Table 23 Labor Force Participation and Employment Bivariate Probit with Selection Results SIPP Combined Years 19861997

Regressor

Labor force participation equation

Employment equation disable = 1

disable x post = 1

Log-likelihood Number of observations

-0.9404*** (0.0105) 0.0293*** (0.0048) -0.1360*** (0.0129)

NOTE: Additional regressors included age; age squared; state unemployment rate;

female, nonwhite, education, regional dummy variables; an indicator for SMSA residence (employment); and non-labor income and marital status (labor force participation). Standard errors are in parentheses.

*** = significant at the 99 percent confidence level.

nondisabled. In addition, the relative decline in labor force participation among the disabled found in the CPS data is also seen using the SIPP data as well.

Along with the reestimation of Equations 2.3 and 2.4, a specification is estimated in which the impact of having a disability post-ADA is allowed to vary by type of impairment:20

(2.5) lfp, = a + yx'xu + (3f musculoskeletal, + ( internal,.

+ (M mental,. +PP other,. + ^jpost,. + 0f musculoskeletal,. X post, + 0{ internal, X post,. + 0M mental, X post, + 00 other,. X post, + eu

(2.6) emp,. = a2 + y2'x21 + (3f musculoskeletal, + (2 internal,.

+ (M mental,. + (0 other,. + ^2post,. + 0£ musculoskeletal, X post,. + 02 internal,. X post, + 0M mental, X post,. + 00 other, X post,. + e2i where lfp, is equal to 1 if person i is in the labor force, 0 otherwise, empi is equal to 1 if person i is employed, 0 otherwise, X is a set of covariates for each person (individual demographic characteristics), musculoskeletal, is equal to 1 if person i has a musculoskel-etal disability,21

iNTERNALi is equal to 1 if person i has a disability involving the internal systems, mental, is equal to 1 if person i has a mental disability, other, is equal to 1 if person i has a disability classified as ''other,'' and post y is equal to 1 if person i is observed in 1992 or later.

Again, these equations are estimated via maximum likelihood as a bi-variate probit with selection, where emp, is only observed if lfp, = 1. In this framework, the type of disability is controlled for by dummy variables indicating whether the individual has a musculoskeletal, internal systems, mental, or other disability; and the time period is controlled for by a dummy variable indicating whether the ADA had been implemented yet or not. The coefficients of particular interest (01 and

02, j = S,I,M,O), therefore, allow us to calculate the labor force participation and employment changes among disabled workers post- versus pre-ADA relative to the changes for nondisabled workers. Table 2.4 provides selected estimated coefficients and regression details.

The estimation results presented in Table 2.4 from the SIPP data set are also consistent with the conclusions drawn using the CPS data: labor force participation declined more for all classifications of disability, relative to nondisability, post- versus pre-ADA. However, employ-

Table 2.4 Labor Force Participation and Employment Bivariate Probit with Selection Results by Type of Disability, SIPP Combined Years, 1986-1997

Regressor

Labor force participation equation

Employment equation

musculoskeletal = 1

-0.8253***

- 0.2798***

(0.0150)

(0.0281)

internal = 1

-0.9597***

-0.1660***

(0.0192)

(0.0409)

mental = 1

-1.2722***

-0.1599***

(0.0237)

(0.0487)

other = 1

- 0.8396***

- 0.3429***

(0.0280)

(0.0508)

post (year= 1992 or later) = 1

0.0305***

- 0.0348***

(0.0048)

(0.0077)

musculoskeletal x post = 1

-0.1416***

0.0542

(0.0187)

(0.0348)

internal x post = 1

-0.1137***

0.1693

(0.0247)

(0.0532)

mental x post = 1

- 0.0850***

0.1187**

(0.0282)

(0.0573)

other x post = 1

- 0.0829**

0.1484**

(0.0360)

(0.0675)

Rho

- 0.4799***

(0.0205)

Log-likelihood

-292,164

Number of observations

500,560

NOTE: See notes to Table 2.3 regarding additional regressors. Standard errors are in parentheses.

*** = significant at the 99 percent confidence level. ** = significant at the 95 percent confidence level.

NOTE: See notes to Table 2.3 regarding additional regressors. Standard errors are in parentheses.

*** = significant at the 99 percent confidence level. ** = significant at the 95 percent confidence level.

ment probabilities (controlling for labor force participation) increased significantly more for the disability classifications of mental and other than for the nondisabled, post- versus pre-ADA. Relative employment probabilities did not change significantly for those with mus-culoskeletal or internal disabilities.

While it is difficult to interpret the employment impact for those with disabilities classified as other, the major role that those with mental disorders play in explaining the overall relative employment improvement is not surprising, given the attention paid to and policies developed for those with mental disabilities in recent years.22 In addition, if we expect costs of accommodation to influence employment outcomes of the disabled, these results might suggest that accommodating workers with mental disabilities (such as through flexible work scheduling) has been relatively less expensive for employers than accommodating workers with musculoskeletal or internal disabilities (for example, through physical modification of the work environment).

Was this article helpful?

0 0

Post a comment