India has recorded high levels of unemployment and low labor force participation rates in recent years even before the onset of the COVID-19 pandemic and the lockdown. How does an episode of unemployment or loss of income affect household consumption expenditure is an important question for designing effective safety nets. We use data on household-specific episodes of job loss and decline in income, from an earlier year (March-April 2019) to estimate the household response to employment shocks. We apply diff-in-diff and quantile regressions to a high-frequency panel data from a nationally representative survey of 1,75,000 households to estimate the impact of a job loss (and change in income) on household consumption expenditure—for urban and rural households, and households across different expenditure levels. We find that loss of employment of an earning member leads to a significant immediate decline in household consumption expenditure. The decline is much larger for urban households and households in the lowest and the highest deciles of monthly per capita. Durable expenses go down the most. Expenditure on health and education also goes down significantly and there is evidence of adjustments in discretionary expenses too, especially for urban households. For households with only one earning member, borrowing does not increase after the job loss, suggesting credit constraints. Government cash transfers help rural households, as the beneficiaries show a smaller reduction in consumption expenditure after the shock. Our findings highlight the high vulnerability of urban households to economic shocks and can inform the design and targeting of income support and other safety-net programs in India and other developing countries.