Credit Analysis and Valuation Methods for Marketplace Lending Loan Portfolios

[Editor’s note: Many of us have questions on marketplace lending loan valuation so when we were approached to see if we would be interested in a Q&A post on the subject we jumped at the chance. The Lend Academy team created the list of questions and Gunes Kulaligil, Director in Houlihan Lokey’s Financial Advisory Services business provided his answers.]

As Marketplace lenders continue to lend at a fast pace, there has been a significant increase in the past several years in non-bank consumer, student and small business lending. Although these platforms operate in a similar fashion, there is a wide variety of underwriting guidelines thereby producing loans with notably different credit profiles and terms. As a result, prepayment and default performance vary significantly based on the platform originating the loan and the credit grade of the loan as determined by the platform. Additionally, loan performance and platform performance have been mixed, with losses sometimes coming in higher than expected and governance issues shaking investor confidence in the sector.

Despite these challenges, origination rates have climbed back towards their prior highs, and funding sources have been expanded with increased interest from whole loan buyers in flow agreements as well as investors seeking to acquire tranches from securitizations. The recalibration of performance expectations has hit many platforms and continues to be an ongoing process, thus highlighting the need to determine the relevant default and prepayment drivers necessary for accurate fair value analyses. While a robust and liquid secondary whole loan trading market has not emerged as some market participants had hoped, secondary transactions are occurring, especially as loans are aggregated into securitizations. Many of these transactions occur at a premium to par and can provide meaningful insight into pricing and relative credit performance of one platform versus others over time.

1. What are the most prevalent methods of valuing loan portfolios today?

Discounted cashflow (DCF) methodology at the loan or cohort level is the most prevalent valuation methodology used today to value marketplace loan portfolios and related assets, including tranches in securitizations and servicing rights, regardless of the lending vertical. I emphasize “today” because marketplace lending is an evolving corner of consumer & business lending that is small and fragmented in general, with over a hundred lenders, both small and large, with a handful of large lenders accounting for the lion’s share of the issuance. Despite its still nascent nature, marketplace lending has caught the eye of nimble credit-focused alternative investment managers and the relatively risk averse large bond funds alike. Credit investors usually acquire whole loan portfolios or subordinate bonds off of securitizations on a levered basis, whereas many of the larger bond funds focus on the senior bonds in securitizations.

The investment thesis from the credit funds’ perspective is the opportunity to achieve mid single to double digit returns on a levered basis with a short duration versus the bond funds’ thesis which could essentially be described as a yield pickup strategy where the buyers expect to achieve unlevered single digit returns with shorter duration than whole loans and ample structural credit support present in securitizations in the form of subordination and excess spread. Non-bank lending owes its ability to attract the attention of such a diverse group of investors not only due to the sector’s capacity to offer a wide range of credit exposures and duration, but also due to its future growth potential with a confluence of support from increasingly tech-savvy borrower demographics, a de-regulatory environment and innovative partnerships in the Fintech ecosystem.

2. Are valuation methods standardized? If not, why not? How does this lack of a valuation standard affect investors?

Marketplace lending is a fragmented space, and it is also diverse, with innovative forms of underwriting and funding methods being deployed. As it is the case with any fledging industry, some growing pains and consolidation are to be expected. However, the year-over-year growth in origination volumes and the trend of increasing institutional involvement (including via more rated securitizations) bode well for the future of the sector. There are not any standard or widely-accepted valuation methodologies in this space yet. Given that marketplace lending is such a diverse lending landscape, and the relative dearth of publicly-available trade color, matrix pricing approaches may be insufficient for all but the most senior bonds in securitizations. Therefore the standard for valuation, at a high level, is more of a well thought out, supportable and transparent discounted cash flow approach. The specifics of the projection methodologies should be considered on a case-by-case basis, depending on the asset, structure, and other factors. Market color is still useful, but may not be directly applied, but rather considered qualitatively when determining the assumptions utilized in the DCF.

In a way, the lack of a single valuation standard dampens and fuels the interest in space concurrently. It dampens because, despite increasing acceptance of the sector as a promising space to achieve yield pickup by real money investors, the lack of a single valuation standard is a negative mark for such investors. At the same time, however, there still is newness premium, partially due to lack of standardization, leading to the level of returns sought by alternative investors thus fueling demand.

3. How do loan valuation methods differ across lending verticals?

Marketplace lending verticals cover a wide spectrum of product, ranging from $500 installment loans, to $100,000 merchant cash advances (MCAs) made to small businesses, to sub-650 FICO unsecured consumer loans to credit impaired borrowers, to student loans extended to borrowers in medical school with high future earning potential. Thus, any methodology that falls short of incorporating all impactful data to project full cashflows does not do justice to the portfolio. In essence, the assumptions used in the DCF are based on loan characteristics that have the biggest impact on prepayment, default, and recovery behavior. These loan characteristics depend on the asset class but often include underwritten payment schedule (e.g. 36 months amortizing term, 60 months amortizing term, daily pay MCA, etc.) credit metrics (e.g. FICO bands, platform ratings, repeat borrower flags), loan size (e.g. <$5K, $20-$30K, etc.), note rate, and more. These assumptions then feed into the DCF model to project principal and interest cashflows generated from the loan portfolio, incorporating prepayments and defaults, net of recoveries.

While the fundamental methodology (i.e. DCF) is the same across different verticals, the derivation of prepayment and default vectors is highly dependent on the lending vertical, including non-standard features of loans originated by that platform (e.g. no stated APR for MCA, holdbacks on disbursement of principal for fix and flip mortgage loans, the specifics of various grace and other non-payment periods for student loan borrowers), the availability of current and historical performance data, the applicability of historical performance data from other data sources, as well as any structure around the loan portfolio such as forward flow or yield maintenance agreements.

4. Will there be a convergence to one single standard?

Not anytime soon. The DCF approach is probably here to stay. Convergence on a more-specific standard than DCF as an overall approach would first require defining that standard and then forming a consensus around that definition, preferably championed by an industry body. The extent to which the valuation standard of any asset class can be “standardized” is also a function of the maturity of the asset class, the homogeneity of loan features, as well as its liquidity and availability of market color. The types of loans we are discussing here cover a wide spectrum of characteristics, with varying non-standard features and, with the exception of certain consumer and student loans, with limited performance history and secondary market color.

5. What are the pros & cons of each valuation methodology?

Other, simpler methodologies – such as marking loans at par and taking loan loss provisions based on average historical experience – is not appropriate given the unique characteristics of the asset class. There aren’t many pros to valuations generated through such methodologies as they lack the rigor of a full DCF analysis. As mentioned previously haircut, or “matrix” type analysis where portfolios are marked by applying rough valuation adjustments to recent trading levels, is not applicable here given the lack of observable trading activity and the unique nature of each portfolio. As an example, generic methodologies often fail to account for seasoning of the loans which can have a sizeable impact on prepayment and default. For seasoned loans, often the DCF analysis should “step in to the curve” i.e. consider the loan age when determining prepay and default expectations.

6. What are the factors that affect valuations of loans? Which measures are the most important in valuing a loan portfolio?

As is the case with most questions regarding marketplace lending, the answer is it depends. Given the wide variety of borrowers with varying credit histories, loan features and structures, all major risk factors (such as prepayment, default, loss severity, duration, convexity, reinvestment, regulatory, liquidity) exist in this space in multitudinous combinations. While senior bonds in securitizations have relatively less credit risk than whole loan portfolios (given the credit enhancement in the form of subordination and excess spread) there is still liquidity risk as well as some duration risk. Furthermore, these senior bonds may have been purchased above par, bringing prepayment risk into the picture as well. For merchant cash advances or fix and flip mortgage loans, on the other hand, factors that affect valuation are almost entirely different. In the case of MCAs for example, whether a borrower is a repeat client or not and the payment history (i.e. number of payments missed) as well as existence of personal guarantees are the driving force behind valuations. While being a repeat borrower is also a strong indicator of performance for fix and flip mortgage loans, given the collateralized nature of the loan, the value of the underlying property and progress towards completion of the envisioned rehabilitation are major drivers.

7. What about the secondary market? How are deals priced relative to what valuation methods tell us they should be priced? How do valuation analysts obtain information about private sales of loans? In the securitization market is there a valuation standard? How are these deals priced relative to the valuation of the underlying loans?

While many new platforms have started originating in the past few years, several lenders have been originating loans since early 2010s, albeit initially at lower volumes. Data on these loans has been normalized and made available for analysis by firms such as PeerIQ and dv01. More established lenders have returned to securitization markets as issuers with sizeable deals. Securitization is one of the many funding sources available to issuers, among venture capital funding, warehouse lines, one-off loan sales, and forward flow purchase agreements. While the execution levels for each of these funding strategies would yield valuable pricing color, most transactions are private. Relatively speaking, securitizations offer some of the most publically available market color both on collateral performance projections and pricing yields. Implied whole loan pricing can be determined by searching for a yield that solves for the total proceeds from the securitization (including residual, net of legal and other securitization related expenses) given the modeled cashflows.

Additionally, structural features of the securitization as well as the characteristics of the underlying loans and the trend of these over time can offer a glimpse into the market’s perception. Are credit enhancement levels for senior bonds getting larger or smaller? Are the most-junior residuals getting thicker or thinner? Has the average credit score and the dispersion changed relative to prior securitizations? Are pricing spreads wider or tighter than previous securitizations or securitizations from platforms originating similar loans? Apples to apples comparison of securitizations among different issuers with the intent of inferring meaningful valuation assumptions requires an in-depth analysis to properly account for differences in collateral and structure but could yield very important insights.

8. Who is providing valuation services in the marketplace lending space? What should one look for when selecting a service provider?

As most marketplace lending assets are considered level 3 (illiquid and valued using unobservable inputs), originators, investors and securitization shelves alike often turn to third party firms to receive valuations for various purposes such as cutting NAV and financial reporting, fair value analyses for risk retention purposes, transferring assets between related parties, to provide management or the board with an independent view, and other reasons. As discussed, marketplace lending assets are illiquid and their values may be cuspy (sensitive to small changes in valuation assumptions), and thus a cookie cutter approach to loan valuation does not work. Even standard definitions of terms, such as when a loan is considered delinquent, do not necessarily exist in this space. For example, the Office of Thrift Supervision (OTS) the Mortgage Bankers’ Association (MBA) delineate definitions of delinquency for mortgages, and these standards are widely used, yet no such definition currently exists for daily pay merchant cash advance loans. Thus the valuation firm should be well versed in not only DCF modeling but the subtleties of the assets. Such deep understanding of an asset class and its market is greatly enhanced by seeing deal flow such as whole loan sales, securitization execution levels (including residuals and issuance costs), and debt raises for a platform in capital markets transaction. In other words, deriving valuation assumptions cannot be done in a vacuum, and thus access to market color is crucial. Additionally, since each portfolio and transaction is unique, careful analysis of the historical performance and current characteristics of the specific portfolio being priced is paramount.

Since these assets are illiquid, may be cuspy, and are valued using unobservable inputs projecting future loan performance, it is best practice to consider a range of values as opposed to a point estimate. In this presentation, any value is in the range is considered reasonable and the width of the valuation range is determined by different sets of collateral performance assumptions and discount rates, representing a high economic scenario and a low economic scenario. Assumptions sensitized in a high low scenario analysis often include prepayment rates, default rates, recovery rates, and discount rates, based on the risk factors most relevant to the asset. Note that the low and high scenarios should not necessarily represent a crisis scenario and the most optimistic scenario , respectively. Rather, the range is an expression of likely loan performance and changes in market sentiment as can be inferred from the loan characteristics and current market conditions. Extreme scenarios (i.e. credit crisis, prolonged recession, etc.) may be covered as part of a scenario analysis, rather than as part of the concluded range of value. Valuation firms should be able to provide sufficient transparency to support the derivation of the cashflow and discount rate assumptions in the low and high scenarios.

Most importantly, it is crucial not to get bogged down in the minutiae, but rather to maintain an inquisitive eye and consider the larger credit and interest rate landscape when analyzing value in such assets.

Subscribe
Notify of
0 Comments
Inline Feedbacks
View all comments