Table A.12:
The Long-Run Effect of the 1980–1982 Recession on Education, Robustness to Specification
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | (10) | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Interaction between 1979–1982 decrease in log real earnings per capita and age in 1979 | ||||||||||||
| Panel A. Dependent variable: HS/GED attainment | ||||||||||||
| 0–10 | 0.065 (0.046) |
0.023 (0.044) |
0.020 (0.043) |
−0.020 (0.047) |
0.002 (0.045) |
−0.043 (0.051) |
−0.003 (0.052) |
−0.020 (0.048) |
−0.052 (0.039) |
−0.012 (0.047) |
||
| 11–19 | 0.041 (0.032) |
0.024 (0.036) |
0.033 (0.036) |
0.055 (0.046) |
0.069 (0.041) |
0.026 (0.047) |
0.073 (0.050) |
0.053 (0.047) |
0.072 (0.049) |
0.060 (0.046) |
||
| Panel B. Dependent variable: Any college attendance | ||||||||||||
| 0–10 | 0.085 (0.091) |
−0.027 (0.095) |
0.003 (0.105) |
−0.269 (0.158) |
−0.192 (0.153) |
−0.287 (0.159) |
−0.259 (0.176) |
−0.263 (0.161) |
−0.334 (0.143) |
−0.254 (0.153) |
||
| 11–19 | −0.060 (0.066) |
−0.106 (0.077) |
−0.081 (0.082) |
−0.141 (0.122) |
−0.077 (0.108) |
−0.156 (0.119) |
−0.126 (0.136) |
−0.146 (0.126) |
−0.134 (0.110) |
−0.133 (0.119) |
||
| Panel C. Dependent variable: Any college degree attainment | ||||||||||||
| 0–10 | −0.253 (0.091) |
−0.358 (0.106) |
−0.284 (0.118) |
−0.444 (0.174) |
−0.376 (0.171) |
−0.459 (0.174) |
−0.474 (0.207) |
−0.443 (0.180) |
−0.566 (0.172) |
−0.425 (0.167) |
||
| 11–19 | −0.220 (0.071) |
−0.276 (0.077) |
−0.241 (0.082) |
−0.297 (0.107) |
−0.236 (0.094) |
−0.311 (0.106) |
−0.314 (0.126) |
−0.309 (0.113) |
−0.279 (0.101) |
−0.291 (0.105) |
||
| Panel D. Dependent variable: Four-year college degree attainment | ||||||||||||
| 0–10 | −0.465 (0.118) |
−0.532 (0.139) |
−0.423 (0.145) |
−0.470 (0.179) |
−0.400 (0.182) |
−0.490 (0.182) |
−0.511 (0.216) |
−0.475 (0.188) |
−0.510 (0.147) |
−0.450 (0.173) |
||
| 11–19 | −0.297 (0.082) |
−0.335 (0.093) |
−0.291 (0.098) |
−0.286 (0.107) |
−0.229 (0.099) |
−0.300 (0.107) |
−0.319 (0.132) |
−0.302 (0.113) |
−0.251 (0.091) |
−0.283 (0.107) |
||
| Panel E. Dependent variable: Two-year college degree attainment | ||||||||||||
| 0–10 | 0.212 (0.069) |
0.175 (0.062) |
0.140 (0.063) |
0.025 (0.059) |
0.024 (0.055) |
0.031 (0.056) |
0.038 (0.059) |
0.033 (0.059) |
−0.056 (0.066) |
0.025 (0.058) |
||
| 11–19 | 0.077 (0.047) |
0.060 (0.046) |
0.052 (0.048) |
−0.010 (0.043) |
−0.007 (0.040) |
−0.010 (0.040) |
0.005 (0.045) |
−0.007 (0.042) |
−0.027 (0.042) |
−0.008 (0.043) |
||
| Panel F. Dependent variable: Years of schooling | ||||||||||||
| 0–10 | −0.743 (0.476) |
−1.315 (0.581) |
−0.985 (0.634) |
−1.967 (0.868) |
−1.658 (0.859) |
−2.103 (0.900) |
−1.982 (1.045) |
−1.993 (0.913) |
−2.311 (0.823) |
−1.859 (0.837) |
||
| 11–19 | −0.532 (0.377) |
−0.864 (0.435) |
−0.743 (0.442) |
−1.065 (0.573) |
−0.803 (0.505) |
−1.206 (0.584) |
−1.038 (0.693) |
−1.150 (0.610) |
−0.871 (0.538) |
−1.030 (0.570) |
||
| Birth state-by-age in 1979 fixed effects | X | X | X | X | X | X | X | X | X | |||
| Birth region-by-age in 1979 fixed effects | X | |||||||||||
| Interaction between age in 1979 and | ||||||||||||
| 1950–70 log median family income growth | X | |||||||||||
| 1950–60, 1960–70, 1970–80 log median family income growth | X | X | X | X | X | X | X | X | X | |||
| 1950–60, 1960–70, 1970–80 log population growth | X | X | X | X | X | X | X | X | ||||
| 1960 covariates | X | X | X | X | X | X | ||||||
| 1950, 1960, 1970 covariates | X | |||||||||||
| Full interaction between age in 1979, race, gender, and 1960 covariates | X | |||||||||||
| Recession severity for 1990–91, 2001, and 2007–09 | X | |||||||||||
| Japanese import competition, 1970–90 | X | |||||||||||
| Log births | X | |||||||||||
Notes: The dependent variable is indicated in the panel title. All regressions include fixed effects for age at time of survey-by-gender, race, birth county, and survey year. The 1960 covariates are log population, log population density, percent urban, percent black, percent foreign, percent with a high school degree, and percent of families with income below $3,000. The 1950 and 1970 covariates are defined similarly, except for 1950 the percent of families with income below $2,000 is available. To control for log births, I include the log number of births of a county’s residents and the log number of births that take place in that county. Recession severity measures are the change in log earnings per capita and change in log employment for 1989–1991, 2000–2002, and 2007–2010. The commuting zone-level change in Japanese import competition comes from Batistich and Bond (2019). Column 4 is the baseline specification reported in Table 2. See notes to Table 2 for details on sample.
Sources: BEA Regional Economic Accounts, Census County Business Patterns, Confidential 2000–2013 Census/ACS data linked to the SSA NUMIDENT file