Let's try this math.
Parents have $100K in trad IRA and $37K invested tax-efficiently in a taxable account.
A. They convert $100K of trad IRA to Roth IRA and use the $37K in taxable to pay the taxes. Roth IRA doubles to $200K. They die and give $200K Roth IRA to heirs. Heirs must take RMDs from Roth IRA, but they stretch it out.
--or--
B. They leave assets alone, tIRA doubles to $200K and taxable account doubles to $74K. Taxes on taxable account are zero since they invested tax-efficiently and unrealized cap gains are not taxed while QDI is offset by tax-loss harvesting and foreign tax credit. They die and give $274K to their heirs. $74K is tax-free and $200K is tax-deferred. Heirs use the $74K to pay the 37% tax on the $200K withdrawal. Net result is that heirs get $200K, same as scenario A. Heirs could be a little more clever about this and not withdraw all $200K (really $274K) at once, they could stretch it out same as inherited Roth IRA.
If there is any chance that the heirs pay lower taxes, I think they would come out ahead. For example, heirs are retired and living off of taxable investments.
Also, I think if the IRA is a mix of stocks and bonds (and perhaps mostly bonds) while the taxable account is equities, then the taxable account may have a higher return than the IRA, but if not, that means more tax-loss harvesting opportunities, concomitant lower taxes, and an overall tax of less than zero on the taxable assets.
My experience is that I don't pay any taxes on my 7-figure taxable account and will not for the next 10 years or so since I am taking into account tax-loss harvesting, tax-efficient investing, and foreign tax credit. How about you?
Unfortunately your examples are full of both errors and omissions.
1) If you are claiming that the ROTH has doubled, then if growth rate assumptions are constant between the ROTH and the tIRA, then the tIRA would NOT have doubled as you suggest. It would have been reduced annually by RMDs, which would then be taxed before subsequently adding the remainder to the tax fund. This is a major point, as there is SOME value in the ability to stretch a taxable IRA, not just a ROTH. The taxable IRA would be a lot better off if it could magically avoid RMDs and double just like the ROTH did. But it won't, and all those RMDs, which if things work perfectly could possibly grow tax free for quite some years, would still be OUTSIDE any kind of account that allows for heirs' enjoyment of many years of further tax deferred growth.
2) While I agree with you that if we're looking at a case where my parents will not touch either the $100k or the the tax fund, and instead leave the appreciated values of both to their heirs, then investing the tax fund in something which you never sell and which pays no taxable distributions would result in a step up for heirs. However, after that, heirs would not be able to make any use of this no-tax-account, unless they wanted to pay their own cg and/or income taxes distributions or sales.
It is an interesting point though about the step-up, one which my dad raised. So, I built it into my model as an alternative scenario. Investing in a perfectly tax-deferred account until death, then receiving the step-up, is better than nothing. However, the calculations clearly show that there are many situations and opportunities (need for huge withdrawal all at once; tax-free compounding for decades with no future tax due) where a ROTH would be immensely more valuable. These are the situations that apply to my case, which is why your assertion that it makes no difference if tax rates are the same is simply wrong mathematically.
3) You are correct that the benefit of a ROTH diminishes if the time before it is tapped is shortened. Thus, your example of the heirs making the ROTH withdrawal IMMEDIATELY and IN ITS ENTIRETY upon inheritance greatly shortens the period of tax free growth and diminishes the argument for a ROTH, especially if parents don't live a long time and thus even THEY don't enjoy much ROTH-deferral-time. However, for one thing if heirs needed the money right away, a lump sum withdrawal of a highly appreciated taxable IRA would likely result in very high taxes for the heirs in my family's case. But here is the most fundamental inaccuracy of your argument: ignoring the specifics of my case, where ROTH can be left to stretch for decades further beyond inheritance. It is completely inaccurate to say that it is economically the same to inherit X dollars in a taxableIRA+taxfund, rather than the same dollars in a ROTHIRA+taxfund. If an immediate withdrawal is made, the ROTH loses some potential, but delivers an enormous benefit in terms of being able to deliver a large sum in a single tax year without a penny of tax. Much more importantly, in specific cases like mine, there will be no immediate need for the ROTH assets, and the fact that some dollars (net of tax-free RMDs) will continue to compound tax free for decades with no tax on any withdrawal at any time, make the ROTH an enormously more valuable asset than inheriting the same value of a taxable IRA and tax payment fund.
You also continue to make the argument that taxes can be essentially zero indefinitely through some kind of perfect process of constant loss-harvesting, foreign tax credits, etc. I believe you when you say that you have enjoyed tax DEFERRAL on your seven figure account. If you need to ACCESS that tax deferred appreciation, however, especially in large chunks, you WILL eventually pay taxes on it...probably at a higher rate than you do now based on my long term view of tax rates. But you will pay taxes, make no mistake about it. Or your heirs will. If your account is worth $100, then yes their tax rates may be very low on withdrawals (assuming base income is also very low, which again is not the case in my scenario). Otherwise, they will have lost the value of tax-free growth that a ROTH offers to those, regardless of income level or tax rate, who are in a position to maximize the time that each dollar of the ROTH can be left intact.
Again, your perspective may well work for your case, although I am unconvinced that you actually have done the math to reveal what your tax will be in each year of your future based on various withdrawal scenarios, and which provides side by side comparisons of ROTH vs taxable variations. Your math and perspectives based on the specifics of my own case are both completely misguided. Assumptions are of course unreliable, but math doesn't lie.
As a former English major, however, I do have to give you some bonus points for the use of "concomitant".