DossiersAnswer as of 5 Oct 2026

Is the market expensive?

Somewhat expensive

The Nifty 50 costs 19.3 times its last year's earnings, higher than 61% of readings since 1999.

Of the seven measures that vote, three say dear, one says cheap and three sit in between.

They disagree: the Nifty 500 CAPE is higher than 75% of readings since 2009, while the Nifty 50 price to book is lower than 90% of readings since 1999.

todaya month agoa year ago

Part of the gap is accounting: NSE switched to consolidated earnings in April 2021, which lowered P/E from that date. Our P/E history is adjusted for the switch; unadjusted, today's P/E is lower than 65% of readings since 1999. Price to book had no such break. The two also measure different things: P/B divided by P/E is the return on book value, which is 14.4% for the Nifty 50 today against a median of 23.5% before 2011. Lower returns on book make the same P/E go with a lower P/B.

The evidence

Each measure against its own history, strongest first. The bar runs from the lowest reading to the highest; the shaded part is the middle half. A measure votes dear at or above the 70th percentile of its history and cheap at or below the 30th, unless its note says otherwise. Primary measures count double.

Nifty 500 CAPE Primary

31.2 5 Oct 2026, 75th percentile since 2009 A year ago 36.4

dear
Market value to GDP Supporting. NSE main board against last year's GDP

135% 30 Sept 2026, 91st percentile since 1999 A year ago 141%

dear
Nifty 50 price to book Supporting. No break at the 2021 switch

2.77 5 Oct 2026, 10th percentile since 1999 A year ago 3.40

cheap
Nifty 50 dividend yield Supporting. A high yield counts as cheap

1.09% 5 Oct 2026, 18th percentile since 2000 A year ago 1.25%

dear
Nifty 50 CAPE Primary. Price over ten years of real earnings

25.5 5 Oct 2026, 68th percentile since 2009 A year ago 30.9

in between
Nifty 500 P/E Primary. Same basis

21.7 5 Oct 2026, 65th percentile since 1999 A year ago 24.1

in between
Nifty 50 P/E Primary. Consolidated basis, adjusted across NSE's 2021 switch

19.3 5 Oct 2026, 61st percentile since 1999 A year ago 22.0

in between
Nifty 50 P/E as NSE publishes it Context. Unadjusted: standalone earnings before April 2021

19.3 5 Oct 2026, 35th percentile since 1999 A year ago 22.0

context

Behind the answer

CAPE, Dividend yield (inverted) and Market value to GDP sit in the dearest third of their history; price to book in the cheapest

Five measures of the Nifty 50's price, each shown as its percentile against its own history up to that quarter, so measures in different units share one scale: the deeper the rose, the dearer; the deeper the blue, the cheaper. Today: P/E 61, CAPE 68, Price to book 10, Dividend yield (inverted) 82, Market value to GDP 91. The typical gap between the highest and lowest of the five was 46 percentile points before NSE's switch to consolidated earnings in April 2021 and 49 since; today it is 81.

cheapest fifthmiddledearest fifth

Source: NSE Indices published PE, PB and dividend yield; CPI from IndiaDataHub (MoSPI, spliced); market value to GDP from NSE and MoSPI.

Table

Date	P/E	CAPE	Price to book	Dividend yield (inverted)	Market value to GDP
Oct 2026	61	68	10	82	–
Mar 2026	66	75	24	78	–
Sept 2025	82	94	35	52	–
Mar 2025	69	90	36	54	–
Aug 2024	90	100	79	64	–
Feb 2024	87	99	70	69	–
Aug 2023	89	94	91	90	–
Jan 2023	82	93	81	45	–
Jul 2022	76	90	77	45	–
Dec 2021	93	98	86	67	99
Jun 2021	98	100	86	59	–
Dec 2020	100	100	77	89	–
May 2020	–	–	–	–	35
Nov 2019	96	90	66	40	–
May 2019	100	97	71	46	–
Oct 2018	89	67	36	30	–
Apr 2018	97	83	62	35	–
Oct 2017	98	77	48	52	–
Mar 2017	94	68	52	37	–
Sept 2016	96	61	46	39	–
Mar 2016	63	24	29	89	–
Aug 2015	–	–	–	–	61
Feb 2015	93	68	63	76	–
Aug 2014	67	54	48	57	–
Jan 2014	53	19	28	60	–
Jul 2013	50	12	34	59	–
Dec 2012	–	–	–	–	63
Jun 2012	37	12	26	56	–
Dec 2011	43	8	24	69	–
May 2011	–	–	–	–	70
Nov 2010	93	83	51	78	–
May 2010	80	45	47	90	–
Oct 2009	75	–	44	75	–
Apr 2009	41	–	25	77	–
Oct 2008	41	–	34	89	–
Mar 2008	–	–	–	–	94
Sept 2007	80	–	91	65	–
Mar 2007	60	–	90	60	–
Aug 2006	–	–	–	–	96
Feb 2006	68	–	94	61	–

Over the last three years prices rose 15% and earnings grew 32%

The Nifty 50's price and its trailing twelve-month earnings, both set to 100 in January 2006. NSE does not publish index earnings; they are implied from its published P/E (index value ÷ P/E), with earnings before April 2021 scaled to the consolidated basis NSE switched to then. Log scale, so equal slopes mean equal growth. When the price line pulls above the earnings line, P/E is rising; when earnings catch up, it falls. Since 2006 prices are 7.7 times their start and earnings 5.9 times. Over three years: 4.7% a year for prices, 9.7% for earnings.

PriceEarnings

Source: NSE Indices: index values and published P/E (consolidated basis from 31 Mar 2021, earlier earnings scaled by the switch-day ratio). Prices exclude dividends.

Table

Date	Price	Earnings
Oct 2026	774	587
Mar 2026	783	574
Sept 2025	862	572
Feb 2025	759	565
Aug 2024	842	541
Feb 2024	750	488
Jul 2023	678	414
Jan 2023	613	417
Jun 2022	539	406
Dec 2021	601	364
May 2021	530	268
Nov 2020	437	223
May 2020	338	267
Oct 2019	400	262
Apr 2019	400	243
Sept 2018	382	249
Mar 2018	351	248
Aug 2017	338	234
Feb 2017	302	228
Jul 2016	296	221
Jan 2016	255	224
Jul 2015	291	218
Dec 2014	282	237
Jun 2014	260	221
Nov 2013	206	205
May 2013	209	202
Oct 2012	194	185
Apr 2012	179	172
Sept 2011	170	168
Mar 2011	184	158
Sept 2010	188	144
Feb 2010	166	142
Aug 2009	154	135
Jan 2009	92	132
Jul 2008	139	141
Dec 2007	209	135
Jun 2007	143	127
Dec 2006	137	113
May 2006	111	106

Nifty 50 companies earn 14.4% on their book value, against 23.5% before 2011

Price to book divided by P/E is the return on book value. With lower returns on book, the same P/E comes with a lower price to book, so P/B reads cheap against its history while P/E does not. Part of the gap between the two measures is profitability, not price.

Source: NSE Indices published PE, PB and dividend yield; CPI from IndiaDataHub (MoSPI, spliced). P/E on the consolidated basis, adjusted across the 2021 switch.

Table

Date	Return on book
Oct 2026	14%
Jan 2026	16%
May 2025	16%
Aug 2024	18%
Dec 2023	17%
Mar 2023	20%
Jul 2022	20%
Oct 2021	17%
Feb 2021	12%
May 2020	15%
Sept 2019	15%
Dec 2018	16%
Apr 2018	17%
Jul 2017	17%
Nov 2016	17%
Feb 2016	18%
Jun 2015	19%
Sept 2014	20%
Jan 2014	19%
Apr 2013	21%
Aug 2012	21%
Nov 2011	20%
Mar 2011	20%
Jun 2010	21%
Oct 2009	20%
Jan 2009	23%
May 2008	30%
Aug 2007	31%
Dec 2006	29%
Mar 2006	31%
Jul 2005	33%
Oct 2004	25%
Jan 2004	24%
May 2003	22%
Aug 2002	25%
Dec 2001	19%
Mar 2001	26%
Jul 2000	25%
Oct 1999	22%
Feb 1999	21%

The broad market costs 12% more than the Nifty 50 per rupee of earnings, above its usual premium

The Nifty 500's P/E divided by the Nifty 50's. Above 1, the 450 smaller companies in the broad index are priced more richly than the 50 largest. The median since 1999 is 1.04; a year ago it was 1.10.

Source: NSE Indices published PE, PB and dividend yield; CPI from IndiaDataHub (MoSPI, spliced)

Table

Date	Nifty 500 P/E ÷ Nifty 50 P/E
Oct 2026	1.12
Jan 2026	1.06
May 2025	1.09
Aug 2024	1.16
Dec 2023	1.05
Mar 2023	1.04
Jul 2022	1.03
Oct 2021	1.04
Feb 2021	1.07
May 2020	1.14
Sept 2019	1.08
Dec 2018	1.13
Apr 2018	1.20
Jul 2017	1.15
Nov 2016	1.18
Feb 2016	1.09
Jun 2015	1.08
Sept 2014	1.03
Jan 2014	1.02
Apr 2013	0.99
Aug 2012	1.00
Nov 2011	0.94
Mar 2011	0.91
Jun 2010	0.88
Oct 2009	0.92
Jan 2009	0.91
Apr 2008	0.88
Aug 2007	0.92
Nov 2006	0.93
Mar 2006	1.08
Jun 2005	1.28
Oct 2004	0.97
Jan 2004	0.86
May 2003	1.19
Aug 2002	0.84
Nov 2001	0.97
Mar 2001	0.95
Jun 2000	1.14
Oct 1999	1.01
Jan 1999	1.05

Dearest against their own history: Nifty Smallcap 100, Nifty Pharma and Nifty Energy. Cheapest: Nifty Realty, Nifty LargeMidcap 250 and Nifty PSU Bank

Each index's P/E as a percentile of its own history, today (filled) and a year ago (open). The shaded middle is the 33rd to 67th percentile. Indices marked * are compared only since 2021, when NSE's switch to consolidated earnings broke their history and the adjustment was not reliable.

Percentile of P/E within each index's own history. Source: NSE Indices published PE, PB and dividend yield; CPI from IndiaDataHub (MoSPI, spliced)

Table

	A year ago	Today
Nifty Smallcap 100 *	71st	94th
Nifty Pharma	51st	76th
Nifty Energy	84th	69th
Nifty 500	79th	65th
Nifty 50	82nd	61st
Nifty Infrastructure *	61st	60th
Nifty Smallcap 250	36th	56th
Nifty MidSmallcap 400 *	63rd	47th
Nifty Auto *	48th	45th
Nifty 100	65th	44th
Nifty FMCG	77th	42nd
Nifty Midcap 100 *	68th	41st
Nifty Metal *	58th	40th
Nifty Midcap 150 *	61st	40th
Nifty 200	63rd	37th
Nifty Next 50	49th	36th
Nifty Bank	55th	33rd
Nifty IT	59th	27th
Nifty Financial Services	50th	27th
Nifty PSU Bank	33rd	21st
Nifty LargeMidcap 250 *	57th	8th
Nifty Realty *	24th	5th

What has changed

A month ago the answer was the same: “Somewhat expensive”.
A year ago the answer was “Expensive”.
Over the month, Nifty 50 CAPE went from dear to in between.
Over the year, three measures went from dear to in between: Nifty 50 P/E, Nifty 500 P/E and Nifty 50 CAPE.
Over the year, Nifty 50 price to book went from in between to cheap.
Over the year, Nifty 50 dividend yield went from in between to dear.
The largest move over the year: Nifty 50 dividend yield, from 1.25% to 1.09% (from the 48th to the 18th percentile of its history).

What history says

Starting from today's CAPE, five-year real returns have ranged from −10% to 20% a year

Each dot is a month since 2006: the Nifty 500's CAPE at the start and its real return (after inflation, dividends reinvested) over the next five years, a year. The line is today's CAPE, 31.2. Lower starting CAPE has gone with higher returns, but the dots come from overlapping windows: about 4.1 independent five-year periods in all, so this is a description of one market's history, not a forecast.

Source: Evidence C1 (docs/research/site_calculations_spec.md); NSE Indices, MoSPI CPI.

Valuation and the next five years

Descriptive, not a test

Since 2006, the Nifty 500's five-year real return (after inflation, dividends reinvested) has depended on its starting CAPE. Median by starting third: cheapest third (CAPE 15.1 to 21.1): 9.7% a year; middle third (CAPE 21.3 to 27.0): 6.2% a year; dearest third (CAPE 27.0 to 50.2): 3.4% a year. Today's CAPE of 31.2 is in the dearest third, where the five-year real return ranged from −10.4% to 19.9% a year. Each third holds about 1.0 independent five-year periods, so read this as a description of one market's history, not a forecast.

Evidence C1, valuation and later returns

How strong the link is

Descriptive, not a test

In non-overlapping samples, a higher starting CAPE went with lower five-year returns 65% of the time. For the trailing earnings yield (earnings over price, one year), a higher starting yield went with higher returns 60% of the time. There are about 4.1 independent five-year periods since 2006, so no t-statistic here is worth quoting. We have not tested valuation against returns over months or quarters, and make no claim about them.

Evidence C1

What would change the answer

It would take two measures moving a step each to make the answer “Expensive”.

One measure moving a step would make the answer “About usual”: Nifty 50 P/E.

Prices move P/E every day; earnings move it each results season, when the trailing year's profits are replaced.

Nifty 50 P/E: 19.3 now. It would count as dear above 20.8 (8% higher), and as cheap below 15.4 (20% lower).
Nifty 500 P/E: 21.7 now. It would count as dear above 22.6 (4% higher), and as cheap below 15.0 (31% lower).
Nifty 50 CAPE: 25.5 now. It would count as dear above 26.4 (3% higher), and as cheap below 18.3 (28% lower).
Nifty 500 CAPE: 31.2 now. It stops counting as dear below 29.2 (6% lower).

Coming up

8 Oct 2026TCS Results, dividend
10 Oct 2026DMART Results, period ended Sep 2026
12 Oct 2026HCLTECH Results, dividend, period ended Sep 2026
12 Oct 2026RBLBANK Results, period ended Sep 2026
14 Oct 2026BHEL Results
14 Oct 2026MRPL Results, period ended Sep 2026
15 Oct 2026ANGELONE Results, period ended Sep 2026
15 Oct 2026HDFCAMC Results, period ended Sep 2026

And 10 more in the next 14 days on the front page.

The readings and the dashboards

19.3usual

21.7usual

25.5usual

31.2above usual

the usual range: the middle half of all readingstodayEach line runs from the lowest reading ever to the highest.

How these answers are made: the dossier method. Historical research, not investment advice.