|
Chapter
3 - Population Projections There
is little need to emphasize the importance of population projections for
countries attempting to plan their economic and social development. The
primary needs of the people, which development programmes aim to satisfy
cannot be gauged rationally without regard to the expected size and
composition of the population, nor can any future development plan be
realistic without due consideration given to future labour force which
in turn depends on population projections. Therefore it can be said that
population projection is a vital tool for planners and policy makers. Three
variants of population projections have been made namely the medium
variant projection, the low variant projection and the high variant
projection. The medium variant projection is the most probable scenario
for the future and therefore is more likely to occur. The low variant
and high variant projections give a lower and upper limit respectively
to a range of probable scenarios that could occur. The
cohort component method was used for projecting the population of
Mauritius. Projections were made for the Island of Mauritius and the
Island of Rodrigues separately and the sum of the two yielded
projections for the Republic of Mauritius. As the name itself implies,
the methodology first requires projections of the components of
population growth namely fertility, mortality and migration. Once the
future levels and patterns of fertility, mortality and migration are
obtained, these are used in conjunction with a base population to obtain
the projected population. The base population used was the Census 2000
figure. The
steps involved in the population projections are as follows: (a)
The age and sex specific survival ratios are applied to the base
population in order to obtain survivors at the end of the five-year
period. Age and sex-specific survival ratios are then applied to these
survivors to obtain survivors at the end of the following five-year
period and the process is repeated until the end of the projection
period. (b)
The expected number of births during a specific five-year period is
estimated by applying the projected age-specific fertility rates to the
projected female population in the reproductive age groups for that
period. The births are disaggregated by sex on the basis of an assumed
sex ratio at birth; they are then survived using the appropriate
survival probabilities. (c)
The projections that have a migration component are obtained by
adjusting the projections, based on fertility and mortality trends for
net migration. The
data required for the projections are: ·
The adjusted 2000 Census population by age and sex ·
The projected total fertility rate (TFR), the age pattern of
fertility and the sex ratio at birth ·
The projected age-specific survival ratios ·
The projected number of migrants by age and sex 3.4.1
The Total Fertility Rates (TFR) In
its 2000 revision of population projections of individual countries, the
United Nations fertility assumption for the Republic of Mauritius is as
follows: Medium
variant projection: TFR
reaches 1.90 by 2000-05 and thereafter remains at that level Low
variant projections:
TFR reaches 1.80 by 2000-05 and then stabilizes to 1.50 by
2015-20 High
variant projections:
TFR reaches 2.10 by 2000-05 and then stabilizes at 2.30 by
2010-15 The
future assumptions were based essentially on past trends in fertility
registered in the country. The
level and trends in fertility in a country is determined by a multitude
of socio-economic factors including adult literacy, school enrolment
levels, level of female employment, levels of infant mortality and
nuptiality as well as the strength of its population policies and
programmes. An
analysis of the trend in TFR in the recent past indicates that after
peaking at 2.31 in 1992, the TFR has been continuously declining to
reach a minimum of 1.96 in 1998. Subsequently, the TFR increased
slightly to 2.04 in 1999 and declined marginally to 2.01 in 2000. For
the year 2001, up to date fertility data indicate a further decline in
fertility thus bringing the TFR further below replacement level. It is
clear however that this declining trend is not stable. In these
circumstances it would be more reasonable to assume that stabilization
to 1.90 would occur a little later than assumed by the UN. It
seems more likely that the TFR will lie within the range 1.90 to 2.15
during the period 2000-05. In
the light of the above, the UN fertility assumptions have been adapted
on the basis of the most recent trends for working out our population
projections. The fertility assumptions used for the projections are as
follows:
The
TFR in the Island of Rodrigues has been declining from 5.16 in 1983 to
3.19 in 1990 and to 2.38 in 1997. Since then, there has been a slight
reversal of the situation whereby the TFR has been gradually increasing
to 2.39, 2.49 and 2.61 in 1998, 1999 and 2000 respectively. Provisional
figures for 2001 indicate a stagnation of the fertility level. The
fertility assumptions for the island of Rodrigues are based on the
observed fertility trends as well as the UN fertility assumptions for
medium fertility countries, i.e countries with declining fertility but
with TFR still above replacement level. It is therefore assumed that
fertility would stabilize to a TFR of 2.1 (replacement level fertility)
in the medium variant projections, to a TFR of 1.6 in the low variant
projections and to a TFR of 2.6 in the high variant projections. 3.4.2 Age patterns of fertility The
age patterns of fertility assumed for the projections are given in table
3.2. The patterns are those that prevailed during the period 1995-2000.
In fact, these patterns have remained unchanged over the last three
decades. Three
models are usually adopted for the childbearing pattern; these are the
early child-bearing pattern, the intermediate pattern and the late
child-bearing pattern. For the Island of Mauritius, the pattern is
roughly equivalent to the replacement level intermediate child-bearing
schedule characterised by roughly the same maximum fertility in
age-groups 20-24 and 25-29. For Rodrigues, the pattern does not match
any of the three above-mentioned models. 3.4.3 Age-specific fertility rates The
age specific fertility rates used in the projections have been derived
from the projected TFR and the assumed age pattern of fertility. The
rates are given in tables 3.3(a) and 3.3(b). The
projected births have been disaggregated by sex by applying the average
sex ratio at birth prevailing during the period 1991-2000. The sex
ratios used are 103.45 male births per 100 female births for the Island
of Mauritius, 99.40 for the Island of Rodrigues and 103.31 for the
Republic of Mauritius. Table 3.3(a) - Assumed age specific fertility rates (ASFR), 2000- 2040 , Island of Mauritius
Table 3.3(b) - Assumed age specific fertility rates (ASFR), 2000-2040, Island of Rodrigues
Mauritius
has experienced a significant improvement in mortality during the last
three decades. Table 3.4
below gives the evolution of life expectancy at birth (e0) from 1972 to
2000 and the average annual change. Table
3.4 - Life expectancy at birth, 1972-2000 - Island of Mauritius The
figures reveal that the average annual changes are erratic. This may be
due to the fact that life tables constructed for inter-censal periods
are affected by errors in population estimates by age and sex. If life
tables for the census years only are compared this would give a better
picture of the situation. Table
3.5 - Intercensal change in life expectancy at birth, 1972-2000 - Island
of Mauritius Table
3.5 gives life expectancy at birth (e0) for 1972, 1983, 1990 and 2000
censuses together with the corresponding average annual increases. It
is expected that gains in e0 would decrease with time because of the
relatively high levels already achieved. It is however observed that the
average annual gain among males is higher during period 1990-2000 than
the period 1983-90. This is attributable to a more rapid improvement in
mortality among adult males during the former period. The average annual
increase among females follows the expected trend. Three
models of mortality improvement were developed by the UN based on the
past experience of a variety of countries. These are the
"slow", "normal" and "fast" mortality
improvement scenarios. The observed annual gains in e0 have been
compared with the UN models of mortality improvement to see which model
is best suited for Mauritius.
The
figures show that improvement in mortality for females during the period
1983-2000 approximately matches the UN "low mortality
improvement" scenario. Among males however, improvement registered
during the same period does not match with the gains even under the UN
"low mortality improvement" scenario. An analysis of survival
ratios during the last two decades was made to better understand the
source of the problem. A
detailed analysis of the figures indicates a deterioration in survival
ratio among males in the age bracket 25-54 years during period 1983-90
while the deterioration shifted to age 15-24 years in 1990-2000. When
the evolution in survival ratio is taken globally between 1983 and 2000,
very little improvement is registered between ages 10 to 44 years. Among
females, although almost no deterioration in survival ratio is
registered during the period 1983-2000, improvement is rather slow in
the age bracket 20-59 years, particularly in the age bracket 45-59
years. The
pattern of higher mortality among adult males compared to females
observed in the past, still persists though to a lesser extent. According
to this procedure, the rate of improvement of age specific survival
ratios for each sex during the period 1983-2000 is first determined. The
rates thus obtained are then used to project survival ratios for each
age group in the projection period. This is done by first locating in
model life tables ( Coale and Guang, "New Regional Model Life table
at High Expectation of life") the current mortality level (survival
ratio) for a particular age and sex group. The survival ratios for the
projection periods are then estimated using the rate of improvement
already determined. In case the projected survival ratio attains a value
equivalent to level 25 in the model, the survival ratio for the
subsequent periods are kept constant. Once the whole set of survival
ratios has been determined, the life expectancy at birth for each period
is then determined. The
trend in the expectation of life at birth is then analysed to verify
whether it progresses smoothly over time. The pattern of mortality by
sex in all the projection period is also analysed to verify whether
there is a smooth transition with age and over time. In the event that
e0 and/or mortality pattern
does not progress smoothly, the survival ratios causing the problem are
localised and then smoothed. The projection of survival ratios for the
Island of Rodrigues was done in a similar fashion. The expectation of
life at birth for each set of age-specific survival ratios was then
calculated. These are given in the table 3.8 above. 3.6.1 Island of Mauritius Prior
to 1979, the embarkation and disembarkation cards of all passengers were
received at the office. They were then coded, processed and then
tabulated to yield estimates of net migration. The
large increase in passenger traffic in the seventies made it difficult
to cope with the 100% coverage, and as from 1979, only a 10% sample was
processed. This manual processing of the passenger traffic cards
continued until the end of 1993. Although this system gave acceptable
estimates of migration for several years, the continued rapid increase
in the volume of passenger traffic seems to have been accompanied by an
exacerbation of the inherent problems in reporting, sampling, coding and
data capture. In particular, the coding of sex is subject to errors
since the question is not explicitly asked, and the answer often has to
be deduced from the name. Uncertainties also arise in distinguishing
residents from non-residents because of inconsistencies between reported
country of residence and permanent address. Hence the reliability of
data on net international migration has suffered somewhat. With
the coming into operation of the computerised system at the Passport and
Immigration Office in 1994, passenger traffic data were obtained on
diskette. The data consisted of records of Mauritian residents which was
filtered from records of all arrivals and departures using the field
"country of residence" = Mauritius. In the data file for
residents, only one field required coding that is "district of
residence". Once the data are coded, tabulations of net migration
by age of migrants and by district of residence are produced. The
data however show inconsistencies in age distribution as well as
distribution by district when compared with the distribution of past
emigrants as well as with data prior to 1994. The net overall migration
also sometimes seems to be unrealistic. The
only period for which net migration data based on passenger traffic data
can be checked against an independent source, is the inter-censal period
1990-2000.
Table
3.9 gives the estimated average annual inter-censal migration for the
Island of Mauritius as well as corresponding estimates based on
passenger traffic data. It is observed that there is no consistency
between the different sets of figures, both with respect to the total as
well as distribution by sex. Passenger traffic data on residents
indicate that males are migrating in while females are leaving the
country. A different picture is obtained if all passengers are
considered, showing an inward movement of 633 males and 3,625 females
during the period 1990-2000. Assuming that non-residents coming would
eventually leave the country and that arrivals and departures of
non-residents would balance out in the long run, net migration of all
passengers should eventually converge towards net migration of
residents. This is not the case however. Inter-censal
migration is estimated using census data. Data by age and sex for two
consecutive censuses are required as well as an appropriate set of
age-sex specific survival ratios. The intercensal estimates for period
1990-2000 shown in table 3.9 appear to be more reasonable than estimates
based on passenger traffic data, as it shows net out-migration of both
males and females. This is in line with intercensal estimates 1983-1990
except that the relative importance of out-migration between the sexes
has changed; there is now relatively less out-migration among females.
The level of out-migration from one inter-censal period to the next has
also decreased considerably (average annual rate of decrease being 18%). Table
3.10 - Projected average annual net migration, 2000-2040 - Island of
Mauritius Assuming
that the fall in migration with time will continue at the same rate in
the future, the projected annual net migration is expected to be as
shown in table 3.10. Migration
data for the Island of Rodrigues is given below. It should be noted that
figures are available only for all passengers because on the manifest of
passengers travelling between the islands of Mauritius and Rodrigues, no
distinction can be made between residents and non-residents. The
passenger traffic data also show inconsistencies as for Mauritius. The
passenger traffic data for the two inter-censal periods are highly
inconsistent, the data for 1983-1990 showing more females (-150) leaving
Rodrigues than males (-100) while the 1990-2000 data showing females
leaving Rodrigues (-625) while males are returning back to Rodrigues
(+19). Table
3.11 - Average annual net migration estimated from Census and passenger
traffic data - Island of Rodrigues
Inter-censal
data however show marked consistency between each other, the 1983-90
data showing equal number of males and females leaving Rodrigues
annually (-300) and the 1990-2000 data showing the same balance between
the sexes but the tempo has lowered ( about 200 of each sex leaving
Rodrigues annually). Table
3.12 - Projected average annual net migration, 2000-2040 - Island of
Rodrigues The
projected yearly out-migration, which has been based on the inter-censal
data, is shown in table 3.12 below. The
projected number migrants for the Republic of Mauritius will be the sum
of migrants for the Island of Mauritius and the Island of Rodrigues. Table
3.13 - Projected average annual net migration, 2000-2040 - Republic of
Mauritius The
age distribution of the migrants was taken to be the same as the that
prevailing during the inter-censal 1990-2000 period. 3.7
Implications of the population projections 3.7.1 Population size and growth According
to the medium variant population projections, the population of the
Republic of Mauritius will grow from 1,186,873 in 2000 to around
1,486,000 in 2040, at an average annual rate of 0.56%. It is projected
that the population of the Island of Mauritius will reach nearly
1,434,000 by 2040 and that of Rodrigues around 52,000. The
sex ratio of the population is expected to continuously decline from
98.3 males per 100 females in 2000 to 95.9 in 2040 due to lower
mortality prevailing among females. The
projections also indicate a continuation in the process of ageing. The
proportion of the population under 15 years of age will be decreasing
from 25.7 % (305,288) in 2000 to around 18.4 % ( 273,261) within the
next 40 years. This is the direct consequence of the fall in fertility
in the projected period. The
proportion of the elderly aged 60 years and over ( i.e eligible for
basic retirement pension) is projected to increase from 9.1 % in 2000 to
around 23.5% by 2040. In absolute numbers, there will be around 350,000
persons eligible for old age pension in 2040 against some 107,500 in
2000, that is, more than three times in number. The increase in the
number of persons aged 60 years and above will mostly be felt as from
year 2011. This corresponds to high fertility cohorts born in the 1950's
onwards who will be attaining their sixties during that period. The
median age of the population is expected to increase from 29.0 years in
2000 to almost 40 in 2040 indicating a shift to an older population age
structure. The
crude birth rate is expected to fall from 16.6 live births per 1,000
population in period 2000-2005 to around 12.2 in period 2035-40. The
crude death rate will however increase from 6.9 to almost 11.1 during
the same period in spite of improving mortality. This is due to an
ageing population whereby there will be higher proportion of population
in the older ages giving rise to more deaths. 3.8
Comparing post-1990 Census population projections for year 2000
with 2000 Census figures The
table below compares projected population figures for year 2000 based on
post 1990-census projections (base year 1993) with 2000 census figures
for the Republic and its component islands. The
figures show that the projected figures for the Republic fall short of
the population actually enumerated at the 2000 census by about 0.6%,
that is, around 7000 persons were missed. Also, the sex ratio at 2000
census was lower (98.3) than projected (100.1). In
order to investigate the reasons for these discrepancies, the
assumptions used in the projections were analysed to find out to what
extent they materialised. It was found that average TFR assumed in the
projection for the period 1993-2000 was 2.20 and was higher than the
actual TFR of 2.11 prevailing during that period. Similarly, expectation
of life at birth assumed for the period 1993-2000 were well above life
expectancy actually prevailing during that period ( 67.7 against 66.9
for males and 75.0 against 74.4 for females). Thus both assumptions tend
to overstate the projected figures. In
contrast, the projections assumed higher level of migration than the
actual level thus understating the projected figures. The figures
assumed were 710 males and 1,130 females outmigrating annually during
the period 1993-2000 as compared to average annual figures of 710 and
483 respectively. The sex ratio of migrants used in the assumption was
lower than the actual sex ratio of migrants thus leading to higher sex
ratio in the projected population. It
can therefore be concluded that the discrepancies observed were mainly
due to the migration assumption while the fertility and mortality
assumptions tended to minimize them. As
regards Rodrigues, the projected population figures fall short of the
2000 census figures by about 1.1%, though in terms of numbers it was
just around 400. Discrepancies in the sex ratio were much less than
observed in the Republic. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
