Added two new Life-Cycle Models to the Intro to Life-Cycle Models. They show how to use ‘experienceasset’ and ‘experienceassetu’ to model human capital (the later is uncertain human capital).
experienceasset is an endogenous state where you cannot choose aprime directly, instead aprime(d,a), that is, next period endogenous state is a function of a decision variable and next period endogenous state. The example in Life-Cycle Model 42 uses experienceasset to model female-labor-force-participation-history’ (a form of human capital).
experienceassetu, is an uncertain version of experienceasset using a ‘u’ shock (between-period i.i.d. shock). The example in Life-Cycle Model 43 uses experienceassetu to model human capital that is accumulated by study (with a shock to the human capital production fn).
There is currently just one example of an actual paper in which these are used to implement the model, namely Attanasio, Low & Sanchez-Marcos (2008) - Explaining Changes in Female Labor Supply in a Life-Cycle Model
Codes and pdf of model: LifeCycleOLGReadingList/AttanasioLowSanchezMarcos2008 at main · robertdkirkby/LifeCycleOLGReadingList · GitHub
The astute reader will notice that for an experienceasset, aprime(d,a), you could just invert this and have d(aprime,a), so you could do it as a standard endogenous state, and then evalute d(aprime,a) inside the return fn (and anywhere else d was needed). But this is computationally way slower (because the space of possible aprime is typically much bigger than the space of possible d; as life-cycle model 42 clearly demonstrates where d is just binary valued). So it is very helpful to have a dedicated feature since human capital is an important economic concept.
I also changed the Life-Cycle Model that deals with semi-exogenous shocks and fertility to be number 29 (it was previously 30). This allowed me to fit the divide-and-conquer as model 30 which is a much better ordering.