The terminal-value problem — where the value hides

When you built the DCF in the last module, something strange happened. The five years you forecast with care — growing the cash, discounting each year — added up to ₹474 crore, barely a quarter of the value. The single terminal value line, resting on one assumption about the distant future, was ₹1,188 crore: about 72% of the whole answer.

This is not a fluke of our numbers. In most DCFs, 60% to 80% of the estimated value sits in the terminal value — the part of the future you can see least clearly. The place where you know the least drives the number the most. This module is about that uncomfortable fact, why it happens, and how fragile it makes every DCF you will ever read.

It seems backwards that the murky far future should outweigh the years you can actually forecast. There are two reasons, and both are just arithmetic.

First, a business lives far longer than any forecast window. Five or ten years of explicit cash is a thin slice of a company meant to trade for decades. The terminal value is a container for all the rest — every year from eleven to infinity, bundled into one figure. Of course the container holding thirty-plus years outweighs the five you listed.

Second, the perpetuity formula that produces the terminal value is extraordinarily sensitive, because it divides by a small number. Recall the shape:

Terminal value = (next year's cash) ÷ (discount rate − terminal growth rate)

The denominator is the gap between two rates that are often close together — say 12% minus 4%, a gap of just 8%, or 0.08. When you divide by a small number, tiny changes in that number swing the result violently. Nudge the gap from 0.08 to 0.06 and you are dividing by three-quarters of what you were — the value leaps. The terminal-value problem is, at bottom, the danger of dividing by a small and uncertain number.

This connects to a deep idea we meet again in Part Three: competition eventually drags extraordinary returns back toward ordinary ones. A terminal growth rate that assumes a company keeps outrunning the field forever is quietly assuming competition never arrives — which history says it always does.

Take the exact company from module 006. illustrative Year-5 free cash flow was ₹161.1 crore; the discount rate is 12%; the present-value factor for year five is 0.567; the forecast years are worth ₹474 crore. Everything is held fixed except the terminal growth rate. Watch what a two-point change on either side does.

Read the arithmetic for the 6% row so you trust it. Next year's cash is 161.1 × 1.06 = 170.7. The gap is 0.12 − 0.06 = 0.06. Terminal value at year five is 170.7 ÷ 0.06 = ₹2,845 crore. Discount it: 2,845 × 0.567 = ₹1,614 crore. Add the ₹474 crore of forecast years: ₹2,088 crore.

Now stand back. A change in terminal growth from 2% to 6% — a band no honest analyst can rule in or out for the far future — moves the whole valuation from ₹1,406 crore to ₹2,088 crore. That is a swing of nearly 50%, driven entirely by one soft assumption. And notice the second column of the table: as the growth rate rises, the terminal value's share of the answer climbs too, from 66% to 77%. The more you assume the future is bright, the more of your valuation rests on the part you can prove least.

Two methods exist for the terminal value, and comparing them is the single best sanity check you can run. The first is the perpetuity growth method we used: assume the cash grows at a steady rate forever and apply the formula. The second is the exit multiple method: assume that at the end of the forecast the business would be worth some sensible multiple of its final-year cash flow or earnings — the price a buyer might pay for a mature company like it.

Run our example both ways. illustrative The perpetuity method at 4% gave a terminal value of ₹2,094 crore at year five. Suppose instead you say "a mature business like this might change hands at 13 times its free cash flow". Then the exit-multiple terminal value is 161.1 × 13 = ₹2,094 crore — almost identical. The two methods agreeing is reassuring: it means your growth assumption and your multiple assumption tell the same story.

But suppose the exit multiple you would actually pay is only 10×. Then the terminal value is 161.1 × 10 = ₹1,611 crore — a fifth lower than the perpetuity number implied. That disagreement is a gift: it is telling you your 4% perpetual growth is quietly more optimistic than the multiple a real buyer would pay. Neither number is "the truth". The discipline is to run both and treat any wide gap as a warning that one of your assumptions is doing too much work.

The terminal value is the softest number in the model, and honesty about its limits is the whole point of this module.

It cannot see competition arriving. A perpetuity growth rate quietly assumes the business keeps its edge forever. In reality, high returns attract rivals, and margins fade. A terminal value built on today's economics may be pricing a moat that will be gone by year fifteen.

It cannot be verified. You will never receive the year-40 cash flows to check your assumption against. Unlike a near-term forecast, which reality eventually tests, the terminal value is a bet that is never fully settled — which is exactly why it is so easy to fudge.

It cannot make the discount rate reliable so far out. The same discount rate applied to year 30 assumes today's view of risk holds for a generation. Small errors compound over long horizons.

1. Setting terminal growth above the economy. Any rate above long-run nominal GDP growth assumes the company eventually becomes the economy. Cap it at, or below, that.

2. Calling a high terminal growth 'conservative'. The word does not make the assumption safe. Test whether the rate is defensible against the fade of competition, not against your hopes.

3. Refining the near term while ignoring the terminal value. Beginners polish year-one revenue to the decimal while the 72% of value in the terminal line goes unquestioned. Effort should follow importance.

4. Never cross-checking the two methods. Running only perpetuity, or only an exit multiple, hides the disagreement that would have warned you. Always run both.

5. Treating the terminal value as a fact. It is the least knowable number in the whole exercise dressed up as the most important. Hold it with the loosest grip, not the firmest.