Risk Theory is the intellectual backbone of the insurance industry. It is the mathematical study of the unknown—specifically, the frequency and severity of claims that could potentially bankrupt an insurer. For actuarial students, this unit represents the transition from basic probability to “Survival of the Fittest” mathematics. You aren’t just calculating the odds of an event; you are modeling the entire financial ruin of a company. It is a rigorous, high-stakes subject that demands absolute precision.
Below is the exam paper download link
PDF Past Paper On Risk Theory For Actuarial Science For Revision
Above is the exam paper download link
To help you prepare for the complex distributions and ruin probabilities found in professional examinations, we have synthesized the core logic of Risk Theory into a high-impact revision guide.
What is the ‘Collective Risk Model’?
In the Collective Risk Model, we look at the total claims ($S$) produced by a portfolio over a specific time period. Unlike individual models that track every policyholder separately, the collective model treats the portfolio as a single engine generating a random number of claims ($N$), each with a random size ($X$). The total claim amount is represented as:
In an exam, you will often be asked to calculate the mean and variance of $S$ using the laws of total expectation and total variance.
How do we define ‘Ruin Theory’?
Ruin Theory is the study of the “Surplus Process.” An insurance company starts with initial capital ($u$), collects premiums at a constant rate ($c$), and pays out random claims ($S$) over time. Ruin occurs if the surplus ever drops below zero. Your revision should focus on the Lundberg Inequality, which provides an upper bound for the probability of ruin ($\psi(u)$) based on the “adjustment coefficient.”
What is the ‘Poisson Process’ in claim frequency?
The Poisson process is the standard way to model how often claims arrive. It assumes that claims happen independently and at a constant average rate ($\lambda$). A key property to remember for your finals is that the time between claims in a Poisson process follows an Exponential Distribution. If the rate $\lambda$ changes over time, you are dealing with a “Non-Homogeneous Poisson Process,” a frequent favorite for advanced questions.
What are ‘Heavy-Tailed’ Distributions?
In Risk Theory, not all distributions are created equal. While a Normal distribution is predictable, Heavy-Tailed distributions (like the Pareto or Log-normal) are dangerous for insurers. They allow for “Black Swan” events—rare but massive claims that can wipe out reserves. In a past paper, you might be asked to justify why a Pareto distribution is more appropriate for catastrophe insurance than a Gamma distribution.
How does ‘Reinsurance’ affect the risk profile?
Reinsurance is insurance for insurance companies. There are two main types you must master:
-
Proportional (Quota Share): The reinsurer takes a fixed percentage of every claim.
-
Non-Proportional (Excess of Loss): The reinsurer only pays if a single claim exceeds a certain “retention” level.
From a Risk Theory perspective, Excess of Loss reinsurance is more effective at reducing the “variance” of the insurer’s total claims, even though it can be more complex to price.
What is the ‘Adjustment Coefficient’ ($R$)?
The adjustment coefficient is a measure of the “safety” of an insurance portfolio. The larger the $R$, the lower the probability of ruin. It is the unique positive solution to the equation involving the moment-generating function of the claim size and the premium loading factor. If an examiner asks you how to decrease the probability of ruin, your mathematical answer usually involves increasing $R$ by raising premiums or buying more reinsurance.
Conclusion
Risk Theory for Actuarial Science is a discipline where “almost correct” is not good enough. It requires you to balance the technicality of compound distributions with the practical reality of financial solvency. The best way to build your “actuarial intuition” is to solve the multi-layered problems found in previous sittings, where one small error in a claim distribution can lead to a massive error in ruin probability.
To help you secure your path toward qualification, we have provided a link to a full set of practice materials below.
Last updated on: March 24, 2026