Life Expectancy

[bbbamI]

A year ago, I used 85....now I'm up to 92.

 

[Mr._johngalt]

I use 60, and I'm 62 now.

Actually I pay no attention to this. Just today I'm talking to my Dad
(89). He says, "I think I might make 100!" Could happen as his father,
grandfather, uncle, etc all passed the century mark. Sooooooo,
I have the good genes but I don't allow them to enter into my
retirement planning.

JG

 

[Moemg]

I use 90 .My Mom's age though she'll probably make it to 100.

 

[lazygood4nothinbum]

though my grandfather & mother died at 75 because of alzheimer's, i've traced it down only that one line. many of the rest live into their 80s & 90s on all my grandparents' lines. as each generation seems to last a little longer, i plan for 100, but i'm ready to rev up my spending in my 60s just in case.

 

[cute fuzzy bunny]

http://www.ssa.gov/OACT/STATS/table4c6.html

Unless you have a lot of long lived relatives, looks like 85-90 is pretty darn safe. You might live longer and you might need more care at that age, but you'll also probably have a few lower spending/improvisational spending years between now and then.

 

[bssc]

I figure 90 because my mother's parents lived to be 94 and 100. My dad's mother only made it into her 60s but she smoked and drank. However, he is turning 75 soon and in good health.

 

[KM]

I use 100. I thought I was odd, but it made me feel better. Guess it is not so odd after all....

 

[LeatherneckPA]

Hard, stress-filled life of fighting crime. The average life-expectancy in my line of work is only 7 years post-retirement. My goal is to make it 35 years, I use 85 in my planning.

Silly reason for that has NOTHING to do with anything other than I decided many years back that I wanted to collect as much in pension as I have in salary over the years.

 

[Bikerdude]

Hard, stress-filled life of fighting crime.

And all those donuts.

Sorry, I couldn't resist.

 

[Goonie]

I've been using 85-95.....but hope to hit the century mark! Actually I tell folks I plan to live to about 120, because they say the good die young.....and since I'm an ornery cuss, I'd like to stick around for a LOOOONG time just to p*ss people off!

 

[bbbamI]

 

[Meadbh]

Women in my family tend to make it to 90, so that's what I use.

 

[Cb]

I've used Vanguard's Joint Life Expectancy web app (there's a link in FIREcalc) to find the 95th percentile ages for my wife and I - that turns out to be about 102 years of age, or a 55 year retirement.

I think it's worth mentioning that modeling retirements longer than ~38 years or so in FIREcalc eliminates much of the late 60's / early '70's rough patch. So if you're using FIREcalc to model lengthy retirements it might also be worthwhile plotting the trendline for 25, 30, 35, 40, 45, 50, 55 year durations.

Cb

 

[sgeeeee]

See the last two links in Section 1.3:

http://www.golio.net/My_Homepage_Files/Page9.html

 

[Peaceful_Warrior]

I don't use one. My presumption is that I'm going for FI more than RE, and therefore should be able to RE indefinitely if I am FI.

 

[peggy]

I use at least age 100. Several relatives on both my mom's and dad's side of the family have lived to 90 and beyond.

 

[cute fuzzy bunny]

Really long firecalc runs sort of screw up your results because you end up with a lot of runs that "succeed" because they ran out of data.

For example...for a 40 year run, ever individual run that starts after 1965 will will use one year less data...in other words the run starting in 1967 will use the next 40 years of data, then the one starting in 68 will use 39 years, etc. There is no "wraparound" or "invention of data" to fill out shortened data sets.

I sort of adopted 20 years as the right "run length". If you've done your budget well and incorporated a realistic set of costs that get you by any 20 year period including the depression and the 65-74 sideways period, you're probably going to get a lot more successive successful 20 year periods in.

If you're a real glutton for punishment (and presuming you got 100% success), try taking the detailed data from your 20 year run, take the lowest successful terminal portfolio size and do another 20 year run with that as your "new" portfolio size. Then you'll have worst-cased your worst-case in creating a scenario that survives TWO trips through both of the worst investment climates in US history.

 

[youbet]

I think your source of retirement income has a pretty profound impact on the age you chose. For example, if you're retiring primarily on a generous COLA pension plus SS, it matters little what age you pick for a life expectancy estimate. You're all set no matter how long you live. If you're retiring primarily on your portfolio, chosing a 100+ year life expectancy estimate would probably mean working longer to accumulate a larger portfolio.

For those of you picking a high number, such as 100+ or infinite, are you actually willing to work longer than you'd have to if you chose a number like 90 yrs old?

DW and I are retired on a combination of sources: small COLA pension, small non-COLA pension, future SS and 401K/IRA. We picked 90 for a life expectancy planning number. Seemed like a reasonable compromise.

 

[lazygood4nothinbum]

just ran 38 years per cb & 20 years per cutefb and my withdrawal rate goes up from the 50 year run. i'm starting to like this retirement thing.

 

[Cut-Throat]

I no longer focus on lifespan, but building income streams from Cola S.S. and Cola annuities that will support us for life. This will ensure that half of my principal is gone by the time I'm age 85. Won't have to be concerned with market movements or lifespans.

The other half of Principal will stay invested for life's emergencies.

 

[Rich_by_the_Bay]

I no longer focus on lifespan, but building income streams from Cola S.S. and Cola annuities that will support us for life. This will ensure that half of my principal is gone by the time I'm age 85. Won't have to be concerned with market movements or lifespans.

I, too, have not written off immediate annuities in that context: limited to a portion of your nest egg only, immediate, fixed only.

When I last ran these numbers, I found that a very cost effective alternative to a COLA annuity was to buy what you need to start with unadjusted/nonCOLA, then supplement every couple of years with additional smaller annuities to keep up with inflation. The benefits are:

1. you keep more of your money in the market and/or available to heirs when you die
2. as you supplement you are older and older and thus get more income for your dollar invested.
3. you have the ability to punt on more annuity if inflation is slow or your portfolio is roaring
4. you spread the applicable interest rate over time a bit
5. if you do become seriously ill and plan to die shortly, you will have kept a little more for you and less for the insurance company compared to buying it all and then expiring young.


Downside:
1. it's a little more maintenance
2. if investments go down during the intervals, you may have to sell more shares to buy more annuity, but this is mitigated by using conservatively invested funds for this purpose, as well as the premium advantage of buying older.

Can you think of any others?

So, you need $3k per month today and buy an SPIA to meet that. In 2 years you need $3.3 per month. Purchase another small annuity to generate the difference of $300 per month. Being two years older, it's a little cheaper than it would have been initially, and you kept your money a couple more years.

Bake for 2 more years, and repeat as needed.

 

[Cut-Throat]

I just plugged in some numbers in Vanguards annuity quote and for a 70 year old with Spouse a few years younger could buy an annuity for $500K with no inflation adjust and get about $49K per year. With an inflation adjustment it would be about $38,400 per year. It would seem that buying without inflation might be a better deal, if you could invest the diff and cover your own inflation! - You'd have the advantage of higher interest rates if inflation did rise and like Rich said, you could always buy another annuity that would be cheape. I'd have to run the numbers some time.

But, Even with the inflation adjustment the $38K would be about a 7.7% SWR on the $500K. The question that everyone would have to ask themselves at age 70 is what Percent of $500K would you be willing to withdraw if you were not buying an annuity with it?

 

[Cb]

Really long firecalc runs sort of screw up your results because you end up with a lot of runs that "succeed" because they ran out of data.

For example...for a 40 year run, ever individual run that starts after 1965 will will use one year less data...in other words the run starting in 1967 will use the next 40 years of data, then the one starting in 68 will use 39 years, etc. There is no "wraparound" or "invention of data" to fill out shortened data sets.

Actually, Dory dropped the incomplete periods altogether in the updated version of FIREcalc. The "% success" is no longer inflated by the successful partial periods as in the case of the earlier version, but it does have the effect of completely eliminating the effects late 60's - early 70's era in longer runs.

 

[sgeeeee]

Really long firecalc runs sort of screw up your results because you end up with a lot of runs that "succeed" because they ran out of data.

For example...for a 40 year run, ever individual run that starts after 1965 will will use one year less data...in other words the run starting in 1967 will use the next 40 years of data, then the one starting in 68 will use 39 years, etc. There is no "wraparound" or "invention of data" to fill out shortened data sets. . . .

I've run FIRECalc for retirement spans between 10 and 40 years, then plotted SWR vs. Time in Retirement. The data assymptotically approaches a value of about 3.0% to 3.6% (depending on allocation). (Figure 2.13 in Engineering Your Retirement). The data beyond ~40 years gets really suspect because of the short number of 40 year intervals available in the data.

Another way to look at longer time periods is to require FIRECalc to end a typical 30 year simulation with a specific amount of money. For example, if you assume the 4% rule is approximately true for a 30 year retirement, you can tell FIRECalc you are going to have a 1-time expense of 25X your annual spending in year 30. This is equivalent to requiring that you end the first 30 years with a minimum nest egg amount required to support another 30 years. Thus, you end up with an approximation for a 60 year retirement.

 

[Cb]

SGeeee, you have some big brains.

I'm going to find the last gift certificate dealie from my old job and order your book.

Cb