Optimal Income Taxation and Job Choice

• Published date: 01 October 2017
• Author: Zhen Song,Jean‐François Tremblay,Robin Boadway
• Date: 01 October 2017
• DOI: http://doi.org/10.1111/sjoe.12199

©The editors of The Scandinavian Journal of Economics 2016.

Scand. J. of Economics 119(4), 910–938, 2017

DOI: 10.1111/sjoe.12199

Optimal Income Taxation and Job Choice*

Robin Boadway

Queen’s University, Kingston, ON, K7L 3N6, Canada

boadwayr@econ.queensu.ca

Zhen Song

Central University of Finance and Economics, Beijing 100081, China

songzhen@cufe.edu.cn

Jean-Fran ¸cois Tremblay

University of Ottawa, Ottawa, ON, K1N 6N5, Canada

jean-francois.tremblay@uottawa.ca

Abstract

In this paper, we study optimal income taxation when different job types exist for

workers of different skills. Each job type has some feasible range of incomes from

which workers choose by varying labor supply. Workers are more productive than others

in the jobs that suit them best. The model combines features of the classic optimal tax

literature with labor variability along the intensive margin, with the extensive-margin

approach where workers make discrete job choices and/or participation decisions. We

find that first-best maximin utility can be achieved in the second-best, and marginal

tax rates below the top can be negative or zero.

Keywords: Extensive margin; intensive margin; skill-type heterogeneity

JEL classification:H21; H23; H24

I. Introduction

In models of optimal income taxation, the focus has been largely on

the labor supply side. The nature of jobs offered has not been explicitly

modeled. Two extreme cases have been studied. The standard intensive-

margin approach, following Mirrlees (1971), assumes that workers of

different skills are perfect substitutes in production. The amount of effective

labor per hour of work reflects skills, and workers receive a fixed wage rate

equal to their skill. In extensive-margin models, jobs are offered that are

suited for each type of skill, and that pay a fixed wage for a given amount

of effort (Diamond, 1980). In some versions of the extensive-margin

*We are grateful to Laurent Simula and the anonymous referees for careful comments on

an earlier version of this paper. Research support was provided by the Social Sciences and

Humanities Research Council of Canada.

R. Boadway, Z. Song, and J.-F. Tremblay 911

model, workers of a given skill can choose the job suited for a less-skilled

worker, but if they do, they receive the same fixed wage (Saez, 2002).

These two approaches have led to important insights into the structure

of optimal income taxes, and how they are affected by the nature of labor

supply decisions, but they set aside two characteristics of jobs that are

potentially relevant for income tax design. First, workers have different

job-specific skills, which can make them more productive than others in the

job that suits them best. Given the division of labor in a modern economy

and the specialization that entails, a worker tends to perform a narrow set

of tasks within a work unit. Each set of tasks requires a corresponding

set of special skills that workers possess in different degrees, such as

innate ability, education and training, and work experience. Workers will

have different comparative advantages at jobs requiring different sets of

skills. Moreover, some workers might have an absolute advantage, over

workers with different skills, at those jobs that suit them best.

Absolute advantages could reflect, for example, the fact that workers

who are best suited for jobs that require high cognitive skills might

not be highly productive in manual jobs that require different types of

skills (i.e., physical strength, the technical know-how required to perform

specific manual operations or routine tasks, etc.) To take a concrete

example, a health professional who has the skills to work in the health-care

sector would generally be relatively unproductive in sectors such as the

construction sector or the natural resource sector where the productivity of

a worker might depend on skills of a very different nature. More generally,

different types of workers have different productivities in different

jobs because of the specific training, experience, and even personal

characteristics they possess. The possibility that higher-skilled workers

are less productive than lower-skilled workers in jobs or tasks better

suited for lower-skilled workers has been considered in some of the

recent optimal taxation literature, including Rothschild and Scheuer (2013),

Scheuer (2014), and Ales et al. (2015). A similar notion of absolute

advantage was also considered in a more extreme form by Diamond

(1980), who assumed that high-skilled workers simply cannot occupy jobs

with lower skill requirements. He argued that individuals with different

types of skills are suited to different types of jobs, and that the social

valuations of different jobs, reflected in the value of marginal products,

determine which workers are high-skilled and which are low-skilled. The

absolute advantage assumption that we adopt below is in the same spirit

as that considered by Diamond, although it takes a milder form.

Second, the assumption used in extensive-margin models that hours

worked and incomes are absolutely fixed seems to be slightly restrictive,

while that used in intensive-margin models – that workers can choose

any income by freely varying their labor supply – can also be restrictive.

©The editors of The Scandinavian Journal of Economics 2016.