Weather fluctuations, such as those caused by the El Niño Southern Oscillation (ENSO), add to the riskiness associated with agricultural production. Improved predictive capacity may help ameliorate negative impacts of climate and weather shocks on agriculture, but it is possible that the benefits of an improved forecast will be distributed unevenly. In particular, poor farmers may not have access to improved forecasts, or they may not have the means to adapt to new weather information.

This paper uses a stochastic computable general equilibrium (CGE) model to examine the distributive effects of improved forecasting of ENSO in Mexico. The particular focus is on agriculture, one of the most vulnerable sectors in the face of ENSO, as well as a sector which provides income to many of the country's poorest households. The model is used to investigate the responsiveness of various sectors of the economy under different degrees of improved predictive capacity and improvements in agricultural technology.

The CGE model used in this study is augmented with a stochastic component, which allows us to simulate a range of stochastic shocks using Monte Carlo methods. With this framework we can compute the mean values and variances of key variables, such as production levels and incomes under stochastic shocks. Given that the model is highly nonlinear, Monte Carlo methods provide information on the sources of volatility in the economy and the built-in shock absorbers that help dampen that volatility.

The results show that while agricultural losses are small as a share of the overall economy, improved forecasting techniques can eliminate these losses. ENSO events harm some regions – particularly the Central, Pacific South, and South East regions – more than others. Agricultural production in these regions benefits the most from improved forecasting. Since these regions also are the regions with higher poverty, they should be targeted by policy makers who are concerned with alleviating the effects of ENSO events on the poor. The simulations also show that poor households are the least able to take advantage of improvements in forecasting, since at higher levels of preparedness agricultural production shifts to sectors from which poor households receive less income. Finally, in Mexico, ENSO events contribute only a small share of overall variability in agriculture. It might be better to focus efforts on the latter problem, in terms of improved agricultural seeds, extension services, and schemes to protect already fragile lands.