Skip to content
Module 10 of 1245 min readIntermediate

Sanity checks that catch the bad DCF

Implied EV/EBITDA, implied EV/sales, implied perpetual growth in the TV, comp-set checks, and the cross-check against precedent transactions.

83%

Listen along

Read “Sanity checks that catch the bad DCF” aloud

Plays in your browser using on-device text-to-speech — nothing leaves the page.

Learning objectives

By the end of this module, you should be able to:

  • 01Compute implied EV/EBITDA, EV/Sales, and perpetual growth from a completed DCF
  • 02Cross-check against trading comparables and precedent transactions
  • 03Solve for implied IRR and compare to cost of equity to gauge over- or under-valuation
  • 04Identify which of the five sanity checks, if disagreeing, signals a structural error in the model

Before you defend a DCF in a meeting, you put it through five sanity checks. Each one cross-references your DCF against an external reference point, and any one of them throwing a discordant signal means you go back to the model.

Implied EV/EBITDA

Compute the implied EV/EBITDA at your DCF answer (EV from the DCF ÷ year-1 EBITDA). Compare to the comparables. If your DCF implies 18x and the comps trade at 10x, your DCF is high — either your forecast is too aggressive, your terminal value is too generous, or your WACC is too low. Conversely, an implied 6x against 10x comps is suspiciously cheap.

Implied EV/Sales

Same exercise with EV/Sales. Particularly useful for businesses where margin is volatile or near zero — early-stage SaaS, miners, transport. The EV/Sales implied by your DCF should be in the same neighbourhood as comparable businesses at similar growth-and-margin profiles.

Implied perpetual growth

If you used the exit-multiple method for terminal value, back out the implied perpetual growth rate using the Gordon-growth formula in reverse. If your TV multiple implies g = 8%, that is unrealistic — long-run growth above 5% in nominal terms is rare for any business. The reverse check often reveals an over-aggressive TV that the multiple alone hid.

Cross-check against precedent transactions

If similar businesses have been acquired recently at certain multiples, your DCF should land in the same neighbourhood — adjusted for control premium (typically 20-30% above public market multiples). A DCF below precedent transactions on a controlling-stake basis is suspicious and may need a strategic-value adjustment.

Cross-check against the implied IRR

Solve for the IRR that equates the DCF price to the cash flows. This is what the equity holder would earn if your forecast is exactly right. Compare to your cost of equity. If implied IRR equals Ke, your DCF says the stock is fairly valued. If IRR > Ke, undervalued. If IRR < Ke, overvalued.

If three out of five disagree, your model is wrong

A single sanity check disagreeing with the DCF is not a problem — it might just reflect a mispriced peer. Three or more disagreeing means your model is wrong somewhere. The most common culprits, in order: terminal-value assumption, WACC build, year-5 margin assumption.

Exercise

Your DCF on a Kenyan-listed bank produces an intrinsic value per share of KES 75. Current market price is KES 40. Run the sanity checks: (1) Year-5 P/E implied by your DCF is 18x; comparable Kenyan banks trade at 6-8x P/E. (2) Implied P/B is 2.5x; peers are at 1.2x. (3) Implied perpetual growth (backed out from exit multiple) is 7%; long-run Kenyan nominal GDP growth is around 10%. (4) Precedent transactions in African banking have happened at 1.5-2.0x P/B with 20-30% control premiums. (5) Implied IRR at current market price is 32%; your Ke is 24%. Walk through what each check says, identify the most-likely error in your model, and recommend what to do next.

Key takeaways

  • If implied EV/EBITDA materially exceeds the comparable set, your DCF is too generous — usually in TV
  • Back out implied perpetual growth from your exit-multiple TV: if it implies g > 5%, you have an unrealistic TV
  • Public-comparable check should adjust for the control premium (20-30%) when comparing to precedent transactions
  • Three or more sanity checks disagreeing means the model is wrong — usually in TV, WACC, or year-5 margin

Further reading

  1. 01

    The Little Book of Valuation: How to Value a Company, Pick a Stock, and Profit

    Aswath Damodaran · Wiley · 2011

  2. 02

    Valuation Multiples: A Primer

    UBS Equity Research · 2001Foundational practitioner reference on relative-valuation discipline.

  3. 03
Loading progress…
LeadAfrikPublic Economics Hub