3
$\begingroup$

I'm studying the Solow growth model from the Acemoglu's book.

Consider the following standard assumptions:

enter image description here

The fact that $F()$ exhibits constant returns to scale means that F() is linear homogenous. Why the book says that $F()$ is concave? Why it is not strictly concave?

$\endgroup$
5
  • 4
    $\begingroup$ I'm not certain that I've understood your question fully. If the book assumes that $F$ is concave, this is a weaker condition than assuming it is strictly concave. Therefore, the analysis conducted in the book, which holds for concave functions, will also hold for strictly concave functions. $\endgroup$
    – Amit
    Apr 1, 2023 at 0:49
  • $\begingroup$ Thanks for your answer. The book says : "Prove that Assumption 1 implies that F(A,K, L) is concave in K and L but not strictly so." $\endgroup$
    – John M.
    Apr 1, 2023 at 13:35
  • $\begingroup$ @JohnM. Are you asking if there any concave functions that are not strictly concave, or what exactly is your question here? $\endgroup$
    – Giskard
    Apr 1, 2023 at 18:45
  • $\begingroup$ @JohnM. Let me ask you a follow-up question. Can you think of any production function $F$ that satisfy assumption $1$? $\endgroup$
    – Amit
    Apr 1, 2023 at 21:52
  • $\begingroup$ Thanks for your replies. Assumption 1, in my understanding implies that $F$ is strictly concave. Strict concavity implies concavity. Why the book says that $F$ is concave only (and not strictly concave as well)? $\endgroup$
    – John M.
    Apr 2, 2023 at 8:38

1 Answer 1

5
$\begingroup$

If you consider the production function $f(K, L) = F(K, L, A = 1)= K^{\frac{1}{2}}L^{\frac{1}{2}}$, this function meets all the conditions specified in assumption 1 within the interior, but not on the axes due to its lack of differentiability on the axes. Typically, this detail is not explicitly mentioned. Check that $f(K, L)= K^{\frac{1}{2}}L^{\frac{1}{2}}$ is concave and not strictly concave. To show that this is not strictly concave, consider $(K_1, L_1)=(0,0)$, $(K_2, L_2)=(2,2)$ and $\lambda = \frac{1}{2}$,

$f(\lambda (K_1, L_1) +(1-\lambda) (K_2, L_2)) = f(1,1) =1 =\frac{1}{2}(0) + \frac{1}{2}(2) = \lambda f(K_1, L_1) + (1-\lambda) f(K_2, L_2) $

$\endgroup$
3
  • $\begingroup$ I understand the definition of concave function as you wrote above, and makes sense. But I also know that if the second (partial) derivatives are strictly negative, then $F$ is strictly concave. Why this is not the case? $\endgroup$
    – John M.
    Apr 2, 2023 at 10:46
  • 4
    $\begingroup$ @JohnM. The example example is a counter-example to your claim "if the second (partial) derivatives are strictly negative, then $F$ is strictly concave". Your claim is wrong which is why you have a counter-example. $\endgroup$
    – user43302
    Apr 2, 2023 at 11:02
  • $\begingroup$ Thanks. I see. Ok, my statement was simply wrong $\endgroup$
    – John M.
    Apr 2, 2023 at 11:06

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.