Does your business have no debt? That’s not a good thing. You’re leaving free tax shields unclaimed.

A business carrying no interest-bearing debt is not being conservative. It is declining a discount that is always on offer, costs nothing to hold, and does not expire.

The Federal Reserve's Small Business Credit Survey puts the share of employer firms with no outstanding debt at 31% in 2025, against 29% in 2024 and 29% in 2019, and 21% in 2020. That number gets read as a discipline story, and a good part of it probably is one. The part that isn't gets skipped.

Modigliani and Miller spent the late 1950s and early 1960s on what looks like a narrow question: does the mix of debt and equity change what a business is worth?

The 1958 answer was that it does not. Strip out taxes, distress costs and information asymmetry and the mix you choose stops mattering. That result is the one most often quoted backwards, as though it recommended debt. It recommended nothing. It said the question had been posed badly.

Then they ran it again in 1963 with corporate taxes switched on, and the answer changed. Interest you pay is a deduction. Dividends you pay are not. The government hands back a portion of every dollar of interest and keeps the full amount of profit distributed to owners. That asymmetry is not temporary and it does not expire. It lowers what the business costs to run, and firm value rises with the present value of the shield.

The counterweight is not a footnote. Every incremental dollar of debt raises the probability of financial distress, and distress has a price: a covenant you miss, a customer who reads about it and moves, equity you sell at a bad moment. Which is why the theory lands on a number rather than on the word more.

But notice what the counterweight is an argument against. It is an argument against too much leverage, which makes it an argument for the right amount. It is not an argument for zero. Zero is what you have when nobody has run the calculation.

There is a number that summarizes most of this: the weighted average cost of capital, which is what a dollar of financing costs once you blend the two sources together.

Say owner money costs 14%, because owners carry risk and expect a return for it, and a lender will say 8%. Then a dollar raised part one way and part the other is cheaper than a dollar raised entirely from owners. That is the whole idea. The tax shield is the reason, and the reason is permanent.

The corollary is what sorts the 31% into groups that are not the same group. A tax shield only has value if there is taxable income underneath it, and that is before you get into how much of the deduction you can actually use. A no-debt business with a healthy pre-tax profit is leaving real money unclaimed. A no-debt business in a loss year is not, and adding interest would have deepened the hole rather than filled it. Same line of the survey. Opposite situations.

"No outstanding debt" is a narrow label. It covers bank loans and lines of credit, notes payable, and finance leases, where the lender calls a truck a lease and you own the truck when it is done.

It does not cover operating leases, which is what a warehouse or a fleet is if you don't buy it. It does not cover deferred revenue, which is customer money already collected for work not yet delivered. It does not cover trade payables.

Consider a supplier moving from 45 days to 75 days. What changes is thirty days of working capital, financed by the supplier, at a rate nobody prints on a rate sheet. A business can hold that gap for years and still report zero, because the number nobody argues about is the number missing from the label. A no-debt business with $1M in the bank and one with $60k report the same thing, and only one of them is choosing.

Worth keeping in view: the survey is self-reported, it covers firms with employees, and it records what firms say they owe rather than what a lender would have offered. A firm that asked for a line and was declined reports zero exactly as honestly as one that never asked.

So the 31% splits at least three ways. Operators who decided. Operators the credit market left out. Operators with no taxable income for a shield to attach to. The first made a decision, the second had one made for it, and the third had nothing to decide, and all three land on the same line.

I don't think the headline moves much. The 31% has sat in the same neighborhood for two years, and the forces on it are pulling against each other.

What changes is that the aggregate stops being interesting. Once a number has been read four or five ways, the question worth asking is the narrow one the survey doesn't ask yet: how much of a company's interest expense sits against a year that could actually shelter it. I suspect that by 2028 the firms nobody worries about will be the ones that picked an amount, and the ones sitting at zero will turn out to be the ones that were chosen for.