iORP Retirement Spending Models

[Brian99]

Hi - I've been running a series of inputs through iORP and find the tool quite useful.

For its "Retirement Spending Plan" it uses "Constant Spending", with inflation adjustments, as its default - understandably the most common approach.

When I select any of the other "Retirement Spending Plan" models (i.e. Reality Retirement, Changing Consumption, Lifecycle Spending or Age Banding) it produces first-year suggested withdrawals substantially higher than the "Constant Spending" model - sometimes as much as 30+% higher. Subsequent growth and/or changes in the future annual spend are, of course, reduced according to each selected spending model.

At this time, I am comfortable with the result produced by "Constant Spending". However, I am wondering if we could/should "live it up" more in the early years as these lifestyle driven spending models indicate. It's the ever-present conundrum: spend more early when you can enjoy it vs. risking a shortfall later.

Like any plan, I'm sure we'll move with the ebb and flow - if we splurge one year, we don't have to the next year. But it is not my goal to leave a huge estate.


Question: What are your views regarding "lifestyle" variable spending plans as shown in the iORP tool? - or in other modelling tools?

 

[Whisper66]

Question: What are your views regarding "lifestyle" variable spending plans as shown in the iORP tool? - or in other modelling tools?

IMHO - These various spending plans reflect observed spending patterns of actual people. So they are a good reflection of what many people end up actually doing. I accept that but don't like it as a planning basis for what I might end up doing.

I want assurance my money won't run out plus I'd like to budget using as much of it as possible. Therefore, I use the constant spending model but rerun I-ORP yearly to set a maximum budget for that year. So after good market years, the following years budget will go up. After bad market years, it will go down. Several people on this board do similar things. Various modifcations are done including for example smoothing out the budget changes by setting a budget on the last 3 yrs of I-ORP run data.

The I-ORP author (James Welch) has documented the version of this method that is very close to the one I use in his white paper located on the I-ORP web site. Look for the white paper at the bottom of https://www.i-orp.com/ . It's called "A Three Step Procedure for Sustainable Retirement Savings Withdrawals" The paper and it's supplement (also on the website) are easy reads and provides back checking for the methodology as well as comparison to other withdrawal methods.

 

[kaudrey]

Hi Brian,


I am not yet retired, but I think they make sense. Anecdotedly, my parents are now just about 80, and due to my mother's health, they don't a lot of things they used to - travel around the country, spend 3 months in a rental in FL (they live in NH), etc. My dad told me they easily are living on SS money now and have been for the last 5 years, which wasn't the case when they retired over 20 years ago at 57. Similar stories for my aunts/uncles and my father's parents before them.


One grandmother was in a nursing home on Medicaid for many years after her assets were depleted....so that is one of the "unknowns", of course.

 

[2017ish]

I am in the same position as kaudrey (until July!).

Until the past year or two I relied upon standard SWR for planning (beginning early with 5% wr, gradually decreasing to 3%--much to DW's chagrin as I kept adding years of work!). Now that we are close to retirement, I've expanded to examine the Bernicke/blanchett spending model, among others. The research behind that model backs up the anecdotes of kaudry and myself--spending does tend to decrease as you age, especially for more affluent retirees. If you haven't delved into the research, this Kitces post is a good place to start: https://www.kitces.com/blog/estimat...penditures-and-the-retirement-spending-smile/

I don't know that we'll spend as much as this model would allow in our early years; but it gives me a lot more confidence in using a fixed percentage withdrawal rate that is higher than I would be comfortable with in traditional SWR.

 

[sengsational]

My recommendation would be to use the constant spending, especially with the market at historical highs right now.

If you read the paper "A Three Step Procedure for Sustainable Retirement Savings Withdrawals", you'll realize that if you plan to run the model every year to determine your allowable spending, that annual spending value jumps around all over the place! In that paper, the allocation is set to 100% equities, so if you had a different AA, it wouldn't jump around as much, but even so, using this process, you should be prepared to accept steep budget cuts in down years. With the constant spending "on", you're making a more conservative model....one that might deal better with those inevitable dips in the market.

 

[Whisper66]

Thanks, that's a VERY important point I should have mentioned. I think that's why some folks use that 3 yr average technique I mentioned in the earlier post. Others takes any "extra" not spent in a year and use it to buffer the "down" years. Others just plan for thin budget years. In any event, that's a critical point to be aware of.

 

[Brian99]

Thanks for all the thoughtful replies. I am still digesting the recommended material...

The I-ORP author (James Welch) has documented the version of this method that is very close to the one I use in his white paper located on the I-ORP web site. Look for the white paper at the bottom of https://www.i-orp.com/ . It's called "A Three Step Procedure for Sustainable Retirement Savings Withdrawals" The paper and it's supplement (also on the website) are easy reads and provides back checking for the methodology as well as comparison to other withdrawal methods.

The 3-PEAT paper refers to RIMO as the linear programming tool used. Is this the same as the iORP tool?

...and is it correct to assume that all the results in the 3-PEAT paper were performed using "constant spending" withdrawal model (re-run each new year to get a new withdrawal rate for that year)? (e.g. the 3-PEAT paper does NOT use any of the lifestyle withdrawal models, correct? - although it does refer to Delorme's for testing.)

There is much additional information on the ORP website that I still have to get through.

If you haven't delved into the research, this Kitces post is a good place to start: https://www.kitces.com/blog/estimat...penditures-and-the-retirement-spending-smile/

I don't know that we'll spend as much as this model would allow in our early years; but it gives me a lot more confidence in using a fixed percentage withdrawal rate that is higher than I would be comfortable with in traditional SWR.

I'm familiar with Kitces/Guyton/Bengen/etc./etc. (with my layman's understanding) -- but will re-read the link you attached.

We're probably feeling the same about it: Maybe we'll spend more than the "constant" model, or maybe not. But it provides some confidence that the base plan should work.

Thanks.

 

[Whisper66]

RIMO as the linear programming tool used. Is this the same as the iORP tool?

Basically yes. RIMO is a generic term the author uses for a "linear programming application that computes the optimal retirement plan; a tax efficient maximization of disposable income" (definition from Glossary). I-ORP is one such program and is the one the author uses for RIMO calculations in this paper. An earlier draft of the report used the term i-ORP. This was changed to RIMO in later drafts.

...and is it correct to assume that all the results in the 3-PEAT paper were performed using "constant spending" withdrawal model (re-run each new year to get a new withdrawal rate for that year)? ...

Yes, the RIMO calculations done for the 3-PEAT procedure were done based on "constant spending". See note 5 in Figure 1 - "The straight line represent's RIMO's constant spending projection from the first RIMO run at age 65....." I suppose if one used an alternative spending model that allowed higher spending during early years, the variance in disposable income year to year would likely increase. Its already pretty high with the constant spending model (see Table 2) so something to plan for.

 

[DawgMan]

Just tried using I-orp for the very first time by plugging in some conservative assumptions on the generic page (not Extended Orp). My model has me retiring at 55 with 60% in tax differed accounts and rest in taxable (upper 7 figures in investable assets). Here are my questions...

- The generic model produced a first year gross income equal to 2.9% of my total investment assets, not say 4%??
- The model shows pulling about 35% of my gross income from my tax differed accounts until age 60, then 100% from tax differed for ages 60 & 61, then a blend thereafter?? I cannot understand how it can assume such a small tax differed withdrawal as I would be subject to rule 72t from 55 - 59 1/2 unless it is assuming I would pay a 10% penalty
- I purposely left SS inputs blank so can I assume the Income output is projected as income only from interest, dividends each year from my taxable account?

I know I need to spend some real time reading/digging through the site/modeling, but was hoping someone could give me the quick Cliff Notes version?

 

[Independent]

Question: What are your views regarding "lifestyle" variable spending plans as shown in the iORP tool? - or in other modelling tools?

When I retired, I had a basic level lifetime spending goal. But, the beginning assets I used in testing that goal were after I had carved out substantial amounts for "extra spending in the early years". Mostly, I wanted to catch up on all the travel I hadn't done while working. But, I also included some for larger gifts to kids.

I assumed I would burn all this excess in the first 10 years of retirement, but I didn't have any plan regarding exactly which years I'd spend which amounts.

I actually set up a tracking worksheet that had my initial carve out amount and then the annual drawdowns from it.

 

[CindyBlue]

I tried to find this information on the iORP site, but couldn't. Do you know if the iORP "disposable income" result means after taxes or before taxes? I'd think "disposable" means after taxes but I'd like to know for sure if it means that on the iORP calculator.

 

[bizlady]

Hi - I've been running a series of inputs through iORP and find the tool quite useful.

For its "Retirement Spending Plan" it uses "Constant Spending", with inflation adjustments, as its default - understandably the most common approach.

When I select any of the other "Retirement Spending Plan" models (i.e. Reality Retirement, Changing Consumption, Lifecycle Spending or Age Banding) it produces first-year suggested withdrawals substantially higher than the "Constant Spending" model - sometimes as much as 30+% higher. Subsequent growth and/or changes in the future annual spend are, of course, reduced according to each selected spending model.

At this time, I am comfortable with the result produced by "Constant Spending". However, I am wondering if we could/should "live it up" more in the early years as these lifestyle driven spending models indicate. It's the ever-present conundrum: spend more early when you can enjoy it vs. risking a shortfall later.

Like any plan, I'm sure we'll move with the ebb and flow - if we splurge one year, we don't have to the next year. But it is not my goal to leave a huge estate.


Question: What are your views regarding "lifestyle" variable spending plans as shown in the iORP tool? - or in other modelling tools?

Probably goes without saying, but your age at retirement makes a difference as well. The younger you are, the better to be conservative but comfortable initially.

 

[pj.mask]

Age but also spending level. If your retirement plan is to live on 50k, that amount may decrease in your golden years but probably not the same % as someone who plans on 250k year income.

As for the comment on spending IRA money before 60, you can do Roth conversions ahead of time to allow it or 72t. Haven’t seen anyone recommend the 10% penalty.

If I didn’t want to leave an inheritance I would be more apt to model realistic spending declines as I age. And I like the idea mentioned of having a constant spending base amount and a separate younger age fun money account. Easy to set up 2 accounts on vanguard to put a dotted line between the $

 

[omni550]

I tried to find this information on the iORP site, but couldn't. Do you know if the iORP "disposable income" result means after taxes or before taxes? I'd think "disposable" means after taxes but I'd like to know for sure if it means that on the iORP calculator.

On the i-ORP Results Report, at the top of the first page, it reads:

"Your projected, maximum, annual disposable income is $XXX,000 in today's after tax dollars."


omni

 

[CindyBlue]

On the i-ORP Results Report, at the top of the first page, it reads:

"Your projected, maximum, annual disposable income is $XXX,000 in today's after tax dollars."


omni

Sheesh, I looked and looked. How could I have missed this, right at the top??
It's been a long few months...maybe I need to retire (grin!)
Thank you!

 

[Fired Up]

Just tried using I-orp for the very first time by plugging in some conservative assumptions on the generic page (not Extended Orp). My model has me retiring at 55 with 60% in tax differed accounts and rest in taxable (upper 7 figures in investable assets). Here are my questions...

- The generic model produced a first year gross income equal to 2.9% of my total investment assets, not say 4%??
- The model shows pulling about 35% of my gross income from my tax differed accounts until age 60, then 100% from tax differed for ages 60 & 61, then a blend thereafter?? I cannot understand how it can assume such a small tax differed withdrawal as I would be subject to rule 72t from 55 - 59 1/2 unless it is assuming I would pay a 10% penalty
- I purposely left SS inputs blank so can I assume the Income output is projected as income only from interest, dividends each year from my taxable account?

I know I need to spend some real time reading/digging through the site/modeling, but was hoping someone could give me the quick Cliff Notes version?

Careful using I-ORP if you have entered different glide path numbers in tax deferred accounts and taxable accounts. For instance, if you have 100% bonds in tax deferred and 100% equities in taxable, it will show that you will need to move funds from tax deferred to taxable. Its trying to maximize your spending and it will do it at the expense of changing your equity/bond ratio of your overall portfolio. Don't know of a workaround, I just take my overall equity/bond ratio and make it consistent across tax deferred, taxable, and roth accounts for the I-Orp.

 

[sengsational]

Are you sure that i-orp will do something different based on bond vs equities? Certainly for taxable vs not, but it doesn't make sense for it to recommend a different plan based on asset allocation because certainly the optimal plan is the one that ignores all that (fewer constraints means more solutions) and you adjust the allocation within the tax category. The presumption I've gone with was that those equities to bond numbers were for calculating return.

 

[Fired Up]

Are you sure that i-orp will do something different based on bond vs equities? Certainly for taxable vs not, but it doesn't make sense for it to recommend a different plan based on asset allocation because certainly the optimal plan is the one that ignores all that (fewer constraints means more solutions) and you adjust the allocation within the tax category. The presumption I've gone with was that those equities to bond numbers were for calculating return.

You are correct the equities to bond numbers are used for calculating return. Based on those returns extended I-Orp will make different plans. Those plans may not agree with what you want for your overall portfolio's equities to bond ratio.

 

[sengsational]

Ok, I think you're saying that i-orp will tell you to spend or move tax advantaged funds at certain times and that by doing so, that would throw your asset allocation off target? I think that would be true if adjustments we're not made. But say it has you pulling and spending out an account with equities when the market happens to be down. You do spend those equities, but you buy equities in the other account. Since you funded that equities purchase with bonds, you have essentially funded your spending with bonds. Another way to look at it is you spend from where it tells you to, and rebalance back to target.

I'm not at my computer right now, but an easy way to test this would be to set after tax to 100 equities and tIRA to 100 bonds with equal amount in each. Then swap it to bonds in after tax and equities in the tIRA. I predict the same result, but if not, I'd have learned something.

 

[Rianne]

iORP is on target. Our VG guy said very close to the same.

 

[Fired Up]

Ok, I think you're saying that i-orp will tell you to spend or move tax advantaged funds at certain times and that by doing so, that would throw your asset allocation off target? I think that would be true if adjustments we're not made. But say it has you pulling and spending out an account with equities when the market happens to be down. You do spend those equities, but you buy equities in the other account. Since you funded that equities purchase with bonds, you have essentially funded your spending with bonds. Another way to look at it is you spend from where it tells you to, and rebalance back to target.

I'm not at my computer right now, but an easy way to test this would be to set after tax to 100 equities and tIRA to 100 bonds with equal amount in each. Then swap it to bonds in after tax and equities in the tIRA. I predict the same result, but if not, I'd have learned something.

I just tried that and the results were significantly different. Also depends on what basis cost you use when you have the equities in an after tax account. And yes it does throw off your asset allocation of your overall portfolio. I like I-Orp but you need to understand what it is doing to insure it is the right thing for you. Dont use it blindly.

 

[sengsational]

Thanks for that research. I'm surprised James has it working like that. I'll play with it when I get back in front of the computer.

 

[sengsational]

Unloading Bonds First

Well, I did a bit of research and testing and DID learn something! The i-orp optimization tends to unload bonds first, and that does effect the result, especially the timing of taxes.

I Set up a scenario with 500K in after-tax and 500K in tax-deferred. No Roth. The basis was set to 1M to keep capital gains from complicating things. The only difference was the asset allocation. In one case, after-tax was all bonds and tax-deferred was all equities, and in the other case, after-tax was all equities and tax-deferred was all bonds. For this sample scenario, the optimization came up with the exact same annual after-tax spending, but through a very different sequence of events.

This scenario showed that i-orp will preferentially spend bonds first. James confirmed that it's because bonds throw off less money, making perfect sense. So if the bonds are in after-tax, tax-deferred with grow and then RMD's will kick-in. If the bonds are in tax-deferred, there's early pulls from tax-deferred, generating taxes early in the life of the model.

One of the things people sometimes don't like about the solutions i-orp offers is that it often "wants" you to pay lots of taxes early in the life of the model. It's justified, in part, by avoiding the RMD "tax torpedo". There is a feature where these high early taxes can be limited. But beyond that, there's another way, I think. I never realized it before, but if i-orp was offering a solution with "too much, too soon" tax-wise, the model inputs could be changed to increase the equity position in the tax-deferred accounts. This would push the optimization toward spending the bonds that were in the after-tax account. Of course if you're going to keep the growth rate accurate, you'd need to keep the weighted average asset allocation across all account types set to what it is in reality.

If you wanted an even-handed pull across all account types, you could calculate your asset allocation for all account types and enter that for all three glide-path entries. That way, the optimization would not favor one account type over another, trying to dump bonds, but you would still preserve the growth rate based on your true equity proportion.

When I need spending money, the driver is tax law; whatever will be more efficient tax-wise is what happens, irrespective of whether those funds are equities or bonds. The reason for this is that asset allocation is perfectly malleable! For those of us who continually rebalance to an asset allocation target, it doesn't matter if I liquidate bonds or equities because I'll be rebalancing. So for example, let's say my direct liquidation action is against equities in after-tax. My concurrent rebalancing action would have me liquidate bonds and purchase equities in the tax-deferred account in order to maintain the asset allocation target. The result of the example is that I bought the same amount of equities that I sold, making it seem like the liquidation was against bonds.

So since I'm continually rebalancing, I'd rather have i-orp take an even handed approach when it comes to deciding which tax category of account to pull from. I suspect that when all the glide paths are the same, the optimization would start being guided more by tax law rather than expected return.

 

[Fired Up]

Just a few comments. I believe we came to the same conclusions. Was not able to get the same results you had with your sample scenarios. Maybe, the starting age you picked just happened to give you the same spending dollars in each case.

When I think about it, I-ORP is constantly rebalancing each account to its respective glide path for that year.

I also set the glidepaths across each account equal to my overall Stock/Bond ratio to help I-Orp be my guide and keep my target Stock/Bond and therefore indirectly growth closer to my actual accounts. Haven't even thought how qualified vs. ordinary dividends would come to play, since I need to fake out I-Orp on the amount of bonds in taxable. Hopefully, its a minor component and can be ignored.

 

[walkinwood]

I can't seem to save the parameters in i-orp. HAving to enter them each time is a pain. any ideas? I'm using firefox.