An independent valuation advisory subscription for SBA 7(a) holdings — guaranteed certificates, retained servicing, and unguaranteed positions. A Monte Carlo framework turns stochastic rate paths, econometric credit models, and option-adjusted spread into recurring, ASC 820-ready marks you can defend.
One methodology, applied to every position a lender or investor holds — each independently valued and documented to the same ASC 820 standard.
The certificated guaranteed portion (GIC) and the COOF interest-only strip — priced with full prepayment and rate optionality, with the IO valued separately over a Treasury benchmark.
The servicing rights asset valued as the net present value of servicing cash flows — fee income less adequate compensation — across the same simulated rate and prepayment paths.
The retained unguaranteed portion and its credit exposure, marked with the same loan-level default, recovery, and discounting framework used for the guaranteed pool.
An independent, subscription-based valuation with the rigor institutional investors, risk teams, and auditors expect — pricing the optionality simple DCF models miss, documented end to end.
Single-speed DCF misses the options embedded in SBA pools. Monte Carlo across hundreds of correlated rate paths captures how prepayment and default shift as rates move — and what that does to value.
Price from a target discount margin to get fair value and dollar price, or solve OAS from an observed market or BWIC price with a precise root-finding algorithm — the same engine, both directions.
Every run records the full assumption set and a 95% convergence interval, exportable to a self-contained Excel workbook. Fair values are market-consistent and ASC 820-ready for external review.
The Dynamic Nelson-Siegel curve paths drive both discounting and the proprietary Q-SBA credit model — which keys off the 10y/2y Treasury spread — capturing the rate-to-behavior interdependence single models ignore.
Not just a number: modified duration, effective convexity, key-rate durations, and a full NPV distribution with 5th/25th/75th/95th percentiles and WAL mean and standard deviation across paths.
A spread sensitivity grid (±25/50/100/150 bps), parallel rate shifts feeding duration and convexity, and a PPC credit multiplier to run faster or slower prepayment — essential for IO strip risk.
A coherent end-to-end pipeline. Each stage feeds the next, carrying raw loan data through simulation and cash flows to a fully documented fair value.
The Dynamic Nelson-Siegel model — calibrated directly to U.S. Treasury data from the Federal Reserve — describes the yield curve by level, slope, and curvature, then evolves those factors as mean-reverting stochastic processes. From one calibration it generates hundreds of correlated curve paths spanning Prime and SOFR through the 2- to 30-year Treasury tenors.
A credible valuation has to account for the full range of future rate scenarios — not a single guess. These paths are the foundation for every calculation downstream.
Three credit models map each rate scenario to loan-level behavior: a Baseline of constant CPR/CDR, a Rate-Sensitive model that shifts speeds with rates, and the proprietary Q-SBA Model — an econometric model calibrated to years of SBA 7(a) performance that keys off the 10y/2y Treasury spread. Output is monthly prepayment (SMM), default (MDR), and recovery vectors for every loan.
Because the credit model is linked to the same curve paths used for discounting, the framework captures the critical interdependence between rates and refinancing — the heart of the optionality.
For each month under each scenario the engine rolls the pool forward loan by loan: remaining balance, scheduled principal, voluntary prepayments, defaults, losses net of lagged recoveries, gross interest, servicing fees, and the COOF interest-only strip. Variable-rate coupons reset on the simulated reference rate, and settlement-date accrued interest is handled at the loan level for a settlement-accurate clean value.
Because most SBA 7(a) loans are variable-rate, interest cash flows are inherently path-dependent — something a fixed-rate model simply cannot capture.
Each path's cash flows are discounted back and averaged into a settlement-adjusted fair value — with the IO strip priced separately over a Treasury benchmark. The engine reports price, implied OAS, weighted-average life, modified duration, effective convexity, and key-rate durations, and retains every path so you can see the full NPV distribution and a 95% convergence interval.
A fair value without its distribution is half an answer. Path-level detail turns a point estimate into a measured, percentile-aware view of risk.
Each valuation in your subscription arrives with the full analytics, the distribution, and the audit trail behind the number.
Term-structure model calibrated to Federal Reserve Treasury data, generating hundreds of correlated curve paths from Prime and SOFR through 30 years.
Baseline, Rate-Sensitive, and the proprietary Q-SBA econometric model — producing monthly SMM, MDR, and recovery vectors per loan, with a PPC multiplier.
Month-by-month, path-by-path waterfall with variable-rate coupon resets, SBA/FTA/lender fees, the COOF strip, and settlement-accurate accrued interest.
Price from a discount margin, or solve option-adjusted spread from an observed market price with a precise root-finder. IO strip priced over Treasury-plus-spread.
Fair value, price, dirty price, WAL, modified duration, effective convexity, key-rate durations, and a complete NPV distribution with percentiles.
Every assumption logged with a 95% convergence interval and full path-level detail, exportable to a self-contained, ASC 820-ready workbook.
Subscribe to ongoing marks, request a one-time valuation, or read the methodology in full.