Growth and inflation risks peppered the headlines this week as the Nasdaq and other equity indices set record highs. The IMF raised the US growth outlook for the US by 50 bps. FinTech is also off to a strong start. This week alone, Funding Circle successfully raised $100 MM from Accel, mobile payments iZettle raised $63 MM, and CompareEuropeGroup raised $21 MM amidst buoyant capital markets.

Today, we conclude our two-part write-up on our pricing and valuation approach. We will discuss this further at Context Summits’ Alternative Lending Summit on January 30th & 31st. (Readers can access Part I here).


The price of a loan is the present value of projected loss-adjusted cashflow for the remaining balance of the loan, discounted by an appropriate rate reflecting the riskiness of the cashflows.

We outline this process in the following subsections:

  1. We apply loan-level credit models to project cashflows over the remaining balance of the loan.
  2. We estimate spread at origination – the incremental compensation an investor earns above the Treasury curve for bearing prepay and default risk.
  3. We incorporate additional market observables to adjust discount rates.

Credit Modeling & Cashflow Projections

The price of a loan depends on the principal repayment and interest rate collected over the loan life. Therefore, the first step in our valuation process is to project the pattern and timing of expected (not historical or realized) cashflows.

Unlike an underwriting model which may be concerned with estimating probability of default, we apply credit models that forecast cashflow (e.g., prepayments, delinquency, and default behavior) over the life of a loan given a loan’s borrower attributes, loan attributes, payment profile, and macro conditions.

The path of expected losses of a portfolio can take on complex shapes and levels which leads to diversity in the price behavior as loans season. Exhibit 1 shows the typical Conditional Default Rate (CDR) curves for personal loans originated in the first quarter of each year from 2009 to 2015. These CDR curves tend to have a back-loaded shape as seen below.

Exhibit 1 Sample Prediction of Eventual Default Rate

Source: PeerIQ

Note borrowers are least likely to default in the first few months after the loan is originated, and the likelihood of default increases as the loan seasons.

PeerIQ has developed a family of loan performance credit models. We use a dynamic transition rate approach where each model is tailored and calibrated for every MPL originator. Further, we focus on modeling the Eventual Default Rate (EDR), representing the probability that a loan will default between any month on book (MOB) and its maturity. The exhibit below shows the modeled EDR profiles for a randomized sample of a loan pool.

Exhibit 2 Sample Prediction of Eventual Default Rate

Source: PeerIQ

Note the model closely tracks the eventual default rate profile in out-of-sample back-testing.

Readers can explore the behavior of loss-timing curves on valuation further in our prior newsletter.

Estimating Spread at Origination

Marketplace lending loans, unlike other fixed income securities such as agency MBS or Treasuries, have both prepayment and default risk. Investors earn compensation for bearing these risks in the form of a spread above the Treasury curve.

In the marketplace lending space, investors typically buy newly originated loans from platforms at par in an arms-length transaction. At origination, we have the first observable “traded” value of the loan. The primary market provides price information for estimating spreads at origination. Credit spread at origination (“crSATO”) can be interpreted as the incremental return an investor earns for bearing default and prepayment risk for a newly issued loan over the Treasury curve.

We determine discount rates from a term structure of credit spreads by solving for the rate that equates projected cashflows with par value. (Further spread adjustments are applied for aged portfolios.)

Exhibit 3 shows boxplots of sample crSATO for loans originated in December 2015 for one specific originator. The box plots are arranged from low to high credit score cohorts (10 credit score point partitions) at the time of origination. As expected, low credit score borrowers exhibit higher variance in crSATO distribution and higher average credit spread, reflecting a higher risk premium.

Exhibit 3 crSATO by Credit Score Buckets for December 2015 Loans

Source: PeerIQ

Incorporating Additional Market Observables

Thus far, we have derived pricing credit spread that reflects repayment risk of cashflow given “on-the-run” origination information. To price seasoned loans, we look into other sources of price discovery to adjust credit spreads. The credit spread inputs can be derived from or corroborated by observable market data. We can observe MPL price discovery through various segments of capital markets, such as credit spreads of ABS transactions or pool sales related to comparable collateral.

We incorporate the above price discovery concepts in our pricing and valuation methodology. In addition to the primary MPL market, we adjust credit spreads by incorporating information retrievable from related capital markets, such as:

  • Secondary MPL whole loan trading
  • Relevant private loan sale transactions
  • Newly issued MPL ABS
  • MPL ABS secondary trading
  • Relevant ABS & RMBS sub-sectors
  • Performance of public MPL-dedicated funds

Exhibit 4 shows a loan pricing profile with a contrived non-constant default experience that underwent the above pricing and valuation process. In this case, the loan price drops to approximately 96% of unpaid principal balance at the lowest point, and improves to slightly above par as the CDR curve decreases.

Exhibit 4 Cashflow Attribution and Prices over Loan Age (MOB) for an Up-Down CDR Curve

Source: PeerIQ

In summary, PeerIQ’s pricing and valuation approach is anchored by projecting loss-adjusted cashflows, and deriving the appropriate discount rate to translate those cashflows into price.

The PeerIQ pricing and valuation framework is transparent, forward-looking, incorporates risk factors at the loan-level, incorporates price discovery in adjacent capital markets, and is consistent with ASC Topic 820.

We are eager to share with our readers further at the Valuation Methodology session at Context Summits’ Alternative Lending Summit on January 30th & 31st.


  • SFIG Vegas on February 26-March 1 in Las Vegas.
  • LendIt on March 6-7 in New York.

PeerIQ in the News:

Industry Update:

  • 10 Who Had a Good Year (American Banker, 1/5/17) Alternative Lender, SoFi, among top 10 noted for their success in 2016.
  • SFIG’s Key Policymaker Tracker (SFIG, 1/13/17) SFIG’s Policymaker Tracker details the new names for major financial services policy making positions, with the potential to impact structured finance.
  • States to Feds: Back Off On New Fintech Bank Plan (WSJ, 1/11/17) In a letter Mondayto Comptroller Thomas Curry, senators opposed the creation of the charter, arguing it would allow fintech companies to avoid state licensing requirements by obtaining a limited-purpose bank charter.

Lighter Fare: