Multivariable functions

Ch 11.1–7, excl 11.4

Prof. Rowan Shi

Week 2
Outline

Functions with two arguments

Visualising
Surface plots
Level curves

Partial derivatives

With many arguments

2/24
Outline

Functions with two arguments

Visualising
Surface plots
Level curves

Partial derivatives

With many arguments

Functions with one argument Ch. 4

• reminder: a function is a rule that

assigns a value to an argument

• example: the welfare function u(x )

• takes the consumption amount x of
x u
some good
u(x) • returns the consumer welfare u(x )
given this consumption

• sometimes:
• argument is called independent or
exogenous variable
• value is called dependent or
endogenous variable

3/24
Functions with two arguments Ch. 11.1

• in general, economic variables (and many other real-world

outcomes) depends on multiple factors!

• for example: the quantity demanded for any good depends on

more than just price
• income
• quality of the good
• whether your friends like it
• ...

• we capture these more complex relationships with multivariable

functions

4/24
Functions with two arguments Ch. 11.1

x
• a function with two arguments takes
a pair (x, y ) and assigns value
f (x, y) f (x, y )
f
• now: the rule needs two inputs to
return one value
y

5/24
Functions with two arguments Ch. 11.1

• an important example: consumer welfare (utility function)

3 1
u(x, y ) = x 4 y 4

x amount of good 1 consumed

y amount of good 2 consumed
u the dependent variable: consumer welfare

• economic intuition: total consumer welfare now depends on

consumption of both goods

6/24
Outline

Functions with two arguments

Visualising
Surface plots
Level curves

Partial derivatives

With many arguments

Graphs Ch. 11.3

• with single-argument functions f (x ), we graph them to see. . .

• if f (x ) increases or decreases with x
• slopes and rates of change
• intercepts
• and more. . .

• would be useful for two-argument functions, like our example

• does welfare increase in x? y?
• what’s the maximum welfare?

• but: how do graph a two-argument function like f (x, y )?

7/24
 Surface plots Ch. 11.3

• plotted: our welfare function

2

• goal: for each amount of good 1 x

and good 2 y, visualise welfare u
1.5
• the domain is 2D – imagine it like the
horizontal surface of a table
1
2
1.5 1.5 2
1 1
y x

partials

8/24
 Surface plots Ch. 11.3

• let’s suppose we consume x = 1.1

2
of good 1 and y = 2 of good 2

1.5

1
2
1.5 1.5 2
1 1
y x

partials

8/24
 Surface plots Ch. 11.3

• let’s suppose we consume x = 1.1

2
of good 1 and y = 2 of good 2

• locate the point in the domain

1.5

1
2
1.5 1.5 2
1 1
y x

partials

8/24
 Surface plots Ch. 11.3

• let’s suppose we consume x = 1.1

2
of good 1 and y = 2 of good 2

• locate the point in the domain

1.5
• represent demand as the height
“above” the domain
1
2
2• welfare is around 1.25
1.5 1.5
1 1
y x

partials

8/24
 Surface plots Ch. 11.3

• overall:
2
• in this picture, welfare is the highest
at x = 2, y = 2
• increases as x rises or y rises
1.5 does that make intuitive sense?

1
2
1.5 1.5 2
1 1
y x

partials

8/24
 Surface plots Ch. 11.3

• overall:
2
• in this picture, welfare is the highest
at x = 2, y = 2
• increases as x rises or y rises
1.5 does that make intuitive sense?

• surface plot is like a piece of fabric

• as you move around in the domain,
1 the fabric curves around as the
2
1.5 1.5 2 function changes
1 1 • hard for non-machines to draw
y x
(e.g. me or you)
partials

8/24
Level curves Ch. 11.3

• easier for humans to draw and also inherently informative

• we can ask: for what consumption bundles (x, y ) is welfare 1?

3 1
u(x, y ) = x 4 y 4 = 1

1
• well, x = 1 and y = 1 gives us u = 1. . . so does x = 2 and y = 8

1
 
u(1, 1) = u 2, =
8

9/24
 Level curves Ch. 11.3

y • we can plot the two points that we’ve

found where u(x, y ) = 1

1
1
8
1 2 x

10/24
Level curves Ch. 11.3

y u=1 • we can plot the two points that we’ve

found where u(x, y ) = 1

• continuing on, we can plot all (x, y )

pairs for which we get welfare u = 1

10/24
Level curves Ch. 11.3

y u=1 • we can plot the two points that we’ve

u=2 found where u(x, y ) = 1

• continuing on, we can plot all (x, y )

pairs for which we get welfare u = 1

• we can repeat for u = 2

10/24
Level curves Ch. 11.3

y u=1 • we can plot the two points that we’ve

u=2 found where u(x, y ) = 1
u=3
u=4 • continuing on, we can plot all (x, y )
pairs for which we get welfare u = 1

• we can repeat for u = 2

• and so on

10/24
Level curves Ch. 11.3

y u=1 • these level curves can also gives us

u=2 an idea of the function’s shape
u=3 • we know that on the u = 1 level
curve, the function is at height 1
u=4
• we know that on the u = 2 level
curve, the function is at height 2

• and so on, increasing as we go

northeast (increasing x and y)

• do these look familiar. . . ?

x

10/24
Level curves Ch. 11.3

• easier to draw and helps visualise the 3D surface plot

• observe: each level curve itself is a single-variable function in

(x, y ) space

• in many scenarios, we are also interested in the actual shape of

these level curves

• let’s try to understand what they mean together:

• what does the level curve 1 = u(x, y ) represent?
• why is the level curve decreasing?
• why does 2 = u(x, y ) lie above 1 = u(x, y )?

11/24
Outline

Functions with two arguments

Visualising
Surface plots
Level curves

Partial derivatives

With many arguments

Intuition Ch. 11.2

• we broadly claimed that u(x, y ) rises as consumption rises

• can we express this intuition formally with math?

• with single variable functions, we can use derivatives:

• the derivative f ′ (x ) is the instantaneous rate of change of f (x ) at x
• if f ′ (x ) > 0, the function is increasing at x
decreasing if f ′ (x ) < 0
• example: we expect u′ (x ) > 0

12/24
Intuition Ch. 11.2

• extending to two arguments f (x, y ). . .

∂f
∂x how f changes with x, holding y fixed

∂f
∂y how f changes with y, holding x fixed
see book for other ways to denote partial derivatives

• important differences to the single variable derivative:

• there are two partial derivatives, one for each independent
variable
• we use ∂ to denote partial derivative, emphasising that we hold
the other constant

• math version of ceteris paribus

13/24
Visually Ch. 11.3

• let’s consider this function: broadly,

8 it’s like a hill, reaching a high at (0, 0)

• what do the partial derivatives at the

4 indicated point mean?

• recall: there are now two

0
∂f ∂f
∂x ∂y
−4
2
0 2
0
−2 −2
y x

14/24
Visually Ch. 11.3

• for x. . .
8
• how the function changes as x
increases, holding y constant
4
• at this point: the function rises as x
increases and y is unchanged
0
∂f
(−1, −1.5) > 0
∂x
−4
2
0 2
0
−2 −2
y x

14/24
Visually Ch. 11.3

• for y. . .
8
• how the function changes as y
increases, holding x constant
4
• at this point: it also increases in this
direction!
0
∂f
(−1, −1.5) > 0
∂y
−4
2
0 2
0
−2 −2
y x

14/24
Visually Ch. 11.3

• the partial ∂x
∂f
(x0 , y0 ) has a natural
8 visual meaning

• stand at the point (x0 , y0 ), face in the

4 x direction (here, “east”)
• sign: does the ground slope
upwards or downwards?
0
• value: how steeply?

• for ∂y
∂f
(x0 , y0 ), face in the y direction
−4 (here, “north”)
2
0 2
0 • let’s try with utility
−2 −2
y x

14/24
Computing Ch. 11.2

• the partial in partial derivative precisely captures: how f changes

while moving only one of the variables (x or y)

• translated to math:
∂f
∂x take the derivative with respect to x, as if y is a constant

∂f
∂y take the derivative with respect to y, as if x is a constant

15/24
Partial derivatives: example Ch. 11.2

3 1
• welfare example: u(x, y ) = x 4 y 4

• there are two partial derivatives: ∂u/∂x and ∂u/∂y

• let’s try together: what’s the interpretation of each?

∂u
∂x

∂u
∂y

• what do you expect the signs to be?

16/24
Partial derivatives: example Ch. 11.2

• let’s try together: compute partial derivaties

3 1
u(x, y ) = x 4 y 4
∂u
=
∂x
∂u
=
∂y

• do the signs make sense?

17/24
Outline

Functions with two arguments

Visualising
Surface plots
Level curves

Partial derivatives

With many arguments

Functions with many arguments Ch. 11.7

• what if you also think a good’s quality should matter for welfare?

• say:
3 1
u = qx 4 y 4

• now: welfare depends consumption amounts (x, y ), but also on

the quality of goods q

• we have now written a function of three arguments

18/24
Functions with many arguments Ch. 11.5

• we can write down a function with any number of arguments

• say there are n arguments: then a function

f (x) ≡ f (x1 , x2 , . . . , xn )

takes the arguments x ≡ (x1 , x2 , . . . , xn ) and returns the value

f (x )
here, we collapse all the arguments into x for ease of writing

• in general, humans cannot visualise multi-dimensional space, so

these functions are hard to draw

19/24
Partial derivatives with many arguments Ch. 11.6

• we can think about f ’s partial derivative with respect to x1 . . .

• it’s how f changes as x1 changes, holding x2 , x3 , . . . , xn constant

• similarly, f ’s partial derivative with respect to x2 . . .

• it’s how f changes as x2 changes, holding x1 as well as

x3 , x4 , . . . , xn constant

20/24
Partial derivatives with many arguments Ch. 11.6

• in general: if i is one of 1, 2, . . . , n, what’s f ’s partial derivative

with respect to xi ?

• it’s how f changes as xi changes, holding xj constant for all j ̸= i

• in practice, we compute the same way: derive with respect to xi ,

as though xj are constants

21/24
Partial derivatives with many arguments Ch. 11.6

• let’s try together: welfare with three arguments

3 1 ∂u
u(x, y, q ) = qx 4 y 4 =
∂x
∂u ∂u
= =
∂y ∂q

• do the signs make sense?

22/24
Summary

• understand functions of two or more variables

• read, graph, and interpret level curves

• understand the meaning of and compute partial derivatives

23/24
System of equations practice

Consider a collection of n individuals, each of whom owns

a definite quantity of m different commodities. Let aij be the
number of units of commodity i owned by individual j. What’s
the total amount of commodity 1? Commodity i? Let pi de-
note the price per unit of commodity i. What is the total value
of the commodities owned by individual j?
Ch 15.1, exercise 5

24/24