Unveiling the Relationship between Economic Growth and Equality for Developing Countries

• Published date: 01 September 2022
• Author: Tsun Se Cheong,Guanghua Wan,David Kam Hung Chui
• Date: 01 September 2022
• DOI: http://doi.org/10.1111/cwe.12435

©2022 Institute of World Economics and Politics, Chinese Academy of Social Sciences

China & World Economy / 1–28, Vol. 30, No. 5, 2022 1

*Tsun Se Cheong, Associate Professor, Department of Economics and Finance, Hang Seng University of

Hong Kong, Hong Kong, China. Email: jamescheong@hsu.edu.hk; Guanghua Wan (corresponding author),

Professor, Institute of World Economy, Fudan University, China. Email: guanghuawan@fudan.edu.cn; David

Kam Hung Chui, Associate Professor, Department of Economics and Finance, Hang Seng University of Hong

Kong, Hong Kong, China. Email: davidchui@hsu.edu.hk. The authors gratefully acknowledge the fi nancial

support provided by the National Natural Science Foundation of China (No. 71833003).

Unveiling the Relationship between Economic Growth

and Equality for Developing Countries

Tsun Se Cheong, Guanghua Wan, David Kam Hung Chui*

Abstract

This study investigates the relationship between economic growth and inequality by

employing the artifi cial neural network approach. There are many important fi ndings.

First, this work reveals the underlying functional form of economic growth and

inequality by using three-dimensional diagrams. Second, the fi ndings show that there

was an inverted-U relationship between economic growth and inequality. This explains

apparent contradictions in research fi ndings in the literature. Third, the optimal level

of inequality, which corresponds to the highest level of economic growth, is computed

for different economies. Our findings were confirmed by the development processes

in many developing countries and also in China in recent years, thereby highlighting

the importance of inequality alleviation in promoting further economic growth. These

fi ndings enable us to derive pragmatic policy implications for other developing countries

at diff erent stages of economic development in achieving sustainable growth with equity.

Keywords: artificial neural network, economic growth, human capital, inequality,

machine learning

JEL codes: D63, N10, O11, O15, O40

I. Introduction

The issue of growth versus equality is a classic socioeconomic issue and draws attention

from different fields, such as economics, sociology, and politics. The relationship

between economic growth and inequality and its underlying mechanism is still subject to

controversy despite a large body of literature. Mixed fi ndings have been obtained – see,

for example, Alesina and Rodrik (1994), Alesina and Perotti (1996), Perotti (1996),

Tsun Se Cheong et al. / 1–28, Vol. 30, No. 5, 2022

©2022 Institute of World Economics and Politics, Chinese Academy of Social Sciences

2

Li and Zou (1998), and Forbes (2000). This paper aims to employ a state-of-the-

art machine learning (ML) technique, namely, the artificial neural network (ANN)

approach, to explore the relationship between economic growth and inequality in

developing countries.

One advantage of the ANN approach is its capability to uncover highly nonlinear

relationships. In one influential paper, Hornik (1991) showed that the multilayer

feedforward network (the version of ANN that we employ in this study) is a universal

approximator provided that sufficient hidden units are used. Given that almost all

research papers on the inequality-growth nexus rely on linear models, the use of ANN

may prove invaluable in reconciling existing and confl icting fi ndings.

Another advantage of ANN lies in the superior accuracy of its forecasts compared

with conventional econometric models. In theory, the ANN can reach any level of

accuracy set by the users. The combination of these two advantages has led to a surge

of ANN applications such as automated car-driving systems, machine translation, and

facial and voice recognition, just to name a few. The performance of ANN is so good

that it can easily supersede human experts. One well-known example is AlphaGo

Zero in playing games such as chess and Go (Silver et al., 2018). However, its use in

economic research is still in the early stage despite the appeal of Mullainathan and

Spiess (2017). Recent attempts include Bajari et al. (2015), Kleinberg et al. (2015), and

Chalfi n et al. (2016).

This study – including the methodology, data, and findings – offers several

signifi cant contributions to diff erent areas of economic literature. First, to the best of

our knowledge, this study is the first attempt to investigate the relationship between

economic growth and equality by using the ML technique. This study therefore off ers

a new perspective and fi lls an important gap in the literature regarding methodology.

Second, by taking nonlinearity into full consideration, this study depicts the underlying

authentic relationship between the variables in great detail. Third, the authors designed

an innovative approach, namely, the regression with ANN and bootstrapping (RAB)

approach, to present our findings by combining ANN models with bootstrapping

techniques. The new approach not only off ers information similar to the conventional

econometric approach but also supersedes it in many ways.

The data are based on the latest Standardized World Income Inequality Database

(SWIID) (Solt, 2016), which has the widest possible coverage across countries and over

time. This study therefore covers most developing countries around the globe, and its

scope is much more comprehensive than previous studies. As the loss of signifi cance of

the parameters resulting from multicollinearity is not a concern for the ANN approach,

an in-depth literature review was conducted to identify all the explanatory variables used