In the solution to the problem below provided by my professor he applies a first order condition to the lagrangian for ct, getting pt for the partial derivative of the argument "$\frac{d}{dc_t(t)} \int p_t(w)*c_t(w)dw$". This is puzzling to me since the derivative of a indefinite integral of ct should be ct not 1.

What's going on with this derivative and is it correct?

$\frac{d}{dc_t(t)} \int p_t(w)*c_t(w)dw$ = $p_t(w)$

Problem description

Edit: Removed hand written notes.

  • $\begingroup$ Please don't post handwritten mathematical formulas as pictures. Use MathJax instead. $\endgroup$
    – Herr K.
    Commented Jun 8, 2019 at 18:53

1 Answer 1


A good way to think about this is to think about the cases where there are a finite number of goods, indexed by $\omega \in \Omega$. In this case, the partial derivative is $$ p(\omega) = \frac{\partial}{\partial c(\omega)} \sum_{\omega' \in \Omega} p(\omega') c(\omega'). $$ The case in which $\Omega$ is a continuum of goods follows as a sort of limiting extension. See that even this finite sum can be expressed as an integral, given the proper measure $\mu$: $$ \sum_{\omega \in \Omega} p(\omega) c(\omega) = \int_{\Omega} p(\omega) c(\omega) \,\mu(\mathrm d \omega). $$

You mention,

This is puzzling to me since the derivative of a indefinite integral of $c_t$ should be $c_t$, not 1.

If you provided your reasoning for this, it would be easier to see where the confusion is. In any case, hope this helps!

Also, this guide may provide further assistance: The basics of “Dixit-Stiglitz lite”


Your Answer

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

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