ENTREPRENEURSHIP AND INCOME DISTRIBUTION DYNAMICS: WHY IS THE INCOME SHARE OF TOP INCOME EARNERS ACYCLICAL OVER THE BUSINESS CYCLE?

• Published date: 01 February 2021
• DOI: http://doi.org/10.1111/iere.12489
• Date: 01 February 2021
• Author: Noh‐Sun Kwark,Eunseong Ma

INTERNATIONALECONOMIC REVIEW

Vol. 62, No. 1, February 2021 DOI: 10.1111/iere.12489

ENTREPRENEURSHIP AND INCOME DISTRIBUTION DYNAMICS: WHY IS THE

INCOME SHARE OF TOP INCOME EARNERS ACYCLICAL OVER THE BUSINESS

CYCLE?∗

By Noh-Sun Kwark and Eunseong Ma1

Sogang University, Korea ; Louisiana State University, U.S.A

In the United States, the income share of the top 5% income group is acyclical over the business cycle.This

study attempts to explain the cyclical behavior of the income distribution over the business cycle, particularly

focusing on the top 5% income earners’ share, using a heterogeneous agent model featuring a choice to be-

come an entrepreneur. The model economy successfully reproduces the acyclical behavior of the income share

of the top 5%. During expansions, relatively more people become entrepreneurs at the top, which offsets the

decline in the income share of the high-income earners from workers’ side.

1. introduction

The “Occupy Wall Street” movement drew much attention to the inequality of distribution,

in particular, to the concentration of wealth and income in the top 1%. It may be surprising

that Piketty (2014)’s 700-page book, Capital in the Twenty-First Century, was listed as a best-

seller on Amazon, but it ref‌lects a lot of interest in the inequality of income and wealth. The

long-run trend that income or wealth distribution becomes more unequal motivated many re-

searchers to identify the causes and effects of the concentration of wealth and income. In fact,

existing papers in the literature have devoted a great deal of effort to studying the causes of

the highly concentrated distributions,2but most of them have mainly focused on reproducing

the steady-state distribution of income or wealth. In addition to the level and long-run trend

of inequality, the issue of income distribution dynamics over the business cycle is also cen-

tral to the discussion of economic policies, such as redistributive as well as stabilization poli-

cies. In spite of its importance, replicating the cyclical behavior of income distribution is in its

infancy in that not only there are only a few studies on this issue, but also the conventional

models trying to explain the business cycle dynamics of income distribution have had limited

success. This study is another attempt to explain the business cycle behaviors of income distri-

bution with an emphasis on the top income earners.

∗Manuscript received December 2018; revised September 2020.

1We would like to thank José-Víctor Ríos-Rull and three anonymous referees for their insightful suggestions and

comments. Wewish to thank Yongsung Chang and Sun-Bin Kim for providing the computational codes for their het-

erogeneous agent model. We also wish to thank seminar participants at the 2018 Asian Meeting of the Econometric

Society, the 41st Annual Conference of the Eastern Economic Association, the Bank of Korea,the Korea Association

of Money and Finance Conference, the Korea Econometric Society Winter Conference, Korea University, Sogang

University, Texas A&M University,Yonsei University, and the Macroeconomics Study Group in Seoul for their valu-

able comments. This work was supported by the National Research Foundationof Korea Grant funded by the Korean

Government (NRF-2014S1A5A2A01015689). Please address correspondence to: Eunseong Ma, Department of Eco-

nomics, Louisiana State University, 2312 Business Education Complex South, 501 South Quad Drive, Baton Rouge,

LA 70803, U.S.A.Phone: +1-225-578-5211. Fax: +1-225-578-3807. E-mail: masilver@lsu.edu.

2For example, Cagetti and De Nardi (2006) and Quadrini (2000) successfully replicate the unequal wealth distribu-

tion by incorporating entrepreneurship into heterogeneous agent models, and Heathcote et al. (2010) investigate an

empirical analysis of various dimensions of inequality in the United States.

321

© (2020) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of So-

cial and Economic Research Association

322 kwark and ma

In the U.S. economy, there is a signif‌icant difference in the cyclical properties of the low-

and high-income groups. As found in Castañeda et al. (1998), the income shares of the f‌irst

three lowest income quintiles are strongly procyclical over the business cycle, strongly coun-

tercyclical for the fourth quintile and the 80–95 percentile, and, interestingly, acyclical for the

top 5%.3That is, during expansions, the income shares of the individuals in the low-income

quintiles tend to increase, as new employment occurs mostly among the low income, whereas

those of the individuals in the high-income groups tend to decrease. However, the income

share of the top 5% group does not show a cyclical behavior, basically unaffected by the busi-

ness cycle. These empirical f‌indings are quite surprising and interesting in that individuals in

the top income groups not only earn a high portion of the total income, but their income share

is also steady over the business cycle. This is a very desirable feature of low risks and high re-

turns, contrasting with the conventional wisdom of high income coming along with high risks.4

We propose a hypothesis that entrepreneurial activities play an important role in the cycli-

cal behavior of the income shares for the top income earners. To support this conjecture, we

f‌irst empirically document the large income share of entrepreneurs in the top 5%, compared

to that of the fourth quintile or the 80th to 95th percentile group, with the large population

share and the high average income of entrepreneurs relative to workers. We also f‌ind empir-

ical evidence that transition between workers and entrepreneurs is very active over the busi-

ness cycle, in particular, from workers to entrepreneurs in booms in the top income group.

More households work as workers or entrepreneurs in booms, but a relatively large portion

of workers switch to entrepreneurs in the top income group, which offsets a decline in the in-

come share of the top income earners from the workers’ side since entrepreneurs earn higher

income than workers on average. Hence, this transition activity between entrepreneurs and

workers may produce the acyclical behavior of the income share of the top income earners.

The purpose of this study is to replicate the cyclical behavior of the income distribution

over the business cycle, focusing on the behavior of the income share of the top 5% in-

come earners. To this end, we incorporate an entrepreneurial choice in a heterogeneous agent

model. Specif‌ically, we build a simple dynamic stochastic general equilibrium (DSGE) model

featuring a large population of households who face idiosyncratic risks to labor eff‌iciency

and entrepreneurial productivity. As in Huggett (1993) and Aiyagari (1994), the capital mar-

kets are incomplete as households cannot fully insure against the two idiosyncratic shocks.

The market incompleteness, together with borrowing constraints, helps produce substantial

heterogeneity across individual households according to their income and wealth. One of

the main features in our model is an occupational decision for households to become en-

trepreneurs, as in Quadrini (2000), Cagetti and De Nardi (2006), and Terajima (2006). In

the model economy, households make an occupational decision among three occupational

choices: employment, entrepreneurship, and no work. A household may decide to provide

labor supply to the market as a worker, to run its own business by investing its labor and

capital as an entrepreneur, or not to work due to unfavorable individual labor productivity

shocks and/or low entrepreneurial productivity shock.5Relative to the previous studies on en-

trepreneurship in the literature, another important feature is that our model economy embeds

aggregate productivity shocks in addition to the two idiosyncratic shocks in an occupational

choice model. In this sense, our model is an extended version of the traditional heterogeneous

agent models of Krusell and Smith (1998) and Chang and Kim (2007), who introduce aggre-

gate uncertainty in an incomplete market model. Incorporating aggregate productivity shocks

allows for f‌luctuations in the income distribution over the business cycles so that we can

3We use the Current Population Survey (CPS) data for the period of 1968–2016, whereas Castañeda et al. (1998)

use the CPS samples from 1948–86.

4We do not talk about micro-level risks but aggregate macro-level risks of the income groups over the business

cycle.

5The previous studies consider either an occupational decision between the two choices of workers and en-

trepreneurs, for example, Quadrini (2000) and Cagetti and De Nardi (2006), or a decision between the two choices

of workers and nonemployed workers (e.g., Chang and Kim, 2007).

entrepreneurship and income distribution dynamics 323

investigate the cyclical behavior of the income distribution. To the best of our knowledge, this

study is the f‌irst quantitative work studying a heterogeneous agent model with occupational

choices in the presence of aggregate shocks.

The main f‌indings of this study are summarized as follows. The transition between work-

ing households and nonemployed workers has been found to be important in reproducing

the overall cyclicality of income shares over income groups (Castañeda et al., 1998). Impor-

tantly, we f‌ind that introducing a decision to become entrepreneurs plays a crucial role in re-

producing the acyclical behavior of the income share of the top income earners. During ex-

pansions, there are more opportunities for entrepreneurship, especially for the top income

earners, which induces more individuals to become entrepreneurs and earn high income. This

offsets a decline in the workers’ income share in the top income group, caused by an increase

in the employment of workers in the low-income quintiles.

1.1. Related Literature. This article is related to two strands of the heterogeneous agent

DSGE literature: the cyclical behavior of income distribution and entrepreneurial activities.

The work of Castañeda et al. (1998) is the f‌irst comprehensive study attempting to account

for the income distribution dynamics over the business cycle in a dynamic general equilib-

rium model with inf‌initely lived agents and unemployment risk. They document the income

distribution dynamics and introduce unemployment spells and a cyclically moving labor share

to account for the cyclical behavior of the income distribution. They f‌ind that unemployment

spells may play a signif‌icant role in the income distribution dynamics, which is supportive of

the indivisibility of labor, but they do not f‌ind the role of a cyclically moving labor share in ex-

plaining the cyclical properties of such dynamics. Even though they are successful in generat-

ing the right signs of the contemporaneous correlations between the income shares and out-

put for most of the income groups, the correlations are very high in absolute terms, and the

income share earned by the top 5% income group shows a strong negative correlation with

output, though acyclical in the data. Heer (2013) also tries to explain the cyclical behavior

of the income distribution with a model of overlapping generations and the savings motive

for retirement, but he f‌inds that the model economy with wage rigidity does not improve the

cyclical behavior of the income distribution over Castañeda et al. (1998), in particular, the

acyclical property of the income shares of the top 5%. The contribution of this article to the

literature is that the model with a choice to become an entrepreneur successfully reproduces

the salient features of the acyclical behavior of the income share of the top 5% income earn-

ers as well as the overall cyclicality of income shares of the income groups.

The second strand of the previous quantitative studies closely related to our work is the en-

trepreneurial choice models by Quadrini (2000), Cagetti and De Nardi (2006), and Terajima

(2006). Quadrini (2000) constructs a dynamic general equilibrium model that incorporates an

entrepreneurial choice and f‌inancial frictions under incomplete markets.6He shows that the

model economy can generate a high inequality in wealth distribution and replicate the em-

pirical patterns of wealth mobility. Cagetti and De Nardi (2006) also study a model economy

with an occupational choice allowing for entrepreneurial activities, similar to Quadrini (2000),

to analyze the effects of borrowing constraints as a source of wealth concentration. In con-

trast to Quadrini (2000), who introduces exogenous f‌inancial frictions, Cagetti and De Nardi

(2006) construct endogenously determined borrowing constraints in equilibrium and f‌ind that

the model economy suff‌iciently reproduces the observed wealth distribution for entrepreneurs

and workers, and more restrictive borrowing constraints generate less wealth concentration.

Terajima (2006) also studies a model considering relations between occupational and educa-

tional choices to analyze the interaction between the changes in earnings and wealth inequal-

ity over time and f‌inds that the model can explain about 33% of the change in the relative

average wealth between different education–occupation groups. This study contributes to the

6Yurdagul (2017) also constructs a model economy which features a f‌lexible hours motive for entrepreneurship

and f‌inds that the model can generate the income difference between entrepreneurs and comparable workers.