Open Social Security Calc - appropriate discount rate

Non-investing personal finance issues including insurance, credit, real estate, taxes, employment and legal issues such as trusts and wills
Post Reply
Topic Author
MNJim
Posts: 1
Joined: Mon Mar 29, 2021 11:45 am

Open Social Security Calc - appropriate discount rate

Post by MNJim »

In looking at the Open Social Securty calculator for the first time, the default discount rate caught my attention. The model uses the current TIPS rate which at the time was a negative .18%. In running different models with even a small positive discount rate, say +1%, the advantage of deferring to age 70 is signifciantly reduced. So this rate could impact your election decision. I am interested in learning if others use the default or modify it to their own situation. Thanks.
User avatar
ObliviousInvestor
Posts: 3933
Joined: Tue Mar 17, 2009 9:32 am
Contact:

Re: Open Social Security Calc - appropriate discount rate

Post by ObliviousInvestor »

You're quite right that discount rate meaningfully affects the analysis.

There are two reasons for using TIPS yields as the default discount rate.

First is that, as a general finance principle, we try to select a discount rate that is the expected return from another investment option with the most similar level of risk. While TIPS do not have exactly the same risk characteristics as Social Security, they are the most similar thing available (inflation-adjusted, backed by Treasury).

The second is that it is the safest assets that will be (or at least, should be) spent down in order to fund the delaying of Social Security. So it is those returns that are being given up.

We discussed this somewhat during the event on Saturday. (Just mentioning that since your username suggests you might be in the MN Bogleheads group.) Here's another related discussion though:
viewtopic.php?p=5362672#p5362672
Last edited by ObliviousInvestor on Mon Mar 29, 2021 2:13 pm, edited 1 time in total.
Mike Piper | Roth is a name, not an acronym.
User avatar
JoeRetire
Posts: 8710
Joined: Tue Jan 16, 2018 2:44 pm

Re: Open Social Security Calc - appropriate discount rate

Post by JoeRetire »

MNJim wrote: Mon Mar 29, 2021 12:09 pmI am interested in learning if others use the default or modify it to their own situation.
I have used the default. And I have also used a custom discount rate.

If you have some reason to believe your personal discount rate is other than the default, you should use your own numbers.
Just remember: it's not a lie if you believe it.
User avatar
Ben Mathew
Posts: 1687
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Open Social Security Calc - appropriate discount rate

Post by Ben Mathew »

TIPS yields make the most sense if you believe that

(A) The US govt's promise of increased SS payouts by delaying

is about as good as

(B) The US govt's promise that it will pay out when the TIPS matures

If you think that (A) is less of a sure bet than (B), then you can ask yourself how much higher of a yield does (A) have to offer for it to equivalent to (B). Use the higher yield in the calculation instead.
Kiana
Posts: 21
Joined: Wed Jul 08, 2020 9:08 am

Yes, Another Open Social Security Discount Question

Post by Kiana »

[Post merged into here --admin LadyGeek]

I have read Mike Piper's paper on why TIPS is the default discount rate on Open Social Security, as well as many recent forum posts on the question, but I am still not getting it.

I understand that TIPS are the closest asset class to Social Security income, and that you should sell bonds to buy Social Security credits, but I can't seem to marry the two. Assume I have a retirement portfolio of 50% VTWAX / 50% VBTLX. I know I should sell VBTLX to defer SS, but would I not use the expected return of VBTLX as the discount rate since that is the return that I am giving up? I don't own TIPS nor would I purchase them with SS income.

Thanks for any help,
Kevin
User avatar
Ben Mathew
Posts: 1687
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Yes, Another Open Social Security Discount Question

Post by Ben Mathew »

Kiana wrote: Thu Jun 17, 2021 5:39 pm I have read Mike Piper's paper on why TIPS is the default discount rate on Open Social Security, as well as many recent forum posts on the question, but I am still not getting it.

I understand that TIPS are the closest asset class to Social Security income, and that you should sell bonds to buy Social Security credits, but I can't seem to marry the two. Assume I have a retirement portfolio of 50% VTWAX / 50% VBTLX. I know I should sell VBTLX to defer SS, but would I not use the expected return of VBTLX as the discount rate since that is the return that I am giving up? I don't own TIPS nor would I purchase them with SS income.

Thanks for any help,
Kevin
You are not necessarily wrong. Let's consider an example:

The price of a bowl of caviar is $100 and the price of a bowl of soup is $10. Clearly some people value a bowl of caviar more than a bowl of soup. They're the ones who go out and buy it. But you like a bowl of soup and a bowl of caviar the same. Someone offers to replace your bowl of soup with a bowl of caviar. How much would you be willing to pay?

Answer:

-If you can resell your caviar to someone who values it more than you, you should be okay with paying up to $100 (value of caviar) -$10 (value of soup) = $90.

- If you can't resell your caviar to someone who values it more, you should be willing to pay only up to $10 (value of caviar) -$10 (value of soup)= $0.

So your personal valuation matters when you can't resell the goods. And since you can't resell Social Security, you can apply your own personal higher discount rate to value it.

If you already hold TIPS, then your value must be the going market rate--that's why you bought it. That's why OSS uses the market value for TIPS--it assumes that you value TIPS as much as the market does. But if you wouldn't otherwise hold TIPS, your personal value might be less than the market value--and that lower value would be the right value to use.

That said, I would suggest that you think about why your personal valuation of TIPS is less than the market's valuation. Is there a reason that you don't like TIPS as much as the other people who have bought them?
LongRoad
Posts: 168
Joined: Mon Dec 31, 2018 2:00 pm

Re: Yes, Another Open Social Security Discount Question

Post by LongRoad »

Kiana wrote: Thu Jun 17, 2021 5:39 pm [Post merged into here --admin LadyGeek]

I have read Mike Piper's paper on why TIPS is the default discount rate on Open Social Security, as well as many recent forum posts on the question, but I am still not getting it.

I understand that TIPS are the closest asset class to Social Security income, and that you should sell bonds to buy Social Security credits, but I can't seem to marry the two. Assume I have a retirement portfolio of 50% VTWAX / 50% VBTLX. I know I should sell VBTLX to defer SS, but would I not use the expected return of VBTLX as the discount rate since that is the return that I am giving up? I don't own TIPS nor would I purchase them with SS income.

Thanks for any help,
Kevin
Also keep in mind the calculator uses a real (rather than nominal) discount rate, since SS benefits are inflation-adjusted.

Estimating the expected real return of VBTLX is less certain than for TIPS, because future inflation is unknown. But isn’t the expected real return of VBTLX currently slightly negative, as well?
Big Dog
Posts: 2734
Joined: Mon Sep 07, 2015 4:12 pm

Re: Yes, Another Open Social Security Discount Question

Post by Big Dog »

Kiana wrote: Thu Jun 17, 2021 5:39 pm [Post merged into here --admin LadyGeek]

I have read Mike Piper's paper on why TIPS is the default discount rate on Open Social Security, as well as many recent forum posts on the question, but I am still not getting it.

I understand that TIPS are the closest asset class to Social Security income, and that you should sell bonds to buy Social Security credits, but I can't seem to marry the two. Assume I have a retirement portfolio of 50% VTWAX / 50% VBTLX. I know I should sell VBTLX to defer SS, but would I not use the expected return of VBTLX as the discount rate since that is the return that I am giving up? I don't own TIPS nor would I purchase them with SS income.

Thanks for any help,
Kevin
When I worked on some finance projects for megacorp, we always tied the discount rate to like projects. More risky projects received a higher discount rate; less risky projects might use AAA corporates or even Treasuries.

And the reason that you would not use Total Bond for a discount rate is that it includes corporates which have higher risk than does SS, as they are not guaranteed. Now one could make the case to use Treasuries, based on your longevity table, but then you'd have to adjust for inflation. TIPS take care of both.

(I'm a big fan of Mike P.)
User avatar
ObliviousInvestor
Posts: 3933
Joined: Tue Mar 17, 2009 9:32 am
Contact:

Re: Yes, Another Open Social Security Discount Question

Post by ObliviousInvestor »

Kiana wrote: Thu Jun 17, 2021 5:39 pm I understand that TIPS are the closest asset class to Social Security income, and that you should sell bonds to buy Social Security credits, but I can't seem to marry the two. Assume I have a retirement portfolio of 50% VTWAX / 50% VBTLX. I know I should sell VBTLX to defer SS, but would I not use the expected return of VBTLX as the discount rate since that is the return that I am giving up? I don't own TIPS nor would I purchase them with SS income.
Sorry for the delay in replying -- I have been traveling and away from internet/cell service for some days.

As Big Dog noted, even still, the expected rate of return from the Total Bond fund is not the appropriate discount rate. Because if you exchange a Total Bond fund for Social Security, you are reducing risk. When choosing a discount rate, you always want to keep risk level the same (or as near as possible).
Mike Piper | Roth is a name, not an acronym.
LivingTheDream
Posts: 100
Joined: Sun Oct 18, 2020 10:30 am
Location: Savannah, GA

Re: Open Social Security Calc - appropriate discount rate

Post by LivingTheDream »

I recently read an article on this subject by Michael Kitces, and he argued (at least for me, successfully) that discount rate should be specific to funds being used in place of collecting early SS benefits. For an early retiree with a 65/35 portfolio, who rebalances annually (does NOT use dedicated "buckets" for the up to 8-year delay), the discount rate should be your personal expected rate of return. I don't own TIPs, nor do I expect to withdraw only bonds from age 62 to 70. With that being the case, it makes sense to me that my "cost" of delaying is the money I would otherwise spend from my annually rebalanced portfolio.

A rising glidepath might complicate things, and I didn't get a great feel for nominal vs. real (I thought Kitces argued in favor of nominal, but I don't recall details).

The article for anyone interested... https://www.kitces.com/blog/net-present ... breakeven/
Living The Dream
User avatar
Ben Mathew
Posts: 1687
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Open Social Security Calc - appropriate discount rate

Post by Ben Mathew »

LivingTheDream wrote: Sat Jun 19, 2021 8:09 pm I recently read an article on this subject by Michael Kitces, and he argued (at least for me, successfully) that discount rate should be specific to funds being used in place of collecting early SS benefits. For an early retiree with a 65/35 portfolio, who rebalances annually (does NOT use dedicated "buckets" for the up to 8-year delay), the discount rate should be your personal expected rate of return. I don't own TIPs, nor do I expect to withdraw only bonds from age 62 to 70. With that being the case, it makes sense to me that my "cost" of delaying is the money I would otherwise spend from my annually rebalanced portfolio.

A rising glidepath might complicate things, and I didn't get a great feel for nominal vs. real (I thought Kitces argued in favor of nominal, but I don't recall details).

The article for anyone interested... https://www.kitces.com/blog/net-present ... breakeven/
Here's why I think Kitces' analysis in that article was deeply flawed.
LivingTheDream
Posts: 100
Joined: Sun Oct 18, 2020 10:30 am
Location: Savannah, GA

Re: Open Social Security Calc - appropriate discount rate

Post by LivingTheDream »

Ben Mathew wrote: Sat Jun 19, 2021 9:42 pm
LivingTheDream wrote: Sat Jun 19, 2021 8:09 pm I recently read an article on this subject by Michael Kitces, and he argued (at least for me, successfully) that discount rate should be specific to funds being used in place of collecting early SS benefits. For an early retiree with a 65/35 portfolio, who rebalances annually (does NOT use dedicated "buckets" for the up to 8-year delay), the discount rate should be your personal expected rate of return. I don't own TIPs, nor do I expect to withdraw only bonds from age 62 to 70. With that being the case, it makes sense to me that my "cost" of delaying is the money I would otherwise spend from my annually rebalanced portfolio.

A rising glidepath might complicate things, and I didn't get a great feel for nominal vs. real (I thought Kitces argued in favor of nominal, but I don't recall details).

The article for anyone interested... https://www.kitces.com/blog/net-present ... breakeven/
Here's why I think Kitces' analysis in that article was deeply flawed.
But doesn't it depend on whether an individual investor (specifically me) would actually adjust risk to match? If I don't hold (& don't plan to) similarly low-risk vehicles, wouldn't the real (rather than academic) analysis be focused on true opportunity cost? I'm not spending TIPs money while deferring SS; instead, I'm spending from the total return of my portfolio.
Living The Dream
Big Dog
Posts: 2734
Joined: Mon Sep 07, 2015 4:12 pm

Re: Open Social Security Calc - appropriate discount rate

Post by Big Dog »

LivingTheDream wrote: Sat Jun 19, 2021 8:09 pm I recently read an article on this subject by Michael Kitces, and he argued (at least for me, successfully) that discount rate should be specific to funds being used in place of collecting early SS benefits. For an early retiree with a 65/35 portfolio, who rebalances annually (does NOT use dedicated "buckets" for the up to 8-year delay), the discount rate should be your personal expected rate of return. I don't own TIPs, nor do I expect to withdraw only bonds from age 62 to 70. With that being the case, it makes sense to me that my "cost" of delaying is the money I would otherwise spend from my annually rebalanced portfolio.

A rising glidepath might complicate things, and I didn't get a great feel for nominal vs. real (I thought Kitces argued in favor of nominal, but I don't recall details).

The article for anyone interested... https://www.kitces.com/blog/net-present ... breakeven/
Personally, I'm skeptical of any recommendation that nearly always makes the wealth advisor wealthier. (A higher discount rate means take SS early instead of spending down investments which an advisor is getting paid to invest.)
LivingTheDream
Posts: 100
Joined: Sun Oct 18, 2020 10:30 am
Location: Savannah, GA

Re: Open Social Security Calc - appropriate discount rate

Post by LivingTheDream »

Big Dog wrote: Sat Jun 19, 2021 10:43 pm
LivingTheDream wrote: Sat Jun 19, 2021 8:09 pm I recently read an article on this subject by Michael Kitces, and he argued (at least for me, successfully) that discount rate should be specific to funds being used in place of collecting early SS benefits. For an early retiree with a 65/35 portfolio, who rebalances annually (does NOT use dedicated "buckets" for the up to 8-year delay), the discount rate should be your personal expected rate of return. I don't own TIPs, nor do I expect to withdraw only bonds from age 62 to 70. With that being the case, it makes sense to me that my "cost" of delaying is the money I would otherwise spend from my annually rebalanced portfolio.

A rising glidepath might complicate things, and I didn't get a great feel for nominal vs. real (I thought Kitces argued in favor of nominal, but I don't recall details).

The article for anyone interested... https://www.kitces.com/blog/net-present ... breakeven/
Personally, I'm skeptical of any recommendation that nearly always makes the wealth advisor wealthier. (A higher discount rate means take SS early instead of spending down investments which an advisor is getting paid to invest.)
I'm a DIY guy (no FA), so I'm hoping to pay more to myself :)

Although, depending on estimated longevity, even a relatively high discount rate still optimizes at FRA (or later). My biggest focus right now is to try to quantify PV so that I can easily compare different claiming strategies. For instance, for me, using a discount rate of 4.5% and life-expectancy of 95 shows optimal claiming to be 70. BUT, the difference is relatively small ($10-20k in PV). Using 92 & 0.5% shows a diff. of $150k+.
Living The Dream
User avatar
Ben Mathew
Posts: 1687
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Open Social Security Calc - appropriate discount rate

Post by Ben Mathew »

LivingTheDream wrote: Sat Jun 19, 2021 10:03 pm
Ben Mathew wrote: Sat Jun 19, 2021 9:42 pm
LivingTheDream wrote: Sat Jun 19, 2021 8:09 pm I recently read an article on this subject by Michael Kitces, and he argued (at least for me, successfully) that discount rate should be specific to funds being used in place of collecting early SS benefits. For an early retiree with a 65/35 portfolio, who rebalances annually (does NOT use dedicated "buckets" for the up to 8-year delay), the discount rate should be your personal expected rate of return. I don't own TIPs, nor do I expect to withdraw only bonds from age 62 to 70. With that being the case, it makes sense to me that my "cost" of delaying is the money I would otherwise spend from my annually rebalanced portfolio.

A rising glidepath might complicate things, and I didn't get a great feel for nominal vs. real (I thought Kitces argued in favor of nominal, but I don't recall details).

The article for anyone interested... https://www.kitces.com/blog/net-present ... breakeven/
Here's why I think Kitces' analysis in that article was deeply flawed.
But doesn't it depend on whether an individual investor (specifically me) would actually adjust risk to match? If I don't hold (& don't plan to) similarly low-risk vehicles, wouldn't the real (rather than academic) analysis be focused on true opportunity cost? I'm not spending TIPs money while deferring SS; instead, I'm spending from the total return of my portfolio.
It only depends on whether you are risk averse. As long as you are risk averse, you will want to adjust for risk when constructing and comparing your options.

The framing:

-- I'm consuming out of a 65/35 portfolio, therefore I should use a 65/35 discount rate

is misleading in a couple of ways, one of which is that it makes an incorrect assumption already--that you should consume out of the 65/35 portfolio (i.e. maintain a 65/35 allocation) like you would have if you hadn't delayed SS.

The question to ask would be:

-- If I delay SS, what should the new asset allocation of my savings portfolio be?

Presumably, since you have less bonds in the SS account for the pre-SS gap years, you should have more bonds in the savings account for those years. So your new allocation for the gap years would be lower than 65/35.
Last edited by Ben Mathew on Sun Jun 20, 2021 11:39 pm, edited 1 time in total.
LivingTheDream
Posts: 100
Joined: Sun Oct 18, 2020 10:30 am
Location: Savannah, GA

Re: Open Social Security Calc - appropriate discount rate

Post by LivingTheDream »

Ben Mathew wrote: Sun Jun 20, 2021 12:19 am
The question to ask would be:

-- If I delay SS, what should the new asset allocation of my savings portfolio be?

Presumably, since you have more bonds in the SS account, you should have less bonds in the savings account. So your new savings allocation would be higher than 65/35. You'd be holding less bonds and more stocks through the pre-SS "gap" years. The allocation and withdrawal strategy you used in the rest of the gap years can apply to the "extra" gap years as well.
This sounds (to me) like the "home equity as a bond" discussion. On that topic, I consider home equity outside of my AA (at least until/unless I liquidate it through either sale or reverse mortgage). Similarly, regarding SS benefits, I consider it a cash/income stream rather than as a bond (at least in relation to my AA).

I approach determining my AA as partly related to my risk tolerance & partly related to the degree to which I want/need my portfolio to produce reliable/consistent/predictable retirement income/withdrawals. Currently, with SS being 7-15 years away, I am comfortable with a 65/35 AA. However, once I begin taking SS benefits, my portfolio will need to produce less annual income; so I'll be comfortable with an AA more in the area of 75/25 [the opposite of what I think you suggest].

Ultimately, I'll likely delay taking SS (perhaps solely in consideration for potential spousal benefits for my wife). And I'm a generally conservative planner (for estimations, not necessarily for risk), so I'll also probably use a fairly low discount rate to try to quantify PV of future benefits as I get closer to 62.

While I understand that you're telling me using my opportunity cost is flawed/incorrect, for now I feel more comfortable with using Kitces analysis. But I do appreciate you sharing your knowledge & insights.
Living The Dream
User avatar
Ben Mathew
Posts: 1687
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Open Social Security Calc - appropriate discount rate

Post by Ben Mathew »

LivingTheDream wrote: Sun Jun 20, 2021 5:55 pm
Ben Mathew wrote: Sun Jun 20, 2021 12:19 am
The question to ask would be:


-- If I delay SS, what should the new asset allocation of my savings portfolio be?

Presumably, since you have more less bonds in the SS account [for the gap years], you should have less more bonds in the savings account [for those years]. So your new savings allocation would be higher lower than 65/35 [for those years]. You'd be holding less more bonds and more [the same amount of] stocks through for the pre-SS "gap" years. The allocation and withdrawal strategy you used in the rest of the gap years can apply to the "extra" gap years as well.
This sounds (to me) like the "home equity as a bond" discussion. On that topic, I consider home equity outside of my AA (at least until/unless I liquidate it through either sale or reverse mortgage). Similarly, regarding SS benefits, I consider it a cash/income stream rather than as a bond (at least in relation to my AA).

I approach determining my AA as partly related to my risk tolerance & partly related to the degree to which I want/need my portfolio to produce reliable/consistent/predictable retirement income/withdrawals. Currently, with SS being 7-15 years away, I am comfortable with a 65/35 AA. However, once I begin taking SS benefits, my portfolio will need to produce less annual income; so I'll be comfortable with an AA more in the area of 75/25 [the opposite of what I think you suggest].

Ultimately, I'll likely delay taking SS (perhaps solely in consideration for potential spousal benefits for my wife). And I'm a generally conservative planner (for estimations, not necessarily for risk), so I'll also probably use a fairly low discount rate to try to quantify PV of future benefits as I get closer to 62.

While I understand that you're telling me using my opportunity cost is flawed/incorrect, for now I feel more comfortable with using Kitces analysis. But I do appreciate you sharing your knowledge & insights.
Sorry--looks like I should have gone to bed last night instead of posting on Bogleheads. "Less" should have been "more" and "more" should have been "less" in the post above. Edited to fix.
User avatar
Ben Mathew
Posts: 1687
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Open Social Security Calc - appropriate discount rate

Post by Ben Mathew »

A graph can help illustrate why the bond yield rather than the portfolio return is the appropriate discount rate to value Social Security.

Option A: Take SS at start of retirement

Consider someone who retires at age 65 and plans to take SS immediately. Planned retirement income is $20,000 each coming in from stocks, bonds and Social Security for a total of $60,000 per year. Assume bonds and SS are of comparable risk.

Expected retirement income looks like this:

Image

Option B: Delay Social Security

If the retiree delays SS to age 70 and uses bonds to replace SS during the gap years, retirement income looks like this:

Image

Because bonds and SS are of equal risk, the risk to retirement income has not changed. It's still $40,000 per year of safe income + $20,000 per year of risky income from stocks.

What the Net Present Value (NPV) Calculation Means:

If you use the bond rate as the discount rate to value Social Security, then the Net Present Value (NPV) of delaying SS (NPV of SS in option B - NPV of SS in option A) is the extra $ you will have in your hand today if you choose option B over option A. If the NPV is $50,000, that means that if you delay SS to age 70 and fill in the gap with bonds, then you will have an extra $50,000 left over to spend as you wish. You can spend it on a home remodel or a vacation or whatever else you want. A natural option is to use it to increase retirement income--i.e. get more than the $60,000 per year that you would have had if you had taken SS early.
Post Reply