Executive Summary: Small claims filings are the cheapest early-warning signal in merchant cash advance underwriting and the one most funders never pull, because any single case looks too small to matter. A $3,200 claim from a produce supplier does not move a credit score, does not appear in most commercial bureau files, and does not survive a manual review queue that is already stretched. The pattern across four such filings in eleven months is a different document entirely, and it usually shows up months before anything reaches a credit report.
Why Do Underwriters Ignore Small Claims Filings?
The neglect is rational at the level of a single case and expensive at the level of a portfolio. Underwriting teams triage by dollar magnitude because that is how every other risk input is scaled, and small claims filings fail that test by design. The jurisdictional ceilings are deliberately low: $10,000 in the New York City Civil Court small claims and commercial claims parts, and $5,000 in city, town, and village courts outside New York City.[1] Florida caps small claims at $8,000 exclusive of costs, interest, and attorney fees.[2] Nothing in that range clears the threshold at which a risk team normally slows a file down.
What Makes a $4,000 Judgment Look Irrelevant?
A $4,000 judgment against a merchant requesting $150,000 in funding is 2.7% of the advance. Read as a balance sheet item it is noise. Read as a behavioral event it is a vendor who exhausted every informal collection option, paid a filing fee, took a day off work, and stood in front of a judge to recover an amount smaller than most funders' underwriting cost per file. Vendors do not do that casually. They do it after the merchant has stopped answering the phone.
The mistake is treating the filing as a claim on assets rather than as a dated record of a payment failure. The amount tells you almost nothing. The filing date tells you when the merchant ran out of room with a counterparty who had no leverage other than the courthouse.
Where Does Credit Bureau Reporting Fall Short?
Commercial bureau files are built from trade lines contributed by companies large enough to have a reporting relationship. The suppliers, landlords, equipment lessors, and staffing agencies that file small claims are, almost by definition, not those companies. A regional food distributor suing a restaurant for an unpaid invoice furnishes trade data to nobody, and the event exists only in a county court index.
That is why small claims activity leads rather than lags. By the time nonpayment appears in bureau data, the merchant has typically defaulted on counterparties large enough to report, which is a later stage of the same deterioration.
Why Does the Dollar Threshold Mislead?
The ceilings are statutory, not economic. A supplier owed $23,000 will often split the claim or sue for the capped amount rather than absorb the cost and delay of a general civil action; New York's commercial claims part exists precisely to give businesses a fast route for claims up to $10,000 without an attorney.[3] The observed amount is therefore a floor set by procedure, not a measure of the underlying dispute. Underwriters who filter court results by amount are filtering on the one field that carries the least information.
What Does a Small Claims Pattern Actually Signal?
A pattern has four readable dimensions, and only one of them is the dollar figure most teams anchor on. The other three are frequency, recency, and plaintiff type, and each answers a different underwriting question.
How Does Frequency Change the Read?
One filing is an event. Two filings from unrelated plaintiffs inside twelve months is a payment posture. Three or more is a merchant who is systematically prioritizing which obligations to service, and an MCA remittance is unlikely to sit at the top of that list.
Frequency also separates disputes from distress. A single contested claim over quality, scope, or a returned shipment is ordinary commercial friction. A merchant does not accumulate four unrelated disputes with four unrelated counterparties in a year because of quality problems. They accumulate them because cash stopped covering obligations.
What Does Recency Tell You That Age Does Not?
A judgment from 2021 that was satisfied and never repeated describes a business that survived something. Two filings inside the last ninety days describe a business currently inside something. The underwriting weight should decay sharply with age and spike sharply with clustering.
The practical rule most risk teams converge on is a tiered window:
• Zero to 90 days. Treat any new filing as an active liquidity event and route the file to manual review regardless of amount.[4]
• 90 days to 12 months. Weight by count. One filing is a note in the file; two or more is a decline factor or a materially reduced advance.
• 12 to 24 months. Useful for establishing whether current activity is a repeat pattern or a first occurrence.
• Beyond 24 months. Context only, unless the same plaintiff appears again in a recent window.
• Any window, same plaintiff twice. Escalate regardless of age, because it indicates an unresolved and recurring obligation rather than a one-time dispute.
Why Does Plaintiff Type Matter More Than Amount?
Who sued tells you which part of the merchant's operation broke first, and the ordering is consistent. Suppliers file first because they extend the shortest terms and feel a slowdown within one billing cycle. Landlords follow because rent is monthly and eviction pathways are slower. Staffing agencies and equipment lessors come next. Professional services firms and lenders come last, and by then the merchant is usually visible in bureau data.
A supplier filing therefore sits earliest in the sequence and gives the most lead time. A landlord filing means working capital has been redirected long enough that fixed obligations are now failing. Neither requires a judgment amount to interpret. The same logic underpins how funders read civil court filings as early warning signals for portfolio risk.
How Should You Read Repeat-Plaintiff Patterns?
Repeat plaintiffs are the highest-value pattern in the dataset and the one that requires the most care to interpret correctly, because court procedure shapes what you are able to observe.
What Does the Same Supplier Suing Twice Mean?
A second filing by the same plaintiff means the first obligation was never resolved, or a new obligation to the same counterparty failed after the first one did. Both readings are bad. The second is worse, because it means the supplier continued to extend terms after suing, which usually happens only when the merchant is operationally dependent on that supplier and cannot switch.
That dependency matters for recovery. A merchant who cannot change suppliers has a fixed claim ahead of your remittance, held by a counterparty already proven willing to litigate. Repeat-defendant analysis is the mirror image of the same technique and is worth running alongside it; the mechanics are covered in detecting repeat defendants in lending applications.
How Do Filing Caps Distort What You See?
New York limits a business claimant to five commercial claims statewide per calendar month, and requires certification to that effect at filing.[5] A claimant's principal office must also be in New York State to use the commercial claims part.[3]
Both constraints suppress volume. A supplier with forty delinquent accounts cannot file them all in one month, so what lands in the index is a rationed subset, and a merchant inside that subset was chosen by the plaintiff as worth one of five monthly slots. Out-of-state plaintiffs are pushed into general civil parts entirely, so the small claims index under-represents disputes with non-New York counterparties. Read any hit as a lower bound on the merchant's dispute volume, never as a complete accounting.
What Happens When the Thresholds Move?
Pending New York legislation would raise the small claims ceiling from $10,000 to $15,000 exclusive of interest and costs.[6] If it passes, disputes now filed in general civil parts will migrate into the small claims index, and filing counts will rise for procedural reasons rather than economic ones. A model that trends small claims volume has to hold jurisdictional thresholds constant or it will read a rule change as a credit event.
What Do the New York and Florida Court Structures Actually Allow?
The two jurisdictions behave differently enough that a single rule set applied to both will misread one of them.
What Are the New York Thresholds and Where Do Cases Land?
New York City Civil Court handles small claims and commercial claims up to $10,000; courts outside the city cap at $5,000.[1] Commercial claims filing fees run $25 plus postage, which is the entire economic barrier a plaintiff faces.[7] New York's trial courts recorded roughly 1.89 million new filings in 2024 across all case types, which is the volume any statewide search has to resolve a name against.[8]
The low fee is the analytically important number. At $25, filing is not a considered financial decision by the plaintiff. It is close to free. That removes the usual objection that only large or well-resourced creditors litigate, and it is why the filing behaves as a clean behavioral signal rather than a proxy for plaintiff sophistication.
What Are the Florida Thresholds?
Florida small claims rules govern county court actions demanding $8,000 or less, exclusive of costs, interest, and attorney fees.[2] County courts carry civil jurisdiction up to $50,000, with circuit courts taking anything above that.[17] The $8,000 line was raised from $5,000 in 2020, the first increase in more than fifteen years, which means Florida filing histories spanning that boundary are not directly comparable year over year.
Practically, a Miami-Dade merchant's supplier disputes sit in county civil records maintained by the Clerk of the Court, and the useful range for stress detection runs well past the $8,000 small claims line into the lower county civil band.[18] Restricting a Florida search to small claims alone will miss the $9,000 to $30,000 vendor disputes that carry the same signal.
Why Do Court Structures Change Where the Signal Lives?
New York routes small commercial disputes into a dedicated part with a filing cap and a residency requirement. Florida routes them by dollar amount into a single county court with a procedural switch at $8,000. The same underlying event, a supplier suing a merchant for $9,500, produces a small claims record in New York City and an ordinary county civil record in Miami-Dade. A pattern rule written around case type rather than around plaintiff type and filing date will produce inconsistent results across the two.
How Does Small Claims Activity Fit the Broader MCA Risk Picture?
Court activity is a leading indicator only if the macro environment is one in which merchants are actually under pressure. It currently is.
What Does the Macro Data Say About Merchant Stress?
The Federal Reserve's 2026 Report on Employer Firms, drawn from 6,525 responses to the 2025 Small Business Credit Survey, found that 77% of firms reported rising costs of goods, services, or wages, increased tariff-related costs, or both, as a financial challenge in the prior twelve months.[4] Operating expense coverage was cited by 49% of firms as a reason for seeking financing, ahead of equipment purchases at 42% and real estate at 21%. Small business optimism weakened even as headline performance metrics held roughly steady.[9]
A merchant seeking capital to cover operating expenses rather than to expand is precisely the merchant whose suppliers are most likely to be waiting on payment. The survey population and the small claims defendant population overlap heavily.
Where Does the Distress Show Up Downstream?
Subchapter V elections within Chapter 11 reached 1,663 in the first half of 2026, a 50% increase over the 1,107 recorded in the same period a year earlier, and commercial Chapter 11 filings rose 28% from 3,595 to 4,589.[10] Total bankruptcy filings rose 11% in the twelve-month period ending December 2025.[11]
Bankruptcy is the terminal event. Small claims filings are what that same merchant generated eight to eighteen months earlier, when a distributor and a landlord each decided the account would not pay voluntarily. In construction, mechanic's lien activity has tracked the same softening: the NCS Credit Lien Index fell to 48 in Q4 2025, below neutral for the first time since Q1 2023 after holding in a 55 to 61 range through early 2025.[12][13]
What Does the Enforcement Record Add?
Court records also carry information about how funders themselves behave, which matters for syndication and participation decisions. New York's January 2025 settlement with Yellowstone Capital canceled $534 million in merchant debt and vacated legal actions the firm had taken against merchants, in a matter the Attorney General valued at more than $1 billion.[14] A federal court separately entered a $20.3 million judgment against merchant cash advance operator Jonathan Braun, including nearly $17 million in civil penalties.[15] Both are public record and both live in the same court indexes that hold merchant-side filings.
Corporate counsel meanwhile report that overall disputes volume declined slightly year over year, while the share feeling very prepared for litigation over the next twelve months fell to 29% from 46%.[16] Litigation is not becoming more common in aggregate; readiness to handle it is thinning. An active case against a merchant is more likely to consume that merchant's attention and cash than it would have two years ago.
Ready to add court records to your pre-funding checks? See what a New York and Miami-Dade judgment search returns on your own applications at cobaltintelligence.com/lp/demo/.
What Are Your Options for Pulling Court Data Into Underwriting?
Knowing the pattern is worth reading is a separate problem from getting the records into a decision flow that already runs on a clock. This is where most small claims initiatives die, and it is worth being direct about the alternatives before discussing any single one.
Manual courthouse search remains the most complete option for any given county, at a cost per file no volume operation can absorb. Unicourt aggregates state and county dockets across a wide footprint and is the tool most MCA shops already know. LexisNexis and Wolters Kluwer sell deep litigation and public records coverage priced for law firms and large financial institutions. CSC covers lien and corporate records with strong UCC depth. PACER covers federal courts only, which excludes essentially all small claims activity, since these are state and county matters.
Each of those solves coverage. None solves the specific problem of a funder running hundreds of files a day through an automated flow where a court check has to return a structured result to a decision engine without a human in the middle. That is an integration problem, not a data problem, and it is where the operating math gets decided.
"We run New York court separately and then we run Unicourt, and then we run UCC searches," said Max Weisz, describing an MCA operation processing roughly 500 files a day and expecting 600 to 700 submissions daily.
Three separate tools, three separate interfaces, three separate outputs to reconcile, at 500 files a day. The check does not fail because the data is unavailable. It fails because nobody can run it that many times.
How Does the Cobalt Court Records API Handle This Check?
Cobalt's Court Case API is a data source, not a decisioning engine. It returns court records; what you do with a repeat-plaintiff pattern remains your credit policy.
What Does the Coverage Actually Include?
Coverage is New York State and Miami-Dade County, Florida. That is the entire footprint. It is not nationwide, it does not include federal courts, and it is not a PACER substitute. If your merchant base is concentrated in Texas or California, this product does not serve that need today, and Cobalt has said so directly in demo after demo.
The reason for the limit is demand, not ambition. Jordan Hansen has put it plainly on customer calls: roughly 80% of Cobalt's funder customers file judgments in exactly these two places. New York is the center of alternative lending, and Miami-Dade became the second hub after a large migration of funders to South Florida during COVID. Broader jurisdictions are on the roadmap, including a planned human-assisted queue for unsupported jurisdictions that would run slower, on the order of an hour, and has not shipped.
Funders have pushed back, and the pushback is fair. Lara Hodgson of RoxWrite said it directly: "Most of our clients are not in New York." Cameron Kelliher at Elementix was blunter: "We'll stay away from the court stuff then. I wish the court stuff was rounded." If that describes your book, the honest answer is that the court product is not for you yet, and the SOS, TIN, and OFAC layers are where the value sits.
For funders concentrated in New York, the overlap is unusually good. New York is the only state where Cobalt's SOS data, UCC filings, contractor licensing, and court records all land together, which makes it the one jurisdiction where a single vendor can carry an entire pre-funding stack.
How Does the Async Call Work?
The endpoint is asynchronous by design. A court search takes 30 to 120 seconds because the data is pulled live from the court site rather than served from a cache, so the API accepts a callback URL and posts results there when the search completes. There is no synchronous mode and there is no subscription or monitoring product; you call it per application.
curl -G "https://apigateway.cobaltintelligence.com/courtCases" \
-H "x-api-key: $COBALT_API_KEY" \
--data-urlencode "businessName=Riverside Provisions LLC" \
--data-urlencode "jurisdiction=newYork" \
--data-urlencode "callbackUrl=https://underwriting.example.com/hooks/court-results"Valid `jurisdiction` values are `newYork`, `miamiDade`, `testNewYork`, and `testMiamiDade`. The two test modes run the full request and response cycle without consuming credits, which is how you should build and regression-test the callback handler before pointing it at production traffic. Results include judgment details, case number, case type and division, filing dates, and parties, with amounts where the underlying record includes them. Not every court record carries an amount, and a pattern rule built on plaintiff and date rather than dollar value is more reliable for that reason as well.
What Does One Credit per Lookup Change Operationally?
Court lookups bill at one credit, the same as an SOS lookup. That single fact is the entire argument for moving the check from exception handling to default behavior.
Yehudah Aron at Cucumber Capital reached the conclusion without prompting: "If courts are cheap enough, then it's worth it to run on every [application] automatically." Gate Rock Capital framed the same threshold from the other direction, describing where a court product earns its keep: "where I would pay $4 a pull is when you have the state index on court search." Anthony at CorFinGroup, running 50 to 60 submissions a day with deep dives on 10 to 15, described the underlying need as "one cost effective solution for KYC and KYB."
The operational shift is from selective to universal. A check that costs enough to require justification gets run on files that already look suspicious, which is exactly the population where it adds the least new information. A check cheap enough to run on everything catches the merchant whose bank statements look fine and whose supplier sued them in March. That merchant is invisible to every other input in the file.
How Do You Build a Small Claims Pattern Rule Set?
The rules should be mechanical enough to run without a human and conservative enough that they route to review rather than auto-decline.
What Thresholds Should Trigger Review?
A workable starting policy, tuned to your own loss data over time:
• Any filing within 90 days. Manual review, regardless of amount or plaintiff.
• Two or more filings within 12 months from unrelated plaintiffs. Reduce advance or decline; this is the core pattern signal.
• Same plaintiff appearing twice in any window. Escalate and require an explanation from the merchant before funding.
• Supplier or landlord as plaintiff. Weight higher than a professional services or lender plaintiff, because it sits earlier in the failure sequence.
• A single filing older than 24 months with no repeats. Note in the file and proceed.
• No court hit in a covered jurisdiction. Absence of evidence only; the merchant may simply operate outside New York and Miami-Dade.
That last line is the one teams most often get wrong. A clean court result from a two-jurisdiction product means the merchant has no filings in those two jurisdictions. It does not mean the merchant has no filings. Encoding coverage limits into the score rather than into a footnote is what keeps the model honest.
How Does This Layer With the Rest of the Stack?
Court records are the third layer, not the first. The sequence that works is SOS status to confirm the entity exists and is in good standing, UCC filings to find existing secured positions and stacking, court records to find the behavioral pattern, then TIN verification to confirm the tax identity matches. Each layer answers a question the previous one cannot, and running court records first on an entity that turns out to be administratively dissolved wastes the call.
Scoring the litigation layer consistently across those inputs is its own build, and the weighting approach is worth reading before you set thresholds: how to build litigation risk scoring with court data walks through the tradeoffs.
What Should You Measure After Deployment?
Segment defaults on funded files by whether a small claims pattern was present at funding. If merchants with two or more recent filings default at a materially higher rate than baseline, the rule is earning its credit cost and thresholds should tighten. If they do not, retire the rule rather than keep it for appearance. Six months of funded-file outcomes is usually enough to tell.
Ready to see what New York and Miami-Dade court records return on your current pipeline? Request a demo.
.jpg){kind=small}
.png)