To Assign, or Not to Assign: Rethinking Assignee Liability as a Solution to the Subprime Mortgage Crisis TO ASSIGN, OR NOT TO ASSIGN: RETHINKING ASSIGNEE LIABILITY AS A SOLUTION TO THE SUBPRIME MORTGAGE CRISIS James Carlson* I. Introduction .............................................................. 1022 II. B ackground ............................................................... 1025 A. Subprim e Lending ............................................. 1025 B . Securitization ..................................................... 1030 III. Current Legal Landscape ........................................ 1034 A. Federal Solutions: The Weak Federal A pproach ............................................................ 1037 B. State Solutions: The Problem of Regulatory C apture ............................................................... 1038 IV. Conceptualizing Assignee Liability ......................... 1040 A. The Costs and Uncertainties of Due Diligence R eview ................................................................ 1042 B. Requiring Solvency: Practical and Legal O bstacles ............................................................ 1043 C. Litigation Barriers: Assignee Liability and Obstacles to Enforcement .................................. 1046 V. Empirical Analysis of Assignee Liability ................ 1048 A. A Note on the Data ............................................ 1049 B. Past Empirical Research ................................... 1050 C. Contributions to the Empirical Study .............. 1052 1. Propensity Score Matching .......................... 1052 2. Using Rate Spread Instead of Subprime Lender List .................................................... 1055 * Abraham L. Freedman Fellow, Beasley School of Law, Temple University; B.A. 1996, Loyola University of New Orleans; J.D. 1999, Columbia University. The author wishes to thank Lester Brickman, Richard Nagareda, David Beck and all of the practitioners representing the distinct constituencies discussed in this article that provided comments and suggestions on previous drafts. 3. Comparing Assignee Liability Provisions ... 1056 4. State-Specific Focus ...................................... 1057 D. Theoretical M odels ............................................. 1059 E . R esults ................................................................ 1060 1. Full Sam ple ................................................... 1060 a. Rate Spread: Interest Rate on Subprime L oan s ....................................................... 1060 2. High Cost: Probability of Originating a H igh Cost Loan ............................................. 1062 3. State Sample: Probability of Securitization 1064 F . D iscussion .......................................................... 1064 VI. The Great Misconceptualization of Subprime L ending ..................................................................... 1065 V II. Conclusion ................................................................ 1068 I. INTRODUCTION No city better epitomizes the subprime crisis than Cleveland. In 2006, the Census Bureau declared Cleveland the poorest big city in America-with 32% of its population living below the poverty level.' Expectedly, nearly 30% of loans originating in the Cleveland region during this time were subprime, many undoubtedly on predatory terms. When the subprime crash engulfed the region, nearly 24,000 people lost their homes, leaving 10,000 abandoned buildings.' By contrast, one of the more devastated regions of New Orleans, in the wake of Hurricane Katrina, lost about 13,700 homes.4 The mortgage foreclosure crisis and subprime meltdown hardly needs further introduction. From 1994 to 2006, the value of all subprime loans increased from $30 billion to $640 billion.5 In 2007, subprime loans plummeted to $50 billion, ' Thomas Ott, Real Estate's Perfect Storm, CLEVELAND PLAIN DEALER, Jan. 20, 2008, at Al. 2 See id. 3 Id. 4 Id. ' See Ted Frank, Prime Target, WALL ST. J., Apr. 25, 2007, at A15. COL UMBIA B USINESS L4 W RE VIE W [Vol. 2008 RETHINKING ASSIGNEE LIABILITY or 2% of all loans.6 Investment in subprime loans, securitization, prompted the expansion of subprime lending. Over 8% of subprime loans in 2006 found their way into securitized pools.' Securitization allowed mortgage originators to shift lending risk to the larger economy, making loans possible to those previously thought too risky. At the same time, it gave investment banks an unprecedented amount of capital to finance new projects. Seven of the largest buyouts of all time occurred in 2006.8 Yet, securitization had consequences. Mortgage originators sold loans to security pools so quickly that they needed to pay little attention to the legal and financial consequences of unsound lending. In 2005, the housing bubble exhibited the first signs of strain.9 Housing prices flattened in 2006.10 This meant homeowners, owing amounts beyond their means, no longer had an ever-increasing amount of equity with which to finance debt. As the number of foreclosures increased, experts predicted a dire future-one predicted that at least 20% of subprime borrowers were likely to default by 2008." Investors panicked, and almost overnight, the subprime credit market collapsed. 6 See Damian Paletta & James R. Hagerty, Fed's New Rules on Mortgages Draw Hostility, WALL ST. J., Dec. 19, 2007, at Al. 7 See Raphael W. Bostic et al., State and Local Anti-Predatory Lending Laws: The Effect of Legal Enforcement Mechanisms 10 (Aug. 7, 2007) (unpublished manuscript, available at http://ssrn.com/abstract= 1005423). ' See Goldman, Citi Top List of 2006 M&A Deal Makers, N.Y. TIMES DEALBOOK, Jan. 3, 2007, http://dealbook.blogs.nytimes.com/2007/01/03/ goldman-citi-top-final-ma-league-tables. ' See Ruth Simon & James R. Hagerty, How American Lenders Shelter Themselves, WALL ST. J., Sept. 22, 2005, at Cl; see also Al Yoon, Housing Bubble Bursts in U.S. Mortgage Bond Market (Update 2), BLOOMBERG.COM, Dec. 6, 2005, http://www.bloomberg.com/apps/news?pid= 10000103&sid=aDSB37OItSJU&refer=us. 10 See Mark Whitehouse, Risk Management: As Home Owners Face Strains, Market Bets on Loan Defaults, WALL ST. J., Oct. 30, 2006, at Al. n See id. No. 3:1021 ] COLUMBIA BUSINESS LAW REVIEW Following the subprime crash of 2007, calls for regulation of subprime lending reached fever pitch. Assignee liability, or abrogation of the "holder in due course" doctrine, emerged as a consistent favorite. Praised by academics and adopted by many states, assignee liability allowed borrowers to bring claims against the trust holding their loan, as well as the originator. This provided a remedy for the homeowner when the originator disappeared and would, in theory, force the banks to police the originators. Regulators believed that by making assignees liable they could prevent some of the risk- shifting that made predatory lending possible. This Note offers a novel empirical rebuke of assignee liability. The use of propensity score matching creates a more robust picture of subprime lending than did previous studies limited to cross-border controls. In doing so, this Note tests three questions: (1) whether assignee laws make subprime lending more expensive, (2) whether assignee laws promote a sense of security in borrowers, and (3) whether predatory lending laws stop predatory lending, or merely cause it to hide at lower interest rates. The results suggest that assignee liability laws may be promoting a false sense of security by appearing to eliminate predatory lending, but not doing so in fact. They demonstrate a consumer confidence effect of predatory lending legislation, one that makes assignee liability particularly dangerous if it is otherwise ineffective. The failure of assignee liability is a failure of conception. In an attempt to remedy the effects of subprime lending, the quest became one of compensation for the victims, rather than deterrence of the practice. The result is a regime that adds a surcharge to the cost of making legitimate subprime loans, limiting the subprime market without preventing predatory practices. Solutions like those used against insider trading-putting the good of the market ahead of compensation for individuals-provide a more sustainable approach. It is least cost avoidance, investor confidence, and deterrence, not compensation, that must form the basis of any solution. [Vol. 2008 RETHINKING ASSIGNEE LIABILITY This Note includes six parts. Part II includes necessary background on subprime lending, a tentative definition of predatory lending, and a brief explanation of how securitization created the present disaster. Part III explains the current federal -and state legal landscape, positing the weaknesses inherent in a state-based approach. Part IV explores how the current conceptualization of the subprime crisis leads to a theoretically ineffective solution like assignee liability. Part V demonstrates, empirically, that assignee liability may not only be ineffective, but actually dangerous. Finally, Part VI suggests a way to properly conceptualize the subprime crisis in the search for a solution. II. BACKGROUND A. Subprime Lending Subprime mortgages fall into several categories. These include borrowers with Fair Isaac & Co. (FICO) scores below 660, borrowers with at least one sixty-day or two thirty-day delinquencies in the past two years, or borrowers subject to a judgment or foreclosure in the last five years. 2 Subprime loans typically carry higher interest rates than prime loans. Not only is the risk of default higher, but the risk of prepayment is higher because prepayment occurs not only 2 Lenders use FICO scores to determine the creditworthiness of a potential borrower. The Fair Isaac & Company offers a full explanation. See generally MyFICO.com, http://www.myfico.com/Downloads/Files/ myFICO_UYFSBooklet.pdf (last visited Dec. 3, 2008) (explaining FICO scores). The subprime market can include people with good credit who have high loan to value ratios, high payment to income ratios, limited documentation of income, or unstable income. See Roberto G. Quercia et al., Assessing the Impact of North Carolina's Predatory Lending Law, 15 HousING POLy DEBATE 573, 574 (2004). Perhaps the easiest way of characterizing subprime lending is that it encompasses all those who are ineligible for prime lending. See Peter Chinloy & Nancy MacDonald, Subprime Lenders and Mortgage Market Completion, 30 J. REAL EST. FIN. & ECON. 153, 164 (2005). No. 3:1021] when interest rates go down, but also if the borrower later qualifies for a prime loan. 13 Households historically excluded from the regular credit market, especially minority households, comprise much of subprime lending. 4 These consumers tend to be unfamiliar with the loan process and often lack the educational skills necessary to comprehend the loan documents they sign. 5 The subprime market tends to consist of the most vulnerable consumers. '" See Ronald H. Silverman, Toward Curing Predatory Lending, 122 BANKING L.J. 483, 528-29 (2005) (suggesting that lenders dislike prepayment because they lose out on interest payments, and that subprime borrowers are particularly likely to prepay). 14 See David J. Weiner, Assignee Liability in State Predatory Lending Laws: How Uncapped Punitive Damages Threaten the Secondary Mortgage Market, 55 EMORY L.J. 535, 542 (2006) (suggesting that subprime lending was three times more likely in low-income neighborhoods, and five times more likely in minority neighborhoods). See also Robert B. Avery et al., Higher-Priced Home Lending and the 2005 HMDA Data, 92 FED. REs. BULL. A123, A126 (2006); DEBBIE GRUENSTEIN BOCIAN ET AL., CTR. FOR RESPONSIBLE LENDING, UNFAIR LENDING: THE EFFECT OF RACE AND ETHNICITY ON THE PRICE OF SUBPRIME MORTGAGES (2006), http://www. responsiblelending.org/pdfs/rrOl1-UnfairLending-0506.pdf; Bostic et al., supra note 7, at 15; Giang Ho & Anthony Pennington-Cross, The Impact of Local Predatory Lending Laws on the Flow of Subprime Credit, 60 J. URBAN ECON. 210, 219 (2006). Moreover Table 4 demonstrates that minorities comprise nearly 40% of all the subprime loans offered between 2004 and 2006, whereas they comprise roughly only 20% of the prime market in the same period. " These are complex transactions. The paper documentation can be up to one inch thick and require signatures in twenty to forty places. One study demonstrated that about 79% of adults have limited quantitative literacy skills and 96% cannot extract and compute credit cost information from disclosure documents. Elizabeth Renuart, Toward One Competitive and Fair Mortgage Market: Suggested Reforms in a Tale of Three Markets Point in the Right Direction, 82 TEX. L. REV. 421, 432 (2003). Moreover, the great effectiveness of credit counseling-homeowners are almost twice as likely to avoid foreclosure when they see a counselor--demonstrates that most subprime borrowers do not have the wherewithal to make sustainable credit decisions. Roberto G. Quercia et al., The Cost- Effectiveness of Community-Based Foreclosure Prevention 31 (Dec. 8, 2005) (unpublished manuscript, available at http://www.fhfund.org/ dnldlreports/MFPFull-Report.pdf). [Vol. 2008COL UMIA B USINESS LA W RE VIE W RETHINKING ASSIGNEE LIABILITY Nevertheless, many experts believe subprime lending is welfare enhancing. 6 Numerous studies link the availability of subprime credit to a national rise in home ownership. 7 Table 1 tracks the dramatic 5% increase in U.S. homeownership since 1995.1 These gains were the most pronounced among African-Americans and Hispanics, with homeownership rates increasing almost 10% for each group. The rise of homeownership closely followed the explosion of subprime lending. 9 One study determined that changes in demographics only accounted for about half of the housing growth.2° Subprime lending provides needed credit to many in the non-prime sector.21 Subprime loans also help credit 16 Ho & Pennington-Cross, supra note 14, at 211; Chinloy & MacDonald, supra note 12, at 154; Keith D. Harvey & Peter J. Nigro, How Do Predatory Lending Laws Influence Mortgage Lending in Urban Areas? A Tale of Two Cities, 25 J. REAL EST. RES. 479 (2003). 17 John Kiff & Paul Mills, Money for Nothing and Checks for Free: Recent Developments in U.S. Subprime Mortgage Markets 4 (Int'l Monetary Fund, Working Paper No. WP07188, 2007); Chinloy & MacDonald, supra note 12, at 163. " Table 1 reflects Census Bureau data through the third quarter of 2007. 19 From 1983 to 1994, subprime lending grew from less than 2% to about 5% of the market. By 1996, it was at roughly 10% of the market. Baher Azmy, Squaring the Predatory Lending Circle: A Case for States as Laboratories of Experimentation, 57 FLA. L. REV. 295, 307 (2005). By 2006, it was well over 20% of the market. If subprime and Alt-A loans are included (loans not technically subprime, but with higher interest rates than regular prime loans), the market-share reaches almost 46%. See, e.g., Kiff & Mills, supra note 17, at 6; Paletta & Hagerty, supra note 6. 20 See Jonas D. M. Fisher & Saad Quayyum, The Great Turn-of-the- Century Housing Boom, 30 ECON. PERSP. 29, 40-41 (2006), available at http://www.chicagofed.org/publications/economicperspectives/ep-3qtr2006- part3_fisher-quayyum.pdf. 21 This real-world result is backed by studies demonstrating that liquidity is quite important. One such study demonstrated that usury lending, lending to the poor at very high rates, was actually welfare enhancing in Africa. Families to whom short-term loans were extended were more likely to keep jobs, eat adequately, and maintain shelter than were those denied the same loans, even with Annual Percentage Rates (APRs) exceeding 200%. See In Praise of Usury; Economics Focus, ECONOMIST, Aug. 4, 2007, at 66. No. 3:1021] consumers avoid many more costly regimes-like credit cards, check cashing, and cash leases.22 While most researchers agree some subprime lending is a net positive, the point at which it becomes a net-negative is a hotly contested and case-by-case inquiry.23 Predatory lending presents the dark side of the subprime market. While difficult to define, one accepted definition is any loan that results in net harm to the borrower.24 Some specific practices include asset-based lending, rent seeking, fraud, non-transparency, loans requiring the waiver of legal redress, lending discrimination, and servicing abuses." A look at state lending laws suggests that other practices like balloon payments, negative amortization, prepayment penalties, and asset-based lending are all predatory.26 22 See generally Silverman, supra note 13, at 483-91 (describing some of the egregious practices in a variety of other alternative credit regimes). These include APRs on payday Loans of over 365% and Cash Leasing services (short-term loans for small amounts of money with proof of ownership of electronic devices) with APRs exceeding 730%. Credit cards, a more mainstream alternative source of credit, have an average APR of almost 15%. See Consumer-Action.org, 2007 Credit Card Survey, http:// www.consumer-action.org/news/articles/2007 credit-cardsurvey. Credit cards with the lowest APRs are unavailable to subprime borrowers, suggesting the average rate for them is much higher. 23 See Daniel S. Ehrenberg, If the Loan Doesn't Fit, Don't Take it: Applying the Suitability Doctrine to the Mortgage Industry to Eliminate Predatory Lending, 10 J. AFFORDABLE HOUSING & COMMUNITY DEV. 117, 119 (2001) (suggesting it is difficult to determine, even looking solely at interest rates, at what level a subprime loan becomes ruinous rather than beneficial); see also Michael J. Pyle, A "Flip" Look at Predatory Lending: Will the Fed's Revised Regulation Z End Abusive Refinancing Practices?, 112 YALE L.J. 1919, 1925 (2003). 21 Some argue that the entire subprime market is predatory. For example, one author argues that all subprime lenders charge interest rates that are unrelated to risk. See Renuart, supra note 15, at 428-30. 21 See Kathleen C. Engel & Patricia A. McCoy, Turning a Blind Eye: Wall Street Finance of Predatory Lending, 75 FORDHAM L. REV. 2039 (2007). 26 California specifically outlawed: prepayment fees after thirty-six months, negative amortization (monthly payments lower than interest), the increase of interest rate following default (late payment), having no reasonable belief that borrower can pay loan (no asset-based lending- COL UMBIA BUSINESS LA W RE VIE W [Vol. 2008 RETHINKING ASSIGNEE LIABILITY Predatory lending also often encompasses "steering," or pushing people toward subprime loans when they could qualify for prime loans.27 Predatory loans entail significant costs. Studies suggest that prepayment penalties or balloon payments boost the risk of foreclosure by up to 50%.28 In turn, foreclosures impose negative externalities. Aside from increasing homelessness, foreclosures lead to abandoned houses. Abandonment creates both direct civic costs and indirect neighborhood cohesion costs.29 Moreover, lenders lose roughly $35,000 on the average foreclosure, a number likely to increase as more homes fall into receivership." The lending with intent to foreclose), brokers' fees over a certain amount, and loan flipping (refinancing with same lender in order to perpetuate someone in a loan they cannot afford). See CAL. FIN. CODE §§ 4973, 4979 (2008). The Texas anti-predatory lending law bans balloon payments (loan payments increase dramatically over life of loan). See TEX. FIN. CODE ANN. §§ 343.201-205 (2007). Finally, Pennsylvania bans call provisions (ability to call for full amount of indebtedness at anytime) and advance payments from the proceeds of the loan itself. See 63 PA. STAT. ANN. § 456.511 (2007). 27 Scholars blame informational asymmetries for some of this problem. Many subprime lenders set up shop in historically unserved areas. Thus, the majority of their customers are subprime, but they can steer even those who are not into subprime loans because they are credit-status naive. See Renuart, supra note 15, at 423; Silverman, supra note 13, at 528-29; Lloyd T. Wilson, Jr., Effecting Responsibility in the Mortgage Broker-Borrower Relationship: A Role for Agency Principles in Predatory Lending, 73 U. CIN. L. REV. 1471, 1478-79 (2005). In fact, Fannie Mae alleges that over half of the people who take out subprime loans could qualify for prime loans but do not realize it. Hunting the Loan Sharks; Predatory Lending in America, ECONOMIST, Aug. 31, 2002, at 55-56. 21 See Engel & McCoy, supra note 25, at 2042. 29 See Silverman, supra note 13, at 528-29 (suggesting foreclosures destroy any sense of neighborhood). These damages include things like extra police and fire costs to cover the risk of vandalism and arson in abandoned homes. See also Gretchen Morgenson, Baltimore is Suing Bank Over Foreclosure Crisis, N.Y. TIMES, Jan. 8, 2008, at A12. Cleveland has recently attempted to recover these damages under a public nuisance theory. See Henry J. Gomez & Thomas Ott, Mayor Takes on Banks in Lawsuit, CLEVELAND PLAIN DEALER, Jan. 11, 2008, at Al. '0 See William C. Apgar et al., The Municipal Cost of Foreclosures: A Chicago Case Study, (Homeownership Preservation Foundation, Policy Paper No. 2005-1, 2005), available at http://www.nw.org/network/ No. 3:1021] problem is not going anywhere. In 2005, experts estimated that almost 15% of all mortgages were to borrowers considered most likely to default." In 2006, experts predicted that by 2008 as many as 20% of all subprime borrowers would be in arrears.12 The cost of foreclosure- induced home turnover necessitates the intervention of a regulatory regime. B. Securitization Securitization has fostered a climate for predatory practices to flourish alongside legitimate subprime lending. However, in an unsolved enigma, legitimate subprime lending only became possible due to securitization 3 Securitization involves the pooling of multiple loans and then the selling of securities backed by the interest and principal payments of the pooled loans. This is no small phenomenon; about 80% of subprime loans are securitized.3 4 Roughly 70% neighborworksProgs/foreclosuresolutions/pdf docs/2005Apgar-DudaStudy- FullVersion.pdf. " See Yoon, supra note 9. 32 See Whitehouse, supra note 10. An early article on securitization pointed out that economies of scale involved in large pooling and reselling of bonds allows the income stream to become far more valuable than it would on its own. See Steven L. Schwarcz, The Parts Are Greater Than the Whole: How Securitization of Divisible Interests Can Revolutionize Structured Finance and Open the Capital Markets to Middle-Market Companies, 1993 COLUM. Bus. L. REV. 139, 167 (1993). " This phenomenon amounts to about $525 billion in the securities market. See Engel & McCoy, supra note 25, at 2045. That said, Wall Street has long experienced securitizing essentially any positive cash flow. These have included credit card debt, automobile loans, commercial loans, equipment leases, and even loans to developing countries. See Christopher L. Peterson, Predatory Structured Finance, 28 CARDOZO L. REV. 2185, 2198 (2006). Even a David Bowie album had its future revenue streams securitized. Lisa M. Fairfax, When You Wish Upon a Star: Explaining the Cautious Growth of Royalty-Backed Securitization, 1999 COLUM. Bus. L. REV. 441,442 (1999). COL UMBIA B USINESS LA W RE VIE W (Vol. 2008 of subprime loans in this Note's study were resold." Creditors utilize securitization to increase their profits from extending credit and share the risk with other parties.36 Pooling the loans reduces the risk from any one individual loan while making "lemons" harder to detect, allowing predatory lenders to package predatory loans with legitimate ones. 37 The complexity of securitization is responsible for its success. Initially, an investment bank buys a selection of subprime loans throughout the country. It pools the loans, takes the receivables from those loans (payments), and converts the cash flows into bonds. The asset-backed securities ("ABSs") are then sold to "special purpose vehicles" ("SPVs") who add credit enhancements in both external (insurance) and internal ("tranching") forms. 9 Tranching refers to a "senior subordinate structure" that should be familiar to any preferred stock holder.4" Next, the SPV picks a servicer to collect payments from the borrowers for transfer " Which makes sense, as the data only include those loans securitized in the same year as the loan was made. While often these happen quickly, one can imagine the case where it does not. 36 See Kathleen C. Engel & Patricia A. McCoy, A Tale of Three Markets: The Law and Economics of Predatory Lending, 80 TEX. L. REV. 1255, 1274 (2002); Peterson, supra note 34, at 2200-07; David Reiss, Subprime Securitization: How Rating Agencies Allow Predatory Lending to Flourish in the Secondary Mortgage Market, 33 FLA. ST. U. L. REV. 985, 1001-02 (2006); Silverman, supra note 13, at 537. " For a more thorough overview of the subprime lending process, see Azmy, supra note 19; Bostic et al., supra note 7; Engel & McCoy, supra note 25; Reiss, supra note 36. 38 Engel & McCoy, supra note 25, at 2045. 9 See Engel & McCoy, supra note 25, at 2045; Reiss, supra note 36, at 1003. This transfer ensures protection of the investor from claims by a creditor against the bank, because the bank no longer holds the assets. When combined with the "holder in due course" rule discussed infra, it makes the actual money paid out of the loan untouchable by the borrower. 40 Essentially, tranching ranks the bonds from "AAA" to "B" with "AAA" being the first to receive payment, and then "AA," and so on. See Engel & McCoy, supra note 25, at 2046. No. 3:1021 ] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LA W REVIEW to the trust and subsequent payment to investors.4 Then, a rating agency rates the bonds based on a cursory, possibly fraudulent, review of the underlying mortgages.42 Finally, the SPV makes the bonds available to the market for investors to purchase. The worst of the tranches, those at the bottom of the subordinate structure, are often pooled together with similar tranches from other pools of bonds. These "collateralized debt obligations" ("CDOs") are usually sold to overseas investors.43 Investors buy many bonds with little to no information about the underlying loans. These securities were initially so popular that they were even sold on a "to be announced" basis.44 The housing bubble made ABSs an unmatched investment vehicle because even the riskiest loans avoided foreclosure. As housing prices appreciated-thus generating new equity-loans were extended to families who could not possibly afford them. These loans were then refinanced as the price of the home increased, taking the appreciated equity immediately before default and continuing payments to investors.45 The subsequent high investor demand and separation from risk, combined with questionable oversight 4" See Engel & McCoy, supra note 36, at 1274. Often the original lender, who no longer holds the loan, is chosen to be the servicer. This can be especially lucrative to the lender, who receives a fee for this service. The origniator gets not only the money from selling the loan, but also gets paid to service the loan. 42 See Engel & McCoy, supra note 36. The rating agency may require further credit enhancement in the form of more tranching or more insurance. Typically the rating agencies rate four aspects: qualitative, quantitative, servicing, and legal risk. Reiss, supra note 36, at 1014. There is significant evidence that rating agencies, who have little incentive to rate tranches lowly (as they are paid by the banks), were doing a completely inadequate, and possibly fraudulent, review. See AAAsking for Trouble; Debt Ratings, ECONOMIST, July 14, 2007, at 83. Collateralized Debt Obligations, ("CDOs"), the worst of the ABSs, made up roughly 40% of rating agency Moody's profits in 2006. See id. 43 See Engel & McCoy, supra note 25, at 2066-68. It should come as no surprise that countries everywhere are concerned by the American foreclosure plight. ' See Engel & McCoy, supra note 25, at 2071. 41 See Kiff & Mills, supra note 17, at 9-10. [Vol. 2008 by the rating agencies, allowed bad loans to be included with good loans while providing no incentive for lenders to make reasonable loans.46 Therefore, the market rewards lenders for quantity, not quality of loans. However, once the bubble burst, housing prices flattened and even fell.47 Now home owners have negative equity and no ability to refinance.48 Unsurprisingly, default and foreclosure followed, with investors at the lower end of the subordinate tranches losing significant amounts of money.49 The securitization market represents a series of incentive mismatches. First, originators can ignore the likelihood of loan repayment when originating new loans because they can hide and sell them before foreclosure. Second, rating agencies are rewarded by banks for rating highly. So long as investors make money, the agencies can justify providing inflated ratings to investors.0 Third, originators can distort the employment history of the borrower, the borrower's 46 See generally Kiff & Mills, supra note 17 (referring to the high demand for ABSs); Reiss, supra note 36 (referring to the questionable rating practices and perhaps outright fraud of the rating agencies); Special Report: Cracks in the Facade, ECONOMIST, Mar. 24, 2007, at 87 (reminding that it was insatiable investor desire driving this trend). 47 See generally Simon & Hagerty, supra note 9 (forecasting that a weak housing market would cause significant problems for the subprime industry); Yoon, supra note 9 (pointing out the slowing of housing price growth). ' There is some suggestion that 18% of people taking out loans in 2006 had negative equity. See Special Report: Cracks in the Facade, supra note 46. 49 See Kiff & Mills, supra note 17, at 10 (estimating that the losses will be almost twenty-five billion dollars if the housing prices remain flat, but that if they drop 5%, losses will be closer to sixty billion dollars). In 2007, some experts predicted prices might fall as much as 10%, which others predict would mean a loss of about one-hundred and twelve billion. See Special Report: Cracks in the Facade, supra note 46. Also, there is a fear that this may have a larger cost in terms of shaking overall investor confidence. See Subprime Subsidence; Mortgage Lending, ECONOMIST, Dec. 16, 2006, at 89. " See Engel & McCoy, supra note 25, at 2055 (pointing out that there were still highly rated securities backed by subprime loans, even when the collapse had already begun in 2006). No. 3:1021] RETHINKING ASSIGNEE LIABILITY income, and home appraisals without much risk of discovery until it is too late.5' To complicate matters further, the "holder in due course" doctrine compounds the securitization incentive mismatch by preventing the borrower from asserting claims against the trusts holding the pooled loans.52 First, securitization allows originators to operate at the edge of solvency with little regard to the quality of originated loans.53 Second, since there is no remedy for predatory lending practices if the originator is insolvent, the borrower may find that the originator has disappeared and that the securitized trust is legally immune. When the trust goes to foreclose after default, the borrower has no remedy and is left without a house, despite having a legitimate claim for fraudulent and abusive conduct. At its simplest, the subprime collapse was a product of misaligned incentives compounded by legal immunity.54 III. CURRENT LEGAL LANDSCAPE Predatory lending laws exist at both the federal and state level and have essentially two important features: (1) damages and (2) coverage. Damages determine predictability, and coverage determines scope. 5 The laws 51 See generally Silverman, supra note 13, at 516; Patrick Barta, Is Appraisal Process Skewing Home Values?, WALL ST. J., Aug. 13, 2001, at Al (interviewing an appraiser to demonstrate the pressure lenders put on them to produce the "right" number). " See infra note 56 (discussing the "holder in due course" doctrine in more detail); see also Peterson, supra note 34, at 2236. Essentially the doctrine means that the holder of a loan purchased for consideration is not liable for anything that could have been brought against the originator. The idea was to encourage liquidity, but it has caused many troubling results. ' As noted, the originator sells the loan so quickly-70% in the first year according to data in this study-that they are very unlikely to be stuck with it when it goes bad. See Kiff& Mills, supra note 17, at 7. 5' Most laws allow for recovery to the extent of the loan plus some extra costs. Some laws, however, allow treble damages and punitive COLUMBIA B USINESS LA W RE VIEW [Vol. 2008 come in several variations under these two dimensions. Some laws cover only refinancing, some laws cover only certain loans based on interest rate, and some laws impose full assignee liability. Assignee lending laws allow consumers to overcome the "holder in due course" doctrine and sue trusts holding the securitized loans.56 Many regulators and academics favor laws with assignee coverage, arguing that it properly realigns market incentives.57 Also, federal regulators have begun to question the wisdom of allowing states to regulate lending. In fact, one recent regulatory proposal is a strictly national assignee liability approach.58 The range of conduct prohibited by predatory lending laws, and therefore, what conduct might also impose liability on assignees, is fairly uniform.59 The targeted predatory practices were already described when defining predatory lending. These include balloon payments, negative amortization, asset-based lending, call provisions, prepayment penalties, and outright fraud.6" Originators are damages. Laws with unpredictable damages can make lending in that market particularly difficult or costly. 56 Normally, the "holder in due course" doctrine prevents any claims from being asserted against the assignee of a promissory note. U.C.C. § 3- 302 (2005). This effectively leaves the borrower with no remedy because the mortgage originators operate constantly at the edge of insolvency. Engel & McCoy, supra note 25, at 2079 (arguing that borrowers are left without full legal relief). " See, e.g., Bostic et al., supra note 7; Engel & McCoy, supra note 25; Kiff & Mills, supra note 17; Peterson, supra note 34; Siddhartha Venkatesan, Abrogating the Holder in Due Course Doctrine in Subprime Mortgage Transactions to More Effectively Police Predatory Lending, 7 N.Y.U. J. INT'L L. & POL. 177 (2003); Weiner, supra note 14. Stuart M. Saft, The Anti-Mortgage Lending Act, WALL ST. J., Nov. 10, 2007, at A10 (reviving the Mortgage Reform and Anti-Predatory Lending Act that was approved by the House Financial Services Committee at the beginning of November). " Currently, the Federal approach covers little, so the state laws generally provide the basis for analysis. What they cover and do not cover is more important than the federal law because they usually are triggered first. 60 See supra notes 26-27 and accompanying text. No. 3:1021] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LAW REVIEW liable for any loans they make that have these characteristics under virtually any predatory lending law. Yet, the important provisions include what practices the law transfers to assignees. The Illinois High Risk Home Loan Act demonstrates a typical predatory lending law with assignee liability. The law allows borrowers to assert all affirmative claims and defenses they would have against the originator also against an assignee, so long as the underlying loan is a "High Risk Home Loan."6' The statute defines the term as loans with interest rates 6% or 8% higher than a comparable Treasury loan, depending on whether it is a first or second lien, respectively.2 So, if a loan has any of the characteristics just mentioned-balloon payments, negative amortization, etc. - the assignee can be liable, but only if the interest rate on the loan exceeds the threshold. However, the law also offers due diligence protection for the assignee. Assignees may avoid liability by demonstrating that they: (1) have a policy against purchasing "High Risk Loans," (2) require the originator, by contract, not to sell such loans to them, and (3) exercise "reasonable" due diligence based on a sampling of the loans. The law offers even more protection in the form of a "shield provision." The "shield provision" allows an assignee to "buy" itself out of liability by paying restitution necessary to remove a loan from the category of "High Risk" and by proving that acceptance of the "High Risk" loan was unintentional.64 Finally, the Illinois Act caps assignee damages at the amount of remaining indebtedness plus costs including attorneys' fees. 5 61 See 815 ILL. COMP. STAT. 137/135(d)(1) (2007). 62 Note that charging fees or closing costs greater than 5% of the value of the loan also makes a loan "High Risk." See 815 ILL. COMP. STAT. 137/10 (2007). See 815 ILL. COMP. STAT. 137/135(d)(1) (2007). See 815 ILL. CoMP. STAT. 137/135(e) (2007). 6' See 815 ILL. COMP. STAT. 137/135(d)(2) (2007). [Vol. 2008 RETHINKING ASSIGNEE LIABILITY A. Federal Solutions: The Weak Federal Approach At the federal level, there are three pertinent laws governing the mortgage market. These laws are ineffective. Primarily, the Home Ownership and Equity Protection Act ("HOEPA") is a narrow measure that prohibits many predatory practices.66 HOEPA includes assignee liability, allowing a borrower to collect against the trusts for the remaining balance of the loan, plus principal paid, attorneys' fees, and costs.67 However, HOEPA assignee liability only applies to refinancing, and even then, only to "high cost" loans. The statute defines "high cost" as loans with an interest rate 8% greater than a comparable treasury security for a first lien or 10% for a second lien. Alternatively, a loan is "high cost" if the total points and fees exceed 8% of the loan or $400.68 "High cost" loans, and thus HOEPA assignee liability coverage, amount to less than 1% of all originations. 9 Second, the Fair Housing Act ("FHA") and the Equal Credit Opportunity Act ("ECOA") prohibit race, gender, or religious discrimination in the provision of housing financing. ° Either disparate treatment or disparate impact is enough to prove a violation under the FHA or the ECOA.7' See 15 U.S.C. §§ 1601-1649 (2006); 12 C.F.R. § 226.32(d) (2007); Peterson, supra note 36, at 2227 (pointing out that HOEPA bans negative amortization, penalty interest increases, balloon payments, and asset- based lending). 67 See 15 U.S.C. § 1641 (2006) (tempering assignee liability by creating essentially due diligence shields for any violation of HOEPA that could not be discovered based on a reasonable review of the loan documents). 6 See 15 U.S.C. § 1602(aa)(1) (2006). 69 Azmy, supra note 19, at 355. Table 3 demonstrates that HOEPA loans are far less than 1% of the standard. Some suggest that many lenders skirt the HOEPA requirements by lending at interest rates and fees just under the HOEPA shield. Pyle, supra note 23, at 1923. 70 42 U.S.C. §§ 3604-3605 (2006) (FHA as implemented by 24 C.F.R. § 100); 15 U.S.C. § 1691 (2006) (ECOA as implemented by 12 C.F.R. § 202); Peterson, supra note 34, at 2228. 71 See Davis v. New York City Hous. Auth., 278 F.3d 64, 81 (2d Cir. 2002) (holding that to establish a violation of the FHA, a plaintiff need not No. 3:1021 ] COLUMBIA BUSINESS LAW REVIEW Because the riskiness of subprime borrowers is hard to calculate, this pair of federal laws makes it difficult for traditional bank institutions to enter the subprime market without incurring disparate impact lawsuits. The federal laws are quite limited. HOEPA applies to such a small percentage of loans that it does little to mitigate abusive practices. 2 The federal laws have already failed to control the market during the subprime build-up and crash. There is little reason to believe they will work now. B. State Solutions: The Problem of Regulatory Capture Finding the federal legislation ineffective, many states began directly regulating lending. The first state to pass a predatory lending law was North Carolina in 1999."3 The vast majority of these states passed "mini-HOEPA" laws." Mini-HOEPA laws usually follow the same basic prohibitions as HOEPA, but with lower interest rate or fee triggers, and often with more draconian remedies. 5 Roughly twenty-three states also adopted, like HOEPA, assignee liability provisions. These range from being no different than HOEPA assignee liability, to Nevada, where the assignee show discriminatory intent but need only prove that the challenged practice has a discriminatory effect). 72 A price ceiling occurs when originators refuse to lend in amounts that put a loan within HOEPA. HOEPA's provisions are sufficiently draconian that the interest rate trigger under HOEPA becomes the market ceiling. Yet, this essentially truncates the credit market. To the extent this ceiling is non-binding, that borrowers would not accept a loan with interest rates that high, there is no inefficiency. However, if the ceiling prevents loans that otherwise would be made, then there must be a determination of whether that loan would have been positive for that particular individual. " See N.C. GEN. STAT. § 24-1.1(e) (2007) (demonstrating an example of a mini-HOEPA law). 71 See Ho & Pennington-Cross, supra note 14, table 2 at 216. 7' For example, the New York law, effective in 2003, triggers at 8% for a first lien and 9% for a second lien. It also imposes assignee liability, but unlike HOEPA that assignee liability is strict-it functions regardless of knowledge or diligence. N.Y. BANKING LAW § 6-1(13) (McKinney 2003). [Vol. 2008 could be liable for up to three times the amount of damage.7 6 Only eleven states have no predatory lending laws of any consequence.7 State-level predatory lending laws are important for two reasons. First, even if only enacting HOEPA at the state level, they enable enforcement by the state Attorney General. Second, in enacting lower triggers or more draconian assignee liability, states can experiment with what combination quashes predatory lending without damaging the subprime market. However, state-level regulation creates unique problems of regulatory capture. States often lack the power to enact mortgage laws on their own. Georgia's attempt to use unlimited assignee liability demonstrates the problem. When Georgia proposed strict assignee liability in 2002, Standard & Poor's ("S&P"), one of the three authorized 76 See NEV. REV. STAT. § 598D.110(2)(a) (2007). States with assignee liability passed in time to be included in this paper: Arkansas, Colorado, Connecticut, District of Columbia, Florida, Georgia, Illinois, Indiana, Kansas, Kentucky, Maine, Massachusetts, Nevada, New Jersey, New Mexico, New York, North Carolina, Ohio, Oklahoma, South Carolina, West Virginia (as of the end of 2006). ARK. CODE ANN. § 23-53-101 (2008); COLO. REV. STAT. § 5-3.5-101 (2008); CONN. GEN. STAT. § 36a-746 (2008); D.C. CODE § 26-1151.01 (2008); FLA. STAT. § 494.00793 (2008); GA. CODE ANN. § 7-6A-1 (2008); 815 ILL. COMP. STAT. § 137/1 (2008); IND. CODE ANN. § 24-9- 1-1 (LexisNexis 2008); KAN. STAT. ANN. § 16a-5-201 (2008); KY. REV. STAT. ANN. §360.100 (LexisNexis 2008); ME. REV. STAT. ANN. tit. 9-A, § 8-101 (2008); MASS. GEN. LAws ANN. ch. 183C, § 1 (West 2008); NEV. REV. STAT. ANN. § 598D.010 (LexisNexis 2008); N.J. STAT. ANN. § 46:10B-22 (West 2008); N.M. STAT. ANN. § 58-21A-1 (West 2008); N.Y. BANKING LAW § 6-1 (McKinney 2008); N.C. GEN. STAT. § 24-1.1E (2008); OHIO REV. CODE ANN. § 1349.25 (West 2008); OKLA. STAT. tit. 14A, § 5-203 (2008); S.C. CODE ANN. § 37-23-10 (2008); W. VA. CODE § 31-17-1 (2008). In 2007, Tennessee and Rhode Island joined the list, making it twenty-three. R.I. GEN LAWS § 34- 25.2-1 (2008); TENN. CODE. ANN § 45-20-101 (2008). " Some articles seem to interpret generic consumer protection laws to be predatory lending laws, despite sometimes having been in effect for many years. Ho & Pennington-Cross, supra note 14, at 216. The eleven states with no predatory lending laws (as of the beginning of 2007) are: Alabama, Alaska, Arizona, Delaware, Hawaii, Iowa, Montana, North Dakota, Oregon, South Dakota, and Washington. 7" See generally Azmy, supra note 19. No. 3:1021] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LA W REVIEW national rating agencies, refused to rate subprime securitizations coming out of Georgia.7 9 Given that unrated loans are impossible to securitize, this essentially would have caused a total collapse of the Georgia subprime market. The Georgia legislature repealed the assignee liability provisions. The rating agencies flexed their muscles with similar results when New Jersey attempted to pass a draconian assignee liability law.8 ° State regulatory capture is especially problematic because there are only three rating agencies: Standard & Poor's, Moody's Investor Services, and Fitch Ratings."1 Due to SEC regulations, these are the only rating agencies with the authority to rate securities before they enter the national securities market, and all securities must be rated before they can be traded." The privileged raters are, of their own admission, on the side of subprime growth and stability or, in other words, more business. 3 They contribute to the campaigns of state political leaders, resorting to direct economic pressure, as they did in Georgia and New Jersey, when necessary. 4 When a state legislature attempts to pass an assignee liability law, it faces a well-organized and wealthy opposition. IV. CONCEPTUALIZING ASSIGNEE LIABILITY Whether at the state or federal level, assignee liability is a hot topic in the regulation of subprime lending. Many view 78 See Engel & McCoy, supra note 25, at 2099; Peterson, supra note 36, at 2243. 80 See Reiss, supra note 36, at 1048-49 (showing that at first New Jersey refused to amend its law under the pressure of the rating agencies, but then eventually was forced to submit); N.J. STAT. ANN. § 46:10B-27 (West 2008). s1 See Reiss, supra note 36, at 987. 82 See generally Reiss, supra note 36, at 987-91 (describing the "privileged" raters). See Reiss, supra note 36, at 988. See Glenn R. Simpson, Lender Lobbying Blitz Abetted Mortgage Mess, WALL ST. J., Dec. 31, 2007, at Al. [Vol. 2008 assignee liability as an integral part of any predatory lending regulatory scheme.8" Yet, while assignee liability has an intuitive appeal, it is not clear that it has been properly characterized to this point. In order to properly characterize the impact of assignee liability, researchers must put themselves in the "shoes" of the banks and trusts. Some studies have begun such an analysis. For example, several academics stress the importance of predictable costs under any regulation.8 The two most likely bank responses to an assignee liability law are either to perform due diligence on loans targeted for securitization or require solvency from the originators. 7 " See, e.g., Engel & McCoy, supra note 25; Venkatesan, supra note 57; Weiner, supra note 14. 86 See Engel & McCoy, supra note 25, at 2052; Peterson, supra note 36, at 2275; Weiner, supra note 14 (finding the real issue is uncertainty in assignee liability laws). However, it is hard to accept the rating agency's arguments that uncertain damages are incalcuable. Punitive damage insurance, although sometimes prohibited by law, is available. This indicates that these damages are not entirely incalculable. MINN. STAT. ANN. § 60A.06 (West 2008) (demonstrating only vicarious liability insurance for punitive damages is allowed in Minnesota); Brown v. Maxey, 369 N.W.2d 677, 686-88 (Wis. 1985) (indicating that punitive damage insurance is allowed). 87 There are essentially four options for an investment bank when faced with an assignee liability law that covers a wider variety of loans: (1) Accept the risk, continue trading as normal; (2) Perform due diligence on the loans; (3) Require solvency from originators; (4) Exit the market. All of these carry certain costs, and we can expect, all else equal, that the investment bank will choose the least costly option. (This would mean opportunity cost in the case of exit or actual cost from legal liability in other cases.) The bank's action depends on a careful calculation of these costs, and thus if the law allows for unlimited punitive damages, or any uncertain damages, many banks are likely to choose exit (for certainty) over the unpredictable cost of any of the other three options. Regulations would not be successful if the banks chose exit, given the benefits to housing and liquidity suggested earlier. See supra Part II(a). Moreover, it is exceedingly unlikely that in any binding legal regime a bank would choose merely to accept the risk. The only way a bank would accept the risk is if the gains to options (2) or (3) minus the cost of implementing (2) or (3) were greater than the cost of option (1). Thus, we are left with options (2) and (3). No. 3:1021 ] RETHINKING ASSIGNEE LIABILITY A. The Costs and Uncertainties of Due Diligence Review While due diligence review appears a plausible response to assignee liability, the remaining uncertainties and considerable costs make it an unlikely solution. Nevertheless, many academics expect banks to take this approach.'8 They cite the fact that automated loan review can cost as little as one dollar per loan.89 However, automated review would not be enough. There are several elements of creditworthiness that cannot be caught on automated review: originator's lying about employment history, home appraisals, or applicant income.9 Since assignees, under most laws, are liable for all of the predatory acts of the originators, this leaves significant, likely unacceptable, uncertainty. Automated review would need to be followed-up with an on-site review of a randomized sample of the loans. Due diligence review would quickly become an expensive and ineffective endeavor. Expanding both the coverage and extent of possible damages for assignee provisions would make review of loans a more high-stakes and, hence, even more expensive, proposition. Even worse, the review would have to be done before the bank securitizes the loans. Since many offerings are done on a to-be-announced basis, this sort of system 88 See, e.g., DELVIN M. DAvis & ELLEN SCHLOEMER, CTR. FOR RESPONSIBLE LENDING, STRONG COMPLIANCE SYSTEMS SUPPORT PROFITABLE LENDING WHILE REDUCING PREDATORY PRACTICES (2005), http://www. responsiblelending.org/pdfs/ipOlO-Compliance-Costs-0705.pdf; Engel & McCoy, supra note 25. Contra Silverman, supra note 13. 89 See DAvis & SCHLOEMER, supra note 88, at 6. Assignees could be liable for the fraud of the originator on the loan documents themselves in that the originator made a loan not based on true ability to repay. Most of the laws deal, in terms of culpability, only with whether the loan meets the interest rate threshold, not whether the assignee knew about the bad acts of the originator. See 815 ILL. COMP. STAT. 137/135(d)(1) (2007). 90 Moreover Fannie Mae and Freddie Mac, the government sponsored entities, have already instituted loan review. See Engel & McCoy, supra note 25, at 2095 (suggesting that there would still need to be manual spot loan reviews). "' See Silverman, supra note 13, at 515-17. COLUAMIA B USINESS LA W RE VIE W [Vol. 2008 would be slower and, therefore, more costly. Finally, after making the process inefficiently slow and costly, the bank would still be left with a certain probability that it missed some very egregious loans that will later come back to haunt it. Loan review is an expensive way to police the subprime market. B. Requiring Solvency: Practical and Legal Obstacles Requiring solvency from originators appears, at first, to be a more realistic option. Presumably, under any assignee liability regime, the assignee, if sued, can implead the originator of the loan. However, the originator is often insolvent; otherwise the borrower would have sued the originator in the first instance.2 One solution is for assignees to require solvency from the originator. This requires less review of loans and could be based on average solvency requirements garnered from readily available characteristics of the loan pool, yet this approach is likely more costly than due diligence. Indemnification is not costless. First, the bank must determine how much solvency to require.93 Second, ensuring solvency has costs. An originator can appear solvent, but then transfer funds, declare bankruptcy, or otherwise disappear after the loans are sold. While larger subprime lenders would be more trustworthy, those outfits are also 92 The difficulties of litigating against a trust are well documented. There are issues of finding who is the right party to sue, ensuring that personal jurisdiction is satisfied, and timing (in a foreclosure case) is of paramount importance. See Peterson, supra note 36, at 2265. " In fact, this question could be quite costly and by itself makes this essentially the same as reviewing loans. Either the bank would need to know exactly how much insolvency was required, and therefore have to conduct the same reviews as before, or the bank would need to require an excess amount to make sure it was covered. This excess solvency requirement would shift the costs to the originator, but would shift them at an inefficiently high level, thus exaggerating the impact of any predatory lending law. Thus, it would be quite important to determine the appropriate level of solvency required, and determining that level may be just as costly as reviewing the loans. No. 3:102 11 RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LAW REVIEW presumably least likely to engage in predatory lending.94 The lenders most likely to be predatory, and therefore most likely to create a need for indemnification, would also be the costliest ones to indemnify. Given the large costs involved, the banks would likely have to create subsidiaries for extending credit to subprime markets or extend subprime credit directly. Yet the extension of subprime credit directly, or even through a subsidiary, has other costs. The first is reputational. Many banks have traditionally avoided subprime lending because it increases the risk of foreclosure. Banks dislike foreclosure because they are expensive and create potential public relations fiascos.9" The second, however, is more serious. Federal laws-the FHA and the ECOA-make it illegal to discriminate in credit transactions.96 Mere disparate impact is enough for liability under these laws; they require no proof of intent.97 When faced with such a suit, it becomes incumbent on the lender to demonstrate that there was a business necessity to the " Although given the action against many of the larger subprime lenders, this might not be true. See Erick Bergquist, Experts Say Lehman Case is Warning, Not Precedent, AM. BANKER, June 18, 2003, at 6 (referring to a case involving predatory lending by First Alliance on large- scale level). " See Engel & McCoy, supra note 36, at 1290-91. Banks, businesses based largely on trust, dislike bad press, which foreclosures tend to create. As subprime loans are more likely to foreclose, traditional banks do not want their prime customers to feel threatened by the negative public image foreclosures can create. Moreover, carrying subprime loans, due to the risk involved, may force the bank to hold a higher loan-loss reserve ratio. See id. at 1291. 96 See 42 U.S.C. §§ 3604-05 (2006) (FHA-which was implemented by 24 C.F.R. § 100); 15 U.S.C. § 1691 (2006) (ECOA as implemented by 12 C.F.R. § 202). " See Davis v. New York City Hous. Auth., 278 F.3d 64, 81 (2d Cir. 2002) (finding discriminatory effect sufficient to prove violation); Charleston Hous. Auth. v. U.S. Dep't. of Agric., 419 F.3d 729 (8th Cir. 2005); Anast v. Commonwealth Apartments, 956 F. Supp. 792 (N.D. Ill. 1997) (finding violation need not be intentional discrimination). [Vol. 2008 RETHINKING ASSIGNEE LIABILITY practice and that there is no less discriminatory means to do it. 98 As demonstrated in Table 4, subprime credit is disproportionately extended to minority applicants. Low to moderate income applicants are very difficult to differentiate based on their individual characteristics.99 Moreover, simply increasing price does not draw in higher quality applicants, but instead brings in riskier ones.'O° Thus, many subprime lending decisions are offered at rates based on the community in which the individual lives. Troubled communities with a higher minority population present a greater risk of stalled housing values, and therefore they systematically receive higher rates.'0 ' Discrimination claims are harder to make out against subprime specialists because they do not make loans to other parts of the population. Instead, it is the large banks and their subsidiaries that invite comparison and, consequently, liability. In January of 2008, the City of Baltimore proved that the FHA is no idle threat when it brought suit against Wells Fargo for reverse- redlining disproportionately African-American neighborhoods of the city.102 Thus, bank subsidiaries or " FDIC statement of policy says that the defense to FHA disparate impact is to demonstrate the practice is justified by "business necessity" and that there is no less discriminatory alternative. See Policy Statement on Discrimination in Lending, 59 Fed. Reg. 18267-18268 (Apr. 15, 1994), available at http://www.fdic.gov/regulations/laws/rules/5000-3860.html. " Engel & McCoy, supra note 36, at 1292; Silverman, supra note 13, at 534-35 (suggesting that things like employment history and credit history often do not tell the whole story). 100 See Michael Klausner, Market Failure and Community Investment: A Market-Oriented Alternative to the Community Reinvestment Act, 143 U. PA. L. REV. 1561, 1566-68 (1995) (noting that if lenders made loans to the highest bidders, the borrowers would be the riskiest borrowers). 101 See Silverman, supra note 13, at 535-36 (noting that race might be used as a proxy, although illegal, as declining neighborhoods present greater risk of falling housing values). 102 Reverse-redlining refers to the practice of targeting certain communities for predatory loans. Complaint for Declaratory and Injunctive Relief and Damages, Mayor & City Council of Baltimore v. Wells Fargo Bank, N.A., No. L08CV062 (D. Md., Jan. 8, 2008), available at http://www.relmanlaw.com/City%20of%2OBaltimore%20v.%2OWells%20 No. 3:1021 ] COLUMBIA BUSINESS LA W REVIEW banks entering the subprime market themselves may face prohibitive costs. Whether it is due to the solvency of originators or reviewing of loans, the securities market will face significant costs from assignee liability. Depending on how high those costs are, the subprime market could disappear, recreating the incomplete credit market of the early 1990s. 10 3 C. Litigation Barriers: Assignee Liability and Obstacles to Enforcement Due to the difficulty of litigating an assignee liability claim, there is likely to be systematic under-enforcement. This not only means the law will entail costs, but that it will also likely do so without completely stopping predatory lending activity. 104 The litigation barriers are of two types. First, there is the difficulty of finding and retaining a lawyer. Many people do not seek attorneys until they are on the verge of foreclosure. At that point, they have little or no money and are unlikely Fargo%20-%2008-cv-62%20-%2oComplaint.pdf; Ben Nuckols, Baltimore Sues Wells Fargo for Subprimes, ABC NEWS ONLINE, Jan. 8, 2008, http://abcnews.go.com/Business/wireStory?id=4105228. 10 See Chinloy & MacDonald, supra note 12, at 11. But see Engel & McCoy, supra note 25, at 2100-01. 104 It would not be possible for companies to adapt diligence standards to exactly match the potential enforcement costs. Most assignee laws apply draconian liability unless a sufficient, set level of due diligence is taken in regards to all the loans made by an originator. See, e.g., Engel & McCoy, supra note 25. A company is subject to serious penalties in all cases based on its generic diligence practice. There could be no "reducing of costs" in this regime. Section 204 of the recently introduced Mortgage Reform and Anti-Predatory Lending Act requires a company to use a specific sampling procedure with all of its loans in order to avoid assignee liability on any individual loan. Mortgage Reform and Anti-Predatory Lending Act of 2007, H.R. 3915, 110th Cong. § 204 (2007), available at http://www.house.gov/apps/list/press/financialsvcs-dem/subprimeleg.pdf. Under this system, assignee is an on-off switch for all loans made by an originator based on the originator's generic diligence practice. So long as the under-enforcement does not make the law meaningless, the costs are likely to exceed the enforcement gains. [Vol. 2008 RETHINKING ASSIGNEE LIABILITY to continue pursuing the claim after the foreclosure ends. 1° The only recompense that attorneys can hope for are their fees, under a statute that requires the paying of those fees, and even those are contingent upon winning. They are not likely to be able to compensate for the chance of losing the claim against the lender. Moreover, it is almost impossible to certify a class to make these cases worthwhile. Not only are the cases time-sensitive and, thus, plaintiffs are unlikely to want to continue long enough to sustain one, but also, since the claim is essentially fraud, it is very difficult to demonstrate the necessary shared reliance element.0 ° The second litigation barrier involves transaction costs for the actual litigation. The Mortgage Electronic Registration System (MERS) has replaced the prior practice of re- registering mortgages with the county records office each time a lender assigns it. When a trust takes control of a securitized pool, it typically registers with MERS and therefore MERS is the only name that appears in the county register. When foreclosures occur, MERS brings the foreclosure in its name.0 7 Thus, even determining who to sue is not easy, raising transaction costs for precisely the type of clients and attorneys already particularly sensitive to them. Moreover, the under-enforcement problems cannot be solved by punitive damages or grossly exaggerated penalties toward the assignees.' 8 105 Peterson, supra note 34, at 2267-68. 'o See id. at 2268. 107 See id. at 2266. 108 There are two reasons for this. First, the industry currently is very hesitant to rate punitive damages, and even though it likely could, it would be a cautious estimate-over-estimating the true probability and leading to deadweight loss. See generally Weiner, supra note 14. Second, as pointed out previously, the current regimes, with the "due diligence safe harbors," make any focus on costs irrelevant to the due diligence performed by the lenders. Mortgage Reform and Anti-Predatory Lending Act of 2007, H.R. 3915, 110th Cong. § 204 (2007), available at http://www. house.gov/apps/list/press/financialsvcs-dem/subprimeleg.pdf. Once they reach a certain level they are protected and have no incentive to perform further review. Moreover, this causes serious problems for loan innovation. The higher the costs go, especially if laws cease to provide due No. 3:10211 Compounding the cost of assignee liability regimes and the under-enforcement problems is the fact that assignee liability laws do not stop predatory lending. Since the laws involve interest rate triggers for assignee liability, they leave open the possibility of predators merely shifting to use of balloon payments, negative amortization, or other practices. Loans can still have predatory characteristics without charging high interest rates. For example, they can still be made to people who cannot possibly afford them, while including balloon payments to maximize the payout before default. In that case, only originators are actionable. Assignees are untouchable, leaving the same problems that caused the 2007 crash. Assignee liability may quash the legitimate subprime market, while leaving predatory lending alive and well.1"9 V. EMPIRICAL ANALYSIS OF ASSIGNEE LIABILITY A study of the impact of assignee liability laws in the United States confirms the theoretical case against assignee liability. By looking directly at interest rates and the prevalence of high-cost loans, and by using states as controlled experimental comparisons, this Note attempts to test three hypotheses. The first hypothesis is that predatory lending laws make subprime lending more expensive. The second hypothesis is that predatory lending laws give consumers more confidence, and therefore offset some of the expense by introducing new, less desperate consumers to the market. Finally, the third hypothesis is that predatory lending laws merely act as interest rate caps on loans, rather than stopping predatory behavior. diligence safe harbors, the more likely the industry will be to adopt specific form transactions. This would mean that all loans would have virtually the same terms in order to ensure the loan did not run afoul of the draconian penalties. Venkatesan, supra note 57, at 191. 109 Even assignee liability's staunchest supporters worry about and, indeed, accept that assignee liability will take a toll on legitimate subprime lending. Venkatesan, supra note 57, at 211. [Vol. 2008_COL UAMIA B USINESS LA W RE VIE W Since the first and second hypotheses push in opposite directions, it becomes clear that just looking at the incidence of subprime lending does not tell the whole story. Further, since the laws target only loans over a certain interest rate, it is unclear whether predatory practices are being prevented or just changed. The story is quite complex, suggesting that predatory lending laws, by including assignee liability provisions, may be crowding out legitimate subprime lending, lulling consumers into a false sense of confidence without actually policing predatory practices. A. A Note on the Data The data set for this project came from required reporting under the Home Mortgage Disclosure Act (HMDA)." ° It includes the vast majority of U.S. mortgage transactions and is the universal standard for mortgage research."' The data are for the years 2004 to 2006. Since there are well over one- hundred million loans for that period, a random sample of roughly 4% is reported here.11 2 This left roughly 3.8 million loans. Some summary statistics of relevant variables are presented in Table 2 through Table 11. While focusing on subprime loans, this Note will refer to them in the data as "high cost" loans. 11' See Home Mortgage Disclosure Act, 12 U.S.C. §§ 2801-10 (2006); 12 C.F.R. § 203 (2007); Federal Financial Institutions Examination Council, Home Mortgage Disclosure Act, http://www.ffiec.gov/hmda/ default.htm (last visited Dec. 3, 2008) (source of data). . See, e.g., Avery et al., supra note 14, at A124; Bostic et al., supra note 7; Paul S. Calem et al., Neighborhood Patterns of Subprime Lending: Evidence from Disparate Cities, 15 HOUSING POL'Y DEBATE 603 (2004); Harvey & Nigro, supra note 16, at 483; Keith D. Harvey & Peter J. Nigro, Do Predatory Lending Laws Influence Mortgage Lending? An Analysis of the North Carolina Predatory Lending Law, 29 J. REAL EST. FIN. & ECON., 435 (2004); Ho & Pennington-Cross, supra note 14, at 218; Wachter et al., Subprime Lending: Neighborhood Patterns Over Time in US Cities, (Institute for Law and Economics, Research Paper No. 06-19, 2006). 112 Similar to the approach used by Bostic et al., supra note 7, this study used a STATA random sampling algorithm to take a 5% random sample. No. 3:1021] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LA W REVIEW B. Past Empirical Research Studies of predatory lending are part of a rapidly growing field. Early studies focused on the first state to enact a major predatory lending law, North Carolina. These studies produced conflicting results. One study, using a proprietary data set, determined that the North Carolina law pushed lenders away from riskier credit extensions. It reduced subprime lending to lower-income borrowers, causing the total amount of subprime lending to decrease."' In 2004, a study using HMDA data concluded that the amount of subprime lending in North Carolina declined because of a decrease in applicants rather than an increase in the rejection rate."4 This suggested that the North Carolina law curbed some abusive practices without harming the flow of subprime credit. A 2004 study determined that a decrease in subprime loans reflected mostly the halting of loans with predatory aspects.' There are also many national predatory lending studies. In 2003, one study compared anti-predatory lending laws in Chicago and Philadelphia."6 In Chicago, where the law targeted banks, the chance of a subprime origination remained unchanged, but the portion of the market serviced by non-bank entities increased.1 7 In Philadelphia, where the law covered a greater percentage of loans, it caused a dramatic decline in subprime lending, virtually removing lines of credit for low-income and minority applicants."8 In 2006, a national study determined that when a law was average or above average in its restrictive effect, there 113 See Gregory Elliehausen & Michael E. Staten, Regulation of Subprime Mortgage Products: An Analysis of North Carolina's Predatory Lending Law, 29 J. REAL EST. FIN. & ECON. 411, 426 (2004). 114 See Harvey & Nigro, supra note 111. 115 See Quercia et al., supra note 12, at 593. 16 See Harvey & Nigro, supra note 16. 17 Id. at 488-89 (indicating it may have actually shifted the market toward predatory lending). 18 Id. at 489. [Vol. 2008 RETHINKING ASSIGNEE LIABILITY was a significant decline in subprime originations.119 Also in 2006, a study using cross-border comparisons examined the impact of subprime laws using an index to account for the strength of the predatory lending law.12 ° The authors found that there was a small or insignificant change in the probability of applying for or originating a loan. However, there was also a large decrease in the probability of being rejected for a loan.121 The authors theorized that while enacting a strong predatory lending law increases the cost of subprime lending, the feeling of security such a law creates draws less risky and previously wary consumers to the market, causing conflicting effects.122 Further, in 2006, one study found that predatory lending laws shifted subprime lending from African-American neighborhoods to slightly higher-income Hispanic neighborhoods.1 23 In a 2006 study using proprietary data, the authors determined that states with stronger laws saw loans with predatory characteristics (prepayment penalties, balloon payments, and steering) decrease 20%.124 However, they found no decrease in overall subprime volumes. 125 Finally, in 2007, a study analyzing data from forty-one states with predatory lending laws was the first to analyze the impact of assignee liability.126 This study formulated an index of the strength of various legal provisions, including an enforcement factor for methods like assignee liability. 27 The 119 See Elliehausen & Staten, supra note 113, at 16. 120 See Ho & Pennington-Cross, supra note 14. 121 Id. at 223. 122 Id. at 212; see generally George A. Akerlof, The Market for "Lemons": Quality Uncertainty and the Market Mechanism, 84 Q. J. ECON. 488 (1970). 121 See Wachter et al., supra note 111. This study introduces the idea that predatory lenders are very adept at shifting practices to avoid law coverage. 124 See WEI LI & KEITH S. ERNST, CTR. FOR RESPONSIBLE LENDING, THE BEST VALUE IN THE SUBPRIME MARKET 11 (2006), http://www.responsible lending.org /pdfs/rrOlO-StateEffects-0206.pdf. 125 Id. at 13. 126 See Bostic et al., supra note 7. 127 Id. No. 3:1021 ] authors found that the presence of a predatory lending law has little impact on applications, originations, or rejections. However, the presence of more restrictive laws reduces originations and increases rejections. They determined that variation in enforcement mechanisms (including the use of assignee liability) has no effect on originations or applications, but does marginally increase rejection rates.12 C. Contributions to the Empirical Study This Note utilizes four novel approaches in an attempt to address many of the problems facing subprime mortgage research. First, this Note utilizes propensity score matching to prove, rather than assume, identical regions are in the treatment and non-treatment conditions. Second, this Note uses the new HMDA identifier for subprime loans to create a more accurate picture of the market. Third, this Note uses a field published index of subprime regulatory severity in order to assess, from the industry's standpoint, the difference between regulatory provisions. Fourth, this Note includes a state-specific focus to more directly study certain impacts of regulatory solutions. All of these innovations offer grounds for a new approach to the study of subprime regulation. 1. Propensity Score Matching First, unlike the cross-border approach, or event approach, used by many other studies, this Note uses a propensity score matching method to compare a broader sample.'29 Selection bias and omitted variables are serious 128 Id. 129 The cross-border approach uses regions on either side of a state border as a proxy for a single economic region, in which the only difference is the predatory lending law. A time series approach looks at the same region both before and after the passage of a law, assuming the only difference is the enactment of the law. See, e.g., GREGORY ELLIEHAUSEN ET AL., CREDIT RESEARCH CTR., MCDONOUGH SCH. OF Bus., GEORGETOWN UNIV., THE EFFECTS OF STATE PREDATORY LENDING LAWS ON THE AVAILABILITY OF SUBPRIME MORTGAGE CREDIT (2006), http://www.business. gwu.edu/research/centers/fsrp/pdf/M38.pdf. See also Bostic et al., supra note 7; Ho & Pennington-Cross, supra note 14. COL UMBIA B USINESS LA W RE VIE W [Vol. 2008 RETHINKING ASSIGNEE LIABILITY problems in any empirical analysis.13 ° This Note suggests propensity score matching may remove selection problems, without sacrificing flexibility of analysis."' Propensity score matching was first developed by H.W. Rubin in the 1970s.132 It provides a method for controlling selection bias in studies. 33 Selection bias occurs in many forms, but in studies such as this, the biggest danger is that comparison groups differ along an unobserved characteristic. That difference makes the delineation between treatment and non-treatment non-random. 34 For example, whether a state decides to enact a subprime lending law may have much to do with the economic conditions of subprime lending in the state itself and, thus, the characteristics of the people who live there. This makes direct comparison of the effect of the different laws passed in two different states difficult. The direct comparison would make as much sense as "0 Selection bias involves the situation wherein inclusion into the treatment (having a predatory law) is not completely random, but rather dependent on characteristics of the subject. Omitted variable bias refers to the problem when a statistical model is missing an independent variable that would otherwise explain some of the variance in the dependent variable. "3 Studies which fail to control for omitted variables include: Avery et al., supra note 14; Calem et al., supra note 111; Elliehausen & Staten, supra note 113; Harvey & Nigro, supra note 16; Harvey & Nigro, supra note 111; KEITH ERNST ET AL., CTR. FOR RESPONSIBLE LENDING, NORTH CAROLINA'S SUBPRIME HOME LOAN MARKET AFTER PREDATORY LENDING REFORM (2002), http://www.responsiblelending.orgpdfs/HMDAStudy-on NCMarket.pdf; LI & ERNST, supra note 124; Quercia et al., supra note 12. 132 See Paul R. Rosenbaum & Donald R. Rubin, The Central Role of the Propensity Score in Observational Studies for Causal Effects, 70 BIOMETRIKA 41, 41-55 (1983). 133 See, e.g., James Heckman et al., Characterizing Selection Bias Using Experimental Data, 66 ECONOMETRICA 1017 (1998). 134 See, e.g., Donald B. Rubin, Estimating Causal Effects of Treatments in Randomized and Non-Randomized Studies, 66 J. EDUC. PSYCHOL. 688, 688-701 (1974); Paul R. Rosenbaum & Donald B. Rubin, Constructing a Control Group Using Multivariate Matched Sampling Methods that Incorporate the Propensity Score, 39 AM. STATISTICIAN 33, 33-38 (1985). No. 3:10211 COLUMBIA BUSINESS LA WREVIEW comparing the effectiveness of two slicers, one with an apple, and one with an orange, to declare which is the better slicer. Researchers combat this problem by trying to find similar "apples" in the two different states. The two most common means of addressing this problem are using a discontinuity approach or analyzing only areas that are across state borders from one another. 3 ' A discontinuity approach compares the same apple to itself-the same group of people in a state before and after the passage of a law. The cross- border approach assumes that economic regions are not contained by state borders, meaning that researchers can test two halves of the same apple by examining the effect of the law on each side of a state line.'36 Either approach limits the universe of possible comparisons significantly. While both discontinuity and cross-border approaches are widely accepted, propensity score matching is a more flexible approach, one that proves, rather than assumes, regions are identical. For each data set used in this Note, the author ran a propensity matching algorithm to calculate the odds whether a particular census tract was covered by an assignee liability law. Essentially, the author attempted to control for economic conditions by equalizing the means of various demographic variables, on a census tract level, based on whether or not that tract received the treatment (an assignee liability law).3 7 Table 9 (unmatched) and Table 10 (matched) demonstrate the outcome. While this approach is not perfect, it demonstrates a novel technique that is more flexible and inclusive than those used previously. 38 135 See, e.g., Bostic et al., supra note 7 (state border approach); ELLIEHAUSEN ET AL., supra note 129 (event series). 136 Ho & Pennington-Cross, supra note 14, at 218. ... This paper first calculated propensity scores and then used "nearest neighbor" matching to match similar observations in the treatment and control. The maximum allowable difference between neighbors is limited in order to acquire demographic means statistically identical (or as close as possible to identical) between the two groups. 138 For example, this analysis assumes that median income, type of housing, percentage of minority groups, wealth disparity, and population, captures regions that are sufficiently similar economically to avoid selection bias. However, this might not be the case for several reasons: (1) [Vol. 2008 RETHINKING ASSIGNEE LIABILITY 2. Using Rate Spread Instead of Subprime Lender List Second, this Note uses the HMDA supplied interest rate data (rate spread) to determine whether a loan is "high cost," instead of the more typical Department of Housing and Urban Development (HUD) list of subprime lenders.139 While either approach introduces considerable noise into the sample, using the rate spread data, which has only been available since 2004, introduces noise that is more controllable. HUD compiles a list each year of lenders that it believes to be subprime lenders. It gathers this list by determining which lenders, based on other characteristics of HMDA, might be subprime, and then contacting them to ask if they are subprime specialists. If subprime loans account for 50% of a lender's business, or if the lender self-identifies as a subprime lender, then all of that lender's loans are counted as subprime. This approach introduces non-random noise, for which other variables cannot control. 4 ° The rate spread interest rate pricing of loans is the approach adopted here. Rate spread refers to the difference between the interest rate on the mortgage in question, and a comparable Treasury note. A rate spread appears in HMDA data only if it exceeds three points on a first lien or five points on a second lien.141 However, Treasury notes are set to a "yield curve," which constantly changes the underlying interest rate. The shape and position of the yield curve the demographics might not be the relevant ones to determine an economic region, (2) there may be other, non-economic related factors that influence subprime lending, (3) the treatment effect specified, assignee liability law, may be too limited for the overall analysis. Nevertheless, propensity score matching could provide an excellent approach for further study. 139 Note the data supplied by HMDA is actually "rate spread," defined as three percentage points above the comparable treasury note. See FEDERAL FINANCIAL INSTITUTIONs EXAMINATION COUNCIL, A GUIDE TO HMDA REPORTING (2003), http://www.ffiec.gov/hmda/pdf/2004guide.pdf; Bostic et al., supra note 7. 140 See Calem et al., supra note 111. 141 See A GUIDE TO HMDA REPORTING, supra note 139. No. 3:1021] determines both the absolute and the relative prices of short and long term notes. 142 Therefore, a shift in the yield curve can cause loans to be, or not to be, reported despite no change in interest rate of the loan itself.' Moreover, in 2004, there was a large shift in the yield curve, causing it to flatten-out and price long-term and short-term loans similarly. 4 1 Yet, since that time, the yield curve has changed very little, making comparisons between 2005 and 2006 relatively accurate and allowing one to control for 2004 with year dummy variables.'45 While this approach still creates noise, it is predictable. 46 3. Comparing Assignee Liability Provisions Third, in order to gauge the severity of assignee liability provisions, this Note uses an index developed by S&P.1 47 142 See HUD Subprime and Manufactured Home Lender List, http://www.huduser.org/datasets/manu.html (last visited Dec. 3, 2008). 143 See Avery et al., supra note 14, at A142 (explaining the HMDA yield curve effect). 14 See id. 141 See, e.g, id. at A143; Frequently Asked Questions About the New HMDA Data, http://www.federalreserve.gov/newsevents/press/bcreg/bcreg 20060403al.pdf (last visited Dec. 3, 2008). 146 Part of the previous reluctance of researchers to use this data was due to suggestions by lenders and HUD that there were reporting errors in 2004, the first year of rate spread information collection. Bostic et al., supra note 7, at 18. However, some of the reported inconsistencies were from government officials at HUD calling to inquire about high levels of subprime loans from a lender that had not identified itself as such. The lenders would generally insist they were not per se subprime lenders, claims of which there is dubious validity. See HUD Subprime and Manufactured Home Lender List, supra note 142. There is precedent for lenders underreporting the amount of subprime lending in which they engage. See Julia Forrester, Still Mortgaging the American Dream: Predatory Lending, Preemption, and Federally Supported Lenders, 74 U. CIN. L. REV. 1303, 1305-06 (2006) (noting that Citigroup admitted in 2005 that it had continued to make subprime loans despite a public policy of not making those loans, and after a large judgment paid in 2001 for making predatory loans). 147 See Standard & Poor's, Criteria: Anti-Predatory Lending Law Update, RATINGSDIRECT, Oct. 20, 2006. COL UMBIA B USINESS LA W RE VIE W [Vol. 2008 This index, based on data used to accurately price securitizations from states with assignee liability laws, is a more consistent approach than the legal indexing used in other studies. 48 The S&P index calculates the "loss severity" for each state based on a calculation of the potential relative cost to the assignee." 9 For example, it calculates that all states have a loss severity index of 119 because of HOEPA. It then proceeds to calculate the loss severity based on each state law. These indexes range from thirty-seven for a weak Ohio law, to 275 for a strong North Carolina law.' 0 In order to analyze the ratings, the variable was "mean-centered," by calculating the mean and subtracting it from each datapoint. This allows for calculation of effects based on a state having stronger or weaker laws than the average state. The S&P index is an important tool because it demonstrates how S&P rates loans coming out of these states. 5' Since the ratings given to bonds based on subprime loans coming out of these states is the basis upon which people choose to invest, using this rating should most accurately reflect actual investor behavior. 52 4. State-Specific Focus Finally, the analysis concludes with an examination of three specific state pairs. This isolation attempts to weed " See, e.g., Bostic et al, supra note 7; Ho & Pennington-Cross, supra note 14. Both studies created a legal index that assigned different normalized values to regimes, based on how draconian the authors believed those regimes to be. 149 See Criteria: Anti-Predatory Lending Law Update, supra note 147, at 6. 150 See id.; N.C. GEN. STAT. § 24-1.1(e) (2007); OHIO REV. CODE ANN. § 1349.25 (West 2008). 151 This approach is unlike the legal index popularized by Bostic et al., supra note 7, at 12 and Ho & Pennington-Cross, supra note 14, at 214, which operates as a post hoc indexing. See generally, ELLIEHAUSEN ET AL., supra note 129, at 12-13. 152 Weiner, supra note 14, at 537 (noting that agencies' refusal to rate such bonds discourages investment in them). No. 3:10211 RETHINKING ASSIGNEE LIABILITY out the "lemon effect" first discovered by George Akerlof. 15 1 The general idea is that it is difficult or impossible for consumers to determine whether they are getting "ripped off' in an unregulated market."' By passing a law, regulators lower that transaction cost, and thus increase the demand for subprime loans.1 5 First, this Note examines the "placebo effect" of subprime laws by comparing Pennsylvania and Delaware. Pennsylvania put into effect an anti-predatory lending law in 2002 that is virtually identical to HOEPA, but without explicit assignee liability. 56 Meanwhile, Delaware has no subprime lending law of its own. The Pennsylvania law changes little legally, but may assuage the consumer into greater confidence and thus create a "placebo effect." Second, this Note compares the Pennsylvania law with a Florida law also put into effect in 2002. The Florida law is essentially a pure "mini-HOEPA" law, which mirrors HOEPA exactly. 7 Thus, this comparison can be used to further test the placebo effect of passing a law with assignee liability that causes little actual legal change. Finally, the Note compares California's approach to its "mini-HOEPA" law with New York's approach. These two laws are identical, except that New York includes an assignee liability provision that is more aggressive than that See George A. Akerlof, The Market for "Lemons": Quality Uncertainty and the Market Mechanism, 84 Q. J. ECON. 488, 488 (1970). "' The "lemon effect" suggests that creditworthy borrowers who were previously not part of any credit market are wary of predatory loans but too unsophisticated to realize the difference. Therefore, these more desirable borrowers stay out of the market until regulators declare it safe by passing a law. 155 See, e.g., Akerlof, supra note 153, at 490-91; Ho & Pennington- Cross, supra note 14, at 212. 156 See 15 U.S.C. §§ 1601-66 (2007) (HOEPA); 63 PA. STAT. ANN. § 456.503 (West 2007) (Pennsylvania Law). 157 See FLA. STAT. ANN. § 494.00793 (West 2007); Bostic et al., supra note 7, at 4-5. COL UMBIA B USINESS LA W RE VIE W [Vol. 2008 included in HOEPA. California specifically excludes assignee liability.158 D. Theoretical Models This paper expands upon existing research by offering essentially four basic specifications at the full sample level, and one basic specification at the state level. In all cases, the data set is restricted to the year-by-year propensity score matched data. The dependent variables are: Rate Spread, High Cost, and Securitized. Rate Spread examines how interest rates on subprime loans change depending on the presence of an assignee liability law. High Cost examines the change in probability of originating a high cost loan (subprime) by using rate spread as a proxy for subprime loans. Securitized is a variable provided by HMDA that indicates whether the initial lender sold the loan within the same year as it was originated. For the full sample, Specification (1) compares states based on whether they ever passed an assignee or predatory lending law.159 Specification (2) compares the respective laws when they are actually in effect. Specification (3) includes controls for the strength of the assignee liability law (assigneeindex and assigneeunrated). Finally, Specification (4) interacts the assigneeineffect with the assigneeindex variable to determine the marginal effect of the strength of the assignee liability law only when it is in effect. 6 ° 158 See CAL. FIN. CODE § 4979.8 (West 2007); N.Y. BANKING LAW § 6- 1(13) (McKinney 2003). "' The basic model looks something like this: (rate spread, high cost) = a + Pl*assigneeliabilty (whether a state ever had an assignee liability law) + 0*predatorylendinglaw (whether a state ever had a predatory lending law) + ,.. (individual characteristics of the loan) + P,. (fixed effects for state and year). 160 The state-level basic model is very similar: (Securitization) = a + Vlawineffect (whether predatory or assignee or both, the measure is just trying to capture the difference between the laws) + P,.. (individual characteristics of the loan) + 0 .,..(fixed effects for year). No. 3:1021] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LAW REVIEW For the state sample, Specification (1) compares Pennsylvania and Delaware, measuring the difference between a state with a law without assignee liability and a state without a law.161 Specification (2) compares Pennsylvania, a state that does not impose assignee liability, with Florida, a state that does, but with no greater penalty than HOEPA. 62 In both Specification (1) and Specification (2), the laws do not exceed HOEPA. Finally, Specification (3) compares California and New York in order to test the difference between a state with no assignee liability and a state with assignee liability that is stronger, as measured by the S&P index, than HOEPA 63 E. Results 1. Full Sample a. Rate Spread: Interest Rate on Subprime Loans Table 12 presents the results for Rate Spread for the entire national sample. The demographic results suggest that African-Americans and Hispanics both face higher rates as compared to White borrowers. Moreover, tractminoritypercent demonstrates that living in a tract with more than the average percentage of minorities increases the likelihood of originating a more expensive loan. It also suggests that applicant income is negatively correlated with higher interest rate. Specification (2) suggests that when an assignee liability law is in effect, the interest rate on high cost loans increases. Moreover, states with predatory lending laws see a decrease in interest rates when compared 161 See 63 PA. STAT. ANN. § 456.503 (West 2007). 16 Compare id. with FLA. STAT. ANN. § 494.00793 (West 2007). 163 Compare CAL. FIN. CODE § 4970 (West 2007) with N.Y. BANKING LAW § 6-1(13) (McKinney 2003); see also Criteria: Anti-Predatory Lending Law Update, supra note 147 (rating New York's law as more aggressive than the HOEPA baseline). [Vol. 2008 to states without predatory lending laws. This suggests the "lemon effect" in full force."M Column (3) lends further support to this hypothesis by demonstrating that stronger assignee liability laws actually reduce interest rates. Finally, in Specification (4), the interaction between assigneeindex and assigneeineffect suggests that marginally increasing the strength of an assignee liability law that is already in effect mitigates the higher costs caused by the assignee law. These interactions describe the conflicting influences of the "lemon effect" and increased costs to securitization. Specification (2) suggests that assignee laws increase the cost of subprime lending, leading to an increase in interest rates. However, a predatory lending law has the opposite effect. The difference is the effect each regime has on the asymmetric information problem.'65 The following illustration may prove useful. The existence of a state law boosts the confidence of consumers, which gain is symbolized by X. The subprime lender incurs a cost, Y. The consumer initially hesitates to enter the subprime loan market due to the transaction costs imposed by the uncertain possibility of deception, the "lemon effect," Z. When the predatory lending law comes into place, the consumer price equation becomes Z-X. If Z-X sufficiently reduces costs such that the consumer receives a net gain from entering the subprime market, he or she will do so. However, enacting a predatory lending law with assignee liability is unlikely to affect the day-to-day behavior of the average consumer any more than a predatory law without assignee liability does.166 Yet, such a law is likely to impact a potential lender. The lender incurs cost Y+A (assignee liability securitization cost), even though the consumer maintains only boost of confidence X. 14 Ho & Pennington-Cross, supra note 14. 16 Akerlof, supra note 153, at 495. As suggested earlier, it seems unlikely that a consumer would understand how an assignee liability law works, or what it does. Thus, while the assignees and originators would be concerned by the law, assignee liability is unlikely to garner more consumers to the market then a traditional predatory lending law. No. 3:1021 ] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LAW RE VIEW Nevertheless, the negative coefficient for the interaction of the strength of an assignee liability law (assigneeindex) with its existence (assigneeineffect) in Specification (4) suggests something more interesting. Weak assignee liability laws create few additional costs for originators because they are no different from the nationwide HOEPA base-level. However, originators under such regimes avoid draconian assignee liability laws. These laws, like HOEPA, typically have interest rate triggers. Most states adopt the same triggers as HOEPA, but the most typical way a law becomes more draconian is by setting a lower trigger rate. '67 The negative coefficient for marginally increasing the strength of an assignee liability law suggests that when a state passes a stronger law, originators lower their interest rates in order to avoid coverage by the law. That originators care about avoiding assignee liability laws should come as no surprise. Given the ease with which predatory laws can be avoided, regulators should be especially concerned about lulling consumers into a false sense of security. 2. High Cost: Probability of Originating a High Cost Loan Table 13 presents the full sample results for the probability of originating a "high cost" (subprime) loan. In Specification (2), assignee liability is insignificant, and Specification (3) demonstrates that only the attitude of the state toward assignee liability-as demonstrated by passing a particularly strong assignee law-reduces high cost (subprime) loans. Specification (4) demonstrates that having the law in effect, however, does not change anything. Table 13 demonstrates the difficulty of assessing the subprime market by looking merely at the incidence of high cost loans. Most assignee liability laws have little effect on 167 See N.M. STAT. ANN. § 58-21A-3 (West 2007) (setting New Mexico cut-offs at 7% and 9%); 815 ILL. COMP. STAT. 137/10 (2007) (setting Illinois cut-offs at 6% and 8%). However, both received ratings from S&P as being less costly than the HOEPA regime. See Criteria: Anti-Predatory Lending Law Update, supra note 147. [Vol. 2008 RETHINKING ASSIGNEE LIABILITY the number of high cost loans, and even the draconian ones change the number of high cost loans very little. Some would declare victory-that the presence of better consumers in the market drives the outcome. That, however, would not be looking at the whole story. Most of the state laws have the same limited coverage as HOEPA, and thus very few affect lending behavior. 6 ' Under these regimes, lenders can insert high balloon payments, engage in negative amortization, and lend in amounts that the lender could never repay without any danger to assignees, so long as they keep the interest rate below a certain amount.'69 Moreover, given the adverse selection problems in the high cost loan market, it is unlikely that demand will change based on the increase in predatory terms. 7 ° Therefore, the laws increase the number of consumers in the market by increasing consumer confidence but, at the same time, only put caps on two methods of predatory lending-interest rates and fees. This leaves consumers a manipulation of interest rates away from having the same problem as before. " This trend may suggest a desire on the part of legislators to create a sort of "placebo effect" by increasing consumer confidence, without really hampering the market. The possible disaster this action may cause is hard to overestimate. 169 HOEPA covers first liens with rate spreads over 8% and second liens with rate spreads over 10%. 15 U.S.C. § 1602(aa) (2007). This meant in February of 2008, HOEPA covered twenty year mortgages with interest rates of 12.5% for a first lien, and 14.5% for a second lien. The Department of Treasury, Daily Treasury Yield Curve Rates, http://www.ustreas.gov/offices/domestic-finance/debtmanagement/interest- rate/yield.shtml (last visited Dec. 3, 2008). Of course, it is important to point out that the originators themselves could still be held accountable for these provisions, but that has been the case for some time, and it has already been shown that holding originators financially responsible for predatory loans does nothing to stop them. 170 Silverman, supra note 13, at 578-79 (suggesting that it is the riskiest, most desperate consumers that will take the worst loans, and since better consumers have lower prices available to them, raising the "cost" of lending-whether through balloon payments, negative amortization, etc.-will increase the number of risky consumers). No. 3:1021 ] COLUMBIA BUSINESS LAW REVIEW 3. State Sample: Probability of Securitization Table 14 demonstrates the change in probability that an individual loan was securitized when comparing states with specific legal situations. Here, we see an increase in Specification (1), an increase in Specification (2), and a decrease in Specification (3). The results are consistent. In Specification (1), comparing Pennsylvania and Delaware, there has been no legal change from an assignee's point of view, only a boost in consumer confidence, resulting in better applicants, and thus more easily securitized loans. This result is consistent with overcoming the "lemon effect." Specification (2) is in line with the situation suggested in the full sample. Since most loans were already avoiding assignee liability under HOEPA, it is unlikely that the investment banks saw this law as changing anything. However, the active role of the New York State Attorney General in touting the new law may have caused a boost in confidence, especially when they could point to the sparse federal enforcement.' 7' Finally, Specification (3) suggests that when an assignee liability law is more draconian, it begins to push away investors and thus increases the costs of subprime lending. F. Discussion Assignee liability laws have an effect on the consumer market. Available data make it difficult to specify that effect, but combined with a state penalty system set to interest rates, the data suggest consumers may be lulled into a market that is no safer after the law was enacted than it was before. To the extent the data demonstrate manipulation of interest rates to avoid laws, subprime lending laws may become entirely wealth-reducing. The state data confirm that the "lemon effect" on consumer confidence is real. Laws that change very little from the 1 Peterson, supra note 36, at 8; Eric Dash, New York Begins Inquiry into Possible Mortgage Bias, N.Y. TIMEs, Apr. 29, 2005, at C2; Special Report: Cracks in the Facade, supra note 46. [Vol. 2008 assignee's point of view increase consumer confidence in the lenders, broadening the market. In states where laws do change the landscape for the assignees, securitization decreases, presumably increasing origination costs. Assignee liability is too blunt. Together, the full sample and state-level analyses demonstrate that assignee liability increases costs (by reducing securitization) at the same time as it encourages consumers to enter the market. In fact, assignee liability may make predatory lending more dangerous by failing to stop the practice while convincing consumers otherwise. At a minimum, assignee liability may be both ill-aimed and too blunt to accomplish its purpose without undue damage to legitimate subprime credit. At worst, it both fails to prevent predatory lending and increases costs for the legitimate market. VI. THE GREAT MISCONCEPTUALIZATION OF SUBPRIME LENDING The failure of assignee liability suggests a need to reconceptualize the problem. Regulators must determine exactly what they are trying to accomplish. There are many possible goals of a predatory lending regime: (1) to make the victim whole; (2) stop predatory lending; (3) soothe jittery investors in order to encourage subprime investment; (4) uphold the principle of free choice in credit decisions; (5) encourage homeownership as a policy goal; and (6) encourage the legitimate subprime market. The current regime tries to accomplish all of these goals, but does so by focusing on the victim with monetary remedies. As this Note has attempted to demonstrate, it does not work. Instead, policy makers must be honest with themselves in determining which goals are most important in the long-run.'72 If society believes that 172 Other academics suggest that, in targeting assignees, society is misconstruing the problem as merely a market failure. One suggests that we need to reconceptualize the broker-borrower relationship from one of equal bargaining to one of agency. Wilson, supra note 27. Another wants to use the "suitability doctrine" from securities law to judge loans. Ehrenberg, supra note 23, at 125. This approach suffers from concern No. 3:1021] RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LA W REVIEW compensating the victim is more important than the availability of subprime credit, then monetary remedies could be the right choice. However, if society determines that stopping the practice and preserving the legitimate subprime market is more important than making individuals whole, the answer may be elsewhere.173 Any approach must address several key problems in the regulation of subprime lending. First, insolvency cannot be a defense. Since near-insolvent mortgage brokers, rather than larger asset-backed banks, make well over half the loans, this is a necessity. Second, experimentation must be restored to the states. Any solution is likely to require fine- tuning; focusing on the originators rather than assignees would remove much of the national rating agency pressure and restore autonomy to the states. Third, the solution should not cause harm to investors. The securitized trust about predictability and does not capture the underlying problem of how to apportion liability. 173 One might pause to consider whether the subprime crisis requires any regulatory action at all. At least two scholars suggest this was a market failure of excess demand-a bubble-that likely will not occur again. See Kiff & Mills, supra note 17. There is ample evidence of an industry-wide self-correction. Investors are starting to sue banks for failing to make adequate reserves or explain the risk of subprime investing. The Finger of Suspicion; Mortgage-industry Lawsuits, ECONOMIST, Dec. 22, 2007, at 118. The SEC has twenty on-going investigations, targeting banks, rating agencies, and insurers. See id. In 2006, there were no subprime securities class actions; in 2007, there were thirty-eight. Karen Donovan, Class Actions Rise, Fueled by Subprime Troubles, N.Y. TIMES, Dec. 21, 2007, at C7. Investors are holding banks accountable in exactly the way predatory laws would have borrowers do. The market, left alone, may fix itself. However, some have suggested that this backlash has gone too far, undermining investor confidence in the market. Special Report: Cracks in the Facade, supra note 46; Subprime Subsidence; Mortgage Lending, supra note 49. By the end of 2007, subprime lending dropped from well over 20% of the mortgage market to roughly 2%. Paletta & Hagerty, supra note 6. What predatory lending laws, even with assignee liability, were not doing, the risk associated with the foreclosure crisis did in one year. If anything, this indicates the need for a law promoting investor and borrower confidence. Taking no action does not preserve the subprime mortgage market. [Vol. 2008 RETHINKING ASSIGNEE LIABILITY must remain independent of the originator to reduce investment risk, especially for those aspects of borrower characteristics that are impossible to verify. Fourth, any regime must not rely on the need for third parties to verify that the originator was truthful about impossible to verify characteristics. 174 Finally, the solution must punish bad actors without providing a windfall for transactions that victims willfully entered. Reconceptualizing requires acknowledgement that the borrower agreed to the contract. Focusing on the perpetrator prevents the regime from encouraging consumers to enter into contracts they cannot afford, while also preventing the law from artificially keeping people in houses they cannot maintain. 175 The solution to predatory lending must separate the culpability of the originator from the negligence of the "victim." Only then can the system punish bad actors without creating a cycle that rewards negative victim behavior. Focusing on deterrence would provide an approach that is victim-neutral, but more effective at both stopping further wrongdoing and maintaining subprime credit. Criminal liability is one possible alternative under this rubric. 76 Criminal liability is no stranger to white-collar offenses. 77 174 Unlike the plaintiffs in a foreclosure case or direct predatory lending case, investors are (relatively) easily pooled into a class and have the wherewithal to bring suits, as evidenced by the thirty-eight brought in 2007. Donovan, supra note 173; Nathan Koppel, Securities Class-Action Lawsuits Climb by 43%, WALL ST. J., Dec. 22, 2007, at B2. 175 See Kiff & Mills, supra note 17, at 15. ... This approach is very much in line with the 10b5-1 insider trading rules. Insider trading remedies run up to disgorgement plus three times that amount. However, insider trading could mean huge damages for a large class of people. Thus, there is virtually no way those people will see more than a "penny on the dollar" for their losses. This approach makes sense not because it is cold-hearted, but because it prevents the practice from causing more damage to society than it has already, both by stopping it, and by not harming anything else in the process. 177 Insider trading is one example. Violations of Rule 10b5-1 can carry not only civil penalties, but also penal consequences. One of the best known examples is the prosecution and one year prison sentence of Ilan Reich for insider trading. Jennifer Reingold, The Rise and Fall (and No. 3:1021] COLUMBIA BUSINESS LA W REVIEW Like insider trading, criminal liability rules that hold mortgage brokers criminally liable for certain, defined lending practices could restore investor confidence without imposing costs on securitization. Criminal liability cannot be thwarted by insolvency, maintains focus on originators, and does not encourage borrowers to enter into unfavorable contracts. Moreover, the Federal Bureau of Investigation ("FBI") already uses criminal investigations against the worst of the predatory lenders.'78 Nevertheless, criminal liability is just one example of a range of possible solutions when regulators reconceptualize the subprime crisis. Instead of focusing on the problem, a focus on the sanctity and importance of homeownership spurred the idea that the victim must be made whole. Yet, that wrong conception creates laws like assignee liability that not only fail to stop predatory lending, but instead lull more unwary victims into the market. VII.CONCLUSION While predatory lending is quite troubling, regulators need to check their instincts at the door when crafting regulations. Regulators' concern for making victims whole and being "fair" has caused them to adopt assignee liability. By compensating the victim, assignee liability paves the way for future injustice, while causing significant damage to the legitimate, welfare-enhancing portion of the subprime industry. Moreover, it may not even make sense to compensate people who freely signed for loans they could not repay. Both theoretically and empirically, this Note attempted to demonstrate problems with assignee liability. Conceptually, assignee liability imposes large costs for assignees. Empirically, these costs are real. Rather than ending predatory lending, the laws may merely be forcing Subsequent Fall and Rise) of Ilan K Reich, FORTUNE, June 11, 2007, at 148. Tax evasion, another form of fraud, would be a further example. "' See Carolyn Said, Plenty of Blame for Lending Mess, S.F. CHRON., Feb. 3, 2008, at C1 (detailing fourteen FBI criminal cases and investigations by the Ohio Attorney General). [Vol. 2008 the practices to change. Assignee lending laws boost consumer confidence, while at the same time failing to adequately police predatory lending practices. Assignee liability laws end up as an ill-fated attempt to "right the wrongs" of predatory lending. They end up not only failing in that attempt, but making the problem even worse. Regulators need a new answer. A first glance at a deterrence-based approach, such as using criminal liability, demonstrates the possibility that regulators can restore consumer confidence in the credit market, restore investor confidence in the securities market, and prevent the victimization of consumers. More important than a specific solution, however, is to attack the problem with the proper deterrence-oriented conceptualization. Perhaps Hamlet was really discussing the perils of assignee liability when he suggested the dangers of jumping into unknown evils, "And makes us rather bear those ills we have/Than fly to others that we know not of?""'9 The subprime mortgage crisis forces lawmakers to face the harsh reality that not all Americans can afford to own a home. In attempting to help those Americans, lawmakers need to be careful they do not foreclose the opportunity for other Americans who can. ' WILLIAM SHAKESPEARE, HAMLET, PRINCE OF DENMARK act 3, sc. 1. No. 3:1021] RETHINKING ASSIGNEE LIABILITY Table 1: U.S. Census Bureau Home Ownership Data by Race U.S. Census Bureau Home Ownership Data by Race (2007 3rd Quarter and National Only) 0.75 0.7 0.65 0.6 _-- US -i -Caucasian 0.55 -*- _ ___ -- a-African American 0 0.5 -- Hispanic 0.45 0.4 i i 1995 1997 1999 2001 2003 2005 2007 Table 2: Applicant Characteristics Mean Median N Applicant Income: 79.29 66 3388181 (in thousands) African-American 9% NA 3762438 Hispanic 11% NA 3762438 Asian 3% NA 3762438 Male 55% NA 3762438 Loan Amount:(in houns 152.93 128 3762438(in thousands) Table 3: Distribution of Originations by Type Percent of Total 2004 2005 2006 High Cost Loans 16.45 28.18 29.35 HOEPA Loans 0.14 0.24 0.11 Conventional Loans 83.51 71.82 70.65 Table 4: Distribution of Ethnicity by Type of Loan Percent of total High Cost Loans HOEPA Loans All Other Loans African-American 15.87 10.37 6.08 Hispanic 18.03 7.41 9.72 Asian 2.60 1.01 4.26 All Minorities 36.13 18.61 19.87 White 63.87 81.40 80.12 COL UMBIA B USINESS LA W RE VIE W [Vol. 2008 RETHINKING ASSIGNEE LIABILITY Table 5: Income by Type of Loan High Cost HOEPA Loans All other Loans Loans Income<25000 4.87 11.31 3.33 Income<50000 28.80 35.34 22.78 Income>50000 66.34 53.35 73.89 Table 6: Percent of Loans Securitized (in same year of origination) 2004 2005 2006 High Cost Loans 68.22 72.10 72.66 HOEPA Loans 27.33 61.82 10.23 All Loans 72.79 69.94 64.81 Table 7: Rate Spread (Point above comparable U.S. Treasury Bond) 2004 2005 2006 4.71 5.19 5.35(1.61) (1.60) (1.66) 4.68 5.26 5.64 (1.51) (1.51) (1.63) 4.78 5.27 5.72 (1.60) (1.47) (1.62) 4.72 5.37 5.90 (1.50) (1.46) (1.59) 4.77 5.23 5.77 Hispanic (1.55) (1.32) (1.47) 4.76 5.23 5.62 (1.48) (1.33) (1.54) 4.74 5.25 5.60 (1.59) (1.55) (1.68) State with Assignee 4.69 5.28 5.69 Liability (1.55) (1.52) (1.65) State with 4.74 5.23 5.66 Predatory Lending (1.53) (1.44) (1.58) No. 3:1021 ] Table 8: Percentage Chance of Denial of an Application 2004 2005 2006 Income<25000 39.42 40.08 39.17 Income<50000 24.33 25.75 26.79 Income>50000 13.13 14.06 15.55 African-American 28.69 28.98 31.15 Hispanic 21.60 21.60 24.78 Asian 13.22 14.86 17.10 White 15.98 16.91 16.85 Average Chance 17.53 18.40 19.09 Table 9: Unmatched Census Tract Comparison, Means 2004 2005 2006 2004 2005 2006 Assignee Liability Law No Assignee Liability Population: 4422 4264 4373 4514 4401 4404 (in (2216) (2151) (2148) (2098) (2038) (2039)thousands) Percent 32.03 29.32 29.33 35.37 31.69 31.62 Minority (30.72) (29.69) (29.71) (30.95) (30.32) (30.26) Median 61678 59745 61242 61237 59272 60846 Income (10533) (11809) (12454) (10957) (12100) (12374) Percent of 101.41 101.02 101.05 102.28 101.62 101.60 Metropolitan (42.14) (39.44) (39.31) (41.29) (37.92) (37.75) Income #of Owner 1101 1114 1116 1076 1085 1086 Occupied (713) (687) (687) (628) (604) (604) Units Number of 1 1457 1511 1514 1446 1498 1500 to 4 family (824) (825) (825) (737) (748) (748) dwellings Observations 23638 28872 28781 27681 34487 34420 COL UMBIA B USINESS LA W RE VIE W [Vol. 2008 Table 10: Matched Census Tract Comparison, Means 2004 2005 2006 2004 2005 2006 Assignee Liability Law No Assignee Liability Population: 4382 4294 4300 4397 4320 4336 (in thousands) (2168) (2072) (2055) (1985) (1967) (1976) 32.51 29.56 29.59 32.12 29.42 29.30 Percent Minority (30.70) (29.66) (29.71) (29.50) (29.27) (29.18) 61592 59433 61001 61640 59591 61115 (10456) (11658) (12368) (10798) (12080) (12357) Percent of Metro 101.05 100.39 100.60 101.37 101.01 101.01 Income (40.47) (37.33) (37.69) (37.08) (34.67) (34.68) # of Owner- 1090 1092 1094 1089 1098 1101 Occupied Units (702) (659) (654) (620) (600) (600) Number of 1 to 4 1456 1500 1501 1448 1500 1502 family dwellings (814) (799) (798) (725) (733) (735) Observations 21440 27175 27365 21440 27175 27365 No. 3:10211 RETHINKING ASSIGNEE LIABILITY COLUMBIA BUSINESS LAW REVIEW Table 11: State Comparisons Ratespread Securitization High Cost California 2004 4.76 76.84% 14.17% 2005 5.17 76.22% 29.74% 2006 5.73 73.99% 32.62% New York 2004 4.73 67.78% 2005 5.24 67.73% 2006 5.68 63.23% Pennsylvania 2004 4.65 51.77% 2005 5.27 53.08% 2006 5.62 52.87% 15.21% 25.23% 26.41% 13.73% 21.12% 22.97% Florida 2004 4.58 74.72% 20.04% 2005 5.24 74.27% 32.25% 2006 5.68 71.15% 36.37% Delaware 2004 4.72 65.78% 17.38% 2005 5.32 68.83% 23.53% 2006 5.68 62.65% 26.01% (Vol. 2008 RETHINKING ASSIGNEE LIABILITY Table 12: Regressions Full Sample: Rate Spread (1) (2) (3) (4) assigneeliability predatorylendinglaw Gender ethhispanic Ethblack Ethasian loanamount applicantincome tractminoritypercent tractmedianincome assigneeineffect predatoryineffect assigneeindex assigneeunrated assigneemagnitudeineffect Constant Observations R-squared -0.464*** (0.051) -0.358*** (0.063) -0.033*** -0.032*** -0.032*** -0.032*** (0.005) (0.005) (0.005) (0.005) 0.028*** 0.028*** 0.028*** 0.028*** (0.007) (0.007) (0.007) (0.007) 0.218*** 0.218*** 0.218*** 0.218*** (0.007) (0.007) (0.007) (0.007) -0.051*** -0.051*** -0.051"** -0.051*** (0.013) (0.013) (0.013) (0.013) -0.000*** -0.000*** -0.000*** -0.000*** (0.000) (0.000) (0.000) (0.000) -0.001*** -0.001*** -0.001*** -0.001*** (0.000) (0.000) (0.000) (0.000) 0.001*** 0.001** 0.001*** 0.001*** (0.000) (0.000) (0.000) (0.000) -0.000*** -0.000*** -0.000*** -0.000*** (0.000) (0.000) (0.000) (0.000) 0.098*** 0.098*** (0.034) (0.034) -0.079*** -0.079*** -0.079*** (0.025) (0.025) (0.025) -0.002*** 0.009** (0.000) (0.004) -0.364*** -0.364*** (0.047) (0.047) -0.011** (0.004) 6.753*** 5.705*** 5.685*** 6.827*** 324288 324288 324288 324288 0.392 0.392 0.392 0.392 Robust standard errors in parenthesesso * significant at 10%; ** significant at 5%; *** significant at 1% 180 Variables not shown: refinancing, tractpopulation, tractmsamdincome, tractowneroccuppiedunits, numberofl4familyunits, firstlien, hoepastatus, and fixedeffectsforyearandstate. No. 3:1021 ] Table 13: Probit Full Sample: Probability of Originating a High Cost Loan (1) (2) (3) (4) assigneeliability -0.155"** (0.022) predatorylendinglaw -0.231*** (0.034) gender -0.120*** -0.120*** -0.120*** -0.120*** (0.003) (0.003) (0.003) (0.003) ethhispanic 0.365*** 0.365*** 0.365*** 0.365*** (0.004) (0.004) (0.004) (0.004) ethblack 0.591*** 0.591*** 0.591*** 0.591*** (0.005) (0.005) (0.005) (0.005) ethasian -0.063*** -0.063*** -0.063*** -0.063*** (0.007) (0.007) (0.007) (0.007) loanamount -0.001"** -0.001"** -0.001"** -0.001"*** (0.000) (0.000) (0.000) (0.000) applicantincome -0.001*** -0.001"** -0.001*** -0.001"** (0.000) (0.000) (0.000) (0.000) tractminoritypercent 0.003*** 0.003*** 0.003*** 0.003*** (0.000) (0.000) (0.000) (0.000) tractmedianincome -0.000*** -0.000"** -0.000"** -0.000*** (0.000) (0.000) (0.000) (0.000) assigneeineffect 0.002 0.002 (0.018) (0.018) predatoryineffect -0.073*** -0.073*** -0.073*** (0.013) (0.013) (0.013) assigneeindex -0.001"** -0.000 (0.000) (0.002) assigneeunrated -0.028 -0.028 (0.025) (0.025) assigneemagnitudeineffect -0.000 (0.002) Constant -0.253*** -0.255*** -0.260*** -0.258*** Observations 1292367 1292367 1292367 1292367 Robust standard errors in parentheses 8' * significant at 10%; ** significant at 5%; •** significant at 1% 181 Variables not shown: Refinancing, homepurchase, tractpopulation tractmsamdincome, tractowneroccuppiedunits, numberofl4familyunits, firstlien, hoepastatus, and fixedeffectsforyearandstate. COL UMBIA B USINESS LA W RE VIE W [Vol. 2008 RETHINKING ASSIGNEE LIABILITY Table 14: State Probits: Securitized California Pennsylvania Pennsylvania and New and Delaware and Florida Ynrk York predatoryineffect gender ethhispanic ethblack ethasian loanamount applicantincome tractminoritypercent tractmedianincome 0.155*** (0.057) 0.140*** (0.054) -0.210 (0.138) -0.002 (0.075) 0.787** (0.336) 0.003*** (0.000) -0.000 (0.001) 0.004** (0.002) 0.000 (0.000) assigneeineffect Constant -1.230"** (0.178) Observations 2897 Robust standard errors in parentheses"8, 0.009 (0.020) 0.039 (0.029) -0.001 (0.032) 0.214** (0.086) 0.002*** (0.000) -0.001** (0.000) 0.001"* (0.001) 0.000"** (0.000) 0.199"** (0.023) -0.614"** (0.069) 20760 0.025** (0.011) -0.017 (0.013) 0.015 (0.019) 0.001 (0.024) 0.001** (0.000) -0.000 (0.000) 0.002*** (0.000) 0.000*** (0.000) -0.129** (0.013) -0.250* (0.139) 78478 * significant at 10%; ** significant at 5%; *** significant at 1% 12 Variables not shown: Refinancing, homepurchase, tractpopulation tractmsamdincome, tractowneroccuppiedunits, numberofl4familyunits, firstlien, hoepastatus, and fixedeffectsforyear. No. 3:1021 ]