Explaining Love

Following my #lovedata challenge this morning, the first few attempts at explaining love around the world are already trickling in.

But first, the finding that blew my mind. Commenter Renars pointed out that the very data I had plotted—the proportion of people feeling love in a country on a typical day, versus a measure of GDP per capita (a measure of average income)–form a heart-shaped cloud. Really. Take a closer look at it. This may be the most amazing chart I’ve ever drawn.

And we’ve also got some other interesting results:

  • For those thinking of using alcohol to loosen inhibitions, Jason Ruspini has news for you: The correlation between love and alcohol consumption is +0.03. Perhaps more striking, the correlation between love and fertility is mildly negative! (-0.04). And there’s good news: Love is negative correlated with mortality (-0.19). 
  • My former student Brian O’Hara thought that religion might be part of the explanation. Alas, the importance of religion is basically unrelated to love:

@justinwolfers on a hunch I tried #lovedata vs importance of religion. (via bit.ly/acYWP8) Not the answer. — Brian O’Hara (@Brian_OHara) February 14, 2013

  • And Tom McDermott found some interesting regional patterns in the data. Latin American and Caribbean countries experience a whole lotta love, as does Sub-Saharan Africa.

@justinwolfers Some regional bias – controlling for GDP per cap, LAC and SSA more loved up than the rest! #lovedata — Tom McDermott (@Tom_Mc_D) February 14, 2013

  • And Pierre-Louis Vezina has a whole blog post of interesting correlations, here. There’s more love in countries with greater inequality and less military spending. 

OK, correlation may not imply causation, but still: Fascinating stuff. Keep ’em coming!


I love that you are analyzing this, but I would worry that the difference may be due to subtle differences in the definition of love in different languages. Some languages may imply romance to a greater degree while others imply more of a caring feeling. Was there any control for that?


What if you control the fertility regression to some measure of the use of contraceptives? Maybe then "love" would have a positive effect...


Easily this first figure is the most important result in Rorschach economics since "Japan's Phillips Curve Looks Like Japan." http://marginalrevolution.com/marginalrevolution/2007/11/rorschach-econo.html




I don't have access to statistics software, but I would be curious if there is a relation between predominantly Christian/Catholic countries and love. Not any religion, but a specific religion. I think Steve is close that it more likely reflects how cultures define love. (Significant other/children vs good friends)


If the heart shape is really a correct characterization of the trend, this is a textbook case of "non-linear correlation", i.e., you should not apply linear regression to this data. I suspect you have already been aware of this, given your expertise as an economist. Would you really partition the data into half at around 6 in the horizontal axis and apply two regressions, or is this just a fluke?


I ran some regressions using Geert Hofstede’s data on national culture, a measure of life satisfaction, WDI data ... It seems that a higher preference for indulgence as well as a higher life satisfaction is positive correlated with love. I also found a negative relationship between love and a measure of long-term orientation.

The results and the dataset are here: https://sites.google.com/site/scbotten/miscellaneous/explaining-love


Man knew how to have sex and raise kids before there was any language, much less the word love. Love is a word created by society that is used to hide all the selfish reasons we get in to relationships for. It makes you and your partner seem noble while your really there to relieve loneliness, raise your kids and have a regular sex partner.