Work Indian Buybacks Decoded

Indian Buybacks Decoded

Every tender and open-market buyback in India from 2005 to 2026, read off the SEBI filings and backtested trade by trade.

Type
Event-driven study
Published
September 2026
Scope
606 buybacks, 2005–2026
Built from
2,337 SEBI filings

Summary

Question
What decides how much a small shareholder actually gets out of an Indian buyback?
Answer
Acceptance, not the premium. Median small-shareholder acceptance fell from 95% in 2012–19 to 43% since 2020, as the crowd arrived. The number that predicts it is coverage: the quota set against the public holding.
Why it matters
The premium everyone quotes is what pulls in the crowd that shrinks acceptance, and the shares that aren't accepted come back to you, usually at a lower price.

The question

A company buys back its own shares and you can sell yours to it. In a tender offer it names a price, you apply, and part of what you offer is bought at that price while the rest comes back to you. On the open market it buys on the exchange over months, and you trade around it.

Everybody quotes the premium: the buyback price over today's close. I wanted to know what actually decides the outcome for a small shareholder, so I read every buyback in SEBI's archive rather than the handful people remember.

What I did

Click a step to see what happens there.

  1. 2,337 SEBI filings, scanned pages included. Acceptance ratios, bids, tendered shares and promoter participation were read off each post-buyback announcement, with the letter of offer alongside.

  2. One row per buyback: 370 tender offers and 236 open-market buybacks. Daily closes come from NSE and BSE bhavcopies, adjusted for the 12 deals that had a split or bonus in the window.

  3. A fixed ₹1 lakh ticket per trade, real delivery charges both ways, whole shares only: more than 8,000 backtested trades across entry days and exit rules.

  4. Find a rule using only 2003–13 and check it on 2014–23, then the other way round. With about 18,000 rule-and-exit combinations tried, this is the check that matters.

  5. Seven steps to run before the next buyback and a one-page pre-trade checklist for each route, tender and open market.

606 buybacks, by yearNumber of buybacks studied. Tender offers by announcement year; open-market buybacks by board-meeting year. 2026 runs to September. Source: the Data Edition datasets; matches the report, pp.7 and 53.

370 deals. Three eras: 10 in 2005–11, 187 in 2012–19 and 173 in 2020–26. The peak was 2018, with 51.

Show the numbers
YearTender offersOpen market
200510
200600
200701
2008131
2009020
201078
2011122
2012117
2013721
201449
2015124
2016278
2017427
20185114
20194312
20203221
20212312
20223323
2023215
2024350
2025120
2026170

What I found

1. Small shareholders' acceptance halved

In a typical tender buyback, small shareholders had 95% of their shares accepted in 2012–19. Since 2020 the median is 43%, and for deals in 2023–26 it is 37%. The share of deals with full acceptance fell from 47% to 26%. The difference between the two periods is statistically clear (Mann-Whitney, p < 0.001).

Median small-shareholder acceptance per tender buybackShare of tendered shares accepted. Source: Indian Buybacks Decoded, pp.15 and 49.
2012–1995%
2020–2643%
2023–2637%
Show the numbers
PeriodMedian acceptance
2012–1995%
2020–2643%
2023–2637%

2. The crowd found buybacks

The median number of small-shareholder bids per buyback rose from 2,724 to 12,662, and from 134 to 766 for every crore of quota. The small category was bid 106% of its quota in a typical 2012–19 deal and 231% since 2020. The median buyback also got smaller, from 3.9% of equity to 2.3%, so more bids chased a thinner quota.

Median small-shareholder bids per buybackValid bids in the reserved category. Source: Indian Buybacks Decoded, p.15.
2012–192,724
2020–2612,662
Show the numbers
PeriodMedian bids
2012–192,724
2020–2612,662

3. The premium is the bait

The median premium over the board-day close rose from 13.3% to 21.6%, and from 8.4% to 16.8% at the cum date. A fatter premium sounds like good news. For acceptance it is the opposite: it is what brings the crowd.

The number that decides most of the acceptance is one almost nobody calculates. Coverage is the small-shareholder quota set against what the public actually holds. Under 1.5% is where the losing trades cluster; above 3% is where full acceptance lives.

Median premium over the board-day closeBuyback price over the close on the day the board approved it. Source: Indian Buybacks Decoded, p.16.
2012–1913.3%
2020–2621.6%
Show the numbers
PeriodMedian premium
2012–1913.3%
2020–2621.6%

4. One grade, tested on years it never saw

For open-market buybacks, a two-condition grade, known on the day you buy, made money in all 63 of its backtested deals. Found using only 2003–13, it went 26 for 26 on 2014–23. Found using only 2014–23, it went 37 for 37 on 2003–13.

Most rules that look perfect on one half of the data shrink when they meet the other. This one held, I think because it is really one idea rather than a stack of lucky filters. Backtested; past results aren't a promise.

63 of 63 deals made money, in two out-of-sample testsEach dot is one deal the grade picked. Source: Indian Buybacks Decoded, p.76.
  • Tested on 2014–23 (rule found on 2003–13) 26
  • Tested on 2003–13 (rule found on 2014–23) 37
Show the numbers
TestDealsMade money
Rule found on 2003–13, tested on 2014–232626
Rule found on 2014–23, tested on 2003–133737
Total6363

Try the arithmetic

The headline spread is only the start. Set the acceptance ratio and the share of profit lost to costs and tax, and see what reaches you on the capital you put in.

Try the tender-buyback maths

Example inputs, not a real deal.

35%
30%
Headline spread0%
After acceptance0%
What you keep0%

Ignores what the unaccepted shares do next. The study measures that.

Skills on show

  • Event-driven analysis
  • Backtesting
  • Data analysis
  • Statistical testing
  • Primary-source research
  • Report writing

Limits

  • Acceptance is a category aggregate. Yours depends on your own entitlement and what you tender.
  • Coverage is a proxy, built from the shareholding pattern rather than the actual register.
  • Survivorship: only buybacks that completed and filed a post-buyback announcement are in the data. 35 open-market buybacks never filed one.
  • Closing prices only, and today's brokerage applied to every year: a like-for-like cost lens, not what a trader paid in 2014. No tax and no dividends.
  • The rule scans are in-sample descriptions of history. Only the out-of-sample checks are tests.

The full study is a 94-page report with an optional data edition. See the report