The High-Risk Underwriting Checklist: The Document Pack That Gets a MID Approved

The application went in eleven days ago. Since then you have answered four rounds of questions from someone you have never spoken to, sent the same bank statements twice, and explained a refund policy already published on your site. Nobody has said no, and nobody has said yes either. Almost none of that delay is risk appetite, and almost all of it is a document pack that arrived incomplete and got underwritten one email at a time. This post covers what underwriters need, why files stall, how to read your own statements before they do, and what to expect once the account is live.

What underwriting is actually deciding

An underwriter is not deciding whether your business is good. They are deciding who eats the loss if you stop shipping tomorrow.

That frame explains every question you will be asked. When a transaction is disputed and the merchant cannot fund the refund, the acquirer covers it, so the underwriter is pricing the size and likelihood of unfunded liability left on their books after you are gone.

Everything in the pack maps to that estimate. Bank statements say whether you can absorb a bad month, and processing history says how often your customers dispute. Fulfillment timelines say how much money you hold before you have delivered anything. Your website and marketing say how likely a card brand or regulator is to make this a problem unrelated to your solvency.

Merchants lose weeks treating the application as a credibility contest. Underwriters do not read narrative. They tick boxes, and a missing box stops the file.

The core pack

Assemble all of this before you apply anywhere, since the same set works across most domestic high-risk ISOs with minor variations.

Entity and ownership

• Certificate of incorporation matching the application's legal name exactly
• EIN letter or equivalent tax registration
• Photo ID for every owner at 25 percent or more, plus the control person
• Ownership breakdown adding to 100 percent, including holding entities in the chain
• Any licenses your vertical requires

Banking

• Voided check or bank letter for the settlement account, in the legal entity's name
• Three to six months of bank statements as complete PDFs, not screenshots of a balance

Processing

• Three to six months of processing statements showing volume, average ticket, chargeback count, and refund count
• If you have never processed cards, a volume projection with the assumptions behind it
• If you were terminated anywhere, the notice and a plain explanation

Operations

• Product list with pricing, plus specifications or ingredient panels for regulated products, and supplier agreements where you do not manufacture
• Fulfillment model: who ships, from where, how long from charge to delivery
• Refund and cancellation policy as it appears live on the site

Website

• The live URL, working, with a real checkout and no password wall
• Terms, privacy, refund, and shipping policies plus a reachable contact page, all in the footer
• Age gates, geo-blocking, or disclaimers where your vertical requires them

Complete files get read in one sitting. Incomplete ones get read one email at a time.

Why files stall

Almost every delay traces back to one of four things.

Mismatch. The legal name on the incorporation certificate, the bank account, the application, and the site footer must be the same string. Three variants across four documents reads as sloppiness at best and an undisclosed entity at worst, and it always triggers a question round.

Undisclosed history. If you have been terminated or listed, say so in the application. The acquirer will find a MATCH listing during onboarding regardless, and a listing you disclosed is a conversation while one they discover is a decline.

A website that contradicts the application. The most common one and the easiest to fix. The application says one-time sales and the checkout pre-checks a subscription. The application says research compounds and a landing page reads like a health claim. Any gap between what you wrote and what they see costs you a round.

Vague projections. A first-time merchant writing 500,000 dollars a month with no basis gets underwritten as if the number is fictional. Give a number you can support from ad spend or existing revenue, and expect a cap near it.

The three things that decide your terms

Once the file is complete, three variables set your reserve, your cap, and your rate.

Variable	What the underwriter looks at	What moves it in your favor
Dispute exposure	Chargeback ratio and count over 3-6 months, refund rate, dispute reasons	Ratio under the program thresholds, a documented prevention stack, recognizable descriptors
Delivery risk	Days from charge to delivery, presale, subscription versus one-time	Short fulfillment, in-stock inventory, no long preorder windows
Financial depth	Bank balances, processing history length, personal guarantee, prior terminations	Real reserves in the bank, a clean statement history, disclosed past problems

Dispute exposure is the one you can move fastest. Getting your ratio under the card brand monitoring thresholds before you apply is worth more than any other single change, and how enrollment works is covered in the VAMP and monitoring programs guide. Fixing your billing descriptor is the cheapest lever there, since unrecognized-charge disputes count against you exactly like fraud disputes.

Delivery risk is structural: charge on order and ship in three weeks and you hold customer money for three weeks, which the acquirer prices with a reserve. That is arithmetic rather than hostility, and the reserve math is in the rolling reserves post.

Read your own statements before they do

Pull the last six months as full PDFs and read them the way an underwriter will, because every number that looks strange to you looks stranger to a stranger.

Find the months where volume jumped or collapsed and write one sentence on why, because a spike from one promotion is a fine answer and no answer reads like a business that does not know its own numbers. Count refunds separately from chargebacks, since a healthy refund rate against a low dispute rate is evidence your support desk works, while near-zero refunds alongside real disputes says customers could not reach you and went to their bank instead. Read the dispute reason mix rather than the total, because unrecognized-charge and cancellation clusters point at fixable mechanics while quality and fraud clusters point at the product or the traffic.

Send those explanations with the file. An underwriter who gets your bad month explained up front is reading a file, and one who finds it alone is opening an investigation.

Fixes to make before you send the file

Do these in the week before you apply, not after the first question round.

Match every name string across entity documents, bank account, application, and site footer. Put the refund policy, shipping timelines, and a working contact method in every page footer. If you rebill, disclose it next to the checkout button: the amount, the interval, and how to cancel. Remove any claim you cannot support, which in restricted verticals means keeping every page on the specification side of the line, as covered in the guide to compliance-safe product page copy. Have your PCI scope and current SAQ answer ready, since what underwriters check on PCI is usually a box rather than an audit, and an unanswered box stops the file.

Then write a one-page cover summary: entity, product, monthly volume, average ticket, current ratio, fulfillment time, and one paragraph on any prior termination. It is required nowhere and consistently shortens the review.

The timeline, and what happens after approval

A complete pack from a merchant with clean processing history is usually a decision in two to five business days at a domestic high-risk ISO. An incomplete pack with rolling question rounds is two to four weeks, because each round waits on someone's queue on both sides.

Apply to two or three providers at once rather than sequentially, and expect terms to differ for the same file, since appetite varies by acquirer and by month. Underwriting is not exclusive, and the second and third approvals are the start of real redundancy, which only pays off if something can route between the accounts, as the payment orchestration guide covers.

Then expect a volume cap, a reserve, and a review near the 90-day mark that asks whether you processed the business you described: same product, same ticket, same fulfillment model, inside the cap. Volume far under the cap reads as projections you could not support, and volume well over it gets held rather than congratulated. Go in with three months of your own ratio, refund rate, and fulfillment times, and ask for one thing rather than three: a lower reserve percentage, a shorter hold window, or a higher cap. A reduction granted on evidence is the normal path, and it resets if you change product mix, ticket, or fulfillment model without telling anyone, since each of those rewrites the application after approval.

Practical takeaway

The path that fails is applying early with a thin file and letting underwriting discover things. Every undisclosed termination, mismatched name, and unmentioned claim turns into a question round, and enough rounds turn into a decline that follows you to the next application.

The path that lasts is boring. Assemble the full pack first, make the website and the application say the same thing, explain your own bad months before you are asked, then send it to three providers on the same day. Underwriters approve files they can finish reading, and the merchants approved fastest are rarely the lowest-risk ones, they are the ones who answered the questions before they were asked. If you want your pack reviewed against what underwriters in your vertical actually ask for, apply for an architecture review.