This document discusses when a constant-elasticity-of-substitution (CES) function is concave. It is concave if the parameters λi are non-negative and the elasticity parameter ρ is between 0 and 1. This is shown by proving the CES function is quasi-concave, and a homogeneous quasi-concave function of degree 1 must be concave. The document generalizes that a CES function of degree k between 0 and 1, defined as the original CES function raised to the power of k, will also be concave.
This document discusses when a constant-elasticity-of-substitution (CES) function is concave. It is concave if the parameters λi are non-negative and the elasticity parameter ρ is between 0 and 1. This is shown by proving the CES function is quasi-concave, and a homogeneous quasi-concave function of degree 1 must be concave. The document generalizes that a CES function of degree k between 0 and 1, defined as the original CES function raised to the power of k, will also be concave.
This document discusses when a constant-elasticity-of-substitution (CES) function is concave. It is concave if the parameters λi are non-negative and the elasticity parameter ρ is between 0 and 1. This is shown by proving the CES function is quasi-concave, and a homogeneous quasi-concave function of degree 1 must be concave. The document generalizes that a CES function of degree k between 0 and 1, defined as the original CES function raised to the power of k, will also be concave.
This document discusses when a constant-elasticity-of-substitution (CES) function is concave. It is concave if the parameters λi are non-negative and the elasticity parameter ρ is between 0 and 1. This is shown by proving the CES function is quasi-concave, and a homogeneous quasi-concave function of degree 1 must be concave. The document generalizes that a CES function of degree k between 0 and 1, defined as the original CES function raised to the power of k, will also be concave.
Consider a constant-elasticity-of-substitution function with constant returns:
n !1 ρ λi xρi X f (x1 , . . . , xn ) = . (1) i=1
This function will be concave if λi ≥ 0 for all i and ρ ≤ 1.. We could
prove this by by showing that the Hessian is negative semi-definite, but let’s try another method. Step 1-Show that f is quasi-concave Let us first show that the this function is quasi-concave. A function is a quasi-concave function if it is a monotone increasing function of a concave function. So let’s look for a simple concave function ”hidden inside” of f . We note that Then 1 f (x) = g(x) ρ (2) where n λi xρi . X g(x) = (3) i=1
Suppose that 0 < ρ ≤ 1. Then we see from Equation 2 that f (x) is a
monotone increasing function of g(x). Now it is easy to verify if 0 < ρ ≤ 1, then g(x) is a concave function, because its Hessian is simply a diagonal matrix with entries of the form
λi ρ(ρ − 1)xρ−2 .
When 0 < ρ ≤ 1, these terms are all non-positive. Therefore f (x) is a mono- tone increasing function of a concave function g(x) and hence f is concave. Suppose that ρ < 0. Then it must be that f (x) = g(x)1/ρ is a monotone decreasing function of g(x). But if it is a monotone decreasing function of g(x), it is a monotone increasing function of −g(x). Now when ρ < 0, the Hessian matrix of −g(x) is a diagonal matrix with entries of the form
−λi ρ(ρ − 1)xρ−2 .
When ρ < 0, these terms are all non-positive and hence −g is a concave function. Then f (x) is a monotone increasing function of −g(x), it must be that f is quasi-concave.
1 Step 2-Show that f is concave We know that f is quasi-concave, but a quasi-concave function that is homogeneous of degree 1 must be concave. You will find a proof of this proposition in the notes on “useful properties of quasi-concave and homoge- neous functions” appearing in week 5. Step 3- Generalize to CES functions that are homogeneous of degree less than 1 Where f (x) is the constant returns to scale function defined in Equation 1, the CES functions that are of degree k less than 1 take the form f (x)k where 0 < k < 1. Now if we have a concave function f : <n → <, and an increasing concave function g → < → <, then if we define the function h : <b → < so that h(x) = g(f (x)), then h must be a concave function. (This has an easy proof that you should be able to supply.)
The Set F May Be Specified by Equations of The Form (1.1) And/or (1.2) - Alternatively, The Term Global Minimiser Can Be Used To Denote A Point at Which The Function F Attains Its Global Minimum
