Social Security includes a spousal benefit, in which the lower earner is able to take the greater of their own SS income or up to 50% of their spouses’.
Upon death of the higher earner, the lower earner assumes the higher earner’s benefit (replacing their own.) A survivor benefit.
Because of these spousal and survivor benefits many choose to delay collecting Social Security until age 70 in order to secure the largest monthly income, which will help to provide their younger / greater longevity spouse financial security throughout their twilight years.
I will do the opposite, and my wife will be better off for it.
Why Early Social Security Provides the Greatest Spousal Benefit
I previously outlined Why I Plan to Collect Social Security as Early as Possible. More money later is no match for less money early just dumped into an index fund, no matter how long I live.
But it isn’t just me that is affected by this decision, as many pointed out in the comments of that post. The odds are good that you have heard people say, “I’m glad my spouse and I delayed collecting social security. Since their death I have been collecting BIGGER NUMBER which is better than collecting SMALLER NUMBER.”
But the question is NOT, “Do you want a smaller or larger income after my death?”
It is, “Do you want a larger income or a smaller income and a GIANT PILE OF CASH?!”
“We know you love him. But if he happens to die, we give you 2 Mercedes and a summer home. Wouldn’t that be nice too?” – Crazy People, 1990
Investing Early Social Security
Where does the pile of cash come from? Via investing the social security income received during the 8 year period between age 62 and 70.
(That cash pile can be even bigger with a term life insurance policy.)
By definition, if you can choose to delay Social Security you can choose to invest it rather than spend it. (We are doing this for the good of our spouse after all!)
After 8 years of such investment we would have accumulated ~$100k+/- based on typical inflation and investment returns with a PIA of $1,000 for easy math (My PIA is ~$3,000 so we would have accumulated ~$300k.)
Now the larger/delayed Social Security income starts to turn the curve around and at some point breaks even*. Of course, that happens after 25 YEARS!!!! (+/-)
Inflation 3% / Investment returns – various (delta is real return)
If inflation is really low it takes a little longer to break even (curves shift to right due to low COLA). If high, a little less (curves shift to left with high COLA.) Even with a 0% investment return (cash just sitting under a mattress) the breakeven period is 17 years, and (a fun data point) there is no historical 20-year period during which the SP500 return was negative (chart.) By delaying SS, the 20-year return is virtually guaranteed to be negative.
Assuming this is all for the benefit of our spouse… 9% of women and 15% of men who are alive at age 62 are expected to pass before age 70, a fairly small number. “My spouse will still be alive in 8 years… seems reasonable to delay.”
But that number increases to 60% deceased for women and 71% for men at age 87 when waiting for the SS delay to break even (based on 2020 SS period life table single life expectancy.)
Maybe all of this is OK… we delay SS for a greater spousal benefit so in cases where we die, inflation is high, our spouse outlives the longevity tables, that they will be secure.
But wait?! Just because an investment turns positive after 25+ years doesn’t meant it was a good investment or the better choice. And what is that curve in the lower right part of the graph where delay never breaks even?!
Investing for Long-Term Spousal Benefit
In the most likely scenario, my wife will substantially outlive me. She is 5 years younger, has good genes, and has an optimistic and joyful demeanor. Securing her financial future is a strong priority.
To do that I need to understand what our joint SS income looks like prior to my death and what her individual SS income looks like. Fortunately, that is easy to do, and I have plotted that combined scenario in the following chart (assuming a delay in my benefits from age 62 to 70.)
If this chart looks the same as that shown in the post Why I Plan to Collect Social Security as Early as Possible, that’s because it is.
We can think of the spousal survival benefit as just the higher earning spouse living longer. Upon my death she takes over my benefits (replacing her own.)
The ROI of delay increases over time but asymptotically approaches some maximum (~6.5%+/- real.) If both spouses live forever (2 SS incomes), then there is a ~0.6% increase in the maximum. The ROI calculations here are the same for everyone as they are fixed by the SS delay rates – same for me, same for you, same for a spouse, same for a couple with a large age delta.
If we can get a greater return on investment elsewhere, then delay never comes out ahead (which is why the 3%/10% curve (7% real) in the previous chart breaks away from the trend.)
A 3% plus real rate of return with no volatility is a nice thing… to get it we just need to have a guaranteed loss for 25 years +/-. And when (if?) we get there, it isn’t really a great rate of return… I waiting 25+ years for 3%? (Over the past 40 years the SP500 annual return is ~11% nominal (~8% real.))
Summary
I want my spouse to have a solid financial foundation in her twilight years after I am gone.
Therefore I plan to collect social security as early as possible and dump that income into an index fund. This will help avoid the virtually guaranteed loss that delayed Social Security offers for the first 25+/- years.
Upon my death, she will have a giant pile of cash and a monthly social security income stream that together are worth more than if I delayed collecting Social Security, providing the greatest spousal survivor benefit.
As usual, your articles are timely and relevant. We are a FIRE couple and my wife is 9 years younger than me, so statistically she will probably live 15-20 years longer than me. Further complicating this is that I am an “old Dad”, meaning two of my kids will still be minors when I turn 62, so under current SS guidelines, I could collect a benefit for them and my wife as a caretaker. I was really struggling with doing the math on this, as taking SS early would mean a max benefit the first couple of years until the kids age out, but then my wife would have the permanently lower benefit because I took at 62 instead of waiting until 70. Plus she will only get about 40 percent of my current pension once I am gone. But after reading your article, it actually looks like it is a no brainer in terms of taking SS early. Anyway, thanks again for producing such great content over such a long period of time. We sort of followed your footsteps and are in Asia right now with two kids under 5. Not sure if eventually we will go back to the States. Private schools are really pricey abroad….
Yeah man, the international school we looked at in Taipei was like $25k/year/kid. Ouch.
The risk you are taking with collecting early is that you don’t get the investment returns… I’m OK with that over 25+ years.
We have a slightly different outlier situation… My wife has never worked in the US so her PIA is $0. She can’t collect her spousal benefit until I collect my primary, so my delay forces her delay. If I die sometime early (<75 or so) the delay actually reduces our maximum potential rate of return. In other words, delay = less money if I die early.
Yes, we are in the same exact boat. My wife is now a US citizen but has no SS record of her own, never having worked in the US.
That is about what international school cost over here in Jakarta too. We have two kids, so would be looking at 50K+ a year for school. So we have a pretty big decision to make in a couple of years.
We bought a house in a good school district back in New England in 2021 that we are renting out right now. We really missed being near family and affordable help with our two small kids, so we came out to Asia for a vacation and never went back because it was so good. I am a bit loathe to sell the house now with the 3 percent mortgage and the possibility that we might come back in a few years to put the kids in school. The transaction costs of selling the house and furniture are high.
I think right now it is about 80/20 percent we stay in Asia but sequence of return risk could definitely affect our plans. We might end up in California as well, as my wife has realized that she likes it more than Vermont. So we have been following your move and house purchase closely – thank you for being so open about it.
$50k/year on schools might still be cheaper than life in the US, depending (higher costs for transportation, food, childcare, anything requiring human labor…)
Yes, this is very true! Plus no school shootings either.
Would it make sense to put the business income (blog, etc) under her name?
Not really, no.
The spousal benefit gives her 50% of my SS income. The amount of US income she would need to earn to just match that is insanely high, so any SS tax she does pay is basically just money we light on fire.
Why not employ your spouse as admin. of this blog? She can then put in her 40 quarters that way. It’s realistic, a friend of mine did this in his self employed IT business.
The math doesn’t really work out. She already gets 50% of my benefit
Even if she worked 10 years / 40 credits, anything she earned during that time is unlikely to yield a benefit equal to or greater than what she will already receive.
Another way to say that: If she earned income in the US, 100% of her SS taxes would be a total loss, whereas if continue to pay it does increase my benefit somewhat (and therefore hers.)
“waiting 25+ years for 3%”:
[62 + 22 = 84 for 3% actually.]
is a coin toss which investment will out perform. Or tea leaf reading.
When 84, 3% real with no volatility and an average life expectancy of (5.88 + 6.93) / 2 = 6.4 y seems quite adequate.
I agree. It is a fine choice and the one most will make, you just have to outlive those decades of negative returns
These posts are why I keep subscribed to GCC. Very thought-provoking and counter to the norm. It would require a pretty robotic commitment to reinvest into SPY without fail (behavior gap) so I’m not sure the average normie comes out ahead, but I can see the merits for sure.
Nice, thank you. I think it is important to flip conventional wisdom around and see it if actually makes sense. No matter the outcome you at least learn something.
That’s the famous Charlie Munger problem solving technique…..
Invert, always invert.
Well done!
You are right about the behavior gap in general, but some of us have been indexers for decades now — and can be confident we’ll follow through and invest the dough.
You don’t have to actually invest the SS into index funds as long as it is replacing funds you would’ve taken out. When I collect SS at 62 (thanks GCC!!), I’ll probably use it for my everyday expenses while keeping more money invested that I would’ve otherwise had to spend.
Exactly. I retired at 64 and came to the same conclusion. I calculated the breakeven point for delaying to age 70 as 87 years old. I could have drawn down my portfolio (which is doing pretty well) for living expenses, or I could take the SS. I took the SS. I’d rather spend the government’s money first and hang on to mine over which I have control.
that last sentence “spend the government’s money first and hang on to mine” is the key
Pure gold, GCC, as usual. I gotta think about this some more. My wife has never worked in the US, so her PIA is $0. We had a beautiful daughter last year when I turned 50. I think my earlier assumption that delaying in this scenario may be wrong….
Congrats Mr Mighty! I would have to think about this too… there are some bumps in SS for parents of minor children that I haven’t really looked into
👍👍👍
You are my financial advisor. Thank you.
I’ve considered selling some merch
Who wants a T-shirt that says, “Financial Advisors hate me!” with a small GCC logo? :p
I still think the sequence-of-returns risk makes the decision not as simple as you claim.
Indeed. This is one of the reasons I included the case where you just stuff the SS income under the mattress with 0% returns. This is a 17-year delay to break even… putting you at age 80-ish.
Your logic on taking SS as early as possible is the same logic I used when selecting between TRS and ORP when I started my career. TRS is a defined benefit plan and ORP is a defined contribution plan.
The question, roughly, was, “Do I want start with $0 and receive $6,000 per month in retirement? Or do I want to start with $1,200,000 and receive $0 per month in retirement?”
It would take about 17 years just for the pension to amount to $1.2m, and if I took $6,000 per month from the $1.2m, then I’d expect to have ~$1.7m in the that account after 17 years, assuming an average return of 7.2% per year.
One of the easiest financial decisions of my life, and I’ll do exactly as you describe with SS.
Seems like a good move.
I am very glad that I was born at a time when 401ks (defined contribution) replaced pensions (defined benefit.) Can you imagine being stuck in a job where you had no escape hatch other than to fully vest?
I need help! Iv’e been going over this exact thing. The spousal benefit does not increase past my FRA so waiting to age 70 doesn’t make sense. Was planning on waiting to FRA 66 and 10 months. I turn 65 this year. I’m doing Roth IRA conversions and don’t want added taxes if I take SS early while doing my last 2 years of Roth conversions. But then I can save the brokerage account I’m living off of if I take SS early and just do less Roth conversions. So many choices and Im not a mathematician.
You are on a different level, bro. Thank you for turning accepted practices on their heads and sharing your results.
Interesting
I think I will try and wait tell 70 if in good health
I look at SS as a what if we get Japanese type results at some point
Same reason I have rentals
I fully support that choice.
Related reading: Lessons From Japan’s Lost Decade(s)
How does the anticipated OASI reserve depletion (circa 2033) affect your calculations?
Dunno without knowing how payouts would change.
If it is just a straight across the board cut, so all ratios stay the same… then the decision is the same.
I agree with your approach in general, but I plan to use 62-70 for low tax Roth Conversion.
That tax arbitrage has to be considered too.
Have you crunched those numbers? What kind of advantage are you able to eke out?
Does a divorced spouse formerly married 10+ years also benefit by getting the higher earning spousal benefit upon the death of the higher earner? I get a bit more retirement Social Security benefit from my ex’s account to add to my own paid benefit already.
These two articles will be a treat to read later and especially comments that follow.
I can see that for you, Jeremy, taking SS very early might be a very good thing to do because you and your wife have converted all or most of your IRA’s to Roth IRA’s, correct? Would you reach a different conclusion if your IRA was say $2MM instead?
We are not completely FIRE’d. We’re in late 40’s and mid-50’s. I quit my corporate rat race almost 2 yrs ago so I can spend more time running our household which includes two teens. My spouse wants to work another 2-3 years (until our youngest finishes HS) as he finds his job quite enjoyable, so he’ll be around 60 then. Well, our issue (I know, it’s a good one compared to people who cannot save or undersave for their retirement) will be two large 401k’s. His is $1.5mln now and mine is around $700k (along with current AGI income in 24% bracket on the federal level + 5% to our state).
I believe, that taking SS at 62 for us would impede IRA to Roth IRA conversions due to serious taxes we will have to pay on those conversions. Hence our plan is waiting until age of 68 at least. Would you agree or am I missing something?
BTW, I’ve also noticed another thing… this is in regards to IRA conversions to Roth. Some people who have large IRA’s are not very convinced if conversions make sense because supposedly the payoff of each annual conversion is only noticed in more than 20 years (some say it will take 23 years for them while some other say 25 years). Funny enough, you say a similar thing about the payoff of taking SS benefits at 62…25 years.
I don’t think delaying SS from 62 to 6x matters very much in terms of long-term tax efficiency.
The never written Part 3 of Is Your 401k Too Big essentially says don’t worry about it because there isn’t much you can do anyway. Big IRAs get bigger faster than you can convert them… our Traditional IRA/401k is at least 2x bigger than when I quit working 11 years ago despite all of the Roth conversions along the way.
Let’s say you do 8 years of Roth conversions while waiting for SS at age 70… and you do that at 12% or 22% instead of paying 24-32% at age 85 with peak RMDs. The amount of taxable Social Security income and the present value of any tax savings is small relative to the portfolio size.
Very thoughtful analysis, as always.
Am I correct that in either scenario (early or deferred SS collection) it assumes that all payments from the SS income stream get and remain fully invested (and never spent), including during the period after your death? Have you examined the impact of spending down any of the SS payments (and the invested proceeds of the SS payments) to cover living expenses during periods after age 70? I’d guess that early collection would still ultimately win out, but it would be interesting to run the numbers with different expenditure assumptions.
Separate question: you say the second chart depicts the combined scenario of pre-death joint SS income and post-death spousal income. But it looks like the chart does not include the pre-death spousal benefit (separate from the post-death survivor benefit). Am I reading it wrong?
You always ask the best questions!
– You are correct and I have not examined any draw down scenarios. In practice you spend the SS income and leave an equivalent amount of your existing portfolio untouched, but that portfolio has to exist and be sufficiently large (for purposes of this conversation, assume 4% * $0.5 million = $20k SS early income stream.)
– The chart does not include pre-death spousal benefit. If both spouses lives forever and the lower earner has PIA = $0 (spousal benefit has max value), then this adds ~0.6% to the maximum.
Right, but for purposes of this cost-benefit analysis I think it’s best to ignore any existing portfolio altogether in order to isolate the value of the tradeoff of delaying SS. In practice, for people in our shoes, all SS income is probably unneeded surplus, and given that assumption your analysis is flawless in showing that collecting early + investing is probably the best choice for maximizing the size of the surviving spouse’s estate upon her death. But if the goal is really examining which choice is better for “securing the surviving spouse’s financial future” (that is, providing for the surviving spouse’s financial needs while she is still alive), then ideally the analysis should check the outcome under both scenarios using some specified spending/draw down assumptions. (Again, I suspect early collection would still win out, but can’t be sure without running the numbers.)
As a sanity check, I just did some quick back-of-the envelope modeling of the two scenarios in cFIREsim ($25,032 income stream starting immediately vs. $44,076 income stream starting after 8-year delay, in both cases investing in 100% equities and otherwise using cFIREsim’s default settings and assumptions (including the use of actual historical data)). I ran various simulations using different variations on spending patterns (all starting at the 8-year mark or various periods thereafter). Under many plausible spending scenarios, early collection still emerged as the clear winner, but in scenarios with high spending early in the period, early collection did not fare quite as well in terms of historical success rates (for example, if you assume spending the full $44k starting at the 8-year mark, the early collection scenario has a historical failure rate of nearly 40%).
Thanks for posting the sanity check calculations. Useful.
In the high failure scenario you are spending $44k with zero income for 8 years?
No, spending $44k per year starting once the 8-year period is over, after having received $25k in income (and spent zero) for 8 years.
The age-at-death distribution has a long tail. For about half of couples, at least one will live past 90. About 20% of couples will have at least one live past 95. The risk of doing something stupid with a nest-egg rises as people get older. I wouldn’t approach this question with just a financial analysis. There are other dimensions.
People are already really good at making decisions based on other dimensions without ever considering the math. They don’t need me to add to the mix.
I agree – pessimistic and cranky probably means a shorter life:) Are you investing the SS after paying taxes on it – or are taxes paid from other spending? Does it matter? I got fired at 58, then realized I was actually FIRE ready so no big deal. Roth accounts are about 55% of total and I expect the last small Roth conversion this year. But that was because I was doing TSP withdrawals until FRA or 70. The relatively small Trad balances could then be managed by charitable contributions in lieu of RMDs that come to me. But what if those balances could still be bigger by taking SS sooner than planned? SS could allow me to switch to dividends only instead of pulling from principal. Serious food to nibble on while enjoying a good GSM….
The way the article is written does make it a little confusing on the whole “investing” the SS payment.
In reality, there is no investing of the payment. It simply allows less money to be taken out of investments that they already own. Therefore no tax issues to worry about, and no action needed.
In other words: SS payment shows up in bank account and they spend that as opposed to money from selling investments. Dividends likely set to not reinvest and also show up in bank account.
Easy is good.
I never considered SS planning as it is so far away. Now I think I will plan to take it at 62. I figure I’ll give it a 30% downward adjustment in my planning to be (very) conservative.
Modeling out a decently diversified portfolio with a healthy portion of stocks puts returns at roughly 7% real. SS is nowhere near breakeven in that scenario, ever.
I guess I could be pessimistic and say that returns are going to be nowhere near that and the real return will be closer to 5% over the next 50-60 years. In that case SS would outperform when I’m 91 *cough, likely dead, cough*. However, if that’s the case I’d have to be pessimistic and think that SS will not be able to keep up with their payments either. Thus it’s a coin flip on what does better. I’ll side with investments. Thanks.
I always wondered if this would be a good choice for me and my wife. I am 76. I draw my SS. My wife turns 62 in January 25, Is this a good game for us to follows. I do not need any monies. As I receive pensions large enough to take care of needs. I think this would be good choice for her to draw at 62 getting half of mine. Reinvest it all. Them when I pass, she will get mine. What do you think.
If I were in your shoes I would suggest to my wife that she start collecting at 62.
She won’t get 50% of yours but a reduced/early amount based on that cap of 50%
I am following your math, but my one concern is that by taking SS early I would eat up a large portion of my 12% tax bracket that I am using for Roth conversions each year. It seems like I will pay more in taxes overall if I take SS early and reduce the amount of Trad IRA conversions I can make before max SS (and RMDs soon after) kick in in my 70s. Did you see any tax considerations in this regard in your calculations of the two options and how would you see that impacting which path wins? I know the tax rate for SS income varies based on other income (I have a military pension that takes up a chunk of my 12% bracket already). I like the thought of having a pile of tax free money in our Roth accounts when I reach RMD land. Thanks for the post — you have me re-thinking my SS strategy. Cheers!
Have you done the math? What percent difference do you see for the present value of all future taxes paid? I don’t think the difference is large.