MORTGAGE SERVICING COLLABORATIVEOptions for Reforming the MortgageServicing Compensation ModelKaran Kaul, Laurie Goodman, Alanna McCargo, and Todd Hill-JonesApril 2019The Mortgage Servicing Collaborative (MSC) is an initiative led by the Urban Institute’s HousingFinance Policy Center that brings together lenders, servicers, consumer groups, civil rights leaders,researchers, and policymakers who appreciate the impact of servicing on the health of the housingfinance system. By calling on a broad range of perspectives and expertise, the MSC has worked toward awell-grounded view of the policy challenges in servicing and a thoughtful, data-driven approach toaddressing them.Since its inception in early 2017, the MSC has published four policy research briefs with specificrecommendations to improve the functioning of the servicing market for consumers and the industry.The first framing brief explained the high-level importance of servicing and outlined key issues in needof reform (Goodman, McCargo, et al. 2018). The second brief recommended enhancements to the loanmodification toolkit for government-insured loans in a rising-rate environment (Goodman, Kaul, et al.2018). The third brief proposed critical improvements to the foreclosure and conveyance processes formortgages insured by the Federal Housing Administration (FHA) to achieve better outcomes forborrowers, reduce costs for servicers, and cut loss severities for the FHA (Kaul et al. 2018a). The fourthbrief focused on the need for uniform mortgage servicing data standards and explained how standardscan improve customer service, reduce costs, and lay the groundwork for future innovation (Kaul et al.2018b). In this final research brief, we turn to the fundamental question of how servicing compensation,especially for nonperforming loans (NPLs), could be structured going forward.

About the Mortgage Servicing CollaborativeThe Housing Finance Policy Center’s Mortgage Servicing Collaborative is a research initiative that seeks to identify and buildmomentum for servicing reforms that make the housing market more equitable and efficient.One core MSC objective is to improve awareness of the role and importance of mortgage servicing in the housing financesystem. Since 2013, the Housing Finance Policy Center’s researchers have studied the landscape, followed the work andpolicies put in place after the crisis, and assessed the impact of the servicing industry on consumers and communities. Thisincludes loss mitigation and foreclosure actions and how servicing practices affect access to credit through tightunderwriting standards. The Urban Institute has analyzed and convened forums on emerging issues in mortgage servicing,including calls for reforms, the impact of mortgage regulation, the rise of nonbank servicers, and the implications forconsumers and communities. We determined that a focused effort that involves external stakeholders and resources couldlead the way in developing policy and structural recommendations and bring visibility to the important issues that lie ahead.The MSC has convened key industry stakeholders—including lenders, servicers, consumer groups, civil rightsorganizations, academics, and regulators—to develop an evidence-based understanding of important factors and to developand analyze solutions and implications with a well-rounded and actionable orientation.The MSC seeks to: bring new evidence, data, and recommendations to the forefront; foster debate and analysis on issues from regulatory reform, technology innovations, cost containment, and consumeraccess to mortgages; and produce and disseminate our research findings and policy recommendations—including perspectives by MSC members—to offer policy options that can clarify and advance the debate and ensure servicing is addressed in broader housingfinance reform.For more information about the MSC or to see other publications, news, and products, visit the MSC program collaborative.Mortgage Servicing Collaborative Participants AmeriFirst: Mark Jones and Greg WarnerBank of America: Terry Laughlin and Larry WashingtonBayview: Rich O’Brien and Julio AldecoceaBlack Knight Financial Services: Joseph NackashiCaliber Home Loans: Tricia Black, Marion McDougall, and Lori FosterColonial Savings: David Motley and Jane LarkinGenworth: Steve Hall and Carol BouchnerGuild Mortgage Company: David BattanyHousing Policy Council and Hope Now: Paul Leonard, Meg Burns, and Eric SelkJPMorgan Chase: David Beck, Ramon Gomez, Erik Schmitt, and Diane KortKatie Porter: Faculty at University of California, IrvineMortgage Bankers Association: Justin Wiseman, Mike Fratantoni, and Sara SinghasMr. Cooper: Jay Bray and Dana DillardNational Community Stabilization Trust: Julia GordonNational Fair Housing Alliance: Lisa RiceNorthern Ohio Investment Corp.: Mark VinciguerraOcwen: John Britti and Jill ShowellPatricia McCoy: Faculty at Boston CollegePennyMac: David Spector and Karen ChangPricewaterhouseCoopers: Sherlonda Goode-Jones, Peter Pollini, and Genger CharlesQuicken Loans: Mike Malloy, Pete Carroll, and Alex McGillisSelf Help Credit Union/Center for Responsible Lending: Martin Eakes and Mike CalhounTed Tozer: Milken Institute, Former President of Ginnie MaeUnion Home Mortgage: Bill CosgroveU.S. Bank: Bryan BoltonWells Fargo: Brad Blackwell, Raghu Kakumanu, and Laura Arce

Why Is the MSC Studying Servicing Compensation?Servicing fees compensate mortgage servicers for managing loans from postclosing until they areterminated, either through payoff (i.e., at the maturity of the loan or earlier, the latter being morecommon), foreclosure, or a foreclosure alternative, such as a deed-in-lieu or short sale. Servicers receivea fixed fee annually, usually 25 basis points (or 0.25 percent)—although it can be as high as 50 basispoints—of the unpaid principal balance for mortgages guaranteed by Fannie Mae and Freddie Mac. Forgovernment-insured loans, which are securitized into Ginnie Mae securities, servicing fees can varyfrom 19 to 69 basis points annually. Although the basis points are fixed for the life of the loan, theabsolute dollar fee declines as the principal balance is paid down. Servicing fees are a component of themortgage rate and are paid by borrowers as part of their monthly payment. The servicer retains theservicing fee portion of the monthly payment and passes the rest through to the investor who owns theloan. 10FServicer responsibilities vary drastically depending on whether a loan is performing ornonperforming (FHFA 2011). Before the Great Recession, servicing consisted primarily of generatingmonthly statements, collecting and processing borrower payments, and keeping track of fees andescrows. Most homeowners made their payments on time. For the small share of borrowers that wentdelinquent from time to time, servicers were required to engage in “loss mitigation” to minimize lossesfor loan investors, insurers, or guarantors. This can be a government entity (the FHA, the USDepartment of Veterans Affairs, or the US Department of Agriculture), a government-sponsoredenterprise (Fannie Mae or Freddie Mac), or a private investor. Low default rates historically meant thatservicers deployed a limited operational infrastructure for loss mitigation.But the housing market crash and the resulting increase in delinquencies and foreclosures createdsignificant stress for the precrisis servicing model. Servicers were still responsible for collecting andprocessing borrower payments and for minimizing investor losses, but the skyrocketing number ofdistressed borrowers in need of assistance required servicers to invest in and rapidly expand theiroperations at a time when the mortgage market was under tremendous financial stress. 21FIn addition, the entities owning, insuring, or guaranteeing the loan (such as Fannie Mae, FreddieMac, government mortgage insurers, or private investors) can make material changes to loan servicingrequirements after origination without changing servicing fees. During the last housing crisis, FannieMae, Freddie Mac, the FHA, the Department of Veterans Affairs, the Consumer Financial ProtectionBureau, and the US Department of the Treasury developed new loss mitigation programs, revampprecrisis servicing policies, and created new regulations for the servicing industry. Some of thesechanges included requirements for every step of the loss mitigation process, such as how often andwhen servicers can contact a struggling borrower, what documentation to request from the borrower,and what loss mitigation options to offer in what order. As a result, servicing responsibilities and costshave increased.OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL3

In light of these changes, some believe that the structure of servicing compensation must bechanged to better align servicing costs and revenues for performing and nonperforming loans in amanner that will improve outcomes for servicers and consumers. Others believe that the presentcompensation model, coupled with postcrisis reforms and recommendations from the previous MSCbriefs, can promote an efficient servicing market with minimal risk of disruption. Whether and whatchanges to the servicing compensation model might be needed is an open question. The MSC studiedservicing compensation and examined three specific options: the status quo, a fee-for-service model,and a central default utility model. In this brief, we discuss pros and cons of each model and assess howeach would perform along the following dimensions: Industry preparedness for downturns Borrower outcomes Alignment of interests Access to mortgage credit Risk of market disruptionBecause of lack of agreement on which model would work best, this brief neither representsconsensus views of MSC members nor recommends any one model. Rather, it provides a window intohow a diverse set of servicing market stakeholders view each option.Structural Options for Servicing Compensation DesignOption 1: The status quoMSC members in favor of preserving the status quo believe that the current model, in conjunction withprior MSC recommendations, would reduce NPL costs and improve borrower outcomes. Today’sservicing market is also much improved than the one going we had into the last recession given thereforms made since. For instance, the current loss mitigation toolkit is more robust than was the casehistorically—it includes loss mitigation programs for various distressed borrower scenarios, such asnegative equity, job loss, and natural disaster. In addition, there is now better operational infrastructurefor nonperforming loans, enabled by the rise of specialty default servicers, as well as stronger serviceroversight by loan guarantors, new consumer protection requirements, 3 and a high-quality postcrisis2Fbook of business.Another argument in favor of retaining the present structure is to avoid the risk of disruption.Servicing compensation is not an isolated issue. Any changes to it will have implications for the scope ofservicer responsibility, mortgage servicing rights (MSR) values and liquidity, servicer revenues,profitability, and customer service. Both bank and nonbank servicers would be affected, but nonbanksmore so. The two have different business models and financing structures, with nonbanks more relianton MSR financing.4OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL

Nonbanks have increased their servicing market presence in recent years as banks have pulledback. Before the housing crisis, banks serviced approximately 70 percent of outstanding unpaidprincipal balance of single-family mortgages. After peaking at 94 percent in 2010, 4 the bank share has3Fdeclined gradually, ending 2018 at 55 percent, with nonbanks accounting for the remaining 45 percent(figure 1). It is unclear what impact new compensation model would have on MSR values. If negative,there would be implications for MSR financing and servicing capacity.FIGURE 1Bank and Nonbank Shares Based on Unpaid Principal Balance Serviced, %68%37%63%44%45%56%55%2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018Source: Urban Institute calculations based on Inside Mortgage Finance data.Note: Shares from 2002 to 2004 are based on top 30 servicers, 2005 to 2007 shares are based on top 40 servicers, and sharesfrom 2008 and onwards are based top 50 servicers.URBAN INSTITUTEMaintaining the present compensation model also comes with a big risk, according to somemembers of the collaborative—it does not provide a formal structure to fund the high cost of servicingnonperforming loans. The cost of servicing performing loans is significantly lower than the annual 25basis points in compensation (for conventional mortgages), but the cost of servicing nonperformingloans is substantially higher. Servicing NPLs is costly, high touch, and labor intensive because servicersmust understand individual borrower circumstances, evaluate multiple loss mitigation options inaccordance with investor guidelines, maintain effective and timely communication with borrowers, andrecommend the best approach to cure delinquencies.Servicers are also generally responsible for advancing principal and interest payments to securitiesinvestors, property taxes to local authorities, and escrows for homeowners’ insurance using their ownfunds while the loan is delinquent. They eventually get reimbursed, but they must have financialresources to absorb significant cost escalations. This situation can create financial stress if more loansthan anticipated default at the same time. NPL costs are high for all loan types but are especially acutefor FHA loans because of inefficient loss mitigation, foreclosure, and conveyance processes and highlyOPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL5

punitive timeline requirements. Servicing FHA NPLs is three times more expensive than servicingconventional NPLs (Kaul et al. 2018a). High NPL servicing cost is one of the reasons why access to creditfor low- and moderate-income households remains tight by historical standards. 54FAnother shortcoming of the current compensation model has to do with the timing of whenservicers get paid. Borrowers pay servicing fees monthly as part of the mortgage payment. Thus,servicers receive full servicing fees while a loan is performing but receive effectively nothing once itbecomes nonperforming. This makes the business highly procyclical—it can be highly profitable in goodtimes, but high costs and lost revenues can exacerbate financial stress during bad times. Ultimately,inadequate financial resources impair servicers’ ability to respond effectively during downturns andlead to poor outcomes for borrowers, neighborhoods, and investors. Accordingly, some members of theMSC believe that the servicing compensation structure should be reformed to better align revenueswith costs and improve preparedness for downturns. The MSC discussed two alternate compensationstructures, each with its own benefits and risks.Option 2: A fee-for-service model for nonperforming loansUnder this model, servicers would handle the servicing of performing loans as they do today. For themost part, their loss mitigation responsibilities would also remain the same. But the way they getcompensated would change to ensure continued revenues, even for delinquent loans. Instead ofreceiving the full 25 basis point fee for performing and effectively nothing for nonperforming loans,servicers would retain less than 25 basis points while the loan is current. The remainder would be savedand tapped to pay for loss mitigation in one of two ways:Servicers could pay a predefined number of basis points on an ongoing basis to the loan insurer, guarantor, or investor (Fannie Mae, Freddie Mac, a government mortgage insurer, or a privateinvestor). In return, these entities would agree to pay servicers flat dollar fees duringdelinquencies tied to specific loss mitigation actions and positive borrower outcomes.Alternatively, servicers could pay the same money into a separate insurance fund, thus bypassing the guarantors. The fund would pay servicers flat fees for delinquent loans based onloss mitigation tasks or outcomes. This approach would avoid concentrating more risk withinloan guarantors, but it would require setting up a new entity.Both alternatives would allow servicers to receive revenue while the loan is in default. A source ofrevenue in stressful times would improve systemic resilience and enable servicers to better servestruggling borrowers. To a limited extent, a version of fee-for-service is used today by governmentsponsored enterprises 6 and the US Department of Housing and Urban Development, 7 generally to5F6Freimburse servicers for certain property preservation–related expenses. During the Great Recession,the federal government offered incentive payments to servicers for certain modifications under theHome Affordable Modification Program while it was in effect. 8 Some recent private-label securitization7F9transactions have gone further by paying servicers a small monthly fee to cover the cost of servicing8F6OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL

performing loans. Once a loan becomes delinquent, these fees increase depending on the state ofdelinquency and loss mitigation success.For example, under JPMorgan Mortgage Trust 2018–9, the servicer receives a base servicing fee of 20 per month per performing loan. Once a loan becomes 60 to 119 days delinquent, the servicerreceives an additional 211 per month per loan. For loans at least 120 days delinquent, the servicerreceives 252 per month. Although these fees help pay for general cost escalations, servicers areeligible to receive additional incentive fees for cured defaults, completion of loan modifications, and thelike. For instance, they can earn 2,500 for completing a loan modification and 1,000 for curing a loan60 or more days delinquent. Although there are multiple ways to structure a fee-for-servicearrangement, the main purpose is the same—it provides servicers access to additional funds to meettheir obligations to borrowers and investors during stressful times. More importantly, knowing they willnot be on the hook for severe cost escalations for delinquent loans, lenders might be more willing toease credit overlays.But a fee-for-service model has drawbacks. It will need to build incentives that encourage proactiveearly intervention to minimize delinquencies. Some MSC members worry that the prospect of earninghigher fees for delinquent loans could discourage servicers from pursuing aggressive early interventionto help prevent defaults in the first place. This could result in more loans becoming delinquentcompared with the current system in which servicers—fearful of revenue loss—have a strong incentiveto keep borrowers from going delinquent. Others believe higher fees would create better alignment ofinterests between servicers (greater incentive to cure defaults to earn higher fees), investors (reducedloan losses) and borrowers (more effective loss mitigation).And because fee-for-service would be paid for using a portion of the 25 basis points strip, it wouldaffect MSR valuation. The impact would depend on the ongoing fees servicers pay and the revenuesthey receive for NPLs, assumptions about expected default rates, and the discount rates used. Althoughthe precise impact on MSR values requires further study, the risk is that if values are lower, servicersthat hold MSRs on their balance sheets will have a less valuable asset compared with today, andservicers that sell or pledge MSRs to raise funds will receive less in return. Both banks and nonbankswould be affected, but the latter more so. Nonbanks depend more on MSR financing because, unlikebanks, they lack access to consumer deposits and other diversified funding sources.The MSC discussed mitigating the risk of disruption by making fee-for-service optional. Servicerscould either keep the full 25 basis points fee for performing loans as today or pay a portion of it to haveloan guarantors (or an insurance fund) absorb NPL costs. Servicers that prefer the status quo couldretain it, while those who prefer fee-for-service could switch over. But it is doubtful this would work,given the risk of adverse selection. With optionality, there would be a temptation to use guarantors (orthe insurance fund) only for loans that are more likely to default. 10 This can significantly raise the cost of9Fseeking protection for servicers and render the whole structure economically unworkable. Additionally,having two compensation models in place could bifurcate the servicing market into submarkets,complicate MSR transfers, and affect liquidity.OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL7

Lastly, under the guarantor-centric approach, many risks and costs associated with NPL servicingwould be shifted from servicers to the entity backing the loan, which ultimately holds credit risk. Someworry this would further increase risk concentration among guarantors (Fannie Mae and Freddie Mac)or government mortgage insurers (the FHA, the US Department of Veterans Affairs, and the USDepartment of Agriculture). The insurance fund structure avoids this risk, but the new entity thathouses the fund would have to be built from scratch. It would have to build the expertise to monitor loanquality, manage credit risk, and price the insurance to ensure reserve sufficiency.Option 3: A central default utility for nonperforming loansSome members of the MSC support the creation of a central utility for default servicing. This would be abigger change to the system than the fee-for-service model. It would change both the scope of workservicers perform and how they get paid for it. Servicers would continue to service performing loans asthey do today but would outsource 11 the loss mitigation function to a central default utility once a loan10 Fbecomes 60 days delinquent.1211FThe utility would have a specialized, mission-like focus on curingdelinquencies and would be financed in a manner similar to the fee-for service model—that is, using aportion of the servicing fee. This arrangement could be structured in two ways:Establish the utility as industry owned to align its interests with servicers. The utility would not necessarily have to build out its own end-to-end default servicing infrastructure. Instead, itwould have the flexibility to do so, rely on specialized subservicers, or use a combination of thetwo. Servicers could fund the utility through direct ongoing contributions or through aninsurance fund housed outside the utility. Using a fund would reduce risk concentration in theutility.Alternatively, servicers could pay the same fee to loan guarantors who would be responsible for loss mitigation. They could hire either the utility or the specialty subservicers.Under both structures, servicers would no longer perform loss mitigation and would be largelyfreed from associated costs and risks. The scope of work for servicers would essentially be reduced tomanaging loans until they become 60 days delinquent. This would mitigate servicer financial andoperational stress during downturns. More importantly, knowing they will not be on the hook forsignificant NPL costs escalations, lenders might be more willing to ease servicing-driven credit overlays.The utility could benefit from better economies of scale. Although NPL servicing is a high-variablecost business, it requires a substantial fixed-cost investment in operational systems and technology.Spreading these costs over a larger number of delinquent loans could mitigate per loan costs to somedegree. A utility would also avoid the misaligned incentive issue with fee-for-service that concerns someMSC members. Because neither the utility nor servicers can receive additional compensation for NPLs,the incentive to earn higher profits from delinquent loans will not exist.Despite these improvements, the utility model has some challenges. The first is the risk of adverseimpact on MSR values, nonbank financing, servicing capacity, and liquidity. With nonbanks composing8OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL

over 80 percent of FHA and VA originations and over half of GSE originations (figure 2), anyrepercussions would be widely felt.FIGURE 2Nonbank Origination Share by rce: Urban Institute Calculations based on eMBS data.URBAN INSTITUTEThe risk of disruption could theoretically be mitigated by making the utility model optional.Servicers would prospectively choose either status quo or pay ongoing fees to have the utility take overloss mitigation at 60 days delinquent. But how this would work is unclear, given that nonbank servicerswill likely opt for status quo. Given their dominant market share, it is unlikely the utility model will workwell without their participation. Optionality would also create an adverse incentive to tap the utilityonly for less creditworthy loans that are more likely to default. The utility could risk-based price, but ifits portfolio is not well diversified across the credit spectrum, the cost of protection will be very high,rendering the model economically unsustainable. Some also worry that layering another risk-basedpricing system on top of government-sponsored enterprise loan-level pricing adjustments andmortgage insurance premiums could price out even more borrowers with low credit scores. Lastly, thereexists the risk of bifurcating servicing into submarkets, complicating MSR valuation and affectingliquidity.There are also fears that the utility would have significant market power, which could lead tomonopolistic behavior. Although structuring the utility as servicer owned would align its interests withthose of servicers, it could still wield too much power over other market stakeholders that do not ownthe utility. This could reduce competition in default servicing, especially if the utility builds its own endto-end platform. Specialty servicers that have built strong loss mitigation expertise over the years couldfind it difficult to compete.OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL9

According to some, anticompetitive fears could be addressed by structuring the utility asgovernment run, similar to the Federal Deposit Insurance Corporation model with an explicit risk-basedfee. It is, however, an open question as to how nimble a government entity would be in responding toborrowers in need of immediate assistance. It is also an open question as to how quickly a utility,whether industry or government owned, can ramp up operations during downturns without carryingtoo much excess capacity during good times.The utility model would also concentrate the entire default servicing function and associatedfinancial risks in a single entity. The amount of funding the utility receives will depend heavily onassumptions about future default rates and estimates of the utility’s costs. If these risks areunderestimated, the utility may not have adequate financial resources to carry out its responsibilitiesduring severe economic downturns, thus increasing the risk of large insolvency.Table 1 below summarizes pros and cons of each of the three options.TABLE 1Comparison of Options 1, 2, and 3Option 1:Status quoOption 2:Fee-for-serviceOption 3:Central utilityAbility to respondduring downturnsNo impactImprovedImprovedCompensation forservicing NPLsNo compensationSome compensationSome compensationProfitability onperforming loansNo impactLowerLowerMSR valuesRisk of marketdisruptionConcentration riskServicing-driven creditoverlaysBorrower outcomesNo impactLowestUnclearHigherUnclearHigherLowestNo impactHigher, can be mitigatedFewer overlaysHigher, can be mitigatedFewer overlaysNo impactBetter, but risk of misalignedincentivesBetter, but risk ofmonopolistic behaviorNote: MSR mortgage servicing rights; NPL nonperforming loans.ConclusionThis brief, the final of the MSC, is an attempt to rekindle the conversation about mortgage servicingcompensation. Some MSC members believe the compensation model should be changed, while otherssupport status quo. Those in favor of change believe the present model leaves us inadequately preparedto respond during downturns, as it does not include a formal mechanism to pay for the high cost ofservicing nonperforming loans. Those in favor of the status quo contend that those costs are best10OPTIONS FOR REFORMING THE MORTGAGE SERVICING COMPENSATION MODEL

mitigated by implementing the reforms outlined in prior MSC issue briefs—that is, by improving the loanmodification toolkit for government loans, reforming the FHA’s foreclosure and conveyance processes,and establishing uniform mortgage servicing data standards.The MSC discussed three compensations options—the status quo, a fee-for-service model, and acentral utility model. Each option has pros, cons, and risks. Some of these issues—such as the potentialimpact on MSR values, servicer revenues, and profitability—need to be studied in greater detail. Giventhese risks and trade-offs and the lack of agreement on how to mitigate them, the Mortgage ServicingCollaborative does not recommend any one option over the other. However, by making its collectivethinking public, the collaborative intends to inform future policy discussions on servicing compensation.Notes1For a basic understanding of mortgage servicing, please see the “Servicing 101” videos produced by the MSC,available at “Help Me Understand Mortgage Servicing,” Urban Institute, accessed April 15, financ

The second brief recommended enhancements to the loan modification toolkit for government-insured loans in a rising-rate environment (Goodman, Kaul, et al. 2018). The third brief proposed critical improvements to the foreclosure and conveyance processes for .