Suppose we wish to solve a simple infinite horizon cake eating problem, such that: $$ \max_{\left\{ c_{t}\right\} _{t=0}^{\infty}}\sum_{t}^{\infty}\beta^{t}u\left(c_{t}\right) $$ subject to: $$ c_{t}+a_{t+1}=a_{t} $$ where $a_{t}$ is the total amount of the cake left over for tomorrow. We can use Lagrangian techniques, obtain the Euler equations, and pin down the optimal levels of $c^{*}.$ Alternatively, we can also write this as a Bellman equation as: $$ V\left(a\right)=\max_{a'}\left\{ u\left(a-a'\right)+\beta V\left(a'\right)\right\} $$ and then either solve it numerically (or even obtain an analytical solution in this case, by guess and verify and matching coefficients). My question is this: what are the advantages of using the Bellman approach? The textbook answer is: it helps breaks down an infinite period problem into sequences of two period problems. But, it is still not obvious to me how this is in fact helpful, preferable over simple Lagrangian based optimality conditions. What am I missing?

  • $\begingroup$ Hi @Kwame. I also found myself unsatisfied with the textbook answer when I was studying these issues. A more complete story accounts for the fact that breaking the model down into stages allows it so be solved computationally. An infinite-horizon Lagrangian by comparison does not share the same benefits. $\endgroup$
    – EB3112
    Aug 26, 2022 at 14:04
  • 2
    $\begingroup$ Does this answer your question? Dynamic optimisation $\endgroup$ Aug 26, 2022 at 19:52


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.