Dutch Auction IPO: How Uniform-Price Bidding Differs From India’s Book Building

A Dutch auction IPO sounds exotic to most Indian investors, and that’s fair. The mainboard IPOs most of us apply for are book-built: a price band, three days of bidding, and a final price chosen by the issuer and its bankers. A Dutch auction takes that last step out of human hands. Bids are stacked from highest to lowest, and the price falls wherever demand meets supply. Same goal, price discovery, but a very different way of getting there.

This piece walks through how uniform-price bidding works, runs one made-up issue through both systems side by side, and looks at why the auction route, for all its fairness on paper, never really caught on with companies.

What a Dutch auction IPO actually is

The original Dutch auction is a falling-price sale. As Wikipedia’s entry on the format describes it, the auctioneer starts high and keeps lowering the price until somebody accepts, or until a reserve price is hit. It’s used for goods that have to move fast, such as flowers and fresh produce, and it’s sometimes called a clock auction.

Share offerings use a modified version. According to Investopedia’s explainer on Dutch auctions, investors submit bids that state two things, how many shares they want and the price they’re willing to pay. Once bidding closes, the shares are handed out from the highest bid downwards until the whole offering is used up. The offering price is the highest price at which the entire issue sells. And here’s the part that matters: every winning bidder pays that same price, the last successful bid, no matter how much higher their own bid was.

That’s why the cleaner name is a uniform-price auction. You might bid ₹240 and end up paying ₹215 because ₹215 is where the book cleared. Your high bid only bought you priority, not a higher bill.

In a uniform-price auction, bidding high buys you a place in the queue, not a bigger bill. Everyone who wins pays the clearing price.

How India’s book building works

Now compare that with what SEBI’s rules ask of an Indian issuer. The SEBI investor FAQ on public issues (May 2025) lays it out plainly. The issuer discloses a price band or a floor price before the issue opens. In an IPO, that has to be announced at least two working days before opening. Investors bid for a quantity at a price inside the band. Based on the demand at various price levels, the book running lead manager (BRLM), in consultation with the issuer, arrives at the issue price.

A few rules from the same FAQ shape how it plays out:

• The spread between the floor and the cap of the price band has to be at least 5% and no more than 20%.
• The band can be revised. If it is, bidding gets extended by three more days, but the total bidding period can’t go beyond ten working days.
• Only retail individual investors, meaning those applying for up to ₹2,00,000, can tick the “cut-off” box, which says they’ll take shares at whatever price is discovered within the band. Cut-off bids always stay valid for allotment, while a price bid below the discovered price can become invalid.
• In a book-built issue under the profitability route, at least 35% of the net offer is kept for retail individual investors, at least 15% for non-institutional investors, and no more than 50% for qualified institutional buyers (QIBs).
• QIB allotment is proportionate. Each retail investor and NII gets at least the minimum bid lot, subject to availability of shares in the retail category, with any balance shared out proportionately.

So book building also collects bids at several prices. The difference is the ceiling, the buckets and who makes the final call. Bids can’t go above the cap. Shares are split by investor category, not by how aggressively someone bid. And the price is a decision taken by the issuer and its bankers, not a number the order book spits out.

One issue, two systems: a worked example

Let’s take a hypothetical company, call it Sahyadri Foods, selling 10,00,000 shares. These numbers are invented purely to show the mechanics; they don’t describe any real issue.

Under a Dutch auction

Suppose these bids come in, sorted from highest price to lowest:

Total demand is 15,00,000 shares, so the issue is 1.5 times covered. Walk down the cumulative column. At ₹220, only 9,00,000 shares are spoken for, which leaves 1,00,000 unsold. At ₹215, cumulative demand hits 11,00,000, which is more than enough. So ₹215 is the highest price at which the whole 10 lakh shares sell. That’s the clearing price.

Everyone who bid above ₹215 gets their full quantity. Bidders at exactly ₹215 asked for 2,00,000 shares, but only 1,00,000 are left, so they’re filled at 50%. The ₹205 bids get nothing. The company raises 10,00,000 × ₹215 = ₹21.5 crore.

Now look at the investor who bid ₹240 for 1,50,000 shares. They pay ₹215, not ₹240. That’s ₹25 a share below their own bid, or ₹37.5 lakh on the full order. Had each winner paid their own bid instead, a pay-as-you-bid system, the company would have collected ₹22.615 crore. Under that system, though, nobody would bid ₹240 honestly; they’d shade their bids down and try to guess the clearing level. The uniform price removes that guessing game. You can bid what the shares are really worth to you, knowing you’ll only pay the market-clearing number.

Wikipedia makes the same point with a US Treasury example: because all filled bids receive the clearing price, the format in theory leads to more aggressive bidding.

Under Indian book building

Run the same demand through a book-built issue with a price band of ₹200 to ₹215. That’s a 7.5% spread, inside SEBI’s 5%-20% range.

The first thing that changes is that nobody can bid ₹240 or ₹232. The investors who’d happily pay those prices simply bid at the cap of ₹215, and retail applicants among them may just tick cut-off. Demand at ₹215 now looks like 11,00,000 shares, everyone from the first five rows, against 10,00,000 on offer. That’s 1.1 times. The ₹205 bidders are still in the book, but below the likely final price.

With demand concentrated at the top of the band, the issuer and BRLM would very likely price at ₹215. The headline number ends up the same as in the auction. The road there is different, and so is the split of shares.

Instead of filling from the highest bidder down, the 10 lakh shares go into buckets. Under the profitability-route split, at least 3,50,000 go to retail, at least 1,50,000 to NIIs and at most 5,00,000 to QIBs. Within the QIB bucket, allotment is proportionate. A fund that would have paid ₹240 gets no priority over one that would have paid ₹215. Retail applicants get at least a minimum lot each, subject to availability. Your willingness to pay more than the cap doesn’t show up anywhere in the allocation.

Book building hides the top of the demand curve behind the cap. A Dutch auction puts the whole curve on the table.

Dutch auction IPO vs book building at a glance

Google’s 2004 listing, the famous test case

The best-known Dutch auction IPO is Google’s, in August 2004. Investopedia notes the company chose the format to avoid a big first-day “pop”. Context helps here. Investopedia cites a first-day pop of 18.8% for US IPOs from 1980 to 2001, jumping to 77% in 1999 and the first half of 2000.

Google’s first plan was 25.9 million shares at $108 to $135. About a week before the offering, after analysts suggested it was overpriced, the range was cut to $85-$95 for 19.6 million shares. The shares were priced at $85, the bottom of that range, and ended the first day at $100.34. Per the Wikipedia history of Google, 19,605,052 shares were offered at $85, the sale raised US$1.67 billion, and it gave the company a market value of more than $23 billion.

So even the poster child didn’t kill the pop. The stock still closed about $15 above the offer price on day one. Investopedia puts part of the lukewarm reception down to negative press before the issue, an SEC inquiry into executive share allocation and secrecy over how the money would be used. The auction did its job of letting ordinary investors bid, but it couldn’t fix sentiment.

Why companies mostly stuck with bankers

Wikipedia notes that auctions have been used for hundreds of IPOs in more than two dozen countries, yet they weren’t popular with issuers and were replaced by other methods. One of the largest uniform-price IPOs was Singapore Telecom’s in 1994. Bill Hambrecht’s OpenIPO, a variation on the Dutch auction, was used for a number of US listings.

The supporters’ case is simple. A 2006 paper in the Stanford Journal of Law, Business & Finance, cited by Wikipedia, describes Dutch auctions as more efficient and fairer, since they can stop underwriters from steering stock to favoured clients. Investopedia adds that the format opens the door to small investors in markets where IPO allocation is usually controlled by investment banks.

The worries are just as concrete:

• Less control over price. When anyone can bid, Investopedia points out, the crowd may do less rigorous homework than bankers, and the price may not reflect the business well.
• Winner’s curse. Bidders who went high may realise after listing that they overpaid and rush to sell, knocking the price down.
• Collusion risk. The same Stanford paper, per Wikipedia, flags the possibility of tacit collusion where the issuer keeps discretion over price and allocation.

There’s also a softer reason that rarely makes it into textbooks. Book building lets issuers and bankers manage the story: who anchors the book, at what price, and how the issue looks on day one. A pure auction gives up some of that control, and most promoters don’t want to.

What this means for an Indian IPO investor

Even if you only ever apply to book-built issues, understanding the auction sharpens how you read them.

First, when an issue is many times subscribed and prices at the cap, remember that the cap may be hiding demand above it. That unmet demand is one possible reason a heavily subscribed issue can list well above its issue price. In an auction, some of that money could have gone to the company instead.

Second, the cut-off option is the closest thing retail investors have to “bid high, pay the clearing price”. Ticking it means you’ll take shares at whatever final price is discovered within the band, and your bid can’t become invalid because you guessed too low.

Third, allotment in India isn’t a reward for bidding higher. In the retail bucket, getting shares in an oversubscribed issue depends on the minimum-lot rule and availability, not on your price. Our guide to the IPO cut-off price goes deeper on that, and the IPO investing in India pillar covers the full application process. If you’re curious about who sets the tone before retail even bids, read our piece on anchor investors in an IPO.

The two systems are trying to answer the same question: what is this company worth to the market on day one? The auction trusts the order book. Book building trusts the bankers to read it. Neither gets it perfectly right, which is why listing day still surprises us so often.

Sources

• SEBI: Frequently Asked Questions on public issues (May 2025)
• Investopedia: Dutch Auction, How It’s Used for IPOs and Treasury Bonds
• Wikipedia: Dutch auction
• Wikipedia: History of Google (financing and initial public offering)

Disclaimer: This article is for information and education only and is not investment advice. IPO investments carry market risk; read the offer document carefully. Please consult a SEBI-registered adviser before making any investment decision.