indirect utility function
maximum utility achievable with a given income
V(p,m) = U(x(p,m))
Marshallian demand function
maximization problem
two optimality conditions:
find the MRS.
MRS = p1=p2
fit that into the budget line m for optimal demand x1 and x2
holds m constant
λ=dV/dm how much more utility if one unit more income
Hicksian demand function
expenditure minimization e(p, U_head)
the solution to the min problem is χ(p, U_head)
first optimality condition: MRS=MOC
second: fix a utility level (a specific indifference curve) and find minimal expenditure necessary
χ is the quantity consumed
holds U_head constant
μ= de(p,U_head)/dU_head increasing U_head by one unit, how much more expenditure
μ=1/λ
linear transformation
x(ap,am)=x(p,m)
V(ap,am)=V(p,m)
χ(ap,Uf)=χ(p,Uf) Uf= U max
e(ap,Uf)=ae(p,Uf)
bliss point
HH cant do any better, even if his m would increase a lot
conditions for duality
continous utility function
local nonsatiation, price vector p is positive (= monotonicity)
formulas in book p. 177ff
expenditure function
increase the price of one good in the goods vector
if the HH consumes the same bundle as before (same utility) he needs to increase expenditure
HH can reshuffle his expenditure minimizing bundle.
Implication: doesnt make sense if small price increase
it is concave p.188 in book not ipad
Envelop theorem without constraints
if you have f(a,x) with a as parameter and x as variable.
x^R(a) is your best response fct.
f_head(a,x^R(a)) is called a reduced fct. A change in a changes f_head through a and indirectly through x^R(a).
-> envelop theorem claims that we can forget about the indirect part
Envelop theorem with equality constraints
imagine f(a,x). x has to obey the constraint g(a,x)=0.
f_head(a, x^R(a)). A change in a influence f directly and indirectly through x^R(a).
now with lagrangian: L(a,x,λ):=f(a,x) + λg(a,x)
if you dL/da, we get the effect of the parameter on the optimal value and can ignore the effect to the optimal response x^R(a).
proofs in the book
Application of enve.. : shephards lemma
how does minimal expenditure vary with a change in p of a good x_g?
basically dV/dp_g
f(a,x) is overtaken by e(p_g,x)=p*x
f_head = f(a,x^R(a)) is translated into e_head(p_g):=p*χ(p_g)
equality constraint is U(x)-U_head=0
the equality constraint does not depend on p_g
forget about indirect effects
de_head/dp_g=χ_g (shephards lemma)
χ_g is the expenditure minimizung bundle
Application of shephards lemma: Roys identity
Roys identity: using U_head=V(p,e(p,U_head))
differentiate both sides by p_g:
0=dV/ …
this yields: dV/dpg=dV/dm(-χ_g)
a price increase increases necessary expenditure to keep the utility level constant by χ_g (shephards lemma).
if the budget is given, the budget for the other good decreases by χ_g. Then the budget restriction becomes a reduction of utility.
Compensated (hicksian) law of demand
h. law of demand: demand moves inversely to prices
dχ_g/dp_g ≤ 0
Concavity and the Hesse matrix
two ways to explain concavity
if the convex combination kx+(1-k)y is greater than the convex combination of the values f(x) and f(y).
or f’’(x)≤0 <- Hesse matrix (l x l matrix whose entries are all f’’(x)
eine steigung kann zwar noch positiv sein, aber wenn sie immer weniger positiv wird, ist sie concave
Hesse Matrix
fij(x):=d(df(x)/dx_i)/dx_j)
symmetric if all f’’ are continous
definite or semidefinite vector
concave or convex function
(strictly) concave if its hesse matrix negative semidefinite (definite)
(strictly) convex if its hesse matrix is positive semidefinite (definite)
concavity of e
its concavity has some implications
a change in p_g changes χ_g negatively
-> the hesse matrix of e is negative semidefinite
off diagonal entries of e’s hesse matrix
a change in p_g changes χ_k equally as a a change in p_k changes χ_g.
χ_g is just the hicksian demand curve (kinda like x_g for marshallian)
if the change is positive goods g and k are substitutes
if change negative they are called complements
Slutsky equations, two different kind of substitutes
two substitution effects for marshallian demand if price changes:
first: how much are you willing to pay for a bahncard 50 in terms of good x_2 to reduce price of good x_1 by 50%?
second: how much are you willing to change bundle to have higher utility
slutsky and money budget
if good g is normal (budget increase leads to more demand), it is also ordinary (price increase leads to less demand). Then a price increase has stronger effect on marshalllian demand than hicks demand.
if income effect (dx_g/dm)=0, hicksian demand does not depend on the utility level attained.
slutsky and endowment budget
remember x_money(p,p*w)=x_endo(p,w)
if a good is normal and HH consumes more than his endowment, it is also ordinary.
compensating and equivalent variations
(good air quality)
variation (sum of money) is equivalent to an event if it leads to the same indifference curve.
you get money if air quality is reduced or vice versa.
two kinds of variations: equivalent and compensating
equivalent (willingness to pay) from b -> a (lower air quality) would mean reduction of income: m_2-m_1.
compensation (compensation cost) would be: m_3-m_2
equivalent variation
willingness to pay that something doesnt happen
something occurs, EV indicates the same loss of utility if it was to occur on the other good
compensating variation
one good given to you in exchange for reducing another good to achieve the same utility
Last changeda year ago