The Senior Analyst level taught you valuation multiples: EV/EBITDA, P/FCF, PEG, return decomposition. But multiples are a shortcut - a DCF condensed into a single number. An Expert Analyst understands the model behind them: discounted cash flow. Not in order to model every stock down to the last line, but because whoever grasps the anatomy of a DCF sees what every valuation truly rests on - and can tell when the model is telling the truth and when it is merely repeating its author's assumptions.
The anatomy of a DCF
Firm value in a DCF has two parts. The first is the explicit period: usually five to ten years of projected free cash flows, year by year. The second is the terminal value - an estimate of the value of everything that comes after the explicit period, most often as a perpetuity growing at a constant rate g. Both parts are discounted to present value: a dollar five years from now is worth less than a dollar in hand today, and the discount rate handles the conversion.
The discounting itself is simple arithmetic: present value = the future flow divided by (1 + r) raised to the number of years, that is PV = CF / (1+r)^t. A hundred arriving in one year is worth 100 / 1.10 = roughly 90.9 today at a 10% required return; the same hundred five years out only 100 / 1.10^5 = roughly 62. The formula says what intuition says - the more distant the money and the higher the rate, the smaller today's value - only precisely.
An uncomfortable truth right at the start: the terminal value makes up most of the result in a typical model - commonly 60 to 80%. That means the value comes mainly from assumptions about the most distant, and therefore least certain, future. This is not a flaw of the model but its essence: the value of a quality company genuinely lies mostly in the years far ahead of us. For practice this implies humility - and the duty to give the terminal assumptions the greatest attention.
WACC: the price of capital
The discount rate of a firm-level DCF is the WACC - the weighted average cost of capital. It has two components. The cost of equity is the return shareholders demand for risk: the risk-free rate plus a risk premium. That equity costs more than debt is not a convention but the logic of the order of claims: the shareholder stands last in line - creditors always receive their interest and principal first, and shareholders own only whatever residue remains, if any. For a residual claim with an uncertain return they therefore rationally demand a higher return than a creditor with a contractually fixed interest rate. The cost of debt is the interest the company pays its creditors, after tax - interest is tax deductible, which makes debt cheaper. The two components are weighted by the shares of equity and debt in the company's financing. Debt is cheaper than equity, but too much of it raises the risk of both components at once - so the WACC cannot be pushed down simply by loading the company with debt.
The terminal value is computed with the perpetuity formula: TV = FCF x (1 + g) / (r - g) - the last projected free cash flow stepped up by one year of growth, divided by the difference between the discount rate and growth. It is precisely the r - g denominator that produces the iron rule: g must be smaller than r. A company cannot forever grow faster than the discount rate - mathematically the value would come out infinite, economically the company would one day outgrow the entire economy. Perpetuity g therefore stays near the economy's long-run growth, typically in the low single digits of percent.
Sensitivity: why the output is a range
A DCF is extremely sensitive to its inputs. Move the WACC from 9 to 8% and terminal growth from 2 to 3% - and fair value can easily jump by tens of percent without anything happening at the company. It is precisely this sensitivity that makes the DCF an easily abused tool: for any target price there exist inputs that will "prove" it. The sensitivity mechanism sits in the r - g denominator: the terminal value scales as 1 / (r - g). Narrow the gap from 7 to 5 percentage points - say by cutting the WACC one point and raising g one point - and the terminal value jumps by roughly 40% (1/0.05 versus 1/0.07) without anything changing at the company. An expert therefore works with a sensitivity table - a matrix of values for combinations of WACC and g - and reads the result as a range, never as a point. A single-figure target price from a DCF is an illusion of precision; the honest output reads "80 to 110 under reasonable assumptions".
From EV to the share price: the bridge
A DCF discounting the cash flow of the whole firm returns enterprise value - the value of the business for all providers of capital. Shareholders, however, own only what remains after the creditors. The conversion is handled by the EV bridge: from enterprise value subtract net debt (debt minus cash) and the value of minority interests, and the result is the value of equity; minority interests come out because consolidated statements always count subsidiaries at 100% even when the parent owns, say, only 80% of them - the outside shareholders' slice of those subsidiaries has to be handed back; divide by the fully diluted share count and you have value per share. Two companies with the same EV and different debt thus have very different share values - the bridge is exactly the place where leverage writes itself into the price.
When DCF, when multiples - and what to do with cyclicals
The DCF excels at companies with predictable flows: infrastructure, consumer staples, established software. It fails where cash flow cannot be predicted or predicting it makes no sense: banks (capital and credit creation do not fit into classic FCF - they are valued via P/B and returns on capital) and cyclicals, whose flows swing with the cycle. Cyclical companies also carry the extrapolation trap: a DCF built on peak-cycle earnings comes out beautifully and at exactly the wrong moment. The remedy is normalization: model the average profit across the whole cycle, not the latest record year. Multiples in turn win on speed and market anchoring - and the combination works best: a DCF to understand the value, multiples as a check against reality.
For conglomerates, sum-of-the-parts comes in handy: value each segment separately with the method that fits it and add up the parts, minus debt. The sum typically comes out above the market capitalization - the market persistently prices conglomerates at a conglomerate discount for complexity and worse capital allocation across unrelated businesses. The gap alone is therefore not an arbitrage: the discount can persist for years, and value usually gets unlocked only by a catalyst, typically a break-up or the sale of a division. And read multiples with a double view: the implied multiple (what P/E falls out of your DCF) against the historical one (what the company traded at in the past) - a large gap means that either your model or the market believes something unusual. Finding out what is exactly the work that separates valuation from plugging numbers into formulas.
Reverse DCF: what is in the price
The most useful trick of expert valuation turns the whole procedure around. Instead of "I estimate growth and compute the value", a reverse DCF takes today's market price as given and backs out what growth and margins justify it. The result is a sentence like: "today's price assumes 15% FCF growth for ten years". The question is different from the usual one - not "what should the stock be worth" but "is the story built into the price believable?" - and it tends to be much easier to answer. For individual stocks you can compare fair value based on fundamentals with the Fair Price Index on Bulios; the reverse logic adds to it what the market has already promised in the current price.
Test yourself
The DCF and its inputs form the largest area of the Expert Analyst exam - the third of four levels of the Bulios certification. If you can explain why the terminal value dominates the result, what exactly the EV bridge does and what question a reverse DCF answers, you are ready for the area.