Table 3. Unemployment rate at the mother city of residence and birth outcomes, robustness checks.
| Ignored as missing | Same as original | Without controlling for race | |||||||
|---|---|---|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | |
| Birthweight | LBW | VLBW | Birthweight | LBW | VLBW | Birthweight | LBW | VLBW | |
| Unemployment rate months 7 to 9 before birth | -0.5242 | 1.0015 | 1.0108 | -0.5014 | 1.0019 | 1.0083 | -0.6140 | 1.0034 | 1.0100 |
| 95% Confidence Interval | (-2.4335–1.3851) | (0.9943–1.0088) | (0.9858–1.0365) | (-2.4040–1.4012) | (0.9948–1.0091) | (0.9835–1.0338) | (-2.4610–1.2331) | (0.9959–1.0110) | (0.9846–1.0360) |
| P value | 0.5774 | 0.6827 | 0.4004 | 0.5927 | 0.5987 | 0.5140 | 0.5005 | 0.3714 | 0.4437 |
| Unemployment rate months 4 to 6 before birth | 0.5697 | 0.9992 | 0.9822 | 0.4524 | 0.9994 | 0.9840 | 0.3750 | 1.0010 | 0.9818 |
| 95% Confidence Interval | (-1.0923–2.2317) | (0.9895–1.0090) | (0.9566–1.0085) | (-1.2511–2.1559) | (0.9897–1.0092) | (0.9586–1.0100) | (-1.3577–2.1077) | (0.9911–1.0109) | (0.9578–1.0064) |
| P value | 0.4873 | 0.8746 | 0.1821 | 0.5898 | 0.9084 | 0.2258 | 0.6601 | 0.8443 | 0.1451 |
| Unemployment rate months 1 to 3 before birth | -1.6919* | 1.0055 | 1.0268** | -1.7921** | 1.0061 | 1.0278** | -1.7360** | 1.0038 | 1.0263** |
| 95% Confidence Interval | (-3.4614–0.0776) | (0.9930–1.0182) | (1.0016–1.0528) | (-3.5706 - -0.0137) | (0.9936–1.0188) | (1.0030–1.0532) | (-3.3702 - -0.1018) | (0.9903–1.0175) | (1.0001–1.0532) |
| P value | 0.0601 | 0.3909 | 0.0372 | 0.0484 | 0.3420 | 0.0278 | 0.0382 | 0.5806 | 0.0493 |
| Observations | 2,457,972 | 2,457,972 | 2,457,972 | 2,474,646 | 2,474,646 | 2,474,646 | 2,594,223 | 2,594,223 | 2,594,223 |
Note: The table shows the association between the mean local unemployment rate in the months 1 to 3, 4 to 6 and 7 to 9 before birth in the city where the mother resided and birth outcomes. Standard errors are clustered at the city level. LBW: low birthweight (< 2500 grams). VLBW: very low birthweight (< 1500 grams). All regressions included as covariates binary variables indicating: the age, marital status, education, and previous children alive or not of the mother; as well as fixed effects for mother city of residence, month of birth and year of birth, and two-way interactions between the fixed effects. Columns 1 to 6 also included the race of the mother as a covariate. In columns 1 to 3, all observations with race, marital status or education classified as ignored were recoded as missing values. In columns 4 to 6, all “ignored” and missing values in those three variables were kept as in the original data. S2 Table presents an extended version of Table 3, including the coefficients estimated for all covariates, except interaction terms.
* p value < 0.1
** p value < 0.05
*** p value < 0.01.