Modern Economy, 2010, 1, 195-205
doi:10.4236/me.2010.13022 Published Online November 2010 (http://www.SciRP.org/journal/me)
Optimal Time and Opportunity Cost of Job Search in
Low-Income Groups: An Out-of-the-Job Search Model
Tamara Todorova, Veselina Dzharova
American University in Bulgaria, Department of Economic s , Blagoevgrad, Bulgaria
Received September 3, 2010; revised October 4, 2010; accepted October 8, 2010
Abstract
Our paper studies the causes of poverty from the perspective of job search. We show that poor people remain
poor because they have less time and initial endowment to search for a better job. Initial endowment is key to
successful job search, as one can afford not to work and search longer for a better job. Having an initial en-
dowment, a worker is able to educate or re-qualify himself. Working long hours and obtaining low pay, poor
people have little time to look for a better job. Low-paid, low-skilled jobs rarely allow on-the-job search like
high-paid positions where with the help of contacts and a lot of idle time professionals seek better jobs. Quit-
ting in order to find a better job increases the opportunity cost of search for poorer people. Since they do not
have any accumulated income, they can only live off their salary. With less income and time, poorer people
are less likely to get educated since education requires both wealth and free time. But being less educated,
they are likely to remain poor as education is a promise for success in contemporary society. Thus, they re-
main in the vicious circle of poverty. In order to prove this hypothesis we investigate optimal search time for
a better job as dependent on factors such as wage rate, individual’s income, education, and skills.
Keywords: Job Search, Optimal Search Time, Initial Endowment, Employment, Educational Level, Povert y,
Low-Skilled Jobs
1. Introduction
Poor, uneducated people are often said to remain in the
vicious circle of poverty and it is a trap for many in the
world economy on the eve of a new century. There are
various ways in which poverty plagues workers around
the globe–in our paper we study just two mechanisms.
More specifically, those are the lack of time to search for
a better job, while overwhelmed by a low-paid one, and
the lack of initial endowment or wealth. Our paper has,
to some extent, been inspired by Barbara Ehrenreich’s
book “Nickel and Dimed: On (Not) Getting by in Amer-
ica” [1], although we have previously been familiarized
with Günter Wallraff’s “Ganz Unten” [2]. Ehrenreich’s
idea is that poor people create much greater value than
what they are paid for, thus, letting others live off their
generosity, opposite to the common belief. The famous
journalist takes low-paid jobs in order to experience the
hardships of poverty. She finds that ordinary low-wage
workers are paid much less than they deserve and than
the subsistence level, barely provid ing for themselves, let
alone a family. Because one job is not enough, a worker
has to take at least two in order to live indoors. Low-
prestige, low-level occupations at th e same time require a
lot of mental and physical efforts. Ehrenreich writes:
“When someone works for less pay than she can live
on ... she has made a great sacrifice for you ... The
‘working poor’ ... ar e in fact the major philanthropists of
our society. They neglect their own children so that the
children of others will be cared for; they live in
substandard housing so that other homes will be shiny
and perfect; they endure privation so that inflation will
be low and stock prices high. To be a member of the
working poor is to be an anonymous donor, a nameless
benefactor, t o everyone.” [1, p. 221]
Intrigued by Ehrenreich’s description of the poor lay-
ers of the American labor force with their struggle for
survival and better life, we study how low pay and hard
labor prevent workers from obtaining wealth and social
status. More specifically, we show that, having less
search time and initial endowment, poor people remain
trapped in a low-level, low-paid po sition with no chance
of escaping the whirlpool of poverty. Working long
hours for low pay, poor people have little time to look

196 T. TODOROVA ET AL.
for a better job. Low-paid, low-skilled jobs do not allow
on-the-job search whereas with the help of professional
contacts and availab le time people in high -paid positions
seek better jobs. In their search for a better job peop le in
low-paid, less prestigious jobs have to quit but quitting
increases the opportunity cost of search for them. Since
they do not have any accumulated income, they can only
live on their salary. With less income and time, poor
people remain less educated and poor. We study optimal
search time for a better job as dependent on factors such
as wage rate, individual’s income, education, and skills.
More specifically, setting to be optimal search time,
we express the comparative static effect of the number of
years to search the minimum wage level the
educational level o, etc., on the optimal search time.
Some of the main findings of our paper can be summa-
rized, as follows:
*
t
,ye,
o
w
1) The higher the initial minimum wage, the smaller
the search time would be and the greater the prospects of
the individual to find a better-paid job. The longer the
time to be at an ideal, better-paid job, the longer one
would search for this job. The greater the total number of
years to work, the longer the job search time.
2) The higher the initial income of an individual, the
greater his prospects of obtaining a better paid job are
and the greater the free time of the job searcher is. Thus,
rich individuals have more time to search for a better job
and are more likely to find it.
3) As common sense dictates, a higher educational
level promises a higher wage. More educated people
need less time to search for a better job and are less
likely to experience long periods of unemployment.
4) As expected, the optimal number of labor hours
depends negatively on the wage rate and non-labor in-
come and positively on the price of free time. Thus, poor
agents, whose labor income should satisfy the subsis-
tence level of consumption, will face higher price of free
time and would work more labor hours to attain their
optimum utility level of co nsumption.
There is extensive economic literature dedicated to job
search. Search theory generally is a branch of microeco-
nomics that studies optimal decision-making of rational
individuals with regard to potential opportunities of dif-
ferent quality, accounting for both the external and in-
ternal costs of search. External costs include those of
acquiring information and the opportunity cost of search
time. External costs are exogenously determined and the
individual may choose only whether to incur them or not.
Internal costs include the mental effort to undertake the
search, sort the incoming information, and integrate it
with what the consumer already knows. Internal costs are
determined by the consumer's ability to undertake the
search, which depends on intelligence, prior knowledge,
education and training [3].
In labor economics, search theory is related to job-
seeking models where a worker searches for a job that
pays a high wage or offers desirable benefits, safe and
desirable working conditions. The search process is re-
lated to the optimization proble m of how much time and
effort to spend searching, and which offers to accept or
reject. In the 1960s Stigler [4,5] and Alchian and Allen
[6] were among the first economists to study the problem
of search for employment in an economy where informa-
tion is costly and uncertainty prevails. McCall [7] put
this economically important problem in a mathematical
framework. His search model obtains an optimal search
policy, where the worker knows the distribution of wages
for his particular skills, the cost of search is a known
constant, and the value of money is constant. Holding
fixed job characteristics, the searcher makes a job search
decision with respect to his reservation wage, i.e., the
lowest wage he is willing to accept. Job offers arrive
periodically and the searcher accepts or rejects them as
they arise. The worker’s optimal strategy is simply to
reject any wage offer lower than his reservation wage,
and accept any wage offer higher than his reservation
wage. The reservation wage may change over time if
some of the conditions assumed are not met. For example,
if the worker is risk averse, the reservation wage will
decline over time if the worker gradu ally runs out of time
and income while searching. The reservation wage will
also differ for two jobs of different characteristics; that is,
there will be compensating wage differentials between
different types of jobs.
McCall’s basic model serves as a basis for later works
among which those of Mortensen [8], Burdett [9], Pis-
sarides [10], Jovanovich [11], and Van den Berg [12].
These authors analyze important issues connected to job
search and the behavior of unemployment [10], wage
determination [9,11,13], job duration, job turnover [12],
quit rates [9], and unemployment insurance and em-
ployment protection [10]. These seminal economic mod-
els have given birth to extensive research in the context
of search theories, and the empirical studies serve to in-
vestigate the stylized facts of the classical theory. Search
theories have discussed optimal search time in the con-
text of other issues–Kahn [14] estimates the relationship
between search time and resulting wage and further the
duration span of unemployment is investigated. Yoon
[15] includes search time as endogenous factor deter-
mining unemployment duration and pays attention spe-
cifically to intensity of the search effort in determining
search costs. Optimal search time is also discussed as the
marginal benefit of not being employed or the value of
unemployment benefits during the unemployment dura-
tion [16].

T. TODOROVA ET AL.
197
Our purpose is to study an individual’s optimal strat-
egy when searching for a job in light of the trade-off in-
volved: the cost of remaining unemployed while search-
ing versus the added value of the option to find a better
job. We are specifically interested in partial search mod-
els, where the focus is on the behavior of the unem-
ployed individuals looking for better jobs rather than the
firms’ behavior in the labor market. We develop a basic
out-of-the job search model that focuses on how optimal
time and opportunity cost differ among low-income and
high-income individuals, ceteris paribus, that is, given
they have the same initial tastes, talen ts, an d ability ren ts.
Low-income searchers who have low or no accumulated
wealth to devote to job-seeking cannot afford to remain
unemployed for a long time and tend to lower their res-
ervation wages below what co rresponds to their skills. In
contrast, more wealthy seekers have more free time and
disposable income to engage in search for a better job.
Given their higher probability to find a better job, the
value of their choice is likely to be high er.
Income inequalities have further implications for job
searchers. Human capital theories study the opportunity
cost of purchased inputs in human capital production.
The basic model formulated on optimal human capital
investment is Becker’s model [17]. This model generates
general description of how ability and family wealth in-
teract to determine the distribution of lifetime earnings.
The opportunity cost of human capital investment de-
pends on family incomes because families should pay for
the post-secondary schooling required to gain additional
skills. In this sense, wealthy agents can finance their
education on better terms and in general can invest more
and boost future earnings. Becker defines this difference
in financing opportunities as “unequal opportunity,” since
poor agents face higher opportunity cost of funds even
when they have the same personal abilities. Following
Becker’s model, we assume that wealthy individuals are
more educated since they face lower costs of financing
and would invest more in developing their skills. As a
result, they would have more specialized skills to per-
form highly-skilled professional labor. Further, we in-
vestigate how the opportunity cost of job search adds to
the opportunity cost of human capital in defining wage
differentials. Poor agents face higher opportunity cost of
delay and tend to choose from a limited range of oppor-
tunities, while rich individu als can afford to reject multi-
ple job opportunities that do not correspond with their
skills. Therefore, wealthy individuals find better jobs
(including higher wage, more benefits, and safer working
conditions) than their lower-income counterparts.
The analysis can further be extended to include impli-
cations for individual careers. Learning models in the
theory of earnings distributions [11,18] state that without
information about their relative talents for different types
of jobs, workers cannot make choices that maximize
earnings. Thus, workers gain from work experience in
two ways: first, they are paid the current wage, and sec-
ond, they receive information about their skills that can
boost future earnings. Since low-income agents accept
low-paid jobs which require low-skilled labor, these
agents would tend to gain less valuable experience at the
workplace. This would decrease the quality of their sort-
ing process and ultimately would lower their lifetime
earnings. On the other hand, high-income agents would
be employed in highly-skilled positions and work ex-
perience would improve their opportunities of finding
even better jobs. We do not elaborate further on theories
of learning, sortin g and matchin g but concentrate on how
the different opportunity costs of search for different
income groups lead to “unequal opportunity” for the
low-income searchers versus the high-income searchers.
The paper is organized as follows: part 1 is a discus-
sion of the rationale for the paper and literature review.
Part 2 represents a simple mathematical model of job
search accounting for the number of years spent at a
well-paid job. In Part 3 we modify this simple out-of-th e-
job search model introducing the total nu mber of years to
work. Part 4 gives a job market equilibrium model, in-
troducing initial income and the level of education as
determinants of job search time. In the last Part 5 of the
paper we turn to consumer theory and investigate the
relationship between optimal job search time and the
maximum utility of a worker who derives his satisfactio n
from two sources, consumption of goods and utilization
of free, non-work time. The paper ends with conclusions.
2. A Simple Out-of-the-Job Search Model
Our partial search model is focused on the behavior of
the unemployed individuals looking for better jobs and
not on firm behavior. We assume that a better job is one
that pays a wage rate beyond the minimal one o be-
cause in the labor market higher wage is associated with
higher labor productivity, i.e., the job is of greater value
both for the individual and for the firm. Another assump-
tion is that searching individuals are unemployed during
the search. This constraint is set by the fact that low-
skilled and low-income individuals have labor-intensive,
exploitative jobs, which leaves them minimal leisure
time to devote to job search. Ours is an out-of-the-job
search model that excludes the possibility of on-the-job
search. Furthermore, the worker decides on an optimal
search policy, given that he knows the distribution of
wages for his particular skills. We also assume constant
value of money, i.e., inflationary effects and time value
of money are ignored for the sake of simplicity. This
w

198 T. TODOROVA ET AL.
assumption would not change the major implications of
the mathematical results. The worker’s total gain from
the job search can be expressed as:

o
wtw y 


where the worker is expected to work number of
years at the well-paid job. The worker spends time
searching for a well-paid job. The wage he would
receive increases with search time, hence, . At
the same time, the job search process is less productive
with time and has diminishing returns, i.e.,
y
()wt t
0
() 0
()wt
wt




.
Without search the worker would receive a lowest pay to
the amount of minimum wage o but with search there
is a differential . Thus, the net benefit from
search to the unemployed individual will be of the form:
w
() o wt w
 
oo
twtwyw  

t
where from the total gain from search for a well-paid job
we subtract the cost of searching o. This is what the
worker sacrifices by not working for minimum wage
during the search period. Maximizing net gain by the
first-order condition,
wt
 
0
o
twtyw




o
wty w


Using the second-order condition,



0twty 
  
proves a maximum for the net gain from search. The
result of the optimization is that the worker will con tinue
his job search until the marginal gain of the search equa ls
his current wage o which represents his marginal cost
of search. To find the effect of minimum wage on opti-
mal search time we write the implicit fu nction and apply
the implicit-function rule:
w
  
,, 0
o
Ftwytwtyw 
  
o
By the implicit-func tion rule,

*10
o
w
ot
F
t
wFwty
 


There is a negative relationship between and o,
that is, the higher the initial minimum wage, the smaller
the search time would be. In the context of unsk illed and
skilled, qualified labor the higher initial payment is con-
nected with qualified labor. Thus, the result shows that
highly skilled workers are likely to find better jobs
sooner than those that hold less qualified jobs and re-
ceive low pay. Since firm demand for qualified labor is
generally higher, these are the two effects of the higher
demand–on the one hand, initial pay is higher and, on the
other, it takes less time for professional labor to search
for a job. This result adds up to the theory of wage dif-
ferentials by which a major factor that contributes to
wage differentials is the fact that the labor market con-
sists of heterogeneous workers and firms. On the supply
side of the market the move from low-paid jobs to
high-paid ones depends on various factors such as skill
differentials, differentials resulting from efficiency wage
payments, as well as non-wage job differentials [19]. We
find one more cause for wage differentials–the unequal
opportunities for job search faced by skilled and un-
skilled workers. Given the cost of information and job
search, some labor market segments are privileged in the
process of finding better jobs, all other things equal,
simply because they can afford the cost of search and
have the time for it. For the effect of the number of years
to work on the optimal search time we obtain:
*
t w
y*
t


*0
y
y
t
Fwt
t
yFwt






The greater the number of years to be at an ideal, bet-
ter-paid job, the longer one would search for this job.
3. A Modified Job Search Model
We then modify the model by specifying to be the
total number of years the worker is expected to work
throughout his lifetime. Hence, the number of years that
he is expected to work at his better-paid job is
y
yt

. We
again assume that the worker is not employed while
searching for an ideal job that pays and there are
positive but diminishing returns to job search, i.e.,
()wt
()wt 0

, and () 0wt



. Without search the worker
would receive a lowest pay to the amount of . The
net gain of search is: o
w



 
o
twtyt  wt
We find the optimal search time by maximizing the
obtained f unction:






0twtytw
 o
tw 








0
o
wty wtt wtw










*oo
wty wtww
ty
wt wt
wt

 


By the second-order condition the net gain of job
search is maximized:






 
20
twtytwtwt
wt ytwt



 

 

We obtain that the optimal time to search for a job
is a discretionary part of the total working-age years.
From analyzing this optimum we see that the faster the
*
t

T. TODOROVA ET AL.
199
wage grows with search time, the long er the search goes.
Thus when is infinitely large and the worker has
the chance to infinitely increase h is income by searching
more, he will continue searching throughout his entire
life, or . However, if search time has little effect
on the wage level, i.e., tends to zero, the worker’s
optimal search time would also tend to zero and the
worker would not be likely to search at all. To find the
effect of minimum wage on optimal search time we use
the implicit-function rule where the implic it function is:
()wt

y

*
t()wt




0F tytwtw
,,wy

o
tw

 o t
and by the implicit-function rule,
 
*10
2
o
w
ot
t
wtytwt
 
 

F
wF

*
t
This relationship again shows that the higher th e initial
minimum wage, the shorter the search would be. Thus,
again the segment of low-skilled workers will incur
higher job search costs, as they will need more time for
job search. Their opportunities for a better-paid job de-
crease. For jobs that have higher initial payment o
search time is smaller. The more slowly the ideal
wage grows with search, that is, the smaller
w
()wt

is,
the greater the effect of the initial wage is. For the effect
of the total number of years to work on optimal
search time we again obtain a positive value:
y

 
*0
2
y
t
F
yF
 wt
t
wtytwt

 
 

Our simple job search model shows that the greater the
total number of years to work, the longer the job search
time. Thus, younger individuals will have more incen-
tives to search than elderly people, as simple logic dic-
tates.
4. A Job Market Equilibrium Model
To study the effect of initial endowment and education
on wage rate we formulate a job market equilibrium
model that accounts both for demand and supply on the
labor market. We assume that the market is in equilib-
rium, i.e., worker supply of labor is equal to firm demand
for it. We also assume demand for labor to be exoge-
nously determined, since we study worker behavior on
the market. Hence, our analysis is focused on the supply
side of the labor market rather than the demand side and
firm behavior. Furthermore, we presume that the supply
of labor depends on the wage rate w, the time available
to workers t and their initial income , such that:
o
m

,, oo SwtmD
The supply of labor is positively related to the wage
level since a higher wage would stimulate people to sup-
ply more labor, hence, . Another assumption is
that the substitution effect prevails, i.e., the worker
would likely substitute his leisure with work when he is
stimulated by a higher wage. This assumption stems
from the fact that for poor, low-paid individuals the sub-
stitution effect prevails over the income effect. Being
highly paid and having substantial accumulated wealth,
professionals in prestigious jobs are less likely to substi-
tute their free time with work for more money. They are
more prone to the income effect. Being less paid and
with few savings, poor peop le are more expressly su bject
to the substitution effect. Thus, the greater the free time
available to the individu al, the greater his supply of labor
and the longer the hours he would be willing to work, i.e.,
. At the same time, supply of labor is negatively
related to the initial endowment or income o that a
worker has. Thus, a richer individual would be less likely
to work and would have more leisure time. As opposed
to him, a poorer person would be willing to supply more
of his labor, therefore, . Furthermore, the wage
rate depends on the minimum wage rate o, the educa-
tional level o
e, and the total free time available to
workers . The function for the wage rate will be of the
form:
0
w
S
0
m
S
0
t
Sm
w
t

,
oo
ww get
A higher educational level would affect the wage rate
positively, as more educated individuals will be em-
ployed in highly paid professional jobs, i.e., . If
the worker ha s a lot of leisure time then he is able to
search longer for a better paid job and wage rate is posi-
tively related to the free time which in this case becomes
job-search time and, therefore, . Thus, we can
formulate the model as:
0
e
g
,t
0
tg


,,00 00
oowtm
Swtm DSSS




,0 0
ooe t
ww getgg0


Assuming the derivatives to be continuous, we can use
the implicit-function theorem where the wage rate
and time t are endogenous. The other variables are as-
sumed to be exogen ous. Solving by the implicit-function
theorem, we analyze the effect of initial wealth or
non-labor income , as follows:
w
o
m
,
10
o wt m
t
o
w
m SS S
gt
m




 


 
 




where the Jacobian is nonzero:
0
wt t
JSgS



200 T. TODOROVA ET AL.
The effect of the initial endowment on wage rate will
be:
00
mt
tmt
ow
SS
gSg
w
mJ SgS




tt

The relationship is positive which means that the
higher the initial income of an individual, the greater his
prospects of obtaining a better paid job. This result
stresses the role of heterogeneous endowments on wage
differentials, which has widely been discussed in classi-
cal economic literature. The positive relationship between
initial endowments and wage rates can be analyzed one
step further if we consider that the initial wealth in the
form of accumulated capital can be readily available for
human capital investment in training and education.
Education and training deserve special attention here
because they are major determinants of personal produc-
tivity. Having in mind that the opportunity cost of human
capital investment depends on the initial income of indi-
viduals (usually in the form of family wealth), rich indi-
viduals can finance th eir schooling and training on better
terms and in general can invest more and boost future
earnings. On the other hand, poor individuals face higher
opportunity cost of funds even when they have the same
personal abilities and talents. This inequality of opportu-
nity among poor and rich individuals translates into cor-
responding wage differentials no matter that their initial
ability rents and productive capacities are equal. Our
equilibrium labor market model implies that rich people,
having an initial endowment or accumulated wealth,
have more leisure time and hence more time for search:
10 0
wm
m
owt
SS
S
t
mJ SgS


t


The higher the initial inco me, the greater the free time
is. Thus, rich individuals have more time to search for a
better job. To trace the effects of the educational level on
wage rate and search time we solve:
00
1
o wt wt t
te
o
w
e
SS JSgS
gg t
e








 





0
0
t
et te
owt
S
gg Sg
w
eJSgS



t

As expected, the relationship between educational
level and wage rate is positive, which confirms our ex-
pectation that a higher educational level promises a
higher wage. Thus, rich people with good prospects of
getting better education actually have the opportunity to
occupy well-paid jobs.
0
t
S
10
w
ewe
owt
gSg
t
eJ SgS


 

This derivation shows that more educated people need
less time to search for a better job and are less likely to
experience long periods of unemployment. At the same
time, the search time for the less educated and less quail-
fied workers increases and they are less likely to be em-
ployed at well-paid jobs. For the effect of minimum
wage we solve analogously:
00
11
o wt wtt
t
o
w SS JS
gS
gt
w



 

w

 
 
 




0t
t
S
10
tt
ow t
gS
w
wJ
SgS





As can be expected, the higher the initial wage, the
greater the prospects of the individual to find a better-
paid job. Poor people who start at a very low wage level
have less chance of getting a higher salary later. This
result has implications about the growing inequality of
opportunity as different groups of workers advance in
their careers. Put in a specific context, if one is employed
as a janitor, he would always be perceived as a janitor
and would rarely be able to jump up the professional
ladder into a better paid and more prestigious job. It is as
if the very perception of the job a person is taking and
the wage rate he accepts determine his future career and
income. By the same token, a secretary may be quite
knowledgeable ab out her boss’s business, be able to per-
form multiple, complex tasks, speak a number of lan-
guages and be computer literate. Yet, a secretary is often
perceived as “the girl who makes the coffee” or as “the
typist” at best and is likely to remain a low-paid staff
member. With regard to the time devoted to search we
find:
0
t
S
11 0
w
w
owt
S
t
wJ SgS


 

The effect of the minimum wage on the search time is
negative, i.e., the higher the initial wage, the less the time
dedicated to search. Thus, high-sk illed workers are likely

T. TODOROVA ET AL.
201
to find qualified job openings sooner than low-skilled
workers f ind correspond ing low-paid jobs. The effects of
the firms’ labor demand on the wage rate and search time
are analyzed as follows:
10
10
o wt wtt
t
o
D SS JS
gS
gt
D



 


 
 




w

Solving for the two effects,
1
t
S

00
t
tt
owt
gg
w
DJ SgS



The effect of demand for labor on the wage rate is
positive, which means that, as firms demand more labor,
the market wage is likely to increase. The effect of labor
demand on search time is:
1
t
S

10 10
w
owt
t
DJ SgS


As expected, the greater the firms’ demand for labor,
the lower the search time for workers and the shorter
spans of unemployment they would experience.
Low-skilled jobs are characterized by long hours of
tedious repetitive activities and are low-paid, as they do
not require professional skills. These jobs are occupied
by poor individuals who rely on subsistence level wages
to live on. On the effort side, these individuals are
time-constrained because they have to dedicate much
more of their time working. On the in come side, they are
also limited–since they earn lower incomes, their oppor-
tunity costs of search are higher. Since the opportunity
cost of unemployment is higher for these individuals, it
is legitimate to assume that they would exert more
work-efforts in order to remain employed. We consider
pay-for-performance wage rate and efficiency wage
theories, which incorporate the possibility that wage in-
creases may stimulate productivity and work effort. The
classical models of efficiency wage theories are widely
developed in the 1980s and 1990s and derive a series of
implications on shirking issues, labor turnover, non-
clearing labor markets and unemployment [20-23]. The
literature defines efficiency wage as the wage “which
minimizes an employer’s wage cost per effective unit of
labor service employed” [19]. We demonstrate this
wage-productivity dependency graphically in Figure 1
where at the optimum firms determine efficiency wage
based on work effort:
The work effort-wage curve is higher for the unskilled
labor, as drawn in Figure 1. Thus, at the optimum, firms
s
e
u
e
s
w
u
w
Work
Effort
Wage
Figure 1. Optimal wage for skilled and unskilled labor.
end up paying less skilled workers a lower wage u,
whereas the wage received by high-skilled workers is
w
s
w. At the same time the effort for the less skilled work-
ers is higher–they are supposed to work harder at the
workplace and shirking is prevented by supervisors.
Amidst the current world economic crisis, it is interesting
to analyze this work effort-wage relationship from the
perspective o f a n economic rece ss i on .
In a recession highly skilled workers are less likely to
be freed. Even if individuals need to switch occupations,
the cost of re-training will be covered by employers or
will be recovered fully by the potential future wage.
There would rarely be any change to their effort-wage
curve so their work effort would likely remain the same.
So would the optimal wage firms pay to them. The
low-skilled workers will be much more exposed to un-
employment since the supply of unskilled labor will re-
main relatively high in times of a crisis, yet, unemployed
workers would not be able to find jobs in more profess-
sional areas as skill training will be extremely costly and
time-consuming for them. Since the opportunity cost of
losing a job is higher for less skilled workers, th ey would
increase their amount of effort in order to keep their jobs.
This effect would result in an upward shift of the effort-
wage curve, as depict ed by Figure 2. Although low-skilled
workers would on average exert more effort in times of a
recession, their opti mum wage is likely to decrease. Thus,
poorer individuals will be required to work even more to
earn sustainable wages and their search time for better
jobs will be limited even more. Their cost of search
would also increase. Thus, even the small pay obtained
would be important for them and they would not opt to
remain unemployed in order to search for a better job.
5. Worker Utility, Leisure and Working
Time
Job search can as well be analyzed from the perspective

202 T. TODOROVA ET AL.


o
CLwLm C
u
e
u
w
s
e
u
e
s
w
u
w
Work
Effort
Wag e
Figure 2. Change of optimal wage in a recession.
of consumer theory where the worker is a consumer of
leisure and other commodities. We investigate the rela-
tionship between wage income and consumer behavior,
i.e., what the optimum labor time and non-labor time that
maximizes consumer utility is. Since we adopt an out-of-
the-job search model for low-income workers, it is le-
gitimate to say that the search time for a better job will
be part of the leisure or non-work time an individual has.
Here we use the terms “non-labor time,” “leisure,”
“non-work time” and “job search time” interchangeably.
Respectively, the “consumer” is a synonym for worker,
as the latter can choose between work and leisure. We
solve a standard Stone-Geary utility problem, where the
consumer derives his utility from a bundle of goods,
consumption and non-work time T, where
is total time available to the worker, and is his
working time in labor hours:
CLT
L
0


C
mLw

CLwLm C
TL* t


,,UCTL   
Consumption must equal the worker’s income less his
positive subsistence level of consumption o. His in-
come consists of non-labor income (accumulated wealth
or some initial endowment) and labor income
where w is the pay p er labor hour and L is the number of
labor hours worked. Thus, consumption can be expressed
as:
o
Job search time will depend on the non-work time a
worker has, i.e., there is a positive relationship between
non-work time and search time . To find the
optimal amount of labor and non-labor hours that maxi-
mize utility to the worker and the effect of wage rate on
the optimal labor hours, we solve the optimization prob-
lem:
Maximize 

o
UCC TL


where the worker’s consumption must equal his income,
or:





ULwL m CTL

(1)
Substituting exp ression (1) in the utility fun ction gives
an unconstrained objective fu nct i on:

Maximize o

 
1
UwwLmCTL



 
Differentiating with respect to labor and setting this
derivative equal to zero,
 
10
Lo
o
wL m CTL




 
 
1
o
wwL m CTL



 
 
1
o
wL m CTL




 

 
o
wTLwLm C  

o
wTwLwLm C   
Thus, the optimum number of labor hours to the
worker can be e xp re ss ed as:


*o
Lw


wTm C  



*o
mC
T


Lw




m
C


(2)
When non-labor income is exactly equal to basic
consumption o, labor is a share of total worker time
measured by the coefficient



. There is a nega-
tive relationship between the amount of labor hours
worked and the amount of non-labor income, i.e., the
higher the non-labor income a worker has, the lower the
amount of labor he will perform to attain his optimal
level of consumption. Adversely, labor is positively af-
fected by a larger amount of subsistence consumption.
Therefore, the worker has to work more hours to meet
his basic needs. Alternatively, the amount of non-work
time that maximizes worker utility can be expressed as:





0 *
0
Lw
mC wT
Tw





 
wTm CwTwT   



0
*mC wT
 
TL
 
This expression shows a positive relationship between
non-labor (search) time and wage rate as well as between
search time and non-labor income. Hence, highly paid
workers face lower opportunity cost of searching for a
new job. In addition, individuals that have more initial

T. TODOROVA ET AL.
203
accumulated wealth will tend to enjoy more free time
and will have more time to search for a new better job.
To see the exact effect of wage rate on optimal amount
of labor hours we differentiate:


2
o
mC
L
ww 


 
mC
mC
C

pCp TLwL m
pwL

The derivative shows a positive effect of wage on the
amount of labor when non-labor income exceeds basic
consumption. In other words, when o, the worker
has some excess income so he needs to work less and
would be willing to substitute labor with leisure. How-
ever, if subsistence consumption exceeds non-labor in-
come, i.e., o, the income effect prevails and the
worker needs to work more to reach his subsistence level.
The sign of this derivative is indicative of the shape of
the labor supply curve. When non-labor income exceeds
minimum consumption the derivative is positive, which
determines a positively sloped labor sup ply curv e, that is,
labor increases with the wage level. This would be the
case with a strong substitution effect when the consumer
(worker) substitutes his free time with labor due to the
wage incentive. Adversely, the supply curve is nega-
tively sloped, if o m or the increase in the wage
reduces the amount of labor. This would be the effect
with a backward-bending labor supply curve when the
income effect prevails over the substitution effect. The
Stone-Geary utility optimization problem can have fur-
ther implications if we introduce a budget constraint for
the worker such th at:
CT
p
where C is the price of consumption, T is the price
of non-labor time, is wage income and m is
nonwage income. The worker will have a logarithmic
utility function of the fo rm:

 
ln1 lnUCCTL   
CT L
1
o
C
where is the worker’s consumption, o is a mini-
mum level of consumption, is total time and is
his working time with 0

 L
  
ln1 lnUCCTL  
. Thus, is the
worker’s non-labor time, which he uses to search for a
better job. To find the optimum number of labor hours,
we solve the following constrained optimization prob-
lem:
T
Maximize
o
Subject to


CT
pCp TLwL m
Introducing a Lagrange multiplier,


 

ln1 ln
o
CT
Z
CC TL
wLmp CpTL



 

and setting the three first-order derivatives equal to zero,

0
CT
ZwLmpCpTL


 
0Zp

CC
o
CC


 (3)
1
L
Zw
TL
 0
T
p



(4)
From the last two equations (3) and (4) we derive:


1
CoT
pCCp wTL
 






 
1
TC
pwTL pC  

o
C   (5)
From the budget constraint we have:

CT
pCwL mp TL


Substituting back in (5),




pwTL
  
1
T
TCo
wL mpTLp C

   







  

1
1
TC
TT
pTLwTLmpC
wL pT LpT L
 

 o

 



 
 
1
1
Co
T
wT Lm pC
wLp TL


 



 
1Co
TT
wTwLmp C
wLwLp TpL
 







1Co T wTmpCwLp Tp  
T
L



 
1
TT
pwp Tmp   

Co
LwC

 


 
 
1
1
TCo
Co
T
Co
T
wwwpT mpC
L
Twp
wTmp C
Twp


  

 

 
11
T
wp
wT mpC 

  




*1Co
T
wTmp C
LT wp
 
 
As expected, the optimal number of labor hours de-
pends negatively on the wage rate and non-labor income
and positively on the price of free time, which is propo r-
tionate to the opportunity cost of working. Thus, poor
agents, whose labor income should satisfy the subsis-
tence level of consumption, will face higher price of free

T. TODOROVA ET AL.
204
 
time and would work more labor hours to attain their
optimum utility level of consumption. We then analyze
the effects that wage rate, non-labor income and the price
of free time have on the amount of free time, as well as
on job search time:
1Co
T
wTmp C
TL wp
 
 
wp
Free time depends positively on income from labor
and non-labor sources, once basic consumption o C is
satisfied. The opportunity cost of leisure is the total
amount of wage lost and the price to be paid for leisure,
or T. Since the individual searches for a job while
not working, search time will be a fraction of non-work
time.
6. Conclusions
Some jobs are more dangerous and onerous than others,
some offer more earning stability than others, some offer
more perks in addition to the wage rate. All these differ-
ences result in inequalities and wage differentials. Our
results show one more cause for wage differentials–the
unequal opportunities for better j obs that low-skilled and
high-skilled workers face. Taking into account the costs
of information and job search, we find that some seg-
ments of the labor market are privileged in the job search
process, ceteris paribus, simply because they can afford
the time to search for these opportunities. We conclude
that initial endowments cause d ifferences in the opportu-
nity cost of job search and result in wage differentials-
the higher the initial income of an individual, th e greater
his prospects to obtain a better paid job. There is a nega-
tive relationship between optimal search time for a b etter
job and the potential minimum wage on that job, i.e., the
higher the initial minimum wage, the smaller the search
time would be.
Wealthy individuals, having higher initial endowment
or accumulated wealth, have more leisure time and hence
more time for search which increases the value-added of
their optimization problem. Wealthy individuals face
lower opportunity cost of investment in education and
professional training. On average, wealthy individuals
are more educated and thus qualified for higher-paid jobs.
More educated people need less time to search for a bet-
ter job and are less likely to experience long periods of
unemployment. Low-income individuals who start at
low-paid jobs have smaller chance of getting high-paid
jobs later. This result h as implications about the growing
inequality of opportunity as different groups of workers
advance in their careers. At the time of a recession
low-income individuals are more prone to unemploy-
ment and their search costs, work effort and income loss
will increase relative to those of wealthy individuals.
7. References
[1] B. Ehrenreich, “Nickel and Dimed: On (Not) Getting by
in America,” Metropolitan Books, May 2001.
[2] G. Wallraff, “Ganz Unten,” Kiepenheuer & Witsch; Au-
flage: Auflage unbekannt, 1985.
[3] G. Smith, M. Venkatraman and R. R. Dholakia,
“Diagnosing the Search Cost Effect: Waiting Time and
the Moderating Impact of Prior Category Knowledge,”
Journal of Economic Psychology, Vol. 20, No. 3, June
1999, pp. 285-314.
[4] G. J. Stigler, “The Economics of Information,” Journal of
Political Economy, Vol. 69, No. 3, June 1961, pp 213-
225.
[5] G. J. Stigler, “Information in the Labor Market,” Journal
of Political Economy, Vol. 70, No. 5, October 1962, pp.
94-104.
[6] A. A. Alchian and W. R. Allen, “University Economics,”
Wadsworth Publishing Company, Belmont, California,
1964.
[7] J. J. McCall, “Economics of Information and Job
Search,” The Quarterly Journal of Economics, Vol. 84,
No. 1, February 1970, pp. 113-126.
[8] D. T. Mortensen, “Unemployment Insurance and Job
Search Decisions,” Industrial and Labor Relations Re-
view, Vol. 30, No. 4, July 1977, pp. 505-517.
[9] K. Burdett, “A Theory of Employee Job Search and Quit
Rates,” The American Economic Review, Vol. 68, No. 1,
March 1978, pp. 212-220.
[10] C. A. Pissarides, “Search Unemployment with on-the-Job
Search,” The Review of Economic Studies, Vol. 61, No. 3,
1994, pp. 457-475.
[11] B. Jovanovic, “Matching, Turnover, and Unemploy-
ment,” Journal of Political Economy, Vol. 92, No. 1,
1979, pp. 108-122.
[12] G. J. Van den Berg, “A Structural Dynamic Analysis of
Job Turnover and the Costs Associated with Moving to
another Job,” The Economic Journal, Vol. 102, No. 414,
September 1992, pp. 1116-1133.
[13] D. T. Mortensen, “Job Search and Labor Market Analy-
sis,” In: O. C. Ashenfelter and R. Layard, Eds., Hand-
book of Labor Economics, Vol. 2, 1986, pp. 849-919.
[14] L. M. Kahn, “The Returns to Job Search: A Test of Two
Models,” The Review of Economics and Statistics, Vol.
60, No. 4, November 1978, pp. 496-503.
[15] B. J. Yoon, “A Model of Unemployment Duration with
Variable Search Intensity,” The Review of Economics and
Statistics, Vol. 63, No. 4, November 1981, pp. 599-609.
[16] G. J. Van Den Berg, “Nonstationarity in job search the-
ory,” The Review of Economic Studies, Vol. 57, No. 2,
April 1990, pp. 255-277.
[17] G. S. Becker, “Human Capital and the Personal Distribu-
tion of Income,” W. S. Woytinsky Lecture, No. 1, Uni-
versity of Michigan, Ann Arbor, 1967.

T. TODOROVA ET AL.
205
[18] H. S. Faber and R. Gibbons, “Learning and Wage Dy-
namics,” Quarterly Journal of Economics, Vol. 111, No.
4, November 1996, pp. 1007-1048.
[19] C. R. McConnell, “Contemporary Labor Economics,”
McGraw-Hill, New York, January 1995, pp. 225-235.
[20] G. A. Akerlof and J. L. Yellen, “Efficiency Wage Models
of the Labor Market,” Cambridge University Press,
Cambridge, 1986.
[21] L. F. Katz, “Efficiency Wage Theories: A Partial Evalua-
tion,” NBER Macroeconomics Annual 1986, Cambridge,
The MIT Press, Massachusetts, 1986, pp. 235-276.
[22] J. Stiglitz, “The Causes and Consequences of the De-
pendency of Quality on Price,” Journal of Economic Lit-
erature, March 1987, Vol. 25, No. 1, pp. 1-48.
[23] A. Weiss, “Efficie ncy Wages: Models of Unemployment,
Layoffs, and Wage Dispersion,” Princeton University
Press, Princeton, New Jersey, 1991.