Market Capitalisation and Free Float: Work It Out

Sarvani Coatings Limited, an invented issuer, supplies every share count, price and percentage worked here. The price carries a stamp of 28 August 2026 and the shareholding pattern carries an earlier quarter end, and that gap matters as much as the multiplication does. The arithmetic is exact. The numbers are assumed rather than read off a market.

Two multiplications, done in whole rupees. Every share in issue, multiplied by the price, gives market capitalisation. Only the shares that are not closely held, being the total less whatever the controlling shareholders sit on, multiplied by that same price, gives the free float version. Both are arithmetic, and neither output means anything until it is scaled against something.

Those two multiplications are the whole of the mechanical part. The rest is three groups of inputs, where each one is physically found, the multiplications themselves, the checks that prove the proportion was applied to the right base, and the step almost every calculator leaves out. That step is what to do when a computed answer disagrees with a figure somebody has already published. An analyst spends more working time on that last step than on the multiplication.

Two things carry over from earlier reading and are assumed here. From the listings material, that a corporate action changes the share count on a stated date and that every per share figure must be restated across it. From the treatment of market prices, that a price is not a fact about a company but a fact about a moment, so it travels with its timestamp or it travels broken.

Where is the share count found?

The share count is the least glamorous input and the one that goes wrong most often. A count is the number of equity shares actually in issue and fully paidShares on which the holder has paid the issuer everything that was called for. The holder of a partly paid share still owes money on it, so partly paid shares are counted separately., as at a stated date, and it lives in two places that should agree with each other.

Put in terms that can be felt. When four cousins split a wedding bill, the first thing settled is how many people are actually splitting it, and that is settled before anybody divides anything. Nobody argues about the per head figure until the head count is agreed. A share count is that head count, and a capitalisation computed on a stale one is the argument that comes afterwards.

Which price, and as at when?

Where is the shareholding pattern found, and what date does it carry?

The third input is a proportion rather than a quantity, and it comes from the periodic shareholding pattern filing that each exchange carries. The filing splits the register into categories: the promoter groupThe controlling shareholders and the persons and entities the rules group with them. Who belongs in it is set by disclosure rules, not chosen by the issuer., then the institutional and retail categories that make up everything else.

Sarvani Coatings has nothing on its locked-in row, so the row sits at zero here and the arithmetic below is unaffected. Plenty of registers do carry one, so the field stays on the calculator. Shares that cannot be sold are not free float, however they came to be locked. The free float proportion can then be reached two ways from this one table: add the three public categories, 18.2 plus 14.6 plus 14.8, for 47.6, or take 100.0 less the 52.4 closely held, also 47.6. When those two routes stop agreeing, a category has been dropped or counted twice, and the table has said so before the multiplication could.

How are the two figures computed?

Two multiplications, in this order, with one check between them. A rounded intermediate figure is the fastest way to make a check fail that should have passed, so the computation runs in whole rupees rather than in crore and converts to crore only when the answer is written down.

1. Multiply the share count by the price. 24,00,00,000 shares multiplied by Rs 486/- gives Rs 1,16,64,00,00,000/-, which is Rs 11,664 crore. That is market capitalisation, as at 28 August 2026 and no other date.
   Check: the count and the price both belong to 28 August 2026.
2. Take the free float proportion from the pattern. 100.0 less the 52.4 per cent closely held gives 47.6 per cent, which is also what the three remaining categories add to.
   Check: 18.2 plus 14.6 plus 14.8 equals 47.6, and the pattern totals 100.0.
3. Compute the free float capitalisation by the share route. 47.6 per cent of 24,00,00,000 is 11,42,40,000 shares, and multiplied by Rs 486/- that is Rs 55,52,06,40,000/-, or Rs 5,552.064 crore.
   Check: 11,42,40,000 plus the 12,57,60,000 closely held shares returns 24,00,00,000.
4. Compute it again by the proportion route, and make the two agree. Rs 1,16,64,00,00,000/- multiplied by 0.476 is Rs 55,52,06,40,000/-. The two routes land on the same rupee, and that agreement is the only evidence available that the proportion was applied to the right base.
   Check: the difference between the two routes is Rs 0/-.
5. Confirm the parts close on the whole. The closely held portion is 12,57,60,000 shares at Rs 486/-, or Rs 61,11,93,60,000/-, being Rs 6,111.936 crore. Added to Rs 5,552.064 crore it returns Rs 11,664 crore exactly.
   Check: nothing has leaked, and no share has been counted twice.

The commonest slip here is multiplying by the closely held proportion instead of the free float one. Step four is the one people skip, and step four is the only one that catches that slip. The wrong proportion returns Rs 6,111.936 crore, a figure that looks entirely respectable and is the wrong half of the register. Two routes that agree cannot both be applying the wrong proportion, and the second route earns its keep on that alone.

The worked instance, in full

Written up for somebody else to use, that last block becomes one sentence: free float market capitalisation of Rs 5,552.064 crore, on a price as at 28 August 2026 and a shareholding pattern as at 30 June 2026. Most published versions round it to Rs 5,552 crore. Rounding at the end is fine. Rounding in the middle is not.

How is a disagreement with a published figure resolved?

A computation returns Rs 11,664 crore, a screen somewhere shows Rs 5,832 crore, and the natural assumption is an arithmetic mistake. The assumption is almost certainly wrong. There are only three real causes, and they are all about dating rather than about multiplication.

The first is a different price date, and the figure moves by whatever the price moved. The second is a different share count, usually because one side is standing before a corporate action and the other after it. The third is a shareholding pattern from a different filing. A different filing changes only the free float figure and leaves the full capitalisation untouched, and that asymmetry is itself a clue about which of the three is at work. A disagreement between a computed capitalisation and a published one is a dating problem far more often than an arithmetic one.

Here is the demonstration. Suppose the published figure was computed before Sarvani Coatings made a bonus issueA corporate action in which existing holders receive additional shares without paying for them. The share count rises and nothing about the business changes, so per share figures must be restated across it., when the count stood at 12,00,00,000 rather than 24,00,00,000. Run the same multiplication on the old count and Rs 486/- gives Rs 58,32,00,00,000/-, or Rs 5,832 crore. The free float version falls the same way, from Rs 5,552.064 crore to Rs 27,76,03,20,000/-, or Rs 2,776.032 crore. Nothing about the company differed between the two computations. Only the record dateThe cut-off date an issuer uses to decide which holders are entitled to a corporate action. The record date is the day on which a share count stops being one number and starts being another. the count was taken from did.

So the sequence when the numbers disagree is fixed, and it starts nowhere near the price. First the price date is confirmed on both sides. Then the share count on both sides, along with whether a corporate action sits between them. Then which shareholding pattern filing each side used. Only when all three match on both sides is a residual difference worth calling an arithmetic error, and by then it has usually been found already.

Who actually runs this calculation, and what for?

A version that can be pictured. A housing society has a hundred flats, and the builder holds fifty two of them and has said publicly that they are not for sale. Valuing the whole society means counting a hundred flats. Asking how busy the resale market in that society will be this year means counting forty eight. Same building, same day, two different denominators, and using the wrong one does not produce a slightly wrong answer, it produces an answer to a different question.

Meghna Iyer, an analyst covering coatings, keeps both figures in the same row of her sheet with both dates beside them, and she recomputes rather than lifting a published number, for one reason: a published figure does not tell her which pattern filing it used. Direct computation does not buy accuracy, which the published figure usually has. Direct computation buys certainty about which dates the number stands on.

Two more uses. The same day produces two very different-looking ratios, so a research analyst scaling a day's traded value picks one of the two figures as the denominator and has to say which. A treasury team sizing a placement wants to know how much stock could realistically be available, and availability is a free float question rather than a capitalisation question. Neither use needs an interpretation of the answer. Both need to know what the answer is made of.

Why can two issuers of the same size have very different amounts to buy?

Hold the share count and the price still and change only the pattern, and the two outputs come apart. Nandivarman Paints Limited also has 24,00,00,000 shares in issue and is also illustrated at Rs 486/- as at 28 August 2026, so its market capitalisation is Rs 11,664 crore, the same figure as Sarvani Coatings to the rupee. Its promoter and locked-in rows come to 88.0 per cent of the register rather than 52.4, leaving 2,88,00,000 shares in public hands instead of 11,42,40,000. Multiply those by the same Rs 486/- and the free float capitalisation is Rs 1,399.68 crore against Sarvani's Rs 5,552.064 crore. Two issuers can be identical on the headline and differ by a little under four times on the part anybody can actually buy.

The count and the price are held identical. The pattern is then the only thing left to explain the gap. The calculator above reproduces it: with the preset for 88.0 per cent closely held, the upper bar does not move at all while the lower one collapses to about a quarter of the length it had. With the promoter slider pushed to 100.0 per cent, the free float capitalisation goes to Rs 0/- while the headline still reads Rs 11,664.00 crore. Nothing states the point more cleanly. A capitalisation of Rs 11,664 crore says nothing about whether there is anything to buy.

The difference is not a defect in either issuer, and a thin free float is a finding rather than a fault. The two multiplications simply answer two different questions, and only one of them is about what is available. Whether a name is picked up by a weighting scheme, and how heavily it sits inside one once it is, follows the free float figure rather than the headline. The second measure exists for that reason alone, and how such a scheme is built is set out under free float weighting. The same asymmetry runs through a placement, a block trade and any calculation that divides a day's traded value by a size measure. Where availability is the question, the headline figure is the wrong denominator no matter how carefully it was computed.

What can neither figure establish?

Neither figure says whether Rs 11,664 crore is large. Sorting an issuer into a size grouping is done by a rule at a body that publishes its own current text, never by the multiplication. Neither figure ranks Sarvani Coatings against Nandivarman Paints Limited, against Kesaria Surface Solutions Limited or against anybody else, and neither says that a free float of 47.6 per cent is high, low, healthy or thin.

Both outputs are inputs to a question, not answers to one. Rs 5,552.064 crore on its own is a quantity with a currency attached. The quantity becomes useful the moment it is scaled against something, a day's traded value, a peer, a weighting scheme, and each of those scalings is a separate step with its own rules and its own way of going wrong. A figure computed the long way can be defended line by line.

Sarvani Coatings Limited, Nandivarman Paints Limited, Kesaria Surface Solutions Limited, Thottam Chemicals Limited and Meghna Iyer are invented.
Educational material. Not advice on any investment, tax, budget or market position.