Wade Pfau on bonds in a retirement portfolio

[Nords]

Wade Pfau's doing some interesting research on whether bonds have a place in an investment portfolio.

We know that bonds can be a reliable (and comforting) source of income, and that they reduce overall portfolio volatility with a smaller reduction in returns.  That's all good.

But his research is indicating that annuities do the same thing, with the added advantages of longevity insurance and portfolio-failure insurance.  Besides the "mortality credits" benefits of an annuity in a portfolio, insurance companies invest their annuity funds largely in... bonds.

Maybe this just suits my confirmation bias, because I've never invested our retirement portfolio in bonds.  Maybe we're in a unique period of history where interest rates are so low that the payouts from annuities and bonds are relatively close.  And maybe this math is not so good for periods longer than the "traditional" 30-year retirement, especially when a bond fund eventually follows prevailing interest rates (while an annuity might not).

But perhaps it's worth annuitizing a part of a portfolio that would have been dedicated to bonds, and that lets investors feel more comfortable about putting the rest of the portfolio in equities.

http://retirementresearcher.com/a-simple-allocation-example/
http://retirementresearcher.com/why-bond-funds-dont-belong-in-retirement-portfolios/
http://retirementresearcher.com/substituting-income-annuities-for-bond-funds-in-retirement-2/
http://retirementresearcher.com/evaluating-investments-vs-insurance-in-retirement/

[sol]

Sure. But social security is already an annuity that every US worker owns.  Ditto for federal pensions.

From that perspective, I'm not sure buying even more annuity is helpful.  In my case (and maybe yours), all expenses past age 62 are already annuitized, including the longevity insurance.  The portfolio only exists to get you to age 62, and over those relatively short investing horizons I think bonds and cash play an important role in reducing volatility on funds you intend to rapidly deplete to zero.

[MoonShadow]

Sure. But social security is already an annuity that every US worker owns.  Ditto for federal pensions.

Or for any pension, since that is basicly what they are.  And I agree with Sol (*cough*, am I sick?) that there really isn't a compelling reason to buy an annuity on the open market, because you already had to buy one with your FICA taxes.  Unless you are unusually wealthy, in which case you should have retired long ago, the SS annuity (and other benefits) should cover those issues until there is a means test added to SS.

As to the cause, yes bonds have been at historic lows, but that is not the main reason annuities  make a poor substitute for bonds.  We keep bonds not for the return, but for the diversification they provide.  Annuities are typically just mutual funds that insurance companies smooth out returns with over time, so they are usually invested in stocks & bonds anyway.  It's the insurance company that takes the risk of principal loss, but they have this stuff down to a science, and will not lose on your payout.  This means that they are reducing the returns you would otherwise see if you owned the same mutual fund, and keeping the difference as profit if their math works out.

[brooklynguy]

And maybe this math is not so good for periods longer than the "traditional" 30-year retirement, especially when a bond fund eventually follows prevailing interest rates (while an annuity might not).

This was my reaction as I've been reading Pfau's latest string of publications on the role of annuities as bond-substitutes.  I haven't looked into it much, but my sense is that private market lifetime annuities are prohibitively expensive for, say, a 35 year old, so Pfau's analyses don't translate well to the super-long retirement period context.

[sol]

It's the insurance company that takes the risk of principal loss, but they have this stuff down to a science, and will not lose on your payout.  This means that they are reducing the returns you would otherwise see if you owned the same mutual fund, and keeping the difference as profit if their math works out.

Right, but they win that game in aggregate, not on any individual policy.  The longevity insurance that Nords mentioned is the benefit of the annuity.  You give up the entire principal amount today so of you die tomorrow your heirs get nothing, but they will continue to pay you if you live to be 200.  They make money because their rates ensure that on average, people will die before getting their money back.  You buy because you hope to be the exception.

For someone who wanted to guarantee income now but leave nothing behind, annuities are a convenient way to trade away investment returns in exchange for stability of (lower) payments.  But I suspect there aren't many such people in this crowd.

The super wealthy already do this, in a way.  Billionaires like Zuckerberg will sometimes fund all of their living expenses with debt and pay the debt upon death, because the interest rate on the debt (secured against a vast fortune so no default risk) is lower than the taxes they avoid by using the estate tax loophole.  They've effectively found an analogous way to fund steady current income by trading away current equity, but they use banks rather than life insurance companies because the rates are better.

[forummm]

Sure. But social security is already an annuity that every US worker owns.  Ditto for federal pensions.

From that perspective, I'm not sure buying even more annuity is helpful.  In my case (and maybe yours), all expenses past age 62 are already annuitized, including the longevity insurance.  The portfolio only exists to get you to age 62, and over those relatively short investing horizons I think bonds and cash play an important role in reducing volatility on funds you intend to rapidly deplete to zero.

I think this is true for many of us here. But for bigger spenders, they may want to spend something beyond what SS provides. I personally don't plan to buy an annuity though. And may never hold bonds.

[brooklynguy]

For someone who wanted to guarantee income now but leave nothing behind, annuities are a convenient way to trade away investment returns in exchange for stability of (lower) payments.  But I suspect there aren't many such people in this crowd.

But one of the points Pfau (quite compellingly) makes in this latest line of research is that partial annuitization, counterintuitively, can actually leave you with a larger total legacy than an equivalent portfolio using bonds in lieu of annuities.

Sure. But social security is already an annuity that every US worker owns.  Ditto for federal pensions.

From that perspective, I'm not sure buying even more annuity is helpful.  In my case (and maybe yours), all expenses past age 62 are already annuitized, including the longevity insurance.  The portfolio only exists to get you to age 62, and over those relatively short investing horizons I think bonds and cash play an important role in reducing volatility on funds you intend to rapidly deplete to zero.

I think this is true for many of us here. But for bigger spenders, they may want to spend something beyond what SS provides. I personally don't plan to buy an annuity though. And may never hold bonds.

But the problem that does exist for many of us here is that the pre-SS period that sol calls "relatively short" is actually relatively long, such that the portfolio required to bridge it is not much (or any) smaller than the portfolio required to bridge an indefinite period.  I don't personally plan to buy an annuity or bonds either (though I do carry a large mortgage balance, which largely serves the function of a bond allocation as well).

[forummm]

(though I do carry a large mortgage balance, which largely serves the function of a bond allocation as well).

How so? Isn't that the opposite of a bond? Especially since the proceeds of the loan are invested in stocks?

[MoonShadow]

(though I do carry a large mortgage balance, which largely serves the function of a bond allocation as well).

How so? Isn't that the opposite of a bond? Especially since the proceeds of the loan are invested in stocks?

In theory, as you pay down your mortgage, the growth in your equity stake is mathmatically similar to buying bonds; they just happen to be your own bonds.  It is similar to a company buying their own issued bonds when they are cheaper on the bond market, thus retiring that debt.  Notice that this doesn't actually happen very often.  There is an argument to be made for paying your mortgage off before developing any significant bond holdings.  The analogy has it's limits.

[brooklynguy]

(though I do carry a large mortgage balance, which largely serves the function of a bond allocation as well).

How so? Isn't that the opposite of a bond? Especially since the proceeds of the loan are invested in stocks?

Carrying the mortgage allows you to invest in more stocks than you otherwise would be able to (using the money that would otherwise be trapped as home equity), but as long as you do have an outstanding mortgage, it operates similarly to holding a bond, where each required repayment gives you a "return" equal to the mortgage interest rate (but the analogy does have limits; you can't really rebalance to maintain a consistent "bond" allocation the way you could if you actually held bonds in your portfolio).

[DavidAnnArbor]

Maybe we're in a unique period of history where interest rates are so low that the payouts from annuities and bonds are relatively close. 

If I were very old, and I was worried about my portfolio failing, or had a surge in medical expenses, it might make sense at that point to get the annuity. For example, I would get the annuity if interest rates were high enough to provide a good yield on an annuity and I was 80 years old, and could foresee a need for hiring nursing assistants for doing activities of daily living, and the portfolio has a good chance of NOT making it for another 10-15 years. 

[Aphalite]

And may never hold bonds.

It's senseless to hold bonds right now, but conditions change and it's all about alternatives - if we start getting 10% interest rates on risk free treasuries (which would also probably mean runaway inflation, at least in the short term), and equities are expensive, I think bonds would be a very sensible holding

[forummm]

(though I do carry a large mortgage balance, which largely serves the function of a bond allocation as well).

How so? Isn't that the opposite of a bond? Especially since the proceeds of the loan are invested in stocks?

Carrying the mortgage allows you to invest in more stocks than you otherwise would be able to (using the money that would otherwise be trapped as home equity), but as long as you do have an outstanding mortgage, it operates similarly to holding a bond, where each required repayment gives you a "return" equal to the mortgage interest rate (but the analogy does have limits; you can't really rebalance to maintain a consistent "bond" allocation the way you could if you actually held bonds in your portfolio).

That doesn't really feel the same to me. It means you will be buying a pseudo bond in the future over time. And in order to sell that bond to live on, you have to 1) have your house maintain its value and 2) either sell your house or take out another loan against its value. Either of the options in #2 is or could be expensive. And you'd have to find another place to live if you sold it. I'm OK with having a mortgage and investing in stocks instead of paying it off. But I think that's a more risky way to go than the safety that holding actual bonds connotes.

And may never hold bonds.

It's senseless to hold bonds right now, but conditions change and it's all about alternatives - if we start getting 10% interest rates on risk free treasuries (which would also probably mean runaway inflation, at least in the short term), and equities are expensive, I think bonds would be a very sensible holding

If we have runaway inflation, investing in bonds is not very attractive. Hence the very high interest rates. They are only risk free if they are TIPS.

[brooklynguy]

That doesn't really feel the same to me. It means you will be buying a pseudo bond in the future over time. And in order to sell that bond to live on, you have to 1) have your house maintain its value and 2) either sell your house or take out another loan against its value. Either of the options in #2 is or could be expensive. And you'd have to find another place to live if you sold it. I'm OK with having a mortgage and investing in stocks instead of paying it off. But I think that's a more risky way to go than the safety that holding actual bonds connotes.

Yes, the relative illiquidity of home equity as compared to a bond holding is where the analogy starts to break down, as both I and MoonShadow mentioned.  But for someone who has a mortgage and will carry the mortgage to maturity, every principal repayment of the mortgage is (disregarding tax consequences and transaction costs) functionally equivalent to purchasing a bond that will be held to maturity with a term equal to the remaining life to maturity on the mortgage that pays an interest rate equal to the mortgage rate.

Leveraged-stock-investing-via-mortgage is definitely more risky in the sense that you have greater exposure to market underperformance (but, as always, the specific risks being referred to need to be clearly defined--leveraged-stock-investing-via-mortgage is less risky than, say, carrying no mortgage and holding 100% cash if the risk you're talking about is inflation risk, for example).  If you do carry a mortgage, there really is no rational reason to purchase a bond that pays less than the "return" you get on mortgage prepayments (after accounting for any tax consequences and transaction costs) (unless the bond's purpose is to operate as a self-restraint tool to prevent you from irrationally panic-selling during market downturns).

[YoungInvestor]

That doesn't really feel the same to me. It means you will be buying a pseudo bond in the future over time. And in order to sell that bond to live on, you have to 1) have your house maintain its value and 2) either sell your house or take out another loan against its value. Either of the options in #2 is or could be expensive. And you'd have to find another place to live if you sold it. I'm OK with having a mortgage and investing in stocks instead of paying it off. But I think that's a more risky way to go than the safety that holding actual bonds connotes.

Yes, the relative illiquidity of home equity as compared to a bond holding is where the analogy starts to break down, as both I and MoonShadow mentioned.  But for someone who has a mortgage and will carry the mortgage to maturity, every principal repayment of the mortgage is (disregarding tax consequences and transaction costs) functionally equivalent to purchasing a bond that will be held to maturity with a term equal to the remaining life to maturity on the mortgage that pays an interest rate equal to the mortgage rate.

Leveraged-stock-investing-via-mortgage is definitely more risky in the sense that you have greater exposure to market underperformance (but, as always, the specific risks being referred to need to be clearly defined--leveraged-stock-investing-via-mortgage is less risky than, say, carrying no mortgage and holding 100% cash if the risk you're talking about is inflation risk, for example).  If you do carry a mortgage, there really is no rational reason to purchase a bond that pays less than the "return" you get on mortgage prepayments (after accounting for any tax consequences and transaction costs) (unless the bond's purpose is to operate as a self-restraint tool to prevent you from irrationally panic-selling during market downturns).

Correct me if I'm wrong, but wouldn't paying your mortgage         be closer to covering a short position on bonds than actually buying them?

To me, then, an outstanding mortgage is a negative allocation to bonds (hence the inflation hedge).

[brooklynguy]

Correct me if I'm wrong, but wouldn't paying your mortgage         be closer to covering a short position on bonds than actually buying them?

To me, then, an outstanding mortgage is a negative allocation to bonds (hence the inflation hedge).

Yeah, a mortgage is a liability (not an asset) which can be thought of as a "negative bond," and each principal repayment of the mortgage (a reduction of the liability) therefore functions similarly to the purchase of bond (but, as discussed above, there are limits to the analogy).  I may have stretched the analogy too far to begin with by introducing the idea of a mortgage as serving bond-like functions in the context of this thread.

[arebelspy]

To clarify brooklynguy's posts, the way I've always heard it is not that the mortgage is like a bond portion of your portfolio, but that paying off the mortgage is like buying a bond (which he did say, but I'll try to illustrate).

For example, say you have a house worth $300k, with a 300k mortgage (no equity) at 4%, and a 1MM portfolio.  You decide you want your allocation to be 90/10 stocks/bonds.  By paying down your mortgage by 100k (and investing 900k in stocks), you've essentially bought a bond that returned 4% (your mortgage interest rate).

Thus the argument is that if you have a mortgage, you shouldn't hold ANY bonds until that mortgage is paid off (assuming the mortgage rate > bond rates), but should be 100% equities, and when you "rebalance" (or, say, are going to buy more bonds via contributions), you pay down the mortgage by that amount.  It's functionally equivalent to buying a 4% bond at that point.

I'm sure that's what he meant, just trying to put it in different words.  Not sure if I succeeded.  :)

[arebelspy]

As far as Pfau's annuities, he's been on that bandwagon for years and years--setting an income "floor."  I also suspect it doesn't work for extreme early retirees due to the actuarial rates, meaning you'd have to annuitize so much to see any significant income, or the amount of bonds you may be annuitizing (say 10-40%) just gives you a way too low of a floor to live on.

I just fundamentally don't like playing against the house.  The insurance company will come out ahead in the aggregate, so why do I think I'll be the one to beat them, and they'll lose money on my policy?  I don't think I can beat the average at stocks, at the casino, or with insurance.  If I think I'll be average (and I have no reason to suspect otherwise), then letting them take their cut out nets me less money over my life than me taking the risks.  As long as you have the stomach to hold on during market rides, I'd think the same would be true of everyone here.

[beltim]

I just fundamentally don't like playing against the house.  The insurance company will come out ahead in the aggregate, so why do I think I'll be the one to beat them, and they'll lose money on my policy?  I don't think I can beat the average at stocks, at the casino, or with insurance.  If I think I'll be average (and I have no reason to suspect otherwise), then letting them take their cut out nets me less money over my life than me taking the risks.  As long as you have the stomach to hold on during market rides, I'd think the same would be true of everyone here.

In my mind, annuitization isn't a bet against the house but instead compares the difference between the 50th and 90th percentile outcomes to the profit of the insurance company.  What do I mean?  Let's look at an example:
Most people on this forum aim for a 4% withdrawal rate because studies have suggested that this is safe for 90+% of historical outcomes.  However, the insurance companies don't need 90% safety - if they have a large enough customer base, then the 50th percentile outcome completely covers their liabilities.  Then the company charges fees (or equivalently reduces payouts) to give them their profit margin.  If the difference between the 90th (or whatever) percentile outcome and the 50th percentile outcome is larger than the insurance company's profit and expenses, then individuals are better off getting the annuity.

However, I agree with you that this probably doesn't work for extreme early retirees due to low rates for lifetime annuities.

[arebelspy]

I just fundamentally don't like playing against the house.  The insurance company will come out ahead in the aggregate, so why do I think I'll be the one to beat them, and they'll lose money on my policy?  I don't think I can beat the average at stocks, at the casino, or with insurance.  If I think I'll be average (and I have no reason to suspect otherwise), then letting them take their cut out nets me less money over my life than me taking the risks.  As long as you have the stomach to hold on during market rides, I'd think the same would be true of everyone here.

In my mind, annuitization isn't a bet against the house but instead compares the difference between the 50th and 90th percentile outcomes to the profit of the insurance company.  What do I mean?  Let's look at an example:
Most people on this forum aim for a 4% withdrawal rate because studies have suggested that this is safe for 90+% of historical outcomes.  However, the insurance companies don't need 90% safety - if they have a large enough customer base, then the 50th percentile outcome completely covers their liabilities.  Then the company charges fees (or equivalently reduces payouts) to give them their profit margin.  If the difference between the 90th (or whatever) percentile outcome and the 50th percentile outcome is larger than the insurance company's profit and expenses, then individuals are better off getting the annuity.

I still view it as a bet against the house, because the 50th% is where they break even, and they say, charge enough to put it at the 55th%, where the rest is their profit.  My outcome should be around 50%, so I'm out their fees.  On average, the average person is out the amount of fees they'd charge, and would come out ahead not having those fees charged (if, like I said, they can stomach volatility).

[beltim]

I just fundamentally don't like playing against the house.  The insurance company will come out ahead in the aggregate, so why do I think I'll be the one to beat them, and they'll lose money on my policy?  I don't think I can beat the average at stocks, at the casino, or with insurance.  If I think I'll be average (and I have no reason to suspect otherwise), then letting them take their cut out nets me less money over my life than me taking the risks.  As long as you have the stomach to hold on during market rides, I'd think the same would be true of everyone here.

In my mind, annuitization isn't a bet against the house but instead compares the difference between the 50th and 90th percentile outcomes to the profit of the insurance company.  What do I mean?  Let's look at an example:
Most people on this forum aim for a 4% withdrawal rate because studies have suggested that this is safe for 90+% of historical outcomes.  However, the insurance companies don't need 90% safety - if they have a large enough customer base, then the 50th percentile outcome completely covers their liabilities.  Then the company charges fees (or equivalently reduces payouts) to give them their profit margin.  If the difference between the 90th (or whatever) percentile outcome and the 50th percentile outcome is larger than the insurance company's profit and expenses, then individuals are better off getting the annuity.

I still view it as a bet against the house, because the 50th% is where they break even, and they say, charge enough to put it at the 55th%, where the rest is their profit.  My outcome should be around 50%, so I'm out their fees.  On average, the average person is out the amount of fees they'd charge, and would come out ahead not having those fees charged (if, like I said, they can stomach volatility).

So the guarantee has no value for you?

[arebelspy]

The "guarantee" has its own issues if the company BKs.  No, putting all my money with them I feel is riskier than keeping it in institutional markets, guarantees aside.

[beltim]

The "guarantee" has its own issues if the company BKs.  No, putting all my money with them I feel is riskier than keeping it in institutional markets, guarantees aside.

I suppose.  All the insurance payouts are backed by state agencies, though.

Anyway, an interesting discussion.  It's an offshoot of the discussion of why pensions are cheaper than everyone funding their own retirement.  And in both cases, it doesn't apply to extreme early retirees.

[Nords]

As far as Pfau's annuities, he's been on that bandwagon for years and years--setting an income "floor."  I also suspect it doesn't work for extreme early retirees due to the actuarial rates, meaning you'd have to annuitize so much to see any significant income, or the amount of bonds you may be annuitizing (say 10-40%) just gives you a way too low of a floor to live on.

The issue is more about timing, not the annuity rate.  The survival of the first 10 years of a retirement portfolio is highly dependent on the sequence of returns.  Jim Otar (and one or two other authors) advocate annuities as a "game over" tactic to save a failing portfolio.  One option that helps many ERs sleep at night is knowing that they have the floor spending taken care of (despite years of inflation).  Others could buy a SPIA as a coningency when it's appropriate.

It'd be interesting to see how Pfau would analyze this situation if rental real estate was in the portfolio.  Maybe that income stream is "good enough" to approximate the reliability of a SPIA.

I think the extreme ER situation is a little more complex due to its generally longer window of  entrepreneurial/employment opportunities.

I just fundamentally don't like playing against the house.  The insurance company will come out ahead in the aggregate, so why do I think I'll be the one to beat them, and they'll lose money on my policy?  I don't think I can beat the average at stocks, at the casino, or with insurance.  If I think I'll be average (and I have no reason to suspect otherwise), then letting them take their cut out nets me less money over my life than me taking the risks.  As long as you have the stomach to hold on during market rides, I'd think the same would be true of everyone here.

I hope that all the money I spend on insurance is wasted.  The whole point of the insurance company is for them to insure the last few percentage points of the SWR that will never be covered by the statistics of the historical returns.  When Bernstein says that any SWR success rate over 80% is meaningless, this is where the SPIA covers the remaining 20%.

You're also depending on the other longevity-insurance suckers to die sooner (mortality credits).  The insurance company is just taking its cut of the deal.

One thing I've learned from my readers is that very few people have the stomach to hold on during market rides.

[Telecaster]

The "guarantee" has its own issues if the company BKs.  No, putting all my money with them I feel is riskier than keeping it in institutional markets, guarantees aside.

I suppose.  All the insurance payouts are backed by state agencies, though.

Some of the insurance payouts are backed by the state agencies.  And you're paying for the re-insurance anyway.  The fees the state charges the insurance companies to maintain the reserve are passed straight onto you.

I'm with arebelspy on this one.  The insurance company has teams of PhD actuaries who think you are on the wrong side of the bet.  Hey, I'll defer to people smarter than I am.  If those guys think I'm a sucker if I buy insurance, then I'm probably a sucker. *

Wade Pfau has gotten a lot of press in the last few years, but I take what he says with pretty large boulder of salt.  I don't know about this specific case, but every time I've looked into his stuff in the fine print he assumes a 1-1.5% management fee.   Fees that large can really screw up your SWR, and no one here pays anything like that.  If you were paying fees that large, what he is saying might make sense, but back out those fees and then not so much. 



*Just because I know this is coming...I happily buy insurance for risks I don't want to take.  I buy homeowners and sleep like a baby.  Glad to do it.  In this case, the deal is to buy an annuity because (hypothetical), (unrealistic fees), (hand waving).  Hand waving is a risk I can afford to take and I can come up with hypotheticals in the other direction. 

[arebelspy]

I hope that all the money I spend on insurance is wasted.

Of course, but only on insurance I'm taking due to not being able to handle the outcome without. With optional insurance it's a false dilemma--getting to use the insurance versus having all the money spent on it being wasted--I'd rather just not waste the money in the first place.

The whole point of the insurance company is for them to insure the last few percentage points of the SWR that will never be covered by the statistics of the historical returns.

Absolutely, but if things go to * so much that we're outside historical returns, I don't have much confidence in the insurance company's guarantees, and I'd probably rather have the liquidity.

To each his own; if it helps someone sleep at night, I have no quibbles about it.  It's just not something I'm a fan of personally.  :)

[Another Reader]

Absolutely, but if things go to * so much that we're outside historical returns, I don't have much confidence in the insurance company's guarantees, and I'd probably rather have the liquidity.

+1.  No one remembers how close a couple of these pillars of insurance were to going under in the Fall of 2008.  Diversification, including the best options of multiple types of investments, is probably the best "insurance."

[MrGreen]

The "guarantee" has its own issues if the company BKs.  No, putting all my money with them I feel is riskier than keeping it in institutional markets, guarantees aside.

There was serious speculation about AIG going belly up in 2008. Man, the mess that would have been.

[YoungInvestor]

The "guarantee" has its own issues if the company BKs.  No, putting all my money with them I feel is riskier than keeping it in institutional markets, guarantees aside.

There was serious speculation about AIG going belly up in 2008. Man, the mess that would have been.

At this point, I think bailouts are pretty much guaranteed. In this situation, being a shareholder sucks, working there probably sucks as well, but I think payouts are pretty close to government bonds as far as likeliness to get paid gets.

And in a situation where this would happen, the 4% rule would likely have been out of the window by a landslide.