VPW - Best Withdrawal Calculator I've seen to date.....

[Lsbcal]

This data is shown in details on the "Path" sheet.

Yes it is but many using this Excel file will only focus on the VPW sheet. Many do not want to explore the ins/outs of Excel. I had to spend some time to understand the relationships between sheets. Don't get me wrong though, you have done an excellent job on this.

I personally think that looking at inflation-adjusted* data is misleading, because CPI is much more volatile than how inflation is felt by us. *(Even if the spreadsheet says "inflation", it's actually a CPI-adjusted amount. In economics, CPI is an indirect measure of inflation).

Good point as Cut-Throat has pointed out inflation to the retiree may be quite different then the CPI. But I do strongly feel I need to apply some sort of inflation adjustment after several years. I may be more sensitive to travel and medical inflation but much less sensitive to home and commute inflation. Tough to quantify.

My salary is in nominal dollars. It doesn't adjust every month. While it adjusts yearly, adjustments sometimes lag CPI and sometimes are ahead of CPI.

In the 1970's you would be getting 10% raises but the taxation was not inflation adjusted so you were pushed into ever higher tax brackets and/or your inflation raise was reduced a lot. This was by design. Government was getting a bigger share of the pie. You had to live this to appreciate it. I think modest inflation is OK now with inflation indexation of taxes

What's important to worry about is the longer-term trend of the cost of living, and whether withdrawals are able to keep up with it. That's why the VPW spreadsheet shows both withdrawals (blue line) and the cost of goods that could be bought by the CPI-adjusted initial withdrawal amount (red line).

I am just suggesting that it is easier on the viewer to not have to compare the red and blue lines. Maybe just add a line that shows how the spending is maintaining its buying power or not.

Going back to your example of a 1966 retiree. In nominal dollars, he never saw a 50% loss in his portfolio (or withdrawals). Actually, had he used a 50/50 portfolio, he would have barely seen any significant reduction in portfolio value before age 90, even while taking withdrawals all along. Starting with $1M, the lowest portfolio value was $830K in January 1975. Not so bad, psychologically!

I revised the AA to be 50/50 (US stock/foreign stock/bond = 30/20/50). I now see that the 1975 portfolio is $759K nominal (or $499K CPI adjusted). Still seems to be a big hit to me. So somewhat worse numbers then you mention, not sure why. I think the real issue here is how much one wants to emphasize CPI adjustment numbers versus nominal. I'm still not convinced of the nominal argument.

 

[Cut-Throat]

(2) I don't want the wild fluctuations the VPW method can produce.

In all probability, VPW will have less fluctuations than someone attempting to employ a 4% SWR adjusted for inflation. Once a bear market hits, they will naturally 'pull in their horns' and reduce spending on a level that will be far more severe than VPW calls for. It's human nature. Especially for the LBYM types on this forum.

So, the question is: Do you want a Withdrawal Plan that you can actually employ and stick to with confidence, or do you want to 'Fly by the seat of your pants'. I don't believe anyone has stuck to the 4% adjusted for inflation method. I am retired and am actually employing VPW.

 

[photoguy]

In all probability, VPW will have less fluctuations than someone attempting to employ a 4% SWR adjusted for inflation. Once a bear market hits, they will naturally 'pull in their horns' and reduce spending on a level that will be far more severe than VPW calls for. It's human nature. Especially for the LBYM types on this forum.

This is just conjecture. However, if one actually follows SWR then it will have no fluctuations the majority of the time.

So, the question is: Do you want a Withdrawal Plan that you can actually employ and stick to with confidence, or do you want to 'Fly by the seat of your pants'. I don't believe anyone has stuck to the 4% adjusted for inflation method. I am retired and am actually employing VPW.

One issue I have with this thread (not with VPW) is that it is titled the "best withdrawal calculator". This raises the question what is meant by "best". If you think about withdrawals in retirement there are really multiple objectives that one may want to optimize:

- failure rate
- volatility of withdrawals
- size of legacy
- amount withdrawn
- minimum portfolio size during retirement (lowest portfolio drawdown)

These objectives are contradictory and you cannot have one method that is the best on all of the criteria (haha raised this point earlier). VPW is at one end of the spectrum of withdrawal methods with 0% failure rate and higher volatility whereas Trinity SWR trades an increase in failure rate to achieve 0 volatility.

One method is not "better" than the other unless you value one objective more than the other. And of course there could be solutions in-between those two that yield a better trade-off for any given individual.

 

[samclem]

VPW and other "% of end of year balance" withdrawal methods >do< have failures, it's just a different failure mode than the "x% plus inflation" withdrawal method (which fail as "now you are flat broke").

The failure mode of VPW and other "% of end of year balance" withdrawal methods is when/if the prescribed withdrawal does not meet the required minimum spending of the retiree. With these systems, this generally happens if the annual withdrawal percentage is too high (e.g. if it exceeds the real growth of the portfolio for too long). A "% of end-of-year balance" model that projects withdrawals in "future-year dollars" (each of which has been diminished by inflation) rather than "present year dollars" will go some distance toward masking this failure mode. In addition, a withdrawal method that accelerates early withdrawal percentages also increases the likelihood that the prescribed allowable withdrawal amounts (as a percentage of the portfolio) in later years will not keep pace with inflation. If you injure that golden goose, she lays fewer eggs. This is particularly troublesome if you're counting on the wounded goose to lay eggs at a fairly precise rate--based on other species of geese feeding in different pastures with different weather a long time ago. Because we are counting heavily on those eggs well into the future, it seems prudent instead to take good care of the goose and make allowances for the case that your single goose, for the coming decades, might lay eggs at a rate significantly different than those other geese did.

I do intend to use a "% of year end portfolio" withdrawal method.

 

[Chester_A._Riley]

I have been fascinated by the VPW for the past several weeks and I have made numerous runs using a number of spending paths. The look-back feature is quite instructive. I think this an essential tool for a retiree drawing on a portfolio.

Also, the comment by Cut-Throat is spot on.

In all probability, VPW will have less fluctuations than someone attempting to employ a 4% SWR adjusted for inflation. Once a bear market hits, they will naturally 'pull in their horns' and reduce spending on a level that will be far more severe than VPW calls for. It's human nature. Especially for the LBYM types on this forum.

I have looked back at my own experience, with nearly seven years in retirement, and my varying 'seat of the pants' withdrawals & spending (and investment 'tweeks') inadvertently track the markets rather closely. I think that using VPW would have assisted immensely.

This spreadsheet will certainly be added to my future financial planning, (but of course one should never be unprepared to adjust spending and withdrawals as demanded by portfolio growth and depletion.)

Again, thanks for a great tool.

 

[Lsbcal]

Sometimes I have to remind myself: "don't forget Lsbcal, this is the internet ... ugh"

I think it would be great to stick to the facts and have an informative discussion, i.e. try to keep our egos in check. I know it's a hard thing to do.

 

[MichaelB]

An informative discussion that is also friendly and polite. Not that hard.

 

[Huston55]

Sometimes I have to remind myself: "don't forget Lsbcal, this is the internet ... ugh"

I think it would be great to stick to the facts and have an informative discussion, i.e. try to keep our egos in check. I know it's a hard thing to do.

An informative discussion that is also friendly and polite.

Absolutely. I concur with this.

I think this is a useful thread in the sense that it's provoked some good discussion of 'what makes a good retirement calculator (analytical tool).' It's clear from this thread that there are many factors that constitute 'good' for the many different situations and preferences even in this self-selected community we call E-R.org.

I've read this and other VPW threads here, and those at BH which provide more background. So, I know how much work went in to VPW's development, especially by 'longinvest'; that's admirable. However, IMO, there's too much "sell" that's worked its way into this thread. And, as we all know, that dog don't hunt in this community. Most here bristle at the first hint of being 'sold' something or told they 'just don't understand.'

So, utility and politeness will be commensurate with lack of salesmanship and acknowledgement of differing goals, circumstances, risk profiles, etc. by those contributing here. I hope that's what we get going forward, and that the thread continues.

 

[longinvest]

I am just suggesting that it is easier on the viewer to not have to compare the red and blue lines. Maybe just add a line that shows how the spending is maintaining its buying power or not.

I have to make difficult choices. There is no place to show more graphs on the main sheet. I've even seen critics that it is already too intimidating for new users.

I do use inflation-adjusted dollars in the summary reports (min, median, etc.), but I think that subjecting portfolio balances and withdrawal amounts to an artificial CPI factor would hide important information. Here's a graph starting in 1916. Look at how CPI was volatile:

I think that hiding this volatility would really be misleading. The blue line represents the amount that a retiree would have withdrawn. The red line represents the cost of a "normalized basket of goods". It does not account for the flexibility that a retiree could have had to deal with inflation, such as delaying major purchases or eating differently (e.g. using a different basket of goods).

I revised the AA to be 50/50 (US stock/foreign stock/bond = 30/20/50). I now see that the 1975 portfolio is $759K nominal (or $499K CPI adjusted). Still seems to be a big hit to me. So somewhat worse numbers then you mention, not sure why. I think the real issue here is how much one wants to emphasize CPI adjustment numbers versus nominal. I'm still not convinced of the nominal argument.

Maybe you played a little too much with your spreadsheet [ ] and changed some default values (e.g. on the "Table" sheet). You could always download a fresh copy of the spreadsheet. Otherwise, here's my setup:

Input numbers (VPW sheet):

Options (Table sheet):

And, here's the result I get:

 

[FUEGO]

I think this is a useful thread in the sense that it's provoked some good discussion of 'what makes a good retirement calculator (analytical tool).' It's clear from this thread that there are many factors that constitute 'good' for the many different situations and preferences even in this self-selected community we call E-R.org.

I've read this and other VPW threads here, and those at BH which provide more background. So, I know how much work went in to VPW's development, especially by 'longinvest'; that's admirable. However, IMO, there's too much "sell" that's worked its way into this thread. And, as we all know, that dog don't hunt in this community. Most here bristle at the first hint of being 'sold' something or told they 'just don't understand.'

So, utility and politeness will be commensurate with lack of salesmanship and acknowledgement of differing goals, circumstances, risk profiles, etc. by those contributing here. I hope that's what we get going forward, and that the thread continues.

I didn't get that sense at all. One long time poster shared a new (to me) resource that's been developed by the online FIRE community. It doesn't appear to be commercial in nature.

I've actually been looking for something like this for a loooong time since playing with variable withdrawal rates in firecalc almost a decade ago (and posting about my experiments here). This excel format is perfect for the tinkerers among us (which is probably half the active posters here).

If the developer (long-invest) doesn't want to incorporate changes, you could always do so and offer an "even better" tool. I'd be happy to take a look at it.

I personally would prefer the backtesting details on the "front page" to be shown in inflation adjusted dollars, because that's how I think. I was checking out the 1929 begin year backtesting and was shocked at how bad the results were, then realized that due to double digit deflation, it wasn't really that bad. A $1 meal in 1929 would have been $0.75 a few years later, so losing half or two thirds of the initial withdrawal in nominal terms isn't as bad as it looks.

 

[longinvest]

I personally would prefer the backtesting details on the "front page" to be shown in inflation adjusted dollars, because that's how I think. I was checking out the 1929 begin year backtesting and was shocked at how bad the results were, then realized that due to double digit deflation, it wasn't really that bad. A $1 meal in 1929 would have been $0.75 a few years later, so losing half or two thirds of the initial withdrawal in nominal terms isn't as bad as it looks.

So, do I understand right that you think that a CPI-adjusted withdrawal path would be more informative to you than the current separate nominal withdrawal and inflation (blue and red) lines in the chart? Even if it hides the significant volatility of CPI in a period like 1916+ ?

(Thanks for noticing that I have no vested interest in all this. I initially developed it for myself and asked the Bogleheads for comments. They helped me improve it and add it to their Wiki. I make no money with it. I do not even sell a book or something. But I do believe that it is a good antidote to fear mongers that use historical SWR failures to sell services and products to misinformed investors).

 

[Lsbcal]

I have to make difficult choices. There is no place to show more graphs on the main sheet. I've even seen critics that it is already too intimidating for new users.

I appreciate you are going to have a lot of enhancement requests and even some simplification requests. So please just take this as a possible idea and I understand if you don't choose to use it.

Inflation is very insidious to retirement and we have had periods of it in the 1940's and again in the 1970's. Seems way back and yes it is volatile but a real possibility. Equities are volatile too but I really want to know about history to understand the possible realities.

I would suggest at least adding a "real balance" (or call it inflation adjusted balance) line to the upper right chart as shown here:

Maybe you played a little too much with your spreadsheet [ ] and changed some default values (e.g. on the "Table" sheet). You could always download a fresh copy of the spreadsheet.

It turns out we had different input values. I used:
start age = 66 (my age)
depletion years = 24 (to age 90)
domestic stocks = 30%
international stocks = 20% (probably high for a lot of retirees but OK for me)
domestic bonds = 50%
start year = 1968 (yours was 1966)

I think that you are right to set depletion years to 35 as personally I would not want to go to zero balance at age 90, i.e. I would always want a healthy liquid asset base at the risk of leaving plenty to heirs. For this calculator going to age 100 is a good idea.

So I revised my start age to 65, and depletion years to 35. I kept my other inputs as shown above. Then the 1975 balance comes to $850K. This looks good to me now.

 

[longinvest]

I would suggest at least adding a "real balance" (or call it inflation adjusted balance) line to the upper right chart as shown here:

I see, now, what you meant before. That's brilliant. The chart scale works really well for the portfolio balance, as it starts at $1M and goes to zero in both cases (nominal & inflation-adjusted). I'll look at what it does for withdrawals; yet, I really like it. Thanks for the idea.

 

[Chuckanut]

Personally, I have found this discussion enlightening to say the least. Some type of variable withdrawal system seems a lot better than a fixed rate of withdrawal or a seat of the pants system which might result in running out of money early or (more likely in my case) not spending enough while I can.

And a big thanks to longinvest. I award him the 2014 annual Chuckanut Boglehead trophy!

 

[Huston55]

Personally, I have found this discussion enlightening to say the least. Some type of variable withdrawal system seems a lot better than a fixed rate of withdrawal or a seat of the pants system which might result in running out of money early or (more likely in my case) not spending enough while I can.

And a big thanks to longinvest. I award him the 2014 annual Chuckanut Boglehead trophy!

Mr. Nut (may I call you "Chuck?" )-

If you like this VPW thread, you will probably find this thread on Pfau's thread on various withdrawal methods interesting

http://www.early-retirement.org/forums/f28/wade-pfaus-yin-and-yang-paper-74407.html#post1516581

I'd also suggest trying a relatively new retirement calculator (********) that evaluates multiple "variable withdrawal" strategies (including VPW, Guyton-Klinger, etc.) here: Crowdsourced Financial Independence and Early Retirement Simulator/Calculator

 

[uofmbish]

In all probability, VPW will have less fluctuations than someone attempting to employ a 4% SWR adjusted for inflation. Once a bear market hits, they will naturally 'pull in their horns' and reduce spending on a level that will be far more severe than VPW calls for. It's human nature. Especially for the LBYM types on this forum.

I'm struggling to follow this one. To expand your scenario if a bear market hits 2 retirees that started with the same initial account balance, AA, etc, but one uses VPW, the other 4%. The person employing VPW would have a lower account balance at the time of the bear market due to the higher initial withdrawals that you have mentioned.

Why do you assume a person would deviate from their plan in the face of a bear market for the 4% SWR plan, but not the VPW plan that would have resulted in an even lower account balance?

FWIW, I agree with a more variable withdrawal strategy, but let's at least call a spade a spade. VPW allows for higher initial withdrawals because it also allows for more fluctuations, not less.

 

[longinvest]

I think that what was meant was that, due to the fear of hitting a worst-case scenario, human behavior could lead a retiree to spend less than 4% of the initial amount, adjusted for inflation, during a major downturn, possibly even less than what VPW would allow for.

Personally, I am too far away from retirement to have an opinion about this. I see VPW as a tool that one can confidently follow without having to worry about any risk of premature depletion.

To be clear, VPW was never meant as a complete retirement plan; it is just a tool to use within one. VPW, by its own nature, cannot insure a spending floor. But, you can combine it with delayed SS (and pensions) and a CD/TIPS ladder to cover the spending floor between retirement and SS, to get a robust retirement plan. VPW is a tool to spend money saved and invested for delayed consumption; it is not meant for growing a large bequest.

So, it is an appropriate tool to use within some retirement plans. It is up to a retiree to evaluate it against other tools and decide whether it is appropriate for him. Personally, I like its robustness against both under-spending and premature depletion, and I intend to use it when I'll retire.

 

[Lsbcal]

After looking at VPW more I was still a little unclear what the withdrawal (speinding) picture really looked like. So I added two columns shown in blue. These are "real balance" and "real spend" (I used the ROUNDDOWN function).

Now I can see that spending started at $48400 in 1968 and declined in real terms to $28100 by 1982 before again increasing. The percent withdrawals were going up to 6.6% by age 79. To me this is a clearer picture of the nasty investment climate of the inflationary 1970's. It relates the balance and spending to the starting portfolio points.

The chart of withdrawals could easily modified to fit these added columns. I'm not suggesting that the tool be changed but this is how I will run it.

 

[longinvest]

Thanks for the feedback, but I do not think that this forum is appropriate to further discuss modifications to the spreadsheet; there's an official BH thread for this. (Please just consider me as a future VPW user, in the context of this forum).

As for the withdrawal variations, I think that the minimal withdrawal should be compared with the amount you would have gotten with another withdrawal strategy. Every VPW user should expect smaller withdrawals than the start withdrawal. This start withdrawal aims to be average between lower withdrawals in bad times, and higher withdrawals in better times.

 

[explanade]

What good is a 6.6 WR at age 79?

 

[longinvest]

What good is a 6.6 WR at age 79?

It's 6.6% of $427k, not 6.6% of $1M.

Here's a simplified explanation. If you wanted to spend $10 in 5 days, you could spend $2 per day.

But you could also spend the same amount with a variable percentage withdrawal as follows:

Day 1: 20% ($10 x 20% = $2)
Day 2: 25% ($8 x 25% = $2)
Day 3: 33% ($6 x 33% = $2)
Day 4: 50% ($4 x 50% = $2)
Day 5: 100% ($2 x 100% = $2)

As you can see, the variable percentage is always applied to the current balance, not to te start balance.

I hope this clarifies things a little.

 

[samclem]

What good is a 6.6 WR at age 79?

FWIW: The IRS RMD for a 79 year old is only 5.13%, so it's more than that.

 

[Running_Man]

After looking at VPW more I was still a little unclear what the withdrawal (speinding) picture really looked like. So I added two columns shown in blue. These are "real balance" and "real spend" (I used the ROUNDDOWN function).

Now I can see that spending started at $48400 in 1968 and declined in real terms to $28100 by 1982 before again increasing. The percent withdrawals were going up to 6.6% by age 79. To me this is a clearer picture of the nasty investment climate of the inflationary 1970's. It relates the balance and spending to the starting portfolio points.

The chart of withdrawals could easily modified to fit these added columns. I'm not suggesting that the tool be changed but this is how I will run it.

I look at this chart and see a reamrkable failure of this withdrawal method.. Too often what occurs looking back and then going forward is you use todays numbers. But to get to how a real person would have applied this, the US average income back in 1968 was $7,700 up from $7,200 the prior year and would mean a nest egg of 156,250 dollars required for this method in the first year of withdrawls. By 1982 you would have been living on $11,850 nominal when median household income was $18,801, that is poverty living and there is no way to hide this with inflation doesn't really matter. I know because I was poor and living on my own with a roommate in 1982 and I made $20,000 that year.

To spend the first 15 years of a retirement in that manner would be devastating and it is due to taking too much out at the start. People I know live on what they have to but this would be extremely harsh

 

[Lsbcal]

I look at this chart and see a reamrkable failure of this withdrawal method.. Too often what occurs looking back and then going forward is you use todays numbers. But to get to how a real person would have applied this, the US average income back in 1968 was $7,700 up from $7,200 the prior year and would mean a nest egg of 156,250 dollars required for this method in the first year of withdrawls. By 1982 you would have been living on $11,850 nominal when median household income was $18,801, that is poverty living and there is no way to hide this with inflation doesn't really matter. I know because I was poor and living on my own with a roommate in 1982 and I made $20,000 that year.
...

I put in a starting portfolio of $160000 to get close to your $7700 spending in 1968. The 1982 nominal spending was $12409. So close to your numbers above.

But don't forget you would be drawing Social Security too in retirement. So for the average single person it might have been a bit hard. For a couple with a payed up mortgage, we have to maybe double these numbers don't we?

My first year as an EE I think I made about $13000 back in 1972. DW was also working too. Perhaps a bit better then the nominal $17000 (2 x $8500 in 1972) for two average earners with the situation you mentioned above. We did not consider ourselves poor and were able to afford a starter house in 1975 with the incredible sum of $5000 down.

 

[longinvest]

We're usually told not to retire before accumulating 25 times expenses (based on a 4% SWR). VPW does not change this, even though its start withdrawal rate is higher.

$7,700 X 25 = $192,500

This is the initial portfolio amount that I would use to evaluate the riskiness of VPW.

In 1982, this gave a withdrawal of $14,929, approximately %20 lower than the median household income of 18,801 (and 30% lower than the inflation-adjusted $7,700 of 1968). After that, the VPW retiree's portfolio not only recovered, but was able to sustain the rest of a 35-year retirement with generous withdrawals. Personally, I see that as a nice success.