Speaking of RMDs (sorry to hijack this thread a little)…

Is there any way to model Spousal RMDs? My spouse and I are 5 years apart and having both worked full time our entire lives we will each own a significant portion of our total 401k savings. RMDs on my wife's account will begin 5 years before they do on mine.

Currently in FRP I am basing it on my Wife's age, with our total 401k balance entered as a singe portfolio...so right now the RMDs are being simulated higher than they need to be. Is there any way to easily account for the fact that we have separate deferred tax accounts and that RMDs start during different years for each of them?

## RMD Calculations

### Re: RMD Calculations

I've never tried this, but maybe you could change things to be based on your age and then use a pair of additional inputs "savings" entries to manually take the RMDs out of tax deferred and put the after tax net back into taxable for those first 5 years.

I think you'd need to use a negative tax deferred savings cash flow and a positive taxable savings cashflow (reduced by estimated taxes due). If you use an average RMD for those first 5 years, I think you could do it with just two additional inputs entries.

I think you'd need to use a negative tax deferred savings cash flow and a positive taxable savings cashflow (reduced by estimated taxes due). If you use an average RMD for those first 5 years, I think you could do it with just two additional inputs entries.

### Re: RMD Calculations

Thanks, Jim.

I might give that a try...should give a reasonable estimate of the RMDs for the first 5 years. After that when both of our accounts are subject to RMDs...the difference in RMD rates between our accounts I would say is close enough for these simulations.

However, I'm modelling our annual expenses based on mu Wife's retirement day mostly because that's when I want to model the start of our retirement spending. I'll still be working for several years ...so during that period when she's retired and I'm not I've added additional inputs for my income, 401k and roth savings , etc.), Using her retirement dates makes that scenario straightforward to model based on how FRP works.

I'll have to think about that...I may just leave it as is and assume that the simulation is a little conservativewith the results due to over aggressive RMDs.

I might give that a try...should give a reasonable estimate of the RMDs for the first 5 years. After that when both of our accounts are subject to RMDs...the difference in RMD rates between our accounts I would say is close enough for these simulations.

However, I'm modelling our annual expenses based on mu Wife's retirement day mostly because that's when I want to model the start of our retirement spending. I'll still be working for several years ...so during that period when she's retired and I'm not I've added additional inputs for my income, 401k and roth savings , etc.), Using her retirement dates makes that scenario straightforward to model based on how FRP works.

I'll have to think about that...I may just leave it as is and assume that the simulation is a little conservativewith the results due to over aggressive RMDs.

### Re: RMD Calculations

That makes sense to me.

While it's good to capture the known inputs as well as you can, there's inherent uncertainty in long-range planning. Using tools like FRP involves accepting that there's a significant amount of built-in error signal that's impossible to ring out. For example, I'd bet 10-20 basis points of difference in your rate of return estimate would completely swamp any error signal from the imprecision in the RMD handling.

So getting as close and you can with the inputs then accepting some built-in fuzziness in the model can be a good reminder that the output isn't a prediction as much as just some extra information to hopefully help you make better decisions.

While it's good to capture the known inputs as well as you can, there's inherent uncertainty in long-range planning. Using tools like FRP involves accepting that there's a significant amount of built-in error signal that's impossible to ring out. For example, I'd bet 10-20 basis points of difference in your rate of return estimate would completely swamp any error signal from the imprecision in the RMD handling.

So getting as close and you can with the inputs then accepting some built-in fuzziness in the model can be a good reminder that the output isn't a prediction as much as just some extra information to hopefully help you make better decisions.

### Who is online

Users browsing this forum: No registered users and 10 guests