Why are Koreans going extinct? — Maybe due to a lack of diversity
South Korea is living through an era with a projected birth rate of 0.73 children per woman in 2023. Discussions in legislation and administration are focused on what financial incentives should be provided to encourage childbirth.
I believe that this policy framework may reinforce the perception that “kids = a loss that needs compensation.” This approach is not economically sustainable. If monetary reasons were the cause, it would be difficult to explain why Korea, with a much lower GDP per capita in the 1970s, had a high birth rate.
In Korea, there are unwritten rules regarding marriage and childbirth. These include the necessity of a legal marriage, impressive weddings, owning a home in your 30s, and sending children to English kindergartens. These implicit rules are perpetuated through social media, causing significant stress for young people, leading them to avoid trying rather than risk being seen as failures. Culture of comparing and being conscious of others have raised the overall cost of parenting(both psycologically and financially).
If in our society, it becomes common to witness things like 22-year-old college students marrying, unwed pregnancies, early marriages in rural areas in the USA, or cohabitting partnerships in France, wouldn’t this help in reducing the fear of childbirth and encourage people to consider it as a viable option for personal happiness?
This article aims to analyze the correlation between the diversity of live and birth rates, categorized by country. Life diversity is defined as the state where one can freely choose their life path without being overly concerned about others’ opinions. While it’s challenging to quantify ‘life’ as a single index, given its multifaceted nature encompassing family, health, career, relationships, and more, it’s still essential to use the term ‘life’ as childbirth decisions are intricately linked to personal values and circumstances. Although we can’t assign a numeric value to life diversity for each country, we can get a comprehensive understanding of it by examining various indicators that together form the concept of “life diversity.”
Method of Data Generation and Processing
I limited my data sources to statistics published or cited by the OECD, excluding population density statistics. The advantages of this limitation are as follows:
First, the credibility of the OECD, a reputable organization, ensures a certain level of trustworthiness.
Second, using data collected by the same statistical agency guarantees consistency in measurement methods and criteria.
Third, it’s essential not to compare birth rates between countries with vastly different socio-economic statuses, which means we can maintain a degree of homogeneity in our subjects.
A total of seven CSV files were used, and they share the following commonalities in processing and generation:
- If no CSV file was available, data from PDF files were manually transferred to a CSV file.
- Data from countries other than the 38 OECD member countries were deleted, and countries not on the list were added manually.
- Manually added data sometimes lacked a corresponding Value for a country (NaN). Assuming that countries in the same cultural region share similar values, we calculated the average for that region and used it to fill in the missing values.
Lesson 2: Cultural Regions — WORLD GEO (weebly.com) - If all countries in a cultural region lacked a Value, resulting in NaN, this error (due to the absence of a value to replace it with) was corrected by researching a representative value through Google searches, and annotations were made in the CSV file to reflect this.
These are the data that I used…
- OECD Fertility Rate
This is the total number of children a woman in an OECD country is expected to have during her fertile years. As the dependent variable, I used the average values from the most recent period, 2019–2021.
Demography — Fertility rates — OECD Data - OECD Diversity Index Based on Country of Birth, 2015
This index indicates the likelihood that two randomly selected people in a country were born in the same country. A score of 0 means everyone was born in the same country, while 10 means everyone was born in different countries. This is an independent variable and is data from a single year, 2015, earlier in the time series than the fertility rate.
Diversity in OECD countries: Population diversity, labour market inclusion and acceptance of diversity | All Hands In? Making Diversity Work for All | OECD iLibrary (oecd-ilibrary.org) - OECD Cohabitation-Forms-Partnership, 2011
This shows the percentage of individuals in a country living alone, in legal marriage, or in cohabitation. Additionally, I calculated the “cohabitation rate/legal marriage rate” for further use in index development.
OECD Family Database — OECD - Population Density (People per sq. km of Land Area)
This is the population density of a country (number of people per square kilometer) and implies the degree of social surveillance. An average from 2014–2016 was calculated.
Population density (people per sq. km of land area) — OECD members | Data (worldbank.org) - Religious Diversity Index in OECD Countries, 2010
This index represents religious diversity within a country. It’s estimated based on eight religions: Buddhism, Protestantism, Hinduism, Indigenous religions, Judaism, Islam, and others. A score of 10 means an equal distribution of populations among these eight religions.
Religious-Diversity-full-report.pdf (pewresearch.org) - Augmented Global Index on Legal Recognition of Homosexual Orientation in OECD & Transgender Rights Index in OECD Countries, 2016
This includes indices measuring the extent to which discrimination based on sexual orientation is prohibited, whether same-sex marriage is legalized (a 10-point scale), and statistics showing the legal recognition of transgender individuals. Since the transgender statistics are on a 5-point scale, researchers doubled these values to calculate the average between the two indices, providing an even assessment of legal recognition for homosexuals and transgender individuals.
LGBTI in OECD Countries : A Review | OECD Social, Employment and Migration Working Papers | OECD iLibrary (oecd-ilibrary.org)
Algorithm Overview
I used python to analyze these data.
- Output dictionaries corresponding to {‘Country Code’ : Value}
- Read 7 CSV data files.
- Recognize country codes and values to produce dictionaries.
- Derive the dependent variable, fertility rate, and 5 independent variables for index development.
2. Replace Missing Values
- Missing values prevent normalization.
- Define cultural regions. Assume similar values within each cultural region.
- Calculate average values for each cultural region using data excluding NaN.
3. Normalization
- Normalization is necessary for comparison on the same scale due to different ranges.
- Create a function for Min-Max Normalization.
- Normalize all dictionaries to a range where the maximum value is 1 and the minimum value is 0.
- For population density, since higher density is assumed to hinder diversity, normalization is reversed, setting the maximum value to 0 and the minimum value to 1.
4. Create the Life Diversity Index
- Combine the 5 dictionaries, excluding the fertility rate (dependent variable), and divide by 5 to create the index.
- A simple arithmetic mean is used for ease of implementation and to avoid the non-scientific approach of subjectively assigning different weights to variables.
- Limitation 1: The Life Diversity Index represents more than just 5 factors.
- Limitation 2: It cannot be said that all variables have the same importance in index creation.
5. Correlation Analysis and Visualization
- Extract common countries (38 OECD countries) between normalized fertility rate and Life Diversity Index dictionaries.
- Calculate Pearson correlation coefficient and p-value.
- Visualize scatter plot and regression line.
if you guys wanna look into the codes and csv files, here’s the link!
So, what was the result?
Normalized Fertility Rate(Dependent Variable)
Israel: 1.000 — Mexico: 0.486 — France: 0.450 — Turkey: 0.437 — Iceland: 0.429 — Czech Republic: 0.423 — Colombia: 0.420 — Denmark: 0.397 — Ireland: 0.391 — Sweden: 0.388 — USA: 0.385 — New Zealand: 0.379 — Australia: 0.377 — Estonia: 0.360 — Slovenia: 0.358 — Slovakia: 0.350 — Belgium: 0.344 — Latvia: 0.341 — Netherlands: 0.341 — Costa Rica: 0.339 — UK: 0.339 — Germany: 0.328 — Hungary: 0.326 — Chile: 0.325 — Norway: 0.314 — Switzerland: 0.297 — Lithuania: 0.297 — Austria: 0.285 — Canada: 0.274 — Portugal: 0.256 — Finland: 0.254 — Greece: 0.251 — Poland: 0.248 — Luxembourg: 0.238 — Japan: 0.224 — Italy: 0.188 — Spain: 0.164 — South Korea: 0.000
※ Note: Israel and South Korea are outliers and should be carefully considered when analyzing correlations and visualizations.
Life-Diversity-Index (Independent Variable)
New Zealand: 0.823 — Sweden: 0.789 — Canada: 0.733 — Estonia: 0.713 — France: 0.689 — Australia: 0.664 — Norway: 0.650 — USA: 0.575 — Denmark: 0.570 — Mexico: 0.557 — Austria: 0.554 — Switzerland: 0.542 — Iceland: 0.539 — UK: 0.536 — Colombia: 0.535 — Finland: 0.525 — Latvia: 0.524 — Costa Rica: 0.523 — Spain: 0.522 — Chile: 0.515 — Germany: 0.507 — Ireland: 0.499 — Belgium: 0.485 — Netherlands: 0.483 — Israel: 0.476 — Luxembourg: 0.469 — Turkey: 0.468 — Slovenia: 0.450 — Hungary: 0.445 — Portugal: 0.434 — Czech Republic: 0.381 — Lithuania: 0.380 — Greece: 0.370 — Italy: 0.346 — Slovakia: 0.324 — Japan: 0.248 — Poland: 0.233 — South Korea: 0.215
Visualization Results (Scatter Plot and Regression Line)
Pearson Correlation Coefficient: 0.294
P-value: 0.073
Conclusion and Suggestion
The Pearson Correlation Coefficient is an indicator that measures the strength of the linear relationship between two variables. This value ranges between -1 and +1, where 1 represents a perfect positive linear relationship, -1 a perfect negative linear relationship, and 0 indicates no linear relationship at all. In this project, the Pearson correlation coefficient is 0.294, suggesting a weak correlation between fertility rate and life diversity. However, the fertility rates of Israel and South Korea appear to significantly distort the normalization process due to their outlier status in a sample of only 38 countries.
The p-value indicates the extent to which the observed data supports the null hypothesis, which states that there is no relationship between the two variables. Generally, a p-value of 0.05 or lower is considered statistically significant. In this project, with a p-value of 0.073, there is a possibility that the relationship between the two variables is not statistically significant. Therefore, it is reasonable to conclude that, while there is a weak linear correlation between the two variables, it is not statistically significant.
This research can be further developed by addressing the following limitations:
- Increase the sample size by including countries outside the OECD.
- Extract more variables that represent life diversity.
- Assign different weights to the variables based on their importance in composing the index.
Interestingly, in this project, South Korea ranked lowest in both life diversity and fertility rate. Although statistical significance needs to be improved, this suggests a need to shift the paradigm of existing policies. Moving away from policies that focus on external factors like financial support and maternity leave, various policies aimed at increasing social diversity could have a trickle-down effect on increasing the birth rate.
Policies such as enhancing legal support for cohabiting couples, recognizing new family forms, easing immigration and naturalization criteria for foreigners, creating incentives for diaspora Koreans to return, and addressing overcrowding issues in the capital region can be keys to solving fertility rate problems that have not been resolved by existing policy approaches.
Ultimately, the decision to have children is a personal choice, and this choice is influenced by how one perceives the world. While financial support and reduced working hours may trigger the choice in the short term and are somewhat necessary, what is crucial is changing the culture itself that individuals have faced for over 20 years before deciding to have children.
This internal change, rather than economic support, provides a stronger foundation for the decision to have children. The reason the older generation had many children was not due to a special cost-benefit analysis; it was because it was a common and accepted culture at the time. If the way to create this culture in the 21st century is through building “life diversity,” then we must move in that direction. Ultimately, the most efficient and effective approach may not be to offer more tangible rewards, but to bring about a change in perception.
I look forward to seeing more creative administration in the future.
