Skip to main content

Full text of "The economic cost of unemployment and underemployment"

See other formats


Historic, archived document 

Do not assume content reflects current 
scientific knowledge, policies, or practices. 











Reserve 

aHD5709 

.2 

• U5Y4 


United States 
Department of 
Agriculture 


Economic 

Research 

Service 


Agriculture 
and Rural 
Economy 
Division 


The Economic Cost of 
Unemployment and 
Underemployment 

Mervin J. Yetley 


An-as Bookplate 


NATIONAL 



LIBRARY 


Want Another Copy? It’s Easy. 

Just dial 1-800-999-6779. Toll free. 

Ask for The Economic Cost of Unemployment and Underemployment (AGES 89-17). 

The cost is $5.50 per copy. For non-U.S. addresses, add 25 percent (includes 
Canada). Charge your purchase to your VISA or MasterCard, or we can bill you. Or 
send a check or purchase order (made payable to ERS-NASS) to: 

ERS-NASS 
P.O. Box 1608 
Rockville, MD 20850. 


We’ll fill your order by first-class mail. 












THE ECONOMIC COST OF UNEMPLOYMENT AND UNDEREMPLOYMENT. By Mervin J. Yetley, 
Agriculture and Rural Economy Division, Economic Research Service, U.S. 
Department of Agriculture. ERS Staff Report No. 89-17. 


ABSTRACT 


Both underemployed and unemployed workers involuntarily forgo personal 
earnings. Those forgone earnings are viewed as a cost to the national 
economy. For nearly all workers, lost earnings from underemployment are 
larger than for unemployment. Rural workers lose relatively more earnings to 
underemployment than do urban workers. The procedure used to estimate the 
magnitude of lost earnings is described in this report. 

Keywords: unemployment, underemployment, labor distress, lost wages, lost 
personal earnings, cost of labor distress. 


This report was reproduced for limited distribution to the 
research community outside the U.S. Department of Agriculture. 



t 


1301 New York Avenue, NW. 
Washington, DC 20005-4788 


May 1989 


iii 



CONTENTS 


Page 


SUMMARY iv 

INTRODUCTION 1 

THE ANALYTICAL FRAMEWORK 2 

THE DATA 6 

THE EMPIRICAL ANALYSIS 7 

THE ESTIMATION RESULTS 9 

RESULTS 10 

DISCUSSION 13 

Interpretation of the Results 18 

Potential Uses of the Results 22 

REFERENCES 24 


SUMMARY 

The magnitude of lost earnings was estimated for workers experiencing labor 
distress: the unemployed and four categories of underemployment (discouraged 
workers, involuntarily part-time employed, the working poor, and those 
occupationally mismatched). The loss of personal earnings due to 
underemployment is greater than for unemployment, with the total loss 
exceeding $205 billion on an annual basis. 

The unemployed in urban areas have the largest individual category loss, while 
this distinction is held by the working poor in rural areas. Females and 
blacks tend to be overrepresented in most categories in both rural and urban 
areas. d&is, happens because they do not contribute proportionately to lost 
earnings' due to lower wage scales. Rural black workers are especially 
affected. 

Information by State indicates considerable variance in the percentage of 
total earnings lost by the unemployed. Comparative data for selected States 
suggest this percentage ranges from 27.5 to 50.7 percent. These findings show 
that reliance on unemployment statistics alone provides an incomplete 
indicator of the extent of labor distress in the Nation's economy and for the 
rural economy in particular. 


iv 


901232 

The Economic Cost of 
Unemployment and 
Underemployment 

Mervin J. Yetley 

INTRODUCTION 

Agriculture was the dominant rural industry and occupation in the United 
States during the 1950's, with 15 percent of the U.S. population living on 
farms. Curing this same period, rural America had widespread poverty and was 
viewed as economically disadvantaged. These problems were eventually 
recognized, and in the mid-1960's, President Johnson created the National 
Advisory Commission on Rural Poverty to address these issues. 

Rural areas experienced economic revitalization during the late 1960's and 
into the 1970's. New jobs were created as rural communities successfully 
competed with urban areas for manufacturing plants. At the same time, 
agriculture experienced booming export markets and rising land values and 
incomes. The results were rural population growth and the hope of a bright 
future for rural America. 

But, even as the population of most rural communities was steady or increasing 
in the 1970's, agriculture was declining in relative importance as non- 
agricultural jobs increasingly provided the major source of income. By the 
mid-1980's, approximately 40 percent of rural residents lived in counties 
primarily dependent on manufacturing for jobs and income (6, 9) . 1 

The 1980's ushered in renewed economic stress in rural America. A number of 
conditions both internal and external to rural economies and communities 
brought about sharp declines in jobs and incomes from agriculture and related 
businesses, mining, energy, and manufacturing. Rural businesses and banks 
began to fail, and, in many areas, population started to decline. The 
optimism of the 1970's plunged as job opportunities disappeared (1). 

These changed economic and structural conditions resulted in limited, or even 
declining, job opportunities in rural areas. Rural workers needed to upgrade 
their skills to maintain employability or learn new skills to move into new 
jobs. Achieving and sustaining a high-quality workforce capable of 
capitalizing on emerging job opportunities in a changing economy is an 
important element in successful rural development. 

Adequate planning and appropriate implementation of remedial programs requires 
accurate and timely information about workers needing new job opportunities. 
Basic information includes age, education, sex, job skills, and the experience 
level of workers currently experiencing employment-related economic distress. 


1 Underscored numbers in parentheses refer to items listed in the 
References. 


1 



Even more basic is the need to determine the extent of labor distress (that 
is, the number of people affected by limited job opportunities, and the 
economic magnitude of lost wages in various sectors of the economy). Measures 
of labor distress are important indicators of the strength of rural and urban 
economies. Recent evidence suggests labor distress in the 1980's is higher 
among rural workers than urban workers. Measures of labor distress, combined 
with information on the incidence by race, sex, age, and geographic groupings, 
would provide a basis for a fuller understanding of the overall economic 
problems in rural America (5). 

This report describes a research methodology developed to estimate the 
magnitude of lost wages due to labor distress in the U.S. economy. The 
methodology estimates lost wages for workers within each labor-distress 
category. Aggregate results are presented for rural and urban, male and 
female, and white, black, and other workers. 


THE ANALYTICAL FRAMEWORK 

The basic assumption underlying this analysis is that it is beneficial to our 
society for individuals of working age who desire work to be able to obtain 
employment, including self-employment, that provides earnings commensurate 
with cohorts having similar skills who work full time. Thus, workers, whose 
earnings seriously lag behind the average earnings of workers with similar 
skills and experience, are defined as labor-distressed workers. The total 
cost to society of labor distress includes both social and economic costs. 
While the total costs are probably incalculable, the major economic costs are 
known. From the perspective of the whole economy, these include the value of 
forgone earnings and public welfare expenditures to assist those with poverty- 
level incomes. An argument can also be made to include the value-added to 
products not produced and, therefore, not available to consumers. This 
section of the report describes the analytical procedure used to estimate the 
dollar value of earnings lost due to labor distress. Personal and family 
financial problems associated with labor distress are undoubtedly contributing 
factors to many other social problems which also produce economic problems for 
the Nation. This report, however, does not attempt to measure these costs. 

Previous work on unemployment and underemployment (2, 3, 4, 7) has identified 
five categories of labor-distressed workers: 

1) The discouraged: workers who have become so discouraged they are no 
longer actively looking for a job, but who would work if an appropriate 
opportunity arose. 

2) The unemployed: those officially out of work and actively looking for 
employment. 

3) The part-time employed: those involuntarily working part-time (less than 
35 hours per week for a majority of the weeks during the past year) 
because they cannot find full-time work. 

4) The working poor: workers who earn less than 1.25 times the individual- 
level poverty threshold, even though they are employed full time. 

5) The mismatched: workers whose job fails to fully use their skills 
because of an occupational mismatch. 

These categories of labor distress are ranked on the basis of financial 
severity to the worker. In this analysis, a worker who fits two categories is 
classified into the more severe category. 


2 


Lichter and Costanzo found labor distress is underestimated in official 
unemployment statistics, especially for rural areas (7, 8). Nevertheless, the 
rate of unemployment is the labor statistic used for allocating Federal funds 
to States for labor retraining or rural development programs (10). 

While the number of workers in the various categories of labor distress is 
useful information, this report takes the next step and estimates the dollar 
value of lost earnings (11). The magnitude of lost earnings, when combined 
with information about the number and characteristics of persons experiencing 
labor distress, provides important information for understanding the total 
economic costs of failing to fully employ the Nation's workforce. 

Estimating lost earnings due to labor distress is an empirically tedious task. 
Initially, the task was to define the universe of individuals to be 
investigated. This analysis, like most studies of labor markets, defines the 
universe of potential workers as noninstitutionalized civilians 16 years old 
or older. From this initial pool, individuals unable to work for reasons of 
disability, retirees, or people voluntarily out of the labor force (for 
example, housewives and full-time students) are defined as "not in the labor 
force" (NILF) (fig. 1). These individuals would be of no further interest, 
except that the NILF category includes discouraged workers who have given up 
looking for work, but who would work if appropriate employment were available 
(category 1, fig. 1). The empirical problem is to identify these individuals 
and estimate expected earnings based on their job-relevant characteristics. 

A basic assumption of this study is that full-time employment is available for 
those desiring to work. While fundamental, this assumption does not imply 
that full employment is achievable. Rather, full employment is used as a 

Figure 1—Labor distress categories 



•This group was used to statistically estimate the parameters ot the expected earnings equation. 


3 















































benc±nnark in the same manner as used in the calculation and reporting of 
unemployment statistics. It is the fully employed workers who provided 
the parameter estimates from which expected earnings were derived for the 
unemployed and those employed part-time. TO estimate an earnings equation 
requires a large pool of full-time workers, including self-employed, with the 
associated personal earned income and characteristics data. 2 Earnings should 
not include dividends and capital gains but should be derived directly from 
the employment of the individual's labor and skills. The list of job-relevant 
personal characteristics can be very long. The characteristics discussed are 
not exhaustive, and they anticipate the variables actually used in the 
analysis. 

Educators and students of labor and labor markets have long been interested in 
the earnings that accrue to additional education and increased skills. Other 
variables of interest include sex, race, geographic region, metro versus 
nonmetro residence, ethnic origin, self-employment, and occupational activity 
within industry. In generic form, the earnings equation is: 

personal earnings = f [ (sex, race, ethnic origin, region, metro versus 

nonmetro, self-employed versus otherwise, industry, 
occupation), (age or experience level, education) ]. 

This equation, statistically fit to full-time workers, provided parameter 
estimates for each variable. These parameter values were then used to derive 
an expected earnings value for workers not employed full-time. 

The cost of labor distress, except as discussed below, is the difference 
between expected earnings and reported earnings on an annual basis. Reported 
earnings for discouraged workers will be zero because these persons are not 
and have not recently been in the workforce. Therefore, the cost of labor 
distress (CLD) for a worker in this category is CLD i = (expected earnings,. -0) 

= expected earnings,, (fig. 2) . For involuntary part-time workers (those 
unable to obtain full-time employment) reported earnings, y ; , are positive 
amounts so that CLD,. = (expected earnings^ - y,) , where y,- > 0. For those few 
individuals whose reported part-time earnings are greater than the expected 
full-time earnings, CLD,- should be zero (that is, if y,- > expected earnings,, 
then CLD,. = 0. The CLD for unemployed persons is calculated the same way, 
where y,- > or = 0, depending upon the duration of unemployment for the person 
in question (fig. 2). 


2 This analysis attempts to follow the procedure of Clogg as closely as 
possible (2, 3, 4). The same categories of distressed workers are used based 
upon the same definitions, except as noted for the occupationally mismatched. 
Nevertheless, because this analysis is an extension of Clogg's previous work, 
there are issues of estimation open to debate. For example, the decision to 
include the working poor and the occupationally mismatched in the earnings 
estimation group can be questioned. Why include categories of distressed 
workers in the group that serves as the standard? The decision to include 
these distressed workers in the estimation group was made on the basis that 
the earnings equation should be based upon all full-time workers. This 
results in a more conservative measure of the cost of labor distress than if 
these labor distress groups had not been included in the earnings estimation 
equation. 


4 



Figure 2—Cost of labor distress 

Discouraged, part-time, and unemployed workers 



Two categories of full-time workers experience labor distress. The first of 
these is the working poor (category 4). The working poor may be thought of as 
persons working below their skill potential. This suggests society has failed 
to develop the inherent potential of these workers, such that they cannot earn 
a living at or above the poverty level. Building on this assumption, the CLD 
for the working poor is the difference between the individual poverty-income 
level and reported earnings, that is, CLD ; = (individual poverty-income 
level - reported earnings) (fig. 3). Because this does not involve the 
statistically estimated expected earnings equation, the calculation of the CLD 
for the working poor is primarily one of identification. 3 

The second category of full-time workers experiencing labor distress is the 
occupationally mismatched. Calculating the CLD for these workers is more 
involved, both conceptually and analytically. The procedure used in this 
analysis follows that developed by Clogg (4), with the addition of an economic 
criterion. It is important to understand that mismatched workers, like the 
working poor, are employed full-time. Thus, if an individual is involuntarily 
employed part-time the mismatch issue is not raised. This is because being 
forced to accept part-time employment is assumed to result in a worse 
financial state for the worker and a greater loss for the economy than full¬ 
time occupationally mismatched employment. This conforms to the ranking on 
severity of labor distress. 


An adjustment to the category CLD could be made to account for those 
with physical or mental handicaps that prevent them form achieving a poverty- 
level income. 


5 








Figure 3--Cost of labor distress 

Working poor and mismatched 



From among the full-time employees, including those self-employed, workers are 
classified as occupationally mismatched if their educational levels are one or 
more standard deviations above the average education and their reported 
earnings are below the mean for the same occupation/industry combination. 

This is a very restrictive definition because the dual criteria require the 
worker to be overeducated for, yet earning less than the average in, this 
presumably less demanding job. Diagrammatically, mismatched workers fall 
within the lower right-hand area of figure 3. Thus, CUD,- = (y - reported 
earnings^ for those classified as mismatched, and is zero otherwise (fig. 3) . 


THE DATA 

The data used are the March 1986 Current Population Survey (CPS) taken by the 
Bureau of the Census. The CPS is carefully drawn to represent the U.S. 
population and is a major source of monthly data on the noninstitutionalized 
civilian labor force. The March CPS also contains information on employment 
status, occupation and industry association, weeks worked, hours worked per 
week, reason(s) for not working, and information on income and components of 
income, including welfare payments. 4 This information allows individuals in 


4 The CPS data are obtained monthly and are used to make the official 
U.S. Government estimates of unemployment. This report uses the March 1986 
CPS data, which contains additional questions not included in other months. A 
CPS code exists for each person with respect to "in the labor force" versus 
"not in the labor force." Coding for full- versus part-time employment is 
based on the past year's work history. Information is also available to 


6 













the CPS sample to be classified into the five labor distress categories 
discussed above, or determined to be not under labor distress. 


THE EMPIRICAL ANALYSIS 

Operationally, the universe was defined as all civilian, noninstitutionalized 
persons 16 or more years old. The CPS data on each person's employment status 
were used to make the initial classification of "in the labor force" (ILF) vs 
"not in the labor force" (NILF). Workers classified as ILF were then further 
classified as full-time employed, voluntary part-time employed, involuntary 
part-time employed (less than 35 hours per week), and unemployed. The full¬ 
time employed workers were then screened for those with earnings below the 
individual poverty level, set at $7,250 per year, and those meeting the 
mismatch criteria. 

Those classified as NILF were screened for discouraged workers, defined as 
those indicating they had looked for work last year, and those indicating the 
main reason they did not work last year was because they could not find work. 
Individuals giving either of these responses were classified as discouraged 
workers and were included in the cost of labor distress calculation (fig. 1). 

The CPS data include information on each respondent's occupation and industry 
affiliation. This information was used to create nine occupational and eight 
industrial classifications. The cross-classification of these variables 
produces a table with 72 cells. All full-time workers were classified into a 
cell for purposes of calculating the educational and earnings criteria used to 
identify mismatched workers. This procedure ensures that an individual is 
compared only with workers in the same industry/occupation cell, thus 
preventing the comparison of a laborer with an executive in the mismatch 
income comparison procedure. 

Those employed full time served as the group on which to estimate the 
statistical earnings equation. Preliminary statistical analysis indicated the 
need to modify the generic equation to fit separate estimates for male and 
female workers. The classification variable, sex, was therefore dropped from 
the estimated equation. After investigating numerous forms of the modified 
generic equation, including various combinations of the independent variables, 
second-order interaction effects, and transformations of both independent and 
dependent variables, equation 1 was found to give the most satisfactory 
statistical fit for both males and females. Equation 1 is specified below and 
was used to calculate expected earnings for involuntary part-time workers and 
the unemployed. A different equation was used for discouraged workers. 

For those initially classified as NILF, any previous industry and occupational 
affiliation may well be irrelevant upon reentry, especially if considerable 
time has elapsed since they last worked. For those who have not yet had a 
job, there is no relevant industry or occupational experience upon which to 


discern voluntary versus involuntary unemployment and part-time employment, 
and to develop indicators for discouraged and occupationally mismatched 
workers. The data on personal earnings were accepted as coded in the CPS, 
including top-coding and estimates for missing data made by the Census Bureau. 
For details of the questions and code definitions, readers are referred to the 
Census Bureau's documentation of the March 1986 CPS. 


7 



base expected earnings. For these reasons, a second earnings equation was 
statistically estimated, again using the full-time workers (equation 2 below). 
The second equation differed from the first in that industry and occupational 
data were deleted. The separate parameter estimates from the second equation 
for both males and females were then used to calculate expected earnings for 
discouraged workers. 


Reported personal 
earnings 

[1] 

= f[ (industry, occupation, race, region, metro/nonmetro, 
Hispanic, self-employed/otherwise, region by 
metro/nonmetro, industry by occupation), (age, age 2 , 
yrs. el. ed., yrs. coll. ed., yrs. coll, ed. 2 , age 
by yrs. el. ed., age by yrs. coll, ed.), error ], 

Reported personal 
earnings 

[2] 

= f[ (race, region, metro/nonmetro, Hispanic, self- 
employed/ otherwise, region by metro/nonmetro), 

(age, age 2 , yrs. el. ed., yrs. coll, ed., yrs. coll. 
ed. 2 , age by yrs. el. ed., age by yrs. coll, ed.), 
error ]. 


Where: 


Industry 

= 1) Farming, fishing, forestry, mining 

2) Construction, durable and nondurable manufactured 
goods 

3) Transportation and communication 

4) Wholesale and retail trades 

5) Finance, insurance, and specialty professionals 

6) Repair and service businesses 

7) Personal services and entertainment 

8) Public administration 

Occupation 

= 1) Executives, administration, managers 

2) Professional specialties 

3) Technicians, precision and crafts workers 

4) Sales 

5) Administrative support 

6) Services 

7) Farming, fishing, forestry, mining 

8) Transportation 

9) Laborers 

Race 

= White 

Black 

Other 

Region 

= Northeast 

Midwest 

South 

West 

Metro/nonmetro 

= (self-explanatory) 

Hispanic origin 

= Yes, no 


8 


Self-employed 


Yes, no 



Years old 


(Years old) 2 


Years of elementary 
education 


Highest grade completed 0-13 (high school) 


Years of college 
education 


Years of college completed 


Years of college 
education 2 


(Years of college completed) 2 


and interaction terms as indicated. 


THE ESTIMATION RESULTS 


The results of the statistical fit of equation 1 for full-time male workers 
are provided in table 1. Note that by including the working poor and 
occupationally mismatched workers in the statistical estimation procedures, 
the results for expected earnings are very conservative (fig. 3). This is 
because the wages of the working poor and the occupationally mismatched are 
lower than the average of other workers. 

To obtain the desired parameter estimates, the regression procedure had to 
incorporate classification variables and continuous variables, as well as 
interaction terms. When classification variables are used, the choice of the 
specific level within the classification to use as the standard for comparison 
is entirely arbitrary. Therefore, the estimated parameter values are not 
unique. However, the aggregate linear effect of the classification variables 
on the dependent variable is unique and unbiased. For this reason, estimated 
individual parameter values for the classification variables are not given in 
table 1. Statistical F values for each classification variable are presented. 

The estimated parameters for continuous variables are both unique and 
unbiased. Table 1 indicates all variables in the estimated equation for men 
were statistically significant except for Hispanic origin, the interaction 
term of region by metro-nonmetro, and the interaction of age by years of 
elementary education. Of the significant variables, the variable accounting 
for the most observed variance in earnings was the square of age, where age is 
the surrogate measure for experience and skill. The second most important 
variable was (the statistical main effect of) age, followed closely by the 
interaction term of age X years of college education. Metro-nonmetro 
residence, years of college education, and race follow in statistical 
estimation importance. 5 The results for men indicate the overwhelming 


5 The inclusion of the metro-nonmetro and race classification variables, 
and the use of separate equations for male and female workers, means that 
white (black or other) rural male workers in labor distress are compared with 
their white (black or other) rural colleagues only, not with their white 
(black or other) urban counterparts, when calculating lost earnings. The same 
procedure was followed for women. This was done to control for differences in 


9 



Table 1—Regression estimation results for men 




Parameter estimates 


Type of variable and name 

F value 

Eauation 1 

B 

T value 

Equation 2 

F value 

Classification variables: 

Occupation 

68.03 

NA 

NA 

NA 

Industry 

17.83 

NA 

NA 

NA 

Race 

132.10 

NA 

NA 

238.01 

Region 

7.10 

NA 

NA 

1.54* 

Metro versus nonmetro 

336.83 

NA 

NA 

436.28 

Self-employed 

291.59 

NA 

NA 

414.42 

Hispanic 

.22* 

NA 

NA 

.03* 

Region X metro-nonmetro 

1.31* 

NA 

NA 

3.58 

Continuous variables: 

Age 

490.34 

1,224.13 

22.14 

607.13 

Age 

1,372.26 

-18.18 

-37.04 

1,742.14 

Years— 

Elementary education 

42.16 

1,222.08 

6.49 

58.86 

College education 

173.92 

-2,411.61 ■ 

-13.19 

62.63 

College education 2 

115.86 

280.86 

10.76 

58.44* 

Age X year— 

Elementary education 

2.26* 

6.14 

-1.50* 

1.36* 

College education 

481.15 

75.10 

21.94 

465.44 


R 2 

= .356 

R 2 = 

30.5 


NA = Not applicable. * = Not statistically significant. 


importance of seniority and the experience and skill learned along the way. 

By comparison, education by itself, metro versus nonmetro, and race fare 
poorly in predictive power. Overall, the estimation equation accounted for 
35.6 percent of the observed variance in workers' earnings. 

The same estimation equation for women accounted for 31.9 percent of the 
observed variance in earnings (table 2). All variables, except the 
interaction tern region by metro-nonmetro, were significant at the 0.05 level. 
The square of age for women was again the single most important variable, 
although less so than for men. Metro versus nonmetro residence and self- 
employed versus not self-employed were the second and third most important 
predictive variables. These results must, however, be interpreted with the 
understanding that when men and women are combined into the same sample and 
sex is included in the earnings estimation equation, sex is the strongest 
predictor variable. This derives from the discrepancy between male and female 
earnings across all groupings tried in this analysis. 

RESULTS 

The cost of unemployment at the national level is considerably less than half 
the grand total of labor distress cost. The cost to the economy in lost wages 


rural-urban labor markets and costs of living that affect earnings. The 
estimates of earnings lost due to labor distress would be considerably larger 
if this procedure had not been used. 


10 








Table 2—Regression estimation results for women 




Parameter 

estimates 

Type of variable and name 

Ecruation 1 

F value B T value 

Equation 2 

F value 

Classification variables: 
Occupation 

56.52 

NA 

NA 

NA 

Industry 

24.46 

NA 

NA 

NA 

Race 

9.17 

NA 

NA 

25.92 

Region of residence 

13.09 

NA 

NA 

12.02 

Metro versus nonmetro 

302.86 

NA 

NA 

364.49 

Self-employed 

223.09 

NA 

NA 

415.63 

Hispanic 

17.16 

NA 

NA 

10.77 

Occupation X industry 

8.58 

NA 

NA 

NA 

Region X metro-nonmetro 

.58* 

NA 

NA 

1.34* 

Continuous variables: 

Age 

139.99 

600.01 

11.83 

80.50 

Age^ 

546.36 

-8.39 

-23.30 

775.11 

Years— 

Elementary education 

27.88 

943.71 

5.28 

39.83 

College education 

28.89 

-719.64 

-5.47 

.13* 

College education 2 

111.81 

200.59 

10.57 

79.58 

Age X years— 

Elementary education 

4.93 

-8.68 

-2.22 

3.97 

College education 

81.37 

23.92 

9.20 

60.99 


R 2 = 

0.319 

R 2 

= 30.5 


NA = Not applicable. * = Not statistically significant. 


due to involuntary part-time employment and the working poor is $90 billion, 
compared with less than $75 billion for the unemployed (table 3). 

The national total estimated cost of labor distress exceeds $205 billion 
annually (table 3). The total for metro workers is in excess of $158.6 
billion, while for nonmetro workers it is nearly $46.5 billion. Unemployment 
contributes most to total cost for metro workers ($59.6 billion), but the 
working poor contribute most for nonmetro workers ($16.4 billion versus $15.1 
billion for the nonmetro unemployed). The total number of metro workers is 
approximately three times the number of nonmetro workers, and the total 
estimated cost of labor distress is also approximately three times greater. 
This pattern does not hold for all categories of labor distress. Among 
discouraged workers, metro workers are about three times more numerous than 
nonmetro workers but incur nearly five times the cost. A similar situation 
occurs for the unemployed, the part-time employed, and the working poor. For 
mismatched workers, both the incidence and cost are about five times higher in 
metro than in nonmetro areas. 

Females predominate in both number and cost among discouraged workers and the 
working poor. There are more females than males among the part-time employed 
workers, but the estimated total cost for females is slightly lower. Males 
outnumber females by two to one in the mismatch category, while their cost to 
the economy is four times greater than for females. The total number of male 
and female workers experiencing labor distress is nearly equal, but males 
contribute nearly $112 billion to the grand total, compared with just over $93 
billion for females (table 3). 


11 







Table 3—Residence, sex, and race: Incidence and cost of labor distress, 
March 1986 


Item 

Discouraged workers 

Unemployed 

Part-time employed 


1,000 

workers 

Million 

dollars 

1,000 Million 

workers dollars 

1,000 

workers 

Million 

dollars 


Residence: 

Metro 

1,134 

15,746 

6,458 

59,609 

3,805 

28,343 

Nonmetro 

368 

3,553 

2,248 

15,092 

1,591 

7,845 

Total 

1,520 

19,299 

8,706 

74,701 

5,396 

36,188 

Sex: 

Male 

522 

8,050 

5,003 

42,302 

2,486 

18,938 

Female 

998 

11,249 

3,703 

32,399 

2,910 

17,249 

Total 

1,520 

19,299 

8,706 

74,701 

5,396 

36,188 

Race: 

White 

997 

13,974 

6,593 

57,585 

4,381 

30,299 

Black 

468 

4,654 

1,861 

14,552 

861 

4,760 

Other 

55 

671 

252 

2,564 

153 

1,129 

Total 

1,520 

19,299 

8,706 

74,701 

5,396 

36,188 


Working poor 

Mismatch 

Total 


1,000 Million 

1,000 

Million 

1,000 

Million 

workers dollars 

workers 

dollars 

workers 

dollars 


Residence: 

Metro 

9,785 

37,407 

2,954 

17,555 

2,4136 

158,659 

Nonmetro 

3,851 

16,431 

559 

3,561 

8,635 

46,482 

Total 

13,636 

53,838 

3,512 

21,115 

32,771 

205,141 

Sex: 

Male 

6,085 

25,608 

2,357 

16,994 

16,452 

111,893 

Female 

7,552 

28,230 

1,155 

4,121 

16,319 

93,248 

Total 

13,636 

53,838 

3,512 

21,115 

32,771 

205,141 

Race: 

White 

11,446 

45,815 

3,019 

18,648 

26,436 

166,321 

Black 

1,736 

6,078 

335 

1,592 

5,262 

31,635 

Other 

454 

1,946 

159 

875 

1,073 

7,184 

Total 

13,636 

53,838 

3,512 

21,115 

32,771 

205,141 


12 



















Results for whites, blacks, and others are also shown in table 3. The 
estimated number of black discouraged workers is roughly half that of whites, 
but they account for approximately 30 percent of all distressed workers. 

Labor distress cost of discouraged black workers is a third that of whites. 

Table 4 presents the same basic information as table 3, but in finer detail. 
Cost estimates are presented for metro and nonmetro whites, blacks, and others 
by sex. 

Estimates of labor distress costs by State are presented in table 5. Five 
States (New York, Pennsylvania, Illinois, Texas, and California) have total 
labor distress costs in excess of $10 billion annually. Only 12 States have 
total cost estimates of less than $1 billion. In only a few States, does the 
cost of unemployment approach 50 percent of the total estimated cost. 

Tables 6 and 7 show comparisons of labor distress costs across worker 
categories. Table 6 shows the percentage each labor distress category 
contributes to the total cost of labor distress for individual categories of 
workers. The total of each worker category as a percentage of the grand total 
is also given. Females, blacks, and other males in nonmetro areas dominate 
the discouraged worker category. Unemployment costs are more evenly 
distributed, except for blacks (especially black females in nonmetro areas). 

Of the five labor distress categories, the costs associated with involuntary 
part-time employment are the most evenly distributed across worker categories. 
Costs associated with the working poor tend to fall most heavily on nonmetro 
and female workers. It is noteworthy that costs associated with blacks are 
underrepresented in the working poor category and are overrepresented in the 
unemployed category. White and other males dominate the costs associated with 
workers occupationally mismatched. 

Table 7 presents information on labor distress costs per labor force 
participant. The labor distress cost per nonmetro worker ($1,885) is higher 
than for metro workers ($1,735). Similarly, the labor distress cost 
per female worker ($1,803) is higher than for males ($1,738). Per capita 
labor force participant costs by race indicate blacks are the most heavily 
affected ($2,547), followed by others ($2,197) and whites ($1,657). 

Table 8 provides additional comparative labor distress cost data for selected 
States. Unemployment cost as a share of total labor distress cost for each 
State varies from 7.4 percent for Alaska to 50.7 percent for West Virginia. 
This variance across States indicates that unemployment is not a constant 
factor in the overall cost of labor distress. Table 8 also provides 
information on total labor distress cost for selected States, as estimated by 
using March 1986 data, as a percentage of each State's 1985 total budget 
expenditure. These data range from 14.8 percent for the District of Columbia 
to 47.8 percent for Mississippi. 


DISCUSSION 

The cost of labor distress to the U.S. economy is large. The estimates show 
the total of forgone wages to be approximately 1.25 times the current annual 
Federal budget deficit. But, even this large figure does not represent the 
whole burden of labor distress. At a minimum, public welfare payments made to 
workers experiencing labor distress must be included as a cost to the economy. 
A strong argument can also be made for including an estimate of the value 


13 


Table 4—Metro versus nonmetro: Incidence and cost of labor distress by 
race and sex, March 1986 


Item 

Discouraged workers 

Unemployed 

Part-time employed 


1,000 

Million 

1,000 

Million 

1,000 

Million 


workers 

dollars 

workers 

dollars 

workers 

dollars 

Metro: 

Whites— 

Males 

237 

4,976 

2,805 

27,085 

1,455 

12,765 

Females 

484 

6,304 

1,917 

17,544 

1,533 

10,499 

Total 

721 

11,279 

4,722 

44,630 

2,988 

23,264 

Blacks— 

Males 

138 

1,462 

826 

5,646 

312 

1,789 

Females 

237 

2,472 

720 

7,085 

382 

2,284 

Total 

375 

3,935 

1,547 

12,731 

695 

4,074 

Others— 

Males 

15 

[249] 

106 

1,175 

54 

477 

Females 

23 

283** 

85 

1,073 

68 

529 

Total 

38 

532 

190 

2,248 

122 

1,005 


Nonmetro: 
Whites— 


Males 

88 

983 

1,079 

7,765 

586 

3,598 

Females 

187 

1,712 

792 

5,191 

808 

3,438 

Total 

275 

2,694 

1,872 

12,956 

1,393 

7,036 

Blacks— 

Males 

34 

291* 

150 

500 

65 

267 

Females 

60 

428 

165 

1,321 

101 

419 

Total 

94 

719 

315 

1,821 

167 

686 

Others— 

Males 

10 

[89] 

37 

131* 

13 

[43] 

Females 

7 

[51] 

25 

185** 

18 

[81] 

Total 

17 

140 

62 

316 

31 

123 


Continued— 


added to products not produced by those under labor distress. Including these 
values would add significantly to the overall cost. 

The cost of labor distress per labor force worker is higher in nonmetro than 
metro areas, even though in estimating the costs, nonmetro workers are 
compared only with nonmetro (not metro) workers. This higher cost suggests a 
structural bias within the economy that operates against nonmetro residents. 

Similarly, the per capita cost is higher for females than for males, despite 
the lower incomes of females. The per capita cost of labor distress for 
blacks and others is substantially higher than for whites. Also, the per 
capita cost of labor distress for blacks and others is substantially higher 


14 











Table 4—Metro versus nonmetro: Incidence and cost of labor distress by 
race and sex, March 1986—-Continued 


Item Working poor Mismatch Total 



1,000 

Million 

1,000 

Million 

1,000 

Million 


workers 

dollars 

workers 

dollars workers 

dollars 

Metro: 

Whites— 

Males 

3,384 

13,409 

1,736 

12,535 

9,616 

70,771 

Females 

4,630 

17,344 

761 

2,843 

9,326 

54,533 

Total 

8,014 

30,753 

2,497 

15,378 

18,942 

125,304 

Blacks— 

Males 

604 

2,328 

147 

891 

2,028 

12,118 

Females 

781 

2,662 

156 

445 

2,276 

14,948 

Total 

1,385 

4,990 

303 

1,336 

4,304 

27,065 

Others— 

Males 

169 

658 

99 

669 

442 

3,227 

Females 

217 

1,006 

55 

173 

448 

3,063 

Total 

386 

1,664 

154 

841 

890 

6,290 

Nonmetro: 

Whites— 

Males 

1,741 

8,566 

352 

2,649 

3,847 

23,561 

Females 

1,691 

6,495 

169 

622 

3,647 

17,457 

Total 

3,432 

15,061 

522 

3,271 

7,494 

41,018 

Blacks— 

Males 

159 

552 

19 

218** 

426 

1,828 

Females 

193 

536 

13 

38** 

531 

2,742 

Total 

351 

1,088 

32 

256 

958 

4,570 

Others— 

Males 

28 

94 ** 

5 

[32] 

92 

389 

Females 

41 

187* 

1 

[ 2 ] 

91 

505 

Total 

68 

281 

5 

34 

183 

894 


* This information is based on a sample size that is too small to provide 
statistically reliable data, but may be used for general policy/program 
guidance. 

** This information is based on a sample size that is too small to provide 
full statistical confidence in the data, but is sufficient for policy/program 
planning. 

[] This information is based on a sample size that is too small for minimal 
reliability for the State. However, the data are included to maintain the 
integrity of the tables and for aggregation to the national level where 
sufficient reliability is achieved. 


15 












Table 5—Region and State: Incidence and cost of labor distress 
March 1986 


Item 


Discouraged workers Unemployed Part-time employed 



1,000 

Million 

1,000 

Million 

1,000 

Million 


workers 

dollars 

workers 

dollars 

workers 

dollars 

Northeast: 

Connecticut 

9 

[91] 

69 

557 

28 

285 

Maine 

5 

[64] 

31 

238 

31 

222 

Massachusetts 

22 

368** 

123 

1,063 

73 

52 

New Hampshire 

2 

[43] 

25 

183 

15 

102 

New Jersey 

32 

340** 

192 

1,801 

90 

739 

New York 

103 

1,280 

600 

5,346 

238 

1,564 

Pennsylvania 

84 

1,179 

448 

4,357 

270 

1,961 

Rhode Island 

4 

[48] 

18 

151 

22 

158 

Vermont 

2 

[35] 

15 

83 

9 

46 

Midwest: 

Indiana 

48 

581* 

211 

1,884 

144 

1,014 

Illinois 

93 

1,396 

505 

4,497 

254 

1,847 

Iowa 

15 

181** 

128 

1,229 

86 

550 

Kansas 

14 

[211] 

85 

721 

50 

239 

Michigan 

100 

1,475 

420 

4,349 

203 

1,401 

Minnesota 

24 

317** 

147 

1,479 

129 

773 

Missouri 

28 

393** 

146 

1,366 

139 

940 

Nebraska 

7 

[81] 

46 

349 

41 

161 

North Dakota 

4 

[53] 

27 

236 

20 

154 

Ohio 

64 

910 

413 

3,935 

266 

1,891 

South Dakota 

4 

[33] 

21 

178 

26 

158 

Wisconsin 

33 

366** 

179 

1,372 

115 

752 

South: 

Alabama 

38 

407* 

198 

1,657 

118 

610 

Arkansas 

20 

142** 

97 

658 

58 

374 

Delaware 

2 

[35] 

15 

72 

21 

100 

District of 

Columbia 

6 

[80] 

27 

187 

12 

62 

Florida 

56 

778 

324 

2,770 

222 

1,458 

Georgia 

41 

356* 

195 

1,295 

176 

1,112 

Kentucky 

29 

325** 

182 

1,151 

104 

593 

Louisiana 

39 

327* 

210 

1,277 

140 

985 

Maryland 

14 

[181] 

102 

833 

72 

569 

Mississippi 

18 

130** 

146 

1,248 

69 

342 

North Carolina 

43 

397* 

181 

1,355 

136 

836 

Oklahoma 

21 

150** 

136 

1,100 

100 

556 

South Carolina 

28 

291** 

107 

873 

69 

391 

Tennessee 

39 

438* 

219 

1,653 

133 

779 

Texas 

94 

1,211 

676 

5,043 

412 

2,773 

Virginia 

21 

180** 

130 

974 

103 

534 

West Virginia 

17 

187** 

89 

974 

55 

342 


—Continued 


16 











Table 5—Region and State: Incidence and cost of labor distress, 
March 1986—Continued 


Item Working poor Mismatch Total 



1,000 

Million 

1,000 

Million 

1,000 

Million 


workers 

dollars 

workers 

dollars 

workers 

dollars 

Northeast: 

Connecticut 

148 

587 

67 

355 

321 

1,875 

Maine 

59 

221 

19 

120 * 

145 

864 

Massachusetts 

283 

1,112 

126 

590 

627 

3,653 

New Hampshire 

51 

198 

22 

118 

114 

644 

New Jersey 

342 

1,332 

120 

667 

776 

4,878 

New York 

891 

3,560 

275 

1,594 

2,107 

13,344 

Pennsylvania 

589 

2,221 

131 

972 

1,522 

10,690 

Rhode Island 

51 

210 

15 

106* 

110 

672 

Vermont 

39 

154 

13 

97 ** 

78 

416 

Midwest: 

Indiana 

317 

1,169 

61 

369 

781 

5,017 

Illinois 

578 

2,375 

139 

846 

1,569 

10,961 

Iowa 

188 

975 

35 

195 

452 

3,130 

Kansas 

162 

935 

32 

191 

344 

2,297 

Michigan 

462 

1,862 

108 

51 

1,293 

9,599 

Minnesota 

226 

1,123 

51 

312 

577 

4,004 

Missouri 

304 

1,223 

6 

466 

679 

4,389 

Nebraska 

120 

561 

19 

135 

233 

1,286 

North Dakota 

44 

210 

8 

46** 

103 

698 

Ohio 

481 

1,745 

138 

761 

1,361 

9,242 

South Dakota 

64 

345 

12 

84** 

127 

797 

Wisconsin 

243 

1,072 

47 

240 

618 

3,802 

South: 

Alabama 

247 

984 

41 

231 

42 

3,888 

Arkansas 

158 

610 

31 

166 

363 

1,950 

Delaware 

41 

143 

11 

59** 

88 

409 

District of 

Columbia 

35 

148 

16 

88 * 

96 

565 

Florida 

711 

2,827 

219 

1,392 

1,532 

9,225 

Georgia 

395 

1,447 

74 

457 

880 

4,667 

Kentucky 

208 

832 

49 

360 

572 

3,261 

Louisiana 

268 

1,025 

60 

374 

718 

3,988 

Maryland 

221 

750 

72 

407 

481 

2,739 

Mississippi 

192 

662 

29 

170 

454 

2,552 

North Carolin 

414 

1,425 

116 

704 

890 

4,717 

Oklahoma 

262 

1,177 

59 

355 

578 

3,338 

South Carolina 

202 

611 

56 

325 

461 

2,491 

Tennessee 

308 

1,066 

73 

455 

772 

4,391 

Texas 

1,012 

4,068 

189 

1,328 

2,382 

14,422 

Virginia 

393 

1,185 

65 

358 

712 

3,231 

West Virginia 

93 

337 

13 

82** 

266 

1,921 


—Continued 


17 











Table 5—Region and State: Incidence and cost of labor distress, 
March 1986—Continued 


Item 

Discouraged workers 

Unemployed Part-time employed 


1,000 

Million 

1,000 

Million 

1,000 

Million 


workers 

dollars 

workers 

dollars 

workers 

dollars 

West: 

Alaska 

3 

[30] 

29 

183 

17 

82 

Arizona 

5 

[65] 

90 

767 

73 

463 

California 

184 

2,599 

939 

8,988 

577 4 

,345 

Colorado 

18 

254** 

134 

1,100 

82 

658 

Hawaii 

4 

[50] 

30 

279 

34 

260 

Idaho 

7 

[87] 

38 

275 

32 

193 

Montana 

7 

[104] 

38 

281 

292 

14 

Nevada 

6 

[89] 

31 

269 

32 

292 

New Mexico 

14 

[147] 

61 

463 

33 

187 

Oregon 

19 

300** 

145 

1,313 

66 

579 

Utah 

7 

[71] 

43 

336 

34 

272 

Washington 

24 

428** 

222 

1,812 

119 

631 

Wyoming 

2 

[17] 

24 

143 

19 

119 

Total 

1,520 

19,299 

8,706 

74,701 

5,396 36 

,188 


* This information is based on a sample size that is too small to provide 
full statistical confidence in the data, but is sufficient for policy/program 
planning. 

** This information is based on a sample size that is too small to provide 
statistically reliable data, but may be used for general policy/program 
guidance. 


than for whites. These data suggest an additional structural bias within the 
economy. 

The unemployed category is the best known of the five components of labor 
distress. The other four categories are virtually unrecognized outside the 
research community. However, the data shew conclusively that underemployed 
workers are the largest contributors to the overall cost of labor distress. 

In fact, involuntary part-time workers and the working poor account for more 
of the total cost of labor distress than do the unemployed. Implications of 
this finding for Federal funding of various development programs are discussed 
in the next section. 


Interpretation of the Results 

This report lays the foundation for exploring the use of labor distress cost 
as an indicator of the magnitude of unemployment and underemployment in the 
economy. The value of wages lost by workers experiencing labor distress 
provides a method for aggregating losses across labor distress categories and 
for assessing the relative share of each category. As an indicator, these 
losses are measured as the difference between a worker's reported earnings and 


18 











Table 5—Region and State: Incidence and cost of labor distress, 
March 1986—Continued 


Item 

Working poor 

Mismatch 

Total 


1,000 Million 

1,000 

Million 

1,000 Million 


workers dollars 

workers 

dollars 

workers dollars 


West: 

Alaska 

19 

86 

6 

36** 

74 

415 

Arizona 

196 

767 

70 

540 

433 

2,601 

California 

1,577 

6,133 

428 

2,531 

3,705 

24,597 

Colorado 

183 

718 

64 

367 

480 

3,095 

Hawaii 

52 

234 

25 

186 

146 

1,010 

Idaho 

74 

330 

17 

115* 

167 

1,000 

Montana 

59 

289 

10 

40** 

144 

928 

Nevada 

71 

265 

11 

69** 

150 

984 

New Mexico 

111 

483 

22 

145 

241 

1,424 

Oregon 

132 

556 

55 

248 

417 

2,995 

Utah 

77 

294 

28 

160 

189 

1,133 

Washington 

253 

1,081 

98 

547 

715 

4,499 

Wyoming 

32 

117 

7 

51** 

85 

446 

Total 

13,636 

53,838 

3,512 

21,115 

32,771 

205,141 


[] This information is based on a sample size that is too small for minimal 
reliability for the State. However, the data are included to maintain the 
integrity of the tables and for aggregation to the national level where 
sufficient reliability is achieved. 


a standard, generally defined in this report as the average wage earned by 
that worker's peers. 

While this is a legitimate measure, it is easily misinterpreted. It should 
not be assumed that if all workers experiencing labor distress were to obtain 
an appropriate job, their earnings would rise to the level of the standard 
against which they were judged. Neither is it implied that the standard used 
will remain constant over time, nor that this standard will remain unchanged 
if all labor-distressed workers obtain work. There is the expectation that 
for small incremental reductions in the level of labor distress, the average 
earnings of workers emerging from distress will equal average earnings of 
their peers. To anticipate the direction and magnitude of earnings that would 
result from large reductions in labor distress would require a simulation 
model of the economy that is beyond the scope of this report. 

In this report, no judgment is made or implied regarding the issue of full 
employment, defined as full-time employment in appropriate jobs of all workers 
experiencing labor distress. Such a full employment economy is probably 
impossible to achieve. Even under the best economic conditions, some level of 
structural unemployment and underemployment will exist to accommodate labor 
adjustments within the economy. 


19 











Table 6—Cost as a percentage of worker category total and category total as 
a percentage of total cost, March 1986 


Item Discouraged 

workers 

Unemployed 

Part-time 

workers 

Working 

poor 

Mismatch 

Total 




Percent - 



Percent of 
total cost 

Residence: 

Metro 

10 

38 

18 

24 

11 

77 

Nonmetro 

8 

33 

17 

35 

8 

23 

Sex: 

Males 

7 

38 

17 

23 

15 

55 

Females 

12 

35 

19 

30 

4 

45 

Race: 

Whites 

8 

35 

18 

28 

11 

81 

Blacks 

15 

46 

15 

19 

5 

15 

Others 

9 

36 

16 

27 

12 

4 

Metro: 

White— 

Males 

7 

38 

18 

19 

18 

35 

Females 

12 

32 

19 

32 

5 

27 

Black— 

Males 

12 

47 

15 

19 

7 

6 

Females 

17 

47 

15 

18 

3 

7 

Other— 

Males 

8 

36 

15 

20 

21 

2 

Females 

9 

35 

17 

33 

6 

2 

Nonmetro: 

White— 

Males 

4 

33 

15 

36 

11 

12 

Females 

10 

30 

20 

37 

4 

9 

Black— 

Males 

16 

27 

15 

30 

12 

1 

Females 

16 

48 

15 

20 

1 

1 

Other— 

Males 

23 

34 

11 

24 

8 

— 

Females 

10 

37 

16 

37 

— 

— 

Total 

9 

36 

18 

26 

10 

100 


— = Negligible. 


20 









Table 7—Labor distress costs per labor force participant 


Category 

Labor 

distress 

costs 

Category 

Labor 

distress 

costs 

Category 

Labor 

distress 

costs 


Dollars 


Dollars 


Dollars 

Residence: 


Sex: 


Race: 


Metro 

1,735 

Male 

1,738 

White 

1,657 

Nonmetro 

1,885 

Female 

1,803 

Black 

2,547 





Other 

2,197 


Table 8—Unemployment cost as a share of total labor distress cost, and this 
cost as a share of 1985 State expenditures 1 


State 


Unemployment Total labor distress 

cost as a share of cost as a share of 

total labor distress 1985 State expenditures 


Percent 


Alaska 

7.4 

44.0 

California 

36.5 

29.2 

Colorado 

33.5 

32.9 

District of Columbia 

33.2 

14.8 

Georgia 

27.8 

32.1 

Idaho 

27.5 

47.0 

Iowa 

39.3 

41.6 

Maine 

27.5 

31.4 

Maryland 

30.4 

23.4 

Massachusetts 

29.1 

21.1 

Mississippi 

48.9 

47.8 

North Dakota 

33.8 

35.1 

Ohio 

42.6 

33.7 

Pennsylvania 

40.7 

36.8 

Tennesse 

37.7 

37.5 

Texas 

35.0 

37.0 

Washington 

40.3 

29.5 

West Virginia 

50.7 

42.7 

United States 

36.4 

31.2 


1 Estimates are derived from March 1986 data. 


21 














It is important not to extend the results beyond the data when using a new 
procedure. Understanding the implicit assumptions in the estimation procedure 
and measures used is very helpful in this regard. 

The estimated earnings equation is based on full-time workers. Most of the 
personal characteristics available in the CPS individual file were used as 
independent variables in the regression equations to achieve the best possible 
estimated earnings. Even so, the R 2 values leave much room for improvement, 
raising the question of whether the critical variables have been included in 
the statistical equation. One may question whether there are unmeasured 
variables that predispose certain workers to fall into one of the labor 
distress categories. If such variables exist, then the estimation equations 
used are incorrectly specified and the reported cost estimates are biased 
upward. 

Stated less technically, if workers in labor distress have any performance- 
inhibiting personal characteristics, which are not shared to the same extent 
by full-time workers, then, on average, distressed workers would not be 
expected to earn the same amount as their fully employed peers. This would 
result in an overestimate of the true amount of labor distress cost. Whether 
distressed workers are predisposed in this manner is not known. Nor are there 
national data to explore this question. Therefore, this report must 
essentially ignore the issue. 

Another note of caution pertains to the working poor. To be classified into 
this category, a full-time worker's reported personal earnings had to fall 
below the individual poverty level. Presumably, an individual living alone in 
his/her own household would need this income as a minimum to avoid living in 
poverty. But, if the worker in question is part of a multiple wage-earner 
family, he/she will likely be able to share in the household income and, thus, 
avoid living in poverty. Whether or not these workers should be compared with 
a different standard is a matter for debate. Strong arguments can be made for 
both viewpoints. 

One can argue that household income should be the critical factor in deciding 
whether a worker is classified as the working poor. If the household income 
is above poverty, no individual worker in that household would be classified 
as working poor regardless of the level of personal income because that 
individual is presumably not living in poverty. But, the implication of this 
line of reasoning is that the level of earnings of the second and succeeding 
wage earners is unimportant so long as the household income remains above the 
poverty line. Minorities, especially working wives, find this reasoning 
especially difficult to accept because it appears to sanction the existing 
lower earned incomes of women. In this report, no adjustments are made for 
workers classified as "working poor" who live in households with incomes above 
the poverty line. 


Potential Uses of the Results 

Underemployment, rather than unemployment, is the largest contributor to the 
total cost of labor distress, a finding that holds for nearly all labor 
distress categories, including State and national results. Current Federal 
assistance to States for programs addressing labor distress use formula 
funding with the unemployment rate as an important variable. Since States 
vary substantially in the percentage contribution the unemployed make to the 


22 


total cost of labor distress, changing the formula to make allocations on the 
basis of total labor distress cost would result in a quite different 
distribution of funds across the States. 

The cost of labor distress, when compared with cost estimates of other social 
problems, provides important information for setting public policy and 
expenditures priorities. The metro-nonmetro comparisons, along with those of 
race and sex, provide insight into the nature, magnitude, and incidence of 
structural differences within the economy. Knowledge of these aspects of 
labor distress can be used to formulate and target appropriate State and 
Federal intervention policies and programs. 

The analysis suggests several program needs. The fact that the rural-urban 
and regional variables are statistically significant suggests the existence of 
systematic structural variations in labor market wages along these geographic 
lines. This suggests a job placement service large enough to span both metro- 
nonmetro and regional areas would be of value by improving the prospects of 
matching workers to available jobs, thus reducing the magnitude of lost wages. 

The success of such a job placement program assumes workers are willing to 
relocate. While this may be true for many workers and their families, it 
would not be true for all who could benefit by moving. Some will not wish to 
leave friends and relatives, while others will resist citing the cost of 
moving. Where the primary consideration is financial, public assistance to 
make a move could be economically justified for those workers with large 
individual costs of labor distress. Identifying these workers would involve 
calculating costs and benefits, where the cost is as derived above and the 
benefit estimated on the basis of anticipated earnings on the new job and any 
savings of welfare payments no longer needed. 

The magnitude of the cost associated with the working poor suggests the 
continuing need for both basic education of new entrants and retraining of 
distressed workers. Given the estimate of labor distress costs associated 
with the working poor developed in this report, appropriate cost-benefit 
ratios can be calculated when estimates of the benefits are provided. Such 
information would be of significant value for policymaking at the State and 
national levels, where the economic impact of various programs is an important 
factor in the allocation of funds. 

But who are the distressed? In addition to the aggregate categories presented 
in this report, more information (such as, the incidence of white/black, 
male/female, metro-nonmetro, and age groups within regions and States) is 
needed on the characteristics of workers within each labor distress category. 
Unfortunately, this information cannot be developed in this report. This is 
because the sample size for many of the geo-political categories of interest 
is too small to support the more detailed analysis. Along with the need for a 
larger or supplemental sample of workers, additional information on worker 
characteristics is also needed. Among the most critical information needed is 
a job skills inventory for distressed workers. Such data would be very useful 
for planning and implementing targeted intervention programs. 


23 


references 


(1) Brown, David L., and Kenneth L. Deavers. "Rural Change and The Rural 
Economic Policy Agenda for the 1980's," Chapter 1 in Rural Economic 
Development in the 1980's: Preparing for the Future. ERS Staff 
Report No. AGES870724. U.S. Dept. Agr., Econ. Res. Serv., July 1987. 

(2) Clogg, Clifford C. "Cohort Analysis of Recent Trends in Labor Force 
Participation." Demography Vol.19. No.4 (Nov. 1982), pp. 459-79. 

(3) Clogg, Clifford C., and James W. Shockey. "Mismatch Between 
Occupation and Schooling: A Prevalence Measure, Recent Trends and 
Demographic Analysis." Demography. Vol. 21, No. 2 (May 1984), pp. 
235-57. 

(4) Clogg, Clifford C., Teresa A. Sullivan, and Jan E. Mutchler. 

"Measuring Underemployment and Inequality in the Work Force," Social 
Indicators Research . T18 (1986), pp. 375-93. 

(5) Danziger, Sheldon, and Peter Gottschalk. "Work, Poverty and the 
Working Poor: A Multifaceted Problem," Monthly Labor (Sept. 1986) 
pp. 17-21. 

(6) Henry, Mark, Mark Drabenstott, and Lynn Gibson. "A Changing Rural 
America," Economic Review (July/Aug. 1986) pp. 23-41. 

(7) Lichter, Daniel T. "Measuring Underemployment in Rural Areas," Rural 
Development Perspectives . Vol. Issue 2 (Feb. 1987), pp. 11-14. 

(8) Lichter, Daniel T., and Janice A. Costanzo. "Nonmetropolitan 
Underemployment and Labor-Force Composition," Rural Sociology . Vol. 
52, No. 3 (fall 1987), pp 329-44. 

(9) Miller, James, and Herman Bluestone. "Patterns of Employment Change 
in the Nonmetropolitan Service Sector, 1969-84." Paper presented to 
the Southern Regional Science Association meeting, Atlanta, GA, Mar. 
26-28, 1987. 

(10) Nardone, Thomas J. "Part-Time Workers: Who Are They?" Monthly Labor 
Review . (Feb. 1986) pp. 13-19. 

(11) Twee ten, Luther. "Rural Labor Market Performance," Symposium on Rural 
Labor Markets Research Issues . ERS Staff Report No. AGES860721. 

Econ. Res. Serv., U.S. Dept. Agr. Sept. 1986. 


24 


^U.S. Government Printing Office 


1989 


241-793/80731