Common LBO Model Mistakes and How to Catch Them

An LBO has a lot of moving parts that all land in two numbers: MoIC and IRR. That's what makes LBO mistakes so annoying. Almost any error, anywhere in the model, shows up as a slightly different return, and a slightly different return doesn't look wrong.

Below are the mistakes that cost the most, measured on a real model: SheetRank's Institutional Multi-Tranche LBO. It's an 11x buyout of a business with $150 million of LTM EBITDA, financed with a term loan, mezzanine debt with a PIK component, preferred equity and management rollover. Built correctly, the sponsor earns 2.12x and a 16.2% IRR over five years.

Here's what each mistake does to that answer.

The scoreboard

Every one of these models runs, balances and produces a believable return. That's the point.

Try it

Work out the LBO sources and uses

Price is EBITDA times the entry multiple. Debt is EBITDA times leverage. The sponsor funds the rest, fees included.

Purchase price: fill B5 (1 of 3)

	A	B
1	LTM EBITDA	$50,000,000
2	Entry multiple	10.00x
3	Debt multiple	5.00x
4	Transaction fees	$15,000,000
5	Purchase price
6	Debt raised
7	Sponsor equity

fxB5=tap cells and keys

789()456×÷123+−0.,:^

fx$

Check

Tap cells to put them in the formula. On a laptop you can type the numbers and operators.

Just need the number? Use the free LBO calculator.

1. Forgetting what sits ahead of the common equity

This model has $200 million of preferred equity accruing at 10%, so it's worth about $322 million at exit. If you take exit equity value straight to the sponsor without redeeming the preferred first, the IRR jumps from 16.2% to 20.9%.

It's an easy miss because the preferred isn't debt, so it doesn't come out in the "less: net debt" step most people build from memory. Anything senior to the common (preferred, PIK notes, earnouts, management promote) has to come out before you calculate the sponsor's share.

2. PIK interest that never compounds

The mezzanine here pays 6% in cash and 6% in kind. PIK interest doesn't get paid; it gets added to the balance, which then earns more interest. If the ending balance just equals the beginning balance, the PIK disappears and the mezz is too small at exit. Returns rise to 2.24x and 17.5%.

The tell is a mezzanine balance that's flat across five years even though the terms include PIK. A $225 million tranche at 6% PIK should be about $301 million by Year 5.

3. Leaving exit cash out of equity value

Equity value at exit is enterprise value minus debt plus cash. If the model keeps a $25 million minimum cash balance, that cash belongs to the equity at sale. Leave it out and the sponsor loses about 0.03x of MoIC. It's a small number, but it's the kind of thing a VP circles in red because it shows you didn't think through the bridge.

4. The exit multiple on the wrong year

Exit value should use the final year of the hold. Point the formula one column left and you're selling the business at Year 4 EBITDA ($195 million) instead of Year 5 ($210 million). Exit enterprise value falls from $2.31 billion to $2.145 billion and the IRR drops to 13.6%.

This one comes from copying a formula across a row and not checking where the last column landed. When you review, click the exit value and confirm it reads the last year, every time.

5. Counting the wrong number of years

IRR from a single entry and exit is MoIC raised to one over the number of years, minus one. Use six instead of five and a 16.2% IRR becomes 13.4%. Nothing else in the model changes, which is why it slips through. Write the hold period as an input and reference it everywhere instead of typing a 5 into formulas.

6. A debt balance that stopped rolling forward

This is the most expensive mistake on the list. Each year's beginning term loan balance should equal the prior year's ending balance. If the Year 2 to 5 beginning balances are anchored with a dollar sign to Year 1's ending balance, the model forgets every paydown after Year 1. Interest stays high, the sweep repays the same balance over and over, and the term loan at exit is $458 million instead of $212 million. Returns fall to 1.77x and 12.2%.

The fix is obvious once you see it. Finding it is the skill. Click the beginning balance in the last year and check it points at the previous column, not the first one.

Smaller mistakes worth checking

Not every error moves returns much, and that's what makes the small ones dangerous. In this model:

• Change in NWC pointing at the wrong row for one year (reading the 25% tax rate instead of the $5 million assumption) moves the IRR from 16.24% to 16.33%. Too small to notice by eye.
• A typed-in revolver draw of zero is correct in the base case and wrong the moment the downside case needs the revolver. You only find it by flexing inputs.
• Taxes that ignore interest overstate taxes and understate cash available for the sweep. Check that the tax line starts from EBT, not EBITDA.

A five-minute LBO review

Then flex the exit multiple and the growth rate and make sure the IRR moves the way you'd expect.

Practice finding them

Reading about mistakes helps. Finding them under a clock helps more. In Audit Mode you get this exact model, fully built, with five mistakes hidden in it, the same kinds as above. The round passes only when every number checks out again. If you want to build the model from scratch first, the PE modeling test guide walks through what a timed LBO test expects.