Allowing for risk and uncertainty in cash flows

Introduction: Navigating the "Wobble" in Financial Planning

Welcome! In your actuarial journey so far, you have spent a lot of time calculating the present value of future cash flows. Usually, you are given a set of "expected" numbers and told to get to work. But in the real world—and in the CP1 exam—nothing is certain. We don't know exactly when someone will die, how many cars will crash, or what the inflation rate will be in ten years.

This chapter is all about how we move from a simple "best guess" to a sophisticated valuation that allows for risk and uncertainty. This is a core skill in the Pricing and valuation of liabilities section because if we ignore the possibility of things going wrong, the providers of financial products could run out of money. Don't worry if the distinction between "risk" and "uncertainty" feels a bit blurry at first—we will break it down step-by-step!

1. Risk vs. Uncertainty: What’s the Difference?

In actuarial practice, we often use these two terms interchangeably in conversation, but they have distinct meanings when we are building models:

• Risk: This refers to variability where we can estimate the probability distribution of the outcomes. For example, we might not know if a specific person will die this year, but based on a large data set, we have a very good idea of the probability that they will.
• Uncertainty: This is "the unknown unknowns." It occurs when we cannot accurately predict the probability of an event or even what the possible outcomes might be. This is often due to a lack of data or a rapidly changing environment (like a new law being passed or a sudden technological shift).

Analogy: Risk is like rolling a six-sided die; you don't know the result, but you know the odds. Uncertainty is like rolling a mysterious object you’ve never seen before—you don’t even know how many sides it has!

2. Methods for Allowing for Risk in Cash Flows

When we value a liability, we want to ensure the provisions (the money set aside) are sufficient. There are three main ways to build "protection" into our cash flow projections:

A. Probability-Weighted Cash Flows (Expected Values)

Instead of assuming one fixed outcome, we look at all possible outcomes and weight them by their likelihood. This gives us the expected value of the cash flow.

The formula for the expected cash flow at time \( t \) is:
\( E[CF_t] = \sum (P_i \times CF_{i,t}) \)

Where \( P_i \) is the probability of scenario \( i \) occurring, and \( CF_{i,t} \) is the cash flow in that scenario.

B. Using Prudent "Best Estimate" Assumptions

We start with our best estimate (the most likely outcome) and then adjust the assumptions to be more prudent. This means we assume things will be slightly worse than we expect.

• For expenses, a prudent assumption would be a higher cost than expected.
• For interest rates (when valuing liabilities), a prudent assumption is usually a lower rate.
• For mortality in a life insurance contract, a prudent assumption is higher death rates.

C. Adding a Margin for Adverse Deviation (MAD)

This is a "buffer" added on top of the best estimate. If your best estimate for a claim is \( \$1,000 \), you might add a 10% margin and value it at \( \$1,100 \). This ensures that even if experience is slightly worse than the best estimate, the provider remains solvent.

Key Takeaway: We allow for risk by either weighting outcomes by probability or by being intentionally "pessimistic" (prudent) with our assumptions.

3. Allowing for Uncertainty in Present Values

Once we have our cash flows, we have to "pull them back" to today's value using a discount rate. The choice of discount rate is a major tool for managing uncertainty.

The Risk-Adjusted Discount Rate (RADR)

We can adjust the discount rate \( i \) to reflect the riskiness of the cash flows:

• In pricing, if a project is very risky, we might use a higher discount rate to ensure the potential return justifies the risk.
• In valuing liabilities, to be prudent, we often use a lower discount rate. A lower \( i \) results in a higher Present Value (PV), meaning we set aside more money today to be safe.

The basic Present Value formula remains our foundation:
\( PV = \sum_{t=1}^{n} \frac{CF_t}{(1 + i)^t} \)

Quick Review: If you want to be safe (prudent) when calculating how much money you need to save for a future bill, do you assume a high investment return or a low one? You assume a low one! That is why a lower discount rate allows for risk in liability valuation.

4. Deterministic vs. Stochastic Approaches

How we calculate these values depends on the complexity of the risk.

Deterministic Modelling

This uses fixed "point" estimates for assumptions. You plug in one set of numbers and get one result. To allow for risk here, you manually change the inputs (e.g., "What if inflation is 5% instead of 3%?").

Stochastic Modelling

Instead of one fixed number, we use probability distributions for the inputs. The model runs thousands of times (using techniques like Monte Carlo simulation) to produce a range of possible outcomes.

• It helps us see the "tail risks" (low-likelihood but high-impact events).
• It is essential for valuing options and guarantees (which we cover in a later chapter).
• It allows us to say, "We are 95% confident that the liability will not exceed \( X \)."

5. Tools for Testing Uncertainty

Even the best model can be wrong. We use these techniques to see how sensitive our "solution" is to changes in the world:

• Sensitivity Analysis: We change one assumption at a time (e.g., increase the discount rate by 1%) to see how much the total value changes. This identifies which risks are the most "dangerous" to our valuation.
• Scenario Analysis: We change multiple assumptions at once to represent a specific event. For example, a "Global Recession" scenario might involve low interest rates, high defaults, and high inflation all at the same time.
• Stress-Testing: We push an assumption to an extreme "breaking point" to see if the provider stays solvent.

Common Mistake: Students often confuse sensitivity and scenario analysis. Remember: Sensitivity = 1 variable changes. Scenario = A whole "story" or event where multiple variables change together.

Summary of Key Concepts

The Goal: To ensure the value of liabilities is high enough to cover future benefits, even if things go wrong.

Methods to Remember:
1. Adjust the Cash Flows: Use probability weighting or add prudent margins to the assumptions.
2. Adjust the Discount Rate: Use a lower rate for prudent liability valuation.
3. Use Stochastic Models: To understand the full distribution of risks, especially for complex products.
4. Test the Result: Use sensitivity and scenario analysis to see what happens when the environment changes.

Did you know? Many regulatory regimes (like those discussed in the General business environment section) actually require actuaries to include these margins to protect the public and ensure that "customers are treated fairly" by making sure the company doesn't go bust!

Cross-reference: While this chapter focuses on the techniques of allowing for risk, the specific market-consistent way of doing this is covered in the chapter on Fair and market-consistent valuation.