I am reading LS book "Recursive Macroeconomics Theory".

In this book, as well as in some online lecture notes on the Internet. They often denote, for example, $\pi_t(s^t)$ for probability that history $s^t$ happens at time $t$.

But I wonder, if we are mentioning $s^t$, it implies time is $t$. So, why don't we just make it simple by writing $\pi(s^t)$ instead?

Thank in advance for any help!


You are right, it is the same thing. Actually, later on the book the density over the history $s^t = [s_t, s_{t-1}, \cdots, s_0]$ is written as

$$ \pi(s^t) = \pi(s_t|s_{t-1}) \cdots \pi (s_1|s_0)\pi(s_0) \tag{2.3.1} $$

where $\pi(s_0)$ denotes the probability of the initial state (or $\pi_0(s_0)$ if you wish), and $\pi(s|s')$ is a transition probability.

Note that Eq. (2.3.1) encapsulates the the probability of obtaining a given history in a Markov process, since ${\rm Prob}(s_t|s_{t-1}, s_{t-2}\cdots,s_0) = {\rm Prob}(s_t|s_{t-1}) = \pi(s_t|s_{t-1})$. This just to show that $\pi_t(s^t)\color{blue}{\equiv}\pi(s^t)$


Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.