Using FRP During Retirement

[srseibs]

Hello! Apologies if this is covered elsewhere, but I sure couldn't find it and FAQ didn't seem to do this justice. I just started using FRP about a month ago, after playing with several other tools. FRP resonates with me - I like its intuitive (to me) interface and power.

Something I cannot figure out though is this: once I enter a plan into FRP I can see the statistical outlook through end-of-life - great!

Each year after creating my initial plan, the projections become a little more certain, since prior years' guesses become this year's historical facts --- is the idea that each year I should update my data with the latest account balances?

What is the recommended way to use FRP during retirement as last year's returns, inflation and spending has taken place? How or if should the initial plan be updated to be most useful to me?

[jimr]

Hello,

People have different ideas about how best to track their progress over time. I'm not in a position to give personalized financial planning advice, but I can tell you what I do myself. Of course, nothing I say should be taken as financial planning advice. You'll need to do your own due diligence and decide for yourself what's best for you.

Basically, each year I update all planner inputs based on the current situation. I update my current age, current account balances and any other updated assumptions about the plan. If I've made any substantial changes to my portfolio's asset allocation, I adjust the portfolio return and standard deviation to reflect those changes. I always use Custom for investing style and use a tool called Simba's spreadsheet from the bogleheads.org website to find the historical return and std deviation to plug into FRP. Simba's spreadsheet lets you to enter your exact asset allocation with specific securities and uses historical results back testing to determine the historical return for the exact assets you have in your portfolio.

Separately, I maintain a spreadsheet that includes a year by year summary of my retirement plan progress over time. This is where I save historical information from prior years so I can see how things are progressing and how they've changed. This spreadsheet tracks things like my current portfolio balances. total portfolio withdrawal amount for each year, the withdrawal percent, and the planner's probability of success as of that year. This allows me to see how I'm doing over time as my retirement progresses. As I said, this is what I like to do, but you may find a different approach works better for you.

Jim

[pacificoast]

Hi Jim,

That's really interesting information, thanks for sharing! As mentioned in my post a couple of days ago, I am one year into my retirement. I was planning to keep all my historical information right in the planner. That is why I asked to fix my retirement year, or starting year of the simulation. I'm assuming I can just adjust each year that passes with actuals and have the planner recalculate the future years. Do you see any issues with using your software in that way?

Thanks,

Nick

[jimr]

My sense is that this would be tricky to do correctly and I've never heard of anyone trying to use the planner this way. That said, over the years the planner has had lots of users trying lots of different things and I don't always hear about every use case.

There's probably a way to adjust the inputs so your approach could make sense, but imo this is a very advanced use of the planner and it's not exactly the way it was intended to be used. To get this right consistently, you'd really need to understand exactly what you're doing. I don't think you'd want the monte carlo variability in the years that have already happened, so for any years that have already gone by, you'd probably want to use the upper table in additional inputs to force the return to a known value with the standard deviation to 0. You'd also want to somehow set up things so the portfolio balance shown for each year that's already passed reflects the actual balance that you experienced (in real dollar terms). That's a lot to get right just in those couple of examples, but I'd guess there's more that I'm not considering.

You know, the more I think about this the more I think it's not really a good way to use the planner. There's too much to adjust and it'd be too fiddly to be reliable. In my opinion, it'd be a lot simpler to just use a separate spreadsheet to pull out highlights from each year's run as it happens. Think of it as sort of doing a year wrap up. At the end of the year, take the simulation results from the previous year, along with the actual results from your portfolio and spending, and just record it in a new columns tacked on at the end of the spreadsheet for the year that just happened. Then as time goes on, you'll have all the important info from each year.

[pacificoast]

Hi Jim,

I probably should have added that my main model has very simple assumptions. First I do not use any standard deviations anywhere. Second my growth rate is set to a flat 5% and 3% inflation (no stnd devs). Also my spending per year is fixed to specific dollar amounts. Therefore I think it would be pretty simple to make sure my annual starting and ending balance reflect history.

So basically it's a very conservative model. If we have big dips in the market, I'll just use cash reserves and not sell other investments and adjust the growth for that down year. If we have booms then I'll just adjust the next years standing balance upwards using the true percentage growth for that year. If I spend too much or too little I'll also make those adjustments for that year. I do use the Monte Carlo modeling for different simulations of what may happen but that is not my primary focus nor usage.

Do you think that'll work, or, maybe I am missing something?

Thanks,

Nick

[jimr]

It sounds like you've got a pretty good handle on things. Keeping std dev at 0 does greatly simplify the challenge.

I'm not sure this is how I'd do it myself, but that's largely because my own plan uses the monte carlo and flexible spending features. For my use case what you're trying to do would be much harder to pull off.

But really, the whole point of building flexibility into this planner was so people could use it in whatever way works best for them and it seems like you've found a way to use that flexibility that's working for you.

[ochotona]

I was planning to keep all my historical information right in the planner. That is why I asked to fix my retirement year, or starting year of the simulation. I'm assuming I can just adjust each year that passes with actuals and have the planner recalculate the future years. Do you see any issues with using your software in that way?

Hi Nick, congratulations on retiring. I'm getting closer and closer... anywhere from 8 - 26 months away. When is a decision that hasn't been made.

So I thought about this exact topic, using FRP in retirement, and then I reealized... you have to keep tracking information on the side. Here's why:

Let's assume you use the 60/40 portfolio, just for making an argument. 60/40 had some down years, of course... the worst single year since 1970 was 2008 which was -18%. So if you just update FRP with actuals on January 1 2009, and if you intended to just use the new spending number you'd be looking at a severe budget cut in the range of -18%. It would be ameliorated somewhat by the fact you'd be one year closer to your end-of-plan...

But having whipsaws in spending like that kind of defeat the point of this type of planning; tying spending in a "mark-to-market" sense to a 60/40 or even 40/60 isn't a fun way to live. I will be using Conservative in retirement, which holds back the COLA if the prior year was a down year in real terms, and the portfolio hasn't grown in real terms since the start-of-plan. In order to make that determination, you have to keep records... not frequent ones, just annual actual portfolio values, and your allowable spending limits for each year.

The hard question to answer is... if you use separate records to drive your annual budgeting, are you at risk of driving off a cliff because when you get down to it, Monte Carlo cannot predict the actual future. The actual future could be bust a 95% confidence interval.

One thing I really want to do is run the Bob Clyatt 95% rule alongside FRP. It's very simple... you get to spend the greater of 4% of your portfolio or 95% of what you were allowed to spend last year. Clyatt designed this for early retirees, to basically help the portfolio last not 20-30 years, but indefinitely. The difference is that this annual budget could actually be cut by 5% a year, so 60/40 has had a few cases in history with back-to-back down years, so potentially 10% budget cut in nominal terms, and more in real terms. If inflation is 3%, then two years of 5% budget cuts and 3% inflation put the retiree down 16%. That's a steep cut if the original budget wasn't flush. But for me Clyatt is confusing because I will have many gap years before Social Security at 70 - I haven't settled on how to run it in that case.

I think I will kick-off retirement with an FRP Conservative plan (Flexible is the same for me, oddly), keep track of whether I get a COLA separately, then definitely at 70 when SocSec begins reassess and start running Clyatt for a different opinion. I would also reassess at major life junctures (widowhood, divorce, marriage, selling a house, buying a house, etc).

There is one way to avoid this, though, and keep using FRP solo, without a side spreadheet... but I don't know how to do it, though I have an idea.

After a large bear market... future returns are always better. That's just math. "Buy when others are fearful". So if there is a 2008-2009 event, you'd have to know by how much to boost portfolio future returns in FRP, this would keep you from cutting your budget severely. I suppose you could use FRP to ask the question, "what future return is required to allow me to keep spending intact for the next decade?" And then you could look at history in PortfolioVisualizer as ask yourself if that added return requirement is justified or just a pipe dream.

example
2009-2010 CAGR for 60/40 was 9.35% 2009 was a down year
2000-2009 CAGR for 60/40 was 2.91% 2000 was a peak year

These concerns are why people do buy annuities. I bought a QLAC as longevity insurance. After 82, my SocSec and QLAC will be a reasonably nice income even if my portfolio is exhausted. The QLAC and my home equity are off the books, not in FRP. They're insurance.

Great question though, I think about it all the time.

Afterthought - you know what would probably be really safe? Set the initial budget with Stable, but run it with Conservative rules. If Stable is 95% probable, then flipping the switch to Conservative ups that to 99%. That's pretty darned safe.

[rpstillman]

Hi Jim,

I have been reading with interest your posts about using FRP during retirement. I like your approach to update the model every year with current information, assumptions, return rates, etc. My question is:

Do you update the anticipated withdrawals to account for the inflation rate for the prior year? Let's say in 2024 my withdrawal rate is $200K. When I update the plan, would I revise this for 2025 (and going forward) for inflation? For example: $200K * 1.03 = $206?

Similarly, would I also update all other variables for inflation?

Boy do I love FRP! I'm very grateful for your continued support of the planner.

Best,

Rich

[jimr]

I usually run a handful of sort of "roughed in" scenarios, each with different withdrawal amounts and maybe some other assumptions. Then I look at the prob of success for each scenario.

For withdrawal amounts, I tend to use round numbers, so instead of $206k, maybe 180k, 200k, 220k, etc. I might also run sensitivity analysis with withdrawal amounts between 150k and 250k and check out the general look of things across that range.

Generally, I'm more playing around to see what kind of probability of success I get with different variations of the plan.

[rpstillman]

That makes sense. Thanks very much.