Tax(i)ing the Poor? Commuting Costs in South African Cities

• Author: Andrew Kerr
• DOI: http://doi.org/10.1111/saje.12161
• Published date: 01 September 2017
• Date: 01 September 2017

TAX(i)ING THE POOR? COMMUTING COSTS

IN SOUTH AFRICAN CITIES

ANDREW KERR*

Abstract

In this paper, I describe the monetary and time costs of commuting to work in South African

cities, and how these have changed in the post-Apartheid era. I interpret these results in light of a

paper by Brueckner, who used a simple urban model to suggest that location and commuting

patterns by race could change as a result of the repeal of Apartheid era legislation such as the

Group Areas Act that made it impossible for black South Africans to live near the centre of cities.

A key finding is that monetary and time costs of commuting in South African cities are high and

have increased in the post-Apartheid period. Journey times are much higher than the OECD

country average and a sample of developing country cities. Part of the explanation for increasing

average commute times is population growth in South African cities, which has been substantial.

Both the population and number of commuters doubled between 1993 and 2013 in the six met-

ropolitan municipalities (metros) analysed in this paper. Since commuting costs a substantial

amount and generates negative externalities through congestion I also explore the financing of

public transport.

JEL Classification: R41, R4, R, R48,

Keywords: Commuting costs, Urban economics, South African cities

1. INTRODUCTION

According to the 2003 National Travel Survey, black South Africans spent an average of

88 min per day commuting to and from work. This was just less than double the average

commute time both in the United States in 2002, a country known for its long commut-

ing times, and Hungary, the European Union country with the longest commute times

(Stutzer and Frey, 2008). White South Africans spent less time commuting than black

South Africans, 54 min per day, but still spent more time commuting on average than

commuters in the United States and all 23 EU countries listed by Stutzer and Frey

(2008). Ten years later, the 2013 South African National Travel Survey suggested that

average commuting times had increased by a further 14 min for both black and white

South Africans.

South African cities have been shaped to a large extent by policies of segregation from

before and during the Apartheid era. This has meant that wage work opportunities are

often located far from home, particularly for black South Africans, and this results in

high monetary and time costs of commuting. Brueckner (1996) used a simple urban

model to show that the repeal of the Group Areas Act and other legislation keeping

black, coloured and Indian South Africans out of central locations in cities might radi-

cally alter racial location and commuting patterns as the land market adjusted to a new

* Corresponding author: DataFirst, University of Cape Town, 3rd floor, School of Economics

Building, Middle Campus, Rondebosch, South Africa. E-mail: andrew.kerr@uct.ac.za

V

C2017 Economic Society of South Africa. doi: 10.1111/saje.12161

321

South African Journal of Economics Vol. 85:3 September 2017

South African Journal

of Economics

equilibrium. The key prediction from this model is that poorer black commuters would

displace whites in city centres where jobs were located, decreasing commuting times for

blacks and increasing them for whites.

Brueckner’s urban model incorporated the intra-urban restrictions on residential loca-

tion in Apartheid South Africa but ignored the constraints on rural to urban migration

that were also a feature of Apartheid, and thus abstracted from population growth in the

post-Apartheid period. The population in the six metros I investigate in this paper

doubled over the 20 years between 1993 and 2013, and the more general pattern was of

rapidly increasing urban population growth in the post-Apartheid period after the repeal

of legislation that made rural to urban migration very difficult for black South Africans.

In this paper, I document commuting times, costs and mode usage in South African

cities, how they have changed during the post-Apartheid period using nationally repre-

sentative household survey data and interpret these changes in light of Brueckner’s

predictions.

Inspired by Hausmann (2013) and Brueckner (1996) I also investigate whether long

and costly commutes on public transport can be thought of as high tax rates on their

hourly earnings, find that these are indeed substantial. Long and expensive commutes to

formal sector employment should make working in informal self-employment (predomi-

nantly an activity carried out near or at home) much more attractive, only adding to the

puzzle of why South Africa has a very small informal sector amidst high unemployment

(Kingdon and Knight, 2004).

Since commuting costs are high and rising, and because commuting imposes negative

externalities, I investigate the subsidisation of public transport. I document the discon-

nection between the direct public transport subsidies paid by the state and the use of

public transport by commuters. Seventy-one percent of public transport users used mini-

bus taxis to get to work in 2013 but minibus taxis received only about 1% of the total

public transport subsidy, none of which is paid for operating expenses.

2 LITERATURE REVIEW

2.1 Comparative Data on Transport Times and Costs

A number of studies have identified South Africa as having very high time costs of com-

muting relative to other countries. The 2011 OECD study “How’s life” used data from

time use surveys in 24 countries to compare commuting times.

1

South Africa had the

highest average commute times, of 56 min per day, compared to the average of 38 min

per day.

The times reported for South Africa in the introduction are much higher than those

in the OECD study using time use survey data. The same thing is true for a number of

other countries when comparing the results from Stutzer and Frey (2008) and OECD

(2011). The three nationally representative South African surveys used in this paper also

suggest that average commuting times in South Africa are actually higher than the times

calculated from the South African Time Use Survey used in the OECD study, between

68 and 94 min per day on average. Rospabe and Selod (2006) also report much higher

1

These were mostly European countries as well as the United States, Canada, Japan, Turkey and

Korea. Surveys were mostly from around 2000 but some were as late as 2009. The South African

data was from the 2000 Time Use Survey.

322 South African Journal of Economics Vol. 85:3 September 2017

V

C2017 Economic Society of South Africa.