Public Finance, 2nd Edition

15

TAXATION UNDER

ASYMMETRIC INFORMATION

LUMP-SUM REDISTRIBUTIONS AND PRIVATE INFORMATION

REDISTRIBUTION THROUGH COMMODITY TAXATION

OPTIMAL TAXATION, PRIVATE INFORMATION, AND

SELF-SELECTION CONSTRAINTS

Elements of the Model

Pareto-Efficient Taxation

An Extension: The Direct-Indirect Tax Mix

OPTIMAL INCOME TAXATION

The Shape of the Tax Schedule

Concluding Observations

TAX EVASION

Increasing the Penalty

Increasing Monitoring

Revenue-Raising Strategies

Tax Amnesties

CONCLUDING REMARKS

The text so far has ignored an important market imperfection, the

presence of private or asymmetric information. Chapter 15 explores the

implications of asymmetric information on taxation, and later chapters

extend the analysis to transfer payments and other public expenditures.

Analysis under the assumption of asymmetric information is inherently

second best because Wrst-best analysis requires that agents have perfect infor-

mation about everything relevant to their economic decisions and exchanges.

The problem of asymmetric information has been a focal point of public

sector analysis for the past 15 or 20 years, just as it has been in almost all Welds

of economics. The recent interest in asymmetric information is understand-

able, Wrst and foremost because it is so common. Agents often possess private

information about themselves that other agents, including the government,

do not or cannot know, at least not without undertaking considerable eVort

485

and cost to monitor behavior. In addition, optimizing agents have obvious

incentives to exploit private information to their own advantage, and econo-

mists quite naturally assume they will do so to the fullest extent possible.

Finally, the assumption of asymmetric information often produces results

that are very diVerent from those obtained under the assumption of perfect

information. This has been especially true in public sector economics.

Regarding the theory of taxation, old standards such as the Ramsey tax

rule for one-consumer equivalent economies or the Diamond±Mirrlees

many-person tax rule are no longer prescriptions for optimal taxation

under private information. This is so even if the government can in principle

tax (almost) everything as the Ramsey and Diamond±Mirrlees models

assume. In fact, governments may be quite restricted in what they can tax

under private information. They may not have suYcient information about

some economic variables to use them as a tax base. What can the government

tax? is an important policy question in a world of private information. A Wnal

point is that private information can severely constrain a government's

ability to redistribute purchasing power through taxes and transfers.

The last comment on redistribution points to a special diYculty with

private information: It is not simply a technological or structural imper-

fection. Rather, it has certain uncomfortable behavioral characteristics that

are absent from other market imperfections such as distorting taxes, or

monopoly power, or legislated budget constraints. Agents who exploit pri-

vate information for their own self-interest at the expense of broader social

goals tear at the fabric of society. They violate the spirit of good citizenship

that is necessary to hold a society together. An obvious example is people

who evade paying taxes that would have been transferred to the poor.

1

Such behavior also strikes at the foundations of normative public sector

theory. What is normative theory to make of the tax evaders, especially when

the norms include a concern for equity as well as eYciency? In formal terms,

should the social welfare function give dishonest taxpayers the same ethical

weight as honest taxpayers who could also exploit private information but do

not? If unequal weights are chosen, how unequal should they be? Should the

dishonest receive zero weight? If the dishonest are to receive a positive weight,

should society be expected to spend its scare resources on monitoring their

behavior or on punishing them? Should the government signiWcantly alter its

tax policies to discourage dishonest behavior? These are diYcult, open ques-

tions that have profound implications for any normative analysis. DiVerent

answers can lead to very diVerent policy recommendations.

As it happens, much of the recent normative public sector analysis has

continued the long-established tradition of using the equal-weight utilitarian

1

The same could be said of agents who exploit monopoly power, but their behavior is not

diVerent in kind from proWt maximizing under perfect competition, unlike the distinction

between honest and dishonest taxpayers. It is also less secretive than something like tax evasion.

486 TAXATION UNDER ASYMMETRIC INFORMATION