Economics 58

Proof of EnvelopeTheoremfor constrainedoptimization problems(from Varian)

Consider a parameterized maximization problem of the form

M(a ): max g(x1,x2,a)

s u ch that h(n1,n2,o) :0.

The Lagrangian for this problem is

E: g ( rr,az,a) - l h(o1, rz,a),

and the first-order conditions are

Idxt -.ril:o
drr
(27.4)
3 -)3 :o
OfrZ Ofiz
h(ay,x2,@ ): 0.

These conditions determine the optimal choice functions (a{a),rz(o)),

which in turn determine the ma;<imum nalue function

M (o) : s@L(a),
r2(a),a). (27.5)

The envelope theorem gives us a formula for the derivative of the

value function with respect to a parameter in the maximization problem.
Specifically the formula is

d.M(a) _ ?E(x,a)
1
da 0a lx=x(o)
_ 0 g (a r,a z,o) | _ ,i l h(xuxz,a) 1
A" l*=x(c) " Ao lx=x(c).

As before, the interpretation of the partial derivatives needs special ca,re:

they are the derivatives of g and h with respect to a holding t1 ond x2
fiied, at their optirnal aalues.
The proof of the envelope theorem is a straightforwa,rd calculation. Dif-
ferentiate the identity (27.5) to get

#:y^*.H**H,
and substitute from the first-order conditions (27.Q to find

d,M ,f Ah dq , 0h da2l , 0s (27.6)

E: " La^ at - a- dol* a"'
Now observe that the optimal choice functions must identically satisfy
the constraint h(a1(a),r2(a),a) : 0. Difierentiatingthis identity with
respectto a, we have
A h d q ,0hd,a2,A h_^
-r = v' (27.7)
a.rE - 6;rE 6;
Substitute (27.7) into (27.6)to find

ry: -t* *P.

aa oa da'

which is the required result.