Estimation, Prediction and Policy Design
for Population and Universal Primary Education in Pakistan
Muhammad Azeem Qureshi
Pal |. Davidsen
System Dynamics Group, Department of Geography, University of Bergen, Norway.
Fosswincklesgt 6, NO-5007, Bergen, Norway
Telephone: +47 47 34 99 00
Muhammad.Qureshi@student.uib.no
Pal.Davidsen@geog.uib.no
The positive relationship between education and sustainable economic development is
now widely recognized. This study develops a dynamic mode! for population and
primary education of Pakistan. Primary education sector in Pakistan is operating in both
public and private sectors. Pakistan has a clear socio-economicdivision of rural and
urban areas that get different priorities in allocations of public funds, believed to be
instrumental for universal primary education. The public financial allocations are
available to some extent from different publicly available sources but that too on country
basis and not on area basis. On the other hand, there is no data available for private
sector investment. Based on its assumptions this study estimates area-wise public and
private sector investment in primary education in Pakistan, predicts development path
of its population and literacy. This study finds that Pakistan may not be able to achieve
universal primary education by 2015 due to alarming dropout rate and suggests
channeling her resources to reduce it.
Introduction
Pakistan, being one of the signatories of UN’s MDGs}, has committed to ensure that by
2015 children everywhere boys and girls alike will be able to complete a full course of
primary schooling. However, her current low educational status, highly fragmented
allocations, and area and gender disparities have created intractable problems to fulfill
her commitment. Such a continued low educational status islikely to hamper human
capital formation and probable sustainable national development as suggested by new
growth theories. Not only the theories, but also the development path followed by the
countries like Malaysia, Korea and Taiwan strongly presents the case that without
education, economic development is not possible.
Pakistan is placed at 135 position out of 175 countries in terms of the Education Index.
The level of public spending is an important indicator of a government’s commitment to
the cause of education. Education, observed to be the least priority in Pakistan, has
suffered not only frompersistent and severe underinvestment by the government, but
also bitty allocations have resulted into so obvious and critical area and gender based
disparities that it will be hazardous to consider a single literacy rate for the country.
The socio-economic importance of universal primary education is unquestioned. A good
educational system, that provides universal primary education, strengthens equality of
opportunities, human capital endowment, and income distribution (Tanzi 1998). H ence
provision of basic education is both an individual and public investment. However ina
country like Pakistan where around 33% people live below poverty line, government
cannot abdicate all responsibility by making education a service made available in the
market for those who could buy it. Moreover, article 37 of The Constitution of Pakistan
makes it one of the core responsibilities of the state to remove illiteracy and provide free,
compulsory and equitable access to basic education with special care of interests of
backward classes and areas.
1 Millennium Development Goals
This study indicates that limited and fragmented financial allocations, otherwise
believed to be instrumental for universal primary education, have been responsible for
current low educational status in Pakistan. Through reduced public allocations,
government presumably is encouraging private sector investmentin education, but
being commercially motivated it focuses more on urban areas not only denying
equitable access to basic education to people of backward classes and areas but also
perpetuating poor educational status that may have implications for population growth
and sustainable development of Pakistan.
Model and its Assumptions
The purpose of this model is to estimate population, primary education rate and
financial allocations by public and private sectors of Pakistan. Based on these estimates
and maintaining trend in financial allocations the mode! will predict development path
of population and primary education in Pakistan.
Themodel, which is a component of a national model under progress, gets its
inspiration from T21 of MI? and comprises of two sectors; Population with a sub-sector
called Fertility and Primary Education. The conceptual framework of this model is based
on the demographic transition theory coined by W. S. Thompson (1929) and refined by
Frank W. Notestein (1945) which suggests a definite relationship between population
change and socio-economic development. Even though the neglect of cultural factors is
one of the shortcomings of this theory, it does provide an effective generalized picture of
the population development on an aggregated level that serves the purpose of the
model. This framework defines the model. The adult literacy rate and level of economic
development are taken as proxies of socio-economicdevelopment. It is commonly
observed that these are the major factors that affect population. Adult literacy rate will
be determined endogenously. However, level of economic development is taken as
exogenous which may be a limitation of this study and will be addressed in future
research.
? Threshold 21 is an integrated simulation model of Millennium Institute, Washington D.C.
The causal assumptions underlying the population sector are the following:
1
The number of births each year is a function of the number of women of
productive age (age 15-49, based on UN classification) in the population and total
fertility rate.
Women of productive age are classified into two types: those who practice birth
control, and those who do not. Each type has its own fertility rate.
Average adult literacy rate is the only factor determining the proportion of
women of productive age practicing birth control.
The fertility rate for those females who do not use birth control is the natural
fertility rate which changes with female life expectancy, and the fertility rate for
those who practice conscious birth control is determined by the desired number
of children per woman. It is assumed that relative per capita disposable income
and average adult literacy rate inversely affect the initial average number of
children per woman to determine the desired number of children per woman.
Age specific fertility rate will remain constant based on Haupt & Kane (2004,
p16)3.
Income, health care and nutrition are most commonly cited determinants of life
expectancy. It is commonly observed that increase in income improves nutrition
intakeand also improves access to health care. Life expectancy is therefore,
assumed to be income based. The number of deaths each year is assumed as a
function of income-based normal life expectancy based on World Resources 1998-
99: p-15, and life table death rates for east developed by Coale and Demeny
(1983).
Migration is treated as exogenous.
Pakistan, like many developing countries, has a clear division of rural and urban areas.
The two sets of areas have not only different demographic characteristics, infrastructure,
° They argue that in reality age-specific rates change and fluctuate from year to year, even if only
gradually. However, year-to-year fluctuations in the TFR may reflect changes in the timing of births rather
than changes in the average number of children women bear.
income sources and per capita income etc. but also get different priorities in allocation of
publicfunds. The population is, therefore, represented by one stock each for rural and
urban population broken down into 81-one year cohorts that are further subdivided into
female and male. Such a representation will better reflect the reality and will also make
the model more transparent.
The birth rate, that increases the population of rural and urban areas, is endogenously
affected by the average literacy rate and per capitaGDP in PPP4 of the twoareas in line
with the demographic transition theory. The impact of these causalities is so estimated
that the model output fits the estimated contraceptive prevalence rate, total fertility rate
(TFR), and crude birth rate of UN World Population Prospects-The 2002 Revision (WPP-
2002), as shown in Figure 1.
contraceptive prevalence rate total fertility rate
06 8
oL 2
19811985 1989 1993 1997 2001 2005 2009 «2013 1981 1985 1989 1993 1997 2001 2005 20092013
Time (¥ ear) Time (Year)
contraceptiveprevalencerate:D:\WPP PSYB —————_— Dmnl__ total fertility rate: D\WWPP_ PSYB. —————___ mal
contraceptive prevalence rte: Base Dmnl total fertility ate: Base. = mal
crude birth rate
60
50
40
30
20
19811985 19891993 1997 2001 +2005 +=«-2009-«:2013
‘Time (Y ear)
crude bitth rate : D:\WPP_PSY B_ ——_____________ ml’ ear
crude birth rate : Base. ——___________________ nl ear
Figure 1. Contraceptive Prevalence Rate, Total Fertility Rate and Crude Birth Rate
(Data till 2001 is estimated and after that are projections. Source: WPP-2002)
4 Gross Domestic Product in Purchasing Power Parity
The model output indicates that given the literacy level as estimated by this study the
TFR in Pakistan is likely to be higher than the TFR projected by WPP-2002. On the other
hand death rate, that decreases the population of two areas, is determined by the Life
Table Death Rates Function developed for eastern countries by MI based on Coale (1983)
mapped over life expectancy. It is commonly observed that lifeexpectancy, and also as
indicated by demographic transition theory, improves with the improvement in
economicconditions that brings in improved access to better nutrition, sanitation and
health facilities. The nature of such a relationship between income and life expectancy is
estimated based on World Resources 1998-99 to fit model output to the estimated
average life expectancy and crude death rate data of WPP-2002, as depicted in Figure 2.
Even though such is simplification, however for the purpose of the model this level of
aggregation and simplification is very much required.
average life expectancy crude death rate
40 0
1981 1985 1989 1993 1997 2001 2005 ©2009 2013 1981 1985 1989 1993 1997 2001 2005 2009 +2013
Time (Y ear) Time (¥ ear)
averagelifeexpectancy :DWPP_PSYB. —————————— Year crude death rate: D:\WPP_PSYB_ ——————__. DmnlVear
average life expectancy : Base ————————_— Year crude death rate Base. —— Dn ear
Figure 2. Average Life Expectancy and Crude Death Rate
(Data till 2001 is estimated and after that are projections. Source: WPP-2002)
The results show that crude death rate, given the assumed GDP, is likely to be higher
than rate projected by WPP-2002. This will be considered in future research by making
GDP endogenous and also adding access to basic health care to the model. Another flow
that affects population is migration from rural to urban areas and from country to
abroad. This flow is governed by the historical migration trend observed in Pakistan.
The Figure 3 shows that thesimulation output of total population fits the
estimated/ projected total population of WPP-2002.
total population
1981 1985 1989 1993 1997 2001 2005 2009 2013
Time (Y¥ ear)
total population: D:\WPP_PSYB_ Person
total population : Base Person
Figure 3. Total Population
(Data till 2001 is estimated and after that are projections. Source: WPP-2002)
In view of thepurpose only thesystem structure of primary education in Pakistanis
modeled5. It assumes that education expenditure, by public or private sector, and
relative literacy rate will proportionately affect the primary gross intake rate and will
inversely affect the primary drop out rate. This causality is so estimated that the model
output may fitthe relevant data of Government of Pakistan and World Bank, which is
discussed in the analysis section.
Data Estimation
The data for this mode is taken from three different sources. TheUN’s World
Population Prospects: The 2002 Revision, The World Bank’s World Development
Indicators 2003 (WDI-2003) and various issues of Pakistan Statistical Year Book (PSY B)
of Government of Pakistan. The population data is taken fromWPP-2002 and itis
reported in WDI-2003 that rural population in Pakistan was 72% of the total population
in 1981. This helps plug-in initial rural and urban population. The net migration per 1000
habitants is taken from WPP-2002. The initial contraceptive prevalence rate is estimated
5 The model could be obtained from authors on request.
based on WDI-2003, age specific fertility distribution and initial desired number of
children is estimated from WPP-2002.
The GDP, current $ in PPP, is taken from WDI-2003 and its three sectors’ composition
like agriculture, industry and services is taken from PSYB. It is assumed that rural GDP
comprises of 100% of agricultural GDP and 20% from industry and services. While,
remaining 80% of industrial and services GDP accounts for urban GDP. The total public
spending on education as % of GDP is taken from WDI-2003. However, the available
data did not provide allocations for rural and urban areas. Based on our empirical
observation we have followed a thumb rule that 70% resources go to 30% population
(urban) and 30% resources go to 70% population (rural) following a fundamental 80:20
principle. The education sector in Pakistan was deregulated in 1979-80 and the
following year brought in investments from private sector especially for primary
education in urban areas being commercially attractive and lately for rural areas. In view
of this event and having the estimated primary education expenditure by government as
% of GDP and assuming that the model and its assumptions represent the system
structure of population and primary education, and we estimated the primary education
expenditure made by the private sector. Such estimation, otherwise necessary due to
non availability of data, was verified by fitting multi-layers of historical data series of
WDI-2003 and average regional literacy rate of PSYB.
Analysis and Conclusion
The simulation results, assuming that the current trend of GDP, literacy and spending
on education continues, show that the population of Pakistan will likely to be around
205 million people by 2015 in line with the population projected in WPP-2002, of which
15-17.5% is primary school age population. Based on the public spending on education
data of WDI-2003 and our assumption on rural-urban allocation, we estimated the
private spending on education in rural and urban areas so that the mode! output fits the
average gross enrolment rate of WDI, presented in Figure 4.
tural private spending urban private spending
oot 0.02
0.0075 - 0.015
0.005 0.01 att
0.0025 0.005 1 ial
O-rTrt | 0
1981 1985 1989 1993 1997 2001 2005 +2009 +2013
Time (¥ ear)
rural private spending : D|WEP_PSYB
rural private spending : Baye,
1981 1985 1989 1998 1997 2001 2005 2009 2013
Time ¥ ear)
urban private spending: D\WPP_PSYB»_
urban private spending Base $$$
Figure 4. Estimated Private Spending on Education in rural and urban areas as percentage of GDP
The Figure 5 presents the public spending on education as percentage of GDP indicating
gradually decreasing role of public sector. The data up to 2000 is taken from WDI-2003
and the data of Government of Pakistan shows that in 2005 they spent 2.86% of GDP on
education. The space created by government was filled in inequitably by private sector
and resultant overall improved spending on primary education attracted students to
schools helped increased gross enrollment rate.
public spending
average gross enrollment primary
0.04
0.0825
0,025 / “a
0.0175
0.01
80
65
50
35
20
1981 1985 1989 1993 1997 2001 2005 2009 2013
‘Time (Y eat)
public spending : Base
public spending DsWPP_PSYB»_
1981 1985 1989 1993 1997 2001 2005 +2009 «2013
Time ¥ ear)
average gross enrollmentprimary :D:\WPP_PSYB —————— mal
average gross enrollment primary : Base mal
Figure 5. Public Spending on Education as percentage of GDP and Average Gross Enrolment Rate
The Figure 6 depicts the negative impact of low and inequitable allocations on the
weaker segments of society.
average gross enrollment rate
average gross enrollment rate
08 a oe ei =a
19811985 198919931997 2001-2005
2009
2013
19811985 198919931997 2001-2005 -«2009 «2013
1981
Time (¥ ear) Time (Year)
average gross enrollmentrate(FEMA LE, rural]: Base Dmal average gross enrollmentratelFEMALE, urban]: Base Dmal
average grossenrollmentrate(MA.LE,rural]: Base Dmal average grossenrollmentrate{MALE urban}: Base Dmal
primary dropout rate primary dropout rate
02
= 015 —— —-
oa | + -
04 eral i “I 0.05
0
7985 1989 19931997 2001 2005 2009 2013 1961 1985 1989 1903 1997 2001 2005 20092013
Time (¥ eat) Time ¥ eat)
primary dropoutrate[FEMALE rural]: Base DmalW’ ear primary dropoutratelFEMALE urban]: Base Dim car
primary dropoutratel MALE, rural]: Base DmalW’ ear primary dropoutratel MALE urban]: Base Din’ ear
Figure 6. Gross Enrolment Rate and Dropout Ratein Rural and Urban areas.
Assuming the current trend of spending on education the gross enrolment rate of
primary education is not likely to exceed 75% rendering Pakistan unable to fulfill its
commitment of universal primary education by 2015. The alarming dropout rate pushes
Pakistan away from its target MDG. We have tested different levels public spending,
assuming that trend of private sector spending on primary education continues, and
found that Pakistan may not be able to achieve universal primary education even if she
spends five times of its current spending level. However, the increased public spending
has substantially brought down the drop out rate |lt would therefore, bemore
appropriate for Pakistan to channel some of itseducational spending to reduce its
primary drop out rate along with ensuring equity in allocations.
References
Coale, A J., and Demeny, P. 1983. Regional M ode Life Tables and Stable Population, Second
Edition, Academic Press.
Notestein, F.W. 1945. Population: The Long View, in T.W. Schultze (ed.) Food for World,
Chicago University Press, Chicago.
The Government of Pakistan, Pakistan Statistical Y ear Book, Various Issues.
The United Nations. 2002. W orld Population Prospects: The 2002 Revision
The World Bank. 2003. W orld D evelopment Indicators 2003
Thompson, W.S. 1929. Population, American Journal of Sociology, Vol. 34, pp. 959-75.
Tanzi, Vito. 1998.Corruption around the World, IMF Staff Papers, Vol. 45, No. 4,
(December): pp. 559-94.
World Resources Institute. 1994. W orld Resources 1994-95, New York: Oxford University
Press.
