Kayanda Capital · Research Note 01

Payer-Side Scoring: why receivables underwriting asks the wrong question

September 12, 2026 · Kayanda LLC · Chevy Chase, Maryland
Abstract

Receivables lenders underwrite the borrower. The borrower does not repay the receivable — the payer does. Every factoring decision is therefore a bet on a counterparty that is never scored, and the bet is priced with a proxy: the applicant's credit, revenue, and time in business.

Federal receivables are the natural place to correct this, because federal payment timing is governed by published statute and regulation rather than by discretion. The Prompt Payment Act and its implementing rule at 5 CFR Part 1315 fix a due date, define constructive acceptance, specify how an improper invoice resets the clock, and set a published interest rate for lateness. These are observable, documented rules — which means payment timing on federal paper can be modelled directly instead of inferred from the borrower.

This note sets out the statutory clock, the four mechanisms that move it, and the five inputs Kayanda scores. It also states plainly what has not yet been demonstrated: the model has not been validated against a funded outcome set, and no agency-level empirical timing study is published here. Section 7 is an open call for the data that would close that gap.

Contents
  1. The wrong question
  2. The statutory clock
  3. Four mechanisms that move it
  4. Lateness has a published price
  5. Assignment changes the payee
  6. Five inputs
  7. What is not yet shown
  8. Sources

1.The wrong question

Invoice factoring and receivables finance are usually described as asset-based: the asset is the invoice, and the invoice is what secures the advance. In practice, underwriting rarely behaves that way. The applicant submits financial statements, bank data, revenue history, sometimes a personal guarantee, and the file is assessed the way a small-business loan is assessed — against the borrower's capacity to repay.

That is a category error, and it is expensive on both sides. The advance is retired when the invoice is paid. The invoice is paid by the customer. The borrower's balance sheet influences recovery only in the failure case, after the primary repayment source has already not performed. Underwriting the borrower prices the secondary source and leaves the primary one unpriced.

The variable that actually determines the economics is when the payer pays. Advance rate, reserve, fee accrual, and recourse exposure are all functions of elapsed days. Almost nobody prices that variable directly, because for commercial payers it is genuinely hard: a private company's accounts-payable behaviour is unobservable from outside, governed by internal policy, and disclosed to nobody.

Federal receivables are different in a way that is worth stating precisely.

2.The statutory clock

Congress passed the Prompt Payment Act in 1982 to require federal agencies to pay on time and to pay interest when they do not. The operating rules live at 31 U.S.C. §§ 3901–3907 and in OMB's implementing regulation at 5 CFR Part 1315, with the contract clauses at FAR 52.232-25 and, for construction, FAR 52.232-27.

Four features of that framework matter for modelling.

The due date is defined, not negotiated

Payment is due 30 days after the later of receipt of a proper invoice or acceptance of the goods or services — 14 days for construction progress payments. The due date is a property of the contract and the statute, not of the vendor's relationship with the contracting office.

Acceptance happens whether or not anyone accepts

Under 5 CFR 1315.4, acceptance is deemed to occur constructively on the seventh day after delivery, unless the contract specifies a longer acceptance period or actual acceptance occurs sooner. This matters because it puts a ceiling on one of the two components of the due date. An agency cannot extend payment indefinitely by declining to act — silence resolves to acceptance.

A defective invoice must be returned quickly

When an invoice is improper, the agency must return it as soon as practicable and no later than seven days after receipt, identifying every defect and every reason it is being returned. If the agency misses that seven-day window, the days allowed for payment of the corrected invoice are reduced by the overrun, and interest is computed against the adjusted date.

Interest accrues automatically

The interest penalty is paid without any request from the contractor. It is not a claim to be filed; it is a consequence attached to the missed date.

The practical implication: for a federal receivable, the distribution of payment dates is shaped by a published rule set. It is not a black box. It can be modelled from the contract, the invoice, and the documentation — before anyone looks at the vendor's financials.

3.Four mechanisms that move it

The 30-day figure is where most discussion of the Prompt Payment Act stops, and it is where the useful analysis begins. Contractors routinely wait far longer than 30 days, and the statute is not being violated in most of those cases. The clock has simply not started, or has started and been restarted.

There are four distinct mechanisms, and they behave differently:

  1. The invoice was never proper. The payment period begins when the agency receives an invoice containing everything required. A defect returns the invoice and restarts the period on the corrected version. Documentation completeness is therefore not an administrative nicety — it is the variable that determines whether the clock is running at all.
  2. The acceptance period was extended by contract. A contracting officer may specify a longer constructive acceptance period where inspection or testing genuinely requires it. Where that clause is present, the due date moves with it, lawfully and predictably.
  3. The payment is contingent on someone else. On a subcontract, the prime's own receipt governs. Pay-when-paid and pass-through structures inherit the federal timeline and add a second party's processing to it.
  4. The amount is disputed or subject to offset. A contested amount is outside the prompt-payment framework entirely until it is resolved.

Each mechanism is visible in the file. A contract clause can be read; an invoice can be checked for completeness against the proper-invoice elements; a pay-when-paid structure is disclosed in the subcontract; a dispute is known to the applicant. None of this requires the borrower's financial statements, and none of it is captured by them.

4.Lateness has a published price

The Treasury Secretary sets the Prompt Payment interest rate semiannually and publishes it in the Federal Register, effective each January 1 and July 1. It is set by reference in 31 U.S.C. § 3902(a) to the rate established under 41 U.S.C. § 7109.

PeriodPrompt Payment rate
Jul 2026 – Dec 20264.750%
Jan 2026 – Jun 20264.125%
Jul 2025 – Dec 20254.625%
Jan 2025 – Jun 20254.625%
Jul 2024 – Dec 20244.875%
Jan 2024 – Jun 20244.875%
Jul 2023 – Dec 20234.875%
Jan 2023 – Jun 20234.625%
Jul 2022 – Dec 20224.000%
Jan 2022 – Jun 20221.625%
Jul 2021 – Dec 20211.125%
Jan 2021 – Jun 20210.875%
Jul 2020 – Dec 20201.125%
Jan 2020 – Jun 20202.125%

Source: Bureau of the Fiscal Service, Prompt Payment interest rates. The July–December 2026 rate was published at 4-3/4 percent per annum (Federal Register, July 9, 2026).

Two observations follow. First, the rate is frequently confused with the Current Value of Funds Rate, which is set annually for a different purpose — 4.00% for calendar 2026 — and substituting one for the other produces a plausible-looking number that is wrong. Second, and more importantly for underwriting: a late federal receivable does not simply age. It accrues a statutory penalty at a published rate, in favour of the holder. That is a materially different risk object from a late commercial invoice, and it is not reflected in pricing that treats all late paper alike.

5.Assignment changes the payee

The Assignment of Claims Act (31 U.S.C. § 3727 and 41 U.S.C. § 6305, implemented at FAR Subpart 32.8) allows a contractor to assign amounts due under a federal contract to a bank, trust company, or other financing institution. Where a valid assignment is filed and the contract contains the no-setoff commitment, payment is directed to the assignee.

For a funder, this is the single largest structural variable in the file. It changes who receives the money, removes the intermediate step where the contractor receives and redirects funds, and narrows the set of events that can divert payment. Two otherwise identical federal invoices — same agency, same amount, same aging — carry materially different risk depending on whether assignment is in place. Borrower-side underwriting does not see this at all.

6.Five inputs

The Payment Certainty Score™ is deliberately small. It reads five variables, all of which are present in a normal receivable file and none of which require the applicant's financial statements:

InputWhat it captures
Payer classWhether the obligation sits under a statutory payment regime (federal direct), inherits one at a remove (federal subcontract), or sits under no timing rule at all (commercial).
Payment mechanismAssignment of Claims filed, standard net terms, pay-when-paid, or milestone-contingent. Determines who is paid and what can interrupt it.
Documentation completenessWhether the invoice is proper. Determines whether the clock has started.
Days outstanding against termsPosition on the clock, measured against the contractual due date rather than the invoice date.
Payer concentrationShare of the applicant's revenue tied to this one payer — the correlation term.

The output is a 0–1000 score, a tier, an estimated payment window, an indicative advance band, and reason codes attributing every point lost. The reason codes matter more than the score: a file scoring 620 because documentation is incomplete is a file that can be fixed in a day, while a file scoring 620 because of a contingent payment structure cannot be fixed at all. A single number conceals that distinction; the codes surface it.

The score makes no credit decision, and Kayanda takes no credit risk. It is an input a funding partner may use inside its own independent underwriting.

7.What is not yet shown

Stated limitations

This note contains no empirical agency timing study. The statutory and regulatory material above is verifiable from primary sources, and the interest-rate series is published by Treasury. Neither establishes how any particular agency, contracting office, or prime actually behaves. Observed payment behaviour and published payment rules are different things, and we do not conflate them here.

The model is in calibration. The Payment Certainty Score™ has not been validated against a funded outcome set. Its weights encode the mechanisms described above; they have not yet been fitted to observed payment dates at scale. No claim of predictive accuracy is made or implied.

Nothing here is an approval, offer, prediction, or financing commitment. Kayanda is not a bank, lender, credit bureau, rating agency, or consumer reporting agency. The score evaluates commercial receivables and institutional payers only, never consumers, and is not a consumer report under the Fair Credit Reporting Act.

The gap between the rules and the behaviour is the entire research question, and it is not closable from public data alone. Agency-level payment timing is partially observable — through agency financial reports, Treasury's Invoice Processing Platform, and federal spending disclosures — but the decisive detail is not: the date a specific invoice was deemed proper, the date acceptance occurred, whether a defect was returned within seven days, and the date funds actually settled.

That record exists in only one place, distributed across the contractors who lived it and the funders who financed it. Kayanda is assembling it one scored file at a time, and reporting outcomes back is a first-class operation in the API (POST /api/v1/outcome) rather than an afterthought.

Open call

We are seeking three things, and will share aggregate findings with contributors:

8.Sources

Contribute or collaborate

Help close the gap between the rule and the behaviour

If you hold closed federal receivables, finance them, or research procurement payment behaviour, we would like to hear from you. Outcome data is reported in aggregate and never identified.

Contact the authors
Disclosure. This research note is published by Kayanda LLC and describes the design of the Payment Certainty Score™ and the reasoning behind it. It is not legal, tax, accounting, investment, or financial advice, and it is not an offer of securities, credit, or deposit products. Kayanda Capital is a product of Kayanda LLC (organized 2018) and is not a bank, direct lender, broker-dealer, investment adviser, credit bureau, consumer reporting agency, or money transmitter. Kayanda makes no underwriting or funding decisions and retains no credit risk; third-party funding partners independently verify, underwrite, price, and decide every transaction, and funding is never guaranteed. Statutory and regulatory descriptions are summaries provided for discussion and are not a substitute for the primary sources cited above or for advice from qualified counsel; readers should verify current rates and regulatory text before relying on them. The Payment Certainty Score™ is a model in calibration that has not been validated against a funded outcome set, and no representation is made as to its predictive accuracy.