LBO Model Explained: Structure, Schedule and Returns

What an LBO is, and why the debt lifts the return

A leveraged buyout is the purchase of a company using a small amount of the buyer's own money and a large amount of borrowed money, where the borrowed money is secured on the company bought rather than on the buyer. The company then services and repays that debt out of its own cash. Five years or so later the buyer sells, and whatever is left after the remaining debt belongs to the buyer. LBO modeling is the arithmetic that turns that purchase into a return figure, and an LBO analysis is the same arithmetic run against a particular price to see whether the return clears a hurdle.

One identity governs the whole thing, and every figure further down this page is an instance of it.

Equity at exit equals enterprise value at exit, minus net debt at exit.

That is why leverage changes the answer, and the mechanism splits in two. Both halves are visible in the Harlow numbers.

The first half is the base the gain lands on. The sponsor writes $378.0 million and controls $720.0 million of enterprise value, which is 720.0 divided by 378.0, or 1.9048 times its own money. When EBITDA growth adds $180.00 million of enterprise value, that $180.00 million lands on the $378.0 million base, which is 47.6 percent of it. A buyer funding the same business with equity alone would need $730.8 million at close, since there are no financing fees without debt but the transaction fees remain. The same $180.00 million of enterprise value growth is then 24.6 percent of the base. Identical operating performance, roughly half the effect.

The second half is the debt paydown, and an all-equity buyer cannot have it at all. Harlow's own cash reduces net debt from $360.00 million to $202.37 million over five years, and every dollar of that reduction lands in the equity by the identity above. That is $157.63 million the buyer did not have to earn in the market. A buyer with no debt has nothing to repay, so this term is zero for them by construction.

Now the cost, because a page that sells only the upside is selling half an argument. Leverage scales both directions off the same smaller base. Hold Harlow's EBITDA flat at $80.0 million for five years and the exit enterprise value is 9.0 times 80.0, or $720.0 million. Flat EBITDA generates less cash every year than growing EBITDA, so net debt at exit must be higher than the $202.37 million below. Equity at exit is therefore strictly below 720.0 less 202.37, which is $517.63 million, so the money multiple cannot exceed 517.63 over 378.0, or 1.369 times. Five years of that is an IRR under about 6.5 percent. The growth did not disappoint by much in that scenario; it simply did not happen.

What the sponsor is actually buying

Start with what the money does. Harlow Packaging has $80.0 million of EBITDA. A sponsor agrees a price of 9.0 times, so the enterprise value is $720.0 million.

The lenders provide 4.5 times EBITDA, which is $360.0 million. The sponsor writes a check for the rest. But the rest is larger than $360.0 million, because two sets of fees have to be funded on day one.

Financing fees run 2.0 percent of the debt raised, so 2% of 360.0 is $7.2 million. Transaction fees run 1.5 percent of the enterprise value, so 1.5% of 720.0 is $10.8 million. Total uses are 720.0 plus 7.2 plus 10.8, which is $738.0 million.

Sponsor equity is $378.0 million, or 51.2 percent of uses. A model that ignores fees puts the check at $360.0 million instead, which flatters the eventual money multiple by 5 percent before anything else happens.

The multiples are ours. Nine times entry and four and a half turns of debt are assumptions about a 2026 market, and they carry our name. Nobody publishes what a lender will accept.

The capital structure, layer by layer

The model below uses two instruments to keep its schedule readable. Real structures stack more, and which layers are present changes the schedule rather than the identity. Here is the full ladder, senior first, with what each one does to the arithmetic.

Two of those nine change the answer more than the rest, and both are worth a number.

PIK trades cash now for principal later, and the exchange rate is the spread. Move Harlow's $90.0 million of notes from cash pay to PIK at the same 9.50 percent. Year one cash interest falls from $30.50 million to $21.95 million, the term loan's share alone. Over five years the accrual compounds the notes from $90.0 million to $141.68 million, which is $51.68 million of extra principal, and that is exactly the cash interest never paid. So before tax the trade is dollar for dollar. Whether it helps depends only on what the freed cash does in the meantime: here it repays a term loan costing 8.13 percent while the notes accrue at 9.50 percent, so the structure loses 137 basis points of spread on every swept dollar.

Rollover is the one layer that changes who owns the gain. It lowers the check and lowers the claim in the same proportion, so it moves the sponsor's dollars of profit without moving the money multiple. It also means the exit equity is split, which a model that tracks only one equity line will miss.

The debt, and where its rate comes from

Two instruments. A term loan B of $270.0 million, floating, amortizing 1 percent a year, with a 75 percent cash flow sweep. Senior notes of $90.0 million, fixed, no amortization, repaid at exit.

The floating rate needs a base. The Federal Reserve put the federal funds effective rate at 3.88 percent and the bank prime loan rate at 7.00 percent on 29 September 2026, checked 1 October 2026 at federalreserve.gov. We price the term loan at the federal funds effective rate plus 425 basis points, which is 8.13 percent.

That is a convention and it deserves a flag. Real credit agreements reference term SOFR, and H.15 does not publish it. We have no government source for term SOFR, so we use the rate that is published and say what we substituted. The senior notes carry 9.50 percent, which is ours.

Year one interest is 8.13 percent on $270.0 million plus 9.50 percent on $90.0 million. That is 21.95 plus 8.55, which is $30.50 million.

The interest cap that bites at four and a half turns

Here is the line most LBO templates leave out. Interest is not automatically deductible in full. The IRS states that "the business interest expense deduction allowed for the tax year is limited to the sum of" three items, which it lists as business interest income, the applicable percentage of adjusted taxable income, and floor plan financing interest expense. It then states that "The applicable percentage is 30% (30% ATI limitation)", checked 1 October 2026 at irs.gov.

The base changed recently, and the change runs in the borrower's favor. The same instructions state that "P.L. 119-21, commonly known as the One Big Beautiful Bill Act, amended section 163(j) to allow an add-back to taxable income for any deduction allowable for depreciation, amortization, or depletion to determine ATI for tax years beginning after 2024." So the cap is now 30 percent of an EBITDA style figure rather than an EBIT style one.

Test year one. Adjusted taxable income is $80.0 million, so the cap is $24.00 million. Interest is $30.50 million. The deduction is limited to $24.00 million and $6.50 million is disallowed. It is not lost: the instructions state that disallowed interest "is carried forward to the following year as a disallowed business interest expense carryforward."

At four and a half turns of leverage with coupons near 8 and 9 percent, the cap binds. That is the finding, and it is checkable in two minutes.

The debt schedule every LBO model turns on

EBITDA grows by $5.0 million a year. Depreciation and amortization is $24.0 million, capital expenditure $20.0 million, and the working capital build $3.0 million. Tax is 21 percent, which the IRS states as the rate corporations apply to taxable income, checked 1 October 2026 at irs.gov.

Walk year one. Taxable income is EBITDA of 80.0 less D&A of 24.0 less deducted interest of 24.0, which is $32.00 million. Tax at 21 percent is $6.72 million. Cash available is EBITDA of 80.0 less capex of 20.0 less working capital of 3.0 less the full cash interest of 30.50 less tax of 6.72, which is $19.78 million. Mandatory amortization takes $2.7 million, the sweep takes 75 percent of the $17.08 million left, which is $12.81 million, and the balance sits in cash.

Note that tax is computed on deducted interest while cash is reduced by paid interest. Confusing the two is the single most common error in a levered model, and it always runs the same way: it overstates cash.

Leverage falls from 4.5 turns to 2.02 turns without a single dollar of new equity. That deleveraging is the engine.

Exit, and where the return actually came from

Exit at the entry multiple, 9.0 times year five EBITDA of $100.0 million, which is $900.0 million. Take off net debt of $202.37 million. Equity proceeds are $697.63 million.

The sponsor put in $378.0 million. The money multiple is 697.63 over 378.0, which is 1.846 times. Over five years that compounds to an IRR of 13.04 percent.

Now split the $319.63 million of gain three ways, and make the pieces add up.

1. EBITDA growth. Twenty million dollars of extra EBITDA at a constant 9.0 times adds $180.00 million of enterprise value.
2. Debt paydown. Net debt falls from $360.00 million to $202.37 million, which hands $157.63 million to the equity.
3. Fees paid at entry. The $18.00 million of fees was equity that bought no asset, so it subtracts $18.00 million.

Add them: 180.00 plus 157.63 less 18.00 is $319.63 million, which is exactly the gain. An attribution that does not reconcile to the dollar is hiding something, usually the fees.

What one turn of entry multiple is worth

Hold the exit at 9.0 times and the debt at 4.5 turns of entry EBITDA. Change only the price paid.

Every row holds the exit at the entry multiple, so none of these returns carries any multiple expansion. A sponsor who underwrites one is assuming a higher price at exit than at entry, which is an assumption about the market in five years rather than about the company.

One turn of entry multiple is worth roughly five and a half points of IRR here. Nothing else in the model comes close. Improving the EBITDA growth assumption, tightening capex, or lifting the sweep to 100 percent all move the answer by a fraction of that.

Read what that shows about the job. The modeling is arithmetic, and the arithmetic says the price is the decision. A candidate who spends the whole answer on the debt schedule and never says what price makes the return has done the easy half.

The four conventions that move an LBO model's answer

Each of these is defensible. Each gives a different number.

1. The interest cap base. With the depreciation add back, cumulative cash tax over five years is $40.95 million and the IRR is 13.04 percent. Under the base that applied for tax years 2022 through 2024, the year one disallowance is $13.70 million rather than $6.50 million, cumulative tax rises to $48.51 million, and the IRR falls to 12.76 percent. The tax line moves 18 percent while the IRR moves 28 basis points.
2. Ignoring the cap entirely. Deducting all interest leaves exit net debt at $201.69 million and an IRR of 13.06 percent. So skipping the cap costs almost nothing in return and misstates the first two years of cash tax, which is precisely when a covenant gets tested.
3. Fees in or out of the equity check. Funding the $18.00 million of fees from equity gives 1.846 times. Leaving them out gives an entry check of $360.0 million and 1.938 times, which is 5 percent of fabricated performance.
4. The sweep rate. At 75 percent, surplus cash splits between repaying the loan and sitting on the balance sheet. At 100 percent it all repays the loan. Net debt barely changes, because unswept cash still nets against debt. The sweep reshapes the closing balance sheet and leaves the return nearly alone.

The build order an LBO model needs, step by step

Every section above and below is one step of this list, and the order is not a preference. Each step consumes a number the previous one produced, and the two places people invert the sequence are both named here.

1. Fix the entry. EBITDA, the entry multiple, and the enterprise value they imply. For Harlow, $80.0 million at 9.0 times is $720.0 million.
2. Size each layer in turns of EBITDA. Not in dollars. Turns travel between deals and dollars do not, and the sizing is what the interest cap will later test.
3. Build sources and uses, with the sponsor check as the plug. Fund the fees on the sources side or the uses will not balance. Harlow's uses are $738.0 million, so the check is $378.0 million.
4. Price the debt, and name the base. Say which published rate you priced off and what spread you added, so the figure is auditable. Ours is the federal funds effective rate plus 425 basis points.
5. Forecast the operating lines. EBITDA, depreciation and amortization, capital expenditure, working capital. These are assumptions and they should be labeled as yours.
6. Compute interest, then test the deduction, then tax. This is the first inversion people make. Tax depends on deductible interest, deductible interest depends on the cap, and the cap depends on a base you have not computed if you jump to tax first.
7. Derive cash available for debt service. EBITDA less capex less the working capital build less interest actually paid less tax actually paid. Note that step 6 gave you two different interest numbers and this line needs the paid one.
8. Run the waterfall in order. Mandatory amortization first, then the sweep on what remains, then cash retained. This is the second inversion: sweeping before mandatory amortization overstates the repayment.
9. Roll the balances and close the year. Opening balance less repayments equals closing balance, every layer, every year. If a year does not close, stop and fix it before building the next one.
10. Exit, then attribute. Enterprise value at the exit multiple, less net debt, equals equity. Then split the gain into growth, paydown and fees, and check that the three reconcile to the total.

Interviewers compress the whole list into four words: walk me through an LBO. The spoken answer is steps 1, 3, 7, 8 and 10, in that order, inside about a minute. Entry price, sources and uses, cash available for debt service, the waterfall, then exit and attribution. The other five steps are keyboard work, and reciting them out loud costs you the minute without adding anything the interviewer was asking for.

Step 10 is the one that catches errors made in steps 1 through 9. An attribution that misses by a few million dollars is telling you a balance did not roll, and it is far easier to find at that moment than after an interviewer asks.

Where an LBO answer loses its points

Our math rubric puts four weights on a calculation, and the weights are ours. No bank or sponsor publishes a marking scheme. The reasoning behind each weight sits on how an answer is assessed.

Drilling the LBO model schedule by hand

1. Rebuild year one. Interest, cap, tax, cash available, sweep, closing balances. You should land on $19.78 million of cash to debt.
2. Break the cap. Raise entry leverage to 5.5 turns, which is $440.0 million of debt. Interest rises and the $24.00 million cap bites harder, so more interest is disallowed and cash tax rises. Work out whether the extra leverage still helps.
3. Flatten the growth. Hold EBITDA at $80.0 million for five years and redo the exit. The ceiling is computed in the first section: the multiple cannot exceed 1.369 times, so an answer above that means a balance has not rolled. Your job is the exact figure, which needs the whole schedule rerun at a flat $80.0 million.
4. Change the exit. At 8.0 times the exit equity is $595.61 million, which is 1.576 times and 9.52 percent. At 10.0 times it is $795.61 million, which is 2.105 times and 16.05 percent.

For the questions an interviewer builds on top of this schedule, including the version you do in five minutes without a spreadsheet, see LBO interview questions. For the discount rate machinery that values the same business without leverage, see the DCF built step by step. And for the accounting that makes the tax line behave, see the statement linkages in investment banking interview questions.