Fertility data, religion and temperature: or how analysts spend their free time

Having worked as an analyst for many years I’ve developed a special relationship with data.

First, I’m always cautious when someone mentions a “fact”, a “study” or a “data point”. Especially, if that someone is a politician or a news / media organization. Remember, that their job is to evoke emotions and tell a story in order to convince you about something. Scientists are the ones who try to find the truth based on data and research (unless they’re pressured into producing a ton of papers, but that’s a whole different story).

Moreover, whenever I see an interesting set of data I try to see if there are any patterns or correlations. That’s I’ve been paid to do for years working for tech companies. Find the data; clean it up and organize it; look for patterns and other findings; produce insights and hypotheses based on them; maybe do some tests (ideally); use the results for better decision-making. This is more of a loop rather than a linear process as you usually go back to the data or the tests and try to validate or update your findings.

The other day I stumbled across the fertility stats for the European Union. It’s a very important topic as the EU is getting older and this will have a whole lot of consequences in the near future. There is a whole range of fertility rates, the lowest one being 1.26 children per woman in Malta (2017), all the way up to 1.9 in France.

When one sees a range of values for a given metric it’s natural to start thinking the causes behind this.

Could the population of each country have to do anything with fertility? Probably not and we don’t see any correlation in the chart below. There is no clear pattern for either big or small countries and how that affects their fertility.

Surely, income plays a big part in the decision to have children, right? GDP per capita doesn’t seem to affect fertility either if we plot the data.

If you look at the list of countries sorted by fertility rate, though, you might notice something.

CountryFertility Rate
France1.9
Sweden1.78
Ireland1.77
Denmark1.75
United Kingdom1.74
Romania1.71
Czech Republic1.69
Latvia1.69
Belgium1.65
Lithuania1.63
Netherlands1.62
Slovenia1.62
Estonia1.59
Germany1.57
Bulgaria1.56
Hungary1.54
Austria1.52
Slovakia1.52
Finland1.49
Poland1.48
Croatia1.42
Luxembourg1.39
Portugal1.38
Greece1.35
Italy1.32
Cyprus1.32
Spain1.31
Malta1.26

The countries with the lowest fertility rate – Malta, Spain, Cyprus, Italy, Greece, Portugal – all belong to the European South. The top ones, on the other hand – France, Sweden, Ireland, Denmark, UK – are situated in the North.

Someone could say that Southern European countries are generally poorer than Northern European countries – which is generally true – but going back to our GDP per capita – Fertility comparison this argument doesn’t hold true. There are countries with higher income and low fertility (Luxembourg, Finland, Austria) and the reverse (Romania, Latvia, Czech Republic).

So, I tried to think what else could be there. The South is warmer than the North. Could the average temperature for each country impact fertility?

Something’s going on here. We have a correlation of -0.6 which is good but not great. And remember correlation does not imply causation!

What else could be there? How about religion? People in the South tend to be more religious in general (or at least that’s what they reply when asked in surveys). Let’s see:

The correlation coefficient is again around -0.6 which shows a certain trend: the least religious countries tend to have a higher fertility.

So, what have we learned from this? Warmer and more religious countries (in the EU; we should be careful not to generalize based on data points from a sample that probably is different than the global population) tend to have a lower fertility.

This is not to say that temperature or religion cause lower fertility! Always be alert not to fall into the trap of spurious correlations.

What we do have is a pattern. In a scientist’s ideal world, we would then take a sample of the population and randomly assign them to different groups (e.g. warm and cold countries) and observe the outcomes. Is there a difference in fertility if the only variable we change is the temperature or the level of religiosity?

Obviously, this is impossible to do for our example in real life. However, a scientist could do some lab research regarding temperature and fertility (maybe this exists, I don’t know) and a social scientist could observe more data regarding religion and fertility.

A policy-maker would benefit from looking into the data and try to implement policies for the countries that have low fertility (if, let’s say, the EU thinks this is a problem).

And an analyst like me, will continue playing around with data, because it’s fun (yes, it can be!) and helps you understand the world around you a bit better.

1 thought on “Fertility data, religion and temperature: or how analysts spend their free time”

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Post