Where Form 8865 Functional 2024

Hi there, in this video, I'm going to be talking about functional misspecification and one of the problems this yields for Econometrica. Let's remind ourselves of what Econometrica is all about. We've got some sort of population, and we don't have the entirety of the population data. We only have a small subsample from that population, and we are trying to make some inferences about the population given our sample of data. So, let's think about, let's say we're interested in finding out what the effect of age is on wages or average weekly wages. We can perhaps think that there are some sort of population parameters with which we can model wages. It is equal to alpha plus beta 1 times age plus beta 2 times age squared plus some sort of error term. In this equation, beta 1 would be greater than zero and beta 2 would be less than zero. That's because you can think about there being sort of an inverted U-shape in terms of the return of age to wages. When someone's very young, they've just left University. Being slightly older means they're slightly more experienced in general, which means that they can command a slightly higher wage. So, in this first half of your life, this term tends to dominate. In the second part of an individual's life, it turns out that perhaps when an individual gets to something like 55, their actual wage on average starts to decline, and that could be for a number of reasons, one of them being that people tend to start retiring at that age. So, there is some sort of population basis which is an inverted U-shape of the effect of age on wages. Obviously, most of the time, we don't actually know what the population process is. We try...