Introduction

Generating Function
Characteristic function
Characteristic Function
Annexe

Chapter II: The probability generating function and

the characteristic function of a real random
variable

4th DS

A.Y: 2025-2026

1 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Plan

1 Introduction

2 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Plan

1 Introduction
2 Generating Function
Definition and Properties
Generating functions of usual discrete laws
Generating functions and moments

2 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Plan

1 Introduction
2 Generating Function
Definition and Properties
Generating functions of usual discrete laws
Generating functions and moments
3 Characteristic function

2 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Plan

1 Introduction
2 Generating Function
Definition and Properties
Generating functions of usual discrete laws
Generating functions and moments
3 Characteristic function
4 Characteristic Function
Definition and Properties
Characteristic Function of Common Continuous Laws
Characteristic Function and Higher-Order Moments

2 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Plan

1 Introduction
2 Generating Function
Definition and Properties
Generating functions of usual discrete laws
Generating functions and moments
3 Characteristic function
4 Characteristic Function
Definition and Properties
Characteristic Function of Common Continuous Laws
Characteristic Function and Higher-Order Moments
5 Annexe

2 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Motivation
The generating and characteristic functions are powerful tools for
describing and studying the probability laws of a random variable:
Generating function: similar to a polynomial, its coefficients are
the probabilities of a discrete random variable.
Characteristic function: always exists (discrete or continuous) and
allows summarizing a distribution, computing moments, studying
sums, and convergence to the normal law.

1 Finance: risk management, pricing financial products.

2 Insurance: estimating claims and calculating premiums.
3 Artificial Intelligence: simulating and comparing distributions to
improve models.
These tools connect probability and analysis to efficiently model,
simulate, and interpret random phenomena.
3 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating Function
Definition
Let X be a random variable taking values in N. The generating function
of X is defined by:
X
GX (z) = P(X = n)z n , −1 ≤ z ≤ 1
n∈X (Ω)

We immediately verify that the generating function depends only on the

law of X , and we can also write GX (z) = E(z X ).
Remark
The generating function of a random variable with finite support X (Ω) is
a polynomial of degree m = max X (Ω):
m
X
GX (z) = P(X = k) z k
k=0
4 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Properties of Generating Functions

Properties
GX (0) = P(X = 0)
GX (1) = 1
Example
Consider a random variable X taking values in N whose generating
function GX is:

∀z ∈ [−1, 1], GX (z) = β(1 + z)3

where β ∈ R.
1 Determine the value of β.
2 Calculate P(X = 0).
3 Write GX (z) as a polynomial.
4 Deduce the law of X .
5 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Solution

1
1 GX (1) = 1 ⇒ β(1 + 1)3 = 8β = 1 ⇒ β = 8
1
2 P(X = 0) = GX (0) = β = 8
3 GX (z) = 18 (1 + z)3 = 1
8 + 3
8z + 38 z 2 + 18 z 3
4 X (Ω) = {0, 1, 2, 3} with probabilities:

1 3 3 1
P(X = 0) = , P(X = 1) = , P(X = 2) = , P(X = 3) =
8 8 8 8

6 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Derivatives of Generating Functions

Definition
Let X be a random variable taking values in N. The generating function
GX (z) is of class C ∞ on [−1, 1] and its n-th derivative is:
+∞ +∞
(n)
X X k!
GX (z) = k(k−1) . . . (k−n+1) P(X = k) z k−n = P(X = k) z k−
(k − n)!
k=n k=n

Example
Let X with X (Ω) = {0, 1, 2} and generating function GX (z) such that:

3 1
GX0 (1) = , GX00 (1) = .
4 2

1
1 Show P(X = 1) = P(X = 2) = 4
2 Deduce GX (z)
7 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Solution

Since X (Ω) = {0, 1, 2}:

GX (z) = P(X = 0) + P(X = 1)z + P(X = 2)z 2

GX0 (z) = P(X = 1) + 2P(X = 2)z, GX00 (z) = 2P(X = 2)

At z = 1:
(
P(X = 1) + 2P(X = 2) = 3/4
⇒ P(X = 1) = P(X = 2) = 1/4
2P(X = 2) = 1/2

P(X = 0) = 1 − (1/4 + 1/4) = 1/2

8 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating Function

The generating function is:

1 1 1
GX (z) = + z + z2
2 4 4

9 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Characterization

Proposition
The generating function characterizes the law of a discrete random
variable X :
1 (n)
P(X = n) = GX (0), ∀n ∈ X (Ω)
n!

Remark
(n)
If GX (0) = 0 for all n > N, then X (Ω) is finite:

X (Ω) ⊂ {0, 1, . . . , N}

10 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Theorem: Characterization of the law

Let X and Y be two r.v.’s taking values in N, then:
L
X = Y ⇔ GX (z) = GY (z), ∀ − 1 ≤ z ≤ 1.

L
with X = Y i.e. X has the same distribution as Y .

Proposition: Sum of independent variables

Let X and Y be two independent r.v.’s taking values in N. The
generating function of X + Y is the product of generating functions:

GX +Y (z) = GX (z) × GY (z), −1 ≤ z ≤ 1

Remark
The converse of the previous proposition is false. That is:

∀z ∈ [−1, 1], GX +Y (z) = GX (z) × GY (z) ; X and Y are independent

11 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating function of the Bernoulli law

Let X be a r.v. following the Bernoulli law with parameter p, then its
generating function is:

GX (z) = (1 − p) + pz

Proof : Let X be a r.v. following the Bernoulli law with parameter p,

then:
X
GX (z) = P(X = n)z n
n∈X (Ω)

= P(X = 0)z 0 + P(X = 1)z 1

= (1 − p) + pz, −1 ≤ z ≤ 1

12 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating function of the Binomial law

Let X be a r.v. following a binomial law with parameters n and p, then

its generating function is:

n
GX (z) = (1 − p) + pz .

Proof : Let X be a r.v. following a binomial law with parameters n, p, then:

X X
GX (z) = P(X = k)z k = P(X = k)z k
k∈X (Ω) k∈{0,...,n}
n
X
= Cnk p k (1 − p)n−k z k
k=0
Xn
= Cnk (pz)k (1 − p)n−k .
k=0

13 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating function of the Binomial law

Using Newton’s binomial formula:

n
X
Cnk (pz)k (1 − p)n−k = ((1 − p) + pz)n
k=0

We get:

GX (z) = ((1 − p) + pz)n , −1 ≤ z ≤ 1.

14 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating function of the geometric law

Proof : Let X be a r.v. following the geometric law with parameter
p ∈]0, 1[, then:
X
GX (z) = P(X = n)z n
n∈X (Ω)
X
= P(X = n)z n
n∈N∗
+∞
X
= p(1 − p)n−1 z n
n=1
+∞
p X
= (z(1 − p))n
1 − p n=1
+∞
X
Using the fact that (z(1 − p))n is a geometric series with ratio z(1 − p) ∈] − 1, 1[,
n=1
it converges and its sum is:
+∞
X 1
(z(1 − p))n = z(1 − p)
n=1
1 − z(1 − p)
We get:
pz 15 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating function of the Poisson law

Let X be a r.v. following a Poisson law with parameter λ > 0 :

X ∼ P(λ)

Its generating function is:

GX (z) = exp λ(z − 1) , ∀z ∈ R.

16 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating function of the Poisson law

Proof : Let X be a r.v. following a Poisson law with parameter λ > 0,

then:
+∞ +∞
X X X λn −λ n
GX (z) = P(X = n)z n = P(X = n)z n = e z
n=0 n=0
n!
n∈X (Ω)
+∞
X (λz)n
= e −λ
n=0
n!
P+∞ (λz)n
Using the fact that n=0 n! is a convergent exponential series whose
sum is e λz , we get:

GX (z) = e −λ e λz = e λ(z−1) , −1 ≤ z ≤ 1.

17 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Exercise

1 Find the generating function of the binomial law B(n, p), with
n ∈ N∗ and p ∈ ]0, 1[.
2 Let X and Y be two independent r.v.’s taking values in N. Let
S = X + Y . Determine the distribution of S using two methods in
the following cases:
1 X ∼ B(n, p) and Y ∼ B(m, p), with n, m ∈ N∗ and p ∈ ]0, 1[.
2 X ∼ P(λ) and Y ∼ P(µ), with λ, µ > 0.

18 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Solution:

Ê Let X be a r.v. following a binomial law with parameters n and p.

Then, X can be written as the sum of n independent r.v.’s X1 , . . . , Xn
each following the Bernoulli law with parameter p. Hence:

GX (z) = GX1 +X2 +···+Xn (z) = GX1 (z) × · · · × GXn (z)

= ((1 − p) + pz) × · · · × ((1 − p) + pz) n times
= ((1 − p) + pz)n , −1 ≤ z ≤ 1

19 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Ë Let X and Y be two independent r.v.’s taking values in N, and let

S = X +Y.

We want to determine the distribution of S in two particular cases using

two distinct methods:
Ì X can be seen as the sum of n independent r.v.’s X1 , . . . , Xn each
following Bernoulli(p). Similarly, Y can be seen as the sum of m
independent r.v.’s Y1 , . . . , Ym each following Bernoulli(p). Moreover, we
L
can say (Y1 , . . . , Ym ) = (Xn+1 , . . . , Xn+m ).
Since X and Y are independent, the r.v.’s X1 , . . . , Xn , Xn+1 , . . . , Xn+m
are independent and all follow Bernoulli(p).
Hence:
n+m
X
S =X +Y = Xi
i=1

Which shows that by definition of a Bernoulli law S ∼ B(n + m, p).

20 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Let X ∼ B(n, p) and Y ∼ B(m, p) be two independent r.v.’s. Using

generating functions:

GX (z) = (1 − p + pz)n
GY (z) = (1 − p + pz)m

Since X and Y are independent:

GX +Y (z) = GX (z)·GY (z) = (1−p +pz)n ·(1−p +pz)m = (1−p +pz)n+m

Therefore:
X + Y ∼ B(n + m, p)

21 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Í Since X and Y take values in N, then S(Ω) = N. To determine the

distribution of S, we compute P({S = n}), n ∈ N. Then:

P({S = n}) = P({X + Y = n}), n ∈ N

For k ∈ Y (Ω) = N, define the event Ak = {Y = k}, then:

∀k 6= k 0 , Ak ∩ Ak 0 = ∅ and ∪k∈N Ak = Y (Ω) = N

so (Ak )k∈N is a complete system of N. Then:

P({S = n}) = P({X + Y = n})
FPT X
= P({X + Y = n}/{Y = k})P({Y = k})
k≥0
n
X
= P({X = n − k}/{Y = k})P({Y = k})
k=0
Xn
= P({X = n − k})P({Y = k})
k=0

22 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

We have:
λn−k −λ
X ∼ P(λ) ⇒ P({X = n − k}) = e
(n − k)!
µk −µ
Y ∼ P(µ) ⇒ P({Y = k}) = e
k!

n
X λn−k −λ µk −µ
P({S = n}) = e e
(n − k)! k!
k=0
n
e −(λ+µ) X n!
= λn−k µk
n! k!(n − k)!
k=0
−(λ+µ) n
e X
= Cnk λn−k µk
n!
k=0

23 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Using the binomial formula:

n
X
Cnk λn−k µk = (λ + µ)n
k=0

We get:

(λ + µ)n −(λ+µ)
∀ n ∈ N, P({S = n}) = e .
n!
Therefore, we deduce that S ∼ P(λ + µ).

24 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Î
Let X ∼ P(λ) and Y ∼ P(µ) be two independent r.v.’s, then:

GX +Y (z) = GX (z) × GY (z) = e λ(z−1) e µ(z−1) = e (λ+µ)(z−1)

In particular, we recover that the sum of two independent r.v.’s following

Poisson laws is Poisson.

25 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Generating functions and moments

Theorem
Let X be a r.v. taking values in N. X admits a p-th order moment,
p ∈ N∗ , if and only if GX is p-times differentiable at 1, and:
 
(p)
GX (1) = E X (X − 1) × · · · × (X − p + 1)

Corollary
Let X be a r.v. taking values in N which admits a second-order moment,
then:
0
E(X ) = GX (1)
00 0
 0 2
V(X ) = GX (1) + GX (1) − GX (1)

26 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Application :

Calculate the third-order moment of a r.v X ∼ P(λ), λ > 0.

Solution:
Knowing that the generating function of X is
GX (z) = e λ(z−1) , −1 ≤ z ≤ 1, we get:
0 0
GX (z) = λe λ(z−1) so GX (1) = E(X ) = λ
00 00
GX (z) = λ2 e λ(z−1) so GX (1) = E(X (X − 1)) = λ2
000 000
GX (z) = λ3 e λ(z−1) so GX (1) = E(X (X − 1)(X − 2)) = λ3

This allows us to calculate:

E(X 2 ) = E(X (X − 1)) + E(X ) = λ2 + λ

Therefore:
 
E(X 3 ) = E X (X − 1)(X − 2) + 3E(X 2 ) − 2E(X ) = λ3 + 3λ2 + λ

27 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Exercise :

A merchant estimates that the demand of a seasonal product is a r.v. X

with probability law:

pk
pX (k) = P(X = k) = , ∀k ∈ N, p ∈]0, 1[
(1 + p)k+1

1 Verify that pX is a probability law.

2 Determine its generating function.
3 Find the expectation E(X ) and variance V(X ) of X .

28 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

Solution:
To ensure pX is a probability law, we need:
+∞
X
pX (k) = 1
k=0
We have:
+∞ +∞ +∞
X X pk 1 X p k
pX (k) = k+1
= ( )
(1 + p) 1+p 1+p
k=0 k=0 k=0
P+∞ p k
Using the fact that k=0 ( 1+p )is a geometric series with ratio
p
1+p ∈] − 1, 1[, it converges and we get:
+∞
X p k 1
( ) = p =p+1
1+p 1 − p+1
k=0
Hence:
+∞
X 1
pX (k) = × (p + 1) = 1
1+p
k=0
29 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

+∞
X pk
GX (z) = E(z X ) = zk ·
(1 + p)k+1
k=0
+∞  k
1 X pz
=
1+p 1+p
k=0
1 1
= · pz
1 + p 1 − 1+p
1
=
1 + p − pz
Hence:
1
GX (z) =
1 + p − pz

30 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Generating functions of usual discrete laws
Characteristic Function Generating functions and moments
Annexe

The generating function GX (z) is differentiable:

 0
1 p p
GX0 (z) = = 2
→ GX0 (1) =
1 + p(1 − z) (1 + p(1 − z)) (1)2
00
2p 2 2p 2

p
GX00 (z) = = → GX
00
(1) = = 2p 2
(1 + p(1 − z))2 (1 + p(1 − z))3 (1)3
This allows us to calculate:
p
E(X ) = GX0 (1) = = p
(1)2

and

V(X ) = GX00 (1) + GX0 (1) − (GX0 (1))2 = 2p 2 + p − p 2 = p(p + 1)

31 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function
Definition
Let X be a real r.v. The complex function defined by:

ΦX :R → C
t 7→ ΦX (t) = E(e itX )

is called the characteristic function of the r.v. X .

32 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function
Definition
Let X be a real r.v. The complex function defined by:

ΦX :R → C
t 7→ ΦX (t) = E(e itX )

is called the characteristic function of the r.v. X .

Consequences
1 If X is a discrete r.v. taking values in X (Ω), then:
X
∀t ∈ R, ΦX (t) = e itn P(X = n)
n∈X (Ω)

2 If X is a continuous r.v. with density f , then:

Z
∀t ∈ R, ΦX (t) = e itx f (x) dx
R
32 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Properties

Properties
Let X be a continuous real r.v., then its characteristic function ΦX
satisfies:

33 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Properties

Properties
Let X be a continuous real r.v., then its characteristic function ΦX
satisfies:
1 ΦX (0) = 1 .

33 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Properties

Properties
Let X be a continuous real r.v., then its characteristic function ΦX
satisfies:
1 ΦX (0) = 1 .
2 ∀t ∈ R, |ΦX (t)| ≤ 1 .

33 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Properties

Properties
Let X be a continuous real r.v., then its characteristic function ΦX
satisfies:
1 ΦX (0) = 1 .
2 ∀t ∈ R, |ΦX (t)| ≤ 1 .
3 ΦX is continuous.

33 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Properties

Properties
Let X be a continuous real r.v., then its characteristic function ΦX
satisfies:
1 ΦX (0) = 1 .
2 ∀t ∈ R, |ΦX (t)| ≤ 1 .
3 ΦX is continuous.
4 ΦX (−t) = ΦX (t), ∀t ∈ R .

33 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Properties

Properties
Let X be a continuous real r.v., then its characteristic function ΦX
satisfies:
1 ΦX (0) = 1 .
2 ∀t ∈ R, |ΦX (t)| ≤ 1 .
3 ΦX is continuous.
4 ΦX (−t) = ΦX (t), ∀t ∈ R .
5 If moreover X is symmetric (i.e. X and −X have the same
distribution), then ΦX is real.

33 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proof
1 ΦX (0) = E(e i0X ) = E(1) = 1

34 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proof
1 ΦX (0) = E(e i0X ) = E(1) = 1
2 |ΦX (t)| = |E(e itX )| ≤ E(|e itX |) = E(1) = 1

34 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proof
1 ΦX (0) = E(e i0X ) = E(1) = 1
2 |ΦX (t)| = |E(e itX )| ≤ E(|e itX |) = E(1) = 1
3 Let t0 ∈ R. For all t ∈ R, we have |e itX | = 1. When t → t0 ,
e itX → e it0 X almost surely. Since |e itX | ≤ 1 for all t, we can apply
the dominated convergence theorem, giving
 
lim ΦX (t) = lim E(e itX ) = E lim e itX = E(e it0 X ) = ΦX (t0 ).
t→t0 t→t0 t→t0

Hence, ΦX is continuous at t0 . Since t0 is arbitrary, ΦX is

continuous on R.

34 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proof
1 ΦX (0) = E(e i0X ) = E(1) = 1
2 |ΦX (t)| = |E(e itX )| ≤ E(|e itX |) = E(1) = 1
3 Let t0 ∈ R. For all t ∈ R, we have |e itX | = 1. When t → t0 ,
e itX → e it0 X almost surely. Since |e itX | ≤ 1 for all t, we can apply
the dominated convergence theorem, giving
 
lim ΦX (t) = lim E(e itX ) = E lim e itX = E(e it0 X ) = ΦX (t0 ).
t→t0 t→t0 t→t0

Hence, ΦX is continuous at t0 . Since t0 is arbitrary, ΦX is

continuous on R.
4 ΦX (−t) = E(e −itX ) = E(e itX ) = ΦX (t)

34 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proof
1 ΦX (0) = E(e i0X ) = E(1) = 1
2 |ΦX (t)| = |E(e itX )| ≤ E(|e itX |) = E(1) = 1
3 Let t0 ∈ R. For all t ∈ R, we have |e itX | = 1. When t → t0 ,
e itX → e it0 X almost surely. Since |e itX | ≤ 1 for all t, we can apply
the dominated convergence theorem, giving
 
lim ΦX (t) = lim E(e itX ) = E lim e itX = E(e it0 X ) = ΦX (t0 ).
t→t0 t→t0 t→t0

Hence, ΦX is continuous at t0 . Since t0 is arbitrary, ΦX is

continuous on R.
4 ΦX (−t) = E(e −itX ) = E(e itX ) = ΦX (t)
5 If X is symmetric, then X and −X have the same law, thus the
same expectation:
ΦX (t) = E(e itX ) = E(e it(−X ) ) = E(e −itX ) = ΦX (−t)
Moreover, ΦX (−t) = ΦX (t) implies ΦX (t) = ΦX (t), so ΦX (t) is
real.
34 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:

1 Determine the characteristic function of the uniform law U([−1, 1]).

2 Determine the characteristic function of the exponential law E(1).

35 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:

1 Determine the characteristic function of the uniform law U([−1, 1]).

2 Determine the characteristic function of the exponential law E(1).
Solution:

¬ Let X ∼ U([−1, 1]), then ∀t ∈ R:

Z Z 1
1
ΦX (t) = e itx f (x) dx = e itx dx
R 2 −1

For t 6= 0, we have:

1 itx 1 e it − e −it sin(t)

ΦX (t) = [e ]−1 = =
2it 2it t
For t = 0, by continuity of ΦX , ΦX (0) = lim ΦX (t) = 1
t→0

35 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

­ Let X ∼ E(1), then ∀t ∈ R:

Z Z +∞
ΦX (t) = e itx f (x) dx = e itx e −x dx
R 0
Z +∞
1
= e x(−1+it) dx = [e x(−1+it) ]+∞
0
0 −1 + it
1
= ( lim e x(−1+it) − 1)
−1 + it x→+∞
1
= ( lim e −x e itx − 1)
−1 + it x→+∞
1
= , (||e −x e itx || ≤ ||e −x ||)
1 − it

36 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Theorem
The characteristic function uniquely determines the law. In other words,
two r.v.’s have the same law if and only if they have the same
characteristic function.

37 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Theorem
The characteristic function uniquely determines the law. In other words,
two r.v.’s have the same law if and only if they have the same
characteristic function.

Proposition
1 Let (a, b) ∈ R2 and X a real r.v., then:

∀t ∈ R, ΦaX +b (t) = e ibt ΦX (at).

2 Let X and Y be two independent real r.v.’s, then:

∀t ∈ R, ΦX +Y (t) = ΦX (t) × ΦY (t).

37 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Exercise:
Let X be a r.v. following an exponential law with parameter 1.
Let λ > 0 and define Y = λ1 X .
1 Determine the probability law of Y .
2 Compute the characteristic function of Y .

38 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:
Ê To determine the probability law of Y , we calculate its cumulative
distribution function FY .
For y < 0:
1
FY (y ) = P(Y ≤ y ) = P( X ≤ y ) = P(X ≤ λy ) = 0
λ
For y ≥ 0:
1
FY (y ) = P(Y ≤ y ) = P( X ≤ y ) = P(X ≤ λy ) = FX (λy ) = 1 − e −λy
λ
⇒ Y ∼ E(λ)
.

39 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Ë Let t ∈ R, then:
t  λ
ΦY (t) = Φ λ1 X (t) = ΦX =
λ λ − it
We conclude that the characteristic function of the law E(λ), λ > 0 is:

λ
ΦY (t) =
λ − it

40 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition
Let X be a continuous r.v. taking values in R with characteristic
function ΦX . If ΦX is integrable, i.e.:
Z
|ΦX (t)| dt < +∞
R

then X admits a probability density f on R defined by:

Z
1
∀x ∈ R, f (x) = e −itx ΦX (t)dt
2π R

41 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition
Let X be a continuous r.v. taking values in R with characteristic
function ΦX . If ΦX is integrable, i.e.:
Z
|ΦX (t)| dt < +∞
R

then X admits a probability density f on R defined by:

Z
1
∀x ∈ R, f (x) = e −itx ΦX (t)dt
2π R
Application:

Let X be a continuous r.v. whose characteristic function is given by:

∀t ∈ R, ΦX (t) = e −|t| .

Find the probability density of X .

41 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:
Z Z
1 1
∀x ∈ R, f (x) = e −itx ΦX (t) dt = e −itx e −|t| dt
2π R 2π R
1 n 0
Z Z +∞ o
= e (1−ix)t dt + e (−1−ix)t dt
2π −∞ 0
h e (1−ix)t i0 h e (−1−ix)t i+∞
= +
2π (1 − ix) −∞ 2π (−1 − ix) 0
1 1
= −
2π (1 − ix) 2π (−1 − ix)
1 1
= +
2π (1 − ix) 2π (1 + ix)
1
=
π (1 + x 2 )

Thus, X follows the Cauchy distribution with parameter 1 (X ∼ C(1)),

and ΦX = e −|t| is its characteristic function.
42 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function of Common Continuous r.v.’s

Proposition: characteristic function of the uniform law

Let X follow the uniform law U(a, b). Then its characteristic function is:
( itb ita
Z b e −e
itX itx 1 if t 6= 0
φX (t) = E[e ] = e dx = it(b−a)
a b−a 1 if t = 0

43 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function of Common Continuous r.v.’s

Proposition: characteristic function of the uniform law

Let X follow the uniform law U(a, b). Then its characteristic function is:
( itb ita
Z b e −e
itX itx 1 if t 6= 0
φX (t) = E[e ] = e dx = it(b−a)
a b−a 1 if t = 0

Application:
1 Determine the characteristic function of U([0, 1]).
2 Determine the characteristic function of U([−a, a]), for a > 0.

43 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:

1 Let X ∼ U([0, 1]), then:

(
e it −1
it if t 6= 0
ΦX (t) =
1 if t = 0

44 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:

1 Let X ∼ U([0, 1]), then:

(
e it −1
it if t 6= 0
ΦX (t) =
1 if t = 0

2 Let X ∼ U([−a, a]), a > 0. Then:

( iat −iat (
e −e sin(at)
i2at if t 6 = 0 at if t 6= 0
ΦX (t) = =
1 if t = 0 1 if t = 0

44 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition: characteristic function of the exponential law

Let X follow the exponential law E(λ), λ > 0. Its characteristic function
is:
λ
ΦY (t) =
λ − it

45 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition: characteristic function of the exponential law

Let X follow the exponential law E(λ), λ > 0. Its characteristic function
is:
λ
ΦY (t) =
λ − it

Proposition: characteristic function of the Cauchy law

Let X follow the Cauchy law C(x0 , a), x0 ∈ R, a > 0. Then:

φX (t) = e ix0 t e −a|t| , t ∈ R.

For x0 = 0 and a = 1, we recover the standard Cauchy C(0, 1):

φX (t) = e −|t|

45 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:
Let X ∼ C(0, 1). Compute its characteristic function.

46 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:
Let X ∼ C(0, 1). Compute its characteristic function.
Solution:
By definition:

1 ∞ e itx
Z +∞ Z
itx 1
φ(t) = e dx = dx
−∞ π(1 + x 2 ) π −∞ 1 + x 2

46 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:
Let X ∼ C(0, 1). Compute its characteristic function.
Solution:
By definition:

1 ∞ e itx
Z +∞ Z
itx 1
φ(t) = e dx = dx
−∞ π(1 + x 2 ) π −∞ 1 + x 2

Using the hint:

+∞
e itx
Z
dx = πe −|t|
−∞ x2 +1

46 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:
Let X ∼ C(0, 1). Compute its characteristic function.
Solution:
By definition:

1 ∞ e itx
Z +∞ Z
itx 1
φ(t) = e dx = dx
−∞ π(1 + x 2 ) π −∞ 1 + x 2

Using the hint:

+∞
e itx
Z
dx = πe −|t|
−∞ x2 +1
We get:
1
φX (t) = · πe −|t| = e −|t| , t ∈ R.
π

46 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:
Let X ∼ C(0, 1). Compute its characteristic function.
Solution:
By definition:

1 ∞ e itx
Z +∞ Z
itx 1
φ(t) = e dx = dx
−∞ π(1 + x 2 ) π −∞ 1 + x 2

Using the hint:

+∞
e itx
Z
dx = πe −|t|
−∞ x2 +1
We get:
1
φX (t) = · πe −|t| = e −|t| , t ∈ R.
π
Note that the Cauchy characteristic function is not differentiable at 0.
Hence, this distribution has no finite mean or variance.

46 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition: characteristic function of the standard normal law

Let X ∼ N (0, 1). Then:
−t 2
∀t ∈ R, ΦX (t) = e 2 .

47 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition: characteristic function of the standard normal law

Let X ∼ N (0, 1). Then:
−t 2
∀t ∈ R, ΦX (t) = e 2 .

Corollary
Let X ∼ N (m, σ 2 ), m ∈ R, σ > 0. Then:
t 2 σ2
∀t ∈ R, ΦX (t) = e itm e − 2 .

47 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Proposition: characteristic function of the standard normal law

Let X ∼ N (0, 1). Then:
−t 2
∀t ∈ R, ΦX (t) = e 2 .

Corollary
Let X ∼ N (m, σ 2 ), m ∈ R, σ > 0. Then:
t 2 σ2
∀t ∈ R, ΦX (t) = e itm e − 2 .

Indeed:
Expressing X in terms of the standard normal Z :
L
X = σ Z + m,

where Z ∼ N (0, 1).

47 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Consequently, using the result from our course, we verify that:

t 2 σ2
∀t ∈ R, ΦX (t) = Φσ Z +m (t) = e itm ΦZ (σt) = e itm e − 2

48 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Consequently, using the result from our course, we verify that:

t 2 σ2
∀t ∈ R, ΦX (t) = Φσ Z +m (t) = e itm ΦZ (σt) = e itm e − 2

Application:
Let X be a standard normal random variable. Define a new random
variable Y by:
Y = 2X + 3

1 Express the characteristic function φY (t) of Y in terms of φX (t).

2 Compute φY (t) explicitly using
2
φX (t) = E[e itX ] = e −t /2
.

3 From the obtained characteristic function, identify the distribution of

Y (type, mean, variance).

48 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:
1 Using the relation between Y and X :

φY (t) = φ2X +3 (t) = e i3t φX (2t)

49 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:
1 Using the relation between Y and X :

φY (t) = φ2X +3 (t) = e i3t φX (2t)

2 Using φX (t):
2 2
φY (t) = e i3t φX (2t) = e i3t e −(2t) /2
= e i3t−2t

49 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Solution:
1 Using the relation between Y and X :

φY (t) = φ2X +3 (t) = e i3t φX (2t)

2 Using φX (t):
2 2
φY (t) = e i3t φX (2t) = e i3t e −(2t) /2
= e i3t−2t

3 Recall that the characteristic function of a normal r.v.

Z ∼ N (m, σ 2 ) is:
1 2 2
φZ (t) = e imt− 2 σ t .
Comparing with
2
φY (t) = e i3t−2t ,
we deduce that Y ∼ N (m = 3, σ 2 = 4).
49 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function and Higher-Order Moments

Proposition
Let X be a real r.v. admitting a moment of order p ∈ N∗ . Then its
characteristic function ΦX is p-times differentiable and:
1 (p)
E[X p ] = Φ (0)
ip X

50 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function and Higher-Order Moments

Proposition
Let X be a real r.v. admitting a moment of order p ∈ N∗ . Then its
characteristic function ΦX is p-times differentiable and:
1 (p)
E[X p ] = Φ (0)
ip X

Corollary
If X admits a second-order moment, then:

E[X ] = −iΦ0X (0)

V[X ] = −Φ00X (0) + (Φ0X (0))2

50 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Characteristic Function and Higher-Order Moments

Proposition
Let X be a real r.v. admitting a moment of order p ∈ N∗ . Then its
characteristic function ΦX is p-times differentiable and:
1 (p)
E[X p ] = Φ (0)
ip X

Corollary
If X admits a second-order moment, then:

E[X ] = −iΦ0X (0)

V[X ] = −Φ00X (0) + (Φ0X (0))2

Remark
If the characteristic function φX of a r.v. X is not differentiable at 0,
then X does not admit a first-order moment.
50 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:

Let X ∼ E(λ), an exponential r.v. with parameter λ > 0. Compute the

third-order moment.

51 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Application:

Let X ∼ E(λ), an exponential r.v. with parameter λ > 0. Compute the

third-order moment.
Solution:
λ
We know that ΦX (t) = λ−it , t ∈ R. Then:

iλ
Φ0X (t) =
(λ − it)2
−2λ
Φ00X (t) =
(λ − it)3
−i6λ
Φ000
X (t) =
(λ − it)4

Thus,
1 000 6
E[X 3 ] = Φ (0) = 3 .
i3 X λ
51 / 55
 Introduction
Generating Function Definition and Properties
Characteristic function Characteristic Function of Common Continuous Laws
Characteristic Function Characteristic Function and Higher-Order Moments
Annexe

Exercise:
Consider a normal r.v. X ∼ N (2, 4), and define:

1
Z= X −1
2
Assume that Z follows a standard normal distribution.
1 Show that Z is symmetric (i.e., Z and −Z have the same
distribution).
2 Give the characteristic function of Z .
3 Verify that φZ is an even function.
4 Conclude that Z is symmetric.
5 Let φX denote the characteristic function of X .
1 Express φX in terms of φZ .
2 Generalize this result for X ∼ N (m, σ 2 ).

52 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Annexe
Fonction caractéristique des v.a discrètes usuelles
1 Bernoulli : X ∼ Bernoulli(p), p ∈ [0, 1]

φX (t) = (1 − p) + pe it

2 Binomiale : X ∼ Bin(n, p), n ∈ N, p ∈ [0, 1]

φX (t) = (1 − p + pe it )n

3 Géométrique : X ∼ Geom(p), p ∈ (0, 1) (nombre d’échecs avant le

premier succès)
p
φX (t) =
1 − (1 − p)e it
4 Poisson : X ∼ Poisson(λ), λ > 0
it
−1)
φX (t) = e λ(e
53 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Séries entières
1 Série géométrique: ∀x ∈] − 1, 1[, on a:
+∞
1 X
= xn
1−x n=0
+∞
1 0 1 X
( ) = = nx n−1
1−x (1 − x)2 n=1
+∞
1 00 2 X
( ) = 3
= n(n − 1)x n−2
1−x (1 − x) n=2

2 Série exponentielle: ∀x ∈ R, on a:
+∞ n
X x
ex =
n=0
n!

54 / 55
 Introduction
Generating Function
Characteristic function
Characteristic Function
Annexe

Useful Formulas

1 Combination:
Let n, k ∈ N such that k ≤ n, then:
n!
Cnk =
k!(n − k)!
2 Binomial Theorem:
Let a, b ∈ R and n ∈ N, then:
n
X n
X
(a + b)n = Cnk ak b n−k = Cnk b k an−k
k=0 k=0

55 / 55