LME protection review, explained.

How a struggling borrower moves value past some of its lenders, and the clauses that stop it

The life of one leveraged loan, in five steps

1. The deal is negotiated and signedCorvane's $420.0 million facility closes in June 2022. The blockers in it are whatever the lenders won that month.
2. Amendments pile upAn incremental loan, a repricing, a covenant reset with a new basket: three amendments by February 2025, each able to widen a route.
3. The book is reviewed for LME protectionEvery agreement read as amended, every blocker graded, the baskets measured. This is where the IP-only blocker and the $96.3 million of room are found, or missed.
4. The borrower asks for consentA waiver request with a deadline: the lenders' best moment to ask for a stronger clause in return.
5. The sponsor moves, or the gap is closedA drop-down or uptier under the clause as it stands, or an amendment that fixes it. Either way the answer was in the documents.

Step 3 is the subject of this guide. The fund, borrower, sponsor and figures in this example are fictional.

Negotiated defaults, and blockers a majority can vote away

Unrestricted subsidiary, blocker, sacred right and basket

Fourteen terms are defined in the one-page summary.

Five ways a stressed borrower moves value away from some of its lenders

In shortA drop-down moves the assets. A guarantee release removes a guarantor. An uptier moves the line, putting new debt ahead of the old. A double-dip or pari-plus deal gives new lenders a second claim the old ones lack. Vote packing changes who holds the majority. Every one uses a permission the agreement already grants.

Why sponsors choose an LME over a bankruptcy filing

The Fifth Circuit summarized the attraction in its Serta opinion: the borrower "needs only majority (versus unanimous) consent", can raise fresh money that might otherwise come only in bankruptcy, and avoids court costs and holdouts.4 The sponsor keeps its equity and the timetable. The cost lands on the lenders who were not part of the majority. In the court's words, the costs of an uptier are "born entirely by the minority lenders, who end up with subordinated debt worth less than before".4

For a lender the timing matters more than anything else. By the time an LME is announced, the question of whether it was allowed is a question about documents signed years earlier. The review this guide describes moves that question to a time when the lender can still change the answer.

Ten named blockers, what each one stops, and how each gets narrowed

In shortTwo guides published in 2025, one by a credit research firm and one by a law firm, describe about ten named blockers. Each answers one move, and each has well-known weak versions: an asset definition that covers only intellectual property, a transfer ban with no ban on designating a subsidiary, a subordination clause that protects liens and leaves payment open, an undefined "open market purchase".2,1

Octus also lists a "Better Health" protection, which requires extensions of maturity to be offered to all lenders pro rata, and names the vote-packing gap Robertshaw showed: a blocker that is a sacred right but still lets the debt be incurred.2

Why a blocker the majority can delete only stops deals with outsiders

In shortMost amendments need the Required Lenders, usually more than half. A few terms, the sacred rights, need each affected lender. A blocker outside that list can be removed by the same majority that wants to do the deal. So every blocker review reads two places: the blocker itself and the amendment section.1

A credit agreement has a voting rule for changes. Most terms change with the consent of the Required Lenders, typically holders of more than half of the loans and commitments. A short list needs more: each lender directly affected must agree to changes in its interest rate, principal or payment dates, to a release of all or substantially all of the collateral, and, in a well-drafted agreement, to changes in pro rata sharing and the payment waterfall. Those are the sacred rights. The Fifth Circuit noted that pro rata sharing is often described as a lender's "sacred right" under syndicated loan agreements, and explained that sacred rights protect individual lenders "from potential machinations by the majority".4

Cleary Gottlieb puts the consequence plainly: if a blocker is not a sacred right, "the borrower could obtain consent from participating lenders that constitute the majority required lenders", and the blocker may only be effective against LMEs with third parties.1 A J.Crew blocker in the negative covenants can be deleted at the same moment as the drop-down it was meant to stop.

One lender's vote in the example

In the example, the fund holds $142.0 million of a $380.0 million term loan. Other lenders holding $238.0 million, 62.6%, are enough to amend anything that is not a sacred right. Corvane's waterfall and pro rata sharing protection passes the fund's checklist because both are sacred rights. The J.Crew blocker is narrow, and the Serta blocker covers liens only. A short blocker that only each lender can remove is worth more to the fund than a long one the majority can delete.

The words that change a vote

What courts decided from Serta and Mitel to the 2026 damages award

In shortNo statute governs blockers. Courts read the contract, almost always under New York law, and the exact words decide. Serta's uptier failed in the Fifth Circuit because the agreement allowed only "open market" purchases; Mitel's survived in New York because its agreement said "purchase" with no qualifier. In July 2026, as law-firm reports describe the ruling, a bankruptcy court ordered Serta's participating lenders to pay the lenders left out of the deal $261.13 million plus 9% interest.4,5,10

Serta and Mitel, side by side

The Fifth Circuit's reasoning is a lesson in reading. The agreement offered two roads around pro rata sharing: a Dutch auction open to all lenders, and open market purchases. If "open market purchase" meant any purchase for value among private parties, the court said, "the § 9.05(g) term could be 'open purchase', not 'open market purchase'", and the Dutch auction exception would do no work.4 Having chosen to deal privately with selected lenders, the borrower "lost the protection of § 9.05(g)".4

What the Serta uptier cost the lenders who ran it

The remand turned on a second clause. As Vinson & Elkins reports the July 7 opinion, Judge Lopez read "payment in respect of" principal to cover the new debt the participating lenders received in exchange for their old loans, "even without a cash payment", so they had to share it pro rata.10 The same report notes Del Monte, decided two months earlier by a New Jersey bankruptcy court: it declined to find that a cashless roll-up of old debt into a bankruptcy loan was, by itself, a payment, and left open whether a later payment on that roll-up would be one.10 KBRA's August 2026 private credit note discusses the award under the headings "the added deterrent" and "Del Monte tempers the result".13

How much value can leave through the investment baskets: the arithmetic

In shortAdd up every basket that can carry value to an unrestricted or non-guarantor subsidiary, each at its larger of a fixed amount or a share of EBITDA, plus the builder basket. Remember that investments are usually counted at book value or cost, so an asset worth more than the total can still fit. J.Crew moved $347 million of trademarks through $250 million of capacity.6

J.Crew, 2016

Buccola and Nini, who studied loan contracts before and after the deal, record that J.Crew "had $250 million of capacity to invest in unrestricted subsidiaries and had received an opinion valuing its intellectual property at $347 million", and that it transferred "72% of an undivided interest in the IP".6 The capacity came from more than one basket, routed through a non-guarantor restricted subsidiary on the way to the unrestricted one. A blocker that bans only direct transfers does not stop that route.

The worked example, Corvane at June 30, 2026

Two details matter as much as the sum. The fixed amounts are floors: if EBITDA falls, the room does not. And the builder basket arrived with an amendment the lenders approved for other reasons, which is why a review reads every amendment in order. Use EBITDA as the agreement defines it, from the latest compliance certificate, and never a projection the agreement does not allow.

How an uptier reorders who gets paid, worked through with numbers

In shortThe majority swaps its loans, at a discount, into new debt that ranks first, often next to fresh money it also provides. Nothing leaves the company, yet the minority's recovery falls because more debt now sits ahead of it. In Serta, the majority exchanged $1.2 billion of loans for about $875 million of second-out debt behind $200 million of new first-out money.4

Serta's 2020 uptier

A $1,000 million term loan before and after an uptier

Take a company with one $1,000 million first-lien term loan. A sale would raise $600 million. Before any deal, every lender would recover 60 cents on the dollar. Now lenders holding 55% swap their $550 million into new super-priority debt at 75 cents ($412.5 million) and put in $100 million of new money that ranks first and is spent on the business. If the company is later sold for the same $600 million, the new money is repaid ($100 million), the swapped debt is repaid ($412.5 million), and $87.5 million is left for the minority's $450 million.

The arithmetic is why the blocker list for uptiers is long. The Serta blocker stops the subordination. The open-market and pro rata clauses stop the selective swap. The NYDJ protection stops the same thing done inside one agreement by rewriting the payment waterfall. The Incora blocker stops a borrower from manufacturing the majority. A loan is open to an uptier if any one of those routes is open.

Guarantee releases, double-dips and pari-plus deals: how they take value without moving it

In shortSome LMEs strip nothing from the collateral and still change who gets paid. A guarantor that stops being wholly owned can fall out of the guarantee. A new lender that holds both an intercompany note and a guarantee has two claims where the existing lenders have one. The Chewy and At Home blockers answer them.7,1

PetSmart and Chewy, 2018

PetSmart's agreement released a subsidiary guarantor automatically once it was no longer wholly owned. Mayer Brown's account: PetSmart "transferred 16.5 percent of its equity interests in Chewy" to an unrestricted subsidiary and "distributed 20 percent of Chewy's equity to its sponsor", 36.5% in all, which led to "a release of Chewy's guaranty and pledged collateral".7 The Chewy blocker that followed allows a release only on a sale to an unaffiliated buyer for a real business purpose.2

How a double-dip gives one lender two claims

Mayer Brown describes the hallmark of a double-dip as "providing new lenders with two means of potential recourse (a.k.a. 'dips') against the borrower and collateral", and notes that double claims do not create double recoveries but can raise the new lenders' share in a downside case.7 The At Home blocker bars non-guarantors from holding debt or liens of the borrower group and stops equal-ranking debt from taking guarantees the existing lenders do not get.1

Where blocker reviews go wrong: the drafting gaps that read as protection

In shortMost misses come from reading a heading and stopping. A clause titled "Unrestricted Subsidiaries; Material Intellectual Property" can protect almost nothing for a services company. A Serta blocker can carve out debt "permitted on the Closing Date". An amendment two years ago may have lifted a cap. Through the first quarter of 2025, more than 40% of the US leveraged loans CreditSights tracks still let a majority subordinate the loans.14

Ten things to check in every agreement

Buccola and Nini found that between 2016 and 2021 only 15% of syndicated loan contracts that permitted unrestricted subsidiaries blocked the transfer of intellectual property, though such blockers grew more common by 2021.6 CreditSights reported that through the first quarter of 2025, "43.4% overall and 44% of PE-backed loans" in its leveraged loan data still had the customary language allowing subordination with majority consent.14

How often LMEs happen and what first-lien lenders get back afterward

In shortMoody's counted most of 2025's corporate defaults as negotiated exchanges, and LMEs are a large share of them. Fitch counted 50 liability management transactions in 2024 and saw first-lien recoveries fall from 76% in 2022 to 39% in 2024. Moody's found about one in four soft defaults later becomes a hard default, most within two years.15,3

Recoveries are where an LME shows up for the lenders left out. Fitch's outlook for 2025 lists first-lien recovery performance as a watch item given how common liability management transactions have become.15 The Federal Reserve's May 2026 Financial Stability Report notes that leveraged loan defaults counting distressed exchanges "remained relatively high" even as the plain default rate stayed below its historical median.16

J.Crew blockers tripled in new private credit deals, while that market's default rate runs from 1.6% to 6% by how you count

In shortPrivate credit and leveraged loans were each about $1.4 trillion at the end of 2025. Default rates depend on what is counted: Fitch's record 6% includes interest deferrals and PIK, Proskauer's 2.41% does not count the same events. Meanwhile J.Crew blockers went from 15% of private credit deals in early 2023 to 45% in the third quarter of 2025, in one data vendor's count.17,18,19,20

Fitch's figure, reported by Bloomberg Law, counted 99 default events in the 12 months to April 30, 2026, 81 of them first-time, with more than half "driven by interest payment deferrals and the introduction of payment-in-kind".18 The Financial Stability Board, in its May 2026 report, cites a study that puts outright private credit loan defaults near 1%, rising to around 5% when selective defaults are included, and notes PIK in about 12% of loans.21

Blockers are spreading, faster in some markets than others

The two markets moved differently. In high-yield deals Noetica counted anti-Serta protections in 85% of deals in the second quarter of 2025, J.Crew blockers in 38%, and omni-blockers in only 4% of 2025's publicly filed high-yield issuances.22 In private credit, lien subordination protection reached 84% of deals in late 2025, up from 42% a year earlier.20 Newer deals are better protected. The older vintages a fund already holds are not, unless an amendment fixed them.

Size makes the gap matter. The Federal Reserve's FEDS Notes estimate private credit and leveraged loans at about $1.4 trillion each at the end of 2025, and, for loans issued in 2025, median debt of 5.0 times EBITDA for private credit borrowers against 3.2 times for leveraged loans.17 Institutional leveraged loans outstanding were $1,380 billion at the end of 2025.16

What supervisors expect of a blocker review, and where AI helps or misleads

In shortNo banking rule says which blockers a loan must have. The 2013 leveraged lending guidance was withdrawn by the OCC and FDIC in December 2025. Model risk guidance was rewritten in April 2026. AI can read hundreds of agreements quickly, and is useful only when every grade quotes its clause and a person decides what to raise.23,24

Where AI earns its place is the reading. A careful reviewer must apply every amendment in order, follow each defined term to its definition and every use, and measure baskets against the right EBITDA. Those are repeatable tasks over long documents. Where it misleads is the confident summary: a grade without the clause behind it, a definition inferred because it was not found, a request for an amendment treated as if it had been signed. A sound tool quotes the clause and page for each finding, says "not found" when it cannot find one, and leaves the asks and the drafting to people.

Private credit funds are not banks, so bank model risk guidance does not bind them directly. Banks that lend to funds and bank-owned direct lenders may apply it anyway, and the third-party guidance keeps the bank responsible for a vendor's work.27

How credit teams check LME protection today: the five roles, the hand-offs and the hours per agreement

In shortCounsel reads the documents when a deal is signed and again when a consent request arrives. In between, the protection of a book lives in deal memos and memory. A full-book question means pulling every agreement and amendment and reading them again. The only published time figure is a vendor's claim of four to six hours per agreement.30

Portfolio manager, in-house and outside counsel, underwriting and workout

How a full-book blocker question reaches a decision

Reading and tracing: the hours in a hand-built blocker review

Hebbia's credit agreement software guide, dated August 26, 2026, quotes another vendor's claim that review times drop "from four to six hours to 30 to 45 minutes per agreement", and reports from its own survey that 62% of teams lose four or more days per deal to manual re-checks and leadership review.30 Both come from software vendors and neither publishes its method. At five hours each, 64 agreements take 320 hours, about 40 working days of reading plus a day to decide, for one question; the next question starts the reading again.

A counsel's blocker grid, outside counsel or credit agreement software

Corvane's consent request worked through: the weak blockers found, the room measured, six asks drafted

In shortThis is the example from the top of the page, inside the product. Six agents read the book as amended and grade every loan against the fund's own checklist. Leah Brandt, the portfolio manager, asks in plain words, opens each finding at its clause, and decides what to raise. Simone Achterberg, deputy general counsel, approves the drafting before anything reaches the borrower.

1

The whole book behind one question box

Monday, October 12, 2026. Leah opens the clause finder over Direct Lending Fund III and two co-invest vehicles: 64 loans, 216 documents, 31,480 pages, read as amended by signed documents only and graded against version 7 of the fund's checklist.

Under the question box sit six saved questions with their counts and when they last ran, and a newly signed amendment waiting to be loaded. The saved answer on screen says 7 of the 64 loans have no J.Crew blocker at all, $455.0 million held.

2

The amendment that lifted a cap

Leah picks Thistlecombe Behavioral Health, $64.8 million held, the widest open of the seven. The agreement opens at Section 7.02(m), page 129. On the amendment walk above it she clicks Amendment No. 2, March 2023: the original $25.0 million cap on investments in unrestricted subsidiaries is struck through, and "unlimited Investments in Unrestricted Subsidiaries" takes its place.

Read as signed in 2021, this loan looked protected. Read as amended, it has no dollar limit at all.

3

One wider question, read as it runs

Then she asks the question the consent request raises: "Which of our loans would let the borrower move assets out of our reach and also prime us without our consent?" The run log shows each agent at work: 216 documents stacked as amended, 1,412 defined terms traced, the drop-down and uptier blockers graded, the baskets measured.

One line matters for the example: Corvane is read as amended by No. 3. Its request for No. 4 is not signed, so it is not applied.

4

Eleven loans, $733.4 million, each finding numbered to its clause

Thirty-eight seconds later: 11 of 64 loans, $733.4 million, 18.2% of the book, would let the borrower do both, with 22 clauses quoted. Corvane comes first: consent due October 23, a J.Crew blocker that covers intellectual property only, a Serta blocker that protects liens and leaves payment open, and $96.3 million that could move today. Four of the 11 belong to Corriebrook Equity Partners, the sponsor that moved Brenmoor Labs' IP last month.

5

Six asks, to counsel first

Leah drafts Corvane's asks from the fund's model clauses: widen the J.Crew blocker to all material assets with a designation prong, extend it to subsidiaries that do not guarantee, take the Serta blocker to payment priority with "effect of" words, define "open market purchase", add a vote-packing blocker, and cut the unrestricted subsidiary basket to $15.0 million. They go to Simone Achterberg, who approves the drafting. Nothing reaches the borrower or the loan's administrator until she does.

The solution's design targets are one question answering the whole book, and 100% of findings shown with the clause and page they rest on. They are targets, to be measured on your own agreements.

The protection map, the loan pages, the gaps to raise and the checklist behind every grade

In shortOne consent request is the story. These screens show the rest: every loan graded on eight protections, one loan clause by clause with its basket room, the clause behind one grade, the gaps to raise in date order, a signed amendment that reruns every saved question, the document stack, and the checklist counsel owns.

Select a screen to open it larger. The fund, borrowers, sponsors, people, clauses and figures are fictional; the clause text is the product's sample drafting and quotes no real agreement.

What each of the six agents does, and where the portfolio manager and counsel take over

In shortEach agent has one job and limits it does not cross. None of them grades without quoting the clause, applies an unsigned document, or contacts a borrower, sponsor or the loan's administrator. The portfolio manager decides what to raise and when; counsel approves every piece of drafting and owns the checklist.

From the amendment stack to the drafted ask: six blocker review agents

The portfolio manager decides which gaps to raise; counsel approves the words

Every question, answer, grade change and ask is on the audit trail, with who or which agent did it and when. Agent steps and human decisions are kept apart, every checklist version is kept, and nothing leaves the fund until counsel approves it.

Leah, Simone and Owen on Direct Lending Fund III's 64 loans

What the review reads and what it hands back: a graded book with every clause quoted

In shortIt reads the documents the fund already holds and grades them against the fund's own checklist. It leaves behind an answer table for every question, the clause and page behind each finding, the room to move for each loan, the asks to raise at each consent, and saved questions that rerun on every signed document.

Credit agreement sources: data room, loan administration, covenant data

For each, check whether it has an API, a file feed, or only screens, and who owns it.

Every blocker grade quotes its clause from signed documents only

What a weak blocker and a slow reread cost you, in recoveries and in hours

In shortThe big cost is a recovery lost to an LME you could have blocked. The steady cost is the hours spent rereading the book every time a borrower wobbles. A few published figures frame both. The number that matters most, how many of your own loans are open today, is yours to measure.

Grading ten older-vintage loans and their basket room yourself

A lost recovery, and a consent waived without a fix

The hours a full-book blocker review gives back, and when they cover the cost

In shortPut in your own numbers. The calculator counts only reading time: how many loans you hold, how long one agreement takes to review, how many full-book questions you ask a year, and what the tool and your own checking cost. Every default is illustrative. The larger value, recoveries protected, is in the levers below and in no formula.

Rereading hours, ready consent asks and basket room closed

What full-book blocker questions cost in counsel and analyst hours

Try your own book. With 64 loans and the defaults, four full-book questions a year release 1,232 hours and clear the cost; a book of 30 loans that asks twice a year does not, on reading time alone, and its case rests on the recoveries the calculator leaves out.

Nine questions that show whether your team is ready to automate its blocker review

In shortPick the answer closest to your team today. Eight questions score 0 to 2; the last is for reflection. The meter adds up your score as you go, and each question says why it matters and what a strong answer sounds like.

Write the blocker checklist first, or pilot on counsel-graded agreements

Seventeen questions that test any credit agreement review software on blockers

In shortAsk every vendor the same questions, us included. The answers show how the review is built: whether it reads amendments in order, quotes the clause behind each grade, grades against your checklist, measures baskets the agreement's way, and can be tested on your own documents.

We found no published independent test of any credit agreement review tool's accuracy on blockers, as of October 2026. A test against counsel's own grades on your agreements is the fairest comparison, and the same test measures your current process.

Start with a ten-week pilot: counsel grades 20 agreements, then the tool runs the whole book

In shortWrite the checklist, have counsel grade a sample by hand, run the tool on the same sample blind, then on the whole book, and decide with three numbers. None of these steps commits you to a vendor, and the first two are worth doing anyway.

1. Write the checklist: the protections you require, the test for strong, the must-haves. Counsel owns it.
2. Grade 20 agreements by hand, as amended, with the clause and page for each grade. That is your answer key and your baseline hours.
3. Run the same 20 blind and compare grade by grade, including loans with no gaps.
4. Run the whole book and use it for real: the next consent requests, the watch list, the committee pack.
5. Decide with three numbers: agreement with counsel by blocker, false alarms, and people's hours per full-book question.

Grading against counsel: misses, false alarms and clauses quoted

J.Crew, Serta, Chewy and omni-blockers, uptiers and double-dips: LME protection questions answered

In shortThe questions lenders, investors and counsel search for most, each answered on its own.