6A.8 IPOs: book building, underpricing and the long run

Motivation

IPOs are the most heavily marketed event in Indian equity. The evidence on them is among the most robust in finance: on average, new issues are priced below their first-day trading value and then underperform comparable stocks for years.

6A.8.1Book building in India#

Setup: book building under SEBI ICDR 2018checked 6 Nov 2026; confirm in the current text

The issuer and its bankers set a price band; investors bid within it; the issue price is set after the book closes.
Allocation for an issuer meeting the profitability track-record test (Reg. 6(1)): at most 50% to qualified institutional buyers (QIBs), at least 15% to non-institutional investors, at least 35% to retail individual investors (bids up to ₹2 lakh). Issuers that do not meet it (Reg. 6(2)) must allot at least 75% to QIBs. Up to 60% of the QIB portion can go to anchor investors the day before the issue opens, with lock-ins (50% for 30 days and the rest for 90 days since 2022).
When the retail book is oversubscribed, retail allotment is by lottery of minimum bid lots, so a retail applicant's allocation depends on demand.
Listing on T+3 (three working days after the issue closes) has been mandatory since 1 December 2023.
Grey market premium (GMP). An informal, unregulated market quotes prices for IPO shares before listing. It has no legal standing, no settlement guarantee, and is thin and easy to manipulate. Press reporting of GMPs is not evidence of anything; this course does not use them.
SME IPOs list on NSE Emerge and BSE SME with lighter eligibility rules. After a period of very high subscription and manipulation concerns, SEBI amended the ICDR in March 2025: an SME issuer now needs operating profit (EBITDA) of at least ₹1 cr in two of the three preceding years, and offers for sale are capped at 20% of the issue¹. Risks specific to SMEs: tiny floats, wide spreads, market makers who exit after the mandatory period, and promoter-controlled price action.

6A.8.2The winner's curse#

Proposition 6A.12Rock's winner's curse

An issue of $N$ shares at price $P$ is worth $V_g$ (good) with probability $\pi$ or $V_b < V_g$ (bad). Informed investors bid only in the good state; uninformed investors bid $U$ shares in both states. In the good state total demand $U + I > N$ and allocation is pro rata, so an uninformed bidder receives a fraction $\alpha_g = N/(U + I)$ of their bid; in the bad state they receive $\alpha_b = \min(1, N/U)$ . If uninformed investors must break even, the issue price is

P = \frac{\pi\alpha_g V_g + (1-\pi)\alpha_b V_b}{\pi\alpha_g + (1-\pi)\alpha_b}, \tag{6A.14}

which is below the unconditional expected value $\pi V_g + (1-\pi)V_b$ whenever $\alpha_g < \alpha_b$ .

Proof

The uninformed expected profit per share bid is $\pi\alpha_g(V_g - P) + (1-\pi)\alpha_b(V_b - P)$ ; set it to zero and solve. The price is a weighted average of $V_g$ and $V_b$ with weights proportional to $\pi\alpha_g$ and $(1-\pi)\alpha_b$ . Rationing in the good state shifts weight toward $V_b$ relative to the weights $\pi, 1-\pi$ , so $P < \mathbb E V$ .∎

Worked example

$V_g = 130$ , $V_b = 90$ , $\pi = 0.5$ ; in the good state the uninformed get a quarter of what they bid, in the bad state all of it. Then $P = (0.5 \cdot 0.25 \cdot 130 + 0.5 \cdot 90)/(0.125 + 0.5) = 98.0$ against an expected value of 110: an average first-day return of 12.2%, while the uninformed investor's allocation-weighted return is zero. Average IPO first-day returns overstate what a retail applicant earns.

6A.8.3Underpricing and long-run returns#

Evidence

Initial returns. Ritter's international table (March 2026 update, 55 countries; built on Loughran, Ritter and Rydqvist 1994) finds average first-day returns positive in every country, with large cross-country differences². Academic Indian samples report large averages: about 78% for 386 firms (Krishnamurti and Kumar, 2002) and about 47% (Sahoo and Rajib, 2010), on different samples and periods [S]. Indian underpricing was historically higher than in the US, partly because of fixed-price issues before book building and long listing delays.
Long-run returns. Ritter (1991): 1,526 US IPOs of 1975–1984 earned an average three-year holding-period return of 34.47% against 61.86% for size- and industry-matched firms. Loughran and Ritter (1995): about 5% a year over five years for IPOs against about 12% for size-matched non-issuers. Explanations include issuers timing the market (issuing when valuations are high) and investor over-optimism about growth. Results are sensitive to the benchmark: with factor-adjusted benchmarks the underperformance shrinks, concentrated in small, low book-to-market issuers.

Practitioner notes

An IPO's offer-for-sale component is insiders selling. Read the offer document's use of proceeds, the pre-IPO transactions (prices paid by investors in the last year), the related-party section and the restated financials (M06 · 6.8). The anchor lock-in expiries (30 and 90 days) are dated supply events.