Collective and individual risk models

Exam FAM – Fundamentals of Actuarial Mathematics · SOA (Society of Actuaries) · 6 min read

Welcome to the World of Aggregate Models!

In your journey to becoming an actuary, one of the most important questions you will ever have to answer is: "How much money, in total, should we expect to pay out in claims?"

Up until now, you may have studied individual claim sizes (Severity) or how many claims happen (Frequency) separately. In this chapter, we bring them together to look at the Aggregate Loss—the "Big Total." We will explore two primary ways to model this: the Individual Risk Model and the Collective Risk Model. Don't worry if these sound a bit intimidating; we’ll break them down using simple analogies and step-by-step logic!

1. The Individual Risk Model

The Individual Risk Model is like looking at a guest list for a private party. You know exactly who is coming (the number of policies is fixed), and you look at each person individually to see if they will "cost" you anything.

What is it?

In this model, we assume we have a fixed number of insurance policies, let's call it $n$. For each policy $i$, there is a loss $X_i$. The total aggregate loss $S$ is simply the sum of all these individual losses:

$ S = X_1 + X_2 + ... + X_n $

Key Characteristics

Fixed Number of Risks: The value of $n$ is known and not random.
Independence: We usually assume that what happens to Policyholder A doesn't affect Policyholder B.
Variable Loss: Each $X_i$ can be 0 (no claim) or some positive amount.

Calculating the Mean and Variance

Because the number of policies is fixed, finding the mean and variance is straightforward:

Mean: $ E[S] = \sum_{i=1}^{n} E[X_i] $

Variance: $ Var[S] = \sum_{i=1}^{n} Var[X_i] $ (Assuming the policies are independent!)

Quick Review:

If you have 100 policies and each has an expected loss of \$50, your total expected loss $E[S]$ is $100 \times 50 = 5,000$.

Summary: The Individual Risk Model is great for groups where the number of participants is set, like a specific group life insurance plan for a company's employees.

2. The Collective Risk Model

The Collective Risk Model is more like a busy coffee shop. You don't know exactly how many people will walk through the door today (Frequency), and you don't know how much each person will spend (Severity). You view the "collection" of claims as a whole.

What is it?

Here, the total loss $S$ depends on two random things:

$N$: The number of claims (a random variable).
$X_i$: The amount of each claim (random variables).

The formula looks like this:

$ S = X_1 + X_2 + ... + X_N $

Notice the big difference: The "top" of the sum ($N$) is now a random variable itself!

Key Assumptions

The claim amounts $X_1, X_2, ...$ are identically distributed (they follow the same "severity" pattern).
The claim amounts are independent of each other.
The number of claims $N$ is independent of the size of the claims $X_i$.

The "Big Formulas" for Collective Models

These are high-priority for Exam FAM! You must memorize these:

Expected Value (Mean): $ E[S] = E[N] \times E[X] $

Variance: $ Var[S] = E[N] \times Var[X] + Var[N] \times (E[X])^2 $

Memory Aid: Think of the variance formula as having two "engines." The first part accounts for the uncertainty in how big the claims are, and the second part accounts for the uncertainty in how many claims there are.

Did you know?

The Variance formula $Var[S] = E[N]Var[X] + Var[N]E[X]^2$ is often called the Law of Total Variance in this context. It’s one of the most frequently tested concepts in the aggregate models section!

Summary: The Collective Risk Model is the "bread and butter" of casualty insurance, where we model the total claims of an entire book of business without tracking every single policyholder individually.

3. Comparing Individual and Collective Models

It can be confusing to decide which model to use. Here is a quick comparison to help you keep them straight:

Individual Risk Model:

Used when the number of potential claimants is small or fixed (e.g., 500 employees).
Each $X_i$ represents the loss from a specific person.
If a person doesn't claim, $X_i = 0$.

Collective Risk Model:

Used for large groups where claims happen "randomly" (e.g., all car insurance policies in a state).
$N$ represents the number of claims, and $X_i$ is the size of the $i$-th claim that actually occurred.
Each $X_i$ is usually $ > 0$.

Common Mistake: Don't confuse $n$ (a fixed number) with $N$ (a random variable). If the problem says "There are exactly 1,000 policies," you are likely looking at an Individual Risk Model. If the problem says "The number of claims follows a Poisson distribution," you are in the Collective Risk Model world.

4. Approximating the Aggregate Loss Distribution

Sometimes, calculating the exact distribution of $S$ is way too hard. In these cases, actuaries use approximations. This is where your knowledge of the Normal Distribution and Central Limit Theorem comes in handy.

The Normal Approximation

If the number of claims is large enough, we can assume that $S$ follows a Normal distribution. We use the mean ($E[S]$) and variance ($Var[S]$) we calculated earlier.

To find the probability that the total loss is less than a certain amount $k$:

$ P(S \le k) = P\left(Z \le \frac{k - E[S]}{\sqrt{Var[S]}}\right) $

The Lognormal Approximation

Since insurance losses are usually positive and often "skewed" (a few huge claims), the Normal distribution isn't always perfect. The Lognormal distribution is often used instead because it doesn't allow for negative losses and handles skewness better.

Step-by-Step: Using the Normal Approximation

Calculate $E[S]$ using the appropriate model (Individual or Collective).
Calculate $Var[S]$.
Find the Standard Deviation: $\sigma_S = \sqrt{Var[S]}$.
Standardize your target value $k$ to get a Z-score: $Z = \frac{k - E[S]}{\sigma_S}$.
Look up the Z-score in the standard normal table provided in the exam.

Key Takeaway: Approximations are your best friend on the exam when the math for an exact distribution gets too messy!

5. Summary and Tips for Success

Individual Model: Sum of a fixed number of independent risks.
Collective Model: Sum of a random number of i.i.d. risks.
Crucial Formula: $Var[S] = E[N]Var[X] + Var[N]E[X]^2$. If you see a Collective Risk problem, you will likely need this!
Units Matter: Always check if the question is asking for the mean of the frequency, the severity, or the aggregate loss.
Stay Calm: If a problem looks complex, start by identifying whether it's Individual or Collective. Once you've done that, write down the components you know ($E[N], Var[N], E[X], Var[X]$) and the rest will start to fall into place.

Don't worry if this seems tricky at first! Aggregate models are just a way of adding up smaller pieces. Once you master the formulas for the mean and variance, you've conquered the hardest part of the chapter.

Quick check

Can you answer these now?

Open each question to check the key ideas from this chapter.

What is defined as the Aggregate Loss in insurance modeling?

Aggregate Loss, often denoted as $ S $, represents the total sum of claim amounts paid out by an insurer over a specific period.

How is the Aggregate Loss expressed in the Individual Risk Model?

$ S = X_1 + X_2 + ... + X_n $, where $ n $ is a fixed number of potential risks or policies.

How are the mean and variance calculated for the Individual Risk Model?

The mean is $ E[S] = \sum_{i=1}^{n} E[X_i] $ and the variance is $ Var[S] = \sum_{i=1}^{n} Var[X_i] $, provided the individual risks are independent.

How is the Aggregate Loss expressed in the Collective Risk Model?

$ S = X_1 + X_2 + ... + X_N $, where $ N $ is a random variable representing the number of claims.

What are the three core independence assumptions in the Collective Risk Model?

The claim sizes $ X_i $ are i.i.d., and the frequency $ N $ is independent of the claim sizes $ X_i $.

What is the formula for the expected value of the aggregate loss in a Collective Risk Model?

$ E[S] = E[N] \times E[X] $

What is the variance formula for the Collective Risk Model?

$ Var[S] = E[N]Var[X] + Var[N](E[X])^2 $

How do you distinguish between the parameter n and the random variable N in exam problems?

Use the Individual Risk Model if the number of policies $ n $ is fixed; use the Collective Risk Model if the number of claims $ N $ follows a distribution (like Poisson or Negative Binomial).

When and how do you use the Normal Approximation for Aggregate Loss?

When the number of claims is large, standardize the target value using $ Z = \frac{k - E[S]}{\sqrt{Var[S]}} $.

Why is the Lognormal distribution often used for aggregate loss approximations?

It accounts for the fact that insurance losses are non-negative and often have a positive skew (a 'heavy tail').

Ready to test yourself?

Turn these notes into exam-style practice. Get unlimited AI questions on this topic with instant marking and explanations.

Practice This Topic

More Exam FAM – Fundamentals of Actuarial Mathematics chapters

* The content provided by thinka is generated by AI and may not always be accurate or up-to-date. Please use it as a supplementary resource and verify with official materials.

Put These Notes into Practice

Reading the notes is a great start. Now practise with unlimited AI-generated questions and get instant feedback. 100,000+ students are already improving their grades.

Start Practising Now View Pricing

Done reading? Test yourself with AI practice questions

Practice This Topic Now