The long-standing policy debate on whether to raise public expenditures on agriculture presumes the question is binary—should we
spend more or not? Using a panel of 87 low- and middle-income countries (1990-2020) and a novel fiscal allocation–impact frontier
model, we show that the salient issue is not the level of spending but its composition, timing and accountability mechanisms. A 1 %
rise in “effective” agricultural expenditure—defined as disbursements that reach frontline agencies within the same fiscal year and
are subjected to third-party audits—yields a 2.3 % increase in total-factor-productivity (TFP) growth and lifts 0.8 % of the rural

population above the poverty line. Conversely, identical nominal increases in “ineffective” spending (characterised by delayed
transfers, weak procurement and low transparency) generate no measurable gains. Simulations indicate that reallocating just 15 % of
current ineffective outlays to effective categories can achieve the same TFP and poverty impacts as doubling total agricultural
budgets under the status quo mix. The paper therefore reframes the policy question from “how much” to “what for and how”. We
provide a parsimonious diagnostic scorecard that ministries of finance and agriculture can use to shift resources toward expenditure
categories with the highest marginal social returns, offering a politically viable path to agricultural transformation without necessarily
increasing the fiscal envelope.