Alabaster Research

India's agricultural value chain: who gains, who stays poor

Agent-based model, version 6. 10,000 rural households over 50 years, real FY2025-26 rupees. Each line is the mean of 30 seeded runs. Shaded bands are 95% confidence intervals. The model shows what follows from its assumptions. It is not a forecast.

Amounts: G01 G03 G07 G09a
Source and range: E01 E03b
Applies to the current package under historical climate.
 
Tiles and the Money Flow and Agent Grid tabs follow this year.

Inequality & Asymmetry

Households below the poverty line

Share of households, %. Consumption against the World Bank $3.00 and $4.20 lines (2021 PPP).

Poverty by household type

Share below the $4.20 line, %.

Inequality

Gini coefficient of household consumption and of income after transfers.

Income shares

Share of total income, %.

Farmer's share of the consumer rupee

Farm revenue as % of consumer food spending.

Mean income by class

Rs thousand per household per year, log scale.
Income shares here are for a rural sample without top urban incomes. They cannot be compared with World Inequality Lab national shares (top 10%: 57.7% in 2022-23).

Savings Dynamics

Financial savings and debt by class. Exits are farm households that left farming after debt stayed above the exit threshold.

Mean financial savings by class

Rs thousand per household, log scale.

Mean debt by class

Rs thousand per household.

Farm indebtedness

% of farm households with debt, and institutional share of their debt, %.

Farm households that exited farming

Cumulative count out of 5,400 farm households.

Wealth shares

Share of net worth (land, savings, less debt), %.

Households by class

Count. Non-farm households move between tiers as consumption crosses the $4.20 line and four times that line.

Money Flow

Every rupee of consumer food spending is paid to a farmer, a trader or a processor. Year .

Where Rs 100 of consumer food spending goes

Value chain flows over time

Rs thousand per household in the model per year.

Government spending by scheme

Rs thousand per household in the model per year.
Institutional interest, trading costs and processing costs leave the model. The rest of the economy is not represented.

Agent Grid

500 households from one run (seed 1, current package, historical climate). Year . Square size shows consumption against the $4.20 line. A crimson outline marks debt above usual annual income.

Households by class, year

Share of the 500 households in the grid, %. Axis runs to %.

Sources & Assumptions

Every parameter in the model code carries the ID shown here. Verified means the value was checked against the publisher or a report of the primary source in this project. Calibrated means the value was set to hit a cited target. Modelling choice means no defensible source was found; these are tested in the sensitivity analysis. Money is in real FY2025-26 rupees, deflated with CPI-Combined annual averages (RBI Handbook of Statistics). Table and page references for the survey figures are still owed for entries marked as verified via secondary report.