Three hands-on checks I use to replay past shocks on today's deal: covenant slack, customer churn, and liquidity gates.

Replay the covenants, not the story

I don't start a big decision by arguing about whether we're headed for a repeat of 2008 or a repeat of 2020. I've watched smart people burn days on that debate, then miss the one line in the credit agreement that turns a normal rough quarter into a forced negotiation. So I do the boring thing first: I ask for the actual covenant definitions and I run a replay.

It sounds small, but definitions are where lenders hide the trap door. Is EBITDA adjusted with add-backs capped, or can management stuff it with "run-rate synergies" for 24 months? Do they net cash in the leverage ratio, and if so, do they exclude restricted cash? I've seen restricted cash sitting pretty on a balance sheet while the revolver availability was pinched, and the covenant math didn't care about the headline liquidity.

Then I take two historical shock templates and I pressure the math, not the narrative. For a 2008-style hit, I assume a demand drop that lasts longer than the team is emotionally prepared to admit. I watch what happens to leverage and fixed charge coverage when revenue slides, gross margin compresses a couple points, and the company tries to cut SG&A but can't cut it fast enough. For a 2020-style hit, I model a sudden stop followed by a lumpy restart. Working capital swings matter more there: inventory builds, customers stretch payables, and your cash conversion cycle becomes the whole plot.

Here's the part people skip: I don't just ask, "Do we trip?" I ask, "How early do we feel it?" A covenant with 35% headroom on the first test date can still be a problem if the next step-down is tight and the add-back sunset lands right when the refinancing market is closed. When the replay shows the first pinch point, I want to see the lender call rights, the cure mechanics (equity cure limits, timing, and whether it increases EBITDA going forward), and any springing covenants tied to revolver usage.

By the time I'm done, I don't need a motivational speech about resilience. I need one clear answer: if the world rhymes with an old bad year, are we negotiating from strength, or are we negotiating because a spreadsheet says we have to?

The customer list test: who vanished last time?

Every deck has a slide that says the revenue is diversified. Fine. I still want to know what happened to the business when the world got weird, because that's usually when concentration shows up with its hands on your throat. Historical decision impact analysis, for me, means getting out of the summary charts and into the customer-level mess.

I ask for a customer list that goes back far enough to catch a real stress window. If they can only give me 24 months because the system changed, that itself is information, and not the kind I like. When I can get it, I want monthly billings by customer, not just ARR snapshots. If it's a product company, I want orders and shipments. If it's services, I want invoice dates and collections. Then I do a simple pass that always creates an awkward pause on the Zoom: I highlight the names that used to be top 20 and aren't anymore.

Those aren't automatically losses. Sometimes a customer got acquired and rolled into another name. Sometimes the company chose to walk away from low-margin work. But I insist on an explanation for each. I want to hear, in plain language, how the relationship ended and what the internal decision was at the time. If the answer is, "They froze spend and never came back," my next question is the part that matters for today: what in the product, contract, or switching cost kept them from returning?

Then I flip it. I look at who stayed and paid. In 2020, I watched certain verticals keep the lights on for businesses that looked diversified on paper. In 2008, I watched sticky customers quietly stretch terms until the supplier became their bank. So I map customers into buckets that line up with old stress behavior: discretionary vs. nondiscretionary, project vs. subscription, short PO cycles vs. annual commits, and any client that has a procurement department that treats net-30 as a suggestion. If the company can't tell me, quickly, which bucket the top 50 sit in, we are not ready to make a high-stakes call.

One more check that sounds petty but saves time later: I scan contracts for termination-for-convenience language and renewal timing clusters. If 40% of renewals land in the same quarter, a mild shock can look like a cliff. That's not theoretical. I've lived the board meeting where everybody pretends a renewal wave is seasonality right up until the quarter closes.

Liquidity under pressure: gates, haircuts, and time-to-cash

If you want one place where history rhymes loudly, it's liquidity. Not the feel-good kind where someone says, "We have cash on the balance sheet." The operational kind. The kind that disappears behind gates, haircuts, and settlement windows the moment volatility shows up.

I learned to be skeptical in 2020 watching perfectly reputable funds limit redemptions, widen bid-ask spreads, and lean on lines of credit that were assumed to be permanent. And in 2008, I watched assets that were supposed to be saleable turn into a phone call to three dealers and an ugly haircut. So before I approve a deal, I lay the liquidity sources out as if I had to make payroll during a stress week, because sometimes you do.

Start with a blunt inventory of cash and near-cash. Is the cash trapped in a foreign sub? Is it pledged? Is it sitting in a sweep account that someone will fix later? Then I ask how quickly receivables become cash when customers are nervous. I don't accept an average DSO number and move on. I want the aging and I want to know the top ten invoices that are always late, because those are the ones that will go from 60 days late to 120 days late the moment a customer CFO starts triaging.

If the plan assumes selling assets, I ask who the buyer is in a stressed market. Not the name, but the buyer type. Strategic buyers vanish when their own stock gets hit. Financial buyers pull back when financing is tight. If the asset is a portfolio holding, I look at any lockups, notice periods, and redemption gates. If the asset is real estate, I ask how long it took to close the last sale, including the parts nobody brags about: lender delays, appraisal re-trades, and insurance questions that land at the worst time.

And yes, I look at margin and collateral rules. Prime brokers and secured lenders don't care that you had a temporary drawdown. They care about the mark today and the haircut schedule in the credit support annex. If the strategy includes any leverage, I want to see how margin calls would stack up during a fast drawdown, plus what collateral is eligible, plus what happens if an asset gets downgraded or becomes ineligible overnight.

When I'm satisfied, it's because the timeline makes sense. Not because the spreadsheet ends in a reassuring green cell.