Facts & Myths about Sri Lankan population growth
The growing tension among some of the ethnic groups in Sri Lanka following easter attacks have triggered a lot of discussions around the population growth especially focusing on the minority fractions. Through social media you see lot of speculations about the composition of the current population and the growth trends. Most of the claims are nothing more than untested hypotheses and therefore it is timely and sensible to resort to a data driven approach in validating some of these popular claims. The purpose of this article is not to drive a certain opinion but more to present facts and forecast the future using standard statistical methods which hopefully would form a credible foundation for forming opinions.
Method
In order to analyze population growth, we need year on year population data. However a census is only done once a decade and therefore we need to use estimation techniques to calculate the year on year population. The method here is to use the “natural growth” derived from birth and death stats (this is available on a yearly basis)and forecast/backcast from 2012 census figures. A method named Logistic Curve is used for forecasting future years where no birth or death figures are available.
Logistic Curve method is used when the growth rate of population due to births, deaths and migrations takes place under normal situation and it is not subjected to any extraordinary changes like epidemic, war, earthquake or any natural disaster, etc., and the population follows the growth curve characteristics of living things within limited space and economic opportunity.
Population Growth
Using the above method we can expect the population distribution to be as follows by the end of 2018. Note that only the top four ethnic groups are considered in all analysis.
We can also compare the year on year population growth rates for the period 2000–2018
The chart above indicates that population growth rates have been on a steady decline throughout the period. This is consistent with the with the growth trends indicated in the census report 2012.
In comparison Sinhalese birth rates have been the lowest in most years dropping below 1% towards 2018 and Muslim birth rates have been the highest throughout. Muslim birth rates seems significantly high compared to others and we can perform a single sample T-test to check the statistical significance.
The T-test results indicates that Muslim birth rates are significantly high compared to the rest of the population. Such t values usually indicates a treatment effect on the particular sample. But what type of treatment are we looking at? Part of the answer is available in the 2012 census report where an analysis is done on the Total Fertility Rate (TFR) across ethnic groups.
As shown above the TFR for Muslims is at 3.3 compared to Sinhalese and Sri Lankan Tamils at 2.3.
What does the future looks like? Following chart demonstrates how the population compositions would shift in the 2018–2098 period.
There are some critical observations here,
- By 2040 Muslim population would overtake Sri Lankan Tamil population to become the largest minority group.
- By 2098 Sinhalese population would drop from 74% today to 64%.
- By 2098 Muslim population would increase from 10% today to 17%.
- By 2098 Sri Lankan Tamil population would only increase slightly from 11% today to 13%
Children below 8 years
There has been some speculations on the composition of children below 8 years. There are no direct stats available but by looking at the birth rates, a reasonable estimation can be done about the children born in last 8 years.
So there is no basis for the estimation which states that the Sinhalese composition is below 50%.
Polygyny and Average age of marriage
There is also a lot of talk about the polygyny and average age of marriage within the Muslim community given the legal provisions provided. The census report 2012 makes some interesting remarks on the two topics.
Table indicates that the average age of marriage for Muslim women is lower compared to others but it is not significantly low as some might have expected.
Regarding polygyny the report makes the following remark.
Closing Remarks
Note that the analysis was done under following constraints
- Birth and death figures are only available till 2015 (online). 2015–2018 figures were extrapolated using forecast functions.
- Migrations are not considered for analysis here. It is assumed that migrations affected consistently across all ethnic groups.
- Logistic curve calculations require a saturation point and this was estimated considering a possible three times growth and adjusted to coincide with known values.
The beauty of data is that it will tell you the brutal truth whether you like it or not. In the next decade data would become the most important asset any organisation or country would own and the ability to make data driven decisions would decide the winners and losers.
I’m happy to share the raw data if someone is interested in further analysis. Also I’m happy to be challenged on the facts presented and the methods used for analysis.
