Confusion Over Retirement Asset Rate Of Return
Hi,
I have been trying to learn to properly use ESPlannerPlus for several months now. I seem to be able to easily confuse myself over some of the nomenclature being used and relating it to the data generated in the reports. I am (hopefully) planning to retire within the next six to twelve months so am becoming a little frantic at getting results I can understand and depend upon out of ESP. I love the flexibility and power of ESP but have difficulty at times understanding what is going on under the covers.
My latest confusion is over the Retirement Asset Rate of Return being used to calculate Retirement Account Asset Income. Let's see if someone can ease my confusion.
I set up a portfolio for Retirement Assets as follows. These roughly match my 401K portfolio classes. I include here the Average Real Return carried by these default assets in ESP. From the Monte Carlo Folder's Build Portfolios tab:
Asset Type Share Avg Real Return
Long Term Corporate Bonds 30% 3.24%
Long Term Inflation Protected Bonds 20% 2.97%
Large Cap Stocks 25% 9.07%
Small Cap Stocks 20% 13.95%
Emerging Markets 5% 23.65%
First, a question of nomenclature. I assume Average Real Return shown by ESP is the return rate corrected for inflation such that the results will reflect results in the future as Todays Dollars. Is that correct? I further assume the nominal rate of return would be higher than this.
If I calculate the composite Average Real Return rate for this portfolio, I get 7.806%. Does that make sense given the numbers above?
Now, when I run a simulation (at an assumed inflation rate of 2.30%) using this portfolio, I get confused very quickly. In the "Inputs and Assumptions" section of the resulting report on the "MONTE CARLO SPENDING BEHAVIOR & PORTFOLIO CHARACTERISTICS" page, it shows this portfolio with the following attributes.
Mean Median Variance Beta
Portfolio 10 0.074621 0.062101 0.393562 0.515386
First, are these Real or Nominal Rates? Why are they different from the Average Real Return rate reported in the Monte Carlo data entry folder? I can understand the median rate being different and coming from the Monte Carlo analysis, but shouldn't the Mean rate agree with (or be closer to) to the Average rate which was assumed upon entry? And why the relatively large difference between the Mean and Media rates in this case?
Now, when I look at the main body of the reports things get even more confusing for me.
In the "Details" section of the reports on the "JOHN'S RETIREMENT ACCOUNTS" page, I see the Retirement Asset Income data table. If I take the value reported in the "Retirement Asset Income" column of the report table and divide it by the "Retirement Account Assets" ending balance of the previous year, I find that the income rate appears to be 3.731%. (The rate is the same for all rows [ages] in the report as I might expect. Just the value is surprisingly low.)
So, I need a lot of help in understanding why the portfolio says it has an Average Real Return Rate of 7.8% in the input folder, shows a calculated Mean Return Rate of 7.46% and median rate of 6.21% in the Input section of the report and results in an apparent calculated income of just 3.73% in the details of the calculated results.
Any help understanding all this would be GREATLY appreciated.
Best Regards,
John
- - - - - - - - -
As a follow-up to the above, I tried one more simulation. I created a portfolio with 100% Long Term Inflation Indexed Government Bonds. (Is this the same as "All TIPS"?)
The Build Portfolios tab of the Monte Carlo folder claims this portfolio has an Average Real Return of 2.97%. The resulting portfolio in the Inputs and Assumptions section of the resulting report says the portfolio has a Mean return rate of 2.84% and a Median of 2.84%. (This makes sense that they are the same because the variance is zero for this asset type.) Again, in the data for the Retirement Asset Income in the details section of the report, the actual return rate is shown to be 1.419%.
Further, in the Monte Carlo section of the report, under the "PERCENTILE DISTRIBUTION OF LIVING STANDARD" table, it lists both the projected income from this portfolio and All TIPS. They differ. The portfolio I created with just the single 100% TIPS asset type is shown to be resulting in 13.25% LESS income than ALL TIPS.
What am I missing in reading these reports?
John
RSS
HI JOHN, I REPLY IN CAPS BELOW. BEST, LARRY
First, a question of nomenclature. I assume Average Real Return shown by ESP is the return rate corrected for inflation such that the results will reflect results in the future as Todays Dollars. Is that correct?
YES, CORRECT.
I further assume the nominal rate of return would be higher than this.
CORRECT.
If I calculate the composite Average Real Return rate for this portfolio, I get 7.806%. Does that make sense given the numbers above?
YES. BUT I DIDN'T DOUBLE CHECK YOUR MATH.
Now, when I run a simulation (at an assumed inflation rate of 2.30%) using this portfolio, I get confused very quickly. In the "Inputs and Assumptions" section of the resulting report on the "MONTE CARLO SPENDING BEHAVIOR & PORTFOLIO CHARACTERISTICS" page, it shows this portfolio with the following attributes.
Mean Median Variance Beta
Portfolio 10 0.074621 0.062101 0.393562 0.515386
First, are these Real or Nominal Rates? Why are they different from the Average Real Return rate reported in the Monte Carlo data entry folder?
THIS IS A SAMPLE AVERAGE, WHEREAS THE OTHER IS A POPULATION AVERAGE. BUT LET ME DOUBLE CHECK ON THIS WITH DICK. CALL ME RE THIS IN A WEEK AT 617 834-2148.
I can understand the median rate being different and coming from the Monte Carlo analysis, but shouldn't the Mean rate agree with (or be closer to) to the Average rate which was assumed upon entry? And why the relatively large difference between the Mean and Media rates in this case?
SAMPLE VARIATION RE THE DIFFERENCE IN THE MEANS AND THE DISTRIBUTION OF RETURNS IS HIGHLY SKEWED, WHICH EXPLAINS THE DIFFERENCE IN MEAN AND MEDIAN.
Now, when I look at the main body of the reports things get even more confusing for me.
In the "Details" section of the reports on the "JOHN'S RETIREMENT ACCOUNTS" page, I see the Retirement Asset Income data table. If I take the value reported in the "Retirement Asset Income" column of the report table and divide it by the "Retirement Account Assets" ending balance of the previous year, I find that the income rate appears to be 3.731%. (The rate is the same for all rows [ages] in the report as I might expect. Just the value is surprisingly low.)
YOU PROBABLY ARE RUNNING THE PROGRAM SPECIFYING CAUTIOUS OR CONSERVATIVE SPENDING, IN WHICH CASE THE MAIN AND DETAILED REPORTS ASSUME THE RETURN IS EITHER ONE HALF THE EXPECTED RETURN OR ZERO. RUN IT WITH AGGRESSIVE SPENDING AND YOU'LL SEE THIS CHANGE. NOTE THAT IF YOU MULTIPLY 3.73 TIMES 2 YOU GET 7.46. SO YOU ARE RUNNING WITH THE CAUSTIONS SPENDING ASSUMPTION SWITCHED ON.
So, I need a lot of help in understanding why the portfolio says it has an Average Real Return Rate of 7.8% in the input folder, shows a calculated Mean Return Rate of 7.46% and median rate of 6.21% in the Input section of the report and results in an apparent calculated income of just 3.73% in the details of the calculated results.
Any help understanding all this would be GREATLY appreciated.
Best Regards,
John
- - - - - - - - -
As a follow-up to the above, I tried one more simulation. I created a portfolio with 100% Long Term Inflation Indexed Government Bonds. (Is this the same as "All TIPS"?)
The Build Portfolios tab of the Monte Carlo folder claims this portfolio has an Average Real Return of 2.97%. The resulting portfolio in the Inputs and Assumptions section of the resulting report says the portfolio has a Mean return rate of 2.84% and a Median of 2.84%. (This makes sense that they are the same because the variance is zero for this asset type.) Again, in the data for the Retirement Asset Income in the details section of the report, the actual return rate is shown to be 1.419%.
Further, in the Monte Carlo section of the report, under the "PERCENTILE DISTRIBUTION OF LIVING STANDARD" table, it lists both the projected income from this portfolio and All TIPS. They differ. The portfolio I created with just the single 100% TIPS asset type is shown to be resulting in 13.25% LESS income than ALL TIPS.
Larry,
Thanks for the explanation. The light bulb finally went on in my head.
You are correct, I had consumption assumptions set for Cautious Spending. That explains the return rate (shown in the outputs) being half of what the portfolio was set to return.
This indirectly also answered the main question I had been puzzling over. I was trying to run a number of simulations using different portfolios as part of a sensitivity analysis. My real question was "How much 'buffer' in consumption do I have?" My intent was to take whatever answer ESP gave me and assume I would only spend up to about 75% of that amount. I was taking the output of ESP as "exact" and wanting to apply a buffer to it when ESP had already applied an even larger buffer through the use of a Cautious Spending profile assumption.
By changing the spending behavior to Aggressive and comparing it to the answer from Cautious spending, I was able to directly see a range of spending that somewhere between "moderately safe" and "risky".
As a check, I worked out the composite return rates for a couple of portfolios and ran a Monte Carlo simulation using an Aggressive Spending assumption. I then compared that to a non-Monte Carlo simulation using the converted Nominal rates of return. The answers I got agreed to within about 0.50% to 0.75% for each scenario I tried. That's plenty close enough to bridge the results.
As a suggestion for a future addition to ESP, might it be possible to have a report that shows how much consumption ESP is buffering or "leaving on the table" when Cautious or Conservative spending is selected? I am not sure how or where you might show this but I find the Monte Carlo reports confusing when running a Cautious or Conservative Spending report. It will show that My percentile distributions may be low when they could be better than shown due to this "buffer" which is not reported. In other words, my consumption may appear risky in a Conservative report even though my portfolio is likely to be returning MUCH more than was assumed (0.00%).
John