THE WORLD B A.NK

EDT Discussion Paper

EDUCATION AND TRAINING SERIES

Report No. EDT46

The Monetary and Non-MonetaryReturns to Education in Africa

Keith Hinchliffe

December 1986

Education and Training Department Operations Policy Staff

The views presented here are those of the author(s), and they should not be interpreted as reflecting those of the World Bank.

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Discussion Paper

Education and Training Series

Report No. EDT46

THE MONETARY AND NON-MONETARY RETURNS TOEDUCATION IN AFRICA

Keith Hinchliffe(consultant)

Policy DivisionEducation and Training Department

December 1986

The World Bank does not accept responsibility for the views expressed herein,which are those of the author and should not be attributed to the World Bank orits affiliated organizations. The findings, interpretations, and conclusionsare the results of research or analysis supported by the Bank; they do notnecessarily represent official policy of the Bank.

Copyright © 1986 The International Bank for Reconstruction and Development!The World Bank

ABSTRACT

This paper surveys the literature on the returns to education in

Sub-Saharan Africa. Income and/or output differentials in the urban

formal and informal sectors and in agriculture are first described and the

implications for rate of return patterns discussed. This is followed by

a discussion of aspects of eduation-migration relationships. Next, survey

results of the effects of education on fertility, infant mortality and

nutrition are presented and competing sets of implications highlighted.

Finally, at a very general level, the effects of education on a variety of

other non-monetary variables are considered.

Table of Contents

Page No.

EXECUTIVE SUMMARY ......................................... i-iii

INTRODUCTION ........... 1 - 2

SECTION ONE: EDUCATION AND ECONOMIC PRODUCTION ........... 3 - 22Education and Earnings in the Formal Sector .......... 5 - 13Education and Earnings in the Urban Informal Sector.. 13 - 18Education and Output in Agriculture .................. 18 - 21Summary and Implications ............................. 21 - 22

SECTION TWO: EDUCATION AND MIGRATION ..................... 23 - 27

SECTION THREE: EDUCATION DEMOGRAPHIC CHANGE AND HEALTH... 28 - 40Education and Fertility .............................. 28 - 33Education and Mortality .............................. 33 - 37Education and Nutrition .............................. 37 - 39Summary and Implications ............................. 39 - 40

SECTION FOUR: SOME OTHER RETURNS TO EDUCATION ............ 41 - 48Income Distribution .................................. 41 - 42Social Mobility ...................................... 42 - 44Savings and Consumption .............................. 44 - 45Localization ......................................... 45Women's Status ....................................... 45 - 47Personal Development ................................. 47 - 48Transmission of Ideology ............................. 48

REFERENCES ................................................ 49 - 54

EXECUTIVE SU1Th4ARY

Discussions of the returns to education in Sub-Saharan Africa

have hitherto largely focused on direct monetary aspects. In this paper,

these aspects are again concentrated on, but, in addition, consideration is

given to a wider set of indirect monetary effects and to a number of

nonmonetary ones.

The first section considers the direct relationship between

education and labor productivity and begins by presenting the results of

existing rate of return studies for 16 African countries. These indicate a

very clear pattern of lower returns to successively increasing levels of

education. The data, on which the individual studies are based, are then

discussed and some of their inadequacies highlighted. Of these, the age of

the data and their general limitation to formal urban sector workers are

emphasized. In the main part of this section, therefore, more recent

African data on trends in wage differentials in the formal sector are

provided together with a presentation of earnings differentials in the

urban informal sector. Finally, reviews of ten studies analyzing the

linkages between farmer education and agricultural productivity or the

willingness to adopt new practices are presented. The overall conclusions

are that rate of return studies utilizing current earnings data across all

economic sectors are likely to result in the returns to primary, secondary,

and higher education being much closer to each other than the existing

studies suggest.

- ii -

Economic returns associated with education may result not only

from higher productivity within jobs, but also from more efficient behavior

in the labor market. One aspect of this is rural-urban migration. In

section two, the results of African studies linking education to migration

and the consequent effects on earnings are discussed. These linkages

appear to be very strong with positive implications. Evidence on levels of

migrants' remittances to rural areas suggest these are also high and this

is again regarded as a positive feature of migration.

In the third section, the effects of education on behavior

influencing fertility, child mortality, and child nutritional status are

discussed on the basis of empirical studies which have recently been

supplemented by the World Fertility Surveys. These included ten

Sub-Saharan African countries. Reviews of the surveys indicate that while

the effect of education on reducing fertility is lower in Sub-Saharan

Africa than elsewhere, it is still significant particularly after the first

few years of schooling. The relationships between parental education and

infant and child mortality rates are similarly strong according to the data

across all regions of the developing world in general and Africa in

particular. Again, the effects of post primary schooling appear to be

particularly strong. In the final part of this section, consideration is

given to the relationship between mothers' education and child nutritional

status. While there are only a limited number of empirical studies of this

relationship for Sub-Saharan Africa, those that exist show it to be

positive. The studies referred to in this section mainly focus on the

- iii -

individual and tend to emphasize the importance of post primary education.

An alternative viewpoint, however, is that changes in fertility and child

health practices are a result of changes in community mores which are

linked to the level of education in the community as.a whole. This

argument has the major implication that a very wide spread of basic

education is necessary for any major demographic transition to occur. The

issue of the necessary levels of threshold requires considerable attention.

The returns from education to individuals and to society are not

confined to the labor market and better health. In the final section, a

limited discussion of the effects of education on income distribution,

social mobility, savings and consumption, localization, women's status,

personal development, and the transmission of state ideology is provided.

The treatment is admittedly partial and aimed solely to pointing to these

areas in very general terms in an attempt to initiate further discussion on

the overall impact of education in Sub-Saharan Africa.

Monetary and Non-Monetary Returns to Education in Africa

Introduction

Twenty years ago, discussions of the 'returns to education'

concentrated almost exclusively on calculations of direct economic

returns focusing narrowly on productivity increases proxied by

incremental earnings associated with educational level. In the

developing countries these calculations, in turn, were based almost

exclusively on earnings in the formal wage earning sector. Today, as a

result of increased empirical knowledge of the relationships between

levels of education and productivity in the urban informal sector and in

agriculture, and an extension of the theoretical framework to cover other

aspects of labor market behavior, theories cf income distribution and

social mobility, incorporation of parts of the new household economics

and the economics of fertility, health and nutrition, the returns to

education can now begin to be placed in a much wider overall framework. 1

In addition, some progress has been made in deciphering the cognitive and

attitudinal changes which accompany education and which then effect a

whole range of social behavior.

Much remains unknown, particularly for the developing countries.

At the theoretical level, the nature of education's precise effects on a

wide variety of outcomes remains speculative and isolation of the causal

mechanisms is still largely at an early stage of hypothesis testing.

Complicating this situation, many of these mechanisms are likely to be

culture-specific. Empirical findings, therefore, for one part of the

developing world may not be generalizable to other parts. As will become

obvious at times in this paper, documentation and quantification of

several of the effects of education are more scarce for African countries

than for those of Asia and Latin America.

The content of this paper covers the monetary and non-monetary

returns to education in Africa. The structure, however, cannot follow

this simple division. Gross increases in productivity caused by

education can, in principle, be directly captured by measures of

increased earnings or increased output. Some of the education-induced

causes of increased earnings and output, however, are indirect and also

For discussions of this wider framework for the developed countries,

see Michael (1982) and Haveman and Wolfe (1984).

-2-

have non-monetary and also non-economic effects (e.g. increases in child

nutritional status resulting from parents' education). Other

education-induced outcomes have economic but non-monetary returns (e.g.

more efficient household production and consumption). As a result it is

not always possible to categorise each outcome as having only monetary or

non-monetary returns. In the paper, therefore, each education-induced

outcome is taken in turn with, where appropriate, both aspects covered

together.

The first section is directly related to economic production and

the effect on an individual's earnings and output of educational

expenditure on that individual. Data is presented for the formal wage

earning sector, the urban informal sector and farming.

Increases in earnings (and for agriculture, output) associated

with additional education result, it is argued, largely from increases in

productivity brought about by skill acquisition, broadly defined. They

also result from more efficient behavior in the labor market via

migration, information gathering and job search. Migration also leads to

remittances flowing to rural areas. Section two discusses these aspects

of migration in Africa.

In section three, the effects of education on what may be broadly

termed health and demographic change are discussed. These *are broken

down into education's impact on fertility, infant and child mortality and

nutritional status. Better health is an end in itself. It also has a

variety of economic implications. The relationships between education

and health may be looked at in three ways. First, educated parents may

adopt practices having positive health effects both for themselves and,

particularly, for their children. Second, healthier, better nourished

children may be more receptive to education. Third, healthier adults may

have a longer working life.

In the final section, several perhaps less obvious areas yielding

potential returns to education are considered. Included in the

discussion are the impact of education on income distribution, social

mobility, savings, consumption patterns, localization of high level jobs,

women's status, personal development and the transmission of state

ideology. The treatment is often admittedly skimpy and somewhat

superficial. The intention is simply to point out that very many aspects

of economic, social and political development are effected by the

widespread provision of education.

-3-

Section One

Education and Economic Production

According to human capital theory, education changes people and in

ways which increase their productivity. This increase, it is argued, can

be proxied by measures of labor incomes. According to Psacharopoulos

(1985), data on labor incomes by education level have been collected now

for 16 African countries and set beside educational costs to calculate

rates of return to education. The results are described in table 1.

Taking the most recent estimates for each country the average social

returns are (in percent):

Primary Secondary Higher

27 17 12

Levels and patterns of rates of return to education can be powerful

indicators of the economic efficiency with which resources are being

used. The present studies we have for Africa suggest that while returns

to all levels of education are high, they are highest at the lower

levels. Most of these studies, however, suffer from data limitations and

are in need of revision in at least two respects. First, many were made

using earnings estimates which today are between ten and twenty years

old. While these estimates may have reasonably reflected the state of

the labor market at that time it is less likely that they reflect the

present state. Second, the studies invariably utilised earnings from

only the formal wage earning sector (even for primary school leavers),

most often assumed full employment and in some cases were simply based on

government salary scales. Very few of the studies used cross sectional

age-education-earnings survey data even confined to the formal sector.

More, and more careful, economy-wide earnings surveys are required,

together with other labor market indicators, to provide a contemporary

picture of the employment prospects of African school leavers in the

1980's and an assessment of the economic returns which can be expected

from investment in the various levels of education. In the absence of

such surveys, the following pages take a necessarily second best and ad

hoc course by extending and adding to the types of data which have been

used so far in the African rate of return studies. First, the results of

recent work by the International Labour Office and World Bank on earnings

structures and trends in the formal sector are discussed. This is

followed by some documentation of education-earnings/output relationships

in the urban informal sector and in the rural sector. Finally, a

tentative discussion of the possible implications of these data for the

pattern of rates of return in Africa is presented.

-4-

Table 1

The Returns to Investment in Education by Level, Africa (percent)

Social Private

Country Year Prim. Sec. Higher Prim. Sec. Higher

Botswana 1983 42.0 41.0 15.0 99.0 76.0 38.0

Ethiopia 1972 20.3 18.7 9.7 35.0 22.8 27.4

Ghana 1967 18.0 13.0 16.5 24.5 17.0 37.0

Kenya 1971 21.7 19.2 8.8 28.0 33.0 31.0

1980 13.0 14.5

Lesotho 1980 10.7 18.6 10.2 15.5 26.7 36.5

Liberia 1983 41.0 17.0 8.0 99.0 30.5 17.0

Malawi 1978 15.1

1982 14.7 15.2 11.5 15.7 16.8 46.6

Morocco 1970 50.5 10.0 13.0

Nigeria 1966 23.0 12.8 17.0 30.0 14.0 34.0

Rhodesia 1960 12.4

Sierra Leone 1971 20.0 22.0 9.5

Somalia 1983 20.6 10.4 19.9 59.9 13.0 33.2

Sudan 1974 8.0 4.0 13.0 15.0

Tanzania 1982 5.0

Uganda 1965 66.0 28.6 12.0

Upper Volta 1970 25.9 60.6

1975 27.7 30.1 22.0

1982 20.1 14.9 21.3

Source: Psacharopoulos (1985) Appendix Table A-1.

- 5 -

Education and earnings in the formal sector

Since wage earning formal sector employment has been the major

motivation for the demand for schooling in African countries, it is

appropriate to begin the consideration of monetary returns in this

sector. Two factors set the context. First, the proportions of the

total labor force employed in this sector remain low and second these are

dominated by the public sector. Table 2 demonstrates these points for

seven countries.

Table 2

Formal Sector and Public Employment in Selected African Countries

Country Year Formal Sector Public Employment

Employment

Total Labor Force Formal Sector Employment

Gambia 1977 10.4 76.0

Ghana 1980 12.3 75.8

Kenya 1979 16.0 46.9

Liberia 1979 22.0 43.6

Somalia 1978 8.0 93.0

Tanzania 1976 7.3 67.0

Zambia 1976 24.0 71.6

Source: International Labour Office (1982) p1 4

Throughout the 1970's, public sector employment grew rapidly in

most African countries as table 3 shows. Private sector employment grew

less rapidly. For example, the 8 percent public employment growth rate

reported for Kenya compares with a 1.1 percent annual growth in the

private sector.

The rather comforting picture of rapid expansion in the sector

dominating the employment aspirations of school leavers in the 1970's is

changing in the 1980's. Public sector employment expansion is slowing

down considerably while the numbers of school graduates continue to

increase. In several countries over the last few years employment

Table 3

Annual Growth Rates of Public Sector Employment

Country Period Annual Growth (%)

Gambia 1976-81 14

Kenya 1971-80 8

Liberia 1973-79 6

Mali 1975-79 23

Somalia 1971-78 17

Togo 1975-79 15

Source: Government and World Bank estimates

guarantees have been rescinded, even for university graduates; and there

is widespread pressure to stabilise rather than expand public employment.

Some governments have gone further. In Ghana, in 1979, a scheme was

announced to turn 1000 government messengers to self employment to be

followed by a similar scheme for clerical assistants.

The Third National Development Plan of Zambia (1979-83) provides an

example of the differences between expected new wage jobs and school

graduates. (Table 4) At each level, graduates exceed wage jobs by a

ratio of three or four to one. Similarly, Kenya's 1979-83 Development

Plan envisaged 60,000 new jobs compared to 68,000 leavers with full

secondary schooling, 50,000 with junior secondary schooling and 100,000

with primary schooling. While exceptions to this pattern exist (Somalia,

Gambia) on the whole it is clear that while formal sector employment is

likely to remain an option for completed secondary school leavers and

above, those with only primary and a few years of secondary schooling are

already increasingly finding that employment has to be taken up outside

the formal wage-earning sector. Whether this then implies a reduction in

the returns to schooling depends on opportunities and incomes in the

informal and agricultural sectors and also on the earnings structure, and

likely changes in this, in the formal wage earning sector. Discussion of

the impact of education in the non-formal sectors is postponed until

later. Below, trends in the formal sector earnings structure are

described.

Table 4

Expansion of Wage Employment and Total Output of the Education Sector

1979-83, Zambia

Education Level Wage Employment Output from

Expansion Education Sector

10 years 24,078 70,000

12 years 12,039 33,500

13 years + 8,026 34,000

Source: ILO (1982) p21

Starting salaries in the public service by education level are

shown in index form in table 5 for seven countries around 1979. Apart

Table 5

Starting Salaries in Public Service by Education Level

('0' level equivalent = 100)

Country '0' level 'A' level Degree

Ghana 100 120 169

Kenya 100 127 269

Liberia 100 - 339

Sierra Leone 100 229 342

Somalia 100 128 171

Tanzania 100 148 323

Zambia 100 116 169

Source: Derived from ILO (1982) pp15-17.

-8-

from Sierra Leone, the two years or so spent in senior secondary

education appear to result in only small increments in earnings. The

addition of higher education, however, has a substantial effect in four

of the countries though in three - Ghana, Somalia and Zambia - the

overall differentials are relatively narrow.

More interesting than these snapshots at one point in time are the

trends in earnings differentials. These have recently been compiled in

World Bank studies of the wage structures in Zambia, Senegal, Nigeria and

the Sudan. The results are described in tables 6, 7, 8 and 9.

Table 6

Wage Structures in the Government, Parastatal and Private Sectors

1970-83, Zambia

(Unskilled Wage = 100)

1975 1983

Government

Entry point univ. grad. 551 272

Lowest paid salaried employee 100 100

Parastatal

Skilled 355 273

Clerical 184 139

Unskilled 100 100

Private

Skilled 157 123

Unskilled 100 100

Source: From Meesook and Suebsaeng (1985) p31.

The evidence for Zambia shown in table 6 implies shows substantial

reductions since 1975 in earnings differentials across education levels

and occupations in both the public and private sectors.

-9-

Table 7

Mean Earnings of Civil Servants by Hierarchy, Senegal, 1976-84

(Hierarchy E = 100)

Hierarchy 1976 1981 1984

A 493 409 323

B 317 282 238

C 203 215 186

D 162 154 130

E 100 100 100

Source: Bloch (1985) p32.

Again, according to table 7, earnings differentials in Senegal,

particularly at higher levels, have appreciably narrowed in the last

decade

- 10 -

Table 8

a) Minimum Salaries, Nigerian Civil Service 1965-1982

(Unskilled = 100)

1965 1979 1982

Graduate entry point 600 464 262

Secondary entry point 165 172 117

Clerical assistant 130 136 105

Unskilled 100 100 100

b) Real Minimum Salaries, Nigerian Civil Service 1965-1982

(1975 = 100)

1965 1979 1982

Graduate entry point 108 56 44

Secondary entry point 83 58 55

Clerical assistant 85 59 64

Unskilled 82 60 76

c) Public:Private Sector Salary Ratios 1976-82

1976 1978 1980 1982

University graduate 92 73 68 57

Clerical 104 88 61 53

Unskilled 150 113 70 95

Source: Suebsaeng (1984) p9, 10, 12.

Several points may be concluded from table 8: differentials in

Nigeria have again significantly narrowed, real incomes have fallen

substantially particularly for university and secondary school leavers

and the greatest erosion has occurred in the public sector.

- 11 -

Table 9

a) Real Basic Salaries in Civil Service, Sudan, 1961-83

(1970 = 100)

1961 1974 1978 1983

University graduate 103 67 55 18

Secondary graduate 99 71 60 20

Unskilled - 78 71 26

b) Salaries of Civil Servants, Sudan, 1970-83

(Unskilled = 100)

1970 1978 1983

University graduate 386 298 260

Secondary graduate 215 180 162

Unskilled 100 100 100

Source: Lindauer and Meesook (1984) p4, 10.

Once again, according to table 9 differentials have narrowed and

real incomes fallen in the Sudan.

These tables all point to substantial reductions in formal sector

differentials associated with educational levels. While this experience

cannot be said to be universal across Africa - and Malawi is one

exception - it does appear to be the norm. To what extent this has

occurred through market forces or through explicit attempts by

governments to decrease earnings inequality is unclear-.

-/ In the case of Kenya, Knight and Sabot (1982) lay the stress on marketforces. "The expansion of the secondary school system in recentyears .. in Kenya reduced the premium on secondary education by roughly20 percent". p.45.

- 12 -

This erosion of differentials does not necessarily mean, however,

that within the total labor force differentials associated with education

have been comprehensively reduced since large numbers are employed

outside of this sector and the proportions differ by educational level.

Table 10 provides an illustration of the distribution of the labor force

by educational level, for Malawi.

Table 10

Occupational Breakdown by Level of Education, Malawi 1977 (percent)

No Primary Secondary University

Education

Professional Tech

Male o.2 1.3 20.6 53.9

Female 0.1 0.9 36.4 71.9

Managerial

Male - 0.1 1.9 13.4

Female - 0.6 4.4

Clerical

Male 0.2 1.9 23.9 10.6

Female - 0.3 28.9 14.4

Sales

Male 2.0 4.3 7.1 4.3

Female 0.6 1.7 6.3 2.6

Service

Male 2.0 3.7 5.8 0.5

Female 0.5 1.4 3.0 -

Agriculture

Male 84.8 70.3 18.1 9.4

Female 96.1 92.4 17.1 1.6

Production

Male 8.1 16.2 18.9 6.0

Female 1.4 2.0 1.9 0.9

Not classified

Male 2.6 2.3 3.8 1.8

Female 1.2 1.3 5.7 3.6

Source: Government of Malawi (1977) taken from Livingstone (1985).

- 13 -

Even in 1977, over 70 and 90 percent of male and female primary

school leavers respectively were working as farmers, as were almost 20

percent of secondary school leavers. Taking the first four occupational

categories and half the 'production' category as roughly constituting the

formal wage earning sector, the percentages of graduates from each

educational level employed in this sector were

No educ. Primary Secondary Higher

Male 10.4 19.3 66.9 82.7

Female 1.9 5.3 75.6 89.4

Since 1977, outputs from the educational system have greatly exceeded the

increase in formal sector jobs and the proportions of primary and, at

least junior, secondary leavers employed in these jobs will have fallen.

An analysis of the monetary returns to education in Malawi as elsewhere

in Africa, therefore, would require knowledge of returns in the urban

informal and rural sectors. Less is known about these than about returns

in the formal sector. Some of what is known is summarised in the

following two sub-sections.

Education and earnings in the urban informal sector

The emergence of the urban informal sector as an area of employment

for school leavers, particularly post primary school leavers, is

relatively recent in many countries. In a period when some formal sector

jobs continue to be filled by those with low levels of schooling,

aspirations of many are encouraged and a lengthy period of waiting or job

search may be common. Only when virtually no formal sector openings

exist and this information has had time to percolate to the schools, can

it be expected that school leavers will immediately look to the informal

sector for first employment. In the transiLional period rates of school

leaver unemployment will be initially high but are likely to fall over

time. In the last few years, tracer studies of secondary school leavers

have been made in a number of countries. Below, we look at the results

for three countries - Malawi, Swaziland and Tanzania.

- 14 -

In Malawi in 1977, primary and secondary school leavers under 24

years experienced urban unemployment rates averaging around 7.5 percent

for males and 11.5 percent for females. Recent tracer study findings

present a more pessimistic picture. The study covered 928 students who

sat for junior and middle school certificates in 1976 and traced their

situation in 1977, 1978 and 1979. Leaving aside those still described as

students and also unpaid workers, unemployment rates were 62, 53 and 50

percent for junior secondary leavers and 35, 16 and 16 percent for middle

secondary leavers over the three years. The indication is that the

junior secondary leavers are being substituted by those with more

schooling and at present they are unwilling to take up employment outside

the sector in which their predecessors have found jobs. We can surmise

that for this category of school leavers, unemployment rates may grow

before they begin to fall.

The period of the Swaziland tracer study also covers the mid

1970's. Information is available on the activities of secondary forms 3

and 5 leavers one year after graduating. An interesting aspect of this

study was the incorporation of examination performance. Table 11

provides the relevant information with Group 1 indicating highest

examination achievement and so on. Proportions in paid employment were

Table 11

Activity in First Year After Leaving School by Level of Education and

Examination Group (%)

Form 3 Form 5

Activity 1 2 3 Total 1 2 3 Total

Further education 25 19 9 15 72 56 16 35

Paid employment 41 46 51 48 21 36 52 43

Unemployed 34 35 40 37 7 8 32 22

Source: Sullivan (1981) Table 7.

- 15 -

similar for both education groups (48 and 43 percent). Proportions

unemployed, however, varied considerably (37 percent for form 3 and 22

percent for form 5). The nature of activity being followed one year

after graduation also varied by examination performance with much larger

proportions of high achievers following further education and in the case

of form 5 much higher proportions of low achievers being unemployed.

The third study is of Tanzania. The main purpose of this study,

conducted over the period 1981 to 1982 was to assess the different

educational and labor force performances of secondary students following

vocational and academic streams. (Psacharopoulos and Loxley 1985) One

year after leaving secondary school form four, only 14.4 percent of

graduates were in paid employment with an additional 7 percent searching

for employment. The rest were either in further training or schooling or

looking for places. The very small number of students in secondary

schooling in Tanzania results in favorable employment prospects or the

ability to take up further education. So far, at least, employment

pressures have not resulted in significant unemployment nor in the

acceptance of non-formal sector employment.

In these three countries, secondary school leavers do not yet

appear to be entering the informal sector in large numbers. Formal

sector employment and participation in further forms of education and

training remain the desired and expected route. The high levels of

unemployment one year after leaving school in Malawi and Swaziland,

however, illustrate the growing difficulties of following this route. As

a result and since large numbers of primary school leavers are already

employed in it, it becomes important to analyse earnings data within the

informal sector. By the nature of the sector, these data are difficult

to obtain and have to be collated by special surveys. For African

countries, these have been relatively few.

From a recent survey in Malawi, Etema (1984) presents gross monthly

earnings of business owners in the informal sector by level of education

(in Kwacha).

No Educ. Std 1-4 Std 5-8 Higher

19 29 36 100

Since no additional information on capital use or any other

income-related factor is provided the interpretive value of these figures

- 16 -

is limited. They do, however, indicate that differentials by educational

level exist, and are particularly wide for post primary graduates.

Hinchliffe's (1975) study of Nigerian workers also shows positive

differentials. In a sample of 397 self employed workers in Kano, 269 had

no schooling and 74 had completed primary school. Average annual

earnings of these two groups are described in table 12.

Table 12

Annual Average Informal Sector Earnings by Age and Schooling, Kano,

Nigeria (£)

Schooling Age Group

15-20 21-25 26-30 31-35 36+

None 96 138 145 133 130

Full primary 124 163 127 189 213

Source: Hinchliffe (1975) p3 12.

In an earlier study, Hinchliffe (1974) compared incomes in farming,

the urban informal sector and in the large textile industry. Due to the

almost total lack of schooling among farmers, comparisons were not made

by educational level but rather between those with no schooling. Hourly

income was calculated at sh.0.7-1.0 for farming, sh.1.0 for the urban

informal sector and sh.1.5 in the formal sector. These results tend to

confirm that earnings are higher in the formal sector than in the

informal sector, although once the comparison is made between similar

groups, not by as much as has often been assumed.

The results of a survey of informal sector employment in Banjal,

Gambia are also of interest. 209 owner/masters were surveyed in 1980.

While no attempt was made to correlate incomes and education, it was

found that certain occupations provided higher incomes than others and

that these were dominated by workers with most schooling. Table 13

-17-

provides this information. It is also interesting to note that in these

occupations, earnings are higher than the average in the formal sector.

Table 13

Gambia: Distribution of School Leavers Among Selected Crafts

and Craft Income, 1980

Gross Secondary Secondary

Craft Monthly All Primary Tech High

0/ ~ ~~~ / 0/0

Welding 3169 5.2 5.0 25

Carpentry 1493 6.6 5.0 - 25

Motor repair 1226 12.9 15.0 50.0 25

Blacksmith 625 3.8 40.0 -

Radio repair 363 2.9 5.0 5.6

Batik 292 7.1 5.0 5.6

Barbering 112 3.3 -

Source: ILO (1982) p44 .

Further evidence that school leavers are becoming substantially

involved in the informal sector in several African countries comes from

Zambia and Ghana. In a study of the informal sector in Lusaka, Todd and

Shaw (1979) found that 64 percent of workers had some schooling and that

the largest group was those with upper primary. During 1976-80 the

National Vocational Training Institute in Ghana reached over 500 rural

and urban artisans. Of these over 80 percent had had a middle school

education (10 years) (ILO, 1982, p46 ). Similarly Aryee's survey of the

informal sector in Kumasi in 1975 found that out of 300 owners, 58

percent had middle schooling and of the 20-25 year age group the

percentage was just under 80 (Aryee 1976). Middle school education was

found to be clearly associated with higher earnings though the effect of

primary schooling alone was much weaker.

- 18 -

On the basis of these few studies of the urban informal sector,

some conclusions can be reached. First, there now appears to be a

readiness to enter the informal sector among primary school leavers and,

in countries with relatively well developed education systems such as

Ghana, by graduates of junior secondary schools. Second, within the

informal sector earnings appear to be positively related to educational

level but most obviously for post primary schooling. Third, while there

is some evidence of earnings overlap between individuals in the formal

and informal sectors, average earnings by educational level are

significantly higher in the formal sector. Overall, both the level of

earnings and earnings differentials, at least for primary school leavers,

appear to be lower in the informal sector.

Education and output in agriculture

So far, the discussion of the monetary benefits from education has

focused on the urban sector. The majority of school leavers in African

countries, however, work in agriculture in rural areas as the

occupational breakdown for Malawi in table 10 showed. Much less is

generally known about the effect of education on productivity in this

sector and for Africa information is particularly scarce. This

sub-section begins by summarising the work of Lockheed, Jamison and Lau

(1980) which brings together education-farmer productivity studies

gathered across developing countries. This is followed by reports of

case-studies of farmer education and rural productivity and innovation

which have been prepared for African countries.

Lockheed et al. analysed 37 data sets from 18 studies in 13

countries chiefly of Asia and Latin America containing data on farmer

education and agricultural production. In 31 of the data sets, education

was found to have a positive and usually significant effect. The mean

annual gain in production for four years of education was 8.7 percent.

Following Schultz's (1975) argument that education is likely to be most

effective in the context of other 'modernising' conditions the analysis

was repeated after dividing the studies according to whether they took

place in the context of 'modern' or 'non-modern' conditions. This time,

four years of schooling was judged to result in annual increases in

output of 1.3 percent in non-modern conditions and 9.5 percent in modern

conditions. From these studies, it is clear that while primary schooling

may have little pay off in a non-supportive environment, alongside

- 19 -

complementary changes the impact of schooling on productivity can be

significant, though still below that implied by formal sector earnings

differentials.

Few of the studies on farmer education and agricultural outcome

which have been made in African countries lent themselves to the form of

analysis performed by Lockheed, Jamison and Lau, since most concentrated

solely on the adoption of innovations rather than productivity. Of the

ten studies reviewed below, three concentrated on productivity and seven

on innovation.-/

Moock (1973, 1976) studied the production of maize in Kenya's

Western Province and concluded that whereas schooling of between one and

three years had no effect on the output of male farmers, increases in

output were associated with four years and more of schooling. Each year

of schooling was found to result in a 1.7 percent increase in output per

year. For female farmers there was a positive association between output

and all levels of schooling.

Vanzetti (1972) also studied productivity in maize production, this

time in Zambia. Two communities were surveyed, one with relatively high

educational levels and one with low. In the area of high educational

levels, farmer productivity was associated with individuals' education.

In the area with low educational levels it was not. The impact of

education on output was seen to be via motivation rather than farming

knowledge and Vanzetti concluded that there is a need for a critical mass

of schooled farmers in the community so as to provide mutual

reinforcement for the adoption of new practices among them.

Hopcroft (1974) using small farm survey data for Kenya measured the

impact of various levels of schooling on the productivity of maize, tea

and cattle farming. No association was found between levels of schooling

and productivity.

The remaining studies of the impact of schooling on farming in

Africa have concentrated on the adoption of improved practices, rather

than directly on productivity.

Naylor and Ascroft (1966) surveying Kenyan farmers found that those

with some formal schooling were differentiated from the others in various

-/ The review of these studies is based substantially on Hanson's (1981)extensive discussion of the effects of schooling on the rural sector inAfrica.

- 20 -

ways, among which were: they knew more about land consolidation, erosion,

irrigation and water supplies; they had more money in the bank and sought

more loans; they employed more labor and used more tools, and they made

more use of extension services.

Heyer and Ascroft's (1970) later Kenyan study along similar lines

produced far less strong conclusions. High percentages of 'high

adopters' were also found among the unschooled.

Almy (1974) studied the adoption of innovations in rural

development in Meru, Kenya. Schooling was found to be important,

particularly via the development of cognitive skills.

Heijen (1967) again found a positive effect of schooling on

innovation in Mwanza, Tanzania. The suggestion in this study was that

schooling plays an important role through creating a receptivity to new

ideas and a motivation to implement them.

The final East African study, by Maris (1971), produced less

positive results. In a survey of Kenyan farmers in Embu District, adult

farmers with relatively high levels of schooling were not found to be

more innovative than farmers with low or no amounts of schooling.

Turning to West Africa, studies of farmer education and

agricultural innovation appear to be even fewer.

Rogers, Ascroft and Roling (1970) studied the extent to which

villagers adopted innovations promoted by the extension service in 71

Nigerian villages. They also attempted to distinguish 'early and late

adopters' and to isolate the determining factors. The conclusion was

that schooling's impact was positive. They also supported Vanzetti's

argument for Zambia that the educational level of the community is an

important factor in stimulating innovation.

More recent work on agricultural innovation has been done by

Bigelow (1978) in Ghana in four different settings. 50 farmers were

sampled in each and assessed against an innovation index. In no case was

farmers' schooling significantly related.

Of these ten studies of schooling and agricultural

productivity/innovation, in Africa, six suggest that farmers with

schooling tend to be more productive and inncvative than those without.

Where this occurs it appears to result from a mixture of greater

motivation, increased cognitive skills and more awareness, and use, of

extension agents. Hanson (1981) concludes his survey of these studies by

emphasising the point made by both Vanzetti (1972) and Rogers, Ascroft

- 21 -

and Noling (1970) that the community's level of education is particularly

important. "It would further appear that where attendance at school has

become a common pattern, the mutual reinforcement which comes from being

with other school leavers with similar motivations encourages a

willingness to innovate and to take a chance in the hope of realizing a

better future". (p7 7).

While data on the effect of schooling on productivity in African

urban informal and rural sectors remains scarce, that which exists points

to a positive relationship. In the formal sector, the relationship is

strongly documented. The final question posed in this section is whether

this relationship results directly from education or whether educational

level merely identifies those who were more productive to begin with.

Hard evidence on this issue is scarce. However, the results of recent

research in Kenya and Tanzania support the argument that educational

skills learned increase productivity (Boissiere, Knight and Sabot, 1985)..

Tests of wage and salaried employees showed that primary school leavers

with higher proficiency in literacy and numeracy earn more than secondary

leavers with lower proficiency in these areas. In addition, using an

ability measure independent of literacy and numeracy skills it was shown

that this measure alone has little effect on earnings. Commenting on

this study, Bowman and Sabot (1982) conclude "These findings support the

common sense view that literate and numerate workers are more productive

and that education is valuable to workers because it gives them skills

that increase their productivity." (pl5).

Summary and implications

This section began by suggesting that the rate of return evidence

available for Africa which shows descending rates for successively higher

levels of education may require some qualification. Many of the existing

studies are based on earnings data from the 19 60's and early 19 70's which

may not reflect recent changes in earnings structures and which were

often collected in ways which led to the urban informal and rural sectors

being ignored. There is, then, a strong need for new rate of return

studies using age-education-earnings data collected across all economic

sectors. In the meantime, the following tentative suggestions are made

about the way in which the actual pattern of returns may differ from the

- 22 -

recorded ones once allowances are made for trends in formal sector

earnings over the past decade and incomes in the informal and rural

sectors:

1) On the basis of the widely available evidence that very few primary

school leavers are now able to find formal sector employment and that

the differentials attributable to this level of schooling, while

still positive, are lower in the informal and rural sectors than in

the formal sector, it is suggested that the returns to primary

schooling are lower than the often very high ones recorded.

2) Since most secondary school leavers still appear to be able to

eventually find formal sector employment in most countries, the

earnings differentials between them and primary school leavers may be

wider and the rate of return higher than existing studies imply.

3) The narrowing of differentials betwe.en secondary and higher education

graduates which appears to have widely occurred across African

countries in the last decade or so would suggest a lowering of the

recorded returns to higher education. However, this may be offset if

allowance is also given to the overestimation of the costs of higher

education (through neglecting the costs attributable to non-teaching

outputs) commonly made in rate of return studies.

While no quantitative assessment can be made of these factors here

it is suggested that differences in the actual rates of return by level

are not as wide as the existing studies suggest and, Ferhaps more

arguably, that it should not be automatically assumed that returns to

secondary schooling are below those to primary schooling. Verification

must await much needed new rate of return studies covering all sectors of

employment.

- 23 -

Section Two

Education and Migration

In the introduction to this paper, it was argued that the

economic returns associated with education may result not only from

higher productivity of graduates within jobs but also from their greater

efficiency in negotiating the labor market. Job search has already been

referred to in general. Here, we take up one aspect of job search,

rural-urban migration. The relationships between education and migration

have produced a huge literature. Here we refer to only a small part of

it.

Analyses of migration differentials have consistently shown

educational level to be one of the main personal characteristics

influencing the propensity to migrate. Caldwell's survey of rural-urban

migration in Ghana 1962-4 is a classic of such studies and concludes

"What education does, more than anything else, is to promote long term

rural/urban migration" (Caldwell 1969, p6 2). However, the past 20 years

in Africa suggest that the educational selectivity of migration varies

over time. A three stage model has been proposed by Conroy (1976):

Stage 1 - where < 30 percent of the age group are in primary

schooling and there are few secondary schools. Demand for the graduates

of these is greater than supply. Primary leavers have a high chance of

urban employment. The steeply rising propensity to migrate begins at low

levels of education.

Stage 2 -- when 50-80 percent of the primary age group are

enrolled and secondary enrollment shows rapid expansion. Urban

unemployment develops and migration propensities for those with low

levels of education fall. Continued expansion of higher level job

opportunities and replacement of expatriates maintains the need for

senior secondary and higher education.

Stage 3 - universal primary education and further expansion of

secondary schooling. A further fall occurs in migration rates for

primary school leavers and also for secondary school leavers.

Sabot's (1979) study of migration in Tanzania between 1955 and

1971 supports at least the early stages of this model. The data in table

14 demonstrate that at each point in time propensities to migrate

- 24 -

Table 14

Rates of Rural-Urban Migration by Education 1955-71, Tanzanian Males (7/)

Arrival Period No formal Std Std Form 1

education 1-4 5-8 and above

1955-61 0.012 0.030 0.106 0.180

1962-66 0.013 0.038 0.144 0.399

1967-70 0.014 0.067 0.277 0.318

Source: Sabot (1979) p137.

increased steeply by educational level and that over time they rose

hardly at all for those with no formal education and most sharply for

upper primary and secondary school leavers.

The probability of finding an urban job within four months was

also analysed by level of education. The results are shown in table 15

and demonstrate that this probability is clearly related to level of

Table 15

Average Probability of Finding an Urban Job within Four Months by

Education and Time of Migration, Tanzanian Males (%)

Date No formal Standard Standard Form 1

education 1-4 5-8 and above

1955-60 0.12 0.22 0.78 1.0

1960-66 0.16 0.21 0.71 1.0

1966-70 0.10 0.19 0.47 1.0

Source: Sabot (1979) p12 6.

- 25 -

schooling, even though there appears to have been a dramatic

deterioration in the success of upper primary leavers over time.

When rural school leavers migrate to urban centers and find

employment, they tend to receive incomes well in excess of those earned

in their rural communities. Rempel's (1976) study of migration in Kenya

strongly demonstrates this point (table 16).

Table 16

Average Cash Income of Migrants per Month by Education, Kenya (shillings)

By Province Prior By Urban Center in First

to Emigration Year after Emigration

No Educ. 1-4 5-8 No Educ. 1-4 5-8

Nyanza 27 44 75 Kisumu 112 139 154

Western 18 23 137 Eldoret 43 82 108

Rift 50 69 116 Nakuru 98 137 106

Central 35 61 93 Nyeri 93 177 225

Eastern 25 18 97 Thika 175 117 166

Coast 20 30 67 Nairobi 148 120 197

Mombassa 137 141 133

Nangahi 105 144 238

Source: Rempel (1976)

Averaging these earnings produces the following results

No Educ. 1-4 5-8

Before migration 29 41 97

After migration 114 132 164

Differentials associated with both migration and educational level are

significant.

- 26 -

The primary motivation behind rural-urban migration is clearly

economic whether the 'push' or 'pull' factors dominate. Individuals

perceive that earnings are higher in the urban centers than in the rural

areas and that these differences increase by level of education. To the

extent that individuals are successful in moving from employment in a low

productivity sector to a higher one, the result is an increase in

economic growth. To the extent that they move to non-employment, the

result is a decrease. However, apart from instances where there are

strong push factors such as land shortage operating in the rural areas,

we would expect that after a certain period of no employment migrants

would return. Unless, then, there are substantial negative externalities

for urban or rural communities, the net impact of migration, encouraged

by education, is likely to be positive.

The overcrowding of urban centers and the additional stresses on

existing services resulting from rural-urban migration cannot be doubted

though the effect is difficult to quantify. However, it is the impact of

migration on rural communities which has aroused most controversy. In

particular, the exit of the young and educated to the towns has often

been said to be detrimental to agricultural output. In Hanson's (1981)

comprehensive search of the African literature for evidence on the impact

of schooling on rural areas he finds little empirical support for this

assertion. On the other hand, thete is considerable evidence that

remittances to rural areas by migrants are often substantial and raise

the standard of living in these areas. Most studies of remittances do

not differentiate by schooling level of the migrant but since migrants

have a higher average level of education than is general for the area

they leave and urban incomes rise with education much of the remittances

can be said to result from educational provision. Short summaries (based

on Hanson, 1981) of studies of African migrants' remittances are provided

below:

Adepoju (1974) studied migrants in Oshogbo, Nigeria. Most

reported sending money and half of these said they did so nine to twelve

times a year.

Caldwell (1969) found, in Ghana, that 40 percent of rural

families with a member in the towns received income. In all, he

estimated that a third of all rural families received money from

migrants.

- 27 -

Dubois (1970) estimated that the average amount remitted by urban

migrants in the Ivory Coast was $80, equivalent to the annual cash income

of a northern farmer.

Byerlee, Tommy and Fatoo (1977) reported remittances in Sierra

Leone to be equal to 5 percent of total urban migrant earnings.

Johnson and Whitelaw (1971) in a study of households in Kenya

estimated that 20 percent of urban wages were remitted to rural areas.

Rempel and Lobdell (1979) found in a survey of recent migrants to

Kenyan towns that 59 percent were making monthly remittances averaging 22

percent of earned income.

Moock's (1978) work in Maragoli, Kenya, went well beyond the

consideration of urban to rural remittances by viewing migrants as

behaving in one rural-urban socio-economic system in which economic life

is determined by two way flows of income and employment (see also Martin,

1982). For our immediate purposes, however, she reported that 83 percent

of migrants surveyed sent home remittances equal to 30 percent of their

earnings.

Not all observers regard remittances as leading to a higher

standard of living in rural areas in the long run, arguing that they

result in delaying those changes required for rural development (Rempel

and Lobdell, 1979). Most writers, however, have taken the opposite view

and argued that remittances lead to higher investment, community services

and/or consumption.

Concluding his survey of studies on the impact on rural areas of

migration, Hanson argues, "In short, evidence suggests that rural-urban

migration, and especially migration of educated youngsters, is paying

substantial dividends to the rural communities from which they come and

to which they often return". (p56).

- 28 -

Section Three

Education, Demographic Change and Health

In the last decade or so an increasing amount of attention has

been given to examining the relationships, if any, between education and

demographic transitions in developing countries - resulting in

particular, from changes in fertility and infant and child mortality

rates. As a sub component of both, but more particularly the latter, the

relationship between education and nutritional status has also been

researched. The results already show that strong associations exist

between education and fertility and mortality rates and between education

and the intervening factors. The precise ways in which education

influences these factors, however, still remain largely unclear.

Similarly, the economic implications of changing fertility and mortality

patterns have hardly begun to be determined but are likely to be

substantial as they alter the size and quality of the labor force in a

variety of ways.

This section is divided into three parts. These examine, in

turn, the relationships between education and fertility, infant and child

mortality and nutritional status. While the emphasis is on African data,

evidence from other parts of the world is also presented.

Education and fertility

The effect of education on fertility has been widely researched

in the past decade with the expectation that increased levels of

education reduce levels of fertility. If this is so, is the result a

benefit, and to whom? Within the family, fewer children allow more

resources to be consumed by existing children (providing there is a

downward flow of resources from parents to children). For parents, only

if fertility reductions produce the number of actual births more equal to

desired births can the result presumably be regarded as a benefit. For

society the net benefits are determined by the differences in resources

that would have been consumed by the (unborn) individual and produced by

it. These net benefits can be expected to vary considerably among

individuals and over time and place. Despite the problems associated

with defining the benefits of reduced levels of fertility, most

developing countries have some form of family planning program aimed at

reducing birth rates. On that basis, the following discussion takes as a

starting point that any effect education has on reducing fertility

constitutes a positive social return.

- 29 -

Education provides an individual with a number of attributes:

literacy and the access to information, socialisation into different

attitudes and behavior patterns, skills which provide occupational

advantages and a certification of status (Cochrane 1979). These affect

fertility by increasing knowledge, changing attitudes towards children,

altering work patterns and so on and can mostly be expected to reduce

levels of fertility.

Analyses of the results of the World Fertility Surveys conducted

throughout the developing world have recently been produced by Cochrane

and Farid (1984) focusing on the ten sub Saharan African countries

included in the Survey and by Singh, Casterline and Cleland (1985).

Levels of fertility are most immediately determined by age of marriage,

length of breastfeeding and contraceptive use which in turn are each

partially influenced by the socio-economic circumstances of individuals.

The relationships between socio-economic circumstances and the proximate

determinants of fertility and then of those determinants on fertility

levels are the subjects of these analyses.

Levels of fertility in African countries are high relative to all

other regions of the world. The reported completed fertility of

ever-married women aged 45 to 49 years ranges from 8.3 in Kenya to 5.8 in

Lesotho. Of the ten African countries surveyed, fertility trends were

downward only in Ghana. This contrasted with falling trends in Asia,

Latin America and the Middle East. The immediate causes of high

fertility are early and virtually universal marriage and very little use

of contraceptive methods. Rates would be even higher were it not for the

very high prevalence and long duration of breastfeeding. Each of these

factors is compared across regions.

- In the African sample 40 percent of women aged 15-19 years have

married compared to 32 percent in the Middle East, 26 percent in

Asia and 20 percent in Latin America.

- Contraceptive awareness and use is very low in Africa. Awareness

averaged 45 percent compared to 84 percent in Asia and 95 percent

in Latin America. Contraceptive use, ('ever used'), averaged 26

percent, 40 percent and 61 percent respectively in the three

regions. For modern or 'efficient' forms of contraception the

averages were 6, 32 and 49 percent.

- Breastfeeding is slightly more prevalent in Africa than in other

regions and with an average duration of 16-19 months, similar to

Asia but well above Latin America and the Middle East.

- 30 -

The high levels of fertility are not unwanted. For the 15-19 year age

group, desired family size in Africa averaged 7.0 compared to 3.3 in

Latin America and 3.2 in Asia. In addition, rates of infant and child

mortality are highest in Africa.

In the second stage of their analysis, Cochrane and Farid (and

also Singh et al.) report differentials in fertility according to various

socio-economic factors. Briefly, urban-rural residence has an effect on

fertility in Africa but this is lower than in other regions. The effect

of husband's occupation produced no generalisations across the ten

African countries and neither did mother's work status as measured by

wage employment. Turning to mother's education, significant differences

in fertility were apparent. These are shown below in table 17 with

equivalent rates for Asia, Latin America and the Middle East.

Table 17

Current Levels of Fertility by Wife's Level of Education

Region Years of Education

None 1-3 4-6 7+

Sub Saharan Africa 6.99 7.43 6.54 4.99

Middle East 8.90 6.91 5.58 4.79

Asia 5.81 5.89 5.66 3.70

Latin America 6.62 6.04 4.71 3.19

Source: Cochrane and Farid (1984), table 8.4.

In both the Middle East and Latin America, there is a consistent

inverse pattern between education and fertility. In Africa and Asia,

fertility increases as the result of a few years of education and

declines later. The difference in fertility between those with no

education and those with seven or more years is 2.0 in Africa, 2.11 in

Asia, 3.43 in Latin America and 4.11 in the Middle East.

- 31 -

From these findings, the question arises of why the differentials

in fertility in Africa are not more effected by education. Several

factors would lead one to expect a more substantial effect:

- differentials in age at marriage by education are larger in Africa

than in other regions,

- breastfeeding does not decline among the more educated as in Asia

or Latin America,

differentials in the use of contraception by education are greater

in Africa.

The answer hinges on the demand for children. There are differentials in

the desired family size by level of education in Africa, but at all

levels the desired size is higher than in other regions. In other words,

at all educational levels women in Africa want more children than women

elsewhere.- /

Cochrane and Farid conclude that compared to other developing

regions, "In general, fertility in SSA tends to be higher among the rural

and uneducated and fertility falls less with increases in urban residence

and education." (p8O) However, "These data indicate that family size

preferences, contraceptive use and marriage patterns do change with

education in ways which will ultimately lower fertility....." (p8 l)

A paper by Ketkar (1978) takes up the question of the inverted U

shape relationship between female education and fertility widely observed

in African countries. The data used are from Sierra Leone. From a

sample of 2000 urban women divided by level of education received,

numbers of children ever born were

No Educ. Primary Secondary College

3.78 4.49 4.17 4.06

An explanation of this in terms of women's low status in most Africansocieties is provided by Faruqee and Gulhati (1983).

- 32 -

Using multivariate analysis, fertility was seen to reach a peak at five

years of schooling and decline thereafter. Ketkar's main purpose was to

explain this threshold. The common explanation is basically a biological

one with the early years of education said to provide better knowledge of

health practices which then enable women to bear more children. Beyond

the threshold the expected negative response becomes effective. Ketkar's

explanation is different (or supplementary) and centers around the

qualitative differences in jobs performed by women with varying levels of

education. 89 percent of uneducated labor-force-participating women and

58 percent of those with some primary schooling are employed in sales

work - mainly petty trading. Very few are in the wage earning formal

sector. Conversely, 57 and 100 percent of secondary school and college

graduates respectively work in the wage earning formal sector. The

compatibility of child rearing in these two sectors varies substantially.

Ketkar concludes that only when more women are absorbed into the formal

sector will the threshold level of education be reduced.

Finally, a rather different approach to the effect of education

on fertility is provided by Caldwell (1980). Although not utilising

empirical evidence from Africa his analysis does have important

implications for the region. Caldwell argues that education effects

fertility not solely through the individual's level of education but also

through the community's level. "The primary determinant of the timing of

the onset of the fertility transition is the effect of mass education on

the family economy." (p22 5) A family morality which encourages high

fertility (within family production) to ensure an upward flow of wealth

from child to parents "cannot survive and is ultimately supplanted by a

new community morality that is eventually necessary for fully developed

non-family production and that is taught, explicitly and implicitly, by

national educational systems." (p2 26) In this argument, the greatest

impact of education on fertility is not direct but occurs through a

restructuring of family relationships and family economies. The policy

implications are important. Whereas the studies focusing on individuals'

education and fertility imply that the greatest effects occur at post

primary levels of schooling, Caldwell's focus suggests that it is the

proportion of the community receiving some schooling rather than the

average duration of schooling among those who have attended schooL which

is the chief determinant of the onset of fertility transition. Also, if

- 33 -

Caldwell's perspective is correct then given that levels of basic

schooling are lowest in Africa compared to other regions, a further

expansion there is likely to Lead to a greater effect than has been

evident in the past.

Education and mortality

Lower rates of mortality are a benefit to social welfare in

themselves. They also have economic returns in terms of extending life

expectancy and allowing greater benefits from early investment in human

capital to be produced.

This sub-section begins by summarising the present state of

knowledge regarding the relationship between parental education and

infant and child mortality, and the causal mechanisms involved across

developing countries as a whole. This is followed by the presentation of

African case studies.

A benchmark study is that by Cochrane, O'Hara and Leslie (1980)

reviewing census and survey tabulations and household data sets across

countries. On average, an additional year of mother's schooling was

found to be associated with a 9 per thousand reduction in child

mortality. Of this, 6 per 1000 was judged to be a direct result of

mother's education and 3 per 1000 a result of a more schooled woman being

married to a more educated, higher income man.

Since this review, additional data on infant and child mortality

have been collected and analysed for a large number of developing

countries through the World Fertility Survey. Using these data Hobcraft,

McDonald and Rutstein (1984b) relate under five mortality and parental

education relationships for 28 countries. (Table 18) The results show

that on average under five mortality falls from 152 to 63 per 1000 over

the range of mother's education and from 148 to 84 per 1000 over the

range of father's education. This corresponds to a fall in child

mortality of 9.9 per 1000 for each year of mother's education and 7 per

1000 for father's. Parental education remained significant after

controlling for other socio-economic variables. Similarly, the results

remain valid after allowing for the negative relationship between

parental education and fertility which influences family size and child

spacing and hence the survival chances of children. In a separate paper,

Hobcraft, McDonald and Rutstein (1984a), using data from 39 countries

- 34 -

Table 18

Under Five Mortality Rates (per 1000 births) by Parental Education

Mother's Education Husband's Education/0 1-3 4-6 7+ 0 1-3 4-6 7+ Total

AfricaSenegal 296 (178) (113) (19) 310 (261) (138) (74) 287Lesotho (224) 215 185 169 185 212 196 169 190Kenya 181 164 128 111 185 208 159 127 162Sudan 146 - 109 - 155 152 104 84 140

AmericaHaiti 226 (176) (206) (89) 230 204 (191) (150) 214Peru 237 171 98 55 241 217 150 75 170Dominican Republic 198 140 122 82 175 134 133 112 139Mexico 153 118 87 50 155 121 96 53 114Columbia 146 127 82 49 139 129 86 68 112Costa Rica 157 100 85 45 138 113 81 44 95Paraguay 110 88 67 32 (87) 88 (76) 41 77Guyana (83) (89) 80 64 (38) (89) 87 66 71Panama 134 90 52 43 100 91 60 45 65Venezuela 79 60 60 35 84 61 58 43 61Jamaica (82) (78) 74 51 (57) (69) 70 55 58Trinidad & Tobago (74) (67) 59 43 80 117 57 43 49

AsiaNepal 261 (204) (157) (136) 270 (216) (234) 167 259Bangladesh 222 198 186 (122) 230 221 191 176 215Pakistan 208 (143) (138) (112) 215 (172) 208 155 202Indonesia 193 194 143 77 199 205 165 99 180Thailand 145 105 110 (38) 151 (162) 109 83 116Jordan 112 83 84 67 112 115 105 77 100Syria 104 (75) 75 50 109 94 86 70 95Philippines 130 118 94 53 118 122 96 62 90Sri Lanka 104 97 80 53 122 98 93 57 84Korea 107 94 74 56 102 (134) 90 67 83Malaysia 67 64 56 18 88 59 60 40 61Fiji 70 67 60 46 59 66 65 55 59

Average 152 122 102 63 148 140 116 84 127

Ia These are current husbands of the mothers and may not be the fatherof the children who may have been born in a previous marriage.

Source: From Hobcraft, McDonald and Rutstein (1984b) table 3.

- 35 -

demonstrated that parental education continues to show a substantial

impact on child survival rates and that the influence of parental

education increases with the age of the child.

The consistent finding of a strong relationship between parental

education and child mortality raises questions of causation. "If the

association is not spurious, then it becomes necessary to understand the

channels through which education has its effects." (Cochrane, 1985) So

far, little is known about this, but a paper by Tekce and Shorter (1984)

found that in Jordan mother's education had significant effects on the

use of a trained birth attendant, DPT immunisation, use of professional

health care when a child is ill and nutritional status of surviving

children.

Turning from data based on the individual - parents - to evidence

at the cross country national level, Caldwell and Caldwell (1985) relate

infant and child mortality rates to percentages of the relevant age

groups in primary and secondary schooling in 1960 and GNP per capita in

1982 to show that proportions educated are more powerful indicators than

are economic measures. The most important single influence, they argue,

is girls' schooling, as comparisons of infant mortality, income and

female schooling in Burma, India, Sri Lanka and China on the one hand and

Saudi Arabia, Libya, Iran and Iraq demonstrate. Reasons for this they

argue, based largely on Indian case-studies, include the new status given

to mothers with education and the greater freedom to allow them to

demonstrate and use those features of the new culture which education

brings with it, such as modern medicine and health centers.

Caldwell and Caldwell also place emphasis on the impact of rising

education levels within the community as a whole, making those

communities more child centered and softening the attitude toward the

work conditions of young children. From Asian experience, they provide

the following, far-reaching conclusion, "Our case-studies do show us

that, in terms of achieving low mortality, economic development can be at

least partly by-passed. They do not show that the education of women can

be similarly short circuited." (pl2)I/

From these general conclusions based on experiences in developing

countries across regions, some studies relating to parental education and

child mortality in Africa are now reviewed.

I/ But see Ware, 1984, "No studies demonstrate that poor but educatedwomen with limited access to effective sanitation or medical facilitiesnevertheless achieve significant reductions in child mortality." (p197)

- 36 -

The association between mother's education and infant and child

mortality has again been well demonstrated in African case-studies: table

19 reports findings for 10 countries. According to Ware (1984, p19 6),

Table 19

Proportion of Children Dying by Age X by Mother's Education for African

Countries

Country No School Elementary Above Primary

Ethiopia .179 .137 .012

Gambia .275 .194 .118

Ghana .129 .116 .082

Kenya .160 .106 .043

Senegal1 .120 .071 -

Sierra Leone .292 .217 .140

Sudan (rural) .212 .151 _

Tanzania .261 .132 -

Uganda .181 .129 .056

Zambia2 .174 .165 .093

Source: Sullivan, Cochrane, Kalsbeek (1982) p41.

Notes: 1. for Senegal the division is illiterate:literate

2. for Zambia the divisions are no school, lower primary,

upper primary+.

studies in other parts of the third world have shown that the fall in

child mortality levels associated with the move from primary to secondary

education is twice as important as the original step to primary

schooling. Cn average, this finding also appears to be borne out by the

African studies, with mortality rates for children of elementary and~

- 37 -

post-elementary schooled mothers averaging 28 and 61 percent respectively

below those of mothers with no schooling.

Individual case-studies fill out the overall picture of table 19.

Chana 1960 Census figures show that the proportion of children dead was

almost twice as high for mothers with no education as for those with

primary schooling and four times greater than for mothers with secondary

schooling (Gaisie, 1969). Evidence that mother's education has a direct

effect and is not simply a proxy for higher household income comes from

Orubuloye and Caldwell (1975) who found that Nigerian women of a similar

economic status exhibited wide child mortality differentials by level of

education. This occurred in both a community with modern health services

and without them. Taking this line of examination further, Caldwell

(1979) analysed data on over 8000 Yoruba women in Nigeria to assess the

impact of a variety of socio-economic variables on child survival. He

concluded that the single most significant determinant was maternal

education and that the other five variables utilised did not together

come close to explaining the effect of maternal education. Similar

results were found for the Sudan by Farah and Preston (1982).

Education and nutrition

Nutritional status is a determinant of life expectancy and

influences mental and physical capacity.

The main area of research into education-nutrition relationships

is of mother's education and child nutritional status. Ideally,

education's effect through factors such as income and health knowledge

shbuld be analysed. The alternative is to measure the net effect of

education, ignoring the ways in which it has an effect, and controlling

for other factors influencing nutrition but unrelated to education.

According to the review made by Cochrane, O'Hara and Leslie (1980), no

studies of the first type have been made and only a few of the second.

Most studies have concentrated on the gross relationship between parental

education and child nutritional status, thereby not controlling for other

relevant exogenous variables. Below, we summarise this review and follow

it by presenting the results of a recent multivariate study by Behrman

and Wolfe (1984). As will be made obvious, very little evidence on the

relationship between parental education and child nutrition exists for

African countries.

- 38 -

Cochrane et al. summarise the findings of 17 studies relating

parental education and child nutritional status. Of these five are from

Africa including two from South Africa and two from Malawi. Most of the

studies show a positive relationship. Of the twelve studies using

mother's education as the measure of education, seven report a positive

and significant relationship while five do not. All four studies using

female literacy as the measure have positive and significant findings as

does the one using education of the household head._/

Given the nature of much of the data, the magnitude of the

effects are difficult to summarise but for five studies which divided

children into malnourished and well nourished the average years of

mothers' schooling were 2.7 and 3.6 years respectively, and for fathers',

3.3 and 4.7 years respectively. Similarly averaging the results of four

studies using the measure of mothers' literacy, 61 percent of mothers of

malnourished children were illiterate compared to 49 percent of mothers

of well nourished children.

As Cochrane et al. argue, these studies underestimate the

magnitude of the effect of education on child nutrition if the latter is

affected by education via any variables used as controls in the matched

samples. The effect is overestimated to the extent that other variables

are excluded which affect nutrition and which are correlated with, but

are not consequences of, education.

A recent multivariate study by Behrman and Wolfe, while not using

African data, takes us a little further in this respect and also

considers the importance of the various ways in which mother's education

might affect child nutritional status. The study is of Nicaragua and

concentrates on the relative importance of income, household size and

mother's education on nutrient demand. The results show that the

elasticities of nutrient demand with respect to mother's schooling are

larger than are the income and household size elasticities.

The channels through which mother's schooling may have an impact

on child nutrition are generally said to be:

through increasing productivity in market activities

through increasing productivity in household activities

Another recent study using mother's literacy as the educationalmeasure is from Sen and Sangupta (1983) for two villages in India. Theresults are differentiated by boys and girls. The positive associationfor boys is much stronger.

- 39 -

through changing their preferences

through increasing their power and influence in household decision

making and thus changing household preferences.

Statistical tests made by Behrman and Wolfe lead them to conclude that

the third and fourth of these are most important. That is, women with

more schooling have a greater direct concern with nutrition and/or are

better able to implement their concerns.

Documenting the effects of parental education on fertility, child

mortality and nutrition is only the first step in determining the

economic significance of education in these areas. Fewer children allow

parents to 'invest' more in them - more food, more toys, more schooling

and so on. In turn this can be expected to increase the cognitive and

non-cognitive outcomes resulting from educational expenditures and,

later, to increase the economic returns to these expenditures. Hard

evidence for these linkages is scarce. While Moock and Leslie (1982)

have demonstrated the importance of a child's height-for-age in

determining school performance in Nepal and Jamison (1981) has reported

similar findings in China, no equivalent studies have been located for

African countries. Heyneman and Jamison (1980), however, have found a

negative relationship in Uganda between indicators of sickness and school

achievement.

Summary and implications

Several studies, including those based on the recent World

Fertility Survey, have demonstrated a negative relationship between a

woman's education and the number of children born to her. Variables

influencing levels of fertility such as age of marriage, contraceptive

use and duration of breastfeeding have similarly each been shown to be

significantly related to levels of education. The precise impact of

education on fertility, however, and the amount required before there is

any impact at all appears to vary between developing regions. At

present, the impact is least in sub-Saharan Africa and while a completed

primary schooling is associated with decreased fertility, those women

with just a few years of schooling appear to give birth slightly more

often than those with no schooling at all. The major impact of schooling

on an African woman's fertility appears to result from post primary

schooling.

- 40 -

A rather different approach to the effects of education on

fertility is provided by those who argue that these result from changes

in family structures and community mores. These changes are then seen to

be linked to the level of education in the community as a whole. This

line of argument has the major implication that a very wide spread of

basic education is necessary for any demographic transition to occur.

For African countries this in turn implies that an expansion of the

relatively low enrolment ratios will result in a strengthening of the

schooling - fertility relationship itself to levels comparable with those

in other regions of the world.

The amount of parents' education, and particularly the mothers',

is also closely related to levels of infant and child mortality across

developing countries. This is particularly the case in sub-Saharan

Africa. The relationship appears to hold strong even after controlling

for other socio-economic variables and the influences of family size and

child spacing. With regard to the impact of different levels of

schooling, the evidence for Africa corresponds with that reported

elsewhere - the fall in child mortality associated with the move from

primary to secondary education is twice as large as that associated with

the original step to primary schooling. The latter, however, is also

substantial and it has also been argued that it is the level of basic

education in the community as a whole which deter-mines whether or not new

health measures are sanctioned.

Turning to the relationship between parental education and child

nutritional status there is less evidence for African countries.

Evidence from other regions suggests the relationship is strongly

positive.

- 41 -

Section Four

Some Other Returns to Education

Previous sections have dealt with the monetary and non-monetary

returns to education in Africa directly through labor market experiences

and indirectly through migration behavior and the impact of parental

education on child welfare. In most areas, available documentation has

been sufficient to indicate the existence of strong relationships even

though the mechanisms by which education has its effects are much less

well understood. The returns from education to individuals and to

society, however, are not limited to the labor market and better health.

In this final section, the possibilities of returns in some other areas

are discussed. The treatment is aimed solely at pointing to these areas

in very general terms and is often cursory. Only a limited amount of

empirical data is provided in support.

Income distribution

A more equal distribution of income with rewards based on

productivity is commonly regarded as a desirable goal for developing

countries.

The relationship between changes in the distribution of income and

educational expansion is complex, and one which has not been

systematically studied in African countries. Results of studies in other

parts of the world are also not definitive. At a theoretical level the

anticipated relationship depends on the underlying theory of labor

markets. To neo-classical economists, educational expansion raises each

individual's productivity and income and also reduces the scarcity

premium paid to educated labor. In these ways, educational expansion is

expected to increase the equality of income distribution. Accoreing to

the more structuralist school of economists, income distribution is

determined by occupational distribution and productivity is associated

with the jobs themselves rather than with personal characteristics of

those in the ,obs. To this group changes in educational distribution,

therefore, may determine who gets jobs but will do little to alter the

distribution of income. Both Carnoy (1979) summarising work done in

Latin America and Richards and Leonor's (1980) analysis of Sri Lanka and

the Philippines, take this second view.

Data produced earlier on the trend in formal sector earnings in

Africa indicate that within this sector income distribution has become

more equal. To what extent this is a result of educational expansion as

opposed to government income policy is not clear. Whether incomes have

- 42 -

become more equal within the whole labor force is also not clear. Very

few detailed studies of income distribution both within the urban

informal sector and the rural sector appear to exist and certainly none

has been made through time. Even when convincing correlational evidence

on the relationship between changes in the distribution of education and

in the distribution of income is eventually collected, it is likely to

differ between countries. "Differences in taxation, income policy,

professional association and union bargaining power, the degree to which

professionals are incorporated into an international job market,

different historical traditions surrounding the appropriate 'price' for

different levels of qualification - are just some of the country specific

factors which are likely to affect the nature of the relationship."

(Lewin, Little & Colclough 1982, p42)

Social mobility

The equalizing of opportunities for individuals to compete in the

labor market thereby increasing the chances of the more able to assume

positions of greatest responsibility would result in social benefits. In

this, the education system may play a part. In Africa, perhaps more than

elsewhere, the route to stable jobs and high incomes has been through

education, though recently this route has become increasingly blocked as

positions have been filled and rates of economic structural change have

declined. This situation has been succinctly described by Foster (1977).

'The upper and middle levels of the occupational structure are peaked and

constricted and grow relatively slowly in relation to the massive base.

Moreover, such opportunities as there are tend to be dominated by the

public sector which usually stresses formal educational qualifications

for access. This implies that aggregate rates of mobility will remain

relatively low... .At the same time there will be a significant. number of

instances of long range individual mobility: in other words, access to

elite roles through education may be relatively open while aggregate

rates of mobility remain low". (p227)

In Africa, education has not been offered to or taken up by, all

sexual, regional, ethnic and social class groups equally. Disparity in

the distribution of educational opportunity in turn accentuates the

economic inequalities. As primary schooling tends towards universality

and formal sector jobs become restricted to those with full secondary and

higher education, selectivity is mainly an issue for secondary and higher

education. While in the past, spatial and sexual differentials were

- 43 -

perhaps most important there is growing evidence that as African

societies become more economically stratified, differentials in access to

education by social class background are becoming the most dominant.

Below we present the results of some studies of differential access

according to parental occupation and education.

According to Weiss (1979) children of professional and clerical

workers in Ghana in the 1960's had 60 times and 11 times more chances of

entering secondary schools than children of unskilled workers and farmers

respectively. The question then posed is whether the expansion of

schooling has reduced this inequality. Comparing enrollments in 1961 and

1974, the numbers of children in secondary form 5 with professional

parents increased from 3.2 percent to 13.6 percent, thus increasing their

overrepresentation. Simiiar evidence for higher education is mixed.

Beckett and O'Connell (1975) report that at Ahmadu Bello University in

Nigeria in 1971, 40 percent of students had fathers whose income was

below the national average. Fields (1975), however, reported a much

different pattern in Kenya. There, 35 percent of university students'

fathers were in middle and high level occupations which constituted only

3 percent of the total male labor force. For Ghana, Weiss (1981)

compared the background of university students in 1951, 1963 and 1974.

Whereas in 1963, 22 percent ok students had professional parents, in 1974

the percentage was 40.

Turning to the parental education of students in secondary and

higher education, a similar case can be made that the educational system

is reinforcing the existing social stratification. Cobbe (1983) provides

selectivity indices from six studies of secondary and higher education in

Africa (table 20).

- 44 -

Table 20

Comparison of Selectivity Indices for Some African Student Populations,

Based on Fathers' Education (percent)

Educational University ABU Ghana Kenya Kenya

Level of Lesotho Nigeria Secondary Sec.4 Sec.4

1976 1971 1961 1961 1968

No education 0.20 0.41 0.4 0.4 0.5

Junior certificate 9.65 2.5

Matriculation 13.85 15.00 8.4 7.1

University 41.50 17.7 6.0 10.0

Source: Cobbe (1983) p235.

Notes: A.B.U. is Ahmadu Bello University.

These indices show, for instance, that in the University of Lesotho,

students whose fathers possessed a degree are present at the university

in 41 times the numbers that they are in the population as a whole,

whereas children of fathers who had no education are present in only

one-fifth of the numbers they are in the population as a whole.

In view of the data presented above, we must remain sceptical

concerning the cufrent impact of the educational system in promoting

aggregate social mobility in Africa. It still remains true, however,

that education is virtually the only mobility mechanism.

Savings and consumption

In a review of nonmarket benefits of schooling directed mainly at

the developed countries, Haveman and Wolfe (1984) suggest there is some

evidence to show that the more schooled save more and are more efficient

in making consumer choices. If capital market distortions cause

aggregate savings to be less than optimal, additional education-induced

savings can be regarded as a social benefit. Unfortunately, we know of

no surveys relating savings to educational level in African countries.

- 45 -

Michael (1982) has hypothesised that schooling leads to greater

efficiency in consumer markets, and home and leisure activities. The

result is an increase in real incomes reflected by differences in

consumption patterns similar to those resulting from an actual increase

in income. Empirical evidence that differences in consumption patterns

between education groups occur due to education itself and not through

additional income resulting from schooling is scarce, even for developed

countries. Again, no such evidence has been located for African

countries.

Localization

The desire to localize positions in the public service, in

particular, and the economy in general was one of the main motivations

for the expansion of higher levels of education in Africa in the 1960's

and 1970's. Jolly and Colclough (1972) showed that for eleven African

countries in the mid and late 1960's expatriates held 63 percent of posts

requiring a university degree or equivalent, 38 percent of posts

requiring senior secondary education and 37 percent of posts requiring

secondary education to an '0' level equivalent. In several African

countries including Liberia and Malawi, expatriates are still employed in

large numbers and in Nigeria a survey in 1977 showed that a fifth of

administrative and managerial workers were expatriate. Cenerally,

however, where expatriate employment remains significant it is now much

more concentrated in professional occupations such as engineers, doctors

and architects. Unlike the 1960's, government policy making positions

are much more firmly in the hands of citizens.

The impact of localization has several features. Undoubtedly, in

some areas it has caused reduced efficiency, at least in the short term.

On the other hand it has probably added to governments' ability to apply

downward pressure on salaries. More important, however, are the

intangible benefits resulting from the increased political and economic

power being exerted by citizens in their own country.

Women's status

Strong associations between mother's education and fertility,

child mortality and child nutritional status have previously been shown

to exist in African countries. Several of the interpretations have

emphasised that these result not so much from specific knowledge which is

learned in school but from an increased awareness of the existence of

- 46 -

'modern' institutions and practices and, importantly, a greater power for

women within the household to use them. Another aspect of power

relationships within the household is women's economic role. This may

also be affected by education. It may alter the quality and efficiency

of non-marketed work in the home and on the 'farm', induce additional

marketable production from within the home, produce changes in activities

carried out in the informal sector outside the home and affect

participation in the wage-earning formal sector.

The extent to which education does raise women's status and

increase their ability to make decisions in the home relative to

uneducated women is the subject of much debate. Similarly, there is

debate over whether increased economic activity is a net benefit to women

or whether it is simply an additional burden not compensated for by any

reduction in existing activities.

These issues are exceedingly complex and cannot be covered at all

adequately in this review. Below we simply point to some aspects of the

literature which refer to Africa. Caldwell's (1979) argument that

education raises the power and confidence of women to take household

decisions has already been noted in the previous section. Commenting on

this, Ware (1984) argues that the power of women in traditional societies

has often been underestimated and in educated households, overestimated.

Citing Peil (1975) and Pellow (1978) she argues that for Africa, the

literature suggests that educated women who marry men whose status and

income are superior to their own actually lose relative power.

The 'impact of education on the work of women including the

question of whether any additional earnings-related activities are

compensated for by reductions in their other activities has not been well

documented in African countries. While aggregate occupational breakdowns

for women by level of education exist, these do not indicate their total

work and its intensity. On the other hand, the few time allocation

studies which have been made have generally not included an educational

dimension.

The double role of women in developing countries in the household

and in economic production is emphasised in a collection of papers edited

by Buvinic, Lycette and McGreevey (1983). Again, the studies

differentiate little by level of education. One of the major points made

in the summary, however, is the need for greater female schooling and the

provision of opportunities for women to take advantage of the increased

productivity education provides. "Less education ..... entails costs to

- 47 -

society, not only because of the loss of women's potential for higher

productivity in market work, but also because women as mothers make the

first investments in the nutrition, health, and education of children -

investments that are critical to future economic growth ..... Some will

become heads of household ..... Others will be widowed or divorced or never

marry. Their poverty is only the most obvious symptom of how development

can bypass women" (Birdsall and McGreevey, 1983, pl3).

Personal development

Schooling entails private costs, directly through the purchase of

materials, clothing, fees and, often, contributions for the erection of

school buildings and indirectly through the reduction in children's time

available for labor inside or outside the household. On the other hand,

opportunities for subsequent benefits resulting from formal sector

employment have become very much reduced and almost non-existent for

those with low levels of schooling in several African countries. Why,

then, do parents continue to demand schooling for their children? What

are the perceived returns? Again, there has been little systematic

research into these questions in Africa.

Commenting on girls' education in Lesotho and Botswana where two

thirds of school leavers will not find employment, Oxenham (1984) states,

"The fact of the matter is that the mothers of Botswana and Lesotho value

schooling as a good in itself, whether or not it leads to a salaried job.

Reading, writing, counting, knowing something of their country and the

world beyond, having a little idea of science, all this is valued simply

as being good for the child and sufficient justification in itself."

Similar conclusions were drawn from a survey of children and

parents in Ghana. Although middle school (10 years) no longer provides

entry to salaried employment (and is terminal) enrollments continue to

grow. After interviews in six Ghanaian towns, Boakye and Oxenham (1982)

concluded that most parents and children value the skills of literacy and

numeracy, the access to a wider world of knowledge and experience and the

socially disciplining effect of the school.

Finally, the most detailed study of parents' perceived returns to

schooling comes from a survey carried out in south India by Caldwell,

Reddy and Caldwell (1985). From interviews with 364 sets of parents two

major points emerged according to the authors. First, education is

overwhelmingly regarded as a good thing for the individual, the family,

the community and the country. Second, to a question asking how

- 48 -

education changes people, most emphasis was given to changes in

personality and behavior. These were seen to result in individuals

having greater chances of deciding their own destiny. The increased

ability to handle bureaucrats and bureaucracies and the reduced

likelihood of being cheated in commercial transactions was particularly

stressed. To a question asking for the primary reason for sending

children to school, 65 percent of boys' parents and 29 percent of girls'

parents answered "to equip the child for a job" while 23 percent and 48

percent respectively answered "for the sake of knowledge and to achieve

literacy". The overall returns to education are obviously perceived in a

wider context than simply the acquisition of a job.

The transmission of ideology

Finally, the point has to be made that educational institutions

are not the purveyors of value free information. They do not simply

impart skills of numeracy and literacy. In addition, through both the

overt curricula and the 'hidden' they convey sets of values. Generally,

in Africa, these values have reflected those of the West in general and

of the western middle class in particular. Since independence, attempts

have been made by governments to promote a greater sense of nationalism

and, in some cases, some form of socialism. To governments, the returns

from achieving these aims may be just as important as those achieved

through education's direct effect on labor productivity.

This final section focusing briefly on various non-monetary

effects of education is admittedly partial in both the number of effects

mentioned and in the amount of detail given to each. Readers will

undoubtedly be able to add to the list of effects and also fill out

conceptually and empirically the treatment of some of those listed. In

some cases, however, there is a genuine lack of conceptualisation and

empirical findings on the effects of education in Africa. This paper has

not intended to add to the total amount of knowledge on this subject but

rather to bring together at least some of the existing knowledge and, in

places, to draw out its implications.

- 49 -

REFERENCES

A. Adepoju (1974) Rural-Urban Socio-Economic Links: The Example ofMigrants in South West Nigeria, in S. Amin (ed) Modern Migrations inWestern Africa, Oxford Univ. Press, London.

S. Almy (1974) Rural Development in Meru, Kenya: Economic and SocialFactors in Accelerating Change. Ph.D. dissertation, StanfordUniversity.

C. Aryee (1976) Effects of Formal Education and Training on theIntensity of Employment in the Informal Sector: A Case Study ofKumasi, Ghana. International Labour Office, Geneva.

A. Basu (1969) The Relationship of Farmers' Characteristics to theAdoption of Recommended Farm Practices in Four Villages of theWestern State of Nigeria. Bulletin of Rural Economics and Sociology,4, 1.

J. Behrman and B. Wolfe (1984) More Evidence on Nutrition Demand.Journal of Development Economics, 14.

R. Bigelow (1978) Impact of Selected Agricultural Schemes on theInnovation Patterns Among Traditional Farmers in the Southern Savannaof Ghana. Ph.D. dissertation, Michigan State University.

N. Birdsall and W. McGreevey (1983) Women, Poverty and Development inBuvinic (et al.).

P. Bloch (1985) Wage Policy, Wage Structure and Employment in the PublicSector of Senegal. World Bank, Washington D.C. (Mimeo)

J. Boakye and J. Oxenham (1982) Qualifications and the Quality ofEducation in Ghanaian Rural Middle Schools. IDS Education ResearchReports (Sussex IDS).

M. Boissierre, J. Knight, R. Sabot (1985) Earnings, Schooling, Abilityand Cognitive Skills. American Economic Review, 75,5

M. Bowman and R. Sabot (1982) Human Resources in Africa: A Continent inRapid Change. Background Paper for the 1982 Conference of AfricanGovernmental Experts on Technical Cooperation among African Countrieson Human Resource Development and Utilisation. Libreville, Gabon,2-12 August.

M. Buvinic, M. Lycette, W. McGreevey (eds) (1983) Women and Poverty inthe Third World. John Hopkins University Press, Baltimore.

D. Byerlee, J. Tommy, H. Fatoo (1976) Rural-Urban Migration in SierraLeone: Determinants and Policy Implications. African Rural EconomyWorking Paper 13. Michigan State University, Department ofAgricultural Economics.

J. Caldwell (1969) African Rural-Urban Migration: The Movement toGhana's Towns. Columbia Univ. Press, New York.

J. Caldwell (1979) Education as a Factor in Mortality Decline. AnExamination of Nigerian Data. Population Studies 33, 3.

- 50 -

J. Caldwell (1980) Mass Education as a Determinant of the Timing ofFertility Decline. Population and Development Review 6, 2.

J. Caldwell and P. Caldwell (1985) Education and Literacy as Factors inHealth. Paper prepared for the Rockefeller Foundation Conference onGood Health at Low Cost. Bellagio, April 29-May 3.

J. Caldwell, P. Reddy, P. Caldwell (1985) Educational Transition inRural South India. Population and Development Review, 11, 1.

M. Carnoy (1979) Can Educational Policy Equalise Income Distribution?Saxon House, Farnborough.

J. Cobbe (1983) The Educational System, Wage and Salary Structures, andIncome Distribution: Lesotho as a Case Study, Circa 1975. Journal ofDeveloping Areas. 17, January.

S. Cochrane (1979) Fertility and Education. World Bank Staff OccasionalPaper 26, John Hopkins University Press, Baltimore.

S. Cochrane (1985) The Effect of Education on Mortality: A Quick Reviewof the Evidence. Paper prepared for the Rockefeller FoundationConference on Good Health at Low Cost, Bellagio, April 29-May 3.

S. Cochrane, D. O'Hara, J. Leslie (1980) The Effects of Education onHealth. World Bank Staff Working Paper No. 405, Washington D.C.

S. Cochrane and S. Farid (1984) Fertility in Sub-Saharan Africa: Levelsand Their Explanation. World Bank, Washington D.C. (Mimeo)

J. Conroy (1976) Education, Employment and Migration in Papua NewGuinea. Australian National University, Development Studies Centre,Monography, 3.

V. Dubois (1970) Population Review, 1970: Ivory Coast. AmericanUniversity Field Staff Reports. 13, 1.

W. Etema (1984) Small-Scale Industry in Malawi: Journal of ModernAfrican Studies. 22, 3.

A. Fareh and S. Preston (1982) Child Mortality Differentials in Sudan.Demographic Transition in Metropolitan Sudan. Department ofDemography, Australian National University, Canberra.

R. Faruqee and R. Galhati (1983) Rapid Population Growth in Sub-SaharanAfrica. World Bank Staff Working Paper No. 559. Washington D.C.

C. Fields (1975) Higher Education and Income Distribution in a LessDeveloped Country. Oxford Economic Papers. Vol.27.

P. Foster (1977) Education and Social Differentiation in Less DevelopedCountries. Comparative Education Review. Vol. 21.

S. Gaisie (1969) Dynamics of Population Growth in Ghana. University ofGhana, Legon.

Government of Malawi (1977) Population Census 1977. Vol.II

- 51 -

Governmenc oE Malawi (1983) Urban Household Expenditure Survey 1979/80.National Statistical Office.

J. Hanson (1981) Is the School the Enemy of the Farm? African RuralEconomy Paper 22, Department of Agricultural Economics, MichiganState University.

R. Haveman and B. Wolfe (1984) Schooling and Economic Well-Being: TheRole of Non-Market Effects. Journal of Human Resources. XIX, 3.

J. Heijnen (1967) Development and Education in the Mwanza District(Tanzania). A Case Study of Migration and Peasant Farming.Rotterdam, Bronder-Offset.

J. Heyer and J. Ascroft (1970). The Adoption of Modern Practices onFarms in Kenya. Da-es-Salaam: Universities of East Africa SocialScience Conference Papers.

S. Heyneman and D. Jamison (1980) Student Learning in Uganda: TextbookAvailability and Other Factors. Comparative Education Review. 24, 2.

K. Hinchliffe (1974) Labour Aristocracy - A Northern Nigerian CaseStudy. Journal of Modern African Studies. 12, 1.

K. Hinchliffe (1975) Screening, De-Schooling and Developing Countries.Higher Education. Vol.4.

J. Hobcraft, J. McDonald, S. Rutstein (1984a) Demographic Determinantsof Infant and Early Childhood Mortality: A Comparative Analysis.International Statistical Institute. (Mimeo)

J. Hobcraft, J. McDonald, S. Rutstein (1984b) Socio-economic Factors onInfant and Child Mortality: A Cross National Comparison. PopulationStudies. 38,2.

P. Hopcroft (1974) Human Resources and Technical Skills in AgriculturalDevelopment: An Economic Evaluation of Educative Investments inKenya's Small Farm Sector. Ph.D. dissertation. University ofStanford.

International Labour Office (1982) Paper Qualification Syndrome andUnemployment of School Leavers. Jobs and Skills Programme forAfrica, Addis Ababa.

D. Jamison (1981) Child Malnutrition and School Retardation in China.Population and Human Resources Division Discussion Paper No. 81-27,World Bank, Washington D.C.

G. Johnson and W. Whitelaw (1972) Urban-Rural Income Transfers in Kenya:An Estimate Remittance Function. Discussion Paper 137. Institute ofDevelopment Studies, University of Nairobi.

- 52 -

R. Jolly and C. Colclough (1972) African Manpower Plans - an Evaluation.International Labour Review. Vol.106.

S. Ketkar (1978) Female Education and Fertility: Some Evidence fromSierra Leone. Journal of Developing Areas. 13 October.

J. Knight and R. Sabot (1985) Educational Expansion, Wage Compressionand Substitution between Education Levels: A Quantitative and PolicyAnalysis. World Bank, Washington D.C. (mimeo)

K. Lewin, A. Little, C. Colclough (1982) Adjusting to the 1980's -

Taking Stock of Educational Expenditure. Paper for the BellagioEducation Group Meeting, Ottawa, 19-21 May.

D. Lindauer and 0. Meesook (1984) Public Sector Pay and EmploymentPolicy in the Sudan. World Bank, Washington D.C. (Mimeo).

I. Livingstone (1985) Youth Employment and Youth Employment Programmesin Africa: Malawi. International Labour Organisation/Jobs and SkillsProgramme for Africa.

M. Lockheed, D. Jamison, L. Lau (1980.) Farmer Education and FarmEfficiency: A Survey. Economic Development and Cultural Change.29, 1.

C. Martin (1982) Education and Consumption in Maragoli (Kenya):Households' Educational Strategies. Comparative Education. 18, 2.

D. Meesook and Suebsaeng (1985) Wage Policy and the Structure of Wagesand Employment in Zambia. World Bank, Washington D.C. (Mimeo)

R. Michael (1982) Measuring Non-Monetary Benefits of Education: ASurvey. In W. McMahon and T. Geske (eds) Financing Education.University of Illinois Press, Urbana.

J. Moock (1978/79) The Content and Maintenance of Social Ties betweenUrban Migrants and their Home-Based Support Groups: The MaragoliCase. African Urban Studies. 3 Winter.

P. Moock (1973) Managerial Ability in Small-Farm Production: An Analysisof Maise Yields in the Vihiga Division of Kenya. Ph.D. dissertation,Columbia University.

P. Moock (1976) The Efficiency of Women as Farm Managers: Kenya.American Journal of Agricultural Economics. 58, 5.

P. Moock and J. Leslie (1982) Child Nutrition and Schooling in the TeraiRegion of Nepal. Population and Human Resources Division DiscussionPaper No. 82-17, World Bank. Washington D.C.

H. Naylor and J. Ascroft (1966) A Baseline Survey of Factors AffectingAgricultural Development in Three Areas of Kenya. East AfricanInstitute of Social Research Conference Papers, No. 345.Da-es-Salaam.

J. O'Connell and P. Beckett (1975) Social Characteristics of an Elite inFormation: The Case of Nigerian University Students. British Journalof Sociology. Vol. 26.

- 53 -

1. Orubuloye and J. Caldwell (1975) The Impact of Public Health Serviceson Mortality: A Study of Mortality Differentials in a Rural Area ofNigeria. Population Studies. 29, 2.

J. Oxenham (1984) New Opportunities for Change in Primary Schooling?Comparative Education. 20, 2.

M. Peil (1975) Female Roles in West African Towns, in (ed) J. Goody,Changing Social Structure in Ghana. London, International AfricanInstitute.

D. Pel.low (1978) Women in Accra: Options for Autonomy. Algonac,Michigan.

G. Psacharopoulos (1985) Returns to Education: A Further InternationalUpdate and Implications. Journal of Human Resources, XV, Fall.

G. Psacharopoulos and W. Loxley (1985) Diversified Secondary Educationand Development: Evidence from Colombia and Tanzania. Johns Hopkins,University Press, Baltimore.

H. Rempel (1976) The Determinants of Rural-to-Urban Labor Migration inKenya. Cambridge, Massachusetts.

H. Rempel and R. Lobdell (1977) The Role of Urban-to-Rural Remittancesin Rural Development. Journal of Development Studies. 14, 1.

P. Richards and M Leonor (1981) Education and Income Distribution inAsia. Croom Helm, London.

E. Rogers, J. Ascroft, N. Roling (1970) Diffusion of Innovations inBrazil, Nigeria and India. Michigan State University.

R. Sabot (1979) Economic Development and Urban Migration. OxfordUniversity Press.

T. Schultz (1975) The Value of the Ability to Deal with Disequilibria.Journal of Economic Literature. Vol. 13.

A. Sen and S. Sangupta (1983) Malnutrition of Rural Children and the SexBias. Economic and Political Weekly, May.

S. Singh, B. Casterline, J. Cleland (1985) The Proximate Determinants ofFertility: Sub-national Variations. Population Studies, 39.

G. Sullivan (1981) Some Aspects of the School Leaver Tracer Project -Swaziland. Paper presented at the Development Studies Associationworkshop on Education and Migration. University of Liverpool, April15-16.

J. Sullivan, S. Cochrane, W. Kalsbeek (1982) Procedures for Collectingand Analysing Mortality Data in LSMS. Living Standards MeasurementStudy Working Paper 16. World Bank, Washington D.C.

P. Suebsaeng (1984) Employment and Salaries in the Nigerian CivilService: Some Preliminary Notes. World Bank, Washington D.C. (Mimeo).

- 54 -

B. Tekce and F. Shorter (1984) Determinants of Child Mortality: A Studyof Squatter Settlements in Jordan, in Child Survival: Strategies forResearch. Eds. W. Mosley and L. Chen. Population and DevelopmentReview (a supplement to Vol. 10).

D. Todd and C. Shaw (1979) Education, Employment and the Informal Sectorin Zambia. Institute for African Studies. University of Zambia 1979.(Mimeo)

N. Vanzetti (1972) Education and the Development of Farming in TwoAreas of Zambia. University of Nottingham Farm Management BulletinNo. 3.

H. Ware (1984) Effects of Maternal Education, Women's Roles, and ChildCare on Child Mortality, in Child Survival: Strategies for Researcheds H. Mosley and L. Chen. Population and Development Review (asupplement to Vol. 10).

L. Weis (1979) Education and the Reproduction of Inequality: The Case ofGhana. Comparative Education Review. 23, Feb.

L. Weis (1981) The Reproduction of Social Inequality: Closure in theGhanaian University. Journal of Developing Areas. 16, October.