1
rd
BS 3 Year (ACT 551)
Multiple state and multiple decrement models, problem of emerging costs
Multiple State and Multiple Decrement Models
Multiple State Models are used in actuarial science to represent the different states an individual can be in
over time, such as alive, dead, disabled, or retired. These models are extensions of the basic life-death
model and allow for more complex transitions between states.
Multiple Decrement Models are a specific type of multiple state model where individuals can exit a given
state via one of several competing risks (decrements). For example, a person might leave the "active
workforce" state due to death, disability, or retirement.
Problem of Emerging Costs
1. The problem of emerging costs refers to the challenge of estimating the future costs associated
with transitions between states in these models. These costs can include benefits paid, premiums
received, and other financial impacts. The complexity arises because:
2. Multiple States and Transitions: With more states and possible transitions, the number of potential
paths an individual can take increases, making it harder to predict future costs accurately.
3. Dependence on Transition Rates: The costs depend on the rates of transition between states (e.g.,
mortality rates, disability rates), which can be uncertain and variable over time.
4. Time Value of Money: Future costs need to be discounted to their present value, adding another
layer of complexity.
5. Policyholder Behavior: In some models, policyholder behavior (e.g., lapse rates, retirement
decisions) can significantly impact emerging costs, and this behavior can be difficult to predict.
Steps to Address the Problem
Model Specification: Clearly define the states and possible transitions in the model. For example, in a
pension plan, states might include "active," "retired," "disabled," and "dead."
Transition Probabilities: Estimate the probabilities of transitioning between states. This might involve
using historical data, actuarial tables, or statistical models.
Cost Identification: Identify the costs associated with each transition. For example, moving from
"active" to "retired" might trigger a pension benefit payment.
Discounting: Apply an appropriate discount rate to future costs to bring them to present value.
Simulation and Sensitivity Analysis: Use simulation techniques to model the range of possible outcomes
and conduct sensitivity analysis to understand how changes in assumptions impact the emerging costs.
Regulatory and Accounting Standards: Ensure that the model complies with relevant regulatory and
accounting standards, which may dictate specific methods for calculating reserves or capital
requirements.
Ms.Ubaida Riaz (Economist, AERC) Actuarial Science, Department of Statistics
2
rd
BS 3 Year (ACT 551)
Example
Consider a pension plan with the following states:
Active: The individual is currently working.
Retired: The individual has retired and is receiving a pension.
Disabled: The individual is disabled and receiving benefits.
Dead: The individual has died.
Transitions:
Active → Retired (retirement)
Active → Disabled (disability)
Active → Dead (death)
Disabled → Retired (recovery and retirement)
Disabled → Dead (death while disabled)
Retired → Dead (death after retirement)
Emerging Costs:
Retirement: Pension payments begin.
Disability: Disability benefits are paid.
Death: A lump sum or survivor benefits might be paid.
To estimate the emerging costs, you would:
Estimate the probabilities of each transition.
Project the cash flows associated with each transition.
Discount these cash flows to their present value.
Aggregate the costs over all possible paths.
Numerical Problem: Multiple State and Multiple Decrement Model
Scenario:
You are an actuary analyzing a pension plan with the following states:
Active (A): The individual is currently working.
Retired (R): The individual has retired and is receiving a pension.
Disabled (D): The individual is disabled and receiving benefits.
Dead (Dead): The individual has died.
Ms.Ubaida Riaz (Economist, AERC) Actuarial Science, Department of Statistics
3
rd
BS 3 Year (ACT 551)
The transitions and associated probabilities (annual) are as follows:
Active → Retired: pAR=0.05
Active → Disabled: pAD=0.02
Active → Dead: pA,Dead=0.01
Disabled → Retired: pDR=0.10
Disabled → Dead: pD,Dead=0.05
Retired → Dead: pR,Dead=0.03
Assume:
There are 1,000 active employees at time t=0.
The annual pension benefit for retirees is $10,000.
The annual disability benefit is $5,000.
A death benefit of $20,000 is paid at the end of the year of death.
The discount rate is 5% per annum.
Question:
Calculate the expected present value of future benefits over the next 2 years for this pension plan.
Solution:
We will calculate the expected number of individuals in each state at the end of each year, the associated
costs, and discount them to present value.
Step 1: Transition Probabilities and State Populations
At t=0:
Active: 1,000
Retired: 0
Disabled: 0
Dead: 0
At t=1:
Active → Retired: 1,000×0.05=50
Active → Disabled: 1,000×0.02=20
Ms.Ubaida Riaz (Economist, AERC) Actuarial Science, Department of Statistics
4
rd
BS 3 Year (ACT 551)
Active → Dead: 1,000×0.01=10
Remaining Active: 1,000−50−20−10=920
Disabled → Retired: 20×0.10=2
Disabled → Dead: 20×0.05=1
Remaining Disabled: 20−2−1=17
Retired → Dead: 50×0.03=1.5
Remaining Retired: 50−1.5=48.5
State Populations at t=1:
Active: 920
Retired: 48.5+2=50.5
Disabled: 17
Dead: 10+1+1.5=12.5
At t=2:
Active → Retired: 920×0.05=46
Active → Disabled: 920×0.02=18.4
Active → Dead: 920×0.01=9.2
Remaining Active: 920−46−18.4−9.2=846.4
Disabled → Retired: 17×0.10=1.7
Disabled → Dead: 17×0.05=0.85
Remaining Disabled: 17−1.7−0.85=14.45
Retired → Dead: 50.5×0.03=1.515
Remaining Retired: 50.5−1.515=48.98
State Populations at t=2:
Active: 846.4
Retired: 48.985+1.7=50.68
Disabled: 14.45
Dead: 12.5+9.2+0.85+1.515=24.06
Ms.Ubaida Riaz (Economist, AERC) Actuarial Science, Department of Statistics
5
rd
BS 3 Year (ACT 551)
Step 2: Calculate Costs
At t=1:
Retired Benefits: 50.5×10,000=505,0005
Disabled Benefits: 17×5,000=85,000
Death Benefits: 12.5×20,000=250,000
Total Costs at t=1: 505,000+85,000+250,000=840,000
At t=2:
Retired Benefits: 50.685×10,000=506,850
Disabled Benefits: 14.45×5,000=72,250
Death Benefits: 24.065×20,000=481,300
Total Costs at t=2: 506,850+72,250+481,300=1,060,400
Step 3: Discount Costs to Present Value
1
Discount Factor for t=1: 1.05 = 0.9524
1
Discount Factor for t=2: (1.05) 2
= 0.9070
Present Value of Costs:
t=1: 840,000 × 0.9524 = 800,016
t=2: 1,060,400 × 0.9070 = 961,682.81
The expected present value of future benefits over the next 2 years is approximately $1,761,699.
Summary of Differences
Term Whole Endowment Universal
Feature ULIP Group Life
Life Life Plans Life
Coverage Specific Varies (group-
Lifelong Specific term Lifelong Specific term
Duration term based)
Death Benefit Yes Yes Yes Yes Yes Yes
Maturity
No No Yes No Yes No
Benefit
Yes (market-
Cash Value No Yes Yes Yes No
linked)
Low (group
Premiums Low High Moderate Flexible Market-linked
rate)
Ms.Ubaida Riaz (Economist, AERC) Actuarial Science, Department of Statistics
