DCF calculator
A discounted cash flow prices a business as the cash it will produce, brought back to what that cash is worth today. The model below is the two-stage version: an explicit forecast for a few years, then a single figure standing in for everything after it. That single figure is usually most of the answer.
The model
Cash flow in year t = FCF x (1 + g) ^ t
Present value of it = cash flow in year t / (1 + r) ^ t
Terminal value = final year cash flow x (1 + g2) / (r - g2)
Value = sum of the present values + terminal value / (1 + r) ^ N
FCF is the most recent free cash flow, g the growth applied through the explicit forecast, N the length of that forecast in years, r the discount rate and g2 the growth assumed for every year after. The terminal formula is Gordon growth, which is the value of a stream growing forever at g2 and discounted at r. It requires r to be above g2, and this page refuses rather than working around it.
Illustrative inputs The calculator loads with invented round numbers: free cash flow of 1000, growing 6 percent for 5 years, discounted at 9 percent, then 2 percent forever, over 500 shares with 1500 of net debt. They are not any company's filed accounts and they are not a view on any security. Three of the seven could never come from a filing, because nothing anyone has filed says what happens next year.
What the model needs from you
Cash from operations less capital expenditure, both from the cash flow statement. Use the annual figure, not a quarter, and not the trailing twelve months if the business is seasonal enough that the two disagree.
Your forecast, not a filed figure. Past growth is in the statements; whether it continues is the thing you are being asked to decide, and this field is where you write your decision down.
Whole years. Five is the common choice and it is a convention rather than a finding. A longer forecast does not add information, it moves guesswork out of the terminal value and into the ladder where it looks like analysis.
What the business grows at forever. Nothing grows faster than the economy forever, so a figure above long-run nominal growth is a claim that this company eventually becomes the economy. It has to sit below the discount rate or there is no terminal value to compute.
The return you require to hold this business instead of the alternatives. Usually a weighted average cost of capital, which is its own calculation with its own estimates inside it. This page will not pick one for you.
Total interest-bearing debt less cash and equivalents, from the balance sheet. Enter a negative number where cash exceeds debt. This is the bridge from what the business is worth to what the shares are worth.
| Year | Cash flow | Discount factor | Present value |
|---|---|---|---|
| Year 1 | 1,060 | 0.92 | 972.48 |
| Year 2 | 1,123.6 | 0.84 | 945.71 |
| Year 3 | 1,191.02 | 0.77 | 919.68 |
| Year 4 | 1,262.48 | 0.71 | 894.37 |
| Year 5 | 1,338.23 | 0.65 | 869.75 |
| The explicit forecast, discounted and summed | 4,602 | ||
| Step | Value |
|---|---|
| The explicit forecast, discounted | 4,602 |
| Terminal value at the end of the forecast | 19,499.86 |
| That terminal value, discounted to today | 12,673.57 |
| The whole business, before debt | 17,275.57 |
| How much of that is the terminal value | 73.36% |
| Less net debt | 15,775.57 |
| Value per share | 31.55 |
- The terminal value is 73 percent of this answer. Most of what you are reading is the perpetual growth rate and the discount rate, not the five figures you forecast.
73.36% of that answer is the two lines you cannot forecast
On the inputs above, the five years you forecast explicitly are worth 4,602. Everything after them is worth 12,674, which is 73.36% of the total. Whatever effort goes into the explicit forecast, roughly three quarters of the number that comes out is produced by two inputs: the perpetual growth rate and the discount rate.
That is not a defect in this implementation. It is the shape of the model, and it holds at any reasonable horizon: extending the explicit forecast moves value from the terminal line into the ladder without making any of it more knowable, because year nine of a forecast is not better information than a perpetuity assumption, it just looks like more work.
The practical consequence is that a single output figure is close to worthless. Run the model at both ends of what you would defend for the discount rate, and read the range. The link below works out how far each input moves the answer.
Three of these seven are filed
Free cash flow, the share count and net debt can be read out of documents. Growth, the horizon, perpetual growth and the discount rate cannot: nothing in any filing states them, and every one of them is a decision you are making. Sorting the seven into those two piles before you start is the difference between a valuation and a number.
| Input | Where it comes from |
|---|---|
| Free cash flow | Net cash from operating activities less purchases of property, plant and equipment, both from the cash flow statement. Watch what the company has put inside operating cash flow: capitalised software and capitalised development costs sit in investing, so a business that capitalises heavily reports a larger operating figure than one that expenses the same spending. |
| Growth through the forecast | Yours. Historical growth is in the statements and is the obvious starting point, but a rate you carry forward because it happened before is an assumption wearing a fact's clothes. |
| Length of the forecast | Yours, and it is a convention rather than a finding. Five and ten are both common. Lengthening it does not add information; it relabels part of the terminal assumption as a forecast. |
| Perpetual growth | Yours, and the most consequential number on the page relative to how little thought it usually gets. Anything above long-run nominal economic growth says this company eventually becomes the whole economy, which is a claim worth making deliberately if at all. |
| Discount rate | Yours. Usually a weighted average cost of capital, which is its own calculation with an unobservable equity risk premium inside it. This page will not fill it in, for the same reason the cost of capital page will not. |
| Diluted shares outstanding | The cover page of the annual report carries a current count, and the equity note carries the detail. Use the diluted figure, not the weighted average used for earnings per share: the weighted average answers a question about the year just gone, and this one is about who owns the future cash. |
| Net debt | Total interest-bearing debt less cash and equivalents, from the balance sheet, including the current portion of long-term debt. Whether operating lease liabilities belong in it is a judgement, and at a retailer or an airline it is a large one. |
Two cases where this page prints n/a instead of a number
The first is a discount rate at or below the perpetual growth rate. Gordon growth divides by the difference between them, so at equality it divides by zero and below it returns a negative terminal value. A business growing forever at least as fast as you discount is worth an unbounded amount, and there is no valuation to print. The tempting shortcut is to set the terminal value to zero and carry on, which returns a figure roughly a quarter the size of the right one and looks exactly like a small valuation. This page refuses and says why.
The second is free cash flow at or below zero. The model compounds whatever it is given, so a company burning cash produces a large negative number in the place a valuation goes. That is arithmetic operating correctly on an input the model was never meant to take. A business that does not yet generate cash may well be worth something; a two-stage discounted cash flow is not the instrument that finds out.
A missing share count produces n/a for the per-share line and leaves the enterprise figure standing, because the two are separate claims. Dividing by zero here would print an infinity, and an infinity next to a currency symbol reads as an answer.
The same engine, with the filings attached
The arithmetic above is imported from the product's own valuation engine rather than written again for this page, so the number here and the number a subscriber sees come from one implementation. What the workbench adds is the three filed inputs already loaded from ten years of accounts, a sensitivity grid across growth and discount rates rather than one cell of it, and every figure expanding into the statement, fiscal year and filing date it came from. It is a subscription and it wants a card for the trial, which is worth saying on a page that has just given you the model for nothing. What it costs.
- WACC calculator builds the discount rate this model is most sensitive to, from the five inputs it takes.
- The two numbers that move a DCF most are the two you cannot look up measures how far each assumption moves the answer, one percentage point at a time.
- Cost of capital is the course on the rate in the denominator: what it is doing, and where each part of it is read from.
What this leaves out
This is arithmetic on assumptions you supply. It is not a valuation of any company, not a view on any security, and not advice. A discounted cash flow computed correctly from poor assumptions is a poor answer arrived at carefully.
One growth rate for the whole explicit forecast is a simplification. Real forecasts fade, and a model that steps from your growth rate to the perpetual rate in a single year overstates value against one that fades gradually to the same place.
Gordon growth is one of two common terminal methods. The other applies a multiple to a final-year figure, which moves the assumption from a growth rate to a multiple without removing it. Neither is more objective than the other.
Free cash flow before any adjustment treats acquisitions, share-based payment and working capital swings as they fall. A year with a large working capital release looks like a better year than it was, and starting a five-year compound from it carries the error through the whole forecast.
The model values the whole business and subtracts net debt. It does not carry preference shares, convertibles, minority interests or pension deficits, and any of those can be large enough to change the per-share answer materially.
Nothing you type here is sent anywhere. The arithmetic runs in your browser and no input is stored, logged or transmitted.