20 years out from retirement, currently hold 10% EDV (Vanguard Extended Duration Treasury ETF)

[stimulacra]

Does anyone here still hold EDV (Vanguard Extended Duration Treasury ETF)? The topic of long term treasuries has been scarce here since 2021.

I currently hold 10% EDV and 5% VIPSX (Vanguard TIPS) in my pre-tax IRA. Just did a rebalance for the year into both when I rolled over my 401k last month.

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

Mostly curious because the majority of the unrealized losses in my portfolio are in my bond fund holdings. I anticipate those eventually going away in 10-15 years but was curious as to the actual mechanics of shifting my allocation over to BND or VBLTX (Vanguard Total Bond Market). Would I just transitioning over 1% allocation per year starting 2035 or just direct new inflows into BND?

For the rest of my portfolio, it's 80% equities and 5% Gold.

[market timer]

Yes, EDV is my top holding. Long term bonds fit in well from a liability matching perspective. If your investment horizon starts to decline, of course it makes sense to migrate to lower duration bonds. As a rule, I think it is best to buy bonds with a duration near when you plan to spend the money.

If you have unrealized losses, consider tax loss harvesting into another long term bond ETF.

[Tyler Aspect]

It seemed to me that you just need to wait for the next economic recession when the Fed lowers short term interest rate near zero. I imagine the long term yield should also be near minimum, and that should recover your net asset value and to sell it without much loss.

Another way is to sell EDV right away, and buy actual 20 year US Treasury notes. Its maturity will shorten every year.

[Catnap]

I hold EDV as my primary bond holding and I am about 15-25 years from retirement. I plan to direct new inflows into a short duration treasury fund very soon. The average duration of my bond portfolio will shorten up as I approach retirement. I will liquidate either EDV or the short bond fund as needed to maintain a duration between 5 to 10 years once I retire. This is roughly the same duration as BND but I prefer treasuries only (no corporates or mbs).

The tax loss harvesting mentioned by market timer makes sense. Directing new inflows into a shorter duration fund will shorten your duration without locking in losses that will probably break even at some point.

[muffins14]

Does anyone here still hold EDV (Vanguard Extended Duration Treasury ETF)? The topic of long term treasuries has been scarce here since 2021.

I currently hold 10% EDV and 5% VIPSX (Vanguard TIPS) in my pre-tax IRA. Just did a rebalance for the year into both when I rolled over my 401k last month.

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

Mostly curious because the majority of the unrealized losses in my portfolio are in my bond fund holdings. I anticipate those eventually going away in 10-15 years but was curious as to the actual mechanics of shifting my allocation over to BND or VBLTX (Vanguard Total Bond Market). Would I just transitioning over 1% allocation per year starting 2035 or just direct new inflows into BND?

For the rest of my portfolio, it's 80% equities and 5% Gold.

20% of my portfolio is in GOVZ.

My goal is to match the bond duration to my investing horizon to what degree I can.

I am 38 and planning to 100. Perhaps retiring in 10 years. So planning for 52 years of withdrawals between 10 and 62 years from now, with a midpoint around 26 years. GOVZ is close enough.

when I do retire and have no income, I will have a TIPS ladder for maybe 15% of my portfolio and will have 15% nominals, again matching my duration. So that will mean long-term bonds until I am around 65, at which point I will start reducing duration annually. If I’m a live at 75 perhaps I’d be mainly intermediate duration.

[WinstonTeracina]

Does anyone here still hold EDV (Vanguard Extended Duration Treasury ETF)? The topic of long term treasuries has been scarce here since 2021.

I currently hold 10% EDV and 5% VIPSX (Vanguard TIPS) in my pre-tax IRA. Just did a rebalance for the year into both when I rolled over my 401k last month.

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

Mostly curious because the majority of the unrealized losses in my portfolio are in my bond fund holdings. I anticipate those eventually going away in 10-15 years but was curious as to the actual mechanics of shifting my allocation over to BND or VBLTX (Vanguard Total Bond Market). Would I just transitioning over 1% allocation per year starting 2035 or just direct new inflows into BND?

For the rest of my portfolio, it's 80% equities and 5% Gold.

20% of my portfolio is in GOVZ.

My goal is to match the bond duration to my investing horizon to what degree I can.

I am 38 and planning to 100. Perhaps retiring in 10 years. So planning for 52 years of withdrawals between 10 and 62 years from now, with a midpoint around 26 years. GOVZ is close enough.

when I do retire and have no income, I will have a TIPS ladder for maybe 15% of my portfolio and will have 15% nominals, again matching my duration. So that will mean long-term bonds until I am around 65, at which point I will start reducing duration annually. If I’m a live at 75 perhaps I’d be mainly intermediate duration.

This +1000

Don't be discurouaged by the unrealized losses of your bond fund or future rises in interest rates. It means that the cost of funding your future liabilities has fallen as well (likely by a much greater amount that your present losses). In fact, if one has plans to liability match in the future and they are still in the accumulation phase, they would LOVE for interest rates to continue to rise (and the PV value of their bond holdings to fall) until the day they retire.

I'd drop your TIPS fund altogether and replace it with more EDV or LTPZ (PIMCO 15+ Year US TIPS Index) which has a duration of 15 years. That's still likely too short for you and the expense ratio is a little higher than I'd like, but it's the only reasonably low-cost long-term TIPS fund out there. 5% Gold is unlikely to make much a difference in your portfolio either, so I'd consider replacing that with either more equities or bonds.

[grabiner]

Yes, EDV is my top holding. Long term bonds fit in well from a liability matching perspective. If your investment horizon starts to decline, of course it makes sense to migrate to lower duration bonds. As a rule, I think it is best to buy bonds with a duration near when you plan to spend the money.

Do you have fixed-dollar liabilities (such as a 30-year mortgage)? If not, then long-term TIPS are better for liability matching, Nominal bonds may cover much more or less than the liability you are trying to match, depending on whether inflation is lower or higher than current expectations over a long time period.

LTPZ is the only long-term TIPS ETF, and it is somewhat expensive, but you can also buy individual long-term TIPS and plan to hold them to maturity.

[Northern Flicker]

Because term risk and inflation risk are correlated, shortening duration of nominal assets also increases their ability to match a future real liability relative to long-term nominal assets. It won't be as close a match to CPI-U as a long TIPS fund would be, but CPI-U is not guaranteed to be a very close match either. A homeowner likely is much more exposed to rising health care expenses than rising housing costs, but the future theoretically could see health care costs rise sharply during a housing bust as a hypothetical example.

[er999]

I’m going to disagree that EDV is a good long term holding. Down by half compared to 2020 (8/21/2020 $169, 9/5/24 $81) compared to BND (8/21/2020 $88 vs $74.96 9/5/24). There was some dividend payments so losses not quite a big as they would be based only on price (unfortunately portfoliovisualizer is gone so can’t get the precise numbers)

Might be worth have a speculative position in EDV now but I don’t think you can buy and hold for decades. If rates drop again you would be rewarded but then have to decide the exit timing. Market timing of the bond market that isn’t recommended on here. Supposedly some life insurance companies leveraged up 8x long term bonds when rates were low according to a podcast I listened to with resulting losses so even the so called professionals make the wrong call.

I would call EDV a high risk asset. Whether a portfolio with some EDV is better than 100% stocks over 20-30 years is questionable. You might have the negative correlation in stock crashes like in 2008 and be glad to have them, or might not like in 2020.

For someone without of lot of extra money (like a 60 year old with 1 million doing a 4% withdrawal rate) hard to recommend EDV over a more conventional bond holding.

Even for the original poster 20 years out, is holding some EDV better than 100% stocks? Vanguard 2045 target fund has 16% bonds, would 16% EDV / 84% stocks instead of 100% stocks have a better return? Will it really dampen volatility?

I don’t know the answer but wanted to at least give a counterpoint to the pro EDV posts.

[muffins14]

I would call EDV a high risk asset. Whether a portfolio with some EDV is better than 100% stocks over 20-30 years is questionable. You might have the negative correlation in stock crashes like in 2008 and be glad to have them, or might not like in 2020.

1) if you have a need that is 24 years into the future, how is a 24-year treasury a high risk?

2) in 2020 long-term treasuries were up massively, like 25%

Yes they went down in 2022 because rates rose. If you needed money in 2022 you should have had it in long-term treasuries.

[grabiner]

I would call EDV a high risk asset. Whether a portfolio with some EDV is better than 100% stocks over 20-30 years is questionable. You might have the negative correlation in stock crashes like in 2008 and be glad to have them, or might not like in 2020.

1) if you have a need that is 24 years into the future, how is a 24-year treasury a high risk?[/qupte]

Unless your need is a fixed-dollar amount, you are taking a lot of inflation risk. A long-term TIPS, or an I-Bond, eliminates that risk.

Long-term Treasuries are especially attractive to pension funds and insurance companies, which do have long-term fixed-dollar liabilities; therefore, they bid up the prices, which makes them less attractive to investors who don't have such liabilities. (Some individuals do; if you have a 30-year fixed-rate mortgage, you can guarantee to cover the mortgage payments by buying a 30-year bond ladder.)

[Northern Flicker]

I would call EDV a high risk asset. Whether a portfolio with some EDV is better than 100% stocks over 20-30 years is questionable. You might have the negative correlation in stock crashes like in 2008 and be glad to have them, or might not like in 2020.

1) if you have a need that is 24 years into the future, how is a 24-year treasury a high risk?

It would be moderately high risk if that need is a real liability.

1.01^24 is about 1.27. So if inflation averages 1 percentage point higher than the yield of a 24-year zero coupon STRIP over 24 years, there will be about a 21% (27/1.27) shortfall.

[muffins14]

I would call EDV a high risk asset. Whether a portfolio with some EDV is better than 100% stocks over 20-30 years is questionable. You might have the negative correlation in stock crashes like in 2008 and be glad to have them, or might not like in 2020.

1) if you have a need that is 24 years into the future, how is a 24-year treasury a high risk?

It would be moderately high risk if that need is a real liability.

1.01^24 is about 1.27. So if inflation averages 1 percentage point higher than the yield of a 24-year zero coupon STRIP over 24 years, there will be about a 21% (27/1.27) shortfall.

Right, I should have specified that this is nominal dollars

[er999]

2) in 2020 long-term treasuries were up massively, like 25%

Yes they went down in 2022 because rates rose. If you needed money in 2022 you should have had it in long-term treasuries.

2018 EDV was 105 11/2018 so it did go up massively to 175 (7/31/2020) but down now to 81 range and I don’t think the dividends made up the difference.

30 year rates 11/2018 were 3.39% and are 4.06% now, down somewhat from the peak 5.04% 10/2023 but looking very good for the past 10 years, not so much for the past 30 years (can find the data on the FRED website from the federal reserve:
https://fred.stlouisfed.org/series/dgs30

Bill Bernstein had written the long term bond is a bet on interest rates before, and seems true to me. If rates cut again could make large profits with equity like shifts of 25-50% up in a year (and vice versa for rates rising).

I still don’t think long term nominal bonds are a great buy and hold for decades but could be great for market timers, even those who do it once every 3-5 years. EDV could 1.5x your money in 1-2 years but I’m not sure how great it is for buy and hold x decades and rebalance periodically.

[muffins14]

2) in 2020 long-term treasuries were up massively, like 25%

Yes they went down in 2022 because rates rose. If you needed money in 2022 you should have had it in long-term treasuries.

2018 EDV was 105 11/2018 so it did go up massively to 175 (7/31/2020) but down now to 81 range and I don’t think the dividends made up the difference.

30 year rates 11/2018 were 3.39% and are 4.06% now, down somewhat from the peak 5.04% 10/2023 but looking very good for the past 10 years, not so much for the past 30 years (can find the data on the FRED website from the federal reserve:
https://fred.stlouisfed.org/series/dgs30

Bill Bernstein had written the long term bond is a bet on interest rates before, and seems true to me. If rates cut again could make large profits with equity like shifts of 25-50% up in a year (and vice versa for rates rising).

I still don’t think long term nominal bonds are a great buy and hold for decades but could be great for market timers, even those who do it once every 3-5 years. EDV could 1.5x your money in 1-2 years but I’m not sure how great it is for buy and hold x decades and rebalance periodically.

The data seem to support that in portfolios like 90/10 or 80/20, having long-term treasuries is more efficient by return / variance than having shorter bonds or corporate bonds.

I would not advocate market timing in bonds.

My PoV is that if you need money in 15 years, a 15 year bond has less risk to that outcome than a 30 year bond you plan to sell early, or a rolling ladder of 1 year bonds. You can approximate this desired goal with a bond fund(s) that you adjust duration on.

[WinstonTeracina]

Unless your need is a fixed-dollar amount, you are taking a lot of inflation risk. A long-term TIPS, or an I-Bond, eliminates that risk.

That's only true if you get your duration right.



Look at the comparison between LTPZ (Long-Term TIPS) and VGLT (Vanguard's Long-Term Treasury Fund which has a similar duration) since 2021, when inflation started to rise.

[muffins14]

Unless your need is a fixed-dollar amount, you are taking a lot of inflation risk. A long-term TIPS, or an I-Bond, eliminates that risk.

That's only true if you get your duration right.



Look at the comparison between LTPZ (Long-Term TIPS) and VGLT (Vanguard's Long-Term Treasury Fund which has a similar duration) since 2021, when inflation started to rise.

The image doesn't load for me. Are you saying that the long-term TIPS did worse than the long-term nominal, due to the duration being "wrong" for one of them?

I'm sure they both went down in 2021->2023 because rates were rising, and I assume TIPS went down less than nominals, since inflation rose and TIPS would have adjusted accordingly.

Getting your duration right should be the base case. If you needed money in 2021-2023, you should not have been holding a long-term bond in 2020. You should have been holding a short-term bond or bond fund with duration 1-3 years

https://www.portfoliovisualizer.com/bac ... fpbmdLKrVX (confirming my statements above)

LTPZ has duration 19 years while VGLT has duration 15 years. So LTPZ was affected about 25% worse by rising rates than VGLT, somewhat offsetting the gains from the adjustment due to inflation

[WinstonTeracina]

Unless your need is a fixed-dollar amount, you are taking a lot of inflation risk. A long-term TIPS, or an I-Bond, eliminates that risk.

That's only true if you get your duration right.



Look at the comparison between LTPZ (Long-Term TIPS) and VGLT (Vanguard's Long-Term Treasury Fund which has a similar duration) since 2021, when inflation started to rise.

The image doesn't load for me. Are you saying that the long-term TIPS did worse than the long-term nominal, due to the duration being "wrong" for one of them?

I'm sure they both went down in 2021->2023 because rates were rising, and I assume TIPS went down less than nominals, since inflation rose and TIPS would have adjusted accordingly.

Getting your duration right should be the base case. If you needed money in 2021-2023, you should not have been holding a long-term bond in 2020. You should have been holding a short-term bond or bond fund with duration 1-3 years

https://www.portfoliovisualizer.com/bac ... fpbmdLKrVX (confirming my statements above)

LTPZ has duration 19 years while VGLT has duration 15 years. So LTPZ was affected about 25% worse by rising rates than VGLT, somewhat offsetting the gains from the adjustment due to inflation

Apologies muffins14, the link should be fixed now.

I was trying to illustrate the fact that long-term bonds (both nominal and inflation-adjusted) got hammered in 2022 at a time when inflation rose rapidly (max drawdown of 40% vs 45%). Getting duration right is indeed the most critical factor when selecting bond funds, certainly moreso than deciding on nominal vs. not.

[Northern Flicker]

Unless your need is a fixed-dollar amount, you are taking a lot of inflation risk. A long-term TIPS, or an I-Bond, eliminates that risk.

That's only true if you get your duration right.



Look at the comparison between LTPZ (Long-Term TIPS) and VGLT (Vanguard's Long-Term Treasury Fund which has a similar duration) since 2021, when inflation started to rise.

The image doesn't load for me. Are you saying that the long-term TIPS did worse than the long-term nominal, due to the duration being "wrong" for one of them?

I'm sure they both went down in 2021->2023 because rates were rising, and I assume TIPS went down less than nominals, since inflation rose and TIPS would have adjusted accordingly.

Getting your duration right should be the base case. If you needed money in 2021-2023, you should not have been holding a long-term bond in 2020. You should have been holding a short-term bond or bond fund with duration 1-3 years

https://www.portfoliovisualizer.com/bac ... fpbmdLKrVX (confirming my statements above)

LTPZ has duration 19 years while VGLT has duration 15 years. So LTPZ was affected about 25% worse by rising rates than VGLT, somewhat offsetting the gains from the adjustment due to inflation

Apologies muffins14, the link should be fixed now.

I was trying to illustrate the fact that long-term bonds (both nominal and inflation-adjusted) got hammered in 2022 at a time when inflation rose rapidly (max drawdown of 40% vs 45%). Getting duration right is indeed the most critical factor when selecting bond funds, certainly moreso than deciding on nominal vs. not.

That is not the correct conclusion, which is that inflation risk and term risk generally trade off, rather than inflation risk can be ignored. If you are taking the inflation risk, you need to shorten duration. The backtest bears that out.

If 19 years is the correct duration of a real liability, then the TIPS portfolio matched the duration. Using a nominal portfolio instead, one should reduce the duration to compensate for taking the inflation risk. Reducing to 15 years was not enough. The final values of the portfolios were $6978 and $7532, and 6978/7532 = 0.92. So reducing the duration from 19 years to 15 years addressed some of the inflation risk, but still left an 8% shortfall in covering the liabilities. The duration needed to be reduced further to use nominal assets to cover real liabilities with a 19-year duration.

For the particular time period, 85% VGLT and 15% cash matched the real liability:

https://www.portfoliovisualizer.com/bac ... h9DGc6jE13

So 0.85 * 15 = 12.75 years is an approximate nominal duration that matched the real liability with a 19 year duration during this time period. Due to the unpredictability of the tradeoff and of inflation in general, I think an even shorter duration is justifiable if nominal assets are used.

As an aside, the durations of both ETFs shortened due to rising rates. I think 21 and 17 were closer to the durations at the start of the backtest. This makes the calculation imprecise, but the change was not dramatic enough to change the conclusion.

[FoundingFather]

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

I think that EDV is a fine fund if you understand it, but I fear that many people don't. I also agree with much of what has been said about it.

I plan to retire in about 30 years, so I bought an individual TIPS bond, not fund, that matures in about 30 years. At the time, it was yielding 2.3% real return. I like that with individual bonds the duration gradually decreases at the same rate that my spending horizon approaches, thus protecting me from interest rate changes causing a shock in the value of the "safe" part of my portfolio. I also like that individual bonds have no ongoing fees associated.

Putting new money, gradually, into a shorter duration bond fund would probably be fine, but buying individual bonds just seems more straightforward to me. Since most of my expenses in 30 years are not nominal, I favor TIPS for those individual bonds.

Founding Father

[Circle the Wagons]

Unless your need is a fixed-dollar amount, you are taking a lot of inflation risk. A long-term TIPS, or an I-Bond, eliminates that risk.

That's only true if you get your duration right.



Look at the comparison between LTPZ (Long-Term TIPS) and VGLT (Vanguard's Long-Term Treasury Fund which has a similar duration) since 2021, when inflation started to rise.

The image doesn't load for me. Are you saying that the long-term TIPS did worse than the long-term nominal, due to the duration being "wrong" for one of them?

I'm sure they both went down in 2021->2023 because rates were rising, and I assume TIPS went down less than nominals, since inflation rose and TIPS would have adjusted accordingly.

Getting your duration right should be the base case. If you needed money in 2021-2023, you should not have been holding a long-term bond in 2020. You should have been holding a short-term bond or bond fund with duration 1-3 years

https://www.portfoliovisualizer.com/bac ... fpbmdLKrVX (confirming my statements above)

LTPZ has duration 19 years while VGLT has duration 15 years. So LTPZ was affected about 25% worse by rising rates than VGLT, somewhat offsetting the gains from the adjustment due to inflation

Apologies muffins14, the link should be fixed now.

I was trying to illustrate the fact that long-term bonds (both nominal and inflation-adjusted) got hammered in 2022 at a time when inflation rose rapidly (max drawdown of 40% vs 45%). Getting duration right is indeed the most critical factor when selecting bond funds, certainly moreso than deciding on nominal vs. not.

That is not the correct conclusion, which is that inflation risk and term risk generally trade off, rather than inflation risk can be ignored. If you are taking the inflation risk, you need to shorten duration. The backtest bears that out.

If 19 years is the correct duration of a real liability, then the TIPS portfolio matched the duration. Using a nominal portfolio instead, one should reduce the duration to compensate for taking the inflation risk. Reducing to 15 years was not enough. The final values of the portfolios were $6978 and $7532, and 6978/7532 = 0.92. So reducing the duration from 19 years to 15 years addressed some of the inflation risk, but still left an 8% shortfall in covering the liabilities. The duration needed to be reduced further to use nominal assets to cover real liabilities with a 19-year duration.

For the particular time period, 85% VGLT and 15% cash matched the real liability:

https://www.portfoliovisualizer.com/bac ... h9DGc6jE13

So 0.85 * 15 = 12.75 years is an approximate nominal duration that matched the real liability with a 19 year duration during this time period. Due to the unpredictability of the tradeoff and of inflation in general, I think an even shorter duration is justifiable if nominal assets are used.

As an aside, the durations of both ETFs shortened due to rising rates. I think 21 and 17 were closer to the durations at the start of the backtest. This makes the calculation imprecise, but the change was not dramatic enough to change the conclusion.

You've traded inflation risk for interest rate risk. Going nominal but shortening duration is not equivalent to duration matching with TIPS.

[Northern Flicker]

I did not say it was equivalent-- quite the opposite. My point was that if you are going to use nominal assets to match a real liability, the duration should be shortened.

[Circle the Wagons]

I did not say it was equivalent-- quite the opposite. My point was that if you are going to use nominal assets to match a real liability, the duration should be shortened.

I don't think that's good advice.

If you're going to skip the (arguably free) inflation protection, at least match duration.

[GAAP]

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

As a general estimate, if you're going to duration-match your fixed income, that duration should be the number of years until you start needing the money, plus half of the period that you will be using the income.

If you're retiring in 20 years and expect to have a 30-year retirement period, the result would be 20 + 0.5 * 30 = 20 + 15 = 35. EDV has a duration of about 25 -- so you would be able to actually duration match about 10 years prior to retirement.

Choice of funds to do that is a wholly separate topic, as is the amount of fixed income you should have 20 years prior to retirement.

For a more in-depth discussion of the topic, see viewtopic.php?t=318412.

[Tyler Aspect]

My view is that the United States could be entering an extended phase of higher than normal inflations, given the level of federal debt there would be incentives to erase it through inflation. However, given a number of program expenses are tied to officially measured inflation rates, there could be incentives to lowball the measured inflation. Since stock is the primary fighter of inflation I have incorporated a +10% over-weight to stocks, to my normal 50% stock / 50% bond weighting.

Since I am guarding the lowballing of official measured inflation, I am sticking to nominal intermediate term bonds.

As to "liability matching" I have never needed to do that. All of my retirement withdrawals have been from stocks, never from bonds.

[Northern Flicker]

I did not say it was equivalent-- quite the opposite. My point was that if you are going to use nominal assets to match a real liability, the duration should be shortened.

I don't think that's good advice.

If you're going to skip the (arguably free) inflation protection, at least match duration.

Then you would be taking even more inflation risk. That is, you risk an even greater mismatch with the real liabilities due to inflation.

If already holding long duration nominal assets with losses, staying the course is probably best.

[Circle the Wagons]

I did not say it was equivalent-- quite the opposite. My point was that if you are going to use nominal assets to match a real liability, the duration should be shortened.

I don't think that's good advice.

If you're going to skip the (arguably free) inflation protection, at least match duration.

Then you would be taking even more inflation risk. That is, you risk an even greater mismatch with the real liabilities due to inflation.

If already holding long duration nominal assets with losses, staying the course is probably best.

I understand your perspective. Especially if inflation risk compounds over time.

But in your example, you're introducing reinvestment risk, and you still face 13 years of inflation risk.

Likely we agree that duration matched TIPS allow the investor to avoid both of these uncompensated risks.

[typical.investor]

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

As a general estimate, if you're going to duration-match your fixed income, that duration should be the number of years until you start needing the money, plus half of the period that you will be using the income.

If you're retiring in 20 years and expect to have a 30-year retirement period, the result would be 20 + 0.5 * 30 = 20 + 15 = 35. EDV has a duration of about 25 -- so you would be able to actually duration match about 10 years prior to retirement.

Choice of funds to do that is a wholly separate topic, as is the amount of fixed income you should have 20 years prior to retirement.

For a more in-depth discussion of the topic, see viewtopic.php?t=318412.

I really don't think much of that duration-matching scheme.

As you say, 10 years to retirement you are in a fund with a 25 year duration. Sure, your spending horizon matches the duration of your portfolio which is 35, but so what.

You need to spend in 10 years.

What you should do 25 years prior to retirement (or whatever the duration of your fund is) is switch to an individual bond that matures. Just remember the higher the coupon means the shorter the duration so if you aren't using treasury zeros, you should take a second to calculate the duration on the individual bond. Simply choosing a 25 year bond could subject you to NAV loss because it's duration may be years less than 25 (again depending on the coupon).

Simply using a long and short fund to try and duration-match will subject you to rate (NAV loss) risk.

[Northern Flicker]

I did not say it was equivalent-- quite the opposite. My point was that if you are going to use nominal assets to match a real liability, the duration should be shortened.

I don't think that's good advice.

If you're going to skip the (arguably free) inflation protection, at least match duration.

Then you would be taking even more inflation risk. That is, you risk an even greater mismatch with the real liabilities due to inflation.

If already holding long duration nominal assets with losses, staying the course is probably best.

I understand your perspective. Especially if inflation risk compounds over time.

But in your example, you're introducing reinvestment risk, and you still face 13 years of inflation risk.

Interest rates generally fall when inflation comes in low. TIPS will underperform duration-matched nominals in that scenario vs reinvestment risk causing lower duration assets to underperform the duration-matched nominal assets in the same scenario, which will be closer to the duration-matched TIPS in the falling rate scenario. The tradeoff of term and inflation risk generally occurs in both directions (approximating each other, not a precise exchange).

Where I would be more leery of the strategy is simulating intermediate duration TIPS with short duration nominals because the yield curve tends to be much steeper in that region (despite being inverted there today). A steep yield curve may not provide the shorter nominals with enough interest income.

[GAAP]

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

As a general estimate, if you're going to duration-match your fixed income, that duration should be the number of years until you start needing the money, plus half of the period that you will be using the income.

If you're retiring in 20 years and expect to have a 30-year retirement period, the result would be 20 + 0.5 * 30 = 20 + 15 = 35. EDV has a duration of about 25 -- so you would be able to actually duration match about 10 years prior to retirement.

Choice of funds to do that is a wholly separate topic, as is the amount of fixed income you should have 20 years prior to retirement.

For a more in-depth discussion of the topic, see viewtopic.php?t=318412.

I really don't think much of that duration-matching scheme.

As you say, 10 years to retirement you are in a fund with a 25 year duration. Sure, your spending horizon matches the duration of your portfolio which is 35, but so what.

You need to spend in 10 years.

What you should do 25 years prior to retirement (or whatever the duration of your fund is) is switch to an individual bond that matures. Just remember the higher the coupon means the shorter the duration so if you aren't using treasury zeros, you should take a second to calculate the duration on the individual bond. Simply choosing a 25 year bond could subject you to NAV loss because it's duration may be years less than 25 (again depending on the coupon).

Simply using a long and short fund to try and duration-match will subject you to rate (NAV loss) risk.

A single bond won't match the duration either, and would still expose you to interest rate risk.

You would really need a ladder of bonds to do this. One of those bonds would mature in 10 years, one in eleven years, etc.

Regardless of whether individual bonds or bond funds are used, the duration will decrease by one until the income need starts, and then by one-half thereafter.

If you really wanted to be careful, you could duration-match to expected longevity -- which would decline at yet another rate, depending upon age.

[typical.investor]

My plan is to retire in ~2045, at which point do I start to transition away from EDV into shorter duration funds?

As a general estimate, if you're going to duration-match your fixed income, that duration should be the number of years until you start needing the money, plus half of the period that you will be using the income.

If you're retiring in 20 years and expect to have a 30-year retirement period, the result would be 20 + 0.5 * 30 = 20 + 15 = 35. EDV has a duration of about 25 -- so you would be able to actually duration match about 10 years prior to retirement.

Choice of funds to do that is a wholly separate topic, as is the amount of fixed income you should have 20 years prior to retirement.

For a more in-depth discussion of the topic, see viewtopic.php?t=318412.

I really don't think much of that duration-matching scheme.

As you say, 10 years to retirement you are in a fund with a 25 year duration. Sure, your spending horizon matches the duration of your portfolio which is 35, but so what.

You need to spend in 10 years.

What you should do 25 years prior to retirement (or whatever the duration of your fund is) is switch to an individual bond that matures. Just remember the higher the coupon means the shorter the duration so if you aren't using treasury zeros, you should take a second to calculate the duration on the individual bond. Simply choosing a 25 year bond could subject you to NAV loss because it's duration may be years less than 25 (again depending on the coupon).

Simply using a long and short fund to try and duration-match will subject you to rate (NAV loss) risk.

A single bond won't match the duration either, and would still expose you to interest rate risk.

You would really need a ladder of bonds to do this. One of those bonds would mature in 10 years, one in eleven years, etc.

Regardless of whether individual bonds or bond funds are used, the duration will decrease by one until the income need starts, and then by one-half thereafter.

If you really wanted to be careful, you could duration-match to expected longevity -- which would decline at yet another rate, depending upon age.

Yeah, I completely agree. I mean to sell part of the fund 25 years (or whatever the duration of your fund) in advance and purchase an individual bond for one year of spending. The next year do the same. And so on and so forth.

[Northern Flicker]

One other point is that laddering, whether implemented as a bond index fund or DIY ladder, diversifies reinvestment risk. It more or less is fully eliminated when duration matches liability duration, but it is not an all or nothing thing. It decreases to zero as asset duration approaches liability duration, but also the risk is averaged out over the terms of the shorter duration assets, further reducing the risk through diversification of maturities.

For a typical stock and bond portfolio with stock at 50-90% of the portfolio, bond reinvestment risk is not a dominant risk. Moreover, it is likely to materialize when stocks are having difficulties, so one is likely to rebalance or withdraw from bonds at that time anyway.