What Disciplined Underwriting Actually Looks Like in Private Credit

The file arrived on a Thursday afternoon, the way they usually do: a deck, a model, a summary from the sponsor and a note asking whether I could have a view by early the following week. What follows is a composite, and every detail of it is invented. The company was a distributor of building products across the Southeast, with a founder close to retirement and a sponsor who wanted to buy the business using a mix of its own equity and a loan from someone like me. None of the people are real. The process is, and I have walked through it in this order more times than I can count.

The first read

I opened the financial statements before the deck, and the model before the summary. The deck is written to be read first, which is exactly why I read it last. The statements showed a business that had grown steadily, earned a reasonable margin on what it sold, and collected its receivables slowly. The model showed the same business growing faster, earning a wider margin and collecting faster, all beginning in the first year of new ownership. The gap between the two was the deal. Everything the sponsor needed to be true lived in that gap.

The lesson I keep relearning is that the first read is not for forming a view. It is for finding the thing the materials were arranged to keep you from noticing. There is always one. Here it was working capital: the model assumed the company could grow without tying up more cash in inventory and receivables, and the history said it never had.

The management conversation

I spent an afternoon with the founder and with the operator the sponsor intended to install after him. I asked few questions about strategy and many about mechanics. Who were the largest customers, and why did they buy here rather than from the national chain down the road? What happened the last time a major supplier raised prices? How many days did it take to collect from a builder, and which builders were slow? What broke in the last downturn, and what did the founder do about it?

The founder answered the early questions quickly and the later ones carefully, and the change of pace told me where the risk was. Builders paid slowly in bad years, and the company had carried them, because carrying them was how it kept them. That was a real competitive advantage and a real drain on cash, and the model had captured the first without the second.

Management conversations are not for hearing the strategy. The deck has already told you the strategy. They are for learning what the company actually does when something goes wrong, because that is the moment your loan will be tested.

Building the downside case

I set the sponsor’s model aside and built my own from the history. I let revenue fall the way it had fallen in the worst stretch the founder described, let the margin compress as suppliers pushed through cost, and let the receivables stretch the way he told me they stretched. Then I ran the cash. In that case the business still made money, but not enough to cover the interest on the loan the sponsor had asked for and also fund the working capital it would need when volumes recovered.

That was the case I feared, and it was not exotic. It was simply the company’s own history, applied to the sponsor’s balance sheet.

The downside case is the underwriting. The base case tells you whether the sponsor is optimistic. The downside case tells you whether you get repaid. I spend most of my time on the second, and I build it from what the company has already survived rather than from what I can imagine.

Structuring for the case you fear

With the feared case in hand, the structure wrote itself. I offered less debt than the sponsor wanted, enough to make the purchase work but not so much that the company would have no room to breathe in the bad year. I asked for a covenant that tested cash flow against interest each quarter, set at a level that would trip early in the downside case rather than late. I asked for a portion of excess cash each year to come back to the loan, so that the good years would reduce the balance before the bad one arrived. I put limits on acquisitions and on distributions to the sponsor while the loan was outstanding. And I asked for monthly reporting on receivables, because that was where trouble would show first.

Every term in a loan agreement should answer to a specific fear.

Terms that answer to nothing are decoration, and decoration gets negotiated away. Terms that answer to a fear you can name will survive the negotiation, because you can explain why they are there.

The decision memo

I wrote the memo in plain language and put the downside case on the first page. I described the business in a paragraph a stranger could follow, named the two or three things that would have to go wrong for the loan to be impaired, and explained how the structure responded to each. I said what I did not know. I said what I would want to see in the first year that would tell me whether the thesis was holding.

If the loan cannot be explained on a page, the lender does not understand it. The memo is not a record of the decision. It is the test of whether a decision was actually made.

After the close

The loan closed. The first year went close to plan, and the cash sweep brought the balance down. The second year brought the slowdown, much as the founder had described it, and the covenant tripped in the quarter I had expected it to. That was the point. It brought the sponsor to the table early, with a smaller balance outstanding and a company that was still profitable. The conversation was uncomfortable and short, and the loan came through it.

Underwriting does not end at the close. It ends when the money is back. The structure you build on a Thursday afternoon is the only thing in the room when the bad year arrives, and disciplined underwriting is nothing more than the habit of building it as if that year were certain.