Population projections and the Financing of Education Dr Wadan Narsey [Presentation to PRIDE workshop, Vila Vanuatu. July 2005]
Summary... The most critical element in the demand for education are the numbers of pupils needing/demanding education which depend on Actual number of children in relevant age groups And participations rates achieved Then can extrapolate on how the costs of education may be shared between taxpayers, parents/communities, and donors Note: need to look at recurrent costs, and capital costs Teachers, teaching materials, classrooms, ancillary facilities
Children needing education versus Effective demand for education Those needing education are the numbers in the appropriate age groups Those effectively demanding education are those whose parents can afford to send the children to school (given all the cost considerations- uniforms, bus-fares, textbooks, school fees) Also availability of school places ( official pushouts??) And provided parents are not driven by poverty to force children to work for the family s sustenance
The participation rates Depend on many factors You in the MOE usually have a good idea of actual participation rates at various age groups in your countries And how they are changing over time Some because of government policy: no of schools, classrooms, teachers, fees Some because of what the communities/private sector are doing themselves, regardless of what governments think or do Leave that issue aside: just focus on the numbers coming through
The numbers of children in various age groups Depends on demographic parameters and characteristics of population under study Some demographic characteristics already decided years ago Some are changing rapidly But with accurate projections of numbers coming through And expected participation rates Can have detailed estimates of actual budgetary requirements from governments, parents, and donors: using unit costs etc (David will be giving a detailed exercise using Tuvalu)
At the national level, from a census count Can do a macro estimate for the future Next year s population is going to be This year s base population + number of births -Number of deaths - net numbers emigrating
At the national level, from a census count Can do a macro estimate for the future eg for Fiji from 1996 census Population Births Deaths EmigrationNext year 1996 776 20 5 8 783 1997 783 20 5 7 791 1998 791 18 5 3 801 1999 801 18 5 2 812 2000 812 18 5 8 817 2001 817 18 5 7 823 2002 823 18 5 7 829 2003 829 18 5 7 835 2004 835 18 5 7 842 2005 842 18 5 7 848 2006 848 18 5 7 854
Look at the sensitivities The most important change comes from??? Then emigration is even more important than deaths And then the deaths- but at what ages? For school age populations: really the younger age cohorts are far more important So need detailed estimates at the one year cohort level
Population projections: get over your fears Basic principle is exactly the same as previous table: At every age group, need base population, add births for first age group, subtract deaths, and add net emigration figure How many children then coming through at Pre-school (ages 5? 4? 3? Or some combination?) Primary: 6 to 13 (what about the 5 year olds or 14?) Secondary (14 to 17? What about 18? Or 13?)
Great predictability From census: know accurately How many males & females at all the age groups By single years All kinds of other data: such as how many children are women having
Can already predict a lot of accurate numbers into the future From the census 1 year old group, can predict the number of 2 year olds next year 3 year olds in two years time 11 year olds in ten years time The 18 year olds in 17 years time
Eg from 1996 census data: can move across diagonally as far as you like 1996 1997 1998 1999 2000 2001 2002 0 18467 1 19673 2 19204 3 18813 4 19168 5 18920 6 17676 7 16446 8 16720 9 18288 10 18376 11 18757 12 19272 13 18725 14 18573 15 18489 16 17987 17 17265 18 15788
What would we need to get accurate numbers for the future? The number of live births every year The numbers dying at every age level: especially Infant mortality rates (less than 1 yr old) Child mortality rates (1 to 5) The number emigrating every year (children esp.) If we can get good estimates of these then you can estimate very accurately into the future
Live births: issue of fertility Can use crude birth rates ie births per 000 population But not the population that bears children It is females Different age groups have different nos of births And how many they choose to have which differs Across countries, regions, communities, religions,
Mortality Infant and child mortality rates differ greatly across countries because of vulnerability to disease, poverty etc And mortality at other age levels also differ but not so greatly
And emigration rates differ also Samoa, Tonga, Niue, Tokelau, Cooks have always had very high rates of emigration Kiribati and Tuvalu citizens have been working abroad Recently, Fiji s coups have accelerated their rate of emigration And Solomon Is also has seen an exodus of skilled and qualified people recently For prediction of education numbers, need to know emigration rates at younger age groups
All projections software work similarly eg PEOPLE software (DOS based) Key in base population by age groups Births: enter parameter for fertility Total Fertility Rate: (per 000 women) Or Age Specific Fertility Rate (ASFR) (per 000 women at particular age groups)
Enter parameter for mortality (separately for males and females) Most simple: Life Expectancy (eg 70 years) And the software uses some standard pattern of mortality from elsewhere in the world eg SE Asia, Europe, Africa etc But really our own countries mortality rates at every age group would be the most useful Usually obtained from Census data, Ministry of Health data
Enter parameters for emigration Your Immigration statistics: at age levels Note difficulty: people often don t want to say they are emigrating Often better to use net departure figures i.e. (total departures total arrivals) of citizens Note: illegal migrants: tourists staying on
Can show you application on my machine There probably are more recent software suitable for WINDOWS applications For later?
Now show an application: case of Fiji Last census: 1996 Next census: 2006 When will census reports come out? 2007? 2008? Some HIES data: but not the details required So still need detailed projections And the Fiji results are unexpected and critical for the financing of education
Overall impressions of PIC population changes? How are total populations changing? Rising rapidly? PNG, Solomon Is, Vanuatu, Kiribati Rising slowly? Samoa, Tonga Depleting? Niue, Tokalau, So what is happening to Fiji?
Total population of Fiji Over next ten years? 2005 2010 2015 Total Population 785 816 838 Change (000) 31 22
Seems like Fiji should plan for normal growth of its school numbers? If Solomons, or Vanuatu: answer may be a qualified yes (see below why qualified ). But as emphasised before: critical to look at disaggregations of key variables And in Fiji: many disaggregations possible with fundamental different impacts on the requirements for financing of education
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 (000) Ethnic components of population (graphs) Projected populations 900 800 700 600 500 400 Total Fijians Indo-Fijians 300 200
Well known to demographers That indigenous Fijian population has high fertility so relatively more children being born Indo-Fijians have low fertility rates: so fewer children And Indo-Fijians have much higher emigration So both emigration and fertility have major impacts on numbers of school age populations
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 17000 So look first of all at those requiring tertiary education: proxy: 18 year olds? 18 year olds (Total and ethnic breakdown) 16000 15000 14000 13000 12000 Total 11000 10000 9000 8000 7000 6000 Fijians Indo- Fijians 5000 4000
The numbers coming through are not rising evenly Total numbers fell from about 17000 some 5 years ago Have been declining- and will reach lower to about 14000 by 2007 It will rise again to 16000 by 2012 Before falling even further to 13000 by 2016
Financing demands: on pure potential entrant numbers alone Short term: slightly less pressure (having already seen big decline from five years ago) Medium term (over next five years) increase of some 3000 extras throughput: must increase pressure on tertiary institutions: Long terms after that: major decline of 3000: some relief of pressure
Eg Indo-Fijians tend towards commercial and science subjects and not towards professions like nursing, or military forces, or agriculture (and vice versa for indigenous Fijians). Are there ethnic differences in demand for tertiary places? Note that Indo-Fijian through-put has plummeted from 8000 some ten years ago to around 5000 now and will decline further to 4000 in another ten years time While indigenous Fijian throughput has risen from 8000 of ten years ago to 9000 currently, but is set to rise to almost 11000 over the next 6 years before falling Do the different ethnic groups have different differing demands for tertiary training?
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 (ooo) Take now the numbers possibly requiring places in secondary education: 14 to 17 year olds Secondary age (14-17) 70 60 50 Total 40 Fijians 30 20 Indo- Fijians 10
A similar story with major resource implications Total throughput has already been declining by about 8000 places over the last ten years BUT is now set to rise by 5000 over the next six years Before falling massively by about 10,000 for the ten years after that. Calculate resource implications: unit secondary cost of around $800 per student (rise of $4 million, followed by fall of $8m) Calculate changes in teacher requirements : ratio of 25 to 1?: Rise of 200 teachers, followed by fall of 500?
But ethnic disaggregation makes planning messier Indo-Fijian numbers have already fallen by 10,000 Will stabilise over the next five years, before falling again by another 6000 in the five years after that While indigenous Fijian student potential numbers are set to rise by 6000 over the next six years before falling over the six years after that Key priority: how cope with net increase of 5000 over next five years? And are net increases good indicators of resource requirements?
Hence: in some locations, the indigenous Fijian demand cannot be met from other schools: has to be met by extra class-rooms, teachers, and funds: so net increases in students will lead to under-estimates of extra resources required Note: that rural:urban disaggregations matter in Fiji Indo-Fijians tend to be in the west of Fiji in the cane belt areas, land leases are expiring, emigration to urban areas To some extent: Indo-Fijian schools are able to take in indigenous Fijians into formerly Indo-Fijian schools: some urban Indo-Fijian schools already have a majority of Fijian students But teachers in cane-belt areas are now facing empty desks: pupil teacher ratios falling naturally
Rural:urban migration will continue in most PICs (SI, Vanuatu, PNG, Kiribati, Fiji, ) If significant numbers are migrating to urban areas Rural areas may be able to improve quality of education But net increases in population will under-estimate resource requirements If rural students are not fully taking up the places left vacant in rural areas Planning must take account of rural:urban migration as well as population increases
So far: potential increase in tertiary and secondary demand: Suggests that Fiji will face major increases in resource demands over the next six years What of potential demands from primary age groups? Well how lucky can you get?
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 (000) Primary age groups: 6 to 13 yr olds? Age 6-13 (primary age) 140 130 120 110 100 Total 90 Fijians 80 70 Indo-Fijians 60 50 40 30 20
Major decline in numbers It would seem that Fiji may face a decline of some 15,000 over the next six years The numbers may stabilise, and rise slightly But definitely significantly reduced pressure from the primary section
But again, note the ethnic differences Indo-Fijian numbers will continue their massive decline: another drop of about 15000 over the next en years; more empty schools and places available for others; with possibilities of reducing pupil:teacher ratios naturally And improving quality of education While the indigenous Fijian numbers is set to rise by about 10000 places over the next ten years Also putting pressure on pupil:teacher ratios: Cost implication for 10,000: roughly $5 millions Teachers required: 400 to 500.
Implications for multiracial schooling These demographic changes suggest greater possibilities of having multi-racial schooling Note the associated benefits of better academic outcomes for Fiji students (refer to my booklet: Resources for Basic Education: an outcomes approach) And also implications for multiculturalism helping to build better race relations
Finally: what of pre-school age groups Age 5 (ethnicity) 18000 17000 16000 15000 14000 13000 12000 11000 10000 Both groups Age 5 Indo-Fijian Age 5 Fijian 9000 8000 7000 6000 5000 4000 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Bad news! The numbers likely to be looking for pre-school places will rise dramatically by around 2000 over the next six years, before falling only slightly The Indo-Fijian number will stagnate over the next five years before falling by 1000 over the last 5 years While the indigenous Fijian numbers coming through will keep rising by some 3000 over the next ten years And if the access rate (or participation rate) rises further, then the actual extra places required will be even more than 3000
Overall: good news and bad news? Bad news is hotch potch patterns of largely increasing demands for resources in schools at all levels except primary (with possibility of long-term relief) But good news: also possibility that short-term savings at the primary level may be able to compensate the increases sought elsewhere Hence overall, MOE in Fiji may be able to cope by reallocating existing resources, rather than making major calls for new additional resources.
Fundamental need for accurate population projections Not only at the aggregate levels for the whole country But disaggregate by variables such as ethnicity, rural:urban movements And what of religious attachments? Catholics, Methodists, SDAs, Anglicans.. Hindus (Sanatan, Arya Samaj), Muslims, Gujeratis, Chinese International
Thank you Questions and comments
PEOPLE software anyone? Based on DOS