Continuity Derivative

Mathematical Economics

Lecture 4

© Hui Xiao Copyrights Reserved

1 / 46
Continuity Derivative

Continuity

Roughly speaking, a function is continuous if it has no ”breaks”.

A function is continuous if it is continuous at each point of its domain.

A function of one variable is continuous at point x = a:

if as one approaches the function from ”the left” (x → a− , i.e. x = a minus

something), or ”the right” (x → a+ , i.e., x = a plus something), the function
approaches the same value,

and this value is the value of the function at x = a (i.e., f(a)).

Formally, function f(x) defined on an open interval including the point x = a, is continuous at that
point if
lim f(x)
x→a

exists, that is limx→a− f (x) = limx→a+ f(x), and

lim f(x) = f(a).

x→a

2 / 46
Continuity Derivative

Discontinuity

Function discontinuous at x = 0.

3 / 46
Function discontinuous at x = 0 since 1/x2 is undefined.
Continuity Derivative

Economic Applications
The possibility of discontinuity is not merely an ”inconvenient” mathematical possibility.
Some economic models exploit the idea of a discontinuity to explain possible characteristics of some Economic
phenomenon.
Example: incentive bonus payments for a critical sales level creates a discontinuity: 20,000 sales threshold creates
”discontinuity” or incentives to sales representatives.

limS→20000− P (S) = 800 + 2000 ̸= limS→20000+ P(S) = 1300 + 2000.

Similarly, governments sometimes create income support programs that are ”all or nothing” in nature creating a
discontinuity in the relationship between hours worked and income.
In above cases, the ”discontinuity” can have important economic implications.

4 / 46
Continuity Derivative

Economic Applications
Bertrand Model of Price Competition (Firms compete in prices or Firms undercut each other’s price):

Assumptions:

2 firms compete in the same market and they produce precisely the same type of
commodity (homogeneous good).
Why firms innovate for product differentiation (iphone vs. Pixel, and Pixel cheaper
than iphone).
All sales go to the firm with the lower price.
If they set the same price, then they share the market equally.

Implications:

Sales for a firm are “discontinuous”:

no sales, when price is above the competitor’s price,
half the market, when price is equal to the competitor’s price,
whole market, when price is “just below” the competitor’s price.

If there is enough product differentiation, the sales could be continuous in pricing

(Always some sales for different market segments, each price point in the demand
curve can be populated by buyers with some matching preferences, making the
5 / 46
Continuity Derivative

Economic Applications
Numerical Example:

If each firm has the same cost function C(Q) = 4Q and market demand is
Q = 20 - 2p, Q > 0, p > 0.
If firm 2 sets its price at 7, we can derive firm 1’s revenue and profit functions.
In fact, for any price p1 < 7, firm 1 captures the entire market, and becomes the single
monopolist for that market.
Firm 1’s revenue function R1 (p1 ):

 p1 × Q = p1 (20 − 2p1 ), p1 < 7, Q1 = Q,
Q Q
R1 (p1 ) = p1 ×Q1 = p1 × 2 = p1 × (10 − p1 ) = 7 × 3 = 21, p1 = 7 = p2 , Q1 = 2 ,

p1 × 0 = 0, p1 > 7, Q1 = 0.
(1)
Firm 1’s profit function π(p1 ):

 (p1 − 4)(20 − 2p1 ), p1 < 7, Q1 = Q,

 Q
(p1 − 4) × 2 = (p1 − 4) × (10 − p1 ) = 3 × 3 = 9,
π(p1 ) = R1 (p1 )−C(Q1 ) = p1 × Q1 − 4Q1 =
Q
| {z }  where p1 = 7 = p2 , Q1 = 2 ,
(p1 −4)×Q1


(p1 − 4) × 0 = 0, p1 > 7, Q1 = 0.
(2)
6 / 46
Continuity Derivative

Market Equilibrium Price

Firm 1’s Revenue & Profit Function:

7 / 46
Continuity Derivative

Derivative
Slope of a Function:
Example: y = f (x) = x2 , where

rise △y y2 − y1 f (x2 ) − f (x1 ) x2 − x21 (x2 − x1 )(x2 + x1 )

Slope = = = = = 2 = = x1 +x2 ,
run △x x2 − x1 x2 − x1 x2 − x1 x2 − x1
(3)
what if x2 and x1 are unbelievably close so that
lim△x = 0 =⇒ limx1 = x2 =⇒ lim Slope = x2 + x2 = 2x2 at x2 .
△x→0
Derivative measures the curvature of the functional curves at various points in the function’s domain.
8 / 46
Continuity Derivative

Slope

For AB’s Slope:

′ △y f (x1 + h1 ) − f (x1 )
AB s Slope = = . (4)
△x x1 + h1 − x1

For AD’s Slope:

′ △y f (x1 + h2 ) − f (x1 )
AD s Slope = = . (5)
△x x1 + h2 − x1

9 / 46
Continuity Derivative

Limit

Thus for any x, we have

f (x + h) − f (x) (x + h)2 − x2 x2 + 2xh + h2 − x2 2xh + h2
Slope = = = = = 2x+h, (6)
x+h−x x+h−x h h

Then if we take the limit as limh → 0,

f (x + h) − f (x)
lim = lim (2x + h) = 2x, (7)
h→0 x+h−x h→0

df (x)
which is the same as dx = f ′ (x) = 2x with y = f (x) = x2 .

To generalize,
f (x + h) − f (x) dy ′
lim = = f (x), (8)
h→0 x+h−x dx
which is also known as the derivative that describes the change in y when the change in x is very
small.

The first order condition (F.O.C) f ′ (x) = 0 (y no longer changes by ∆x or function no longer
curves at x) is necessary but insufficient to guarantee Equilibrium Solutions (need to check
f ′′ (x)).

10 / 46
Continuity Derivative

Function’s Differentiability
Generally, a Function’s derivative exists at x = x1 if

The Function f (x) is well defined at x = x1 . Counter Example: If f (x) = 1

, then the derivative does not exist at
x
x = 0.

Right Hand Side Limit=Left Hand Side Limit. Example: y = f (x) = x2 ,

Left Hand Side Limit at x (or x1 , but use x for generality) and x − h (h > 0), is at the Left Hand Side of x:

f (x) − f (x − h) x2 − (x − h)2 x2 − x2 + 2xh − h2

lim = lim = lim
h→0 x − (x − h) h→0 x − (x − h) h→0 h
(9)
2xh − h2
= lim = lim(2x − h) = 2x.
h→0 h h→0

Right Hand Side Limit at x, and x + h is at the Right Hand Side of x:

f (x + h) − f (x) (x + h)2 − x2 x2 + 2xh + h2 − x2

lim = lim = lim
h→0 x+h−x h→0 x+h−x h→0 h
(10)
2xh + h2
= lim = lim(2x + h) = 2x,
h→0 h h→0

f (x+h)−f (x) f (x)−f (x−h)

So, Right Hand Side Limit lim = 2x = lim = 2x, which is the Left Hand Side
x+h−x x−(x−h)
h→0 h→0
Limit at x for f (x) = x with x being any element in the domain of f (x).
2

11 / 46
Continuity Derivative

Function’s Indifferentiability
If a function is not continuous at a point, then it is not differentiable at that point because besides left- and right-hand limits
being equal, we also need
lim f(x + ∆x) = f(x).
∆x→0

Example: The threshold of $20,000 sales creates ”discontinuity” or incentives to sales representatives, and the Salary Function is
not Continuous at S= 20,000, thus Non-differentiable at S= 20,000.

Function discontinuous at x = 0 since 1/x2 is undefined. 12 / 46

Continuity Derivative

Function’s Indifferentiability

Functions that are Continuous but non-differentiable:

Watch out for kinks in the functions (usually non-differentiable but still continuous at the kink
points), so watch out for corner solutions.

Example:

13 / 46
Continuity Derivative

Function’s Indifferentiability
Example:

This function is continuous at x = 1 but non-differentiable.

For x < 1, use f (x) = x to generate the left-hand side limit,

14 / 46
Continuity Derivative

Function’s Indifferentiability
The Progressive Tax Function T(y) is Continuous at y= 5,000 & 15,000, but
Non-differentiable at y= 5,000, & 15,000.

15 / 46
Continuity Derivative

Derivative
Graphically,
A tangent to a curve is a straight line that touches the curve at a given point.
The line lp below is the tangent to the curve defined by y = f(x) at the point P.
The derivative of the function y = f(x) at a given point P is the slope of the tangent line.
Its slope is defined as the change in y, ∆y = f(x2 ) − f(x1 ), divided by the change in x,
∆x = x2 − x1 :

′ ∆y f(x2 ) − f(x1 )
lp s Slope = =
∆x x2 − x1

16 / 46
Continuity Derivative

Economic Applications

Derivatives’ Economic Applications:

How one variable (cost) changes when another variable (demand) changes is an example of an
important economic relationship.

Economists use ‘Marginal Analysis’ to describe such relationships.

Marginalism describes both an economical method of analysis and a theory of value.

According to this theory, individuals make economic decisions on the margin.

In more general and abstract terms:

How does a small change in one variable x affect another variable y?

This formalizes and analyzes a proposition such as

A profit maximizing firms expand output as long as the marginal revenue (M R) exceeds the
marginal cost (M C).

We use the functional derivatives to characterize such Economics problems.

17 / 46
Continuity Derivative

Economic Applications
Let C = C(y) be a firm’s total cost function, where y is output.
Thus,
∆C C(y + ∆y) − C(y)
=
∆y ∆y
is the rate of change in cost per unit of output produced.
Taking the limit, as ∆y → 0, the instantaneous rate of change or the marginal cost of
production becomes the derivative of the total-cost function:

∆C C(y + ∆y) − C(y) ′

lim = lim = C (y).
∆y→0 ∆y ∆y→0 ∆y

18 / 46
Continuity Derivative

Rules of Differentiation
Although one can derive the derivatives of simple functions from the definition of the derivative (i.e., finding the limit
expressions), this is tedious.
Even for moderately complex functions, doing so can be a very challenging and time-consuming exercise.
Thus, learning the rules of differentiation for functional forms is required. These rules themselves can be generated from first
principles (i.e., from the definition of the derivative).

Derivative of a constant function:

If f(x) = c, a constant, then f′ (x) = 0.

f (x) = 6, then f′ (x) = 0.

Derivative of a linear function:
If f(x) = mx + b, with m and b constants, then f′ (x) = m.

f (x) = 6x + 5, then f′ (x) = 6.

Derivative of a power function:
If f(x) = xn , then f′ (x) = nxn−1 .

f (x) = x3 , then f′ (x) = 3x3−1 = 3x2 .

Derivative of the constant multiple of a function:
If g(x) = cf(x), with c a constant, then g′ (x) = cf′ (x).

g(x) = 2f (x) = 2x3 , then g′ (x) = 2f ′ (x) = 2 × 3x3−1 = 6x2 .

19 / 46
Continuity Derivative

Rules of Differentiation
Derivative of the sum or difference of functions:

If h(x) = g(x) + f(x), then h′ (x) = g′ (x) + f′ (x).

Example: h(x) = 5x + 3x2 , then h′ (x) = 5 + 2 × 3x2−1 = 5 + 6x.

Derivative of the sum of an arbitrary but finite number of functions:

n n
X X
′ ′
If h(x) = gi (x), then h (x) = gi (x).

i=1 i=1

Example: h(x) = 4x + 3x2 − x5 , then

h′ (x) = 1 × 4x1−1 + 2 × 3x2−1 − 5 × x5−1 = 4 + 6x − 5x4 , where x0 = 1.
Derivative of the product of two functions:

′ ′ ′
If h(x) = f(x)g(x), then h (x) = f (x)g(x) + g (x)f(x).

Example: h(x) = [x3 − 4x2 ][5 − 20x8 + 6x4 ].

h(x) = f(x)g(x), with f(x) = x3 − 4x2 and g(x) = 5 − 20x8 + 6x4 .

Note: f′ (x) = 3x2 − 8x and g′ (x) = 0 − 160x7 + 24x3 .

So, h′ (x) = [3x2 − 8x][5 − 20x8 + 6x4 ] + [160x7 + 24x3 ][x3 − 4x2 ].
20 / 46
Continuity Derivative

Economic Applications
Marginal Utility Functions:
1 1 1 1
1 −2 ∂2 U −3 1
U = x12 x22 where ∂U
∂x1 = 2 x1 x22 > 0, ∂x21
= − 14 x1 2
x22 < 0, as x1 > 0, x2 > 0.
If a monopolist wishes to increase quantity (q) sold, then it must reduce price (p), so its price is a
function of q (trade-off between q and p).

21 / 46
Continuity Derivative

Economic Applications
Recall: Economists use ‘Marginal Analysis’ to describe such relationships.

Marginalism describes both an economical method of analysis and a theory of value. According to this theory, individuals
make economic decisions on the margin.

Monopolistic Firm’s problem:

Extra (marginal) revenue from one more unit sold = unit price - revenue loss from selling ”all units”, q, at lower price.
An optimal level of production could be MR=MC.
MR > MC, produce and sell more, as the extra revenue > the extra cost.
MR < MC, producing and selling more reduces profit, as the extra revenue < the extra cost.

In Maths,

For MR(q):

MR(q) from Differentiating TR(q) with respect to q.

TR(q) = p(q)q, the product of functions p(q) and q.
MR(q) = dTR(q)/dq = p′ (q)q + p(q) [dq/dq] = p′ (q)q + p(q).
| {z }
=1
In particular, if p(q) = 40 − 2q we have p′ (q) = −2, then
MR(q) = −2q + (40 − 2q) = 40 − 4q.
To check, expand TR(q) = [40 − 2q]q = 40q − 2q 2 .
We see TR′ (q) = MR(q) = 40 − 4q.
22 / 46
Continuity Derivative

Economic Applications

For MC(q):

If C(q) = aq 2 + bq + c, then C ′ (q) = M C(q) = 2aq + b.

For the optimal production level q ∗ :

Conditional Equation: M R(q) = M C(q),

Conditional Equations are found in Economic Models where optimization based on

certain conditions that are either derived intrinsically from the Economic Models or
directly applied based on existing theories.

M R(q) = M C(q) =⇒ 40 − 4q = 2aq + b =⇒ (2a + 4)q = 40 − b =⇒

q∗ = 2a+4 .
40−b

For q ∗ to be Economically Meaningful =⇒ q ∗ has to exist first, and q ∗ ≥ 0 =⇒

2a + 4 ̸= 0, and 2a + 4 and 40 − b have the same sign.

23 / 46
Continuity Derivative

Black-Scholes/Merton Option Pricing

The most famous equation in finance, the Black-Scholes/Merton equation launched an industry (Chicago Board Options
Exchange (CBOE) https://www.cboe.com/us/options/overview/) worth trillions and led to the world’s best investments.
∂2 V
∂V
+ rS ∂V + 1
σ2 S 2 − rV = 0, where the financial instruments’ valuation V = f (S, t, . . .), r is the market
∂t ∂S 2 ∂S 2
risk-free rate, S is the asset’s spot price (stock price), t is time to maturity, and σ is asset’s volatility (standard deviation).
24 / 46
Continuity Derivative

25 / 46
Continuity Derivative

Black-Scholes/Merton
BSM:
The Black-Scholes-Merton (BSM) equation is a pricing model for financial instruments, which is used to determine the fair
prices of financial instruments such as exchange traded options, over-the-counter derivative securities, securitized debt,
credit default swaps, etc., based on volatility, asset type, underlying stock price, strike price, time, and risk-free rate.

Black-Scholes posits that financial instruments, such as stock shares or futures contracts, will have a lognormal
distribution of prices following a random walk with constant drift and volatility.

The Black-Scholes equation requires five variables: volatility, underlying asset’s price, option’s strike price, time to the
option’s expiry, and the risk-free interest rate, which yields theoretically rational prices for financial instruments.

BSM Benefits:

Framework: provides a theoretical framework for pricing options, which help investors and traders determine the fair price
of an option or financial instruments by a tried and tested methodology.

Portfolio Optimization: optimize portfolios by providing a measure of the expected returns and risks associated with
different options, allowing investors to align investment with their risk-return preferences.

Risk Management: investors use a fair price to manage risk exposures to different assets by understanding portfolio’s
weakness and deficiencies.

Enhances Market Efficiency: improve market efficiency as traders are able to better price and trade options, which was
difficult before especially for different underlying assets in a financial instrument.

In summary, BSM is widely accepted in the financial industry, which allows greater consistency and comparability across different
markets and jurisdictions.
BSM Limitations: Not all types of options (American), lack cash flow flexibility based on the future projections of a security,
inaccurate assumptions on constant future volatility.
26 / 46
Continuity Derivative

Medallion Fund
Each fund is essentially 1 or many investment strategies. Prof. Andrew Lo (MIT): Possible to predict and beat the financial
market with the right model and computations. Medallion represents the ultimate challenge to the efficient market
hypothesis. Buffett’s Annualized Return 19.8%.

27 / 46
Continuity Derivative

Long-Term Capital Management

Long-Term Capital Management (LTCM) was a highly-leveraged massive hedge fund including Nobel laureates, and yielded ultra
returns for three years, then suddenly collapsed. In 1998, it received a $3.6 billion bailout from a group of 14 banks, in a deal
brokered by the Federal Reserve Bank of New York. Tragic : assets’ correlations in market crashes approach 1.

Portfolio volatility (standard deviation) 10% reduced to 1% by diversifying into 100 uncorrelated assets.
28 / 46
Continuity Derivative

Derivative of the quotient of two functions

f(x) ′
f′ (x)g(x) − f(x)g′ (x)
If h(x) = , then h (x) = .
g(x) [g(x)]2

[6x5 +2x3 ]
Consider h(x) = ,
[3x2 −5x]

h(x) can be written as h(x) = f(x)/g(x) where f(x) = 6x5 + 2x3 and g(x) = 3x2 − 5x, which have
derivatives f′ (x) = 30x4 + 6x2 and g′ (x) = 6x − 5.

Thus,
′
[30x4 + 6x2 ][3x2 − 5x] − [6x5 + 2x3 ][3x2 − 5x]
h (x) = .
[3x2 − 5x]2

29 / 46
Continuity Derivative

Derivative of the quotient of two functions

If y = f(u) and u = g(x), so that y = f(g(x)) = h(x), then

′ ′ ′
dy dy du
h (x) = f (u)g (x) or = .
dx du dx

Consider the function y = (12x5 + 3x2 )12 .

It would be tedious to expand the function (i.e., multiply (12x5 + 3x2 ) by itself 12 times!

The chain rule allows for a simple way to find the derivative of this function.

Note, we can write y = u12 where u = g(x) = 12x5 + 3x2 .

So,
dy dy du 11 4
= = (12u )(60x + 6x)
dx du dx
or
dy 5 2 11 4
= 12(12x + 3x ) (60x + 6x)
dx

30 / 46
Continuity Derivative

Derivative of the inverse of a function

If y = f(x) has the inverse function x = g(y), that is, if g(y) = f−1 (y) and f′ (x) ̸= 0, then

dx 1 ′
1
= or g (y) = where y = f(x)
dy dy/dx f′ (x)

The linear function y = cx provides a simple illustration of this rule. Since x = (1/c)y is the inverse function, it is
clear that dx/dy = 1/c or 1/(dy/dx).

This rule is especially useful when one cannot (easily) find the inverse of a function, such as y = x3 − 6x9 .
dy
In this case, with = 3x2 − 54x8 , we can simply note that dx
= 1
= 1
.
dx dy [dy/dx] [3x2 −54x8 ]

Economic Applications:
Relationship between production function and cost function:

If q(L) is the production function where L is Labor, L(q) is it’s inverse.

If c0 is fixed cost, w is unit cost of labor (L), then cost function is C(q) = c0 + wL(q) = c0 + 2wq.

Thus marginal cost for production:

dC(q) dL(q)
= w = 2w.
dq dq

By the inverse function rule for differentiation:

dC(q) 1 1 1 dL(q)
= w = w = w = 2w = w .
1 1
dq dq/dL(q) dL(q) dq
2
dq

31 / 46
Continuity Derivative

Derivative of the exponential function

If y = ex , then dy/dx = ex .

The derivative of the function y = ex is the same as the value of the function itself (see below).

For most economics applications, the exponential function is with an exponent that itself is a
function of x; i.e., f(x) = eg(x) .

Use the chain rule so that f′ (x) = [eg(x) ]g′ (x) or g′ (x)eg(x) .

Example:
f(x) = e−rx , f′ (x) = −re−rx .

32 / 46
Continuity Derivative

Derivative of the logarithmic function

dy
If y = lnx, then dx = x.
1

The slope of the function lnx gets smaller (but remains positive) as x gets larger.
As x → ∞ the slope (dy/dx = 1/x) approaches zero (i.e., the function becomes ‘flat’ in the
limit).
As x → 0+ (i.e., as x approaches zero from the right - through positive values) the slope
approaches positive infinity (i.e., the function becomes ‘vertical’ in the limit).

Example:
f(x) = ln[3x2 − 2x], then f′ (x) = (1/[3x2 − 2x])(6x − 2) = [6x − 2]/[3x2 − 2x]. 33 / 46
Continuity Derivative

L’Hôpital’s Rule
L’Hôpital’s Rule:
f(x) f′ (x)
lim = lim ′
x→a g(x) x→a g (x)

This rule is really an application for differentiation rather than a rule for finding derivatives.
It is useful when the functions f(x) and g(x) are such that the ratio f(x)/g(x) isn’t defined at a
point x = a.
However, that doesn’t mean the limit of the ratio doesn’t exist as x → a, and this rule allows one
to find this limit when it exists.
Example:
the ratio [x2 − 1]/[x − 1] isn’t defined at x = 1 (as it is 0/0).

However, in the limit x → 1, the ratio has a value of 2 according to

L’Hôpital’s Rule:

x2 − 1 f′ (x) 2x 2
lim = lim ′ = lim = = 2.
x→1 x−1 x→1 g (x) x→1 1 1

In this example, the result can also be found by factoring the expression, with

2 [x2 − 1] (x − 1)(x + 1)
x −1 = (x−1)(x+1), so that = = x+1, for x ̸= 1.
[x − 1] x−1 34 / 46
Continuity Derivative

Higher-Order Derivatives: Concavity and Convexity

A derivative of a function is also a function (e.g., for f(x) = x2 we have f(x) = 2x, a function of
x).
So we can find the derivative of the first derivative, which is called the second-order derivative.
Since the second-order derivative will also be a function of x, we can find its derivative, which we
call the third-order derivative, and so forth to find the nth-order derivative of y = f(x), noting it
by f(n) (x), for n = 1, 2, 3, 4, . . .. or dn y/dxn .
Sometimes one uses successive derivative signs (e.g., f ′′′ (x) to denote the third derivative of
f(x)).
Finding higher order derivatives is no different than finding the first derivative of a function.
One simply differentiates derivatives successively to find successively higher order derivatives.
Eg., for f(x) = x5 , we have f′ (x) = 5x4 , f′′ (x) = 20x3 , f′′′ (x) = 60x2 , f(4) (x) = 120x,
f(5) (x) = 120, f(6) (x) = 0.
Note that for this function all derivatives of order higher than 5 take on the value zero; that is,
f(n) (x) = 0 for n = 6, 7, 8,. . ., or change in x no longer result in change in y .

d6 y 6 6
= 0 =⇒ d y = 0 × dx ,
dx6
where dx6 (change in 5 five more such changes in x) results in no change in d6 y.
35 / 46
Continuity Derivative

Importance of the sign of the second derivative

Whether the second derivative of a function is positive-valued or negative-valued tells us

something important about the function.

If a function’s second derivative is positive at all points of its domain, then the function is convex.

If the second derivative is negative at all points of its domain, then the function is concave.

A positive second derivative for a function f(x), i.e., a convex function, means the slope of the
function (i.e., its first derivative) is getting higher as x gets larger.

A negative second derivative for a function f(x), i.e., a concave function, means the slope of the
function (i.e., its first derivative) is getting smaller as x gets larger.

36 / 46
Continuity Derivative

Convex Function

Convex Function:

A twice differentiable function f(x) is convex if, at all points on its domain, f′′ (x) ≥ 0.

A twice differentiable function f (x) is strictly convex if f′′ (x) > 0, except possibly at a single
point.

37 / 46
Continuity Derivative

Concave Function
Concave Function:

A twice differentiable function f(x) is concave if f′′ (x) ≤ 0 on all points of its domain.

A twice differentiable function f (x) is strictly concave if f′′ (x) < 0 on all points of its domain
except possibly at a single point.

which is strictly concave, and so also concave.

38 / 46
Continuity Derivative

Convex Function and Concave Function

f′′ (x) ≥ 0 for convexity and f′′ (x) ≤ 0 for concavity means that a linear function (f′′ (x) = 0) is
both concave and convex.
Functions which are strictly convex or strictly concave cannot be linear or have linear segments in
them.
Functions may be convex or strictly convex in some regions, but concave or strictly concave in
other regions of their domains.
Example:
f(x) = −(1/3)x3 + 3x2 − 5x + 10 has second derivative f′′ (x) = −2x + 6.
So f′′ (x) > 0 if x = 0 and 3, and f′′ (x) < 0 if x > 3.

39 / 46
Continuity Derivative

Taylor Series Expansion

Recall that the differential dy = f′ (x)dx can be used to approximate the change in the y variable
(dy ≈ ∆y) for a given change in the x variable (dx = ∆x):

Earlier, the approximation of dy for ∆y can be ‘as accurate as one wishes’ provided one is willing
to consider sufficiently small changes in the x variable (i.e., ‘small’ ∆x).
The Taylor Series Formula allows us to explore this relationship further, and consider changes in
the x variable that are not necessarily ‘small’.
Moreover, the Taylor Series Formula allows us to better understand the second order conditions of
optimization problems - a very important tool in Economics (Diminishing Marginal Utility).
The Taylor Series Formula are widely used in economic models.
40 / 46
Continuity Derivative

Taylor Series Expansion Example

The example shows that using the Taylor Series Formula is not necessarily so complicated.

Moreover, as in this case as n becomes ‘large enough’, the remainder term gets ‘small’ and so not
knowing exactly what value of x to compute f(n) (x) in the remainder term isn’t necessarily
problematic.

41 / 46
Continuity Derivative

Taylor Series Expansion Example

Many important economics applications require considering only the first two terms of the formula
(i.e., choose n = 2):

′ f′′ (ξ)(x1 − x0 )2
f(x1 ) = f(x0 ) + f (x0 )(x1 − x0 ) + .
2
Rearranging this formula gives:

f′′ (ξ)(x1 − x0 )2
∆y = dy + .
2
42 / 46
Continuity Derivative

Taylor Series Expansion Example

So, if f′′ (x) < 0 for all x (i.e., the function is strictly concave), then using dy as estimate of the
actual change in y or ∆y will be an overestimate.

43 / 46
Continuity Derivative

Taylor Series Expansion Example

If f′′ (x) > 0 for all x (i.e., the function is strictly convex), then using dy as estimate of the actual
change in y or ∆y will be an underestimate.

44 / 46
Continuity Derivative

Taylor Series Expansion Applications

Taylor series expansion can be a useful tool in economics, especially for functions that are not
differentiable.

In economics, many real-world problems involve non-smooth, non-differentiable functions, and

Taylor series can provide approximate solutions or insights in such cases.

Approximation for Non-Differentiable Functions:

When dealing with non-differentiable functions, such as those involving discontinuities or kinks
(e.g., utility functions with corner solutions), you can use Taylor series to approximate these
functions around specific known points x0 . By taking a low-order Taylor expansion, you might
obtain a locally valid approximation for these functions.

Optimization and Numerical Methods:

In economics, optimization problems are common. Even if the objective function is not
differentiable, you can use Taylor series to approximate the function near a critical point as initial
values (e.g., a maximum or minimum). This can be helpful in numerical optimization algorithms
to find solutions to complex nonlinear economic models.

45 / 46
Continuity Derivative

Taylor Series Expansion Applications

Dynamic Economic Models:
In dynamic economic models, the behavior of agents or economic variables is often governed by difference or differential
equations. In many cases, these equations may not have closed-form solutions. Taylor series expansion can be used to
linearize these equations around an equilibrium point to facilitate analysis or computation.

Risk and Uncertainty:

In financial economics, risk management, and decision theory, Taylor series expansion is used to approximate nonlinear
utility functions. This is valuable when modeling risk aversion, decision-making under uncertainty, or portfolio optimization.

Heterogeneous Agents and Behavioral Economics:

In cases where agents have varying degrees of rationality or face bounded rationality, Taylor series expansions can be used
to approximate how these agents respond to changes in economic conditions, which can be valuable in behavioral
economics.

Note:

Taylor series approximations have limitations as they are local approximations that work well only near the point around
which you are expanding.

The accuracy of the approximation diminishes as you move away from that point. Additionally, the quality of the
approximation depends on the order of the expansion and the behavior of the function.

For non-differentiable functions, economists often rely on numerical methods, simulations, or other advanced techniques to
address complex problems.

Nevertheless, Taylor series expansions can provide valuable insights, and serve as a first step in understanding the behavior
of economic models when dealing with non-differentiable functions.

46 / 46