A Summary of Actuarial Formulae and Tables: Curtin Student Actuarial Society Last Updated: February 20, 2019

A Summary of Actuarial Formulae and Tables
Curtin Student Actuarial Society

Version 2
Last Updated: February 20, 2019
Contents
Introduction 1
Document conventions 3
Disclaimer 3
Acknowledgments 3
A Actuarial Formulae 4
1 Mathematical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Solving Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Gamma Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Bayes’ Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Statistical Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1 Discrete Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Continuous distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Compound Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Truncated Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Relationships between statistical distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Statistical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1 Sample Mean and Variance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Parametric Inference (Normal Model) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 Maximum Likelihood Estimators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.4 Linear Regression Model With Normal Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.5 Analysis of Variance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.6 Generalised Linear Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.7 Bayesian Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.8 Empirical Bayes Credibility – Model 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.9 Empirical Bayes Credibility – Model 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4 Compound Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5 Survival Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.1 Mortality “Laws” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.2 Empirical Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.3 Mortality Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.4 General Markov Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.5 Graduation Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.6 Multiple Decrement Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.7 Population Projection Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6 Annuities and Assurances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.1 Approximations for Non Annual Annuities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.2 Moments of Annuities and Assurances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.3 Premiums and Reserves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.4 Thiele’s Differential Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7 Stochastic Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7.1 Markov “Jump” Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7.2 Brownian Motion and Related Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
7.3 Monte Carlo Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
8 Time Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8.1 Time Series – Time Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8.2 Time Series – Time Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
8.3 Time Series – Box-Jenkins Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
9 Economic Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.1 Utility Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.2 Capital Asset Pricing Model (CAPM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.3 Interest Rate Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
10 Financial Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.1 Price of a Forward or Futures Contract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.2 Binomial Pricing (“Tree” Model) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
10.3 Stochastic Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
10.4 Black-Scholes Formulae for European Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
10.5 Put-Call Parity Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
B Non-Statistical (Mostly Life) Tables 41

Compound Interest Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Population Mortality Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Assured Lives Mortality Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Pensioner Mortality Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
International Actuarial Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Sickness Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Example Pension Scheme Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Sample Time Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
C Statistical tables 46
Standard Normal probabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Standard Normal percentage points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
t percentage points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
χ2 probabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
χ2 percentage points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
F percentage points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Piosson Probabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Binomial Probabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Critical values for the grouping of signs test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Pseudorandom values from U (0, 1) and from N (0, 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
D Reference by particular units 48

ACCT1000 Accounting - The Language of Business . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
ECON1000 Introductory Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
ECON1001 Actuarial Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
ECON2001 Macroeconomics Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
FNCE2000 Introduction to Finance Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
MATH1016 Calculus 1* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MATH1017 Accelerated Mathematics 1* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MATH1018 Accelerated Mathematics 2* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MATH2004 Theory of Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MATH2009 Calculus 2* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MATH3005 Survival Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MATH3006 Life Contingencies 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
MATH3007 Life Contingencies 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
STAT1000 Regression and Non-Parametric Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
STAT1001 Statistical Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
STAT1002 Statistical Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
STAT2001 Mathematical Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
STAT3001 Statistical Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
STAT3002 Risk Analysis and Credibility Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
STAT3005 Stochastic Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
STAT3006 Investment Science 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
STAT3007 Investment Science 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
ACTUARIAL FORMULAE BREAKDOWN Version 2 Curtin Student Actuarial Society
Introduction
This document is designed to provide information on the formulae which are relevant to actuarial science units at Curtin
university. Actuarial students are provided with selected formulae prepared by the UK Institute and Faculty of Actuaries
for use in all examinations that follow their curriculum, which Curtin (and Australia as a whole) adhere to. Students
are typically provided with a pdf document including the required formulae on Blackboard in any units where they are
provided in the exam but this is simply an abridged version a publication which is available for purchase (Curtin students
need not purchase it) known as the “orange book” (for the fact that it has an orange cover). In addition to excerpts
from the orange book, Curtin students are often also provided with a copy of separate publication called ‘Mathematical
Formulae and Statistical Tables for Tertiary Institutions’ in exams as an alternative to the statistical tables that appear
in the orange book. This book is often used in first year statistics units so having a copy and being familiar with it is
recommended. Also note that in spite of the orange book being created in 2002, the actuaries institute are not introducing
a new formula and tables for the new curriculum (at least for now).
These formulae are not provided in all units. The units for which these formulae are provided in the exam are those
for which actuarial exam conditions are adhered to. These conditions include:
• 3 hour exams
• Exams weighted higher than 50% (typically 60% or 70%)
• Actuarial tables and formulae provided
• Students are not permitted to bring any notes of their own into actuarial exams
• Only approved calculators allowed

• Content for these units follows closely with the actuarial curriculum and thus the official ActEd CT notes are often
the prescribed textbook.
The units for which these conditions are adhered to are always CT exemption units, but the converse is not always
true. This is because some of the earlier units are not exclusively part of the actuarial science program. The nature of
each CT exemption unit is displayed in the table below.
1 Updated : February 20, 2019

CT subject Curtin Exemption Unit Actuarial Actuarial Separate statistical Run by

conditions formulae tables
CT1 MATH2004 Theory of Inter- X X X Maths and Stats
est (Actuarial)
CT2 ACCT1000 Accounting - The Business School
Language of Business
CT2 FNCE2000 Introduction to Business School
Finance Principles
CT3 STAT1000 Regression and X Maths and Stats
Non-Parametric Inference
CT3 STAT1001 Statistical Proba- X Maths and Stats
bility
CT3 STAT1002 Statistical Data X Maths and Stats
Analysis
CT3 STAT2001 Mathematical (p) X X Maths and Stats
Statistics
CT4 STAT3005 Stochastic Pro- X X X Maths and Stats
cesses (Actuarial)
CT4 MATH3005 Survival Analysis X X X Maths and Stats
(Actuarial)
CT5 MATH3006 Life Contingen- X X X Maths and Stats
cies 1 (Actuarial)
CT5 MATH3007 Life Contingen- X X X Maths and Stats
cies 2 (Actuarial)
CT6 STAT3001 Statistical Mod- X X X Maths and Stats
elling (Actuarial)
CT6 STAT3002 Risk Analysis and X X X Maths and Stats
Credibility Theory (Actuarial)
CT7 ECON1000 Introductory Business School
Economics
CT7 ECON1001 Actuarial Eco- (p) Maths and Stats
nomics (Actuarial)
CT7 ECON2001 Macroeconomics Business School
Principles
CT8 STAT3006 Investment Sci- X X X Maths and Stats
ence 1 (Actuarial)
CT8 STAT3007 Investment Sci- X X X Maths and Stats
ence 2 (Actuarial)
X Completely satisfied
(p) Partially satisfied (students are sort of eased into it in earlier units).

Document conventions
• In orange boxes, are the formulae and tables which are actually provided in actuarial exams. The color orange was
chosen because these are the formulae that appear in the ‘orange’ book (see Introduction).
• In the green boxes are some extensions made to the formulae and tables you will be provided with that you may
wish to remember. That is to say anything in the green boxes is not in the formula sheet you are given in the exam
but they are formulae you might be expected to draw upon in the exam.
• Red text is used when a formula or table is described instead of being explicitly given in this document.
• Blue text will be used to indicate the actuarial units in which a formula or table is used.
• An asterisk* after a unit name indicates that the actuarial tables and formulae are not provided in the unit but they
are still involved in this document in some way (usually because they are a unit in which a formula is introduced or
explained).
Disclaimer
The validity of the information in this document cannot be guaranteed. Curtin units and the actuarial curriculum change
over time.
This document was typeset in LaTeX in an attempt to make all the formulae appear as similarly as possible to the
ones in the orange book. Subtle and superficial differences may exist.
Should you find any errors in this document please let us know via our website. Link:
https://www.curtinactuary.com/actuarial-formulae/corrections.
Acknowledgments
This document was devised and prepared as an extension to the SAS Guide by Tim Gummer, our Honours Representative
for 2019. Assistance and suggestions for revisions were provided for versions 1 and 2 of this document by 2018 secretary
Tom Marshall.
This document was created using document preparation and typesetting language LATEX by way of the online editor
Overleaf.

A Actuarial Formulae
1 Mathematical Methods
1.1 Series
Formula 1.1.1: Exponential Function
x2 x3
exp(x) = ex = 1 + x + + + ···
2! 3!
This formula may be familiar to some students from high school but it is also taught in MATH1016 Calculus 1* and
MATH1018 Accelerated Mathematics 2* under the topic of Taylor series.
You may find some use for this formula in the exam for the following actuarial unit(s):
• STAT3002 Risk Analysis and Credibility Theory
Extension 1.1.1.1: Exponential Function Using Sigma Notation

P
A more precise version of formula 1.1.1 would be to use notation.
∞
x
X xn
e =
n=0
n!
Formula 1.1.2: Natural log Function
This part of the formulae and tables is not used in any Curtin units in which actuarial formulae and tables are
provided in the exam and has thus been omitted.
Formula 1.1.3: Binomial expansion

n n−1 n n−2 2
(a + b)n = an + a b+ a b + · · · + bn
1 2
where n is a positive integer
p(p − 1) 2 p(p − 1)(p − 2) 3

(1 + x)p = 1 + px + x + x + ···
2! 3!
• STAT3001 Statistical Modelling
1.2 Calculus
Formula 1.2.1: Taylor series (one variable)
h2 00
f (x + h) = f (x) + hf 0 (x) + f (x) + · · ·
2!
Taylor series are covered in MATH1016 Calculus 1* or MATH1018 Accelerated Mathematics 2*. You will have likely
seen it in a different form however.
• STAT3006 Investment Science 1

Formula 1.2.2: Taylor series (two variables)
1 2 00
f (x + h, y + k) = f (x, y) + hfx0 (x, y) + kfy0 (x, y) + 00
h fxx (x, y) + 2hkfxy 00
(x, y) + k 2 fyy (x, y) + · · ·
2!
This formula is covered in MATH2009 Calculus 2*.

Formula 1.2.3: Integration by parts
Z b Z b
dv
u dx dx = [uv]ba − v du
dx dx
a a
Integration by parts is covered in MATH1016 Calculus 1* or MATH1017 Accelerated Mathematics 1*. It is a common
integration technique. Most students will likely have remembered this formula by this time but . . .
• MATH3006 Life Contingencies 1
Formula 1.2.4: Double integrals (changing the order of integration)
!
Z b Z x Z b Z b
f (x, y) dy dx = f (x, y) dx dy or
a a a y
Z b Z x Z b Z x
dx dy f (x, y) = dy dx f (x, y)
a a a a
The domain of integration here is the set of values (x, y) for which a ≤ y ≤ x ≤ b
These are covered for the first time in STAT1001 Statistical Probability, but are given more detail in MATH2009
Calculus 2*.
• STAT2001 Mathematical Statistics
Formula 1.2.5: Differentiating an integral
1.3 Solving Equations
Formula 1.3.1: Newton-Raphson method

Formula 1.3.2: Integrating factors

dy
The integrating factor for solving the differential equation dx + P (x)y = Q(x) is:
Z
exp P (x) dx
This formula is explained in MATH1016 Calculus 1* or MATH1018 Accelerated Mathematics 2*.

• STAT3005 Stochastic Processes
Formula 1.3.3: Second-order difference equations
The general solution of the difference equation axn+2 + bxn+1 + cxn = 0 is:
if b2 − 4ac > 0 : xn = Aλn1 + Bλn2 (distinct real roots, λ1 6= λ2 )
if b2 − 4ac = 0 : xn = (A + Bn)λn (equal real roots, λ1 = λ2 = λ)
if b2 − 4ac < 0 : xn = rn (A cos nθ + B sin nθ) (complex roots, λ1 = λ2 = reiθ )
where λ1 and λ2 are the roots of the quadratic equation aλ2 + bλ = 0.
This formula is derived in MATH1017 Accelerated Mathematics 1*, but simply taken as a fact in STAT3001 Statistical
Modelling.
1.4 Gamma Function

Formula 1.4.1: Definition
Z∞
Γ(x) = tx−1 e−t dt, x > 0
0
Formula 1.4.2: Properties
Γ(x) = (x − 1)Γ(x − 1)
Γ(n) = (n − 1)!, n = 1, 2, 3, . . .
√
Γ(1/2) = π

1.5 Bayes’ Formula
Formula 1.5.1: Bayes’ formula
Let A1 , A2 , . . . , An be a collection of mutually exclusive and exhaustive events with P (Ai ) 6= 0, i = 1, 2, . . . , n.
For any event B such that P (B) 6= 0:
P (B|Ai )P (Ai )
P (Ai |B) = P
n , i = 1, 2, . . . , n.
P (B|Aj )P (Aj )
j=1
2 Statistical Distributions
Formula 2.0.1: Notation
PF = Probability function, p(x)

PDF = Probability density function, f (x)
DF = Distribution function, F (x)
PGF = Probability generating function, G(s)
MGF = Moment generating function, M (t)
Note. Where formulae have been omitted below, this indicates that (a) there is no simple formula or (b) the
function does not have a finite value or (c) the function equals zero.
These conventions will have been first encountered in STAT1001 Statistical Probability.
This notation is used throughout 2.1 and 2.2
Extension 2.0.1.1: Common alternative conventions
• Referring to the “probability function” in 2.0.1 as a “probability mass function” (PMF).

• Referring to the “distribution function” in 2.0.1 as a “cumulative distribution function” (CDF).
• A subscript is often used to indicate which random variable the function corresponds to. E.g. fX (x) denotes
the PDF of some random variable X.
Extension 2.0.1.2: Definitions that correspond to these conventions
• The definition of a PF is p(x) := P (X = x).

d
• The definition of a PDF is f (x) := dx F (x).
• The definition of a DF is F (x) := P (X ≤ x).
• The definition of a PGF is G(s) := E(sX ).

• The definition of a MGF is M (t) := E(etX ).

Extension 2.0.1.3: Moment generation using the MGF
If M (t) is r times differentiable at t = 0, then

(r)
MX (0) = E(X r )
Extension 2.0.1.4: Cumulant Generating Function
The cumulant generating function (CGF) is defined as
C(t) := log M (t)
This is a commonly used transformation for the following properties which follow from 2.0.1.3 and some basic
differentiation techniques:
0
CX (0) = E(X)
00
CX (0) = var(X)
h 3 i
000
CX (0) = E X − E(X)
(note that higher order derivatives do not continue this nice pattern).
2.1 Discrete Distributions

Formula 2.1.1: Binomial distribution
Parameters: n, p (n = positive integer, 0 < p < 1 with q = 1 − p)

n x n−x
PF: p(x) = p q , x = 0, 1, 2, . . . , n
x
DF: The distribution function is tabulated in the statistical tables section.
PGF: G(s) = (q + ps)n
MGF: M (t) = (q + pet )n
Moments: E(X) = np, var(X) = npq
Coefficient
q−p
of skewness: √
npq
• MATH3005 Survival Analysis


Formula 2.1.2: Bernoulli distribution

The Bernoulli distribution is the same as the binomial distribution with parameter n = 1.

Formula 2.1.3: Poisson distribution
Parameter: µ (µ > 0)
e−µ µx
PF: p(x) = , x = 0, 1, 2, . . .
x!
PGF: G(s) = eµ(s−1)
t
−1)
MGF: M (t) = eµ(e
Moments: E(X) = µ, var(X) = µ
Coefficient
1
of skewness: √
µ


Formula 2.1.4: Negative binomial distribution – Type 1
Parameters: k, p (k = positive integer, 0 < p < 1 with q = 1 − p)

x − 1 k x−k
PF: p(x) = p q , x = k, k + 1, k + 2, . . .
k−1
k
ps
PGF: G(s) =
1 − qs
!k
pet
MGF: M (t) =
1 − qet
k kq
Moments: E(X) = , var(X) = 2
p p
Coefficient
2−p
of skewness: √
kq
Formula 2.1.5: Negative binomial distribution – Type 2
Parameters: k, p (k = positive integer, 0 < p < 1 with q = 1 − p)
Γ(k + x)
PF: p(x) = pk q x , x = k, k + 1, k + 2, . . .
Γ(x + 1)Γ(k)
k
ps
PGF: G(s) =
1 − qs
!k
pet
MGF: M (t) =
1 − qet
k kq
p p
Coefficient
2−p
of skewness: √
kq


Formula 2.1.6: Geometric distribution

The geometric distribution is the same as the negative binomial distribution with parameter k = 1.
Formula 2.1.7: Uniform distribution (discrete)
Parameters: a, b, h (a < b, h > 0, b − a is a multiple of h)
h
PF: p(x) = , x = a, a + h, a + 2h, . . . , b − h, b
b−a+h
!
h sb+h − sa
PGF: G(s) =
b−a+h sh − 1
!
h e(b+h)t − eat
MGF: M (t) =
b−a+h eht − 1
1 1
Moments: E(X) = (a + b), var(X) = (b − a)(b − a + 2h)
2 12
2.2 Continuous distributions

Formula 2.2.1: Standard normal distribution – N (0, 1)
Parameters: none
1 1 2
PDF: f (x) = √ e− 2 x , −∞ < x < ∞
2π
1 2
MGF: M (t) = e 2 t
Moments: E(X) = 0, var(X) = 1
1 Γ(1 + r)
E(X r ) = , r = 2, 4, 6, . . .
2r/2 Γ(1 + 2r )



Formula 2.2.2: Normal (Gaussian) distribution – N (µ, σ 2 )
Parameters: µ, σ 2 (σ > 0)
( 2 )
1 1 x−µ
PDF: f (x) = √ exp − , −∞ < x < ∞
σ 2π 2 σ
1 2 2
MGF: M (t) = eµt+ 2 σ t
Moments: E(X) = µ, var(X) = σ 2


Formula 2.2.3: Exponential distribution
Parameters: λ (λ > 0)
PDF: f (x) = λe−λx , x > 0
DF: F (x) = 1 − e−λx
−1
t
MGF: M (t) = 1 − ,t<λ
λ
1 1
λ λ
Γ(1 + r)
E(X r ) = , r = 1, 2, 3, . . .
λr
Coefficient
of skewness: 2


Formula 2.2.4: Gamma distribution
Parameters: α, λ (α > 0, λ > 0)
λα α−1 −λx
PDF: f (x) = x e ,x>0
Γ(α)
DF: When 2α is an integer, probabilities for the gamma

distribution can be found using the relationship:
2λX ∼ χ22α
−α
t
MGF: M (t) = 1 − ,t<λ
λ
α α
λ λ
Γ(α + r)
E(X r ) = , r = 1, 2, 3, . . .
Γ(α)λr
Coefficient
2
of skewness: √
α

Formula 2.2.5
The chi-square distribution with ν degrees of freedom is the same as the gamma distribution with parameters
α = ν2 and λ = 21 .
The distribution function for the chi-square distribution is tabulated in the statistical tables section.


Formula 2.2.6: Uniform distribution (continuous) – U (a, b)
Parameters: a, b a < b
1
PDF: f (x) = ,a<x<b
b−a
x−a
DF: F (x) =
b−a
1 1 bt
MGF: M (t) = (e − eat )
(b − a) t
1 1
Moments: E(X) = (a + b), var(X) = (b − a)2
2 12
1 1
E(X r ) = (br+1 − ar+1 ), r = 1, 2, 3 . . .
(b − a) r + 1

Formula 2.2.7: Beta distribution
Parameters: α, β (α > 0, β > 0)
Γ(α + β) α−1
PDF: f (x) = x (1 − x)β−1 , 0 < x < 1
Γ(α)Γ(β)
α αβ
Moments: E(X) = , var(X) =
α+β (α + β)2 (α + β + 1)
Γ(α + β)Γ(α + r)
E(X r ) = , r = 1, 2, 3, . . .
Γ(α)Γ(α + β + r)
Coefficient
s
2(β − α) α+β+1
of skewness:
α+β+2 αβ


Formula 2.2.8: Lognormal distribution
Parameters: µ, σ 2 (σ > 0)
( )
1 1 1 log x − µ
PDF: f (x) = √ exp − ,x>0
σ 2π x 2 σ
1 2 2
2

Moments: E(X) = eµ+ 2 σ , var(X) = e2µ+σ eσ − 1
2
1
σ2
E(X r ) = erµ+ 2 r , r = 1, 2, 3, . . .
Coefficient
2
p
of skewness: eσ + 2 eσ 2 − 1


• MATH2004 Theory of Interest
Extension 2.2.8.1: Variance of the Lognormal distribution
It can be useful identify the following relationship between the mean and variance of the lognormal distribution:
2 σ2
var(X) = E(X) e −1
This is useful whenever the parameters of a lognormal distribution are estimated by way of the method of moments.
This is done in MATH2004 Theory of Interest and STAT3002 Risk Analysis and Credibility Theory.

Formula 2.2.9: Pareto distribution (two parameter version)
Parameters: α, λ (α > 0, λ > 0)
αλα
PDF: f (x) = ,x>0
(λ + x)α+1
α
λ
DF: F (x) = 1 −
λ+x
λ αλ2
Moments: E(X) = (α > 1), var(X) = (α > 2)
α−1 (α − 1)2 (α − 2)
Γ(α − r)Γ(1 + r) r
E(X r ) = λ , r = 1, 2, 3, . . . , r < α
Γ(α)
Coefficient
r
2(α + 1) α−2
of skewness: (α > 3)
(α − 3) α
This is also known simply as the Pareto distribution.

This is mentioned as an example but not actually assessed in STAT2001 Mathematical Statistics.

Extension 2.2.9.1: Variance of the Pareto distribution

It can be useful identify the following relationship between the mean and variance of the Pareto distribution:
2 α
var(X) = E(X)
α−2
Formula 2.2.10: Pareto distribution (three parameter version)
Parameters: α, λ, k (α > 0, λ > 0, k > 0)
Γ(α + k)λα xk−1

PDF: f (x) = ,x>0
Γ(α)Γ(k)(λ + x)α+k
kλ k(k + α − 1)λ2
Moments: E(X) = (α > 1), var(X) = (α > 2)
α−1 (α − 1)2 (α − 2)
Γ(α − r)Γ(k + r) r
E(X r ) = λ , r = 1, 2, 3, . . . , r < α
Γ(α)Γ(k)
This is also known as the generalised Pareto distribution.



Formula 2.2.11: Weibull distribution
Parameters: c, γ (c > 0, γ > 0)
γ
PDF: f (x) = cγxγ−1 e−cx , x > 0
γ
DF F (x) = 1 − e−cx

r r 1
Moments: E(X ) = Γ 1 + r = 1, 2, 3, . . .
γ cr/γ
Formula 2.2.12: Burr distribution
Parameters: α, λ, γ (α > 0,λ > 0 , γ > 0)
αγλα xγ−1
PDF: f (x) = ,x>0
(λ + xγ )α+1
α
λ
DF F (x) = 1 −
λ + xγ
r λr/γ

r r
Moments: E(X ) = Γ α − Γ 1+ r = 1, 2, 3, . . . , r < αγ
γ γ Γ(α)
2.3 Compound Distributions
Formula 2.3.1: Conditional expectation and variance
E(Y ) = E[E(Y |X)]
var(Y ) = var[E(Y |X)] + E[var(Y |X)]
This is also known as the “tower property”.


Formula 2.3.2: Moments of a compound distribution
If X1 , X2 , . . . are IID random variables with MGF MX (t) and N is an independent nonnegative integer-valued
random variable, then S = X1 + · · · + XN (with S = 0 when N = 0) has the following properties:
Mean: E(S) = E(N )E(X)
Variance: var(S) = E(N )var(X) + var(N )[E(X)]2
MGF: MS (t) = MN [log MX (t)]
Formula 2.3.3: Compound Poisson distribution
Mean: λm1
Variance: λm2
Third central moment: λm3
where λ = E(N ) and mr = E(X r )
Formula 2.3.4: Recursive formulae for integer-valued distributions
2.4 Truncated Moments

Extension 2.4.0.1: Notation used in truncated moments formulae
It is not stated anywhere in the formula section of the orange book (but it is in the Standard Normal probabilities
table, which Curtin students aren’t actually provided with in exams since they are replaced with ‘Mathematical
Formulae and Statistical Tables for Tertiary Institutions’) but it is common to use the following notation for a
standard normal random variable (Z ∼ N (0, 1)):
PDF: φ(z) ≡ fZ (z)

DF: Φ(z) ≡ FZ (z)
Note that the φ and Φ are the lowercase and uppercase versions respectively of the greek letter ‘phi’.

Formula 2.4.1: Normal distribution
If f (x) is the PDF of the N (µ, σ 2 ) distribution, then

Z U
xf (x)dx = µ[Φ(U 0 ) − Φ(L0 )] − σ[φ(U 0 ) − φ(L0 )]
L
L−µ U −µ
where L0 = and U 0 =
σ σ
This formula is derived in STAT3002 Risk Analysis and Credibility Theory.

Formula 2.4.2: Lognormal distribution
If f (x) is the PDF of the lognormal distribution, then

Z U
2
σ2
xk f (x)dx = ekµ+1/2k [Φ(Uk ) − Φ(Lk )]
L
log L − µ log U − µ
where Lk = − kσ and Uk = − kσ
σ σ
This formula is derived in STAT3002 Risk Analysis and Credibility Theory.

Relationships between statistical distributions
Formula 2.4.3: Statistical distributions diagram
This diagram shows the relationships between all the Discrete Distributions and Continuous distributions above in
the form of a flowchart. It has been omitted due to its cumbersome nature.

Formula 2.4.4: Explanation of the distribution diagram
A description of the conventions used in 2.4.3. It has been omitted due to its cumbersome nature.


3 Statistical Methods
3.1 Sample Mean and Variance
Formula 3.1.1: Sample mean and variance
The random sample (x1 , x2 , . . . , xn ) has the following sample moments:

n
1X
Sample mean: x= xi
n i=1
 
n
1 X 
Sample variance: s2 = x2i − nx2
n − 1  i=1 
3.2 Parametric Inference (Normal Model)
Formula 3.2.1: One sample
Formula 3.2.2: Two samples
3.3 Maximum Likelihood Estimators

Formula 3.3.1: Asymptotic distribution
If θ̂ is the maximum likelihood estimator of a parameter θ based on a sample X, then θ̂ is asymptotically normally
distributed with mean θ and variance equal to the Cramér-Rao lower bound
h i
∂2
CRLB(θ) = −1 E ∂θ 2 log L(θ, X)


Formula 3.3.2: Likelihood ratio test

 
max L
H0
−2(`p − `p+q ) = −2 log   ∼ χ2q approximately (under H0 )
max L
H0 ∪H1
where `p = max log L is the maximum log-likelihood for the

H0
model under H0 (in which there are

p free parameters)
and `p+q = max log L is the maximum log-likelihood for the

H0 ∪H1
model under H0 ∪ H1 (in which there

are p + q free parameters)
3.4 Linear Regression Model With Normal Errors

All of the below formulae are explained/derived in STAT1000 Regression and Non-Parametric Inference.
Formula 3.4.1: Model
Yi ∼ N (α + βxi , σ 2 ), i = 1, 2, . . . , n
Formula 3.4.2: Intermediate calculations
n
X n
X
2
sxx = (xi − x) = x2i − nx2
i=1 i=1
n
X n
X
syy = (yi − y)2 = yi2 − ny 2
i=1 i=1
n
X n
X
sxy = (xi − x)(yi − y) = xi yi − nx y
i=1 i=1
Formula 3.4.3: Parameter estimates
sxy
α̂ = y − β̂x, β̂ =
sxx
n
!
2 1 X 1 s2xy
σ̂ = (yi − ŷi )2 = syy −
n − 2 i=1 n−2 sxx

Formula 3.4.4: Distribution of β̂
Formula 3.4.5: Variance of predicted mean response
Formula 3.4.6: Testing the correlation coefficient
Formula 3.4.7: Fisher Z transformation

Formula 3.4.8: Sum of squares relationship
3.5 Analysis of Variance

Analysis of variance is covered in STAT1002 Statistical Data Analysis but it is not revisited in later CT subjects which
adhere to actuarial exam conditions at Curtin.
Formula 3.5.1: Single factor normal model
Formula 3.5.2: Intermediate calculations (sums of squares)
Formula 3.5.3: Variance estimate

Formula 3.5.4: Statistical test


3.6 Generalised Linear Models

Formula 3.6.1: Exponential Family
For a random variable Y from the exponential family, with natural parameter θ and scale parameter φ:

yθ − b(θ)
Probability (density) function: fY (y; θ, φ) = exp + c(y, φ)
a(φ)
Mean: E(Y ) = b0 (θ)
Variance: var(Y ) = a(φ)b00 (θ)
Formula 3.6.2: Canonical Link Functions
µ
Binomial: g(µ) = log
1−µ
Poisson: g(µ) = log µ
Normal: g(µ) = µ
1
Gamma: g(µ) =
µ
3.7 Bayesian Methods
Formula 3.7.1: Relationship between posterior and prior distributions
Posterior ∝ Prior × Likelihood

The posterior distribution f (θ|x) for the parameter θ is related to the prior distribution f (θ) via the likelihood
function f (x|θ):
f (θ|x) ∝ f (θ) × f (x|θ)
Formula 3.7.2: Normal / normal model
If x us a random sample of size n from a N (µ, σ 2 ) distribution, where σ 2 is known, and the prior distribution for
the parameter µ is N (µ0 , σ02 ), then the posterior distribution for µ is:
µ|x ∼ N (µ∗ , σ∗2 )

!, ! , !
nx µ0 n 1 2 n 1
where µ∗ = + 2 + 2 and σ∗ = 1 + 2
σ2 σ0 σ2 σ0 σ2 σ0

This result is derived in STAT3001 Statistical Modelling.


3.8 Empirical Bayes Credibility – Model 1
Formula 3.8.1: Data requirements
{Xij , i = 1, 2, . . . , N, j = 1, 2, . . . , n}
Xij represents the aggregate claims in the j th year from the i th risk.

n N
1X 1 X
Xi = Xij , X= Xi
n j=1 N i=1
Formula 3.8.3: Parameter estimation
Quantity Estimator
E[m(θ)] X
 
N n
1 X 1 X 
E[s2 (θ)] (Xij − X i )2
N i=1  n − 1 j=1 
 
N N  n
1 X 1 X 1 X 
var[m(θ)] (X i − X)2 − (Xij − X i )2
N − 1 i=1 N n i=1  n − 1 j=1 

Extension 3.8.3.1: Simplification of parameter estimation
If the following additional intermediate calculations are performed:

n N
1 X 2 1 X
Si2 = (Xij − X i )2 , S = (X i − X)2
n − 1 j=1 N i=1
These calculations as well as those for the the means in 3.8.2 can be performed quickly and accurately by making
use of the statistics functionality of most scientific calculators.
We can then make the following simplifications to 3.8.3:
Quantity Estimator
N
1 X 2
E[s2 (θ)] S
N i=1 i
2 1
var[m(θ)] S − E[s2 (θ)]
n
Formula 3.8.4: Credibility Factor

n
Z= E[s2 (θ)]
n+ var[m(θ)]
Extension 3.8.4.1: Credibility formula
When applying formula 3.8.4 (and 3.9.4), it is important to remember that the credibility premium is:
Zx + (1 − Z)X
Where
x is an estimate of the credibility premium based on direct data. If we are interested in the k th risk, then in
the notation above, we would set x = X k .
X is an estimate of the credibility premium based on collateral data. Note that this is the same as X in the
formulae above.
This is usually presented as ZX + (1 − Z)µ in the CT notes, however this is inconsistent with the notation used in
the formulae above, so the notation has been adjusted accordingly.
3.9 Empirical Bayes Credibility – Model 2
Formula 3.9.1: Data requirements
{Yij , i = 1, 2, . . . , N, j = 1, 2, . . . , n}, {Pij , i = 1, 2, . . . , N, j = 1, 2, . . . , n}

Yij represents the aggregate claims in the j th year from the i th risk;
Pij is the corresponding risk volume.
Note that Yij here is analogous to Xij in 3.8.1



n N N
!
X X
∗ 1 X Pi
Pi = Pij , P = P i, P = Pi 1−
j=1 i=1
N n − 1 i=1 P
n N X
n
Yij X Pij Xij X Pij Xij
Xij = , Xi = , X=
Pij j=1
Pi i=1 j=1
P
Extension 3.9.2.1: Quantities of interest
Here Xij represents number of claims per year per risk volume. Whether we are more estimating X or Y depends
on the particular problem. We need to use the relationship in 3.9.2 if we want to recover an estimate for Y .
Formula 3.9.3: Parameter estimation
Quantity Estimator
E[m(θ)] X
 
N  n
1 X 1 X 
E[s2 (θ)] Pij (Xij − X i )2
N n − 1 
i=1 j=1
  
n
N X N  n
1  1 X 1 X 1 X 
var[m(θ)] Pij (Xij − X)2 − Pij (Xij − X i )2 

P∗ Nn − 1 N n − 1


i=1 j=1 i=1 j=1
Formula 3.9.4: Credibility Factor

Pn
j=1 Pij
Zi = Pn E[s2 (θ)]
j=1 Pij + var[m(θ)]
4 Compound Interest
Formula 4.0.1: Increasing/decreasing annuity functions
ä n − nv n n − an
(Ia) n = , (Da) n =
i i

Formula 4.0.2: Accumulation factor for variable interest rates

 
Zt2
A(t1 , t2 ) = exp  δ(t) dt
 
t1
5 Survival Models
5.1 Mortality “Laws”
Formula 5.1.1: Survival Probabilities

 
Zt
t px = exp  µx+s ds
 

Formula 5.1.2: Gompertz’ Law

x
(ct −1)
µx = Bcx , t px = gc where g = e−B/ log c
Formula 5.1.3: Makeham’s Law

x
(ct −1)
µx = A + Bcx , t px = st g c where s = e−A
Formula 5.1.4: Gompertz-Makeham formula
The Gompertz-Makeham graduation formula, denoted by GM(r, s), states that
µx = poly1 (t) + exp[poly2 (t)]

where t us a linear function of x and poly1 (t) and poly2 (t) are polynomials of degree r and s respectively.

5.2 Empirical Estimation
Formula 5.2.1: Greenwood’s formula for the variance of the Kaplan-Meier estimator
h i2 X dj
var[F̃ (t)] = 1 − F̂ (t)
nj (nj − dj )
tj ≤t
Formula 5.2.2: Variance of the Nelson-Aalen estimate of the integrated hazard
X dj (nj − dj )
var[Λ̃t ] =
n3j
tj ≤t
5.3 Mortality Assumptions
Formula 5.3.1: Balducci Assumption
1−t qx+t = (1 − t)qx (x is an integer, 0 ≤ t ≤ 1)
5.4 General Markov Model

Formula 5.4.1: Kolmogorov forward differential equation
∂ gh X gj jh gh hj

t px = t px µx+t − t px µx+t
∂t
j6=h
5.5 Graduation Tests

Formula 5.5.1: Grouping of signs test
If there are n1 positive signs and n2 negative signs and G denotes the observed number of positive runs, then:

n1 − 1 n2 + 1
t−1 t
P (G = t) = and, approximately,
n1 + n2
n1
!
n1 (n2 + 1) (n1 n2 )2
G∼N ,
n1 + n2 (n1 + n2 )3
Critical values for the grouping of signs test are tabulated in the statistical tables section for small values of n1
and n2 . For larger values of n1 and n2 the normal approximation can be used.

Formula 5.5.2: Serial correlation test
m−j
1
P
m−j (zi − z)(zi+j − z) m
i=1 1 X
rj ≈ m where z = zi
1
P m i=1
m (zi − z)2
i=1
√
rj × m ∼ N (0, 1) approximately.
Formula 5.5.3: Variance adjustment factor
i2 πi
P
i
rx = P
iπi
i
where πi is the proportion of lives at age x who have exactly i policies.
5.6 Multiple Decrement Tables
Formula 5.6.1
For a multiple decrement table with three decrements α, β and γ, each uniform over the year of age (x, x + 1) in
its single decrement table, then
α α 1 β γ 1 β γ
(aq)x = qx 1 − (qx + qx + qx qx )
2 3
5.7 Population Projection Models
Formula 5.7.1: Logistic model

6 Annuities and Assurances

6.1 Approximations for Non Annual Annuities
Formula 6.1.1: Approximations for Non Annual Annuities
m−1
ä(m)
x ≈ äx −
2m

(m) m−1 Dx+n
äx:n ≈ äx:n − 1−
2m Dx
6.2 Moments of Annuities and Assurances

Formula 6.2.1
Let Kx and Tx denote the curtate and complete future lifetimes (respectively) of a life aged exactly x.
Formula 6.2.2: Whole life assurances
E[v Kx +1 ] = Ax , var[v Kx +1 ] = 2 Ax − (Ax )2
E[v Tx ] = Āx , var[v Tx ] = 2 Āx − ( Āx )2
1
Similar relationships hold for endowment assurances (with status · · ·x:n ), pure endowments (with status x : n ),
1
term assurances (with status x : n ) and deferred whole life assurances (with status m| · · ·x ).
Formula 6.2.3: Whole life annuities
2
Ax − (Ax )2
E[ä Kx +1 ] = äx , var[ä Kx +1 ] =
d2
2
Āx − ( Āx )2
E[ā Tx ] = āx , var[ā Tx ] =
δ2
Similar relationships hold for temporary annuities (with status · · ·x:n ).


6.3 Premiums and Reserves

Formula 6.3.1: Premium conversion relationship between annuities and assurances
Ax = 1 − däx , Āx = 1 − δ āx

Similar relationships hold for endowment assurance policies (with status · · ·x:n ).
Formula 6.3.2: Net premium reserve
äx+t āx+t
t Vx =1− , tV x = 1 −
äx āx
Similar formulae hold for endowment assurance policies (with statuses · · ·x:n and · · ·x+t:n−t ).
6.4 Thiele’s Differential Equation
Formula 6.4.1: Whole life assurance

Formula 6.4.2: Multiple state model
7 Stochastic Processes
7.1 Markov “Jump” Processes
Formula 7.1.1: Kolmogorov differential equations
X
∂
Forward equation: ∂t pij (s, t) = pik (s, t)σkj (t)
k∈S
X
∂
Backward equation: ∂s pij (s, t) =− σik (s)pkj (s, t)
k∈S
P
where σij (t) is the transition rate from state i to state j (j 6= i) at time t, and σii = − σij .
j6=i

Formula 7.1.2: Expected time to reach a subsequent state k
1 X σij X
mi = + mj , where λi = σij
λi λi
j6=i,j6=k j6=i
7.2 Brownian Motion and Related Processes

Formula 7.2.1: Martingales for standard Brownian motion
If {Bt , t ≥ 0} is a standard Brownian motion, then the following processes are martingales:
1
Bt , Bt2 − t and exp(λBt − λ2 t)
2
Formula 7.2.2: Distribution of the maximum value

Formula 7.2.3: Hitting times
Formula 7.2.4: Ornstein-Uhlenbeck process
dXt = −γXt dt + σ dBt , γ>0
7.3 Monte Carlo Methods

Formula 7.3.1: Box-Muller formulae
If U1 and U2 are independent random variables from the U (0, 1) distribution then
p p
Z1 = −2 log U1 cos(2πU2 ) and −2 log U1 sin(2πU2 )
are independent standard normal variables.

Formula 7.3.2: Polar method
If V1 and V2 are independent random variables from the U (−1, 1) distribution and S = V12 + V22 then, conditional
on 0 < S ≤ 1,
r r
−2 log S −2 log S
Z1 = V1 and Z2 = V2
S S
are independent standard normal variables.
Pseudorandom values from the U (0, 1) distribution and the N (0, 1) distribution are included in the statistical tables
section.
8 Time Series
8.1 Time Series – Time Domain
Formula 8.1.1: Sample autocovariance and autocorrelation function
n n
1 X 1X
Autocovariance: γ̂k = (xt − µ̂)(xt−k − µ̂), where µ̂ = xt
n n t=1
t=k+1
γ̂k
Autocorrelation: ρ̂k =
γ̂0
Formula 8.1.2: Autocorrelation function for ARMA(1, 1)
For the process Xt = αXt−1 + et + βet−1 :
(1 + αβ)(α + β) k−1
ρk = α , k = 1, 2, 3, . . .
(1 + β 2 + 2αβ)

Formula 8.1.3: Partial autocorrelation function
ρ2 − ρ21
φ1 = ρ1 , φ 2 =
1 − ρ21
det Pk∗
φk = , k = 2, 3, . . . ,
det Pk
 
1 ρ1 ρ2 ··· ρk−1
ρ1 1 ρ1 ··· ρk−2 
 

ρ2 ρ1 1 ··· ρk−2 
 
where Pk = 
 .. .. .. .. .. 

 . . . . . 

ρk−1 ρk−2 ρk−3 ··· 1
and Pk∗ equals Pk , but with the last column replaced with (ρ1 , ρ2 , ρ3 , . . . , ρk )> .
Formula 8.1.4: Partial autocorrelation function for MA(1)
For the process Xt = µ + et + βet−1 :
(1 − β 2 )β k
φk = (−1)k+1 , k = 1, 2, 3, . . .
1 − β 2(k+1)
8.2 Time Series – Time Frequency
Formula 8.2.1: Spectral density function
Formula 8.2.2: Inversion formula

Formula 8.2.3: Linear filters

8.3 Time Series – Box-Jenkins Methodology
Formula 8.3.1: Ljung and Box ”portmanteau” test of the residuals for an ARMA(p, q) model
m
X rk2
n(n + 2) ∼ χ2m−(p+q)
n−k
k=1
where rk (k = 1, 2, . . . , m) is the estimated value of the k th autocorrelation coefficient of the residuals and n is
the number of data values used in the ARMA(p, q) series.

Formula 8.3.2: Turning point test
In a sequence of n independent random variables the number of turning points T is such that:
2 16n − 29
E(T ) = (n − 2) and var(T ) =
3 90
9 Economic Models
9.1 Utility Theory
Formula 9.1.1: Utility functions
Exponential: U (w) = e−aw , a > 0
Logarithmic: U (w) = log w
Power: U (w) = γ −1 (wγ − 1), γ ≤ 1, γ 6= 0
Power: U (w) = w + dw2 , d < 0
Formula 9.1.2: Measures of risk aversion
U 00 (w)
Absolute risk aversion: A(w) = −
U 0 (w)
Relative risk aversion: R(w) = wA(w)
9.2 Capital Asset Pricing Model (CAPM)
Formula 9.2.1: Security market line
cov(Ri , RM )
Ei − r = βi (EM − r) where βi =
var(RM )

Formula 9.2.2: Capital market line (for efficient portfolios)
σp
Ep − r = (EM − r)
σM
9.3 Interest Rate Models

Formula 9.3.1: Spot rates and forward rates for zero-coupon bonds
Let P (τ) be the price at time 0 of a zero-coupon bond that pays 1 unit at time τ.
Let s(τ) be the spot rate for the period (0, τ).
Let f (τ) be the instantaneous forward rate at time 0 for time τ.
Spot rate
1
P (τ) = e−τs(τ) or s(τ) = − log P (τ)
τ
Instantaneous forward rate
Z τ
d
P (τ) = exp − f (s) ds or f (τ) = − log P (τ)
0 dτ
Formula 9.3.2: Vasicek model

β −ατ
f (τ) = e−ατ R + (1 − e−ατ )L + e (1 − e−ατ )
α
Price of a zero-coupon bond

β
P (τ) = exp −D(τ)R − (τ − D(τ))L − D(τ)2
2
1 − e−ατ
where D(τ) =
α

Extension 9.3.2.1: Alternative parameterisation of the Vasicek model
9.3.2 is not the way in which this model is presented in the CT8 notes (also the formula students are expected to
used in the STAT3007 exam).
Price of a zero-coupon bond
B(t, T ) = ea(τ)−b(τ)r(t)
σ2 σ 2 −ατ
f (τ) = e−ατ r(t) + (µ − )(1 − e −ατ
) + e (1 − e−ατ )
2α2 2α2
where:
1 − e−αt
b(τ) =
α
" #
σ2 σ2
a(τ ) = (b(τ) − τ) µ − 2 − b(τ)2
2α 4α
We can get to this from 9.3.2 by knowing that:
τ=T −t
R = r(t)
D(τ) = b(τ)
σ2
β=
2α
L=µ−β
Also, 9.3.2 implicitly assumes that t = 0.
10 Financial Derivatives
Formula 10.0.1
Note. In this section, q denotes the (continuously-payable) dividend rate.
10.1 Price of a Forward or Futures Contract

Formula 10.1.1
For an asset with fixed income of present value I:
F = (S0 − I)erT
For an asset with dividends:
F = S0 e(r−q)T

10.2 Binomial Pricing (“Tree” Model)
Formula 10.2.1: Risk-neutral probabilities
er∆t − d
Up-step probability =
u−d
√
where u ≈ eσ ∆t+q∆t
√
and u ≈ e−σ ∆t+q∆t
Also known as ‘calibration’ of the binomial model.

10.3 Stochastic Differential Equations
Formula 10.3.1: Generalised Wiener process
dx = a dt + b dz
where a and b are constant and dz is the increment for a Wiener process (standard Brownian motion).

Formula 10.3.2: Ito process
dx = a(x, t) dt + b(x, t) dz
Formula 10.3.3: Ito’s lemma for a function G(x, t)

1 2 ∂G
dG = a ∂∂xG + b2 ∂∂xG2 + b ∂∂tG dt + b dz
2 ∂x

Formula 10.3.4: Models for the short rate rt
Ho-Lee: dr = θ(t) dt + σ dz
Hull-White: dr = [θ(t) − ar] dt + σ dz
Vasicek: dr = a(b − r) dt + σ dz
√
Cox-Ingersoll-Ross: dr = a(b − r) dt + σ r dz
10.4 Black-Scholes Formulae for European Options
Formula 10.4.1: Geometric Brownian motion model for a stock price St
dSt = St (µ dt + σ dz)
Formula 10.4.2: Black-Scholes partial differential equation
∂f ∂f 1 ∂2f
+ (r − q)St + σ 2 St2 2 = rf
∂t ∂St 2 ∂St
Formula 10.4.3: Garman-Kohlhagen formulae for the price of call and put options
Call: ct = St e−q(T −t) Φ(d1 ) − Ke−r(T −t) Φ(d2 )
Put: pt = Ke−r(T −t) Φ(−d2 ) − St e−q(T −t) Φ(−d1 )
log(St /K) + (r − q + 21 σ 2 )(T − t)

where d1 = √
σ T −t
log(St /K) + (r − q − 21 σ 2 )(T − t) √

and d2 = √ = d1 − σ T − t
σ T −t

10.5 Put-Call Parity Relationship
Formula 10.5.1
ct + Ke−r(T −t) = pt + St e−q(T −t)
B Non-Statistical (Mostly Life) Tables

Compound Interest Table
(Mostly) Life Table 1: Compound Interest
This table is relevant to, but never used in MATH2004 Theory of Interest.
Population Mortality Table

(Mostly) Life Table 2: ELT15 (Males)
The following quantities are tabulated for ages (x) 0 to 109:

• `x
• dx
• qx
• µx
◦
• ex
You may find some use for this table in the exam for the following actuarial unit(s):

(Mostly) Life Table 3: ELT15 (Females)

• `x
• dx
• qx
• µx
◦
• ex


Assured Lives Mortality Table

(Mostly) Life Table 4: AM92

• `[x] , `[x+1]−1 , `x
• d[x] , d[x+1]−1 , dx
• q[x] , q[x+1]−1 , qx
• µ[x] , µ[x+1]−1 , µx
• e[x] , e[x+1]−1 , ex
The following commutation factors are tabulated for ages (x) 17 to 110, using a rate of interest of 4%:
• D[x] , D[x+1]−1 , Dx
• N[x] , N[x+1]−1 , Nx
• S[x] , S[x+1]−1 , Sx
• C[x] , C[x+1]−1 , Cx
• M[x] , M[x+1]−1 , Mx
• R[x] , R[x+1]−1 , Rx
The following quantities are tabulated for ages (x) 17 to 120, using a rate of interest of 4% and 6%:
• ä[x] , äx
• A[x] , Ax
• 2 A[x] , 2 Ax
• (Iä)[x] , (Iä)x
• (IA)[x] , (IA)x
• The following for n = 60 − x and n = 65 − x:

– ä[x]:n , äx:n
– A[x]:n , Ax:n

Pensioner Mortality Table

(Mostly) Life Table 5: PMA92Base and PFA92base

• qx
(Mostly) Life Table 6: PMA92C20 and PFA92C20

• `x
• dx
• qx
• µx
◦
• ex
The following quantities are tabulated for ages (x) 50 to 105 using a rate of interest of 4%:
• äx
• 2 Ax
• For male (x), female (y), let d = y − x.

For integer d between −5 and 5 (inclusive) as well as ±10 and ±20:
äxy
International Actuarial Notation

This component of the orange book is for some reason never provided to students in actuarial examinations at Curtin.
Perhaps students should request it given it is in the official orange book (and could come in handy during exams for
example it contains the meaning of standard commutation functions). The knowledge in this section is taught and used
throughout MATH2004 Theory of Interest, MATH3006 Life Contingencies 1, MATH3007 Life Contingencies 2.
Sickness Table
(Mostly) Life Table 7: S(ID)

Example Pension Scheme Table

(Mostly) Life Table 8: PEN
The following service table and relative salary scale quantities are tabulated for ages (x) 16 to 65:
• `x
• wx
• dx
• ix
• rx
• s∗x
• sx = (1.02)x s∗x
• zx = 13 (sx−3 + sx−2 + sx−1 )
• zx+ 12 = 12 (zx + zx+1 )
The following contribution functions quantities are tabulated for ages (x) 16 to 65 at an interest rate of 4%:
• Dx = v x `x
• Dx = 21 (Dx + Dx+1 )
P
• Nx = Dx
• s Dx = sx Dx
• sN x =
Ps
Dx
• s Dx = sx Dx
The following ill health retirement functions quantities are tabulated for ages (x) 16 to 65 at an interest rate
of 4%:
• āix+ 1
2
1
• = v x+ 2 ix
Cxi
• Mxi =
P i
Cx
i
• Rx = (Mxi − 12 Cxi )
P
• Cxia = Cxi āix+ 1

2
• Mxia = Cxia
P
ia
• Rx = (Mxia − 12 Cxia )
P
ia
• s M x = sx (Mxia − 12 Cxia )
ia ia
• s Rx =
Ps
Mx
• z Cxia = zx+ 12 Cxia
• z Mxia =
P z ia
Cx
ia 1 z ia
• z Rx = (z Mxia −
P
2 Cx )

(Mostly) Life Table 9: PEN cont.
The following age retirement functions quantities are tabulated for ages (x) 16 to 65 at an interest rate of 4%:
• ārx+ 1 (ar65 at 65)
2
1
• Cxr = v x+ 2 rx (v 65 r65 at 65)
• Mxr =
P r
Cx
r P r 1 r
• Rx = (Mx − 2 Cx )
• Cxra = Cxr ārx+ 1 (v 65 r65 ār65 at 65)
2
• Mxra = Cxra
P
ra
• Rx = (Mxra − 12 Cxra )
P
ra
• s M x = sx (Mxra − 12 Cxra )
ra P s ra
• s Rx = Mx
ra
• z Cxra = zx+ 21 Cxra (z65 C65 at 65)
• z Mxra =
P z ra
Cx
ra
• z Rx = (z Mxra − 21 z Cxra )
P
The following Functions for return of contributions, accumulated with interest at 2% p.a., on death
quantities are tabulated for ages (x) 16 to 65 at an interest rate of 4%:
1 1
• j Cxd = v x+ 2 (1 + j)x+ 2 dx
• j Mxd = j Cxd
P

d P j Mxd − 21 j Cxd
• j Rx = x+ 1
(1+j) 2

j
sj d P Mxd − 21 j Cxd
• Rx = sx 1
(1+j)x+ 2
1 1
• j Cxw = v x+ 2 (1 + j)x+ 2 wx
• j Mxw = j Cxw
P

j w
P j Mxw − 12 j Cxw
• Rx = x+ 1
(1+j) 2

j
sj w P Mxw − 12 j Cxw
• Rx = sx 1
(1+j)x+ 2
Sample Time Series

(Mostly) Life Table 10
This table is relevant to, but never used in STAT3001 Statistical Modelling.

C Statistical tables
Note that this part of the orange book is not provided to Curtin students in examinations. In this section we will indicate
which statistical tables are provided to students in another form in various units.
Standard Normal probabilities

Statistical Table 1
Instead of this table, curtin students are provided with “Cumulative probabilities for the standard normal distri-
bution” from ‘Mathematical Formulae and Statistical Tables for Tertiary Institutions’ in exams which includes:
Probabilities for values of Z between −3.49 and 3.39 at increments of 0.01.
• MATH2004 Theory of Interest MATH3005 Survival Analysis


Statistical Table Extension 1.1: Notation that appears in the orange book version
The following appears above the standard normal probabilities table in the orange book:
This is not given to students in exams despite the fact that formulae 2.4.1 and 2.4.2 use this notation. This is
indicated in extension 2.4.0.1.
Standard Normal percentage points

Statistical Table 2
Instead of this table, curtin students are provided with “Cumulative probabilities for the standard normal distri-
bution” from ‘Mathematical Formulae and Statistical Tables for Tertiary Institutions’ in exams which can be used
to obtain the percentage points (critical values) in reverse. The method of doing this is first taught in STAT1002
Statistical Data Analysis.
• MATH2004 Theory of Interest MATH3005 Survival Analysis



t percentage points
Statistical Table 3
χ2 probabilities
Statistical Table 4
Since this distribution is only really used as a distribution for a test statistic in certain hypothesis tests, it’s probabilities
are redundant information if we have access to critical values. Indeed ‘Mathematical Formulae and Statistical Tables for
Tertiary Institutions’, does not contain a probabilities table for the chi-squared distribution.
χ2 percentage points
Statistical Table 5
Instead of this table, curtin students are provided with “Critical points of the chi-squared distribution” from
‘Mathematical Formulae and Statistical Tables for Tertiary Institutions’ in exams which includes:
Critical values for degrees of freedom between 1 and 100 and for right tail probabilities of 0.995, 0.99, 0.95, 0.9,
0.1, 0.05, 0.025, 0.01, 0.005.
F percentage points
Statistical Table 6
Piosson Probabilities
Statistical Table 7
Instead of this table, curtin students are provided with “Cumulative probabilities for the Poisson distribution”
from ‘Mathematical Formulae and Statistical Tables for Tertiary Institutions’ in exams which includes:
Probabilities for parameter values between 0.1 and 3 at increments of 0.1, between 3.5 and 10 at increments of 0.5
and x values over a range for which significant changes occur.
In addition, “Individual probabilities for the Poisson distribution” has the same values available.

Binomial Probabilities
Statistical Table 8
Instead of this table, curtin students are provided with “Cumulative probabilities for the binomial distribution”
from ‘Mathematical Formulae and Statistical Tables for Tertiary Institutions’ in exams which includes:
Probabilities for n values between 2 and 15, as well as n values of 20, 25 and 30, p values between 0.1 and 0.9 with
increments of 0.1 and x values between 0 and n.
In addition, “Individual probabilities for the binomial distribution” has the same values available.
Critical values for the grouping of signs test

Statistical Table 9
This table contains critical values of the grouping of signs test for n1 and n2 between 1 and 25.
This is the only table which is actually provided to students directly from the orange book.
Pseudorandom values from U (0, 1) and from N (0, 1)

Statistical Table 10
D Reference by particular units

ACCT1000 Accounting - The Language of Business
This is a CT exemption unit for which no actuarial formulae are provided.
ECON1000 Introductory Economics

ECON1001 Actuarial Economics

ECON2001 Macroeconomics Principles

FNCE2000 Introduction to Finance Principles


MATH1016 Calculus 1*
Although this unit does not provide the actuarial tables and formulae in the exam, it does introduce/ derive formulae
from the following sections:
• 1.1 Series
• 1.2 Calculus
• 1.3 Solving Equations
MATH1017 Accelerated Mathematics 1*

• 1.1 Series
• 1.2 Calculus
MATH1018 Accelerated Mathematics 2*

• 1.1 Series
MATH2004 Theory of Interest

Formula from the following sections are useful in the exam of this unit.
Section Topic(s)
4 Compound Interest Interest rates, Annuities
The following statistical tables are useful in exam for this unit:
• Standard Normal probabilities
• Standard Normal percentage points
MATH2009 Calculus 2*
• 1.2 Calculus
MATH3005 Survival Analysis

Section Topic(s)
2.1 Discrete Distributions The Binomial and Poisson models
2.2 Continuous distributions Survival models, Graduation
3.3 Maximum Likelihood Estimators Estimating the lifetime distribution function, Proportional hazards models
5.1 Mortality “Laws” Survival models
5.2 Empirical Estimation Estimating the lifetime distribution function
5.3 Mortality Assumptions Survival models, The Binomial and Poisson models
5.5 Graduation Tests Graduation
The following life tables are useful in exam for this unit:
• Population Mortality Table


• χ2 percentage points
• Critical values for the grouping of signs test
MATH3006 Life Contingencies 1

Section Topic(s)
1.2 Calculus Life assurance contracts, Life annuity contracts
5.1 Mortality “Laws” ALL
6.1 Approximations for Non Annual Annuities Evaluation of assurances and annuities
6.2 Moments of Annuities and Assurances ALL
6.3 Premiums and Reserves Net and gross premiums and reserves
• Assured Lives Mortality Table
MATH3007 Life Contingencies 2

Section Topic(s)
1.2 Calculus Payments involving two lives
5.1 Mortality “Laws” Payments involving two lives, Contingent and reversionary benefits
5.6 Multiple Decrement Tables Competing risks
6.1 Approximations for Non Annual Annuities Payments involving two lives, Contingent and reversionary benefits
6.2 Moments of Annuities and Assurances ALL
• Assured Lives Mortality Table
• Pensioner Mortality Table

• Example Pension Scheme Table
STAT1000 Regression and Non-Parametric Inference

STAT1001 Statistical Probability

• 1.2 Calculus
• 2.1 Discrete Distributions
• 2.2 Continuous distributions
STAT1002 Statistical Data Analysis


STAT2001 Mathematical Statistics

Section Topic(s)
1.4 Gamma Function Probability distributions
2.1 Discrete Distributions Probability distributions, Expectation, Generating functions
2.2 Continuous distributions Probability distributions, Expectation, Generating functions
2.3 Compound Distributions Conditional probability, Functions of random variables
3.1 Sample Mean and Variance Estimation
3.3 Maximum Likelihood Estimators Estimation
STAT3001 Statistical Modelling

Section Topic(s)
1.1 Series Time Series
1.3 Solving Equations Time Series
1.4 Gamma Function Bayesian statistics
1.5 Bayes’ Formula Bayesian statistics
2.1 Discrete Distributions Bayesian statistics, GLMs
2.2 Continuous distributions Bayesian statistics, GLMs
3.3 Maximum Likelihood Estimators Bayesian statistics, GLMs
3.6 Generalised Linear Models GLMs
3.7 Bayesian Methods Bayesian statistics
7.3 Monte Carlo Methods Monte Carlo simulation
STAT3002 Risk Analysis and Credibility Theory

Section Topic(s)
1.1 Series Ruin theory
1.2 Calculus Ruin theory, Reinsurance
1.4 Gamma Function Loss distributions, Reinsurance, Credibility theory
2.1 Discrete Distributions Credibility theory, Risk Models, Ruin theory
2.2 Continuous distributions Loss distributions, Reinsurance, Credibility theory, Risk models, Ruin theory
2.3 Compound Distributions Risk models, Ruin theory
2.4 Truncated Moments Loss distributions, Reinsurance
3.3 Maximum Likelihood Estimators Loss distributions, Credibility theory
3.7 Bayesian Methods Credibility theory
3.8 Empirical Bayes Credibility – Model 1 EBCT
3.9 Empirical Bayes Credibility – Model 2 EBCT
• Piosson Probabilities

STAT3005 Stochastic Processes

Section Topic(s)
2.1 Discrete Distributions Stochastic processes
2.2 Continuous distributions Stochastic processes
5.4 General Markov Model Markov Jump Processes
7.1 Markov “Jump” Processes Markov Jump Processes
STAT3006 Investment Science 1

Section Topic(s)
1.2 Calculus Stochastic calculus
1.3 Solving Equations Stochastic calculus
2.2 Continuous distributions Stochastic models of security prices
2.3 Compound Distributions Brownian motion and martingales
3.4 Linear Regression Model With Normal Errors Asset pricing models
7.2 Brownian Motion and Related Processes Brownian motion and martingales, Stochastic calculus
9.1 Utility Theory Utility theory
9.2 Capital Asset Pricing Model (CAPM) Asset pricing models
10.3 Stochastic Differential Equations Stochastic calculus
10.4 Black-Scholes Formulae for European Options Stochastic models of security prices
STAT3007 Investment Science 2

Section Topic(s)
1.3 Solving Equations The term structure of interest rates
2.1 Discrete Distributions The Binomial model
2.2 Continuous distributions Black-Scholes, 5-step Method
2.3 Compound Distributions 5-step Method
2.4 Truncated Moments 5-step Method
7.2 Brownian Motion and Related Processes The term structure of interest rates, 5-step Method
9.3 Interest Rate Models The term structure of interest rates
10.1 Price of a Forward or Futures Contract Derivative securities
10.2 Binomial Pricing (“Tree” Model) The binomial model
10.3 Stochastic Differential Equations Black-Scholes, 5-step Method, Term structure of interest rates
10.4 Black-Scholes Formulae for European Options Black-Scholes, Credit risk
10.5 Put-Call Parity Relationship ALL
• Binomial Probabilities

A Summary of Actuarial Formulae and Tables: Curtin Student Actuarial Society Last Updated: February 20, 2019

Uploaded by

Copyright:

Available Formats

A Summary of Actuarial Formulae and Tables: Curtin Student Actuarial Society Last Updated: February 20, 2019

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Summary of Actuarial Formulae and Tables: Curtin Student Actuarial Society Last Updated: February 20, 2019

Uploaded by

Copyright:

Available Formats

What mathematical and statistical methods are covered in the document?

What mathematical and statistical methods are covered in the document?

What actuarial topics are discussed in the document?

What actuarial topics are discussed in the document?

A Summary of Actuarial Formulae and Tables

Curtin Student Actuarial Society

B Non-Statistical (Mostly Life) Tables 41

D Reference by particular units 48

• Only approved calculators allowed

1 Updated : February 20, 2019

CT subject Curtin Exemption Unit Actuarial Actuarial Separate statistical Run by

2 Updated : February 20, 2019

3 Updated : February 20, 2019

• STAT3002 Risk Analysis and Credibility Theory

Extension 1.1.1.1: Exponential Function Using Sigma Notation

Formula 1.1.2: Natural log Function

Formula 1.1.3: Binomial expansion

p(p − 1) 2 p(p − 1)(p − 2) 3

• STAT3006 Investment Science 1

4 Updated : February 20, 2019

Formula 1.2.2: Taylor series (two variables)

This formula is covered in MATH2009 Calculus 2*.

Formula 1.2.3: Integration by parts

Formula 1.2.4: Double integrals (changing the order of integration)

• MATH3007 Life Contingencies 2

Formula 1.2.5: Differentiating an integral

1.3 Solving Equations

Formula 1.3.1: Newton-Raphson method

5 Updated : February 20, 2019

Formula 1.3.2: Integrating factors

This formula is explained in MATH1016 Calculus 1* or MATH1018 Accelerated Mathematics 2*.

Formula 1.3.3: Second-order difference equations

if b2 − 4ac > 0 : xn = Aλn1 + Bλn2 (distinct real roots, λ1 6= λ2 )

if b2 − 4ac = 0 : xn = (A + Bn)λn (equal real roots, λ1 = λ2 = λ)

if b2 − 4ac < 0 : xn = rn (A cos nθ + B sin nθ) (complex roots, λ1 = λ2 = reiθ )

where λ1 and λ2 are the roots of the quadratic equation aλ2 + bλ = 0.

1.4 Gamma Function

Formula 1.4.2: Properties

6 Updated : February 20, 2019

• STAT3001 Statistical Modelling

• STAT3002 Risk Analysis and Credibility Theory

1.5 Bayes’ Formula

Formula 1.5.1: Bayes’ formula

Let A1 , A2 , . . . , An be a collection of mutually exclusive and exhaustive events with P (Ai ) 6= 0, i = 1, 2, . . . , n.

For any event B such that P (B) 6= 0:

PF = Probability function, p(x)

• Referring to the “probability function” in 2.0.1 as a “probability mass function” (PMF).

Extension 2.0.1.2: Definitions that correspond to these conventions

• The definition of a PF is p(x) := P (X = x).

• The definition of a DF is F (x) := P (X ≤ x).

• The definition of a PGF is G(s) := E(sX ).

7 Updated : February 20, 2019

Extension 2.0.1.3: Moment generation using the MGF

If M (t) is r times differentiable at t = 0, then

Extension 2.0.1.4: Cumulant Generating Function

The cumulant generating function (CGF) is defined as

C(t) := log M (t)

2.1 Discrete Distributions