Paradigm shifts in demographic components in the new millennium and possible implications: A case study of Sri Lanka

This paper, using Demographic and Health Surveys, population censuses, and results of two population projections, attempts assess future population dynamics of Sri Lanka. Since the turn of the new millennium, fertility demonstrated an increasing trend where the average number of live births per woman increased from 1.9 in 1995-2000 to 2.4 in 2012. Although there was an increase in life expectancy of both sexes, after the cessation of civil war in 2009, male life expectancy increased significantly. At the same time departures of labour migrants dropped from 300,000 in 2014 to 200,000 in 2019. As per the population projection of 2015, which captured fertility upturn, the size of the population of the country was expected to increase significantly to 24 and 25 million by 2032 and 2042 respectively. On the other hand, the population projection of 2007, which did not capture the fertility upturn, indicated that the country was reaching population stability by the early 2030s with a population even below 22 million. However, with the increase in fertility and life expectancy, and drop in labour migration, significant growth of the country’s population is now expected to be eminent. The aging process and the unexpected fertility upturn have contributed to overall dependency on these increases significantly, which would generate more pressure on the working age population and the State. It is not only that Sri Lanka at present stands at the last phase of the demographic dividend, other factors required for economic take-off such as political stability, investment, savings, and productivity, also not demonstrating an adequate improvement to harness this window of opportunity. Failing to implement appropriate policies and programmes, the rapid growth of the population and aging would also create serious socio-economic and environmental issues in

Indices related to fertility transition in Sri Lanka, particularly the TFR valued at 1.9 of the DHS 2000, aroused the general expectation among the researchers that Sri Lankan fertility could remain stable or even go below the level reported in the DHS 2000 (De Silva, 1997a;2007;Abeykoon, 1998).
Results of the DHS 2006-07 were to identify first an unexpected increase in the TFR. The hypothesis of remaining at the replacement level fertility or even declining to below that level in Sri Lanka was therefore not to be in fact realized. The expectations of the population planners were not fulfilled. The TFR of 1.9 reported in DHS 2000 was increased to 2.3 by DHS 2006-07 (Table 1).
The TFR value of 2.4 obtained from population census 2012 data has confirmed this unexpected increasing trend of fertility in Sri Lanka. The fertility increase as observed in Sri Lanka has not been observed in any of the other South Asian countries so far, where the fertility has an upturn from below to well above replacement level (Perera, 2017). Interestingly, the TFR levels of Bhutan and Bangladesh in 2012 stand even lower than the Sri Lankan level and India at 2.6, only slightly higher than Sri Lanka (ESCAP, 2012). Although since the year 2012 there was a marginal decline in Sri Lankan fertility, the DHS 2016 also reported a TFR of 2.2, which is still stand above the replacement level (Table 1). This unforeseen increase in fertility would create significant impacts on the population size, age Fertility increase has been reported in the recent past not only in Sri Lanka but also in a few other developing countries. For instance, a number of Central Asian countries, namely Kazakhstan, Uzbekistan, and Kyrgyz Republic reported a significant increase in fertility since the late 2000s (Spoorenberg, 2013). For an example Kazakhstan's fertility increased by almost 50 percent between the lowest fertility in 1999 and peak in 2009. In Egypt after declining TFR to three children in mid 2000s, it has risen back to 3.5 children in recent years (Goujon and Al Zalak, 2018).
As per Egyptian researchers the reasons for this increase were women marring in younger ages and an increase in religiosity. Even in Algeria, after dipping to 2.2 in the early 2000s, TFR has increased to 2.9 in 2010 (Ouadah-Bedidi et al, 2012).

Comparison of Sri Lankan fertility trends with South Korea
In the context of the economic transition taking place in Sri Lanka, it is logical to compare the fertility decline that had taken place in a country with that of a Newly Industrialized Economy (NIE). Congruent patterns of fertility decline can be observed when analysing changing fertility levels and patterns of Sri Lanka with South Korea, a newly industrialized economy.
As shown in figure 1, similarities in the fertility of both countries are clear. Although in general over the past many decades, both Sri Lanka and South Korea show fertility decline, before the year 1973 Sri Lanka's fertility level (TFR) has always remained lower than those of South Korea.
In 1963 when Sri Lankan women recorded a TFR value of 5.3 live births, the corresponding figure of South Korea was 5.6, and even in 1971 also the same difference was maintained. However, since 1973 to date Sri Lanka's fertility is at a higher level than South Korea.  (Ban et al. 2002;De Silva, 1994; 1963 1971 1981 1991 1995 1998 2005 2011 2016 No. of live births per woman (TFR)

Factors contributed to fertility upturn
Fertility, a naturally complex process is also affected by several socioeconomic and biological variables. Additionally, the decrease or increase of the fertility level in any given country, is largely determined by three non-genetic factors such as changes in the age at marriage, use of contraception, and the practice of induced abortion. However, the impact of these changing agents on fertility also varies. For instance, an increase in the average age at marriage of females could have a negative impact on fertility, while a decrease in the age at marriage could affect the level of fertility in an opposite way (Jones, 2007).

Decline in mean age at marriage
The postponement of marriage was a significant contributor to the decline in fertility in Sri Lanka during the 20 th Century.   The declining trend in the mean age at marriage of females during the past decade has contributed significantly to the recent increase in fertility from below replacement to well above the replacement in Sri Lanka.
During the past few decades in Sri Lanka divorce and separation rates have increased significantly and the majority of their entry into the re-marriage has contributed to the fertility increase in Sri Lanka (Siriwardhane and De .

Decline in contraceptive use
The contraceptive prevalence among Sri Lankan women increased rapidly in the period following 1970. There was increasing awareness of contraception not only for the purpose of ceasing childbearing but also for the spacing of births.   Census and Statistics (2002;2009;2017) One of the important reasons for the drop in the proportion of women relying on permanent methods is the family disputes in Sri Lanka.
During the past many years, the stability of the Sri Lankan families had deteriorated (De Silva, 2003;2005). Consequently, many married men and women would not have opted for permanent methods to plan their families as there was the possibility of changing marital partners during the life cycle of their marriages. This was further aggravated with the Tsunami and other natural disasters, where sterilized men and women who lost their children, had no chance of replacing them.
Although modern temporary methods have increased over the past few years, due to the significant drop in the acceptance of permanent contraceptive methods, some of the women would have experienced unwanted pregnancies, primarily due to the user or method failure (De Silva et al., 2006). Since the abortion services were hardly available during the past few years most of the unwanted pregnancies would have endedup as live births and this phenomenon might have partly contributed to the recent fertility increase.

Withdrawal of abortion service
Before 2006, despite the highly restricted abortion laws in Sri Lanka, a large proportion of women with unwanted pregnancies were able to terminate their pregnancy (De Silva, 1992 and. Fertility decline observed in 1980s and 1990s was significantly related to this practice. However, by the beginning of the year 2006, the government launched a well-coordinated drive to bring to halt the abortion services provided in the country by (I)NGOs and some private facilities. As a result, almost all abortion service centres were closed down, and therefore the majority Having experienced natural (Tsunami) and man-made (civil strife) disasters, most couples in Sri Lanka has revised their fertility preferences towards higher side and the outcome is the increase of fertility. In addition to the large scale disasters due to changes in socio-economic environment in Sri Lanka, most couples also would have changed their fertility preferences (De Silva et al., 2010).

Increase in life expectancy
Prior to 1962, the female life expectancy at birth had been lower than the male which can be largely attributed to high maternal mortality rates which stood at two digit levels.  Silva, 2008).

Table 3: Life expectancy at birth (in years)
Source: De Silva (2008) and Department of Census & Statistics (1991;undated) During

Drop in labour Migration
The mass contract labour migration from Sri Lanka commenced in 1976,  (Table 4)  would be an additional component of the population growth of Sri Lanka, which would be realized in the next five years time. to identify the significant upturn in fertility, due to the unavailability of data, that projection results tend to under estimate the future size of the population. However, the latest projection by De Silva and de  captured the upturn in fertility well, thus it tends to provide more reliable estimates on the volume of future population.

www.ncas.ac.lk
National Centre for Advanced Studies in Humanities and Social Science

Growth of total population
After obtaining independence from the British, the first census of Sri Lanka was carried out in 1953, which recorded a population of 8.1 million. Since then censuses were conducted in 1963, 1971, and 1981. The enumerated population of the country in 1981 was reported as 14.8 million ( Table   5).The size of the total population of Sri Lanka in census 2001 stood at 18.7 million and by 2012 census it has reached 20.4 million mark.   stability by the early 2030s with a population slightly below 22 million (De Silva, 2007). This deviation is mainly due to the increase in fertility from below replacement to above replacement level in the past 10-15 years. As of the more reliable latest projection, it is more likely that in after another 20-25 years the country's population would be over 25 million mark by adding another 5 million to the 2012 population base.   and De Silva (2007) Apart from the change in projected size of the population, due to the fertility and life expectancy increase, population dynamics such as age structure, dependency ratios, and demographic dividend would be expected to change in coming years and such changes are discussed in subsequent sections.    and De Silva (2007) The projected volume of child population for the year 2022 in two projections demonstrates a significant difference (Figure 3). Although the difference in child population is noted to be 1.2 million (5.3-4.1 million) in 2022 the corresponding difference in 2032 is 1.6 million (5.1-3.5 million).

Changing age structure
The reason for this difference is originated from the fertility upturn  (Table 1).
Although only 12.4 percent of the enumerated population is noted as elderly in 2012 census, in both projections, the proportion of elderly projected for the year 2022 and 2032 is significantly higher (Table 7).

Unlike the child proportion, the elderly proportion reported by De Silva
(2007) is marginally higher than the corresponding figures by De Silva and de . This highlights the fact that in the earlier projection (De Silva, 2007), the upturn in fertility was not captured as that child proportion was significantly lower. Consequently, the proportion of the elderly population reported to be higher.

Ageing of the population is well demonstrated in the both projections
and more-or-less same volume of the elderly population (60+ years) is projected for 2022 as well as for 2032 (Fig 3). Although the projected results are fairly consistent and related to the ageing process what is important to note is during 2022 and 2032 one could expected addition of one million elders; in 2032 the volume of elderly would be 4.6-4.7 million.
Compared with a 2.5 million elderly observed in 2012, doubling of that would occur by another 20-25 years. It is important to note that as of the more reliable latest projection (De Silva, 2007), around 2034 the volume of child and elderly population would equilibrate at around 5.0 million.

Figure 3: Projected population in three broad age groups by two different projections done in 2007 and 2015
Source: De Silva (2007); De  Of to 13.9 million). This confirms the fact that during the coming next 10-12 years Sri Lanka would have a significantly large working age population, which will have a significant impact on the country's labour force -the growth of the labour force would be a positive one.

Change in the dependency ratio
The age structure changes of a population determine the change in the level of demographic dependency. Changes in the different components in age structure have major implications on the country's socio-economic and development activities. Thus, the economic effect of the changes in the functional age groups, namely children (0-14 years), adults of working age (15-59 years), and the elderly (60+ years) can be presented by a summary measure known as the dependency ratio. Three dependency ratios, comprising the child dependency, old age dependency and, the sum of these two fractions known as the total dependency, are presented in table 8.  Silva (2007); De .  child dependency ratios presented by De  for the subsequent periods, except 2017, is declining (Table 8), however each figure is higher than the corresponding child dependency ratios reported by De Silva (2007). According to the latest projection (2015) child dependancy will remain more-or-lessat the same level during 2012 to 202, however beyond 2027 a sizable decline could be observed.
Old age dependency ratios projected by De Silva (2007) and De  do not vary significantly until the year 2037, however beyond that the projected ratios of the latest projection are lower compared to the corresponding ratios of the 2007 projection.
Although the overall dependency ratio derived from the 2012 census data stand as high as at 60.2 dependents per 100 working age population, the projected corresponding figure by De Silva (2007) is 54.9 only ( Table 8).
As per the latest projection, the overall dependency ratios during 2017 to 2047 are significantly higher than the corresponding ratios of 2007 projection. The increase in fertility over the past 10-15 years and assumed slow decline in fertility for coming years has contributed to the increase in overall dependency of the Sri Lankan population.

Demographic dividend
The demographic dividend is the potential accruable gain by the society during the period of demographic transition in which there is a high ratio of individuals in the working age, in relation to the old and children in the dependent age categories (children below age 15 and elders above 60 or 65 years). This phenomenon (age structure transition) is defined as the first demographic dividend. Physical capital accumulation could generate sustainable economic growth, which is defined as the second demographic dividend (Mason, 2005).
As of the latest projection (De Silva and de Silva, 2015) the age structure transition, covering the period of 1991 to 2034, has produced a demographic dividend, which is conducive for an economic take-off. (first diagram of Figure 4). During the period of demographic dividend, the proportion of the people in the working ages (aged 15-64) is noted to be significantly larger than the proportion in the dependent age categories www.ncas.ac.lk National Centre for Advanced Studies in Humanities and Social Science (aged less than 15 and 65 and above). Nevertheless if the working age is defined as 15-59 years, the dividend would fade away by 2019.

Figure 4: Age structure transition and prevalence of demographic dividend
Source: De  It is important to note that in either 15-59 or 15-64 years of the population considered as the working age category, the least dependency ratio is noticeable around the year 2001 (De . In fact, beyond 2001 the ratio has been increasing significantly. Apart from the rapid ageing process the increase of the child population, due to fertility up turn, has contributed for this demographic environment. This highlights the fact that Sri Lanka has already missed the best era of the demographic dividend. In other words, although according to the UN definition the dividend would be available for another 13 years (2021-2034), the remaining period is relatively not attractive to the period that the country has already consumed, which is 30 (2021-1991) years ( Figure 5). Source: Calculated by using data from De Silva (2007); De  What is the impact of fertility upturn on demographic dividend in Sri Lanka? As of the standard definition of demographic dividend this window of opportunity will be available for another 13 years (De . However, as of the earlier projection (De Silva 2007), which did not capture the fertility upturn, reported only another 09 years as the remaining time period of dividend (Table 8).   (Bloom et al., 2003).
In a situation where the expected improvements in the socio-economic environment are not introduced, the existence of a window of opportunity alone will not be productive. In fact, along with a rapid aging process, Sri Lanka will enter into a period of demographic turbulence. With the onset of demographic turbulence, population and labour supply will begin to decline and the dependency burden will increase significantly along with population ageing.

Demographic dividend in South Korea and India
Newly industrialized countries such as the Republic of Korea, Singapore, Hong Kong, and Taiwan had effectively utilized the window of opportunity offered by the best demographic environment. In each of the respective countries, the rapid economic take-off had taken place when the dependency burden was the least and the highest proportion of the population age structure, was in the working ages (Bloom et al., 2003).
The demographic dividend of South Korea spread-out from 1970-1995 and experienced significant economic growth.
India in South Asia, which will be the most populous country in the world in a few years has experienced the onset of the demographic dividend in (Wlliam et al., 2018. The dividend of India would continue up to the middle of the present century, depending on how India's demographic dynamics would change during the following few decades. Due to the diversity of India, the overall dividend is expected to spread thinly over a long period of time.
India would face a significant challenge in providing suitable employment opportunities to its enormously expanding working age population.
Despite that due to COVID-19 pandemic situation in India, which is becoming as a world epi-centre, the benefits of demographic dividend will not be realized in the coming few years in India. The dividend would millennium and possible implications: A case study of Sri Lanka not be realized without a massive effort to create new jobs, especially those that could put illiterate and semi-skilled labour force to work. It will be extremely difficult to generate a large volume of employment opportunities in a short period of time and significant numbers will search for suitable employment opportunities in foreign countries including Sri Lanka. If this proposition is realised in subsequent years, then Sri Lankans would face a serious challenge in local and foreign labour markets to secure suitable employment opportunities for locals.

Conclusion and policy implications
Demonstrating the continuation of the fertility decline, the DHS 2000, reported a TFR of 1.9 births for the period 1995-2000. Nevertheless, unexpectedly during the past 10-15 years, Sri Lankan fertility has increased from below replacement to well above replacement -Simply a reversed trend has occurred. Consequent to this new trend, Sri Lanka's fertility is closely on par with or even higher than the fertility levels of many South Asian countries. However, the matter of concern is that in all other South Asian countries fertility had been gradually declining during the new millennium, whereas Sri Lanka is the only country that indicated an increasing trend.
Key factors that contributed to the fertility upturn include a decline in age at Such a significant growth of the population would create serious implications, including demand for housing, health, food production, education, energy, water, transportation, and infrastructure facilities in the near future. Problems related to environmental degradation particularly, deforestation, landslides, soil erosion, water and air pollution, garbage disposal, etc. would also likely to emerge at serious levels in coming decades. All these issues are expected to impact implicitly and explicitly on the socio-economic progress of the nation.
Apart from the increase in total population, due to fertility upturn and changes in other demographic components (life expectancy and migration), population dynamics such as age structure, dependency burden and demographic dividend would be expected to change in coming years.
Consequently to fertility increase, by 2020s child population, and 2030s youth population would increase significantly compared to 2012 figures.
On the other hand although the elderly was only 12.4 percent in 2012, millennium and possible implications: A case study of Sri Lanka what is important to note is that compared to 12.7 million working age population in 2012 by 2022 and 2032 their size would increase to almost 14 million. This confirms the fact that at least during the coming 10-12 years Sri Lanka would have a significantly large working age population, which will have a significant impact on the country's labour force the growth of the labour force would be a positive one.
With the recent increase in fertility observed in Sri Lanka, the child dependency ratio is likely to increase in the near future together with the ageing of the population. A future demographic scenario for Sri Lanka indicates a difficult outlook at both ends of the population pyramid. In other words, a contradictory situation is emerging in which the ageing of the population would continue along with an increase in child dependency, thus increasing the dependency burden for the working population at both ends of the population pyramid.
As of the latest projection of 2015, by taking 15-64 years as working age population, the age structure transition, covering the period of 1991 to 2034, has produced a demographic dividend, which is conducive for an economic take-off. Although the dividend would be available for another 13 years (2021-2034), the remaining period is relatively not attractive to the period that the country have already consumed, which is 30 years.
However when assessment is made with the alternative definition (15-59 years as working age) the dividend would be available only up to year 2019.
The fertility increase has been able to further increase the demographic www.ncas.ac.lk National Centre for Advanced Studies in Humanities and Social Science dividend-a longer period of window of a opportunity for economic progress. As of the standard definition (15-64 years) and alternative definition (15-59 years), the dividend has increased by further 4 years and 2 years respectively.
Demographic dividend, which emerges due to age structure transition, does not provide automatic benefits. However, in a congenial environment of political stability, adequate savings, investment potential (FDIs), human capital including enhanced female labour force participation, productivity, and the knowledge economy, the optimum utilization of the dividend to gain economic acceleration would materialize. Failure to utilize the dividend successfully in the reaming short period with the required adjustments in the socio-economic arena would push the country into demographic turbulence.