T. Rowe Price Retirement Calculator - 20%/80% Allocation? (2024)

T. Rowe Price has a Retirement Income Calculator posted on their website.

http://www3.troweprice.com/ric/ric/public/ric.do

This tool now handles Savings, Transitioning to Retirement, and Retirement. It uses Monto Carlo analysis and performs a thousand market simulations in determining results.

The calculator seems to show the best results with an Asset Allocation of 20% Equity and 80% Fixed Income for the retirement case. I believe they use an approximate 4% withdrawal strategy, and recent T. Rowe Price publications show the same result.

Does this result surprise anyone on the board?

I'm also wondering if their assumptions are valid going forward. The tool assumes stock returns of 10%, intermediate term bond returns of 6.5%, and short term bond returns of 4.75%.

dbr wrote:By what margin of what property is 20:80 best? One consideration would be the assumed volatility of the different asset classes.

I was looking at the case of someone born in the late 1940's and just entering retirement. The calculator provides a monthly income figure through Age 95 with a 90% success rate.

The monthly income figure was the highest at an asset allocation of approximately 20/80.

The calculator is very easy to use, and one only needs to enter birth date, assets, and asset allocation which is adjusted with a slider.

After the simulation, the slider can be adjusted to try a different asset allocation.

From the results page of the Web site; emphasis added. Kinda makes you not want to trust the thing.

Material Assumptions Include:
• Underlying long-term expected annual returns for the asset classes are not based on historical returns, but estimates, which include reinvested dividends and capital gains.
• Expected returns plus assumptions about asset class volatility and correlations with other classes are used to generate random monthly returns for each class over specific time periods.
• These monthly returns are then used to generate hundreds of scenarios, representing a spectrum of possible performance for the modeled asset classes. Success rates are based on these scenarios. Success rate is defined as the percent of market simulations that result in a positive balance at the end of the time horizon.
• Taxes on withdrawals are not taken into account, nor are early withdrawal penalties. But fees average expense ratios for typical actively managed funds within each asset class are subtracted from the expected annual returns.
• Required minimum distributions (RMDs) are included. In the simulations, if the RMD is greater than the planned withdrawal, the excess amount is reinvested in a taxable account.

KAWill70 wrote:

dbr wrote:By what margin of what property is 20:80 best? One consideration would be the assumed volatility of the different asset classes.

I was looking at the case of someone born in the late 1940's and just entering retirement. The calculator provides a monthly income figure through Age 95 with a 90% success rate.

The monthly income figure was the highest at an asset allocation of approximately 20/80.

The calculator is very easy to use, and one only needs to enter birth date, assets, and asset allocation which is adjusted with a slider.

After the simulation, the slider can be adjusted to try a different asset allocation.

Thanks, but I am still wondering by how much does the income vary over the range of AA. It is classic in these studies that while there may be a maximum, the actual difference from best to worst is not very large and would be better summarized by saying AA had little effect on the outcome. This is usually a result when withdrawal rate is not overwhelming, such as 4%.

The calculator results seem to agree with two studies I've seen from TRPrice. Here's a link to one of them. See the table on page 16.

http://individual.troweprice.com/static ... nGuide.pdf

TRPrices studies are based on having at least $1 left, so while the results show the probability of assets lasting 30 years, they do not account for spending power losses due to inflation in most cases. Accounting for that, in the study I linked, it looks like the sweet spot is around 40% equity. That would still provide a high probability of success with enough equity to preserve spending power.

Paul

Thanks to everyone who replied. Here are some numbers to answer the question asked by dbr.

Sample inputs to the T. Rowe Price Retirement Income Calculator:

Birth date - January 1948
Retirement Assets - $500K
No other income or assets

Calculator Results:

Equity/Bond Asset Allocation

10/90 $1536 Monthly Income - 90% Success Rate - through Age 95
20/80 $1556
25/75 $1563
30/70 $1556
40/60 $1543
50/50 $1517
60/40 $1484
70/30 $1439
80/20 $1393
90/10 $1328

The highest income is in the area of 25/75 Equity Bond Asset Allocation.

Note that an initial 4% withdrawal on $500K is $1666 per month.

T. Rowe Price has a great deal of expertise in this kind of analysis so I think it deserves some thought. It appears that this calculator tries to go beyond what others have done. For example, I believe that Vanguard Financial Planning uses simulations from actual market experience starting in 1926. T. Rowe Price is apparently going beyond that and modeling a larger range of possible outcomes.

Personally, I was surprised at the optimum Asset Allocations. However, maybe the explanation is that they are coming up with a monthly income figure based on a 90% Success Rate. Most of us have not thought in those terms. In addition, the model may include very challenging periods in the market and economy.

Lastly, I still question whether 10% stock returns and 6.5% bond returns are the best numbers for the model. Bond returns of 6.5% would seem unlikely at least in the next decade or so. The assumption for stock returns may be more reasonable.

pkcrafter wrote:The calculator results seem to agree with two studies I've seen from TRPrice. Here's a link to one of them. See the table on page 16.

http://individual.troweprice.com/static ... nGuide.pdf

TRPrices studies are based on having at least $1 left, so while the results show the probability of assets lasting 30 years, they do not account for spending power losses due to inflation in most cases. Accounting for that, in the study I linked, it looks like the sweet spot is around 40% equity. That would still provide a high probability of success with enough equity to preserve spending power.

Thanks for the link. I'll review the document in some detail and see how it compares.

The assumptions used by the calculator include a constant 3% inflation and the spending amount presented is the initial withdrawal rate. That amount is then increased each year to account for inflation.

KAWill, here is the other study. See the chart on pg 15 and the article.

http://individual.troweprice.com/static ... Report.pdf

Higher equity allocationsgenerally translated to higher purchasing power. For example, after 30 years ofwithdrawals, a fixed portfolio of 40% stocks and the rest in bonds had a median purchasing power of 131%of the original balance at retirement,while a portfolio of 60% stocks had amedian purchasing power of 162% ofthe original balance.

Every decision in investing is a compromise, except costs.

Paul

While the calculator is easy to use it starts with a very bad premise, i.e., income needed in retirement = 70% of income while working. Income needs in retirement should be tied to spending in pre retirement, not earnings.

It also does not account for defined benefit plans.

I like the features, it is the premise behind the calculations that is flawed.

Thanks Paul and Peter for your notes and references.

The Calculator actually has three modes - Saving, Transitioning, and Retired. Initially I used the Retired option.

This afternoon I tried the Transitioning option and had slightly different results. I need to explain that because there should hardly be any difference between telling the Calculator that you are retired now or retiring in January.

In any event, with a retirement date specified for January, I received the following results:

Birth date - January 1948
Retirement Assets - $500K
No other income or assets

Equity/Bond Asset Allocation

20/80 $1747 Monthly Income - 80% Success Rate - through Age 95
30/70 $1767
40/60 $1771
50/50 $1765
60/40 $1747
70/30 $1723

Note that the highest income is now at an Asset Allocation of 40/60. However, they now model for an 80% Success Rate. That probably explains the difference.

In retirement they model for a 90% Success Rate. You can specify your desired monthly income. I'm not sure where the 70% figure comes in unless it is an assumption in the Savings part of the Calculator.

The Transitioning mode does accept more inputs and it is easy to change the retirement age to see the effect.

The 70% figure is the basis for their projection of how much you will need; "how did we arrive at this?". It is a fixed amount in the savings mode - in the transitioning mode more user input is allowed and the numbers become much more realistic.

Having worked through this again in "transitioning", I would give the tool a much more favorable review. Allowing more inputs provides greater accuracy.