1
$\begingroup$

In every standard growth model, we suppose labor=1.
I learnd that this is because optimization solution of labor is equal to 1.
How to derive that optimazation solution?

For example, consider that standard CASS growth model.
$max \sum_{t=0}^{\infty} \beta^t u(c_t)$ s.t $c_t+i_t=y_t$, $k_{t+1}=(1-\delta)k_t +i_t$, $y_t=A_t k_t^\alpha (E_t L_t)^{1-\alpha} $
I tried to use lagrangian but failed because there is no labor term in utility function.
Every FOC condition shows $MPL=0$. Is that right?
How can I derive the optimal solution for labor if there is no labor disutility term in utility function?
Or it's just a simple answer that there is no disutility of labor we chose the maximum labor time(=1)?

$\endgroup$
1
  • $\begingroup$ It should be Ramsey-Cass-Koopmans standard growth model. Let’s give due credit :) $\endgroup$
    – erik
    Apr 13 at 16:57

1 Answer 1

2
$\begingroup$

Formally, you would need to set up a Kuhn-Karush-Tucker-Problem with the two inequality constraints $N_t<1$ and $N_t>0$. Plugging in for consumption, you get the Lagrangian

$ L = \sum\limits_{t = 0}^\infty {{\beta ^t}\ln \left( {\left( {1 - \delta } \right){K_t} + K_t^\alpha N_t^{1 - \alpha } - {K_{t + 1}}} \right)} + {\beta ^t}{\lambda _{1,t}}\left( {1 - {N_t}} \right) + {\beta ^t}{\lambda _{2,t}}\left( {0 + {N_t}} \right) $

with the relevant first order conditions:

\begin{gathered} \frac{{\partial L}}{{\partial {N_t}}} = {\beta ^t}\frac{1}{{\left( {\left( {1 - \delta } \right){K_t} + K_t^\alpha N_t^{1 - \alpha } - {K_{t + 1}}} \right)}}\left( {1 - \alpha } \right)\left( {K_t^\alpha N_t^{ - \alpha }} \right) - {\beta ^t}{\lambda _{1,t}} + {\beta ^t}{\lambda _{2,t}} = 0 \hfill \\ {\lambda _{1,t}}\left( {1 - {N_t}} \right) = 0 \hfill \\ {\lambda _{2,t}}\left( {0 + {N_t}} \right) = 0 \hfill \\ \end{gathered}

The last two are the complementary slackness conditions. You need to check the three cases of

  1. no constraint is binding (interior solution) or
  2. the lower bound is binding ($N=0$) or
  3. the upper bound is binding ($N=1$)

We do this in turn:

  1. If the inequality constraints were non-binding, the first condition implies

\begin{gathered} \frac{{\partial L}}{{\partial {N_t}}} = {\beta ^t}\frac{1}{{\left( {\left( {1 - \delta } \right){K_t} + K_t^\alpha N_t^{1 - \alpha } - {K_{t + 1}}} \right)}}\left( {1 - \alpha } \right)\left( {K_t^\alpha N_t^{ - \alpha }} \right) = 0\\ \Rightarrow \frac{1}{{N_t^\alpha }} = 0 \\ \Rightarrow {N_t} = \infty \end{gathered}

This is a contradiction.

  1. Next, if $N_t=0$, then $\lambda_{1,t}=0$ and $\lambda_{2,t}=-\infty$. The multiplier must be positive, thus this is not a solution.

  2. Finally, if $N_t=1$, then $\lambda_{2,t}=0$ and $ {\lambda _{1,t}} = \frac{1}{{\left( {\left( {1 - \delta } \right){K_t} + K_t^\alpha - {K_{t + 1}}} \right)}}\left( {1 - \alpha } \right)\left( {K_t^\alpha } \right) > 0 $ which is the solution.

This formally proves that with no disutility of labor, the maximum feasible amount will be chosen.

$\endgroup$

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.