at a premium yield. Liar’s loans made it easy to do both – and prevented the creation of an incriminating underwriting paper trail documenting that the lender knew the information on the loan application was false when it made the loan. The resultant deniability is implausible to anyone that understands fraud mechanisms, but it does fool the credulous.
Liar’s loans were overwhelmingly sold by the issuers and the fee the issuer could obtain was increased if the issuer could make the loan appear to be less risky. There were two key ratios that could be fraudulently manipulated to make the loan appear to be less risky. By inflating the appraisal, the issuer could make the reported loan-to-value (LTV) ratio appear lower. By inflating the borrower’s income the issuer could make the debt-to-income ratio appear lower. Appraisal fraud, which inherently comes from the lenders and their agents, was the key to producing a more desirable LTV while liar’s loans were the perfect device to inflate the borrower’s income. Because the loan broker’s fee could be much larger with a reduced debt-to-income ratio, loan brokers had a powerful, perverse incentive to inflate the borrower’s income on the loan application. Investigations, to date, have confirmed this logic. The fraudulent nonprime lenders and brokers typically initiated, directed, and sometimes even directly created the lies on the liar’s loans. The testimony of Thomas J. Miller (Miller, 2007), Attorney General of Iowa, at a 2007 Federal Reserve Board hearing began by describing the Gresham’s dynamic that the interaction of accounting control fraud and modern executive compensation produces:
Over the last several years, the subprime market has created a race to the bottom in which unethical actors have been handsomely rewarded for their misdeeds and ethical actors have lost market share…. The market incentives rewarded irresponsible lending and made it more difficult for responsible lenders to compete. Strong regulations will create an even playing field in which ethical actors are no longer punished. (p. 3)
Despite the well documented performance struggles of 2006 vintage loans, originators continued to use products with the same characteristics in 2007. (note 2)
[Many originators invent] non-existent occupations or income sources, or simply inflat[e] income totals to support loan applications. A review of 100 stated income loans by one lender found that a shocking 90% of the applications overstated income by 5% or more and almost 60% overstated income by more than 50%. Importantly, our investigations have found that most stated income fraud occurs at the suggestion and direction of the loan originator, not the consumer. (p. 10)
A small sample review of nonprime loan files by Fitch, the smallest of the three large rating agencies, adds support for the view that fraud became endemic in nonprime mortgage lending. Fitch’s analysts conducted an independent analysis of these files with the benefit of the full origination and servicing files.
The result of the analysis was disconcerting at best, as there was the appearance of fraud or misrepresentation in almost every file.
[F]raud was not only present, but, in most cases, could have been identified with adequate underwriting, quality control and fraud prevention tools prior to the loan funding. Fitch believes that this targeted sampling of files was sufficient to determine that inadequate underwriting controls and, therefore, fraud is a factor in the defaults and losses on recent vintage pools. (Pendley, Costello, & Kelsch, 2007, p. 4)
Fitch did not investigate these loans. It simply reviewed the loan files and servicing files to identify frauds obvious on the face of the documents. They were able to identify likely frauds “in almost every file.” Any honest, mildly competent review of the loan files by the loan brokers and lenders would have prevented these loans from being closed. The logical conclusion is that the lenders and brokers encouraged fraudulent loans.
Hudson also explains the tactics that loan officers use to intimidate borrowers to ensure that they did not read the false disclosures that the officers had fabricated (p. 157).
Recent studies by criminologists show the leading role that lenders and loan brokers took in creating fraudulent loan applications. Tomson H. Nguyen and Henry N. Pontell recently published an article reporting the results of their interviews with lender personnel and loan brokers. (I published the responsive policy essay on their article.)
Appraisal fraud was endemic and it is a “marker” of accounting control fraud
There is no honest reason why a mortgage lender would inflate the appraised value and the size of the loan. Causing or permitting large numbers of inflated appraisals is a superb “marker” of accounting control fraud by the lender because the senior officers directing an accounting control fraud do maximize short-term reported (fictional) income (and real losses) by inflating appraisals and stated income. Lenders and their agents frequently suborned appraisers by deliberately creating a Gresham’s dynamic to try to induce them to inflate market values, leaked the loan amount to the appraisers, drove the appraisal fraud, and made it endemic. A national poll of appraisers in early 2004 found that 75% of respondents reported being subjected to coercion in the last 12 months to inflate appraisals. A follow-up survey in 2007 found that the percentage that had been subjected to coercion had risen to 90 percent. Appraisers reported that when they refused to inflate appraisals 68% had lost at least one client and 45% were not paid for at least one appraisal in the prior 12 months. In 2005, Demos warned of an “epidemic” of appraisal fraud.
As with inflating income in order to minimize the reported debt-to-income ratio, inflating the appraisal allowed everyone with a financial stake in the lies to minimize the reported loan-to-value (LTV) ratio and allowed everyone to pretend that the loan was far less risky because it had such a large (but yet again fictional) equity cushion. Given that we know that appraisal fraud was endemic, that endemic appraisal fraud is impossible without being led or permitted by the lenders and their agents, and that no honest lender would permit or cause widespread inflated appraisals, the logical inference is that the lenders and their agents led both the stated income and the appraisal fraud. Appraisal fraud is particularly pernicious because the borrower does not know it has occurred. He may be told that the home he offered to pay $400,000 to acquire (subject to an appraisal contingency) has a market value of $480,000 when its true market value is $350,000. This constitutes fraud in the inducement.
The New York Attorney General’s investigation of Washington Mutual (WaMu) (one of the largest nonprime mortgage lenders) and its appraisal practices supports this dynamic.
New York Attorney General Andrew Cuomo said [that] a major real estate appraisal company colluded with the nation’s largest savings and loan companies to inflate the values of homes nationwide, contributing to the subprime mortgage crisis.
“This is a case we believe is indicative of an industrywide problem,” Cuomo said in a news conference.
Cuomo announced the civil lawsuit against eAppraiseIT that accuses the First American Corp. subsidiary of caving in to pressure from Washington Mutual Inc. to use a list of “proven appraisers” who he claims inflated home appraisals.
He also released e-mails that he said show executives were aware they were violating federal regulations. The lawsuit filed in state Supreme Court in Manhattan seeks to stop the practice, recover profits and assess penalties.
“These blatant actions of First American and eAppraiseIT have contributed to the growing foreclosure crisis and turmoil in the housing market,” Cuomo said in a statement. “By allowing Washington Mutual to hand-pick appraisers who inflated values, First American helped set the current mortgage crisis in motion.”
“First American and eAppraiseIT violated that independence when Washington Mutual strong-armed them into a system designed to rip off homeowners and investors alike,” he said (The Seattle Times, November 1, 2007).
Note particularly Attorney General Cuomo’s claim that WaMu “rip[ped] off … investors.” That is an express claim that it operated as an accounting control fraud and inflated appraisals in order to maximize accounting “profits.” A Senate investigation has found compelling evidence that WaMu acted in a manner that fits the accounting control fraud pattern.
Pressure to inflate appraisals was endemic among nonprime lending specialists.
Appraisers complained on blogs and industry message boards of being pressured by mortgage brokers, lenders and even builders to “hit a number,” in industry parlance, meaning the other party wanted them to appraise the home at a certain amount regardless of what it was actually worth. Appraisers risked being blacklisted if they stuck to their guns. “We know that it went on and we know just about everybody was involved to some extent,” said Marc Savitt, the National Association of Mortgage Banker’s immediate past president and chief point person during the first half of 2009 (Washington Independent, August 5, 2009).
Inducing endemic appraisal fraud is an optimal strategy for a lender that is engaged in “accounting control fraud.” Accounting control frauds drove the second phase of the S&L debacle, the Enron era crisis, and the ongoing crisis.
Hudson notes that:
One former loan officer and branch manager testified that inflating property appraisals served the “dual purpose of both making sure the loan was approved by the home office as well as making the loan more attractive to sell to investors” (p. 156).
The amount of liar’s loans made was staggering
By 2006, we believe that 30% of U.S. mortgage issuances made during the year were liar’s loans. That represents millions of liar’s loans, with a fraud incidence in the range of 90%. It also explains why the housing bubble was hyper-inflated.
Liar’s loans hyper-inflated the housing bubble
Rajdeep Sengupta, an economist at the Federal Reserve Bank of St. Louis, reported in 2010 in an article entitled “Alt-A: The Forgotten Segment of the Mortgage Market” that:
[B]etween 2003 and 2006 … subprime and Alt-A [loans grew] 94 and 340 percent, respectively.The higher levels of originations after 2003 were largely sustained by the growth of the nonprime (both the subprime and Alt-A) segment of the mortgage market.
These figures greatly understate the role of “Alt-A” loans (the euphemism for “liar’s loans”) for they ignore the fact that by 2006 half of the loans called “subprime” were also liar’s loans. (Credit Suisse: 2007). Fraudulent liar’s loans hyper-inflated and greatly extended the life of the bubble.
Competent investigations find endemic fraud at entities making, purchasing, and selling liar’s loans and CDOs backed by liar’s loans
The recent FHFA complaints against many of the world’s largest banks allege that the banks defrauded Fannie and Freddie. The complaints set forth the investigations that have confirmed the existence of endemic fraud by the named defendant. The FHFA complaints also allege that the lenders knew that they were engaged in defrauding Fannie and Freddie. (The senior Swiss regulator later conceded to me that he had not read the complaint against Credit Suisse and that if the complaint’s allegation (which I described for him) that Credit Suisse knowingly sold Fannie and Freddie enormous volumes of paper it knew through reviews to be non-complying with the underwriting guidelines that the bank purported to be following were true he would have to concede that such acts were fraudulent.
The key facts that the FHFA alleges that are new are that the banks created a paper trail establishing that the banks knew that the loans did not comply with the underwriting standards that the bank promised it followed. The divergence between the promises and the actual terrible quality of the loans delivered is so great that the fraud must have been endemic at the banks if the complaint is accurate.
The paper trail establishing intent to defraud raises the question – why hasn’t the DOJ prosecuted these cases? The DOJ’s excuses about how tough it is prosecute elite frauds always rang hollow to those who built the system that prosecuted the S&L control frauds successfully. But the paper trail the FHFA alleges exists is more than a smoking gun – it is a flaming rocket launcher. Either the FHFA is wrong about the facts or Attorney General has run out of excuses for his failure to prosecute the fraudulent plutocrats that drove this crisis.