13 February 2016

Income and working hours

A few days ago I saw on the Facebook interesting bar graph that illustrated averaged annual working hours across countries included in OECD data collection. What intrigued me was that countries I knew to be poorer had more working hours than the richer countries.

Just to satisfy my curiosity, I decided to head to the OECD site, collect some data and do the analysis myself, just to see is there a correlation between working hours and income.

I’m not really experienced with the OECD database so I could not find a median income per country, so I decided to use minimum income per country because average income is skewed towards larger values due to the huge income gap.
Since I wished just to glance in the answer and not to involve myself in an extensive study, I picked only years 2010, 2011, 2012, 2013, and 2014. This period is similar, shows recovery in the economy. Yeah, I know we’re now in the front another bubble, but that is not subject to this little analysis.
Another reason why I decided to pick several years is to make sure that results from one year are not some kind of fluke.
So here are the data for 2010:

But now, are those data correlated. I had to remove some countries from one dataset because income data has only 25 countries while hours data has 40.
So for only 25 countries result is:
I was quite pleased to see that correlation is negative, as I suspected. The value of -0.73 shows quite satisfying negative relation between variables.
Meaning, more work does not bring more money to a human. How much money an individual earns depends on other factors that form the labor situation in the country. The factors I did not consider in this little analysis.
R-squared value is also quite decent; 0.53 shows that linear fit is a good approximation. And of course, the p-value is 2.2 e-16, adding another confirmation of the significance of this result.
The rest of the analyzed years gave similar results.

All years had similar values for correlation, a coefficient of determination and p-value. I do not know much about the economy to conclude why I got this sad result.
The only thing I can say from this little analysis is that it does not matter how much someone works, but in which country they live in.
Details of the analysis you can find in my repository at GitHub.

No comments:

Post a Comment