Lane Kenworthy, The Good Society
Most of what we call social policy (or the “welfare state”) is actually public insurance.1 Social Security and Medicare insure against the risk of having little or no money in your retirement years. Unemployment compensation insures against the risk of losing your job. Disability payment programs insure against the risk of suffering a physical, mental, or psychological condition that renders you unable to earn a living.
Other government services and benefits also are insurance programs, even if we don’t usually think of them as such. Public schools insure against the risk that private schools are unavailable, too expensive, or poor in quality. Special education services insure against the risk of having a disability that inhibits participation in school. Retraining and job-placement programs insure against the risk that market conditions make it difficult to find employment. The Earned Income Tax Credit (EITC) insures against the risk that your job pays less than what’s needed for a minimally decent standard of living. Social assistance programs such as the Supplemental Nutrition Assistance Program (SNAP, or “food stamps”) and Temporary Assistance for Needy Families (TANF) insure against the risk that you will find yourself unable to get a job but ineligible for unemployment or disability compensation. Even affirmative action programs are a form of insurance; they insure against the risk of being in a group that is, or formerly was, discriminated against.
In providing a backstop against risk, public insurance safeguards individuals and families who otherwise might find themselves in dire straits. That reduces anxiety, improving mental well-being and perhaps also physical health. Public insurance also boosts the living standards of the poor. It increases their income, and it provides them with services for which they bear relatively little of the cost.
But it’s possible that this boost to living standards is merely a near-term one. Critics charge that public social programs tend to hurt the poor in the long run by reducing employment and economic growth. Are they correct?
In addition to this important question, I consider several others. Does public insurance erode self-reliance? Is a large private safety net as helpful to the least well-off as a large public one? Are universal programs more effective than targeted ones? Does decentralized government authority result in less generous social policy? Are income transfers the key, or are services important too?
- Is public insurance good for the poor?
- Does public insurance turn us into moochers?
- Is a private safety net just as effective as a public one?
- Universalism or targeting?
- Centralization or decentralization?
- Are public services pro-poor?
IS PUBLIC INSURANCE GOOD FOR THE POOR?
Figure 1 shows average market income and disposable income for households on the bottom fifth of the income ladder in the United States. Market income is earnings plus other private cash income (gifts, alimony, capital gains). Disposable income adds government transfers received and subtracts taxes paid. On average, public transfers nearly double the income of households in the bottom fifth.2
But notice that average market income among the poor hasn’t increased over time. Some argue that this is because government social programs weaken work incentives, disrupt markets, and waste resources. In this view, public insurance helps in the short run, but it hinders progress for the poor in the long run by reducing employment and economic growth.3
Any program that allows people to survive without employment will cause some reduction in work effort. This is unavoidable. Even America’s social assistance benefits, though low by affluent-country standards, induce some adults who could work to choose not to.4 Similarly, any tax will cause some individuals or firms to cut back on investment. And any government program will feature some inefficiency and waste. But unless these effects are large, the damage is likely to be outweighed by public insurance’s benefits for the least well-off. Plainly there is some level of generosity at which public insurance will do more harm than good. Have existing rich nations passed this tipping point?5
In a few instances they have. The Netherlands’ disability program in the early 1980s created strong incentives for people to drop out of the labor force. In Denmark, prior to the mid-1990s there was no de facto limit on the duration of eligibility for unemployment compensation. In Sweden, sickness insurance was the chief culprit, as Jonas Agell details: “According to the rules in place by the end of the 1980s, employees were entitled to a 90% compensation level from the first day of reporting sick. Due to supplementary insurance agreements in the labour market, however, many employees had a compensation level of 100%. For the first seven days of sickness leave, a physician’s certificate was not required. If individuals ever respond to economic incentives, work absenteeism ought to have been widespread in Sweden. The increase in the average number of sickness days per insured employee from 13 days in 1963 to 25 days in 1988 can hardly be attributed to a deteriorating health status of the population.”6
But each of these problems was recognized and eventually corrected. And a number of countries, led by the Nordics, have introduced and expanded programs such as childcare and retraining that encourage employment.7 Also, by limiting the risk of personal financial catastrophe, public insurance programs facilitate entrepreneurship and mobility.8
What does the comparative experience tell us about the effect of public insurance generosity on economic output and employment? Figure 2 considers output. On the horizontal axis of both charts is the share of GDP spent on public social programs.9 The measure is adjusted for the share of the population that is elderly and the share that is unemployed, as expenditures will automatically be greater when either of these is higher, irrespective of the generosity of the programs. In the first chart we see no association between countries’ level of GDP per capita and the generosity of public insurance, and in the second we see a similar lack of association for change in per capita GDP (economic growth).10
Figure 3 does the same for employment, measured as the share of the working-age population that is employed. Once again we see no indication that public insurance generosity has had a damaging effect.11
Note also that the employment rate increased in nearly all of the countries during this period. On average, it rose by nine percentage points between 1979 and 2013. That’s not what we would expect to see if generous public insurance programs were inducing large numbers of able adults to withdraw from the labor market.12
What do we see when we look across countries at a measure that gets directly at the living standards of the least well-off? The vertical axis of figure 4 has a widely-used indicator: the relative poverty rate. This is calculated, for each country, as the share of people living in households with an income below 60% of the country’s median income. On the horizontal axis is public insurance expenditures as a share of GDP. There is a strong correlation; countries with a bigger welfare state tend to have lower relative poverty rates.
Some think the living standards of the least well-off are best assessed via an absolute measure, rather than a relative one. The vertical axis in the next chart, figure 5, shows the income of a household at the tenth percentile of the distribution (90% of households have larger incomes, and 10% have smaller ones). The incomes are adjusted for cost-of-living differences across countries. The incomes of low-end households tend to be higher in nations with more expansive and generous social programs.
A more direct indicator of material well-being is people’s responses to questions about their living conditions. Since 2007, the Gallup World Poll has asked a representative sample of adults in each country whether there has been a time in the past year when they didn’t have enough money to (a) buy food that they or their family needed or (b) provide adequate shelter or housing. On the vertical axis in figure 6 is the average share of households responding yes to these two questions. The share ranges from 5% in Denmark to 15% in the United States and 20% in South Korea. Here too we see a tendency for countries with a bigger welfare state to do better.
So public insurance programs do tend to help the poor. But there is more to the story. Recall from figure 1 that average market income among the America’s least well-off hasn’t increased since the late 1970s. It turns out that the US isn’t alone in this. Economies in the world’s affluent countries have experienced profound changes since the 1970s. With globalization, the advance of computers and robots, increased pressure from shareholders for short-run profit maximization, union weakening, and other shifts, wages have been under pressure. Couple this with the fact that many people at the low end of the income ladder have labor market disadvantages — disability, family constraint, geographic vulnerability to structural unemployment — and we have a recipe for stagnation in the market incomes of the poor.
In most of these countries that is exactly what’s happened. Consequently, the chief source of income gain for the least well-off has been increases in government transfers.13
This helps us to answer a question frequently raised by welfare state critics: If public insurance programs help the poor, why has there been so little decrease in poverty in the United States since the 1970s? The reason is that the market incomes of low-end households have been flat, due to wage and employment stagnation, and in the US, unlike in many other affluent nations, income transfers to the working-aged poor haven’t increased much. The result has been very little progress in boosting the incomes of the least well-off.14
DOES PUBLIC INSURANCE TURN US INTO MOOCHERS?
Set aside social policy’s economic impact for a moment. Another prominent concern is that it has troubling moral consequences — specifically, that it breeds dependency. Mitt Romney, the 2012 Republican presidential nominee, famously quipped that there are “47 percent of Americans … who are dependent upon government, who believe that they are victims, who believe that government has a responsibility to care for them, who believe that they are entitled to health care, to food, to housing, to you name it — that that’s an entitlement and the government should give it to them.”15
Nicholas Eberstadt makes a more detailed case for this sentiment in his book A Nation of Takers.16 Eberstadt notes that over the past half century, the share of Americans who receive a government cash transfer and/or public health insurance — Social Security, Medicare, Medicaid, unemployment compensation, and so on — has grown steadily. The United States, he concludes, is now “on the verge of a symbolic threshold: the point at which more than half of all American households receive, and accept, transfer benefits from the government.” According to Eberstadt, growing reliance on government for help is undermining Americans’ “fierce and principled independence,” our “proud self-reliance.”
Is this a genuine cause for concern? In Eberstadt’s view, people are either givers or takers — taxpayers or benefit recipients. But that’s mistaken. Every adult who doesn’t live entirely off the grid pays some taxes: anyone who is an employee pays payroll taxes, and anyone who purchases things at a store pays sales taxes. Likewise, virtually every American receives benefits from government, whether it’s attending a public school, driving on roads, drinking tap water or taking a shower in your dwelling, not being stricken by polio, not having a band of thugs remove you from your home at gunpoint, visiting a park or lounging on a beach or hiking a mountain trail, using the internet, and on and on. And nearly all Americans receive a government cash or near-cash transfer at some point in their life: more than half of families with children receive the EITC; about 70% of us get income from SNAP (food stamps), TANF, SSI, or unemployment insurance; and more than 90% of the elderly get a Social Security check.17
It also isn’t clear that receiving a cash transfer from government renders people less self-reliant than receiving the myriad public goods, services, and tax breaks that government provides. Once upon a time, individuals and privately organized militias ensured the public safety. Then we shifted to government police forces and armies. At one point humans got their own water and disposed of their own waste. Then we created public water and sewage systems. The education of children was once a family responsibility. Then we created public schools. There’s a good reason for these shifts: government provision offers economies of scale and scope, which reduces the cost of a good or service and thereby makes it available to many people who couldn’t or wouldn’t get it on their own. Did Americans’ character or spirit diminish when these changes occurred? Is there something different about the more recent shift from individual to government responsibility in how we deal with retirement saving, healthcare, unemployment, and other risks?
Government provides more insurance now than it used to. All of us, not just some, are dependent on it. And life for almost everyone is better because of it.18
IS A PRIVATE SAFETY NET JUST AS EFFECTIVE AS A PUBLIC ONE?
Spending on social programs doesn’t have to be done by the government. In fact, total social spending is greater in the United States than in social democratic archetypes Denmark and Sweden, because even though its public programs are far less generous, the US has a lot more private spending on transfers and social services.19 But America’s large private safety net tends to be less helpful to the poor than a public safety net.20
How can it be that social expenditures are larger in America than in Denmark and Sweden? The standard measure is gross public social expenditures as a share of GDP. The first row in figure 7 shows that on this measure Denmark and Sweden are much higher than the United States. But this leaves out some important things. Private social expenditures, such as those on employment-based health insurance and pensions, are greater in the United States. Also, the US government distributes more social benefits in the form of tax reductions than do Denmark and Sweden, those two countries tax back a larger portion of public transfers than the United States does, and America’s per capita GDP is larger than Denmark’s or Sweden’s.
If we shift to net (rather than gross) public and private (rather than public alone) expenditures per person (rather than as a percentage of GDP), we get a different picture. According to the calculations of OECD researchers Willem Adema and Maxime Ladaique, by this measure the United States is the biggest spender of the three.21 These numbers are in the second row in figure 7.
This ought to be good news for America’s poor. Unfortunately, it isn’t. Private social spending accounts for roughly two-fifths of the US social expenditure total shown in row 2 of figure 7. It consists mainly of employer health insurance and pension contributions. These expenditures are encouraged by government tax advantages.22 But they do little to help people on the bottom of the ladder, who often work for employers that don’t provide retirement or health benefits. Another version of the private safety net approach is tax-advantaged individual accounts, such as individual retirement accounts (IRAs) and health savings accounts (HSAs). These rely heavily on individual capacity and initiative to contribute, so the poor end up with inadequate protection.23
What about tax clawbacks? Public transfer programs in Denmark and Sweden tend to be “universal” in design: a large share of the population is eligible for the benefit. While this boosts public support, it makes the programs very expensive. To make them more affordable, the government claws back some of the benefit by taxing it as though it were regular income. All countries do this, including the United States, but the Nordic countries do it more extensively. Does that hurt their poor? Not much. The tax rates increase with household income, so much of the tax clawback hits middle- and upper-income households.
So how well-off are the poor in the United States, with its private welfare state, relative to their counterparts in Denmark and Sweden? One measure is average posttransfer-posttax income among households in the bottom income decile. The third row in figure 7 shows my calculations using the best available comparative data, from the Luxembourg Income Study (LIS). There is a sizable difference, not in America’s favor.24
This cross-country difference in the incomes of low-end households is a function of government transfers. I’ve calculated averages among households in the bottom income decile for the three chief sources of household income: earnings, net government transfers (transfers received minus taxes paid), and “other” income (money from family or friends, alimony, etc.). Average earnings are virtually identical across the three countries. The same is true for “other” income. Where bottom-decile Danish and Swedish households fare much better than their American counterparts is in net government transfers, as shown in the fourth row of figure 7.
Not only are incomes in the bottom decile higher in Denmark and Sweden; they also have increased more rapidly over the past generation.25 In those two countries net government transfers have risen more or less in line with economic growth. Not so in the United States.
Another difference is that public services such as schooling, childcare, medical care, housing, and transportation are more plentiful and of better quality for the poor in the Nordic countries. Public services reduce deprivation and free up income to be spent on other needs. It’s difficult to measure the impact of services on living standards, but one indirect way is to look at indicators of material deprivation, such as the measure used in figure 6 above. Material deprivation rates for Denmark, Sweden, and the US are shown in the fifth row of figure 7. The gap between the countries in material deprivation is larger than the gap in low-end incomes, which is what we would expect to see if public services help the poor more in the Nordic countries than in the United States.
So while the United States spends more money on social protection than is often thought, that spending doesn’t do nearly as much to help America’s poor as we might like.
The private safety net model has another important weakness: it fits poorly with employers’ need for flexibility and workers’ need for mobility. Tying a person’s health insurance and pension to a job doesn’t make much sense in a modern economy.26 Either is fine as a supplement, but people’s main health insurance and retirement pension should be independent of their employer.
UNIVERSALISM OR TARGETING?
Generous public insurance is threatened by population aging, modest productivity growth, downward pressure on tax rates stemming from capital mobility, and, in some countries, persistent budget deficits.27 One possible response is to make greater use of targeting in social policy.28 Targeted transfers are directed (sometimes disproportionately, sometimes exclusively) to those with low incomes and assets, whereas universal transfers are provided to most or all citizens. Targeted programs are more efficient at reducing poverty; each dollar or euro or kroner transferred is more likely to go to the least well-off. Increased targeting therefore could be an effective way to maintain or enhance public insurance in the face of diminished resources.
But targeted programs tend to have political constituencies that are smaller and less cohesive, engaged, and influential, so targeted programs may be less generous than universal programs, which have their broader political base.29 Consequently, while the needy get a smaller proportion of the transfers in universal programs, they may get a larger quantity of transfers from such programs than from targeted ones.30 This led Walter Korpi and Joakim Palme to posit a paradox of redistribution: “the more we target benefits to the poor … the less likely we are to reduce poverty and inequality.”31
Is this correct? Do nations that rely more heavily on targeting achieve less redistribution? Korpi and Palme found that to be the case in the 1980s. Their measure of targeting-universalism is an index of concentration; it ranges from -1 if the poorest household gets all of the government transfer income (targeted) to 0 if all households get an equal amount of transfer income (universal). Their measure of redistribution is the percentage difference between pretransfer-pretax and posttransfer-posttax income inequality. Korpi and Palme found that the pattern across eleven affluent nations supported the hypothesis that greater use of targeting in transfers yields less redistribution.32
Subsequent studies have found that this correlation no longer holds.33 As figure 8 shows, as of the mid-2000s, countries that rely more heavily on targeting tend to achieve just as much redistribution as those with more universalistic policies.34
What has changed? Recall that there are two steps in the hypothesized causal process. First, universalism is thought to increase the size of the redistributive budget. Second, larger redistributive budgets are said to increase redistribution. The quantity of government social expenditures, a measure of the size of the redistributive budget, remains a major determinant of how much redistribution takes place, but the universalism of transfer programs no longer seems to have much impact on the quantity of government social expenditures.35
What if we look over time within countries? All of the rich countries have faced pressure to reduce social policy generosity over the past several decades, due to economic globalization and to changes in the balance of power between unions and left parties on one side and employers and right parties on the other. If universalism is better for redistribution, nations with more universal social policy should have fared better in resisting this pressure for cutbacks.
It turns out, though, that they haven’t. An early skeptical assessment came from Robert Greenstein, who examined the pattern of attempted cuts and successful cuts to targeted programs by the Reagan administration in the United States in the 1980s.36 Greenstein concluded that these programs fared surprisingly well. Paul Pierson reached a similar conclusion in an analysis of social policy developments under the Reagan administration in the United States and the Thatcher government in the United Kingdom.37 These are cases in which, according to the conventional view, we have might have expected significant cutbacks. Christopher Howard has updated the US story through the mid-2000s, and his conclusion echoes those of Greenstein and Pierson.38 A comparative analysis of eighteen rich countries by Kenneth Nelson finds little difference between the trajectories of means-tested benefits (mainly social assistance) and social insurance benefits (old-age pensions, unemployment insurance, and sickness insurance) during the 1990s and early 2000s.39 And my own cross-country analysis yields a similar conclusion; nations with greater targeting haven’t experienced larger declines (or smaller increases) in redistribution in recent decades.40
The hypothesis that targeting in social policy reduces political support and thereby lessens redistributive effort is a sensible one. Yet the experience of the rich countries in recent decades suggests little support for it. Countries with more universalistic social policy don’t (any longer) tend to be more redistributive. Nor do we observe a systematic tendency for universal programs to grow and targeted programs to shrink over time.41 Targeting has drawbacks relative to universalism: more stigma for recipients, lower take-up rates, and possibly less social trust.42 But targeting is less expensive. As pressures to contain government expenditures mount, policy makers may therefore turn to greater use of targeting. That may not be a bad thing.
CENTRALIZATION OR DECENTRALIZATION?
Subnational governments sometimes have considerable authority in rule setting, administration, and/or financing of public insurance. This authority can apply to the conditions for benefit eligibility, the implementation of those conditions (outreach efforts, ease of application), the benefit level, and the duration of benefit receipt.
Where subnational authority is greater, we would expect more inequality in public insurance provision across regions or localities, due to differences in affluence and in policy preferences.43 Massachusetts is much richer, and much more progressive, than Mississippi. That is indeed what we observe. Sarah Bruch, Marcia Meyers, and Janet Gornick examined 10 American safety net programs that differ in the degree to which subnational governments have authority over rule setting, administration, and/or financing. They found that policies with more state government authority tend to vary more across the states: “The weaker the federal role, the further apart are the states with respect to both the share of the needy they help and the level of assistance they provide.”44
Is decentralization bad for the poor? Not necessarily. For instance, if rule making is centralized and national policy makers worry that a generous benefit level will create problematic incentives in low-income regions, they may set the benefit at a level that is quite low for recipients in richer, more expensive locations. In this situation, it might be better for the poor to have a national floor with local governments allowed to go higher than the floor rather than to have a uniform policy. Also, allowing subnational governments to experiment with different strategies may lead to helpful policy learning, with local governments able to borrow the most successful approaches.
The conventional view, though, is that decentralization will produce a race to the bottom in public insurance benefits. To avoid encouraging in-migration of low-income individuals and households, or to permit lower tax rates, richer areas will reduce benefit access, benefit levels, or benefit duration to match less affluent areas.45 That will tend to be bad for the poor.
The US experience offers mixed evidence. Consider first the main cash social assistance program: AFDC-TANF. The 1996 welfare reform significantly increased states’ authority over rule setting and administration, and this is widely thought to have caused a subsequent reduction in benefit generosity and accessibility. However, benefit levels for this program began declining around 1970, and while they continued to fall after the 1996 reform, there is no indication of an acceleration in the decline, as we see in figure 9. On the other hand, figure 10 shows that access to the benefit did begin to decrease in the mid-1990s, and this was a shift from the prior pattern. Then again, the reduction in access to cash social assistance may have been mainly a product of the new time limits imposed on recipients, and this change came from the federal government rather than from state governments.
Next, consider the statutory minimum wage — not technically a public insurance program, but one of the key government policies aimed at achieving economic security and shared prosperity. As we see in figure 11, the federal government increased the minimum wage steadily from its inception in the late 1930s until the late 1960s. That continued in the 1970s, though higher-than-expected inflation rates during that decade nullified those increases. Since 1980 the federal government has boosted the minimum wage only a few times, and its inflation-adjusted value hasn’t risen at all.
In the late 1990s, however, states and a few cities began enacting statutory minimum wages above the level of the federal minimum, sometimes also indexing their minimum wage to prices or increasing it regularly. As a result, the average minimum wage across the country has risen, and since 2014 the rise has been quite sharp. By 2019 the difference between the federal minimum and the actual minimum was quite large.
This suggests that the ability of state and city governments to set their own minimum wage yielded a significantly higher minimum than would have obtained with a more centralized policy. It’s conceivable that this state and local freedom was itself a key reason why the federal minimum didn’t increase more — that national policy makers chose not to raise the minimum wage because they expected state and city policy makers to do so. But the pattern in the 1980s and 1990s suggests grounds for skepticism about that hypothetical counterfactual; though there were no separate subnational minimum wages during those decades, the federal minimum wasn’t increased.
Further evidence comes from cross-country comparison. Among the rich democratic nations, the Nordic countries have some of the most expansive and generous public insurance programs. Yet many of these programs are characterized by a national framework law that provides guidance and sets minimum standards coupled with extensive autonomy for local governments.46 And subnational governments’ share of total government expenditures is larger in Denmark and Sweden than in the United States.47
We need more research on the impact of centralization and decentralization of public insurance. At this point we can conclude that decentralization tends to produce more geographic inequality. When it comes to benefit generosity, decentralization may result in less, no difference, or even more.
ARE PUBLIC SERVICES PRO-POOR?
Governments in affluent nations provide or subsidize a host of services and public goods.48 Here’s a partial list:
- Physical safety: policing, military
- Assurance of basic liberties: freedom of thought, speech, political participation, religious practice
- Enforcement of property rights and contracts
- Financial safeguards: limited liability for passive investors, bankruptcy, bank deposit insurance, protection against unauthorized use of credit cards
- Clean air and water
- Street cleaning, removal and disposal of sewage and garbage
- Health care
- Disability services
- Elderly services
- Workplace safety
- Consumer safety
- Disaster prevention and relief: firefighting, levies, cleanup, compensation to uninsured victims
- Schooling: early education, K-12, university
- Child care
- Job training
- Job search and placement assistance
- Antidiscrimination enforcement
- Public transportation
- Facilitation of private transportation: roads, bridges, stoplights, enforcement of speed limits, air traffic control
- Public spaces: sidewalks, museums, parks, sports fields, forests, campgrounds, beaches, oceans, lakes, swimming pools, zoos
- Communication, information, and entertainment: support for phone lines, broadband, the Internet, public television and radio programming, subsidization of free private TV and radio networks, libraries, festivals
- Free time: work hours regulations, statutory holidays, mandated vacations, mandated paid parental or family leave
Public services help the poor in two ways. First, they boost living standards directly. Governments subsidize or provide a wide array of services and public goods so that the cost to consumers is small or nil. Second, government services boost the earnings of those at the low end by enhancing human capital, assisting with job search and placement, and facilitating work-family balance.
When governments provide or subsidize public goods and services, they expand the sphere of consumption for which the cost is zero or minimal. This lifts the living standards of the least well-off, and it frees up their limited income for use in purchasing other goods and services.
On one view, however, public services are a highly ineffective way to help the poor, because they are just as widely used by the middle class and the affluent. It turns out, though, that this seemingly sensible intuition is wrong. Here’s why.
For any given type of government service, consumption may be progressive, equal, or regressive. That is, the service or public good may be used more by poor households than by affluent ones, about the same by each, or more by the affluent than by the poor. We have no precise measures of the distribution of service consumption, but estimates suggest that it tends to be similar across the income scale or slightly progressive.49 Let’s assume, perhaps conservatively, that the distribution is equal rather than progressive.
Services can be thought of as akin to a cash transfer. If a person receives government-funded health care worth $10,000, it is as though she has been given a $10,000 cash transfer that she then uses to pay for those health services. The same holds for schooling, roads, policing, and virtually every other type of public service. If the rich, the middle class, and the poor each consume roughly the same total dollars or euros or kroner worth of public services, it is equivalent to the government providing a large flat-rate (equal number of dollars or euros) cash transfer to all households.
In a society in which the market distribution of income is unequal, a flat-rate benefit — one that goes in equal amount to all citizens — is redistributive. It boosts the consumption, and hence the living standards, of the poor more than it does for the rich. If my market income is $100,000 and I receive public services worth $20,000, my consumption has increased by 20%. If my market income is only $10,000 and I receive government services worth $20,000, my consumption is boosted by 200%.
We also need to consider the financing of public services. If the tax system that funds government services is regressive, the provision of services may fail to redistribute. In fact, most rich countries have a tax system that is roughly proportional; people in low-income, middle-income, and high-income households each pay a similar share of their market income in taxes, so the effective tax rate is approximately the same for each group.50 But because the rich get so much of the market income, they pay a much larger share of the tax dollars (or euros or kroner or pounds) than the poor.
Services that are financed proportionally and consumed equally are redistributive. Figure 12 illustrates this. The tax payment data are estimates for the United States. When all types of taxes are taken into account — income, payroll, consumption, and others — the US tax system is proportional; each income group pays approximately the same share of its market income in taxes.51 But because the market distribution of income is quite unequal, those with higher incomes pay much more in dollars. The distribution of public service consumption shown in the chart is hypothetical. It is assumed to be equal across the income scale. In this illustration, households in the bottom income fifth pay about 3% of the taxes and consume about 20% of the government services and public goods.
How large is the redistributive impact of public services? One way to think about this is in relation to government transfers. Estimates by the OECD suggest that services reduce income inequality by only one-quarter to one-half as much as transfers do, depending on the measure used.52 However, this type of estimate includes only expenditures on health, education, and other public “social” services. It leaves out government spending on safety, infrastructure, public spaces, provision of free time (via holidays, paid parental leave, and regulation of work hours), and some of the other public goods and services listed above. If these were included, the estimated redistributive value would be larger.
The second way public services boost the living standards of the poor is by helping them into employment. Here we tend to think mainly of the K-12 school system, but governments do far more than this. The contribution includes early education, health care, job training, job placement assistance, special services for the disabled, language assistance for immigrants, targeted programs for the young and the elderly, assistance with transportation, and provision of temporary employment when few or no private-sector jobs are available.
All rich countries provide services in each of these areas, and in terms of expenditures there is considerable similarity in the two biggest areas — health care and education. The largest difference is in services other than health care and education.53 The Nordic countries in particular stand apart from other rich nations. Since the mid-1960s, Sweden, Denmark, and Norway, and later Finland, have put in place and steadily expanded an array of services that enhance the capabilities of people who grow up in relatively poor households and help them balance employment with family commitments throughout the life course. At the center of this is paid parental leave and government-funded early education. Parents typically are able to take a paid leave during a child’s first year. They can then put the child in high-quality low-cost early education centers after the first year and up to kindergarten. This facilitates employment for parents, especially mothers, and it boosts the cognitive and noncognitive skills of children from disadvantaged homes.54 After finishing formal schooling, young, middle-aged, and older adults can take advantage of various supports collectively referred to as “active labor market programs” — from specialized training to job placement to assistance with geographical relocation, among others.
Figure 13 offers one piece of evidence that these policies help. On the vertical axis is the employment rate among 25-to-64-year-olds with less than upper secondary schooling. On the horizontal axis is the share of GDP spent on public social services other than health care and education. Denmark, Norway, and Sweden are the biggest spenders, and apart from New Zealand and Portugal their employment rates among the less-educated are the highest. Given that the Nordic countries have generous social assistance benefits and other supports that make it possible to live reasonably well without employment, this performance is impressive.
The value of public services in boosting the living standards of the poor is difficult to measure, so we tend to overlook it. We shouldn’t.
Public insurance programs boost the incomes of the least well-off and improve their material well-being. If such programs are too generous, this benefit could be offset by reduced employment or economic growth, but the comparative evidence suggests that the world’s rich nations haven’t reached or exceeded the tipping point.
Spending lots of money on social protection is not in and of itself helpful to the poor. Total social expenditures in the United States are greater than in Denmark and Sweden, because the US has a large private welfare state. But relatively little of America’s private social spending reaches the poor.
Universalism in social policy can be beneficial for poverty reduction, but it is by no means necessary. Countries that make heavier use of targeting have tended to be as successful at income redistribution as those with less targeting.
We know less than is often thought about the impact of centralization versus decentralization of public insurance. Decentralization almost certainly tends to produce more geographic inequality. But its effect on benefit generosity may be good, bad, or neither.
Public services are an important antipoverty tool. Their benefit doesn’t show up in income data, but they appear to play a key role in reducing material hardship. Services expand the sphere of consumption for which the cost is zero or minimal. And they help to boost the earnings and capabilities of the poor by enhancing human capital, assisting with job search and placement, and facilitating work-family balance.
- Michael J. Graetz and Jerry L. Mashaw, True Security: Rethinking American Social Insurance, Yale University Press, 1999; Nicholas Barr, The Welfare State as Piggy Bank, Oxford University Press, 2001; David A. Moss, When All Else Fails: Government as the Ultimate Risk Manager, Harvard University Press, 2002; John Quiggin, “The Risk Society: Social Democracy in an Uncertain World,” Centre for Policy Development, 2007; Lane Kenworthy, Social Democratic Capitalism, Oxford University Press, 2020. ↩
- See also Hilary W. Hoynes, Marianne E. Page, and Ann Huff Stevens, “Poverty in America: Trends and Explanations,” Journal of Economic Perspectives, 2006; Yonaton Ben-Shalom, Robert A. Moffitt, and John Karl Scholz, “An Assessment of the Effectiveness of Anti-Poverty Programs in the United States,” in The Oxford Handbook of the Economics of Poverty, edited by Philip N. Jefferson, Oxford University Press, 2012; Robert Haveman, Rebecca Blank, Robert Moffitt, Timothy Smeeding, and Geoffrey Wallace, “The War on Poverty: Measurement, Trends, and Policy,” Journal of Policy Analysis and Management, 2015, figures 4 and 5. ↩
- Arthur M. Okun, Equality and Efficiency: The Big Tradeoff, Brookings Institution, 1975; Milton Friedman and Rose Friedman, Free to Choose, Harcourt Brace Jovanovich, 1979; Charles Murray, Losing Ground: American Social Policy, 1950-1980, Basic Books, 1984; Vito Tanzi, Governments versus Markets, Cambridge University Press, 2011. ↩
- Kathryn Edin and Laura Lein, Making Ends Meet, Russell Sage Foundation, 1997; Jason DeParle, American Dream, Penguin 2004; Diane Whitmore Schanzenbach, “Can Benefits and Incentives Promote Work?,” Journal of Policy Analysis and Management, 2018. ↩
- For earlier assessments, see Walter Korpi, “Economic Growth and the Welfare State: Leaky Bucket or Irrigation System?,” European Sociological Review, 1985; Ian Gough, “Social Welfare and Competitiveness,” New Political Economy, 1996; Anthony B. Atkinson, The Economic Consequences of Rolling Back the Welfare State, MIT Press, 1999; Lane Kenworthy, Egalitarian Capitalism, Russell Sage Foundation, 2004, ch. 6; Peter Lindert, Growing Public: Social Spending and Economic Growth since the Eighteenth Century, volume 2, Cambridge University Press, 2004; Lane Kenworthy, Jobs with Equality, Oxford University Press, 2008, ch. 7. ↩
- Jonas Agell, “Why Sweden’s Welfare State Needed Reform,” Economic Journal, 1996, p. 1767; Jelle Visser and Anton Hemerijck, A Dutch Miracle, Amsterdam University Press, 1997; Mats Benner and Torben Bundgaard Vad, “Sweden and Denmark: Defending the Welfare State,” in Welfare and Work in the Open Economy, volume 2, edited by Fritz W. Scharpf and Vivien A. Schmidt, Oxford University Press, 2000; Anders Bjorklund, “Going Different Ways: Labour Market Policy in Denmark and Sweden,” in Why Deregulate Labour Markets?, edited by Gøsta Esping-Andersen and Marino Regini, Oxford University Press, 2000; Jørgen Goul Andersen, “Denmark: From the Edge of the Abyss to a Sustainable Welfare State,” in Europe’s New State of Welfare, edited by Jørgen Goul Andersen, Jochen Clasen, Wim van Oorschot, and Knut Halvorsen, Policy Press, 2002; Gordon B. Dahl and Anne C. Gielen, “Intergenerational Spillovers in Disability Insurance,” Working Paper 24296, National Bureau of Economic Research, 2018. ↩
- Janet C. Gornick and Marcia K. Meyers, Families That Work: Policies for Reconciling Parenthood and Employment, Russell Sage Foundation, 2003; Jingjing Huo, Moira Nelson, and John Stephens, “Decommodification and Activation in Social Democratic Policy: Resolving the Paradox,” Journal of European Social Policy, 2008; Kenworthy, Jobs with Equality; Gøsta Esping-Andersen and John Myles, “Economic Inequality and the Welfare State,” in The Oxford Handbook of Economic Inequality, edited by Wiemer Salverda, Brian Nolan, and Timothy M. Smeeding, Oxford University Press, 2009; Lane Kenworthy, “Labor Market Activation,” in The Oxford Handbook of the Welfare State, edited by Francis G. Castles, Stephan Leibfried, Jane Lewis, Herbert Obinger, and Christopher Pierson, Oxford University Press, 2010; Jonas Pontusson, “Once Again a Model: Nordic Social Democracy in a Globalized World,” in Futures of the Left, edited by James Cronin, George Ross, and James Shoch, Duke University Press, 2011; Anton Hemerijck, Changing Welfare States, Oxford University Press, 2012; Henrik Jacobsen Kleven, “How Can Scandinavians Tax So Much?,” Journal of Economic Perspectives, 2014. ↩
- Moss, When All Else Fails; Gareth Olds, “Entrepreneurship and Public Health Insurance,” 2014; David Sraer, David Thesmar, Antoinette Schoar, and Johan Hombert, “Can Unemployment Insurance Spur Entrepreneurial Activity?,” Working Paper 20717, National Bureau of Economic Research, 2014. ↩
- Government spending isn’t the only way to measure the generosity of public insurance. Another option is to focus on program structures rather than expenditures. See Gøsta Esping-Andersen, The Three Worlds of Welfare Capitalism, Princeton University Press, 1990; Lyle Scruggs, Detlef Jahn, and Kati Kuitto, “Comparative Welfare Entitlements Dataset 2,” version 2014-03. Using this type of measure yields patterns very similar to what we see in figures 2 and 3. ↩
- Controlling for other determinants of GDP per capita doesn’t alter this conclusion. See Lane Kenworthy, “Economic Growth,” The Good Society. ↩
- Controlling for other determinants of employment doesn’t alter this conclusion. See Lane Kenworthy, “Employment,” The Good Society. ↩
- Lane Kenworthy, “Employment,” The Good Society. ↩
- Lane Kenworthy, Progress for the Poor, Oxford University Press, 2011, ch. 2; Kenworthy, “A Decent and Rising Income Floor,” The Good Society. ↩
- Kenworthy, “A Decent and Rising Income Floor.” ↩
- “Full Transcript of the Romney Secret Video,” Mother Jones, September 2012. ↩
- Nicholas Eberstadt, A Nation of Takers: America’s Entitlement Epidemic, Templeton Press, 2012. See also Charles Murray, Losing Ground; Murray, Coming Apart: The State of White America, 1960-2010, Crown Forum, 2012. ↩
- Council of Economics Advisors, “The War on Poverty 50 Years Later: A Progress Report,” 2014, p. 28, using data from the IRS and the National Longitudinal Study of Youth; Social Security Administration, Annual Statistical Supplement to the Social Security Bulletin. ↩
- Kenworthy, Social Democratic Capitalism. ↩
- Willem Adema, “Revisiting Real Social Spending Across Countries: A Brief Note,” OECD Economic Studies, 2001; Willem Adema and Maxime Ladaique, “How Expensive is the Welfare State? Gross and Net Indicators in the OECD Social Expenditure Database (SOCX),” OECD Social, Employment, and Migration Working Paper 92, 2009; Price V. Fishback, “Social Welfare Expenditures in the United States and the Nordic Countries: 1900-2003,” Working Paper 15982, National Bureau of Economic Research, 2010. See also Willem Adema, “What Do Countries Really Spend on Social Policies? A Comparative Note,” OECD Economic Studies, 1997; Christopher Howard, The Hidden Welfare State, Princeton University Press, 1997; Jacob S. Hacker, The Divided Welfare State, Cambridge University Press, 2002; Christopher Howard, The Welfare State Nobody Knows, Princeton University Press, 2007; Irwin Garfinkel, Lee Rainwater, and Timothy Smeeding, Wealth and Welfare States, Oxford University Press, 2010; Neil Gilbert, “Comparative Analyses of Stateness and State Action: What Can We Learn from Patterns of Expenditure?,” in United in Diversity? Comparing Social Models in Europe and America, edited by Jens Alber and Neil Gilbert, Oxford University Press, 2010; Suzanne Mettler, The Submerged State, University of Chicago Press, 2011. ↩
- This section draws from Kenworthy, Progress for the Poor, ch. 9. See also Kimberly J. Morgan, “America’s Misguided Approach to Social Welfare: How the Country Could Get More for Less,” Foreign Affairs, 2013. ↩
- Adema and Ladaique, “How Expensive is the Welfare State?,” table 5.5; Fishback, “Social Welfare Expenditures in the United States and the Nordic Countries,” table 5. ↩
- One important tax benefit for low-income households is the EITC, but it is already included in the standard OECD data on government social expenditures. Another is the Child Tax Credit, but it is only partially refundable and thus of limited value to low-income households, many of whom don’t owe any federal income tax. ↩
- They also tend to cost more, due to higher administrative costs and management fees. See Graetz and Mashaw, True Security; Steven Attewell, “Competing Visions of the Past: Learning from History for the Future of American Social Policy,” New America Foundation, 2012. ↩
- Consumption tax rates are higher in the Nordic countries than in the United States. But these are incorporated in the purchasing power parities (PPPs) used to convert incomes to a common currency, so the income numbers in the third row of figure 7 are adjusted for differences in consumption taxes. ↩
- Kenworthy, “A Decent and Rising Income Floor.” ↩
- Graetz and Mashaw, True Security; Anne Kim, Adam Solomon, Bernard L. Schwartz, Jim Kessler, and Stephen Rose, “The New Rules Economy: A Policy Framework for the 21st Century,” ThirdWay, 2007. ↩
- This section draws from Kenworthy, Progress for the Poor, ch. 6. ↩
- Rebecca M. Blank, It Takes a Nation: A New Agenda for Fighting Poverty, Russell Sage Foundation and Princeton University Press, 1997, ch. 6; Neil Gilbert, Transformation of the Welfare State: The Silent Surrender of Public Responsibility, Oxford University Press, 2002, ch. 5; Peter H. Schuck and Richard J. Zeckhauser, Targeting in Social Programs, Brookings Institution, 2006. ↩
- Harold Wilensky, The Welfare State and Equality, University of California Press, 1975; Walter Korpi, “Approaches to the Study of Poverty in the United States: Critical Notes from a European Perspective,” in Poverty and Public Policy, edited by V.T. Covello, Schenkman, 1980; Lee Rainwater, “Stigma in Income-Tested Programs,” in Income-Tested Transfer Programs, edited by Irwin Garfinkel, Academic Press, 1982; Stein Ringen, The Possibility of Politics: A Study in the Political Economy of the Welfare State, Clarendon Press, 1987; Esping-Andersen, The Three Worlds of Welfare Capitalism; Theda Skocpol, “Targeting within Universalism: Politically Viable Policies to Combat Poverty in the United States,” in The Urban Underclass, edited by Christopher Jencks and Paul E. Peterson, Brookings Institution, 1991; Jonah B. Gelbach and Lant H. Pritchett, “Does More for the Poor Mean Less for the Poor?,” Working Paper 1523, Policy Research Department, Poverty and Human Resources Division, The World Bank, 1995; Korpi and Palme, “The Paradox of Redistribution”; Bo Rothstein, Just Institutions Matter: The Moral and Political Logic of the Universal Welfare State, Cambridge University Press, 1998; Karl Ove Moene and Michael Wallerstein, “Targeting and Political Support for Welfare Spending,” Economics of Governance, 2001; Wim van Oorschot, “Targeting Welfare: On the Functions and Dysfunctions of Means Testing in Social Policy,” in World Poverty, edited by Peter Townsend and David Gordon, The Policy Press, 2002; Jonas Pontusson, Inequality and Prosperity, Cornell University Press, 2005; Andrea Louise Campbell, “Universalism, Targeting, and Participation,” in Remaking America: Democracy and Public Policy in an Age of Inequality, edited by Joe Soss, Jacob S. Hacker, and Suzanne Mettler, Russell Sage Foundation, 2007; Christian Albrekt Larsen, “The Institutional Logic of Welfare Attitudes: How Welfare Regimes Influence Public Support,” Comparative Political Studies, 2008. ↩
- Korpi and Palme, “The Paradox of Redistribution”; Hwanjoon Kim, “Anti-Poverty Effectiveness of Taxes and Income Transfers in Welfare States,” International Social Security Review, 2000; Pontusson, Inequality and Prosperity. ↩
- Korpi and Palme, “The Paradox of Redistribution,” p. 663. ↩
- Korpi and Palme, “The Paradox of Redistribution,” p. 677. ↩
- Kenworthy, Progress for the Poor, ch. 6; Ive Marx, Lina Salanauskaite, and Gerlinde Verbist, “The Paradox of Redistribution Revisited,” Discussion Paper 7414, Institute for the Study of Labor, 2013. ↩
- Counting public pensions in a measure of targeting-universalism or redistribution may be misleading. In retirement many people have no income from employment, so the pension they receive appears in the calculations as though it is going to a very poor household. But this is an illusion, as pension programs in effect are forced saving; the government requires employed citizens to put money away during their working years and then returns it to them (with interest) in their retirement years. The measures therefore, according to this view, overstate the degree of targeting and the degree of redistribution achieved by transfers. Peter Whiteford has some calculations of targeting-universalism and redistribution that address this concern. He uses households’ position on the income ladder after transfers are added and taxes subtracted, rather than before. If a retired couple’s income consists solely of a public pension payment, they will be at the very bottom of the ladder according to the calculations in figure 8. In Whiteford’s calculations they instead might be at the twentieth percentile or even higher, depending on how large their pension check is. According to Whiteford’s data, as of the mid-2000s the degree of universalism correlates negatively with redistribution; nations that score higher on universalism tend to score lower on redistribution. This suggests additional reason to rethink the notion that targeting is an impediment to effective redistribution. See Peter Whiteford, “How Much Redistribution Do Governments Achieve? The Role of Cash Transfers and Household Taxes,” ch. 4 in Growing Unequal?, OECD, 2008; Whiteford, “Transfer Issues and Directions for Reform: Australian Transfer Policy in Comparative Perspective,” Social Policy Research Center, University of New South Wales, 2009. ↩
- To measure the size of the redistributive budget I use government social expenditures as a share of GDP, adjusted for the size of the elderly population and the unemployment rate, as in figure 4 above. This is similar to the measure used by Korpi and Palme (“The Paradox of Redistribution,” table 3). ↩
- Robert Greenstein, “Universal and Targeted Approaches to Relieving Poverty: An Alternative View,” in The Urban Underclass, edited by Christopher Jencks and Paul E. Peterson, Brookings Institution, 1991. ↩
- Paul Pierson, Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment, Cambridge University Press, 1994. ↩
- Howard, The Welfare State Nobody Knows, p. 106. ↩
- Kenneth Nelson, “Universalism versus Targeting: The Vulnerability of Social Insurance and Means-Tested Minimum Income Protection in 18 Countries, 1990-2002,” International Social Security Review, 2007, figure 1. ↩
- Kenworthy, Progress for the Poor, ch. 6. ↩
- A softer version of the universalism-is-better-for-redistribution hypothesis says that once a country creates large universal public insurance programs, it can then shift toward greater targeting while still maintaining redistributive generosity, because it will have conveyed to its middle class a sense that the country’s welfare state mainly serves to provide insurance against risk rather than to redistribute money from rich to poor. Unfortunately, this is difficult to test. The problem is that there are other factors apart from the structure of social programs — union strength, left party influence, government structure, public opinion, and perhaps others — that might account for why Denmark has been able to make greater use of targeting without experiencing a shrinking of its welfare state and the US has become more universalistic without a noteworthy increase in the size of its redistributive budget. ↩
- van Oorschot, “Targeting Welfare”; M. Matsaganis et al, “Child Poverty and Family Transfers in Southern Europe,” Euromod Working Paper EM2-04, 2004; Chris de Neubourg, Julie Castonguay, and Keetie Roelen, “Social Safety Nets and Targeted Social Assistance: Lessons from the European Experience,” SP Discussion Paper 0718, Social Protection and Labor, The World Bank, 2007; Bo Rothstein, “Corruption, Happiness, Social Trust, and the Welfare State,” Social Research, 2010. ↩
- Aaron Wildavsky, “Federalism Means Inequality,” Society, 1985; Robin Broadway and Anwar Shah, Fiscal Federalism: Principles and Practice of Multiorder Governance, Cambridge University Press, 2011. ↩
- Sarah K. Bruch, Marcia K. Meyers, and Janet C. Gornick, “The Consequences of Decentralization: Inequality in Safety Net Provision in the Post–Welfare Reform Era,” Social Service Review, 2018. In contrast, a study of changes in centralization of social assistance in the Nordic countries in the 1990s and 2000s found little or no effect on inequality of benefit generosity or provision. However, this may have been because the changes were relatively small compared to the variation across the US states. See Renate Minas, Vibeke Jakobsen, Timo Kauppinen, Tomas Korpi, and Thomas Lorentzen, “The Governance of Poverty: Welfare Reform, Activation Policies, and Social Assistance Benefits and Caseloads in Nordic Countries,” Journal of European Social Policy, 2018. ↩
- Wallace E. Oates, Fiscal Federalism, Harcourt Brace Jovanovich, 1972; Sanford F. Schram and Joe Soss, “Making Something Out of Nothing: Welfare Reform and a New Race to the Bottom,” Publius, 1998; Craig Volden, “The Politics of Competitive Federalism: A Race to the Bottom in Welfare Benefits?,” American Journal of Political Science, 2002. ↩
- Renate Minas and Einar Overbye, “The Territorial Organization of European Social Assistance Schemes,” in Rescaling of Social Welfare Policies, edited by Yuri Kazepov, 2010. ↩
- OECD, OECD Regions and Cities at a Glance. ↩
- This section draws from Kenworthy, Progress for the Poor, ch. 7. ↩
- OECD, Growing Unequal?, 2008, ch. 9; Garfinkel, Rainwater, and Smeeding, Wealth and Welfare States, table 4.1; Alari Paulus, Holly Sutherland, and Panos Tsakloglou, “The Distributional Impact of In-Kind Public Benefits in European Countries,” Journal of Policy Analysis and Management, 2010. ↩
- Lane Kenworthy, “Taxes,” The Good Society. ↩
- Kenworthy, “Taxes.” ↩
- OECD, Growing Unequal?, ch. 9. ↩
- OECD, Growing Unequal?, figure 9.4; Garfinkel, Rainwater, and Smeeding, Wealth and Welfare States. ↩
- Lane Kenworthy, “Early Education,” The Good Society. ↩