Where every number comes from

Each series names its publisher and the table it was read from. The three indices are computed by one script, data/compute_indices.py, for every year from 2004 to 2026, and the replication code reproduces the scorecard from the published dataset.

Cite this work. Raman, V.S. (2026). Some Perspective: India's Political Economy, 2004-2026. someperspective.info. Data and analysis CC BY 4.0.

MethodologyHow each series and index is built.Technical appendixIndex construction, robustness tests and full result tables.The working paperGrowth Without Accountability, with a PDF.Research packageThe dataset as CSV and Excel, the replication code and the briefs.data.jsonThe site's source of truth, every series and every source note.Master datasetOne row per year, every column the scorecard and the indices use.

The source of each series

Real GDP growth

Real GDP growth, MoSPI national accounts. Year Y is the fiscal year that starts in April of Y (2014 = 2014-15), the same convention as the World Bank series used for 2004-2013. 2014-2022: the latest MoSPI estimates on the 2011-12 base (2017-18 6.8, 2018-19 6.5, 2019-20 3.9, 2020-21 -5.8, 2021-22 9.7, 2022-23 7.6). 2023-2025: the new series on base 2022-23. MoSPI's release of 27 February 2026 gives 2023-24 7.2 and 2024-25 7.1; the Provisional Estimates of 5 June 2026 give 2025-26 7.7 (7.8% was the fourth quarter alone). MoSPI's API returned 7.3, 7.2 and 7.8 for the same years and is not used; on the old base 2023-24 had been 9.2 and 2024-25 6.5. BASIS BREAK at 2023. 2026 (2026-27) is empty: the first annual estimate is due in January 2027. Quarterly growth for April to June 2026 was 7.8%. The World Bank series for 2004-2013 is on the 2011-12 base.

Consumer price inflation

MoSPI CPI (Combined), calendar-year change in the average index, the World Bank convention (FP.CPI.TOTL.ZG) used for 2004-2013. 2014-2025 on base 2012=100; 2025 is 2.4 on this convention (the mean of the twelve monthly year-on-year prints is 2.2). 2026: the average of the monthly year-on-year prints, because no annual index exists yet published for 2026 (January to August 2026: January 2.74, February 3.21, March 3.4, April 3.48, May 3.93, June 4.38, July 4.45, August 4.82). BASIS BREAK: from the January 2026 print (published 12 Feb 2026) MoSPI moved to a new CPI series on base 2024=100, with weights from HCES 2023-24 and the COICOP 2018 classification, replacing base 2012=100. The 2026 value is on a different basis from the years before it. RBI target band 2-6%.

Union fiscal deficit, share of GDP

Union Budget documents; central government fiscal deficit as % of GDP, for the fiscal year starting in the year shown (2014 = 2014-15). 2025 is the provisional actual for 2025-26 released by the Controller General of Accounts on 1 June 2026 (Rs 15,19,169 crore, 4.4% of GDP); 2026 has no outturn yet. The Budget Estimate for 2026-27 (4.3, Union Budget of 1 February 2026) is held in the estimates block and drawn as a hollow marker; no average uses it.

Rupees per US dollar

Calendar-year average rupees per US dollar: the mean of the monthly averages in the Federal Reserve's H.10 release (FRED series EXINUS), rounded to the nearest rupee. 2026 is January to September (93.9, shown as 94); 2025 averages 87.2. The record low of 96.8 is the Fed's daily rate for 20 May 2026 (DEXINUS 96.82).

Unemployment rate, ages 15 and over

2015: Labour Bureau, Fifth Annual Employment-Unemployment Survey (2015-16), usual principal status, 5.0. 2014 and 2016 are empty: the Labour Bureau published no survey for 2014-15 or 2016-17, and its 2013-14 survey (4.9) belongs to an earlier year. The Labour Bureau series is a different survey from PLFS. 2017-2025: annual rates on usual status (ps+ss). From 2017 a year is the PLFS survey year that starts in July of it (2017 = 2017-18), through 2023 (2023-24); 2024 and 2025 are MoSPI's calendar-year rates. 2026: the average of the 8 PLFS Monthly Bulletins published for 2026 (January to August 2026: January 5, February 4.9, March 5.1, April 5.2, May 5.5, June 5.5, July 5.1, August 5), all-India, persons 15+, rural + urban. BASIS BREAK: the monthly bulletins report current weekly status (CWS), which counts a person as unemployed on the past week's activity and runs well above usual status. The 2026 value is therefore a different measure from the years before it, and most of the rise from 3.1 per cent in 2025 is the change of measure. On the same weekly measure MoSPI's calendar-year rate for 2025 was 5.3 per cent.

Youth unemployment rate, ages 15 to 29

MoSPI PLFS, usual status (ps+ss), persons aged 15-29, all-India, rural + urban, from api.mospi.gov.in (retrieved 10 October 2026). 2017-2023 are the PLFS survey years starting in July of the year shown (2017 = 2017-18); 2024 and 2025 are MoSPI's calendar-year rates. 2014-2016 are left empty because PLFS began in 2017-18 and no comparable published figure exists. 2026 is left empty until an annual figure is published: the 2026 monthly bulletins average 15.4% for January to August, but they report current weekly status, which runs well above usual status.

Unemployment among graduates

MoSPI PLFS, usual status (ps+ss), persons aged 15 and over with a graduate degree, all-India, rural + urban, from api.mospi.gov.in (retrieved 10 October 2026). 2017-2023 are the PLFS survey years starting in July of the year shown (2017 = 2017-18). MoSPI publishes no figure for graduates for 2014-2016 (before PLFS) or for 2024 onward yet, so those years are left empty.

Women's labour force participation, ages 15 and over

MoSPI PLFS, labour force participation rate, usual status (ps+ss), women aged 15 and over, all-India, rural + urban, from api.mospi.gov.in (retrieved 10 October 2026). 2017-2023 are the PLFS survey years starting in July of the year shown (2017 = 2017-18); 2024 and 2025 are MoSPI's calendar-year rates. 2014-2016 are left empty because PLFS began in 2017-18, and 2026 until an annual figure is published. PLFS counts a woman who helps without pay in a household enterprise as a worker.

Workers in regular jobs with social security

Derived from MoSPI PLFS published tables (api.mospi.gov.in, retrieved 10 October 2026), usual status (ps+ss), all-India, persons: the share of regular wage/salaried workers among all workers, multiplied by the share of regular wage/salaried employees with any social security benefit (100 minus the published share without any). 2017-18: 22.8% x 50.4% = 11.5; 2018-19: 23.8 x 48.1 = 11.4; 2019-20: 22.9 x 45.8 = 10.5; 2020-21: 21.1 x 46.2 = 9.7; 2021-22: 21.5 x 47.0 = 10.1; 2022-23: 20.9 x 46.1 = 9.6; 2023-24: 21.7 x 46.6 = 10.1. The social security table covers regular wage/salaried employees in the non-agricultural sector; regular wage work in agriculture is about 1% of workers. Years are PLFS survey years starting in July of the year shown. 2014-2016 (before PLFS) and 2024 onward (no annual table yet) are left empty.

Workers without a regular job and social security

100 minus formalEmployment, for the years formalEmployment has a value.

Gig and platform workers, millions

NITI Aayog, India's Booming Gig and Platform Economy (June 2022), Table 14, gig workers in millions, usual principal and subsidiary status. 2017-2019 are NITI's actual estimates for the PLFS survey years starting in July of the year shown (2017-18 5.3, 2018-19 5.4, 2019-20 6.8). NITI's table labels 2020-21 (7.7 million) a projection, though the Ministry of Labour cites it as the current estimate, so it is held in the estimates block with the 2025-26 projection (14.3 million) and drawn as hollow markers. 23.5 million for 2029-30 is outside the site's range. NITI gives no figure for 2014-2016. NITI's 2011-12 estimate is 2.5 million.

Income share of the top 1%

World Inequality Lab: Bharti, Chancel, Piketty and Somanchi (2024), 'Income and Wealth Inequality in India, 1922-2023', WIL Working Paper 2024/09, Statistical Appendix, table of per-adult pre-tax national income shares, 1951-2022. The series ends at 2022 (2022-23); 2023-2026 are empty because no annual figure has been published for them. Top 1% share.

Income share of the bottom 50%

Real growth in rural household income

NABARD All India Rural Financial Inclusion Survey (NAFIS), average monthly income of rural households: Rs 8,059 in 2016-17 and Rs 12,698 in 2021-22, a nominal rise of 57.6%. Deflated by MoSPI's CPI-Rural (base 2012=100; fiscal-year averages 132.43 and 164.53, a rise of 24.2%), the real rise is 26.8%, about 4.9% a year. Indexed 2016-17 = 100, so 2016 = 100 and 2021 = 126.8. NAFIS has two rounds only; every other year is left empty. The same survey found the share of rural households with outstanding debt rose from 47.4% to 52%.

Press freedom rank, of 180 (1 is best)

Reporters Without Borders, World Press Freedom Index (annual). RSF published a single 2011-12 index; its rank (131) is shown for both 2011 and 2012. RSF ranked 167 to 179 countries in 2004-2013 and 180 since.

Statistical Suppression Index, 0 to 10

Constructed: SSI = sum(weight × graded severity) on a 0-10 scale, with persistence; 6 documented suppression streams incl. COVID-mortality undercount and the NSO merger. See data/compute_indices.py

Fiscal Centralisation Index, 0 to 1

Constructed: mean of six components on a relative 0-1 scale. Four are published fiscal ratios, min-max normalised over the years each is published: the cess and surcharge share of gross tax revenue (15th Finance Commission Table 3.8 for 2011-2018; Union Receipt Budgets for 2019-2026), tax devolution as a share of gross tax revenue (16th Finance Commission Annexure 7.1; Budget at a Glance 2026-27), states' own revenue as a share of their revenue receipts (RBI Handbook 2025-26 Table 94) and non-Finance-Commission grants as a share of Union transfers (16th Finance Commission Annexure 7.1). Two are coded: GST from 2017 and the reform-linked borrowing conditions of 2020-21. Rebuilt from these published figures on 10 October 2026; see data/compute_indices.py

Democratic Quality Index, 0 to 1

Constructed: geometric mean of V-Dem Liberal Democracy Index (v2x_libdem) × (Freedom House/100) × ((180-RSF rank)/180) × V-Dem Core Civil Society Index (v2xcs_ccsi); see data/compute_indices.py. The two V-Dem columns are the published series verbatim; earlier approximations that overstated the pre-2014 baseline were corrected on 15 August 2026.