Government should do AB testing
Government should do AB testing
In a very simplified view:
A Governments goal is to manage a country to improve it, for that it:
-
Creates / changes rules and processes
-
Makes decisions on how to collect and spend the money
-
etc.
That is, it creates decisions that when applied to a country hopefully leads to a better country.
But then the question is, once a Gov. decision is applied how to tell its real impact? that is:
-
How can it be sure is not going to be damaging?
-
How can it be sure that the impact is not because of an external factor, like global economy fluctuation?
-
How can it quantify exactly the impact of each decision?
Imagine a world where:
-
Each decision is quantified, an exact number, like: policy A resulted in +5% GDP.
-
Success of a ruling government is measured by a well quantified impact number.
-
We can compare different initiatives across the government and learn from the ones that work and from the ones that didn’t. That leads to better strategies in future.
Turns out, this is in fact a very common problem across many domains:
How does medicine makes sure a new treatment works?
Experimentation, also known as AB testing !
Pick a group of people with the same disease put part of the people on the new drug and the rest of group without it.
Give it some time.
Compare: Are the ones taking the drug getting better? exactly by how much?
“Science !”
Caveats
Ok, real world is not as simple, sometimes is not feasible to run a clean experiment. Or it might take to long, etc…
Nevertheless, we should try as much as possible go towards that direction.
If the experimentation would part of a Government’s DNA, am convinced over the years infrastructures would be created to allow run experimentation.
Betting in future this (or some variation of it) will be commonplace.