McQ wrote: ↑Mon Jul 26, 2021 10:58 pm
Hello all: Woodspinner asked me to comment on his metrics here. Apologies to those who also participated in the earlier thread; it is not likely I will say anything new here.
Thank you very much for taking the time and thought in your reply. Its much appreciated and I have been pondering it (and FiveK's) input for the last day or so.
First, as to metrics, I generally agree with the line by line response by FiveK. See especially his/her concise demonstration earlier today that tax paid cannot be the metric.
Next, if I may put the “how to measure outcomes” question in my own language: there is no substitute for preparing a year by year spreadsheet. Results will always be better for Roth conversions over longer time frames. Since you have a bequest motive, I think you should run the spreadsheet out to your age 110. I know you don’t think you will live past 90, but … I’m somewhat older than you, and my mother is still going strong at 96. There can be no final age from which IRR calculations are done. That’s not given to us humans to know.
Next, the absolute most crucial spreadsheet to construct is the “do nothing” spreadsheet: all your current portfolios, no conversions, how do you like the disposable income each year and the bequest? If there is clearly surplus—you can’t / won’t spend it all—then Roth conversions become attractive, up to the surplus amount.
WoodSpinner wrote:
So let's put some real numbers and see if it helps. The numbers reflect the deltas of the
Total Tax Adjusted Portfolio Value which is the same as "spendable after taxes" at 90.
I would certainly eliminate
No Conversions and
Balance Effective Taxes based on your suggestions..
This leaves the other scenarios in-play and the variance between them is fairly small (under 2%).
I don't think this metric is enough to make a decision on.
I have run full Year-by-Year scenarios for each alternative including the Do Nothing. I have presented the actual Deltas from the Do Nothing up thread for consideration. I hesitate to post the details for each scenario since I don't want to overwhelm the members and wanted to focus the discussion more on the metrics to use rather than the specifics.
There should be plenty of disposable income based on any of the scenarios under consideration. Pension(s) + SS(s) will cover most of our expenses and our WR is expected to be under 2% (including Health Care, Taxes, Travel etc.). Bequests and Charitable giving goals will also be met by any of these plans.
That said, I do want to be wise in my choices and choose to incur additional taxes carefully.
Second, in version two of the spreadsheet you must create an exact counterfactual for, say, the first year’s conversion amount. You segregate the unconverted amount, invest it the same as the Roth, take RMDs from it, subtract taxes, reinvest the RMDs, and track how the Roth conversion slowly then quickly catches up and surpasses the counterfactual. Because it will, eventually.
You can also test after-tax liquidation values year by year; the Roth is ahead from the first RMD here, but gets farther ahead the longer you live. I always test both, pretax value and post-tax liquidation value, per the several comments upthread, re charitable donations, LTC and deductible medical expenditures, etc.
I think I am doing the same thing although I am using a different approach. What I do is run a full model for each Scenario, copy it to a spreadsheet and then compare the results between each scenario. For instance this allows me to compare
No Conversions to
Convert to top of 24% Bracket on a year by year basis. This allows me to compare scenarios using the same Cashflows, Inflation, Tax Rates, IRMAA etc. Taxes are always paid from the Conversion itself since there are no funds currently in Taxable (Taxable starts building when RMDs begin at 72).
Here is a sample (fictitious data) of a model run:
This is then loaded into another spreadsheet which allows me to compare across the scenarios:
Wrinkle: when I adapted my spreadsheet to my own situation, I had some low return short term bond funds as well as some stock funds available to convert—let’s say $24,000 of the one and $76,000 of the other. My counterfactual now has two columns, one for each asset, which stay invested as they had been. For the Roth account, I assumed that I liquidated the bond funds to pay the tax, and had the $76,000 of stock now in the Roth. This asset location strategy will generally boost Roth outcomes; but it presumes you want to increase your stock allocation relative to your current AA, again as described up thread.
To be clear, my overall Portfolio Allocation (non Tax Adjusted) remains the same througout all of my scenarios (60/40, 45% VTI, 15% VXUS, 40% VGIT). The Asset Location does change but the model handles the rebalancing to insure that the overall Portfolio Allocation remains constant.
The idea of a Tax-Adjusted Portfolio allocation is not something I have dealt with before (its amazing what is in the Wiki!) and I am comfortable with the notion that this may indeed be changing as assets are moved from the IRA to Roth or Taxable.
I will continue to think on this.....
That said, I do see Roth Conversions as a way to better optimize my Asset Location and in general will be better off if they are in Roth (via Conversions) versus Taxable (via RMDs). This provides a great deal more resilancy to our Retirement plan--the question comes down to how much does it cost in Taxes.
PS: after the first conversion you must never, ever leave California for Nevada, Washington, Texas, or Florida. Your spreadsheet will blow up. Not a problem for me; I’m never leaving California except feet first.
PPS: It is my understanding that California does not tax social security. Be sure your state tax calculations post age 70 reflect that fact.
I think we will likely be in CA for the rest of our lives as well. Moving elsewhere is financially attractive but I believe family, friends and social networks are more important. That said, you do have a good point, the 9.3% Marginal Tax Rate hit for Conversions by CA is painful.
My modeling does account for how CA handles SS taxes (which is a nice benefit), thanks for the reminder.
Last, and to cut to the chase: I think you are converting too much. I would run an alternative scenario where you convert only up short of the first IRMAA level (near top of the 22% bracket in AGI terms).
Instead of your current max-convert plan, if it were me, I would pay the taxes on the RMDs (from the portion you failed to convert), and with the remainder, each year take your daughter and family to New Zealand or other bodacious destination, and leave her a much lower Roth balance as an inheritance (the future value of that New Zealand trip, if the expense was instead invested in VTI within the Roth… breathtaking!). But you see how quickly personal values supplant rational calculation once bequests come into the picture, and your values may not be the same as mine.
And, with this last recommendation I stand revealed as far more hedonistic, or Epicurean, than the typical BH. Imagine…spending money on pleasure today rather than scheming how to pass the future value of those funds on to heirs tax-free. Might get me placed on Boglehead probation!
To be clear, I am still deciding on my approach for Roth Conversions/ Focusing on the a foundation in modeling and metrics to use each year to review and course correct.
I appreciate your thoughts on being too aggressive in Conversions given there are so many unknowns involved and will keep them in the forefront of my consideration.
I will add a IRMA Tier-1 Scenario as well, now that I have the foundation built adding scenarios is fairly trivial. That should be interesting!
We are definitely travelers and experience enthusiasts -- that is definitely where we have decided to spend our money (although my wife does maintain a significant clothing budget!
). I totally get that part of the equation.
PPPS: since Celia got you started on this path, would love to see her response to the dollar level of conversions that you currently intend in your plan.
Celia wrote:
Think it would be interesting as well.
This discussion is certainly helping me crystalize my thinking on how to figure out if Roth Conversions are worthwhile and to develop a metric based approach for choosing between scenarios.