@jkandell Thanks for the link! I saved it for future reference too. I'm just coming to the same conclusion about VTIAX. The "conventional wisdom" always touted the Foreign Tax Credit; however, I just learned this credit will be reduced/non-existent for the 2025 tax year for me (first year fully retired)...and probably others who have very little income besides interest/dividends. I now know to think about the FTC NOT as an offset foreign taxes paid, but as a credit against US taxes to avoid double taxation. And if you are at a low US tax %, you don't get the credit since there's no/less double taxation. All this to say I have regrets about tax-(in)efficient placement.
My calculations for VTIAX look a little different. I'm using last year, which was 61.16% qualified. But I get growth (same 8% ROR) that's 19% qualified, 12% non, and 69 cap gains. I also notice your figures only add up to 90%...
@jlee Thank you for catching my math error, corrected in the post (adding up to 90%). I'm guessing our difference is our figures for dividend yield. You must be using about 2.5% dividend yield, whereas I was using 3% (and using the four year average for qualified). I have it even worse to be honest, since I also have some small and mid cap value funds.
In any case, both of us are likely 'good enough' for Pralana taxation modeling.
In any case, both of us are likely 'good enough' for Pralana taxation modeling.
Agreed! I'm trying to avoid the sin of false precision.
Just bought Online yesterday, no matter what I enter in Growth Taxation, I get the same number for 3a Qualified Dividends and 3b Ordinary Dividends. And my 2b taxable interest is way overstated. That has never been the case for my investment/tax returns. Obviously I am doing something wrong, any idea what?
@midpack Did you determine what was causing (3a) Q-Divs to equal (3b) 0-Divs?
Here's what I do:
I look at last year's 1099 to get the qualified and non-qualified dividends as a percent of the average account value. Then I divide those by my projected nominal rate of return. So last year, our stocks had about 2% total dividends, 1.6% qualified and 0.4% non-qualified. So I divide each by my projected nominal rate of return for stocks (5% real + 3% inflation), giving me 1.6/8 = 20% qualified dividends taxed as capital gains, 0.4/8 = 5% non-qualified dividends taxed as ordinary income and the remaining 75% is taxed as capital gains when sold.
I started with my 1099's and investment balance, so if the stock market goes up exactly as I forecast and dividends move exactly in line with price appreciation, the dividends will match too. That's good enough for long term modeling.
I have other income that is variable, so I only do Roth Conversions late in the year. By mid December, the brokerage posts estimates of the forth quarter dividends and I adjust the dividend percentages in Pralana to make the current year match.
@ricke I gather each year you create a distinct "portfolio time period" for the current year (only) in order to tailor returns/taxation for the current year as it unfolds?
I don't really have multiple time periods with different returns, it's just that I do my planning near the end of the year, mainly for Roth Conversions. Makes sense in our case as I work a variable amount, so can't really plan Roth Conversions for the year ahead of time. At that point, I might as well use what's known about dividends as use that to help tune Roth Conversions to stay under IRMAA tiers. Pralana is powerful enough to have multiple time periods, but my vision of the future isn't.
Note that when you manually do an intra-year update like I do, you have to remember how Pralana works, which is to use start of the year balances and have them increase by whatever amount you told it. So if your account is worth $1M near the end of the year and you told the program that it was going to grow by 5%, then you need to tell Pralana that the starting balance was $1M/1.05 =$952K, regardless of what the actual starting balance was. So doing it that way requires a bit of side calculations. I think someday they plan to add the mid-year update feature to Online, but it's not in yet. Even when/if they add the mid-year update, folks should think through any lumpiness in income/expenses/asset sales/tax loss harvesting and make some manual adjustments.
You have to remember how Pralana works, which is to use start of the year balances and have them increase by whatever amount you told it. So if your account is worth $1M near the end of the year and you told the program that it was going to grow by 5%, then you need to tell Pralana that the starting balance was $1M/1.05 =$952K, regardless of what the actual starting balance was. So doing it that way requires a bit of side calculations. I think someday they plan to add the mid-year update feature to Online, but it's not in yet. Even when/if they add the mid-year update, folks should think through any lumpiness in income/expenses/asset sales/tax loss harvesting and make some manual adjustments.
Aha. I am online but could still backtrack all the entries; but I’m not willing to put in that much work.
i hope they switch up the online implementation to simply ask for ones current amounts, and then either (1) prorate the equations and save last years entries to calculate taxes. Or, second option is (2) one can’t update $ amounts till 1/1 of next year but Pralana has option to enter projected % returns for the current year which overrides the overall average return entries for current year only. Worst option (3) would be the program back-calculates the jan 1 $ "pseudo" figure for you (the calculations you explained above) after you enter current $ amounts at any point in the year.
Then again, maybe pralana isn’t designed to do this year's taxes, and perhaps one best uses free turbotax or a spreadsheet or some such to estimate this year's roth conversions and gain harvesting and one's ACA subsidy. My issue is that a big tax error in the current year will due to compounding cause numerous other monte carlo and conversion errors in the years to come.
Here’s the problem with allowing current account balances to be entered as the tool’s starting point: We don’t know what puts and takes have actually occurred for each account in the current year that have resulted in the balance you’ve specified. Have some RMDs already been taken? Has a Roth conversion already been done? Have some scheduled withdrawals already occurred? Have quarterly tax payments been made? How much of your annual income has already been received? Has some unexpected windfall occurred? How much of your annual expenses has already been paid? Have you encountered some large, unexpected expense? Trying to deal with this can lead to a high degree of additional complexity in the tool design and its user interface. Prorating everything for the first year is an option but probably inaccurate. We believe that it’s far better to simply ask users to specify start-of-year account balances and then calculate all projections from January 1.
Regarding Pralana doing this year's taxes, please remember that at least for now, it does not do iterative tax calculations, so unscheduled withdrawals from your tax-deferred account(s) are not taxed until the subsequent year.
Stuart
@smatthews51 Thank you for the clarification on the design considerations. (The transparency of Pralana, thanks to you and Charlie, is such a breath of fresh air in this arena.) I can see how having a common anchor point on 1/1 prevents multiple double-counting errors in the common "puts and takes" you identify. One large roth conversion alone could set things off my huge amounts if data was entered willy nilly. Then again, anchoring on 1/1 with a long-term average return in years with a huge drop or positive return can also set things off by huge amounts, but at least it avoids things cascading into other areas areas of the program.
Appreciate the reminder that tax deferred accounts are not taxed until next year (to avoid looping errors, if I recall). That reinforces my view that Pralana is designed for long-term planning, including long-term tax-planning--but not intended to get a handle on "this year's taxes", and by implication "this year's" roth conversion, tax gain and loss harvesting etc.
My mind now wonders:
1) Is the Pralana Gold mid-year growth methodology the way you all anticipate Pralana Online rolling out eventually? ("In the Mid-Year growth mode, Annual Growth = [Account Balance (at the end of the prior year) + annual deposits/2 - annual withdrawals/2] x ROR"). It does seem to be a good enough solution that avoids complexity.
2) I wonder now if/how Boldin (aka "the lesser Pralana" ) avoids this problem since it asks people to "keep your accounts up to date" and even optionally links accounts directly to stay accurate. It looks like for better and worse they may ask to make continuing adjustments to avoid the "puts and takes". (Posing for others, not asking you for an answer Stuart!)
Not a comment on the last post, but more in regards to the original post... I find the Growth Taxation input also confusing, you can tell just by reading this post back and forth. A suggestion might be... it is very likely that many users of the tool have Boglehead-like assets, i.e. VTI, VXUS, BND i.e. a conventional 3 fund portfolio. It would be useful if someone knew the actual data for these 3 for 2023, and 2024... and showed the proper values and how they were calculated in an Excel spreadsheet table. These would show a few examples, of actual data and a reality check for newbies trying to input data correctly. By adding in Muni Bonds, and Cash to the table that would give a pretty complete example set for people to refer to.
2) I wonder now if/how Boldin (aka "the lesser Pralana" ) avoids this problem since it asks people to "keep your accounts up to date" and even optionally links accounts directly to stay accurate. It looks like for better and worse they may ask to make continuing adjustments to avoid the "puts and takes". (Posing for others, not asking you for an answer Stuart!)
I am a former Boldin Pro user. It pro-rated balances and expenses by month...so if it's March, you could only plan for April and beyond. Anything beyond the current year was tied to the user's age--so you could schedule a windfall for when you turn 73. Not sure when RMDs were assumed, but Boldin has so much less power, I'm not sure all the issues raised for Pralana apply to Boldin. It was not robust with taxes, nor very transparent. The power Pralana has seems to, in this case, make this a much more difficult task than it was for the Boldin folks.
I started with my 1099's and investment balance, so if the stock market goes up exactly as I forecast and dividends move exactly in line with price appreciation, the dividends will match too. That's good enough for long term modeling.
@jkandell Yes, Pralana Online will implement a mid-year growth algorithm similar to that of Pralana Gold in the near future.
Stuart
