LBO Interview Questions: The Paper LBO, Worked Twice

The transaction, and why free cash flow is just net income

A sponsor buys a business with $100 million of EBITDA at 8.0 times, so the enterprise value is $800 million. Lenders provide 5.0 times, which is $500 million, and the sponsor puts in $300 million. Fees are ignored here on purpose, and nothing below adds them back.

The operating assumptions. EBITDA rises by $8 million a year, so it runs $108, $116, $124, $132 and $140 million. Depreciation and amortization is $25 million and capital expenditure is $25 million, deliberately equal. The blended coupon is 9 percent on the opening debt balance, set against a federal funds effective rate of 3.88 percent and a bank prime loan rate of 7.00 percent on 29 September 2026, checked 1 October 2026 at federalreserve.gov. Tax is 21 percent, which the IRS states is the rate corporations apply to taxable income, checked 1 October 2026 at irs.gov.

Setting depreciation equal to capital expenditure is the trick that makes this doable by hand. Net income plus depreciation less capital expenditure collapses to net income, so free cash flow is just net income.

The paper LBO interview question, done in five minutes

Year one out loud. Interest is 9 percent of $500 million, which is $45.00 million. EBIT is 108 less 25, which is $83 million. Taxable income is 83 less 45, which is $38.00 million. Tax at 21 percent is $7.98 million. Net income is $30.02 million, all of which repays debt, leaving $469.98 million.

Exit at 8.0 times year five EBITDA of $140 million, which is $1,120 million. Take off debt of $259.13 million and equity proceeds are $860.87 million. The multiple is 860.87 over 300, which is 2.870 times. The IRR is 2.870 to the power of one fifth, less one, which is 23.47 percent.

What revenue has to do for that EBITDA line

The $8 million a year is a postulate, and it covers the only part of this model an operator actually controls. Put a small income statement under it before accepting it. Assume a 20 percent EBITDA margin at entry, which is our assumption rather than the case's: $100 million of EBITDA then sits on $500 million of revenue, and the $800 million price is 1.6 times that revenue.

From there the postulate forks, and both branches are arithmetic.

Hold the margin at 20 percent and revenue carries the whole thing. Each year's EBITDA divided by 0.20 gives revenue of $540, $580, $620, $660 and $700 million. Year one needs $40 million of new revenue on a base of $500 million, which is 8.0 percent, and five years of it is 700 over 500, or 1.4 times, compounding at 6.96 percent a year.

Hold revenue growth at 4 percent instead and the margin does the work. Revenue runs $520.0, $540.8, $562.4, $584.9 and $608.3 million, and the margin each year has to reach 20.8, 21.4, 22.0, 22.6 and 23.0 percent, against 20.0 at entry.

Say which branch you are underwriting. Both produce the same 2.870 times, and one of them asks a sales force for 7 percent a year while the other asks operations for 300 basis points of margin.

The five minute shortcut costs 80 basis points of IRR

Almost nobody recomputes interest on a falling balance in their head. The standard shortcut holds interest at $45.00 million for all five years.

Do that and year five closing debt is $286.70 million rather than $259.13 million. Equity proceeds are $833.30 million, the multiple is 2.778 times, and the IRR is 22.67 percent.

So the shortcut costs 80 basis points of IRR, and it costs them in one direction every time. Holding interest at the opening level overstates the interest bill, understates cash flow, and understates the paydown. An interviewer who knows the model knows the sign of your error before you finish.

Say the shortcut as you take it. "I will hold interest flat, which understates paydown, so treat 2.78 times as a floor." That sentence is worth more than the extra precision, and it takes four seconds.

Converting a multiple to an IRR in your head

You will be asked to convert between a multiple and a rate with no calculator. Five pairs cover most of what gets asked.

Two facts do most of the work. Doubling in five years is roughly 15 percent. Tripling in five years is roughly 25 percent. Between them, two and a half times in five years is roughly 20 percent, which is the number most funds quote as a target.

The direction of the error matters more than the decimal. A candidate who says "2.87 times over five years, so call it a bit under 25 percent" has answered the question. A candidate who says 23.47 percent without showing how they got near it has shown a memorized number.

Leverage is worth 9.9 points of IRR, and it reverses

The standard answer is that debt is cheaper than equity and magnifies the return. That is true and it is not an answer, because it does not say how much. Price it.

Buy the same asset at the same 8.0 times for all equity, so $800 million of cash. Without interest, net income is EBIT after 21 percent tax, and free cash flow equals net income again. Over the five years that accumulates $65.57, $71.89, $78.21, $84.53 and $90.85 million, which sums to $391.05 million.

Exit at the same $1,120 million and add the accumulated cash. The unlevered sponsor collects $1,511.05 million on $800 million, which is 1.889 times and a 13.56 percent IRR.

Leverage is worth 9.9 points of IRR here. Two things produce it. The enterprise value gain of $320 million lands on a $300 million equity base instead of an $800 million one. And the interest deduction shelters income, so cumulative cash tax is $64.03 million levered against $103.95 million unlevered.

Now say the part that makes it a real answer. The same magnification works in reverse. Hold EBITDA flat at $100 million and exit at 7.0 times, giving $700 million against debt that barely moved, and the equity is close to wiped out while the unlevered buyer is merely disappointed.

Interest coverage and the covenants lenders write

"How much debt can this business support" has a cash answer and a tax answer, and candidates give only the cash one.

The cash answer is coverage, and it moves every year here. At $500 million of debt and a 9 percent coupon, year one interest of $45.00 million against $108 million of EBITDA is 2.4 times covered. By year five, interest has fallen to $29.41 million and EBITDA has reached $140 million, so coverage is 4.8 times. Leverage falls on the same two movements: $500 million against $100 million of EBITDA is 5.0 times at entry, and $259.13 million against $140 million is 1.85 times at exit.

Lenders write covenants around both ratios, and the year that has to survive them is the first one, because that is where this model is tightest. Coverage of 4.8 times at exit is the easiest number in the table, and a candidate who quotes it has answered a question nobody underwrote.

The 30 percent interest cap costs seven basis points

The tax answer is the cap. The IRS states that "the business interest expense deduction allowed for the tax year is limited to the sum of" a list of three items, of which the one that binds here is the applicable percentage of adjusted taxable income. It then states that "The applicable percentage is 30% (30% ATI limitation)", checked 1 October 2026 at irs.gov. For tax years beginning after 2024 that base adds depreciation back.

Test it here. Adjusted taxable income in year one is $108 million, so the cap is $32.40 million against $45.00 million of interest. Twelve million six hundred thousand dollars of interest is disallowed and carries forward.

Model it properly and cumulative cash tax rises from $64.03 million to $64.89 million, year five debt lands at $261.51 million, the multiple is 2.862 times and the IRR is 23.40 percent. Seven basis points. So the honest verdict is that the cap barely touches the return at this leverage and it does move the early cash tax, which is where a covenant test lives. The full year by year treatment sits on the LBO model explained.

One turn of exit multiple, and two extra years of hold

Two sensitivities come up constantly and each has an arithmetic answer.

1. What if you exit at a lower multiple. At 7.0 times, year five enterprise value is $980 million. Take off $259.13 million of debt and equity is $720.87 million, which is 2.403 times and 19.16 percent. One turn of exit multiple cost 4.3 points.
2. What if you hold for seven years instead of five. A 2.870 times outcome reached in seven years is 16.25 percent rather than 23.47 percent. Time is the denominator, and extending the hold is the cheapest way to destroy an IRR while leaving the multiple intact.

Lifting the exit multiple pays less than cutting it costs. At 9.0 times, year five enterprise value is $1,260 million, equity is $1,000.87 million, which is 3.336 times and 27.25 percent. A turn up is therefore worth 3.8 points while a turn down costs 4.3. The equity steps are equal, $140 million in each direction, and the multiple steps are equal too, 0.467 each way. The rate is the uneven part, because a fifth root flattens as the multiple it is taken from rises.

The dividend recap moves the multiple and the rate apart

Taking cash out of the business mid hold is the question that separates answers, because it moves the two measures in opposite directions. Work it.

Pay out $100 million at the end of year three, funded by adding $100 million of debt back on, with the interest cap of the previous section set aside here as it is in the base schedule. Closing debt was $383.98 million, so it goes back up to $483.98 million, which is 3.90 times the year three EBITDA of $124 million and still inside the 5.0 times the lenders wrote at entry. Year four interest is 9 percent of $483.98 million, which is $43.56 million, net income is $50.12 million, and closing debt is $433.86 million. Year five interest is $39.05 million, net income is $60.00 million, and closing debt is $373.86 million.

Exit at the same $1,120 million and equity proceeds are $746.14 million. Add the $100 million already banked and the sponsor collected $846.14 million on $300 million, which is 2.820 times against 2.870 times without the dividend. The rate moves the other way, to 24.61 percent against 23.47 percent, because $100 million of the proceeds arrived two years early.

The multiple fell five hundredths and the rate rose 1.1 points. Name the measure that improved before anyone asks, because a fund judged on multiples will not thank you for the trade.

Where paper LBO interview questions go wrong

Four errors account for most wrong paper LBOs, and none of them is a multiplication.

1. Taxing paid interest instead of deducted interest. When the cap binds, the two differ. Tax follows the deduction and cash follows the payment.
2. Forgetting that capital expenditure is not depreciation. Setting them equal is a simplification we declared. In a real model capital expenditure usually exceeds depreciation for a growing business, and the gap comes straight out of debt paydown.
3. Sweeping cash that has already been spent. Cash available for the sweep sits after interest, tax, capital expenditure and working capital. Sweeping pre tax cash flow is a factor of a fifth.
4. Netting cash against debt twice. Unswept cash accumulates on the balance sheet. Counting it as a paydown and as a cash balance double counts it.

The fourth is the quiet one. Our own recheck of this model caught a version where surplus cash reduced the loan and sat in the cash line at the same time, which flattered exit equity by tens of millions. The arithmetic was correct on every line.

How an LBO answer gets marked

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