DSCR Covenants: How $400M of Interest on $550M EBITDA Killed Toys R Us

Most PE associates can recite the leverage ratio on a deal in their sleep. Ask the same associate what the maintenance covenant headroom looks like in year three of a downside case, and you'll usually get a blank stare.

That's the gap that killed Toys "R" Us.

The deal closed at 5.7× leverage on $550M of EBITDA — by 2007 LBO standards, not especially aggressive. The cap stack looked fine on a static spreadsheet. What it couldn't survive was a covenant test in a year when EBITDA fell faster than the debt could be paid down. Twelve years after close, KKR, Bain and Vornado walked away from a 0.0× MOIC, and ~33,000 employees lost their jobs.

This post is about the covenant mechanic that quietly does most of the killing in failed LBOs — the debt service coverage ratio, or DSCR — and how it played out at Toys "R" Us specifically.

What DSCR actually is

DSCR is the cleanest single test of whether a leveraged company can keep paying its lenders out of operating cash flow. The headline formula:

DSCR = Cash Available for Debt Service / Total Debt Service

In an LBO context, "cash available for debt service" is usually EBITDA minus capex and cash taxes — sometimes minus a working-capital reserve. "Total debt service" is cash interest plus scheduled principal amortization in the period.

The cousin metric, interest coverage, drops the principal component and just compares EBITDA to cash interest:

Interest Coverage = EBITDA / Cash Interest

PE credit agreements typically include one or both. Bank deals (Term Loan A, revolvers) tend to carry maintenance covenants — coverage and leverage thresholds the company must test against every quarter. High-yield notes typically carry incurrence covenants — tests that only apply when the borrower wants to do something (issue more debt, pay a dividend, sell an asset).

The two are not equivalent. A maintenance covenant breach is a default. An incurrence covenant breach is a no.

That distinction matters because it determines whether the lender has a steering wheel before the wheels come off, or only an emergency brake when they already have.

How DSCR breaks in a downside case

A static cap table at close tells you nothing about whether the deal survives a bad year. The math is asymmetric:

• Cash interest is largely fixed. Floating-rate debt resets on a benchmark schedule, but the spread is locked. A SOFR + 450 term loan stays a SOFR + 450 term loan whether EBITDA is up 10% or down 30%.
• EBITDA is the variable. Operating businesses fluctuate; consumer retail fluctuates a lot.

So when EBITDA compresses, every dollar of decline lands almost one-for-one on coverage. If interest is $400M and EBITDA falls from $550M to $400M, your interest coverage drops from 1.4× to 1.0× — and on the second day below covenant, you are formally in default.

The lenders' rational response, then, is to either (a) waive the covenant in exchange for fees and a higher rate, (b) demand additional collateral or amortization, or (c) push the borrower toward restructuring. By the time you're at the table, the equity has already lost most of its negotiating leverage. The capital structure decides the outcome before the conversation starts.

This is why, when you size leverage on an LBO, the binding constraint is rarely the leverage cap. It's almost always the coverage covenant in a stress case. The free debt capacity calculator makes this visible in 30 seconds: set leverage and the interest rate, and watch which covenant binds first as you flex EBITDA down.

The Toys "R" Us setup

Here are the numbers from close, July 21, 2005:

KKR, Bain Capital, and Vornado Realty Trust split the equity. Vornado was the real-estate hand — Toys owned a substantial block of store-level real estate, and the thesis baked in unlocking that value through a property-co spinoff.

The headline number to fixate on is the $5.3B of debt translating to ~$400M/year in cash interest. Against $550M of EBITDA, that's interest coverage of roughly 1.4×. In a flat year, fine. In a year where EBITDA drops 10%, you're at 1.2×. In a year where EBITDA drops 30% — closer to what actually happened during the secular Amazon decline — you're at sub-1.0×.

There's no version of that cap table that survives an EBITDA recession in retail.

What happened when EBITDA actually fell

Amazon launched its dedicated toy warehouse two years before the buyout. The sponsors did not seriously model e-commerce disruption — the case file at close was a traditional retail base case extended forward. By 2010, Amazon had become the structural competitor of record, and Toys' EBITDA had begun the slow grind down that retail consultants spent the rest of the decade calling "transition."

Here's the mechanic of what coverage compression looks like when interest is sticky and EBITDA is the variable:

The "down 30%" line isn't a precise quarterly snapshot — it's the shape. The line that matters is the third column: cash left over for capex, working capital, and, critically, reinvestment in e-commerce. By the time EBITDA had compressed enough to break coverage, the cash that would have funded a real digital transformation simply wasn't there. The business needed to invest $500M+ in distribution centers and digital infrastructure to compete with Amazon. The cap structure permitted ~$0.

The real-estate spinoff (Propco) that was supposed to be the relief valve also got jammed. Bank covenants prevented asset sales during distress, and "distress" was defined exactly when the proceeds would have mattered most.

The 2017 refinancing failure

The thing about a covenant-heavy cap structure isn't just that it can trigger a default. It's that it forecloses the workaround options that healthier companies use to buy time.

When 2017 arrived with a $400M-plus debt wall coming due, Toys' coverage profile was too compressed to refinance into the high-yield market on workable terms. Lenders priced incremental capital at rates the EBITDA couldn't support. The holders of the existing debt — sitting on covenant trip-wires — had every reason to negotiate from a position of strength rather than concede new flexibility. The asset sales that might have raised liquidity were blocked by the same covenants.

There was nothing left to negotiate with. Toys filed Chapter 11 in September 2017. Liquidation was announced in March 2018. Final store closings completed in June 2018, 12 years and 11 months after the buyout closed.

Final returns to LPs: a 0.0× MOIC. A -100% IRR. The full equity check, gone.

The asymmetric footnote: the sponsors collected approximately $470M in management fees and dividend recaps over the 12-year hold. The fund-level economics — what the LPs experienced — was a complete wipeout. The sponsor-level economics, less so. (We dug into that decoupling at length in the dividend recap post, but the short version is: incentive design in PE does not, by default, align the GP's downside with the LP's.)

Why DSCR is the covenant that does the killing

In post-mortems, people tend to point at the leverage multiple — 5.7× sounds too high in retrospect. But the deeper lesson is about which covenant binds and when.

• A leverage covenant breach is a slow trigger — it requires either rising debt or falling EBITDA over several quarters before it formally trips.
• A DSCR or interest coverage covenant is a fast trigger — one bad quarter of EBITDA can put you over the line.

DSCR is also the covenant lenders care most about in distress, because it's the most direct test of "are you producing enough cash to pay me?" Leverage is a snapshot of the balance sheet; DSCR is a flow test on the income statement. In a refinancing conversation, the lender doesn't care about the static cap table — they care about whether you generate the cash to service whatever they extend you.

For PE diligence, that translates into a different stress test than the one most associates run. The standard sensitivity is EBITDA × exit multiple → MOIC/IRR. The covenant stress is EBITDA × duration → quarter-by-quarter coverage path. The first one tells you what the deal returns if it works. The second one tells you whether the deal survives if it doesn't.

You can preview both flavors of stress at high level — leverage limit vs coverage limit — in the debt capacity calculator. The deeper version belongs in your model: a DSCR schedule by quarter, with covenant lines drawn in, against a downside EBITDA case that takes the asset out 20-30% from base.

The practical takeaway

Three rules that fall out of the Toys "R" Us file:

1. Size leverage against a downside EBITDA case, not the base case. If your model only shows coverage above the covenant line in the base case, the deal is one operating miss away from default.

2. Pay attention to which covenants are maintenance vs incurrence. A capital structure heavy on bank debt (maintenance) is brittle in a downturn. A structure that pushes more risk into high-yield notes (incurrence) buys flexibility — at the cost of a higher coupon.

3. Build the coverage schedule into the LBO model. Quarterly DSCR projection, with covenant trip-lines, against a real downside. Most templates skip this entirely and present coverage only at the annual level — which is exactly the resolution at which a covenant breach gets missed.

The clean S&U test in the sources & uses post covers the mechanical integrity of the cap table at close. DSCR covers what happens to the same cap table when EBITDA moves the wrong direction.

Both matter. Sponsors who only check the first one end up running the Toys "R" Us scenario in some other industry, on some other vintage.

What this looks like in a real model

Our All-in-One PE Model carries a full coverage schedule alongside the debt waterfall — DSCR and interest coverage by period, with configurable covenant thresholds, against a downside EBITDA toggle. The LBO tab surfaces both the leverage-cap and coverage-cap binding constraints when you size the cap table, so you can see which one is actually limiting you. See the sheet preview before buying.

Get the DSCR analysis right at diligence and you set the cap structure for the business you're going to actually operate in years 2-5. Skip it, and you build a structure for the world you wish you were operating in — which is the same place the Toys "R" Us deal ended up.

---

Sources:

1. Toys "R" Us Chapter 11 Petition (S.D. NY 2017), Prime Clerk docket — original capital structure, debt at close, interest burden
2. House Financial Services Committee Hearing, "The Aftermath of the Toys R Us Bankruptcy" (2019) — EBITDA, interest expense, layoff totals
3. Bloomberg, "How Vulture Capitalists Ate Toys R Us" (March 2018) — sponsor fee history, deal close details
4. Center for Popular Democracy, "Pirate Equity: How Wall Street Firms Are Pillaging American Retail" — aftermath and fee/dividend totals