How an IPO is priced: fixed price and book building
The answer
There are 2 pricing methods. In a fixed price issue, the company names 1 price in advance and you take it or leave it. In a book building issue, the company names a range, collects bids inside it, and sets the final price from the demand it received.
Book building looks like price discovery. It is price discovery inside a range that the seller drew.
Why this costs you money
The price band does something to your thinking before you notice it.
A band of ₹380 to ₹400 makes ₹380 feel like the cautious end and ₹400 like the optimistic end. You then think about whether the issue is worth ₹380 or ₹400. You have already accepted that the answer is somewhere between those 2 numbers. It might be ₹120. Nothing in the band rules that out.
This is the most reliable trap in the whole IPO process, and it works on professionals too. The seller drew the range, and the range set the terms of your thinking.
The second loss is a document nobody opens. Every Indian prospectus contains a section called Basis for the Issue Price. It states the earnings per share, the return on net worth, the net asset value per share, the price-to-earnings ratio the issue implies, and a comparison with named listed peers. It is the only place in the process where the seller must justify the number in writing. It is usually 3 or 4 pages. Most applicants have never read one.
The third loss is mechanical. In a book-built issue, a bid below the final price is rejected and gets nothing. Retail investors who try to be clever by bidding at the floor in a heavily subscribed issue are not being cautious. They are buying a lottery ticket for a lottery they have already left.
How it works
Fixed price. The company and its bankers decide a price and publish it. You apply at that price. Demand is only known after the issue closes. It is simple, it is used mostly for small issues now, and it removes the pretence that the market participated in setting the number.
Book building. Five steps.
- The band is set. A floor price and a cap. In India the cap is normally within a defined percentage of the floor.
- Anchor investors are allotted first, 1 working day before the issue opens, at a price fixed in that process.
- The issue opens for at least 3 working days. Investors bid a price and a quantity in lots. Retail investors may instead bid at cut-off, meaning they accept whatever the final price is.
- Demand is collected. In India, the exchanges publish subscription figures by category through each day. In the United States the book is confidential and only the underwriters see it.
- The final price is set. The issuer and the lead managers pick a price at which the book is comfortably covered. When demand is strong this is at or very near the cap.
The band can be revised, and if it is, the issue period is extended. That revision is itself information: a band raised means the book is strong, a band cut means it is not.
A word on lot sizes. In India you bid in multiples of a lot, and the lot is sized so that 1 lot costs a little under the retail application limit. This matters more than it sounds, because it sets the minimum cheque and therefore who can apply at all.
What it tells you, and what it does not
The final price tells you where demand met supply, given the range on offer. That is a real fact and it is not nothing.
It does not tell you the company is worth that. The band was drawn by the seller. If the seller draws it too low, the shares jump on listing. If the seller draws it too high, the shares fall on listing. Both happen regularly, to companies of every quality.
Subscription numbers tell you less than they appear to. An issue subscribed 80 times is not 80 times better than an issue subscribed 1 time. Subscription depends on the size of the issue, the reserved quotas, and how much borrowed money is being deployed into the non-institutional category to chase an allotment. It is a measure of enthusiasm and of the plumbing. It is not a measure of value.
And the grey market premium tells you nothing that is worth acting on. It is an unofficial, unregulated quote in a small opaque market with no obligation to be right and no mechanism for you to enforce anything. It is widely reported because it is a number and numbers are easy to publish.
The decision rule
Convert the price band into a multiple yourself, then compare it with a company you can buy tomorrow without a lottery.
Step 1. In the prospectus, open Basis for the Issue Price. Write down the price-to-earnings ratio the issue implies at the cap of the band, not at the floor.
Step 2. In that same section, the issuer names its listed peers and their ratios. Write those down.
Step 3. Open your own app and check the current ratio for 1 of those peers. The number in the prospectus was calculated on a date that has passed.
Step 4. If the issue's implied multiple is higher than the peers, finish this sentence out loud: "I am paying a premium to a listed company because ______." If you cannot finish it with something specific about the business, you have the answer.
The peer is available on Monday, in any quantity, with no lottery and no deadline. That is the alternative you are actually choosing against.
Try this now
Five minutes. Use a live IPO if there is one, or the most recent one you can find.
- Search for the company name plus "RHP". Open the document. It will be several hundred pages, which does not matter, because you need 1 section.
- Search inside the document for "Basis for the Issue Price". In a US prospectus, the closest equivalents are the Dilution section and the discussion of how the offering price was determined.
- Write down 3 numbers: the earnings per share, the price-to-earnings ratio at the cap of the band, and the return on net worth.
- In the same section, find the table of listed peers. Pick 1 peer you can look up. Ideally pick a peer you already hold.
- Open your broker app, find that peer, and read its current price-to-earnings ratio from the fundamentals or overview tab.
What you should see. The comparison is usually not close. New issues are frequently offered at a higher multiple than their listed peers, and the prospectus says so in a table, in the seller's own document.
Where you already own the peer, the exercise becomes concrete in a useful way. You are being asked to pay more for a company with a shorter public record than you paid for one you already know. That may still be the right decision. It is a different decision from the one people think they are making.
If the company has no earnings, the section will compare something else, or the peer table will be thin. That is also information, and it is the reason India has a separate quota structure for companies without a profit record.
Three real cases
1. Google, 19 August 2004 (United States) — the auction that tried to remove the bank's discretion Google went public through a modified Dutch auction, in which investors of any size submitted the price and quantity they wanted, and the clearing price was derived from the bids. The company's stated aim was to reduce underpricing and to give individual investors the same access as institutions. The initial indicated range of $108 to $135 was cut, and the shares were sold at $85 . The stock still rose sharply on its first day. The auction reduced the bank's discretion. It did not remove the gap between the offer price and the market price, which tells you that gap is not only about who allocates.
2. Paytm, November 2021 (India) — priced at the cap, and what happened next One 97 Communications set a band of ₹2,080 to ₹2,150 and priced the issue at ₹2,150, the top of the band. The issue was subscribed about 1.89 times overall, which is covered but not enthusiastic. The stock fell more than 27% on its first trading day, 18 November 2021. Pricing at the cap of the band is normal and is not evidence of anything by itself. What this case shows is that a fully subscribed book at the cap is not a market verdict on value. It is the point at which the seller stopped raising the price.
3. Reliance Power, January and February 2008 (India) — record demand, immediate loss The issue was priced at ₹450 per share, with a lower price for retail investors, and attracted extraordinary demand, including a very large number of retail applications. It listed on 11 February 2008 and fell below its issue price on the first day, then kept falling as the wider market turned. The company later issued bonus shares to non-promoter shareholders to reduce their effective cost. Subscription had been enormous. Subscription was not a valuation.
The question that resolves it
A novice sees a price band and asks: is the low end or the high end more likely?
An expert asks: what multiple does the high end imply, and what does a company I could buy on Monday trade at?
The first question has 2 possible answers and both are inside the seller's range. The second question uses a number the seller had to publish and a number you can look up yourself.
What would make this wrong
If book building produced accurate prices, first-day price changes would be small and random. They are neither. Average first-day returns are positive in almost every market that has been studied, and the size of that average moves with market conditions rather than with company quality.
The honest limits:
Book building genuinely does aggregate information. Institutions with analysts covering the sector put real money behind a view, and their bids shape the price. A book-built price is better informed than a price a founder picks alone.
The peer comparison in the prospectus is also not a neutral document. Issuers choose which peers to list, and a flattering peer set is a legal document with a persuasive purpose. Check whether the peers named are really comparable in size, margin and growth, and add 1 of your own.
And a higher multiple than the peers is not automatically wrong. Faster-growing or higher-return businesses deserve higher multiples. The point of the exercise is not to reject a premium. It is to make you state, in a sentence, what the premium is for.
In India
Book building dominates. The fixed price route survives mainly for smaller issues.
The band. A floor and a cap, with the cap within a defined percentage of the floor. The band is announced a short time before the issue opens, and it can be revised, which extends the bidding period.
Categories and quotas. For a company meeting the profitability route, the allocation is broadly up to 50% to qualified institutional buyers, at least 15% to non-institutional investors, and at least 35% to retail investors. For a company using the alternative route, at least 75% must go to qualified institutional buyers and the retail share is much smaller. That difference exists precisely because the second kind of company is harder to price.
Cut-off bidding. Retail investors may bid at cut-off and accept the final price. This removes the risk of bidding below the clearing price and receiving nothing.
Discounts. Issuers may offer a discount to retail investors and to employees within limits. LIC's 2022 offer used exactly this, with separate discounts for retail applicants and for policyholders.
Public demand data. The exchanges publish live subscription figures by category during the issue. This is genuinely unusual and it is free.
Basis for the Issue Price is a mandatory section. SEBI requires the justification to be disclosed. It does not require the justification to be convincing, and it does not approve the price.
In the United States
Negotiated pricing is standard. The company files a Form S-1 with an estimated price range. Underwriters run a roadshow, collect indications of interest from institutions into a confidential book, and then the company and the banks agree the final price, usually the evening before trading begins.
The range moves, and the movement is public. Amendments to the registration statement disclose changes to the range and to the number of shares. A range raised and a deal upsized means the book is strong. A range cut, or a deal shrunk, means it is not. Because the book itself is secret, these revisions are the main public signal of demand in a US IPO.
No subscription ticker, no cut-off bid, no retail quota. An individual investor does not see demand building and cannot see how many times the deal is covered.
Dutch auctions are permitted but rare. Google's 2004 auction remains the best-known attempt, and the method did not become standard.
Direct listings do not have an offer price at all. The exchange publishes a reference price, which is not a price anybody paid, and the opening price is set by matching buy and sell orders on the first morning. Spotify's reference price on 3 April 2018 was $132 and its first trade was at $165.90.
SPACs invert the sequence. Shares are sold at a standard $10 with a redemption right, and the price of the business is negotiated later, when the shell merges with a real company.
Where they differ, and what that tells you
India publishes demand in real time. The United States hides it.
That sounds like an advantage for the Indian investor, and it is a small one, but it is smaller than it looks and it comes with a cost.
What the Indian subscription ticker actually contains: the institutional figure is the useful part, because those are informed buyers committing capital. The non-institutional figure is heavily influenced by funded applications, where investors borrow for a few days purely to chase a larger allotment. The retail figure is a measure of public enthusiasm, and public enthusiasm has been highest before some of the worst outcomes.
The cost of publishing it: a visible, rising number during a 3-day window creates urgency in exactly the people least equipped to price the issue. That is what the grey market premium amplifies. The United States has neither, and American retail investors are, for this specific reason, marketed to far less aggressively during an offer.
What the American investor should watch instead is the amendment history. A company that raised its range twice and increased the deal size has a book the underwriters could not fill at the original terms — in the good direction. A company that cut its range has the same information in the other direction. It is slower than a live ticker and it carries more meaning.
The transferable rule. In both markets, demand data tells you about the auction, not the company. The valuation work is the same in Mumbai and New York, and it is done in the prospectus, not on the ticker.
Carry this
- The band is the seller's asking range. Do your work at the cap, not the floor.
- Basis for the Issue Price is 3 pages and it contains the whole argument.
- Compare the issue to a listed peer you could buy on Monday with no lottery.