TIME TO WORK:
Job Search Strategies and Commute Time
for Women on Welfare in San Francisco
University of Minnesota
ABSTRACT: The major policy approaches to welfare-to-work attempt to facilitate the tran-
sition into the workforce by providing job search assistance and transportation subsidies.
Although these policies help some women on welfare, they fail to respond to the needs of
most, who rely disproportionately on social contacts to find jobs, seek to minimize com-
mutes, and lack the educational attainment that would help them penetrate the regional la-
bor market. This article uses in-depth interviews with 92 women on welfare in San Francisco,
as well as a binomial logit model, to examine the relationship between job search strat-
egies and employment characteristics. The findings suggest that low-income women with
children are more likely to rely on contacts than women without children, because they seek
to work close to home. For most women, building connections to employers, improving hu-
man capital, and increasing the density of neighborhood economic and social activity will
make jobs more accessible.
The President, in his 2000 budget, has called for full funding—$150 million to provide the
needed resources to help welfare recipients get to—and keep—jobs. Transportation is about
so much more than concrete, asphalt and steel: it is about people and their pursuit of hap-
piness. It is about getting parents to training and to work, getting children to day care, mov-
ing families from the dependence of welfare rolls to the independence of payrolls.
Rodney E. Slater, U.S. Secretary of Transportation, April 13, 1999.
The passage of the Personal Responsibility and Work Opportunity Reconciliation Act
(PRWORA) of 1996 mandates that 50% of women receiving welfare (Temporary Aid to Needy
Families) work by mid-2002. The assumptions underlying the new welfare-to-work require-
ments are that jobs are available and that with improved access to opportunities, in the form
of transportation and job search assistance, local residents will obtain jobs and withdraw from
the welfare program.
*Direct correspondence to: Karen Chapple, Urban and Regional Planning Program, Humphrey Institute of Public
Affairs, University of Minnesota, 301-19th Avenue South, Minneapolis, MN 55455. E-mail: email@example.com
JOURNAL OF URBAN AFFAIRS, Volume 23, Number 2, pages 155–173.
Copyright © 2001 Urban Affairs Association
All rights of reproduction in any form reserved.
But access to employment opportunities only provides a structural context. It is in the pro-
cess of social interaction—in the form of formal or informal searches for jobs in the local
labor market—that matches between workers and employers occur. Local labor markets are
not simply containers of jobs that require appropriate levels of human capital and are located
within an economically rational commute distance. Despite a strong economy, some welfare
recipients are far less likely to obtain jobs, either because of the lack of social contacts who
can help them find and keep employment, or because of spatially constrained labor market
The major policy approaches to welfare-to-work attempt to facilitate the transition into the
workforce by providing job search assistance and instruction in workplace norms (the “work-
first” or job readiness approach) and transportation subsidies. Although these policies are
undoubtedly helpful for some women on welfare, they fail to acknowledge the complexity
of both the job search method used and transportation decision making for most low-income
mothers. If such women rely disproportionately on social networks, it may be important to
improve their connections to employers; and if they only search for jobs close to home, policy
strategies might better focus on community economic development than transportation.
This article examines the job search methods and commuting patterns of women on wel-
fare, exploring the hypothesis that low-income mothers rely heavily on informal job search
strategies and tend to work within a limited labor market area, resulting in incongruity be-
tween welfare-to-work policy and practice. In essence, this research asks two questions: why
do low-income mothers rely on contacts in the job search and what does that reliance mean
The study relies on in-depth, semi-structured interviews with 92 women on welfare in San
Francisco. Interviews elicited full job histories, in order to compare the spatial extent of em-
ployment and job search methods used before and after having children. The remainder of this
article explores the relationship between the use of social contacts in the job search and the
location and other characteristics of the job found. It begins with a description of the litera-
ture examining the role of social networks in the job search, the relationship between the job
search and space, and the role of other (nonsearch) barriers in the transition to work for women
on welfare. After a description of the research methodology, the following section describes
the job search strategies used, the associated employment characteristics, and commuting pat-
terns. Abinomial logit model is then constructed to explore the reasons women have turned to
contacts in the job search. After a look at other transportation considerations, the evaluation
of policy implications examines in more detail the potential for network strategies and trans-
portation assistance for facilitating the transition to work for women on welfare.
JOB SEARCH STRATEGIES, SPACE, AND THE WELFARE-TO-WORK
TRANSITION: A LITERATURE REVIEW
Job Search and Social Networks
Although almost no studies examine the job search process for women on welfare in par-
ticular, a rich body of literature examines job search strategies for different population sub-
groups. Job search may occur via direct application, when the potential employee contacts an
establishment in the hopes that an opening is available; impersonal institutional intermediar-
ies, such as employment agencies, union hiring halls, or newspaper advertisements; or per-
sonal contacts, such as relatives, friends, acquaintances, or teachers. Because roughly half of
all jobs, whether low-skill or high-skill, are found through contacts, having extended social
networks can help employment outcomes (Granovetter, 1995). Weak ties (e.g., work contacts,
156 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
casual acquaintances, or other non-intimate associates) play an important role in the job search.
These ties provide access to more varied sources of information, from a greater variety of peo-
ple, than that provided through a more dense personal network of family and friends (Grano-
vetter, 1995; Lin, Ensel, & Vaughn, 1981). Although it is still debated, weak ties in particular
seem to lead also to higher wages and greater job satisfaction (Granovetter, 1995; Marsden &
Hurlbert, 1988). Little research actually evaluates the network structures of jobseekers, but
there is also some evidence that larger networks are associated with greater use of contacts
(Calzavara, 1983; Lin & Dumin, 1986).
Research suggests that low-income and minority mothers will obtain relatively fewer of their
jobs via contact, although the differences are slight (estimates range from 40% to 50% for
women, with men a few percentage points higher) (Campbell & Rosenfeld, 1985; Corcoran,
Datcher, & Duncan, 1980; Granovetter, 1995; but see Hanson & Pratt, 1995). Lower socio-
economic status is associated with smaller networks and smaller proportions of weak ties. Women
have fewer weak ties than men and women use ties in the job search less than men do (Camp-
bell & Rosenfeld, 1985; Corcoran, Datcher, & Duncan, 1980; Fischer, 1982; Fischer et al.,
1977; Moore, 1990). Minorities such as AfricanAmericans are underrepresented in the work-
force, which places them at a disadvantage in using informal channels of job information (Brad-
dock & McPartland, 1987; Granovetter, 1995). In fact, research has found that AfricanAmericans
rely slightly less on contacts in the job search than do other ethnic groups, and AfricanAmer-
ican women less than all other ethnic or gender subgroups (Corcoran, Datcher, & Duncan,
1980; Green, Tigges, & Diaz, 1999; Wilson, 1996). Members of low-income groups may also
experience difficulty in the job search because of the frequency of labor market entry and exit—
the lack of job stability or experience of long-term employment. Finally, mothers experienc-
ing labor force interruptions with childrearing may be unable to take advantage of the snowballing
effect of longer careers, which helps to develop a contact network (Granovetter, 1988).
Job Search and Space
Much of the empirical work on local labor markets has centered on the spatial mismatch
debate, exploring whether the relocation of firms to outlying areas has adversely impacted em-
ployment prospects for inner city residents (Kain, 1968; for a review see Ihlanfeldt & Sjoquist,
1998). Yet for women, the spatial mismatch phenomenon interacts with spatial entrapment or
containment, or the shorter commuting times that women experience relative to men (Cooke,
1997; Wyly, 1998). Thus, research on job accessibility for women has focused on explaining
this difference, hypothesizing that it occurs because of differential investments in human cap-
ital, the role of household responsibilities, and/or the location of female-dominated employ-
ment (see, for instance, Hanson & Johnston, 1985; Johnston-Anumonwo, 1992; Madden, 1981;
Preston, McLafferty, & Hamilton, 1993).
However, a growing body of research suggests that shorter commute times may stem from
the matching of workers to jobs through localized job-search networks (Hanson & Pratt, 1995;
Wyly, 1998). The high costs of job search—obtaining information about different opportuni-
ties distributed widely across a metropolitan area—dissuade many low-wage workers from an
extensive search (Clark & Whiteman, 1983). Space also shapes the job search because of the
social opportunities related to neighborhood structure, as women may fall into a job in the
neighborhood without commencing a formal job search (Hanson & Pratt, 1995).
Barriers to the Employment Transition for Women on Welfare
In general, about 15% to 20% of welfare clients are considered essentially job-ready (as
long as they have adequate support services) and simply need job placement; about 50% need
Job Search Strategies and Commute Time
long-term training along with the job search program, and need assistance as they move through
a series of jobs and a process of education and skills upgrading; and the remainder (about 30%)
face a variety of severe challenges and may not be able to reach unsubsidized employment
(Pavetti, 1996). Other severe challenges faced by such women include illness, family crises,
emotional or mental problems, alcoholism or drug addiction, and legal difficulties (Pavetti,
Olson, Nightingale, Duke, & Isaacs, 1997).
In studies that analyze the role of different barriers in hindering the transition to work for
women on welfare, there is some indication that both human capital and lack of transportation
are important. However, few studies examine their relative importance, and none examine the
barrier created by the lack of social ties. For instance, Zedlewski (1999) and Loprest &
Zedlewski (1999), using data from the National Survey of America’s Families, find that the
most important barriers are the lack of work experience, the lack of a high school diploma or
GED, and poor mental or physical health. They also find that one of the most significant dif-
ferences between current and former welfare recipients is that current recipients experience
multiple barriers; however, they do not examine the role of transportation.
Similarly, transportation studies generally fail to examine other barriers systematically,
although they do demonstrate the importance of transportation: longer commutes for welfare
recipients tend to increase turnover rates and are not compensated by higher earnings (Ong &
Blumenberg, 1998), and auto ownership improves the likelihood that welfare recipients are
employed (Ong, 1996; Cervero, Sandoval, & Landis, 1999). However, it is unclear how much
work experience these welfare recipients have, and thus prior labor market attachment may be
improving auto ownership (as well as employment outcomes), rather than vice versa.
Although many welfare agencies have conducted surveys of welfare recipients indicating
that transportation is important, few quantify its effects on employment outcomes. However,
one survey from Massachusetts indicates 6% of jobs held by welfare recipients are lost
because of a problem with transportation (Nightingale, Wissoker, Burbridge, Bawden, &
Only one study systematically compares the relative role of various barriers in hindering
the transition. Danziger et al. (1999) find that the most important barriers to working 20 or
more hours per week are perceived discrimination, drug dependence, lack of job skills, less
than a high school education, and a transportation problem, defined as lack of access to a car
and/or lack of a driver’s license.
The 92 study respondents were drawn from the universe of the 8,800 female welfare recip-
ients in San Francisco in 1998, just as the new welfare-to-work regulations were instituted.
Approximately one-third of the sample was recruited via flyers posted at local community-
based organizations and churches, another third via recruitment at a local work-first welfare-
to-work program, and another third by snowballing. Because of the different sampling methods,
the results were tested for differences. The most significant difference was that women at the
work-first program tended to have less work experience and smaller networks. Ideally, the sam-
ple would be weighted to reflect these differences; but without knowing the actual propor-
tions in the population, this is not possible. But because women participating in work-first
programs comprise only a very small percentage of women on welfare in San Francisco (no
more than 10% of all recipients), the sampling method probably results in oversampling of
women less likely to obtain jobs with the assistance of a contact. In other words, the results
probably understate the use of contacts.
158 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
The sample included only adult mothers currently or recently on welfare, from age 19 to
59, with an average age of 32 (the same as in the San Francisco welfare population as a whole).
Although the women in the sample had an average of four and one-half years of work expe-
rience in four jobs, only a handful were working at the time of the interview. Overall, the study
sample was 74% AfricanAmerican, 3% Asian, 12% Latina, 8% Caucasian, and 3% other (mixed
race/ethnicity)—proportions roughly representative of the universe of English-speaking wel-
fare recipients in San Francisco. The welfare population in San Francisco is extremely diverse
ethnically and only about 60% of all welfare recipients speak English at home; of these women,
65% are African American, 12% are Latina, 10% are Asian, and 13% are Caucasian. Alto-
gether, the findings are generalizable to the universe of English-speaking non-immigrant adult
female welfare recipients, particularly African Americans, in large cities with service economies.
Because of racial, class, and other differences between some of the respondents and the in-
terviewer (the author), one of the AfricanAmerican respondents (Sierra) was hired and trained
to conduct some of the interviews. A comparison of the results revealed substantial differ-
ences for only one variable, work experience, which was lower for Sierra’s interviews, possi-
bly because they were shorter.
Interview respondents were paid $20 in cash for participating, and the interviews lasted one
to two hours. The interview format was semi-structured: the same open-ended questions with
the same wording were asked of all respondents. However, the order of the questions varied
as the interviews took the form of a conversation. Interviews ranged over a number of topics.
A series of social network-related questions focused on eliciting how many friends and family
members are located in the Bay Area. Another series of questions drew out the respondent’s
job history, including journey-to-work mode choice, job satisfaction, location, duration, and
remuneration, as well as number of children in the household when each job was held. This
facilitated the comparison of labor market experience before and after becoming a mother and
getting public assistance. Finally, a series of background questions were asked to determine
the respondent’s education, time spent on welfare, and other factors.
To examine job search strategies, the 328 jobs (a figure which only includes jobs in the
formal labor market) in the sample were weighted by respondent case to control for individ-
ual differences in networks, human capital, and strategies. Because welfare policy targets in-
dividuals, not jobs, the appropriate unit of analysis is respondents. To construct the weights
for each job, the total number of jobs was divided by the total number of women in the sam-
ple and then divided by the number of jobs held by a particular respondent.
SEARCH STRATEGIES, JOBS, AND SPACE
The following describes the research findings in three areas: the job search strategies used
in the past to find jobs; the employment characteristics associated with use of social contacts;
and the geography of low-skill jobs, which affects commute times. Then, a binomial logit model
is constructed to explore the reasons women have turned to contacts in the job search. Acon-
cluding section looks at the relationship of job search strategies to transportation considerations.
Job Search Strategies
Contacts play a surprisingly important role in helping female welfare recipients in San Fran-
cisco find work, a role equally or slightly more important than that shown by other studies
examining low-income groups or women (Figure 1). When children are present in the house-
hold, women are more likely to rely on contacts (52% versus 45% for women without children).
Women with children also rely relatively less on direct application, which usually consists
of walking in to an establishment, perhaps because of a sign in a window or a flyer. Using
Job Search Strategies and Commute Time
public or private agencies or the newspaper are relatively less common job search strategies
for low-income women, with or without children.
Relationship Between Search Strategy and Employment Characteristics
The job search strategy used is closely connected to the type of employment gained, espe-
cially for women with children. Table 1 presents difference of means tests for use of contacts
and other job search methods, for women with and without children. For women with chil-
dren, jobs obtained through a social contact pay more and are significantly more satisfying
and longer lasting. Jobs found through contacts are less likely to be temporary (or seasonal),
part-time, or retail jobs, but are more likely to be in the institutional sector (at government
agencies, schools, or hospitals). Interestingly, the use of contacts by women without children
differs in several respects: contacts do not lead to longer job tenure or more satisfaction, or to
jobs in the institutional sector.
Another key difference is in travel time to work. Women with children have significantly
shorter travel times when they use a contact to find work, while search method is not related
to travel time for women without children. It is possible that the type of contact explains these
differences. For instance, while both women with and without children find about 30% of their
jobs through family, women without children are significantly more likely to find jobs through
friends (31%, versus 16% for women with children, p5.02), while slightly less likely to find
jobs through acquaintances (39%, versus 53% for women with children, p5.11). Friends may
be more dispersed than family and acquaintances, leading to more distant jobs for women with-
out children. The use of acquaintances may lead to the greater job satisfaction and stability of
Job Search Strategies for Sample Women With and Without Children
160 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
Difference Between Using a Contact or Other Means for Selected Job Characteristics, Women with Children Versus Women without Children
Women with children Women without children
Contact Other means Contact Other means
Variable N Mean/% (S.D.) Mean/% (S.D.) N Mean/% (S.D.) Mean/% (S.D.)
Wage (1998$) 212 $8.00 ($3.15) $7.55 ($2.53) 104 $8.51 ($3.68) $7.85 ($3.25)
Job length (years) 220 1.27 (1.44) 0.92 (1.62)*108 1.02 (0.96) 1.10 (1.51)
Job satisfaction (dummy, 1 = satisfied) 178 64.1% (48.2%) 30.9% (46.5%)*** 83 38.4% (49.3%) 42.4% (50.0%)
Temporary job (dummy, 1 = temporary) 220 18.2% (38.8%) 29.5% (45.8%)** 108 17.6% (38.5%) 28.8% (45.7%)
Fulltime job (dummy, 1 = fulltime) 220 39.6% (49.1%) 67.2% (47.2%)*** 108 53.5% (50.4%) 82.9% (38.0%)***
Institutional sector job (dummy, 1 = institutional) 220 22.5% (41.9%) 2.0% (14.2%)*** 108 2.8% (16.5%) 9.1% (29.0%)
Office sector job (dummy, 1 = office) 220 11.1% (31.5%) 18.9% (39.3%) 108 18.6% (39.3%) 13.2% (34.2%)
Retail sector job (dummy, 1 = retail) 220 15.3% (36.2%) 37.0% (48.5%)*** 108 27.7% (45.2%) 42.0% (49.8%)
Service sector job (dummy, 1 = service) 220 37.3% (48.6%) 27.4% (44.8%) 108 36.6% (48.7%) 17.4% (38.2%)**
Manufacturing sector job (dummy, 1 = manufacturing) 220 13.8% (34.7%) 14.7% (35.6%) 108 14.4% (35.4%) 18.4% (39.1%)
Travel time (minutes) 181 21.4 (13.0) 27.5 (14.0)*** 82 28.2 (18.9) 28.3 (14.2)
< .01, **
< .05, *
Job Search Strategies and Commute Time
women with children because these weak ties are often in similar occupations and/or located
in the same residential neighborhood. For women with children, the concentration of jobs found
in the institutional sector probably occurs because of their more extensive acquaintance net-
works in local schools and government agencies. But before examining the reasons why women
rely on contacts in the job search, the next section explores another reason for short com-
mutes, the geography of low-skill jobs sought by the women.
Geography of Low-Skill Jobs
Job search strategy is not simply a function of the availability of different resources to help
in the job search, but is also related to the geography of low-skill jobs. Figure 2 maps the
location and density of existing entry-level jobs that are not male-dominated in the Bay Area.
Density of Entry-Level, Non Male-Dominated Jobs in the Bay Area
162 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
The types of low-skill jobs sought by the women in the sample (57% of whom have only a
high school degree or less) are concentrated in downtown and central San Francisco, Oak-
land, and Berkeley. This suggests that low-skill jobs are relatively accessible for women on
welfare in San Francisco.
The local labor market for the women in the sample is essentially limited to San Francisco
and municipalities immediately south; women report looking for few jobs outside this radius.
However, in limiting their job searches essentially to San Francisco, women eliminate sub-
stantial employment opportunities—accessible by mass transit—in adjacent cities. As ex-
plored further below, this may occur in part because travel time is closely related to job search
strategy, particularly for women with children.
Why do women rely on contacts in the job search? Using a binomial logit model with con-
tact as the dependent variable, it is possible to evaluate the role of different factors associated
with the job obtained and the job seeker, for both women with children and women without.
The evidence presented above, as well as in the literature review, suggests that a number of
variables associated with the type of job desired, as well as the individual’s human capital and
network resources, may lead women to use contacts to get jobs.
One set of variables looks at the role of the job characteristics described above, including
wage, tenure, hours (temporary vs. permanent, part-time vs. fulltime), industry sector, and travel
time. For these variables, it is hypothesized that certain types of employment—in particular,
part-time, well-paying, permanent jobs close to home and in the institutional or service sectors—
increase the probability of having used a contact.
A second set of variables examines the role of individual and social network characteristics
to determine the relationship between resources, particularly human capital and network con-
tacts, and use of contacts. Individual variables include a dummy variable for African Ameri-
can race; continuous variables for age and work experience at the time of job search; and dummy
variables for low educational attainment (no high school diploma) or high educational attain-
ment (some college or vocational training) at the time of job search. It is expected that African
Americans are less likely to use contacts, as shown in the literature. Previous research has in-
dicated that job seekers with either high or low educational attainment may rely on contacts,
so no significant relationship is expected. Network variables include number of close friends
in the region and number of family members in the region; since network composition at the
exact time of job search was not available, overall network composition proxies for network
contacts at a particular time. However, because it is very difficult to measure networks of
acquaintances, this measure may correspond only to the use of strong, not to weak ties, except
to the extent to which family and friendship networks are correlated with acquaintanceship
Finally, in order to determine how reliance on contacts differs specifically for women with
children, the model interacts each variable with a dummy for children present in household.
As suggested in Table 1, it is expected that women with children are particularly likely to use
contacts for certain types of jobs, particularly those in the institutional sector that are close to
The results from the logistic regression are shown in Table 2. Overall, the model predicts
75.3% of the cases, with a x
of .00. In general, the only variables that affect the decision to
use a contact are the hours worked per week (fulltime or part-time) and the work experience
of the job seeker at the time of the job search. For women with children, institutional jobs and
travel time to work are also highly significant. That most of the job, individual, and network
Job Search Strategies and Commute Time
characteristics are not significant is not surprising, given that research has found wide use of
contacts among groups of different race or ethnicity, income and skill levels, and network com-
position. In other words, the results suggest that contacts are an important job search method
for everybody, regardless of the type of job sought, network size, race or ethnicity, and education.
Binomial Logit Model: The Decision to Use a Contact
Analytical Variables Parameter
estimate S.E. Exponentiated
Variables interacted without child present
Wage (1998$) 0.103 0.116 1.108
Job length (years) −0.415 0.394 0.661
Temporary job (dummy, 1 = temporary) −0.989 0.826 0.372
Fulltime job (dummy, 1 = fulltime) −2.078*** 0.850 0.125
Institutional sector job (dummy, 1 = institutional) −2.077 1.579 0.125
Office sector job (dummy, 1 = office) −0.065 0.942 0.937
Retail sector job (dummy, 1 = retail) −0.311 0.820 0.733
Service sector job (dummy, 1 = service) 1.172 1.067 3.228
Manufacturing sector job (dummy, 1 = manufacturing) [omitted]
Travel time (minutes) 0.035 0.022 1.035
African-American (dummy, 1 = African-American) −0.270 0.862 0.763
No high school diploma at time of job search
(dummy, 1 = no diploma) −0.252 0.750 0.777
Some college or vocational training at time of job search
(dummy, 1 = college) −0.308 0.999 0.735
Work experience (in weeks) at time of job search 0.009** 0.004 1.000
Age at time of job search −0.023 0.077 0.977
Number of close friends living in the Bay Area −0.123 0.184 0.885
Number of family members living in the Bay Area −0.053 0.130 0.948
Variables interacted with child present (dummy)
Wage −0.023 0.132 0.977
Job length 0.221 0.416 1.247
Temporary job 0.151 0.965 1.164
Fulltime job 1.844*0.982 6.322
Institutional sector 5.125*** 2.058 168.095
Office sector −0.762 1.141 0.467
Retail sector −1.494 1.031 0.224
Service sector −0.828 1.238 0.437
Manufacturing sector [omitted]
Travel time −0.079*** 0.028 0.924
African-American −0.276 0.987 0.759
No high school diploma 0.310 0.948 1.363
Some college or vocational training −0.500 1.137 0.606
Work experience (in weeks) −0.010** 0.000 0.990
Age 0.050 0.004 1.051
Number of close friends 0.240 0.198 1.271
Number of family members 0.012 0.155 1.012
Constant 1.258 1.432 3.519
N = 232; x
= .00; predicted = 75.3%; −2 log likelihood = 251.43
< .01, **
< .05, *
164 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
In the overall model, the coefficient for fulltime is negative and significant (p,.01), in-
dicating that using a contact is less likely when looking for fulltime work. The exponentiated
coefficient for fulltime indicates only a small effect, an increase of 12.5% in the likelihood of
using a contact to find work that is not fulltime. However, the coefficient for work experience
is positive and significant (p,.05), indicating that more work experience—which may lead
to more extensive acquaintanceship networks—increases the likelihood of having found a job
through a contact.
Likewise, for women with children alone, fulltime jobs and work experience are significant
(at p,.10 and p,.05, respectively). However, the direction of the sign changes for both. If
it is for fulltime, rather than part-time jobs, that women with children use contacts, contacts
may be more tolerant of the special needs of women with children who are working fulltime.
More work experience does not lead to the use of contacts, perhaps because contacts met through
work are more likely to lead to jobs outside of the local neighborhood. Women with children
also are significantly likely to use a contact to find jobs in the institutional sector; again, this
probably occurs because of their familiarity with workers at local schools and government agen-
cies, who help them with finding jobs.
Most significant, at p5.00, is the influence of travel time on the likelihood of having used
a contact for women with children. Women with children have significantly shorter travel times
when they use contacts to find work; interestingly, this contrasts with the overall model, in
which travel time to work is insignificant. This confirms the work of Hanson & Pratt (1995),
indicating that for women with children in particular, the decision to use a contact in the job
search is related closely to the desire to work close to home. Most of the other variables are
not significant factors in finding a job through a contact, because use of contacts is not con-
sistently associated with job characteristics like higher wages or longer job length, nor with
individual or network characteristics such as human capital or strong tie network strength. But
for women with children seeking to work close to home, use of contacts becomes important.
Women with children are more likely to rely on contacts than women without children be-
cause they prefer to work close to home. The desire to work close to home is accommodated
by relying more on local acquaintances than friends, who are more spatially dispersed, in the
job search. Thus, these findings suggest that the desire for shorter travel times increases the
probability of having used a contact, rather than commute times decreasing incidentally as a
result of having used a contact. This interpretation is plausible for two reasons: women with
children are more likely overall to find jobs through contacts, and they only use the contacts
who are likely to help them work close to home. Before turning to the policy implications of
these findings, the next section looks in more detail at the importance of job accessibility and
the transportation needs of women on welfare.
The Role of Transportation in Job Search and
Retention for Low-income Mothers
Although mothers may use contacts to work close to home, transportation considerations
also shape the job search overall: women look for jobs in areas to which they are able to com-
mute. The importance of job location varies depending on several factors, including career
goals, auto ownership, and degree of familiarity with other areas of the city, but probably the
most important factor is educational attainment. Women with low educational attainment (no
high school diploma or GED) have significantly shorter travel times (22 minutes, p5.01)
than women who have either finished high school or have some college or postsecondary voca-
tional education (27 minutes). This suggests that the women falling into local jobs do so in
part because they lack the skills to compete in the regional labor market. Rosa is uncomfort-
Job Search Strategies and Commute Time
able outside of her own neighborhood, in part because she quit school in the ninth grade. Down-
town is close to her house, about a mile, but asked if she has applied for retail jobs, she responds
No. Every time I like look I don’t see nothing. Lot of things are like out in the Financial
District. I don’t want to go all the way out there . . . I would feel too uncomfortable down
that way, so I never tried . . . The people, they feel like snobs. It’s like they’re better than
For the women with relatively high educational attainment, the type of work they do is much
more important than job accessibility, and they will not sacrifice their job goal—or their ex-
pected pay—in order to work closer to home. Asked if she would take approximately 20%
less than her reservation wage to work at a job closer to home, Grace’s response was typical:
Interviewer: Where would you want to work, if you could work anywhere?
Grace: Umm, work? I don’t really care about that. I mean, as long as I have the job that I
really want, I work in the medical setting, right in the hospitals....
Interviewer: If you were paid a little less [than your reservation wage], but it was closer to
home, would you take it? If it were like, $9.50 an hour, but it’s like, you could walk from
Grace: I think I would umm, I mean, I would think about other circumstances, like does it
have job advancement, and things like that. Because I could always get a car and if the job
is better, I could advance...
Examining the reasons why employment ends provides another perspective on the role of
transportation. As Figure 3 shows, only 4% of the jobs in the sample ended because of trans-
portation problems. In contrast, almost half of all jobs were lost for personal reasons (dissat-
isfaction, pregnancy, or getting fired); and a third of the jobs ended for demand-side reasons
(temporary employment, layoffs, or company relocation). Although undoubtedly the minor role
of transportation in job retention stems mostly from the accessibility of low-skilled jobs in
San Francisco and the location of many jobs close to home, Althea provides another explanation—
that transportation problems can be solved—“If there’s a problem with the bus, I can get a
ride from just about anybody—so long as I have a good reason, like a paying job to go to, so
I can return the favor somehow.”
This suggests that traveling to jobs may be of minor concern relative to getting and keeping
a job. Not only do most mothers seek to work near home, but also the obstacles they face in
retaining jobs are more complicated than the logistical problem of transportation access. How-
ever, those with greater levels of human capital seek jobs that offer career ladders.As a result,
they are more likely to try to find jobs further from home. For these women, welfare-to-work
policy will need to focus more on transportation assistance.
Social contacts play an important role in the job search for low-income mothers. Most crit-
ically, they offer a way for women to work near home; but they also provide more stable and
satisfying jobs, perhaps because using a contact as an intermediary between employer and job-
seeker helps to build the mutual trust so important for helping women on welfare enter and
166 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
remain in the labor market. Because contacts are such a vital search strategy, improving the
connections of women on welfare to the formal labor market will clearly help their transition
into the workforce. While many women rely on social contacts to find jobs, some lack the
kinds of social ties that are connected to well-paying, stable, and satisfying jobs in the formal
workforce. This indicates the importance of network strategies to assist in the job search.
Although most low-income mothers try to minimize their commute time, they still need trans-
portation assistance. As in many other counties, women on welfare in San Francisco receive
discounted or free local bus passes, in essence an indirect wage subsidy. But beyond such sub-
sidies, transportation assistance such as reverse commute programs is unlikely to address the
employment problems of women on welfare. Only with the educational attainment that allows
them to obtain jobs consistent with their career goals will low-income mothers commute lon-
ger, thus requiring such transportation assistance.
States have generally developed welfare-to-work programs without much concern for net-
work strategies or the actual commuting patterns of low-income mothers. With limited re-
sources available and inadequate knowledge of what works, the vast majority of states have
chosen to develop “work-first” programs, many of which are modeled on the California Greater
Avenues to Independence (GAIN) program in Riverside County. Likewise, many localities have
quickly taken advantage of funding from the Transportation Equity Act for the 21
(TEA-21) to expand existing transit route coverage and schedules. The following examines in
more detail the potential for network strategies and transportation assistance for facilitating
the transition to work for women on welfare.
Two different approaches to workforce development fall under the category of network strat-
egies: work-first programs and employment brokering. Although work first is a short-term
Reason Why Job Ended
Job Search Strategies and Commute Time
approach to job search emphasizing soft skills, it also serves as a networking strategy. Work-
first participants become network resources through sharing job leads, and networking also
occurs as employers recruit workers directly from the program. Employment brokering is a
more long-term network approach linked to human capital development, as potential employ-
ers and job seekers get to know each other during training programs.
Role for the Work-First Approach
Although early research on work-first programs criticized their limited impact on earnings
and poverty levels, a recent series of studies of pre-PRWORAwelfare-to-work programs shows
significant improvement in earnings (Freedman et al., 2000; Grubb, 1995). Whatever their long-
run effectiveness, work-first programs play a valuable role in terms of motivating women to
look for jobs, providing them with peer and technical support, and connecting them with em-
ployers willing to hire them.
Like other women experiencing labor force interruptions due to child rearing, women on
welfare often find jobs when they aren’t actively looking. Yet because such opportunities arise
only infrequently, many women on welfare find the work-first program a convenient way to
jump-start the job search and ready themselves for the formal labor market. To Kay, the work-
first approach was very helpful:
They have really inspired me to, you know, when I first came here I guess, I can say, hon-
estly say I was just a little low on self-esteem, not that much . . . I needed just a little pick-up
because due to the fact that I was just in and out of temporary jobs and that stuff gets old.
Other women value the support the program provides, both in terms of peer encouragement
and technical assistance. Women like Peaches and Sharon, who have lived in housing projects
for most of their adult lives, distrust their neighbors too much to speak with them, but found
women they met at the program to be invaluable resources for job leads and daily motivation
to continue the job search.
Work-first programs also provide opportunities to network with local employers. Firms that
experience chronic low-skilled labor force shortages, such as retailers during the Christmas
season, often recruit through the program. In this way work-first programs may shield job seek-
ers from discrimination, because firms recruiting through the program are expressing their will-
ingness to hire women on welfare. At the same time, these firms are particularly likely to be
offering short-term, dead-end jobs at low wages. Yet this is one way to get some quick expe-
rience on the light resumes of women returning to the workforce after childrearing interruptions.
However, work-first will only help a certain group of women make the transition into the
workforce—in particular, younger women and women who have experienced long-term labor
force interruptions and need help only with job readiness skills, such as resume preparation.
For women with more complicated needs than obtaining some quick work experience, the work-
first approach will be ineffective. For such women, employment brokering may be a more
effective network strategy.
Employment brokering is a network strategy that offers job training closely linked to em-
ployer needs. In employment brokering programs, such as the Center for Employment Train-
ing in San Jose (and replicated elsewhere), employers are introduced to potential employees
early in the job training process, in effect creating social networks (Giloth, 1995; Melendez &
Harrison, 1998). Employment brokering programs identify areas of labor demand, develop train-
ing curricula in cooperation with industry, and conduct aggressive placement programs.
168 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
Many women re-entering the workforce have needs and expectations that work-first pro-
grams cannot fulfill. In particular, older women dislike the soft skills emphasis of work-first
and the quick placement in dead-end jobs, whether because they can’t find the type of work
they want to do, they can’t acquire the technical skills they need, or they simply resent the
Women like Althea, who is almost 50 years old and has been doing child care on the side
for almost all her life, have a hard time finding jobs through the Work-First Program, where
most job leads are for retail jobs; the posted job openings in the fields that interest Althea,
child and elderly care, require certification. Eve would like to work as a security guard, but
has been unable to find a position, despite trying for seven months, through the work-first pro-
gram. Instead, she has been getting jobs in fields for which job leads are more common, like
telemarketing—only to quit the jobs quickly and return to the program to begin the search
anew. She is frustrated:
Plenty of times I wanted to just tear up everything, throw my book bag in the garbage, and
just go home and be like I’m going to stay on welfare for the rest of my life until they kick
At age 32, Eve has a GED, a year of college, and four children, but very little work expe-
rience. Whether because nobody has informed her of the availability of job training opportu-
nities, or because of her own lack of interest in advancing her education, she continues to try
for work through simple job search, gambling that she can get any training she needs on the
job. Yet the minimal kind of networking provided by the local work-first program is not enough
to get her foot in the door at the type of job that interests her.
An employment brokering approach would help women like Althea and Eve to learn how
to do the kind of work they’d like to do; with the assurance of a solid job prospect at the end
of training, such discouraged women are more likely to agree to participate in training pro-
grams. By linking job seekers to well-paying jobs in sectors traditionally difficult to access,
such as office work, employment brokers may counter the tendency of women on welfare to
work close to home through social ties from the neighborhood. If women can realize a payoff
for longer commutes in the form of higher wages and/or jobs with career ladders, they are
more likely to accept jobs with longer travel times from home. At this point, transportation
assistance will become more critical.
Role for Transportation Assistance
TEA-21 provides $750 million in funding over five years for Job Access projects, which
are to provide new transportation services or extend existing services to connect welfare re-
cipients and low-income workers to jobs in distant areas, and Reverse Commute projects, which
are meant specifically to transport the workers from urban, suburban, and rural areas to sub-
urban employment opportunities.
Although funding for accessibility is abundant, as little is known about the effectiveness of
transportation strategies in improving job accessibility as is known about the role of transpor-
tation as a barrier to the employment transition. However, some initial results are available
from a four-year demonstration project called Bridges to Work. Bridges to Work is intended
to provide 3,100 job seekers in five cities (Denver, Milwaukee, Chicago, Baltimore, and St.
Louis) with placement assistance and transportation to suburban job opportunities. To date,
it has succeeded in placing 61% to 67% of enrollees. The problem with placement is due to
unanticipated barriers to matching workers to jobs: these early results indicate that physical
Job Search Strategies and Commute Time
access to work is a simple logistical problem to solve, while workforce development, in terms
of recruitment, assessment, preparation, placement, and retention, is the major policy chal-
lenge (Elliott, Palubinsky, & Tierney, 1999). These findings are not surprising, given the low
levels of success that reverse commute programs have experienced historically (O’Regan &
Quigley, 1999).As Waller & Hughes (1999) point out, reverse commuting strategies may also
be less cost effective than policies facilitating auto ownership, but states have been slow to
provide assistance for auto purchase.
However, improved accessibility to jobs may help the women on welfare who are job-
ready to access distant jobs with career ladders, for instance the clerical opportunities that are
increasingly concentrated in suburban areas. While auto ownership may not increase employ-
ment significantly, programs that facilitate auto commuting are one way to give low-income
mothers the flexibility to sustain labor market attachment while bearing substantial household
responsibilities. Finally, although many women on welfare are unlikely to work far from home,
most receive very low wages, which makes it important to continue subsidizing the commute
costs of all women through indirect wage subsidies such as transit passes.
Need for Multiple Strategies
Networking and transportation assistance alone are insufficient: to ensure that women can
find and keep stable employment, multiple strategies are necessary. Community economic
development strategies helping to make well-paying jobs available in the neighborhood can
improve job accessibility more effectively than reverse commute programs, given the prefer-
ence to work nearby. Traditional education can help to address problems such as the inability
to access jobs with career ladders (e.g., temporary clerical jobs obtained through agencies),
the lack of appropriate credentials for the formal workforce, and the lack of familiarity with
new technology that occurs with lengthy labor force interruptions. Direct and indirect subsi-
dies to increase wages improve employment outcomes for jobs obtained outside of networks
(as for retail jobs obtained through direct application), ease the transition into the workforce
for women with complicated child care obligations, and motivate women to leave informal
for formal work (Edin & Lein, 1997). Micro-enterprise programs, although insufficient to pull
women out of poverty, can provide self-employment training and small amounts of credit (gen-
erally less than $1,000), and thus provide a first step toward self-sufficiency or at least the
formal labor market (Servon, 1997).
A variety of different policies will help women on welfare make the transition into the work-
force, depending on how connected they are to employers, how much education and work ex-
perience they have, where they live relative to job opportunities, and of course, the wages
employers offer. But the question remains which policies to apply to which women. Based
upon the educational attainment and use of contacts by this sample, not more than 40% of the
women—those with some advanced education and/or work experience—may take advantage
of transportation assistance beyond simple subsidies. Approximately 15 to 20% of the women—
those with very little work experience and few social ties to help in the job search—are likely
to benefit from the work-first approach. About half—interestingly, the same percentage as sug-
gested by Pavetti (1997)—will require some sort of employment brokering program to build
better connections to employers.
Since the passage of PRWORA, most states have adopted a “one-size-fits-all” approach to
the transition to work for women on welfare. The emphasis on both work-first and transpor-
170 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
tation assistance presupposes that job accessibility is a simple matter of helping women learn
how to search for and then travel to jobs. Yet job accessibility is related so strongly to the
social connectedness and the human capital attainment of women that training in search tech-
niques and assistance in transportation are secondary concerns. For most women, building con-
nections to employers and human capital is what will make jobs accessible; then, for a few,
transportation assistance will help with access and retention.
Yet concerted network strategies such as job training offer only a short-term fix for women
on welfare. In the long term, if mothers are going to rely on contacts in order to work close to
home, the most effective approach may be to increase the density of both economic and social
activity in their neighborhoods. By facilitating local job creation, particularly in local schools,
hospitals, and agencies, policy will ensure that these valued neighborhood job opportunities
are available. With more opportunities for social interaction, mothers are more likely to be
able to use friendship and acquaintanceship networks for assistance in the job search.
Some of these findings may not be easily generalized to women on welfare in other parts of
the country. For instance, few labor markets have the job densities of San Francisco, although
with an average travel time of 26 minutes for women in the sample the overall job accessibil-
ity is arguably not very high. But relative to other cities, particularly in the south and west,
San Francisco probably offers a relatively high number of neighborhood-based job opportunities.
Yet despite the greater density of opportunities in San Francisco, the experience of these
women is instructive and familiar: the differences in job search and commuting patterns be-
tween women with and without children are likely to occur anywhere, although the degree
may vary. No matter what the spatial pattern of non-male-dominated low-skill employment in
their regions, many women with children will try to minimize their commutes, and are likely
to use social networks in order to do so. Although local social institutions and spatial context
shape the struggle of poor women to manage work and child rearing in particular ways, the
influence of such social and spatial structures on economic outcomes is universal.
ACKNOWLEDGMENTS: The work that provided the basis for this publication was supported by disser-
tation grant funding from the U.S. Department of Housing and Urban Development. I am grateful to
Rachel Weinberger, Joe Ritter, and the three anonymous reviewers for their help and comments.
Braddock, J. H. II, & McPartland, J. M. (1987). How minorities continue to be excluded from equal em-
ployment opportunities: Research on labor market and institutional barriers. Journal of Social Is-
Calzavara, L. (1983). Social networks and access to jobs: A study of five ethnic groups in Toronto (Cen-
tre for Urban and Community Studies Research Paper No. 145). Toronto: University of Toronto.
Campbell, K., & Rosenfeld, R. (1985). Job search and job mobility: Sex and race differences. Research
in the Sociology of Work,3, 147–174.
Cervero, R., Sandoval, J. O., & Landis, J. (1999). Transportation as a stimulus to welfare-to-work: Ev-
idence from the San Francisco Bay area. Unpublished paper.
Clark, G. L., & Whiteman, J. (1983). Why poor people do not move: Job search behaviour and disequi-
librium amongst local labour markets. Environment and Planning A,15, 8–104.
Cooke, T. (1997). Geographic access to job opportunities and labor-force participation among women
and AfricanAmericans in the Greater Boston metropolitan area. Urban Geography,18(3), 213–227.
Corcoran, M., Datcher, L., & Duncan, G. (1980). Information and influence networks in labor markets.
In G. Duncan & J. Morgan (Eds.), Five thousand American families: Patterns of economic progress
(Vol. VIII, pp. 1–37). Ann Arbor, MI: Institute for Social Research.
Job Search Strategies and Commute Time
Danziger, S. K., Corcoran, M., Danziger, S., Heflin, C., Kalil, A., Levine, J., Rosen, D., Seefeldt, K.,
Siefert, K., & Tolman, R. (in press). Barriers to the employment of welfare recipients. In R. Cherry
& W. Rodgers (Eds.), The impact of tight labor markets on black employment problems. New York:
Edin, K., & Lein, L. (1997). Making ends meet: How single mothers survive welfare and low-wage work.
New York: Russell Sage Foundation.
Elliott, M., Palubinsky, B., & Tierney, J. (1999). Overcoming roadblocks on the way to work: Bridges to
work field report. Philadelphia, PA: Public/Private Ventures.
Fischer, C. S., Jackson, R. M., Stueve, C. A., Gerson, K., Jones, L. M., & Baldassare, M., (1977). Net-
works and places: Social relations in the urban setting. New York: The Free Press.
Fischer, C. S. (1982). To dwell among friends: Personal networks in town and city. Chicago, IL: Uni-
versity of Chicago Press.
Freedman, S., Friedlander, D., Hamilton, G., Rock, J., Mitchell, M., Nudelman, J., Schweder, A., & Storto,
L. (2000). Evaluating alternative welfare-to-work approaches: Two-year impacts for eleven pro-
grams. New York: Manpower Demonstration Research Corporation.
Giloth, R. (1995). Social investment in jobs: Foundation perspectives on targeted economic develop-
ment during the 1990s. Economic Development Quarterly,9(3), 279–289.
Granovetter, M. (1988). The sociological and economic approach to labor markets: A social structural
view. In G. Farkas & P. England (Eds.), Industries, firms, and jobs: Sociological and economic
approaches. New York: Plenum Press.
Granovetter, M. (1995). Getting a job: A study of contacts and careers (2nd ed.). Chicago: University of
Green, G., Tigges, L., & Diaz, D. (1999). Racial and ethnic differences in job-search strategies in At-
lanta, Boston, and Los Angeles. Social Science Quarterly,80(2), 263–278.
Grubb, W. N. (1995). Evaluating job training programs in the United States: Evidence and explana-
tions. Berkeley, CA: National Center for the Study of Vocational Education.
Hanson, S., & Johnston, I. (1985). Gender differences in work-trip lengths: Explanations and implica-
tions. Urban Geography,6, 193–219.
Hanson, S., & Pratt, G. (1995). Gender, work, and space. New York: Routledge.
Ihlanfeldt, K. R., & Sjoquist, D. L. (1998). The spatial mismatch hypothesis: Areview of recent studies
and their implications for welfare reform. Housing Policy Debate,9(4), 849–892.
Johnston-Anumonwo, I. (1992). The influence of household type on gender differences in work trip dis-
tance. Professional Geographer,44, 161–169.
Kain, J. F. (1968). Housing segregation, Negro employment, and metropolitan decentralization. Quar-
terly Journal of Economics,82(2), 175–97.
Lin, N., & Dumin, M. (1986). Access to occupations through social ties. Social Networks,8, 365–385.
Lin, N., Ensel, W., & Vaughn, J. J. (1981). Social resources and strength of ties: structural factors in
occupational status attainment. American Sociological Review,46, 393–405.
Loprest, P. J., & Zedlewski, S. R. (1999). Current and former welfare recipients: How do they differ?
Washington, DC: The Urban Institute.
Madden, J. (1981). Why women work closer to home. Urban Studies,18, 181–194.
Marsden, P., & Hurlbert, J. (1988). Social resources and mobility outcomes:A replication and extension.
Social Forces,66(4), 1038–1059.
Melendez, E., & Harrison, B. (1998). Matching the disadvantaged to job opportunities: Structural expla-
nations for the past successes of the center for employment training. Economic Development Quar-
Moore, G. (1990). Structural determinants of men’s and women’s personal networks. American Socio-
logical Review,55, 726–735.
Nightingale, D., Wissoker, D., Burbridge, L., Bawden, L., & Jeffries, N. (1990). Evaluation of the Mas-
sachusetts employment and training (ET) choices program. Washington, DC: The Urban Institute.
Ong, P. (1996). Work and car ownership among welfare recipients. Social Work Research,20(4), 255–262.
Ong, P., & Blumenberg, E. (1998). Job access, commute and travel burden among welfare recipients.
Urban Studies,35(1), 77–93.
172 6JOURNAL OF URBAN AFFAIRS 6
Vol. 23/No. 2/2001
O’Regan, K., & Quigley, J. (1999). Accessibility and economic opportunity. In J. Gómez-Ibáñez, W.Tye,
& C. Winston (Eds.), Essays in transportation economics and policy. Washington, DC: Brookings
Pavetti, L. (1996, December). Testimony to the San Francisco Mayor’s Welfare Reform Task Force.
Pavetti, L. (1997). Against The Odds: Steady Employment Among Low-Skilled Women. Washington, DC:
Annie E. Casey Foundation.
Pavetti, L., Olson, K., Nightingale, D., Duke, A., & Isaacs, J. (1997). Welfare-to-work options for fami-
lies facing personal and family challenges: Rationale and program strategies. Washington, DC:
The Urban Institute.
Preston, V., McLafferty, S., & Hamilton, E. (1993). The impact of family status on Black, White and
Hispanic women’s commuting. Urban Geography,14, 228–250.
Servon, L. J. (1997). Microenterprise programs in US inner cities: Economic development or social wel-
fare? Economic Development Quarterly,11(2), 166–180.
Waller, M., & Hughes, M. A. (1999). Working far from home: Transportation and welfare reform in the
ten big states. Philadelphia, PA: Progressive Policy Institute.
Wilson, W. J. (1996). When work disappears: The world of the new urban poor. Princeton, NJ: Prince-
ton University Press.
Wyly, E. (1998). Containment and mismatch: Gender differences in commuting in metropolitan labor
markets. Urban Geography,19(5), 395–430.
Zedlewski, S. (1999). Work activity and obstacles to work among TANF recipients. Washington, DC:
The Urban Institute.
Job Search Strategies and Commute Time