A Libertarian Perspective on Economic and Social Policy Lecture 15 Poverty and Redistribution ©2007 Jeffrey A. Miron.

A Libertarian Perspective on Economic and Social Policy

Lecture 15

Poverty and Redistribution

©2007 Jeffrey A. Miron

Introduction

• Most modern societies attempt to alleviate poverty through a variety of programs.

• Relatedly, most modern societies attempt to redistribute income from higher to lower income households, using a variety of methods such as cash transfers, in-kind transfers, progressive taxation, and the like.

Introduction, continued

• Relatively few people object to these policies in principle:– Instead, critics mainly note that the magnitude should

be moderate because of disincentive effects on labor supply and saving and distortions from taxation.

– Or, they note that current methods of redistributing income are badly designed.

• And, some people suggest that even libertarians should accept government intervention in this arena because redistributing income is a public good: everyone benefits if one person does it.

Introduction, continued

• This lecture challenges the key components of this view, based on the following propositions:– Alleviating poverty and redistributing income are

different things, with only the former having a reasonable degree of justification.

– There are costs of redistribution that are potentially more important than the strictly economic effects.

– The public goods case for government intervention, while not wrong, is overstated.

– The current level of anti-poverty provision is generous.

Introduction, continued

• Thus, if there is a case for government attempts to alleviate poverty, the amount should be less than currently provided.

• In addition, the design should differ radically from current practice:– Conducted by state rather than federal

governments;– Based on replacing all existing programs with

a negative income tax.

Outline

• Redistribution Versus Alleviating Poverty

• The Argument for Anti-Poverty Programs

• The Costs of Anti-poverty Programs

• The Design of Anti-Poverty Programs

• The Negative Income Tax

• What is the Right Amount of Transfers?

• Federal or State Anti-Poverty Programs?

General Redistribution Versus Alleviating Poverty

• Before discussing the benefits and costs of income redistribution, it is useful to distinguish two possible goals:

• One is helping people at the low end of the income distribution; – For example, transferring cash to families with

income below $5,000 per year (“welfare”).

• Refer to this goal as alleviating poverty.

General Redistribution Versus Alleviating Poverty, continued

• A different goal is systematically transferring income from higher income to lower income persons at every point in the income distribution:– For example, progressive income taxation,

under which taxes as a fraction of income increase with income.

• Refer to this as general redistribution.

General Redistribution Versus Alleviating Poverty, continued

• There is a credible case for policies to alleviate poverty:– Few poor people “choose” to be poor;– Instead, most are poor due to bad luck.

• But this characterization is less true, or not true at all, for people who are “non-poor:”– Some Harvard law grads choose to work at low-

stress, low-effort jobs with moderate salaries;– Others choose to work at high-stress, high-effort jobs

with larger salaries;– There is no reason one group is more deserving than

the other.

General Redistribution Versus Alleviating Poverty, continued

• More generally, differences in income across the entire distribution are partly luck, but they also reflect skill, effort, risk-taking, and preferences for leisure.

• Thus, the case for general redistribution based on compassion or fairness is not particularly convincing.

• The remaining discussion, therefore, takes as given that the question is whether to have anti-poverty programs, i.e., redistribution aimed at the very poor.– Given that the conclusion will be somewhat skeptical about the

value of this intervention, it follows there is not much of a case for general redistribution.

Arguments for Anti-Poverty Programs: Compassion

• The first possible rationale for anti-poverty programs relies on two assumptions.

• The first is that everyone in society is adversely affected by the existence of poverty due to feelings of compassion.

• The second is that the private sector does not provide sufficient anti-poverty transfers because of a classic public goods problem:– Everyone benefits when any given individual provides

aid to the poor, so each individual donor captures only a portion of the benefit.

Arguments for Anti-Poverty Programs: Insurance

• The second possible rational for anti-poverty programs also rests on two assumptions:

• The first is that many people would like to purchase insurance against the possibility of ending up in poverty;

• The second is that the private sector does not supply this income insurance due to adverse selection and moral hazard problems.

Further Arguments for Anti-Poverty Programs

• Diminishing Marginal Utility:– A dollar means more to a starving person than

to Bill Gates.

• Preventing Insurrection and Civil Unrest:– In a society with “too much” inequality, the

probability of revolution might increase.

Arguments Against Anti-Poverty Programs: Disincentives

• The first possible objection to anti-poverty programs is that they reduce recipients’ incentives to work and save.

• This prediction follows from standard economic theory; the possibility of such an effect is not really in dispute.

• Whether this effect is large depends on a range of auxiliary issues.

• In particular, it depends critically on how much income is transferred by anti-poverty programs.

Disincentives, continued

• Consider a policy that guarantees every person an income of $1,000 per year. How many people will reduce their labor supply substantially and simply live off the transfer?– Very few.

• Consider a policy that guarantees every person an income of $50,000 per year. How many will reduce their labor supply substantially?– Lots.

• This point is obvious but often swept under the rug. It is impossible to talk rationally about whether to redistribute without also discussing the question of how much.

Disincentives, continued

• The evidence on disincentive effects of current anti-poverty programs is mixed:– Unquestionably evidence of non-trivial responses.– But it is difficult to argue this is a major efficiency loss

to the U.S. economy.

• This does not mean arbitrarily high levels of redistribution make sense.

• But, in terms of reasons to oppose existing U.S. anti-poverty programs, the direct disincentives created are probably not the main reason.

Arguments Against Redistribution:Deadweight Loss of Taxation

• In order to transfer income to recipients, the government must impose taxes to raise the necessary revenue.

• Standard economic theory explains that taxation imposes deadweight losses by distorting incentives (Harberger triangles).

• As a matter of theory, there is no real dispute about this issue, either.

Deadweight Loss of Taxation,continued

• The critical question is again about magnitudes; – If raising a dollar of taxes imposes a DWL of 5 cents,

this is relevant but not a huge issue.– And, on the one hand, there is a theoretical reason to

believe the effect is this magnitude.

• There are also plausible assumptions, however, under which the DWL from taxation is large:– Existing evidence is mixed, but it certainly raises the

possibly of significant DWLs from existing levels of taxation in the U.S. economy.

Arguments Against Anti-Poverty Programs: Crowding Out

• One key component of the argument for government anti-poverty programs is that the private sector will not supply such transfers because of the public goods problem.

• This is a defensible perspective, but despite the free-rider issue, the private sector engages in substantial aid for the poor:– Churches, Cub Scouts, Soup Kitchens, Shelters, …

• And, most of these pre-date government anti-poverty programs.

Crowding Out, continued

• Further, it is possible and indeed likely that government anti-poverty measures to some degree crowd out private measures.– Thus, elimination of government anti-poverty

programs would decrease aid to the poor by less than the cut in government aid.

• Also, the existing evidence is incomplete, but it does not suggest the government anti-poverty programs have been particularly effective at reducing poverty.

Arguments Against Anti-Poverty Programs: Envy

• Any policy that transfers income to persons below a given cutoff means there are persons above this cutoff who do not get transfers.

• This fact generates envy and stimulates a demand for transfers from those above the cutoff:– The result is substantial redistribution beyond what

can be plausibly considered alleviation of poverty.• Assuming these transfers cause labor supply

and savings distortions, and that raising taxes to pay for these transfers causes DWL’s, this means substantial efficiency losses.

Arguments Against Anti-Poverty Programs: Self-Reliance

• Finally, but perhaps most importantly, government anti-poverty programs promote the view that low income is “someone else’s fault.”– This is unquestionably true in some instances;

consider children born into poverty.

• Nevertheless, this attitude, if accepted broadly, reduces society’s work ethic and productivity, potentially to a degree that harms everyone:– Society devotes its efforts to “transfer-seeking” rather

than to innovation and productivity.

Summary So Far

• Redistribution can mean general transfers from richer to poorer, or it can mean transfers to the very poor:– There is a good argument only for the latter.

• There is a plausible argument for government anti-poverty programs, but there are non-trivial costs associated with these as well.

• The magnitude of these costs depends on the level of generosity:– There is an inevitable tradeoff between helping and

hurting.

The Design of Anti-Poverty Programs

• If the government runs anti-poverty programs, how should it do this? • Currently, anti-poverty programs are numerous and complicated:

– Temporary Assistance to Needy Families– Food Stamps– Medicaid– Subsidized Housing– Unemployment Insurance– Social Security– Supplementary Security Income– Disability Insurance– Earned Income Tax Credit– Educational Loans– Public Schools– Energy Subsidies– General assistance and other state progams

Design of Anti-Poverty Programs,continued

• Note several things about this approach:– In-kind transfers are plausibly less efficient than cash

transfers. – The amount being transferred is difficult to see.– The interactions are often complicated.– In many cases there are extremely high marginal tax

rates on the transfers.– The structure of some components potentially

encourages single family households.– There is lots of bureaucracy involved.

An Alternative Approach:The Negative Income Tax

• Consider the following policy:– Guarantee a minimum level of income per person, G.

Transfer this amount to every person with zero earnings.

– Reduce the size of the transfer by t dollars for every dollar of income earned (with t<1).

• For concreteness, assume G = $5,000 and t = .5. Then – If Y=0, B = $5,000;– If Y=$5,000, B = $2,500 (Y+B = $7,500)– If Y=$10,000, B = $0 (Y+B = $10,000)

The Negative Income Tax, continued

• Is this approach perfect?• No:

– There is still a disincentive to work/save.– There is still the DWL from taxation to fund.

• But the NIT has many advantages:– The rules are clear.– The amounts are clear.– There is far less bureaucracy.– There are no distortions of particular markets involved with in-

kind transfers, in particular, health care.– There are no political issues such as whether programs like

Medicaid should fund specific items like abortion.• The amount transferred can be higher or lower than

under current policy.

Aside: Transfers to Families or Individuals?

• An issue that arises for most anti-poverty programs is whether to define the beneficiary as a family or an individual.– Current practice is a mix, but tends to focus on the

family.

• In particular, key programs like TANF, EITC, increase benefits as the number of children increases, or they give benefits only to single moms:– This potentially encourages additional children and/or

single parent households.

Transfers to Families or Individuals?, continued

• An alternative approach is to define the target of transfer programs as individuals and limit the set of individuals considered to, say, persons 18 and over.

• This might harm children relative to adults, for those children that exist.

• But it might also reduce the number of children born into poor families by eliminating any subsidy for children.

• It also avoids any incentive for single parenting. • And it’s way simpler than current practice.

Summary, Updated

• Redistribution should be only for the poor.• The costs of this approach depend on the

magnitude of transfers and on the design.• The current design of anti-poverty programs is

ridiculous:– A NIT would be substantially better, although not

perfect either.

• Society must accept a basic tradeoff:– Generous programs cause greater distortions;– Stingier programs help fewer people.

How Generous Are Current Anti-Poverty Programs?

• The current level of “anti-poverty” support is substantial.– $450 billion on programs for the non-elderly.– $500 billion on Social Security plus Disability Insurance– $275 billion on Medicare– $40 billion on Unemployment Insurance

• If this amount were transferred to the poorest 10% of the population, it would work out to $45,000 per person.

• This is a huge amount. – The potential for disincentives is substantial.– The potential for DWLs from taxation is substantial.– The potential for envy, reduced work ethic, is substantial.

Aside: CPI Bias and Absolute Versus Relative Poverty

• Most programs adjust the level of benefits for inflation using the CPI.

• It is now well understood that the CPI overstates inflation:– Substitution effects– Quality improvements– New products

• Existing estimates suggest the bias is roughly 1% per year.

• The implication is that benefit levels have in fact increased substantially in real terms.– This make intuitive sense. Most poor in 1950 did not have, e.g.,

TVs, phones, indoor plumbing, access to modern medical care, and innumerable other results of productivity growth.

Federal Versus State Provision of Anti-Poverty Programs

• Currently, most anti-poverty programs are funded / run by the federal government.

• The standard defense of this approach asserts that, if such programs were left to the states, there would be a “race-to-the-bottom” in which each state attempted to “export” its poor by setting a low level of transfers to the poor.

Federal versus State, continued

• This argument is plausible as far as it goes, but there is more to consider.

• The generosity of anti-poverty programs has unquestionably grown over time, and there are reasons why this is more or less inevitable:– Demand for transfers from the near poor;– Mission creep by the bureaucrats who run

anti-poverty programs.

Federal versus State, continued

• Thus, while the race-to-the-bottom effect is probably relevant, so is the tendency for anti-poverty programs to expand.

• There is no guarantee that leaving anti-poverty programs to the states gets this balance exactly right, but it at least has a chance;– Having the federal government run these programs

allows the expansion effect to operate unchecked.• The U.S. experience with TANF since 1996 is

consistent with this view:– The transfer of authority to states has moderated but

certainly not gutted welfare.

Should States Operate Anti-Poverty Programs?

• Imagine everyone had agreed that – Income redistribution should be limited to anti-poverty

programs – The method of redistribution should be a negative

income tax;– The appropriate level of government to operate these

programs is states.– The appropriate amount is “small.”

• Given all this, should states operate anti-poverty programs?

Should States Operate Anti-Poverty Programs?, continued

• Libertarians can reasonably come down on either side of this question.– On the one hand, Libertarians accept government interventions

to protect people who can truly not help themselves. – One the other hand, there is no convincing evidence that

government intervention in this area has improved matters substantially relative to what would otherwise have occurred, and these interventions have costs.

• Thus my own hunch is that the optimal level of government anti-poverty programs is zero, even if done as well as possible.– But the case is not definitive. If the programs satisfied all the

conditions above, there is probably not a substantial cost.

Conclusions

• General income redistribution is difficult to justify; anti-poverty programs are perhaps defensible.

• The current design of anti-poverty programs is abysmal; a Negative Income Tax would be far better.

• There should be no federal redistribution of income.

• The current level of anti-poverty programs is too high; the right level is substantially smaller, and plausibly zero.