Quantifying the Impact of Standardized Payment Data
All data is not created equal. The standardized payment datasets that Orchard provides to investors offer more accurate insights into the performance of pools of loans and allows them to make more informed evaluations and projections of future performance. In this article, I present a quantified analysis of the effects of Orchard’s payment standardization on the pricing of a hypothetical Prosper loan pool. Orchard’s standardized payments will serve investors well for pricing the purchase and sale of loan pools, as well as for ongoing reporting of results.
Quantifying Pricing Differences Across Grades
Exercise and Brief Methodology
For this exercise, we wanted to quantify the effects of Orchard’s payment standardization for the pricing of a hypothetical pool of Prosper loans. We went through the following steps in order to determine the impact:
- Begin with pool of loans with an initial balance of $100MM
- Season this pool of loans for one year
- Determine outstanding balance by delinquency status for remaining loans after one year
- Apply factors to outstanding delinquent loans to flush delinquency pipeline
- Project cashflows using CPRs & CDRs calculated from raw and standardized payments
- Discount cashflows to calculate valuation/Calculate IRR for cashflows
The primary impact came from prepayment (CPR) and delinquency calculations, while charge-offs (CDR) had a minimal effect.
Using raw payments to calculate CPRs tends to overstate the level of prepayments. The primary reason this happens is due to payment bunching, or the tendency for two separate monthly payments to be reported in the same month. For example, if a borrower makes a payment on March 1st and another payment on March 31st, those are most likely two separate payments to different payment cycles, but in raw form, this will often be reported as a double payment in March. The double payment in March would be categorized as a prepayment, when in reality there is no prepayment. Orchard’s standardized payments aim to correct this type of miscategorization, and therefore generally show a slight reduction in prepayments. The effect is usually 1-3 CPR. Cumulative 36M prepayment curves for 2013 Prosper loans are shown below to illustrate the impact.
As the above graph demonstrates, standardized payments represent a 2-3% reduction in cumulative reported prepays relative to raw payments after 36 months.
When buying loans at par, the valuation effect of changes in prepayment speeds is minimal. When buying at prices other than par, however, the effect becomes more pronounced. In particular, for loan purchases below par, higher prepayments increase expected returns to the purchaser. Because you are paying less than $1 for every dollar of principal, the faster that principal is returned to you, the higher your return. If buying loans above par, the opposite is true, and higher prepayments decrease expected returns. In our example pricing exercise, we assign a price of $0.97, which means that the faster prepayment speeds associated with the raw payments leads to a higher IRR, or alternatively a higher price assuming a constant yield.
The core difference between standardized and raw payments with respect to delinquencies is that Prosper’s raw payment statuses are based on Prosper’s reported days past due, which represents the maximum amount of days past due a loan reached during the month, while standardized days past due are based on the end-of-month value. Generally, this manifests in a slightly higher percent of current loans and slightly lower percent of 30-days late loans for standardized data than for raw, although this is not always the case.
Once we have calculated outstanding balances for each delinquency status, we apply long-run transition probability factors to the outstanding balances to reflect the probability of prepayment. A loan that is 30 days late is less likely to continue paying than a loan that is current. Similarly, a loan that is 60 days late is less likely to continue paying than a loan that is 30 days late. Using 12-month roll rates we estimated the probabilities that loans would cure and continue to pay, and then applied factors to the outstanding delinquent balances to reflect this probability. For example, we assumed that loans 60+ days past due would only cure 20% of the time, and we therefore scaled down those outstanding balances to 20% of the outstanding balance.
As we might expect, this effect is strongest for the lower-grade (C, D, and E) loans, where a larger percentage of the loan pool is delinquent, and the effect of any change is therefore magnified.
Below we present two tables quantifying the effects of Orchard’s payment standardization on the pricing of the loan pool. We used two different methods to calculate the impact of the change. First, we used a flat 10% annualized yield to determine the pricing on the pool of loans. Second, we assumed at $0.97 price on the outstanding balance and calculated the IRR.
In each grade, the raw payment values result in a higher valuation and a higher IRR. The discrepancy is larger for the C, D, and E grade loans, where the delinquency effect is most pronounced due to the higher overall level of delinquencies in those loans.
As we’ve shown, the process of payment standardization does result in material changes to prepayments, delinquencies, and loan pricing. An understanding of both methods is important to any pricing decision. We believe access to Orchard’s standardized payments gives buyers an advantage in a purchase negotiation, as an understanding of the discrepancies gives the purchaser additional information. If the standardized payments show lower pricing, as above, this can be used as leverage in a discussion of sales terms. If standardized payments show higher pricing, the investor can keep this information to himself and bid based on the lower raw price, capturing the difference for himself.
To read more about Orchard’s approach to standardized data, check out our latest whitepaper, here.