In my blog last week (A Financial Plan’s Two Most Important Assumptions) we talked about the role of Monte Carlo simulation (MCS) and how to use it to determine the most important assumptions in a financial plan.
This week we’ll discuss a list of the assumptions on which I use MCS. First, let’s look at an overview of the information required to create a comprehensive plan.
The information required to create a comprehensive financial plan can be categorized as follows:
Debt (balances > 12 months)
Current and retirement
Fixed (i.e. debt payments)
Current and retirement
- The Client’s Goals and Dreams
When gathering client data, some advisors prefer to use a legal pad and pen while others opt for a standardized questionnaire. You can easily use the first method if you follow an outline such as the one above. However, to be sure you capture all the necessary information, you might want to use a standardized questionnaire. I’ve used both methods and have developed my own standardized questionnaires. These are very detailed but I prefer to get into the weeds when planning. After all, the output will only be as good as the data you input.
You should consider using MCS on assumptions which contain the greatest variability. I have three categories for assumptions.
- Low Variability
- MCS Assumptions.
One example of a static assumption is a corporate pension. Other data may not be static but the degree of variability is low and adding MCS to these assumptions will have little impact on the results. The final category includes assumptions which are most suitable for MCS.
If the results don’t change much by using MCS on a particular input, then it probably doesn’t warrant using MCS. After testing this over the past several years, I have compiled a list of assumptions on which I use MCS. They are listed in the table below.
Although I am still testing number seven, intuitively it’s clear that mortality is a very dynamic and important assumption. As a rule, the financial plan should test mortality at an age beyond the client’s life expectancy. Therefore, using MCS on this should be considered.
There are several additional issues to consider when using MCS such as the number of trials you run, the type of MCS parameter you utilize, etc. We’ll discuss these and other planning issues in the coming weeks.