Richard Stanton’s papers

This paper reviews the recent literature on algorithmic fairness, with a particular emphasis on credit scoring. We discuss human vs. machine bias, bias measurement, group vs. individual fairness, and a collection of fairness metrics. We then apply these metrics to the U.S. mortgage market, analyzing HMDA data on mortgage applications between 2009 and 2015. We find evidence of group imbalance in the dataset for both gender and (especially) minority status, which can lead to poorer estimation/prediction for female/minority applicants. Loan applicants are handled mostly fairly across both groups and individuals, though we find that some local male (non-minority) neighbors of otherwise-similar rejected female (minority) applicants were granted loans, something that warrants further study. Finally modern machine-learning techniques substantially outperform logistic regression (the industry standard), though at the cost of being substantially harder to explain to denied applicants, regulators, or the courts.

Download PDF

U.S. fair-lending law prohibits lenders from making credit determinations that disparately impact minority borrowers if those determinations are based on characteristics unrelated to creditworthiness. Using an identification under this rule afforded by the GSEs’ and FHA’s pricing of mortgage credit risk, we show that risk-equivalent Latinx/African-American borrowers pay significantly higher interest rates on both GSE-securitized and FHA-insured loans, with particularly high rates being paid in high-minority-share neighborhoods. In total, we estimate that these rate differences cost minority borrowers over $450 million yearly. These rate disparities were substantially lower for FinTech lenders for loans issued between 2009 and 2015, but there is no significant difference between FinTech and other lenders for loans issued in 2018 and 2019.

Download PDF

This paper reviews the dramatic growth of non-bank mortgage lending after the Global Financial Crisis, especially to borrowers with lower credit scores, and the related importance of mortgage-backed securitization. Our literature review suggests that the existing theoretical and empirical work on securitization is more relevant to bank than to non-bank lenders, thus leaving outstanding questions on why non-bank market shares have increased to their current levels and how best to structure non-bank oversight. To highlight key differences in the mortgage-lending incentives of banks and non-banks, we build a simple theoretical model of bank versus non-bank mortgage lending and use it to generate and test empirical hypotheses. We find, in particular, that loans issued by non-banks are more likely to prepay early than loans issued by banks, the difference not explainable by non-bank borrowers prepaying more rationally. Using regulatory filings from non-banks that are typically unavailable to academic researchers, we examine the balance sheets and liquidity and capital positions of large Ginnie Mae non-bank servicers, which face and pose more risk in the current mortgage system. We find that on average these servicers have reasonable liquidity and capital positions relative to standard regulatory thresholds, particularly in 2022:Q1 after a few quarters of elevated profits. However, some large Ginnie Mae servicers appear to have inadequate capital as gauged by risk-based capital measures. If defaults rise on a large scale, the liquidity and capital positions of these servicers may amplify the disruption in the mortgage and housing markets.

Download PDF

The disproportionate burden of COVID-19 among communities of color, together with a necessary renewed attention to racial inequalities, have lent new urgency to concerns that algorithmic decision-making can lead to unintentional discrimination against members of historically marginalized groups. These concerns are being expressed through Congressional subpoenas, regulatory investigations, and an increasing number of algorithmic accountability bills pending in both state legislatures and Congress. To date, however, prominent efforts to define algorithmic accountability have tended to focus on output-oriented policies that may facilitate illegitimate discrimination or involve fairness corrections unlikely to be legally valid. Worse still, other approaches focus merely on a model’s predictive accuracy—an approach at odds with long-standing U.S. antidiscrimination law.

We provide a workable definition of algorithmic accountability that is rooted in the caselaw addressing statistical discrimination in the context of Title VII of the Civil Rights Act of 1964. Using instruction from the burden-shifting framework, codified to implement Title VII, we formulate a simple statistical test to apply to the design and review of the inputs used in any algorithmic decision-making processes. Application of the test, which we label the input accountability test, constitutes a legally viable, deployable tool that can prevent an algorithmic model from systematically penalizing members of protected groups who are otherwise qualified in a legitimate target characteristic of interest.

Download PDF

We develop an empirical model of employee stock option exercise that is suitable for valuation and allows for behavioral channels. We estimate exercise rates as functions of option, stock, and employee characteristics using all employee exercises at 88 public firms, 27 of them in the S&P 500. Increasing vesting frequency from annual to monthly reduces option value by 11% to 16%. Men exercise faster, reducing value by 2% to 4%, while top employees exercise slower, increasing value by 2% to 7%. Finally, we develop an analytic valuation approximation that is more accurate than methods used in practice.

Download PDF

This paper presents an empirical analysis of the relationship between energy factor markets, leasing structures and the transaction prices of office buildings in the U.S. We employ a large sample of 15,133 office building transactions that occurred between 2001 and 2010. In addition to building characteristics, we also include information on the operating expenses, the net operating income, and the market capitalization rates at sale to estimate an asset pricing model for commercial office real estate assets. A further set of important controls in our analysis is the one-to-twelve month forward contract prices and the shape of the forward contract price curve, using auction data for the regional electricity trading hubs in which the building is located and auction data from the Henry Hub for natural gas. We also include weather metrics in the form of the variance in the last twelve months of minimum and maximum temperature and precipitation for each building’s location and sale date. Our final set of controls in- cludes information on the dominant contractual leasing structure of the buildings. Our empirical results suggest that Energy Star labels do not explain additional variance in property prices once the key asset pricing factors of expenses, income and market capitalization rates are included. Energy factor market prices, the shape of the energy forward price curves, and weather metrics are consistently shown to be statistically sig- nificant determinants of office building transaction prices, suggesting that commercial office building prices are likely to be exposed to shocks in these markets.

Download PDF

We develop a theoretical model of a network of intermediaries, which gives rise to heterogeneous financial norms and systemic vulnerabilities. The optimal behavior of each intermediary regarding its attitude toward risk, the quality of the projects that it undertakes, and the intermediaries it chooses to interact with, is affected by the behavior of its counterparties. These strategic network effects influence the financial strength and systemic vulnerability of individual intermediaries, as well as in aggregate. We apply the model to the mortgage-origination and securitization network of financial intermediaries in the U.S., using a data set containing all of the more than three million private-label, fixed-rate mortgages originated and securitized between 2005 and 2007, for which we track the ex-post default performance using loan flows to define linkages. We find that default risk was closely related to network position and evolved in a predictable manner among linked nodes, and document a positive relationship between a firm’s estimated quality and its labor intensity. This suggests that network effects are of vital importance in the U.S. mortgage market.

Download PDF

Non-banks originated about half of all mortgages in 2016, and 75% of mortgages insured by the FHA or VA. Both shares are much higher than those observed at any point in the 2000s. We describe in this paper how non-bank mortgage companies are vulnerable to liquidity pressures in both their loan origination and servicing activities, and we document that this sector in aggregate appears to have minimal resources to bring to bear in a stress scenario. We show how the same liquidity issues unfolded during the financial crisis, leading to the failure of many non-bank companies, requests for government assistance, and harm to consumers. The high share of non-bank lenders in FHA and VA lending suggests that the government has significant exposure to the vulnerabilities of non-bank lenders, but this issue has received very little attention in the housing-reform debate.

Download PDF

Using detailed origination and performance data on a comprehensive sample of CMBS deals, along with their underlying loans and a set of similarly rated residential MBS, we apply reduced-form and structural modeling strategies to test for regulatory-capital arbitrage and ratings inflation in the CMBS market. We find that the spread between CMBS and corporate-bond yields fell significantly for ratings AA and AAA after a loosening of capital requirements for highly rated CMBS in 2002, while no comparable drop occurred for lower-rated bonds (which experienced no similar regu- latory change). We also find that CMBS rated below AA upgraded to AA or AAA significantly faster than comparable residential RMBS (for which there was no change in risk-based capital requirements). We use a structural model to investigate these results in more detail, and find that little else changed in the CMBS market over this period except for the rating agencies’ persistent reductions in subordination levels between 1997 and late 2007. Indeed, had the 2005 vintage CMBS used the subordination levels from 2000, there would have been no losses to the senior bonds in most CMBS structures.

Download PDF

We show that the U.S. residential single-family mortgage-origination market is highly concentrated once account is taken of the contractual coordination that arises from the correspondent- and warehouse-funding channels. We represent these channels as a network, using the flow of loans through three strata of the loan origination market: origination, aggregation, and securitization. We develop a network representation of the origination market and demonstrate that it is a “small world,” in that most nodes are close in the network. We then rank-order the interlinked aggregators and securitizers using ex post mortgage foreclosure rates as a proxy for performance. Our findings suggest that these significant interlinkages in the mortgage origination network represent a previously under-appreciated source of systemic risk. Many apparently atomistic mortgage underwriters are, in fact, coordinated to act in parallel because of their funding relationships with the large, “too-big-to-fail” bank holding companies.

Download PDF

This paper reviews the current state of residential-mortgage data structures from origination through the securitization supply chain. We discuss the various uses of these data, their limitations in mortgage-risk management, and the current lack of transparency in important segments of the mortgage market. We conclude that despite the size and importance of the mortgage market in the overall U.S.\ economy, current data-management practices make it difficult or impossible for borrowers, lenders, investors and government regulators to perform the oversight and analysis functions necessary to maintain an orderly market and to ensure fair pricing of securities backed by those mortgages.

Download PDF

The law governing the United States’ $13 trillion mortgage market is broken. Courts and legislatures around the country continue to struggle with the fallout from the effort to build a 21st century global market in mortgages on a fragmented, arguably archaic legal foundation. These authorities’ struggles stem in large part from the lack of clarity about the legal requirements for mortgage transfer, the key process for contemporary mortgage finance.

We demonstrate two respects in which American mortgage transfer law is unclear and offer suggestions for fixing it. Revisions to the Uniform Commercial Code adopted around the turn of the century may be interpreted as doing away with preexisting laws arguably requiring parties to record their ownership interests to protect them. But the interaction of these revisions and preexisting state recording laws is most unclear.

Moreover, it is not clear just what parties have to do to invoke the provisions in question: the rules require that the mortgages be transferred “in return for any consideration sufficient to support a simple contract,” but that concept is notoriously malleable. In fact, our study indicates that in many transactions the requirement may not have been met because the parties used only nominal or other questionable consideration. At least in some states, the use of questionable consideration in existing transactions makes foreclosure more difficult and makes mortgage investments less secure. Consequently, the use of nominal consideration seems to strengthen mortgage investors’ claims that transactions were not executed properly.

We suggest an approach to law reform that would provide needed clarity and bring about an appropriate balance between private and public. The Article 9 revisions reflect a preoccupation, prevalent in the 1990s, with reducing the cost of mortgage transfers to the transacting parties. Obviating public recording, as the Article 9 revisions purport to do, does reduce cost, but it also tends to eliminate public records of mortgage ownership. As we show, these public records have value not just for parties that may transact in mortgages, but for the public more generally. A more balanced approach would clearly require transacting parties to record their interests in order to protect them, but would adopt this change in tandem with an expansion of low-cost digital recording. This approach provides the public benefits of high-quality mortgage records while reducing the cost and inconvenience of recording to transacting parties.

Download PDF

We study the finite-sample properties of some of the standard techniques used to estimate modern term structure models. For sample sizes similar to those used in most empirical work, we reach three surprising conclusions. First, while maximum likelihood works well for simple models, it produces strongly biased parameter estimates when the model includes a flexible specification of the dynamics of interest rate risk. Second, despite having the same asymptotic efficiency as maximum likelihood, the small-sample performance of Efficient Method of Moments (a commonly used method for estimating complicated models) is unacceptable even in the simplest term structure settings. Third, the linearized Kalman filter is a tractable and reasonably accurate estimation technique, which we recommend in settings where maximum likelihood is impractical.

Download PDF

Mortgage Electronic Registration Systems, Inc. (“MERS, Inc.”) owns legal title to some thirty million mortgages in the United States via its role in the Mortgage Electronic Registration System (“MERS”). The company, which was a key part of the mortgage securitization apparatus in the late 1990s and 2000s, is now under intense pressure from public and private lawsuits and investigations and faces a very real threat of insolvency. Policymakers are looking ahead to potential replacements for MERS, as a recent Fed staff proposal for a substitute system indicates. This Article examines what might happen to the mortgages that MERS, Inc. at least nominally owns in the event that the company enters bankruptcy, a question that apparently has never been explored in a publicly available analysis.

Although the legal analysis underlying the design of MERS does not appear to be publicly available, a key assumption seems to have been that if the company ever entered bankruptcy, the mortgages in its hands would not enter the company’s bankruptcy estate and would not be available to creditors. This Article challenges that assumption, pointing to the broad authority the Bankruptcy Code confers on the bankruptcy trustee with respect to interests in real property, such as mortgages. Most courts that have considered the issue have found that the bankruptcy trustee can bring into the estate any real property interest that the debtor could have conveyed to a good-faith purchaser. There is a risk that MERS, Inc. can convey MERS mortgages to a purchaser acting in good faith.

The risk arises in large part from the company’s conduct in making and acquiescing in claims in court that the company can sell the mortgages, has constitutionally protected property interests in the mortgages, is a creditor of mortgage borrowers, and owns a beneficial interest in the mortgages. But under current law, part of the risk is inherent in any mortgage recording system that operates nationally, is subject to the Bankruptcy Code, and holds mortgages as an agent. Policymakers should give serious consideration to creating a transparent, authoritative national mortgage registry with a solid legal foundation, and should consider the bankruptcy risk we identify as they do so.

Download PDF

We develop a parsimonious general-equilibrium model of banking and asset pricing, in which intermediaries have the expertise to monitor and reallocate capital. The model allows us to study the connection between financial development, intra-economy capital flows, the size of the banking sector, the value of intermediation, expected returns in the market, and the risk of bank crashes. We show that realized capital flows are informative about financial development, whereas other common empirical proxies, such as the size of the banking sector, may be only weakly related. Our model also has strong asset-pricing implications; for example, a market’s dividend yield is related to its financial flexibility, and capital flows should be important in explaining expected returns and the risk of bank crashes. The model’s predictions are broadly consistent with the aggregate behavior of the U.S. economy since 1950.

Download PDF

We show that several well-known asset pricing puzzles are largely mitigated if we endow the representative agent with an arbitrarily small minimum consumption level. This allows us to solve the model for parameter values where the standard “Lucas tree” model is not defined. For these parameters, disasters become more important, and the market risk premium therefore higher, even though consumption is less risky. Our model yields reasonable risk premia, Sharpe ratios and discount rates; excess price volatility; and a high market price-dividend ratio. We derive closed-form solutions for all variables of interest.

Download PDF

ABX.HE indexed credit default swaps on baskets of mortgage-backed securities are now the main benchmark used by financial institutions to mark their subprime mortgage portfolios to market. However, we find that current prices for the ABX.HE indices are inconsistent with any reasonable assumption for mortgage default rates, and that ABX.HE price changes are only very weakly correlated with observed changes in the credit performance of the underlying loans in the index. These results cast serious doubt on the suitability of the ABX.HE indices as valuation benchmarks. We also find that ABX.HE price changes are related to short-sale activity for publicly traded builders, the commercial banks, the investment banks and the government sponsored enterprises (GSEs). This suggests that capital constraints, limiting the supply of ABS insurance, may be playing a role here similar to that identified by Froot (2001) in the market for catastrophe insurance.

Download PDF

In an economy with perfectly competitive capital and labor markets, we derive the optimal labor contract for firms with both equity and debt, and show that it implies employees will become entrenched and therefore face large human costs of bankruptcy. The firm’s optimal capital structure emerges from a trade-off between these human costs and the tax benefits of debt. Our model delivers optimal debt levels consistent with those observed in practice without relying on frictions such as moral hazard or asymmetric information. In line with existing empirical evidence, our model implies persistent idiosyncratic differences in leverage across firms, and shows that wages should have explanatory power for firm leverage.

Download PDF

This paper conducts a comprehensive study of the optimal exercise policy for an executive stock option and its implications for option cost, average life, and alternative valuation concepts. The paper is the first to provide analytical results for an executive with general concave utility. Wealthier or less risk-averse executives exercise later and create greater option cost. However, option cost can decline with volatility. We show when there exists a single exercise boundary, yet demonstrate the possibility of a split continuation region. We also show that, for CRRA utility, the option cost does not converge to the Black-Scholes value as the correlation between the stock and the market portfolio converges to one. We compare our model’s option cost with the modified Black-Scholes approximation typically used in practice, and show that the approximation error can be large or small, positive or negative, depending on firm characteristics.

Download PDF

This paper presents a general, nonlinear version of existing multifactor models, such as Longstaff and Schwartz (1992). The novel aspect of our approach is that rather than choosing the model parameterization out of “thin air”, our processes are generated from the data using approximation methods for multifactor continuous-time Markov processes. In applying this technique to the short- and long-end of the term structure for a general two-factor diffusion process for interest rates, a major finding is that the volatility of interest rates is increasing in the level of interest rates only for sharply, upward sloping term structures. In fact, the slope of the term structure plays a larger role in determining the magnitude of the diffusion coefficient. As an application, we analyze the model’s implications for term structure pricing, focusing on the conditional distribution of interest rates and the term structure of term premiums and volatilities.

Download PDF

This paper develops a rational, liquidity-based model of closed-end funds (CEFs) that provides an economic motivation for the existence of this organizational form: They offer a means for investors to buy illiquid securities, without facing the potential costs associated with direct trading and without the externalities imposed by an open-end fund structure. Our theory predicts the patterns observed in CEF IPO behavior and the observed behavior of the CEF discount, which results from a tradeoff between the liquidity benefits of investing in the CEF and the fees charged by the fund’s managers. In particular, the model explains why IPOs occur in waves in certain sectors at a time, why funds are issued at a premium to net asset value (NAV), and why they later usually trade at a discount. We also conduct an empirical investigation, which, overall, provides more support for a liquidity-based model than for an alternative sentiment-based explanation.

Download PDF

The high persistence of interest rates has important implications for the preferred method used to estimate term structure models. We study the finite-sample properties of two standard dynamic simulation methods—efficient method of moments and indirect inference—when they are applied to an AR(1) process. For data that are as persistent as interest rates, the finite-sample properties of EMM differ dramatically both from their asymptotic properties and from those of indirect inference and maximum likelihood. EMM produces larger confidence bounds than indirect inference and maximum likelihood, yet they are much less likely to contain the true parameter value. This is primarily because the population variance of the data (which is extremely sensitive to the autocorrelation coefficient for persistent data) plays a much larger role in the moment conditions for EMM than in the moment conditions for either indirect inference of maximum likelihood. These results suggest that indirect inference is preferable to EMM when working with persistent data such as interest rates.

Download PDF

Despite the importance of leases in the US economy, and the existence of several theoretical lease pricing models, there has been little systematic attempt to estimate these models. This paper proposes a simple no-arbitrage based lease pricing model, and estimates it using a large proprietary datset of leases on several property types. We find sizeable pricing errors that cannot be explained using interest rates, lease maturity, or information on the options embedded in the contracts, suggesting the presence of significant mispricings in the market for real estate leases. We also propose a new measure, the Option-Adjusted Lease Spread, or OALS (analogous to an option’s implied volatility, or a mortgage-backed security’s Option-Adjusted Spread), that allows leases with different maturities and contract terms to be compared on a consistent basis.

Download PDF

This paper shows that the existence of managerial ability, combined with the labor contract prevalent in the industry, implies that the closed-end fund discount will exhibit many of the primary features documented in the literature. We evaluate the model’s ability to match the quantitative features of the data, and find that it does well, although there is some observed behavior that remains to be explained.

Download PDF

This paper develops a two-factor structural mortgage pricing model in which rational mortgage-holders choose when to prepay and default in response to changes in both interest rates and house prices. We estimate the model using comprehensive data on the pool-level termination rates for Freddie Mac Participation Certificates issued between 1991 and 2002. The model exhibits a statistically and economically significant improvement over the nested one-factor (interest-rate only) model in its ability to match historical prepayment data. Moreover, the two-factor model produces origination prices that are significantly closer to those quoted in the TBA market than the one-factor model. Our results have important implications for hedging MBS.

Download PDF

The standard VaR approach considers only terminal risk, completely ignoring the sample path of portfolio values. In reality interim risk may be critical in a mark-to-market environment. Sharp declines in value may generate margin calls and affect trading strategies. In this paper we introduce the notion of MaxVaR, analogous to VaR in every way except it quantifies the probability of seeing a given loss on or before the terminal date rather than at the terminal date. Under standard set of assumptions we provide a simple formula for MaxVaR and examine the ratio of MaxVaR to VaR. For reasonable parameterizations MaxVaR may exceed VaR by over 40%. MaxVaR exceeds VaR by as much as 80% or more for high Sharpe Ratio hedge-fund-like sets of portfolio return distribution.

Download paper from JOIM Web site

The regulations governing asset distributions from many retirement plans give plan participants the option to time retirement or rollovers from the plan strategically. They possess a long-lived put option, whose exercise price resets periodically to the current value of the assets in the plan. We derive a recursive closed-form valuation formula for the option, and develop a numerical algorithm for implementing the result. We find that, for reasonable assumptions about volatility and life expectancy, the option’s value may approach 40% of the value of the assets in the plan, financed entirely by those still contributing. This wealth transfer can, however, be easily avoided by making a simple change to the current regulations governing valuation and payout of these retirement plans.

Download PDF

This chapter presents a non-parametric technique for pricing and hedging mortgage-backed securities (MBS). The particular technique used here is called multivariate density estimation (MDE). We find that MBS prices can be well described as a function of two interest rate factors; the level and slope of the term structure. The interest rate level proxies for the moneyness of the prepayment option, the expected level of prepayments, and the average life of the MBS cash flows, while the term structure slope controls for the average rate at which these cash flows should be discounted. We also illustrate how to hedge the interest rate risk of MBS using our model. The hedge based on our model compares favorably with existing methods.

Download PDF

This paper analyzes the dynamics of the commonly used indices for Adjustable Rate Mortgages, and systematically compares the effects of their time series properties on the interest rate sensitivity of adjustable rate mortgages. Our ARM valuation methodology allows us simultaneously to capture the effects of the dynamics of the index, discrete coupon adjustment, and caps and floors, and can either calculate an optimal prepayment strategy for mortgage holders, or use an empirical prepayment function. We find that the different dynamics of the major ARM indices induce significant variation in the interest rate sensitivities of loans based on different indices. We also find that changing assumptions about contract features, such as loan caps and coupon reset frequency, has a significant, and in some cases unexpected, impact on our results.

Download PDF

This paper shows that, in the presence of transaction costs payable by borrowers on refinancing, it is possible to construct a separating equilibrium in which borrowers with differing mobility select loans with different coupon rate/points combinations. We also show that, in the absence of such costs, no such equilibrium is possible. This provides a possible explanation for the large menus of mortgages typically encountered by potential borrowers, and suggests that the menu of contracts available at the time of origination should be an important predictor of future prepayment. We numerically implement such an equilibrium, developing the first contingent claims mortgage valuation algorithm that can quantify the effect of self-selection on real contracts in a realistic interest rate setting. Our algorithm allows investors to account for self-selection when valuing mortgages and mortgage-backed securities. It also, for the first time, allows lenders to determine the optimal points/coupon rate schedule to offer to a specified set of potential borrowers, given the current level of interest rates.

Download PDF

This paper uses multivariate density estimation (MDE) procedures to investigate the pricing of mortgage-backed securities (MBS) in a multifactor interest rate environment. The MDE estimation suggests that weekly MBS prices from January 1987 to May 1994 can be well described as a function of the level and slope of the term structure. We analyze how this function varies across MBSs with different coupons and investigate the sensitivity of prices to the two factors. An important finding is that the interest rate level proxies for the moneyness of the option, the expected level of prepayments, and the average life of the cash flows, while the term structure slope controls for the average rate at which these cash flows should be discounted. Though the origination and prepayment behavior of mortgages differ substantially across coupons, there remains an unexplained common factor which explains 80-90% of the remaining variation of MBS prices. This factor does not seem to be related to the usual suspects, and therefore presents a puzzle to financial economists.

Download PDF

This paper presents a technique for nonparametrically estimating continuous-time diffusion processes which are observed at discrete intervals. We illustrate the methodology by using daily three and six month Treasury Bill data, from January 1965 to July 1995, to estimate the drift and diffusion of the short rate, and the market price of interest rate risk. While the estimated diffusion is similar to that estimated by Chan, Karolyi, Longstaff and Sanders (1992), there is evidence of substantial nonlinearity in the drift. This is close to zero for low and medium interest rates, but mean reversion increases sharply at higher interest rates.

Download PDF

Previous mortgage prepayment and valuation models assume that two mortgage pools with the same observable characteristics should behave indistinguishably. However, even pools with apparently identical characteristics often exhibit markedly different prepayment behavior. The sources of this heterogeneity may be unobservable, but we can infer information about a pool from its prepayment behavior over time. This paper develops a methodology for using this information to calculate pool-specific mortgage-backed security prices. Knowledge of these prices is important both for portfolio valuation and for determining the cheapest pool to deliver when selling mortgage-backed securities. We find that unobservable heterogeneity between mortgage pools is statistically significant, and that pool-specific prices, calculated for a sample of outwardly identical mortgage pools between 1983 and 1989, may differ greatly from any single representative price.

Download PDF

This paper develops a new strategy for dynamically hedging mortgage-backed securities (MBSs). The approach involves estimating the joint distribution of returns on MBSs and T-note futures, conditional on current economic conditions. We show that our approach has a simple intuitive interpretation of forming a hedge ratio by differentially weighting past pairs of MBS and T-note futures returns. An out-of-sample hedging exercise is performed for 8%, 9% and 10% GNMAs over the 1990-1994 period for weekly and monthly return horizons. The dynamic approach is very successful at hedging out the interest rate risk inherent in all of the GNMAs. For example, in hedging weekly returns on 10% GNMAs, our dynamic method reduces the volatility of the GNMA return from 41 to 24 basis points, whereas a static method manages only 29 basis points of residual volatility. Moreover, only 1 basis point of the volatility of the dynamically hedged return can be attributed to risk associated with U.S. Treasuries, which is in contrast to 14 basis points of interest rate risk in the statically hedged return.

Download PDF

This article presents a new model of mortgage prepayments, based on rational decisions by mortgage holders. These mortgage holders face heterogeneous transaction costs, which are explicitly modeled. The model is estimated using a version of Hansen’s (1982) generalized method of moments, and shown to capture many of the empirical features of mortgage prepayment. Estimation results indicate that mortgage holders act as though they face transaction costs that far exceed the explicit costs usually incurred on refinancing. They also wait an average of more than a year before refinancing, even when it is optimal to do so. The model fits observed prepayment behavior as well as the recent empirical model of Schwartz and Torous (1989). Implications for pricing mortgage-backed securities are discussed.

Download PDF

This paper analyzes adjustable rate mortgages (ARMs) based on the Eleventh District Cost of Funds Index (EDCOFI). We study the behavior of EDCOFI over the period 1981-1993, and find that adjustments in this index lag substantially behind term structure fluctuations. We also find that the seasonality and days-in-the-month effects noted by previous authors are really symptoms of a “January effect”.

Due to the lag in EDCOFI, if interest rates fall, mortgage holders may want to refinance their mortgage loans to avoid paying a coupon rate that exceeds the market rate. We develop a finite difference valuation algorithm which accounts for all usual ARM contractual features, in addition to the dynamics of EDCOFI. The advantage of our pricing algorithm over commonly used simulation strategies is that it allows us to determine endogenously the optimal prepayment strategy for mortgage holders, and hence the value of their prepayment options. We find that the dynamics of EDCOFI give significant value to this option, typically around 0.5% of the remaining principal on the loan. Our algorithm permits issuers and investors in ARMs based on EDCOFI to quantify the effects of the many interacting contract features, such as reset margin, coupon rate caps and reset frequency, that determine mortgage value.

Download PDF

This paper is a study of continuously resettled contingent claims prices in a stochastic economy. As special cases, the relationship between futures and forward prices is analyzed, and a preference-free expression is derived for these prices, as well as the price of a continuously resettled futures option, whose formula differs from Black’s futures option pricing formula due to the effects of marking-to-market the changes in the futures option premium.

Download PDF

This paper studies the effects of wildfires on housing and mortgage markets. We motivate our empirical investigation with a game-theoretic model of homeowners’ decisions to rebuild or improve their homes, considering both neighborhood externalities and insurance. We test the model’s implications using California data from 2001 to 2015. We find an increase in house prices and square footage in wildfire treatment areas five years post-fire. We also find decreases in mortgage terminations, but little evidence of gentrification. Our analysis of expected wildfire losses challenges the ability of insurance companies to absorb these losses without a serious reconsideration of property- and casualty-insurance pricing in California.

Download PDF

Using a newly available panel data set containing property-specific, time-varying hedonic characteristics and sales prices, we develop a new dynamic house-price model that is suitable for out-of-sample forecasting applications such as mortgage valuation and bank stress-testing. The model is set up in a classical state-space framework and includes common factors that are univariate structural time series models scaled to form linear combinations specific to locations. Our common factors include trend, seasonal, and autoregressive components. The equations are linear and errors are Gaussian; however, the unbalanced nature of our panel data means that standard Kalman-filter smoothing algorithms are not suitable. Instead, we apply an alternative four-block Markov Chain Monte Carlo algorithm (Strickland, Turner, Denham and Mengersen, 2009). We find significant in- and out-of-sample forecasting differences between our model and standard repeat-sales and hedonic regressions. We also test, and reject, two assumptions of repeat-sales estimators: a single common trend for all regions and a unit root in the index.

Download PDF

Energy efficiency is a key factor to the future of the U.S. economy, and commercial buildings are among the largest users of energy. However, existing commercial mortgage underwriting practices provide little incentive for building owners to make their buildings more energy efficient. In this paper, we extend standard mortgage valuation methods, which account for the expected dynamics of interest rates and building prices, by including the expected dynamics of the electricity and gas forward prices as well as average building-level energy consumption. This allows us explicitly to incorporate energy risk and efficiency measures, which depend on both the energy efficiency of the building and the characteristics of its location, into commercial mortgage valuation and underwriting. We apply our valuation methodology to price a sample of 1,390 mortgages, originated between 2005 and 2007 on office buildings located in 28 cities across the U.S. We find that, relative to the traditional mortgage valuation methodology, our proposed strategy leads to an 5% reduction in the mispricing of the default risk of commercial mortgages.

Download PDF

Combining insights from the contingent claims and the asset-backed securities literatures, we study the economics of value creation in the asset management business. In particular, we provide a theoretical model and a closed form formula for the value of fund fees in the presence of the well known flow-performance relation, giving rise to interesting nonlinearities and volatility-related effects. The theoretical model sheds light on the role of fees, asset growth, asset and benchmark volatility, and the intensity of the flow-performance relation. To better understand the role of changing fund characteristics such as age and size on the fund value and fund risk, we estimate the empirical relation between returns and flows conditional on these characteristics for various asset classes. We study these effects using Monte Carlo simulations for various economically meaningful parameter values for specific asset classes. Measuring value as a fraction of assets under management, we find that both value and risk, systematic and idiosyncratic, decline in size and age. In addition, value is a complex, non-monotonic function of the fee charged on the fund.

Download PDF