Break-even SS 62 vs 66 vs 70 calculators ?

[Independent]

I'd be interested to see the math behind that conclusion.

If you're saying I was being a little too flippant, I'd agree.

Here's the math:

Run FireCalc twice. In both cases, set the investments to 100% equities.
The first time, input SS of $20,000 starting in 2018.
For the second, input SS of $26,200 starting in 2022.
That is consistent with someone with a NRA of 66 deferring to 70.

Download the year-by-year detail for both runs (two 117 x 30 matrices).
For each projection year, count the number of times that the year end values from the first run exceed the second.
Of course, for the first year it will be 117. And, that has to continue for the next three.

The first projection year where the second run is higher for any of the start years is the 14th. There are 2 where the second wins, corresponding to start years of 1969 and 1970. The 15th year has 10, then 12, then 16, eventually 55 in the 30th year. Calculate the corresponding annual percentages e.g. 115/117=98.3%.

Now find a mortality table. Referring to this thread http://www.early-retirement.org/forums/f28/longevity-77386.html , I picked the RP-2014 Male, White Collar. It has the lowest death rates (hence most favorable to deferring) of the tables there.

The death detail is not in the thread, but it says that out of 1,000 males aged 66, about 8 will die in the first year, 9 in the second, 10 in the third, ... with a peak of 45 in the 27th year (age 91), and declining numbers thereafter.

Now, the probability of dying in the first year is 0.8%, and the probability that the "start at 66" will be ahead at that point is 100%, for a compound probability of dying while ahead at 0.8%

The math is equally boring for the next 13 years. Finally, in the 14th year we get a probability of dying of 2.24%, a probability of being ahead of 98.3%, for a compound probability of 2.20%.

Or, in the 27th year, 4.45% times 64.1% = 2.85%.

Summing these probabilities, for the first 30 years only, I get 60.8%. This would be the chance that a man dies in the first 30 years after age 66, and the the start-at-66 value exceeds the start-at-70 value when that man dies.

But, there are still 21% of our beginning group alive at the end of 30 years, and 53% of the start years still favor at-66. I hate to run FireCalc for more than 30 years, but just eyeballing trends, I'd think about a third of those 21% would die soon enough to favor the start-at-66. This gets up to 67% or a 2-1 advantage for start-at-66.

And, yes, I was cutting it too close when I confidently said that starting early would usually win. I thought it would be more like 80+%. (And, a female would be lower than a male.)

 

[Independent]

Shirley and I found a very relevant organization called the SKI club which was started in Australia. There are no membership fees, just follow the objectives, S.K.I. Spending your Kids Inheritance!

Between us we have put 4 children through college with no college debt, our turn!

Since Shirl is a widow, we can play this game from both sides. Start early, before 62, on Survivor Benefits, then switch to her max benefits at 70!

I read the post to say that they were planning to invest the money, not spend it.

 

[ScottS]

Pretty thorough analysis, Independent!

Another data point in the 62 / FRA / 70 debate. MarketWatch has a decent retirement visual calculator. I have used it recently to get a rough idea of future monies and affordability of lifestyle. It doesn't have the past market data to use in it's calculations like Firecalc does so I simply use long term market return percentage for investment returns.

What I did notice when using this and using SS at 65, and then at 70 with every other item staying the same (funds invested, yearly expenses, rate of return, etc) that retiring at 65 vs 70 resulted in a larger nest egg still available at the final years.

My assumption would be that this is due to the compounding of investment returns. If you are living on your investments from 65 to 70 that money is not available for growth and reinvestment.
I need to plug the same information into Firecalc to see what results it provides.

 

[Independent]

I need to plug the same information into Firecalc to see what results it provides.

Here are some FireCalc numbers for the very simple default run. The $29,815 is what I get if I just take the default assumptions and as it to solve for a withdrawal amount.

Then I made the assumption changes indicated in the table, and again asked it to find the maximum withdrawal for the target confidence levels.

I varied the time horizon, confidence target, equity/bond mix, and Social Security benefit ($20,000 at 66 or $26,400 at 70).

Horizon	30 yrs	30 yrs	23 yrs
Confidence	95%	50%	50%
Equity/Bnd	75/25	75/25	75/25
NoSS	29,819	45,344	54,672
SS@66	49,812	65,347	74,670
SS@70	55,991	65,585	74,359
…......…	…......….	…….......	…......….
Equity/Bnd	100/0	100/0	100/0
NoSS	28,403	49,274	59,806
SS@66	48,402	69,272	79,805
SS@70	49,823	69,551	78,644

The first group, 95% confidence level at 30 years, gives the result I've seen before with FireCalc. Deferring SS allows you to spend more money each year. That's because the 95% means your spending is controlled by the 111th scenario out of 117. Presumably, very low return.

I assumed that moving to a lower confidence would allow the median scenario to control, and I thought that would have a high enough return to favor starting early. I also thought that shortening the horizon (23 years is roughly the 50th percentile age at death) would favor starting early.

I'm surprised that these were virtually a wash between the two SS options, especially since I used FireCalc runs to do my earlier post.

I don't have time tonight to research any more, so I'm just posting some un-analyzed numbers. Maybe someone else will see something.

 

[ERD50]

....

I'm surprised that these were virtually a wash between the two SS options, especially since I used FireCalc runs to do my earlier post.

I don't have time tonight to research any more, so I'm just posting some un-analyzed numbers. Maybe someone else will see something.

Interesting, but I think 30 years (so 25 or 26 years of collecting delayed SS) is too short a time frame. As many of us have said, the decision to delay SS is not based on a break-even analysis, but as longevity insurance. Individuals cannot count on being 'average' or 'median', but we can insure/prepare in case we live a long time.

With any insurance, you shouldn't expect to 'gain' - you are paying to offset risk. Few people do a break even analysis on the home or car insurance, they look for the best value, because they know they need to protect against risk, and plan to pay for it, in the long run.

Though I can't explain why the advantage dwindled when you looked at 50% success. Probably just different market/inflation conditions shift things. Regardless, all those people who delayed (especially if they have a spouse eligible for SS survivor benefits), gained the longevity insurance. Maybe they never needed it, but that's how insurance works.

-ERD50

 

[RunningBum]

I don't think the comparison with car or home insurance is a good one. "Insurance" a broad term and comparing SS to car insurance doesn't make sense because they are totally different kinds of insurance. We hope to never collect on home or car insurance. For myself, it's only to protect against major damage or a lawsuit, not breaking even.

With SS, even as longevity insurance, I fully expect to collect regularly, and it makes sense to see when I'll break even. If the break even point isn't until I'm 110, it doesn't make sense to me to defer collecting. If it's at my life expectancy, then it does, because I'm very unlikely to run into financial trouble if I die early, but I could be running low if I live a lot longer than my life expectancy. What you really want is to get the best use out of your SS benefits, and knowing your break-even point is part of that calculation.

There are other factors besides the break-even point, but no factors that compare with home or car insurance.

 

[ERD50]

I don't think the comparison with car or home insurance is a good one. "Insurance" a broad term and comparing SS to car insurance doesn't make sense because they are totally different kinds of insurance. We hope to never collect on home or car insurance. For myself, it's only to protect against major damage or a lawsuit, not breaking even.

With SS, even as longevity insurance, I fully expect to collect regularly, and it makes sense to see when I'll break even. If the break even point isn't until I'm 110, it doesn't make sense to me to defer collecting. If it's at my life expectancy, then it does, because I'm very unlikely to run into financial trouble if I die early, but I could be running low if I live a lot longer than my life expectancy. What you really want is to get the best use out of your SS benefits, and knowing your break-even point is part of that calculation.

There are other factors besides the break-even point, but no factors that compare with home or car insurance.

But the actuarial b-e point isn't 110, it is close to the life expectancy of most of us (some people may have more info on this than others, like a known life-shortening condition). W/o that info, or a crystal ball, it seems like a moot point.

So most of us turn to looking at it from the longevity aspect. If we don't need it now, we can delay to help in the case that one of us (of a married couple) live long. Though I don't worry about the odds, if you have a surviving spouse who collects SS, I think that really does put the b-e in your favor, as they don't provide separate odds for single/married. So (either) one of a couple is highly likely to live longer than a single.

And as Cut-Throat has pointed out, one actually can spend more in those delay years, because they don't need to reserve as much against a long life.

-ERD50

 

[RunningBum]

But the actuarial b-e point isn't 110, it is close to the life expectancy of most of us (some people may have more info on this than others, like a known life-shortening condition). W/o that info, or a crystal ball, it seems like a moot point.

So most of us turn to looking at it from the longevity aspect. If we don't need it now, we can delay to help in the case that one of us (of a married couple) live long. Though I don't worry about the odds, if you have a surviving spouse who collects SS, I think that really does put the b-e in your favor, as they don't provide separate odds for single/married. So (either) one of a couple is highly likely to live longer than a single.

And as Cut-Throat has pointed out, one actually can spend more in those delay years, because they don't need to reserve as much against a long life.

-ERD50

Well, there is the actuarial b-e point, and the personal b-e point based on investment return assumptions. Right or wrong, some people plan on greater returns. If there's a big market drop, I'm probably going to guess a recovery is coming, and that my future returns are going to be better. This would push my b-e point further out, and if the odds of me actually living past my calculated b-e point is slim, I'm going to start taking SS. I haven't put a number on "slim" yet. Maybe 20%?

 

[ERD50]

Well, there is the actuarial b-e point, and the personal b-e point based on investment return assumptions. Right or wrong, some people plan on greater returns. If there's a big market drop, I'm probably going to guess a recovery is coming, and that my future returns are going to be better. This would push my b-e point further out, and if the odds of me actually living past my calculated b-e point is slim, I'm going to start taking SS. I haven't put a number on "slim" yet. Maybe 20%?

OK, I'll admit I haven't looked in detail using some positive assumption for returns on the early SS. If a reasonable (but rosy) assumption makes it attractive, that is worth considering.

But there are still other factors. As haha has pointed out from time to time, the 'cost' of delaying SS provides a very attractive purchase of additional full COLA'd annuity for a person and maybe their survivor-eligible spouse.

And if I think in terms of diversification, a larger SS gets me more of it. None of this makes it wrong/right, even the most sub-optimal decision probably won't be 'bad'. Similar to the mortgage debate.

-ERD50

 

[joeea]

set the investments to 100% equities.

Why?

For the second, input SS of $26,200 starting in 2022.

Shouldn't it be $26,400?

And, yes, I was cutting it too close when I confidently said that starting early would usually win. I thought it would be more like 80+%. (And, a female would be lower than a male.)

And it appears you ignore all cases for married folks?

 

[pb4uski]

If you're saying I was being a little too flippant, I'd agree.

Here's the math:

Run FireCalc twice. In both cases, set the investments to 100% equities.
The first time, input SS of $20,000 starting in 2018.
For the second, input SS of $26,200 starting in 2022.
That is consistent with someone with a NRA of 66 deferring to 70.

Download the year-by-year detail for both runs (two 117 x 30 matrices).
For each projection year, count the number of times that the year end values from the first run exceed the second.
Of course, for the first year it will be 117. And, that has to continue for the next three.

The first projection year where the second run is higher for any of the start years is the 14th. There are 2 where the second wins, corresponding to start years of 1969 and 1970. The 15th year has 10, then 12, then 16, eventually 55 in the 30th year. Calculate the corresponding annual percentages e.g. 115/117=98.3%.

Now find a mortality table. Referring to this thread http://www.early-retirement.org/forums/f28/longevity-77386.html , I picked the RP-2014 Male, White Collar. It has the lowest death rates (hence most favorable to deferring) of the tables there.

The death detail is not in the thread, but it says that out of 1,000 males aged 66, about 8 will die in the first year, 9 in the second, 10 in the third, ... with a peak of 45 in the 27th year (age 91), and declining numbers thereafter.

Now, the probability of dying in the first year is 0.8%, and the probability that the "start at 66" will be ahead at that point is 100%, for a compound probability of dying while ahead at 0.8%

The math is equally boring for the next 13 years. Finally, in the 14th year we get a probability of dying of 2.24%, a probability of being ahead of 98.3%, for a compound probability of 2.20%.

Or, in the 27th year, 4.45% times 64.1% = 2.85%.

Summing these probabilities, for the first 30 years only, I get 60.8%. This would be the chance that a man dies in the first 30 years after age 66, and the the start-at-66 value exceeds the start-at-70 value when that man dies.

But, there are still 21% of our beginning group alive at the end of 30 years, and 53% of the start years still favor at-66. I hate to run FireCalc for more than 30 years, but just eyeballing trends, I'd think about a third of those 21% would die soon enough to favor the start-at-66. This gets up to 67% or a 2-1 advantage for start-at-66.

And, yes, I was cutting it too close when I confidently said that starting early would usually win. I thought it would be more like 80+%. (And, a female would be lower than a male.)

I dunno... seems like a lot of work.

Here's another approach using opensocialsecurity.com that computes expected present values (cash flow * probability and then discounted).

Using default assumptions (single male born 4/15/60 with $1,000 PIA), 2015 SS mortality and default discount rate (20 yr TIP rate or 0.84%):

Age 67 $156,038
Age 70 $155,803
Optimal 68y/5m $157,381

And with no discounting (discount rate = 0%)

Age 67 $176,638
Age 70 $179,249
Optimal 69y/1m $179,838

And with a 5% real return

Age 67 $89,584
Age 70 $82,068
Optimal 62y/1m $98,341

 

[Independent]

I dunno... seems like a lot of work.

Here's another approach using opensocialsecurity.com that computes expected present values (cash flow * probability and then discounted).

Yep, that is an approach that many people have used. Though opensocialsecurity doesn't actually show the math, that was the question.

We know that "historic" investment returns have varied all over. We don't ignore that fact when we think about withdrawal rates, maybe we shouldn't ignore it when we're talking about SS, either. I thought it would be interesting to see what happened if we used FireCalc's data set to ask questions about end of life values.

(I wouldn't use the 2015 SS period table for this question, for reasons I gave in the other thread.)

 

[Independent]

Interesting, but I think 30 years (so 25 or 26 years of collecting delayed SS) is too short a time frame. As many of us have said, the decision to delay SS is not based on a break-even analysis, but as longevity insurance. Individuals cannot count on being 'average' or 'median', but we can insure/prepare in case we live a long time.

With any insurance, you shouldn't expect to 'gain' - you are paying to offset risk. Few people do a break even analysis on the home or car insurance, they look for the best value, because they know they need to protect against risk, and plan to pay for it, in the long run.

Though I can't explain why the advantage dwindled when you looked at 50% success. Probably just different market/inflation conditions shift things. Regardless, all those people who delayed (especially if they have a spouse eligible for SS survivor benefits), gained the longevity insurance. Maybe they never needed it, but that's how insurance works.

-ERD50

I agree with your first two paragraphs. That's why I deferred to 70.

I think the question came from a poster who didn't want/need longevity insurance. He was simply looking to maximize end of life values. I used the word "expectation" in my first reply, intending it to mean the mathematical definition. So, that's kind of what I tried to design my calculation to use.

I can explain why the advantage dwindled when I went to the 50% probability - the middle scenarios controlled the result, and they have higher investment returns than the extremely low scenarios that control for 95%.

I just expected the "early" SS option to show a clear advantage in those scenarios.

 

[ERD50]

...
I can explain why the advantage dwindled when I went to the 50% probability - the middle scenarios controlled the result, and they have higher investment returns than the extremely low scenarios that control for 95%. ....

Ahhh, that makes sense, thanks. -ERD50

 

[eytonxav]

I have not read through this whole thread, but the other consideration regarding breakeven will be the impact of the year SS is no longer able to pay full benefits. Just another variable to think about.

 

[Dtail]

I have not read through this whole thread, but the other consideration regarding breakeven will be the impact of the year SS is no longer able to pay full benefits. Just another variable to think about.

My mindset is they will not reduce benefits if one is past a certain age or already collecting benefits. We shall see.

 

[braumeister]

My mindset is they will not reduce benefits if one is past a certain age or already collecting benefits. We shall see.

That would mean an even bigger haircut for younger folks. I'm counting on an across the board cut.

 

[Dtail]

That would mean an even bigger haircut for younger folks. I'm counting on an across the board cut.

Maybe yes, maybe no. My thoughts are they can clearly raise the contribution limits to unlimited instead of ~ 130k, plus raise the W/H% somewhat.

 

[REWahoo]

I'm counting on an across the board cut.

Based on the history of how our government (mal)functions, our elected officials will wait until the very last minute - or maybe beyond - to solve the SS funding issue. Only when faced with job termination at the ballot box by a very unhappy bunch of seniors will they at last find the courage to do so. The survival instinct is a powerful force.

 

[eytonxav]

It might also be another opportunity for means testing. You place your bets and take your chances.

 

[Chuckanut]

It might also be another opportunity for means testing. You place your bets and take your chances.

I think it's more likely that those above a certain income level (IOW, many people here) will simply pay more income tax on the SS benefits they get.

IIRC, the current number over which SS is taxable was chosen decades ago and is not adjusted for inflation, so this process is already in effect.

 

[eytonxav]

I think it's more likely that those above a certain income level (IOW, many people here) will simply pay more income tax on the SS benefits they get.

Sure that is just another way to implement means testing.

 

[beowulf]

"I think it's more likely that those above a certain income level (IOW, many people here) will simply pay more income tax on the SS benefits they get."

We can only pay more income tax on 15% of our SS - for those of us who are already paying the top marginal rate on 85% of our SS. Unless a new, separate tax rate is enacted for SS benefits, which I doubt. In addition, unless I am mistaken, taxes on SS do not go to SS, they go to the general fund, which does not help the coming SS negative balance.

 

[Nature Lover]

"I think it's more likely that those above a certain income level (IOW, many people here) will simply pay more income tax on the SS benefits they get."

We can only pay more income tax on 15% of our SS - for those of us who are already paying the top marginal rate on 85% of our SS. Unless a new, separate tax rate is enacted for SS benefits, which I doubt. In addition, unless I am mistaken, taxes on SS do not go to SS, they go to the general fund, which does not help the coming SS negative balance.

Good point on where these taxes go. I wonder how much tax is actually collected from benefit recipients on their SS benefit? And how many years would that add to SS's remaining intact if directed back to the program instead of the general fund?

 

[pb4uski]

That would mean an even bigger haircut for younger folks. I'm counting on an across the board cut.

+1 Everything that I have read suggests cuts will be across-the-board so even then current recipients will get benefits cut.... I read something recently that suggested that across-the-board is mandated somehow but I have never been able to find a cite for it.

I think that grandfathering would be unfair and cause such political turmoil that there is no chance of it happening.... also, unless they announce it after the
fact, it would cause a huge amount of people who would be impacted to sign up that day.