EARNING MORE AND RECEIVING LESS:
THE ECONOMIC DISINCENTIVES FOR COLORADO FAMILIES
MOVING OFF AFDC
Jane Cecile Venohr
B.S., William James College, 1982
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirement for the degree of
Master of Arts
Department of Economics
This thesis for the Master of Arts degree by
Jane Cecile Venohr
has been approved for the
Venohr, Jane Cecile (M.A., Economics)
Earning More and Receiving Less: The Economic Disincentives for Colorado
Families Moving off AFDC
Thesis directed by Professor Suzanne W. Helbum
Despite the changes provided by the 1988 Family Support Act,
Colorado single parents receiving public assistance will still encounter
economic disincentives to increase wage income. These disincentives are
caused by the interaction of the implicit tax rates in the public assistance
system and income tax system. An increase in wage earnings, due to a pay
raise or increase in hours worked by the single parent, results in a decrease
in spendable income and benefits for typical Colorado AFDC families moving
toward self-sufficiency because the dollar value decrease in public assistance
is more than their increased earned income after taxes.
High implicit marginal tax rates would occur with any Colorado
single parent family moving off AFDC and toward self-sufficiency if they
receive a multitude of public assistance programs and have work-related child
care costs. More severe child care needs lead to stronger economic
disincentives to increase wage earnings. Families living in regions of
Colorado with higher living costs face greater economic disincentives
especially if they receive a housing subsidy. If health insurance is considered
a necessity for the family, and they cannot receive subsidized health
insurance, there is no incentive for this family to become long-run indepen-
dent of AFDC, nor to work full-time.
Overall, high implicit marginal tax rates create economic incentives
for poor, single parent families to remain public assistance dependent. This
is contrary to the goal of self-sufficiency pursued in the 1988 Family Support
Act. Alleviation of these economic incentives cannot be achieved through
targeted programs aimed at curing problems with child care or medical
insurance. A comprehensive solution that takes the entire multitude of public
assistance programs in to account is necessary.
The form and content of this abstract are approved,
I recommend its
This thesis is dedicated to the memory of
Dr. Marcile N. Wood.
This research is indebted to several people. John Morris provided
the original idea of researching the implicit marginal tax rate for single parent
families moving off poverty and toward self-sufficiency. The original idea was
expanded, nurtured, and once it worked its way into paper was endlessly
edited by Suzanne Helbum. This project could not have been completed
without her immense support and boundless energy. I would also like to
thank Jim Smith for always believing this was a good project and strongly
encouraging me to continue in this area of research.
My fellow graduate students, Gaylene Martinez and Daurie
Augostine, helped me put this research in a thesis form. Anita Venohr kept
me company late nights in an otherwise vacant office, and PSI allowed me
access to their computers.
Final thanks goes to Gregg Swayze who financially and morally
supported this thesis.
I. INTRODUCTION .................................... 1
Purpose of Study................................ 3
Arrangement of Thesis .......................... 4
Notes .......................................... 6
n. PUBLIC ASSISTANCE: REFORM AND WORK
INCENTIVES ........................................ 7
The Need to Reform AFDC ........................ 7
Welfare Reform through Work ................... 10
The 1988 Family Support Act ................... 13
A Multitude of Public Assistance............... 18
Previous Research on Implicit Marginal Taxes .. 21
Chapter Summary................................ 27
Notes ........................................ 29
III. METHODOLOGY: MEASURING ECONOMIC
DISINCENTIVES . ................................ 30
Defining Implicit Marginal Tax Rate ........... 32
The General Model .......................... 32
Independent Variables ...................... 33
The Prototype Family and Variations ........... 38
The Prototype Family......................... 39
Set One: Varying Child Support Payments......... 41
Set Two: Varying Child Care Needs ........... 41
Set Three: Two Children........................... 41
Set Four: Three Children ....................... 42
Set Five: Geographical Differences ............... 42
Set Six: Medicaid Cut-off ........................ 44
Set 7a and 7b: Increasing Hours Worked ......... 45
Set Eight: Months Employed........................ 46
Chapter Summary...................................... 47
Notes ............................................... 49
IV. RESULTS: WHERE ECONOMIC DISINCENTIVES
Graphing the Results ................................ 52
Economic Incentives to Increase Wages ............. 53
Set One: Child Support Payments ................ 53
Set Two: Varying Child Care Needs .............. 57
Set Three: Two Children........................... 57
Set Four: Three Children ....................... 58
Set Five: Geographical Differences ............... 58
Disincentives Associated with Medicaid Cut-off....... 62
Set Sue: Medicaid Cut-off ........................ 62
Economic Incentives to Increase Hours Worked ........ 64
Set Seven: Increasing Hours Worked................ 64
Set Eight: Increasing Months Worked .......... 66
Summary of Results............................... 67
Notes ........................................... 71
V. RESTORING WORK INCENTIVES AND ADDITIONAL
CONSIDERATIONS ..................................... 72
Restoring Economic Incentives.................... 73
The Negative Income Tax....................... 73
Piecemeal Solutions .......................... 78
Changes in Future Economic Incentives............ 82
Child Support................................. 83
Inflation and Increases in the Price Level.... 85
Success Depends on Earnings ................. 86
Child Care Costs ............................ 89
Proposed Legislation.......................... 92
Chapter Summary................................. 95
Notes ........................................... 97
VI. CONCLUSION . .................................... 99
Chapter Summary................................. 105
A: 1988 Public Assistance Grant and Tax Determination . 119
B: Calculating Child Care Costs.............................. 124
C: Calculating Fair Market Rent for Colorado Families . 126
D: Marginal Implicit Tax Tables .............................128
2.1 Benefits Potentially Available to a Four-Person,
Female-Headed Family in Chicago, 1971 ................. 23
2.2 The Effect of Transfer Benefits and Taxes on the
Incentive of a Pennsylvania Mother with Two
Children to Earn Income................................ 25
2.1 Poverty Families Headed by Females.......................... 9
4.1 Set 1: Increasing Child Support ........................... 54
4.2 Set 2: Varying Child Care Needs ........................... 55
4.3 Set 3: Two Children ....................................... 56
4.4 Set 4: Three Children ..................................... 59
4.5 Set 5: Geographic Location .............................. 60
4.6 Set 6: Medicaid Cut-Off.................................... 63
4.7 Set 7a: Increasing Hours Worked (Changes in the Wage
4.8 Set 7b: Increasing Hours Worked (Changes in Child Care
Costs) .................................................... 68
4.9 Set 8: Increasing Months Worked.......................... 69
The 1988 Family Support Act replaces the Aid to Families with
Dependent Children program (AFDC) with a mandatory job training and work
program. The intention of this legislation is to make AFDCs families self-
sufficient by encouraging single parents to bring home a paycheck. However,
this legislation does not eliminate the work disincentives in the public
assistance system available to Colorado single parents and their families.
These parents encounter economic disincentives to increase wages, economic
incentives to work only part-time, economic incentives to remain Medicaid
eligible thus AFDC eligible1, and economic incentives which encourage a cycle
of working a few months, returning to AFDC, and through the cycle again.
This study defines these economic disincentives and explains how these
disincentives contradict the goals of the 1988 Family Support Act.
These economic disincentives are caused by high implicit marginal tax
rates which exceed 100 percent. There are several cases where a familys
spendable income would be less if the parent would have accepted a pay raise,
or would have increased her hours worked. For example, consider the
situation of a Colorado single parent with one child needing full-time child
care. The family receives a Title XX child care subsidy, low-income energy
assistance, food stamps, child care tax credit, and earned income tax credit
when applicable. If the parent works full-time at $4.65 per hour2, the monthly
spendable income and benefits would be $822 per month. However, if the
parent receives a pay raise and is now earning $5.00 per hour the familys
spendable income decreases to $779 per month. The implicit marginal tax
rate on this pay raise would be 170 percent because 170 percent of the pay
raise disappeared into implicit taxes; that is increased federal, state, and social
security taxes, and decreased food stamps, decreased child care subsidy and
decreased low-income energy assistance.
This occurs because welfare recipients face another tax system in
addition to what all working denizens face. Not only is their earned income
subject to federal, state, and social security taxes, but their earned income is
also subject to the marginal tax rate implicit in public assistance programs.
Each public assistance program has its own implicit marginal tax rate. For
instance, for every dollar earned by an AFDC parent3 their AFDC grant is
reduced by one dollar; this can also be noted as a 100 percent tax rate. The
tax rate for food stamps is 80 percent; for every dollar earned food stamps
are reduced 80 cents. The tax rate on housing assistance is 67 percent; and,
there are other tax rates for other types of public assistance. Since most
AFDC4 families receive more than one type of public assistance, parents
attempting to move their families off public assistance by going to work and
increasing their wage earnings will not always find incentives to do so.
Impoverished single parent families with more children, greater child care
needs, families living in more expensive regions of Colorado, families receiving
larger child support awards, and families with a greater need for health
insurance face stronger disincentives to become self-sufficient. Despite the
1988 Family Support Acts goal of self-sufficiency the Act does not correct
these disincentives, nor does any currently proposed legislation. To restore
consistent work incentives in the public welfare system massive reform may be
Purpose of Study
This thesis demonstrates that economic disincentives faced by
impoverished Colorado single parents to increase income and increase hours
worked will exist under the 1988 Family Support Act. These disincentives are
caused by high implicit marginal tax rates, and these disincentives are contrary
to the 1988 Family Support Acts goal which is to replace AFDC with a self-
sufficiency program. The Family Support Act only changes the structure of
the AFDC program and not the other public assistance programs and taxes
available to single parent families such as food stamps and earned income tax
credits. Since most AFDC families receive more than one form of public
assistance, it is necessary to measure the implicit marginal tax rate for the
most typical Colorado AFDC families. Although this study is most concerned
with whether a multitude of public assistance programs always encourages low-
income single parents to accept pay raises, it also examines how child support
payments, family size, varying child care needs, geographical location, medical
insurance needs, and hours worked affect incentives.
This study does not end with defining economic disincentives which
interfere with the self-sufficiency goals of the Family Support Act. Realizing
that the economic disincentives for Colorado AFDC parents change quickly
with inflation and proposed legislation affecting child care and medical
insurance and other factors, this study discusses what will happen to implicit
tax rates due to these changes. Such a discussion is necessary since it provides
evidence that piecemeal solutions are inadequate and total reform of the
entire public assistance system may be necessary.
Arrangement of Thesis
Chapter Two provides background information. It explains why AFDC
is in desperate need of reform and why work programs are believed to be the
solution. An historical account of previous attempts to reform AFDC with
work is given. Also, in the second chapter, the 1988 Family Support Act is
described and a literature review on previous studies relevant to implicit
marginal tax models is provided.
To measure the implicit marginal tax rates a microsimulation of
Colorado public assistance programs is developed in Chapter Three. There
are two major components of this microsimulation. First, there is the model
itself, and secondly there is the development of typical Colorado AFDC
families. The model calculates all of the public assistance available to a
family at a particular wage, and what their total taxes will be. From the
model, spendable income and implicit marginal tax rates can be calculated.
These results are presented in Chapter Four.
The second major component of Chapter Three is necessary since
family characteristics vary from family to family: family size varies, age of
children varies, child care needs vary, where a family lives varies, medical
needs vary, average child support receipts vary, and the types of public
assistance a family receives varies from household to household. In Chapter
Three, a prototype family which defines the most typical Colorado AFDC
family will be developed. The spendable income and marginal implicit tax
rates will be calculated for the prototype family using this model. Family
characteristics will then be varied, and spendable income and marginal implicit
tax rates will be calculated for these variations in Chapter Four.
In summary, Chapter Three builds the methodology which includes the
model, the prototype family, and the variations of the prototype family; and,
Chapter Four calculates the results.
Chapter Five discusses the possibilities of eliminating work
disincentives revealed in Chapter Four. Proposed legislation, inflation, and
other factors could change the implicit marginal tax rate. For instance, the
federal government is proposing legislation that could greatly affect child care,
hence change the model used in this study and the implicit marginal tax rate.
Finally, Chapter Six provides a summary and a conclusion.
1. Medicaid and AFDC have the same eligibility requirements.
2. This is both a wage rate near the newly proposed minimum
wage and a wage rate near what job programs typically place their
3. This assumes the family has been on AFDC for a year.
4. See Marine (1988b).
REFORM AND WORK INCENTIVES
This chapter provides background information necessary to understand
how high implicit marginal tax rates interfere with the 1988 Family Support
Act. As a prelude to a thorough discussion of the 1988 Family Support Act
this chapter begins with an explanation of why AFDC reform is necessary and
a historical account of previous attempts to reform AFDC with work.
Once the Family Support Act is defined, other forms of public
assistance that are not affected by the Family Support Act can be defined.
Recall, it is the multitude of public assistance system programs that create
high implicit marginal taxes. Finally, a review of previous research similar to
this project is given.
The Need to Reform AFDC
For over a decade now, anti-poverty programs have been perceived as
a heavy burden on taxpayers dollars. Furthermore, these programs are
criticized for creating welfare dependency, and discouraging employment that
could lead to self-sufficiency. The bulk of the criticism focuses on the only
direct cash transfer program available to single-parent families, AFDC.
AFDCs budget is about $13 billion dollars a year (Elwood and Summers
1986), and according to the Public Welfare Association there are over three
million AFDC recipients nationally. A recent Colorado survey estimates that
approximately 30,000 AFDC families live in the state of Colorado.
The AFDC program did not start out large. First established in 1938
as a short-term relief measure for widows with children and orphans, the
AFDC program1 assisted 372,000 families through a budget of $133 million in
1940 (Abramovitz 1988). At that time the majority of AFDC heads of
households were women and they were widows, the type of families the AFDC
program hoped to assist. Today, the majority of AFDC head of households
are still women; yet, almost half of them have never been married. A 1987
Senate Subcommittee estimates the median length of time on AFDC is 26
months. Most importantly, the numbers of poor female-headed households
are increasing rapidly (see figure 2.1). "By the year 2000, The National
Advisory Council on Economic Opportunity estimated that female headed
families will make up 100 percent [sic] of Americas poor" (Dunn 1984, 89).
Proposals to remedy the AFDC problem have been innovative and
silly. Just 20 years ago, mandatory birth control for all AFDC households was
seriously considered. Now the suggestion seems ridiculous, even without
considering the personal freedom violations, given that todays2 average AFDC
family has 1.9 children (U.S. Senate subcommittee 1987). Innovative solutions
such as the Negative Income Tax program (NIT) would have replaced current
Poverty Families that are Female-Headed
POVERTY FAMILIES HEADED BY FEMALES
Source: Economic Report of the President. January. 1987.
U.S. Government Printing Office, Washington, D.C.
public assistance programs with a guaranteed minimum income close to the
federally established poverty line, yet an economic incentive to work. NIT
experiments conducted in the late 1960s and early 1970s decreased work
effort, rather than increasing it. Regardless of NITs failures, the key element,
work, continued to be considered the ideal solution to the AFDC problem.
Now according to the mass media AFDC has experienced its most
important reform since its origin (Rocky Mountain News 1988). Under the
1988 Family Support Act AFDC will be transformed into a work and training
program mandating AFDC parents participation. This study focuses on the
effects of the Family Support Act, aiid its economic incentives to increase
work efforts. As an introduction to the study it is important to discuss the
work ethic ideology on which the new policy is based.
Welfare Reform through Work
Today, in the United States, there exists a fundamental ideology that
able-bodied individuals should provide for their own livelihood. This ideology
holds doubly true for AFDC parents. When Governor of California, former
President Ronald Reagan stated, "We are not going to perpetuate poverty by
substituting a permanent dole for a paycheck" (quoted in Abramovitz 1988,
This ideology has guided AFDC policy. Initially, AFDC funds were
for the purpose of encouraging the care of dependent children in their
own homes or in the homes of relatives by enabling each state to
furnish financial assistance and rehabilitation and other services as far
as practicable under the conditions in each state to needy children and
their family with whom they are living to help maintain and strengthen
family life and to help parents or relatives to attain or retain capability
for self-support and personal independence consistent with the
maintenance of continuing parental care and protection (Title IV of
the Social Security Act, 1938, quoted in Dunn 1984, 88).
At the time of AFDCs initiation, working mothers were not common
to the American family norm; mothers were expected to stay home to care for
their children. That attitude has changed. Today, according to a 1987 Senate
Subcommittee Hearing on welfare reform, almost two thirds of all mothers
work. This senate hearing also indicated that only five percent of AFDC
mothers work; but, 22 percent are either actively seeking work, in school, or
obtaining job training. The ubiquitous attitude of today is that welfare
mothers should work just as much as non-welfare mothers do.
The first attempt toward mandatory work for AFDC parents was the
Work Incentive (WIN) Program established in 1967. It required all
employable AFDC parents to register for work, engage in job training
programs, and take any suitable job offered. However, due to severe under-
funding and federal administrative problems the WIN initiative achieved little
in the 1970s; rather, the welfare rolls escalated.
WIN was resurrected in 1981 by the Omnibus Reconciliation Act
(OBRA). It, also, added WIN demonstration projects that would be
administered solely by state or local welfare agencies. Concurrently, four
other programs were developed. Community Work Experience Program
(CWEP) allows states to require mandatory, unpaid, community workfare
programs as a job training measure for AFDC parents. The work
supplementation program allows AFDC funds to be transferred from direct
cash assistance to jobs creation for AFDC parents. Second, the work
supplementation program is on a voluntary basis. Third, the Job Training
Partnership Act (JTPA) provides job search skills and short-term training to
prepare welfare recipients for immediate, private-sector employment. Finally,
there is a job search program which provides computerized job-opening
announcements at welfare offices throughout a state.
Only the WIN and JPTA program are mandatory; WIN demonstration
projects, CWEP, work supplementation programs, and job search programs
are up to the states discretion. The State of Colorado has a WIN program,
JPTA program, and job search program for the entire state. In addition,
select Colorado communities have instigated WIN demonstration projects and
On a more national level, many states and communities that initiated
a WIN demonstration project claimed it a success. The most famous being
the Massachusetts Employment Training Program (ET) whose participation
was voluntary, rather than mandatory. ET evolved into the most
comprehensive job training and placement program for AFDC parents ever.
ET emphasized real jobs and decent wages; it paid special attention to the
transitory phase between welfare and secured employment by guaranteeing
child care for one year after leaving the welfare rolls. Many praise ET for
reducing Massachusettss AFDC roles by 9.5 percent in its first two and an
half years of operation (Wiseman 1988).
Others criticize work training programs in general for placing AFDC
parents in jobs not paying self-sufficiency wages (Keen 1989). E.T. graduates
averaged over $6.00 an hour. Additional criticism from Whitman (1987)
suggests that these work training programs only train and place AFDC parents
who would have found work despite the program. Regardless, what has
become self-evident through the voluntary recruitment nature of ETs
participants and other programs is that poor families have a real desire for
jobs, and prefer work over welfare3.
Massachusetts welfare reform measure along with Californias Greater
Avenue for Independence (GAIN) have become the role models for the
country. As a result, a national welfare reform measure was sought and
became a reality in 1988.
The 1988 Family Support Act
The $3.6 billion dollar Family Support Program added one more
clever acronym4 to the slew of existing job training programs agencies, JOBS
which stands for the Job Opportunities and Basic Skills Training program.
JOBS consolidates previous work programs such as WIN, CWEP, work
supplementation, JTPA, and job search programs into one job training
program. It also guarantees child care, Medicaid, and work-related expenses
to enable AFDC parents participation in the JOBS program. The purpose
of the JOBS program is "to assure that needy families with children obtain
the education, training, and employment that will help them avoid long-term
welfare dependence" (U.S. House of Representative Report 1988,110). Each
state is responsible for coordinating its current job training programs into one
JOBS program and developing its own state plan to encourage self-sufficiency
and discourage welfare dependency.
Certain specifications are strictly regulated through the federal
government. For example, in order to avoid reduced federal funding states
must target the JOBS program at specified AFDC groups. Parents under the
age 24 and who have not completed high school, parents with little to no work
experience, parents whose youngest child is within two years of becoming
AFDC ineligible due to age, and long-term AFDC recipients will be heavily
targeted for job training programs. Thus, skimming the cream off the top by
placing the easiest or the most employable AFDC parents in jobs and training,
will be curtailed.
The state of Colorados plan will rely on a case management system.
That is, each non-exempt AFDC case will be required to follow an
individualized plan toward self-sufficiency co-developed with their county-
level Department of Social Services case worker. It will be based on the
recipients needs, circumstances, employability, and the local labor market.
This individualized plan may include employment training, preparatory classes
toward obtaining a high school diploma equivalency (G.E.D.), job search
training, and/or job placement. Employment training can take the form of
on-the-job training, work supplementation, and community work experience.
AFDC parents exempt from participating in the JOBS program
include parents who cannot participate due to second trimester pregnancy,
illness or a physical disability, parents who care for another household
member who is ill or physically disable, parents under twenty who are enrolled
in an educational program and are making successful progress toward a high
school degree, parents who already work at least 30 hours a week, and parents
who have a child under the age one. Parents who have children under the
age of six are only required to participate in the JOBS program 20 hours a
week, and only if child care is guaranteed.
Parents who are required to participate in the JOBS program and fail
to comply by not participating in the education and training programs, or by
refusing a directive to accept employment, may have economic sanctions taken
against them. No sanction will take in effect without a fair hearing. A parent
does not have to abide by a directive to accept employment if it results in a
net loss of cash income; furthermore, parents of children under age six do not
have to accept a directive requiring them to work more than twenty hours per
week. Parents who are required to participate in a community work
experience need not comply if they are gainfully employed at least twenty
hours per week, or if they have children under age six. There are several
other exemptions, but these are irrelevant to this study.
The Colorado JOBS program shall be in place by October 1, 1990,
and operating in all subdivisions of Colorado where it is feasible by October
1, 1992. All Colorado AFDC parents must participate in the JOBS program
beginning October 1,1992. Prior to this date, the state of Colorado will most
likely target AFDC parents specified as high priority by the federal guidelines.
For fiscal year5 1990 and 1991 Colorado must place an average of seven
percent of their monthly non-exempt cases in the JOBS program. For fiscal
years 1992 and 1993 this increases to eleven percent; the participation rate
must be 15 percent for 1994, and 20 percent in 1995. Any Colorado county
which has developed a plan complying with the JOBS program may begin
their JOBS program January 1, 1990.
Beginning April 1, 1990, Colorado will provide transitional benefits
for AFDC families moving from public assistance to self-sufficiency.
Subsequently, these transitional benefits will be available to JOBS participants.
Medicaid and child care will be guaranteed during job training and education
programs. When the AFDC parents achieve employment and are no longer
eligible for AFDC they will be guaranteed child care for the first year.
However, they will be required to contribute to child care costs based on an
income-determined, sliding fee scale. Medicaid will also be guaranteed for the
first year. Whether or not there will be a premium charged for Medicaid
after the first six months has still not been determined. All of these
transitional benefits will be denied if a familys average gross monthly income
minus child care costs exceeds 185 percent of the federal poverty level for
The JOBS program is only one half of the 1988 Family Support
Program. The other half concentrates on increasing child support
enforcement. The goal is to make absent parents more economically
responsible for their childs well-being. Currently, less than 50 percent of
those on AFDC receive child support payments.
Overall, the Family Support Act changes only AFDC and no other
public assistance program. It makes minor changes in AFDC grant
determination by slightly increasing the child care and work-related expense
deductions. It makes major changes by forcing AFDC parents to participate
in the JOBS program. Essentially, it will put thousands of Colorado AFDC
parents into the labor market. It intends to put all non-exempt, AFDC
parents to work, but it has set its goal more realistically by aiming for a
smaller percentage of non-exempt, AFDC parents. Therefore, all non-
exempt, AFDC parents may not be required to participate in the JOBS
program. Who, exactly, will participate in the JOBS program is not entirely
clear. Certainly, federally targeted AFDC subgroups6 will be prime
candidates. Volunteers will also most likely be prime candidates.
The transitional benefits provided by the Family Support Act for the
first year ease the costs of child care, and the loss of Medicaid an AFDC
family incurs when becoming AFDC ineligible due to increased earnings. But,
what happens to AFDC families after a year when there is no medicaid and
no guaranteed child care? This question will be addressed in this study.
A Multitude of Public Assistance
Essentially, the Family Support Act is mandating that AFDC parents
go to work. Forcing these parents into the labor market could cause several
problems. This study analyzes whether a parents increased wage earnings will
always make a family financially better off. As single, AFDC parents7 enter
the workforce and increase their wages their dollar amount of public
assistance decreases, and their taxes increase. For AFDC grants a dollar
earned reduces the AFDC grant a dollar after the first 12 months of
employment. There are other types of public assistance available to poor
single-parent families that are also affected by increased wage earnings.
These include food stamps, Medicaid, subsidized housing, energy assistance,
and other food assistance. Each program has its different eligibility
requirements, different method for determining assistance levels, and different
rate of reduction given earned income or income received through child
support. Given this multitude of public assistance there exists some instances
where increased earnings are totally offset by their decreased public assistance
and increased taxes. For example, a mother may earn one dollar more a
month, but because reduced public assistance and increased taxes her actual
disposable income remains unchanged. In this particular case, the implicit
marginal tax rate on her marginal earnings is 100 percent, since 100 percent
of her earnings disappeared into taxes and reduced public assistance. She
retained absolutely none of her additional earnings.
A welfare system that sometimes provides little to no economic
incentives to work was not the intention of any of these programs. Each
program was developed to alleviate a specific poverty symptom. The federal
Housing Act of 1949 established the goal of a decent home and suitable living
environment for every American family. For impoverished families this was
carried out by Section 236 housing vouchers, and Section Eight public housing
projects. The housing vouchers offer a housing subsidy for low-income
families living in privately owned apartments or houses.
Other anti-poverty programs such as food stamps were a Great
Depression remanent of food Surpluses redistributed to the poor. In the
1930s orange stamps were bought for a dollar, and could be exchanged for a
dollar worth of food. Blue stamps came free of charge with the orange
stamps and could be used to buy surplus farm food at discounted price. In
the 1960s the food stamp program was restructured as "an exclusive anti-
hunger policy" (Champagne and Harpham 1984,112). The 1960s war against
poverty, also, produced the Medicaid program to assist poor people with
medical coverage. Medicaid was added into 1966 legislation at the last
moment and without much contemplation.
Additional food supplement programs through Woman, Infant, and
Children (WIC) and the school lunch programs were developed to insure
nutritional food consumption by the poor. WIC started in 1972 through the
Department of Agriculture. Low-income energy assistance, a relatively new
program developed during the 1970s energy crisis, assists low-income families
with their heating bills. Finally, there is Title XX which provides child care
subsidies to low-income families through the Social Security Act.
For several reasons AFDC families may not receive all forms of public
assistance. First, there are some public assistance programs that cannot be
received simultaneously. A family cannot receive AFDC and Title XX child
care assistance simultaneously. In most cases, a family cannot receive both
housing subsidies and low-income energy assistance because housing subsidies
are also housing subsidies for utilities. Secondly, some of these programs are
not entitlement programs available to all eligible families. For instance, there
is a limited amount of funding available for housing assistance, Title XX child
care subsidies, and WIC. Thirdly, families do not consider all public
assistance programs desirable. It is embarrassing for some families to pay for
their groceries with food stamps. Some AFDC families avoid public housing
projects because of their reputation as being unsafe, unkept and crime
festering. Finally, an AFDC family may not receive all of the public
assistance and tax credits available to them because they did not know they
Taxes compound the complexity of a multiple public assistance
program. There are progressive state and federal income taxes, and a flat
rate social security income tax. The earned income tax credit gives working
families a tax credit if their income was under $18,576 in 1988. Since this
earned income tax credit is refunded even if the family owes no federal
income taxes, it is called a negative tax. Finally, the child care tax credit
allows a maximum credit of $2,400 a year for one child, and $4,800 for more
than one child. The child care tax credit is not a negative tax; it only applies
toward owed federal taxes.
The interaction of these programs and tax systems affect each other
in their fight against poverty and their strive for fairness. Food stamp
allotment and housing subsidies include child care deductions, but if all of the
child care is being provided by Title XX child care subsidy there can be no
child care deduction. Furthermore, a Title XX child care subsidy interferes
with the child care tax credit. Food stamps are complicated by public housing
allowance because of their shelter deduction. The number of ways these
programs and taxes interfere with each other is endless. The interaction
sometimes causes little economic incentive to increase earnings because the
implicit marginal tax is over 100 percent. This study analyzes when high
implicit marginal taxes occur under the rules established by the 1988 Family
Previous Research on Implicit Marginal Taxes
The implicit marginal tax rate is commonly defined as the reduction
rate of public assistance due to a one dollar increase in income. This study
measures the implicit marginal tax rate for a family receiving a multitude of
public assistance programs. James Gwartney and Thomas McCaleb (1985),
and John Morris and Suzanne Helbum (1988) have completed similar studies.
Other studies have measured the implicit marginal tax rate for only one public
assistance program such as AFDC, and still others have examined the
interaction of multiple public assistance programs without looking at implicit
marginal taxes. For simplicity, this section reviews previous research in
chronological order, rather than the three categories developed in this
Sar Levitan, Martin Rein, and David Marwick (1972) calculated the
potential public assistance available for a four-person, single-parent family
living in Chicago during 1971. Public assistance included AFDC, food stamps,
Medicaid, and public housing. Since medical vendor payments were not
available Levitan, Rein, and Marwick used average Medicaid payments made
to AFDC families. Although their research made no direct mention of
marginal implicit taxes, it can be deduced from their calculations that an
implicit marginal tax rate over 100 percent occurred when the family was cut-
off AFDC and Medicaid. Table 2.1 modifies their original work by including
implicit marginal tax rates. Levitans research is extremely significant since it
demonstrates that economic disincentives to increase wage earnings (high
implicit marginal tax rates) were occurring as early as 1971.
AFDC and its potential work incentives were the focus of Eric
Solberg and Frederick Langille (1974). Although not the main point of their
study, they suggested there is not a dramatic decrease in spendable income,
known as a notch effect, when families become ineligible for AFDC. They
suggest these notch effects are exaggerated by in-kind transfers such as public
housing and Medicaid. In a footnote, Solberg and Langille claim that
Benefits Potentially Available to A Four-Person,
Female-Headed Family in Chicago, 1971
Earnings AFDC Income Tax FICA Food Stamps Medic- aid Public Housing Total Implicit Tax Rate
None $1,000 3,384 _ 408 910 1,416 6,118
$2,000 3,384 . - 52 288 910 1,269 6,799 32%
$3,000 2,890 - 104 288 910 1,089 7,073 73%
$4,000 2,224 156 288 910 908 7,174 90%
$5,000 1,557 - 208 288 910 728 7,275 90%
$6,000 890 164 260 288 910 547 7,211 106%
$7,000 224 349 312 288 910 367 7,128 108%
$8,000 545 364 ~ 186 6,277 185%
$9,000 723 416 6 6,667 61%
- 908 468 -- - 7,624 4%
Source: Sar Levitan, Martin Rein, and David Marvick, Work and Welfare Go Together. Johns Hopkins
University Press, Baltimore, MD, 1972.
previous studies had used medical expenditures averages that were not
indicative of the typical AFDC household because the median was no where
near the average; suggesting that the mean is not a good proxy of medical
expenditures. Furthermore, they conclude that what little notch effect occurs
at a wage that is much higher than low-skilled workers can typically obtain.
Another study (Hutchens 1977) focused solely on the AFDC implicit
marginal tax rate, but suggested that the implicit marginal tax rate would be
much higher if other transfer programs and cumulative tax rates were
There are several other studies that looked at the multitude of public
assistance programs, but not in the context of implicit taxes. Lewis (1983)
reviewed the interaction of AFDC, Title XX child care subsidies, food stamps
and taxes. Mainly, he explored whether single-parent families would be better
off financially on AFDC or Title XX; he concluded they would be better off
on Title XX. MacDonald (1985) looked at the interaction of food stamps
with other forms of public assistance to see if food stamp families still lived
below the federally established poverty level. A study conducted by Fraker,
Moffitt, and Wolf (1985) found that the effective marginal tax rate is lower
than in the AFDC benefit formula because work and child care deduction.
Gwartney and McCaleb (1985) view the effects of increased earnings
on transfer benefits and taxes for a Pennsylvania mother with two children
(see table 2.2).
The Effect of Transfer Benefits and Taxes on the
Incentive of a Pennsylvania Mother with Two Children
to Earn Income (September, 1983)
Annual Transfer Income & Employment Spendable
Gross Wage Benefits Taxes Income
0 7,568 0 7,568
2,000 6,525 134 8,391
4,000 5,482 268 9,214
5,000 3,040 346 7,694
6,000 2,059 611 7,448
7,000 1,719 810 7,909
8,000 1,378 1,021 8,357
9,000 1,038 1,240 8,798
10,000 698 1,469 9,229
Source: James Gwathey and Thomas S. McCaleb. 1985. Cato Journal
Table 5, page 11.
vol. 5, no. 1,
They summed up the AFDC grant, dollar value of food stamps award,
average Medicaid payments, earned income and the earned income tax credit,
and subtracted the federal, state, and social security taxes. This gave them
spendable income. The implicit tax was derived by subtracting the
percentage of retained earnings from one. The percentage of retained
earnings is calculated by dividing the change in spendable income by the
change of earned income. The implicit tax was close to fifty percent for
incomes below $4,000 and over $7,000. That is for every additional dollar
earned, families loss about 50 percent to taxes and forgone public assistance.
They retained fifty percent of their earnings. They measured a 252 percent
implicit tax rate when income rose from $4,000 to $5,000. That is, for the
additional $1,000 earned the family received $2,520 less in spendable income.
Moving from $5,000 to $6,000 the family encountered a 124 percent implicit
tax. After reaching annual earnings of $6,000, the implicit tax went back to
being around fifty percent. According, to Gwartney, "such high implicit
marginal tax rates pose a significant disincentive to work for those individuals
whose potential earnings are relatively low." (Gwartney and McCaleb 1985,
In 1988, Morris and Helbum also studied the implicit tax; they looked
at spendable income for a single-parent with two children living in Denver
while receiving public housing assistance, school lunch subsidy, food stamps,
low-income energy assistance, Medicaid, and AFDC or Title XX child care
subsidy. Working Coloradan single-parents may choose between AFDC
(assuming their income as not reached the cut-off level) and a child care
subsidy, but cannot receive both. According to Morris and Helbum the family
is financially better off if they choose the child care subsidy. AFDC has
inadequate child care deductions, $160 a month per child in 1988, and cuts a
family of three off when they are employed full-time at $4.71 an hour.
Nevertheless, Helbum and Morris found implicit taxes greater than 100
percent while taking the Title XX child care subsidy option; the implicit tax
in their model varied from 27 percent to 8,450 percent. The worst case
scenario, 8,450 percent occurred when wages increased from $6.97 an hour to
$6.98 an hour because the family was no longer eligible for subsidized child
care. Other points where the implicit tax reached above 100 percent also
mirrored cut-off points of various public assistance programs such as food
An alteration of the Morris and Helbum model will be used as a
general model in this study.
The reason AFDC is in desperate need of reform is because it has
grown into a costly multi-billion government programs that politicians along
with others would like to scale back. Their solution is to replace AFDC with
a self-sufficiency program through the 1988 Family Support Act. However,
the Family Support Act merely consolidates existing job training programs
such as WIN, CWEP, and work supplementation into one program. It does
not reform all public assistance programs available to single parents such as
food stamps, and housing assistance. Most AFDC recipients receive another
type of assistance in addition to AFDC.
Each of the public assistance programs in operation were created to
cure a symptom of poverty. Food Stamps insure against starvation, housing
assistance programs provide shelter, WIC provides important nutrients, and
the other public assistance programs also have a narrowly defined goal. Each
program has its own eligibility requirement and own implicit marginal tax
rates without regard to the impact of other programs or taxes. It is the
interaction of these programs that create work disincentives through high
implicit tax rates. These implicit tax rates will be measured in this study for
Colorado single parent families moving off AFDC.
Gwartney and McCaleb (1985) and Morris and Helbum (1988) have
done previous research on implicit marginal tax rates. The model used in this
study will be based on their work.
1. At this time it was known as Aid to Dependent Children (ADC).
It only assisted the children, rather than the entire family. This was
changed in 1962 when it became AFDC.
2. However, in the 1960s, AFDC families averaged 3 children.
Interestingly enough, the average AFDC family size has decreased more
rapidly than average family size for all families.
3. In a 1988 survey conducted by Susan Marine 60 percent of the
Coloradan AFDC parents surveyed indicated that a job was their hope for
4. Nathan Glazer noticed this much earlier. In 1986, he wrote,
"The mere roll call of the initials of different programs, each of which has
had its own evaluations in number, reduces one to dumbness: ARA,
MDTA, JC, NYC, WIN, JOBS, CEP, PEP, CETA, PSE, STIP, HIRE,
YEDPA, PSIP, PICs, YCCIP, YIEPP. They all stand for something that
is substantial and of some consequence or that once excited enthusiasm,
but very little emerges that can be asserted with confidence."
5. Government fiscal years run from October of the initial year to
October of the following year.
6. These target groups were discussed earlier in this section. They
include parents who are under age 24, parents of long-term welfare
dependent families, and parents whose family is about to become AFDC
ineligible due to youngest child reaching adulthood.
7. Currently, Colorado only gives AFDC grants to single-parent
households. This will change by 1990 due to the 1988 Family Support Act
which mandates an AFDC program for poor, two-parent households.
FOR MEASURING ECONOMIC DISINCENTIVES
Implicit marginal tax rates over 100 percent provide no incentive for
a single parent moving off AFDC to increase her wage earnings. The
microsimulation of the public assistance described in this chapter will be used
to find out where implicit marginal tax rates are over 100 percent. There are
two major components to this microsimulation. First is the model which
measures the spendable income and benefits and the implicit marginal tax
rate. Second is defining a prototype family which represents a typical AFDC
family moving off AFDC due to the Family Support Act, and defining relevant
The model simulates public assistance programs and taxes: it
simulates eligibility requirements, award determination, tax formulas, and tax
tables in order to compute public assistance amounts and taxes for a
particular wage rate given specified family characteristics. After public
assistance amounts and taxes are determined, all of the public assistance
amounts are added together, and taxes and work-related child care expenses
are subtracted to derive spendable income and benefits. From spendable
income and benefits the implicit marginal tax rates are derived. The model
is a equation. Spendable income and implicit marginal tax rate are the
dependent variables. The independent variables are family characteristics.
The equation parameters are the eligibility requirements, methodologies to
determine public assistance amount, tax formulas, and tax schedules.
Since this study analyzes the disincentives to increase wage earnings
all of the independent variables except for wage earnings will be held constant
for the prototype family. Therefore, a relationship between wage earnings and
monthly spendable income can be developed and represented by a line
graphed in two-dimensions with wage on the x-axis, and spendable income and
benefits on the y-axis. The implicit marginal tax rate is the slope of this line.
Higher wages which lead to higher spendable income would be represented
by a line sloping upward. If implicit marginal tax is greater than 100 percent
the slope of the line will be negative.
The second component of the microsimulation is specifications of the
family characteristics to be used in the model. Public assistance levels and tax
levels depend on such individual family characteristics as size, child care needs
and costs, and housing costs. Public assistance levels and tax levels also
depend on the gross income of the family. Gross income is a function of the
parents wage rate, hours worked by parent, and child support receipts. A
prototype family which approximates the most-likely Coloradan family will
be developed. Eight sets Of variations of this prototype family will also be
developed. The implicit marginal tax rate for the prototype family and
variations will be calculated using the model. The results are in Chapter
Defining Implicit Marginal Tax Rate
The General Model
The prototype model is adapted from the Morris and Helbum (1988)
model. Monthly spendable income is calculated by summing monthly wages,
child support, AFDC grant, food stamps allotment, average dollar value of
Women, Infants, and Childrens supplement, school meal subsidies, low-
income energy assistance, housing subsidies, and subtracting average monthly
federal, state, and social security taxes, and work-related child care expenses.
It uses Fall, 1988 public assistance eligibility rules and grant determination
methods except for AFDC1 which will include the changes made by the
Family Support Act. Taxes are calculated from the 1988 tax tables. The
model can be summarized by the following general equation:
inc = wage + sup ccare + afdc + fs + leap + hous
+ wic + sch tax..................(1)
The equation variables are defined below:
inc....monthly spendable income and benefits
wage...monthly earned income
sup....monthly child support
ccare..monthly child care cost to the family
afdc...monthly AFDC grant
fs..monthly food stamp allotment
leap...monthly low-income energy assistance
hous...monthly housing voucher
wic....value of monthly women, infant, and children
sch....value of monthly school lunch subsidy
tax....average monthly taxes.
The percent of income retained is the ratio of change in spendable
income to change in earned income. By subtracting this ratio from one the
implicit marginal tax rate is derived. If an increase in wages resulted in a
decrease in spendable income the ratio would be a negative number;
subtracting a negative from one yields to a number greater than one. In
percentage form, this would be an implicit marginal tax rate greater than 100
percent. This equation is summarized below:
rate = 1 (Inc, Inc,.,) / (wage, wage,.,)..(2)
rate....implicit marginal tax rate
inc,....monthly spendable income and benefits at time t
wage,...monthly earned income at time t
Equation one states that spendable income depends on the variables:
monthly wages, child support, AFDC grant, food stamps allotment, average
dollar value of Women, Infants, and Childrens supplement, average school
meal subsidies, low-income energy assistance, housing subsidies, average
monthly federal, state, and social security taxes, and child care expenses.
Equation one can be further broken down into equations which calculate
monthly wages, AFDC grant, food stamps, and others. Thus, these variables
(such as monthly wages, AFDC grant, and food stamps) are interpolated
variables. These interpolated variables depend on the independent variables:
family size, child care needs, and other family characteristics. In this chapter
it is only necessary2 to identify the independent variables. The exact
determination of these interpolated variables (sub-equations) are explained in
Appendix A. In the next section each interpolated variable is examined
separately to identify which independent variable(s) affect it. Then, a list of
independent variables is compiled; a prototype family that typifies the average
AFDC familys characteristics and variations of this family can be developed
to define these independent variables.
Identifying the Independent Variables
Monthly earned income3 is a function of the wage rate the parent
earns, and the hours the parent works a month.
wage = w (wage rate, hours worked)............(3)
For purposes of this study the child support award is considered an
independent variable. In Colorado, the child support award is based on a
formula which involves the income of both the custodial parent and the absent
parent. Although the income of the custodial parent is known in this study,
the absent parents income is not. Therefore, child support award can not be
determined; it will be assumed the child support award is given at $50, $100,
and $200 per month. This will be discussed in greater detail in the next
Not only do child care costs depend on the number of children, and
the number of hours per month they are cared for, but it also depends on
whether or not the family receives subsidized child care. Recall that a family
cannot receive AFDC and child care subsidies simultaneously. There are two
different subsidies available. One is funded under the Family Support Act,
and the other is funded through the Social Security Acts Title XX program.
The eligibility requirements and method for determining subsidy amount are
identical for both these programs. Eligibility is based on income. The
amount a family must pay for child care is determined from a table
considering income4 and family size. Child care costs can be summarized in
the following equation:
ccare = c (number of children, hours of care, child care subsidies
(income, family size, AFDC status))....................(4)
AFDC grants are determined using a formula considering family size,
income, child care costs, and length of time on AFDC. Eligibility is based on
income. Also, an AFDC grant cannot be made to a family receiving
government financed child care subsidies.
AFDC = a (family size, income, child care costs,
length of time on AFDC, child care subsidy recipient).........(5)
Food stamp allotment is determined by a formula relying on family
size, income5, child care costs, and housing costs. Families are not eligible
if their income is above 130 percent of the federally established poverty line.
Furthermore, housing costs can be affected by housing subsidies, and low-
income energy assistance.
food stamps = f (family size, income, child care
costs, housing costs (housing subsidies, leap)).........(6)
No housing subsidy will be made without a voucher. There are more
families eligible for housing subsidy vouchers then there are vouchers
available6. If the family is able to obtain a voucher the amount of housing
subsidy depends on a formula considering family size, income, child care costs,
and the average rent and utilities paid. Eligibility is based on income, it
cannot exceed 80 percent of the local median income.
hous = h (availability of housing vouchers (family size, income,
median income of geographic location, child care costs, rent )).........(7)
Low-income energy assistance level is determined by using a formula
which considers the family size, income, average monthly heating costs in the
winter, and housing subsidy recipient status. Eligibility is based on income;
income cannot exceed 150% of the federally established poverty level.
leap = 1 (family size, income, average heating costs, housing subsidy
The average monthly value of the school meal subsidy depends on
income, family size, number of children school age, and the price of school
sch = s (income, family size, number of children
school age, price of school meals)..........(9)
The average monthly value of Women, Infant, and Children
Nutritional (WIC) supplement depends on income, family size, age of
children, and dollar value of targeted foods. Targeted foods are those which
can be purchased with WIC vouchers. They include milk, juice, cereal, peanut
butter, and other food items. Eligibility depends on income and age of the
children; income cannot exceed 130 percent of the federally established
poverty level, and currently children must be under three years old.
wic = w (income, family size, age of children,
dollar value of targeted foods)........(10)
Average monthly taxes are made up of average monthly federal, state
and social security taxes. The social security tax only depends on earned
income. Federal and state taxes depend on earned income, and family size.
Federal taxes also depend on child care costs. Therefore, total taxes can be
summarized as following:
taxes = t (earned income, family size, child care costs)......(11)
Summary of Independent Variables
Equations 3 through 11 define all the interpolated variables as
independent variables. Interpolated variables compose spendable income and
benefits. In summary, these equations show that the implicit marginal tax
rate, and spendable income and benefits for a single-parent family moving off
AFDC depends on these independent variables: wage rate, hours worked,
child support award, family size, AFDC recipient status, months on AFDC if
receiving it, market child care cost rates if receiving AFDC, number of
children in child care, hours children are in child care, availability of
subsidized housing vouchers, rent and utilities, median income of geographic
location, housing subsidy recipient status, average winter heating bill, value of
school meals, and value of WIC food supplement vouchers.
Thus, equation one can be rewritten as:
inc = i (wage, hours, sup, fsize, afdcstat, afdcmos, mktccare, children,
ccarehrs, rent, med, vouchers, houstat, heat, schmeals,
inc.......monthly spendable income and benefits
wage......hourly wage rate
hours.....average hours worked per month
afdcstat..AFDC recipient status (1 = receives AFDC, 0 = does not
afdcmos.....months on AFDC
mktccare....market hourly child care rate
children....number if children
ccarehrs....number of hours children are in care
rent......rent and utilities
med.......median income of geographic area
vouchers....housing vouchers available (1 = available, 0 = not available)
houstat...housing subsidy recipient status (1 = receives housing
assistance, 0 = does not receive housing assistance)
heat......average monthly heating bill in winter
schmeals.....average monthly cost of school meals
wicfood.....average value of WIC food supplement per month
The Prototype Family and Variations
From these independent variables listed in equation 12 a prototype
family will be developed. This prototype family exemplifies a typical Colorado
AFDC family whose head-of-household will move into the labor market due
to the 1988 Family Support Act. In order to examine how different family
characteristics affect the implicit marginal tax rate, eight sets of families which
vary slightly from the prototype family will be analyzed. The first five sets
examine the differences in implicit tax rates caused by increased child support,
different child care needs, increased family size, and different geographical
location. The sixth set examines the economic disincentives associated with
Medicaid cut-off. The seventh set examines the economic incentives to
increase hours worked, and the eighth set examines what happens to an
AFDCs spendable income during the first 15 months of working.
This section defines the prototype family, and the eight sets of
prototype family variations. In all, there are 30 different fictional families.
The Prototype Family
Only wage rate of the parent will be allowed to vary for the prototype
family. All of the remaining variables are selected to exemplify the typical
Colorado AFDC family moving into the labor force due to the 1988 Family
Support Act. The prototype family consists of one adult working full-time and
one child under the age six. The majority of AFDC families both on the
national and state level have only one child7. Although no figures are
available at the state level, national figures show a large percent of AFDC
children, 44 percent, (U. S. Senate Subcommittee 1987) are under age six.
This is significant since in most incidences a child under age six would need
full-time child care while the parent works full-time. The percent under age
six is far greater than the age brackets which would need before and after
school child care and the age brackets which would need no care.
Age of the child can be further specified. Since AFDC parents with
a child under the age of one year old are not required to work, it can be
assumed that the child in the prototype family is one to six years old. Thus,
the average age would be three and a half years old. This age would make
the family ineligible for WIC food supplements8 and school meal subsidies.
This age would also classify the child as a preschooler; the average monthly
subsidized child care rate for a preschooler is $48.34 (CAEYC 1988) for child
care provided in a home.
The prototype family receives subsidized child care over AFDC
because the parent has been through the JOBS program, and has been
working for over a year. They also receive food stamps and low-income
energy assistance. They do not receive child support payments; currently most
AFDC families do not. It is assumed they do not receive housing assistance
for two reasons. First, in actuality only a small percent9 of AFDC families
receive both AFDC and housing subsidies. Secondly, not many housing
vouchers are available; the waiting time for a housing voucher can be up to
two to three years.
It is assumed that the prototype family lives in a two bedroom
apartment which costs $4051 for rent and utilities. Their winter heating bills
average $42 a month11.
Set One: Varying Child Support Payments
Besides putting AFDC parents to work the 1988 Family Support Act
will step up child support enforcement. The effect child support payments
have on economic incentives to increase earnings will be analyzed in this set.
Three variations of the prototype family will be compared with the prototype
family. Family A will receive $5012 a month in child support payments,
Family B will receive $100, and Family C will receive $200. All other
characteristics of the prototype family, Family A, Family B, and Family C will
Set Two: Varying Child Care Needs
In this set we examine the differences in spendable income and the
implicit marginal tax rate due to different child care needs. Two variations
of the prototype family will be compared with the prototype family. Family
D will have one child in before and after school care13. Family E will have
one child with no child care needs. All other family characteristics are
identical to the prototype family, Family D, and Family E.
Set Three: Two Children
In this set it is assumed the family has two children14, rather than
one. Family F has two children in full-time care; they receive WIC
supplement since it is assumed one child is under six. Family G has two
children: one in full-time care, and one in part-time care. They receive school
meal subsidies for one child. Family H has two children in part-time care
and receive school food subsidies for both. No child care is necessary for the
two children in Family I and they receive school meal subsidies for both
children. Families F, G, H, and I differ from the prototype model only in that
they have two children with varying child care needs, and they may receive
WIC and/or school meal subsidies.
Set Four: Three Children
There are four families each with three children15 in this set. Family
J has two children in full-time care, and one child in part-time care; they
receive WIC and a school meal subsidy.. Only school meal subsidies are
received by Family K which has one child in full-time care and two children
in part-time care. Family L has three children in part-time care and receives
school meal subsidies for all of them. Family M has no child care needs and
receives school meal subsidies. Other than family size, varying child care
needs, WIC food supplement eligibility, and school meal subsidies eligibility,
these four families do not differ from the prototype family.
Set Five: Geographical Differences
In this set spendable income arid the implicit tax rates for four
different families living in different regions of the state are compared. Family
N lives in Boulder, Family O lives in Pueblo, Family P lives in Denver, and
Family Q exemplifies a typical Colorado family. Family Q is identical to the
prototype family except they receive a housing subsidy rather than leap.
Family N, Family O, and Family P also receive a housing subsidy; however,
their housing subsidies differs with each geographic location. Furthermore,
their child care costs differ with each geographical location. Other than these
two characteristics, these four families are identical to the prototype family.
There are several reasons why these cities were selected for
comparison. Boulder is one of the most expensive Colorado areas to live in
with monthly child care costs averaging $83.57 for a preschool child age two
to five years old (City of Boulder Childrens Services Division 1988, 1), and
boasting a fair market rent of $54216 (Federal Register 1988, 36707) for a
two bedroom apartment with utilities included. Nevertheless, less than three
percent of Colorados AFDC population live in Boulder (Marine 1988a, 4).
Other counties that may have living expenses greater than Boulders are the
counties hosting Colorados alpine ski industry such as Pitkin County (Aspen
area) and Summit County (Dillion and Breckenridge). Both counties contain
less than one percent of the total Coloradan AFDC population; this percent
is also common for other counties whose economy is heavily dependent on the
ski and tourist industry.
On the other hand, Pueblo is a relatively inexpensive area to live in.,
Yet, it has a relatively large AFDC population. The largest percent of AFDC
cases in Colorado live in the five county, Denver metropolitan area.
Metropolitan Denvers 1988 fair market rent for a two bedroom including
utilities is $502; this is slightly higher than the FMR rent for Colorado as a
whole. Child care costs for a Denver area preschool age child averaged
$57.95 for home care in 1988 (Mile High Child Care).
Set Six: Medicaid Cut-off
Over 80 percent of the Colorado AFDC clients surveyed (Marine
1988b) indicated that loss of Medicaid benefits is the biggest fear for them
getting off AFDC. This set will provide further insight in the Medicaid cut-
off issue. In this set we compare the spendable income and implicit marginal
tax rates for four different families. Families R, S, and T are identical to the
prototype family except Family R pays $50 per month in health insurance,
Family S pays $100 per month, and Family T pays $20017 per month. Family
U differs from the prototype family in that the head-of-household only works
20 hours a week18, and the family remains on AFDC rather than taking the
child care subsidy. By remaining on AFDC Family U also remains Medicaid
eligible. To compare Families R, S, and T with Family U the cost of medical
insurance is subtracted from spendable income and benefits. Cost of medical
insurance is subtracted since it is an additional going-to-work expense that
would not have occurred if the family would have remained Medicaid eligible.
Thus, Medicaid is treated like a health insurance policy in this
comparison. Medicaid is difficult to put a dollar value on because it is not
used unless a family member becomes ill or injured19. Yet, Medicaid
eligibility provides a security that catastrophic health care needs will be met;
it acts like a health insurance policy.
Set 7a and 7b: Increasing Hours Worked
The 1988 Family Support Act does not explicitly mandate a parent
must work full-time, 40 hours per week. It does say that parents who already
work at least 30 hours per week are exempt from the JOBS program, and that
parents with children under age six are only required to work a minimum of
20 hours per week. To analyze the economic incentives to increase hours
worked, two sets of families will be analyzed. First, the spendable income for
Families V, W, and X will be compared. These families follow the prototype
family identically except in four areas. First, they receive AFDC over the
child care subsidy. Secondly, the hours worked per week will vary, and the
wage rate is held constant: $3.35 for Family Vs head-of-household, $4.55 for
Family Ws, and $6.00 for Family Xs. The current minimum wage is $3.35
per hour, minimum wage proposed in the 101th Congress was $4.55 per hour20
which also happens to be very close to the wage most job training programs
place participants at, and $6.00 per hour is the typical goal of many job
Working part-time will also affect child care costs. Therefore, the
proxy21 for hourly child care rate, $1.58 per hour, will be multiplied by the
hours worked per month to get monthly child care costs. This brings us to the
second set used for analyzing economic incentives to increase hours work. In
this set wage is held at $4.55 for Families Y, Z, and AA. Hours worked are
allowed to vary and the families prefer AFDC over Title XX. However,
hourly child care rates vary. Family AA pays $1.00 per hour, Family Z pays
$1.58 per hour, and Family AA pays $2.00 per hour. This set not only
indicates the economic incentives associated with increasing hours worked, but
also how different wage rates and different child care rates may affect these
Set Eight: Months Employed
In this set no implicit tax rates are measured. Only months working,
which affect AFDC grant determination, are allowed to change. Wages are
held constant; Familys BB head-of-household earns $3.35 an hour, Familys
CC head-of-household earns $4.55 per hour, and Familys DD head-of-
household earns $6.00 per hour. Other than that, these families follow the
prototype family identically.
The dollar value of both the AFDC grant and Food Stamps allotment
vary drastically from month to month for the first fifteen months a family
receives public assistance. There are several reasons. First, AFDC allows a
$30 and one third earned income disregard for the first four month of
employment; it is reduced to a $30 income disregard for the following eight
months of employment. After one year, all income disregards are eliminated.
The income disregards in the first four months mean that the family gets to
keep the first $30 they earn with no reduction in AFDC, and for every dollar
they earn their AFDC grant is only reduced by 67 cents. For the next eight
months, only the $30 dollar disregard applies, and for every dollar earned
after that, their AFDC grant is reduced a dollar.
The lag is lengthened to two more months due to income reporting
methods. A familys previous month income is used to determined a familys
AFDC grant for next month. Thus, the complete elimination of the work
disregard would not occur until after the fourteenth month of employment.
The complexities multiply when food stamps are included because food stamp
allotment depends on the previous months earned income and the previous
months AFDC which is dependent on another previous months income.
Thus, this months food stamp allotment depends on income from two and
four months ago.
Additional problems results from child care deductions. To receive a
child care deduction, a family must provide a receipt. Therefore, they can not
claim child care deduction until after the first month of employment. This
makes the current months AFDC check and food stamp allotment dependent
on time lagged two, three, four and five months.
Essentially, this set will critique AFDCs and food stamps income
reporting method, it critiques AFDCs changes in earned income disregard.
The model, equation 1, simulates public assistance programs and the
tax system in order to compute spendable income and benefits. Equation 2
uses a change in spendable income and a change in wage earnings to calculate
the marginal implicit tax rate. Equation 1 can be further broken down into
interpolated equations; these interpolated variables contain the independent
variables, or family characteristics. By defining these family characteristics for
a typical Colorado AFDC family (a family moving into the labor market due
to the Family Support Act) a prototype family is built.
The prototype family consists of a single parent family with one child
preschool age, and a mother working 40 hours a week. They receive a Title
XX child care subsidy, food stamps, and low-income energy assistance. The
family receives no child support payments. The parent has already been on
AFDC for a year and through the JOBS program. Child care costs, housing
costs, and heating costs when appropriate are based on average Colorado
There are eight sets of family variations. The first discusses what
happens to the implicit tax rate as child support payments are increased. The
second, third, and fourth variation look at various levels of child care needs
(i.e., full-time care verses part-time care verses no care) for a family with one
child, a family with two children, and a family with three children.
Geographical differences are discussed in the fifth set. The economic
disincentives associated with the Medicaid cut-off are examined in the sixth
set. The seventh set discusses the economic disincentives to increase hours
work while on AFDC. Finally, the last set examines the economic incentives
to continue working while on AFDC during the first 15 months of
1. The employment expense deduction will increase from $75 to
$90 due to the Family Support Act. Also, the child care deduction will
increase from $160 per child to $175 per child.
2. These sub-equations are long, complicated formulas which
describe the public procedures for estimating public assistance awards and
taxes for a particular earned income. The formulas of these interpolated
variables depend on eligibility requirements, methodology to determine
grant award, tax schedules, tax formulas, and/or family characteristics.
Only family characteristics vary, thus the rest of the formula is irrelevant
to this chapter. These sub-equations will be of greater interest to those
who wish to duplicate this model.
3. This is referred to as Svage in equation one.
4. Income typically refers to child support award, AFDC grant,and
wage earnings. When the family receives a child care subsidy income
refers only to child support award, and wage earnings.
5. This was previously discussed in Chapter One.
6. Nationally 44 percent of the AFDC families have one child
(Senate Subcommittee 1987). For the state of Colorado, 46 percent of the
AFDC families had one child (Marine 1988a).
7. The cut-off age is three in many Coloradan counties.
8. In Colorado, 28 percent of the AFDC population receives both
AFDC and housing subsidies (Marine 1988a). Nationally, 25 percent of
the AFDC population receives both AFDC and housing subsidies.
9. The weighted average fair market rent (FMR) for AFDC county
populations in Colorado was $476 in 1988. This amount is multiplied by
85 percent since the Department of Housing and Urban Development
suggests that 85 percent of the FMR represents the lowest rent range at
which a supply of units would be available. See Appendix B for further
10. The average electric heating costs were $42 a month during
1988 (Low Income Energy Assistance Program, 1988).
11. No good statistics on average child support payments per
AFDC child in Colorado are available. However, Barbara Bergmann
(1987: 46) cites the national average child support was $104 for an AFDC
12. The monthly market rate is estimated to be $158 a month for
child care provided in the home during 1988. This is estimated in
13. Nationally, 30 percent of the AFDC families have two children
(Senate Subcommittee 1987). This statistic is identical for the state of
Colorado (Marine, 1988a).
14. Nationally, 16 percent of the AFDC families had three children
(Senate Subcommittee 1987). For the state of Colorado this was 17
percent (Marine 1988a).
15. Although, for this model only 85 percent of the FMR is used.
This is in accordance with HUDs policy.
16. According to a recent Wall Street Journal article by Albert
Karr and Mary Lu Camevale the average medical insurance premium for
employers per employee was $2,555 a year in 1988, $213 a month, and
$3,117 a year in 1989. Robert Kuttner states a good health insurance for
a family of four costs $150 to $200 in his article on welfare reform.
17. If Family Ts head-of-household worked full-time, even at
minimum wage, the family would become Medicaid ineligible.
18. See Gwartney and McCaleb (1985) for past use of this method;
and see Solberg and Langille (1974) for criticism.
19. The 102nd Congress has just passed a new minimum wage of
$4.25 per hour.
20. Refer to Appendix B.
WHERE ECONOMIC DISINCENTIVES OCCUR
In Chapter Three a model to measure monthly spendable income and
benefits and the implicit marginal tax rate was developed. Eight sets of
families were developed to analyze the effects of increased child support,
different child care needs, larger family sizes, geographical differences, the
economic disincentives associated with Medicaid cut-offs, the economic
incentives to increase hours work, and the economic incentives to increase
months Worked. Monthly spendable income and benefits, and the implicit
marginal tax rate for these eight sets of families are presented in this chapter.
Overall, the results show that there is a critical range between $5.00
and $7.00 where there may be no economic incentives to increase earnings for
most family scenarios analyzed. Also, increased child support payments cause
economic disincentives (to increase earnings) at much lower wages. Larger
families and families with greater child care needs face greater economic
disincentives to increase wage earnings. Families that live in more expensive
areas such as Boulder, also, face greater disincentives. If the family needs
medical insurance there is no incentive to get off medicaid unless employer
subsidized health insurance can be provided. Finally, there are economic
disincentives to work more than a few months for AFDC head-of-household
who have just gone to work.
Graphing the Results
For all of the sets except two, the results will be presented in graph
form with wage rate on the x-axis and monthly spendable income and benefits
on the y-axis. The slope is the implicit marginal tax rate. If there is a
positive relationship between wage rates and spendable income the slope or
line will always move in a positive direction. If a pay raise (increase in the
wage rate) results in a decrease in spendable income (an implicit marginal tax
rate above 100 percent) the line will move in a negative direction. For
simplicity, a line that moves in a positive direction and then reverses direction
to move in a negative direction is referred to as a cliff effect. This is
identical to the term notch effect discussed by Solberg and Langille (1975)
only they focused on AFDC and Medicaid. Sets one through six are
presented this way.
Set seven is presented with hours worked on the x-axis and monthly
spendable income on the y-axis. Set eight is presented with months worked
on the x-axis and monthly spendable income on the y-axis. All of the results
are presented in graph form accompanied by a table listing spendable income
for each wage, or in the case of set seven and eight hours worked and months
worked, respectively. Appendix F gives complete tables which list wage rate,
hours worked, AFDC grant, child Care costs and subsidies, food stamps
allotment, low-income energy assistance, housing subsidies, WIC food
supplement, school meal subsidies, social security taxes, federal income taxes,
and state income taxes.
Economic Incentives to Increase Wages
This section examines the work incentives for the 31 families described
in the previous chapter. Families with greater child care needs have less work
disincentives along with families living in areas of Colorado with higher cost
Families with medical insurance needs have an incentive to work part-
time and alternate between work and AFDC.
Set One: Child Support Payments
Cliff effects occur at lower wages for families receiving child support
payments (figure 4.1). Child support is treated as income in most eligibility
programs. With no child support the cliff effect occurs by $5.00 because a two
member family becomes ineligible for food stamps at 130 percent of the
poverty level ($4.83 an hour) and at $6.50 because they are no longer eligible
for child care subsidy. When the Colorado prototype family receives $50 per
month in child support the cliff effect occurs by wages of $4.65 and $6.00. If
they receive $100 per month it occurs by $4.50 and $6.00 per hour and at $200
as month child support there is only one cliff effect at $5.00 per hour. In all
these cases the lower cliff effect is caused by food stamps being cut-off at a
lower wage rate and the child care subsidy being cut-off at a lower wage rate.
Monthly Spendable Income & Benefits
Set 1: Increasing Child Support
Prototype Family Family A Family B Family C
(no child support) (child support = $50) (child support = $100) (child support = $200)
Monthly Spendable Income & Benefits
Set 2: Varying Child Care Needs
Prototype Family Family D Family E
(full-time child care) (part-time child care) (no child care)
Monthly Spendable Income & Benefits
Set 3: Two Children
Family F Family G Family H Family I
(two In full-time) (one In full, one In part-time) (two In part-time) (none)
Set Two: Varying Child Care Needs
A family with full-time child care needs (the prototype family) will
have the same spendable income as a family with one child in full-time care
(family D) until wages near $6.50 per hour (see figure 4.2). This is because
the amount a family is required to pay toward child care is the same
regardless of the amount of child care they receive when they are enrolled in
a child care subsidy program. Furthermore, the child care disregard for food
stamps is adequate for part-time child care, but inadequate for full-time child
care. By $6.50 per hour, both families are cut-off the child care subsidy
program and now must pay the full market cost of child care.
A family with one child with no child care needs always fares better
than a family with child care needs.
Set Three: Two Children
Increased wage earnings for a parent with two children, neither of
them requiring child care, will always result in increased spendable income
and benefits except when getting a $0.35 raise after earning $4.65 per hour
(see figure 4.3). At this point school meal subsidies are reduced in half.
As child care needs increase for a family with two children, the cliff
effects get worst. Cliff effects begin to occur by $5.00 an hour for a family
with two children both in part-time care due to school meal subsidy reduction.
They also occur at $6.50 due to food stamps cut-off and $7.00 per hour due
to child care subsidy cut-offs. Cliff effects due to food stamps cut-off and
child care subsidy cut-offs occur at identical wages for a family with two
children, one child in part-time care and the other in full-time care, and for
a family with two children in full-time care.
Set Four: Three Children
A parent with three children will always experience economic
incentives to increase earnings regardless of their familys child care needs
until about $6.00 per hour wage. At this point, school meal subsidies are
reduced (see figure 4.4). Shortly, after that low income energy assistance,
WIC food supplements, and school meal subsidies are cut-off. The big cliff
effect occurs near $7.50 per hour when the family is cut-off child care
Set Five: Geographical Differences
This section analyzes the implicit tax rate with respect to geographical
differences (see figure 4.5). Child care costs and housing costs vary
geographically along with Title XX child care subsidies and housing assistance
eligibility. A general Colorado family with one child in full-time care
receiving Title XX subsidy, food stamps, and housing assistance does not
encounter a cliff effect until the head of households wage rate reaches $5.00
per hour for full-time work; this is a very small disincentive to earn more.
Food stamps are cut of by this point. Additional cliff effects occur at $6.50
and $7.00 per hour. Child care subsidy is cut-off by $7.00 per hour. Prior to
$7.00 per hour the implicit tax rate varies from 68 to 82 percent. Thus for
Monthly Spendable Income & Benefits
Set 4: Three Children
0 $3.35 $4.00 $4.25 $4.50 $4.65 $5.00 $5.50 $6.00 $6.50 $7.00 $7.50 $8.00 $9.00 $10.00 $11.00
Family J Family K
(two in full, one in part-time) (two in part, one in full-time)
(three in part-time)
Monthly Spendable Income & Benefits
Set 5: Geographic Location
Family N Family O Family P Family Q
(Boulder) (Denver) (Pueblo) (General Colorado)
every additional dollar earned this family gets to keep $0.18 to $0.32 of it.
After $7.00 per hour the tax rate ranges from 28 to 39 percent.
The implicit marginal rate for the Denver family is almost identical to
the general Colorado familys. The cliff effects are at $5.00, $6.00, and $6.50
an hour, and are due to cut-offs from food stamps, child care subsidy, and
housing assistance. The implicit tax rate prior to $5.00 an hour wage rate are
some what smaller than Colorado as a whole. They range from 68 to 80
Boulder, one of the more expensive counties to live in the state,
contends with much more severe cliff effects at $6.00 and $6.50 wages due to
higher child care costs and housing costs relative to Denver and the remaining
state. A Boulder family moving from $6.00 to $6.50 wage has 283 percent
implicit tax rate compared to 186 percent for a general Coloradan family.
This cliff effect may be exaggerated because the model used average
negotiated Title XX rate for families eligible for Title XX subsidy, and
average market rates for families ineligible for Title XX subsidy. But, there
is no rationale for assuming that all families coming off Title XX would be
forced to accept average market rates. If the county, a large entity, can
negotiate child care rates below market costs why could not a small entity, a
single family, negotiate a child care rate lower than the market average?
Unlike the Colorado, Denver, and Boulder family a Pueblo family
with one child receiving Title XX does not experience a cliff effect at $7.00
per hour. Although, they experience a cliff effect at $5.00 and $6.50 wages for
identical reasons: at $5.00 food stamps cut-off, and at $6.50 Title XX cut-
The eligibility requirements for these two programs are identical
throughout the state. However, housing assistance eligibility rules differ in
each region. In Pueblo the family would be cut-off at $6.58 per hour from
housing assistance. Housing costs are relatively lower in Pueblo than in
Colorado, $370 compared to $405; therefore, the magnitude of the cliff effect
Disincentives Associated with Medicaid Cut-off
Set Six: Medicaid Cut-off
Being cut-off Medicaid is a major fear for many AFDC families1. If
the parent can obtain employer provided health insurance this may alleviate
that fear. In this section we compare a family that remains Medicaid eligible
with a family whose head-of-household has been in the workforce for over a
year. Figure 4.6 demonstrates that a family is better off if the head-of-
household works full-time and can receive employer subsidized health
insurance. This would mean the insurance premium would be less than $200
per month2. Otherwise, the family is better off while the parent works part-
time, and therefore the family remains AFDC eligible and remains Medicaid
eligible. If the latter option is chosen, their Medicaid and AFDC would be
cut-off by earning $6.50 per hour 20 hours per week.
Monthly Spendable Income & Benefits
Set 6: Medicaid Cut-Off
Family R Family S Family T Family U
(insurance premium = $100) (insurance premium = $150) (insurance premium = $200) (no premium, Medicaid)
Economic Incentives to Increase Hours Worked
Set Seven: Increasing Hours Worked
In this section several disincentives to increase hours worked for an
AFDC family are examined. In this set the parent has been working for over
a year and still remains on AFDC if eligible. At a wage rate of $3.35 and
$4.55 parents experience cliff effects at 25 hours a week (see figure 4.7a). If
the parent earns $3.35 per hour, monthly spendable income and benefits
continue to decline as hours worked per week increases. At $4.55, spendable
income increases as hours worked per week increases after the parent has
worked 30 hours a week. At $6.00 per hour, the parent has an economic
disincentive to work 30 hours to about 40 hours a week due to a large cliff
effect. These cliff effects reflect food stamps and/or AFDC cut-offs. In this
case, the parent may choose the Title XX child care subsidy option, over the
AFDC option. Almost all families are financially better if they go the child
care subsidy route if medical insurance is not a problem.
The second part of this section analyzes the effects of varying child
care rates. If cheaper child care rates are available, such as $1.00 per hour,
there will always be an economic incentive to increase hours worked assuming
the wage is $4.55 per hour. If the child care rates are $1.58 per hour or $2.00
per hour the economic incentives to increase hours become hazy to
nonexistent until about 38 hours per week. This is because the AFDC child
care deduction does not cover the full cost of market child care.
Set 7a: Increasing Hours Worked
Changes in the Wage Rate
Family W Family V Family X
(wage rate = $4.55) (wage rate = $3.35) (wage rate = $6.00)
Set Eight: Increasing Months Worked
AFDC grant determination for next month is based on the previous
months income. For example, the amount of AFDC grant awarded in
November is determined by using Septembers income. This is also true for
food stamps3. This lag creates a major fluctuation in spendable income for
AFDC and food stamps recipients in their first 12 months of employment. It
becomes smooth after a year if their wage income along with their child
support income remains constant. To further complicate things, AFDC
changes its method of grant determination twice during the first year an
AFDC client goes to work; once after four months of employment from a
$30 and one third earned income disregard to a $30 disregard for the next
eight months. After 12 months there is no earned income disregard. So, a
dollar earned reduces the familys AFDC grant by a dollar. The two month
lag between income earned and receipt of that AFDC grant (based on those
earnings) delays the changes of earned income disregards to the sixth, tenth
and fourteenth month of employment. The 1988 Family Support Act will not
Since AFDC is considered income in determining food stamps
awarded the lag problem multiplies. A change in income will take two
months to be reflected through AFDC, and another two months to be
reflected in food stamps. Therefore, food stamps are determined by income
earned two and four months prior. The picture becomes even more
complicated when child care disregards are included since a receipt must be
present. Thus, child care costs disregards are not seen until the third month
in which they accrue.
The result of these program idiosyncracies are shown in Figure 4.8.
Spendable income is at its highest during the first two months of employment
and then plummets. Spendable income rises again in the fifth and sixth
month, but not as high as the first two months. This pattern is identical for
wages at $3.35, $4.55, and $6.00 while working only twenty hours a week.
Full-time work is not considered because a two member family would not be
AFDC eligible after the fourth month of employment if the parent earns more
than $3.44 an hour. The JOBS program hopes to place participants at wages
higher than that; most recent surveys indicate job training programs average
over $4.00 an hour4.
Summary of Results
Several economic disincentives to increase wages have been pointed
out in this chapter. Most of these points occur due to cut-offs from public
assistance programs. The most severe instances stem from wage increases
that moved a family from food stamp eligible to food stamp ineligible and
Title XX child care subsidy eligible to Title XX child care subsidy ineligible.
There were other salient results. Increased child support awards force cliff
effects to occur at lower wages. Cliff effects, also, occur at much lower wage
levels for larger families if they have child care costs. More expensive cities
such as Boulder have higher implicit tax rates. Smaller cities such as Pueblo
Set 7b: Increasing Hours Worked
Changes in Child Care Costs
(hourly child care cost = $1.00)
(hourly child care cost = $1.58)
(hourly child care cost = $2.00)
Set 8: Increasing Months Worked
Family BB Family CC Family DD
(wage rate = $3.35) (wage rate = $4.00) (wage rate = $6.00)
have a more consistent non-decreasing relationship between wage earnings
and spendable income.
There are economic disincentives associated with Medicaid cut-offs if
the family must pay full health insurance cost. There are economic
disincentives to increase hours work. Parents working minimum wage and
remaining on AFDC and Medicaid have no economic incentive to work more
than 25 hours a week. If their child care costs are over $1.58 per hour the
family will be worst off when the parent works 30 hours per week.
1. See Marine (1988a).
2. See page 44 for discussion on average medical insurance costs
3. This is not true for housing assistance, nor leap. Housing
assistance reassesses income yearly and low-income energy assistance
assesses income twice: that is once for November through January, and
again for February through April.
4. For example, see Proctor (1988).
RESTORING WORK INCENTIVES
AND ADDITIONAL CONSIDERATIONS
A public assistance program that has economic disincentives to accept
pay raises, economic disincentives to increase hours work, economic incentives
to stay Medicaid eligible, thus AFDC eligible, and economic disincentives to
remain employed more than a few months is contrary to the goals of the 1988
Family Support Act. This public assistance system does not emphasize work,
nor does it encourage self-sufficiency. Long-run welfare dependency will not
end, rather it will manifest itself. The 1988 Family Support Act is inadequate;
additional welfare reform is necessary. This can be done via a comprehensive
solution such as the negative income tax or each public assistance program
and tax program could be restructured to eliminate economic disincentives.
If a comprehensive solution is taken it must be done at the federal level.
Piecemeal solutions can take place at the federal, state and local level.
There are additional considerations that must be taken before only
solution is sought. For instance, if the parent is placed in a job that pays a
wage above wages with cliff effects they bypass the portion of the pay-scale
with little incentives to earn more. They face only positive economic
incentives to increase earnings. Increased child support payments, as
demonstrated in this study, force economic disincentives to occur at lower
wages. It is unknown how much additional child support will be received.
The likelihood of increased child support payments is worthy of further
discussion. Furthermore, the implicit marginal tax rate will change with time
due to inflation1. Food stamps, housing assistance, leap, school meal subsidy,
and WIC food supplement eligibility requirements are based on the federally
established poverty level which changes each year with the general price level.
Finally, there are reform measures currently proposed that could affect the
economic disincentives revealed in this study. These include proposals which
affect child care and medical insurance. Child care policy is tantamount since
all of the families with no child care needs modelled in this study have
consistent economic incentives to increase earnings. Whether or not these
measures, especially in the area of child care, could alleviate current economic
disincentives is worth further discussion.
Restoring Economic Incentives
To restore economic incentives additional welfare reform measures
are necessary. This section discusses the advantages and disadvantages of a
comprehensive solution such as the Negative Income Tax, along with the
advantages and disadvantages of reforming each public assistance program
The Negative Income Tax
The food stamps program is called a straightforward negative income
tax (Weinberg 1986). It guarantees a specified level of food stamps and then
reduces that amount by $0.80 for each dollar earned. The cut-off level is 130
percent of the poverty level. The only difference between the NIT and the
food stamps program is that food stamps is an in-kind transfer, it can only be
used for food; whereas, the NIT program was direct cash assistance. The
multitude of public assistance programs, also, acts as a negative tax. However,
sometimes it reduces the amount of public assistance by more than a dollar
for every dollar earned. This was demonstrated over and over again in
Review of the Negative Income Tax Experiments
The Negative Income Tax (NIT) was touted in the early 1960s as a
solution to the handicaps of a multiple public assistance program. It was
assumed that one program which provided direct cash transfers rather than
in-kind transfers would be cheaper to manage. It would also be beneficial
to recipients because they prefer cash to in-kind transfers2, and there would
always be an economic incentive to earn more. NIT experiments were
conducted in the 1960s and 1970s in five areas of the country. Countless
studies3 analyzing the results show that rather than increasing work efforts,
NIT caused work efforts to decline.
Most research analyzing the NIT experiments indicates that the
guarantee level4 had a much larger effect on work efforts than did tax rates;
furthermore slight variation of tax rates had little effects. In one of the most
widely quoted articles on the NIT experiments, Irwin Garfinkel and Larry L.
Orr (1974) take it one step further. They suggest that short of reducing
guarantees nothing will motivate work.5
Yet, the evidence may prove them wrong. Essentially, the inflationary
decrease in AFDC need levels6 are a reduction of the guaranteed level. In
Colorado, AFDC benefits have actually decreased 38.8 percent from 1970 to
1987 (Shapiro 1988). The percentage of AFDC recipients who work held
steady in the range of 13 to 16 percent from 1969 to 1979 (U.S. Senate
Subcommittee 1987). Then, in 1983 and 1984, it dropped to about five
percent (Ibid). What accounted for this sudden decrease in work effort?
According to Garfinkel and Orr, the work effort should have increased due
to a reduced guarantee level7. The tax rate on earned income was increased
in 1981, could this explain the reduced work effort predicted by Danzinger
(1983)? If so, this indicates the tax rate level does matter.
On the other hand, work effort was not reduced by AFDC recipients
surveyed in Georgia due to the 1981 AFDC tax rate changes on earned
income (Wodarski, et. al. 1986). However, this was a survey conducted
amongst households who became AFDC ineligible due to increased earnings.
Only 23 percent of the women surveyed considered quitting or reducing hours
to remain AFDC eligible. They cited the need for independence as the
greatest reason for continuing work.
The effect of wage rates on work efforts is interesting. The Gary and
Denver NIT experiments both had decreased hours work for female head-
of-households, but not by identical amounts. In an explanation offered by
(Moffit 1979), Garys single-parents reduced their labor supply by a larger
amount then Denvers because Denver single-parents had higher paying and
better jobs than their Gary counterparts; therefore, Denver parents were less
willing to give up their jobs.
Research by Robert Plotnick (1983) indicates that lower AFDC
guarantee rates decrease the monthly duration on AFDC, and higher wages
increase the duration off AFDC substantially. The results from these studies8
may suggest that wage rates might matter more than tax rates or guarantees.
This thought contradicts some of the conclusions made about the NIT
There are other reasons why the NIT experiments results should be
used with caution. Burtless (1978) hints that all taxes such as social security,
unemployment insurance, and public assistance programs may not have been
included in NIT calculations; if so, not all variables were included in the
modelling. Robert Lerman (1976) puts it more blatantly.
It is now clear that the NIT would coexist with other programs or taxes
that reduce the marginal return from earnings. Even if food stamps
were eliminated, state supplements, Social Security taxes, state and
local taxes, subsidized housing, child care, higher education, and
national health insurance (or Medicaid) would raise the effective tax
rate. (Lerman 1976, 3)
Certainly, the New Jersey results may have been skewed since many NIT
participants simultaneously received other forms of public assistance (Kershaw
and Fair, 1976).
NIT as a Solution
NIT offers an alternative to the current public assistance system. It
is straightforward; therefore the problems associated with the interaction of
public assistance programs could be alleviated. Yet, on the surface, the
results from the NIT experiments appear somewhat discouraging; on a more
in-depth level the results are ambiguous. More research as to what happened
in the NTT experiments would need to be done. Also, more research on the
labor supply for female heads-of-households is needed to get a better
understanding of whether or not female head-of-households have a backward
bending labor supply. Several approximations of the labor supply curve of
female household heads such as Frank Levy (1977) have been done. A study
by Garfinkel and Masters (1974) points out that greater social pressures to
work significantly influence a womens decision to work. Social pressures
mitigate income effects which normally reduce work efforts.
Even if NTT does not have negative work effects, if it were introduced
as a comprehensive reform measure, would it give AFDC parents the job
training they may need? It could afford AFDC parents the opportunity, but
it would not have the degree of job training guidance imposed by the Family
In conclusion to this discussion on NIT and work incentives, it can
be said that the current welfare reform legislation has eliminated the problem
of whether work efforts are decreased by higher guarantees or higher tax rates
by making work compulsoiy. Nevertheless, the real impacts welfare reform
will have on self-sufficiency will be small since work is not the equivalent of
self-sufficiency. The Family Support Act will merely shift poor AFDC non-
working households into the working poor. Perhaps, a combination of
compulsory work with education and job training options and a negative
income tax could provide a solution.
The advantage of having specialized and individualized in-kind
transfers and public assistance programs such as one for housing and another
for child care is that they insure transfers are used only for their stated
purpose. The disadvantage is that interactions of these different public
assistance programs create high implicit marginal tax rates.
Even if all public assistance programs were structured to encourage
work effort, they would be unfair for those incapable of work such as the
handicapped and aged. For example, although 80 percent of the AFDC
recipients receive food stamps and may be capable of work, over forty percent
of Food Stamps households receive social security due to old age, survivors
insurance or disability (Macdonald 1985). Only thirty percent of the Food
Stamps households receive AFDC (Ibid). The implies that a work program
or a self-sufficiency program may be unreasonable for the share of food
stamp households that are unable to work as determined by Social Security
However, there is no doubt that food stamps needs reform9. Its cut-
off has a consistent cliff effect. This requires federal intervention since food
stamps are a federal program.
Limited Possibilities for State Initiatives
Essentially, the state of Colorados ability to change the public
assistance system is tied to the federal government. The federal government
strictly regulates what a state may do. For example, federal matching funds
would be reduced to 50 percent unless 55 percent of the JOBS funds is spent
on federally specified AFDC targeted population10.
Food stamps are 100 percent funded through federal dollars whereas
federal dollars for AFDC is closer to two-thirds. Thus, a welfare dependent
family saves the federal government more than the state government when the
parent goes to work and becomes ineligible for AFDC and food stamps. An
empirical study done by David Long (1988) found that the federal government
does far better than the state government in budgetary savings through
Last year the state of Colorado sought a welfare reform measure
which was revenue neutral: it cost neither the federal government nor the
state government more. This plan was curtailed by the 1988 Family Support
The state does have discretion to enact legislation which would
improve child care and medical care for AFDC parents moving toward self-
sufficiency. They could also raise the need standard for AFDC. Yet, most
improvements call on a financial commitment from the State which they are
unwilling to make.
Linking Official Poverty to Self-Sufficiency
There is something very seriously wrong with the federal definition
of poverty and its linkage to self-sufficiency. The linkage occurs through the
public assistance program. Currently, the official poverty is defined as the
Thrifty Food Plan (TFP) multiplied by three. The TFP is based on a market
basket of food required for a nutritionally balanced diet consumed for a short-
term period. This poverty level is an inadequate measure of self-sufficiency.
The housing cost used in this study, which was based on HUD statistics and
represents average rent for a modest two-bedroom apartment in Colorado,
was $405 per month in 1988. If a family of two living at the 1988 federally
established poverty level paid $405 per month for rent this would be 63
percent of their income; for a family of three living at the 1988 federally
established poverty level11 rent this would be 50 percent of their income. A
recent Census announcement (1989) reported that the poor pay 22 percent of
their income to child care. Summing this up for a family of two living at the
poverty level, they would need 33 percent of their income for food12, 63
percent for rent, and 22 percent for child care. This totals to 118 percent of
the poverty level just to exist. This does not include transportation, medical
insurance nor medical expenses, clothes, laundry, and other basic household
goods. A family of three would need 105 percent of their 1988 federally
established poverty level.
The point suggested by these percentages is that the federally
established poverty level is not a self-sufficiency level. So, why is public
assistance eligibility for low-income energy assistance, food stamps, Women,
Infant and Children food supplement, and school lunch program dependent
on the federally established poverty level? Certainly, if these programs used
self-sufficiency incomes for eligibility requirements which are higher than the
poverty level more people would be eligible; thus, the public assistance
programs would become larger and constitute a greater share of the
government budget. But, economic incentives to become self-sufficient cannot
exist if public assistance programs cut-off before a family achieves a self-
In 1988, food stamps, one of the largest dollar amounts of public
assistance, cut-off for a family of two at $4.83 if the parent works full-time
and $6.06 an hour for a family of three. If food stamps cut-off at a self-
sufficiency wage the cliff effect associated with food stamps cut-off would be
mitigated. However, it would still exist because of the considerable amount
of deductions permitted. For instance, a standard deduction and other
deductions for child care, shelter, and earned income rarely allow food stamp
allotment to approach zero. The positive attribute to these deductions is they
form a linkage with other public assistance programs. The shelter deduction
is smaller for those receiving housing assistance. The child care deduction is
smaller for those receiving Title XX child care subsidy. Therefore, housing
and child care subsidies reduce the food stamps allotment and prevent a
household from receiving the maximum public assistance available for each
program. On the negative side, if all programs were tied to one and another
it would be extremely difficult to develop a combination of public assistance
programs with consistent economic incentives.
Most importantly, tying together programs that have income eligibility
requirements based on official poverty levels maintain families at an income
near the eligibility requirement. Since the eligibility requirement is the
poverty level poverty is encouraged, not self-sufficiency. If self-sufficiency is
the goal of public assistance programs, their income eligibility standards
should be associated with a self-sufficiency income.
Changes in Future Economic Incentives
The economic disincentives exposed in this study provide evidence for
additional reform. However, reform does not happen overnight, and there are
other factors which may affect implicit marginal tax rates in the near future.
This includes pending legislation which assists low-income families with child
care, and medical insurance. If this legislation is passed will it improve
economic incentives or make them worse? Additionally, if the JOBS
program is extremely effective it could place AFDC parents in jobs paying
high wages. In general, cliff effects disappear by $7.00 to $9.00 per hour.
What is the probability that JOBS will place parents at these high wages?
Thirdly, there are almost13 no economic disincentives associated with families
that have no child care costs. Are child care costs exaggerated in this study?
Fourthly, it was noted that larger child support payments will force economic
disincentives to occur at lower wages. How successful will the Family Support
Act be at increasing child support payments and what will they be? Finally,
inflation erodes the safety net; increases in the general price level alter where
cliff effects will occur. Will cliff effects occur at higher or lower wages as the
general price level increases?
This section attempts to answer all of these questions.
The effects of increased child support on the implicit marginal tax rate
should not be ignored. If the new welfare reform measures are efficacious
more AFDC families will be receiving more support.
Empirical research by Philip Robins (1985) indicates that child support
enforcement does increase child support receipts; it reduces AFDC program
costs. However, it does not have a dramatic impact on reducing welfare
dependency. This is consistent with the findings in the implicit tax model. An
increase of child support, like an increase of income, could cut-off families
from public assistance program sooner, thus yielding to a situation where the
implicit tax could be greater than 100 percent on child support. To
compensate for economic disincentives occurring earlier in the wage scale due
to increased child support an economically motivated parent could work at a
lower wage or decrease their hours.
Another empirical study by Andrea Beller and John Graham (1988)
also indicates that child support enforcement has increased child support
payments. However, they contend that inflation has eroded the dollar value
of these child support payments; thus real child support receipts declined from
1978 to 1983 when they conducted their study. This suggests that any child
support payment schedule must have an automatic inflationary adjustment in
There are two major questions concerning the new aggressive child
support enforcement. First, how much will these child support payments be?
Gan they really be enough to make a considerable difference between welfare
dependency and self-sufficiency? This question depends on how the dollar
amount of child support payments will be determined, and what income is
necessary to be self-sufficient. If the Colorado Child Support Guidelines are
strictly followed a non-working mother with one child would be ineligible for
AFDC through the child supports payments required by the childs absent
father if he earns more than $4.45 an hour working full-time.
Secondly, will absent parents be able to pay? A empirical study
conducted at the University of Michigan (Bergmann 1987) indicated that ten
percent of the children living with poor, single-parent families had absent
parents who were also in poverty. Only two percent of the absent parents
surveyed would become poor if they were forced to share their income with
their ex-spouse and children. The Family Support Act recognizes that some
absent parents are unemployed and underemployed; thus they have made
provisions for some demonstration programs which include the absent parent
in the JOBS program.
Inflation and Increases in the Price Level
Official poverty levels are used to determine food stamp and leap
eligibility. The official poverty level, defined by the United States Department
of Agricultural (USDA), is three times the Thrifty Food Plan (TFP). The
TFP is a market basket of food deemed necessary for short-term subsistence;
it varies in accordance with family size. The USDA updates the TFP
annually, thus it reflects increases in the general price level.
Housing assistance has similar automatic adjustors built into its
eligibility requirements. Families are eligible for housing assistance if their
income falls below fifty percent of the local, (county or metropolitan area)
median income. This is updated annually.
Potential AFDC and Medicaid recipients face another criterion: their
earned income combined with their received child support (excluding the
earned income deductions and $50 child support deduction) must not exceed
185 percent of the need level established by their state. Need levels are
derived from a normative view of what an adequate level of living is in a
particular state. AFDC grants cannot exceed this need level. For the state
of Colorado, this need level has not changed since 1979 when it was
established at $421 per month for a family of three (Augostine 1988). It is
determined from a market basket of goods. The Family Support Act will
require this to be updated every three years from now on.
No state has a need level equal to the poverty level; as a matter of
fact, on average states cut families off of AFDC and Medicaid when their
income becomes greater then fifty-three percent of the poverty level for a
Inflation erodes the safety net. Any effective public assistance
program must include features which automatically adjust eligibility standards,
grants, and assistance levels for increases in the general price level. If the
public assistance level continues as is economic disincentives15 associated with
AFDC and Medicaid cut-offs will be occurring at much lower wages.
Whereas, economic disincentives associated with food stamps, and housing
assistance will occur at much higher wages.
Success Depends on Earnings
The success of the new welfare reform program depends on the
success of the JOBS programs ability to place women in jobs paying self-
sufficiency wages or more. Jobs and wage earnings are factor outside the
realm of government control with the exception of the minimum wage. Wage
earnings are determined by the labor market. The labor market can be
divided into two sectors that of the primary labor market and that of the
secondary market. Janice Peterson summarizes the differences between the
primary and secondary labor force acutely in her article, "The Feminization
In the primary sector, jobs are characterized by good pay and
fringe benefits, job security, a high degree of unionization, and
good working conditions. In contrast, jobs in the secondary
sector are low paying, often seasonal or sporadic, and less
likely to be unionized. The majority of "womens jobs" are
concentrated in this secondary sector, offering little
opportunity for these women to "work their way out of
poverty." (Peterson 1987: 333)
As the United States becomes more of a service sector economy, more and
more jobs in the secondary labor market are created. According to
Workforce 2000 most of the new jobs, about 70 percent will be female-
dominated; furthermore, over 50 percent of them will pay below the poverty
level (Remington 1989). Employees of the secondary labor market will not
be able to earn a livable wage until the minimum wage is brought up and
society recognizes comparable worth.
However, Peterson points out that human capital theorists such as
Gary Becker have another explanation for womens lower earnings: many
women make a conscious decision to spend more of their time in home
responsibilities rather than investing their time obtaining education and job
training which is integral for higher wages. Beckers studies are disputed,
and the discussion as to whether women have lower wages due to the
primary/secondary labor market, or because of less training and job
experience is on-going.
Average wages for jobs placements of job training programs vary.
For Weld County, Colorados Diversion program it is $4.70 an hour, and
Boulder Countys is $4.71. For the E.T. program it is over $6.00 an hour.
Jefferson County, Colorado, one of the more wealthier counties in Colorado,
it is over $5.00. These wages are in the range where there are major cliff
effects associated with food stamps cut-offs. Furthermore, another cliff effect
occurs near $6.50 per hour due to child care subsidy cut-off. Parents earning
these average wages will have little economic incentives to increase their
earnings at these points.
On the whole, job training programs administrated in the last decade
have been successful16. Taxpayer dollars have been saved. The impact has
not always been tremendous. For example the difference between AFDC
payments for recipients in job training programs and AFDC payments for
recipients not in job training programs during a 15 month period in Baltimore
was $6.00 (Gueron 1987). Programs with heavy state financial involvement
such as Massachusetts Employment and Training (ET) have the best results.
The long-term savings cannot yet be calculated. Those who criticize job
training programs17 question their ability to place participants in jobs that
pay self-sufficiency wages. Bureaucrats have done little more to define self-
sufficiency as a family off public assistance. Yet, if a parent achieves a wage
between $4.70 and $6.00, which is the norm for job training graduates, they
will still be eligible for many public assistance programs such as child care
subsidies. A clear, dollar definition of self-sufficiency is in order. Studies by