How to Input Possible Market Fluctuations

I assume when using Monte Carlo Planning that the Regular Assets Income on the Total Income Report shows the expected mean return for each year, taking into consideration years when historical stock returns are higher in some years and lower in others. I have not found any specific way to input times in the future when someone could guess when returns could possibly be very low, such as they were in 2008, or high, like 2017. Instead, to try and simulate what might happen this year if things keep going south, for example, I used Special to input a $50,000 Special Expenditure in 2018 and then $25,000 in Special Receipts for the following two years. I realize this wouldn’t make for a precise calculation of actual returns, but does this make sense, or is there another way to input a market drop and recovery?


dan royer's picture

No, the MC does not allow you to simulate that kind of sequence of return risk. The way you are trying to model it will not do what you want. We are doing some work on the MC now, but the basic idea is that it's doing the analysis to show the variance in bottom-line living standard rather than focusing on any one pool of money.

Just another user here. You might try creating your own asset classes and then applying them to a specific year. For example: Under the Build Portfolios tab, create an asset called "Bear Market", give it a real return of -20% for each year in the annual returns table (ten years minimum in order to add it) and then make it the only asset in the portfolio. Then, under Implement Portfolios, apply this portfolio to a specific year for regular assets and retirement account assets. When you run the report and look at retirement asset income, you should see big negative amounts for that year. And if you look at the Total income page, you should see a big negative amount for regular asset income. So you could do the same thing with other return assumptions for various years. Just don't bother looking at the Monte Carlo reports because they won't really mean anything. But at least you can force specific returns onto the main report sections. Maybe create a set of assets with specific real returns to apply to specific years. Of course, you are imposing your own bias for expected returns rather than using long-term average returns. But this might be a way to at least see some worse case scenarios.

That's creative. That way you can test sequence of returns risk for recessions early in your retirement vs later.

I've done this in MC. You can set up say a 40% decline in year 5, then another 20% decline in year 12 or however you choose. It works fine for me.


Exactly. I also use MaxiFi and I'm hoping they will add the same sort of flexibilty. And maybe even the option to set specific returns for each year (independent of Monte Carlo analysis) rather than just one future rate change. It might also be helpful within the Monte Carlo mode to include, in addition to customized portfolios, the ability to choose from different pre-set generic portfolios (e.g., Aggressive 70/30, balanced 50/50, capital preservation 20/80, etc.). And I would assume the same ability to assign portfolios (customized or pre-set) to each year.

These are all excellent suggestions, and as Chris noted, creative. I'll give these ideas a spin. Thanks for the feedback.

Thinking about this, rather than pick an arbitary decline % to represent a Really Bad Year, I'll test my desired asset allocation using Portfolio Visualizer over 2007-2009. I'll use that outcome as the amount of decrement in value.

I'll test other bad periods like the bust and other historical recessions, for comparison of results.

Don't forget the commensurate recovery back to mean from the drop in the year or two following the market collapse. I think this is where Monte Carlo does the work for you and trumps the need for this kind of exercise. You certainly will have a 30-40% drop from time to time, but it's unlikely the market will follow with 10 years of 6% returns for a slow and steady climb back to the mean.

I understand the value of MC. What I'm trying to simulate is sequence of returns risk. I'm deferring SS to 70 and will be drawing heavily on my investments in the interim. Bad years at the outset of retirement would have a greater impact than if I was spending my investments smoothly across my entire life, or if the bad period happened later in my life.

I think of this approach as being similar to contingent reports. I can rely on ESP for determining the median and mean incomes, and show me the probable ranges. By using the negative return years I can control when bad years happen, for testing purposes.