Average predicted probability

Being in the labor |
Being in the labor | |||

force and employed |
Employment |
force and employed |
Employment | |

(1) |
(2) |
(3) |
(4) | |

Year |
Prob(lfp = 1, emp = 1) |
Prob(emp = 1) |
Prob(lfp= 1, emp = 1) |
Prob(emp = 1) |

1981 |
0.64271 |
0.84244 |
0.47840 |
0.79271 |

1982 |
0.62609 |
0.81519 |
0.47742 |
0.77384 |

1983 |
0.61727 |
0.79255 |
0.47474 |
0.76856 |

1984 |
0.64824 |
0.85387 |
0.50693 |
0.82802 |

1985 |
0.65675 |
0.85665 |
0.50837 |
0.82475 |

1986 |
0.66467 |
0.85364 |
0.51167 |
0.81465 |

1987 |
0.67498 |
0.86818 |
0.51993 |
0.83475 |

1988 |
0.68827 |
0.87833 |
0.53576 |
0.82908 |

1989 |
0.69070 |
0.88346 |
0.53884 |
0.85364 |

1990 |
0.69150 |
0.87427 |
0.52609 |
0.82427 |

1991 |
0.67530 |
0.84653 |
0.52123 |
0.81285 |

1992 |
0.67426 |
0.82397 |
0.51998 |
0.76743 |

1993 |
0.67738 |
0.83780 |
0.53733 |
0.80476 |

1994 |
0.68448 |
0.84973 |
0.49882 |
0.81405 |

1995 |
0.68820 |
0.87372 |
0.49520 |
0.83877 |

1996 |
0.69305 |
0.87742 |
0.50769 |
0.84971 |

1997 |
0.70309 |
0.87734 |
0.51408 |
0.84607 |

1998 |
0.71169 |
0.89380 |
0.51334 |
0.86027 |

1999 |
0.71529 |
0.89326 |
0.50755 |
0.83657 |

2000 |
0.72085 |
0.91141 |
0.52832 |
0.86493 |

NOTE: Estimates obtained from a bivariate probit model with selection. Regressors for both labor force and employment determination included the state unemployment rate; age; age squared; female, nonwhite, education, and regional dummies; and a disability dummy. The labor force participation equation also included nonlabor income, marital status, and a worked-last-year indicator. The employment equation included the number of weeks worked last year. The probabilities for each column are the average across the entire sample obtained by calculating the probability for each person (varying the disability dummy variable between 0 and 1), then averaging across the sample.

Table C.4 Relative Predicted Probabilities of Working in Each Firm Size, CPS 1987-1999

Ratio of predicted probabilities of a disabled worker relative to the predicted probability for a nondisabled coworker being employed by a firm by size

Year |
Small firm |
Medium firm |
Large firm |

1987 |
1.1842 |
1.0756 |
0.9189 |

1988 |
1.1836 |
1.1063 |
0.9134 |

1989 |
1.1524 |
1.0403 |
0.9409 |

1990 |
1.1372 |
1.0174 |
0.9507 |

1991 |
1.1682 |
0.9417 |
0.9493 |

1992 |
1.2127 |
0.8759 |
0.9501 |

1993 |
1.0873 |
1.0410 |
0.9512 |

1994 |
1.0567 |
0.9999 |
0.9787 |

1995 |
1.1772 |
0.9384 |
0.9399 |

1996 |
1.1811 |
0.8598 |
0.9575 |

1997 |
1.1341 |
0.9900 |
0.9511 |

1998 |
1.0110 |
0.9319 |
1.0106 |

1999 |
1.0963 |
0.9840 |
0.9675 |

Change over | |||

time period |
-0.0879 |
-0.0917 |
0.0487 |

NOTE: Small firms have fewer than 25 employees; medium firms have 25-99 employees; and large firms have 100 or more employees. Regressors (in addition to a disability dummy variable) in the multinomial logit regression included age; age squared; and regional, education, female, nonwhite, and central city dummy variables.

Was this article helpful?

## Post a comment