Welcome to Ruin Theory!

Welcome to one of the most important parts of the CM2: Economic Modelling course! In the section on Liability Valuations, we don't just care about how much money an insurance company makes; we care deeply about whether they stay in business. Ruin Theory is the mathematical way of asking: "How much money do we need to keep in the bank to make sure we don't go broke?"

Don’t worry if the math looks intimidating at first. We are essentially just balancing a checkbook where the income is steady but the expenses (claims) are unpredictable. Let's break it down together!

1. The Surplus Process: The "Money in the Bank" Equation

At the heart of Ruin Theory is the surplus process. Think of this as the balance of the insurance company's bank account at any time \( t \).

The standard formula for the surplus at time \( t \), denoted as \( U(t) \), is:

\( U(t) = u + ct - S(t) \)

Let's break down these pieces:

  • \( u \): The initial surplus. This is the amount of cash you start with on Day 1.
  • \( c \): The premium income rate. This is the money flowing in from policyholders. We usually assume this is a constant stream.
  • \( t \): The time elapsed.
  • \( S(t) \): The aggregate claims up to time \( t \). This is the money flowing out. This is the "random" part because we don't know exactly when claims will happen or how big they will be.

Quick Review: If \( U(t) \) ever drops below zero, "ruin" has occurred. It doesn't matter if the company gets a massive premium payment the next day—once you hit zero, the game is technically over!

Analogy: The Leaky Bathtub

Imagine a bathtub. You are filling it with a steady stream of water (Premium \( c \)). However, there is a drain that occasionally opens and lets out random amounts of water (Claims \( S(t) \)). The amount of water currently in the tub is your Initial Surplus \( u \). Ruin happens if the tub ever becomes completely empty.

2. Defining Ruin

In your exams, you will encounter two main types of ruin probabilities:

1. Finite Time Ruin \( \psi(u, T) \): This is the probability that ruin occurs at any time between now and a fixed future date \( T \).
2. Infinite Time Ruin \( \psi(u) \): This is the probability that ruin occurs at any point in the future, even 100 years from now.

Key Point: Naturally, \( \psi(u) \) (infinite time) will always be greater than or equal to \( \psi(u, T) \) (finite time). Why? Because there's more time for things to go wrong!

3. The Adjustment Coefficient \( R \)

The adjustment coefficient (often denoted by the letter \( R \)) is a very special number that measures the "safety" of the insurance portfolio. It depends on the size of the claims and the premium being charged.

For a Poisson Process (where claims arrive randomly), \( R \) is the unique positive solution to the following equation:

\( 1 + (1+\theta)\mu r = M_X(r) \)

Where:

  • \( \theta \): The relative security loading. This is the "extra" bit of premium you charge above the expected claims to cover risk.
  • \( \mu \): The mean of an individual claim.
  • \( M_X(r) \): The Moment Generating Function (MGF) of the claim size distribution.

Why do we care about \( R \)?
The larger \( R \) is, the safer the company is. A high \( R \) means the ruin probability is low. Think of \( R \) as the "strength" of your financial shield.

Did you know? If the claim sizes follow an Exponential distribution with mean \( 1/\alpha \), the adjustment coefficient is much easier to calculate: \( R = \alpha - \frac{\alpha}{1+\theta} \).

4. Lundberg’s Inequality

Calculating the exact probability of ruin is often incredibly difficult. This is where Lundberg’s Inequality comes to the rescue! It gives us a "worst-case scenario" estimate.

The formula is:

\( \psi(u) \leq e^{-Ru} \)

This tells us that the probability of ruin is always less than or equal to \( e \) raised to the power of negative \( R \) times the initial surplus \( u \).

Common Mistake: Students often forget that this is an upper bound, not the exact probability. It tells you the ruin probability won't be higher than this value, but it could be lower!

The "Logic" Trick: Look at the formula. If you increase your starting cash (\( u \)), the right side of the equation gets smaller. This makes sense: more starting money equals less chance of going broke!

5. The Impact of Reinsurance

In the context of Liability Valuations, insurers use reinsurance to protect themselves. Reinsurance changes the ruin probability by altering the "net" claims the insurer has to pay.

Proportional Reinsurance (Quota Share):
The insurer keeps a percentage \( \alpha \) of every claim and every premium. While this reduces the total dollar amount of risk, it doesn't always change the ruin probability in the way you might expect, because the premium income drops at the same rate as the claims.

Excess of Loss Reinsurance:
The reinsurer pays for any claim amount above a certain "retention level" \( M \). This is very effective at cutting out "tail risk" (huge claims). Usually, adding Excess of Loss reinsurance increases the adjustment coefficient \( R \), which lowers the ruin probability.

Key Takeaway: Reinsurance is like buying an umbrella. It costs a bit of your premium income, but it keeps you dry when the "storm" of large claims hits.

6. Summary and Quick Review

Let's recap the most vital points for your revision:

  • The Surplus Equation: \( U(t) = u + ct - S(t) \). Keep this memorized!
  • Ruin definition: Ruin occurs the moment \( U(t) < 0 \).
  • The Role of \( u \): Increasing initial capital \( u \) exponentially decreases the ruin probability.
  • Adjustment Coefficient \( R \): It is the "index of safety." Find it using the MGF of claim sizes.
  • Lundberg's Inequality: \( \psi(u) \leq e^{-Ru} \). This is your best friend for finding maximum risk levels.
  • Security Loading (\( \theta \)): You must have \( \theta > 0 \) (i.e., you must charge more than the expected claims) for the probability of ruin to be less than 100% in the long run.

Quick Tip for the Exam: If a question asks you to calculate the effect of a change in premium on the ruin probability, first find the new adjustment coefficient \( R \), then plug it into Lundberg’s Inequality.

Don't worry if this seems tricky at first! Ruin theory is all about understanding the relationship between the money coming in and the randomness of the money going out. Practice a few MGF calculations, and you'll be a pro in no time!