## Representation and Growth

A high-growth indicator for each occupation (and industry) in the sample is constructed as follows:

(1 / K) = |^emp]+1 - emp] ) / emp] 0 otherwise where occupation (or industry) * is defined as high-growth if the percentage change in employment between year t and t + 1 for that occupation exceeds the average percentage change in employment in all occupations represented in the sample (K corresponds to the total number of occupations or industries). The probability of being employed in a high-growth occupation (and/or industry) is determined as a function of individual characteristics, including disability status.

The sample was split into three time periods, and the employment growth of each occupation and industry represented in the sample was determined by a source external to the data file.16 Table 4.4 contains the growth rates of each occupation and industry represented in the sample. For example, for the 1983-1989 period, service occupations, managerial and professional specialty, and technical, sales, and administrative support are considered ''high growth,'' since their growth exceeds the average for all occupations. Similarly, for the same period, mining and construction; transportation and public utilities; retail trade; finance, insurance, and real estate; services; and public administration are all considered high-growth industries.

Simple probit models were estimated to determine whether disabled workers are more or less likely to be employed in growing occupations or industries:

where individual /'s job is in a growing occupation/industry if Y > 0. Since Y is unobserved, a binary variable, Yp, is defined as

The set of parameters, p, were estimated via maximum likelihood probit. X comprises various individual and job characteristics for

Table 4.4 Employment Growth Rates for Industry and Occupational Classifications

Growth rate (%) 1983-89 1987-93 1991-98

Occupation

Growth rate (%) 1983-89 1987-93 1991-98

Occupation

 Managerial and professional specialty 28.8 16.2 25.9 Technical, sales, and administrative support 15.6 5.6 6.1 Service 12.3 11.7 9.7 Precision production, craft, and repair 12.1 - 1.0 8.8 Operators, fabricators, and laborers 12.0 - 0.8 4.6 Farming, forestry, and fishing -7.5 - 3.6 - 0.1 Average Growth Rate 12.2 4.7 9.2 Industry Agriculture -9.7 - 2.9 3.3 Mining and construction 18.8 - 3.9 16.1 Manufacturing 8.6 -5.8 0.7 Transportation and public utilities 15.8 8.2 13.0 Wholesale and retail trade 14.6 7.5 11.4 Finance, insurance, and real estate 22.7 2.7 10.2 Services 23.6 17.7 18.4 Public administration 17.9 10.2 4.1 Average growth rate 14.0 4.2 9.7

SOURCE: Author's calculations from Jacobs (1998).

SOURCE: Author's calculations from Jacobs (1998).

worker i, including a dummy variable indicating whether the worker is disabled or not. The model was estimated on a sample of workers only; therefore, the results are generalizable to workers only. The marginal effect of disability on being employed in a high-growth occupation or industry was calculated as the marginal benefit for each worker, then averaged over the entire sample. Table 4.5 presents the estimated marginal effects of a work-limiting disability on having a job in a high-growth occupation or industry.

In each of the three years analyzed, the probability of a disabled worker being employed in a high-growth occupation was from 2 to 5 percentage points less than the probability of a nondisabled worker being employed in a high-growth occupation. In addition, the marginal (negative) effect was the highest during the post-ADA years, suggesting that the ADA has not improved the opportunity of disabled workers to move into high-growth occupations. On the other hand, disabled

 dPROB (HIGH GROWTH dPROB (HIGH GROWTH Year occupation)A5disable industry)A5disable 1989 -0.0299 0.0123 (0.0092) (0.0108) 1993 -0.0194 0.0469 (0.0085) (0.0121) 1998 -0.0453 0.0261 (0.0140) (0.0129) All three years - 0.0381 0.0268 (0.0068) (0.0070)

NOTE: Probit estimations included the following regressors, in addition to a disability dummy variable: age; age squared; regional, education, marital, female, and nonwhite dummy variables; and occupation (for the industry probit) and industry (for the occupation probit) dummy variables. Standard errors are in parentheses.

NOTE: Probit estimations included the following regressors, in addition to a disability dummy variable: age; age squared; regional, education, marital, female, and nonwhite dummy variables; and occupation (for the industry probit) and industry (for the occupation probit) dummy variables. Standard errors are in parentheses.

workers have been more likely to be employed in high-growth industries. Unfortunately, a worker's industry does not reflect as much on an individual's job opportunities as one's occupation does. For example, for someone with skills suited to a secretarial job, a decline in manufacturing as an industry is not as devastating to the person's opportunities as a decline in the administrative support occupation. Of course, occupational representation within an industry, such as the proportion of those in the precision production occupation in manufacturing industry, could be an important consideration.