The Dodd-Frank Act and Housing Finance: Can It Restore Private Risk Capital to the Securitization Market?
Levitin, Adam J., Pavlov, Andrey D., Wachter, Susan M., Yale Journal on Regulation
Private risk capital has virtually disappeared from the U.S. housing finance market since the market's collapse in 2008. This Essay argues that private risk capital is unlikely to return in any scale until the informational problems in housing finance are resolved so that investors can accurately gauge and price the risks they assume.
The Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010 represents a first step in reforming U.S. housing finance. It takes a multi-layered approach by regulating both loan origination and securitization. Dodd-Frank's reforms, however, fail to address adequately the opacity of credit-risk information in mortgage markets and thus are insufficient to restore private risk capital.
This Essay argues that such Dodd-Frank reforms as "skin-in-the-game" credit-risk retention fail to solve the informational problems in the housing finance market, as they merely replace one source of informational opacity with another. Instead, this Essay argues, it is necessary to institutionalize structural changes in the housing finance market, particularly the standardization of mortgage securitization, that force the production of information necessary for accurate risk-pricing.
(ProQuest: ... denotes formulae omitted.)
I. Unraveling the Dodd-Frank Act's Regulatory Layers ................................. 160
A. The Defense-in-Depth Approach................. ................................. .160
B. Factors That Will Determine the Success of Dodd-Frank. ...........163
II. The Information Problem in Mortgage Markets ......................................... 165
A. Information Problems in Securitization...... ..................................165
B. Information Problems in Portfolio Lending ................................. 166
C. The Limitations of Risk Retention .................................................167
D. The Limitations of Market Discipline in Informationally Shrouded Markets..........................................................................169
E. The Limitations of Diagnosis from Market Symptoms.................. 172
III. Degrees of Standardization ........................................................................173
Appendix.......................................................... ................................................. 178
The U.S. housing finance system is currently on government life support. In the wake of the housing bubble's collapse, the federal government has stepped into the breach that was created by both the failure of the financial institutions that created the secondary market for mortgage financing and the withdrawal of private risk capital from the housing finance system. Prior to 2008, private risk capital fueled the U.S. housing finance system. Almost all residential mortgages were funded through either the government-sponsored enterprises (GSEs) - Fannie Mae and Freddie Mac - private-label mortgage securitization (PLS), or portfolio lending by depositories.
The GSEs were federally chartered entities owned by private shareholders and were subject to federal oversight. The GSEs purchased mortgages, which they either held in their own portfolios or securitized. The GSEs financed their portfolio operations by issuing widely held corporate debt. The credit risk on the GSEs' mortgage-backed securities (MBS) was retained by the GSEs, which guaranteed their own MBS.1 Private-label securitization involved exclusively private risk capital, with credit risk assumed by investors and private insurers, while the credit risk on loans funded by depositaries was retained by the depositaries.
Formally, then, at least, the federal government only bore the credit risk on the small segment of the market financed through loans originated by the Federal Housing Administration (FHA) or the Veterans Administration (VA) (many of which were then securitized with a guarantee from Ginnie Mae). In addition, the federal government bore risk contingently on depository-funded loans and depositories' investments in mortgage-backed securities through federal deposit insurance. Thus, for everything except the small percentage of the pre-2008 market that was comprised of FHA-insured/VA-guaranteed products, private-risk capital was in the first-loss position.
Today, over ninety percent of new mortgages are either guaranteed directly by the government through Ginnie Mae and FHA/VA or indirectly through the federal government's eighty-percent ownership of Fannie Mae and Freddie Mac and its open-ended financial backstop of the GSEs.2 This ad hoc federal intervention in the housing finance market has always been presented as a temporary intervention until the market can find its feet again.3
Regardless of whether the current federal involvement in housing finance is primarily a change in formalities, rather than economic realities, any sort of restoration of a private housing finance system in the United States requires the return of private risk capital to the system. Thus, the fundamental question for housing finance reform is how to get private risk capital to return to the mortgage market. This is, first and foremost, a question of restoring investor confidence in a once-burned, twice-shy world.
The centerpiece of any effort to restore private risk capital to the U.S. mortgage market is designing regulatory and market structures that give investors confidence in the level of credit risk they are assuming. Unless investors can accurately gauge the credit risk they are assuming, they will remain concerned that they are underpricing for risk. Underpricing for risk can, in turn, fuel a housing bubble and further exacerbate credit risk, thus making underpricing even more severe.
Credit risk has two components: default risk and loss severity upon default. The former is a matter of underwriting quality, while the latter is a combination of underwriting quality - especially loan-to-value (LTV) ratios - and mortgage servicing quality. Both are major areas of investor concern. The housing bubble was an object lesson for investors in the perils of poor underwriting. And the aftermath of the bubble has been an object lesson hi the importance of mortgage-servicing quality in determining loss severities.4 Investors have discovered that loss severities on defaulted loans are heavily dependent on servicer behavior regarding loan modifications and foreclosures.5 Investors, however, have little ability to monitor servicer conduct or discipline wayward servicers.6
Already, legislative and regulatory attempts to restore investor confidence in the market have taken shape. There are three stages in attempts to restore private risk capital to the mortgage market. The first stage was the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, which took a multilayered approach to ensuring the quality of mortgage lending and mortgage securitization - the major financing channel for U.S. mortgages.
The second stage, still yet to unfold, involves the restructuring of Fannie Mae and Freddie Mac - and the secondary mortgage market in general. These two stages are inextricably linked. The design of secondary mortgage markets affects behavior in mortgage-origination markets by controlling what risks remain with loan originators and what risks are passed on to secondary-market purchasers.7 But the regulation of primary origination markets also affects the risks that inhere in secondary markets.
The third stage is the reform of mortgage servicing. Servicing reform is likely to occur piecemeal through regulatory enforcement actions, litigation, and agency rulemaking, rather than through legislation.8 It remains unclear whether private risk capital will return to the U.S. housing finance market in any volume following the completion of any or all of these stages of reform.
This Essay examines the Dodd-Frank Act's approach to restoring investor confidence in mortgage markets. The Essay argues that the Dodd-Frank Act's reforms of the mortgage market represent a fundamental misunderstanding of what went wrong. It is now widely understood that the change in mortgage financing from an originate-to-hold (portfolio) lending model to an originateto-distribute (securitization) lending model played a critical role in the housing bubble. But the problem with the originate-to-distribute market is often believed to be the lack of skin in the game by securitizers. Lacking a stake in the performance of mortgages, securitizers were willing to securitize poorly underwritten loans. Therefore, the reasoning goes, requiring skin-in-the-game via risk-retention requirements will correct the market by incentivizing securitizers to underwrite only sound loans.
This explanation for what went wrong is an agency problem story. Securitization is ultimately a form of financial intermediation between (consumer) borrowers and capital-market lenders. The financial institutions that intermediate between these end-borrowers and lenders in the securitization process act as economic (although not legal) agents. The Dodd-Frank Act sees the problems in the securitization market as agency problems first and foremost and attempts to correct them by forcing the agent into a partnership with the principal.
We do not dispute that agency problems played an important role in the financial crisis and in the development of the housing bubble in particular. But we believe that the agency problem in securitization was able to become so virulent because of an information problem. Securitization investors suffered from an information problem that led them to rely excessively on financial institution agents (securitizers) despite limited monitoring ability, including the inability to identify the formation of an asset bubble in real time. Put another way, the information problem in securitization amplified the endemic agency problem, and, as we argue, it is easier to deal with this information problem than it is to root out the agency problem.
Lack of skin in the game might well have contributed to the housing bubble, but it is important to recall that housing bubbles have been inflated and burst before with portfolio lending. The savings-and-loan (S&L) crisis, for example, involved portfolio lending, not securitization. Having complete skinin-the-game did not prevent the S&Ls from making bad loans. Similarly, many troubled commercial-real-estate loans today are held in depositaries' portfolios. One does not need securitization to engage in poor credit underwriting. And indeed, one does not need securitization to have agency problems; they are rampant in portfolio lending too.
The real importance of the shift in the financing channel from originateto-hold to originate-to-distribute was that it enabled a massive (and poorly underwritten) expansion of the capital base for mortgage lending. This had two effects. First, a larger capital base facilitated cheaper credit, which resulted in housing price inflation as buyers bid up prices with ever more leveraged bids. And second, the expansion of the capital base brought in a new class of less sophisticated mortgage investors (including collateralized debt obligations [CDOs] and CDO investors) who were placed at an informational disadvantage by their distance from the actual lending, a risk not compensated for by higher rates.9 These two factors set the stage for the mispricing of mortgage credit on a massive scale.
The Dodd-Frank Act is likely insufficient to ensure against future housing bubbles because it does not rectify the information problems inherent in housing finance markets that were at the center of the financial crisis. As things currently stand, neither regulators nor the market can determine - in anything close to real time - whether there is a housing bubble, and thus whether risk is being adequately priced.
At best, they can look to increases in the volume of housing finance or a decline in the price of credit. As this Essay shows, however, neither an increase in the volume of housing finance nor a decline in the price of credit is an adequate indicator of increased risk in the market. Both phenomena may simply reflect the removal of credit constraints or lower interest rates. Instead, the volume and cost of housing credit are useful indicators of market risk only if they can be tracked in a way that relates them to the level of risk in the market.
It is currently extremely difficult, however, to track changes in credit conditions in real time. Absent a market structure that enables the tracking of risk, market surveillance is not feasible. Thus, structural changes that force the production of information needed for accurate risk-pricing, particularly standardization of mortgage-backed securities, are the key to avoiding a repeat of the bubble and crash.
This Essay proceeds as follows: Part I explains the Dodd-Frank Act's multi-layered approach to regulating risk in the mortgage market. Part II explains the deficiencies in this approach. It shows that information problems in mortgage markets are not limited to securitization. It also shows that the Dodd-Frank Act failed to improve either investors' or regulators' ability to gauge credit risk in the mortgage market in real time. Part III discusses how standardization of securitization could improve the ability to gauge credit risk in real time. This Essay then concludes, and includes an Appendix describing a cause of real estate underpricing that affects commercial real estate, but not residential real estate.
I. Unraveling the Dodd-Frank Act's Regulatory Layers
A. The Defense-in-Depth Approach
The Dodd-Frank Act takes a multi-layered approach to regulating the mortgage market. First, title XFV of the Dodd-Frank Act, the Mortgage Reform and Anti-Predatory Lending Act, mandates significant reforms in the mortgage origination market.10 The centerpiece of title XIV is the requirement that mortgages only be originated if the lender has verified the borrower's ability to repay." Failure to do so is a defense against foreclosure.12 The Dodd-Frank Act provides a safe harbor for lenders with respect to the ability-to-repay requirement, which does not apply to "qualified mortgages" (QMs),13 as defined by Federal Reserve Board regulations. Non-QMs do not benefit from a presumption that the borrower is able to repay14 and are also prohibited from bearing prepayment penalties.15
Second, title IX, subtitle D of the Dodd-Frank Act undertakes a reform of asset-backed securities markets.16 The focal point of this title is a risk-retention requirement for asset securitizations, including mortgage securitizations, known as the "skin-in-the-game" requirement.17 Under regulations promulgated by a consortium of federal financial regulators, securitizers must retain a certain portion of credit risk on assets securitizations unless the securitized assets fall into certain exempt categories. The most important of those exemptions is for "qualified residential mortgages" (QRMs), again a term left to definition by the federal financial regulatory consortium.
The skin-in-the-game requirement is based on the theory that the "originate-to-distribute" mortgage lending model, meaning the origination of loans with the intention of reselling them in the secondary market, generally as part of securitizations, is deficient in providing sufficient incentive for originators to make good loans. The premise of the Dodd-Frank Act is that if the parties engaged in securitization are required to retain some credit risk on the securitized loans, they will be incentivized to ensure that the securitized loans are of higher quality, and that this will dry up the financing for shoddily underwritten loans.
The premise that originators who securitize will take care if they hold credit risk on their securitization reflects the belief that these originators/securitizers can take better care with underwriting. This belief is likely correct, but it is important to note the foundation of this belief, namely that originators/securitizers have the information necessary to engage in good underwriting. Related to this assumption is that originators/securitizers have an informational advantage over investors because they are closer to the underwriting of the loans. Thus, by aligning the interests of originators/securitizers and investors, the Dodd-Frank Act seeks to enable investors to piggyback on the informational advantages of originators/securitizers.
There is an intuitive appeal to the skin-in-the-game theory. Yet, if the problem with the securitization model is the exploitation of inherent information failures by financial intermediaries, then the Dodd-Frank Act response will not be sufficient to right the course of housing finance going forward. Dodd-Frank relies on skin in the game to raise the cost - or at least the risk - to financiers that make risky mortgages. But it may simply fail to do so sufficiently to be an effective deterrent. If the bank is willing to "bet" the bank on mortgages whose long term performance is risky, for the purpose of booking short-run gains, a "tax" on those short-run gains in the form of a requirement to hold some of the long-term risk in house will not matter.
Ultimately, the conceit of skin in the game is that banks provide investors with a meaningful bonding function when they retain credit risk on securitized assets.18 Skin in the game is supposed to have a bonding function by assuring investors that a party with superior information is willing to assume the same or similar risks.
The problem with this bonding function is that investors cannot determine whether the bank is competent at evaluating the risk on mortgages. Banks' financial reporting presents a composite picture of the banks' overall performance (or often of a financial conglomerate's overall performance). The performance of the banks' securitization is just one factor out of many and may well be overshadowed by the performance of the banks' other assets. Thus, a bank might be a poor judge of credit risk on mortgages but still be quite profitable overall. Indeed, the opacity of bank balance sheets19 mutes market discipline, and for larger banks, market discipline is further muted because of the moral hazard of investing in a too-big-to-fail financial institution.
The Dodd-Frank Act nonetheless anticipates that investors will make a judgment on the future credit risk of the mortgages based on the signals sent by the bank's overall historical performance. The Dodd-Frank Act thus attempts to fix the securitization market by mandating bonding by an entity of unknowable quality. In essence, then, the Dodd-Frank Act merely replaces one informational problem - that of securitization - with another - that of financial conglomerates. Investors are left incapable of adequately gauging and pricing for risk. Accordingly, investors are unlikely to put much stock in the bonding function of skin-in-the-game.
Finally, titles X and I of the Dodd-Frank Act create a pair of backstops to the QM and skin-in-the-game concepts. Title X of the Act creates a new, independent Consumer Financial Protection Bureau (CFPB) within the Federal Reserve that is charged with regulating the mortgage lending among other areas and broadly empowered to do so.20 The CFPB has the ability to regulate mortgage origination practices beyond the reforms in title XIV of the DoddFrank Act. It has the power to make rules under the Real Estate Settlement Procedures Act, the Truth in Lending Act, the SAFE Mortgage Licensing Act, the Interstate Land Sales Act, and Fair Debt Collection Practices Act,21 and also a broad power to proscribe unfair, deceptive, and abusive acts and practices.22
Title I of the Dodd-Frank Act creates the ultimate catchall, the Financial Stability Oversight Council (FSOC), a "Justice League" of federal financial regulators that is to act as a financial über-regulator, charged with keeping the world safe from systemic risk, and supported by a new Office of Financial Research.23 Thus, any problems that make their way through the sieve of titles IX, X, and XIV, should be caught by the FSOC as the ultimate line of defense.
The Dodd-Frank Act's design of defense in depth, through specific statutory rules coupled with broad regulatory authority to regulate products and practices not specifically restricted by the Act might well work. Unfortunately, we will only know if it does not if and when it fails. Victories of regulatory deterrence are never visible.
B. Factors That Will Determine the Success of Dodd-Frank
The success or failure of the Dodd-Frank Act apparatus depends on three factors. First, it depends on whether the QM and QRM provisions are effective at crowding out non-QM and non-QRM products from the market by making them too expensive. Put differently, the Dodd-Frank Act relies on tilting the playing field sufficiently to make non-QM and non-QRM products noncompetitive, and therefore limited in market share, which reduces their systemic importance.
Second, and relatedly, the Dodd-Frank Act's success depends on the definition of QM and QRM, because this will determine the extent of the loans and MBS that are affected by Dodd-Frank. Broad definitions will subject most mortgages to Dodd-Frank, but will allow "riskier" mortgages; narrow definitions will mean that the Act has little impact. If the "hard" lending characteristics of QMs and QRMs, like FICO score and loan-to-value (LTV) ratio, are defined too narrowly, then Dodd-Frank will be ineffective, as the QMs and QRMs will not crowd out the unregulated mortgage securitization market.24 Thus, if QRMs have LTV requirements under eighty percent, meaning down payments of twenty percent or more would be required, a large non-QM/QRM market in much riskier, higher-LTV mortgages is more likely to develop.25 At the same time, however, if the "soft" lending characteristics, such as servicing standards, are defined too broadly, the QM and QRM concepts will be ineffective at controlling risk, as everything will be QM and QRM, so neither will affect mortgage or MBS underwriting standards.
Third, if the QM and QRM provisions fail to crowd out poorly underwritten, riskier mortgage products, the fallback under Dodd-Frank is regulatory wisdom to identify and somehow contain risky practices. In terms of regulatory wisdom (and will), it is important to note that the CFPB's actions are subject to a regulatory veto by a two-thirds majority of the FSOC, so even if CFPB can identify and somehow contain risky practices, other regulators need to be on board.
More importantly, however, even if the CFPB and FSOC had the inclination to act when faced with a problematic practice or product, they would first have to identify the problem. 6 Therein lies the central problem with the Dodd-Frank Act in regard to the mortgage market: the Dodd-Frank Act fails to recognize that the housing bubble was caused by informational problems in the mortgage market that prevented investors and regulators from knowing the extent of credit risk in the market and responding appropriately.
The Dodd-Frank Act does nothing to rectify these informational problems, and as a result, the FSOC (and CFPB) are left with the same insufficient information as before the housing bubble. Investors are left with the same problem - post-Dodd-Frank they have no better ability to gauge credit risk than they did pre-Dodd-Frank. In such circumstances, it is difficult to imagine that private-risk capital will quickly return to U.S. mortgage markets.
II. The Information Problem in Mortgage Markets
The price theory of demand postulates that demand for an asset or service is correlated with the price of that service. In an efficient, complete market, price is transparent to all potential buyers, so demand is set at the efficient level. This means that information is a sine qua non of market efficiency. Absent sufficient information, mispricing is likely to occur, and, if market participants are concerned about information quality, they may simply withdraw from the market.
Mortgage finance markets are beset with a variety of informational problems that make it difficult for mortgage investors to price risk. In the United States, mortgages are financed primarily through two methods - portfolio lending and securitization. Portfolio lending involves loans made by a depository institution. The depository institution keeps the loans on its balance sheet and funds them through its deposits and other borrowings. Securitization involves a chain of financial institutions that link capital market funders with consumer borrowers. With securitized mortgages, the mortgages' risk lie with the securitization investors and the guarantors, if any, of their securities.27
A. Information Problems in Securitization
Both securitization and portfolio lending suffer from informational problems. The informational problems in securitization have been long noted.28 Ultimately, horneo wncrs know more about their risks than mortgage originators, who in turn have information advantages over other mortgage securitization intermediaries and MBS investors.29 MBS investors receive limited information about the mortgages backing an MBS issuance, and the quality ofthat information is dubious.
A typical MBS is supported by several thousand home mortgages. An MBS investor will have access to the prospectus and, more importantly, the prospectus supplement. These documents tell investors about the weighted averages of a variety of loan, borrower, property characteristics, and underwriting characteristics - LTV, FICO, state, loan size, amortization, loan term, loan type, for example. It might also give some distributional breakdown, by quintiles. What is impossible for investors to tell, however, in most cases, is the relationship between these terms.30
Consider two securitizations, each of two loans. One has a loan with a high LTV and low FICO and another loan with a low LTV and high FICO. The other has a loan with high FICO and high LTV and a loan with low FICO and low LTV. The average FICO and LTV for both securitization pools may be the same, but it cannot be assumed that the pools present equal risks. Absent the ability to correlate loan characteristics, weighted averages are of limited use.
A further problem exists in terms of the sophistication and market -specific knowledge of securitization investor base. A significant portion of the MBS investor base consisted of foreign investors whose knowledge of the U.S. mortgage market may have been limited. Instead, most investors relied on informational proxies - rating agencies' ratings first and foremost, but also market analysis produced by sell-side firms - and exhibited significant herding to follow dominant investment strategies.31
B. Information Problems in Portfolio Lending
Portfolio lending has its own set of informational problems. International experience in recent decades is replete with examples of portfolio lenders making disastrous lending decisions. Market discipline cannot be relied upon to keep portfolio lenders in check in real time because of the opacity of most lending institutions.32
Consider an entity like Bank of America, the largest consumer lender in the United States. The level of mortgage credit risk on Bank of America's balance sheet is impossible to determine with any certainty. Bank of America's financiáis do not provide sufficiently detailed information about its mortgage lending. Indeed, the more complex financial institutions become, the less transparent their financial reporting becomes relative to, say, a monoline lender.33 The reporting opacity of large financial conglomerates frustrates market discipline.
As with securitization, portfolio lenders suffer from internal agency problems - executives are often compensated based on periodic, rather than long-term performance, which creates incentives to boost deal volume now and worry about the consequences later. Moreover, loan officers and brokers are typically compensated based on production. This creates an incentive for them to originate as many loans as possible.34 Agency problems incentivize shortterm gain at the expense of long-term risk.
The agency problem, however serious, is not the critical problem in either portfolio lending or securitization. The agency problem is one that can only grow in an environment with opaque information, which encourages principals to rely on agents and frustrates attempts at monitoring - including market-wide monitoring for the formation of asset bubbles. The agency problems in portfolio lending and securitization cannot be easily fixed. The informational problems, however, can be addressed; and ensuring that principals are able to receive better information will prevent economic agents from exploiting their principals' information problems. Fixing the information problems in lending, whether through portfolio lending or through securitization, enables the endinvestors to deal better with the agency problem.
C. The Limitations of Risk Retention
Irrespective of the agency problems in the financing channel, the implicit put option on nonrecourse or functionally nonrecourse mortgages also creates a systemic risk problem that no amount of skin in the game can fix. On a nonrecourse mortgage, a homeowner has the option to "put" the house back to the bank by defaulting and surrendering the house. In other words, the strike price on this option is the value of the house. When the value of the house falls below the unpaid balance on the mortgage (meaning that the house is "upside down" or "under water" or there is "negative equity"), then the "put" option is "in the money." The de facto nonrecourse status of many U.S. mortgages means that when there is current or expected negative equity, homeowners have an incentive to default on their mortgages.
Skin-in-the-game risk-retention requirements would, at first blush, seem to mitigate against a decline in underwriting standards. Loan originators would demand lower LTV ratios, for example, to decrease the likelihood of negative equity. But the upfront fees that lenders (and their agents) make on loans push for greater lending volume, both in terms of number of loans and size, and this pressure can overwhelm skin-in-the-game incentives if the put option is "in the money." In fact there is no expected "skin in the game" if loans are made that are not expected to perform long term. If origination fees are high enough, the skin-in-the-game can merely be a loss leader.
The likelihood that the put option is or will be in the money creates an anticipatory change in lending behavior that exacerbates declining credit standards.35 Savvy financial intermediaries can foresee the put option coming into the money because credit risk levels are unsustainable and have led to an inflation of housing prices. If those intermediaries are compensated through short-term fees, they are incentivized to "make hay while the sun shines." Thus, Goldman Sachs pushed forward with synthetic CDOs like Abacus 2007 -ACl despite an email from the Abacus team leader Jonathan Egol to his underling "Fab" Fabrice Toure urging that deals in progress be completed because "the cdo biz is dead we don't have a lot of time left."36
Reputational constraints on credit quality dissipate if repeat business is not anticipated. Instead, there is a rush to expedite as many loan applications as possible, regardless of credit standards, before the crash comes. This phenomenon keeps bubbles going longer. Because the incentives are so strong at this point to make any and all loans before the market collapses, the volume of bad lending and the bubble supported by the lending may increase substantially.37 At this point mortgages will be originated with long-term risks underpriced, since originators do not expect the loans' long-term survival.
The smart-money response will, of course, be to go short in the above situation. The most attractive method of going short on the real estate market is through credit default swaps (CDS) on MBS.38 But the existence of credit derivatives means that going short does not necessarily create market discipline in the short-term. Synthetic CDS do not themselves exert market pressure on MBS except through cash-synthetic CDS arbitrage.39 Moreover, long-short strategies that involve taking out a long junior position and a short senior position in the same deal - such as those used by the Magnetar hedge fund - can effectively short-circuit market discipline.40 Instead, it can actually fuel a bubble.41
D. The Limitations of Market Discipline in Informationally Shrouded Markets
The risk of an undetected bubble, together with an undetected decline in credit standards, remains a major deterrent to the return of private risk capital to the housing finance market. The basis of the problem is a lack of information in assessing future credit risk and the lack of a signaling device for indicating heightened credit risk. If there is a signal in the market that points to heightened credit risk, then the market will sell off that product. This will cause rates to rise, stemming the continued issuance, origination, and acceptance of risky lending and stemming the growth of risk in that market.
In the market for government bonds, for example, "bond vigilantes" should, in theory,42 sell a country's bonds when inflation risk (or interest rate risk) increases, causing a higher yield. This corrective acts as a deterrent to inflation-generating lax monetary policy. The response of credit generation to heightened inflation risk stops the generation of inflation risk.
In a market with shrouded or incomplete information on credit risk, there are no such signals. Without signals, market participants cannot know when to withdraw funding from a market that is increasing hi risk. As a result, risk can snowball because asset prices themselves are increased by the issuance of debt to increasingly risky borrowers. Thus, the key question is: What informational factors are available to signal unsustainable increases in asset prices and credit quality?
The financial crisis of 2008 was due to the correlated risks of an asset bubble, and excess leverage and credit deterioration. Indeed, financial crises derive from asset bubbles hi which the financial system is implicated and which have the possibility of a lending freeze, which has spillover effects across the economy. The severity of the future crisis is directly related to the buildup of the asset bubble and deteriorating credit standards, which are permitted to continue by the absence of information on the building up of these joint risks. From 2003 to 2006, the possibility of an asset bubble was acknowledged, but the deterioration of credit underwriting was shrouded. Thus, observers lacked the information necessary to explain the asset price bubble.
What factors could have been known to identify the "excess" increase in housing prices? And could these factors have been known at the time that the "bubble" hi housing prices was generated?
Changes in volume in the financing channel could serve as such an indicator. Figure 1 shows for the United States the actual purchase price-to-rent ratio using Federal Housing Finance Agency data for price and consumer price index data for rent. It graphs that ratio against an estimation of the price -to-rent ratio based on the principal outstanding on mortgages in private-label (nonGSE) securitization pools. Figure 1 shows that when the ratio of PLS to overall volume of lending is used as an explanatory factor, the spike in the price-to-rent ratio is "explained."
By itself, however, volume is an imperfect warning sign of increased risk. For instance, easing credit in an otherwise constrained market can drive growth without increasing unpriced risk or creating an asset bubble.43 Volume alone, then, is not a conclusive indicator of risk.
Indeed, if volume of lending increases and if lending gets riskier, there may simply be a new, stable equilibrium if risk is appropriately priced. If, on the other hand, the increased lending is at increasing risk and markets do not recognize the increased risk and therefore do not raise the risk premium on this lending, the system will not be stable. Once the increased non-priced risk is recognized, the impact will be a spike in the risk premium. As interest rates rise to include the higher risk premium, there will be a decline in housing prices, as higher financing costs will keep buyers out of the market. Moreover, to the extent that existing borrowers need to roll-over their loans, higher interest rates will lead to higher default rates, which result in an increase in foreclosure inventory, which further pushes down prices.
Lenders who have made loans during the bubble period without increasing the risk characteristics of their own loans will nonetheless be exposed to increased credit risk as the value of the collateral backing the loans declines: that is, as the consolidated LTV ratios rise. Thus investors need to know to what degree loans that they will make may be subject to deterioration in lending standards. Ex ante, such potential deterioration in lending standards creates correlated risk of housing price declines because the mispriced lending is recognized, leading to the scramble to "make hay while the sun shines." Loan originators and securitizers - the intermediaries connecting capital market funders and home mortgage borrowers - are compensated largely based on the volume of loans they originate and securitize. When credit risk is underpriced, loan originators and securitizers can originate and securitize more loans than otherwise, thus increasing their own revenue. They will be incentivized to do so as long as the mispricing remains. Ex post, this results in increased defaults and foreclosures, which bring excess inventory of homes to the market putting further downward pressure on prices.
The crisis emerged when there was growth in lending volume and housing prices, and a continuing correlation between these two factors beyond the period of economy-wide interest rate declines. What mattered, however, in interpreting this correlation is whether higher housing prices and lending volume were sustainable. If the credit risk were correctly priced, even as constraints were overcome, the lending and the lending rates could be sustained. This demonstrates the importance of the question of the sustainability of the pricing of risk.
In 2005, as the crisis evolved, there were those who pointed to the possibility of a bubble emerging. We know now that this included economists at the Federal Reserve Board who estimated that prices were too high by about twenty percent in 2005 .44 hi their deliberation of what to do, they ruled out the raising of interest rates to "prick" the bubble as this might throw the economy into recession. Moreover, according to the released record, they believed at the time that price acceleration would ease and that economic growth would allow rents to catch up to prices without a major price decline.
The Federal Reserve Board economists did not have information at the time on the increase in the risk of lending and the decrease in the price of this risk.45 In other words, they did not track the risk premium that investors were placing on MBS lending. Nor were they aware, it appears, except anecdotally, of decreases in lending standards.
While in theory Federal Reserve Board economists could have known the evolution of the pricing of MBS, this was in fact not easy to track, much less in real time. We have done so ex post, but only with difficulty.46 Moreover, loan pricing should be tracked while holding quality of lending constant for a given risk. This raises difficult issues of how to identify the quality of lending in real time when the mix of loan products itself shifts, hi any event, the Federal Reserve economists did not know, as we do now, that the underpricing of risk would in fact get worse and would occur along with a continued growth of a housing bubble. Unfortunately, the Dodd-Frank Act provides them with no additional tools in this regard.
E. The Limitations of Diagnosis from Market Symptoms
Having observed frequent and widespread deterioration of lending standards around the world, economists have developed symptoms of deteriorating lending standards based on observable market indicators. One such indicator is the Pavlov-Wachter symptom of underpricing in the lending market.47 Specifically, Pavlov and Wächter use the correlation between asset prices and lending rates to distinguish between legitimate positive demand shocks and price increases due to lax lending standards.48
The Pavlov-Wachter symptom is based on the fact that the price of an asset depends not only on the value of the asset itself but also on any mispricing of the loan through which the asset is financed. If a mortgage is priced correctly - that is, the market has properly accounted for expected future volatility with a sufficient yield spread (risk premium) - its market value equals its face value, and the transaction price equals the fundamental value of the asset. If a mortgage is mispriced, then the transaction price reflects not only the fundamental value of the asset, but also the mispricing of the mortgage.
According to this relation, in a market with diversified investors, an increase in asset volatility should have little effect on the transaction price, since it is fully reflected in the yield spread. If, however, the yield spread changes due to underpricing rather than an increase in asset volatility, the transaction price of the asset changes - in the opposite direction of the yield spread. A mortgage with underpriced risk therefore results in an overpriced asset.
Pavlov and Wächter verify empirically that markets that exhibit their underpricing symptom do experience larger price declines when a negative demand shock occurs.49 This larger decline is because the new (lower) prices reflect not only the reduced demand but also compensate for the disappearance of underpriced financing.
But the relation Pavlov and Wächter find between demand shocks and pricing applies only to markets with diversified investors.50 In the real world, real estate investors - particularly homeowners - are likely to be undiversified. hi this case, the relation indicates that the transaction price and the yield spread tend to move in opposite directions, regardless of the cause.51
It is, therefore, very difficult for market participants to determine whether the increase in asset prices is due to an underpricing of the yield spread or a true increase in the asset's value. Thus, we cannot rely on the correlation of asset price increases and yield spread declines to point to a likely bubble.
Nonetheless, the Pavlov-Wachter underpricing symptom can still be useful. Major residential market players are also likely to extend loans to commercial real estate investors. Because most of these investors are diversified, the underpricing symptom works as originally developed. Therefore, one can use the correlation between commercial real estate prices and commercial lending spreads to detect underpricing in the economy.
Emphatically, the Pavlov-Wachter symptom is only an indirect method of detecting bubbles. The ability to detect accurately underpriced financing in the residential real estate markets depends on tracking the relationship between the levels of lending standards and credit constraints and the cost of credit and housing prices. Housing prices be sustainable only when the cost of credit accurately reflects the impact of lending standards on credit risk.
III. Degrees of Standardization
If private risk capital is to return to the U.S. mortgage market, it is necessary to rebuild investor confidence by improving the information available about credit risk. One route, proposed by the SEC, would simply be to require securitization sponsors to provide investors with enormous amounts of loan-level data disclosure in XML format.52
The problem with such an approach, however, is that a deluge of information is of limited use absent proven credit risk models that can make sense of the relationships between the different variables disclosed. Those models do not exist. While it may be possible to design effective multivariate risk models, excess information and variables reduce the predictive ability of such models, especially when new terms are introduced. Moreover, it would require multiple business cycles to hone such models in real market conditions. Multiple - and changing - products, each with many potentially different loan characteristics, make the development of reliable credit models difficult, if not impossible.
The corrective to disclosure of data without information is product standardization. Standardization decreases the number of variables investors must account for and thereby facilitates analysis; it is much easier to solve a two-variable equation than a hundred-variable equation. We have discussed standardization as a response in detail in other work,53 but essentially there are two ways by which it could be achieved.
First, it could be done either through t