Headroom / Market data

What 200 credit agreements say about covenants

Public-company credit agreements filed with the SEC, 2021 to 2025. By Tyler Morton. Published October 28, 2026.

Add-back caps: common, and usually 10 to 25%

145 of the 200 agreements define EBITDA. Most of the rest use debt-to-capitalization or similar balance-sheet tests instead. Of the 145, 94 (65%) cap at least one category of add-back, most often restructuring and integration costs, projected cost savings and non-recurring charges.

Where the cap is a percentage, the most common level is 15%, and the median is 15%. How the cap is measured matters as much as the level: in 56 of the 70 percentage caps it is measured against EBITDA before the capped add-backs, and in 6 against EBITDA after them, which is looser (a 20% cap measured after add-backs allows up to 25% of pre-add-back EBITDA). 8 don't say. One example:

…the aggregate amounts added to Consolidated EBITDA pursuant to this clause (b) in any such period shall not exceed 15% of Consolidated EBITDA for such period (calculated before giving effect to the adjustment set forth in this clause (b))…

TechnipFMC plc, credit agreement filed with the SEC

Run-rate savings. 54 of the 145 agreements (37%) let the borrower add back cost savings or synergies it expects but hasn't realized yet. 36 of the 48 that state a window give 12 or 18 months to realize them, and 9 allow them with no cap at all.

Leverage limits cluster at 3.50x to 4.00x

106 of the 200 agreements (53%) have a maximum leverage covenant. The most common initial limit is 3.50x (34 of the 107 leverage covenants; one agreement has two) and the median is 3.75x. Some limits start higher to allow for an acquisition and step down later; using each covenant's final level, the median is 3.50x. These are public companies, many of them large, and the levels sit below the 4.25x median reported for leveraged loans.

Interest coverage is even more standardized: 70 agreements have a minimum, 30 of them at exactly 3.00x (median 3.00x).

Acquisition step-ups are the norm, not the exception

76 of the 106 agreements with a leverage covenant (72%) let the limit rise after a qualifying acquisition. This is an acquisition step-up, sometimes called a leverage holiday: the covenant is still tested, at a higher limit. The typical version adds 0.50x (51 of the 66 step-ups measured against the initial limit) for four quarters (65 of 75). In at least 18 of the 76 agreements the borrower has to elect it. Acquisitions have to clear a size threshold first; the median is about $263 million.

…elect to increase the maximum Net Leverage Ratio permitted under this Section 6.10(a) to 4.00 to 1.00 for a period of four consecutive fiscal quarters commencing with the fiscal quarter in which such Specified Acquisition occurs…

Tennant Company, credit agreement filed with the SEC

9 step-ups raise a limit that has already stepped down, so they aren't in the size chart. We found no published count for public-company agreements; covenant research services track this for leveraged loans. For a borrower, the practical point is that a step-up that must be elected has to be elected on time, and it expires on schedule. Headroom builds step-ups into its covenant math.

What's rare in public-company agreements

• Scheduled step-downs: only 18 of the 106 agreements with a leverage covenant tighten the limit over time, mostly acquisition financings. Step-downs are far more common in leveraged loans and private credit.
• Conditional tests: 9 agreements test a covenant only when a condition is met, such as revolver usage or availability crossing a threshold, and 4 apply a covenant only while the borrower is below a credit rating level.
• Equity cures: 5 of 200 (2.5%). Every one allows two cures in any four quarters and five over the life of the loan.

How this compares with published studies

The pattern is what you'd expect: public companies carry less debt than leveraged and private-credit borrowers, so their limits sit lower, with fewer step-downs and almost no equity cures. Sponsor-backed private loans look different, and testing on them is next.

Method and limits

• Sample. 200 credit agreements filed as exhibits on SEC EDGAR, drawn at random, 40 per year from 2021 to 2025, from agreements matched to a loan record that lists at least one financial covenant, so covenant-lite deals are excluded. They are public companies, mostly with revolving credit facilities, and some borrowers appear more than once. Private, sponsor-backed loans are not in this sample and look different.
• Coding. Headroom extracted each agreement's covenants and EBITDA definition, and each term was then re-read against the filing. Every figure traces to a verbatim quote found in the agreement. That shows where each number came from, but it is not a second independent coding: a spot-check of 10 random rows against the filings matched in all 10.
• Headline cap. When an agreement has several percentage caps, we count the one on run-rate savings, or else the highest. Leverage figures use each covenant's initial limit unless noted.
• Counts. Percentages use the agreements where the term can apply: add-back figures use only the 145 agreements that define EBITDA, and step-ups use only agreements with a leverage covenant.

Download the data (CSV, 200 agreements, each with its SEC link)

This report describes contract terms in public filings. It is not legal, credit or investment advice.