Week 1

1. Simple interest !:
" = $ + & = $ + $!' = $(1 + !')
"
⟺ $=
1 + !'
2. Simple discount -:
$ = " − / = " − "-' = "(1 − -')
$
⟺ "=
1 − -'
3. With same " & same $, conversion between ! and -:
$
" = $(1 + !') =
1 − -'
1
1 + !' =
1 − -'
- !
!= ⟺ -=
1 − -' 1 + !'
4. Simple interest vs compound interest:
- (1 + !') ≥ (1 + !)! when ' ≤ 1,
- (1 + !') < (1 + !)! when ' > 1.
 Week 2

1. Compound interest: i e↵ective per year

• (1 + i)n > (1 + in) for n > 1

• A = P (1 + i)n () P = A(1 + i) n = Av n where v = (1 + i) 1

2. Nominal interest rate: i(m) payable / convertible m-thly per year

i(m)
• equivalent to m e↵ective per 1
m year
i(m) i(m)
• A = P (1 + m )mn () P = A(1 + m )
mn

3. Force of interest: continuous compounding at rate = limm!1 i(m)

• A = Pe n () P = Ae n

4. With the same P and same A:

⇣ i(m) ⌘mn
(1 + i)n = 1 + =e n
m
⇣ 1
⌘
i(m) = m (1 + i) m 1 = m(e m 1)
i(m)
= log(1 + i) = m log 1 +
m
1 ⇣ i(m) ⌘ m
v= = 1+ =e
1+i m

1
 Week 3

1 Annuities
1. Annuity in arrears:
1 vn
an = v + v2 + v3 + · · · + vn =
i
• 1 = ia n + v n
• Reasonableness check:
(i) a n  n
(ii) a n  1i
n+1
(iii) a n ⇡ nv 2

2. Annuity in advance:

ä n = 1 + v + v 2 + v 3 + · · · + v n 1
= (1 + i)a n = 1 + a n 1

3. Deferred annuity:

m|a n = v m+1 + v m+2 + · · · + v m+n = v m a n

4. m-thly annuity:
(m) 1 1 2 mn 1 vn 1
an = v m + v m + · · · + v m )= (m) = ajmn
m i m
(m) 1 1 2 mn 1 1 (m) (m)
ä n = 1 + v m + v m + · · · + v m )= (1 + i) m a n = (1 + j)a n
m

5. With the same e↵ective annual interest rate,

(m) (m)
ä n > ä n > an > an

2 Accumulated Values of Annuities

1. Accumulated value of annuity in arrears:
(1 + i)n 1
s n = (1 + i)n 1
+ (1 + i)n 2
+ ··· + 1 = = (1 + i)n a n
i

2. Accumulated value of annuity in advance:

(1 + i)n 1
s̈ n = (1 + i)n + (1 + i)n 1
+ · · · + (1 + i) = (1 + i) = (1 + i)s n
i
= (1 + i)n+1 a n = (1 + i)n ä n

1
 Week 4

1 Treasury Bonds

1. Treasury bonds:

• Coupon rate per payment period is C

F fixed.
• If P = F , the bond is sold at par, yield = coupon rate.
If P < F , the bond is sold at a discount, yield > coupon rate.
If P > F , the bond is sold at a premium, yield < coupon rate.
• Annual rate of return from this investment is called yield to maturity (YTM).
Yield per period depends on the price.
• Bond price:
P = Cajn + F (1 + j) n

2. Estimating bond yields

C F P
(i) Guess the yield by reasoning: j ⇡ P + Pn
(ii) Find two estimated values of j, j1 and j2 , such that P (j1 ) > 0, P (j2 ) < 0
(iii) Use linear interpolation to refine estimate of j

2 Housing Loans

1. The amount of principal / loan outstanding (after payment):

L0 = P = Xa n
Lt = Xa n t = P (1 + i)t Xs t (0 < t < n)
Lt+k = Lt (1 + i)k Xs k (t + k  n)
Ln = 0

2. Decomposition of instalments:

X = I0 + C 0 = · · · = It + C t = · · · = In + C n
It+1 = i ⇥ Lt (i = e↵ective rate per payment period)
Ct+1 = X It+1
Lt+1 = Lt Ct+1

1
 Week 5

1. Summary statistics of a population:

• Sex ratio: number of males

number of females
• Child-woman ratio: 100 ⇥ number of children under 5
number of women aged 15 49

• Total dependency ratio: 100⇥ number of children aged (0,14] + number of seniors aged 65+
number of individuals aged [15, 64]
number of children aged (0,14]
Youth dependency ratio: 100 ⇥ number of individuals aged [15, 64]
number of seniors aged 65+
Age dependency ratio: 100 ⇥ number of individuals aged [15, 64]

2. Change in population:

P (t + 1) = P (t) + Births(t, t + 1) Deaths(t, t + 1) + Immigrants(t, t + 1) Emigrants(t, t + 1)

= P (t) + natural increase + net migration

3. Crude birth/death rates:

total number of births/deaths

1000 ⇥
population size

4. Specific death rates:

total number of deaths from the specific group
death rate for a specific group = 1000⇥
total population of the specific group

5. Standardized death rates: apply specific rates for both populations to the reference
population, compare the standardized crude rates.

6. Population (age-gender) pyramids:

• Triangular-shaped: undeveloped countries, rapid population growth

• Beehive-shaped: developed countries, slow population growth
• Rectangular-shaped: aging, stationary population
• Upside-down triangular-shaped: shrinking population

1
 Week 6

1. Survival function: s(t) = P (T (0) > t),

• s(0)=1
• limt!1 s(t) = 0
• s0 (t)  0

2. Lifetime distribution: F (t) = P (T (0)  x) = 1 s(t),

• F(0)=0
• limt!1 F (t) = 1
• F 0 (t) 0

3. Future lifetime:
⇣ ⌘ ⇣ ⌘ s(x + t)
P T (x) > t = P T (0) > x + t | T (0) > x =
s(x)
⇣ ⌘ ⇣ ⌘ s(x + t)
P T (x)  t = 1 P T (x) > t = 1
s(x)
⇣ ⌘ ⇣ ⌘ ⇣ ⌘ s(x + t) s(x + t + r)
P t < T (x)  t + r = P T (x) > t P T (x) > t + r =
s(x)

4. Force of mortality / Hazard rate:

P (T (x)  h) s0 (x)
µx = lim =
h!0 h s(x)

Approximation: P T (x)  h ⇡ hµx when h is very small

5. Survival function in terms of force of mortality:

⇢ Z t
P (T (0) > t) = exp µx dx
0
⇢ Z y+t ⇢ Z t
P (T (y) > t) = exp µx dx = exp µx+r dr
y 0

6. Life table functions

• Life table elements:

– l0 : radix of the table
– lx : expected number of lives surviving to age x out of l0 births
– dx : expected number of deaths at ages x out of lx lives,

dx = lx lx+1
1
X 1
X
l0 = d x , lx = dy
x=0 y=x

1
 – qx : mortality rate, probability that a life aged x dies before x + 1,
dx
qx =
lx
– px : probability that a life aged x survives to x + 1,
lx+1
px = 1 qx =
lx
• Actuarial notations of survival probability and lifetime distribution:
t 1
s(x + t) Y lx+t
t px = P (T (x) > t) = = px+r =
s(x) lx
r=0
s(x) s(x + t) lx lx+t
t qx =1 t px = P (T (x)  t) = =
s(x) lx

• Life expectancy / complete expectation of life:

Z 1 Z 1 ⇢ Z t
e̊x = t px dt = exp µx+r dr dt
0 0 0

• Average number of survivors between age x and x + 1:

Z x+1 Z 1 Z 1
lx + lx+1 dx
Lx = ly dy = lx+t dt = lx t px dt ⇡ or lx
x 0 0 2 2

• Expected total future life time of a group of lx lives:

Z 1 Z 1 1
X
Tx = ly dy = lx+t dt = Ly = Lx + Tx+1 = lx e̊x
x 0 y=x

7. Stationary population: number of entrants = number of exits, population size and

mortality rate are unchanged for every age group.

• Lx is the population size between ages x and x + 1 at any time

• Tx is the population size above age x at any time during the stationary period

2
 Week 7

1. Mathematical models of mortality:

• Gompertz law: µx = Bcx where B > 0, c > 1
• Makeham’s law: µx = A + Bcx where A, B > 0, c > 1
2. Curtate future life time at age x: K(x) is the number of whole years of future life
time from age x.
P (K(x) = k) = k px k+1 px
Curtate expectation of life at age x:
1
X 1
X 1
X
ex = E[K(x)] = kP (K(x) = k) = k k px k+1 px = k px
k=0 k=0 k=1

3. Fertility rates
• General fertility rate:
number of live births
GFR = 1000 ⇥
number of women aged [15, 49]
• Age specific fertility rate:
number of live births by women aged [x, x + 1)
ASFRx = 1000 ⇥
number of women aged [x, x + 1)
number of live births by women aged [a, b]
ASFR[a,b] = 1000 ⇥
number of women aged [a, b]
b
X wx
= Pb ASFRx
x=a x=a wx

• Total fertility rate:

49
X
TFR = ASFRx
x=15
• Gross reproduction rate:
49
X 1
GRR = ASFRfx = TFR
2.05
x=15

• Net reproduction rate:

49
X 49
lx+0.5 1 X lx+0.5
NRR = ASFRfx = ASFRx
l0 2.05 l0
x=15 x=15

4. Population projections
P
• Polynomial model: Pt = P0 + ni=1 ai ti (linear: Pt = P0 (1 + rt))
• Geometric population growth: Pt = P0 (1 + r)t
• 1
Logistic model: Pt = A+Be rt

• Component method: Pt = P0 + natural increase + net migration

1
 Week 8

P
1. Expected Present Value: EPV = i Amounti ⇥ Discount factori ⇥ Probi

• EPV of a series of payments X1 , X2 , . . . , Xn made at times 1, 2, . . . , n:

n
X ⇣ k
X ⌘
EPV = ⇡k Vk ⇡k = P (k payments are made exactly), Vk = v j Xj
k=1 j=1
Xn ⇣ ⌘
= v j X j pj pj = P (kth payment is made)
j=1

1
 Week 9

1. Types of life insurance: Term insurance, whole life insurance, endowment insurance.

2. Principle of equivalence: the EPV of all premium income at the time of issuing the
policy = EPV of future benefits and expenses.

3. Reserving: t V = EPV of future benefits and expenses at t - EPV of future premiums

at t

4. 3 main sources of profit:

• Interest surplus
• Mortality profit
• Lower actual expenses than assumed

5. Whole life insurance: death benefit payable on the death of the policyholder,
essentially term insurance with an infinite term.

P äx = SAx
P
• äx = 1 t=0 v · t px is the EPV of $1 paid annually in advance by a life (x)
t

contingent on survival
P P1 t
• Ax = 1 t
t=1 v · t 1 px qx+t 1 = t=1 v (t 1 px t px ) is the EPV of $1 paid at
the end of the year of death of (x)

1
 Week 10

1. Whole life insurance: P äx = SAx

• EPV of $1 paid annually in advance contingent on (x)’s survival:
1
X
äx = v t · t px
t=0

• EPV of $1 paid at the end of the year of (x)’s death:

1
X 1
X 1
X
t t K(x)+1 dx+t
Ax = v ·t 1 px qx+t 1 = v (t 1 px t px ) = E[v ]= v t+1
lx
t=1 t=1 t=0

2. Endowment insurance: P äx:n = Ax:n S

• EPV of $1 paid annually in advance contingent on (x)’s survival, for a maximum
of n payments.
n
X1
äx:n = v t t px
t=0

• EPV of $1 payable at the end of the year of (x)’s death if death occurs within
n years, or at time n if (x) survives for n years.
n
X1 dx+t 1
Ax:n = v t+1 + v n n px = Ax:n
1 + Ax: n
lx
t=0
1
3. Term insurance: P äx:n = S Ax:n
4. Recursive formula for äx :
äx = 1 + v · px äx+1

5. Premium conversion relationships:

Ax:n + (1 v)äx:n = 1
Ax + (1 v)äx = 1

By the principle of equivalence:

S Ax:n ⇣ 1 ⌘ S Ax:n
P = =S (1 v) = (1 v)
äx:n äx:n 1 Ax:n

6. Reserving: t V = EPV (expenses + benefits) - EPV (premiums)

• Endowment: t V = S Ax+t:n t P äx+t:n t (t = 1, 2, . . . , n 1)
– 0 V = 0, n V = 0
– n 1 V = Sv P
• Term: t V = S Ax+t:n
1
t P äx+t:n t (t = 1, 2, . . . , n 1)
• Whole life: t V = SAx+t P äx+t (t = 1, 2, . . . )

1
 Week 11

[email protected]

1. Profit analysis for a life insurance policy: define random variable L0 to be the PV
of future loss:
⇣ 1 + i⌘ 1+i
L0 = Sv K(x)+1 P ä K(x)+1 = v K(x)+1 S + P P
i i
The insurer makes a profit if L0 < 0, which is when
ln Ax
K(x) > 1
ln(1 + i)

2. Risk pooling: n policies are sold, premiums P1 , P2 , . . . , Pn collected, claims amounts

X1 , X2 , . . . , Xn , assume Pi > E[Xi ] for every policy.

(a) If the policies sold are identical, same premiums charged P1 = P2 = · · · =

Pn = P , and i.i.d claims X1 , X2 , . . . , Xn
• The Law of Large Numbers (LLN): for a sequence of i.i.d random variables
{Xn }1 2 , the sample mean Y = X1 +X2 +···+Xn
n=1 with mean µ and variance n n
converges to the distribution mean E[Yn ] = µ as n goes to 1

lim Yn = E[Yn ] = µ
n!1

• According to LLN,
⇣ n
X ⌘
lim P nP > Xi = 1
n!1
i=1

(b) If we let Pi and Xi be the gross premium and the aggregate claims amounts
from the ith insurance lines, Pi ’s have di↵erent values and Xi ’s are independent
but not identical, Pi > E[Xi ] = µi for every line. Then according to extended
LLN,
⇣X n Xn ⌘
lim P Pi > Xi = 1
n!1
i=1 i=1

3. Aggregate claim amount from a risk:

N
X
S= Xi
i=1

where N is the claim frequency, Xi ’s are claim severities.

1
4. Reinsurance: insurance to insurers.

Retained by Insurer Ceded to Reinsurer

Proportional aX (1 a)X
Stop loss min(X, M ) max(0, X M)

5. Superannuation:

Defined Contribution Defined Benefit

Contributions Defined percentage of salary Variable percentage of salary
Total balance of the account Guaranteed amount (either a
Benefits net of fees and taxes, depend lump sum or annuity), no
on investment performance investment risk to member