Working Paper: NBER ID: w19808
Authors: Mark R. Rosenzweig; Christopher Udry
Abstract: We look at the effects of rainfall forecasts and realized rainfall on equilibrium agricultural wages over the course of the agricultural production cycle. We show theoretically that a forecast of good weather can lower wages in the planting stage, by lowering ex ante out-migration, and can exacerbate the negative impact of adverse weather on harvest-stage wages. Using Indian household panel data describing early-season migration and district-level planting- and harvest-stage wages over the period 2005-2010, we find results consistent with the model, indicating that rainfall forecasts improve labor allocations on average but exacerbate wage volatility because they are imperfect.
Keywords: rainfall forecasts; agricultural wages; migration; NREGA; India
JEL Codes: J2; J31; J43; O1; O13; Q12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Forecast of above-average rainfall (Q47) | Reduction in village outmigration (R23) |
| Reduction in village outmigration (R23) | Decline in planting stage wages (J39) |
| One standard deviation increase in forecasted rainfall (C53) | 0.7 percentage point reduction in migration (J69) |
| Actual rainfall shocks (Q54) | Increase in planting stage wages (J39) |
| NREGA implementation (D78) | Increase in planting stage wages (J39) |
| Bad rainfall shock following a good forecast (Q54) | Decline in harvest stage wages (J39) |
| Increased labor supply due to reduced outmigration (J69) | Decline in harvest stage wages (J39) |
| Decreased demand for labor by farmers facing adverse weather conditions (J43) | Decline in harvest stage wages (J39) |