Explaining Africa's public consumption procyclicality: Revisiting old evidence

• Author: João T. Jalles
• DOI: http://doi.org/10.1111/infi.12365
• Published date: 01 August 2020
• Date: 01 August 2020

International Finance. 2020;23:297–323. wileyonlinelibrary.com/journal/infi © 2019 John Wiley & Sons Ltd.

|

297

DOI: 10.1111/infi.12365

ORIGINAL ARTICLE

Explaining Africa’s public consumption

procyclicality: Revisiting old evidence

João T. Jalles

1,2,3

1

Portuguese Public Finance Council,

Lisbon, Portugal

2

REM/UECE, Lisbon, Portugal

3

Nova School of Business and Economics,

Centre for Globalization and Governance,

Lisbon, Portugal

Correspondence

João T. Jalles, Portuguese Public Finance

Council, Praca de Alvalade 6, 1700‐036

Lisbon, Portugal.

Email: joaojalles@gmail.com

Abstract

This paper compiles a novel data set of time‐varying

measures of government‐consumption cyclicality for a

panel of 46 African economies between 1960 and 2014.

Government consumption has, generally, been highly

procyclical over time in this group of countries.

However, sample averages hide serious heterogeneity

across countries with the majority of them showing

procyclical behaviour despite some positive signs of

graduation from the “procyclicality trap”in a few cases.

By means of weighted least squares regressions, we find

that more developed African economies tend to have a

smaller degree of government‐consumption procyclical-

ity. Countries with higher social fragmentation, and

those that are more reliant on foreign aid inflows, tend

to have a more procyclical government‐consumption

policy. Better governance promotes countercyclical‐

fiscal policy while increased democracy dampens it.

Finally, some fiscal rules are important in curbing the

procyclical behaviour of government consumption.

KEYWORDS

financial constraints, government consumption, inequality,

institutions, time‐varying coefficients, weighted least squares

JEL CLASSIFICATION

C22; C23; H50; H60; H62

1

|

INTRODUCTION

The Global Financial Crisis and its legacies have put fiscal policy at the centre of the debate on the

policy mix required to help steering economies towards a stable growth path. In addition to the

allocation and distribution roles, fiscal policy is also responsible for the macroeconomic stabilization

against business cycle fluctuations (Musgrave, 1959). Pressing policy challenges have revived the

debate on the effectiveness of fiscal policy as a stabilizer (Botman, Laxton, Muir, & Romanov, 2006;

van der Ploeg, 2005).

1

A growing number of countries have turned to fiscal policy as their primary

stabilization tool either because of changes in their monetary regime, or because financial

conditions deteriorated to the point of making monetary policy ineffective (Spilimbergo, Symansky,

Blanchard, & Cottarelli, 2008). As economic conditions normalize, policymakers expect to rely on

fiscal stabilizers to prevent short‐term setbacks and dampen any future volatility, which in turn is

known to negatively affect medium‐term growth (Ramey & Ramey, 1995).

2

Against this background, understanding how government expenditure fluctuates with business

conditions is extremely relevant from a policy‐making perspective. Expenditure patterns may

change due to policymakers’discretionary actions (even though the majority of the empirical

literature suggests a procyclical bias of these measures—van den Noord, 2002) or as a result of the

operation of automatic stabilizers (by lessening the effects of the liquidity constraints faced by

households and alleviating the impact of exogenous shocks on aggregate current consumption and

output—Granado, Gupta, & Hajdenberg, 2013). Government spending has a stabilizing effect if it

increases when output growth rises and falls when output growth declines (Furceri, 2010).

3

The

higher the countercyclical government spending is, the greater its stabilizing effect.

Most of the empirical literature looking at the cyclical properties of government expenditure

typically uncover (a) an acyclical or countercyclical behaviour in advanced countries (see e.g.,

Hallerberg and Strauch, 2002); and (b) a procyclical pattern in developing countries (Alesina &

Tabellini,2005;Gavin,Hausmann,Perotti,&Talvi,1996;Kaminsky,Reinhart,&Vegh,2004).A

number of explanations have been advanced to justify the different cyclical patterns in various

groups of countries (see Section 2 for more details). Moreover, most empirical studies on this topic

can be split into two: (a) those that document the cyclical properties of fiscal policy and/or its

components and (b) those that inspect their determinants. The overwhelming majority of papers

have focused (due to data‐related issues) on empirical analyses of European or Organisation for

Economic Co‐operation and Development (OECD) countries, with a few exceptions (see e.g.,

Thornton (2008) for African countries, which serves as the key reference for this paper).

In this paper, we ask two main questions. First, how stabilizing is the government’s

consumption in African countries and how has its cyclicality varied over time, between

countries and around business cycles’turning points? Second, which macroeconomic, financial,

institutional and political variables determine the degree of the cyclicality of government

consumption in this group of African countries? We make a positive contribution to the debate

by answering the two research questions using a novel empirical strategy. Specifically, we

estimate time‐varying measures of government‐consumption cyclicality for a panel of 46

African countries between 1960 and 2014.

4

To the best of our knowledge, this is the first paper

that estimates time‐varying measures of government‐consumption cyclicality for a large set of

African economies.

5

In addition, we examine which are the most relevant determinants of time‐

varying measures in government‐spending cyclicality. The use of time‐varying measures in

government‐spending cyclicality overcomes the key limitation of previous studies assessing the

drivers of fiscal cyclicality that rely on cross‐country regressions which are not able to account

for country‐specific, as well as global, factors.

298

|

JALLES