0“
Jul
. x.‘ I...».
la ....U:

 

.l ‘n. I‘MQ .

 

Ari/o; \o.»csu ...

 

. . V 1L0..-
3 .-.! .l.--¢{

. . .th‘WI
. . I . f ‘ .....- ‘ . {Les wlt.hv.1n‘ll‘l.;-l
4......u...ka.....?tL:fl {Vitamnnmmrhwk .....1.. Janinnuuu.‘ . .1 In. A

is v‘ .7 iv Julu‘; , .1 V!

 

TYUBHARIES

12““le

 

 

 

 

 

 

 

“E“ Illlllllllll“WWW\Illl‘lllll

C" ‘ 3 1293 01712

l i 2;.

This is to certify that the

dissertation entitled

Determinants of Credit Constraints on Micro and Small
Enterprises in the Northern Province of South Africa

presented by

Charles Lepepeule Machethe

has been accepted towards fulfillment
of the requirements for

Ph.D. Agricultural Economics

degree in

 

 

(initiate-E

Major professor

DateJSéFr-CIJQQ’.’ 7 WW

 

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

{LIBRARY

Michigan State
University

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINE return on or before date due.

 

DATE DUE

MTE DUE

DATE DUE

 

q “l h .2 u A.
We ‘3“ U 2%

DEC 22 2003

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1!” W969.“

 

DETERMINANTS OF CREDIT CONSTRAINTS ON MICRO AND SMALL
ENTERPRISES IN THE NORTHERN PROVINCE OF SOUTH AFRICA

By

Charles Lepepeule Machethe

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

1 997

ABSTRACT

DETERMINANTS OF CREDIT CONSTRAINTS ON MICRO AND SMALL
ENTERPRISES IN THE NORTHERN PROVINCE OF SOUTH AFRICA

By

Charles Lepepeule Machethe

There is a widespread belief in South Africa that the contribution of micro and
small enterprises (MSEs) to employment and income generation is limited by credit
constraints. However, there is limited knowledge of the significance and
determinants of such credit constraints. This study is concerned with determining
the proportion of MSEs that are credit constrained and identifying determinants of
credit constraints on MSEs in the Northern Province of South Africa. A logit
framework, using data obtained from 270 peri-urban and rural MSEs, is employed
to identify determinants of credit constraints on MSEs. The estimation of the
proportion of credit constrained MSEs and identification of determinants of credit
constraints are done for the overall credit market, formal credit market and informal
credit market.

An important finding of this study is that many but not most MSEs included
in the analysis are credit constrained. Forty-eight percent of all the MSEs in the

sample are credit constrained in the overall credit market. The proportions of credit

constrained MSEs in the formal and informal credit markets are 30 and 42 percent,
respectively.

The results of the study indicate that the most important determinants of
credit constraints in the overall credit market are household/business wealth,
location of the business, and the economic sector in which it operates. MSEsthat

are more lilgelxi9939r99it.999.§tr.ained inihe éIElljereditmadeiare (a) from poor

 

households; (b) located in rural areas; and (c) in the manufacturing sector. The

most important determinants of credit constraints in the formal Icredit market are

—-_.__.. ..v—puu-o
Hb;.,_+_.._' _,«

éducatio>andéender of the MSE operator, and the(economic sectchn which the
MSE operates. MSEs in the<rnanufacturing sectoll. and those operated (y less
educated and male entrepreneurspre more likely to be credit constrained in the
Wefidit mariget. Determinants of credit constraints In the Woredlt markeL
are @usehold/business wealth) whether the firm is 6ffmially registered/ as a
business, and (cationflf the business. MSEs front poor households, hot officially

registered/as business concerns and located in rural areasare more likely to

experience credit constraints in the"infornEl‘_credit market.

 

ACKNOWLEDGMENTS

The task of writing a dissertation is accomplished only with the assistance
and guidance of a number of people. I want to thank all those who have contributed
to the completion of this long and, sometimes, lonely journey.

I express my deepest appreciation for the support and guidance provided by
my dissertation advisor, Dr Donald Mead. His unfailing assistance and dedication
are greatly acknowledged. The encouraging remarks he made on the many draft
versions of the dissertation helped to keep me going. I have also benefitted
tremendously from Dr Thomas Reardon, who played the role of co-dissertation
advisor. He helped me to reach a decision on what the focus of my research should
be by asking his most favorite question “What is the research question?” He was
always ready and willing to read and comment on the draft versions of the
dissertation. It is difficult to find appropriate words to express my appreciation for
the support and guidance I received from Dr Carl Eicher, my major professor. He
was more than just a major professor. He was very helpful from the first day I met
him in Namibia until the last day of my stay in East Lansing. I also thank his wife,

Shirley, for her support to me and my family.

I acknowledge my sincere appreciation to the other members of my guidance
committee for their assistance and advice, especially during the final stages of the
preparation of the dissertation. They are Drs Carl Liedholm, Allan Schmid and John
Strauss. Their challenging questions made me think more carefully about things that I took
for granted.

This dissertation would not have been completed without the support, sacrifice and
understanding of my family, especially my wife, Nobesutu. On many occasions, she has
had to manage the family alone. Our children, Thabo, Phomolo and Hlaudi, took a while
to understand how I could be a student when it was their turn to go to school. However,
they eventually understood and accepted that for several years they would see little of me
at home.

I am grateful to The W K Kellogg Foundation for providing financial support for me
and my family during my graduate studies. I also thank my employer, University of the
North, for granting me a study leave to pursue my graduate studies.

During the last stages of my doctoral studies when I had to return to Michigan State
University, I was fortunate to be offered accommodation by Tom Reardon and Bonnie
Banks in their house. I am grateful to them for their generosity and hospitality. I also thank
Dr Larry Hamm, Chairperson of the Department of Agricultural Economics, for his efforts
to assist me to complete the dissertation.

I acknowledge the assistance received from Pat Eisele during my graduate studies.
I also wish to thank Pat and Sherry Rich, for attending to the technical aspects of the
dissertation. Thanks a million Pat and Sherry. Last, but not least, I thank Vera Burger for

typing parts of the dissertation.

TABLE OF CONTENTS

CHAPTER ONE — INTRODUCTION ................................. 1

1.1 General problem ...................................... 1

1.2 Specific problem ...................................... 5

1.3 Research questions ................................... 7

1.4 Hypotheses .......................................... 8

1.5 Outline of subsequent chapters ......................... 11
CHAPTER TWO — BACKGROUND ON THE NORTHERN PROVINCE, THE

SMALL BUSINESS SECTOR AND SOURCES OF CAPITAL ............. 13

2.1 Geographical location ................................. 13

2.2 Socio-economic characteristics .......................... 14

2.3 The small business sector in South Africa ................. 14

2.4 Sources of capital for MSEs ............................ 17

2.4.1 Credit sources ................................. 17

2.4.1.1 Formal lenders ...................... 18

2.4.1.2 Semi-formal lenders .................. 20

2.4.1.3 Informal lenders ..................... 20

2.4.2 Noncredit sources ............................... 21

CHAPTER THREE — CONCEPTUAL FRAMEWORK .................. 23

3.1 Credit in the context of other sources of capital ............. 23

3.2 What is a credit constrained firm? ........................ 24

3.3 Determinants of firm-specific demand for credit ............. 29

3.4 Determinants of firm-specific supply of credit ............... 33

3.5 Determinants of firm-level credit constraints ................ 35

3.5.1 Imperfect information ............................ 36

3.5.2 Transaction costs ............................... 40

3.5.3 Institutional constraints ........................... 41

3.5.4. Insufficient collateral ............................. 42

3.5.5 Credit constraints and economic efficiency ........... 42

CHAPTER FOUR — DATA SOURCES AND CHARACTERISTICS

OF THE SAMPLE ............................................... 44
4.1 Data sources ........................................ 44
4.2 Characteristics of the sample ........................... 47

vi

4.2.1 Households and MSE operators .................... 47

4.2.2 MSEs ........................................ 49
4.2.3 Households and MSEs according to location .......... 51
4.2.5 Households and MSEs according to economic sector . . . 53
4.2.6 Sources of capital for MSEs ....................... 55
4.2.6.1 Sources of capital for establishment
of business ......................... 55
4.2.6.2 Sources of capital for financing assets . . . . 56
4.2.6.3 Sources of capital for purchasing
inputs/raw materials .................. 57

CHAPTER FIVE -— REVIEW OF METHODS FOR IDENTIFYING

CREDIT CONSTRAINED FIRMS ................................... 59
5.1 Introduction ......................................... 59
5.2 Full-sorting approaches ................................ 62
5.3 Partial-sorting approaches ............................. 67
CHAPTER 6 — REGRESSION SPECIFICATION, CLASSIFICATION OF
FIRMS AND ESTIMATION PROCEDURE ...................... 70
6.1 Regression model and specific hypotheses ................ 70
6.2 Method used in this study for classification of MSEs ......... 79
6.2.1 Steps in classifying MSEs as credit constrained
or unconstrained ..................................... 80
6.3 Estimation procedure ................................. 84
CHAPTER SEVEN — RESULTS OF THE STUDY ..................... 88
7.1 Proportion of credit constrained MSEs .................... 88
7.1.1 Overall credit market ............................ 88
7.1.2 Formal credit market ............................. 90
7.1.3 Informal credit market ............................ 91
7.2 Proportion of credit unconstrained MSEs .................. 92
7.3.1 Peri-urban areas ................................ 95
7.3.2 Rural areas ................................... 98
7.3.3 Construction sector ............................. 99
7.3.4 Manufacturing sector ........................... 103
7.3.5 Service sector ................................ 103
7.3.6 Trade sector ................................. 103
7.3.7 Ranking of sources of capital for MSEs ............. 104
7.4 Determinants of credit constraints ....................... 107
7.4.1 Overall credit market ........................... 107
7.4.2 Formal credit market ............................ 114
7.4.3 Informal credit market ........................... 116

CHAPTER EIGHT — SUMMARY, CONCLUSIONS AND POLICY ..
IMPLICATIONS ............................................... 12%

8.1 Summary of the study ................................ 123
8.1.1 Background .................................. 123
8.1.2 Purpose of the study ........................... 124
8.1.3 Data sources and methodology ................... 124
8.1.4 Significance of credit constraints .................. 125
8.1.5 Determinants of credit constraints ................. 126
8.2 Implications for policy ................................ 128
8.2.2 Peri-urban versus rural MSEs .................... 130

8.2.3 Differentiated approaches for different credit markets . . 132
8.3 Shortcomings of the study and implications for future

research .......................................... 134
8.3.1 Lack of representativeness of the sample ........... 134
8.3.2 Lack of information on leverage (debt information) . . . . 135
8.3.3 Little emphasis on explanations for determinants
of credit constraints ............................ 135
8.3.4 Notional versus effective demand ................. 136
8.3.5 Further breakdown of economic sector .............. 136
8.3.6 Effects of credit constraints ....................... 136
APPENDIX A ................................................. 138
BIBLIOGRAPHY ............................................... 163

viii

Table 4.1:
Table 4.2:
Table 4.3:
Table 4.4:
Table 4.5:

Table 4.6:

Table 4.7:

Table 4.8:

Table 4.9:

Table 4.10:
Table 7.1:
Table 7.2:

Table 7.3:

Table 7.4:

LIST OF TABLES

Number and proportion of respondents (MSEs) per region . . . . 46
MSEs according to economic sector ...................... 46
Household and MSE operator characteristics ............... 48
Characteristics of MSEs ............................... 50

Characteristics of households and MSEs according to location . 52

Characteristics of households and MSEs according to
economic sector ..................................... 54

Ranking of the most important source of capital for
establishing MSEs ................................... 56

Participation in the credit market to purchase assets ......... 56

Participation in the supplier credit market to purchase

inputs/raw materials .................................. 57
Reasons for not applying for supplier credit ................ 58
Proportion of MSEs according to credit status .............. 89
Proportion of credit unconstrained MSEs .................. 94

Characteristics of MSEs and households according to
credit status ......................................... 96

Characteristics of MSEs and households in peri-urban
areas according to credit status ......................... 97

Table 7.5:

Table 7.6:

Table 7.7:

Table 7.8:

Table 7.9:

Table 7.10:
Table 7.11:

Table 7.12:

Table 7.13:

Table 7.14:

Table 7.15:

Table 7.16:

Characteristics of MSEs in rural areas according to
credit status ........................................ 100

Characteristics of MSEs and households in the construction
sector according to credit status ........................ 101

Characteristics of MSEs and households in the manufacturing
sector according to credit status ........................ 102

Characteristics of MSEs and households in the service
according to credit status ............................. 105

Characteristics of MSEs and households in the trade sector

according to credit status ............................. 106
Ranking of sources of funds for establishing MSEs ......... 107
Determinants of credit constraints in the credit market ....... 112

Determinants of credit constraints in the credit market,
including region variables ............................. 113

Determinants of credit constraints in the formal credit market . 118

Determinants of credit constraints in the formal credit
market, including region variables ....................... 119

Determinants of credit constraints in the informal credit
market ........................................... 1 21

Determinants of credit constraints in the informal credit
market, including region variables ....................... 122

LIST OF FIGURES

Figure 3.1. Access and quantity rationing ........................... 27
Figure 3.2. Credit constrained and unconstrained firms ................ 28

Figure 3.3. Credit unconstrained when demand equals zero and
supply equals demand ................................ 29

xi

CHAPTER ONE

INTRODUCTION

This chapter outlines the socio-economic problems confronting South Africa
and highlights the emphasis placed on the promotion of micro and small enterprises
(MSEs) as a partial solution to the problems. The research problem, research

questions and hypotheses of the study are also described in this chapter.

1.1 General problem

South Africa is experiencing problems of high unemployment, skewed
income distribution and poverty. The problem of unemployment can be attributed
mainly to previous government policies that favored capital-intensive production
(World Bank, 1994b) and inadequate attention to employment generation in the
former homelands. Apartheid policies that discriminated against the majority of the
population and favored the white population have contributed to poverty and
created the most unequal distribution of income in the world.

Among Africans, only 47 percent are employed in the formal sector (World
Bank, 1994b). The formal sector is no longer able to provide employment for a
large number of people seeking employment. The labor absorption capacity of the

formal sector in urban areas has declined rapidly. In 1965-70 the urban formal

2

sector could absorb 75 percent of annual labor entrants but the absorption capacity
dropped to only 12.5 percent in 1985-89 (Ligthelm and Kritzinger-van Niekerk,
1990). The inability of the urban formal sector to absorb a large proportion of labor
entrants contributes significantly to unemployment in the former homelands. Many '
of the unemployed are in the former homelands, with women experiencing higher
rates of unemployment than men. Men who are less than twenty-five years of age
have greater difficulty in finding employment (World Bank, 1994b).

Approximately 50 percent of South Africa's population of forty million can be
classified as living below the poverty datum line (SAIRR, 1993). More than 36
percent of all households are classified as poor (Data Research Africa, 1995).
Poverty is more pervasive in rural areas, with the incidence of poverty being the
same in the rural areas of the former white areas and the former homelands (World
Bank, 1994b). The proportion of the poor living in rural areas of South Africa is
estimated at 70 percent (Government of South Africa, 1995). Approximately 50
percent of all rural households and 68 percent of rural people are poor. In urban
and metropolitan areas, 39 and 17 percent of the population, respectively, can be
classified as poor (Data Research Africa, 1995). The incidence of poverty in rural
areas is highest among Africans: 71 percent of poor rural households are African.
Carter and May (1997) note that 77 percent of the poorest decile of the population
of South Africa are Africans living in mral areas and control just over one percent
of household and adult equivalent expenditure. The wealthiest ten percent of South

African households account for 40 percent of the household and adult equivalent

3

expenditure and only four percent of these households are Africans living in rural
areas.

South Africa has the w'orld’s most unequal distribution of income (Deininger
and Squire, 1996). The per capita incomes of whites in 1987 was on average 9.5
times that of Africans. Gini coefficients are estimated at 0.69 for aggregate income
in 1980, 0.82 for farmland ownership in 1989, and 0.82 in the manufacturing sector
in 1982 (Bureau for Market Research, 1993; World Bank, 1994b). In 1993, the Gini
coefficient is estimated to be 0.65 (see Blumenfeld, 1997). These coefficients are
among the highest in the world.

The alleviation of poverty, unemployment and income inequalities (problems
inherited from the previous apartheid government) are major priorities of the South
African government. The government's plans to deal with these problems are
outlined in a document entitled WWW
(RDP) (African National Congress, 1994). The RDP emphasizes the role of micro
and small enterprises in poverty alleviation and narrowing the income gap between
different race groups through employment and income generation. The potential
contribution of MSEs to equitable income distribution and employment and income
generation in South Africa is also acknowledged in the White Paper of the
Department of Trade and Industry (Department of Trade and Industry, 1995).

The role of MSEs in employment creation and poverty alleviation is, thus,
undisputable. However, to date, the contribution of MSEs to employment and
income generation and poverty alleviation in South Africa is considered to be

inadequate. The limited contribution of MSEs may be attributed to constraints they

4

face. These constraints include insufficient capital and tend to be severe in rural
areas, especially among women (Department of Trade and Industry, 1995).
Capital for MSEs can be accumulated from various sources of household
income, including wage employment in farm and nonfarm sectors, remittances from
migrants, agriculture, self-employment in farm and nonfarm sectors, old-age
pension, and loans from informal, formal and semifonnal lenders. Carter and May
(1997) identify remittances from migrant workers, employment in the secondary
market‘, and agricultural production as the three most important (in terms of the
proportion of households that indicated that they derive income from the various
sources) sources of income for mral households. However, income from agriculture
for rural households is relatively small (Carter and May, 1997; May 1996). Old-age
pensions is also an important source of capital for MSEs. Ardington and Lund
(1996) note that old-age pensions are of critical importance to household income.
Some MSE operators manage more than one business and income from one
business can be reinvested in the other business. Another source of capital for
MSEs is credit from formal (e.g. commercial banks, development corporations, etc),
semi-formal (e.g. NGOs) and informal lenders (family and friends, traders, etc).
The high unemployment rate and the declining labor absorption capacity of

the formal sector in South Africa imply that the proportion of households generating

 

I Burawoy (1975) distinguishes between primary and secondary labor

markets. Primary labor market refers to that labor market in which workers are well
paid and have job security and prospects of career advancements. The secondary
labor market is defined as that labor market with low-paying jobs and little job
security and career advancement.

5

income from wage employment and remittances is likely to decline. Thus, the
relative importance of wage employment and remittances as sources of household
income will decrease. The dependence of smallholder agriculture on erratic rainfall
means that rural households cannot rely on agriculture as the main source of
income. Income from old-age pensions is inadequate for the financing needs of
households and MSEs. Capital accumulation from self-employment in other
businesses is likely to be limited because of the small number of MSE operators
who can afford to own more than one business. Therefore, in addition to other
efforts to solve the problem of insufficient capital, it becomes important to pay
attention to credit as a possible tool for removing capital constraints on MSEs.
Credit has the potential to alleviate capital constraints on MSEs and, thus, enable
them to increase their contribution to income and employment generation and

poverty alleviation.

1.2 Specific problem

Many analysts and policy makers in South Africa believe that credit
constraints are among the major constraints limiting the contribution of MSEs to
employment, income generation and poverty alleviation (African National Congress,
1994; Department of Trade and Industry, 1995; Coetzee et al., 1994). Credit
constraints occur in a firm when its effective demand for credit exceeds its supply
of credit at the rate of interest prevailing in the market. Effective demand for credit
is demand which reflects wants translated into action (which is borrowing coupled

with the ability to repay a loan), not mere intention or desire to borrow (notional

6

demand). The supply of credit to the firm refers to the schedule of quantities of
credit lenders are willing to make available to the firm at prevailing rates of interest.

Credit constraints are reported to be the most persistent complaint of
operators of small businesses in South Africa (Moore and Schoombee, 1995; World
Bank, 1993; Zeidler, 1994). The belief in the existence of credit constraints is often
based on results from surveys in which MSE operators are asked to indicate
whether lack of credit is a problem. These surveys usually do not ask whether
respondents did attempt to obtain credit and, if so, whether they obtained the
requested amount of credit. Ideally, determining whether credit constraints exist
should entail comparing the firm’s supply of credit and its effective demand for
credit. However, due to difficulties often encountered in gathering data to determine
the effective demand for credit, the distinction between effective and notional
demand for credit is usually not made in the literature. To determine whether there
is effective demand or not requires assessing the ability of potential borrowers to
repay loans. This is difficult because it requires judgments about the future
behavior of borrowers and the future profitability of an investment project.

Results from surveys which only ask whether lack of credit is a problem for
MSE operators are not very useful for policy because they do not separate people
who are credit constrained from those that are not credit constrained. Government
and donor agencies may expend resources to remove credit constraints on people
who do not face such constraints. This may result in misallocation of society’s

resources which could be used to remove other important barriers to firm growth.

7

Therefore, there is a need for research to determine whether credit
constraints on MSEs in South Africa exist recognizing that such constraints occur
only when the firm’s demand for credit exceeds its supply of credit. Determining
whether credit constraints exist will enable policy makers and others to know the
proportion and type of MSEs whose credit constraints should be removed.
However, to draw conclusions about which MSEs are credit constrained requires
knowledge of the viability of projects people want to use credit for. This is very
difficult to determine from a survey and involves guesses about the future.

Currently, there is a paucity of information on what determines credit
constraints on MSEs in South Africa. Research focusing on determining whether
credit constraints exist among MSEs in South Africa and, if so, what determines the
constraints would be interesting to policy makers in other countries because of the
historical exclusion of blacks from the mainstream economy. Knowledge of what
determines credit constraints is important because it sheds light on possible policy

measures that could be taken to remove the constraints.

1.3 Research questions
This study addresses the following questions:

(a) What is the proportion of MSEs that are credit constrained?

(b) What are the determinants of credit constraints on MSEs and, therefore,
which types of MSEs are likely to be credit constrained?
These questions will be addressed in a specific study area, namely, the

Northern Province of South Africa.

 

1.4 Hypotheses

The hypotheses of this study focus on the determinant variables that reflect
key problems associated with credit constraints in the literature. These problems
are common characteristics of the credit market of the Northern Province. They
include high transaction costs of borrowing and lending, insufficient collateral for
loans, imperfect information, and institutional constraints. These problems are
described in more detail in chapter 3.

The specific hypotheses of the study are:

Hypothesis 1: MSEs from poor households are likely to experience credit
constraints.

Many formal lending institutions in South Africa and elsewhere require
collateral for loans. This means that firms that cannot provide collateral or can only
provide insufficient collateral are likely to be credit constrained, especially in the
formal credit market. Poor firms are the ones that fail to provide sufficient loan
collateral and, therefore, are likely to be credit constrained.

However, there is some ambiguity in this hypothesis. It is possible that MSEs
from poor households may not face credit constraints because they do not demand
credit. A possible reason for this is that MSEs from poor households may lack
investment opportunities which generate the demand for credit. Thus, the richer
may actually face more credit constraints than the poorer.

Hypothesis 2: MSEs located in rural areas are more likely to be credit

constrained than those in peri-urban areas.

9

Peri-urban areas are defined to include townships in which black people
reside in the former homelands and some of the areas that were designated for
occupation by white people located near cities and towns. A possible reason to
expect MSEs in rural areas to be more likely to be credit constrained Include high
transaction costs of borrowing and lending associated with credit transactions
involving lending and borrowing in rural areas. Transaction costs include actual
monetary expenses and the opportunity cost of time involved in executing credit
transactions (see chapter 3 for further discussion).

However, it may also be possible that MSEs in rural areas are less likely to
be credit constrained because of little or no demand for credit. This could happen
if there are limited investment opportunities to generate demand for credit.

Hypothesls 3: MSEs operated by less educated persons are more likely to
be credit constrained.

Less educated MSE operators are not likely to obtain the amount of credit
they ask for because lenders tend to associate repayment ability with the level of
education achieved by the borrower. In rural areas, less educated persons are not
visible to lenders and, therefore, the cost of information to evaluate their
applications for credit is relatively high.

Hypothesis 4: Younger businesses are more likely to be credit constrained
than older ones.

Information required by lenders to evaluate and process loan applications
from younger businesses is not likely to be readily available because these

businesses often do not have an established track record . Thus, transaction costs

10

associated with lending to younger businesses are likely to be relatively high.
Furthermore, younger businesses are not likely to meet the loan collateral
requirements of lenders because they have not accumulated sufficient assets.

Hypothesis 5: MSEs in the manufacturing sector are more likely to be credit
constrained.

MSEs operating in the manufacturing sector require relatively large
investments in assets. These asset requirements coupled with problems
manufacturing MSEs experience in obtaining credit are likely to lead to these MSEs
being credit constrained. Assistance programs for MSEs tend to neglect
manufacturing MSEs in favor of MSEs in other sectors. However, there is also a
view that assistance programs are biased towards manufacturing MSEs.

Hypothesis 6: Female-operated MSEs are more likely to experience credit
constraints.

Rural women in South Africa were considered to be minors in terms of past
laws of the country. This means that they could neither own land nor enter into loan
agreements with lenders without the consent of their husbands. Furthermore,
women tend to be less educated than men. These factors are expected to make
it more difficult for women to obtain the amount of credit they request.

Hypothesis 7: MSEs operated by younger people are more likely to be credit
constrained.

Younger people are expected to have little business experience. In addition,
according to the life-cycle hypothesis, younger people are likely to have

accumulated little wealth and, therefore, they cannot provide their own finance.

11

Hypothesis 8: MSEs registered with government as business concerns are
less likely to be credit constrained.

Evidence from other countries indicates that formal registration is a
requirement for obtaining credit from formal lenders. This implies that unregistered
businesses may experience difficulty in obtaining credit from formal lenders.
Furthermore, formal registration may overcome problems of imperfect information
and this could make formal lenders interested in lending to registered firms.
However, there is some ambiguity in the hypothesis. Unregistered businesses are
likely to be engaged in activities that do not require credit. Thus, they may be

expected to be credit unconstrained because their demand for credit is zero.

1.5 Outline of subsequent chapters

This study is organized into eight chapters.

Chapter 2 provides background information on the Northern Province. This
chapter also describes the small business sector and sources of capital for MSEs
in South Africa.

Chapter 3 focuses on explaining the meaning of a credit constrained firm.
This chapter also provides a theoretical framework for the determinants of credit
constraints.

The data source and characteristics of the sample analyzed in this study are
described in Chapter 4.

Chapter 5 reviews methods used in previous studies to identify credit

constrained firms/households.

12

Chapter 6 specifies the regression model and describes the estimation
procedure. This chapter also outlines the approach adopted in this study to classify
firms as credit constrained or unconstrained.

The results of the study are presented in Chapter 7. This chapter begins with
a description of the characteristics of MSEs according to their credit status (l.e.
whether credit constrained or not). The findings of the study on determinants of
credit constraints on MSEs are then presented. Determinants of credit constraints
are outlined for each credit market (i.e., overall credit market, formal and informal
credit markets).

Chapter 8 presents a summary of the study and also draws implications of

the findings of the study for policy and future research.

CHAPTER TWO

BACKGROUND ON THE NORTHERN PROVINCE, THE SMALL BUSINESS
SECTOR AND SOURCES OF CAPITAL
This chapter provides background information on the Northern Province
Including a description of the small business sector in South Africa and sources of

capital for MSEs.

2.1 Geographical location

Northern Province is one of the eight provinces in South Africa. The province
is located in the northern part of South Africa and shares borders with Botswana,
Mozambique and Zimbabwe. Three former homelands and former white areas
comprise the province? The three former homelands are Gazankulu, Lebowa, and

Venda. Each former homeland area is divided into magisterial districts. A

 

2 Former homelands refer to areas that were set aside for black people to
“govern themselves” in terms of past apartheid policies. Some of these homelands
were later granted political “independence” by the apartheid regime. Venda was
one of these “independent” homelands. White areas were meant for white people
and people belonging to other racial groups were not allowed to live or do business
in these areas.

13

14

magisterial district consists of a number of villages. Some of the magisterial districts
may consist of only rural villages while others include both rural and pen—urban
vMages.

Peri-urban areas are usually located not too far from towns and cities
because blacks were supposed to commute from there to cities and towns to work
and shop. Towns and cities are located in the former white areas. Some of the
major towns and cities in the Northern Province include Pietersburg, Louis Trichardt,

and Tzaneen.

2.2 Socio-economic characteristics

Northern Province is one of the poorest provinces in South Africa. Poverty
and unemployment are more pervasive in the former homelands which are
predominantly rural. About 70 percent of the population in the province live below
the poverty line. Kirsten (1996) notes that 62 percent of households and 69
percent of individuals live in poverty. It is estimated that 49 percent of the
economically active population in the formal economy in 1994 were unemployed.
More than 90 percent of the economically active population who are unemployed
are in the mral areas. The promotion of small businesses is considered to be one

of the ways to alleviate the problems of unemployment and poverty in the province.

2.3 The small business sector In South Africa
The small business sector in South Africa is comprised of survivalist, micro,

small, and medium enterprises (Department of Trade and Industry, 1995). The

15
Department of Trade and Industry (1995) defines these four categories of
enterprises as follows: Survivalist enterprises encompass activities undertaken by

poor people in an attempt to survive. Microenterprises are . very small
businesses, often involving only the owner, some family member(s) and at the most
one or two paid employees." Small enterprises are considered to be part of the
formal economy, employing between five and fifty people per enterprise. Medium
enterprises, often difficult to distinguish from small and large businesses, include
businesses that employ between fifty and two hundred persons and with capital
assets not exceeding five million randsa.

Microenterprises in this study are defined as any enterprise employing less
than six persons“. Microenterprises are, thus, defined to include survivalist and
non-survivalist enterprises. Defining microenterprises in this manner is not meant
to suggest that we ignore the fact that survivalist and non-survivalist enterprises are
different and, therefore, may require different strategies for growth. This definition
is adopted mainly because, in practice, it is difficult to separate survivalist from non-

survivalist enterprises. Small enterprises in this study are as defined by the

Department of Trade and Industry above.

 

3 Rand (R) is the South African monetary unit equivalent to about US$0.20

4 This definition is different from that in USAID guidelines where
microenterprises are defined as any enterprise employing less than ten persons
(see Boomguard et al., 1989; Boomguard, 1991). By South African standards, a
firm employing more than five persons would be considered too large to be
described as a microenterprise.

16

The small business sector in South Africa is dominated by the retail sector,
followed by service and manufacturing sectors (Zeidler, 1994). However, Tendler
(1996) notes that the number of manufacturing MSEs is underestimated. The
majority of retailers are street vendors. The sectoral composition of the small
business sector is as follows: retail (40-70%), manufacturing (14-36%), and service
(9-38%) (Zeidler, 1994). These statistics should be viewed as rough estimates
because there is no agreement among researchers on their validity. The sectoral
breakdown of the small business sector in the Northern Province as provided by
Kirsten (1996) is as follows: trade (48%), manufacturing (13%), service (6%),
construction (4%), transport (19%), repairs (5%) and food processing (4%). These
figures may not provide a representative picture of the sectoral composition of the
small business sector in the Northern Province because they are derived from a
case study.

Based on survey results involving two townships in South Africa, McPherson
(1992) finds that the majority of MSEs are run by females. Female-run MSEs tend
to be smaller than those run by males.

McPherson (1992) observes that the majority of MSE activity in Southern
Africa is in the rural areas. In South Africa, the disproportionate share of rural and
township MSEs is the result of restrictions on blacks to live and own businesses in
former white areas. Restrictions on blacks date back to the 19203 when they were
regarded as only temporary residents in former white areas and were there solely
to provide labor for the white-owned industry (Krige, 1988). These restrictions were

lifted in the early 19905. Legislation became more restrictive in the 19603 when

17

black entrepreneurs living in urban areas were encouraged to move to newly
established “homelands”. Legislation required that black businesses be located in
black residential areas and their market was limited to areas in their vicinity. Black
businesses were not permitted to serve customers from other racial groups (Krige,

1988; Kahn, 1983; Southall, 1980).

2.4 Sources of capital for MSEs

Sources of capital for MSEs may be divided Into credit and noncredit
sources. Credit sources include loans from informal lenders and formal and semi-
formal lenders. Noncredit sources consist of remittances from migrant workers,
old-age pensions, self-employment in farm and nonfarm sectors and wage

employment.

2.4.1 Credit sources

The credit market may be subdivided into formal, semi-fonnal and informal
credit markets. The formal credit market is comprised of all lenders whose activities
do not fall within the regulatory framework of the country’s central bank. The semi-
formal credit market includes all lenders that cannot be easily classified as formal
or informal lenders. This is because some of the lending activities of these lenders
may resemble those of formal lenders while others appear to be informal. The
informal credit market encompasses lenders whose activities are not regulated by

the country’s central bank.

18

2.4.1.1 Formal lenders

The formal credit market is comprised of commercial banks in the private
sector and parastatals (development corporations and development banks) in the
public sector.

Commercial banks are located mainly in urban areas (Commission of Inquiry
into the Provision of Rural Financial Services, 1996) and have traditionally directed
their lending mainly to big businesses. Because commercial banks are located
mainly in the former white areas, MSE operators have to travel long distances to
obtain financial services such as credit. Some of the major banks and other lending
institutions have outlets in the townships. This means that the distance traveled by
MSE operators in townships to banks is shorter than for rural MSE operators. There
are virtually no banks in rural areas although some commercial banks provide
mobile banking services (perhaps once a week and provide mainly withdrawal and
deposit services). Commercial banks are not well adapted to the needs of
disadvantaged people and have been reluctant to lend to the small business sector
(Zeidler, 1994). The reluctance may be attributed to a number of factors, Including,
high transaction costs associated with lending to the small business sector,
problems related to loan security and the perceived high probability of default on
loans.

Commercial banks usually require collateral for loans and many businesses
in the small business sector are not able to meet this requirement. These banks
also charge market rates of interest on loans. Some of the assets owned by the

poor such as land are usually not acceptable to commercial banks as collateral.

19

This is because land in the anal areas is largely communally owned and cannot be
sold.

Development corporations include former homeland development
corporations and the Small Business Development Corporation. These
corporations were established to provide financial and other services for black
people. The Small Business Development Corporation is located in a former white
urban area and has branches in smaller towns, including, pen-urban areas in the
former homelands. Former homeland development corporations are located in peri-
urban areas in the former homelands. Development corporations provide loans to
small black businesses at subsidized rates of interest. These interest rates are
usually fixed and do not vary with the characteristics of the borrower. Loan
collateral requirements for some of the development corporations have not been as
stringent as those for commercial banks. Loans may be granted on the basis of
potential to repay rather than on the availability of collateral.

Development banks include the Development Bank of Southern Africa and
former homeland development banks. The Development Bank of Southern Africa
was established to act as a multi-sectoral institution providing financial and other
services to the former homelands. The bank is located in a former white urban
area. Former homeland development corporations obtained substantial amounts
of money for on-lending from the Development Bank of Southern Africa. Former
homeland development banks are registered as commercial banks and are located

in peri-urban areas in some of the former independent homelands. They were

20

established to provide financial services in predominantly rural areas (Commission

of Inquiry into the Provision of Rural Financial Services, 1996).

2.4.1 .2 Semi-formal lenders

Semi-formal lenders include nongovemment organizations (NGOs) and
savings and credit cooperatives. These organizations became involved in the
financing of MSEs in South Africa recently and little is known about their lending
activities. Zeidler (1994) notes that these organizations are concentrated in black
townships and are less engaged in rural areas. Zeidler (1994) estimates that the

formal credit sector and NGOs reach only one percent of MSEs in South Africa.

2.4.1 .3 lnfonnal lenders

lnforrnal lenders include individuals (e.g. relatives, friends, neighbors,
moneylenders, traders - input or raw material suppliers, etc.) and groups (e.g.
rotating and savings associations or ‘stokvels’, burial societies, etc). The informal
credit market is the main source of credit for rural households in South Africa.
Coetzee (1988) found that 71 and 57 percent of MSE credit transactions in the
former homelands of Lebowa and KaNgwane, respectively, occurred in the informal
financial market.

Loans from friends and relatives are a significant source of finance in many
developing countries. This is also true for South Africa’s rural areas and townships.
Friends and relatives are the biggest source of credit for MSEs in South Africa

(Coetzee et al., 1994). Liedholm and Mead (1987) and Liedholm and McPherson

21

(1991) observe that family and friends are the most important source of funds for
working capital and fixed assets for MSEs.

Rotating and savings associations or ‘stokvels' (as they are called in South
Africa) also play a significant role in informal finance in South Africa. However, they
are more important in the townships. Moore and Schoombee (1994) note that a
third of the black population in metropolitan areas participate in ‘stokvels'.

Traders (including input and raw material suppliers) are also important role
players in rural financing in many developing countries. Fafchamps (1997) notes
that trade credit provided by suppliers is the principal source of external finance that
firms owned by Africans have access to. However, this conclusion does not seem
to be valid for South Africa. Results from several studies indicate that the most
important source of credit for MSEs in South Africa are family and friends. In South
Africa, firms from which MSEs purchase their supplies are located mainly in cities
and towns. These firms are owned mainly by whites. A large number of these

firms sell only on cash basis.

2.4.2 Noncredit sources

Rural households derive income from various sources such as wage
employment, remittances, self-employment in the farm and nonfarm sectors, old-
age pension and agriculture. Since MSEs and households are intertwined, these
income sources also serve as sources of capital for MSEs. The composition of
income for black rural households in South Africa is as follows: wages (34.1%),

remittances (22.1%), transfers (22.4%), informal sector/other (15.3%) and

22

agriculture (6.1%) (see Ardington and Lund, 1996). Remittances from migratory
workers and old-age pensions (transfers) are the second and third most important
sources of household income, respectively. Income from MSE (informal) sector
activities and agriculture rank fourth and fifth, respectively, in terms of their
contribution to household income. However, if importance of household income
sources is measured in terms of the proportion of households that derive income

from a particular source, agriculture ranks as one of the top three sources.

CHAPTER TH REE

CONCEPTUAL FRAMEWORK

This chapter describes the meaning of a credit constrained firm. Previous
studies are reviewed to find out how other researchers define a credit constrained
firm. A theoretical framework is presented to provide an understanding of
determinants of credit constraints. Because whether a firm is credit constrained or
not depends on its demand for and supply of credit, this chapter also provides

theoretical explanations of determinants of demand for and supply of credit.

3.1 Credit in the context of other sources of capital

Firms require fixed and working capital for operation and growth. The capital
may be derived from various sources, including credit and noncredit sources, as
outlined in previous chapters. It is within the context of multiple sources of capital
that credit as another source must be placed. Noncredit and credit sources of
capital may substitute or complement each other. On the one hand, access to
capital from credit sources may eliminate the demand for capital from noncredit
sources. On the other hand, having access to capital from noncredit sources may
make it unnecessary for the firm to borrow. It is typical for capital from noncredit

sources to serve as collateral for a loan.

23

24

Firms usually use their own resources to finance their operations and
projects and resort to external sources of capital when their resources are
inadequate. Thus, noncredit sources will be sought first and only when capital from
these sources are inadequate for their financial needs will firms consider credit
sources. Also within the credit market, MSEs usually borrow from informal lenders
first, especially family and friends, and approach formal lenders when they cannot
raise sufficient capital from informal lenders.

A firm faces capital constraints when the amount of capital is inadequate for
its financing requirements. The amount of capital may be inadequate because
capital raised from either credit or noncredit sources or both is not sufficient. Credit

constraints are a sufficient condition for a firm to be capital constrained.

3.2 What is a credit constrained firm?

A firm is credit constrained when it obtains less credit than it requested at
the prevailing rate of interest and cannot secure more credit by offering to pay a
higher interest rate (Feder et al., 1990; Jaffe and Russell, 1976). Carter and Olinto
(1994) describe a credit constrained household or firm as one that is willing to
borrow more than what the lender is willing to supply at the given rate of interest.
In the economists language, the interest rate is ‘sticky' because it does not adjust
to allow market clearing in the credit market. The ‘stickiness’ of the interest rate is
an important explanation for the divergence between the amount of credit the firm
demands and the amount it obtains. The interest rate may be sticky because

lending institutions find it unprofitable to raise the rate of interest above a certain

25

level in the presence of imperfect information (see section 3.4.1 ). That is, lenders
do not use the price of credit (interest rate) to ration credit and market clearing does
not occur.

Stiglitz and Weiss (1981) define credit rationing as a situation in which either
(3) among loan applicants who appear to be identical some obtain loans while
others do not - those who do not obtain loans are unable to do so even if they offer
to pay a higher Interest rate or (b) some identifiable group within a population are
unable to obtain a loan at the given supply at any given interest rate even though
they would with a larger supply of credit (see section 3.4.1 for further discussion).
Thus, we may distinguish between two situations under which credit rationing might
arise. First, a firm or household may be credit rationed even though the lender has
adequate funds to lend. Second, the lender may have a fixed amount of money to
lend and, therefore, only a certain number of firms or households obtain credit.

Kochar (1992:17) distinguishes between ”access rationing" and "quantity
rationing.” Access rationing refers to a situation in which the amount of credit
supplied to a household/individuallfirrn is zero at the given rate of interest. This
situation is depicted in Figure 3.1(a) and corresponds to the definition of credit
rationing as provided by Stiglitz and Weiss (1981). In Figure 3.1(a), firm-level
supply of credit (l.e., borrower's access to credit) at the rate of interest (r) is zero
although demand for credit is positive. The firm (borrower) faces a vertical supply
curve for credit which coincides with the y-axis. Quantity rationing occurs when the

borrower gets less credit than requested at the given rate of interest, is. when

26

demand exceeds supply. Quantity rationing is illustrated in Figure 3.1(b). The
focus in this study is on access rationing.

Conning (1995) suggests that to determine whether a borrower is rationed
or not the demand for and the largest supply of credit at the current rate of interest
should be compared. In Figure 3.1(b), the largest quantity of credit supplied to the
firm at the current rate of interest (r) is 08 while the quantity demanded is Go. The
amount of credit received by the firm is less than the desired amount by 0500. The
downward-sloping demand curve reflects the inverse relationship that exists
between the amount of credit demanded and the rate of interest on a loan. At lower
rates of interest, there are a number of possible projects that a borrower can invest
in and earn a good rate of return to his/her investment. Thus, more credit is
demanded. However, as the rate of interest rises, the number of projects with a
good rate of return declines. This causes the borrower to demand less credit. The
upward-sloping supply curve for credit indicates the positive relationship that exists
between the rate of interest on a loan and the amount of credit the lender is willing
to supply to the firm. As the interest rate rises, the lender expects to earn more
from lending and will be willing to supply more credit to the firm.

A distinction is often made between price and nonprice credit rationing. Price
rationing occurs when the lender allocates the amount of credit by varying the rate
of interest. Nonprice credit rationing refers to credit allocation that occurs not as the
outcome of varying the rate of interest. In credit markets with imperfect loan
information, information costs may be significant. In such situations, the lender

may find it optimal to ration credit through nonprice means rather than raise interest

27

rates even though there is excess demand for credit (see section 3.4.1 for
explanation). Such nonprice means may include measures taken by the lender to
make it difficult or impossible for loan applicants to obtain a loan (e.g. requiring a
loan collateral, loan application procedures requiring lots of documentation,
soliciting bribes, etc.). Carter and Olinto (1996) note that quantity rationing occurs
in formal credit markets where imperfect information problems are more relevant.
Imperfect information problems are less relevant in informal credit markets because
informal lenders deal with a small clientele known to them. lnforrnal credit is, thus,

often price rationed.

  

 

 

 

 

 

Interest S a. Access rationing Interest b. Quantity rationing
rate rate S

1' T I

I I

I I

I I

I I

I I

I I

I I

. QS QD

Loan SIZE Loan size

Figure 3.1. Access and quantity rationing

Determining whether a firm is credit constrained requires comparing the finn's
effective demand for credit and its supply of credit at a given rate of interest. The
firm's supply (firm-specific supply) of credit refers to the schedule of quantities of
credit available to the firm at current rates of interest. A firm whose demand for

credit exceeds the supply of credit would be considered to be credit constrained.

28

The gap between the amount demanded and received is a measure of the extent
of the credit constraint. Barham et al. (1996) classify firms into three categories:
fully constrained, partially constrained, and unconstrained (see Figure 3.2). A fully
constrained firm has a zero supply of credit and a positive demand for credit
(access rationing). A firm is partially constrained when it receives some credit and
its demand for credit is positive and greater than its supply of credit (quantity
rationing). A partially constrained firm fails to obtain a loan large enough for it to be

on its demand curve.

 

 

 

 

Credit Constrained Credit Unconstrained
(D>S) (D=S, or D<S, or D=0)
Partially Constrained Fully Constrained
(0<S<D) (D>O an? S=O)
l
Applied for Credit Did not Apply for Credit

Note: D = Demand for Credit
S = Supply of Credit

 

 

 

Figure 3.2. Credit constrained and unconstrained firms

A credit unconstrained firm is a firm whose demand for credit is less than or
equal to its supply of credit. Carter and Olinto (1994) describe a credit-
unconstrained firm or household as one whose demand for credit is less than or

equal to the amount offered by the lender at the prevailing rate of interest.

29

A firm or household may be credit unconstrained because it has no demand
for credit or its demand for credit equals its supply at the prevailing rate of interest.
These situations are depicted in Figure 3.3. In Figure 3.3(a), the quantity of credit
demanded at the prevailing rate of interest (r) is zero and at this rate the lender is
willing to supply 03. Figure 3.3(b) depicts a situation in which the quantity of credit

demanded and supplied to the firm at the prevailing rate of interest is the same (Q).

Interest! a. Dem and equals zero S Interest b. Demand equals supply S
rate / rate \/
r

‘ 7

 

 

 

63 .

Q

Loan size Loan size

Figure 3.3. Credit unconstrained when demand equals zero and supply equals
demand

Since the firm’s demand for and supply of credit determine whether the firm
is credit constrained, an insight into what determines the firm’s supply of and

demand for credit is crucial to understanding determinants of credit constraints.

3.3 Determinants of firm-specific demand for credit

The firm’s demand for credit refers to the schedule of quantities of credit the
firm is willing to obtain (and able to repay the loan) at different rates of interest,
ceten's pan'bus. The firm’s demand for credit is derived from its demand for capital.

Firm-specific demand for credit may be expressed as a function of household

30

wealth, cost of borrowing, asset requirement, and expected return to investment in
the economic activity to be financed. Algebraically, the relationship may be

expressed as follows:

Dc = f(C,W,A,R) ........................................................................................ (1)
where:

Dc = amount of credit demanded by the firm;

C = cost of borrowing (includes interest on loan and

transaction cost);

W = household wealth;
A = asset requirement of the firm; and
R = expected return to investment in the activity for which

is credit sought.

Determinants of firm-specific demand for credit may be divided into those
that (1) cause a movement along the demand curve and (2) shift the demand curve.
These two categories of determinants are discussed below. In explaining the effect
of each determinant on demand for credit, we shall assume that all other factors
which may affect credit demand are constant.

A change in interest rate results in a movement along the demand curve and,
thus, affects the quantity of credit demanded. Interest rate is the price paid by the

borrower for credit. Therefore, the quantity of credit demanded by the firm and the

31

rate of interest are inversely related. An increase or decrease in interest rate
causes the quantity demanded to decrease or increase, respectively.

Demand curve shifters include transaction costs, household wealth, expected
return to investment, and asset requirement.

Transaction costs of borrowing include the actual monetary expenses
incurred by the borrower and the borrower's opportunity cost of time involved in
acquiring a loan (e.g., traveling to and negotiating with the lender and completing
application forms). Transaction costs of borrowing determine the position of the
demand curve. This means that a borrower facing higher transaction costs will have
his/her demand curve more to the left relative to the demand curve for a borrower
with lower transaction costs. If the borrower's transaction costs are currently high
and something happens lowering the transaction costs, the demand curve for credit
for the borrower will shift to the right indicating a rise in the demand for credit.
Anything that increases the borrower’s transaction costs would shift the demand
curve for credit to the left implying a decrease in demand for credit. This means that
at a given rate of interest, a smaller quantity of credit will be demanded.

The firm’s demand for credit changes with household wealth and resource
endowment. Household wealth could affect demand for credit positively or
negatively. Wealth might cause the firm not to demand more credit because it no
longer needs credit (i.e. economic activities can be financed from household
wealth). On the other hand, knowing that there is sufficient collateral for loans made
possible by a higher level of household wealth might encourage the firm to demand

more credit. Thus, an increase in household wealth may shift the demand curve for

32

credit either left or right. However, in conventional analysis of the relationship
between wealth and demand for credit, it is usually assumed that an increase in
wealth shifts the demand curve for credit to the right.

The firm’s requirement for assets affects its demand for credit. An increase
in the firm’s asset requirement results in an increase in the firm’s demand for credit.
Asset requirement is related to the economic sector in which the firm operates.
Firms in certain sectors will require more money to invest in equipment, machinery,
buildings, labor and raw materials than firms in other economic sectors. That is, the
nature of economic activities the firm engages in determines the level of investment
required and, thus, asset requirement of the firm. Firms engaged in economic
sectors which require a high level of investment are likely to demand more credit.

The expected return to investment in the activity for which credit is to be used
and demand for credit are positively related. Changes in expected return to
investment can result in a movement along the demand curve for credit or a shift in
the demand curve. At a lower rate of interest, a borrower in a particular economic
sector might expect a higher pay-off from his/her project, and, therefore, the quantity
of credit demanded increases (a movement along the demand curve). If the
borrower decides to switch to another sector which promises to yield a better return
to investment, the demand curve for credit shifts to the right. This implies that at
any given rate of interest the borrower demands more credit than before the change
to the sector with a higher return to investment. If the borrower shifts to a sector
with a lower expected return to investment, the demand curve for credit shifts to the

left.

33

3.4 Determinants of firm-specific supply of credit

Firm-specific supply5 of credit is the schedule of quantities of credit the lender
is willing to supply to the firm at the current rates of interest, ceteris paribus. The
firm’s supply of credit is a function of the cost of lending, firm’s ability to repay the
loan as perceived by the lender, and the lender’s expected return on the loan. Firm-

specific supply of credit may, thus, be expressed as follows:

sc = f(C,,B,R) ............................................................................................ (2)

where

(D
0
II

amount of credit supplied to the firm;

cost of lending;

firm’s ability to repay a loan; and

23030
II

= lender’s expected return on a loan.

As in the case of demand for credit, it is useful to distinguish between credit
supply determinants that shift the supply curve from those that cause movement
along the supply curve. In the discussion below, we shall explain the effect of each
determinant of firm-specific supply of credit by assuming that all other factors

affecting the firm's credit supply are constant.

 

5 Firm-specific or individual supply of credit is not the same as market supply
of credit. Market supply of credit refers to the schedule of quantities of credit the
lender is willing to make available to all firms in the credit market at the current rate
of interest.

34

Interest rate is the price the lender receives for the loan. Therefore, interest
rates are positively related to supply of credit. A higher interest rate encourages the
lender to provide more credit. Changes in interest rates result in movements along
the supply curve.

Whether the firm obtains credit depends on the lender's assessment of the
firm’s ability to repay the loan. If the lender perceives the firm’s ability to repay the
loan to be good, the lender will be willing to make credit available to the firm. The
firm’s ability to repay the loan as perceived by the lender is a function of wealth,
debt obligations, expected return to investment in the project to be financed and
characteristics of the operator (e.g. business experience, level of education, etc.)
and business. Changes in these variables cause a shift in the firm’s supply curve
for credit.

An important component of the cost of lending is transaction cost.
Transaction costs of lending include the actual monetary expenses incurred by the
lender and the time cost involved in evaluating, disbursing and collecting loans. The
higher the transaction costs of lending to the firm, the less willing will be the lender
to make credit available to the firm. Transaction costs of lending determine the
position and slope of the firm’s supply curve for credit. Higher transaction costs
mean that the firm’s supply curve for credit will be more to the left. Lower
transaction costs imply that the firm's supply curve for credit will be more to the
right. This means that at a given rate of interest, lower and higher transaction costs
of lending result in larger and smaller amounts of credit, respectively, received by

the firm.

35

Lenders are interested in earning a good return on money they lend and,
therefore, consider the expected return on loans to be an important factor in
considering whether to lend or not. A higher expected return on a loan is an
incentive for lending. The expected return on a loan is affected by the profitability

of the activity to be financed.

3.5 Determinants of firm-level credit constraints
Credit constraints at the firm level are a function of firm-level supply of and
demand for credit, given the interest rate. Algebraically, the relationship between

credit constraints and their determinants may be expressed as follows:
CONSTR = f{Dc= h(C, W, a, R); Sc =g(C,, B, R)} ................................... (3)

where CONSTR is the credit status of the firm (i.e. whether the firm is credit
constrained or not).

The existence of credit constraints is evidence of a mismatch between the
firm's demand for and supply of credit. Credit constraints occur when there is
excess demand for credit. Thus, both demand for and supply of credit are important
in determining whether credit constraints occur.

The role of demand for credit in determining whether a firm is credit
constrained is often ignored. Kochar (1992) notes that participation in borrowing is
often modeled as being determined solely by the lender's decision with no

consideration of the household's demand for credit. Such an approach assumes

36

that demand for credit is always positive (and high) and, therefore, only the lender’s
decision supply credit will determine participation in borrowing. Kochar (1992)
shows that ignoring demand for credit results in the overestimation of the level of
credit rationing. In a model where demand for credit is ignored, Kochar (1992) finds
that the probability of access to credit is only 19 percent compared with 62 percent
when demand is allowed to determine participation in borrowing. Thus, 81 percent
and 38 percent are classified as credit rationed in the formal credit market under the
first and second models, respectively.

Robinson (1995) provides two possible explanations for the mismatch
between supply and demand for credit for MSEs. First, she believes that there are
deeply-held misconceptions about local financial markets that the delivery cost of
financial services at the local level is too high for financial institutions and that
informal financial services satisfy local demand. Second, Robinson (1995) notes
that lenders often do not recognize the demand for financial services by small
households .

For many economists, the main reasons for the gap between demand for and
supply of credit are imperfect information, lack or insufficiency of collateral,

transaction costs (Hoff and Stiglitz, 1990) and institutional constraints.

3.5.1 Imperfect information
The problem of imperfect information arises mainly in formal credit markets.
Formal lenders are usually located far from potential borrowers and often do not

have perfect information regarding the characteristics of potential borrowers and the

37

projects for which credit is to be used. Informal lenders, especially family and
friends, have good information regarding the characteristics of potential borrowers
(Udry, 1990) and their projects. Udry (1995) points out that it would be incorrect
to conclude that imperfect information problems are unimportant for the structure
of the credit market when it is comprised mainly of loan transactions among family
and friends. Rather, these problems tend to be most severe when formal credit
transactions are involved.

An information gap may also arise from the side of the potential borrower.
It is not uncommon for potential borrowers in developing countries not to have
adequate information about formal lenders and their operations (Llanto, 1990). In
such situations, formal lenders forfeit the opportunity to do business with people
who may turn out to be good clients.

There are two sources of informational imperfection from the lender's side.
Firstly, the lender may be unable to identify the probability distribution of returns
associated with the borrower's/applicant's projects. The probability distribution of
returns is known to the borrower (i.e., differences in the riskiness of projects are
known to the borrower). In such situations the lender treats all borrowers as if they
were homogeneous and resorts to contracts offering identical terms to borrowers
(Barham et al.,1996). The second concerns the inability of the lender to monitor the
behavior of the borrower and credit use after the loan has been granted.

Raising the rate of interest beyond a certain level as a mechanism for
rationing credit in the presence of imperfect information leads to a risky pool of

borrowers (adverse selection effects) or creates an incentive for borrowers to divert

38

credit to riskier uses (adverse incentive or moral hazard problems) (Stiglitz, and
Weiss,1981; Stiglitz, 1987; Besley, 1994; Fry, 1995). Because of adverse selection
and moral hazard problems arising from the use of the rate of interest as a rationing
mechanism in situations of imperfect information, lenders will be inclined to employ
nonprice means to ration credit. Thus, in markets with imperfect information, price
is an inadequate tool for rationing credit.

As the lender increases the rate of interest on a loan (beyond the rate which
maximizes the expected return a loan) in the presence of imperfect information,
borrowers with a low probability of default drop out of the credit market because
they are unwilling to pay a higher rate of interest. That is, they are most
discouraged from borrowing because they are most likely to repay the loan. This
leaves borrowers with risky projects and a high probability of default. Such
borrowers are less concerned about the interest rate they have agreed to pay than
those who undertake safe projects because they are most likely to default on their
loans (Stiglitz, 1987). If the lender continues to offer credit at the rate of interest
beyond that which maximizes the expected return, the expected return declines.
This discourages lenders to use the rate of interest to allocate credit in situations of
imperfect information. Lenders, thus, fix the rate of interest and credit rationing
occurs.

Increasing the rate of interest when information is imperfect may create an
incentive for borrowers to undertake risky projects (adverse incentive or moral
hazard). This means that a borrower who has to choose between two projects, one

with a high risk and the other with a low risk of failure, may undertake the risky

39

project because of the higher rate of interest charged on the loan. This may reduce
the lender's expected return on the loan. So, the lender will fix the rate of interest
and use tools other than the rate of interest to allocate credit.

Nonprice credit rationing arising from imperfect information may occur
independent of wealth of the borrower. Figure 3.4 illustrates credit rationing for poor
and rich borrowers under imperfect information conditions.

In Figure 3.4, 03283 and 08485 represent credit supply curves for poor and
rich borrowers, respectively. S3 and S5 represent points beyond which increases
in interest rate result in a decrease in expected return to the lender due to adverse
selection and incentive problems. Therefore, the lender will not charge a rate of
interest above r2 to ration credit.

Points S2 and S4 represent the threshold on the credit supply curves where
the borrower is overleveraged? At 82 and 8,, the value of the collateral provided
by poor and wealthy borrowers is combined with the rate of interest to ensure that
the expected return on the loan is equal to its opportunity cost. This means that the
lender can provide loans at the contract rate of r1 without risk.

Due to imperfect information regarding the characteristics of the borrowers,
the lender is unable to tailor loan contracts to their risk profile. Therefore, poor and
rich borrowers are charged the same rate of interest (r2) and none of them is on

their demand curve (i.e., there is excess demand for credit). Demand curves for

 

6 The amount of financial leverage in a firm’s capital structure may be
expressed as debt/equity ratio. An overleveraged firm is one that has a high
debt/equity ratio implying that a relatively large amount of the firm’s assets is
financed through borrowing (Miller, 1977).

40

poor and rich borrowers are D(KP) and D(KR). respectively. Excess demand for

credit for poor and rich borrowers is represented by LS,L°1 and LSZLDZ, respectively.

3.5.2 Transaction costs

The role of transaction costs as a credit rationing mechanism is well
documented in literature (Ahmed, 1982; Cuevas, 1984; Ladman, 1984; Inter-
American Development Bank, 1983; Cuevas and Graham, 1984). Higher
transaction costs associated with lending to small borrowers are an important
explanation for the bias in lending towards large borrowers. Transaction costs of
lending are higher for transactions involving rural borrowers are involved (Pohlmeier
and Thillairajah, 1989).

Lending to small borrowers involves relatively higher information collection
and administrative costs per currency unit. Such non-interest fees raise the cost of
borrowing per currency unit for small borrowers. Thus, transaction costs are similar
to interest costs of borrowing because both result in a higher price of credit to the
borrower. Guia-Abiad (1993) finds that transaction costs as a percentage of loan
amount received is regressive. Therefore, the small borrower is likely to be
transaction-cost rationed.

In Figure 3.4, a poorer household could face a supply curve for credit such
as SOS1 or even have its supply curve coinciding with the y-axis if transaction costs

are too high. This is because, with transaction costs incorporating a fixed

41

component, the leverage ratio7 is higher for poorer households. Higher transaction
costs of lending, thus, result in the supply curve for credit to be positioned more to
f the left. This implies that, at a given rate of interest, the quantity of credit received
is lower. In the presence of adverse selection and incentives, lenders are less able
to use the interest rate as a rationing device for poorer borrowers and the supply

curve is truncated at rc rather than r2.

3.5.3 Institutional constraints

Credit may be rationed as the result of legal restrictions requiring financial
institutions to lend to firms meeting certain requirements. For example, Branch
(1996) found that formal credit to informal firms in Peru was discouraged by legal
restrictions requiring formal institutions to lend only to formal (registered) firms. In
such a case the legal status of the firm determines credit access or rationing.
Interest rate ceilings and credit allocation quotas may also result in credit rationing.
The response of formal lenders may be to concentrate lending on a few large
borrowers to minimize transaction costs and default risks (Heidhues, 1990).
Gonzalez-Vega (1984) uses the phrase “the iron law of interest-rate restrictions”

to describe this credit rationing behavior.

 

7 The effective leverage ratio for a borrower can be written as L/(C-TC,)
where L = loan size, C = value of loan collateral, and TC, = transaction costs. The
denominator may be interpreted as the net collateral offered by the borrower
(Barham et al., 1996)

42

3.5.4. Insufficient collateral

Formal lenders often require collateral for loans (when there is imperfect
information) which they can seize in case of default on the loan. Insufficient
collateral may produce credit rationing that would not otherwise exist. Numerous
studies indicate that credit in agrarian settings is rationed according to the ability to
offer collateral (von Pischke et al., 1983; Rudra, 1982; Binswanger et al., 1989).
Insufficient collateral is also an important explanation for the existence of credit
rationing in non-agrarian settings. Credit rationing caused by insufficient collateral
is demonstrated in Figure 3.4 where the supply curve for credit for poorer borrowers
is more to the left than the supply curve for credit for richer borrowers. This means

that poorer borrowers are more credit rationed than richer borrowers.

3.5.5 Credit constraints and economic efficiency

The existence of credit constraints implies that the market fails to bring about
an efficient allocation of credit. Quarcoo (1979) describes an efficient credit system
as one in which (a) the cost of capital is uniform - investors pay the same price for
capital and (b) the purchase price of financial services is equal to the marginal cost
of providing the services. An inefficient credit market system which is characterized
by credit rationing results in nonoptimal allocation of society’s capital resources and
under-consumption of financial services (Quarcoo, 1979). Misallocation of society’s
capital resources may also occur when credit is granted to people without any

repayment ability. This can happen when people receiving credit do not have

43

investment opportunities which would generate enough income for them to repay
the loans.

In an economy where there is no market failure, the interest rate (the price
of credit) charged on a loan allocates credit. Such allocation is efficient (Pareto
efficient) when it is not possible to make someone better off without making
someone else worse off (i.e. no Pareto improvement is possible). Pareto efficiency
is achieved when there is no incentive for the person who obtains a loan to on-lend
to another person (Besley, 1994).

Pareto efficient outcomes are possible in markets which are perfectly
competitive and without extemalities. Credit markets in developing countries are,
however, not perfectly competitive. These markets are characterized by imperfect
information and problems related to repayment of loans. Besley (1994) suggests
that the test of whether a credit allocation is efficient or not should be that a Pareto
improvement is not possible but taking into account the problems of imperfect
information and enforcement - constrained Pareto efficiency. Using constrained
Pareto efficiency as the standard, it is possible for a credit allocation which is less
than what the borrower requested (because of imperfect information problems) to

be efficient.

CHAPTER FOUR
DATA SOURCES AND CHARACTERISTICS OF THE SAMPLE

This chapter provides information on the data used in the study and

describes the characteristics of the sample.

4.1 Data sources

The data for this study were obtained from a survey on MSEs conducted in
1994 in the Northern Province. The survey was part of an investigation into the
provision of rural financial services in South Africa. The investigation was carried
out by a commission of inquiry. The commission’s findings and recommendations
are contained in a report titled Elnalfiepmmflhefiommlssmgflmmmtojbe
EmvisiontuLaLEinanciaLSentices. A questionnaire designed by researchers
from Ohio State University and University of Pretoria was used to collect the data
(see Appendix a). In the Northern Province, the questionnaire was administered
by researchers from the University of the North.

The main purpose of the survey was to determine the nature of financial
transactions and behavior of MSEs in selected economic subsectors. The
subsectors include tailoring and alterations, building and construction, metalworking,

block and brick making, retailing and repair services. Data were collected from 270

44

45

MSEs located in 79 villages in the former homeland and former white areas of the
Northern Province. The sample was designed to include a minimum number of
MSEs engaged in each of the selected economic subsectors. The MSEs included
in the sample ranged in size from very small to relatively large (employing up to
fifteen persons). The reason for selecting MSEs differing in size was the
expectation that the nature of financial transactions would be different depending
on firm size and economic activity. The sample is not designed to be representative
of the MSE sector in the Northern Province in a statistical sense.

The sample was taken from the former white magisterial districts of
Pietersburg, Soutpansberg and former homelands of Lebowa and Venda.
Magisterial districts in the former homeland of Lebowa include Bochum, Mokerong,
Sekgosese, Seshego, and Thabamoopo. Thohoyandou is the only magisterial

district included from the former homeland of Venda.

The magisterial districts are subdivided into fourteen regions. Table 4.1
shows the number and proportion of respondents per region.

The MSEs are engaged in diverse business activities. Because one of the
purposes of this study is to determine whether the economic sector in which MSEs
are engaged has any bearing on their credit status, the MSEs are categorized into
four economic sectors, namely, manufacturing, construction, services and trade.
Table 4.2 shows the number and proportion of MSEs according to economic sector

and business activity.

46

Table 4.1: Number and proportion of respondents (MSEs) per region

 

 

 

I Region Respondents Percentage fl
Bochum 16 6
Louis Trichardt 11 4
Mankweng 32 12
Mokerong 15 6
Mothapo 4 1
Phokoane 5 2
Pietersburg 1 1 4
Praktiseer 1 0
Sekgosese 1 3 1
Sekgosese 2 18 7
Seshego 1O 4
Thabamoopo 1 48 18
Thabamoopo 2 24 9
Thohoyandou 50 1 8
Tzaneen 2 1
Zebediela 20 7

Total 270 100

 

 

 

 

 

Table 4.2: MSEs according to economic sector

  

 

  

 

 

 

 

F— L
Sector Percentage
Construction (Building & construction) 44 16
Manufacturing 89 33
Brickmaking 39 (44%)
Carpentry 4( 5%)
Metalworking 39 (44%)
Shoemaking 4( 5%)
Other 3 (3% )
Service 86 32
Electronic repair 10 (12%)
Vehicle repair 33 (37%)
Tailoring 4O (47%)
Other 3( 4%) II
Trade 48 18
Shopkeeping 44 (92%)
Liquor trading 4( 8%)
Other 3 1
I Total 270 100

 

 

 

 

47

4.2 Characteristics of the sample

4.2.1 Households and MSE operators
The characteristics of households and MSE operators included in the sample

are presented in Table 4.3. These characteristics may be summarized as follows:

. The average household size is relatively large.

. The majority (80%) of households are headed by men.

. Income from business activities comprises the largest part of household
income.

. Wage employment is the second most important source of income.

. Old-age pensions and remittances are of only minor importance as sources

of household income.

. The major of MSE operators are in the middle-age group as indicated by the
average age of 44 years.

. The average level of education of MSE operators is relatively high (9 years

of schooling).

48

Table 4.3: Household and MSE operator characteristics

 

 

 

 

Average
Household size (persons) 5.5
. Gender of household head (male=1, female=0) 0.82
Monthly income (R) 5918
Business 4653 (79%)
Remittances 57 ( 1%)
Pension 88( 1%)
Wages 1 120 (19%)
Value of assets (R) 6663
Level of education (years) 8.6
Age of operator (years) 43.8

 

The gender composition of MSE operators included in the sample does not
reflect the fact that most MSEs in Southern Africa are operated by females. It also
does not reflect the gender composition of MSE operators in South Africa. Zeidler
(1994) notes that, in South Africa, 62 percent of microenterprises are operated by
females and 70 percent of small enterprises run by men. The fact that the
proportion of females included in the sample is not representative of the tme gender
composition of MSE operators in South Africa could limit the usefulness of the
results of the study for policy purposes.

Official statistics on rural household income sources in South Africa indicate
that wages are the most important component of rural household income (see
Central Statistical Services, 1996). These statistics also indicate that remittances
and transfers (pensions) are important sources of income for rural households;
income from MSEs and other sources is the fourth most important source of rural

household income. lnforrnation on household income sources from the sample

49

seems to contradict official statistics. This could be attributed to the
unrepresentativeness of the sample and poor measurement of the level of income
from the various sources.

The relatively high level of education of MSE operators included in the study
is supported by findings from other studies. For example, Zeidler (1994) finds that

two-thirds of MSE entrepreneurs in South Africa are functionally literate.

4.2.2 MSEs
Table 4.4 contains characteristics of MSEs included in the sample. These
characteristics may be summarized as follows:
. Three-quarters of all the MSEs in the sample are located in rural areas.
. The majority of MSE operators included in the sample are male. Only 18
percent of MSE operators are female.
. The majority (65 percent) of the MSEs in the sample are in manufacturing

and service sectors.

. Of all the MSEs in the sample, only 33 percent are officially registered as
businesses.
. The average business age of seven years indicates that most of the MSEs

in the sample are young.
. On average the MSEs in the sample employ three persons per firm.

The geographical composition of MSEs included in the sample reflects the
fact that most MSEs in South Africa are located in rural areas. However, the

sectoral composition of the MSEs contradicts findings from other studies and official

50

statistics which indicate that the MSE sector in South Africa is dominated by the

trade (retail) sector.

Table 4.4: Characteristics of MSEs

 

 

Firms in

 

 

Manufacturing 89 (33%)
Construction 44 (16%)
Service 86 (32%)
Trade 48 (18%)
Other 3( 1%)
Firms located in
Rural area 202 (75%)
Peri-urban area 68 (25%)
Formally registered firms 89 (33%)
Male operators 222 (82%)
Female operators 48 (18%)
Average age of business (years) 7.4
Workers per firm 3.6

 

 

The small proportion (33%) of formally registered MSEs is a typical
characteristic of the MSE sector in many developing countries. For example,
McPherson and Liedholm (1996) note that only four and eight percent of rural firms
are formally registered in Swaziland and Niger, respectively.

The average number of workers (3.6 persons) per firm for MSEs included in
the sample indicates that on average the MSEs are small. The average number of

workers is twice that estimated by Zeidler (1994) for rural MSEs in South Africa.

51

4.2.3 Households and MSEs according to location

The characteristics of households and MSEs included in the sample

according to whether they are located in rural or peri-urban areas are presented in

Table 4.5. These characteristics can be described as follows:

In both peri-urban and rural areas, most MSEs in the sample are in
manufacturing and service sectors.

Wages from employment outside the firm is the most important non-
business source of income for both peri-urban and rural households; but it
is far smaller than income earned from the MSE.

Rural households derive more income from remittances than peri-urban
households.

Rural MSEs have a higher average value of assets and savings than peri-
urban MSEs.

The proportion of formally registered MSEs is higher in rural areas than in
peri-urban areas. This is surprising because other studies indicate that the
proportion of registered firms is higher in urban areas than in rural areas

(McPherson and Liedholm, 1996).

52

 

Eoqu 2 2e Engage n .

2:023 m we EmoEfifi u +

2683 a E 2:8»...ch u n

 

fig

 

 

94.0 E. 3 me 3.83 8<
.9930 80:35

83 $4 $838: $02 Z8 88;

+23 SN 3.; I: $5 I meccgsom

Sad 85 31 V8 32.8 :5 Bacon.

mmwd «No $02 384 §8 V85 $868
9: oEooE cameo;

83 a: 8.0 «3 one:

Bod mod .85 «no 8E8

.806 2.. F- and Ned 3:203:52

38 28 Ed Ed 88.528
8: u o ”mg n 5 228.9. 289.com

8:8...5.» 38qu 5qu
:84 3.23 .23. 59.38

 

 

 

 

 

 

 

5:82 2 9.6608 mum—2 use 3.9.830; we 85:28.20 ”m6 03m...

53

Rural MSEs in the sample have a higher average value of assets, annual
savings in the bank, and business income than peri-urban MSEs. These
averages are distorted by major outliers in the data on assets, savings and

income in rural areas.

4.2.5 Households and MSEs according to economic sector

Table 4.6 presents the characteristics of households and MSEs in the sample

according to economic sector. The main points from Table 4.6 are:

MSEs in the manufacturing sector have the highest average monthly
business income.

MSEs in the construction sector have the highest proportion of monthly
income derived from business.

Wages are the second most important source of household income.
Old-age pension and remittances are of only minor importance as sources
of household income.

Operators of MSEs in construction and manufacturing sectors are
predominantly male.

MSEs in the trade sector are the wealthiest as indicated by the value of
assets and savings.

The trade sector has the highest proportion of formally registered firms.
The manufacturing sector has the highest proportion of firms located in peri-

urban areas.

54

  

 

 

 

 

 

 

 

 

 

o3» 88 m 333 MB o2m>
Ema m Rm EV 3528
N. or A283 one 39:96
and end cmSfio E35423" S 5:80..
396669 >=m§8
wad m P .o Vocno ”85662 265.8" 5 539639 .mcto...
ad NM 35903 Loam. peg:
#6 0m $52qu 8.5. ..ono. Vo mum
to Nm 3:883 cum 2028.51
8.0 8.0 EmEoVuo 6.9:" PV .5950
w a $503 cozmoaum
we we A283 om<
.2830 8833
§- V 9.9
32.8 V E: 3°qu 83. Asa? V m8 882,
33 V R §~ V «.2 £6 V 8 38 V 2 moocmefimm V
§~ V a: 3% V E 3.; V 3.. 3: V mu c228 _
380 V 83 3.: V 83 Assn. V 83 303 V «2.4 $233
$08: .88 ago: 83 §8: 32. ego: $8 2me58:5882
34qu .3qu .8an .3qu
one; 3.20m 2:308:55 EOVSESceo

  

 

 

 

5.03 28280 2 95:88 mums. use «Rosanne: ho 8.63508ch 5% 032.

55

4.2.6 Sources of capital for MSEs

4.2.6.1

Sources of capital for establishment of business

The most important source of capital for establishing a business as ranked

by the respondents is own funds or savings (see Table 4.7). The following

observations can be made:

More than 70 percent of the respondents identified own funds/savings as the
most important source of funds for starting or buying a business. This finding
oonfirrns the observation that most MSE operators tend to self-finance most
of their working and fixed capital (Coetzee et al., 1994).

The second most important source of capital is severance payment”. The
proportion of operators who identified severance payment as the most

important source of funding for establishing their businesses is 12 percent.

Parastatals (development corporations) were identified by four percent of the
respondents as the most important source of capital for establishing or
buying a business.

Family and friends and moneylenders were identified as the most important
sources of capital for establishing or buying a business by a small proportion

of the respondents.

 

8 Money paid to a person who loses his/her job when the employer is forced

to reduce the number of employees for economic reasons.

56

Table 4.7: Ranking of the most important source of capital for establishing MSEs

 

 

 

 

 

 

 

[- MSEs Percentage Ram
Own funds/savings 191 71 1
Severance payment 31 12 2
Parastatals 12 4 3
Family and friends 9 3 4
Pension 8 3 5
Remittances 5 2 6
4.2.6.2 Sources of capital for financing assets

The majority of MSE operators use their own funds to purchase assets such
as tools and equipment. Only a few approach lenders for credit to purchase assets.
Table 4.8 provides information on the number and proportion of MSE operators

who participate in the credit market to obtain funds to purchase assets.

Table 4.8: Participation in the credit market to purchase assets

 

 

 

 

 

 

I Formal or
Formal Informal
Informal
Number did not request credit 182 (67%) 246 (91%) 98 (36%)
Number requested credit 88 (33%) 24 ( 9%) 83 (85%)
Number refused credit 81 (92%) 9 (37%) 15 (15%)
Number received credit 7 ( 8%) 15 (63%L a 172 (64%)

 

Most MSE operators requesting credit to purchase assets approach formal

lenders but only a few obtain formal credit. Thirty-three percent of all MSE

57

operators requested credit from formal lenders to purchase assets but only eight
percent were successful. A smaller proportion of MSE operators (9%) applied for
credit from informal lenders and sixty-three percent were successful. Overall, 36
percent of the MSE operators asked for credit to purchase an asset from formal or

informal lenders, or both, but only 15 percent received credit.

4.2.6.3 Sources of capital for purchasing inputs/raw materials

The majority of MSE operators use their own funds to purchase inputs/raw
materials. Only thirty-percent of MSE operators indicated that they asked to
purchase inputs/raw materials from suppliers on credit. Table 4.9 presents

information on participation of MSE operators in the supplier credit market.

Table 4.9: Participation in the supplier credit market to purchase inputs/raw materials

 

 

 

Number (7.) ll
Did not apply for supplier credit 163 (67%)
Applied for supplier credit 81 (33%)
Received supplier credit 62 (77%)
Did not receive supplier credit 19 (23%)

 

 

Note: Information missing on 26 MSEs

Reasons given for not asking for supplier credit are presented in Table 4.10.
The majority of MSE operators indicated that they did not ask for supplier credit

because the supplier did not provide credit facilities (sold on cash basis only).

58

Twenty percent of MSE operators did not ask for supplier credit because they do not

like to be in debt.

Table 4.10: Reasons for not applying for supplier credit

 

 

 

Reason Number :ercent I
Felt request would be turned down 8 5
Do not like debt 33 20
Had sufficient savings 11 7
Credit costs too much 8 5
Insufficient collateral 3 2
Had access to another credit source 1 0
Supplier is cash and carry 99 61
Total 163 100 I

 

 

 

 

CHAPTER FIVE

REVIEW OF METHODS FOR IDENTIFYING CREDIT CONSTRAINED FIRMS

This chapter reviews literature on approaches for determining whether a
firm/household is credit constrained. The difficulty of estimating demand for and
supply of credit, separately, at the firm level is highlighted. Two approaches for
categorizing firms/households as credit constrained or unconstrained are described.
These approaches are categorized into two: (1) those in which information from
respondents is used to completely sort them into credit constrained and
unconstrained categories (full-sorting approaches); and (2) approaches that entail
using information from respondents to partially sort them into credit constrained and

unconstrained categories (partial-sorting approaches).

5.1 Introduction

Identifying a firm as credit constrained essentially requires that we show that
the firm’s effective demand for credit exceeds its supply of credit. This means that
we need to estimate the firm’s effective demand for credit (equation 1 in Chapter 3)
and its supply of credit (equation 2 in Chapter 3) and compare them. A number of
problems confront the analyst who is attempting to estimate the firm’s/ household’s

demand for credit and its supply of credit separately.

59

60

A major problem encountered in estimating credit demand and supply
functions separately (stmcturaI-form approaches) arises from the difficulty of finding
sufficient variables that enter into the firm’s demand model but do not
simultaneously affect the firm’s supply of credit. This is the difficulty of identifying
demand and supply equations because many variables which determine demand
for credit also determine supply of credit. If credit demand and supply equations are
not identified, it means that when we attempt to estimate credit demand or supply
functions we will not know whether we are estimating a demand or supply function.

Another problem relates to the difficulty of determining the firm’s effective
demand for credit. This is because we need to determine not only how much credit
the firm would like to have but also its ability to repay the loan. Determining the
ability of a potential borrower to repay a loan is very difficult, especially in the
presence of imperfect information which characterizes credit markets in many
developing countries. In particular, to determine the loan repayment ability of a
potential borrower involves judgements about the future behavior of the borrower
and the future profitability of the project for which the loan is sought. Different
people may have widely divergent views as to how that future is likely to evolve.
The difficulty of estimating effective demand for credit explains why most analyses
of credit markets only consider notional demand for credit. This study also
considers notional demand rather than effective demand for credit in determining
whether a firm/household is credit constrained.

Thirdly, the dependent variable in the demand equation is difficult to

measure because demand for credit is not directly observable. While we can

61

observe whether a firm or household receives credit or not, we cannot tell whether
the fimtlhousehold is on its demand curve or not. That is, there is usually no
information available to indicate whether the observed amount of credit corresponds
to a point on the demand curve or supply curve (Carter and Olinto, 1996).

Because of the difficulty of estimating demand for and supply of credit at the
firm-level, many analysts use an approach in which credit demand and supply
variables are collapsed into one equation (reduced-form approach) with excess
demand as the dependent variable (equals to 0 or 1 depending on whether demand
is considered to exceed supply of credit). This is the approach adopted in this
study.

Two approaches are widely used to assign the values of 0 or 1 to a particular
flan/household. These approaches can be categorized into two: complete-sorting
and partial sorting approaches (Carter and Olinto, 1996).

Complete-sorting approaches entail using ancillary information obtained from
respondents regarding their credit market experiences to classify them as credit
constrained or unconstrained. In complete-sorting approaches firms/households
are fully sorted into credit constrained and unconstrained regimes. Two
subcategories of complete-sorting approaches may be identified. The first
subcategory involves requesting respondents to indicate whether they applied for
credit or not. Those who applied for credit are considered to have demand for credit
and, therefore, are either credit constrained or unconstrained. Nonapplication is
considered to imply zero demand for credit and absence of credit constraints (e.g.,

Kochar, 1992). The second subcategory includes those approaches in which, in

62

addition to asking the respondent to indicate whether he/she applied for credit and
the amount of credit received, the respondent is also requested to give reasons for
not applying for credit (e.g. Feder et al., 1990; Barham et al., 1996; Zeller, 1994;
Mushinski, 1995). Respondents are then classified as constrained or unconstrained
depending on their reasons for not applying for credit.

Partial-sorting approaches involve the use of ancillary information to partially
sort firms or households into credit constrained and unconstrained regimes (see
Carter and Olinto, 1996; Coming, 1995). In these approaches, respondents are
sorted into different categories depending on observed transactions in the credit
market. For example, one such category might include those respondents who
indicate that they did not receive credit from either formal or informal lenders.
These nonborrowers can be either credit constrained or unconstrained. This means
that they cannot be placed unambiguously into one of the two credit regimes (credit
constrained or unconstrained). They are only partially sorted into the category of
nonborrowers from formal and informal lenders. Categorization of respondents into
credit constrained and unconstrained where only partial sorting is possible requires
application of econometric methods (econometrics of unobserved switching) to
estimate credit demand and supply parameters plus the probability of being credit

constrained (i.e. the probability that demand for credit exceeds supply of credit).

5.2 Full-sorting approaches
Feder et al. (1990) use an approach that explores the latent demand for

credit in China. The approach involves asking borrowers in the formal credit market

63

whether they received all the credit they wanted or if they wanted more credit than
they received at the going interest rate. Nonborrowers are requested to explain why
they did not borrow. Respondents are classified as credit constrained if they
obtained part of the credit they wanted or could not obtain credit at all. Considering
the whole sample, respondents are classified as fully constrained or unconstrained
in the formal credit market.

Feder et al. (1990) find that 37 percent of the households included in their
analysis are credit constrained. An important observation from the study is that not
all rural households are credit constrained in their farming activities. The approach
recognizes that credit constrained firms can be found among both credit applicants
and nonapplicants. The approach has been adopted by several analysts (Barham
et al., 1996; Zeller, 1994)

Barham et al. (1996) build on the method used by Feder and others to
explore the latent demand for credit in Guatemala. In addition to asking
nonborrowers why they did not borrow in the formal credit market, Barham et al.
(1996) ask why their applications were rejected. Unlike Feder et al. (1990) who
classify households as credit constrained or unconstrained, Barham et al. (1996)
classify households as partially or fully credit constrained or unconstrained. The
results of their study indicate that 34 percent of the households are fully credit
constrained and none is partially credit constrained in their dealings with private
banks. Of those that are credit constrained in their dealings with private banks, 55
percent were either partially or fully credit constrained in their dealings with credit

unions. Barham et al. (1996) conclude that poorer households are more likely to be

64

tightly credit constrained in their dealings with private banks. The study also
concludes that credit unions are important in relaxing credit constraints faced by
low-wealth households but leave the poorest households still quantity rationed.

Zeller (1994) adopts an approach similar to that of Feder et al. (1990) to
classify loan applicants in Madagascar rationed by their lenders as supply-
constrained individuals. Unlike the studies reviewed above, Zeller’s study analyzes
credit rationing in both formal and informal credit markets. Respondents are
classified as credit rationed either in the formal or informal credit market or both.
However, formal credit lenders in Zeller’s study refer to formal lending groups that
obtain loans from formal lending institutions for on-Iending to members of the
groups. Respondents whose applications for membership in a formal lending group
were rejected are classified as credit constrained even though they did not apply for
credit directly. Zeller (1994) proceeds to use probit models to estimate the
determinants of loan application (demand) and credit rationing. He finds that the
proportions of constrained households in formal and informal credit markets are 24
and 16 percent, respectively.

One problem often encountered in classifying respondents as credit
constrained or unconstrained is how to classify the group that did not apply for
credit. In the studies reviewed above, the problem was dealt with by asking
nonapplicants to provide reasons for not applying for credit. In a study of rationing
constraints in the formal credit market (allowing for household participation in the
informal credit market) in India, Kochar (1992) deals with the problem differently.

The approach used by Kochar involves determining whether credit was applied for.

65

Application for credit is considered to be evidence of the existence of demand for
credit. Nonapplication is considered to imply a lack of interest in obtaining credit or
absence of a credit constraint. According to this approach, all nonapplicants are
considered to be credit unconstrained. Applicants for credit are considered to be
either constrained or unconstrained depending on whether demand for credit is
greater or less than supply, respectively. The decision on whether demand for
credit exceeds supply of credit is based on responses to questions asked of
applicants for reasons for not obtaining the desired amount of credit.

Kochar (1992) estimates demand and access functions and uses a univariate
probit model to estimate the probability of access to credit. She estimates the
probability of access to formal credit to be 81 percent when demand for credit is
considered to be high (i.e., when all households are considered to demand credit
at the prevailing rate of interest) and, therefore, the lender determines whether one
gets credit or not (i.e., only supply determines the amount of credit received and
demand is irrelevant). In a bivariate probit model modified to incorporate both
demand for and supply of credit as determinants of access to credit, the estimated
proportion of credit constrained households drops to only 19 percent.

A fundamental problem with Kochar's approach is that it fails to distinguish
between nonapplicants who have no demand for credit and those who do not apply
for credit for reasons other than zero demand. Empirical analysis has often
assumed that only those who apply for credit have demand for it. Acceptance of
such an assumption is likely to result in erroneous conclusions about demand for

credit. Nonapplication for credit does not necessarily imply zero demand for credit.

66

For example, many people do not apply for credit because either transaction costs
are too high or they believe that their applications will be rejected for reasons of
insufficient collateral (Baydas et al., 1994). Mushinski (1995) labels firms or
individuals that feel they cannot obtain credit despite their willingness to borrow at
prevailing interest rates as being "subjectively rationed." Zeller (1994) found that
individuals that never applied for formal credit had their applications for formal group
membership in a credit association rejected or did not apply for group membership
because they perceived no chance of success. The individuals were categorized
as being supply-constrained in the formal (credit) market although they did not apply
for formal credit.

Mushinski (1995) expands Kochar's model in which both demand and supply
are considered to estimate the probability of household access to formal credit and
the effect of credit unions thereon in Guatemala. In Mushinski's approach,
households are requested to indicate whether they have outstanding loans from
formal and informal sources. Those who indicate that they have outstanding loans
are classified as loan recipients. Households that do not have outstanding formal
loans are asked to indicate whether they applied for loans and were rejected. If
they did not apply for formal credit, they are requested to provide reasons for not
having done so. The latter group are classified as "subjectively rationed” if their
reasons for not applying were insufficient funds, high transaction costs or fear of risk
(of losing collateral or wealth). Those who indicate that their reason for not applying
for formal credit is lack of interest, uninformed or unconstrained are classified as not

rationed. Households that obtained credit from moneylenders at interest rates

67

‘ exceeding the average interest rate offered by formal credit sources are also
considered to be rationed in the formal credit market.

Mushinski (1995) finds that credit unions can supply credit to a segment of
the credit market not reached by private and public banks. Credit unions are found
to raise credit supply and demand probabilities for their members. Carter et al.
(1994) find that credit unions may possess informational and cost advantages which
enable them to provide credit to households at lower cost. Credit unions may also
access moral suasion for loan repayment and, thus, lower default risks (Barham et
al., 1996). The lower cost of delivering credit and moral suasion enable credit
unions to make loans readily available to households. These advantages of credit
unions may also raise demand for credit in two ways. Firstly, the collective
knowledge available to credit unions may reduce transaction costs of applying for
a loan. Secondly, households may become more interested in applying for loans
if they perceive their chances of obtaining credit from credit unions to be good

(Mushinski, 1995).

5.3 Partial-sorting approaches

Conning (1995) uses a similar approach to Kochar’s to partially sort
respondents into credit constrained and unconstrained categories. The two
approaches are similar because both base their classification on whether
respondents participate in borrowing or not. Conning (1995) develops a model
(bivariate probit model) to predict whether farmers are credit rationed based on

participation decisions in borrowing (i.e. whether or not a borrower is observed in

68

a particular credit market) rather than examining the details on actual amounts
transacted and interest rates. Although Kochar's and Conning’s approaches are
similar, there are some differences between the approaches. One of the main
differences is that Kochar divides households between those borrowing either from
the formal credit market or informal credit market while Conning's model allows
households to borrow from the formal (commercial banks) and informal (including
only traders and excluding other lenders in the informal credit market) credit
markets, or neither source.

The procedure in Conning’s model involves specification of demand and
supply functions and estimation of the probabilities of households being in any of
the three regimes (borrowing from traders only, both trader and formal lender, and
none of the two). The estimated probabilities are then used to estimate the
parameters of the demand and supply functions from a log likelihood function.
Households participating in the credit market are then subdivided into credit
rationed and nonrationed categories depending on the value of the predicted
amount of credit received. One of the main conclusions of the study is that
households that borrow from informal lenders (traders) only rather than the poorest
or nonborrowing households are the most credit rationed.

Carter and Olinto (1996) employ an approach similar to Conning's to analyze
the impact of land titling and other tenure security measures on agricultural
performance in Paraguay. Households are categorized into those that borrow from
informal credit market only, formal credit market only, formal and informal credit

markets and those that do not borrow. Using information on the amount of credit

69

received from each source, credit demand and supply together with the probability
of being credit rationed in a particular credit market or combination of credit markets

are estimated.

CHAPTER 6
REGRESSION SPECIFICATION, CLASSIFICATION OF FIRMS AND
ESTIMATION PROCEDURE

This chapter specifies the regression model used in this study to identify
determinants of credit constraints on firms. The variables included in the model are
described and the hypothesized relationship between being credit constrained and
the explanatory variables is outlined. The approach used to classify firms included
in the analysis as credit constrained or unconstrained is explained. A logit

framework is used to estimate the odds of being credit constrained.

6.1 Regression model and specific hypotheses
The general form of the function depicting the relationship between the finn's
credit status and its supply of and demand for credit in the credit market (without

distinguishing between formal and informal credit markets) is specified as follows:

CONSTR = f[Supply=g(X); Demand=h(Y)] ............................................ (4)

where

CONSTR = qualitative variable taking the values 0 or 1 if the firm is
unconstrained or constrained, respectively;

70

71

determinants of firm-specific supply of credit; and
determinants of firm’s demand for credit.

The general form of the function expressing credit status of the firm as a

function of its demand for and supply of credit in formal and informal credit markets

is the same as for the overall credit market except for the left-hand side variables.

The left-hand side variables in formal and informal credit sectors are CONSTRf

(equals 0 or 1, if unconstrained or constrained in the formal credit market,

respectively) and CONSTRi (equals 0 or 1, if unconstrained or constrained in the

informal credit market, respectively).

The regression model for determinants Of credit constraints in the overall

credit market is specified as follows:

CONSTR = 3,, + 3, AGE + 32 ASSETS1 + 3,, AGE2 + 3, BUSAGE + 35
CONSTRUC1 + 36 MANUF1 + 3, SERVICE + 3,3 LOCATION + 39 EDUC +
310 LABOR + 3,1 SEX + 3,2 WHITE + 3,3 LABOR2 + 3,, FORMREG + 3,5
REG1 + 3,6 REG2 + 3,, REG3 + 3,8 REG4 + 3,9 REG5 + 32,, REG6 + 32,

REG7 + 322 REGB ................................................................................... (5)
where

AGE = age of the business operator (years).

AGE2 = age of the business operator squared.

ASSETS1 = value of assets owned by the firm/household
(Rands).

BUSAGE = age of business (years).

CONSTRUC1 = construction sector (equals 1 or 0 if business is
in construction or not, respectively).

EDUC = educational level of business Operator (years of

schooling).

FORMREG
LABOR

LABOR2
LOCATION

MANUF1

REGI

REGZ
REG3
REG4
REG5

REGG

REG7
REGB
REC-)9

SERVICE
SEX
WHITE

Bo
3, (i =1,2, ,22)

72

formal registration of business (equals 1 or 0 if
formally registered or not, respectively);
number of workers in the business, including
family workers.

number of workers in the business squared.
location of business (equals 1 or 0 if business is
located in peri-urban or mral area, respectively).
manufacturing sector (equals 1 or 0 if business
is in manufacturing sector or not, respectively).
whether firm is in region 1 or not (1=region 1,
0=otherwise). Region 1 includes Pietersburg,
Mothapo, Mankweng, and Seshego.

whether firm is in region 2 or not (1=region 2,
0=otherwise). Region 2 includes Thabamoopo1.
whether firm is in region 3 or not (1=region 3,
0=otheMise). Region 3 includes Thohoyandou.
whether firm is in region 4 or not (1=region 4,
0=othen~ise). Region 4 includes Louis Trichardt.
whether firm is in region 5 or not (1=region 5,
0=otherwise). Region 5 includes Zebediela,
Phokoane, and Praktiseer.

whether firm is in region 6 or not (1=region 6,
0=otherwise. Region 6 includes Tzaneen,
Sekgosese1 and Sekgosese2.

whether firm is in region 7 or not (1=region 7,
0=othenIvise). Region 7 includes Mokerong.
whether firm is in region 8 or not (1=region 8,
0=otherwise). Region 8 includes Bochum.
whether firm is in region 9 (1=region 9,
0=othen~ise). Base variable and includes
Thabamoopoz.

service sector (equals 1 or 0 if business is in the
service sector or not, respectively).

gender of business operator (male = 1, female =
0);

white area (1 or 0 if business is in a former white
area or not, respectively).

constant.

regression coefficients.

73

Regression models for credit constraints in formal and informal credit markets
are the same as above with CONSTRf and CONSTRi as dependent variables,
respectively.

The value of household and firm assets (ASSETS1) is used to capture the
effect Of wealth on the firm’s credit status. The value of assets was computed as
follows: (1) The original purchase price and expected lifespan of each asset were
determined. (2) If the age of the asset exceeds its expected lifespan, a value of
zero was assigned to the asset item. (3) If the age of the asset item is less than its
expected lifespan, depreciation on the asset was calculated to determine its current
value. To determine the total value of assets for a household or firm, the value of
individual items owned by each firm or household were summed up. The assets
included mainly tools and equipment (see question 7 in Appendix a). The value of
assets in the regression model is divided by 10000 for sealing purposes.

Wealthier MSEs as measured by the value of household or firm assets are
expected to experience less difficulty in obtaining credit because Of their ability to
offer collateral. Household or business wealth is identified in many studies as an
important determinant of the credit status of a flrrn or household. Mushinski (1995)
finds that business wealth is statistically significant in deciding who obtains credit
from formal lenders. Conning (1995) concludes that the value of nonland assets
(machinery and equipment) is positively related to the probability of access to credit.
He also finds that the most constrained farmers are not necessarily the poorest but
those operating small to medium-sized farms and borrowing from traders but not

from private banks. This implies an inverted U-shape function expressing the

74

relationship between being credit constrained and wealth. Zeller (1994) concludes
that total assets owned by the household are significant determinants of the formal
lender's decision to lend. Barham et al. (1996) find that nonprice rationing is
common in the formal credit market and likely to affect low-wealth households.

Location of the business is included as a proxy for transaction costs of
borrowing and lending. The effect of location on the credit status of the firm is
captured by two variables, namely, LOCATION and WHITE. MSEs located far from
cities or towns (in rural areas) are likely to experience more difficulty in obtaining
credit than those located near cities or towns (in peri-urban areas). This is because
of high transaction costs associated with lending and borrowing in rural areas.
Transaction costs have been shown to be important in determining the credit status
of households.

Because the variables LOCATION and WHITE do not capture the effect of
region- or area-specific factors on the credit status of the firm, it was decided to
include dummies for regions (REG1 , REG2, REGQ) (see question 1 in Appendix
A). To minimize the loss in degrees of freedom, it was decided to reduce the
number of regions to nine. This was also done to ensure that areas which are
similar with regard to physical infrastructure are grouped together. Respondents
from some of the regions were very few and it was decided to merge those with
region(s) that are near to them. Anderson (1990) found that regional characteristics
were important in determining who gets credit among small farmers in Brazil. It is
possible for certain areas within a rural or peri-urban area to have better

infrastructure than other areas and, therefore, have

75

lower transaction costs. Regions with poor physical infrastructure are expected to
be more likely to be credit constrained.

More educated operators are expected to be less credit constrained. The
level of education (in terms Of the number of years of schooling) achieved by the
business operator is expected to improve the chances of obtaining credit for the
operator. The probability of being credit constrained is expected to be inversely
related to the number Of years of schooling of the MSE operator. Kochar (1992)
observes that the extent Of education Of the household head influences the
probability of borrowing from the formal sector through its positive effect on demand
for credit. The extent of education of the household head was also found to be a
significant determinant of access to credit. The level of education Of the household
was also found to be a statistically significant determinant of loan demand by Guia-
Abiad (1993). Anderson (1990) finds that households with a higher proportion of
adults who are literate are more likely to receive credit. The level of education
achieved by the household head may be perceived by formal lenders to be an
indicator of managerial capacity and, therefore, increases the probability of access
to credit. Coming (1995) concludes that farmers are more likely to borrow and
obtain credit as the number of years of formal education increases.

Older businesses are likely to experience less credit constraints than younger
ones. lnforrnation required by the lender to decide whether to lend to MSEs is likely
to be available from Older businesses. The problem of information asymmetries is,
thus, likely to be less severe in Older businesses. The age of the business could

also serve as an indicator for business experience. Liedholm (1992) notes that

76

access to lenders grows as firms age and evolve. Kilby et al. (1984) conclude that
supplier credit for fixed capital becomes available when a microfirm becomes well
established and develops a good repayment record. A firm that has been in
business for many years is likely to have survived crises in the past and is,
therefore, more successful. Older businesses are, thus, expected tO be less likely
to be credit constrained.

Economic sector is included in the model to capture the effect of expected
return from investing in the activity to be financed and asset requirement of the firm
on credit status. Four variables are included for construction (CONSTRUC1),
manufacturing (MANUF 1), service (SERVICE1) and trade (TRADE1) sectors.
Expected return to investment in the activity to be financed is expected to be an
important factor affecting lending and borrowing decisions. The lender is likely to
be interested in lending money to a firm engaged in a business activity that
promises a high return to investment because the probability of repayment is
perceived to be high. Kochar (1992) notes that the profitability Of lending varies
(among households) with the probability of repayment. It is hypothesized that firms
in certain economic sectors earn a higher return to investment and, therefore, are
likely to be less credit constrained. Firms may also have different risk levels
depending on the economic sector in which they operate. Prospective borrowers
expecting to earn a higher return to investment from business activities are likely to
be more interested to apply for credit. Expected return to investment in business

activities affects the demand for credit (Kochar, 1992).

77

Asset requirement is expected to vary according to the economic sector in
which the firm is engaged. Firms engaged in economic sectors with relatively high
asset requirements to carry out their activities are likely to be credit constrained as
they will demand more credit. Thus, the probability of being credit constrained can
be expected to increase with the level of asset requirements. Asset requirement in
manufacturing and service sectors are expected to be higher because Of the need
to purchase relatively expensive tools and equipment. Therefore, firms in these
sectors are likely to be credit constrained. Zeidler (1994) Observes that MSEs
engaged in the service sector such as taxi operators are more credit constrained
due to high investment requirements. Given that firms in manufacturing and service
sectors are Often avoided by MSE assistance organizations in favor of retail
businesses (Tendler, 1996), their demand for credit is likely to grow more rapidly
than their supply of credit. Techniques employed by MSEs in the construction
sector (mainly building contractors) tend to be labor intensive and asset
requirements are likely to be lower.

Gender of the business Operator is expected to be an important determinant
of whether the business is credit constrained. In South Africa, the poor as a group
and women in particular have limited access to credit (Department of Trade and
Industry, 1995). Women also tend to have limited collateral. Wickrama and Keith
(1994) note that rural women have limited or no access to formal credit. Zeller
(1994) concludes that women in Madagascar are neither discriminated against by
nor encouraged to obtain credit from formal lenders. It is hypothesized that female-

operated MSEs are more likely to experience difficulty in obtaining credit.

78

Age of the MSE operator is expected to be inversely related to the odds Of
being credit constrained. This means that as the age of the operator increases, the
operator becomes less likely to be credit constrained. Asset accumulation occurs
over time and, therefore, Older operators are likely to have accumulated more
assets than younger ones. Older operators are not likely to require credit because
they can provide their own funding and, therefore, are expected to be less likely to
be credit constrained. Also, as people become old, they are likely to be more
concerned about retirement than obtaining credit to invest in business. However,
Zeller (1994) concludes that the loan demands of older people are more frequently
rationed in Madagascar.

Formal registration with government as a business is likely to be inversely
related to the odds of being credit constrained. This means that Officially registered
firms are expected to be less likely to be credit constrained. Evidence from Peru
indicates that firms that were officially registered had better chances of obtaining
credit because formal lenders were required to lend to registered businesses only
(Branch, 1995). It is also expected that official registration will alleviate the problem
Of imperfect information because information required by formal lenders would be
available from official records. This would make formal lenders interested to lend
to Officially registered firms.

Larger businesses are expected to be less credit constrained than smaller
ones. The transaction cost of lending and firm size tend to be inversely related.
Formal lenders would, therefore, find it cost-effective to lend to larger businesses.

Furthermore, imperfect information problems should be less important in larger

79

firms than in smaller ones. However, it is also possible that larger firms may
experience more credit constraints than smaller firms because of the greater
demand for credit associated with larger businesses. The number of workers
employed by the firm (family and hired workers) is included in the model to capture

the effect Of firm size on credit status.

6.2 Method used in this study for classification of MSEs

The approach adopted in this study to classify MSEs as credit constrained
or unconstrained is similar to that in Feder et al. (1990), Barham et al. (1996) and
Zeller (1994). The approach is desirable because it enables us to use available
data to sort credit nonapplicants into those with and those without demand for
credit. As indicated in earlier sections, some of the methods used by researchers
to classify firms/households as constrained or unconstrained assume that those
who do not apply for credit have zero demand for credit. Nonapplication for credit
does not necessarily imply zero demand for credit.

The approach used in this study to classify firms as credit constrained or
unconstrained involves four steps. The first and second steps are concerned with
identifying firms that have demand for credit to purchase assets. In these steps, we
recognize that demand for credit does not exist only among firms that have applied
for credit to purchase assets but can also occur among nonapplicants. Firms with
demand for credit to purchase assets are identified by considering the overall credit
market without distinguishing between formal and informal credit markets. The third

step entails classifying firms that have applied for credit as credit constrained or

80

unconstrained. This is done for (a) the overall credit market (formal and informal
credit markets together), and (b) formal and informal credit markets separately. The
fourth step involves classifying credit nonapplicants as credit constrained or
unconstrained. This is also done for the overall credit market and informal and

formal credit markets separately.

6.2.1 Steps in classifying MSEs as credit constrained or unconstrained

Stem]; Identify firms that have tried to obtain a loan to purchase assets
either from formal or informal lenders or both.

In the survey, respondents were requested to indicate whether they tried to
secure a loan to purchase an asset during the past two years from the following
sources: moneylender, family or friends, parastatal, savings club, commercial bank
or other source (refer to question J in Appendix a). Attempting to Obtain a loan is
considered to be a signal for demand for credit. But we are unable to separate
those with only notional demand from those with effective demand for credit. From
step 1 we can identify firms that have demand for credit based on whether they
attempted to Obtain a loan to purchase assets. We can also determine whether the
firm got all or none of the credit requested. Respondents were given two options
to choose from: to indicate whether the application was successful (received all
credit requested) or not (received no credit). However, among firms that did not
attempt to Obtain credit to purchase assets, we do not know which of these have

demand for credit and which ones have zero demand for credit.

81

The issue which remains to be resolved is how to subdivide nonapplicants
for credit to finance an asset into those with demand and those with zero demand
for credit. To do this we make use of information obtained from the respondents
regarding their participation in the input/raw material supplier (trader) credit market.
We use this information because nonapplicants for input/raw material supplier credit
were asked to give reasons for not applying for credit (refer to question L in
Appendix A).

Step]; Subdivide nonapplicants for credit (from input/raw material supplier
or to finance assets) into those with and without demand for credit.

We determine whether nonapplicants for credit to finance assets did apply
for input/raw material supplier credit. Ninety percent of firms that did not try to
obtain credit to finance an asset also did not ask for input/raw material supplier
credit. This means that we can establish for most nonapplicants for credit to finance
an asset whether they have demand for input/raw material supplier credit based on
the reasons provided for not asking for input/raw material supplier credit. A
respondent was considered to have demand for input/raw material supplier credit
if any Of the following reasons was given for not asking for input/raw material
supplier credit : (a) felt request would be rejected; (b) input/raw material supplier
sells only on cash basis; (c) credit costs too much. The reason provided for not
applying for input/raw material supplier credit is assumed to hold for not applying for
credit to purchase an asset. Thus, we can establish for 90 percent Of nonapplicants

for credit to purchase assets whether they have demand for credit or not based on

82

the reason provided for not applying for input/raw material supplier credit. The
remaining ten percent Of nonapplicants for loans to purchase

assets who have applied for input/raw material supplier credit are considered to
have demand for credit.

Since nonapplicants for input/raw material supplier credit were asked to
provide reasons for not asking for credit, we are able to subdivide them into those
with demand and those without demand for input/raw material supplier credit based
on reasons provided. Nonapplicants for input/raw material supplier credit were
considered to have no demand for supplier credit if they gave as reasons for not
asking for supplier credit any of the following: (a) do not like incurring debt; and (b)
had sufficient savings. .

Step3; Determine for firms with demand for credit whether their demand for
credit was satisfied or not, i.e. whether firms are credit constrained or
unconstrained. From the above steps we know whether demand for credit exists but
cannot tell whether it was satisfied. Therefore, this step is concerned with
determining whether there is excess demand for credit.

First, we identify firms that have tried to obtain credit but were unsuccessful
(unsuccessful firms include those that got none of what they requested). This is
done for firms that tried to borrow to finance an asset and those that tried to obtain
input/raw material supplier credit. Provision was made for firms to indicate whether
their request for a loan to finance an asset was turned down (question J in Appendix
a). Firms were also requested to indicate whether their request for input/raw

material supplier credit was successful (question L147 in Appendix a). Step 3

83

enables us to determine whether demand for credit was satisfied for firms that tried
to borrow. Firms that indicate that their request for a loan (from input/raw material
supplier or for purchasing an asset) was turned down are considered to have their
demand for credit not satisfied (credit constrained).

Second, we identify firms that applied for credit and were successful. These
are firms that were successful in Obtaining input/raw material supplier and credit to
purchase assets. Such firms are considered to have their demand for credit
satisfied (credit unconstrained).

Credit constrained and unconstrained firms are further subdivided into those
that are constrained or unconstrained in the formal or informal credit market. Firms
are classified as credit constrained in the formal credit market if their applications
for credit to purchase assets were turned down by formal lenders. Only credit
applicants are considered for classification in the formal credit market because we
have no way of finding out whether those that did not apply for credit to purchase
assets from formal lenders (banks and development corporations) have demand for
such credit or not. In the informal credit market, firms are classified as credit
constrained if (1) their request for input/raw material supplier credit or credit to
purchase assets from informal lenders was rejected; or (2) they did not apply for
input/raw material supplier credit but were identified as having demand for credit.
Because Of the small number of applicants for credit to purchase assets from
informal lenders, the credit constrained category in the informal credit market is
dominated by those that are classified as constrained in the input/raw material credit

market.

34

StepA: Classify credit nonapplicants as credit constrained or unconstrained.
All credit nonapplicants (for input/raw material supplier credit or for
purchasing assets) who were identified as having demand for credit are classified
as credit unconstrained. Credit nonapplicants without demand for credit are

classified as credit unconstrained.

6.3 Estimation procedure

The purpose of this section is to estimate the specified regression model
(equation 5). Estimation of the regression model will enable us to identify
determinants of credit constraints on MSEs.

A two-stage procedure is followed to identify determinants of credit
constraints. In the first stage, we ask whether the firm is credit constrained by
considering the total firm-level credit supply and demand without distinguishing
between formal and informal credit markets. We then proceed to find out what
determines credit constraints in the credit market. In the second stage, we
subdivide the credit market into formal and informal credit markets and identify
determinants of credit constraints in each of the credit markets.

The formal credit market in South Africa is comprised of commercial banks
and parastatals or development corporations (see Chapter 2 for more information).
In the formal credit market, only credit transactions involving the purchase of assets
are considered in the analysis of determinants of credit constraints. Our concern
is to find an explanation for the existence of excess demand for credit to finance the

purchase Of assets. Our analysis is restricted to credit transactions involving assets

85

because these are the only transactions in our data set from which we are able to
determine whether a firm is credit constrained or not in the formal credit market.

The informal credit market in South Africa is comprised of numerous types
of lenders from which MSEs may obtain credit (see Chapter 2 for description).
These include traders, moneylenders and family and friends. Our analysis of
determinants of credit constraints in the informal credit market focuses on whether
there is excess demand for credit from all three types of lenders combined. Our aim
in the informal credit sector is to provide an explanation for the existence of excess
demand for credit in the form of input/raw materials and cash from moneylenders
and family and friends to purchase assets. However, credit transactions involving
input/raw material suppliers dominate in the informal credit market.

The model to be estimated (equation 5) involves a binary dependent variable.
Two of the most commonly used procedures for estimating the relationship between
a binary dependent variable and explanatory variables are logit and probit. Logit
is used in this study to estimate the relationship between the right—hand side
variables and the credit status of the firm.

Logit and probit models give similar results and the choice between the two
is one of convenience and availability of computer programs. Kennedy (1992) notes
that the logistic function (logit) is easier to calculate than the cumulative normal
distribution (probit). For this reason, logit is often preferred to probit (Gujarati,
1995). The estimates of the parameters of logit and probit models are not directly

comparable. TO compare these Amemiya (1981) suggests that logit estimates be

86

multiplied by 0.625 and this gives a good approximation of the relevant probit
estimate.

The logit model for the overall credit market is estimated first. This is
followed by the estimation of the logit models for formal and informal credit markets.

The logit model for the overall credit market is specified as follows:

log [prob (CONSTR=1)/prob (CONSTR=0)] = 30 + 3, AGE + 3, ASSETS1
+ 3, AGE2 + 3, BUSAGE + 3, CONSTRUC1 + 3, MANUF1 + 37 SERVICE
+ 3, LOCATION + 3, EDUC + 3,0 LABOR + 3,, SEX + 3,, WHITE + 3,3
LABOR2 + 3,, FORMREG + 3,, REG1 + 3,, REG2 + 3,7 REG3 + 3,, REG4
+ 3,9 REG5 + 3,0 REGG + 3,, REG7 + 3,, REG8

................................................................................................................ (6)

where log [prob (CONSTR=1)/prob (CONSTR=0)] = logarithm of the odds of being
credit constrained.

The specification of logit models for formal and informal credit markets is the
same as for the overall credit market with CONSTRf and CONSTRi replacing
CONSTR in the models for formal and informal credit markets, respectively.

Equation 6 can be written in terms of Odds rather than log odds as follows:

,nen a sauce

Prob (CONSTR=1)/Prob (CONSTR=0)= e 3" e B

where eB is the factor by which the odds of being credit constrained change when

the ith explanatory variable increases by one unit.

87

A factor greater or less than one means that the odds Of being credit
constrained are increased or decreased, respectively, when the corresponding
variable increases by one unit. If the factor is zero, this means that the odds remain
unchanged (Norusis, 1993). a positive or negative Sign attached to Bi indicates

whether the odds of being credit constrained increase or decrease, respectively.

CHAPTER SEVEN

RESULTS OF THE STUDY

This chapter presents the results of the econometric analysis and is divided
into two sections. In the first section, the results are presented mainly in a tabular
form. The focus is on describing the characteristics Of credit constrained and
unconstrained MSEs and households with a view to isolating the main differences.
The second section presents the findings of the study on the determinants of credit

constraints on MSEs based on results from regression analysis.

7.1 Proportion of credit constrained MSEs

This section provides information regarding the proportion of credit-
constrained MSEs in the credit market. This is done by (1) considering the overall
credit market and (ii) subdividing the credit market into formal and informal credit

markets.

7.1.1 Overall credit market
The results of this study indicate that 48 percent of MSEs included in the
sample are credit constrained in the overall credit market (see Table 7.1). The

majority (82 percent) of credit constrained MSEs are in the rural areas. Within the

88

89

rural areas, 52 percent of MSEs are credit constrained and the corresponding

proportion for peri-urban areas is 34 percent. This finding supports the observation

that constraints faced by MSEs in South Africa are more severe in rural areas. It

also supports one of the hypotheses of the study that MSEs in rural areas are more

likely to be credit constrained. However, the bias of the sample toward rural areas

(75 percent of the respondents are from rural areas) could also contribute to the

higher proportion of credit constrained MSEs in the rural areas.

Table 7.1: Proportion of MSEs according to credit status

 

 

r== Constrained WW
Number of firms located in*
Rural area 106 (52%) 96 (48%) 202 (100%)
Peri-urban area 23 (34%) 44 (66%) 67 (100%)
Number of firms in*
Formal credit market 81 (30%) 188 (70%) 269 (100%)
Informal credit market 1 12 (42%) 1 57 (58%) 269 (100%)
Overall credit market 129 (48%) 140 (52%) 269 (100%)
Number of firms in”
Manufacturing 53 (60%) 36 (40%) 89 (100%)
Construction 16 (36%) 28 (64%) 44 (100%)
Services 35 (41%) 50 (59%) 85 (100%)
Trade 23 (48%) 25 (52%) 48 (100%)
Number of formally registered firms 46 (52%) 43 (48%) 89 (100%)
&= —=

 

 

 

 

 

 

 

* = missing information on credit status of firm(s)

90

Table 7.1 also provides information on the proportion of credit constrained
MSEs stratified by economic sector. The economic sector with the highest
. proportion of credit constrained MSEs is manufacturing. Forty-one percent of all
credit constrained MSEs are in the manufacturing sector and 60 percent of all MSEs
engaged in manufacturing activities are credit constrained. Corresponding
estimates for the service sector are 27 and 41 percent. Credit constrained MSES
in the construction sector comprise only 12 percent of all credit constrained MSEs
and 36 percent of MSEs in this sector are credit constrained. These findings lend
support to the hypothesis that firms in manufacturing and service sectors are likely
to be credit constrained because Of the higher capital investment required in these

sectors and the tendency for lending institutions to pay less attention to this sector.

7.1.2 Formal credit market

Table 7.1 indicates that 30 percent of all the MSEs are credit constrained in
the formal credit market (includes commercial banks and former homeland
development corporations). The estimated proportion Of credit constrained MSEs
in this study compares with estimates from other studies. For example, Barham et
al. (1996) find that 34 percent of households are fully credit constrained in the
formal credit sector in Guatemala. Zeller (1994) estimates the proportion of credit
constrained households in the formal credit sector in Madagascar to be 24 percent.
The proportion of credit constrained households in the formal credit sector

estimated by Kochar (1992) for India is 19 percent.

91

The majority (52 percent) Of MSEs that are credit constrained in the formal
credit market are constrained in their dealings with former homeland development
corporations. If it is taken into consideration that these corporations were
established to provide financial and other services to black people, one would have
expected the proportion of MSEs that are credit constrained in their dealings with
these corporations to be lower than that for commercial banks. As indicated in
previous chapters, commercial banks in South Africa have traditionally directed their
lending to big businesses and are not well adapted to the needs of MSEs which are
operated mainly blacks. A possible reason for this higher proportion could be that
most people who apply for loans to development corporations do not meet the

requirements for loans and, therefore, their applications are rejected.

7.1.3 Informal credit market

The results presented in Table 7.1 Show that the proportion of MSEs that are
credit constrained in the informal credit market is 42 percent and all Of them are
credit constrained in their dealings with input/raw material suppliers. Only three
percent of all MSEs are credit constrained in their dealings with other informal
lenders (family and friends, moneylenders and savings clubs). Most of the MSEs
that are credit constrained in their dealings with input/raw material suppliers did not
apply for input/raw material supplier credit even though they indicated that they
would have liked to do so. The most important reason for not applying for supplier
credit is that the supplier sells only on cash basis. The small proportion of MSEs

that are credit constrained in their dealings with other informal lenders may be

92

interpreted to mean that MSEs experience few credit constraints in their dealings
with informal lenders other than input/raw material suppliers. But because of the
bias towards supplier credit transactions in the sample, the conclusion may not be
correct.

The proportion of MSEs estimated to be credit constrained in the informal
credit market is higher than the estimated proportion for the formal credit sector.
This result is unexpected because the literature on credit markets suggests that
firms are more tightly constrained in the formal credit market due to nonprice credit
rationing which is not a characteristic of informal credit markets. The result could
arise from the fact that credit transactions involving input/raw material suppliers
dominate in the informal credit sector. As Fafchamps (1997) notes for Zimbabwe,
black entrepreneurs experience difficulties obtaining trade credit from traders who
are predominantly white. MSE enterpreneurs in South Africa are likely to be facing
similar problems. This could mean that MSEs are tightly credit constrained in their
dealings with suppliers. The exclusion Of credit transactions involving other lenders
who may be more willing to lend to MSEs than input/raw material suppliers
contributed to the higher proportion of credit constrained MSEs in the informal credit
market. Although there are not many studies in South Africa from which results
could be obtained and compared with the estimated proportion Of credit constrained
MSEs in the informal credit market, it would seem that the estimated proportion in
this study is higher. For example, Zeller (1994) finds that the proportion Of credit

constrained households in South Africa is 16 percent.

93

7.2 Proportion of credit unconstrained MSEs

Table 7.2 provides information on the proportion Of MSEs classified as credit
unconstrained according to the reason for being unconstrained (i.e., whether
demand is zero or not). The proportions are presented according to credit market,
location of the firm, economic sector and formal registration as a business.

Credit unconstrained MSES are almost equally divided between those with
zero demand and those with demand greater than zero. However, the proportions
of credit unconstrained MSEs with and without demand for credit differ within and
between credit markets, economic sectors and location. There is also a difference
between the proportion of credit unconstrained MSEs with and without demand
between formally registered and unregistered MSEs.

Considering the overall credit market there is little difference between the
proportions Of credit unconstrained MSEs with and without demand for credit. But
within the formal credit market, the proportion of credit unconstrained MSEs whose
demand was satisfied is very small. This could point to difficulties in obtaining
formal credit. There is little difference between the proportions of credit
unconstrained MSEs with and without demand for credit in the informal credit
market.

Subdividing unconstrained MSEs according to whether they are located in
pen-urban or rural areas indicates that most MSEs in rural areas are unconstrained
because their demand for credit is zero. In urban areas, the majority of credit

unconstrained MSEs are not constrained because they received the amount of

94

credit requested. Therefore, it may be concluded that the proportion of MSEs
whose demand for credit is satisfied is higher in peri-urban than in rural areas.

Stratifying credit unconstrained MSEs according to the economic sector they
operate in indicates that the manufacturing sector has the smallest proportion of
MSEs that are unconstrained because their demand for credit was satisfied. This
provides some indication that MSEs in the manufacturing sector may be more credit
constrained than MSEs in other sectors. The construction sector has the highest
proportion of MSEs that are credit unconstrained because their demand for credit
was satisfied. Thus, MSEs in the construction sector face few credit constraints.
Considering all the sectors, it may be concluded that, with the exception of the
manufacturing, MSEs are unconstrained because their demand for credit was
safisfied.

Subdividing credit unconstrained MSEs into those with and without demand
for credit according to whether they are formally registered with government as
businesses shows that the majority of MSEs are unconstrained because their
demand was satisfied. This suggests few credit constraints for formally registered

firms.

95

Table 7.2: Proportion of credit unconstrained MSEs

 

 

 

 

 

 

 

1* I

Unconstralned Unconstralned Total I

with zero with positive I

demand demand 5

I

Number of firms located in 5

Rural area 50 (52%) 46 (48%) 96 (100%) I

Peri-urban area 17 (39%) 27 (61%) 44 (100%) '

Number of firms in :

Formal credit market 181 (96%) 7 ( 4%) 188 (100%) '

Informal credit market 86 (55%) 71 (45%) 157 (100%) '
Overall credit market 67 (48%) 73 (52%) 140 (100%)

Number of firms in l

Manufacturing 19 (53%) 17 (47%) 36 (100%) I

Construction 3 (1 1%) 25 (89%) 28 (100%) ;
Services 13 (26%) 37 (74%) 50 (100%)

Trade 7 (28%) re (72%) 25 (100%) I

I

Number of formally registered firms 17 (40%) 26 (60%) 43 (100%) '

 

 

 

96

7.3 Characteristics of households, MSEs and MSE operators according to
credit status

This section is concerned with describing the characteristics of credit
constrained and unconstrained households, MSEs and MSE operators to find out
if there are any Significant differences. We first describe the characteristics and
then consider whether the MSEs differ in their characterstics according to location,
economic sector and sources of capital. Characteristics of households, MSEs and
MSE operators are presented in Tables 7.3.

The main difference between credit constrained and unconstrained MSEs is
in the value of household/firm assets. The difference between the value of assets
Of credit constrained and unconstrained households/firms is statistically significant
at the five percent significance level with credit unconstrained household/MSEs
having a higher value of assets. This supports the hypothesis that poorer
households are most likely to be credit constrained. Unconstrained households also
have more annual savings than constrained households, but the difference is not

statistically significant.

7.3.1 Peri-urban areas

Characteristics of MSEs according to credit status in peri-urban areas are
presented in Table 7.4. Except for household size, the difference between credit
constrained and unconstrained MSEs and households is not statistically significant.
The average household size for unconstrained MSEs is larger than that for

constrained ones. MSEs in peri-urban areas are, thus, fairly homogeneous.

97

 

.0023 2 .0 3000.00.00 u .

.0023 m .0 000.5090 u +

.0023 h .0 3000.00.00 u 0

 

 

53 mm. 0. mm 8 3 EV 3:30» 29.88:
EV 0.0000 .0 0:_0> 30.05.

0303

020003

0 5.0 5.0. 33V 0000 $9. 000000000”.
- - mm mm 00 EV £00E .3 0E000_ 000063-000 30.0><
- - m: now «am EV £00E .3 0005030 E00 0E000_ 30.03..
20.0 «No- 0.00 0.3 0.00 0.90.30 0_0E0._ .0 .3052
3&0 n F... 0.0 0.0 0.0 0.0.0.30 0_0E .0 .00Ezz
28.0 00.0- N.» ms v.» 3.005 80.30 *0 3<
3.003 .9830 V0 00:00:00 V0 05..
wand mmd Em m0 Qm 3.00.5 000503 .0 30 30.024.
$000.03 00... .3 0.00003 V0 .00E30 30.024
550 and- 0.0 0.0 0.0 8n0_0E0. .Fu0_0EV0000 0.000300 00 .00000
03.0 00.? oh Qm Om $000.03 00..» 2000030:

000005090 0:_0> 00.000000; 00.000000: 00.000030;
0050300000: 005050000 =<

 

 

 

 

 

 

 

 

 

0390 500.0 2 006.0000 020000000 000 «Mm: V0 02.000.00.000 ”0.x. 0.00...

 

98

0.0000 0. .0 0000.090 u ..

0.00.00 0 .0 3000.090 u +

E00000 . .0 3000.090 u ...

 

 

 

 

 

 

000.0 00.0 0.0 00.0 . ..0 000...
000.0 ...0. .00 00.0 00.0 83.00
.000 0.0 000 00.0 00.0 00.500000:
0000 00.0. 0.0 0.0 00.0 00002.28
6.0.000 Ev... ..9000 E Hocuov .9000 0_Eoc00m
690.050.
30000.. .00n0 ”00.2060. $0000." 5
000.0 0 ... 0.0 0.0 00.0 00000000. 05.0“.
0000 N0. T 0.0. 0 0. $000.00. 000 0005000
0000 00.0 0.. ... 0.. .2800. .000 00.:
0000 00.0 0.0 ..0 0.0 $000.00. 8.0. .002 .0 0.00
+0000 00.0- 0.0 0.0 0.0 $000.00. 000 0.000000:
000.0 .00 0000 E00 0000 .0. 0.0000 .0 80>
«000 .00 .000 .000 .000 E. 000300
0....0 00.. 000 000 000. 00002.
000.0 00.. 0 0 . 0800.00.00
.000 00.9 00 00 00 00.0000
000.0 00.0- 0000 .000 00.0 0000000
9.. 00.005 20.000. 000.0><
0.0.0 .00- .0 00.0 .00 .205". 0050.0. 00000
0.0.0 .00. 0 0 0 0.00.0 0008000
000.0 00.0- 00 00 00 .2000 00<
8.0.000 0000000
8:80.00.» 020., =0 .00 u z. .00 u z.
zuaufl a .00 00000>< Bosch—0:00.03 UGEEWCOO

 

 

 

00.0.0 .008 0. 000.0000 000.0 0005-000 0_ 020000000 000 000.2 .0 02.000.00.000 ”0.0 0.000

 

99

7.3.2 Rural areas

Table 7.5 presents characteristics of MSEs according to credit status in rural
areas. There are significant differences between credit constrained and
unconstrained MSEs and households in rural areas. The main differences occur in
the level of education of the MSE operator, value of household assets, and
proportion of MSEs in manufacturing and service sectors.

Operators of credit unconstrained MSEs are more educated than those of
credit constrained ones. This supports the hypothesis that MSEs operated by more
educated persons are less likely to be credit constrained.

The value of assets for unconstrained households/firms is more than twice
that for constrained ones. This means that unconstrained MSEs in the mral areas
are richer as measured by the value of household/firm assets. This supports the
hypothesis that poorer households are more likely to be credit constrained and is
in agreement with findings by other researchers regarding the relationship between
credit constraints and wealth.

There are statistically significant differences between the proportion of credit
constrained and unconstrained MSEs in the service and manufacturing sectors. In
the manufacturing sector, the proportion of credit constrained MSEs exceeds that
of credit unconstrained ones. The proportion of credit unconstrained MSEs in the
service sector exceeds that of credit constrained MSEs. This means that MSEs in
the manufacturing sector are more likely to be credit constrained while those in the
service sector are less likely to be credit constrained. The need to make large
investments in equipment and tools in firms operating in the manufacturing sector

is likely to be an important explanation for the likelihood of being credit constrained

100

in that sector.

7.3.3 Construction sector

Characteristics of credit constrained and unconstrained MSEs and
households in the construction sector are shown in Table 7.6. There is virtually no
significant difference between credit constrained and unconstrained
MSEs/households in the construction sector. The only difference between these
MSEs/households is with regards to the amount of income derived from old-age
pension. Credit unconstrained households derive no income from old-age pension
while income from old-age pension for constrained households is R70 per month.
Because of the low quality of information collected from the survey on income
derived from the variuos sources, it is difficult to say whether this difference is

indeed significant.

101

~5qu 2 am 3onsz u ..

“cocoon m .m EmouEQm n +

38ng F E E8556 u u

 

 

 

38.0 2am mud 2d :3 9:35:52
E to an. F- :d «No «to 8:2528
c203
Eu... ...900m 5 «ocnov ..ouomm 289.com
396359
2.25. “ocuo ”8.0569 bags" 5
mad 85 mud and 26 cocaine ass".
«who mud- K Na To. 389% .82 8::
32. Rd. S m.» on economy 8.8 .on. so mum
83 and- men 3 mm 3:853 one 2238:
was 8.— x. x. m A983 mum $368
83 mm. 7 83 em 5 3R 9: 3.38
+ .85 SN. fit m 52 mm? E 333 a mam>
~85 mod 2: F «2 P we r 882,
:3 a? E 3” v: moocgsmm
ES :1 3 no 2. 83:8
$3 :5 8% 88 38 $238
Ev 0885 25:9: 08.524
89o Ed 85 86 :3 3553 Queens 528
.de 8. F- m m m $53 8:838
~33 Rd 3. 9. 9. Emma om<
.2930 30:65
8:355.» am u z. 63 u z.
=3..N o:_u> u ¢GE¢>< Dofigamccuca top—3.5230

 

 

 

 

 

 

 

35¢ :88 2 @5288 macaw .93.. E mme “.o 853onsz ”ms aim...

102

35qu 2 E EmoEszn ..

25.83 m .5 3:83.29» u +

3583 s .5 3:85:93. n u

 

and mo. F- 25 mm: 2.: 2925 smegma": 8:80.. _
€957.52
2.58:2 «ocno 6923mm: 2.58:2" 5 _
55.: Ed 2.: 3.: 8.: 532539 555".
58.0 8.? N55 85 £8. 3: 8:38 n
83 8.0 2 9 : A2553 8: 85:33 A
.83 Ed 3 N tm 3:853 58. 8:: _
58.: mm; 5.3 «.3 3 E3 8:2 58. 5 55w _
28.: N to 5 av 88 E 3. 33 3585 5 35>
2a.: ta «.3 m 3.3 3:853 5% 22.88:
33.: 3.: N8 58 85 3582,
55.: 2.9 2 a m: 38:§_Emm
58.: of 8 o E 835:
855 :3. 29. Nam 88 85:38
9: 9:83 25:08 59953..
85.: 33 85 8.: a 355$ 5.2553 55:50
83 85. 3V 8 m 3553 8:838
85.0 3.: 9 N: me 3553 52
:25an mmoEmam
* GUCSECG_Q 0=_fl> AON fl 2v AmP fl 2v
_ __8.N a 3823 3583525 0558.230

 

 

 

 

 

 

 

 

£353 :35 2 932085 56% 5302350 55 E wEocomaoc new ma: .0 85:909sz um.» 035.?

103

3:83 2 8 EmoEcuB u ..

:3qu m 8 E8353 u +

3:83 5 2: EmotEuB u u

 

 

 

 

 

 

 

 

m
88: 8:- 8: B: 2.: 85:": ”8335:": 8:83 _
3:92:65: m
:8: :N: :N: 8.: 8.: 2555: 8:": ”8535:: 2555:" s A
5.833%: .mctou
8:.: 8. F- 8 P: «:8 :8 E: 383:: W
8:: 3:. N h 5 35:3 5:: 85:58 _
wand :md ad ad EN 85:53 .58... 3::
:8: 8:. :.: :.: N: 3:853 :22 58. 5 5:: m
8:.: 8. P- «:8 :89 8:: E: :58: 5 85> fl
+:::: 8:.- ::.: :.: : 3:853 :8 2888: _
_
8:: :5- 8: :E 882, n
8:.: 8:. R. 83 2 38:88:: _
:8: 8.7 8 :: 8 85:5: ‘
:8: 8:. :8: 8:: 8:: 85:38 ‘
$3 0585 omm:o>< _
A
8:: 8:. 8: 8: ::: 355". 652553 5:80
2:: 8:. : : : A983 8:88:
3:: 8:- :: :: :: 383 ::<
.9986 $0595
8:83:33 33> 3: u z: 3: u z.
:3.“ . ou2e>< 3:33:53: 0:523:00
fl “[Illllln h

 

:33: :62: 2 9:808: :98: 9:383:58 on. c_ :u_oco::o:_ ocm mum—2 :o 85:20:..sz K.» 28¢

104

7.3.4 Manufacturing sector

Table 7.7 describes characteristics of credit constrained and unconstrained
MSEs and households in the manufacturing sector. The main differences between
these MSEs and households are with regards to household size, value of assets,
and location of business. The household size for credit unconstrained households
is larger than for constrained households. Unconstrained households are also
wealthier than constrained ones. This is indicated by the value of assets which is
larger for unconstrained than for constrained households. The proportion of MSEs
that are rural-based is higher for credit constrained than for unconstrained

households.

7.3.5 Service sector

Credit constrained and unconstrained MSEs and households in the service
sector differ significantly with regards to the number of workers employed per
business as shown in Table 7.8. Credit constrained households have a larger labor

force than unconstrained households.

7.3.6 Trade sector

Characteristics of credit constrained and unconstrained MSEs and
households in the trade sector are contained in Table 7.9. These MSEs and
households differ significantly only with regards to income derived from wages.
Income from wages for credit constrained households is more than twice that for

unconstrained households. This could mean that households that are not able to

105

meet their requirements for capital from borrowing are compelled to rely on wage

employment outside their businesses.

7.3.7 Ranking of sources of capital for MSEs

Sources of capital for establishing a business for credit constrained and
unconstrained MSEs according to importance are shown in Table 7.10. There is
little difference in terms of the importance of sources of capital for establishing a
business between credit constrained and unconstrained MSEs. Both types of MSEs
ranked own funds/savings and severance (retrenchment) pay as the first and
second most important sources of capital for establishing a business. These results
highlight the importance of own savings as a source of capital for MSEs. Results
from other surveys also indicate that own savings is the most important source of
capital for MSEs. The results also indicate that although family and friends are an
important source of capital, they are not very important role players in providing
capital for establishing businesses. However, credit uncenstrained MSEs ranked
family and friends higher than parastatals (development corporation) while
parastatals were ranked higher than family and friends by credit constrained MSEs.
The results in Table 7.10 also indicate that savings clubs and moneylenders are of
only minor importance as sources of capital for establishing a business for both

credit constrained and unconstrained MSEs.

106

 

 

:::.qu E 2: E8529: u .

“383 m 2: “:::Scu‘: u +

3:83 : ...: Emochm u u

 

] J
:::.: ::.? ::.: ::.: ::.: 89:8:fl .888 8:83 _
2:99:62 _
gusto: 6:": 69:86:. 2.958": _
:::.: ::.: ::.: ::.: ::.: 8888:. 8:8“. _
:::.: ::. F- :3: :8: :::F E: 858: ‘
mood 3:. o x. x. ~. 82:93 om: ::o:_::m _
:F F: ::. w ::. a: N: 38:83 88. 8:: _
.:::.: ::. F- N: :.: :.~ 38:83 88. 88. 8 :5: A
:::.: ::. F- 8:: N: E. :8: E: 98:: 8 28> _
:::.: 8.9 :. : :. : :. : 38:83 :8 22.88: _
:::.: ::.: 5:, :t F :::F 883
:::.: ::.: ::F : ::: :8858m
:::.: ::. F ::. :: ::v 8:8:
.8: ::. P : 8: :8: 3:: 8:58:
9b 9:85 3903:.
:::.: ::.:- s: N: ::.: Easfl :::.::ue 8:8:
:::.: ::.:- :.: : : E8: 8:88:
:::.: ::.: :: :: :: .28: 8:.
.2:an ::oc_::m
3:85:53 33> 8n n 2v 3:... r 5
=3.N a ou0ha>< 005059-0095- 350.3950 .

 

 

 

 

 

 

 

 

:22: .69: 2 @5908: 838: :5 5 8.02830: ac: :mws. :o 85:20:.ch 5.x. :38.

107

 

 

.2880 0: :: E::S:u..: u .

232:0 m :: :::0S:m..: u +

::::::Q s :: :::oEzg: n u

 

                       

 

:::.: ::.: :::: :.:: :::: :92": 83.78:" s 8:83
:::.: :F: :F: ::.: t: .8888.
2.958 «ocuo 62:35:: 2.958" S
:::.: :3 ::.: ::.: ::.: 88:33:: 8:8“.
::.: ::.:- 3:: :::F :::.: E. :::.::
oomd 3.: :0 m o 33:90 :u: ::::_::m
:::.: E: :: :: 3 38:83 8:: 8:....
::.:.o mm: 6: :N ::.. 3:883 :80: 50:. :0 ::.m
:::.: ::.? :::: :::: :8: E. 9:33: 8 28>
:::.: ::.: :.: :.: w: 389:3 ::.: 88:81
+:::.: t: :::F :::F 8:: 3:82,
:::.: ::.: :: :: :2 808:8:
:::.: ::.: :3 ::P :2 838:
:::.: ::.:- ::F: :::: :8: 33:53::
E. :E8:. :o:..:><
:::.: ::.:- ::.: ::.: ::.: 3:5": 6:883 8:8:
:::.: ::.: :.: : 2 38:3 8:88:
:::.: ::.: :: :: :: 38:: ::<
.9830 ::::.::m
8:838.» 28> 3: u z. 3.: u z.
..3.~ 0:583:08: 0:588:00

 

 

 

 

 

 

 

 

 

:35: :::.: 2 9.0.000: 88:: :09. :5 :. :0.0::::0... 0:: :mws. n.0 ::.::tfloflmco um.» :30...

108

Table 7.10: Ranking of sources of funds for establishing MSEs

 

 

 

All MSEs Constrained MSEs Unconstrained

Source MSEs

No. % Rank No. % Rank No. % Rank
Own funds/savings 191 71 1 91 70 1 100 71 1
Severance pay 31 12 2 16 12 2 15 1 1 2
Parastatals 9 4 3 7 5 3 5 4 4
Family and friends 8 3 4 2 2 4 7 5 3
Pension 5 3 4 3 2 4 5 4 4
Remittances 5 2 5 2 2 4 3 2 5
Other business 4 2 5 3 2 4 2 1 6
Commercial bank 2 2 5 2 2 4 2 1 6
Other 1 1 6 1 1 5 1 1 7
Savings Club 1 0 7 1 1 5 O O 8
Moneylender 0 7 1 1 5 0 0 8

 

 

 

 

 

 

 

 

 

 

 

 

7.4 Determinants of credit constraints

The purpose of this section is to describe the determinants of credit
constraints on MSEs based on regression results. This is done in two steps. In the
first step, we identify the determinants of credit constraints in the overall credit
market. The second step is concerned with identifying the determinants of credit

constraints in formal and informal credit markets.

7.4.1 Overall credit market

The regression results concerning the determinants of credit constraints in
the overall credit market are presented in Tables 7.11 and 7.12. In Table 7.11, the
regression results are presented without dummy variables for regions. The
regression results with dummy variables for regions included are presented in Table

7.12.

109

The results in Table 7.11 indicate that the main determinants of credit
constraints in the overall credit market are location of the firm (i.e. whether located
in peri-urban, rural, or white area), value of household/firm assets, and economic
sector in which the MSE operates (i.e., whether the firm is in the manufacturing
sector or not). The coefficients of the variables ’location of the firm’, and ‘value of
assets of the household/firm’ are negative and statistically significant at the five
percent significance level. The coefficient of the variable ‘whether the firm is located
in a former white area’ is negative and statistically significant at the ten percent
significance level. The variable ‘whether the firm is in the manufacturing sector or
not’ has a coefficient with a positive sign and is statistically significant at the five
percent significance level.

The negative signs of the coefficients indicate that the there is an inverse
relationship between the odds of being credit constrained and the variables
concerned. For example, the negative sign attached to the coefficient of “value of
assets of the household/firm’ indicates that as the value of assets increases, the
odds of being credit constrained for the firm decrease. That is, as the firm gets
richer the likelihood of being credit constrained declines. The sign of the coefficient
of ‘whether the firm is in the manufacturing sector or not’ is positive and statistically
significant. This means that firms in the manufacturing sector are more likely to be
credit constrained.

Wealthier firms as measured by the value of assets owned by the
household/firm are less likely to be credit constrained. This supports hypothesis 1:

MSEs from poorer households are more likely to experience credit constraints. The

110

result confirms the observation in many studies that wealthier households are able
to supply collateral for credit and, therefore, are likely to receive the amount of credit
requested. The result could also be interpreted to mean that wealthier households
are in a position to provide their own finance and, therefore, do not require credit.

Firms located in peri-urban areas are less likely to be credit constrained.
This result supports hypothesis 2: MSEs in rural areas are more likely to be credit
constrained than those in pen-urban areas. A possible explanation for this are
lower transaction costs due to better physical infrastructure, shorter distance to
towns and cities where formal lending institutions are located. Peri-urban areas are
found near towns and areas designated as townships in former homeland areas
where only black people were permitted to live. Former white areas are located
near towns and cities and only white people were permitted to live there. The
quality of physical infrastructure in pen-urban areas, especially former white areas,
is better than in rural areas. Thus, transaction costs are lower in peri-urban areas.
The better-quality of physical infrastmcture which contributes to lower transaction
costs also makes peri-urban areas more attractive to invest in. The attractiveness
of pen-urban areas to investors could also explain the likelihood of firms located in
these areas to be less likely to be constrained.

The nature of land ownership rights in rural and peri-urban areas could also
explain the difference in credit status of firms. Formal credit institutions do not
recognize land in rural areas as collateral for loans because land is communally
owned. On the other hand, land in peri-urban areas is acceptable to formal credit

institutions as collateral. This means that firms in rural areas are disadvantaged by

111

the nature of property rights in land when it comes to borrowing from formal lending
institutions.

The economic sector in which the firm operates is a statistically significant
determinant of the firm’s credit status. Manufacturing firms are more likely to
experience credit constraints. This partially supports hypothesis 5: MSEs in
manufacturing and service sectors are more likely to be credit constrained. MSEs
in the manufacturing sector require a relatively large investment in equipment and
tools. Therefore, manufacturing MSEs will require more capital relative to firms in
other sectors. Because the demand for credit is derived from the demand for
investment, manufacturing MSEs are more likely to demand more credit relative to
MSEs in other sectors. The sign of the coefficient of ‘whether the firm is in the
service sector or not’ has the expected sign (negative) but the coefficient is not
statistically significant. The negative sign indicates a decreasing probability of being
credit constrained for a firm which is in the service sector.

The bottom part of Table 7.11 provides information that can be used to
assess the performance of the logit model. There are two ways of assessing the
performance of the model. We can look at the value of -2log likelihood or the
percentage of cases in each credit regime correctly or incorrectly predicted by the
model. A model that fits the data perfectly has -2log likelihood value of zero or the
percentage of cases correctly predicted by the model will be 100 percent. The
overall percentage of cases correctly predicted by our model and the value of -2log
likelihood are 62.95 percent and 314.66, respectively. So, our model does not

perfectly fit the data. Because the statistical significance level for the model’s chi-

112

square is 0.0033, we can reject the null hypothesis that the coefficients of all
variables included in the model are zero.

Table 7.12 shows the results of the regression analysis with dummy variables
for regions included. None of the coefficients of the dummy variables for regions
is statistically significant. However, four of the dummies have negative coefficients
while coefficients of the other four dummy variables are positive. This implies that,
holding other variables in the model constant, MSEs located in some of the regions
are more likely to be credit constrained (positive-signed coefficients) while those
located in regions with negative coefficients are more likely to be credit
unconstrained.

With the addition of dummy variables for regions to the model, the
coefficients of the variables ‘age of the business operator’ and ‘age of business
operator squared’ become statistically significant at the 10 percent significance
level. This implies that MSEs operated by older people are more likely to be credit
unconstrained. However, the odds of being credit constrained are high for MSEs
operated by very old persons as indicated by the positive coefficient of ‘age of the
business operator squared’. This supports the finding by Zeller (1994) that older

people were more likely to be credit rationed in Madagascar.

113

.823 2 3 2:85:90 u ..

300000 0 we ”cmoficgm u +

$00.60 w .0 “0000.505 u u

.xib. I. u
mmood u
mod ”
hodvm ”
00.3.” “
..mw H

l i ll ll IllillllJ

-.Bagmm..- 300:8 =90>O _

022005 $09.00 00595000 __

02060.5 >=00t00 0050300000: .
00:85:90 0.020045 6002

90:00-30 .0022 M

E .6 38080 M

“085.95 03 N)

0005020000 00 80:52 M

 

waned.
ooood
ooood
owned-
ooood
ooood
300.0-
ooood
5506
©0006
ooood
mumod
waned.
ooood

..mvwd
+ ..mmod
wound
53.0
...nmmod
awed
oommd
.onood
mmnwd
+0036
mmmwd
Send
5.. ...o
Nam—no
wand

omooN
o 5 _. .o
mmood
Nmm rd
nomad
$.de
33.0
mmm¢d
mwmmd
mvad
.3de
unwed
wooed
Fomod
mmwod

mowed
mmsmdl
good-
me ...o
Nikko-
0mm _..o.
mNNVd
.33.?
named.
N road
mwmmd-
$18.0-
v 25.0
wmm v.0-
wmood

M
29200 M
30000 “_0 0:_m> _
00.0300 29:03 v.0 50:52 M
$00.03 *0 000052 __
EE *0 00:000.. .

020.000 “.0 50:00
“0: .0 022050. 2.0.58 :5“. _
“0: .0 new 923 0. FE“. M
«0: .0 02300 c_ :5... _
“0: .0 9:203:09: 5 .EE M

00: .0 3:023:00 5 Es...
.2930 00 0030033 _
00.0300 092000 000503 .0 09‘ _
029000 000503 *0 00.0. M

000503 ho cad. M

 

 

00:20:00
Etna—

 

000005.53

 

00:0
200005

a.
«00.05000

 

 

0.00to> bouucaixm M

II I |l||M

 

.9.me £020 05 5 350.3000 .690 *0 90058880 #2. 0.00...

114

300.00 00 .0 3000.090 u .

00300 .

00.0.00... 2.0008 .005

0000000 0 .0 000000005 u +

0000000 0 .0 ~000£09m u u

 

 

 

 

 

 

 

_ 00.3 . 0.000000 .0022 M
000.00 . 00.0.00... 0.00008 005000000 00.00 . .0 .0 00000000

50.00.00 . 000.00... 000008 000.000.0800 00.000 . 080.00.: 03 0- M

008.0 . 000000.090 20:00-20 .0005. 000 . 0000020000 00 .00052 M

M - - 0000.0 0000.0 300.0 0.0.0000 M

0000.0 0000.0 0000.0 0000. 0000.. 0 00.000 M
0000.0 0000.0 0000.0 0000.0 0000.0 0 00.000

M 0000.0 0000.0 0000.0 0000.0 0000.0. 0 00.000 M

00.00 0000.0 0000.0 0000.0 000.0. 0 00.000 M

_ 0000.. 0000.0 0000.0 0000.. 0000.0 0 00.000 M

M 0000.0 0000.0 0000.0 0000.0 0000.0. 0 00.000 M

M 0000.0 0000.0 0000.0 00000 00000.- 0 00.000 M
M 0000.0 0000.0 003.0 0000.. 0.000 0 00.000

_ 0000. 0 0000.0. + .0000 0000.0 0000.0. 0.0000 .0 00_0> M

0000.0 0000.0 003.0 0000.0 0000.0- 00.0000 0.0000; .0 .3052 _

.00 0. 0000.0 0000.0 00000 5.000 0.0000; .0 000.02 M

00000 0000.0. 0000.0 0000.0 030. 0- E... .0 00008.. M

0000.0 0000.0 0000.0 .0000 0000.0. 0.0.000 .0 .00000 M

M 00.00.. 0000.0 00000 00000 300.0 .9. .0 00.20.00. 0:08.00 ...... M

0000.0 0000.0 .0000 0000.0 0000.? .9. .0 00.0 00...; ... E... .
0000.0 0000.0 0000.0 0 .000 0000.? .9. .0 00.200 0. E...
0000.. 0000.0 0000.0 0000.0 0000.0 .9. .0 05.200300... ... ......
0:00 0000.0 0000.0 0000.0 030.0. .9. .0 000000008 ... ......
0000.0 0000.0 0000.0 0000.0 0000.0- 5.0.000 .0 00000000

0.00.. 0000.0 0000.0 0000.0 0.00.0 00.0000 80.000 0000.000 .0 00< M

0000.0 0000.0 .3000 0000.0 003.0. 5.0.000 0000.000 .0 00<

M 0000.0 0000.0 0000.0 0000.0 0000.0. 0000.000 .0 00< M

M 00.00.0000 00.00 An. __

00 .0300 000005006 0000000» 00205000 0.00.00> #3000033 M

 

003000., 00.00. 00.00.00. 000000. 0.00.0 05 0. 0.0.0.0008 0.00.0 00 0.000.000.00 NE. 030...

115

Another change in the regression results due to the addition of dummy
variables for regions is that the coefficient of “whether the firm is located in a former
white area or not’ becomes statistically not significant, but the sign of the coefficient
remains the same (negative). This means that although being located in a white
area or not is no longer an important determinant of credit status of the firms, the
negative sign indicates that firms located in former white areas are more likely to be
credit unconstrained compared to those located in other areas.

Since none of the coefficients of the regional dummy variables is significant,
we may conclude that regional characteristics are not important in determining the
credit status of the firms in the overall credit market. What is important in
determining the credit status of the firms is whether the firm is located in a peri-

urban or rural area.

7.4.2 Formal credit market

The determinants of credit constraints in the formal credit market are shown
in Table 7.13. Table 7.13 also provides information about the performance of the
fitted model. The model does not do well in predicting the proportion of constrained
MSEs. However, the overall prediction is 70 percent correct and the model is
statistically significant as shown by the chi-square significance level of 0.0022.

Statistically significant determinants of credit constraints in the formal credit
market are education level of the operator (years of schooling), sex of the operator
and whether the firm is in the manufacturing sector or not. The coefficient of
education level of the operator is negative and statistically significant. Sex of the
MSE operator and whether the firm is in manufacturing or not have positive

coefficients.

116

Education reduces the odds of being constrained in the formal credit market
as indicated by the negative sign and statistical significance of the relevant
coefficient. The more educated the MSE operator is, the less likely he/she will be
credit constrained in the formal credit market. Formal credit institutions seem to
regard years of schooling of the operator as an indicator of repayment ability.

Household wealth does not affect the odds of being credit constrained in the
formal credit as shown by the value of 1 for e”. This is contrary to expectation as
formal lenders usually require collateral for loans. The result could reflect the
tendency for some former homeland development corporations to waive the
requirement for loan collateral and to base their decision to lend on potential of the
borrower to repay a loan. The result could also indicate that formal lenders
experience difficulty in determining the value of household/firm assets, especially
in rural areas due to imperfect information.

MSEs in the manufacturing sector are likely to be credit constrained in the
formal credit market. This is indicated by the positive sign and statistical
significance of the coefficient of ‘whether the firm is in manufacturing sector or not’.
The value of 3.1159 for e‘3 indicates that the odds of being credit constrained for
MSEs in the manufacturing sector are increased. The explanation for this
relationship is the same as that provided in the previous sections: asset
requirements and problems related to supply of credit.

Male operators are more likely to be credit constrained than female operators
in the formal credit market. This is contrary to the expectation that women are more
credit constrained than men. A satisfactory explanation for this unusual relationship

is not available from the model.

117

Variables such as location of the firm, value of household/firm assets, and

number of workers in the business are not statistically significant determinants of

credit status of the firm in the formal credit market but are statistically significant

when the whole credit market is considered.

The regression results for the formal credit market with dummy variables for

regions included are shown in Table 7.14. The inclusion of dummy variables for

regions has the following effects:

The sign of the coefficient of ‘age of business' changes to negative
(expected sign), but the coefficient is not statistically significant. The
negative sign indicates that older businesses are less likely to be
credit constrained.

The coefficient of “age of business operator squared’ becomes
statistically significant at the 10 percent significance level. This
means that older people are less likely to be credit constrained and
supports our hypothesis 7: MSEs operated by younger people are
more likely to be credit constrained.

The coefficient of ‘gender of operator’ is no longer statistically
significant. That is, male operators are no longer more likely to be
credit constrained in the formal credit market as was the case without
dummy variables for regions.

‘Region 7' has a positive and statistically significant coefficient at the
5 percent significance level. This suggests that firms located in region
7, which is one of the regions with poor infrastructure, are more likely

to be credit constrained.

118

7.4.3 lnforrnal credit market

Table 7.15 presents determinants of credit constraints in the informal credit
market and information regarding the performance of the fitted model. The model
is overall statistically significant as shown by the chi-square significance level of
0.0007. The overall prediction of the model is 65 percent correct.

Statistically significant determinants of credit constraints in the informal credit
market are value of household/firm assets, location of the business and whether the
firm is formally registered.

The value of household/firm assets has a negative and statistically significant
coefficient. This means that wealth is an important determinant of the firm's credit
status in the informal credit market. Wealthier firms are less likely to be credit

constrained in the informal credit market.

:3qu 3 E ”28$ch u . :5qu m do .coofizgm n + 2588 h E 38$:de n u

119

 

 

$3.2. u 3869 . acctoo =Eo>o M
__ «81% u 86605 289:8 859380 W
m $8.8 H 2585 269:8 859.285 _
m Nmodd ” cocooEcgm 933-20 .322 ,
W 2.8 M $83.20 .322 d
m mwdmm ” E do 39580 h
m oath” 82:3... 8.. N;
W EN N 2263030 do .3532 ,
H - - 885 $03 $2.? 2928 ,
coco; oddod 82d 58d 85d «.33 do m2m> W
fl $8.0 ooood modmd 35d 35d. 85:3 89:03 do L.onEzz _
W cm E; 88.0 nomad one Pd end I. £9.83 do .2E32 _
W ammo... doodd or 5d amend named at: .o cosmos .r
m mocha «Edd .dmmod mm Ed ommdd .2930 .0 5300 k
‘ omen; doood dosed 35d 85d do: .6 857.69 2.9th ::E A
moomd ddood Sued mmmvd 39d. 8: do men 923 5 .EE _
_ moan; doodd «End mmmvd mmwmd do: 3 022mm 5 .EE _
~ amp Pd odood +ms~od mmrmd men: do: 3 dctsfifidcmE 5 .EE w
_ 2.8... ooood 55d 35d Rood “o: .o 5:03.680 5 .EE W
m mummd $8.0- +oomod N Fvod Budd- .2925 do con—83m _
« wwddd named. momwd 38d «Food- 0933 .2926 30:63 .0 cd<_
mum : 88d do; Pd oBod ddmpd .2803 89:25 co co<
V mm 5... 88d omdmd «cued 55d «8:63 ..o om<
w 5:22.30 .23 E. _
, no Eran. 350555 Enocfim «5.0530 o3atu> bongo-Sm “

Fl; .Iull ll'll .l

 

 

 

 

 

uoxcmE =38 $5.2 05 c_ £59350 .620 do 9558880 5:. min...

30qu o. .0 .cooScofi n . E0800 0 .0 ::gficgm u + E0200 0 .0 3003.56 u a

I I|| I |. II II l|| I II | | ||Il || ||||| IlilllJ

0.. 5. t.” 88.8... 2.8:8 .085 00.?

120

 

 

20:00.20 .0022 .
080.00” 88.85 28:8 859.080 8.80 .... :o 08880 .
$8.5. 00.0.00... 2.00:8 005000305 modem. 305.00.... 03 «L
38.9 00:85:90 90:00-20 .0022 EN. 05.52008 :o 09:22 _
- - :33 «000.0 500.? .8080 H
EBN ooood oncwd mwmod 900d 0 :o.o0m _
. 000.20 ovood +oo~od oomod 030.2 s :o.o0m. _
Nwowé ooood Bond wmmnd oovmd o 8601 _
t 5.? ooood ooood 93d or Pod m :o.o0m
omood ooood ooood meow; moomd. v 5.001 _
Nommd ooood Emvd voood 050d. m 560”.
momod ooood oompd «wood 300d N 5.001
Sm v. F ooood ooood nomoé maid v :o.o0m
ooooe ooood 0 50d 03 2d moood 0.0000 “.0 0:.0>
moood ooood mm ..~.o mood so Ed- 00.0300 0:00.83 .0 :0nEaz
moon; ooood 3.0 to So rd ovmmd 0:00.53 ..o 00:52 .
doom; ooood hood omnod oNSd ct: .0 5:003 _
owmo; ooood oowwd onomd oomod 3.0000 .6 00:00 .
out; ooood Read or ......o moomd .o: :0 00:20.00: 2.05:0. ::.“. A
comm; ooood nosed 003d wound .o: :o 00:0 0.23 :. ::.“. _
oommé ooood 500d mommd nomad .o: :0 00.200 :. ::.“. W
ovmod. moood .Noood Eomd two; .0: :0 05.300358 :. ::.".
moi: ooood momod Rood «mono .o: :o 5:02.050 :. ::.”.
050d 5.. «d. #9 Pod oovod 00.. ed. .2908 .6 5.50:3. .
_ 500d «mood .oonod :ood opood. 00.0300 6.0000 000:.03 00 00¢. .
_ ..oo: ooood mmmfio ooovd 000 rd 3.0000 000:.03 .0 oo<
moood ooood ~..oo.o moood 0 Food. 000:.03 ..0 0o<
_ 3.3.2.50 ..oho Amy _
.r 0 0 tad 00:00- :o.m L035 «:0 0 . 0o 0. . at; 3805.0

 

 

 

 

 

 

Ill l| |.||| |.|ll || .|||L

00.00:? :o.o0: 9.3.0:. ..0x:0E «60:0 .958 05 :. 0.503050 .690 .0 0.:0:.::0.0o ”v: 0.00...

121

Location of the firm is also an important determinant of the firm's credit status
in the informal credit market. The coefficient for location is negative and statistically
significant. Firms located in peri-urban areas (especially those in former white
areas) are less likely to be credit constrained.

Significant determinants of credit constraints in the informal credit market are
almost the same as for the overall credit market. The differences are that (a)
whether the firm is in manufacturing sector or not is not an important determinant
of credit constraints in the informal credit market ; and (b) formal registration of the
firm is a statistically significant determinant of credit status of the firm in the informal
credit market.

Education and sex of the operator are not important determinants of the
firm's credit status in the informal credit market although they are important in the
formal credit market.

Table 7.16 shows the regression results for the informal credit market with
dummy variables for regions included in the model. The inclusion of dummy
variables in the regression model for the informal credit market has a marginal
effect. None of the coefficients of the dummy variables is statistically significant.
The only notable changes are: (a) the coefficient of “age of business’ changes from
positive to negative (expected sign); and (b) the overall prediction of the model

improves.

122

E028 o. .0 E00555 n .

E0200 0 .0 E00355 u +

E023 . .0 ESEEG u a

 

 

 

0.8.8 8.0.8. . 82.8 .0881. .

$9.00 00.0.00... 2.00:8 00:.0..0:oo _

.... ...... 8.0.8... .828 8:08:85 __

..oood 00:35:90 90:00-30 .0022 .

"u 8:... 0.88.8 .08.... .

. 3.0.0 .0 .o 08880 .

h 8.08 88.0.... 8.. ~-_

_ .mm . 050020000 “.0 00:52 .

. _

_ - - ~80... ......~ 80... .8080 U

88.. 8.....- + ...o... 83... ~00... 0.88 ..o 80> .

_ 28... 88... 0.8... 8...... .05...- 8.88 0.9.8; .o .882 m

. 88.. 88... 8. .... 80.... 8-... 0.9.3; .o .882 .

A 80.... R8..- +0.8... 0.8... .000? :8 ..o 8.08.. _

. ..-.. ....8... .80... 8 ...... .8 .... .2088 ..o .880

88.. 8...... .88... 0 .00... .....0... .8 .5 8.0.0.8. 2.05.... E... u

__ ..-.. 0.8...- +0.8... 88... 08.. .- .8 .o 8:0 3...; ... E... .

_ .000... 88... 8...... 88... 88...- .8 .o 8.28 ... E... M

_ :8. 88... 000.... 800... .80... .8 5 88.8.80... ... E... _
. ~ .00... 88... 0 .8... ..30... .08...- .8 .o 882.88 c. E...

. R00... 88... 88... 88... ~08? .9908 ..o 80880 _

u 8...... 88... 008... 88... 88... 8.88 5.0.08 08508 .o 0.... H

..-...o ....8... .08... ~08... 88...- .2088 008.08 .o 0...... .

_ 88.. 88... 080... 008... 88... 008.08 .0 09.. _

. 500.230 :23 Am. __

00 30.0.. 00:35:05 0:00:05 2.0.0530 030...: Boa-0:293 .

 

 

 

 

 

.900... ..008 .0E..o.:. 0:. :. 0.:.0...0:o0 .080 .o 0.558200 .0 E. 0.00...

E0800 0. .0 E33090. u . E0800 0 .0 E000..00..m u + E0200 . .0 E80056 n ...

20:00-20 .0022 _

 

123

 

0.000. n 00.0.00... 2.00:8 ..0.0>0 .00.. H

03.0.. M 000.00... .8008 00500080 ....0..~ H ... .o 000880 _

0000.00 M 000.00... .8008 000000080: ~000~ . 88.0.... 00.. ~- .
0000.0 . 000000.090 20:00-20 .0005. .mw . 0000020000 .0 .00052 _

.. - 0~...0 000~.~ 0..0.. 0.0.080

. ~000.. 88... 00.00 0000... .000... 0 8.00m .

_ 00000 800.0 ~80... .~..00 00...... . 8.00m

. 00000 88... 0.0.... 0.~.... 00000- 0 8.00m

. 0080 88... ..-... 0...... 0.00...- 0 8.000

_ 0%.... 008... .00.... .0~0.. 0000... .. 8.00m

. 0..00 ....8... 0.000 0000... 0.00.? 0 8.00m _

. 00000 .08.? .... ..0 .000... 008. .- ~ 8.000 _

_ .08.. 008.0 0.000 0000.. 00...... . 8.000 n

_ 88.. 00.0- 0...... 0.0.0 00000. 0.0000 .0 0:.0> _
0000... 008... .~0~0 00...... 0. .00. 00.080 0.0.5; .0 .0852 _

m 0 .0~.. 800... .0 .~... .00.0 008... 0.00.0; .0 .8202

_ 00..~0 ~030- .....0..... 0000... ~000. .- E... .o 8008-.

. . .00.. ....8... 00000 0.30 0000... 8.0.0.... .o .0800

. ~.00.. 00.00 .00.. .... .0000 $000 .8 .0 8.00.00. 2.05.... E...

W ~80... 000...? .0000... 0000... .00... .- .9. .o 00.0 0...... ... E...
:00... 080... ....0... 003.... 00.0.9 .8 .0 00.200 ... 0......
~80. 008... 00.0... 0.~00 000v... .8 .0 000300.80... ... ......

_ .000... 008... 0.000 .0000 ..000- .8 .o 8.02.008 ... E... _

. ....0... 080... 0.0.... 00.00 0800. 3.0.0.... .0 800080

_ ..8.. ....8... ~0¢~0 080... ..8... 00.0000 8.0.08 0000.08 .o 00<

_ 0~.00 0800 00000 ~08... 0 .00.? 8.0.08 0000.08 .o 00<

. .0000 080... 00.00 008... 0.8.? 0000.08 .o 00..

02.0.0000 .000 Am. _
.0 0.00 - - 0.0:..00 . :0 , 0.0. :0. .molwg 00 0 .0 0> 30.000 0 |._

 

 

 

 

 

 

 

 

 

00_00_.0> 00.00. 00.0200. .0me 000.0 0000.0. 0:. 0. 0.0.8.0000 ..00.0 .0 0.000.000.00 .0...- 0.00...

CHAPTER EIGHT
SUMMARY, CONCLUSIONS AND POLICY IMPLICATIONS

8.1 Summary of the study

8.1.1 Background

South Africa faces a major challenge of solving the problems of high
unemployment, poverty and skewed income distribution. The rate of unemployment
is high and the labor absorption capacity of the urban formal sector has declined
significantly over last three decades. About fifty percent of South Africa’s population
can be classified as living below the poverty datum line and poverty is pervasive,
especially in the rural areas. South Africa’s income distribution is the most skewed
in the world. A number of strategies are being considered by the South African
government to solve these problems. One of these strategies involves the
promotion of micro and small enterprises (MSEs).

While it is generally accepted that MSEs can play a major role in addressing
the problems outlined above, their role is limited by constraints they face. These
constraints are numerous. One of the constraints that has been singled out as
significant is insufficient capital. MSEs can generate capital from various sources
which may be categorized as noncredit (e.g., wages, remittances, pension, etc.) and
credit (from informal and formal lenders). Capital constraints can be alleviated by

improving access to these sources for MSEs. However, improving access to credit

124

 

125
for MSEs as a way of alleviating capital constraints has received more attention
over the years.

Many analysts and policy makers in South Africa believe that MSEs are
facing credit constraints. Although this belief is widespread, there is limited
knowledge about the existence and significance of such constraints in terms of the
proportion of MSEs that are credit constrained. Furthermore, there is paucity of
information on which types of MSEs are experiencing great difficulty in obtaining the
desired amount of credit. There is also little information on what determines

whether MSEs are credit constrained.

8.1.2 Purpose of the study

This study is concerned with determining the proportion of credit constrained
MSEs and identifying determinants of credit constraints on MSEs in the Northern
Province of South Africa. The study also investigates whether certain MSEs are
more likely to experience credit constraints than others. The hypotheses tested in
the study are that MSEs more likely to be credit constrained are (1) from poor
households; (2) located in rural areas; (3) operated by less educated entrepreneurs;
(4) younger; (5) in the manufacturing sector; (6) female-operated; (7) operated by

younger persons; and (8) officially not registered MSEs.

8.1.3 Data sources and methodology
Data used in this study for testing these hypotheses were obtained from a
sample of 270 MSEs taken in 1994 from 79 peri-urban and rural villages in two

former homelands and three former white areas in the Northern Province. The data

126

were collected as part of an investigation into the provision of rural financial services
by the Commission of Inquiry into the Provision of Rural Financial Services.
Information obtained from respondents was used to categorize MSEs into
credit constrained and unconstrained regimes. The MSEs were first divided into
credit constrained and unconstrained regimes in the overall credit market (i.e.,
formal and informal credit markets combined). The next step involved sorting MSEs
into credit constrained and unconstrained categories in the formal and informal
credit markets, separately. A logit framework was used to identify determinants of

credit constraints in the overall credit market and formal and informal credit markets.

8.1.4 Significance of credit constraints

The study finds that many but not most of the MSEs included in the analysis
can be described as credit constrained. Considering the overall credit market, the
results of the study indicate that 48 percent of MSEs are credit constrained. The
proportions of credit constrained MSEs in the formal and informal credit markets are
30 and 42 percent, respectively. The estimated proportion of credit constrained
MSEs in the formal credit market compares with estimates from studies in other
countries. However, the estimated proportion of credit constrained MSEs in the
informal credit market is higher than estimates in other studies. This may be
attributed to the dominance of credit transactions involving supplier credit in the
informal credit market. This is despite the dominant role of credit from family and
friends in South Africa as demonstrated in several studies. The majority of MSEs
that are credit constrained are located in rural areas. Credit constrained MSEs
based in rural areas comprise 86 percent of all credit constrained MSEs. Although

this result may reflect the bias in the sample toward rural areas, the proportion of

127

MSEs that are credit constrained in rural areas is higher than for MSEs in peri-urban
areas.

The results of the study also indicate that MSEs in the manufacturing sector
are more likely to be credit constrained than MSEs in other sectors. Sixty percent

of all manufacturing MSEs are classified as credit constrained.

8.1.5 Determinants of credit constraints

The results of the study indicate that the most important determinants of
credit constraints on MSEs in the overall credit market are household/firm wealth
(measured by the value of household/firm assets), location of the business (whether
the business is located in a mral or peri-urban area), and economic sector (whether
the firm is in manufacturing or not).

This study finds that MSEs from poor households (as measured by value of
household/firm assets ) are more likely to experience credit constraints. A possible
reason for this is that poor households do not have sufficient collateral for loans
and, therefore, their applications for loans are often turned down. Poor households
are also unlikely to obtain the amount of credit they request because they do not
have viable investment opportunities which lenders find worthwhile to provide credit
for. The results of the study throw new light on the observation by Conning (1995)
that the poorest households are not necessarily the ones that are most credit
constrained. The poorest MSEs (measured by value of assets and income) in this
study are in the construction sector, but they are not the most constrained. Instead,
MSEs in the manufacturing sector (which are more wealthy) are the most credit

constrained. Yet when economic sector is taken into account as a separate

128

determinant in multi-variable analysis, the poorest households are the most likely
to be credit constrained.

MSEs located in rural areas are likely to experience more credit constraints
than those in peri-urban areas. A possible explanation for this is relatively high
transaction costs associated with credit transactions involving rural borrowers. Low
returns to investment in rural areas may also discourage lenders from providing
credit for projects in these areas and, thus, make it difficult for people in rural areas
to obtain the amount of credit they request.

The results of the study also indicate that MSEs in the manufacturing sector
are more likely to experience credit constraints compared to those in other sectors.
A possible reason for this is the relatively high financing requirement for
manufacturing MSEs and their inability to obtain sufficient credit. Furthermore,
because of the importance of economies of scale in manufacturing, there may be
fewer projects that are worth financing in this sector. This could result in many
credit applications for manufacturing projects being turned down by lenders.

Determinants of credit constraints on MSEs found to be statistically
significant in the formal credit market include education of the MSE operator, gender
of the MSE operator, and whether the firm is in the manufacturing sector or not.
MSEs with more educated operators are more likely to obtain the amount of credit
they request from formal lenders. Contrary to expectations, male-operated MSEs
are more likely to experience formal credit constraints. This is probably because
men are more likely to apply — unsuccessfully - for credit, while women are less likely
to seek loans from formal credit institutions. Firms in the manufacturing sector
have higher odds of being credit constrained in the formal credit market than those

in other sectors.

 

129

The most important determinants of credit constraints on MSEs in the
informal credit market are household wealth, whether the firm is formally registered
or not, and whether the firm is located in a rural or peri-urban area. MSEs from
wealthier households are more likely to obtain the amount of credit they request
from informal lenders, particularly input/raw material suppliers. This finding
contradicts the observation in many studies that informal lenders do not place much
emphasis on collateral but use collateral substitutes. MSEs that are formally
registered with government as business concerns are found to be more likely to
experience credit constraints in the informal credit market. This finding also
contradicts findings from other studies indicating a positive relationship between
formal registration and access to credit. Again this may reflect the fact that
registered enterprises are more likely - unsuccessfully - to seek credit from their

input suppliers, the main component of informal credit in our data set.

8.2 Implications for policy
The findings of the study have several policy implications. Although the
study focuses on the Northern Province, policy implications of the findings may also

be applicable to other provinces in South Africa.

8.2.1 Credit in the context of other sources of capital and credit constraints
in the context of other constraints
An important implication of the findings of this study is that, to design
appropriate policies for raising the contribution of MSEs to employment and income
generation by removing credit constraints, it is necessary to first determine whether

MSEs are credit constrained. Despite credit constraints being singled out as the

130

most significant constraint by MSE operators interviewed in many surveysfltot)

every MSE operator who identifies lack of credit as the most important constraint

Ifigconstrama There Is a need to separate those who are credit constrained
from those who just desire to have credit regardless of whether they can use the
credit productively and generate a good return to investment which will enable the
borrowers to repay their loans/.l An approach which attempts to improve access to
credit for every small business in South Africa i539 likely to be successful.
Improving access to credit can alleviate capital constraints on MSEs but this
is by no means the only way to lift capital constraints on MSEs. Credfiitflshouldflbe
viewed' In the context of other sources of capital for MSEs. In some cases, it may

be more effective to focus onl raising the amount of capital generated from noncredit

sources/rather than improving access to credit. For example, survey results in

 

several countries indicate that the mainsgurce ofyfinan‘cing for MSEs isrovvn savings
within the household or the enterprise. The provision of secure places to save

-_.-.«»...A

money may contribute more to the accumulation of assets than Improving access
the need for government to improve access to credit for MSEs because of the
complementary relationship which exists behNeerI/Capital from noncredit sources)
and/“955393192" Furthermore, removal of credit constraints will not necessarily
lead to an increase in the contribution of MSEs to employment and income
generation. The removal of other constraints is also important and, in some
instances, may be more important than removing credit constraints. Thus, in
addition to identifying MSEs that are credit constrained, efforts should be made to
identify other constraintswvhich if not removed might render efforts to alleviate credit

rfi—WW’ '

constraintsu éeless-

 

 

131

8.2.2 Peri-urban versus rural MSEs

One of the findings of the study is that MSEs located in rural areas are more
likely to be credit constrained than those in pen-urban areas. Based on this finding,
it is tempting to recommend that government and nongovemment efforts to remove
credit constraints should focus on rural areas. However, making such a
recommendation without an insight into the reasons for rural-based MSEs to
experience more difficulty in obtaining credit is unwise because it is the reason(s)
that will point to the most appropriate action to be taken.

Possible explanations for MSEs in rural areas to face greater difficulty in
obtaining the amount of credit they requested include (1) high transaction costs of
lending and borrowing; (2) lower returns to investment in rural areas; and (3) lack
of recognition of land in rural areas as collateral for loans by formal lenders.

Transaction costs of lending and borrowing may be high due to poor physical
infrastructure and imperfect information in rural areas. Therefore, reducing
transaction costs would require improving the quality of physical infrastructure and
increasing the amount of information available to both potential borrowers and
lenders who seek to engage in credit transactions. One way to alleviate imperfect
information problems is for formal lenders to lend to groups rather than to
individuals. Lending through groups has proved successful in lowering transaction
costs in other countries. The most cited example of an institution that has been
successful in lending through groups is the Grameen Bank. Another way to lower
transaction costs is by establishing formal lending institution branches in rural areas.
This may not necessarily reduce the lender’s transaction cost but would lower

transaction costs of borrowing significantly.

132

If low returns to investment in rural areas is the reason for MSEs in these
areas to face more credit constraints than MSEs in pen-urban areas, then it is good
resource allocation for lenders to focus their lending efforts in peri-urban areas.
Removing credit constraints caused by low returns to investment in rural areas
would require increased policy efforts to make rural areas more attractive for
investment. Reducing transaction costs and improving physical infrastructure as
suggested above are ways to increase returns to investment in rural areas.

The problem of lack or insufficiency of collateral caused by lack of recognition
of assets (e.g., land) of rural households as collateral by formal lenders can be
alleviated by paying attention to property rights in land in the rural areas. Land is
by far the most widely-held asset among poor people in rural areas. Formal lenders
would recognize land as collateral if they could sell the land to recover their money
in case of default on loans. This means that ways should be sought to make it
possible for land in rural areas to be sold. However, there is a concern that if this
were to happen, many rural people would be displaced from their land. Some have
suggested that the government should rather stand in as a guarantor for loans
made to rural people by formal lenders.

Given that credit constraints on MSEs exist in rural areas, their removal could
result in a significant rise in the contribution of MSEs to employment and income
generation in these areas. Since rural development is known to have spillover
effects on urban development, an increase in the contribution of MSEs to
employment and income generation would also contribute to development in urban

areas. Investing in rural areas is likely to have the greatest impact on the economy.

 

 

133

8.2.3 Differentiated approaches for different credit markets

The study finds that statistically significant determinants of credit constraints
in the formal and informal credit markets are not the same. In the formal credit
market, significant determinants are education and gender of the MSE operator,
and whether the firm is in manufacturing or not. Significant determinants of credit
constraints in the informal credit market are householdffirm wealth, whether the firm
is formally registered or not and whether the firm is located in a rural or peri-urban
area. These findings imply a differentiated policy approach to alleviating credit
constraints in the formal and informal credit markets.

ln the formal credit market, policy efforts to alleviate credit constraints should
focus on manufacturing MSEs and improving the education level of MSE operators.
Although the results of the study would suggest that male-operated MSEs should
receive more attention than female-operated MSEs, it is doubtful whether such an
approach would have the desired effect on poverty and unemployment because
women are usually affected by poverty and unemployment more than men.

A possible reason for MSEs in manufacturing to experience credit constraints
in the formal sector was that these MSEs have a greater need for financing than
other MSEs and their supply of credit grows less rapidly than their demand for
credit. The little emphasis on assisting manufacturing MSEs by formal lending
institutions and MSE assistance agencies is an important factor contributing to the
slow growth in credit supply. Since manufacturing activities in rural areas tend to
be labor intensive, alleviation of credit constraints faced by MSEs in manufacturing
has the potential to contribute significantly to employment and income. Therefore,
more attention should be focused on improving access to credit for MSEs in

manufacturing sectors.

 

134

The importance of education of the business operator as a determinant of
credit constraints in the formal credit market implies that access to formal credit
could be enhanced by improving the level of education achieved by the business
operator. Thus, adult education programmes could indirectly contribute to
alleviation of credit constraints in the formal credit market. Since household wealth
is not a statistically significant determinant of credit constraints in the formal credit
market, it could be concluded that formal lenders regard education of the operator
as a good indicator of repayment ability and, thus, a substitute for household/firm
wealth.

The finding that wealth (measured by value of household/firm assets) is not
an important determinant of credit constraints in the formal credit was also
explained in terms of the difficulty of formal credit institutions to determine the value
of MSE/household assets. Therefore, measures to alleviate credit constraints in the
formal credit market should focus on inter alia improving the ability of formal credit
institutions to determine the market value of MSE/household assets. Since
business/household wealth is found to be an important determinant of credit
constraints in the informal credit market, it could be concluded that informal lenders
employ better methods to determine the value of MSE/household assets.
Collaboration between formal and informal lenders could thus improve the ability of
the former to determine the market value of household/MSE assets.

Another finding of the study is that, male operators are more likely to be
credit constrained than female operators in the formal credit market. The
implication of this is that efforts to alleviate credit constraints in the formal credit

market should focus on male-operated MSEs. However, acceptance of such a

 

 

135

recommendation might be unwise. The lack of representativeness in the sample
in terms of gender composition may have resulted in this conclusion.

The results of the study indicate that economic sector in which the firm
operates is not an important determinant of credit constraints in the informal credit
market. This implies that there is no need to put emphasis on any particular
economic sector to alleviate credit constraints in the informal credit market. Instead,
the focus should be on poor households, rural-based MSEs and MSEs that are

formally registered as business concerns.

8.3 Shortcomings of the study and implications for future research

8.3.1 Lack of representativeness of the sample

Perhaps the greatest shortcoming of the study is the lack of
representativeness in terms of gender, diversity of the MSE sector, geographical
area covered and composition of the informal credit market.

The proportion of female-operated MSEs in the sample is much smaller than
the proportion of MSEs operated by women in the small business sector in South
Africa. Female-operated MSEs in South Africa comprise a significant proportion of
the MSE sector and actually dominate the microenterprise sector. Thus, female-
operated MSEs are under-represented in the sample.

The selection of subsectors included in the sample was not based on their
importance or size. The main criterion was whether inclusion of a subsector would
result in some variation in the nature of financial transactions. This means that

some subsectors may be over-represented while others are under-represented.

136

Almost half the number of MSEs included in the study were sampled from
one region. Thus, MSEs from other regions were either left out or under-
represented resulting in some bias.

Family and friends are known to dominate the informal credit market for
MSEs in South Africa. Yet, in the sample, transactions involving family and friends
were of only minor importance. Input/raw material suppliers were accorded more
importance in the sample than is the case in the real situation.

Therefore, there is a need for studies using data from more representative
samples to gain a better understanding of what determines credit constraints in the
Northern Province. Such studies should be representative of the MSE sector in
terms of the size of the various subsectors, geographical and gender composition,

and should more effectively cover other components of the informal credit market.

8.3.2 Lack of information on leverage (debt information)

The amount of outstanding debt is an important consideration in deciding
whether to grant a loan or not among both formal and informal lenders. However,
there was no information from the data to enable us to include a variable reflecting
the level debt of each MSE. This is a major shortcoming and, therefore, the study

does not give a complete picture of determinants of credit constraints on MSEs.

8.3.3 Little emphasis on explanations for determinants of credit constraints

This study focuses on what determines credit constraints on MSEs but shed
only limited light on why those constraints occur. As pointed out in eadier sections,
it is important to know the reasons for the existence of credit constraints to

determine what action to take to remove the constraints. Therefore, future research

 

 

137
should focus more on finding explanations for the existence of credit constraints on

MSEs in South Africa.

8.3.4 Notional versus effective demand
Like other studies, a major shortcoming of this study is the focus on notional
rather than effective demand for credit. Future research should seek ways to

resolve this difficult issue of measuring effective demand for credit.

8.3.5 Further breakdown of economic sector

The classification of MSEs according to economic sector is probably too
broad. A further breakdown of the categories would be useful because it would
show the types of businesses within the broad categories that are particularly more

credit constrained. This implies a need for a larger sample.

8.3.6 Effects of credit constraints

The study sheds light on what determines credit constraints but does not go
further to examine the effects of credit constraints. Further research is needed to
determine the effects of credit constraints on the activities of MSEs and on their

ability to generate employment and income.

APPENDIX A

APPENDIX A
September 16, 1997 Questionnaire for Micro and Small Enterprises
Micro-Entrepreneur Identification Number: (for office use only)

Ask to speak to the owner/operator of the business. Explain that the information provided
will be treated as confidential, and will be used by researchers to assist the government in
designing programmes aimed at improving the economic climate for micro and small
businesses. The respondent should be the entrepreneur/owner of the business. If the
respondent does not know the answer to a question code as OK, if the respondent prefers
not to answer a question, code as -1. If the question does not apply, code the response
as not applicable (NA).

 

 

 

 

 

 

 

 

1. Name of interviewer: 1.
2. Date of interview: 2.
3. Region/Province: 3.
4. District: 4.
5. Village/Area: 5.
A. Demographic Information
6. What is your first name and your surname? 6.
7. What is your local name? 7.
8. What is your marital status?

1. Single; 2. Married; 3. Widowed 4. Divorced 8.

If II I | . . | . | .
9. In what year were you married? 9. 19

 

10. Do/did you have a formal marriage contract (YIN) 10.
11. Can/could you sign contracts without getting

 

 

 

 

 

 

 

 

permission from your husband? (YIN) 11.
12. When did you start living in this area? 12. 19
13. Do you have electricity in your home? (YIN) 13.
14. If so, is it:
1. Grid; 2. Generator 14.
15. Do you have electricity in your business? (YIN) 15.
16. If so, is it:
1. Grid; 2. Generator _ 16.
17. Does your family own the house it lives in? (YIN) 17.
18. How many rooms are there in your house? 18.

 

19. What is your main source of drinking water?
1. Unprotected source (such as an open spring
or river); 2. Protected source 3. Public standpipe
or 4. Pipe into house 19.

 

138

139

.5 mm 38
5:85.30 2me “won .5... .0: mm 38 c: 35956 $6283me m ..o < mm ouoo m was < 95 Lo”. 0

.2858 Earn 9 95% 02 .0 E85 .0 ”6:5 mEom 6028-85 wz

5:03 956% .0: Sn vm>o_an:3 z: 5:058.an mciwmm Sn “39635.23 3 xcoficwa m Ema .m_ 65

cocoficmmv n. 5.55 3:586 m 23 .w_ 35 60383: D ”ccwuam .o 5.05% .00ch SE23 ms uccouam

Lo 50:8 .828. .259 mm ”3mm; 58 a he >__E& $595 Go mEo: $5 5 9:03 on; @8928 ..oxcoz

ozmmEoov a ”3mm; m 505:5 mEoc m.>__E£ :26 5: E 9:03 on; 5an m .2 $5 ...mmemEoIV I ”€96.an
-mmmé m>> x90 .LomeiooE 58380 N mm anmxm Lo“— .um>o_an.=mmV mm Ecumasooo .2 mmuoo m

Emcmiwmug .m ”9:201:02 .w 6259 550 N 6mm;

ho EEoucSo .m 68: co Ewan. .m 68; co Ema ..o .0505 .v Home: ho EEO .m 6mm: Go 330% .N 6mm: .F <

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.R at: $8288. 9: e 96% 9: Co 5255 .5

NF
:
9
m
m
N
m
w
v
m

I N 82 93.8 .N

F 82 222 a

mm mm «N mm «a a an
59:86
overland E ..o s: Lune... 2 29.330: .0989:
Eaucflm _oocow mcozuasuuo 23> E om< Eon ..no> xom aim—3:23. Co .3252 20:332..

 

 

 

 

 

 

 

 

 

 

:oEmonEoo Eosomso: .m

140

 

 

 

 

.mm ”0 mmmEmsm m. .m:>> .8
.5 ”m 89.35 m. 85.5 ...m
.8 ”< 89.65 m. 85.5 .mm

”3me .ch co. 9. cosmos". E9. £68 9.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NF
5
or
m
w
n
o
m
v
m
N umoc o_mEou. .N
F one: mu.ms. 9
mm 3 mm mm ., u. ..n . on _ mu .. .3.
E. E. E. E.
.585 .585 .585 E. .585 as E. E. .859:
$9.9m. ommLo>mv $993 EEoE 58 2:535 2.8:. .03 EEoE .53 29.329. .389:
083.95 m. 96:63 <mmm£m=m 05.50.55. wmcoacon. moocmEEom moon; .0 .mnEsz EocomsoI

 

 

 

 

 

 

 

 

 

 

9.5.5.2 NF Hum... 05 ..o>o 9:02.. .o moosom 22.330... .0

141

D. Business Activity

39. What is your most important business activity? 39.

 

 

1. Sewing/alterations; 2. Building/construction; 3. Contracting

and haulage; 4. Vehicle repairs and/or panel beating;

5. Block/brick making; 6. Shopkeeper (Shop, spaza/tuck shop)

7. Hawking; 8. Shoe making/repair; 9. Mat making/basketwork;

10. Pottery; 11. Metalworking; 12. Tanning leather and leather work;

13. Thatching; 4. Electronic/TV/radio repair; 15. Carpenter;

16. Water-haulage; 17. Sculpture/curios; 18. Bottle store; 19. Shebeen;
20. Tavern; 21. Hairdresser; 22. Child minder; 23. Agriculture (cane,
timber, livestock, vegetables); 24. Other (please specify in answer space)

 

 

 

40. How many other businesses do you own and operate? 40.
41. What is the activity of your most important other
business? 41.
(Use above codes)

42. Nature of the specific business to be discussed: 1. Sewing/
alterations; 2. Building/construction; 3. Contracting
and haulage; 4. Vehicle repairs and/or panel beating;
5. Block making; 6. Retailing/Shopkeeping (shop,

 

 

 

spaza/tuck shop) 42.
43. In which year did you start this business activity 43. 19
44. On average, how many months do you do this activity

per year? 44.
45. How is the business organized? 45.

 

1. Sole proprietor; 2. Partnership; 3. Close
corporation (CC); 4. Family business; 5. Working group
46. Is the business formally registered? (YIN) 46.

 

47. Where is your place of business?
1. In the family’s home; 2. Adjacent to the
family's home; 3. In a location outside the family’s
home; 4. On the site of service delivery 47.

 

48. If the place of business is in a location outside the
family’s home, is it 1. Owned/held by the respondent; 2. Rented
3. Used through communal arrangement;

 

 

 

 

4. Other (please specify) 48.
49. Did you inherit the business (YIN) 49.
50.Did you buy the business? (YIN) 50.
51. How much money did you use to start or buy the

business? (Rands) 51 .
52. Did you use assets/money from another business to

start this business? (YIN) 52.

 

53. How many people (including yourself) worked in the
business when it first started or when you
took it over? 53.

 

54. How many of these people (including yourself) were
family members? 54.

 

 

142

E. From what sources did you get the money to startlbuylbulld up your business?
(Please rank the sources in the order of importance according to the respondent.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source Source code Used this source? Order of
(YIN) ImportanceA
55 56 57
Own funds/own savings 1
Remittances received 2
Retrenchment package ‘ 3
Pension 4
Local moneylender 5
Loan from family/friends 6
Loan from KFC/SBDC/ 7
VDC/STOK/LDC
Savings club/stokvel 8
Loan from commercial bank 9
Income from another business 10
(include agricultural enterprises)
Other (please specify) 11

 

 

 

 

 

 

A Rank from 1 upwards

143

F. Employment in the business
Who works in your business, what are they paid, and how often do they work? (Include
the respondent).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Employee Sex Year Wage How Family StatusC Hours Days Weeks
(M/F) began in often is member per per per
work RandsA worker (YIN) day hour year
here paidB
518' _ 59 60 61 A 62_ 63 64 65 66 67

1 (respon.)
2
3
4
5
6
7
8
9
10
11
12
13
14
15

A Calculate Rand value of in kind payments.

B 1. Daily; 2. Weekly; 3. Fortnightly; 4. Monthly; 5. On task basis

C 1. Full-time; 2. Part-time; 3. Seasonal; 4. Occasional; 5. Apprentice

 

144

3.0000 0000.0. .050 .0 ..0__0000 E9. 000:0.00 051 N 5:00 0.206800 .0 50590306
09.300 .0 .5858 05...: ..V 09290002000005 .0 0085.050. .0 amazesocoefioom...” m
.0 0
.050 .8 ”0.5.0.0... 30. 0000:: .NN ”0.00090 0.00:: ..N .0:_:00E 05300 .8 090.050 .0. 0:05.530
.08. .0. 08.02, .t 0.0.20 0 0.8.0 .0. .50 .850 .0. 0522.80 .3 00...... .8500 .2 5.8.350...
3500.0. 05:00E 050.0E .005 .N. 500:0E. 05:00E 050.08 .005 .... 0:0.0 .0. 0000.1 .0 5000.0 .0
000003 ... 0030:01 .0 00:0: .0 ”.200; .... .30. . .0>0 0.. .00.... .0 0.3.0.0.). .N ”0000.00 0:0 0.00 .905. .F <

 

.22 «0.85.0.0...
5 >00 00> 0.0

 

 

m.00.:05
:00. .0 00.000

 

 

m.0000.0.
:00. .0 00.000

.22

0500.0 .0 :000 .:00.
0 .0. >.000 00> 0.0

 

 

 

 

.0200. 008. .o.
._ >2. 8.. 0.08 .055

 

 

0.0.0:. 8...
000:0.00 .0595

 

00005.00 .00>

 

 

(0.80 .002

 

Or .0004 a «0004 a ~000< h «002 o ~000< m .0004 v ~000< n ~000< N .0004 —. «000< £23030

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0E0... 0000.... 00> 0.0 30.. 0:0 :30 >..:0..00 0005000 .00> 0000 0.0000 0532.0. 0:. .0 >:0E 30...
000500.. 0... .0 B000< .0

145

H. Cash loans to finance assets for the business
If a casthan was used to finance any assets listed in table E, please respond to the
following questions concerning the most recent asset financed by a cash loan.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Question 0# Asset
76. Asset code from the asset code table (see page 24) 76
77. How many Rands did you request for the loan? 77
78. How many Rands did you receive? 78
79. How many kilometres away from you is the lender? 79
80. How long do you have to travel to get to the lender? (hours) 80
81. What means of transportation do you use to get to the lender? 81
82. How many times did you have to visit the lender to get the loan? 82
83. Did you sign a written application? (YIN) 83
84. Was the application written in your language? (YIN) 84
85. How many days passed between the application and receipt of the 85
money?
86. What guarantee/security was required? 86
87.What was the interest rate in percent per year? 87
88. How many Rands on average was each installment/repayment? 88
89. How many days passed between receiving the money and the final 89
repayment?
90. How many repayments/installments were required? 90
91. Does the lender come to your home to receive the payments? (YIN) 91
92. Does the lender send regular, on-time statements? (YIN) 92
93. How many Rands do you still owe? 93
94. Had you borrowed from this lender before? (YIN) 94
95. Is it possible to delay repayments if there is a shortage of cash? (YIN) 95
96. Have you ever reduced or stopped making loan repayments? (YIN) 96
97. Will the guarantee/security be enforced if you do not repay? (YIN) 97
98. Does any household member work for the lender? (YIN) 98
99. Does the lender buy any of your products/services? (YIN) 99
100. Does the business buy inputs/raw materials from the lender? (YIN) 100
A 1. Taxi; 2. Bus; 3. Walk; 4. Train; 5. Own vehicle; 6. Other (please specify)
B 1. None; 2. Livestock; 3. Asset purchased; 4. Share of output; 5. Other
asset; 6. Other (please specify)
C Code-2 if the number of days was not specified when the loan was made
D If mailed by post = Y

 

146

I. Credit (hire-purchase) used to finance assets for the business.
If an asset was purchased through a credit (hire-purchase) arrangement, please respond
to the following questions concerning the most recent asset purchased.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Question Q# Asset
101. Asset code from the asset code table (see page 24) 101
102. How many kilometres away from you is the lender? ' 102
103. How long do you have to travel for to get to the lender? (hours) 103
104. What means of transportation do you use to get to the lender?A ‘ 104
105. How many times did you have to visit the lender to get the loan? 105
106. Did you sign a written application? (YIN) ' 106
107. Was the application written in your language? (YIN) _ 107
108. How many days passed between the application and delivery of the ' 108
asset? -

109. What guarantee/security was required? ' 109
110. How many Rands would the item cost if you had paid for it in cash? 110
111. How was the final purchase price determined? 111
112. How many Rands did you have to pay as deposit before the item was 1 112
delivered? '

113. How many Rands was each repayment after delivery? 113
114. How many repayments/installments did you make after delivery? 114
115. How many days passed between the delivery and the final repayment? _ 115
116. What was the interest rate you paid in percent per year? _ 116
117. Does the lender come to your home to receive the payments? (YIN) - 117
118. Does the lender send regular, on-time statements? (YIN) 118
119. How many Rands do you still owe? 119
120. Had you borrowed from this lender before? 120
121. Is it possible to delay repayments if there is a shortage of cash? (YIN) 121
122. Have you ever reduced or stopped making loan repayments? (YIN) 122
123. Will the guarantee/security be enforced if you do not repay? (YIN) 123
124. Does any household member work for the lender? (YIN) 124
125. Does the lender buy any of your product/services? (YIN) _ - 125
126. Does the business buy inputs/raw materials from the lender? (YIN) 126

 

 

 

 

 

A 1. Taxi; 2. Bus; 3. Walk; 4. Train; 5. Own vehicle; 6. Other

B 1. None; 2. Livestock; 3. Asset purchased; 4. Other asset; 5. Other (please
specify)

C 1 Matched to market price; 2. Set by seller; 3. Set by buyer; 4. Negotiated

D Code-2 if the number of days was not specified when the loan was made

E If mailed by post = Y

147

J. Sources of credit refused.
Have you tried to take a loan from any of the following sources to finance an asset during
the past 2 years? Indicate whether you were successful or not.

 

 

 

Source Source Code Asked this Refused by
source for this source?
loan? (YIN) (YIN)

127 i 128 ‘ 129
Local moneylender 1

 

Family or friends

 

KFC/BDCNDC/STOK/LDC

 

Milling company

 

Savings club/stokvel

 

Commercial bank

 

 

 

 

 

 

NO)01#(JDN

Other (please specify)

 

148

3.0000 0000.0.

.050 .0 302.00 .0. 0.003 0.00 .m 0.0.20> :30 .0 .x0. .0 00m .0 0.00.. 00.... .N .00. :0 002.00 .F o
.20E>00 0020>0< .v 0.003 02. E9. 500.0 .m :00. 0 EB. 2000 .N 00:0. :30 2.0.. 2000 .F 0
5.0000 0000.0. .050 .0 000000 .0 0020.... .0 030.2. .0 >_.E0n. .N 000500m .. .0

3.0000 0000.0. .050 N .2000 >.0>__0o .0 05:05:00.0 00.300 .80.. .m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.0f0E.0Q:0t0=90m .v 02.0.0080 .m .0_000_0£>> .N .0030... 003.000.003.095 _000_ =me .—. <
.23. .0... 50> .o.
03. 20000. 202005 .00E 02. 2.0203 .002. 0.2. .0 00020.00 02. 00>> .03
000002800
«3 .0 2000200020.. :0 02000 2.0203 00> 0.0 00:00. >202. 30... .NE
.3 00.0200... 30.0002. 000020.00 02. 200020.. 00> 0.0 30... .2;
03 0000002200 02. .0. >00 00> 0.0 30... .03
00.. 0202000202 2000 .0. 200000005 0.2. .0 .000 .00_0>. 02. 003 .02>> .00.
00.. . 000200 5 ..20 .00 0020 .00.0>. 02. 003 .02>> .00.
...,02 .. 0200000202 .00 00020.00 >..00.0>. 00> 00 302000 .02>> Km.
0.00>
00.. .00. 02. 5 .002. 0.2. .0. 000E 00> 0302 000020.00 >:0E 30... .00.
00.. 000.0000 0.2. 253 0.2020000. .00> .0 0.0.0: 02. 0. .02>> .00.
3... 0.0..0000 0.2. 253 0002.002 0:00 00> 0>02 02.202. >:0E 302 .0n. .00..
00.. .z\>. 0.0..0000 0.2. E0.. 00020.00 2.0.000. 00> on. .00..
N9 00020.00 .0 .500 0. 00202.05. .Nm.
..0. 10.003 .00.0:..n. .5.
on. VN 0000 000. 0.20. 2.0.. 0000 .002...0..0.0E 30m. .00..

v 3.0.0.2 n .0..0.0.2 N .0220: F 0.0.0....

335002. 3010002. 3000002. 3010002. 00 20.00000

 

 

 

 

 

 

 

"0.0..0000 ..02. 0:0 02.20E 0.. .000 02. 5 0000 0>02 00> 0.00:. 020 0.0.0.0.: 30. 00.050 02. 000.02.

0.0..0000 ..02. 020 0.2.0.0.: 30. 020 0.00:. .v.

149

650000 0000.0. .050 .0 .....00 .0 2000 0. .0..0000 N 5.00.0 .0 00.000 00.0.0.0
.050:0 0. 000000 00...... 5.5000 .:0.0E00.0.0..00.000 0>02 .0: 0.0 .m .200... 00. 0.000 ..00.0 ..V

 

 

 

 

 

 

”00.005. 0.. 0.003 .00000. 0:. :0”. .m .....00 05.50:. 00... .0: on. .N .00:.>00 E06500 00... .3 <
.23.
03 005.000 .0..0000 0:. .0 00:0000. 02. 003 .200 0.0 00> ._ N3
03 $00.00 00.. ..:0>0: >23 000.00 .0: 0>02 00> ._ .03
Az\>. .....00.0 :0 000020.00
0.3 00.0... 0.000 00.. .. .0..0000 .00.. 000.00 00.. 0>0... .03
....N 0000 00w.
#3 0.20. .0..0.0E 30500:. 2.0.. 0000 .0..0.0E 30m .03
v .0..0000 m .0..0000 N .0..0000 3.2.0000 *0 :030000

 

 

 

 

 

 

 

"0.0.3.0... 32.0.00... .0 0.2.0000 .20... 000.0 0. 00000< ...

150

 

 

 

 

 

 

 

 

 

n? 0003309 903 mEoE=Smc<£coE>mqe >58: 30: .8?
30393332

mm. 3c: 9: 93 >39: 05 :0 3302 05 c0923 ummmma m>mn >:mE 30: .m9
5... $c0E>mamScwE=EmE 5030 mm; 09203. c0 20ch >53 30: .EF
8? «30> .03 300.03 E 03.. 3225 05 mm; 3&5 .09
m9. 00003309 mm; 3:3033033333 3z>> .9:
«>39:

09 05 :0 3309 0.3 832333 05 c0053 08me m>mu >cmE 301 .9:

V 5,9 ‘ >z\>v «033393. ..30> E 5:25 8:30:33 05 mm>> Km?
33 22 «832.3% 83:; m :33 30> 33 .9:
. $30. 05 33 8 33:0. 05 gm; 2 03; 30> 03 $8: >56 30: .mmw

 

@0335. 05 2 “mm 2 mm: 30> 00 3033:0323: :0 930E 3c>> .va

 

3:305 0.03:9 05 2 3m 2 .2 33: 2 03; 30> 03 3:0. 30: .mmw

 

“:03ch 05 m_ 30> Eot >323 wwhmEoi >53: 30: .Nfl

 

002000.. 30> 03 3:31 >cmE 301 .5?

 

«:32 05 .8 30330. 30> Eu mucmm >cmE 261 .09

 

 

fiance; .91.

 

Ava mama 003 039 3:03E 393335 E0: 0300 3:036 23m .9;

 

 

w 3:035.
3353:.

 

n 3:035.
231330:—

 

N 3:032
Samba...

 

F 3:035.
33530:.

 

 

c2330

 

 

”2333 3532.2 05 03393 $303 30> «mm. 05 .m>o £5.83... 23.. :0 3305 >:m 003:: 9 333 903 230. .330 =

.m3:03& BE ..0 3305 023:: 8 230. .330 .s.

151

 

 

 

 

 

 

 

 

 

 

> - .83 .3 003:. ._ m
:00. 0:. :0 000.90 000.03 0.2 0:. :0:>> 0050030 .0: 003 0>00 .0 .0083: 0:. x «-0000 0
30033 .050 .0 333.30 .0 03:0 41.000420. 0.000023 .N ”0:02 .0 0
3.00%. .050 .0 00:0., 550 .0 :0: .... 03.5 .0 0:0 .0 as» .F 0
330030 000030 .050 K 2:0: 30.08600 .0 003.9033:
0m:_>0w .m ”>:03E00 @255. .v u00.5.0.55020009005— .m ”00:20:23.0... .N u.00:0_>0:0E .003 .F <
0 m: 2.5 30022000303093 .30> .0 >:0 >3: .00:0_ 0:: 0000 N:
E Etc 0.00:2 05.0.1.0; .00E0E 0.0:0030: >:0 0000 A:
0.2. Etc «>030. .0: 00 30> : 000.8:0 0: >._.3000\00.:0.03m 0:. ___>> .0:
>z\>v 00.:0E__0.0:_\0.:0E>030.
mm... :00. 3:03... 00330.0 .0 000300. .0>0 30> 0>0I .09
02 5:6 0:000 .0 030020 0 0_ 0.0:. ._ 0.:0E>030. >0_00 0. 0.90003 3 0_ .09
>3 00.0.0: .00:0_ 0:. E0: 0039.0: 30> 00: >0.
me. 0030 :00 30> 00.00:0m >:0E 301 .09
m9. 205 00.:0E033 0E:-:0 ..0_3m0. 0:00 .00:0_ 0:. 0000 .m?
«or £06 000.:0E>03 0:. 02000. 0. 0E0: .30> 0. 0800 .00:0_ 0:. 0000 .09

 

 

 

 

 

 

 

3.0030 0000.3. .050 .0 533.30 .0 03:0 .v ”00005.33 0.33:. .0 0.00.0020 .N ”0:02 .0
:00. 0:. :0 000.00 000.03 02.. 0:. :05... 00....0030 .0: 002. 0>00 .0 .0053: 0:. ._ «.0000
00.050002 .0 ”.0>30 >0 .00 .0 3:00 >0 .00 .N .00:3 .0:0E 0. 00:03.2 ..

.050 .0 .0_0.:0> :30 m :3... ... 0.33 .m .030 .N 0.0.... .

<moo

 

 

 

 

 

 

 

5!... 3000.5 0. 003 0030.03 . :3000 .0:>> .0.
00.. 0.03330 0.5 5:... 0005030 0:00 30> 0>0: 05:0E >:0E 2.0: ow.
00.. 0.00> .03 .:00.03 :_ 0.03 30> 0.0. .00.0.:_ 0:. 003 .0:>> .00.

00.:0E>030.
0.0.. _0:_.. 0:. 0:0 >.0>__00 05 :002000 000003 0>00 >:0E 30: .00.
00,... 0>.02_00 0.3 00.0.: 30> 0_0 0.:0E__0.0::0.:0:.>030. >:0E 30... .00.

 

0.0.. _ 0>.02_00 .000 .:0E>030. :000 003 0m0.0>0 :0 00:00. >:0E 30... .00.

 

000.0230 0.0.... 060.. 0:.
, mm. . 0.0.00 ..00300 00 >03 0. 0>0: >__00_3>. 30> 0.0 00:00. >:0E 30... .00.

 

 

 

 

152

 

 

00,... a M.000:_E.0.00 00:3 000:0.33 3:... 0:. 003 .50... .00..

,. 00... 002.8 0:. .o 020., :000 02.... 9.. 0 .0:>> .00.
«an. Azi. 0:000 >03 30> ._ 03:00 :000 .0 35:03.. 0 .00 30> 00 .Nm.
.0.... 0.3: 0.0.0. 0:0..000:0.. :000 0 .0. ._:3 .03 00:3 0:. 0. .0:>> ..0.

, an: .. 0.33 0.0.0. 05500035 200.0 0 .0. ._:3 .03 00:3 0:. 0. .0:>> .00.
, 0.00> .00. 0:. :_ 8:30:05

000 . .03 ._00.0 :0 0000:0.33 05.0 03_0> _00_3>. 0:. 00.5 .0:>> .00.

 

0300.330 0.055 ._:3 0.0.0. 0:0_.000:0.. 200.0
00.. H :000 5.? >30 >__00_3>. 30> 00 03:03:. 3050.335 .0 >533... .0:>> .00.

 

03:03... 30533:.
, 00. . 0_:. .0. .00> .00. 0:. :. 0>0: 30> 0.0 0:0..000:0.. ..00 .0 >:0E 2.0... .00.

 

 

 

 

00.0 «0.00:0. 05 0. .00 0.003 30> 00 53.03035 .0 0:00.: .0:>> .00..
0.... 0.30:. 0.029 0:. o. .03 0. .0>0: 0. 0>0: 30> 00 0:0. 2.01 .00.
a 0.00:0. 0:. 0. 30> 60.. >030 00500.03. >:0E 30: .00.
0.0.. $0 0003 000. 050. E9. 0000 3:03.: 30.0335 .00.

 

 

03:035. 03:03.2 03:03.2 20.3.0.2
2.00.235 2.00.233... 2.00.2335 2.00.2335 #0 :0..0030

 

 

 

 

 

 

 

”0:0.00 msflgo. 0:. 025.3 00003 ..00> .00.. 0:. ... 0.0..3330 E9. 200.0 :0 0000:0.33 0.02. 03:03.: 2.050.335 ._
0005030 0... .0. 03:03.: 2.0.3.33... 00:0.... 0. 000: 200.0 .2

153

320030. .050 .0 ”05.0000 ... .00:0>00 5 .:0E>0n. .0 200.0 .N ”:000 .. <

 

30:0... 5. 0:0..000:0..
.03 0003.00 .0 0030 .30> .0 033.. 3.0. 0:. 0. .0:>> .00.

 

 

32:3 00300203093 00:.
05:30:05 .03 ..00 >__00_3>. 30> 00 35:03: .055 .00.

 

 

<00003.00 3.03023
0005 .0. .:0:.>03 .0 ::0. 85:00 .00.: 0:. 0. .0:>> .00.

 

 

 

30:0 .0 080: 80.. 00.03050
0. 0005030 2 .5. o 00 0.00.05 .>.000000: 2 000.0>0 00:.
0003.00 .0 0.00 .0 .503 0. 00:00.00. E2. 0052.05. .3.

 

 

0.00> .03 002200383093
0005.0. 0>0: 30> 00 05:30:05 >:0E 2.0... .00.

 

3.0 0003 :0 000500303093 .0 0.00. 000.
00030.3 30> 0.0 003.00 .0 .03093 .0:>> .00.

 

 

 

0 003.00
20300.3

 

0 003.00
20300.3

 

0 003.00
20300.3

0 003.00
20300....

 

 

. 003.00

20300... 5.3030

 

 

 

 

 

”0...:0E 0.. .003 0:. 0::30 0:0..000:0.. .3o> 00:0000 0000.3
000.200 0:0 0.0300... .0

3.0030 0000.3. .050 ... ..000< .0

.0.0..30 0..0._3330 0330.0. 00.0w .m ..33.30 .0 0.05 ..V 000000.33 E0: .0 .x00.00>... .N 0:02 .. m
.m ..0..3330 >0 .0m .N .0003 .909: 0... 0. 0000.02. F <

00.0

.0302 .v ..0>30 >0 .0w

 

 

m.00..3.00. 30 0.0 ..3000\00.:0.03 .0:>> NE

 

0.00.. .03 .:00.03 :. 003.000 0. 0.0. .00.0.:. .0:>>. Cm

 

0.00200
.0..0 .:0E>030. :000 0. 0m0.0>0 :0 00:03 >:0E .50... .owm

 

3.0200
.0..0 00.0.: >__00_3>. 0.:0..0 00. 00 0.:0E>030. >:0E 30... .38

 

0.:003030.
.05. 0:. 0:0 202.00 :00500 0003 >__00_3>. 0>00 >:0E 30... .wow

 

0..00300 00 >03 >__00_3>. 0.5.8 0... 00 00:03 >:0E 30: .BN

 

<0.00 >__00_3>. 003.00 .0 0.00 0... .0 0003 00000.33 0:. 0. .50... .000.

 

00000
:. .0. 0.03.. 02200303093 0... .0 00:3 0:. 00 0.303 .0:>> .mom

 

003:00 0.... .0 03_0> 0:03 0m0.0>0 0:. 0. .0:>> .vom

 

.23. 0.80.33
0000 .0. 03:00 :000 .0 20:03.0 0 023 >__00_3>. 30> on. .mom

 

0:0..000:0.. ..00.0 0000 .0. 00030.3
30.. .00. 0003.00 .0 00.00 .0 03.0> 0:03 0:. 00.5 .0:>> .Now

 

.0..:3 02200303093 003. 0:0..000:0.. ..00.0
.03 0030.3 >__00_3>. 30> 00 000.200 .0 00.00 .0 20:03.0 .0:>> .SN

 

00m

..0E00 0.0..
00.8080 0. 000:.030 ._ E0. 0 00 0.8.0:. 2000000: .. 0m0.0>0
003. 000.200 .0 0.00 .0 .:.03 0. 00:00.00. E0.. 0000605. .08

 

00.

0..00.0 :0 .00.. .03 .030:00 30> 00 0:0..000:0.. >:0E 301 .mm.

 

”aw

Avm 0303 :0 000.200.0.0300.3
.0 0.00. 00% 00030.3 30> 0.0 00.200 .0 .0300.3 .0:>> .wm.

 

 

0. 0003.00
30300.3

 

0 0003.00
30300.3

 

0 0003.00
30300.3

N 0003.00

 

..03 00.3

 

0 000.200
:03 00.3

 

#0

 

 

:0..0030

 

”05:00.. «P .003 0.... 3:..30 0.00.0 :0 0.00 0>00 30.. 00.02, 0003.00 .0 0.03023 00. 2.0030 0000.3
A20300x3000.0.0_000 .0.0:0m >0 00.030c0 00 0. .02 ...—00.0 :0 00.0w .3

155

029.082 .v um>=n >9 .mm .m 5:33 >9 .3 .N ”8:3 .35: 9.. o. 3:052 .F <

 

30.. m5 :0 5.6.388
:0 >3 >..8_a>. 9:28 on. on 85m. >:mE 26.... .VNN

 

 

ocozommcmb :3 8:39,...
:. >3 2.8.33 3:0..0 50> on mocmm >:mE 26: 63

 

 

8.8 8.28 :0 2mm 9.. S 8.3 89.9.3 9.. m. 26... .mFN

 

 

«58 :_ >3
2 2oz. 35:0 9.. t 8.3 390%.. m5 8 2:03 .33 .mvm

 

 

38:33 :_ EoE>3 5.3 80.28 ucm
8.8 :o 5.83:9. :3 3.? 0:3. .833 05 m. .33 K K

 

 

ABE: 8.288333 8:. 35:53 8:33 :3 8.33
30> on @855 :o m8_>:8 :o 8.3 :o 33:33 .m...>> .o E

 

 

$59. Ea: 863:8
m. 88.3: 2 E. o mm 286:. 53800: ___ 389m 83.
«8.28 :0 2mm .0 .53 on. 2 mngo..x >:mE Bo... .m K

 

 

853.3v :_ EoE>3 5.2, m8.2om :o 8.3 >:mE 26... .vrw

 

2% 33 :0 2:3 68. 8.28 :o 8:85 .m E

 

 

 

m «8.28 e 80.23 n 3033.» N 80.28 F8223 , ......m...“
Eugen. 3033.5. Eugen. 38:69:. 38:09:. «3 55830

 

 

 

 

 

 

 

.9358 up «m3 9.. acts: 8:523 :. 23:33 5.? Bow 93 30> 5...; 80.28 :0 3025.5 .3855 m5 >:0QO .8me
€38.85?ng .833 >n 8.033.“: 3 o. .02. .8523 :. «cacti. 5.3 33m .0

 

156

 

 

 

 

 

 

 

 

 

n- 8:830

 

 

$383 550 .m 635E .v 2838 new .m "58:. 5 Eu .N in“. on .P m

368$ 550 N ”mmmmogmsn .050 .m
”mhmEmm wmm>> .m

”muocmEEm. 5.3me 9609 0:3 maomn. 6 65:0?ch .m ”$210.83 .83 .N ”2E5 .mucm_..u_ .F <

 

«wLoEoumzo :88 § a: £9: $8528 50> w_ “an; .3 w. .mmm

 

Etc $690 :o 20m .0680 L2 8% 5:9: m o>mc 30> cu .6: t .mmm

 

«29 69.95 “as; am .mm> = sum

 

22 $855 $920 a; 8 .mNN

 

manom: w_ EoEomcmtm :uob ho 25 “93> .mmm

 

9590 :0 >33 95:0 50> be $95er «95> .vmm

 

$.53 $305 2 328 8me $3 <2qu co :8 30> on E23 oh .mmm

 

A25 285 S $80 =3 a; on .NNN

 

 

amcoamom

 

owconmwm

muconmam

 

 

wmconmam

 

cozmazc

 

 

2239.205 >3 E35 :0 mofim .m

157

S. Mechanisms for survlving misfortunes.
230. Has your business suffered any misfortunes

during the past Five years? (Y/N) 230.
231. If yes, in what year did the worst misfortune occur? 231.
232. If yes, what was the misfortune that occurred? 232.

1. Stolen equipment and/or goods; 2. Vandalised
equipment or workshop; 3. Loss of client base;
4. Loss of key personnel; 5. Breakdown of equipment;
6. Violence in the area; 7. Natural disaster; 8. Family
related misfortune; 9. Other (please specify)

233. Did you have insurance covering this misfortune?

(YIN) 233.

 

 

 

 

How did you deal with this misfortune? (Mark the strategies, then ask the respondent to

rank them in order of importance).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Action Action YIN Order of
Code importance
1_ {34 235 236

Took a loan from KFC/SBDC/DCISTOK/LDC or NGO 1

Took a loan from commercial bank 2

Took a loan from family/friends 3

Took a loan from stokvel/savings club 4

Sold personal belongings 5

Sold assets 6

Scaled back business operations 7

Got a grant from government or an NGO 8

Self or family member took extra work locally for wages 9

Self or family member migrated to seek extra work for 10

wages

Radically reduced the family’s food consumption 11

Radically reduced the family’s consumption of other goods 12

Delayed paying debts - 13

Drew on personal savings 14

Other (specify) 1 5

 

 

158

T. Savings and Investment Behaviour.
If you had money to save, what did you do with it? (Please rank the responses in order
of importance.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Action Action YIN Order of
' Code * importance
237 238 239
Made a deposit in a bank 1
Made a deposit to a stokvel or a savings club 2
Bought livestock 3
Sent it to a relative or family member who needed it 4
Bought assets that maintained their value 5
Made loans 6
Entrusted the money to someone for safekeeping 7
Kept the cash in a safe place at home 8
Expanded the business 9
Settled outstanding debts 1O
Bought a vehicle or bakkie 11
Other (please specify) 12

 

U. Deposit accounts.
If you have money in a bank account(s) please provide the following information:

 

 

 

 

 

 

 

 

Question 01! Account 1 Account 2

At what bank is the account?A 240

How many Rands are currently in the account? 241

Is the account a: 1. fixed deposit 2. savings or 3. current 242

account?

What is the interest rate in percent per year? 243

How many kilometres away is the bank? 244

How long does it take to get there? 245

What means of transportation do you use to get to the 246

lender?a '

 

 

 

 

 

 

A 1. Volkskas; 2. United; 3. Trust Bank; 4. Standard Bank; 5. First National
Bank; 6. NBS; 7. Perm; 8. ltala; 9. African Bank; 10. Allied; 11. Nedbank; 12.
Post office; 13. Other (please specify)

B 1. Taxi; 2. Bus; 3. Walk; 4. Train; 5. Own vehicle; 6. Other

159

V. Current Accounts.

If you have a current account please provide the following information:

 

 

 

 

 

 

 

 

Question Q# Response
Do you have a current (or cheque) account at a bank? (YIN) 247

Do you have an overdraft facility at a commercial bank? (YIN) 248

What guarantee/security did you offer to secure the overdraft?A 249

Do you use the overdraft to purchase inputs/supplies for the business? 250

(YIN)

 

A 1. Insurance policy; 2. Fixed property; 3. Asset; 4. Other

W. Credit Cards.

If you have an ATM or credit card account(s) please provide the following information:

 

 

 

 

 

 

 

 

Question Q# Response
Do you use an electronic bank card (YIN) 251

Is it: 1. credit card 2. ATM card 3. both? 252

Do you use your ATM/credit card to finance your business ‘ 253

operations? (YIN)

If you have a credit card, what is the credit limit? 254

 

 

 

160

X. Financial lntermediation through participation in Informal savings groups.
Have you contributed to a stokvel, savings club, rotating club, or other type of savings
group during the past year?

 

 

 

 

 

 

 

 

Question Q# Club 1 Club 2 Club 3
255. What type of group was it? y 255

256. How many people belong to the group? 256

257. How many of the members of the group are women? 257

258. How many Rands was the last contribution to the 258 '

group?

259. Does the group function during all times of the year? 259 ‘

(WM

260. How often are contributions made? 260 .

261. How many contributions are made in one year? 261

 

262. In what form do group members receive their savings 262
when it is returned?

 

263. Can you request cash at any time you need it? (YIN) 263

 

 

 

 

 

264. How many times per year does a group member 264
receive cash from the group?

265. How does the group decide who will receive the 265
money collected?

266. How many Rands are paid to the organizer? 266
267. Kilometres to the place where contributions are l 267
collected?

268. What common bond links the group members? (I 268

 

269. Is a social gathering usually associated with collecting 269
contributions? (YIN)

270. Are there problems with collecting contributions? (YIN) 270

 

 

 

 

 

 

 

 

 

271. How many people left the institution during the past ‘ 271
year?
A 1. Stokvel; 2. Rotating stokvel; 3. Savings club; 4. Burial society; 5. Other
(please specify)
B 1. Weekly; 2. Twice monthly; 3. Monthly
C 1. Cash, 2. Goods
D 1. No set rule; 2. Rotation by age 3. Rotation by seniority in the savings

group; 4.Rotation by seniority in social status; 5. By lottery; 6. By negotiate
among group members; 7. The organizer decides; 8. Whoever needs it the
most; 9. Other (please specify)

E 1. No common bond exists; 2. Members of the same neighborhood; 3.
Same economic activity; 4. Same sex; 5. Same family; 6. Same age 7.
Colleagues; 8. Same area; 9. Other (please specify)

161

Y. Burial Societies.

 

 

 

 

 

 

 

 

 

Question 01! Society 1 Society 2
1. Are you a member of a Burial Society? (YIN) 272
1. If yes, how much have you contributed to date? 273
2. How often are contributions made“? 274
3. What is the regular amount you have to pay? 275

 

A 1. Weekly; 2. Twice monthly; 3. Monthly

276. What is the address of the business?

277. What is the name of the neighbour you know best? 277.

Thank you.

276.

 

 

 

162

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Asset Code Table:

Asset Name Asset Asset Name Asset
Code Code
Motor cars and bakkies 1 Block making machine 12
(electric/mechanical)
Motorbike 2 Cement mixer 13
Truck (ie. over 1 ton) 3 Scaffolding 14
Tractor 4 Spray gun 15
Trailer 5 Block & tackle 16
Harrower 6 Welder 1 7
Weeder 7 Tools/equipment 18
Plough 8 Generator 19
Ridger 9 Sewing machine 20
Crane 10 Unsold products 21
Block making machine 11 Unused raw materials 22
(manual)
Other (specify) 23

Raw materials/Inputs Table:
Raw material/Input Number Raw material/Input Number
Fuel (petrol/diesel) 1 Timber/timber products 7
Vehicle 8 equipment maintenance 8 2 Cloth/sewing materials 8
servaces
Cement 3 Paint/solvent 9
Sand/ash 4 Vehicle parts 10
Bricks/blocks 5 Stock for shop 11
Hardware/fitting: 6

Products/services Table:
Products and services # Products and services #
Sewing 1 Vehicle repairs 9
Garment alterations 2 Vehicle servicing 10
House building 3 Panel beating and spray 11

painting

House alterations/repairs 4 Block making 12
Other building jobs (eg. Shops, clinics, 5 Brick making 13
classrooms, churches, halls etc)
Haulage 6 Retail trading 14
Ploughing 7
Cane cutting 8

 

 

 

 

BIBLIOGRAPHY

BIBLIOGRAPHY

African National Congress, 1994. IbLBemnstmctionjndDelelopmenLEmgramme.

Johannesburg: Umanyano Publications.

Ahmed, Z.B., 1982. Transaction Costs in Rural Financial Markets in Bangladesh: a Case
Study of a Rural Credit Market. Ph.D Dissertation, University of Virginia.

Amemiya, T., 1981. “Qualitative Response Model: a Survey.” JnumaLoLEQQanlc
Litenatute, 19: 481-536.

Anderson, J., 1990. "Does Regulation Improve Small Farmers' Access to Brazilian Rural

Credit." JoumaLoLDexelome-mtficonomim. 33:67-87.

Ardington, E. M. and Lund, F. 1996. “Questioning Rural Livelihoods.” In: Lipton, M. Ellis,

F. and Lipton, M. (eds..) WWW
WWW. Durban: Indicator Press

Areetey, E. and Udry, C., 1995. “The Characteristics of Informal Financial Markets in
Africa.” Paper Presented at the Bi-annual Research Conference of the African
Economic Research Consortium, Nairobi, Kenya.

Barham, B.L., Boucher, S. and Carter, MR, 1996. "Credit Constraints, Credit Unions, and
Small-Scale Producers in Guatemala." MW 24(5): 1-14.

Baydas, M., Meyer, R.L. and Aguilera-Alfred, N., 1994. “Discrimination Against Women in
Formal Credit Markets: Reality or Rhetoric?" WorldJlelelmeent 22(7): 1073-1082.

Besley, T., 1994. “How do Market Failures Justify Interventions in Rural Credit Markets?”
IbeflofliBankflbsemer 9(1): 27-47.

Binswanger, H. McIntire, J. and Udry, C. ,.1989 “Production Relations In Semi-Arid African
Agriculture.” In: Bardhan, P. (ed..) IbeEmnomiclbeonLoLAgradaanstitutions.

Oxford: Clarendon Press.

Blumenfeld, J., 1997. “From Icon to Scapegoat: The Experience of South Africa’s
Reconstruction and Development Programme.” DexelopmenLEolflReyiew 15: 65-
91.

Boomgard, J.J., De Santis, D., Malhorta, M.,Timberg, T. and Tzavaras, a., 1989. ”Taking
Stock of A.I.D.'s Microenterprise Portfolio: Background and Conceptual Overview."

163

 

164

Draft A.I.D. Evaluation Occasional Paper, Washington, D.C.:U.S. Agency for
International Development.

Boomgard, J., 1991. "A.I.D. Microenterprise Stock-Taking: Synthesis Report." A.I.D.
Evaluation Occasional Paper. Washington, DC: US. Agency for lntemational
Development.

Branch, B., 1996. Underground Production in Peru: The Informal Sector. Ph.D
Dissertation, University of Wisconsin, Madison.

Burawoy, M., 1975. “The Functions and Reproduction of Migrant Labor. Comparative
Material from Southern Africa and the United States." Americamdnumalfl

Sociology 81(5): 1050-1087.

Bureau for Market Research, 1993. "Personal Income of the RSA-TBVC Countries by
Population Group and Magisterial District." Paper Presented at the Finance 2000
Workshop, May 27-28, Pretoria.

Carter, M., 1989. "The Impact of Credit on Peasant Productivity and Differentiation in
Nicaragua." JoumaLoLDexelopmenLEconomiss 31:13-36.

Carter, M., Barham, B., Boucher, S. and Zegara, E., 1994. ”Closing the Financial Efficiency
Gap: The Role and Design of Cooperative Financial Intermediaries in Low Income
Countries." Unpublished Mimeo, University of Wisconsin, Madison.

Carter, M.J. and May, J., 1997. “Poverty, Livelihood and Class in Rural South Africa.”
Agricultural and Applied Economics Staff Paper 408. University of Wisconsin,
Madison.

Carter, M. and Olinto, P. 1996. "Getting Institutions Right for Whom? The Wealth-
Differentiated Impact of Property Rights Reform on Investment and Income in Rural
Paraguay.“ Unpublished Paper, Department of Agricultural Economics, University
of Wisconsin, Madison.

Carter, M. R. and WIebe, K. D., 1990. ”Access to Capital and its Impact on Agrarian

Structure and Productivity In Kenya." AmericanJoumaLoLAgriculturaLEconomjcs
72(5). 1146- 1150.

Central Statistics Services, 1996. Population Census 1996. Central Statistical Services,
Pretoria.

Coetzee, G.K., 1988. "Die Finansiering van Landbouproduksie in die Kleinboersektor van
Suid Afrika." Unpublished Master’s Thesis, University of Stellenbosch.

Coetzee, G.K., Kirsten, MA. and Christodoulou, N.T.,1994. “Financing Entrepreneurs in
South Africa: a New Approach.” Paper Presented at the Development Bank of
Southern Africa, Halfway House, South Africa.

165

Coming J., 1995. "Mixing and Matching Loans: Credit Rationing and Spillover in a Rural
Loan Market in Chile." Working Paper, Department of Economics, Yale University.

Cuevas, CE, 1984. Interrnediation Costs and Scale Economies of Banking under Financial
Regulation in Honduras. Ph.D Dissertation, Ohio State University, Columbus, Ohio.

Cuevas, OE. and Graham, DH, 1984. "Rationing Agricultural Credit in LDCs: The Role
and Determinants of Transaction Costs for Borrowers." Agricultural Finance
Program, Ohio State University, Columbus, Ohio.

Data Research Africa, 1995. "The Composition and Persistence of Poverty in Rural South
Africa." Unpublished Report. Durban: Data Research Africa.

Deininger, K. and Squire, L., 1996. “a New Data Base Measuring Inequality.” Mimeo.
Washington, DC: World Bank.

Department of Trade and Industry, 1995. "National Strategy for the Development of and
Promotion of Small Business in South Africa." White Paper. Department of Trade
and Industry, Pretoria.

Fafchamps, M., 1997. “Trade Credit in Zimbabwean Manufacturing.” Worldflexelopmem
25(5): 795-815.

Feder, G. Lau, L..,J Lin, J..Y and Luo,X.,1990. "The Relationship between Credit and
Productivity in Chinese Agriculture. a Microeconomic Model of Disequilibrium."

AmericanJoumaLoLAgriculturaLEconomics72(5)z 1151- 1157.

Final Report of the Commission of Inquiry into the Provision of Rural Financial Services,
1996. Pretoria.

Fry, M., 1995. Money, Interest, and Banking in Economic Development. Baltimore: The
Johns Hopkins University Press.

Gonzalez-Vega,C.,1984. “Credit-rationing Behavior of Agricultural Lenders. The Iron Law
of Interest Rate Restrictions. " In: Adams, D. W. Graham, D. H. and von Pischke,

J. D. (eds) UnderminingliuraLDexelopmemmnmeapCredit. Boulder, Colorado:

Westview Press.

Government of South Africa, 1995. “Key Indicators of Poverty in South Africa.” Ministry in
the Office of the President, Reconstruction and Development Programme. Pretoria.

Guia-Abiad,V.,1993. "Borrower Transaction Costs and Credit Rationing In Rural Financial

Markets: The Philippine Case. " IbeDexelopingEconemies 31(2): 208-219.
Gujarati, UN, 1995. W. New York: McGraw-Hill, Inc.

 

166

Inter-American Development Bank, 1983. "Agricultural Development Bank: Evaluation of
Global Agricultural Credit Programs in Panama." Washington, DC: Inter-American
Development Bank.

Inter-American Development Bank, 1983. "Banco Agrario del Peru: Evaluation of Global
Agricultural Credit Programs in Peru." Washington DC: Inter-American
Development Bank.

Heidhues, F. , 1990. “Cameroon.” In: Masini, M. (ed.). BuLaLEinamELQflleanAfflcan
Countries. Vol. 2. Milano, Italy: Finafrica-Caprilo.

Hoff, K. and Stiglitz, J.E., 1990. "Introduction: Imperfect Information and Rural Credit
Markets - Puzzles and Policy Perspectives." WW 3.

Jaffe, D.M. and Russell, T., 1976. "Imperfect Information, Uncertainty, and Credit
Rationing.” OuarterlicumalgLEconomics. 90: 651-666.

Kahn, M., 1983. Regional Shopping Centres Research Project: a Study of Regional
Shopping in Metropolitan Durban. Unpublished Report. Pietermaritzburg: Natal
Town and Regional Planning Commission.

Kennedy, P., 1992. afiuideJQEccncmetcim. Cambridge, Massachusetts: The MIT Press.

Kilby, P., Liedholm, C. and Meyer, R., 1984. "Working Capital and Nonfarrn Rural
Enterprises." In: Adams, D..,W Graham, DH. and von Pischke, J..D (eds.).

UndenmnmgfluraLDexelcpmenthfibeapfiredrt. Boulder, Colorado. Westview

Press.

Kirsten, J.,1996. “The Potential for Creating Additional Rural Livelihoods in Agriculture and
the Rural Non-Farm Sector in Semi-Arid Areas: a Case study in the Northern
Province.” In: Lipton, M., Ellis, F. and Lipton, M. (eds..) LaniLabQLandLnLelihmfi

I: - o I;I.- 0 II‘ .o' .1. -L-.- IOLOII "'0 ..Durban
Indicator Press.

Kochar, A., 1992. "An Empirical Investigation of Rationing Constraints in Rural Credit
Markets in India." Unpublished Paper, Department of Economics, Stanford
University.

Krige, DJ, 1988. “The Prospects for lnforrnal Small Businesses in KwaMashu.”

DevelopmentfioutbemAfrica 5(2): 168-185.

Ladman, J. R, 1984. ”Loan Transactions Costs, Credit Rationing and Market Structure: The
case of Bolivia. " In: Adams, D..,W Graham, D. H. and von Pischke, J. D. (eds..)

UndennmmgfiuraLDexelcpmenmItbfibeapfiLedn. Boulder, Colorado. Westview

Press.

Liedholm,C.,1992. "Small-scale Enterprise Dynamics and the Evolving Role of Informal
Finance." In: Adams, D..W and Fitchett, D.A. (eds..) mIQLmaLEinancejnm
incomchuntries. Boulder, Colorado: Westview Press.

167

Liedholm C. and McPherson, M.A., 1991. “Small-scale Enterprises in Mamelodi and
Kwazakhele Townships, South Africa: Survey Findings.” GEMINI Technical Paper
No. 16. Prepared for the Office of Marketing Development and Investment, Bureau
for Africa, USAID, Growth and Equity through Microenterprise Investments and
Institutions (GEMINI), Maryland.

Liedholm, C. and Mead, DC, 1987. "Small-scale Industries in Developing Countries:
Empirical Evidence and Policy Implications." lntemational Development Paper no.
9. Department of Agricultural Economics, Michigan State University, East Lansing,
MI.

Ligthelm, AA. and Kritzinger-van Niekerk, L., 1990. “Unemployment: The Role of the Public
Sector in Increasing the Labor Absorption Capacity of the South African Economy.”

DexelcpmenLScutbemAfrica 7(4): 629-641.

Llanto, GM, 1990. “Asymmetric lnforrnation in Rural Financial Markets and Interlinking of

Transactions through the Self-help Groups." Sayingiandfleyfilnmnam. 14(2).

May, J. 1996. “Assets, Income and Livelihoods In Rural KwaZqu-Natal. ” In: Lipton, M.

Ellis, F. and Lipton, M. (eds..) WWW
WWW Durban: Indicator Press.

McPherson, M.A., 1992. Growth and Survival of Small Southern African Firms. Ph.D
Dissertation. Michigan State University, East Lansing, Michigan.

Miller, LR, 1977. W New York: Rockefeller

Foundafion.

Moore, B.J. and Schoombee, GA, 1995. “Bank Credit to the Informal Sector: Challenge
and Reward.” DevelapmentficutbmAfrica 12(3): 347-361.

Mushinski, D., 1995. "Credit Unions and Business Enterprise Access to Credit in
Guatemala." Unpublished Paper, The College of William and Mary, Williamsburg,
Virginia.

Norusis,M.,1993. Wm Illinois, Chicago: SPSS Inc.

Pohlmeier, L. and Thillairajah, S., 1989. “Review of Rural Financial Services in Sub-
Saharan Africa." Africa Technical Division Paper, World Bank, Washington, DC.

Quarcoo, P., 1979. “Credit Allocation in Segmented Capital Markets of Underdeveloped
Economies.” Savingsandfielelopment, 3(3): 199-213.

Robinson, MS, 1995. "Microfinance: The Paradigm Shift from Credit Delivery to
Sustainable Financial Interrnediation."

168

Rudra, A., 1982. Indian Agricultural Economics. New Delhi: Macmillan Company of India.

SAIRR, 1993. W. Johannesburg: South African Institute of Race
Relations.

Southall, R., 1980. “African Capitalism in Contemporary South Africa.” JnumamLSQutbem
Africanfitudies 7(1):38-70.

Stiglitz, J. E. and Weiss, a., 1981. "Credit Rationing in Markets with Imperfect Information."
AmericamEccncmmRexiew, 72: 912-927.

Stiglitz, J. E., 1987. “The Causes and Consequences of the Dependence of Quality on
Price.” JoumaLcLEccncm'LcLiteLatuLe, 25: 1-48.

Tendler, J., 1996. “Support for Small Enterprises in South Africa: Impressions from
Interviews with Firms, their Corporate Customers, and their Service Providers.”
Paper for the South African Department of Trade and Industry of the Ministry of
Trade and Industry. Department of Urban Studies and Planning, Massachusetts
Institute of Technology, Cambridge, Massachusetts.

Udry, C., 1990. “Credit Markets in Northern Nigeria: Credit as Insurance in a Rural
Economy.” WorldlBanILEccnomicfleview 4(3):251-269.

von Pischke, J. D., 1983. “a Penny Saved: Kenya’ s Cooperative Savings Scheme. ” In: von
Pischke, J. D., Adams, DW. and Donald, G. (eds). BuraLEinanmaLMarkeIfln
DelelcmngfiuntuesteLUsLandAbuse. Baltimore: The Johns Hopkins

University Press.

von Pischke, J. D. and Adams, D. W.,1983. "Fungibility and the Design and Evaluation of
Agricultural Credit Projects. " In. von Pischke, J. D. Adams, D. W. and Donald, G.

(eds..) BuraLEinanciaLMarketumllexelcpingmumnes. Baltimore: The Johns

Hopkins University Press.

Wickrama, K. A. S. and Keith, P. M. 1994. "Savings and Credit: Women' s Informal Groups

as Models for Change In Developing Countries." IheJnumaLoLQexelnpingALea.
28: 365-378.

World Bank, 1993. Characteristics of and Constraints facing Black Business in South
Africa: Survey Results. Draft Report. Washington, DC: World Bank.

World Bank, 1994a. W. New York: Oxford University Press.

World Bank, 1994b. "South African Agriculture: Structure, Performance and Options for the
Future." Discussion Paper. Washington, DC: World Bank.

Zeidler, M., 1994. “Financial Services and Small and Microenterprises in South Africa:
Status, Problems and Potential." Paper Prepared for Friedrich Ebert Stiftung,
Johannesburg, and The Ohio State University, Department of Agricultural
Economics and Rural Sociology, Columbus, Ohio.

 

169

Zeller, M., 1994. "Determinants of Credit Rationing: a Study of lnfcrrnal Lenders and Formal
Credit Groups in Madagascar." Mddfiexelnpmem, 22(12): 1895-1907.

MICHIGAN smre UNIV. LIBRARIES
IIHI “III II III ”I ”ll" Ill ll WI Ill llll “ll lllll HI llllll
31293017129242