Discounted Cash Flow Model 2026: Every Line Recalculated

Six discounted cash flow inputs, and only three are public

Start with the part that is public. The Federal Reserve publishes the risk free curve daily. On 29 September 2026 the ten year Treasury constant maturity stood at 5.26 percent and the ten year inflation indexed yield at 2.91 percent, checked 1 October 2026 at federalreserve.gov. The gap between them, 2.35 points, is the inflation the bond market expects over ten years.

The tax rate is public too. The IRS states that corporations "figure their tax by multiplying taxable income by 21% (0.21)", checked 1 October 2026 at irs.gov. That is the federal statutory rate. A company's effective rate differs because of state taxes and permanent differences, and choosing between the two is already a convention.

Three inputs are not public. The beta, the equity risk premium and the perpetuity growth rate are ours. A beta for a private company is levered up from a set of listed peers, so it inherits whatever peer set you chose, and that choice is an assumption too. We take a beta of 1.10, an equity risk premium of 4.60 percent, and a perpetuity growth rate of 2.35 percent, which is the breakeven inflation above. The reference dataset for implied premiums was last fully updated on 9 January 2026, checked 1 October 2026 at stern.nyu.edu, so we do not present a premium as a current market reading.

Assemble the rate. Cost of equity is 5.26 plus 1.10 times 4.60, which is 10.32 percent. Pre tax cost of debt is 6.40 percent, so after tax it is 5.056 percent. At 75 percent equity and 25 percent debt the weighted rate is 7.740 plus 1.264, which is 9.004 percent. We round it to 9.0 percent and say so.

Calder Instruments, five years of free cash flow

Calder Instruments makes industrial measuring equipment. Revenue last year was $840.0 million. The drivers below are all ours.

1. Revenue growth: 6.0, 5.0, 4.0, 3.5 and 3.0 percent.
2. EBIT margin: 14.0 percent, held flat.
3. Tax: 21 percent, the federal statutory rate.
4. Depreciation and amortization: 4.5 percent of revenue.
5. Capital expenditure: 5.5 percent of revenue.
6. Change in net working capital: 12 percent of the revenue increase.

Year one revenue is 840.0 times 1.06, which is $890.4 million. EBIT is 14 percent of that, which is $124.7 million. NOPAT is 124.7 times 0.79, which is $98.5 million. Add D&A of $40.1 million, take off capex of $49.0 million, and take off working capital of $6.0 million. Free cash flow is $83.5 million.

Read what the working capital line does. It shrinks every year, from $6.0 million to $3.6 million, because growth is slowing. A model that holds working capital at a percent of revenue instead of a percent of the increase would charge $124.4 million in year five, and the whole forecast would collapse. The unit matters more than the percentage.

Discounting the cash flow, and the 4.4 percent a convention is worth

Discount each year at 9.0 percent. Year one is 83.53 times 0.9174, which is $76.63 million. Year two is 88.71 times 0.8417, which is $74.67 million. Years three, four and five give $72.07 million, $68.83 million and $65.42 million. The explicit period is worth $357.6 million.

That used end of year discounting. Cash arrives on 31 December and nowhere else. It never does. The mid year convention discounts each year at t minus 0.5, which lifts the explicit period to $373.4 million.

Where the two conventions really separate is the terminal value. Discount the terminal value at 4.5 years as well and the enterprise value rises from $1,364.4 million to $1,424.5 million, a gain of 4.40 percent. Discount the cash flows at mid year but leave the terminal value at year five and it rises to $1,380.2 million, a gain of 1.15 percent. Three defensible treatments, three answers, and the spread between the outer two is $60 million.

We use end of year discounting throughout this page. It understates value on any real cash profile, and the understatement is known and small. Whichever you pick, say it out loud before anyone asks.

Terminal value is three quarters of a discounted cash flow answer

Gordon growth first. Year five free cash flow is $100.65 million, so year six is 100.65 times 1.0235, which is $103.02 million. Divide by the discount rate less growth, 0.09 less 0.0235, which is 0.0665. The terminal value is $1,549.1 million. Discount it five years at 0.6499 and it is worth $1,006.8 million today.

Enterprise value is 357.6 plus 1,006.8, which is $1,364.4 million. The terminal value is 1,006.8 over 1,364.4, or 73.8 percent of the total. Five years of careful forecasting decides a quarter of the answer.

Now the exit multiple. Year five EBITDA is EBIT of 145.1 plus D&A of 46.6, which is $191.8 million. At 9.0 times, the terminal value is $1,725.8 million, worth $1,121.7 million today, for an enterprise value of $1,479.3 million. That is 8.4 percent above the Gordon answer.

Here is the check almost nobody runs. Divide the Gordon terminal value by year five EBITDA: 1,549.1 over 191.8 is 8.08 times. So our Gordon assumption already embeds an exit at 8.08 times. Choosing 9.0 times instead is choosing multiple expansion, and a candidate who does that without naming it has smuggled in a market call.

From enterprise value to a price per share

Calder's last reported balance sheet, dated 30 June 2026, shows $340.0 million of debt and $95.0 million of cash. Net debt is $245.0 million. The group also consolidates a subsidiary it owns 70 percent of, carrying $40.0 million of non controlling interest.

Take the Gordon enterprise value of $1,364.4 million. Subtract net debt of $245.0 million and non controlling interest of $40.0 million. Equity value is $1,079.4 million.

The non controlling interest matters and gets dropped constantly. The EBITDA in the forecast is consolidated, so it includes all of that subsidiary's earnings. The enterprise value derived from it therefore covers the whole subsidiary. Leaving the minority out would hand $1,119.4 million to group shareholders instead of $1,079.4 million, which is 3.7 percent too much.

One more convention hides in the net debt line. We used the balance sheet as reported. A buyer would use net debt at the transaction date, which includes cash generated in the meantime. Calder throws off roughly $20 million a quarter, so a two quarter gap moves the equity value by about $40 million, which is about 65 cents a share.

Three share counts, three discounted cash flow answers

Calder has 62.0 million basic shares outstanding. It also has 5.0 million options struck at $10.00 and 1.8 million restricted share units with no exercise proceeds.

Basic first. 1,079.4 over 62.0 is $17.41 per share.

Now the treasury stock method. The options bring in 5.0 times $10.00, which is $50.0 million of proceeds. At $17.41 those proceeds buy back 2.872 million shares, so the net new shares are 2.128 million. Add the 1.8 million units and the diluted count is 65.928 million. Equity value per share is 1,079.4 over 65.928, which is $16.37.

But $16.37 is now the share price, so the buyback should have been priced there. Redo it at the lower price and the buyback shrinks, the count rises, and the price falls again. The loop settles at 65.754 million shares and $16.42 per share.

Three ways of handling the same options give $17.41, $16.37 and $16.42. The spread between the highest and lowest is 5.7 percent. None of them is wrong. Only one of them is the one you meant.

What the discounted cash flow sensitivity grid admits

Hold the forecast. Move the discount rate half a point either way and the growth rate half a point either way.

The grid runs from $14.82 to $20.90 on basic shares. That is a spread of 41 percent against the low end, produced by moving two assumptions by half a point each. Every number in the grid follows from the same forecast and the same arithmetic.

This is the most useful thing a sensitivity table does, and most candidates present it as reassurance. It is the opposite. It says the forecast barely matters next to two inputs nobody can observe. An interviewer who asks you to defend a DCF is asking you to defend those two numbers, and the honest answer names the range before it names a point.

Where a DCF answer loses its points

Our math rubric puts four weights on a calculation, and the weights are ours. No bank publishes a marking scheme, which is one of the few constants in this market. The reasoning behind each weight sits on how an answer is assessed.

Drilling this by hand

Rebuild Calder from the drivers without looking at the tables. Say the unit before every product.

1. Rebuild year one. Revenue, EBIT, NOPAT, then the three cash adjustments. You should reach $83.5 million.
2. Change one input. Take the EBIT margin to 15 percent and redo year five. EBIT becomes $155.5 million, NOPAT $122.8 million, and free cash flow $112.8 million.
3. Reprice the terminal value. With year five free cash flow at $112.8 million, Gordon at 2.35 percent gives $1,780.2 million, worth $1,157.0 million today.
4. Check the implied multiple. Year five EBITDA becomes $202.1 million, so the implied exit is 8.81 times. One margin point moved the implied exit by three quarters of a turn.
5. Redo the bridge. Subtract net debt and the minority, divide by each of the three share counts, and note which convention you used.

For the arithmetic drill that sits underneath all of this, the math practice set times mental products and percentages. For the growth arithmetic a terminal value rests on, the CAGR calculator will check your compounding. Neither knows what a terminal value is, and neither will tell you your growth rate is indefensible.

If you want the same discipline applied to a capital structure rather than a forecast, the LBO model explained runs a debt schedule line by line. For the reconciliation step that turns three valuation methods into one recommendation, see valuation interview questions.