Equity with Employment*
Robert H. Haveman
Over the past decade, the economies of western, industrialized countries
have moved along two distinct tracks, neither of which offers a compelling
model for current public policy. The economies of North America, the United
Kingdom, and Australia have been creating lots of jobs, but suffering from
wage stagnation and growing inequality; economies in Western Europe have featured
growing wages and more modest income gaps, but have been far less successful
at job generation. Must we simply decide between North American wage stagnation
and Western European double-digit unemployment? Or, might there be a third
way that combines jobs with decent compensation for workers, low poverty rates
with fiscal control--equity with employment? These are arguably the fundamental
questions of contemporary social and economic policy.1
I think better performance is possible, and will sketch a simple, stylized
strategy--I call it an "Employment-Centered Social Policy"--that will help
overcome current trade-offs. Focused on redirecting welfare state programs
in both Western Europe and North America, it brings together a variety of
current reform proposals, each of which addresses only a single aspect of
the problem. If successful, this new, multi-pronged strategy would--as the
name indicates--get social policy and labor markets working in tandem, rather
than at cross purposes: Social policy would protect people from destitution
without undermining incentives to work; labor markets would generate jobs,
with augmented compensation for the most disadvantaged workers; and work would
be the principal source of income.
Before getting to the strategy, though, I need to explain my focus on the
welfare state. If the objective is to move economic performance onto a better
track, why not concentrate on trade, technology, education, regulation, or
fiscal policy? The reason, in brief, is that the two current tracks of economic
performance are associated with two distinct structures of welfare state,
and these structural differences go a long way to explaining the associated
disparities in performance: Western European countries, with the most accessible
and generous welfare state systems, also tend to have the slowest employment
growth and most serious poverty/joblessness traps, while more tight-fisted
North American welfare state systems have more rapid job growth, but also
rising inequality largely caused by eroding wages especially among less-skilled
To make the case for the employment-centered strategy, then, I start by outlining
the basic components of contemporary welfare states, describing national differences
in generosity and accessibility of support from income protection and labor
market programs, and showing some of the connections between social policy
and economic performance. Once the problem is more sharply in focus, I will
outline my proposed alternative, note some likely hurdles, and suggest an
approach that may reduce resistance to large-scale change.
All larger and more industrialized western nations now have complex and financially
costly systems of income protection in place. Despite substantial variations
across countries, these systems share certain basic elements:2
1. An unemployment benefit program replacing 40-80 percent of the
earnings lost when a worker involuntarily loses a job. In many countries,
unemployment benefits well in excess of a year are available; in others, the
expiration of benefits triggers eligibility for either a new round of eligibility
or early retirement or disability benefits.
2. A disability benefits program providing long-duration replacement
of 50-90 percent of lost earnings (at the median) to handicapped workers.
3. Provision for early retirement from the work force, in several
countries at a replacement rate of up to 70 percent after age 60.
4. A safety-net welfare system with benefit levels that assure recipients
of an income level from roughly one-third to over one-half of the median income
level in the country. These benefits range widely across the countries, and
in some cases vary significantly across various regions within a country.
5. A legal minimum wage policy (or bargaining arrangements) applying
to employees irrespective of education, skill, training, or productivity.
6. Employment regulations that constrain the ability of employers
to alter the size of their work force in response to changes in the demand
for their output, resulting in caution in adding permanent workers to the
enterprise and disguised unemployment in periods or places of slack demand.
As this brief enumeration indicates, social policies vary considerably in
breadth of coverage, generosity, and accessibility. Moreover, these variations
influence economic performance. For example, a comprehensive and generous
social safety net reduces income poverty, moderates the wage disparities resulting
from the unfettered markets, and protects workers from income losses due to
illness, disability, unemployment, retirement, and family breakups. Moreover,
employment protection fosters long-term employer-employee relationships (and
associated training-productivity benefits), as well as job search and efficient
But these same policies also have an important downside. They reduce the
demand for labor--especially low-wage, low-skilled labor--and encourage employers
to substitute temporary for permanent workers, thereby reducing the level
of job training that firms offer to employees. Moreover, recipients (or potential
recipients) of long-duration and generous income support benefits have reduced
incentives to work or even search for a job, and often end up caught in what
are described as poverty or joblessness "traps." The financing of such systems
also increases firm costs, which in open economies reduces foreign demand
for goods and services. These arrangements also create rigidities in the labor
market (keeping the demand for labor from expanding to its true potential),
and erode the incentive for diligence and effort for current job holders as
both possibility of job loss and the cost of such loss if it occurs is reduced).3 And these troubles are even more serious when income protection
policies are--as is often the case--loosely structured, poorly integrated,
As I stated earlier, the contrast in economic performance--that is unemployment,
wage growth, and inequality--among developed countries, especially between
North America and Europe, appears to be systematically related to differences
in such labor market and social protection policies among these nations. While
relative demand for low-skilled workers has fallen in all nations, policy
responses in these two regions suggest a trade-off between two packages of
policies and consequent performance: On the policy side, the North American
package consists of low minimum wages, a relatively modest income protection
system, and few institutional barriers to hiring, firing, and geographic mobility.
And it is accompanied by rapid employment growth, especially for low-wage
workers, high and rising wage inequality, and slow wage growth. In contrast,
a European-style policy package comprises generous and accessible social benefit
programs, high minimum wage levels, and relatively stringent labor market
regulations and constraints. It is accompanied by high unemployment and joblessness,
slow employment growth, low and steady wage inequality, and relatively high
This simple contrast may overdramatize the linkages between the structure
of social protection measures and economic performance. Still, the data underscore
the force of the connections. Relative to other industrialized regions, North
America has been a "job creation machine." Since 1960, North American employment
has doubled, while jobs in the European Community have increased by less than
20 percent (see fig. 1). Over the 1980s, the number of new jobs created in
North America and Australia was more than 20 percent of 1979 employment levels.
The numbers for the remainder of major country groupings range from essentially
zero to about 10 percent.
This differential pattern of job growth is also reflected in official unemployment
rates among countries (those rates count workers as unemployed only if they
are actively looking for work but cannot find it). In North America, the standardized
unemployment rate from 1990-93 was 6.9 percent, which is about 110 percent
of its 1960-90 level. In the European Community, the 1990-93 unemployment
rate of 9.2 percent was equal to its value in the 1980s, but more than 300
percent of its level during the 1960s and 1970s.
The story is much the same if we look at joblessness (people count as jobless
if they are not working, whether or not they are looking for work). Consider
males aged 25-54--a group that excludes those who have chosen early retirement
or continued schooling in the face of poor labor market prospects. From the
1970s to the 1990s, the jobless rate in this group in the United States increased
by about one-third--from about 9 percent to about 12 percent. In the European
Community, that rate has doubled over the same period, increasing from about
7 percent to 14 percent.
These differential changes in patterns of unemployment and joblessness appear
to be related to changes in the generosity and accessibility of public income
support programs. Take the case of unemployment benefits. Between 1961 and
1991, 6 of 12 European Community countries increased the maximum period for
which unemployment benefits are available; there was no increase in North
America. A composite index of overall unemployment benefit generosity--taking
account of both benefit levels and accessibility--provides an even crisper
comparison. In 1961, the simple average of this generosity index in the European
Community countries was 17.4; by 1991, it had climbed to 31.5--an 80 percent
increase. In the Scandinavian countries (plus Austria and Switzerland), this
same index increased more than four-fold, from 8.3 to 34.5. In contrast, the
simple average indicator for North America stood at 19.5 in both 1961 and
1991, while for the United States alone, it fell from 17 to 11.
Other income support programs--for example, early retirement and disability
benefits--also enable working-age individuals to opt for nonwork/recipiency
status. Consider, then, the "replacement rate" for disability benefit programs.
(The replacement rate is the percentage of previous income that is paid to
disabled workers: a 50 percent replacement rate means that disabled workers
get half their previous salary.) Between 1974 and 1993, the simple average
of replacement rates in European Community disability benefit programs increased
slightly, from the mid-40s to the upper 40s. For the Scandinavian countries
plus Austria and Switzerland, the 1981 rate was a very high 66 percent; by
1993, it had fallen slightly to 61 percent. For North America, the replacement
rate increased from the mid-20s in 1974 to about 30 percent in 1993. These
differences in the absolute level of the replacement rate are notable and
appear to be behaviorally important: in OECD countries the generosity of disability
benefits as measured by the replacement rate substantially predicts the percentage
of those aged over 55 who receive those benefits.5
With minimum wages the picture is much the same. European Community countries
with legislated minimum wages set this value much higher than in North America.
For example, in the early-1990s the minimum wage in the Netherlands and France
stood at over 50 percent of the average wage, while in the North American
countries it was roughly 35 percent of the average wage.
These patterns in income support, unemployment, and joblessness are striking,
but not shocking. One would expect that countries with the highest social
protection and minimum wage levels would also face the strongest pressures
to substitute capital and productivity-improving technology for labor. Such
substitution would in turn improve productivity and lead to more rapid wage
growth in the high benefit-high unemployment countries of the European Community
than in North America, with predictable results for inequality and poverty.
Here, again, the pattern is visible in the statistics: In the early post-war
period in the United States, the real wage of the average worker grew roughly
2 percent per year; since the early 1970s, the average worker's real hourly
earnings have been virtually constant. The basic cause of the sharp decline
in average earnings growth is the fall in the rate of productivity growth,
to which gains in average compensation are ultimately tied.6
Since 1970, the 10 percent real wage growth of the United States contrasts
with real wage growth of nearly 60 percent in the European Community; conversely,
over the same period the nearly 60 percent employment increase in the United
States compares with a small 10-15 percent increase in the European Community.
Most high-income Western economies have experienced growing labor market
inequality since the late-1970s. Once more, however, the variations are striking.
In North America, Australia, Japan, and the United Kingdom, the relative wage
of low-skilled workers fell by 10-25 percent during the 1980s (see fig. 2).
In contrast, the relative wage of low-skilled workers in most continental
European countries was roughly the same in 1980 and 1990, probably reflecting
the higher relative minimum wage in these countries established either through
legislation or bargaining.
The high and accessible level of income protection benefits and high minimum
wages in Europe have successfully maintained a relatively low level of family
income inequality, even in the face of rising unemployment and joblessness.
Consider the simple average Gini coefficient, a standard measure of income
dispersion: In North America (Canada and the United States) in the late 1980s,
it was 31.5; for the European Community countries (excluding Denmark, Spain,
Portugal, and Greece) it was 27.9; for the EFTA group (Scandinavia plus Austria
and Switzerland) the level stood at 22.0.
Poverty rates among the high-income economies also vary widely. North America
and Australia all have poverty rates well above the average, while Austria,
Belgium, Germany, Luxembourg, and the Netherlands have the lowest rates. Among
the European countries, France, Ireland, Italy, Sweden and the United Kingdom
have rates close to the average, with France and Sweden falling at the low
end of this group.7
Evidence on changes in poverty rates is not available for most of these countries.
For the United States, however, official income poverty has risen substantially
since the early 1970s, from an already high level. At the beginning of the
1970s, the rate of poverty defined by the United States government stood at
about 11 percent; by 1992, over 14.5 percent of the nation's population lived
in poverty. Between 1973 and 1992, the number of people officially classified
as poor increased from 23 million to 37 million persons. This pattern is consistent
with the growing earnings inequality of the past two decades, and again reflects
the lack of generosity and accessibility of income protection programs, the
erosion of market wage rates for low-skilled workers, and the relatively low
In sum, then, wage rates in Europe have been maintained by a variety of measures,
including policies and collective bargaining arrangements that result in a
high effective minimum wage (as a percent of the average wage). Moreover,
income protection policies--unemployment benefits, disability benefits, early
retirement policies, and welfare benefits--are more accessible and generous
in Europe than in North America. As a result, the declining relative European
demand for new labor market entrants and other lower-skilled workers has been
revealed in high unemployment and joblessness rates, and in the increasing
prevalence of long-term unemployment. Given the pattern of labor demand growth,
the structure of income protection programs in the European countries has
tended to create a poverty or joblessness "trap" for workers with low potential
In North America, a lower minimum wage, less generous and accessible social
benefits, and, in general, lower hurdles to job creation have encouraged job
growth for low-skilled workers. The overall unemployment rate has not shown
an upward trend, and a high percentage of all workers occupy minimum wage,
low-earnings jobs. Wage rates and earnings have fallen substantially for low-skilled,
low-education, young, and minority workers, and as a result wage rate and
earnings inequality has grown substantially. Low minimum wages and weak social
protection, among other factors, have led to a more than 30 percent jump in
the US poverty rate since the 1970s, and a more than 60 percent increase in
the number of poor people.
Finally, in nearly all countries, unemployment and joblessness has become
increasingly concentrated among youths and other low-skilled workers. While
the incidence of joblessness and part-time work (especially for younger and
minority male workers) has increased substantially in North America since
1973, both of these indicators have increased more in the European Community.
A Third Way?
Is there, then, an alternative approach to social protection that can simultaneously
address the European dilemma--high aggregate unemployment, growing joblessness,
and poverty traps in a system with generous, accessible, and in some cases
ill-managed income protection programs--and the North American dilemma--stagnant
wages, high rates of nonemployment among low-skilled workers, and increasing
inequality and poverty in a system with relatively modest income protection
I suggest a common strategy for addressing both kinds of labor market problems.
In brief, it would aim to provide a minimum income floor under all individuals
and families; eliminate the serious work disincentives implicit in existing
social policies, while creating positive incentives for increases in the labor
supply of low-skilled workers; and stimulate demand for the services of low-skilled
workers. Stated more fully, the objectives would be to:
1. Guarantee income support, but at a level that will not inhibit the willingness
of people to work at prevailing market wages.
2. Increase employment opportunities for low-skilled labor and stimulate
the willingness of workers to accept work at prevailing wages.
3. Increase the monitoring of recipients in terms of unreported work and
job search efforts, and improve aspects of program integration and administration.
A wide variety of individual measures now under discussion would advance
these objectives: from imposing more stringent eligibility conditions and
less generous benefits, to replacing the existing constellation of often ill-integrated
measures with a costly, universal, high-level Basic Income Guarantee (BIG)
program. Other proposals would replace categorical and difficult-to-administer
systems with a Negative Income Tax (NIT) and Credit Income Tax (CIT) programs;
or increase the earned income tax credit or other programs that supplement
earnings; or use wage rate and employment subsidies to improve both the demand
and supply sides of the low-wage labor market.8
Each of these individual measures--often in conjunction with drastic surgery
on existing programs--could contribute to more effective social policy, particularly
in high-benefit welfare states. But no single measure will accomplish all
three objectives set out above. A CIT or an NIT could provide a decent minimum
income floor, and reduce the serious disincentives to work implicit in existing
programs with relatively high benefit levels and few restraints on duration.
But these policies would be unlikely to do much for the labor market to induce
a sizable increase in overall employment levels or in the supply of or the
demand for low-skilled workers. Conversely, the work-conditioned policies--earnings
supplements, wage rate subsidies, and employer-based marginal employment subsidies--are
capable of increasing the returns to labor supply and work, increasing the
effective demand for the services of low-skilled workers, and contributing
to the reduction of poverty through increasing the rewards from working. But
they will not provide a minimum income floor that ensures against destitution.
These observations suggest that what is needed is a judicious combination
of a moderate minimum income guarantee, programs designed to increase both
the supply of and demand for low-skill labor, and measures to tighten program
monitoring and administration. The minimum income guarantee would eliminate
economic destitution, but not be so high as to seriously diminish the attractiveness
of earnings from work. Labor market policies stimulating demand for and supply
of low-skilled labor would increase the employment and earnings of this high
joblessness/high unemployment group. And the increased stringency of program
administration would seek to ensure that those able to work would remain in
the labor market.
Consider the following Employment-Centered Social Policy strategy, designed
to ensure a decent social minimum and a high-employment and high-growth economy,
and relevant to countries on both tracks.9 I emphasize that the aim of the proposal is to stimulate
discussion leading to policy changes with this goal, and not to argue the
superiority of this prototypical approach relative to alternatives with a
1. A two-pronged employment subsidy program for low-skilled workers.
This component of the proposal rests on the view that most advanced countries
face a situation where minorities, youths, older workers, disabled workers,
and single mothers--all characterized by low levels of skill and education--face
relatively bleak labor market opportunities. The sources of these poor prospects
are complex, and involve a variety of constraints on labor market flexibility.
Such constraints involve prevailing wage standards (including minimum wages),
union wage contracts, and the fringe benefits and payroll taxes that businesses
are required to pay for every standard worker. Because of these arrangements,
employers do not find hiring low-skilled workers a profitable proposition.
The program that I propose here is aimed both at disadvantaged workers and
those who hire them--at the supply and demand sides of the labor market. Its
effect would be to alter the terms on which workers could be hired: in effect,
it would make hiring low-skilled workers a more profitable and attractive
proposition than it is now. Both prongs of the policy are designed to offset
constraints on labor demand from market rigidities (including minimum wages),
to increase the employment of less-skilled workers, and to increase the returns
to them from the work that they do. In the process, business costs would tend
to fall, while output would tend to increase.10
The first component is an Employer-Based Marginal Employment Subsidy. It
would provide financial incentives to employers who hire low-skilled workers
over and above the numbers they would otherwise hire. Such a subsidy might,
for example, work as follows: The government would provide a tax credit (or
other financial subsidy) to any enterprise equal to 50 percent of the first
$10,000 of wages paid to the 50 workers hired in a firm above 102 percent
of the firm's previous year's employment. The numbers are of course somewhat
arbitrary: the essential point is that the subsidy would be marginal in nature,
affecting the decisions of firms regarding both the number of workers to hire
and the composition of those new hires.
While this arrangement does not directly target workers according to their
unemployment or poverty status, the subsidy is a higher percentage of the
wages of low-skilled than it is for more skilled workers. As a result, firms
have incentive to hire low-skilled workers, substituting them for higher skilled
employees or capital. By directly expanding the demand for low-wage workers,
this program will contribute to a reduction of poverty and unemployment.
The New Jobs Tax Credit that was in place in the United States during the
late-1970s could serve as a possible model for this program. Evaluations of
this program concluded that it was a potent and cost-effective measure to
increase employment of low-skilled workers. Analysts have suggested several
modifications to its structure that would increase its employment-generating
The second prong focuses on the low-skilled workers themselves. An employee-based
wage rate subsidy program would be instituted for low-skilled and hence low-wage
workers. Some portion of their wages would be subsidized by the government,
increasing their take-home pay, and hence providing an incentive to work more
if already employed, or to seek work if not yet working.
The first step in establishing such a program is to stipulate a target wage
rate through legislation. Then, a per hour subsidy equal to some percentage
(the subsidy rate) of the difference between the target wage rate and the
actual wage rate would be offered to all workers. Clearly, only low-wage workers
would have their wage rate subsidized. For example, if the target wage rate
were set at $8.00/hour, and the subsidy rate at 0.5, a worker earning $5.00/hour
would receive a subsidy of $1.50/hour ($1.50 is 0.5 of the difference between
the target wage rate of $8.00 per hour and the actual $5.00 per hour wage).
The "take-home" wage rate of that person would then be $6.50/hour.
A wage rate subsidy is but one example of what have become known as "in-work
benefits," all of which are characterized by the provision of support only
when accompanied by market work. Other examples include the Family Credit
in the United Kingdom, the Danish and Dutch subsidies for work in markets
that are intensive in employing low-skilled labor (e.g., home cleaning and
lawn and garden work), and the Earned Income Tax Credit (EITC) in the United
The EITC provides a supplement of about 26 percent to earnings below $8,000
per year for families with one child, and 30 percent for families with more
than one child. Beyond $8000, the size of the supplement remains constant
until earnings of $11,000, when it reduced to 16 to 17 percent. The credit
is totally phased out when earnings reach about $24,000. Clearly, the EITC
has strong work incentives for lower-income families--especially those experiencing
the `negative' marginal tax rate. But it reduces work incentives for the large
number of families whose earnings are in the range where the subsidy is being
phased out--$11,000 to $24,000. Offsetting this disincentive effect is the
positive work effect of the incentive to move from joblessness to employment.12
The new labor market environment generated by this two-pronged supply-side
and demand-side policy proposal will improve the employment prospects of disadvantaged
workers by generating ongoing job creation pressures at reasonable cost. By
targeting the additional employment on segments of the labor market with the
most severe unemployment problems and the greatest susceptibility to "poverty
trap" problems created by existing tax and transfer programs, employment and
output can be increased without significant inflationary pressure. This labor
market program will fundamentally alter the wage structure in private labor
markets, raising the take-home pay of low-skilled workers relative to those
with more secure positions in the labor market. It will reduce inequality
in employment and earnings in a way that encourages independence, work, and
2. A Credit Income Tax Program. In this program, a family's income
would be defined comprehensively, and a tax credit would be awarded to each
living unit (or taxpaying unit) according to how large it is and who lives
in it. This credit would effectively guarantee a modest minimum income to
families of perhaps two-thirds of the explicit or implicit poverty line in
the country. Families with no other means of support would receive the full
amount of the credit as a grant; those with low income would receive smaller
net payments. Better-off families receive no net payment, and would pay positive
taxes. This program would be integrated with the positive income tax, so as
to yield a smooth marginal tax rate pattern. (Implementing this CIT would
require changes in the current income tax structure, but I will not pursue
these changes here.)
This low-guarantee plan would eliminate "hard-core" poverty, as the bottom
tail of the nation's income distribution would be cut off. But the assured
minimum income would be a relatively low percentage of median income--indeed,
well below accepted benefit levels for unemployment, disability, and retirement
pension programs (and safety net welfare programs) in several European countries.13
A central gain from this component of the proposal is the substantial increase
in work incentives relative to those that now prevail in most countries with
developed welfare state systems. A universal CIT would also strip away the
complexity of current programs, and eliminate much of the stigma associated
with welfare. With a CIT integrated into the standard income tax structure,
incomes would be taxed and support provided in a simple, open, universal,
and just manner.
Reform proposals with such a CIT at their heart have been recently debated
in the United Kingdom. In his 1995 book Full Employment Regained, Nobel
laureate James Meade suggests such an arrangement.14 Roger Douglas , former Finance Minister of New Zealand,
has also included a Guaranteed Minimum Family Income in recent reform suggestions.15 Anthony Atkinson has proposed a basic income guarantee
for the nonelderly population, conditional on "participation" (defined to
include those who are working, sick though employed, unemployed though seeking
work, in education or training, and caring for young, elderly and disabled).16 Commenting on these proposals, Samuel Brittan states
that: "If people are to be priced into work, whether through free market forces
or direct intervention, there must be an adequate safety net for those whose
earnings in the market place do not provide a reasonable standard of living."17 The employment-centered social policy reform suggested
here rests on just such a sentiment.
3. Elimination of existing programs for disability income transfers, unemployment
compensation benefits, and welfare benefits (or scaling back the benefit
levels in these programs, and establishing them as supplements to the CIT
described above). The elimination or scaling back of these programs would
free up budgetary resources to support the new programs, and would reduce
the high work disincentives and the poverty/joblessness trap problems that
constrain employment and growth.
An employment-centered social policy would simultaneously provide income
support to poor families and encourage work effort and individual responsibility.
But, as with any new approach to policy, this proposal confronts several obstacles.
A first concern is that its income guarantee falls well below the implicit
income minimum provided by existing programs in countries with generous and
accessible benefits. Clearly, these high implicit income minimums could not
be sustained were a universal income support measure such as the CIT substituted
for categorical benefit programs. If the minimum income guaranteed under a
CIT (or NIT) were as high as even 50 percent of per capita median income,
then marginal tax rates imposed on all earnings would have to rise markedly
to keep the proposal budget neutral.18 While the poverty-unemployment trap would be reduced
for current benefit recipients, the higher marginal tax rates imposed on the
remainder of the population would be a serious disincentive to work. Such
a policy substitution seems economically unworkable.
Two options seem possible. First, countries seeking to attain employment-with-equity
objectives must recognize that the replacement rates and eligibility standards
now in force in some of their programs need to be scaled back if incentives
to retain or to seek work are to exist for lower-skilled workers. While this
is a bitter political pill to swallow, the "iron law" of income support policy
indicates that high guaranteed incomes and strong work incentives are incompatible
The second alternative is for governments to set the level of the universal
income guarantee in the CIT (or NIT) at a rather low level (say, one-third
of per capita median income), but then allow people included in restricted
categories--for example, the disabled or unemployed--to be eligible for supplemental
benefits in categorical programs. This two-tiered arrangement would establish
a modest income floor and maintain work incentives. However, workers receiving
benefits from programs in the special categories would not face the desirable
work incentives that confront the remainder of the population; for them, something
of a poverty trap would persist. To address these concerns, eligibility restrictions
on benefits from these programs would have to be tightly enforced, recipients
judged able to work would have to be monitored for unreported work and job
search activities, and program administration would have to be geared to returning
beneficiaries to the work force. Over time, the remaining unemployment/poverty
trap could be further eroded by incrementally reducing the replacement rates
in these programs, and simultaneously restricting eligibility criteria.
A second obstacle to the employment-based scheme is political. Individuals
now receiving benefits in the unemployment, disability, and early retirement
programs of countries with generous and easily accessible benefits--as well
as those who expect to be beneficiaries in the future--would be adversely
affected by the change. There is no easy way to assess the strength of this
potential hurdle, though efforts to retrench social benefit programs by reducing
replacement rates or tightening eligibility standards inevitably generate
heated opposition from those who reflect the interests of beneficiaries, or
who otherwise support the status quo. Road blockades in France by truck drivers
demanding full retirement benefits at age 55, and mass strikes in Germany
by workers who oppose modest reductions in very generous sick leave with pay
provisions, are but two recent illustrations.
Another obstacle to an employment-centered strategy rests on the argument
that generous and accessible systems (especially if tightly administered)
provide all citizens with insurance against downside risks to future income.
A corollary claim is that systems providing universal income support to all
citizens (including those in higher income categories) have strong and resilient
systems of social protection against poverty and inequality. While such arrangements
may, indeed, reduce both poverty and inequality, they come at the cost of
higher unemployment and reduced prosperity. Perhaps the only way to blunt
the force of such arguments for generous income transfers and public benefits
is to make as vivid as possible the employment and efficiency toll that such
systems and their incentives impose on the society as a whole. Such arguments,
unfortunately, tend not to be persuasive in the face of those private interests
that benefit from the generous systems in place. Here, then, may lie the ultimate
* This paper is drawn from my presentation at the OECD
Conference "Towards 2000: The New Social Policy Agenda" held November 12-13,
1996, in Paris.
1 While this North America/Western Europe contrast is overstated,
and neglects important distinctions within each group, it helps to sharpen
the basic contours of the problem.
2 I focus on benefit programs providing income support
to the working-age population, and do not discuss taxation or retirement programs.
3 See Richard Layard, Stephen Nickell, and Richard Jackman,
Unemployment: Macroeconomic Performance and the Labour Market (Oxford:
Oxford University Press, 1991), Assar Lindbeck, "The West European Employment
Problem," Seminar Paper No. 616, Institute for International Economic Studies
(August 1996), and Edmund S. Phelps, Structural Slumps: The Modern Equilibrium
Theory of Unemployment, Interest, and Assets (Cambridge, Mass.: Harvard
University Press, 1994). A helpful discussion of the implications of alternative
theoretical frameworks in analyzing the effects of the welfare state on economic
welfare is found in Anthony Atkinson, "The Welfare State and Economic Performance,"
National Tax Journal 48, 2 (1995): 171-98.
4 For example, unemployment benefit programs sometimes
fail to verify if job search activities are being pursued, or if job loss
was voluntary, or if beneficiaries are concealing employment. Other benefits
are sometimes entirely eliminated (taxed at 100 percent) once the beneficiary
has some wage income. Or others have, in effect, infinite duration because
beneficiaries are allowed to cycle between getting benefits and participating
in active labor market programs, or among various benefit programs. If the
incidence of such problems is as widespread as anecdotal evidence suggests,
the availability of income protection benefits can increase program costs,
increase measured unemployment rates beyond true rates of unemployment, promote
the demand for benefits and its resulting joblessness, and undermine the effectiveness
of active labor market policies. These examples, and numerous others, are
discussed in The OECD Jobs Study: Evidence and Explanations (Paris:
OECD, 1994), Chapter 8.
5 Data in this paragraph are from Sveinbjorn Blondal and
Mark Pearson, "Unemployment and other Non-Employment Benefits," Oxford
Review of Economic Policy 11, 1 (1995).
6 From 1948 to 1973, output per worker in the US business
sector increased nearly 2.9 percent per year; since 1973, the average gain
has been about 1 percent per year.
7 See Michael Forster, "Measurement of Low Incomes and
Poverty in a Perspective of International Comparisons," OECD Labour Market
and Social Policy Occasional Paper 14 (1994).
8 Elsewhere, I have discussed these various approaches
as potential reforms for existing welfare state systems, and assessed their
advantages and disadvantages. See Robert Haveman, "Reducing Poverty while
Increasing Employment: A Primer on Alternative Strategies, and a Blueprint,"
OECD Economic Studies (1996).
9 A recent Australian comprehensive reform package contains
many of the elements emphasized in the prototypical plan presented here. They
include: subsidies to youth, the long-term unemployed, and those at risk of
long-term unemployment for job search, employment and training; employer subsidies
for hiring/training people in these categories; reform of income support (unemployment
benefit) programs so as to provide incentive for job search and work; intensive
case management and a "compact" between workers and the public sector indicating
mutual obligations. See Commonwealth of Australia, Working Nation: Policies
and Programs (Australian Government Publishing Service, Canberra, 1994).
10 The case for such a labor market strategy is made in
Robert Haveman, Starting Even: An Equal Opportunity Program for the Nation's
New Poverty (New York: Simon and Schuster, 1994). See also the paper by
David Betson and John Bishop in Robert Haveman and John Palmer, eds., Jobs
for Disadvantaged Workers: The Economics of Employment Subsidies (Washington:
Brookings Institution, 1982), and Phelps, Structural Slumps.
11 For evidence on the effect of the New Jobs Tax Credit
on employment levels, see Robert Haveman and John Bishop, "Selective Employment
Subsidies: Can Okun's Law Be Repealed?" American Economic Review 69
(May 1979): 124-30; and Jeffrey M. Perloff and Michael Wachter, "The New Jobs
Tax Credit: An Evaluation of the 1977-78 Wage Subsidy Program." American
Economic Review 69 (May 1979): 173-79.
12 On the design and effects of the EITC see John Karl
Scholz, "The Earned Income Tax Credit: Participation, Compliance, and Antipoverty
Effectiveness," National Tax Journal 47 (March 1994): 59-81.
13 In countries with very generous and accessible benefit
systems, poverty rates would be increased if these programs were replaced
with only the universal CIT, even though "hard core" poverty would be eliminated.
In part, income losses by recipients of these generous benefits would be offset
by increased take-home pay associated with the employment subsidy component
of the proposal. Alternatively, supplemental categorical unemployment, disability,
and retirement benefits could be adopted (see below).
14 James Meade, Full Employment Regained (Cambridge:
Cambridge University Press, 1995).
15 Roger Douglas, Unfinished Business (Wellington:
New Zealand Random House, 1995).
16 Anthony Atkinson, "The Distribution of Income: Evidence,
Theories, and Policy," de Economist 144, 1 (1996): 1-21.
17 Samuel Brittan, "No, not the next Budget," Financial
Times, November 9, 1995.
18 See OECD, The OECD Jobs Study.
19 Viewed this way, the proposed approach would seem more
feasible for countries such as the United States with lower benefit levels
and replacement rates (relative to median income), than for those European
countries with higher benefits and replacement rates.