On 8 of 9 measures, India's record since 2014 is worse than in the decade before

The page is split where the governments changed. The UPA decade sits on the left and the NDA period on the right, each at its own width in years, so every chart below changes side between 2013 and 2014. Inflation is the one measure that improved. Growth was lower overall and about the same once each period's crisis year is left out.

By Dr. Varna Sri Raman · Updated 2026-10-19 · Version 2.58.0

The scorecard

Nine measures that can be compared across both periods

Each figure is the average of the years with a published value in that period. Under each pair is the series itself, year by year. Dashed lines are the two averages. Press freedom is a rank, drawn so that up is better.

UPA 2004-2013 NDA 2014-2026

GDP growth: UPA average 6.8%, NDA average 6.2%. Worse after 2014.

CPI inflation: UPA average 8.0%, NDA average 4.8%. Better after 2014.

Fiscal deficit, % of GDP: UPA average 4.6%, NDA average 5.0%. Worse after 2014.

Top 1% income share: UPA average 20.4%, NDA average 21.9%. Worse after 2014.

Bottom 50% income share: UPA average 16.8%, NDA average 14.7%. Worse after 2014.

Press freedom rank (of 180): UPA average 120, NDA average 145. Worse after 2014.

Statistical Suppression Index (0-10): UPA average 0.00, NDA average 4.78. Worse after 2014.

Fiscal Centralisation Index (0-1): UPA average 0.30, NDA average 0.43. Worse after 2014.

Democratic Quality Index (0-1): UPA average 0.57, NDA average 0.37. Worse after 2014.

8of 9 measures worse after 2014

Only measures with a published figure in both periods are compared, so series that begin in 2017-18 with the labour force survey have no earlier counterpart on the same basis and are left out. Each average uses the years with a published figure, and the periods differ in length: ten years for the UPA and thirteen for the NDA, with the two income shares ending in 2022.

The count records the direction of each difference and gives no weight to its size. Some gaps are small, and the growth row closes to within a tenth of a point once each period's crisis year is left out, as the first finding shows. The comparison describes how the two periods differ. It does not say that a government caused each change, and the paper makes no causal claim. The research package holds the dataset and the code that reproduces every average on this page.

What the numbers show

Six findings, each with the series behind it

Each finding gives the claim, the series behind it and the publisher of the data. Where a result turns on a choice of years or on the author's coding, the text says so. All the charts share one time axis, so the change of government falls on the dividing line in every one.

Finding 1 of 6

Growth was lower over the whole period, and about the same once each period's crisis year is left out

The economy grew 6.2% a year on average from 2014-15 to 2025-26, against 6.8% from 2004-05 to 2013-14. The difference comes from 2020-21, when output fell 5.8%. Without that year the later period averaged 7.3%, above the earlier decade's 6.8%. The earlier decade had a crisis year of its own, and without 2008-09 it averaged 7.2%, so with each period's crisis year set aside the two are within a tenth of a point of each other. Setting aside the rebound years as well, 2009-10 and 2021-22, gives 7.15% for the UPA decade and 7.05% for the NDA period. The growth row therefore depends on how the crisis years are treated. The series joins two bases, 2011-12 (with its back-series for the earlier years) and 2022-23 from 2023-24, and the 2015 rebasing first moved 2013-14 from 4.7% to 6.9%, so a gap of under one point is smaller than past revisions.

MoSPI national accounts; World Bank for 2004-2013. Year shown is the fiscal year that starts in it.

Finding 2 of 6

Inflation is the one measure on which the later period does better

Consumer prices rose 4.8% a year on average after 2014, against 8.0% in the decade before, when inflation reached double digits in 2009, 2010 and 2013. Inflation targeting began in 2016, and the oil-price collapse of 2014-16 came at the start of the later period. The site's inflation explainer sets out how much of the fall each explanation can carry.

World Bank for 2004-2013; MoSPI CPI (Combined) after. Calendar years; 2026 is January to September on the new 2024 base.

Finding 3 of 6

The top 1% took a larger share of income than at any time since 1922

The richest 1% of adults received 22.6% of national income in 2022-23, up from 18.4% in 2004, the highest in the World Inequality Lab's series, which begins in 1922. The bottom half's share fell from 18.8% to a low of 13.9% in 2017 and was 15.0% in 2022-23. No annual figure has been published after 2022.

Bharti, Chancel, Piketty and Somanchi (2024), World Inequality Lab Working Paper 2024/09.

Finding 4 of 6

Press freedom and democratic quality fell furthest

India ranked 157 of 180 on the RSF index in 2026, against an average rank of about 120 in the UPA decade, when RSF ranked between 167 and 179 countries. The Democratic Quality Index, which combines V-Dem, Freedom House and RSF scores, fell from 0.49 in 2014 to 0.28 in 2026. V-Dem first classified India as an electoral autocracy in its March 2021 Democracy Report.

Reporters Without Borders, yearly index files. RSF published no separate 2011 index; its 2011-12 rank is used for both years.

Finding 5 of 6

Official statistics were withheld, delayed or disputed in ways they were not before 2014

The Statistical Suppression Index scores documented events: the census postponed in 2020, the 2017-18 consumption survey scrapped in 2019, the jobs survey held back before the 2019 election, the merger of the statistical offices and the COVID death count. It is zero for every UPA year, peaked at 9.0 in 2021-22 and was 4.65 in 2026, when the census, now on a notified schedule, counts as a delay. The severities are the author's coding and the most contestable part of the study.

Author's coding of dated events, set out in the methodology and the research package.

Finding 6 of 6

The fiscal index is higher after 2014 because of GST and the 2020 borrowing conditions

The Fiscal Centralisation Index averaged 0.30 in the UPA decade and 0.43 after 2014, with a peak of 0.87 in 2020. On the four measured fiscal ratios alone the earlier period was slightly more centralised (0.49 against 0.45), and states' share of gross tax revenue rose after the 14th Finance Commission. The higher score after 2014 comes from the two coded components: GST, which folded most state indirect taxes into a shared tax (states still levy VAT on petroleum and alcohol, stamp duty and electricity duty), and the borrowing conditions of 2020-21.

Finance Commission reports, Union Receipt Budgets, RBI Handbook Table 94; GST and borrowing coded.

The three indices

One method applied to both periods

All three are computed for every year from 2004 to 2026 under the same rules for both periods. The inputs and their sources are set out in the methodology, and the research package holds the code that computes them. The SSI severities are the author's coding against a published rubric, and a different coder could score some events differently.

SSI

4.65of 10, 2026

Statistical Suppression Index. Graded severities of documented events across six streams, with a decay after an event is resolved. UPA average 0.00, NDA average 4.78.

FCI

0.35of 1, 2026

Fiscal Centralisation Index. Four published fiscal ratios and two coded events, GST and the 2020 borrowing conditions. UPA average 0.30, NDA average 0.43.

DQI

0.28of 1, 2026

Democratic Quality Index. The geometric mean of V-Dem's liberal democracy and civil society scores, Freedom House and RSF. UPA average 0.57, NDA average 0.37.

What would change this

What would change these conclusions

The paper is descriptive and identifies no causal effect. Global conditions, demographic change and technology could move both the institutional and the economic series. Below is the evidence that would weaken each finding. The paper sets out every case in full in section 9, section 6.10.2 and section 10.7.

Growth

The growth row is the one most sensitive to treatment. With each period's crisis year left out the two periods are level to within a tenth of a point. The national accounts were rebased in 2015 and again in 2026, and past revisions have been larger than the present gap, so a further revision could move this row either way.

Statistical suppression

The zero for every UPA year is the most contestable number in the study. A documented UPA-era event that meets the index's threshold would raise it, and the published code lets a critic recompute the index. A generous re-coding that gave the UPA decade an average of 2.0 would still leave 2.0 against 4.78.

Fiscal centralisation

On the four measured fiscal ratios alone the UPA years average 0.49 and the NDA years 0.45, so the NDA years are slightly less centralised. The higher NDA index comes from counting GST and the 2020 borrowing conditions as centralisation, and a reader who rejects that reading rejects this finding.

Democracy and press freedom

A demonstration that the movements in the V-Dem, Freedom House and RSF scores come from drift in the instruments would weaken the democracy result. It would also have to account for an independent synthetic-control study, Grier and Grier (2026), which finds the same direction by a different method.

Income and jobs

The income shares are World Inequality Lab estimates, which are revised from time to time, and no annual figure exists after 2022. The jobs finding rests on a direct survey measure. It would weaken if later labour force survey rounds showed regular-job coverage with social security back above its 2017-18 level of 11.5%.

Inflation

Inflation is the one measure that improved. A consistent core-inflation series, which leaves out food and fuel, showing a fall of comparable size would be strong evidence of something beyond the oil-price cycle. The paper does not carry such a series back to 2004 and has not run that test.

Read further

The rest of the site

Explore the dataThe interactive dashboard: every series, the scorecard, scenario lab and trajectories.Read the paperGrowth Without Accountability, the full working paper (HTML, with a PDF).The walkthroughEighteen slides that take the argument one step at a time.AnalysesShort pieces, each on one finding and the data behind it.The inflation explainerWhy the one measure that improved improved, and what that can and cannot show.What GDP does not countUnpaid household work, measured in the Time Use Survey and missing from GDP.Data and downloadsThe dataset, replication code, technical appendix and briefs.Corrections and updatesEvery change to the data and the findings, dated, with the reason.