Endogeneity alert!

An advertisement for Multigrain Cheerios Cereal says: “Studies show that people who eat more whole grain tend to have healthier body weights”.

Well, very clever. Who said that this is causal? This is what we call endogeneity. It is impossible to know what causes what. Usually, people who buy multi-grain cereals are healthier and eat healthier before buying the cereal. That is why, in fact, they bought a boring healthy cereal. A correlation is not causation!

Something similar we can see here: a study that “shows” that people who drink more alcohol tend to earn higher salaries (specially if they drink in bars) because of the social network that they create. Well, so people that drink in bars are in most cases already wealthier. This is a classical case of simultaneity.

The same happens with police and crime (cities with more police tend to be more dangerous… Mmm, it doesn’t make sense, right?).

Many of these studies fail to compare to the counterfactual - what would have happened in the absence of the “treatment”. Meaning, what would be the difference in the salary of the same exact person if he/she drinks in a bar or if he/she doesn’t. Of course ,this is impossible to see, because the person either drink in bars or not (and cloning is still forbidden in all countries). Many econometric techniques try to assess the counterfactual by using larger samples and treatment and control groups that are statistically identical, as in medical research.

In any case, everytime that you see a study like that, ask yourselves if this is just a simple correlation, or it is really the case…

Keep drinking and eating cheerios, but not to be healthier or wealthier… but just because you like it.