Three methods on one asset produce three different answers, and the disagreement is the most useful output. Averaging it away destroys the only thing the exercise produced.
10 min read · Level 3
The instinct to reconcile three methods into one figure is not laziness - it comes from a real and reasonable belief that an asset has a value, and that three different answers therefore indicate two of them are wrong. That belief is worth examining, because the evidence points the other way: an asset's worth genuinely differs depending on who holds it, what they can do with it, how long they can wait and what else they could buy instead. A buyer who can fold the business into an existing operation is not mistaken to value it higher; a buyer who needs to sell in three years is not mistaken to value it lower. If that is right, then the spread between methods is not error to be eliminated but a description of a real property of the asset, and the person who collapses it into one number has substituted a tidier claim for a truer one. The practical consequence is that the output worth producing is never the value - it is your maximum price, which is a fact about you as much as about the asset, and is the only one of the two you are ever in a position to know.
Each method answers a different question, and knowing which question you are asking is most of the skill.
Comparables answer: what are people paying for things like this, right now. Discounted cash flow answers: what is this worth if my own view of its future is correct. Yield-based valuation answers: what return does this price buy me, expressed so I can compare it with everything else I could do. Asset or liquidation value answers: what is the floor if the business stops working.
None of them is the true method with the others as approximations. They are four different instruments, and the reason to run more than one is not to average the readings - it is to find out where they disagree, because a disagreement is a specific, answerable question about the asset.
Because agreement and disagreement carry completely different information and both are worth having. When a market price, your own forecast and the income yield all point to a similar number, the price is probably a fair reflection of ordinary expectations, and any case for acting has to rest on something none of the three methods captured. When they diverge sharply, one of them is carrying an assumption the others are not, and finding out which one that is tells you what the decision actually turns on.
And because the alternative - averaging - looks like rigour and is the opposite. An average of a market view, a private forecast and an income calculation is a number with no interpretation at all, and it conveniently sits between two figures that were telling you something.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
Harbour Press, valued three ways. The business: 2,000,000 of EBITDA, 1,600,000 of operating profit, 975,000 of net profit, 4,000,000 of net debt, 1,000,000 of free cash flow to the firm.
Comparables. Peers at 7.0 to 14.5 times EBITDA, median 10.25. Enterprise value from 14,000,000 to 29,000,000, median 20,500,000. Less 4,000,000 of net debt, the owners' share is 10,000,000 to 25,000,000, with a median of 16,500,000.
Discounted cash flow. The base case gives an enterprise value of 15,928,151, and the four corners of the sensitivity table run from 12,893,770 to 21,143,254. Less the same 4,000,000, the owners' share is 11,928,000 at the base case, in a band from 8,894,000 to 17,143,000.
So the two bands overlap between 10,000,000 and 17,143,000, which is the honest answer: a range whose top is 1.71 times its bottom. Anything outside that overlap is a figure only one method supports.
Now the disagreement, which is the interesting part. The comparables median of 16,500,000 is 38.3% above the discounted cash flow base case of 11,928,000. That is not a contradiction to be averaged away; it is a question with an answer. Look back at the sensitivity table: an enterprise value near 20,500,000 corresponds roughly to the corner using a 9% discount rate and 3.5% perpetual growth, which produced 21,143,254.
So the market's median multiple is, in effect, assuming Harbour Press grows at around 3.5% forever and that a 9% return is adequate compensation for owning it. Your own base case assumed 2.5% and 10%. The whole 38% gap is those two assumptions, and now it is a question you can actually hold a view on rather than a difference between two spreadsheets.
The third lens turns both numbers into something comparable with everything else you could buy. On net profit of 975,000, the comparables median of 16,500,000 is an earnings yield of 5.9% - a price-to-earnings ratio of 16.9. The discounted cash flow base case of 11,928,000 is an earnings yield of 8.2%, or 12.2 times earnings.
That is the form in which the decision is finally makeable. The question is no longer 'what is Harbour Press worth', which has no single answer, but 'is a 5.9% earnings yield on a business with these characteristics attractive against the alternatives available to me' - which is a question a person can answer, defend, and be judged on afterwards.
The first mistake is averaging the methods. Three answers averaged into one discards the disagreement, which was the only genuinely new information the exercise produced.
The second is running a second method only after the first gave an unwelcome answer. A method introduced to rescue a conclusion is not a cross-check; it is an appeal, and it will be selected for how favourable it is.
The third is treating the market's method as the right one because it is the market's. A price is the most recent transaction, not a verdict, and the entire case for doing this work is that occasionally the most recent transaction is wrong.
The fourth is skipping the floor. Asset and liquidation value are often ignored because they produce a low number, and a low number is exactly what makes them useful - they say what happens if the story does not work, which is the scenario every other method assumes away.
The fifth is finishing with a valuation rather than a price. The purpose of the work is a maximum you will pay and the reasoning behind it. A band with no decision attached to it is analysis that was never converted into anything.
Predictable cash and no peers: discounted cash flow, with comparables as a sanity check. Plenty of similar transactions: comparables first. Held for income: yield-based. Struggling or asset-heavy: the floor matters more than either. Write the choice and the reason before starting.
Never reduce any method to a point before comparing them, or the comparison happens between three numbers that have each already thrown away their uncertainty.
The region where the bands agree is the defensible range. A figure only one method reaches needs an explicit argument for why that method is the reliable one here.
Work out what growth rate and discount rate would make the discounted cash flow agree with the comparables. That converts a difference between two methods into a specific claim about the business, which is something you can accept or reject.
Express the price as a return - earnings yield, cash yield, whatever the asset produces - so it can be compared with every other use of the same money. A valuation that has not been converted into a return cannot be weighed against anything.
On any decision large enough to justify more than one method, and particularly where a single method has produced an answer far from the asking price - that is exactly when a second instrument is worth having.
Do not run three methods on a decision too small to warrant them, and do not let the existence of three answers become a reason to postpone. The work is meant to produce a maximum price, and a valuation that never reaches one has cost time and bought nothing.
A shape that appears often in investment discussions is a second valuation method appearing only after the first produced an inconvenient answer. The sequence is recognisable: a discounted cash flow suggests the price is high, comparables are then introduced and turn out to be more supportive, and the comparables become the headline. Nothing about the second method is illegitimate, and it may well be the better instrument for that asset. What makes the sequence unsound is that the choice was made after the answers were known, so the method was selected for its conclusion rather than for its fit.
Choose the methods and write down why before running any of them. A method chosen after the answers are visible is a vote, not a cross-check.
When comparables and a discounted cash flow produce similar answers it reads as confirmation, and it is often less than that. Analysts build forecasts in an environment where the market price is already known, and growth assumptions tend to settle near whatever makes the model roughly agree with it - not through any decision to do so, but because an output far from the price invites a re-examination of the inputs, and an output close to it does not. Two methods can therefore converge because one quietly used the other as a reference rather than because two independent routes arrived at the same place.
Agreement counts as evidence only if the second method could not see the first. Build the forecast before looking at the price, or record that you did not.
Create a free account to read the rest and apply it to your own investing.