How to Calculate Intrinsic Value

To calculate intrinsic value, estimate the cash a business will produce in future and discount it back to today, which is a discounted cash flow (DCF), then cross-check it against what similar companies trade at or against what the company owns minus what it owes. Treat the result as a range rather than a precise figure, and require a margin of safety between that value and the price you pay. It is useful for judging whether a price is sensible for the business behind it.

Intrinsic value is what a business is worth based on the cash it can produce, as distinct from what the market is currently willing to pay for it. The two are often different, and the gap is the entire reason valuation is worth learning.

This page covers the three approaches that are actually used, a worked discounted cash flow, and - most importantly - an honest account of how wrong these numbers can be and what to do about that.

The idea underneath all of it

A business is worth the cash it will hand to its owners over its life, adjusted for the fact that money arriving in ten years is worth less than money arriving today.

Everything else is an attempt to estimate that with imperfect information. Discounted cash flow tries to model it directly. Multiples estimate it by comparison. Asset-based methods work from what the company owns. All three are approximations, and knowing which approximation suits which business is more valuable than being fluent in any one of them.

Method 1: Discounted cash flow

DCF projects the cash the business will generate, then discounts each year back to today.

The inputs

• Free cash flow. Operating cash flow minus capital expenditure - cash left after keeping the business running. Both figures come straight from the cash flow statement.
• Growth rate. How fast that cash grows during the forecast period, usually five to ten years. Anchor it to the company's actual history and the size of its market, not to the last two good years.
• Discount rate. The return you require, reflecting risk. Higher for a volatile business, lower for a stable one.
• Terminal growth rate. The rate at which the business grows forever after the forecast period. It must be modest - no company outgrows the economy indefinitely. Something at or below long-run nominal GDP growth is the discipline.

A worked example

Take a company generating Rs 100 crore of free cash flow. Assume 10% growth for five years, then 4% forever, discounted at 12%.

Years one to five, each year's cash flow discounted back to today at 12%:

Year	Free cash flow (Rs crore)	Worth today at 12%, roughly (Rs crore)
1	110	98
2	121	96
3	133	95
4	146	93
5	161	91
Total	-	474

That is about Rs 474 crore in total for the forecast years.

The terminal value takes year five's Rs 161 crore, grows it once more at 4% to Rs 167 crore, and divides by (12% - 4%) = 8%, giving about Rs 2,094 crore as of year five. Discounted back five years at 12%, that is roughly Rs 1,188 crore today.

Total: about Rs 1,662 crore. Subtract net debt, divide by the number of shares, and you have a per-share value.

Now notice the problem

Of that Rs 1,662 crore, about 72% is terminal value - a number resting on a guess about growth beyond year five, which nobody can know.

And it moves violently. Change the terminal growth from 4% to 5% and the denominator falls from 8% to 7%, lifting terminal value by about 15%. Change the discount rate from 12% to 11% and it moves again, in the same direction, by more. Two reasonable analysts using the same statements can produce values 40% apart without either doing anything wrong.

This is the honest position on DCF: it is a good way to understand what a price implies, and a poor way to produce a precise number. Its best use is often backwards - take the current market price, and solve for the growth rate required to justify it. If the market is implying 25% growth for a decade in a mature industry, you have learned something concrete without pretending to forecast.

Method 2: Relative valuation

Comparison to similar companies. Faster, more widely used, and honest about being a comparison rather than an absolute.

• P/E - price divided by earnings per share. The default, and unreliable where earnings are cyclical or distorted by one-offs.
• EV/EBITDA - enterprise value against operating earnings. Better for comparing companies with different debt levels, because enterprise value includes debt.
• P/B - price to book value. Most meaningful for banks and financials, where the balance sheet is the business.
• P/S - price to sales. A last resort for companies that are not yet profitable; it ignores whether the revenue is worth anything.

Two rules make these usable. Compare against genuine peers - same industry, similar size, similar growth - not against a market average. And compare against the company's own history: a stock at 45 times earnings that has traded at 20 to 25 for a decade is expensive relative to itself, whatever the sector average says.

The limitation is structural. Relative valuation tells you whether something is cheap compared to these other things. If the whole sector is overvalued, every member of it looks reasonable.

Method 3: Asset-based

What the company owns minus what it owes. Most useful for asset-heavy businesses, holding companies, and situations where the business is worth less as a going concern than in pieces.

Book value is the accounting version and understates assets carried at historical cost - land bought decades ago sits on the books at a fraction of its value. For most operating businesses this method is a floor rather than a valuation, because it captures nothing about the ability to earn.

The margin of safety

Given that every method above has meaningful error bars, the practical response is not to seek a better model. It is to require a gap between value and price large enough to absorb being wrong.

If your estimate is Rs 500 a share, buying at Rs 490 leaves no room for a mistaken assumption. Buying at Rs 350 means the thesis can be partly wrong and the investment still works.

Two habits follow. Produce a range rather than a point - run the DCF with pessimistic, base and optimistic assumptions and see how wide the spread is. And require a larger margin where the business is harder to predict: a cyclical commodity producer deserves more room than a utility with contracted revenue.

Points specific to Indian companies

• Use consolidated numbers. Subsidiaries often hold much of the actual business; standalone statements can miss it entirely.
• Check promoter pledging. High pledged promoter holding is a risk that no cash flow model captures, because a falling price can force sales that push it lower.
• Holding companies trade at persistent discounts. A company whose main asset is stakes in other listed companies typically trades well below the sum of those stakes, and the discount rarely closes on schedule.
• Adjust the discount rate for the risk-free rate you actually face. Indian government bond yields have generally sat well above those in developed markets, and a discount rate borrowed from a US textbook will overvalue an Indian business.
• Cyclicals need a full cycle. Valuing a commodity producer on peak-cycle earnings produces a P/E that looks cheap precisely when the stock is most expensive.

Common mistakes

• False precision. Reporting a value of Rs 487.30 implies a confidence the inputs cannot support. Say Rs 450 to Rs 550 and mean it.
• Reverse-engineering the answer. Adjusting growth until the model justifies the price you already wanted to pay. Easy to do accidentally; set assumptions before opening the price chart.
• Extrapolating a good run. Three strong years are not a growth rate. Look at a decade, including whatever went wrong in it.
• Comparing multiples across sectors. A bank at 2 times book and an FMCG company at 12 times book are not comparable statements about value.
• Valuing a business you do not understand. If you cannot explain how it makes money in two sentences, the model is arithmetic performed on guesses.
• Treating value and price as the same question. A fairly valued business can stay unloved for years. Valuation tells you what to own; it says very little about when. That second question is covered in how to combine technical and fundamental analysis.

Before any of this is worth doing, the statements have to be read properly - that groundwork is in how to read financial statements. Valuation applied to numbers you have not questioned is confident arithmetic on an unexamined foundation.

Frequently asked questions

What is the intrinsic value of a stock?

Intrinsic value is what a business is worth based on the cash it can hand to its owners over its life, adjusted for the fact that money arriving later is worth less than money today. Divided by the number of shares, it gives a per-share value. It is distinct from the market price, and the gap between the two is what valuation is for.

How do you calculate intrinsic value with a DCF?

Take free cash flow, which is operating cash flow minus capital expenditure. Grow it at a rate anchored to the company's history for five to ten years, then at a modest terminal growth rate forever. Discount each year's cash flow and the terminal value back to today at the return you require, add them up, subtract net debt and divide by the number of shares.

How accurate is a DCF valuation?

Less accurate than its decimal places suggest. In the worked example on this page, about 72% of the value is terminal value, which rests on a guess about growth beyond year five. Small changes to terminal growth or the discount rate move the answer sharply, and two reasonable analysts can produce values 40% apart. Treat a DCF as a range, and as a way to see what a price implies.

What discount rate should I use for an Indian company?

There is no single correct figure. The discount rate is the return you require, so it should be higher for a volatile business and lower for a stable one. It should also reflect the rates you actually face: Indian government bond yields have generally sat well above those in developed markets, so a discount rate borrowed from a US textbook will overvalue an Indian business.

What terminal growth rate should I use?

A modest one, at or below long-run nominal GDP growth, because no company outgrows the economy indefinitely. It matters more than it looks: in this page's example, moving terminal growth from 4% to 5% lifts the terminal value by about 15%, and terminal value is most of the total. If a model only works with a generous terminal rate, that is itself the finding.

What margin of safety should I use?

There is no universal percentage. The margin should be large enough to absorb being wrong, and larger where the business is harder to predict: a cyclical commodity producer deserves more room than a utility with contracted revenue. If your estimate is Rs 500 a share, a price of Rs 490 leaves no room for a mistaken assumption, while Rs 350 lets the thesis be partly wrong.

Which valuation method is best: DCF, P/E or book value?

None is best for every business. DCF models the cash directly and suits businesses whose cash flows can be reasonably forecast. Multiples such as P/E and EV/EBITDA compare a company with genuine peers and with its own history. Asset-based methods suit asset-heavy businesses and holding companies, and are a floor rather than a valuation for most operating businesses. Using more than one is a sensible cross-check.

Does a low P/E mean a stock is undervalued?

Not by itself. P/E is unreliable where earnings are cyclical or distorted by one-off items, and a commodity producer valued on peak-cycle earnings looks cheap precisely when the stock is most expensive. A low P/E is also only low compared with something: if the whole sector is overvalued, every member of it looks reasonable. Compare against genuine peers and the company's own history.

Does intrinsic value tell me when to buy a stock?

No. Valuation tells you what is worth owning; it says very little about when to act. A fairly valued business can stay unloved for years, and a price below your estimate can keep falling for reasons your model does not capture. Timing is a separate question, covered in how to combine technical and fundamental analysis.