Roth, Long term care & tax avoidance

[tadamsmar]

One of my hobbies is going all Hamlet on doing Roth conversions and choosing Roth investment in general. To be or not to be?

Anyway, there are some Boglehead comments who raise the specter of regretting Roth conversions because your tax deductions get large due to LTC expenses. So I am noodling that one.

The first thing that pops out is that qualifying for the tax deduction requires that you have a high level of need of assistance. You need assistance on 2 of the big 3 (eating, dressing, bathing) for 90 days as certified yearly by a physician or other licensed health-care practitioner. Or certified to need monitoring for health and safety.

So, how likely is this? Is there any way to estimate the financial impact of this consideration?

I figure that the IRS must have some real data on current LTC deductions that might help estimate this.

It's really a more general issue of having large qualifying medical expenses that allow you to draw lots of tax-free money from trad IRAs.

Any real estimates on this?

[financeperchance]

I suggest not letting such a specific scenario like this guide your planning.

[tadamsmar]

I suggest not letting such a specific scenario like this guide your planning.

You appear to be suggesting a 100% focus on specificity and zero focus on likelihood and financial impact. I don't think that is a good idea.

But I suspect this scenario does not have a large estimated financial impact. It's probably low risk.

Part of my argument with those who bring up this scenario is that it is more specific than they realize due to the qualification requirements.

[tadamsmar]

This might help with estimation:

https://www.pwc.com/us/en/industries/in ... vices.html

It provide a percentile distribution of costs.

[cas]

https://www.pwc.com/us/en/industries/in ... vices.html

That is an interesting paper. Thanks for finding it.

Did you download/read the whole 29 page paper? It does a lot more data slicing and dicing on the comment in the executive summary (the page you linked) that

This average cost varies considerably by factors such as gender, type of service, cognitive status and geographical location.

Thus far I've had time only to scan (barely) the whole paper, so I probably shouldn't say much of anything before having time to absorb what was their dataset, assumptions, funding source etc. etc.

I do notice that the dataset consists of people who made long term care claims to 8 long term care insurance companies and (I think upon initial scan) is not adjusted for (excludes) people who never make long term care claims.

One thing that I don't see that it addresses, that is probably of interest to the bogleheads/financial planning crowd, is what is the distribution of costs for a *couple* and/or for someone who was single to start with? I'm trying to get at the scenario I've seen among my older relatives where a (relatively) healthy spouse serves as caregiver for the first spouse to decline, allowing them to stay at home (without paid care) for much longer (especially in cases of cogntive decline) than they would have been able to if they were single. But then the surviving spouse (or someone who was single to start with) doesn't have a spouse to provide that same care, so they end up paying for LTC for much longer than the first spouse (especially in the case of cognitive decline). I'm not 100% sure, but my initial scan is that all the slices and dices of the dataset have all the first spouse, second spouse, single-to-start-with cases lumped in together.

[tadamsmar]

https://www.pwc.com/us/en/industries/in ... vices.html

That is an interesting paper. Thanks for finding it.

Did you download/read the whole 29 page paper? It does a lot more data slicing and dicing on the comment in the executive summary (the page you linked) that

This average cost varies considerably by factors such as gender, type of service, cognitive status and geographical location.

Thus far I've had time only to scan (barely) the whole paper, so I probably shouldn't say much of anything before having time to absorb what was their dataset, assumptions, funding source etc. etc.

I do notice that the dataset consists of people who made long term care claims to 8 long term care insurance companies and (I think upon initial scan) is not adjusted for (excludes) people who never make long term care claims.

One thing that I don't see that it addresses, that is probably of interest to the bogleheads/financial planning crowd, is what is the distribution of costs for a *couple* and/or for someone who was single to start with? I'm trying to get at the scenario I've seen among my older relatives where a (relatively) healthy spouse serves as caregiver for the first spouse to decline, allowing them to stay at home (without paid care) for much longer (especially in cases of cogntive decline) than they would have been able to if they were single. But then the surviving spouse (or someone who was single to start with) doesn't have a spouse to provide that same care, so they end up paying for LTC for much longer than the first spouse (especially in the case of cognitive decline). I'm not 100% sure, but my initial scan is that all the slices and dices of the dataset have all the first spouse, second spouse, single-to-start-with cases lumped in together.

Thanks for pointing out the link, I overlooked it.

Using today's prices...

Nursing home care is about 96,000. Assume SS is about 36000. So, you need 60,000 to cover the rest.

So, maybe it make sense to have $60,000 in a Trad IRA to cover a year of nursing home care.

But if you plan for having $60,000 in a Trad IRA late in life then you will have a lot of RMDs before you get there, so you perhaps have to pay taxes in speculation that you are going to save taxes.

Not sure that retaining a Trad IRA to cover LTC expenses makes sense.

But if you locked in a big Roth conversion and ended up permanently in LTC the next day due to an stroke or something, then you might regret that conversion I guess.

Edit: But if you do it in the same marginal tax bracket then it does not matter if you pay taxes for RMDs or Roth conversions.

[TN_Boy]

https://www.pwc.com/us/en/industries/in ... vices.html

That is an interesting paper. Thanks for finding it.

Did you download/read the whole 29 page paper? It does a lot more data slicing and dicing on the comment in the executive summary (the page you linked) that

This average cost varies considerably by factors such as gender, type of service, cognitive status and geographical location.

Thus far I've had time only to scan (barely) the whole paper, so I probably shouldn't say much of anything before having time to absorb what was their dataset, assumptions, funding source etc. etc.

I do notice that the dataset consists of people who made long term care claims to 8 long term care insurance companies and (I think upon initial scan) is not adjusted for (excludes) people who never make long term care claims.

One thing that I don't see that it addresses, that is probably of interest to the bogleheads/financial planning crowd, is what is the distribution of costs for a *couple* and/or for someone who was single to start with? I'm trying to get at the scenario I've seen among my older relatives where a (relatively) healthy spouse serves as caregiver for the first spouse to decline, allowing them to stay at home (without paid care) for much longer (especially in cases of cogntive decline) than they would have been able to if they were single. But then the surviving spouse (or someone who was single to start with) doesn't have a spouse to provide that same care, so they end up paying for LTC for much longer than the first spouse (especially in the case of cognitive decline). I'm not 100% sure, but my initial scan is that all the slices and dices of the dataset have all the first spouse, second spouse, single-to-start-with cases lumped in together.

For what it is worth, I think your scenario of spouse 1 cares for spouse 2, spouse 2 dies, then spouse 1 winds up in LTC, possibly longer than spouse 2, is not uncommon; I've also seen this exact scenario.

So I wanted to point out this is where a paid off house provides useful LTC firepower; with spouse 2 gone, if spouse 1 is moving to a care facility, spouse 1 can sell the house and that sale could easily pay for all or a good chunk of spouse 1's LTC expenses.

[shess]

It's really a more general issue of having large qualifying medical expenses that allow you to draw lots of tax-free money from trad IRAs.

Well, it stands to reason that as you get into the age where you'll have to do RMDs from your tIRA, you're also more likely to have large healthcare expenses. You can't possibly predict the specifics. A couple random links which seem relevant:
https://www.geripal.org/2010/08/length- ... t-end.html
https://www.aaltci.org/long-term-care-i ... m-care.php
Makes me think there's a good chance that a given elderly person might make use of long-term care, BUT it will most likely not be for more than a year, either because they die or because they recover enough to return to independent life. This matches my experience with relatives and their neighbors, there are a few people who are in nursing homes for a VERY long time for one reason or another, but there's a fair bit of turnover reducing the median.

Put another way, it seems likely that if you were retired for 20 years, you might only benefit from this for two or three years out of the 20, assuming your long-term care doesn't map directly to tax years.

Keep in mind that deduction of healthcare-related expenses only happens if over a threshold percentage of income. So you might have substantial chronic healthcare expenses which do _not_ rise to the level of long-term care, which might not be deductible at all if you have substantial RMDs. Planning to spend $15k/year or $20k/year on medical expenses dealing with chronic-but-manageable health issues doesn't seem outlandish to me, especially once you're into your 70's.

[tadamsmar]

It's really a more general issue of having large qualifying medical expenses that allow you to draw lots of tax-free money from trad IRAs.

Well, it stands to reason that as you get into the age where you'll have to do RMDs from your tIRA, you're also more likely to have large healthcare expenses. You can't possibly predict the specifics. A couple random links which seem relevant:
https://www.geripal.org/2010/08/length- ... t-end.html
https://www.aaltci.org/long-term-care-i ... m-care.php
Makes me think there's a good chance that a given elderly person might make use of long-term care, BUT it will most likely not be for more than a year, either because they die or because they recover enough to return to independent life. This matches my experience with relatives and their neighbors, there are a few people who are in nursing homes for a VERY long time for one reason or another, but there's a fair bit of turnover reducing the median.

Put another way, it seems likely that if you were retired for 20 years, you might only benefit from this for two or three years out of the 20, assuming your long-term care doesn't map directly to tax years.

Keep in mind that deduction of healthcare-related expenses only happens if over a threshold percentage of income. So you might have substantial chronic healthcare expenses which do _not_ rise to the level of long-term care, which might not be deductible at all if you have substantial RMDs. Planning to spend $15k/year or $20k/year on medical expenses dealing with chronic-but-manageable health issues doesn't seem outlandish to me, especially once you're into your 70's.

The medical expenses depend on your insurance. We have pretty good health insurance coverage in retirement.

True that you can't let your RMDs get too big. That's part of the analysis.

[sawdust60]

Too much Roth? Or, what is the optimal pretax amount?

Years of high medical expenses and/or qualified charitable distributions can have lower marginal tax rates. So when you develop a strategy, remember to consider them.

One approach is to consider the components of your income in retirement, and then vary the tIRA and Roth components.

So how much tIRA?
1. Roth conversions in early retirement to utilize low tax brackets and qualify for ACA subsidy.

2. With SS income, enough non-SS income to utilize 0% (or perhaps 12% should you find that 85% of SS income is always taxable).

The idea here is that you can have a limited amount of non-SS income and capture the benefit of untaxed SS.

-----

If you have reasons to have tIRA balances, you may find that they will be high enough to cover LTC, etc. And the lack of precision means that an estimated LTC amount doesn't make much difference when you're figuring out how to build the balances.

Do the analysis. Try to find the right mix for your plans. High tIRA balances can result in high marginal tax rates and Medicare IRMAA and unknown taxes for your heirs.

Roth at 22% and higher can certainly be part of the plan. If you're building Roth accounts at 10% or 12%, there is less risk than if your rate is higher.

And when you pay the tax on the Roth with after tax funds (as required before age 59.5), you are effectively moving those funds into the Roth (as compared with the same balance in tIRA).

[sandramjet]

No idea how likely it is but both of my parents have had 24/7 assistance in home for the last 10 years to the tune of about 175k per year. They have not had to pay income taxes on their IRA wihdrawals yet