Going Beyond Correlations Introduction HOW does an increase in the relative price of food (RPF) affect rural poverty in India? This question begs an analytical and empirical understanding, and many pages of this journal have been recently devoted to a discussion and debate over this question. Sen (1996) claims that econometric models which include RPF1 along with other explanatory variables such as agricultural productivity and public development expenditure result in a much better explanation of pre-reform and post- reform poverty, than econometric models which ignore RPF. Furthermore Sen shows that RPF does better in tracking poverty than the inflation rate. Ravallion (1998a) uses 24 observations of the NSS rounds from 1958 to 1993-94. and obtains a correlation of 0,76 between poverty (the head count ratio, HCR) and RPF,2 Although this confirms the claim made by critics of the economic reforms of a strong positive correlation between measured poverty and RPF, Ravallion rejects the explanation that this correlation is driven by the adverse distributional effect of changes in relative food prices, and maintains that the correlation is due to the mean effect, via, depressed mean per capita consumption. In rebutting Mohan Rao's [Rao 1998] objection that econometric tests of distributional impact, which ignore d ecile specific changes in real income lead to misleading conclusions, Ravallion (1998b) claims that his conclusions are on solid ground si nee (a) EngeI expenditure curves3 are flat for the poorest four deciles and (b) using the Consumer Price Index for Agricultural Labour (CPIAL) actually overestimates the loss of real income to the poor, since their budget shares on food are lower than What the CPIAL assumes.4 While the last word might not have been said in the debate on the distributional consequence of a rise in RPF, in this paper while we comment on the correlation aspect, we wish to nudge the debate a bit away from the correlation issue and into a territory far less contentious. Specifically we will discuss (a) who among the rural poor are unambiguously hurt by rising RPF-the 'net purchasers of food;;(b) what is the received wisdom on the mechanism by which rising RPF is supposed to benefit the rural poor (the 'terms of trade' and the supply response), and why it might not work, and finally (c) what should be the appropriate mix of policy that goes along with the reforms induced rising food prices. In this paper we take it as an incontrovertible observation that economic reforms have led to an increase in food prices. The explanations for this may be varied5 and sometimes opposites of each other, but the crux of the matter is that economic liberalisation along with structural adjustment has meant that input subsidies are lower, and output (food) prices closer to international prices, which are typically higher than domestic prices.6 One of the sources of controversy surrounding the question of higher food prices and benefits to agriculture, is the theoretical possibility that the resultant surge in demand for rural labour, and hence wages, might more than compensate for the rise in food prices.