Trading Multiples: Every Multiple and What Each One Assumes

A trading multiple divides a traded value by a measure of what produces it, and the two halves must belong to the same claimants. Enterprise value pairs with earnings before interest, tax, depreciation and amortisation (EBITDA), with earnings before interest and tax (EBIT) and with sales; a share price pairs with earnings, book value and cash flow for each share. Sankalp Industrial Systems Limited, invented, trades at 7.78 times EBITDA, 1.87 times sales, 13.04 times earnings and 2.00 times book. Each answers a different question.

Almost everybody has either bought a flat or listened to somebody argue about one. Start there. Suppose a flat changed hands at Rs 60,00,000. The buyer put in Rs 15,00,000 of her own money and a bank put in the other Rs 45,00,000. She lets the flat out. The rent brings in Rs 3,00,000 a year before she pays the bank anything, and after the instalment there is Rs 1,20,000 left in her hand.

There are two honest answers to what she paid for that income, and one dishonest one. Rs 60,00,000 divided by Rs 3,00,000 is 20 times: the whole price of the flat against the whole of what the flat earns, before anybody has been paid out of it. Rs 15,00,000 divided by Rs 1,20,000 is 12.5 times: her own money against what is left for her after the bank has taken its share. Both of those are real numbers. Rs 60,00,000 divided by Rs 1,20,000 is 50 times. The top half of that third figure includes the bank's money and the bottom half has already paid the bank. So the 50 times is not a worse ratio or a more conservative one. It is not a ratio of anything at all. The loan has been counted on one side and removed on the other.

A listed company is that same flat, dressed in a different vocabulary. There are six multiples in common use, they divide into two shapes, and the whole of the discipline is knowing which shape is in hand.

What is a trading multiple, and what are its two halves?

A trading multipleA ratio of a traded value to a measure of what produces it. is a ratio of what a company currently trades at to some measure of what the company produces. The word trading is doing real work in that sentence: the value on top is the value at which shares are changing hands right now, in ordinary volumes, between buyers and sellers who are taking no control of anything. The value on top is not a price somebody paid to acquire a whole business, and that difference is set out under what a trading multiple is not.

The top half is the numeratorThe value half of the multiple, either an enterprise value or a share price. and there are only two candidates for it in practice. One is the enterprise value, being the value of the operating business as a whole, on which every provider of capital has a claim. The other is the market value of the equity, expressed either as the total or as one share's price. For Sankalp Industrial Systems Limited, an invented listed manufacturer of industrial valves, precision castings and aftermarket parts and service, the two are Rs 22,40,00,00,000 and Rs 18,00,00,00,000, the second of which is 20,00,00,000 shares at Rs 90.00 each.

The bottom half is the denominatorThe measure the value is divided by, such as EBITDA, sales or earnings. and the candidates for it are many rather than two. Revenue, EBITDA, EBIT, profit attributable to owners, book value, cash flow. Each denominator is a different point on the way down a set of accounts, and where the descent stops decides what the multiple is measuring. Choosing a multiple is not choosing a formula; it is choosing how far down the accounts to travel before stopping and dividing.

What is the matching rule, and whose money sits in each half?

Here is the rule, and it is one sentence. The claimants named in the numerator must not already have been paid inside the denominator. That is the whole of it, and the flat above is the proof that the rule was already understood before anybody gave it a name.

Work it through on the two numerators. An enterprise value is the value of the operating business before anybody's claim on it has been settled. The lenders have a claim on it, worth Rs 6,00,00,000 for every crore of gross debt and Rs 6,00,00,00,000 in total for this company. The minority holder in Sankalp Coatings Private Limited, the invented subsidiary that is 75.0 per cent held and fully consolidated, has a claim worth Rs 60,00,00,000. The owners have the rest. Every one of those claimants is inside the top half. The bottom half therefore has to be a measure taken before any of them has been paid, and that means before interest and before tax. Revenue qualifies. EBITDA qualifies. Depreciation is not a payment to anybody in the queue, so EBIT qualifies.

A share price is narrower. A share price is the value of one claim only, the owners', on what is left after everybody else has been dealt with. So the denominator underneath it must also be a figure from which the lenders and the tax authority and the minority holder have already been removed. Earnings for each share qualifies. Book value of the equity for each share qualifies. Cash flow available to the owners for each share qualifies. Revenue does not, EBITDA does not, and EBIT does not.

One question is the entire matching ruleThe requirement that the numerator's claimants and the denominator's timing agree.. Under time pressure nobody consults a principle, so the rule is better said out loud than nodded at: has the denominator already paid the people who are in the numerator? If the answer is yes, there is no multiple.

Where does each multiple stop on the way down the accounts?

The cleanest way to see all six at once is to write the company's Year 0 income statement out as a ladder and mark where each multiple gets off. Two of the rows are worth pausing on: the interest charge of Rs 48,00,00,000 is the blended 8.00 per cent contracted rate on Rs 6,00,00,00,000 of gross debt across three tranches, and the tax charge is struck at a 25.0 per cent effective rate which is this company's own assumption rather than any country's statutory figure.

The ladder reconciles in both directions, and it is worth checking rather than trusting. EBITDA of Rs 2,88,00,00,000 less depreciation and amortisation of Rs 48,00,00,000 gives EBIT of Rs 2,40,00,00,000. Taking off Rs 48,00,00,000 of interest leaves Rs 1,92,00,00,000 before tax. The same Rs 1,92,00,00,000 makes interest cover exactly 5.00 times. A quarter of that in tax is Rs 48,00,00,000, leaving Rs 1,44,00,00,000. The record states that Rs 1,38,00,00,000 is attributable to the owners, so the Rs 6,00,00,000 difference is the minority holder's share of the subsidiary's profit, and it falls out of the arithmetic rather than being asserted anywhere.

When is Enterprise Value to EBITDA the right measure, and when is it not?

Enterprise value to EBITDA is the default in an industrial segment, and a reason rather than a habit sits behind it. EBITDA stops above depreciation, and depreciation is the line most contaminated by choice. Two companies can buy the identical machine on the identical day and charge it off over different lives under different methods, and the profit they report will differ without a rupee of difference in what either of them earned or spent. Stopping above that line removes the contamination.

Two identical tea stalls stand on the same road. One rents its cart by the month and one has bought a cart outright. The cash each stall clears after every cash cost is comparable. Only one of them has a cart to write down and the other has a rent line instead, so what each reports after writing down the cart is not comparable. EBITDA is the measure that puts the two stalls back on the same footing, and the price of that is that it pretends carts are free.

For the subject company the calculation is a single division. Traded enterprise value Rs 22,40,00,00,000 divided by Year 0 EBITDA of Rs 2,88,00,00,000 gives 7.78 times. The median of the six invented peers is 7.8 times, so the company trades close to the middle of that invented set on this one measure. The arithmetic supports nothing beyond that. Nothing about being close to a median establishes whether the peers were correctly priced, or whether this business is like them in the ways that matter.

One assumption inside the multiple never announces itself, and it is the part worth naming. The measure removes depreciation from both sides of the comparison and puts nothing back, so using enterprise value to EBITDA assumes that the capital the business consumes to stay alive is broadly similar to the capital the peers consume. Where that is not true, the multiple flatters whichever company eats the most capital. Sankalp Industrial Systems Limited's forecast has capital expenditure running at Rs 1,34,80,00,000 in Year 1 against depreciation of Rs 52,80,00,000. The business is investing well above the rate at which its assets are wearing out. An EBITDA multiple cannot see any of that.

What does one line lower change, and why does enterprise value to EBIT exist?

Enterprise value to EBIT exists precisely for a business that eats capital faster than its peers do. EBITEarnings before interest and tax, so after depreciation and amortisation. is EBITDA after depreciation and amortisation, and it is still above the interest line, so it keeps the enterprise value numerator legitimate while putting the cost of the asset base back into the comparison. For the subject: Rs 22,40,00,00,000 over Rs 2,40,00,00,000 is 9.33 times.

The relationship between the two multiples is not a coincidence and it is worth carrying. Depreciation and amortisation is Rs 48,00,00,000 against EBITDA of Rs 2,88,00,00,000, or 16.67 per cent, so 83.33 per cent of EBITDA survives to become EBIT. The EBITDA multiple divided by that survival share gives the EBIT multiple exactly: 7.7778 divided by 0.8333 is 9.3333. The EBIT multiple is the EBITDA multiple divided by the share of EBITDA that survives depreciation. The gap between the two is therefore a direct reading of how capital hungry the business is.

Now see what that does to a comparison. Imagine a business trading at the identical 7.78 times EBITDA whose depreciation runs at 30.0 per cent of EBITDA rather than 16.67 per cent. On the same Rs 2,88,00,00,000 of EBITDA its depreciation would be Rs 86,40,00,000 and its EBIT Rs 2,01,60,00,000, so its enterprise value to EBIT would be 11.11 times against the subject's 9.33. The two companies looked identical on one multiple and are 1.78 turns apart on the other, and no valuation opinion changed between the two lines. Only the asset base did.

Why does Enterprise Value to Sales exist at all, and what is really inside it?

Revenue is the top line, and stopping there feels like giving up. Revenue is the crudest place on the ladder and it takes no account of whether the revenue is profitable. It is also the only denominator that still exists for a company losing money, and that is the honest reason it is used. But the interesting thing about a sales multiple is not why people reach for it. The interesting thing is what it is made of.

There is an identity underneath it, and it is exact rather than approximate. Enterprise value to sales equals enterprise value to EBITDA multiplied by the EBITDA margin. The identity falls straight out of the algebra: divide the enterprise value by EBITDA and then multiply by EBITDA over revenue, and EBITDA cancels. The derivation is trivial and absorbing it is not. A sales multiple is not a cruder cousin of an EBITDA multiple. A sales multiple is an EBITDA multiple with a margin assumption multiplied into it, and every use of one imports somebody else's margin whether that was intended or not.

Check it across the whole invented peer set rather than taking it on trust. Peer 1, Aravalli Flow Controls Limited, trades at 6.6 times EBITDA on a 20.0 per cent margin, and 6.6 times 0.200 is 1.3200. Its enterprise value of Rs 11,88,00,00,000 over revenue of Rs 9,00,00,00,000 produces exactly the 1.32 times sales printed for it. Peer 2, Satpura Engineering Works Limited: 7.1 times 0.210 is 1.4910, printed as 1.49. Peer 3, Kaimur Industrial Limited: 7.6 times 0.230 is 1.7480, printed as 1.75. Peer 4, Girnar Precision Limited: 8.0 times 0.245 is 1.9600. Peer 5, Shivalik Systems Limited: 8.7 times 0.260 is 2.2620, printed as 2.26. Peer 6, Nallamala Components Limited: 13.8 times 0.300 is 4.1400. And the subject: 7.7778 times 0.240 is 1.8667, printed as 1.87. Seven cases, seven exact agreements on the unrounded arithmetic, with three of the printed figures rounded to two decimals in the usual way.

One consequence of the identity is large enough to work through. Holding the EBITDA multiple absolutely still, at the peer median of 7.8 times, and letting only the margin move across the range the six invented peers actually occupy, from 20.0 to 30.0 per cent, the sales multiple runs from 1.56 times to 2.34 times. The spread is 50 per cent, produced by nothing except the margin, on companies the market is pricing identically on EBITDA. A sales multiple that varies by half again across a peer set may say nothing about value and everything about the mix of businesses inside those companies.

What does the Price-to-Earnings Ratio carry that an enterprise multiple does not?

The price to earnings ratio is the multiple everybody has met, usually before they met any of the others, and the familiarity hides how much is loaded into it. The numerator is the share price, Rs 90.00. The denominator is earnings for each share, Rs 6.90, being Rs 1,38,00,00,000 of profit attributable to owners across 20,00,00,000 shares. The ratio is 13.04 times.

Look back at the ladder and count what happened between EBIT and that Rs 1,38,00,00,000. Three things left, and each of them is a decision or a circumstance rather than a fact about the operating business. A price to earnings ratio carries the capital structure through the interest charge, the tax position through the tax charge, and the ownership structure through the minority's share, and an enterprise value to EBITDA multiple carries none of the three.

The interest charge of Rs 48,00,00,000 exists because this company has chosen to fund itself with Rs 6,00,00,00,000 of debt at a blended 8.00 per cent across three contracted tranches. A company with the identical operations and no debt would report higher earnings and, at the same share price, a lower price to earnings ratio, without one machine or one customer being different. The tax charge of Rs 48,00,00,000 reflects a 25.0 per cent effective rate that is this company's own assumption; a company with a different mix of locations or reliefs would show a different figure. The minority's Rs 6,00,00,000 exists only because Sankalp Coatings Private Limited is 75.0 per cent held rather than wholly held.

There is a fourth item, and it sits in the numerator rather than the denominator. The market capitalisation of Rs 18,00,00,00,000 includes the value of Rs 1,20,00,00,000 of cash and Rs 1,00,00,00,000 of non-operating assets, being a surplus land parcel and a 26.0 per cent holding in Aruna Tooling Private Limited that is equity accounted. Neither produces a rupee of the EBITDA in the denominator of the enterprise multiple, and both are inside the price of the share. So the price to earnings ratio is quietly pricing assets whose earnings are largely not in its own denominator.

When does the Price-to-Book Ratio mean something, and when does it mean almost nothing?

Price to book is the one multiple in this guide that never touches the income statement. The denominator is the book value of equityWhat the accounts say the owners' stake is worth, here Rs 9,00,00,00,000., and for this company it is Rs 9,00,00,00,000, being Rs 45.00 for each of the 20,00,00,000 shares. At Rs 90.00 the ratio is exactly 2.00 times.

Ask what that denominator is actually counting and the answer is uncomfortable: broadly, what was paid for the assets, less what has been written off them since, plus profits that were retained rather than paid out. Book value is a historic record, not a current valuation. On this company the surplus land parcel sits in the accounts at a historical cost of Rs 12,00,00,000 while its current value is stated at Rs 45,00,00,000, so a single asset is understated by Rs 33,00,00,000 in the very denominator the ratio divides by. The full exercise of marking a balance sheet to current values is covered under asset revaluation.

So when does the ratio mean something? Price to book carries information in proportion to how much of what the business earns is actually recorded on its balance sheet. The proportion is why the ratio is a workhorse for lenders and financial businesses and close to useless for businesses whose earning power is not an asset anybody bought. A bank's assets are loans and its liabilities are deposits, and both are recorded at amounts closely related to what they are worth, so book value is a real measure of something. Now take the aftermarket parts and service business inside Sankalp Industrial Systems Limited. The business turns Rs 1,80,00,00,000 of revenue at a 30.0 per cent margin. An installed base of valves in other people's plants and a set of customer qualifications are what produce that margin. Nobody bought either, so neither is on the balance sheet.

Here is the household version. Two neighbours each have a small business. One runs a rental shop full of equipment that cost Rs 40,00,000 and is worth roughly that. The other is a tailor whose entire business is a sewing machine worth Rs 30,000 and thirty years of people trusting her with wedding clothes. Divide what each business would fetch by what its equipment cost. The first ratio means something real. The second is a division by a number that has almost nothing to do with why the business earns.

What does the Price-to-Cash-Flow Ratio fix about earnings, and what does it leave alone?

A price to cash flowA share price against a cash flow measure for each share rather than earnings. ratio replaces the earnings denominator with a cash flow one, and the fix it is aiming at is real. Earnings carry a depreciation charge that is a policy decision, and they carry every other accrual judgement made in preparing the accounts. Cash does not care about any of that.

One version matches the numerator properly. A share price sits above free cash flow to equity: the cash left for the owners after the lenders have been paid and after the business has funded its own reinvestment and borrowing. For Sankalp Industrial Systems Limited the Year 1 forecast figure is Rs 87,00,00,000, being Rs 4.35 for each share. Rs 90.00 over Rs 4.35 is 20.69 times.

Now compare 20.69 times with the 13.04 times on earnings and ask why they are so far apart. The answer is the whole of what a cash flow multiple does and does not fix. Earnings for each share are Rs 6.90 and cash flow for each share is Rs 4.35, a difference of Rs 2.55, and it is not an accounting artefact. The gap is reinvestment. The Year 1 free cash flow to equity of Rs 87,00,00,000 is struck after Rs 1,34,80,00,000 of capital expenditure and Rs 18,00,00,000 of increase in working capital, and after adding back Rs 25,00,00,000 of new borrowing. A cash flow multiple removes the depreciation policy from the comparison and replaces it with the company's actual spending plan. A company investing heavily for growth therefore looks worse on it than a company harvesting a mature business, and neither position is a valuation opinion.

The trade is plain. Earnings hide the spending behind an averaged charge; cash flow shows the spending in the year it happens and is therefore far lumpier. Neither is more honest than the other, and a table using one of them without saying which has left the reader unable to tell.

How Multiples Reflect Growth, Margins and Risk: can the three be read back out?

Growth, margins and risk cannot be read back out of a finished multiple. The three can be watched going in, and the six invented peers show it.

Aswath Damodaran makes the underlying argument better than anybody and it is his: a multiple is a compressed valuation, so everything a full model would argue about line by line is still present inside the single number, folded flat. Koller, Goedhart and Wessels put the same thing in a different frame, expressing value in terms of growth and the return on invested capital together, and the frame stops growth being treated as a good thing on its own.

Take the six peers and rank them by their EBITDA multiple, then rank them by margin, then by forecast revenue growth, then by return on invested capital. The order is identical in all four cases, and it is exactly identical rather than roughly so. Peer 1 is lowest on all four, peer 6 highest on all four, and the four peers in between hold their positions row for row. The identical ordering is not a discovery about the world; it is a property of an invented set. Drawing the ordering is still worth doing, because it shows the shape of the claim: growth, margin and the return on capital are not factors sitting outside a multiple to be considered separately, they are what the level of the multiple is made of.

The fourth dimension, leverage, is the one that behaves differently and it is worth being honest about it. Net debt to EBITDA across the six runs 1.8, 2.2, 1.5, 1.2 and 0.6 times, with peer 6 carrying net cash of 0.4 times EBITDA. Broadly, the higher multiples sit with the lower leverage, but the order is not exact: peer 2 carries more leverage than peer 1 and trades at a higher multiple. Three ladders line up perfectly and the fourth does not. Noticing that is more useful than four ladders that all agreed.

Which multiple should be reached for, and what decides it?

The choice is not a matter of taste and it is not settled by what the last analyst on the desk used. Each branch is a statement about what the denominator can carry, and the branches can be named.

Where the companies being compared charge depreciation on policies that cannot be reconciled, the comparison stops above depreciation and uses EBITDA. Where the amount of capital each business consumes differs materially, it goes one line lower to EBIT so the asset base is back inside the comparison. Where the profit is negative, or where the argument actually being had is about what margin the business can reach, sales is the denominator, and the imported margin has to be stated out loud. Where the balance sheet is more or less the business, book value serves. And where the capital structure, the tax position and the ownership structure genuinely belong inside the number, earnings serves, but then all three come with it rather than only the one that was in mind.

Why does the same company land in two different places on two multiples?

Run the subject against the identical six peers on two multiples and watch what happens. On enterprise value to EBITDA it is at 7.78 times against a peer median of 7.80. On price to earnings it is at 13.04 times against peer price to earnings ratios of 11.0, 12.5, 14.0, 15.0, 17.0 and 28.0 times, whose median is 14.5, being the average of the third and fourth of six.

Nothing about the company changed between those two lines. The same day, the same share price of Rs 90.00, the same six peers. The change is in which lines of the accounts each multiple includes, and those lines are all named above: the interest charge of Rs 48,00,00,000, the tax charge of Rs 48,00,00,000, the minority's Rs 6,00,00,000, and the cash and non-operating assets sitting inside the market capitalisation. Two multiples including different lines of the accounts will place the same company differently, and the difference is a mechanical consequence of what each one includes rather than a message about the company.

The strongest pull anywhere in this arithmetic appears exactly here, and the pull is worth naming. A number below a median invites one word and a number above it invites another, and neither word is available from this arithmetic. A company at a lower multiple than its peers is either priced for something the multiple does not capture or it is not, and the multiple by itself cannot separate those two situations. A multiple is a question rather than an answer, and its whole usefulness is in indicating which question to ask next.

What is missing from a multiple when nobody states the period?

Everything, very nearly. A multiple is a value divided by a measure covering a stated period, and a figure with no period attached has not finished being a number. There are two conventions and they are not close together. A trailingComputed on the last completed period rather than on a forecast one. multiple uses the last completed period, so it is built entirely from figures that have been reported. A forwardComputed on a forecast period, which must be stated wherever it is used. multiple uses a forecast period, so it is built from somebody's estimate and inherits every assumption in that estimate.

Watch the size of it on this company. The traded enterprise value is Rs 22,40,00,00,000 either way. Year 0 EBITDA, the last completed year, is Rs 2,88,00,00,000, giving 7.78 times trailing. Year 1 forecast EBITDA is Rs 3,16,80,00,000, giving 7.07 times forward. The same company, at the same price, on the same day, is 0.71 of a turn apart depending only on which period the denominator covers. And the gap is not random: the forward figure is exactly 9.09 per cent lower, because EBITDA is forecast to grow 10.00 per cent and one divided by 1.10 is 0.9091.

The consequence for a comparison table is severe and mostly invisible. A growing company always looks lower on a forward multiple than on a trailing one, and the faster it grows the bigger the drop. So a table that mixes the two produces a ranking in which part of the ordering is an artefact of which column each row was pulled from, and nothing in the table shows it.

What is a trading multiple not?

A trading multiple is not a price somebody paid for a whole company. It is the level at which a small parcel of shares changes hands between a buyer and a seller, neither of whom is taking control of anything or planning to change how the business is run. A multiple derived from a completed acquisition is a different object: it is a price a specific buyer paid, on a specific day, to take control of the whole thing, usually for reasons of their own.

Mixing the two inside one statistic produces a figure that answers neither question, and no amount of care with the arithmetic afterwards repairs it. Deal multiples are a separate set of numbers with a separate meaning, and they are covered separately.