Split, bonus and dividend adjusted stock prices for India
How split, bonus and dividend adjusted stock prices for NSE and BSE are calculated: the factor formulas, worked examples on Reliance and Tata Steel, raw vs adjusted close, ISIN changes and what does not get adjusted.
Reliance closed at 1,075.50 on 4 January 2010 and at 1,324.40 on 9 November 2026. Read straight off the exchange files, that is a 23.1% gain in almost seventeen years. The adjusted series says 452.0%. The difference is two 1:1 bonus issues and every dividend in between. Any return, moving average or backtest computed on the raw close is wrong across those dates.
This post explains how adjusted prices for Indian stocks are calculated, with the formulas we use and examples you can check against the API.
Raw and adjusted, and when to use each
The NSE and BSE bhavcopy has the raw close: the price that actually printed that day. When a company splits its shares or issues bonus shares, the price drops overnight while holders are no worse off, because they now hold more shares. A dividend does a smaller version of the same thing on the ex date.
Back-adjustment leaves the latest prices alone and scales everything before each corporate action by a factor, so that the series moves only when the value of holding the stock moved. Use the raw close to match a contract note, compute turnover or check against the exchange. Use the adjusted close for returns, indicators, volatility and backtests.
The factors
Each corporate action has a factor. The cumulative factor on a date is the product of the factors of every action with an ex date after it, and the adjusted close is the raw close times that cumulative factor. For rows after the last action the factor is 1.
split new face value / old face value
bonus N:M M / (N + M) N new shares for every M held
dividend (prev_close - cash) / prev_close
prev_close = close on the session before the ex date
adj_factor_cumulative(d) = product of factors with ex_date > d
adj_close(d) = close(d) * adj_factor_cumulative(d)Split. Tata Steel split its Rs 10 shares into Re 1 shares with ex date 28 July 2022. The factor is 1 / 10 = 0.1. The raw close went from 959.40 on 27 July to 100.35 on the 28th, and the adjusted series steps by exactly that factor, so the adjusted close moves from 87.20 to 91.21.
Bonus. Reliance issued 1 bonus share for every 1 held, ex date 28 October 2024. N = 1, M = 1, so the factor is 1 / 2. A 1:2 bonus (one new share for every two held) would be 2 / 3, and the raw price drops by a third.
Dividend. Reliance went ex a Rs 10 dividend on 19 August 2024. The previous session was 16 August, close 2,956.40. The factor is (2,956.40 - 10) / 2,956.40 = 0.996618. Use the close of the session before the ex date, even if that session was in the previous calendar year. Getting that lookup wrong across a year boundary is one of the bugs we describe in the split that broke our adjusted prices. When several payouts go ex on the same day, such as a final and a special dividend, the cash is their sum.
A worked example you can check
After the October 2024 bonus, Reliance went ex two dividends: Rs 5.50 on 14 August 2025 (previous close 1,382.60) and Rs 6 on 5 June 2026 (previous close 1,303.70). So every row from the bonus until 13 August 2025 carries:
(1382.60 - 5.5) / 1382.60 * (1303.70 - 6) / 1303.70 = 0.991438Rows before the bonus carry that times 0.5, and rows before the August 2024 dividend carry that times 0.996618 as well. The API returns exactly these numbers:
import os
from tej import Client
c = Client(api_key=os.environ["TEJ_API_KEY"])
for r in c.adjusted("RELIANCE", "nse", from_="2024-08-14", to="2024-10-29"):
print(r["date"], r["close"], round(r["adj_factor_cumulative"], 6), round(r["adj_close"], 2))date close factor adj_close
2024-08-14 2923.70 0.494042 1444.43
2024-08-16 2956.40 0.494042 1460.59
2024-08-19 2976.80 0.495719 1475.66 dividend Rs 10 ex
...
2024-10-25 2655.70 0.495719 1316.48
2024-10-28 1334.35 0.991438 1322.93 bonus 1:1 ex
2024-10-29 1340.00 0.991438 1328.530.494042 / 0.495719 is the dividend factor 0.996618, and 0.495719 / 0.991438 is the bonus factor 0.5. The raw close halves on 28 October, the adjusted close moves 0.5%.
Only the close is adjusted in the response. If you need adjusted open, high and low, multiply them by adj_factor_cumulative yourself. Volume is as traded.
What does not get adjusted
Demergers, rights issues, buybacks and mergers carry a factor of 1. We follow the NSE convention here, and it differs from Yahoo Finance, which scales history for spin-offs. Reliance is the example again: on 20 July 2023 it went ex the Jio Financial Services demerger, and the raw close fell from 2,841.85 to 2,619.85, or 7.8%. The adjusted series shows the same 7.8% drop. Holders received Jio Financial shares, so for them the adjusted Reliance series understates the return from that day. If your strategy held through a demerger, account for the spun-off shares separately.
ISIN changes and renames
In India a face value split usually issues a new ISIN. Tata Steel traded as INE081A01012 until 28 July 2022 and as INE081A01020 from the 29th. If you key a price history on ISIN, which is the usual advice because symbols change, a split cuts the history in two and the split factor never reaches the old rows. Around 590 NSE symbols have changed ISIN under the same ticker since 2010. We link ISIN intervals into one instrument chain when the same symbol moved from one to the other within 30 days, and apply factors along the chain. The isin column still shows what the exchange printed that day.
A rename is the opposite case: same ISIN, new symbol. Zomato traded as ZOMATO until 8 April 2025 and as ETERNAL from 9 April, both on INE758T01015. Our endpoints take the symbol as the exchange used it on each day, so look up the intervals with /v1/symbols/nse?isin=INE758T01015 (or c.symbols) and fetch each one.
Sometimes the exchange publishes no corporate action at all. The NIFTYBEES ETF went from a raw close of 1,292.54 on 18 December 2019 to 130.20 the next day, a 10:1 split with no action in NSE's feed. Our pipeline records it as a split derived from the price step and says so in the action's raw_subject. The adjusted close for 18 December 2019 is 129.25.
Getting adjusted prices
curl "https://api.tejhq.dev/v1/actions/RELIANCE"
curl -H "Authorization: Bearer $TEJ_API_KEY" \
"https://api.tejhq.dev/v1/adjusted/nse/RELIANCE?from=2010-01-01&to=2026-11-10"/v1/actions/SYMBOLneeds no key and lists splits, bonuses, dividends and the other action types with ex dates, ratios, face values, cash amounts and the original announcement text./v1/adjusted/EXCHANGE/SYMBOLneeds a free key (email sign-in at /keys, 1,000 calls a day). Each row has the raw OHLCV plusadj_factor_cumulativeandadj_close.- For every symbol at once,
prices_adjusted/on HuggingFace has one parquet file per exchange per year.
Coverage is NSE from 4 January 2010 and BSE from 8 July 2024. One practical point: adjusted history changes every time a new action goes ex, because every earlier row gets a new factor. If you cache adjusted prices, refresh them after ex dates, or store the raw close and the actions and apply the factors yourself.
How far to trust it
Against Yahoo Finance's adjusted close for the 50 most traded NSE names from January 2024 to May 2026, 96.88% of daily comparisons agree within 1%, and 46 of the 48 symbols we could compare sit between 98.96% and 100%. The two exceptions are a Yahoo bonus date and the demerger convention above. That covers large caps. Small caps with unusual actions are where we expect errors to be most likely, on our side and in the exchange files.
The data is provided as is, from the files NSE and BSE publish openly, and the data terms say what you can do with it. None of this is investment advice. If you find a split, bonus or dividend we have wrong, send the symbol to github.com/tejhq/tej-bazaar.