The Invisible Hand of Rain in Spending: Effect of Rainfall‐Driven Agricultural Income on Per Capita Expenditure in Ghana

• DOI: http://doi.org/10.1111/saje.12131
• Date: 01 March 2017
• Author: Eric Akobeng
• Published date: 01 March 2017

THE INVISIBLE HAND OF RAIN IN SPENDING: EFFECT OF

RAINFALL-DRIVEN AGRICULTURAL INCOME ON PER

CAPITA EXPENDITURE IN GHANA

ERIC AKOBENG*

Abstract

This paper uses a nationally representative household pseudo-panel dataset for Ghana, a

rain-fed agriculture economy, to investigate whether there is a positive relationship between

rainfall-driven agricultural income and household per capita expenditure. By using the Two

Stage Least Squares Instrumental Variable (2SLS-IV) estimator, it is found that a fall in

rainfall-driven agricultural income leads to a decrease in per capita expenditure. The results

show that the gender and the locality of the household head matter in the response of per

capita expenditures to rainfall-driven agricultural income. Female-headed and rural house-

holds are more vulnerable to rainfall-driven agricultural income changes. The expenditure

disaggregation indicates that female-headed households significantly reduce per capita non-

food expenditure in times of rainfall-induced agricultural income decrease whilst the response

of male-headed households focuses more on reducing per capita food and remittance

expenditures.

JEL Classification: D12, Q18

Keywords: Per capita expenditure, agricultural income, rainfall

1. INTRODUCTION

Shocks affecting agricultural production such as rainfall variation have been considered

to be a major factor having a negative effect on the welfare of households in developing

countries (Dercon and Krishnan, 2000). How to tackle poverty emanating from house-

holds vulnerability to risks associated with income changes is a major issue in developing

economies. Agriculture has long been an important sector in the Ghanaian economy. It

contributed 40% of Ghana’s GDP in the 1990s (GSS, 2013). In 2013, agriculture, serv-

ice and the industrial sectors contributed 21.6%, 50.6% and 27.7% to GDP, respectively

(GSS, 2015). The provisional estimates for 2014 have 19.9%, 51.7% and 28.4% as the

contribution of the agriculture, service and the industrial sectors to GDP, respectively

(GSS, 2015).

1

The contribution of agriculture to foreign exchange earnings averaged

45% in the 1990s and 40% in the 2000s (GSS, 2013).

* Corresponding author: PhD Student/Graduate Teaching Assistant, Department of Economics,

University of Leicester, University Road, Leicester, LE1 7RH, United Kingdom.

E-mail: ea211@leicester.ac.uk or eakobeng@yahoo.co.uk

1

The agricultural sector has experienced a decline of its contribution to total GDP. This is due to

increased performance of the transport and storage sub-sectors of the service sector. The improved

performance of the industrial sector is also due to the growing importance of petroleum in Ghana

(ISSER, 2012; MOFA, 2013).

V

C2016 Economic Society of South Africa. doi: 10.1111/saje.12131

98

South African Journal of Economics Vol. 85:1 March 2017

South African Journal

of Economics

The total land area of Ghana is 23,884,245 hectares with 14,038,224 hectares (59%)

being agricultural land. The agricultural land area is made up of 7,847,300 hectares

(56%) under cultivation and 6,190,924 hectares (44%) not under cultivation. Only

0.4% of the agriculture land area under cultivation is irrigated (MOFA, 2013). Thus,

changes in rainfall levels are associated with variations in production and farm incomes.

The instability of agricultural production due to rainfall shocks implies a huge burden

for the low-income households, in the environment of inadequate national social security,

credit and insurance systems provided by the government (Asiimwe and Mpuga, 2007).

Some studies consider the direct effect of shocks on consumption.

2

Other microeco-

nomic studies use weather as an instrument for income in looking at the effect of income

shocks on the outcomes of remittance income, murder and savings.

3

Some macroeco-

nomic studies also adopt rainfall as an instrument for income or growth in looking at

civil conflict and democracy.

4

However, the literature on the effect of agricultural income

on per capita consumption and remittance expenditures is lacking.

This paper contributes to the existing literature by providing an empirical evidence of

how rainfall-driven agricultural income affects per capita consumption and remittances

expenditures from the gender of the household and locality perspectives. Exogenous vari-

ation in rainfall is used as an instrument for agricultural income in order to estimate the

effect of agricultural income on per capita expenditure. By using daily rainfall data from

50 rainfall stations and household pseudo-panel datasets for 1991/1992, 1998/1999 and

2005/2006, the results provide that agricultural income is positively related to per capita

household expenditure.

The paper is organised as follows. Section 2 provides the overview of agriculture, flood

and expenditure in Ghana. Section 3 reviews the findings of related studies. Section 4

outlines the hypothesis and research questions. Sections 5 and 6 present the dataset and

the econometric methodology. Sections 7 and 8 present the results and conclusion,

respectively.

2. OVERVIEW OF AGRICULTURE, FLOOD/DROUGHT AND EXPENDITURE IN

GHANA

The 2010 population and housing census of Ghana put the country’s population at

24,658,823 with a population growth rate of 2.5% per annum (GSS, 2013). The Ghana

Statistical Service

5

projected the population of Ghana in 2014 at 27,043,093. The popu-

lation distribution varied across the 10 administrative regions of the country with 51%

and 49% living in the urban and rural areas, respectively. Ghana has a total of 5,467,136

households, with 2,503,006 of these being agricultural households, representing 46% of

the total households.

6

Urban and rural households are 56% and 44% of the total

2

For discussion on this see Dercon and Krishnan (2000), Dercon (2004) and Dercon et al.

(2005).

3

For an interesting discussion on these see Paxson (1992), Miguel (2005) and Yang and Choi

(2007).

4

For a review see for example Miguel et al. (2004) and Bruckner and Ciccone (2011).

5

Refer to http://www.statsghana.gov.gh/pop_stats.html.

6

Household is defined as “a person or group of persons, who live in the same house or com-

pound, share the same housekeeping arrangements and recognise one person as the head of the

99South African Journal of Economics Vol. 85:1 March 2017

V

C2016 Economic Society of South Africa.