Modi vs Manmohan Singh on GDP with Keshav Bedi

What this is

Comparisons of India's GDP growth across governments circulate constantly, and almost all of them differ from each other. This workbook explains why. It rebuilds the comparison from the World Bank's World Development Indicators series (NY.GDP.MKTP.KD, GDP in constant 2015 US$, updated July 2026) and isolates each choice an analyst has to make, showing how much each one moves the result.

Every number is a live formula traced back to source data. Change a window year or a level and the entire workbook recalculates. Nothing is hardcoded except the source series itself, which is colour-coded so you can see what came from the World Bank and what was derived.

What's inside — seven sheets

1. Data. Annual real GDP for India, the World, Low & middle income, China, Lower middle income, and four derived peer aggregates, 2002–2025, with growth rates and excess-over-benchmark columns.
2. Headline growth. CAGR across four window definitions, with the tenure overlap at FY2014-15 handled explicitly.
3. Crisis adjustment. Six treatments of FY2020-21 and FY2008-09, from raw through matched-sample. Includes the COVID level-effect counterfactual: what the FY2025-26 output level would have been without the contraction.
4. Peer benchmarks. India measured against four peer groups, including Lower middle income excluding India, plus sensitivity tests for graduation bias and China's classification.
5. Rolling CAGR. Ten-year rolling growth, 2013–2025, as an endpoint-sensitivity check.
6. Adjustments register. Ten adjustments ranked by how much each moves the answer, including the ones this dataset cannot resolve.
7. Claims tested. Seven common assertions checked against the computed figures.

What it concludes

On raw CAGR the 2004–13 decade is ahead by 0.69pp. Under symmetric crisis treatment the two are close to indistinguishable. On growth relative to comparable economies, 2014–25 is ahead. All three are correct, and they answer different questions. The workbook is built to let you see that rather than to argue a side.

What it does not do

It cannot resolve the 2011-12 base-year revision and the contested back-series, which is flagged as the largest unresolved issue. Per-capita conversion, employment elasticity and terms-of-trade adjustment need data outside this series, and are marked as such rather than approximated.

Who it's for

Analysts, journalists, policy researchers, students and anyone who has to evaluate a growth comparison and wants to know where its numbers came from.