I don't understand the case for EE bonds

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: something odd regarding EE bonds

Post by Kevin M »

jdilla1107 wrote:
nisiprius wrote: but anyway, $10,000 for 20 years at 2.69% compounded gets you $17,000 instead of $20,000.
You are forgetting tax deferral, which makes that gap significantly wider for people who are high tax bracket now and low at retirement. EE bonds expand my tax deferral space somewhat significantly. Try adding a 35% tax rate over that 20 years with 0% in retirement.
The original thread had petered out, but has been revived with some new blood in the new, now merged thread, and going from 35% tax rate to 0% is a new twist. How do you get to a marginal tax rate of 0% in retirement after being at 35% while working? With a decent savings rate, it's more likely that your marginal tax rate will be at least 25%, especially after RMDs and SS kick in, or if you do Roth conversions before then. I'm at 15% without any Roth conversions, and I do Roth conversions to the top of the 15% bracket, so any more income would be taxed at 25% (federal).

It's possible if you are able to take the qualified educational expense (QEE) exemption on 100% of your EE Bond redemption, but that requires a number of conditions. You must still have kids in school, your income must be below the exemption limit, and you must not have consumed the QEE with education tax credits.

The last one is what got us for 2016. We had used up the AOTC in previous years, and the LLC consumed $10K of QEE, leaving only a small amount of QEE to apply toward our $20+K I Bond redemption. Didn't amount to much tax savings.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
letsgobobby
Posts: 12073
Joined: Fri Sep 18, 2009 1:10 am

Re: something odd regarding EE bonds

Post by letsgobobby »

Deleted
Last edited by letsgobobby on Sat Apr 27, 2019 12:38 pm, edited 1 time in total.
letsgobobby
Posts: 12073
Joined: Fri Sep 18, 2009 1:10 am

Re: something odd regarding EE bonds

Post by letsgobobby »

Deleted
Last edited by letsgobobby on Sat Apr 27, 2019 12:38 pm, edited 1 time in total.
User avatar
flamesabers
Posts: 1848
Joined: Fri Mar 03, 2017 11:05 am
Location: Rochester, MN

Re: something odd regarding EE bonds

Post by flamesabers »

willthrill81 wrote:
jdilla1107 wrote:
willthrill81 wrote:
The interest rate risk is not to be underestimated and neither is inflation risk. If inflation averages at the Fed's target then the pre-tax real return of EE bonds would be about 1.5%.
People on this board keep trying to compare EE bonds to stocks and other investments. They should only be compared to long term bonds. If you don't want to hold long term bonds, then don't. But, it's not like liquidity and rebalancing save you from the risks of long term bonds. In fact, EE bonds are lower risk than a 20 year treasury as they essentially have a free put option in the beginning, which is also quite valuable.
Why? Should we only compare one deferred annuity to another? Should we only compare REITs with REITs? I don't follow this logic.
I think it's a matter of doing an apples-to-apples comparison in certain circumstances.

If someone wanted to invest in a fixed income investment such as a 7-year bank CD that has a 3% interest rate, would you bring up the returns you can get by investing in equities for 7 years?

If on the other hand the person was wondering what the best investment was for the ideal rate of return for say 7 or 10 years, then it makes sense to compare bonds vs. equities, assuming the person has the necessary tolerance for volatility.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: something odd regarding EE bonds

Post by willthrill81 »

jdilla1107 wrote:Because you should decide on what your investment plan is with an appropriate AA first. Then you want to decide on what the best implementation of that plan and AA is going to be.
You should start with your investment goals and risk tolerance. Nobody should want stocks just because their AA calls for it, for instance. At best, that's circular logic. If there were another investment vehicle that could achieve the returns of stocks without their volatility, you had better believe that I would be all over it.

No investment vehicle should be considered in isolation from an investment plan, and an AA is not an investment plan. It's the product of a plan.
The Sensible Steward
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: something odd regarding EE bonds

Post by willthrill81 »

flamesabers wrote:
willthrill81 wrote:
jdilla1107 wrote:
willthrill81 wrote:
The interest rate risk is not to be underestimated and neither is inflation risk. If inflation averages at the Fed's target then the pre-tax real return of EE bonds would be about 1.5%.
People on this board keep trying to compare EE bonds to stocks and other investments. They should only be compared to long term bonds. If you don't want to hold long term bonds, then don't. But, it's not like liquidity and rebalancing save you from the risks of long term bonds. In fact, EE bonds are lower risk than a 20 year treasury as they essentially have a free put option in the beginning, which is also quite valuable.
Why? Should we only compare one deferred annuity to another? Should we only compare REITs with REITs? I don't follow this logic.
I think it's a matter of doing an apples-to-apples comparison in certain circumstances.

If someone wanted to invest in a fixed income investment such as a 7-year bank CD that has a 3% interest rate, would you bring up the returns you can get by investing in equities for 7 years?

If on the other hand the person was wondering what the best investment was for the ideal rate of return for say 7 or 10 years, then it makes sense to compare bonds vs. equities, assuming the person has the necessary tolerance for volatility.
I agree. Historically, there's been a lot more potential for downside risk with equities over a 7 year period than a 20 year period. Comparing stocks to bonds over a one year period is nonsensical. Over a twenty year or longer period, it's prudent, with the stipulation of tolerance for volatility.
The Sensible Steward
jwhitaker
Posts: 70
Joined: Thu Jun 16, 2016 12:57 pm

Re: I don't understand the case for EE bonds

Post by jwhitaker »

Thank you all for the discussion. I am leaning towards buying some and the arguments on here have really challenged me. Here are my thoughts:

-I looked at CAPE and subsequent 20 year nominal returns on stocks. I'm eyeballing about 5 or 6%. Not a prediction, just the input to my mental math. So a guaranteed 3.5% versus a risky 5 or even 6% seems like a decent tradeoff to me.

-I also looked at historical long term interest rates (this is the Schiller data from Yale website and its 10 yr bonds I think). Last time it was at 2% or so in 1940, it took until 1956 to get above a 3.5% rate and the average from 1940 to 1960 was 2.7%. I realize inflation and stocks did stuff during that time. Also history. So it's possible that the waiting for better yields thing doesn't work out.

-I'm not worried about the 20 year lockup. I also don't plan on touching IRA/401k assets and view the penalty as in the same range (35% marginal tax + 10% penalty), versus 50% at 19.9 years for an EE. In a full liquidation, nuclear scenario on my portfolio where I need to convert every last dollar to cash I have bigger fish to fry than missed yield. This isn't my emergency fund, it's part of my bond allocation.

-I want to have a total 20-30% bond allocation. I'm at 20% now. I max tax advantaged space elsewhere, 401k, Roth, 529. I like the deferral of the income, even if there's a possibility it's not tax rate arbitrage.

So I'm trying to park 10k in taxable meant for long term investment, and I want to increase my bond allocation. I'm also in a high marginal bracket. So what are my alternatives? There's high yield municipal which is yielding 3%. I bonds are at 2.7%, but they could go to 0. Long term treasuries are below 3.5%. I also already do crowdfunding (real estate and personal loans). I expect returns in the 4-5% range for that. Those are very tax inefficient, but you have to add back in the fun factor. I can't do Swedroe's CAT bonds because that is highly correlated to my income, plus you have to go through an advisor.

I don't love the EE bond, but it's an option among other not-that-great options.
Last edited by jwhitaker on Mon Aug 07, 2017 8:39 am, edited 1 time in total.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: something odd regarding EE bonds

Post by Kevin M »

willthrill81 wrote: I agree. Historically, there's been a lot more potential for downside risk with equities over a 7 year period than a 20 year period. Comparing stocks to bonds over a one year period is nonsensical. Over a twenty year or longer period, it's prudent, with the stipulation of tolerance for volatility.
Volatility is a vague description, but I think most people use it to describe variance or standard deviation of short-term returns--say 1-month to 1-year. From an emotional perspective (aka, willingness to take risk), short-term volatility is an appropriate measure, even though your investment horizon might be much longer than that. Believing in your head that stocks are not risky in the long run does you no good if your gut tells you otherwise as you watch your portfolio decline by 50% or more, and you dump your stocks when the pain gets to be too great, and then you hesitate to get back into stocks until they have recovered most of their value.

From the perspective of ability to take risk, ignoring the emotional component of risk tolerance, for a 20-year investment horizon, the volatility of 20-year returns is what matters. Some refer to this as terminal wealth dispersion (TWD). This is where we really don't have enough historical data to draw valid statistical conclusions. There are only about four independent 20 year periods since 1926. If we start the clock in 1929, one out of four of those periods had an annualized return of less than 3.5%. From this perspective, EE Bonds don't look so bad.

The stocks-are-not-risky-in-the-long-term folks like to use rolling periods instead of independent non-overlapping periods, but statisticians tell us that that is not valid statistically, since rolling periods have much data in common. But it makes stocks look a lot less risky over 20-year periods, so the stocks-are-not-risky-in-the-long-term folks like to do it. You'll see lots of this if you review the entire thread.

We're covering ground that's already been covered, but it's a long thread, and the stocks-are-not-risky-in-the-long-term folks have not addressed this point.

Implicit in the investment policy of anyone with a 20-year investment horizon (and guts of steel) who owns bonds of any type, whether EE or not, is the belief that there's a reasonable probability that stocks will underperform bonds over that investment horizon. If we view coming up with an AA as a decision tree, whether or not you believe this is the first decision to make. We can eliminate a lot of the debate here by simply agreeing to place ourselves into one camp or the other. There are plenty of other threads in which this has been debated.

If you do believe it, then the next decision involves comparing EE bonds to other fixed-income choices. A lot of the thread has been about this, and I think it's the more interesting part of the discussion.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

jwhitaker wrote:So a guaranteed 3.5% versus a risky 5 or even 6% seems like a decent tradeoff to me.
That's a pretty big difference after compounding.
I want to have a total 20-30% bond allocation.
It still doesn't make sense to me to set a generic bond allocation then hunt for bonds regardless of whether they actually serve your purposes for having a bond allocation in the first place.
Last edited by NiceUnparticularMan on Fri Apr 28, 2017 3:36 pm, edited 1 time in total.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

Kevin M wrote: There are only about four independent 20 year periods since 1926. If we start the clock in 1929, one out of four of those periods had an annualized return of less than 3.5%. From this perspective, EE Bonds don't look so bad.
There can be good reasons to separate out completely independent periods. This isn't one of them. Based on the historic record, there is not in fact a 25% chance that starting in 2017 will be like starting in 1929. That's just an artifact of when we happened to start collecting the data.

If you take this problem seriously, you'll still end up concluding that there is only a very small chance of 2017 being like 1929, but you might get a somewhat wider spread of the estimate of possible outcomes, some on the good side and some on the bad. This will somewhat increase your odds of it being 1929 or WORSE, but not really that much since 1929 was on the far end of the curve to begin with.

Now personally, for the reasons given before I think this is a bounded range UNLESS there is a truly epic disaster in the United States. That sort of fat tail risk does concern me, but I think it pushes to international investment, not EE bonds.
If you do believe it, then the next decision involves comparing EE bonds to other fixed-income choices. A lot of the thread has been about this, and I think it's the more interesting part of the discussion.
I agree, and I'll just note that the scenarios in which it is 1929 or WORSE, and you haven't invested internationally, don't really favor EE bonds over other fixed-income choices. That's because those scenarios are quite likely to include one or more events that will make you wish your EE bonds were liquid.

So the thing I can continue to struggle with is people who believe there is a good chance stocks will underperform 3.5% nominal (possible, but this is an extremely bad scenario), and then suggest the lack of liquidity doesn't matter to them. I think that requires imagining a truly unlikely scenario.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

nisiprius wrote:I guess you can make an argument that they beat Treasuries... what is the 20-year Treasury rate now? 2.69%. I think it's hard to reinvest Treasury interest and no guarantee of getting the same effective rate as the initial coupon... but anyway, $10,000 for 20 years at 2.69% compounded gets you $17,000 instead of $20,000. Still, I think that in most cases you'd be better off with the Treasury if you needed the money before the twenty years were up, and I don't think there are 20-year bank CDs.
Yeah, I wouldn't buy 20-year Treasuries with a 100% certainty I would hold them to maturity. I'd only buy them if I thought there was a chance I might actually want to sell them at some point earlier. But I wouldn't buy EE bonds either with a 100% certainty I would them to maturity--I'd buy something else entirely.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: something odd regarding EE bonds

Post by willthrill81 »

Kevin M wrote:
willthrill81 wrote: I agree. Historically, there's been a lot more potential for downside risk with equities over a 7 year period than a 20 year period. Comparing stocks to bonds over a one year period is nonsensical. Over a twenty year or longer period, it's prudent, with the stipulation of tolerance for volatility.
Volatility is a vague description, but I think most people use it to describe variance or standard deviation of short-term returns--say 1-month to 1-year. From an emotional perspective (aka, willingness to take risk), short-term volatility is an appropriate measure, even though your investment horizon might be much longer than that. Believing in your head that stocks are not risky in the long run does you no good if your gut tells you otherwise as you watch your portfolio decline by 50% or more, and you dump your stocks when the pain gets to be too great, and then you hesitate to get back into stocks until they have recovered most of their value.
If you can't take the heat, don't go into the kitchen. Buy EE bonds or something else instead.
Kevin M wrote:From the perspective of ability to take risk, ignoring the emotional component of risk tolerance, for a 20-year investment horizon, the volatility of 20-year returns is what matters. Some refer to this as terminal wealth dispersion (TWD). This is where we really don't have enough historical data to draw valid statistical conclusions. There are only about four independent 20 year periods since 1926. If we start the clock in 1929, one out of four of those periods had an annualized return of less than 3.5%. From this perspective, EE Bonds don't look so bad.

The stocks-are-not-risky-in-the-long-term folks like to use rolling periods instead of independent non-overlapping periods, but statisticians tell us that that is not valid statistically, since rolling periods have much data in common. But it makes stocks look a lot less risky over 20-year periods, so the stocks-are-not-risky-in-the-long-term folks like to do it. You'll see lots of this if you review the entire thread.

We're covering ground that's already been covered, but it's a long thread, and the stocks-are-not-risky-in-the-long-term folks have not addressed this point.
Stocks are volatile no matter how long they are held, but this does not mean that returns over twenty year holding periods are as volatile as annual returns, which is what some espousing this argument imply. The returns for twenty year periods have indeed been volatile themselves, but the volatility is all on the positive side (the 50th percentile for S&P 500 20 year returns is 6.886%).
viewtopic.php?t=210755

Certainly if one focuses on non-overlapping periods the results can change just as they change if one examines 10 year vs. 20 year rolling returns. But saying that those are more valid than rolling periods seems to ignore the significant variation that can occur from from overlapping period to the next. Sequence of returns risk is very real as retirees from 1973, 1974, and 1975 all had very different outcomes (see http://www.firecalc.com homepage for an illustration). When the totality of the data are examined and all potential investment points are considered, it becomes clear that holding equities for two decades has not only always had positive returns, but nearly always returns far exceeding 3.5%.
Kevin M wrote:Implicit in the investment policy of anyone with a 20-year investment horizon (and guts of steel) who owns bonds of any type, whether EE or not, is the belief that there's a reasonable probability that stocks will underperform bonds over that investment horizon.
For one to believe this, one must believe that there is a strong enough possibility that the future will look so different from the past as to turn the equity premium on its head for two decades such that you would be willing to sacrifice a historical 10.1% nominal return for a 3.5% return. That smacks of market timing and gambling to me.
The Sensible Steward
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: something odd regarding EE bonds

Post by willthrill81 »

NiceUnparticularMan wrote:So the thing I can continue to struggle with is people who believe there is a good chance stocks will underperform 3.5% nominal (possible, but this is an extremely bad scenario), and then suggest the lack of liquidity doesn't matter to them. I think that requires imagining a truly unlikely scenario.
I'll take that a step further and say that it would likely be a rare investor indeed for whom liquidity truly did not matter but volatility did.
The Sensible Steward
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

jdilla1107 wrote:People on this board keep trying to compare EE bonds to stocks and other investments. They should only be compared to long term bonds.
They are actually like none of those things, all in very important albeit different ways. On this logic you should compare them to nothing. But the practical investor should feel free to compare them to anything, given that if you invest in one thing you can't invest the same money in something else.
In fact, EE bonds are lower risk than a 20 year treasury as they essentially have a free put option in the beginning, which is also quite valuable.
We showed the free put has a very limited range of scenarios in which it would be valuable, and that CDs would dominate in those scenarios.

Generally, that just showed there are SOME liquidation scenarios in which EE bonds would beat 20 year Treasuries. There are other liquidation scenarios in which 20 year Treasuries would beat EE bonds. Now that we are considering liquidation scenarios and associated risks, we should consider all of them, not just the ones that happen to favor EE bonds.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

willthrill81 wrote:I'll take that a step further and say that it would likely be a rare investor indeed for whom liquidity truly did not matter but volatility did.
Yep. Really, the scenario would have to be something like a known nominal liability in 20 years, at which point you could maybe argue any volatility around that final value was a huge issue, and yet liquidity was not.

Which sounds pretty far-fetched, at least for normal individual investors.
letsgobobby
Posts: 12073
Joined: Fri Sep 18, 2009 1:10 am

Re: I don't understand the case for EE bonds

Post by letsgobobby »

Deleted
Last edited by letsgobobby on Sat Apr 27, 2019 12:39 pm, edited 1 time in total.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

I don't have a strong statistics background, so can't really argue the case, but I don't think our two vocal EE-bond critics have presented a compelling argument that justify using rolling 20-year periods for a valid statistical analysis from which we can determine a statistically valid 20-year expected return or variance of the probability distribution. The true mean of the distribution might be significantly lower than the sample mean from the limited history, and the dispersion of future returns could be much wider.

EDIT: Nisiprius often cites an academic paper that demonstrates this, especially the point about our uncertainty about the true mean.

I agree about the big difference changing the start or end date by just a little can make, and this is one intuitive reason I think looking at rolling periods still has value. But I can't argue it statistically, and my limited understanding is that rolling periods can't be used to determine valid statistical sample means.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: I don't understand the case for EE bonds

Post by willthrill81 »

Kevin M wrote:I don't have a strong statistics background, so can't really argue the case, but I don't think our two vocal EE-bond critics have presented a compelling argument that justify using rolling 20-year periods for a valid statistical analysis from which we can determine a statistically valid 20-year expected return or variance of the probability distribution. The true mean of the distribution might be significantly lower than the sample mean from the limited history, and the dispersion of future returns could be much wider.

EDIT: Nisiprius often cites an academic paper that demonstrates this, especially the point about our uncertainty about the true mean.

I agree about the big difference changing the start or end date by just a little can make, and this is one intuitive reason I think looking at rolling periods still has value. But I can't argue it statistically, and my limited understanding is that rolling periods can't be used to determine valid statistical sample means.

Kevin
I see little point in getting into the statistical weeds as even those with Ph.D.s in finance argue about this. (I will point out that rolling periods of market returns on average have higher returns than the simple geometric (CAGR) mean because most of these rolling periods do not include the Great Depression.)

But this I do know, and I've said it several times before. There has only been one twenty year period in which investing in stocks would have had very slightly worse performance than EE bonds; the rest of them beat it hands down. Regardless as to what the variance, probability distribution, or anything else are, that's very telling.
The Sensible Steward
jdilla1107
Posts: 849
Joined: Sun Jun 24, 2012 8:31 pm

Re: something odd regarding EE bonds

Post by jdilla1107 »

NiceUnparticularMan wrote:
jdilla1107 wrote:People on this board keep trying to compare EE bonds to stocks and other investments. They should only be compared to long term bonds.
They are actually like none of those things, all in very important albeit different ways. On this logic you should compare them to nothing. But the practical investor should feel free to compare them to anything, given that if you invest in one thing you can't invest the same money in something else.
This doesn't say anything meaningful. We are now at "How many angels can dance on the head of a pin"?
There are other liquidation scenarios in which 20 year Treasuries would beat EE bonds.
There are no scenarios of this that don't involve doing something wrong or being able to see the future. It's like saying there are some scenarios where $10 is worth more than $20.

Words have meaning and this thread has none.
Last edited by jdilla1107 on Fri Apr 28, 2017 8:04 pm, edited 3 times in total.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

willthrill81 wrote: I see little point in getting into the statistical weeds as even those with Ph.D.s in finance argue about this.
It's important if you're going to base your expected return and risk estimates on historical data. Expected value is a statistically determined number, as is variance or standard deviation. If you're using flawed statistics, whether explicitly or implicitly to determine these numbers, then your conclusions are flawed.

I just think it's worth pushing back on so cavalierly basing one's estimates of future probabilities on a potentially flawed statistical basis. I appreciate some of the statistics jocks here having pointed out this flaw in using rolling periods, and although I'm not 100% convinced about the merits one way or the other, I think it's worth being aware of in a discussion like this.
But this I do know, and I've said it several times before. There has only been one twenty year period in which investing in stocks would have had very slightly worse performance than EE bonds; the rest of them beat it hands down. Regardless as to what the variance, probability distribution, or anything else are, that's very telling.
And you're right back to using rolling periods without having addressed the potentially flawed statistical basis on which you're basing your conclusions. There's a big difference whether it's one out of four or one out of 70 periods we're talking about. You can't just wave away the fact that a probability distribution based on four independent samples is essentially meaningless, and conclude that your results are "very telling".

For the purposes of this thread, why don't we just acknowledge that some people aren't 100% confident that stocks will outperform EE bonds over the next 20 years? There are plenty of threads on whether or not stocks are risky in the long run in which this kind of things has been discussed ad nauseam, and it's clear that there is no consensus on it here.

So the simple answer to the OP's question is, "I think that there's a reasonable probability that stocks will return less than 3.5% over the next 20 years." You, the OP, and the other vocal critics of this view can disagree, but I don't see you changing any minds of the vocal EE bond proponents. We don't see letsgobobby or market timer saying, "You know what, you've convinced me that stocks are not risky over a 20-year period; I'm going to sell my EE bonds tomorrow and use the money to buy stocks."

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
knpstr
Posts: 2894
Joined: Thu Nov 20, 2014 7:57 pm
Location: Michigan

Re: I don't understand the case for EE bonds

Post by knpstr »

Kevin M wrote: So the simple answer to the OP's question is, "I think that there's a reasonable probability that stocks will return less than 3.5% over the next 20 years." You, the OP, and the other vocal critics of this view can disagree, but I don't see you changing any minds of the vocal EE bond proponents. We don't see letsgobobby or market timer saying, "You know what, you've convinced me that stocks are not risky over a 20-year period; I'm going to sell my EE bonds tomorrow and use the money to buy stocks."
Reminds me of Cialdini's book Influence!: the consistency and commitment principle, regarding human misjudgments.

Very difficult for a human to change their mind once they've made a commitment to a position. Due to self-created justifications they stay consistent with it even in the face of new and/or conflicting information.
Very little is needed to make a happy life; it is all within yourself, in your way of thinking. -Marcus Aurelius
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

knpstr wrote: Reminds me of Cialdini's book Influence!: the consistency and commitment principle, regarding human misjudgments.

Very difficult for a human to change their mind once they've made a commitment to a position. Due to self-created justifications they stay consistent with it even in the face of new and/or conflicting information.
And that applies as well to the stocks-are-not-risky-in-the-long-run crowd, who continue to use overlapping, non-independent samples as their primary justification on which to base their confidence, despite having the statistics repeatedly challenged (not so much in this thread, other than by me, but many, many times in other threads on this topic).

There is nothing new in the "information" that U.S. stocks have the best long-term track record if you use rolling periods as your basis. I guarantee you that letsgobobby and market timer are well aware of the historical record.

What may be new information for some folks reading this thread is that using overlapping, non-independent samples may not be a valid basis on which to draw your statistical conclusions. I certainly wouldn't gamble my financial well being on a potentially flawed analysis once this was pointed out to me. I'd want to at least dig into the statistics of it to see how valid the criticism is. It would be interesting for someone with a strong statistics background to weigh in on this (as they have done in other threads on the topic, and all I've ever heard them say is that using rolling periods is not a valid way to come up with mean and variance estimates).

What you may find interesting is that I don't have a dog in this fight. I don't own EE bonds, and haven't yet been convinced that I should. But I do see a very specific way in which they could outperform the fixed income products that I do own, and I think it's worth challenging my own views on this. Hence, I like to straddle the fence and debate both sides.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: I don't understand the case for EE bonds

Post by willthrill81 »

Kevin M wrote:
knpstr wrote: Reminds me of Cialdini's book Influence!: the consistency and commitment principle, regarding human misjudgments.

Very difficult for a human to change their mind once they've made a commitment to a position. Due to self-created justifications they stay consistent with it even in the face of new and/or conflicting information.
And that applies as well to the stocks-are-not-risky-in-the-long-run crowd, who continue to use overlapping, non-independent samples as their primary justification on which to base their confidence, despite having the statistics repeatedly challenged (not so much in this thread, other than by me, but many, many times in other threads on this topic).
People repeatedly challenge passive vs. active stock picking too. That doesn't necessarily mean that the former is incorrect.
Kevin M wrote:There is nothing new in the "information" that U.S. stocks have the best long-term track record if you use rolling periods as your basis. I guarantee you that letsgobobby and market timer are well aware of the historical record.

What may be new information for some folks reading this thread is that using overlapping, non-independent samples may not be a valid basis on which to draw your statistical conclusions. I certainly wouldn't gamble my financial well being on a potentially flawed analysis once this was pointed out to me. I'd want to at least dig into the statistics of it to see how valid the criticism is. It would be interesting for someone with a strong statistics background to weigh in on this (as they have done in other threads on the topic, and all I've ever heard them say is that using rolling periods is not a valid way to come up with mean and variance estimates).
Very few financial experts believe in the 'non-overlapping' data argument as evidenced by their analyses covering specific years rather than merely four blocks. And for those that adhere to the 'non-overlapping' argument, I suppose they want us to wait until we have hundreds of years of data before we reach an adequate sample size on which to base a conclusion.

Below is a recent thread discussing this issue. In comparing ten year blocks and ten year rolling averages, the mean was identical.
viewtopic.php?t=212785

But again, I'm not getting into the weeds because that's OT. No one, to my knowledge, has reached a definitive conclusion on this matter, so it's pointless to argue over it.
Last edited by willthrill81 on Fri Apr 28, 2017 8:54 pm, edited 1 time in total.
The Sensible Steward
User avatar
knpstr
Posts: 2894
Joined: Thu Nov 20, 2014 7:57 pm
Location: Michigan

Re: I don't understand the case for EE bonds

Post by knpstr »

Kevin M wrote: And that applies as well to the stocks-are-not-risky-in-the-long-run crowd, who continue to use overlapping, non-independent samples as their primary justification on which to base their confidence, despite having the statistics repeatedly challenged (not so much in this thread, other than by me, but many, many times in other threads on this topic).

There is nothing new in the "information" that U.S. stocks have the best long-term track record if you use rolling periods as your basis. I guarantee you that letsgobobby and market timer are well aware of the historical record.

What may be new information for some folks reading this thread is that using overlapping, non-independent samples may not be a valid basis on which to draw your statistical conclusions. I certainly wouldn't gamble my financial well being on a potentially flawed analysis once this was pointed out to me. I'd want to at least dig into the statistics of it to see how valid the criticism is. It would be interesting for someone with a strong statistics background to weigh in on this (as they have done in other threads on the topic, and all I've ever heard them say is that using rolling periods is not a valid way to come up with mean and variance estimates).

What you may find interesting is that I don't have a dog in this fight. I don't own EE bonds, and haven't yet been convinced that I should. But I do see a very specific way in which they could outperform the fixed income products that I do own, and I think it's worth challenging my own views on this. Hence, I like to straddle the fence and debate both sides.

Kevin
To be clear, the misjudgment tendency can and does apply to all of us.

When facing uncertain future, unfortunately, all we can go off of is past performance, even while admitting it may not indicate future performance. Even those that argue the future will be such in such way they come to those conclusions by looking at periods similar in the past and what happened then - which may or may not be true in the future.

So the correct bet (which is what all of us are doing, betting) on what is going to have better returns over a 20 year time period is to put your money in what usually works out best in 20 year time periods. Even if the results are poor, that was still the correct bet.

Another misjudgment "trap" is with what you said about EE bonds earlier - if you don't hold them for 20 years it isn't 3.5% but 0.1% - that will provide another very difficult hurdle for someone to psychologically jump to sell out early and "to lose their return" to switch to more advantageous investments.
Very little is needed to make a happy life; it is all within yourself, in your way of thinking. -Marcus Aurelius
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

letsgobobby wrote:We're nearing 10 years of low growth, low inflation.

No reason we won't have 10, or 20 more.
And compounded earnings has returned way more than 3.5% nominal in that period.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

Kevin M wrote:I don't have a strong statistics background, so can't really argue the case, but I don't think our two vocal EE-bond critics have presented a compelling argument that justify using rolling 20-year periods for a valid statistical analysis from which we can determine a statistically valid 20-year expected return or variance of the probability distribution. The true mean of the distribution might be significantly lower than the sample mean from the limited history, and the dispersion of future returns could be much wider.
I don't disagree with the dispersion part of this (I am not sure I am one of the EE bond critics, but if so). As I suggested before, if you do this correctly, you won't get a different central expectation, but you will get a wider dispersion. That won't result in 25% chance of 1929 or worse, but it will result in something higher than the about 1% you would get if you treated these as non-overlapping periods.

I do think for fundamental reasons this is a bounded distribution, but that goes beyond just statistical analysis. For that matter, I also believe the central expectation should be a bit lower, but that also goes beyond just statistical analysis. However, on no reasonable analysis will you end up with the odds of stocks returning less than 3.5% nominal being anything close to as high as 25%.
lazyday
Posts: 3849
Joined: Wed Mar 14, 2007 10:27 pm

Re: I don't understand the case for EE bonds

Post by lazyday »

[after dividends]
NiceUnparticularMan [url=https://www.bogleheads.org/forum/viewtopic.php?f=10&t=217081&start=200#p3345893]here[/url] wrote:, I'd rather just take out the remainder of my desired earnings through selling a few shares
If you’re taking out all of the earnings, then I think you need to worry about dilution. I believe Bill Bernstein and Rob Arnott estimated (pdf) 2% dillution, and that Bernstein later posted here that dilution more recently has been 1%. If you want to see the post, let me know I can probably find it. I think it included the words Arnott (or Rob) and “peek”. Also the January AQR paper I mentioned earlier might estimate net buybacks, after dilution I suppose. I can easily find that link.
So you can instead just keep rough track of the overall dividend payout ratio of your portfolio, and periodically sell shares to get out the additional earnings you want, and I think that will be good enough (
In fact, even if there is some massive devaluation event between the time the earnings are retained and reinvested and when you sell those few shares, say as much as a year later, at most you are getting the devaluation on one year's worth of earnings--and only for the ones retained and reinvested. In the greater scheme, this should be no big deal as long as you do have a cushion for spending. ....

I like this framework for thinking about safely retiring on an equity heavy portfolio. Of course you would need to be more conservative than those who use a similar concept but with dividends instead of earnings.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

jdilla1107 wrote:This doesn't say anything meaningful. We are now at "How many angels can dance on the head of a pin"?
I quite disagree. The huge penalty for early withdrawal from EE bonds is a very important consideration. That makes them unlike other long bonds.
There are no scenarios of this that don't involve doing something wrong or being able to see the future. It's like saying there are some scenarios where $10 is worth more than $20.
There is a large deflationary event. Long bonds significantly increase in market value while stocks crash. You want to sell some of the long bonds to rebalance into stocks.

What did I do wrong?

The put scenario is the opposite sort of scenario. Rates increase a lot, and I decide I would like to withdraw something out of my long bonds. If there is enough of a rate increase and it happens soon enough after buying, it might be better to withdraw from EE bonds than from other long bonds. Note that both of these are bad outcomes, but withdrawing from normal long bonds is indeed a worse outcome than withdrawing from EE bonds in a limited range of such circumstances.

If you are going to consider the second scenario as a valuable feature of EE bonds, why are you refusing to consider the first scenario as a valuable feature of normal long bonds?

Now I might note it is pretty odd to set things up such that you have to withdraw from long bonds in the event of a rate increase. If you think that is a significant possibility, you are much better off with something like CDs. And personally, I'd say the whole purpose of having long bonds at all is precisely so that you can withdraw from them in the event of a rate decrease, most particularly in the event of a big deflationary scenario.
lazyday
Posts: 3849
Joined: Wed Mar 14, 2007 10:27 pm

Re: something odd regarding EE bonds

Post by lazyday »

jdilla1107 wrote:People on this board keep trying to compare EE bonds to stocks and other investments. They should only be compared to long term bonds. ....
If Larry were here, he might say to think about each within the context of the whole portfolio, not in isolation.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

Kevin M wrote:So the simple answer to the OP's question is, "I think that there's a reasonable probability that stocks will return less than 3.5% over the next 20 years." You, the OP, and the other vocal critics of this view can disagree, but I don't see you changing any minds of the vocal EE bond proponents. We don't see letsgobobby or market timer saying, "You know what, you've convinced me that stocks are not risky over a 20-year period; I'm going to sell my EE bonds tomorrow and use the money to buy stocks."
I agree there isn't much chance of changing any minds among the participants in the debate, particularly since once you have bought EE bonds, it is very hard to undo that decision. But since this is a public forum, there likely will be silent readers who may not be committed yet to a position on EE bonds.

For such people, I think it is indeed worth explaining why U.S. stocks normally return much more than 3.5% nominal, and what it would likely take for them to fail to do so. You can then use that understanding to ask whether EE bonds are really the best insurance against those scenarios, or rather alternatives like normal long bonds and/or international stocks.

And who knows--I might actually be persuaded myself if someone came up with a good explanation of what practical role EE bonds could serve. Having not bought EE bonds yet doesn't mean I couldn't in the future.
jdilla1107
Posts: 849
Joined: Sun Jun 24, 2012 8:31 pm

Re: something odd regarding EE bonds

Post by jdilla1107 »

NiceUnparticularMan wrote:
jdilla1107 wrote:
There are no scenarios of this that don't involve doing something wrong or being able to see the future. It's like saying there are some scenarios where $10 is worth more than $20.
There is a large deflationary event. Long bonds significantly increase in market value while stocks crash. You want to sell some of the long bonds to rebalance into stocks.

What did I do wrong?
You held too many EE bonds by being forced to use them to rebalance.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

lazyday wrote:If you’re taking out all of the earnings, then I think you need to worry about dilution. I believe Bill Bernstein and Rob Arnott estimated (pdf) 2% dillution, and that Bernstein later posted here that dilution more recently has been 1%.
Absolutely! I think in practice this means you should in fact not take out ALL the earnings but reinvest at least some percentage (1% at a minimum, could be more). I also think these are grounds for "tilting" your investments to, say, small value and EM, as a way of trying to capture a bit more of that "new enterprise" effect (note that small growth is often assumed to be the best proxy for such a strategy, but I think that has it backward when you think about what "value" and "growth" really mean).

At current valuations, you might think you could draw as much as 6% from a global portfolio before accounting for dilution. My target in withdrawal would actually be to average more like 3-4%, which is hopefully more than low enough to account for this effect.
I like this framework for thinking about safely retiring on an equity heavy portfolio. Of course you would need to be more conservative than those who use a similar concept but with dividends instead of earnings.
Indeed. Relying on dividend ratios is very risky, in my view.
jdilla1107
Posts: 849
Joined: Sun Jun 24, 2012 8:31 pm

Re: something odd regarding EE bonds

Post by jdilla1107 »

lazyday wrote:
jdilla1107 wrote:People on this board keep trying to compare EE bonds to stocks and other investments. They should only be compared to long term bonds. ....
If Larry were here, he might say to think about each within the context of the whole portfolio, not in isolation.
It's obvious to me that both positions are right.

This is how I designed my portfolio:

1) How much risk can I handle taking in stocks without ever freaking out?
2) Decide on my AA between stocks and bonds
3) Decide how to implement the bond and stock side of my portfolio separately.

These are the big questions. I believe it's important to do it in this order. The biggest mistake people make is selling stocks when they go down cause they can't take the pain anymore.

4) On the bond side, I next decide to implement a diversified bond portfolio
5) I decide not to use TBM, because I am going to use munis, optimize for taxes, avoid short treasuries, use CDs, use EE and I bonds.

By #5, I am at optimization stage. The big questions were already answered.

But, to go back at #5 and say "maybe I should hold more stocks because bonds are worse", is a recipe for failure. I went through 2008 with a large portfolio and it was hard. Questioning your risk tolerance during portfolio optimization is how people confuse themselves, in my opinion.
Last edited by jdilla1107 on Sat Apr 29, 2017 8:42 am, edited 1 time in total.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

jdilla1107 wrote:
NiceUnparticularMan wrote:
jdilla1107 wrote:
There are no scenarios of this that don't involve doing something wrong or being able to see the future. It's like saying there are some scenarios where $10 is worth more than $20.
There is a large deflationary event. Long bonds significantly increase in market value while stocks crash. You want to sell some of the long bonds to rebalance into stocks.

What did I do wrong?
You held too many EE bonds by being forced to use them to rebalance.
Well, in this scenario I didn't hold any EE bonds at all.

What happened is I benefited from holding some normal long bonds. And for what it is worth, this possible purpose is precisely why I would ever hold long bonds. I would not otherwise see them worth the risk.
jdilla1107
Posts: 849
Joined: Sun Jun 24, 2012 8:31 pm

Re: something odd regarding EE bonds

Post by jdilla1107 »

NiceUnparticularMan wrote:
jdilla1107 wrote:
NiceUnparticularMan wrote:
jdilla1107 wrote:
There are no scenarios of this that don't involve doing something wrong or being able to see the future. It's like saying there are some scenarios where $10 is worth more than $20.
There is a large deflationary event. Long bonds significantly increase in market value while stocks crash. You want to sell some of the long bonds to rebalance into stocks.

What did I do wrong?
You held too many EE bonds by being forced to use them to rebalance.
Well, in this scenario I didn't hold any EE bonds at all.

What happened is I benefited from holding some normal long bonds. And for what it is worth, this possible purpose is precisely why I would ever hold long bonds. I would not otherwise see them worth the risk.
If you don't want to hold long term bonds, then EE bonds make no sense. But, if you hold TBM or another bond fund, then one might want to check if they are holding long bonds in their funds.

The way I look at it is TBM holds about 15% long bonds. So, EE bonds should likely be kept under 10% of a bond portfolio. For some people, this means "don't hold them".
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

jdilla1107 wrote:This is how I designed my portfolio:

1) How much risk can I handle taking in stocks without ever freaking out?
2) Decide on my AA between stocks and bonds
3) Decide how to implement the bond and stock side of my portfolio separately.
So this is indeed the approach that doesn't make sense to me. Stocks certainly do have risks. So you allocate some funds to non-stocks to address those risks. But then you forget about what stock-related risks you were trying to address when deciding which non-stocks to hold?

I really can't understand why that makes sense. I get it is a pretty common attitude, but I really do think people should be thinking about how the disconnection introduced at Step (3) may result in your portfolio failing to satisfy the risk-management goals established in Steps (1) and (2).
But, to go back at #5 and say "maybe I should hold more stocks because bonds are worse", is a recipe for failure. I went through 2008 with a large portfolio and it was hard. Questioning your risk tolerance during portfolio optimization is how people confuse themselves, in my opinion.
At least personally, that's not really what I am saying. As before, I am saying that disconnection you introduced at Step 3 means your bond mix may not be serving the risk-management goals established after Steps (1) and (2). Asking "why not stocks" isn't supposed to motivate the conclusion "you should hold stocks instead," but instead is a way of recalling exactly what risks warranted not holding all stocks in the first place. What you actually do next could be many different things--the point would be to re-establish the connection between your risk-management goals and your bond mix, not to dictate a particular end result.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: something odd regarding EE bonds

Post by NiceUnparticularMan »

jdilla1107 wrote:If you don't want to hold long term bonds, then EE bonds make no sense. But, if you hold TBM or another bond fund, then one might want to check if they are holding long bonds in their funds.
To be frank, TBM doesn't make sense to me either. Again, bonds in my view are risk-management tools. Different entities use different bonds for different risk-management purposes. Holding TBM in my view is therefore just an arbitrary mish-mash of a bunch of different tools being used by a bunch of different entities, many of whom are nothing like me.

It would be like determining that for a certain purpose, I needed a hammer. But then I buy a tool that is made up 10% of hammer parts, but 90% of other tool parts.
The way I look at it is TBM holds about 15% long bonds. So, EE bonds should likely be kept under 10% of a bond portfolio. For some people, this means "don't hold them".
Again, EE bonds have quite different features from normal long bonds. So even if for some risk-management purpose you thought it was worth holding 15% nominal long bonds, that wouldn't necessarily mean it made sense to instead hold 5% normal long bonds and 10% EE bonds. Because they really aren't the same things.
letsgobobby
Posts: 12073
Joined: Fri Sep 18, 2009 1:10 am

Re: I don't understand the case for EE bonds

Post by letsgobobby »

Deleted
Last edited by letsgobobby on Sat Apr 27, 2019 12:41 pm, edited 1 time in total.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

letsgobobby wrote:
NiceUnparticularMan wrote:
For such people, I think it is indeed worth explaining why U.S. stocks normally return much more than 3.5% nominal, and what it would likely take for them to fail to do so.
irrelevant, except for demonstrating once again that after 250 posts, you have completely missed the point.
letsgobobby, why is it irrelevant? The OP asked why invest in EE Bonds when stocks are likely to return more than 3.5% over a 20-year period. So it seems that a discussion of the likelihood of stocks doing that is relevant.

If we believe that historical U.S. stock returns form a reliable data set on which we can base a probability distribution of future 20-year returns, and if we believe that using rolling 20-year periods is a statistically valid way to do that, then determining the mean and variance of the historical, rolling 20-year returns is something we'd want to do.

I think it's worth challenging the two "ifs" in that statement, but if someone believes both those things, the data does show that the mean and variance of the distribution would leave little to no reason to hold bonds of any kind, including EE bonds, for a 20-year holding period. One reason might be low willingness to take risk because one finds monthly or annual volatility too painful, and that's a valid reason that's really kind of hard to argue against.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
jdilla1107
Posts: 849
Joined: Sun Jun 24, 2012 8:31 pm

Re: I don't understand the case for EE bonds

Post by jdilla1107 »

Kevin M wrote: letsgobobby, why is it irrelevant? The OP asked why invest in EE Bonds when stocks are likely to return more than 3.5% over a 20-year period. So it seems that a discussion of the likelihood of stocks doing that is relevant.

Kevin
There are two questions buried in this one question:

1) Why should I hold bonds over a 20 year period?
2) Why should I hold EE bonds over other long term fixed income investments?

It's relevant if you are answering question #1 and irrelevant if you are answering question #2. This thread is constantly bouncing between those questions. Some people are answering #1 and some #2.

Then there are the people who are going to argue that those two questions are the same question. :oops:
letsgobobby
Posts: 12073
Joined: Fri Sep 18, 2009 1:10 am

Re: I don't understand the case for EE bonds

Post by letsgobobby »

Deleted
Last edited by letsgobobby on Sat Apr 27, 2019 12:41 pm, edited 2 times in total.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

NiceUnparticularMan wrote:
Kevin M wrote:I don't have a strong statistics background, so can't really argue the case, but I don't think our two vocal EE-bond critics have presented a compelling argument that justify using rolling 20-year periods for a valid statistical analysis from which we can determine a statistically valid 20-year expected return or variance of the probability distribution. The true mean of the distribution might be significantly lower than the sample mean from the limited history, and the dispersion of future returns could be much wider.
As I suggested before, if you do this correctly, you won't get a different central expectation, but you will get a wider dispersion.
Well, the part I underlined is easy to prove wrong.

For rolling 20-year periods since 1929, the mean (expected value) of 20-year cumulative returns is 833% and the standard deviation of those returns is 470%. For the four independent 20-year periods since 1929, the mean is 599% and the SD is 591%.
That won't result in 25% chance of 1929 or worse, but it will result in something higher than the about 1% you would get if you treated these as non-overlapping periods.
If we assume a normal probability distribution, and if we assume that our analysis is statistically valid (which it probably is not, either due to using rolling periods or because four independent periods isn't enough to have much statistical validity), then the independent periods indicate roughly 1/6 probability that our 20-year cumulative return would be 8% or less (compared to a 100% 20-year return for EE bond). So about a 17% chance that the stock return will be much worse than the EE bond return, not a 25% chance that it will be slightly worse.

That's a lot worse than the 1/6 chance that the 20-year return would be 181% or less you get with the rolling returns.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

jdilla1107 wrote:
Kevin M wrote: letsgobobby, why is it irrelevant? The OP asked why invest in EE Bonds when stocks are likely to return more than 3.5% over a 20-year period. So it seems that a discussion of the likelihood of stocks doing that is relevant.
There are two questions buried in this one question:

1) Why should I hold bonds over a 20 year period?
2) Why should I hold EE bonds over other long term fixed income investments?

It's relevant if you are answering question #1 and irrelevant if you are answering question #2. This thread is constantly bouncing between those questions. Some people are answering #1 and some #2.
Yes, that's exactly right, and I even mentioned in an earlier reply that I thought #2 was the more interesting question. However, the OP seemed to be more interested in #1:
sometimesinvestor wrote:My main reason for doubt is the unlikelihood of not getting a higher return in the stock market if I am in it for 20 years. I have read there is no 15 year period where one lost money in the stock market so there must be a decent probability of success.
So the emphasis of the OP's question definitely was on comparing EE bonds to stocks for a 20-year horizon.

I almost wrote in an earlier reply that we should just agree to disagree about #1, so we can focus on the more interesting (to me) #2, but then I re-read the OP, and saw that #1 was the crux of the concern, and that there was no explicit mention at all about comparing EE Bonds to other fixed income.

So I'm still waiting for letsgobobby to explain why it's irrelevant to be discussing the 20-year expected return of stocks vs. EE Bonds, since that's exactly what the OP mentioned as the primary concern.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: I don't understand the case for EE bonds

Post by willthrill81 »

Kevin M wrote:
NiceUnparticularMan wrote:
Kevin M wrote:I don't have a strong statistics background, so can't really argue the case, but I don't think our two vocal EE-bond critics have presented a compelling argument that justify using rolling 20-year periods for a valid statistical analysis from which we can determine a statistically valid 20-year expected return or variance of the probability distribution. The true mean of the distribution might be significantly lower than the sample mean from the limited history, and the dispersion of future returns could be much wider.
As I suggested before, if you do this correctly, you won't get a different central expectation, but you will get a wider dispersion.
Well, the part I underlined is easy to prove wrong.

For rolling 20-year periods since 1929, the mean (expected value) of 20-year cumulative returns is 833% and the standard deviation of those returns is 470%. For the four independent 20-year periods since 1929, the mean is 599% and the SD is 591%.
As I said before, that's due to the fact that most 20 year rolling periods do not include the Great Depression. That one event has a huge impact on the data when viewed in their entirety.

The question then becomes which one is more useful for investors. If we use mean returns from beginning to end (very similar to the non-overlapping period argument), that would be analogous to someone who put a lump sum in at, say, 1926, and held it until now; note that dollar cost averaging would improve the returns significantly since so little would be invested during the GD.

Alternatively, 20 year rolling periods are analogous to investors who put in a lump sum and held it for twenty years (again, no DCA). Among all of those investors, only one would have come out with less than a 3.5% nominal return (and only slightly at that).

There's little disputing that, despite its potentially questionable statistical properties, the second approach is more akin to the issue brought up by the OP. For instance, if we look at this as randomly 'dropping an investor' into any year in the available data and they invested a lump sum, the rolling period data seem quite reasonable.
The Sensible Steward
letsgobobby
Posts: 12073
Joined: Fri Sep 18, 2009 1:10 am

Re: I don't understand the case for EE bonds

Post by letsgobobby »

Deleted
Last edited by letsgobobby on Sat Apr 27, 2019 12:41 pm, edited 1 time in total.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

letsgobobby wrote:this thread like many others has expanded beyond the original scope. that's fine. I don't own EEs instead of stocks. I own them as part of my bond allocation, as an alternative to other bond choices. I would never recommend EE bonds as an alternative to stocks; and I would never recommend stocks as an alternative to EEs. they are simply different animals.
Yeah, but you own some bonds instead of 100% stocks for an investment horizon of at least 20 years, so you must believe that there's a reasonable probability that stocks will underperform bonds over a 20-year horizon. I'm with you on that, but obviously some posters disagree that that's a rational position. If someone thinks that there's only a 1% chance that EE bonds or any other bonds will outperform stocks over a 20-year horizon, then neither EE bonds more any other bonds make sense for that person.

That seemed to be the point of view the OP was coming from, and others obviously agree. You and I, and others that aren't in the stocks-are-not-risky-in-the-long-run camp, must have some reasons that we don't agree. My main reason is that I'm unwilling to bet my financial well being on what I believe to be flawed statistical analysis of historical data. I'm curious about what your reasoning is on this.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: I don't understand the case for EE bonds

Post by Kevin M »

willthrill81 wrote: There's little disputing that, despite its potentially questionable statistical properties, the second approach is more akin to the issue brought up by the OP. For instance, if we look at this as randomly 'dropping an investor' into any year in the available data and they invested a lump sum, the rolling period data seem quite reasonable.
The problem with this thinking is that we're not talking about randomly dropping into any year in the available data. We're talking about a 20-year period that starts now, and this future 20-year period shares no data with the historical data, unlike rolling 20-year periods that share 19 years of data with adjacent rolling periods.

So the questionable statistical properties of using rolling periods is absolutely critical to the issue at hand. And that's not the only issue. There's also the issue that there is lots of uncertainty about the true mean, mean reversion, and other parameters relevant to the future distributions of returns.

In a previous reply, I mentioned an academic paper that Nisiprius often cites to help us understand this issue. Here is a link to that paper: ARE STOCKS REALLY LESS VOLATILE IN THE LONG RUN?. From the abstract:
According to conventional wisdom, annualized volatility of stock returns is lower when computed
over long horizons than over short horizons, due to mean reversion induced by return predictability.
In contrast, we find that stocks are substantially more volatile over long horizons from an investor’s
perspective. This perspective recognizes that parameters are uncertain, even with two centuries of
data, and that observable predictors imperfectly deliver the conditional expected return. Mean reversion
contributes strongly to reducing long-horizon variance, but it is more than offset by various uncertainties
faced by the investor, especially uncertainty about the expected return.
The authors identify five components of long-run predictive variance.
We show that mean reversion is only one of five components of long-run
predictive variance:

(i) i.i.d. uncertainty
(ii) mean reversion
(iii) uncertainty about future expected returns
(iv) uncertainty about current expected return
(v) estimation risk.

Whereas the mean-reversion component is strongly negative, the other components are all positive,
and their combined effect outweighs that of mean reversion.

Of the four components contributing positively, the one making the largest contribution at long
horizons reflects uncertainty about future expected returns. This component (iii) is often neglected
in discussions of how return predictability affects long-horizon return variance
And here is a link to a discussion of this by Ken French and Gene Fama: Q&A: Are Stocks Safer in the Long Run? . Just a taste:
KRF: The Pastor-Stambaugh result is driven by uncertainty about the true expected return. The volatility or standard deviation of returns is usually defined as the expected variation relative to the true mean of the process generating returns - as if we knew the true expected return. But, as Pastor and Stambaugh emphasize, we never actually know the true mean. When they include uncertainty about the true mean (as well as uncertainty about other true parameters) in the analysis, they find that long-run returns are indeed more volatile than short-run returns.
Kevin
If I make a calculation error, #Cruncher probably will let me know.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

letsgobobby wrote:But at least you are consistent. You don't own TBM, you don't own long bonds, and I believe you said you don't own any fixed income outside of a cash emergency fund. I grant you this: EE bonds clearly make no sense for you.
Actually, I own a small portion of long TIPS through a Real Return fund. I also have a non-stock allocation which is larger than we need for just emergency funds, and instead is also serving to reduce short-term volatility and provide for rebalancing.

But sure, generally I put my money where my mouth is. However, I am open to being persuaded I should consider new ideas.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

Kevin M wrote: Well, the part I underlined is easy to prove wrong.

For rolling 20-year periods since 1929, the mean (expected value) of 20-year cumulative returns is 833% and the standard deviation of those returns is 470%. For the four independent 20-year periods since 1929, the mean is 599% and the SD is 591%.
I'm quite confident for the purpose of establishing the central expectation, the second methodology is erroneous because it depends on the arbitrary time at which we start collecting the necessary data.
If we assume a normal probability distribution, and if we assume that our analysis is statistically valid (which it probably is not, either due to using rolling periods or because four independent periods isn't enough to have much statistical validity), then the independent periods indicate roughly 1/6 probability that our 20-year cumulative return would be 8% or less (compared to a 100% 20-year return for EE bond).
To be honest, I didn't follow your math here. But I would again note that while it is true you have to be careful with overlapping periods, you also have to be cautious with the significance of when we happen to start collecting data. And I'm not actually aware of any 20-year period in which stocks returned less than 59.8% total, which is what you get if you buy in 1928 and sell in 1948:

http://pages.stern.nyu.edu/~adamodar/Ne ... retSP.html

So a 1/6th probability of only 8% nominal for the entire 20 year period seems really high to me--but maybe you can identify the period you are talking about.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: I don't understand the case for EE bonds

Post by NiceUnparticularMan »

willthrill81 wrote:As I said before, that's due to the fact that most 20 year rolling periods do not include the Great Depression. That one event has a huge impact on the data when viewed in their entirety.
I'm just going to note again that if you are worried about that scenario, you really want long bonds you can use for rebalancing.
Post Reply