Common DCF Mistakes

DCF is the most rigorous valuation framework, and the most abused. Because it accepts any inputs you choose, it can be used to "prove" almost any valuation you want. The same DCF model can justify ₹500 or ₹2,000 per share depending on what assumptions the analyst makes, or is incentivised to make.

Most DCF errors fall into three categories: mechanical errors (wrong formula, wrong sign), assumption errors (inputs don't reflect economic reality), and conceptual errors (wrong framework for the situation). Understanding the failure modes makes you a better analyst of others' models and a more rigorous builder of your own.

The most important discipline: always reverse-engineer the market price. Instead of computing an intrinsic value and comparing to market, ask "what assumptions does the current market price require?" Then judge whether those assumptions are realistic.

Ten common DCF mistakes:

1. Terminal growth rate > WACC. TV = FCFF ÷ (WACC − g). If g ≥ WACC, terminal value blows up to infinity. This is mathematically and economically wrong.

2. Using unadjusted reported FCFF without stripping working capital distortions, exceptional items, or one-off capex.

3. Not adjusting WACC for changing capital structure over the forecast period. A company that's deleveraging rapidly has a different risk profile in year 10 vs year 1.

4. Discounting at mid-period vs end-of-period without consistency. Mid-period discounting (cash flows received throughout the year) is more realistic but must be applied consistently.

5. Double-counting the tax shield. If using WACC (which uses after-tax Kd), don't also add the PV of tax shields separately.

6. Forgetting minority interest / associates. EV from DCF is for the whole enterprise. Must add associate values and subtract minority interest and net debt to get equity value.

7. Stale beta. Using a 5-year monthly beta from a period when the company's business has fundamentally changed.

8. Not stress-testing. A DCF without a sensitivity table (WACC × g) is incomplete.

9. Anchoring to the current stock price. The goal is independent analysis. If your model keeps "coming out" near the current price, check whether you're unconsciously fitting inputs to justify it.

10. Treating DCF output as precise. A value range (e.g., ₹800–₹1,200) is more honest and useful than a single point estimate.