Achieving level consumption with deferred social security

I am retired at age 62. I want to plan for level consumption. Since I am planning to defer social security until 70, I expect to draw more heavily on my retirement accounts until social security kicks in. In ESPlanner, the draw is the same each year. So my spending is less now and then spikes up when social security starts. How can I get a level plan?


I think the easiest way is to increase your Maximum Indebtedness to an arbitrarily large amount so that you get a smooth standard of living.
If that requires borrowing, i.e. a negative Regular Assets balance, just realize that rather than actually borrowing you will in fact take special retirement withdrawals to meet your needs.
You can also actually input those special withdrawals. But if other things change, so will the withdrawals.
If the nominal rates of return are the same on regular and retirement assets, then the resulting standard of living will be the same with or without entering the detailed special withdrawals.



I am in the same situation as you. I have previously done what Mike is suggesting -- increasing the Maximum Indebtedness and covering it with a reserve fund. I have experimented with other methods also such as entering Special Withdrawals from the Retirement Accounts Folder. (Using multiple years of Special Withdrawals messes up the Monte Carlo calculations, however, so after talking to Larry K. last year, I no longer do it that way.)

Currently I am modeling this using the Smooth Withdrawals tab of the Retirement Accounts Folder. I start with Percent to be Annuitized at 0% and Percent on Non-Annuitized to be Spent at 100%. Next on the Choice of Annuity tab, I enter the number of years until I will take Social Security and set the Annuity to stop after the Guarantee Period.

Now, I run a report and check the smoothness of the Standard Of Living. At first it will be way off. I then go back to the Smooth Withdrawals tab and bump up the Percent to be Annuitized and run another report. I do this until I reach a smooth Standard of Living. Once smooth, look at the "Retirement Accounts" page in the Details section of the report. It will show the base Retirement Account Withdrawals as well as a Retirement Annuity amount. (The Retirement Annuity amount will stop in the year you start Social Security if you picked the right number of years for the annuity.) I have not actually purchased an annuity; I just use the amount reported as the "extra amount" I need to withdraw each year to smooth my Standard of Living until SS starts.

I have looked at this in detail and compared it to other methods I have seen recommended and not found a problem yet. If anyone knows of any big drawbacks with this method, I'd really appreciate a head's up. The numbers reported generally seem to be a good match to the Special Withdrawals method but this method doesn't cause erroneous results in the Monte Carlo calculations. Still, I haven't used it long enough to have complete faith in it.

Best Regards,

I thought this wasn't working, but I realize I was looking at the Non-Asset Income report in Details (where Annuity column shows $0 every year) rather than the Retirement Accounts report...just to flag this for others who may be trying this method.

Hi Nick,

I favor the special withdrawals method myself. After some experimentation, this worked fine.

Also, per John's comment, you can adjust the "smooth withdrawals % of non-annuitized assets to be spent" for each spouse. This required some sensitivity testing, but I now use a figure less than 100%. Of course, this depends on your specific profile.

Best regards,

Hi Brian,

If you have only a few years of Special Withdrawals from your Retirement Accounts, the method works well. I found that when I used as little as five years, I started to get some odd results in my Monte Carlo reports. I did some testing and found multiple years of Special Withdrawals can really throw off the Monte Carlo calculations.

If you wish to see what happens, try this extreme stress test of Monte Carlo.
1) Run a normal profile with no Special Withdrawals from your Retirement Accounts.
2) Look at the flat withdrawal amount reported.
3) Copy the Profile to a new name.
4) Create a Special Withdrawal for 15, 20 or to end of plan with the same Spoecial Withdrawal amounts as was reported by ESPP in step 2. (You will have to adjust for your assumed inflation rate for the amounts after the first year.)
5) Run this new report.

You should find that if the Special Withdrawals were entered to match the first report where they were not used, the standard of living projections will be within a couple of dollars of the first report all the way to the end of plan. The two reports will look the same until you get to the Monte Carlo reports.

However, now take a look at the Monte Carlo reports. You may find them interesting or amusing. Check out the Regular Assets Monte Carlo section. You will probably notice the Regular Assets take a nosedive into the red. I talked to Larry about this when I came across it. He researched what was going on and it is a side effect of how Special Withdrawals are handled. It is not reasonable to handle the withdrawals in other ways. This is an extreme example that you wouldn't normally hit in real life. However if you have more than a few years of Special Withdrawals, it will still affect the Monte Carlo simulations to some extent.

With just a few years, the effects aren't noticeable. Special Withdrawals work great for "Special" situations. Using them for multiple years can cause confusion in the Monte Carlo reports. That is why I have changed to the "Fake Annuity" method for covering shortfalls when delaying Social Security.

Best Regards,

Hi John,

Thanks for sharing. I ran a few tests with ~20 years of special withdrawals and different amounts, but didn't find anything out of the ordinary in MC results. I'm running the 2.26.0 beta so that may be the reason. Or our profiles are different enough so it didn't show up.

Best regards,

I've pretty much given up on this. Even as a single person, this is a tough problem. Even with a huge amount per year in Special Withdrawals until age 70, ESP Pro still has a huge jump after 70. I'm a little surprised that this software is not able to account for this without a lot of messy kluges. Ironically, Maximize my Social Security recommended I wait until 70 to collect.

Hi John,

You may want to consider the "One on One Services". Profile optimization can be tricky and single people have fewer levers than married couples which can make it tougher to postpone to age 70. See for details.

Since MMSS doesn't see your entire situation, it may recommend age 70 to maximize those benefits. However, this may not fit in with other factors that ESPlanner takes into consideration.

Still, if you are not asset constrained, it may be possible to do what you want in the software. There is even a "Direct Calculation Service" although that costs more, but may be worthwhile if you get exactly what you are looking for.

Optimizing profiles can be certainly be complex. The two links below have some ideas (in lengthy threads) that could be of interest if you continue trying to do this yourself. I'm hoping more people take an interest in optimization so we can build up the forum knowledge in this area.


dan royer's picture

Each situation is different. Sometimes that liquidity constraint is something you just live with if it's slight. The problem is you don't have enough regular assets to smooth over. So I've not seen a situation where just taking extra special withdrawals didn't solve the liquidity constraint. Take more money out now and you raise living standard in the near term at some expense to the long term.

Ignoring the need to eat in the meantime, any SS optimization calculator will tell a single person to wait until age 70 to collect. As Dan says, you can smooth your standard of living with increased special withdrawals. But even that assumes you have enough in retirement savings to bridge the gap.

Experiment with eliminating special withdrawals and, instead, changing the age of your last smooth withdrawal to something in your 70s. If that works, start stretching it out, as long as regular assets remains greater than 0. That's a requirement for ESP to be able to smooth your standard of living.

In the end, it might not work for you at all if you don't have enough retirement savings. Adjust your SS benefits date forward a year, then go through the same process with age of last smooth withdrawal.

As Brian suggested, you might find paying for one-on-one very worthwhile if it helps you figure out how to get to your goal. I certainly did.

Chris Cowles
(a user)

Reviving this old thread and responding to Mike's original reply, is there not a cost to borrowing?

ESPlanner assumes that the cost of borrowing, i.e. the interest rate that you pay on regular asset loans, is the same as the rate of return on regular assets.

Borrowing regular assets and taking special withdrawals from retirement assets are two distinctly different, but viable strategies. You need to decide which strategy that you are actually going to follow and then model that strategy accurately in ESPlanner.

If you're actually going to take special withdrawals, then you can:
1) model it directly with special withdrawals determined by iterating until you get a smooth standard of living, this will automatically incorporate income taxes on the special withdrawals, or
2) you can model it with regular asset loans, i.e. negative regular asset balances, but then you'd have to use special expenses to account for the income taxes that would have been due on the retirement account special withdrawals. This works if the interest rates for regular asset loans are the same as the rate of return on retirement accounts. If they are different, then you'd have to use special expenses/receipts to account for the different rates.

If you're actually going to borrow regular assets, i.e. take out a signature loan or a home equity loan, then just model that directly by allowing a negative regular asset balance. Again, this assumes that the interest rates on loans are the same as the rate of return on regular assets, if they are not, then you'd need special expenses to account for the different rates.


Considering the taxable issue for retirement withdrawals and the interest expense for borrowing, I think I'll use borrowing to determine the level of deficit. Having done that I'll add special withdrawals in the deficit years, increasing the withdrawal amount by the marginal tax rate, plus the marginal tax rate on the extra amount withdrawn to cover the taxes.

Once I get to the point where there are no deficit years I could disable borrowing. Or, since there will be no deficits, ESP won't need to borrow. In that case it doesn't matter if I allow borrowing or not, since there will none, either way.


We use cookies to deliver the best user experience and improve our site.