Reading this this blog post:
I was wondering if someone could explain how to get J_{k + 1, k + 1} , when I attempt I get:
J_{k + 1, k + 1}
= \frac{ \partial p_{k + 1} }{ \partial c_{k + 1} }
= -\rho(\phi-c_{k+1}) = -\rho(c_{k+1} - \phi)
Thanks
Reading this this blog post:
I was wondering if someone could explain how to get J_{k + 1, k + 1} , when I attempt I get:
J_{k + 1, k + 1}
= \frac{ \partial p_{k + 1} }{ \partial c_{k + 1} }
= -\rho(\phi-c_{k+1}) = -\rho(c_{k+1} - \phi)
Thanks
As you and I have discussed in PM but also should be noted here: the last equality only holds when the latent distribution is symmetric (around 0).
The math here is a bit sloppy – the arguments to all of the rho terms should be \phi - c_{k} not c_{k}. See the code implementation that immediately follows that defines everything correctly.