Intermediate Microeconomics : Math Concepts

Andrew Caplin and Srijita Ghosh

Fall 2021

1 / 33
M.1 Convex Combination

M.1 Convex Combination

2 / 33
M.1 Convex Combination M.1.1 Vectors

M.1.1 Vectors

3 / 33
 M.1 Convex Combination M.1.1 Vectors

M.1.1: Vectors

I For first half of course arbitrary quantities of commodities with

fixed units (expressed as vectors)
I R is the real line

Figure: R

4 / 33
 M.1 Convex Combination M.1.1 Vectors

M.1.1: Vectors

I Given two real numbers x < y :

I the closed interval [x, y ] comprises all numbers that lie between
them,
[x, y ] = {z ∈ R|x ≤ z ≤ y };
I the open interval (x, y ) excludes end points.

(x, y ) = {z ∈ R|x < z < y };

I half open (x, y ] and [x, y )

I “Element of”notation ∈
I “set of all”notation {..}
I “Such that” notation |

5 / 33
 M.1 Convex Combination M.1.1 Vectors

M.1.1: Vectors

I RN comprises all ordered lists (“vectors”) of N elements of R

I Will use x̃ for vectors and subscripts xn for components

RN = {x̃ = (x1 , ..., xN )|xn ∈ R, 1 ≤ n ≤ N };

RN
+ = {x̃ = (x1 , ..., xN ) ∈ R |xn ≥ 0, 1 ≤ n ≤ N }.
N

I Need strict inequality of vectors

I Write x̃ >> ỹ for xn > yn all n and x̃ >> 0 for xn > 0 all n .
I x̃ > ỹ implies xn ≥ yn for all n but there exists at least one n0
such that xn0 > yn0

6 / 33
 M.1 Convex Combination M.1.1 Vectors

M.1.1: Vectors

I Consider the standard case of N = 2

Figure: x̃ >> ỹ

7 / 33
 M.1 Convex Combination M.1.1 Vectors

M.1.1: Vectors
I Scalar multiples are defined component by component: given
α ∈ R and x̃ ∈ RN , the vector αx̃ ∈ RN is defined by,
αx̃ = (αx1 , .., αxN )
I Addition of vectors is component by component:
(x1 , ..., xN ) + (y1 , ..., yN ) = (x1 + y1 , ..., xN + yN );
(1, 2, 4) + (3, 7, 12) = (4, 9, 16).
I Given two vectors p̃, x̃ ∈ RN , the dot product is defined as,
N
p̃.x̃ = ∑ pn xn .
n =1
I Think prices, quantities, and value:
p̃ = (1, 2, 3), x̃ = (3, 1, 2) =⇒ p̃.x̃ = 3 + 2 + 6 = 11

8 / 33
M.1 Convex Combination M.1.2 Convex Combination

M.1.2 Convex Combination

9 / 33
 M.1 Convex Combination M.1.2 Convex Combination

M.1.2: Convex Combination

I “Convex combinations” (weighted averages) important

I Abstractly given x̃, ỹ ∈ RN , the corresponding line is:

{λx̃ + (1 − λ)ỹ ∈ RN |λ ∈ R}.

I The corresponding line segment between x̃ and ỹ is:

{λx̃ + (1 − λ)ỹ ∈ RN |λ ∈ [0, 1]}.

I Elements of this set are called convex combinations of x̃ and ỹ .

I Geometrically mid-point is λ = 12 ; points closer to x̃ involve
λ > 12 ; points closer to ỹ involve λ < 12 .

10 / 33
 M.1 Convex Combination M.1.2 Convex Combination

M.1.2: Convex Combination

I Mid-point of line segment from (1, 0) to (0, 1) is (0.5, 0.5) :
0.5(1, 0) + 0.5(0, 1) = (0.5, 0.5)

Figure: Mid-point, λ = 0.5

11 / 33
 M.1 Convex Combination M.1.2 Convex Combination

M.1.2: Convex Combination

I Convex Combination 1 3
4 weight on (1, 0), 4 weight on (0, 1) is
(0.25, 0.75) :
0.25(1, 0) + 0.75(0, 1) = (0.25, 0.75)

Figure: λ = 0.25
12 / 33
 M.1 Convex Combination M.1.2 Convex Combination

M.1.2: Convex Combination

I Convex Combination 3 1
4 weight on (1, 0), 4 weight on (0, 1) is
(0.75, 0.25) :
0.75(1, 0) + 0.25(0, 1) = (0.75, 0.25)

Figure: λ = 0.75
13 / 33
 M.1 Convex Combination M.1.2 Convex Combination

M.1.2: Convex Combination

I Convex Combination 0.6 weight on (1, 0), 0.4 weight on (0, 1)
is (0.6, 0.4) :
0.6(1, 0) + 0.4(0, 1) = (0.6, 0.4)

Figure: λ = 0.6 14 / 33
 M.1 Convex Combination M.1.2 Convex Combination

M.1.2: Convex Combination

I Mid-point of line segment from (1, 2, 4) + (3, 7, 12) is
(2, 4.5, 8) :
0.5(1, 2, 4) + 0.5(3, 7, 12) = (2, 4.5, 8)
I Convex Combination 1 3
4 weight on (1, 2, 4), 4 weight on
(3, 7, 12) is:
0.25(1, 2, 4) + 0.75(3, 7, 12) = (2.5, 5.75, 10)
I Convex Combination 3 1
4 weight on (1, 2, 4), 4 weight on
(3, 7, 12) is:
0.75(1, 2, 4) + 0.25(3, 7, 12) = (1.5, 3.25, 6).
I Convex Combination 0.6 weight on (1, 2, 4), 0.4 weight on
(3, 7, 12) is:
0.6(1, 2, 4) + 0.4(3, 7, 12) = (1.8, 4, 7.2).
15 / 33
M.1 Convex Combination M.1.3 Convex Set

M.1.3 Convex Set

16 / 33
 M.1 Convex Combination M.1.3 Convex Set

M.1.3: Convex Set

I Set X ⊂ RN is convex if, given x̃, ỹ ∈ X , the corresponding

line segments joining x and y are subsets of X .

Figure: panel a: convex set, panel b: not a convex set

17 / 33
 M.1 Convex Combination M.1.3 Convex Set

M.1.3: Convex Set

I Technically, if x̃, ỹ ∈ X and λ ∈ [0, 1] , then,

λx̃ + (1 − λ)ỹ ∈ X .

I Sets we work with are generally convex sets.

I Standard budget set is convex (how will you show?)

18 / 33
M.2 Concave Functions

M.2 Concave Functions

19 / 33
M.2 Concave Functions M.2.1 Concave Functions

M.2.1 Concave Functions

20 / 33
 M.2 Concave Functions M.2.1 Concave Functions

M.2.1: Concave Functions

I Given a convex set X ⊂ RN , f : X −→ R is concave if, given

x̃, ỹ ∈ X and λ ∈ [0, 1] ,

f (λx̃ + (1 − λ)ỹ) ≥ λf (x̃) + (1 − λ)f (ỹ);

I Note the cave shape underneath it to remember name!

I It is strictly concave if the inequality is strict for λ ∈ (0, 1) .
I It is convex if the weak inequality is reversed for λ ∈ [0, 1]
I It is strictly convex if the inequality is strictly reversed for
λ ∈ (0, 1) .

21 / 33
 M.2 Concave Functions M.2.1 Concave Functions

M.2.1: Concave Functions

Figure: panel a: concave function, panel b: convex function

22 / 33
 M.2 Concave Functions M.2.1 Concave Functions

M.2.1: Concave Functions

I In one dimension given X ⊂ R convex and f : X −→ R :

I concavity corresponds to always non-positive second derivative;
f 00 (x ) ≤ 0
I strict concavity corresponds to always strictly negative second
derivative; f 00 (x ) < 0
I convexity corresponds to always non-negative second derivative;
f 00 (x ) ≥ 0
I strict convexity corresponds to always strictly positive second
derivative; f 00 (x ) > 0

23 / 33
 M.2 Concave Functions M.2.1 Concave Functions

M.2.1: Concave Functions

I Examples:
I f (x ) = x =⇒ f 00 (x ) = 0 is linear, hence both concave and
convex, but not strictly either;
I f (x ) = x 2 =⇒ f 00 (x ) = 2 , hence strictly convex
I f (x ) = e x =⇒ f 00 (x ) = e x , hence strictly convex
I f (x ) = ln x on x > 0 =⇒ f 00 (x ) = −x −2 < 0 , hence strictly
concave
I f (x ) = x 12 on x > 0 =⇒ f 00 (x ) = − 41 x − 32 < 0 , hence

strictly concave

24 / 33
 M.2 Concave Functions M.2.1 Concave Functions

M.2.1: Concave Functions

I If the functions f : X −→ R and g : X −→ R for X ⊂ RN

are concave and a, b ≥ 0 , then function h = af + bg is
concave
I If the functions f : X −→ R and g : X −→ R for X ⊂ RN
are strictly concave and a, b > 0 , then function h = af + bg is
strictly concave

25 / 33
M.2 Concave Functions M.2.2 Optimization

M.2.2 Optimization

26 / 33
 M.2 Concave Functions M.2.2 Optimization

M.2.2: Optimization

I Reason we work with strictly concave functions is purely

mathematical: simplifies optimization
I In single variable calculus a maximum, x̂ is such that
f (x̂ ) ≥ f (x ) all x ∈ R .
I Here a necessary condition given f : R −→ R is that the
derivative is zero,
f 0 (x̂ ) = 0.
I Sufficiency condition (SOC) f 00 (x̂ ) ≤ 0 satisfied for concave
functions

27 / 33
 M.2 Concave Functions M.2.2 Optimization

M.2.2: Optimization
I With a bounded interval f : [0, B ] −→ R :
I x̂ = 0 can be optimal provided limx ↓0 f 0 (x ) ≤ 0 (right limit):
sufficient if f is concave.
I x̂ = B can be optimal provided limx ↑B f 0 (x ) ≥ 0 (left limit):
sufficient if f is concave.

Figure: panel a: 0 optimal, panel b: B optimal

28 / 33
 M.2 Concave Functions M.2.2 Optimization

M.2.2: Optimization

I Usually use two goods

I Introduce new techniques in section 1.6

29 / 33
M.3 Partial Derivatives

M.3 Partial Derivatives

30 / 33
 M.3 Partial Derivatives

M.3: Partial Derivatives

I Need to know a little multi-variable calculus

I Define unit vectors en with 1 in position n , zeros elsewhere.
I Define partial derivatives of f : RN
+ −→ R as in SVC

f (x̃ + een ) − f (x̃) ∂f (x̃)

lim ≡ fn (x̃) ≡ ,
e *0 e ∂xn
∂f (x̃)
I Use either notation fn (x̃) or ∂xn whichever is more clear

31 / 33
 M.3 Partial Derivatives

M.3: Partial Derivatives

I Examples with N = 2 :
I With f (x1 , x2 ) = 3x1 + 5x2 :

f1 (x1 , x2 ) = 3; f2 (x1 , x2 ) = 5.
I With f (x1 , x2 ) = 3x1 + 5x2 + x1 x2 :

f1 (x1 , x2 ) = 3 + x2 ; f2 (x1 , x2 ) = 5 + x1 .
I With f (x1 , x2 ) = v (x1 ) + x2 :

f1 (x1 , x2 ) = v 0 (x1 ); f2 (x1 , x2 ) = 1.

I With f (x1 , x2 ) = α ln x1 + (1 − α) ln x2 :
α 1−α
f1 (x1 , x2 ) = ; f2 (x1 , x2 ) = .
x1 x2

32 / 33
 M.3 Partial Derivatives

M.3: Partial Derivatives

I If implicit function often need total derivatives

f (x1 , h (x1 )) = c; x2 = h ( x1 )

I Total derivative

f1 + f2 h 0 (x1 ) = dc = 0
dx2 f1
h 0 ( x1 ) = =−
dx1 f2

33 / 33