Table 3.
Coefficient estimates from linear probability models of first internal migration in 93 agricultural communities in the Mexican Migration Project testing heterogeneity of precipitation effects
| Variables | Precipitation effects by household wealth |
Precipitation effects by community irrigation |
|---|---|---|
| (1) | (2) | |
| Precipitation (ref: normal) | ||
| Dry in (t-1) | 0.0010(0.0006)∗ | 0.0012(0.0006)∗ |
| Household properties (ref: none) | ||
| Medium | 0.0006(0.0007) | 0.0011(0.0008) |
| High | -0.0013(0.0010) | -0.0008(0.0010) |
| Precipitation x Properties | ||
| Dry in (t-1) x Medium | 0.0048(0.0024)∗ | |
| Dry in (t-1) x High | 0.0049(0.0040) | |
| Community has no irrigation (0/1) | 0.0032(0.0024) | 0.0029(0.0023) |
| Precipitation x No community irrigation | 0.0065(0.0027)∗∗ | |
| Weather indicators | yes | yes |
| Controls | yes | yes |
| State x year fixed effects | yes | yes |
| N (person-years) | 436,978 | 436,978 |
| R2 | 0.010 | 0.010 |
∗p<0.1, ∗∗p<0.05, ∗∗∗p<0.01. Standard errors (corrected for clustering at the community level) are in parentheses. Precipitation (temperature) deviation equals rainfall (maximum number of consecutive days over 30°C) in a community during corn season last year minus the mean value in community in 1980-1990, divided by standard deviation in that period. A community-year is wet (dry) if rainfall is one standard deviation or higher (lower) than its baseline mean, and normal otherwise. Temperature deviation categories are computed similarly. Corn season is June–February in Yucatan; September–March in Baja California, Chihuahua, Nayarit, Sinaloa and Sonora; and May–December in other states. All models control for weather (dry, wet, hot, cool), individual (age, sex, whether person in household head, years of education), household (former internal and US migrants), community characteristics (share in agriculture, share with internal and US migration experience) and state-by-year fixed effects.