Need help understanding the outputs

A retirement planning tool is only as good as its assumptions and inputs. Share your thoughts or ask questions about the internals of the simulation, built in planner assumptions, or planner inputs.
Post Reply
FRPJunkie
Posts: 23
Joined: Wed Feb 24, 2016 8:17 am

Need help understanding the outputs

Post by FRPJunkie » Wed Feb 24, 2016 11:02 am

Longtime user, first time poster, love this program. I use the standalone version but the question pertains to both versions.

My "additional input" settings (which override the main page settings, as I understand it) have my Portfolio Return Rate at 7%, with a Standard Deviation of 10%.

When I run a simulation and click on "Show All Runs," under the "Input Avg Return (Std Dev)" column, the 7% is shown, but in the parenthesis it says (2%). Should it not say 10%? This is the case for virtually all of the runs, except one. That run reads: 10% (15.8%)"

Can someone explain what I am looking at here? Perhaps it is pulling the info from somewhere else?

Also, I assume each "run" represents many thousands of Monte Carlo simulations. Is there a way to see them? I guess I'm a little unclear on what is happening behind the scenes, because the planner output shows that for each run, my probability of success 99.9%. It seems like there must be some scenario somewhere where it would be less than this.

Even when I right click and look at the details, it shows every year with an 8% average return. This seems highly unlikely. Why does it not show some years with a loss? This "consistent returns" assumption is the problem with most retirement calculators, and my understanding is that FRP does not do this...it actually assumes market/return fluctuations. Right?

jimr
Posts: 518
Joined: Thu Feb 28, 2008 6:48 pm

Re: Need help understanding the outputs

Post by jimr » Wed Feb 24, 2016 8:56 pm

Hi FRPJunkie,

I suspect the program's terminology might be part of the confusion. The 'Show all runs' table is a table of information from each of the previous times you hit the "run" or 'run all' buttons. It's a history table of prior times you ran the planner. Perhaps you already knew that.

Anyhow, I'm not sure why you'd see a standard deviation of 2% (in parens) if you set the standard deviation to 10%. I can't reproduce that myself (If I create an entry of 7%/10%, I see that in the show all runs summary table. Is the year-by-year information showing what you expect for return/std dev? Since the show all runs summary table has to choose just one return/std dev. pair to show for the entire run, it shows either the entry on the main page, or the last active entry in the additional inputs table (if if the last entry is for part of the plan or even just one year).

In terms of showing the information for each of the 10,000 simulation iterations or paths, that's a bit of a data visualization nightmare. An early version of the planner actually drew lines showing the year-to-year portfolio value for each simulation iteration. but it turned into a mess pretty fast. Now I just show key information that's derived from the 10,000 runs including median values and the year-by-year 10%/90% portfolio values (eg see the checkbox on the graph).

The best way to see the impact of high portfolio volatility is to set up a test run with a reasonable, but not too solid plan. Try this:

1) After saving your data, click the 'New' button to reset all inputs to the defaults.
2) Enter $800,000 for tax deferred portfolio value
3) Click Run
4) Next, change the investing style to custom and change the standard deviation from 9.9% to 15%
5) Click Run
6) Change the standard deviation from 15% to 20%
7) Click Run

After doing this, go to the 'Show all runs' window and check the 'Show Portfolio value bands' checkbox on the graph. Now use the down arrow and up arrow to switch between the different runs you just did and notice how the higher volatility changes the chances of success and also how it changes the range of portfolio values that might be experienced.

Finally, if you have more questions, ask away...

Jim

FRPJunkie
Posts: 23
Joined: Wed Feb 24, 2016 8:17 am

Re: Need help understanding the outputs

Post by FRPJunkie » Thu Feb 25, 2016 8:21 am

Jim thanks so much. For the first part, I was unclear what I was looking at in terms of the historical run table. This time I clicked on detailed view and did in fact see the correct investment percentage and the deviation for each year as I had input them. So that answers the first part of my question.

Perhaps what my other question is getting at is this: What is going on behind the scenes to produce these results? I do see that the bottom and top 10% values are different for each year...meaning the program is running scenarios where the portfolio loses money (or makes less). But is it actually backtesting the portfolio in the simulation?

For example, if you go to Vanguard's retirement nest egg calculator, it states that your investment is backtested against 10,000 past market periods. Does FRP actually backtest against real historical data? And if not, what is it actually doing to produce the results?

Simple example: If I note that I have a $10 portfolio invested $7 stocks and $3 in bonds, when I hit run, does it take this 70/30 portfolio and test its performance against 10,000 past market periods over x number of years to produce the probable results? Or does it make assumptions about the future rather than use real data?

jimr
Posts: 518
Joined: Thu Feb 28, 2008 6:48 pm

Re: Need help understanding the outputs

Post by jimr » Thu Feb 25, 2016 8:38 am

FRP uses the return and standard deviation to run 10,000 Monte Carlo simulations, each a possible path through your retirement plan. In each of these simulation iterations, for each year of the plan, the tool chooses a random number for the portfolio return based on the return and standard deviation that you input. The tool doesn't 'understand' the details of your portfolio mix, only the return and standard deviation that you tell it to use. Also, the tool doesn't consider historical returns or make any predictions about future returns. It only uses the parameters that you told it to use to generate returns.

There are other tools that use a historical returns approach. These tools run through your retirement assuming you started in each of the different years represented in the historical returns sequence that they use. For example, say the historical data goes back to 1900, the first pass assumes your plan starts in 1900, the second pass assumes a start in 1901, the third pass assumes 1902, and so on. This approach uses the actual returns that the market delivered (based on the portfolio mix you told it to use). Examples of this type of tool include OnTrajectory.com and cfiresim.com.

The two approaches are very different buy they usually return results that are very similar. Generally, it's a good idea to try out multiple tools and compare the results.

FRPJunkie
Posts: 23
Joined: Wed Feb 24, 2016 8:17 am

Re: Need help understanding the outputs

Post by FRPJunkie » Thu Feb 25, 2016 8:41 am

Thanks Jim. Appreciate all your efforts.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest