Concepts and Measures of Saving: Selected Issues for South Africa

• Author: Anna Orthofer
• Published date: 01 June 2017
• DOI: http://doi.org/10.1111/saje.12129
• Date: 01 June 2017

CONCEPTS AND MEASURES OF SAVING: SELECTED

ISSUES FOR SOUTH AFRICA

ANNA ORTHOFER*

Abstract

South African household savings rates have been declining steadily over the last f‌ive decades,

raising concerns that the population structurally under-saves. Against the background of new

saving-enhancing policy initiatives, this paper asks to what extent the concern is founded, and

whether the measurement of saving is really appropriate to guide economic policy. Comparing

different macroeconomic concepts and measurements of saving, we show that the measure of sav-

ing in the national accounts (the residual between income and expenditure) understates the

household savings rate compared to other measures. Specif‌ically, an alternative measure from the

balance sheets (the change in wealth) yields a signif‌icantly higher and non-declining f‌igure. While

households have not been “putting aside” their incomes, they have nevertheless grown richer,

driven largely by the appreciation of asset valuations. We also examine the impact of taking non-

f‌inancial saving and wealth into account, and conclude that household sector saving on the aggre-

gate is signif‌icantly higher than the national accounts suggest. However, these adjusted measures

are most relevant for the upper tail of the income and wealth distribution, raising important dis-

tributional concerns.

JEL Classif‌ication: E01, E21

Keywords: Saving, wealth, measurement and data, national income accounting

1. INTRODUCTION

A look at the South African national accounts suggests that South African households save

worryingly little. Household savings rates have been declining steadily over the last 50

years, from about 10% of GDP to almost nothing today. In net terms, household savings

rates have been in negative territory for almost a decade, meaning that they are insuff‌icient

to replace even existing capital, let alone fund new investment (see Fig. 1). While low lev-

els of saving are widespread across sub-Saharan Africa, the South African situation is in

stark contrast to that of the fast-growing Asian economies, where household savings rates

in excess of 20 or 30% of GDP have often been associated with the successful transition

to more dynamic economies (Commission on Growth and Development, 2008).

In response to these trends, the South African government has launched a number of

initiatives to enhance the country’s saving culture, most recently through the introduction

of tax-free savings accounts in March 2015. The government’s concern about saving is

* Corresponding author: PhD Candidate, Department of Economics, Stellenbosch University,

Matieland, Stellenbosch, Western Cape 7602, South Africa. E-mail: anna.orthofer@gmail.com

This work was supervised by Stan du Plessis, Dieter von Fintel and Monique Reid, whom I

wish to thank for extensive discussions and invaluable comments. I am grateful for f‌inancial

support from Economic Research Southern Africa (ERSA-2015-0084).

V

C2016 Economic Society of South Africa. doi: 10.1111/saje.12129

222

South African Journal of Economics Vol. 85:2 June 2017

South African Journal

of Economics

founded on both micro- and macro-level considerations. On the household level, the

low savings rate means that households are vulnerable to unexpected losses or expenses,

and will face challenges to maintain their living standards during retirement. On the

aggregate level, the failure of the household sector to contribute to the aggregate savings

pool increases the reliance on foreign capital inf‌lows to f‌inance domestic investments.

This is thought to carry the risks of higher macroeconomic volatility and a less competi-

tive market structure, as larger companies have better access to foreign capital than

smaller ones (National Planning Commission, 2012).

Given the prominence of the savings debate in South Africa, it is surprising how little

attention is currently paid to understanding what exactly we are measuring as saving in the

national accounts. Saving is by no means an unequivocally def‌ined concept, and different

measures yield vastly different results. The most commonly used metric (according to which

household saving is nil or negative) stems from the national income statements, where sav-

ing is calculated as the residual between disposable income and consumption expenditures.

If, instead, we look at the balance sheet side of the national accounts, we f‌ind that real

household wealth increased by 5-10% of GDP per year over the last decades. Our assess-

ment of the level of household saving also changes when we consider investments in physi-

cal, human and environmental capital in addition to pure f‌inancial saving and wealth; an

adjustment that would add another three to seven percentage points to the conventionally

measured savings rate.

Although one specif‌ic measure of saving dominates contemporary empirical analy-

ses, it is not clear that it is the one measure that corresponds most closely to the theo-

retical concepts we try to investigate. Nor would that be true for the alternative

measures of saving. Instead, the variety of questions asked by economists and eco-

nomic policymakers – whether low levels of household saving cause a reliance on for-

eign capital, to what extent households are prepared to absorb unexpected losses and

expenses in the short run or retire in the long run, and whether their savings behav-

iour corresponds to our models, to name but a few – calls for different concepts and

measures of saving.

Figure 1. Saving rates

Notes: Saving rates by institutional sector, net, % of GDP. Annual data.

Source: South African Reserve Bank, author’s calculations.

223South African Journal of Economics Vol. 85:2 June 2017

V

C2016 Economic Society of South Africa.