A useful thumb rule:
The government should only spend an additional one rupee when the gains to society from that action exceed ₹3.
Origin
From In Service of the Republic: The Art and Science of Economic Policy (2019) by Vijay Kelkar and Ajay Shah. The underlying concept is the Marginal Cost of Public Funds (MCPF).
What it says
Taxation is not frictionless. Every rupee the government raises costs society more than a rupee because of deadweight loss — exchanges that disappear due to tax-induced price distortions. In OECD countries, the MCPF ranges from 1.25 to 2. In India, bad taxes, exemptions, and distortions push the MCPF to roughly ₹3. That means the bar for public spending is high: only fund activities where the marginal social benefit exceeds ₹3 per rupee spent. Public goods with large externalities — health, defence, basic research — clear the bar; private goods do not. As a corollary, governments should avoid spending on private goods where market provision is viable.
Applied
- When evaluating whether the government should continue operating loss-making PSUs or exit.
- When deciding whether to subsidise a private good or invest in a public good with large positive externalities.
- When setting the threshold for cost-benefit analysis of new public projects in India.
- When reviewing ongoing subsidies or transfers that may have once been justified but now deliver social returns well below the MCPF.
- When arguing for tax simplification — reducing exemptions and bad taxes lowers the MCPF and expands the fiscal space for legitimate public goods.
When it falls short
- The ₹3 estimate is a rule of thumb, not a precise figure; it varies by tax instrument and over time.
- It does not capture political economy constraints; even inefficient spending may be necessary to maintain coalition support.
Further reading
Originally explored in A Framework a Week: A Useful Thumb Rule on Anticipating the Unintended.