Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Have a question about your personal investments? No matter how simple or complex, you can ask it here.
Post Reply
User avatar
Topic Author
Ketawa
Posts: 2192
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by Ketawa » Fri Oct 25, 2019 2:12 pm

I will be investing in a taxable brokerage account soon. With the availability of free ETF trades, I am considering whether to add an allocation to long-term Treasuries in order to extend the duration of my fixed income holdings for theoretical correlation or "flight to quality" benefits, if they manifest. My portfolio is heavily tilted. Here is a breakdown of my portfolio and what I am considering.

Current Portfolio
75% equities, 15% fixed income, 10% alternatives
equities regions split at market weight (domestic/intl developed/emerging)
TSP: 63% of portfolio (G Fund, S Fund, I Fund)
IRA: 33% of portfolio, all AQR multifactor funds (QSMLX, QICLX, QEELX, QSPIX)
taxable: 4% of portfolio, prepaid cards earning 5%+

My fixed income is earning 2.5% with zero volatility. 25% is in prepaid cards earning 5%+, but this isn't scalable any further. The remaining 75% is in the G Fund, which earns 1.75% this month.

The new taxable brokerage account will represent over 10% of my portfolio. I should be able to add it to it further in the future. I was originally planning to throw the whole thing into Vanguard Total World Stock ETF (VT) and not bother with tilting in taxable, but then I remembered that I have long considered extending the duration of my fixed income holdings. I never did it before because it would have taken away from the tilts in my IRA.

Because my prepaid cards holdings are not scalable, the decision at the margin is between continuing to use only the G Fund, or extending duration with something like one of these options.

Code: Select all

Target Duration = 5.0 years
                                                Duration   Yield   % of FI in ETF   Blended Yield
Vanguard Extended Duration Treasury ETF (EDV)       24.3   2.20%              21%           1.84%
Vanguard Long-Term Treasury ETF (VGLT)              17.9   2.09%              28%           1.84%
iShares 20+ Year Treasury Bond ETF (TLT)            18.0   2.02%              28%           1.83%
iShares 10-20 Year Treasury Bond ETF (TLH)          12.4   1.74%              40%           1.75%
G Fund                                               0.0   1.75%
The top 3 options are effectively the same. The long-term Treasury holding would be 3-4% of my portfolio if I target 5 years duration. My primary holding would be EDV for efficiency. I could use VGLT and TLT as secondary holdings for tax loss harvesting.

The G Fund yield should rise to ~1.875% at the start of November, so a comparison to today's yields would be a 35 bp advantage for the G Fund/ETF blend over actual 5-year Treasuries, which earn 1.58% today.

To head off the people who will recommend the Three Fund Portfolio, I will not use something like the F Fund or Vanguard Total Bond Market when they have large allocations to short- and intermediate-term Treasuries that earn less than the 1.75% of the G Fund.

My marginal federal income tax rate is 22% and I do not pay state income tax.

Thoughts on extending duration from 0 to 5 years? Is it worth adding 5 years (or any number of years) of duration for potential correlation or "flight to quality" benefits in exchange for only 10 bp in higher expected return?

dbr
Posts: 30798
Joined: Sun Mar 04, 2007 9:50 am

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by dbr » Fri Oct 25, 2019 2:18 pm

An obvious reaction is that what you do with a few percent of your portfolio won't matter enough to worry about. A related reaction is that what you buy in bonds at 15% of your portfolio won't matter very much.

Over against that is the presumed result of portfolio theory that the optimum bond diversification for high equity allocations is long term bonds only, so that would mean getting out of the G fund and your cards and putting the 15% in EDV.

A different piece of advice is that Larry Swedroe has sometimes suggested that extending duration in fixed income is justified if one can gain 20bps per year of extra duration. I doubt that condition holds today.

Sorry I can't be more helpful.

User avatar
grabiner
Advisory Board
Posts: 25361
Joined: Tue Feb 20, 2007 11:58 pm
Location: Columbia, MD

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by grabiner » Fri Oct 25, 2019 9:12 pm

One note: the G fund rate resets monthly, so the 1.75% is a stale number; current yields suggest that it would be about 1.9% if reset for today, and probably about the same when it resets next week.

The problem with moving money to long-term Treasury bonds is that you increase your inflation risk. Normally, you get some compensation for taking that risk, as long-term Treasury bonds yield more than Treasury bills. The G fund already gives that compensation without the risk. (At the moment, with the flat yield curve, the compensation for the risk is very low; one-month Treasuries yield 1.73%, while 10-30 year Treasuries yield 1.8%-2.2%.)
Wiki David Grabiner

michaelsieg
Posts: 571
Joined: Mon Jan 07, 2013 11:02 am

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by michaelsieg » Fri Oct 25, 2019 10:38 pm

I also think that the term risk you are considering is not worth the additional yield at this time. Just about a year ago, it would have been worth considering, with intermediates yielding more than 3%. Obviously I have no idea which way interest rates will go but there have been fairly large movements over the past 2 years and I think anything is possible.
Quick question, what are the prepaid cards earning 5%? - sorry for my ignorance.
Good luck
Michael

fujiters
Posts: 283
Joined: Tue Mar 06, 2018 2:17 pm

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by fujiters » Sat Oct 26, 2019 2:38 am

michaelsieg wrote:
Fri Oct 25, 2019 10:38 pm
I also think that the term risk you are considering is not worth the additional yield at this time. Just about a year ago, it would have been worth considering, with intermediates yielding more than 3%. Obviously I have no idea which way interest rates will go but there have been fairly large movements over the past 2 years and I think anything is possible.
Quick question, what are the prepaid cards earning 5%? - sorry for my ignorance.
Good luck
Michael
Netspend and Mango, probably. I have a couple of those. Netspend used to offer 5% on $5k, but were limited to $1k about a year ago. Insight was another card that offered this, but they've completely stopped offering 5% on anything.

Doctor of Credit keeps a list of high return accounts like this (there are always hoops to jump through for high rates though): https://www.doctorofcredit.com/high-int ... gs-to-get/
“The purpose of the margin of safety is to render the forecast unnecessary.” -Benjamin Graham

User avatar
Topic Author
Ketawa
Posts: 2192
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by Ketawa » Sat Oct 26, 2019 10:06 am

dbr wrote:
Fri Oct 25, 2019 2:18 pm
An obvious reaction is that what you do with a few percent of your portfolio won't matter enough to worry about. A related reaction is that what you buy in bonds at 15% of your portfolio won't matter very much.

Over against that is the presumed result of portfolio theory that the optimum bond diversification for high equity allocations is long term bonds only, so that would mean getting out of the G fund and your cards and putting the 15% in EDV.

A different piece of advice is that Larry Swedroe has sometimes suggested that extending duration in fixed income is justified if one can gain 20bps per year of extra duration. I doubt that condition holds today.
Agreed, a few percent is probably not going to make much of a difference. I need to compile some data series to run the G Fund through Portfolio Visualizer to see what would have happened over a period with little end-to-end change in the yield curve to see what might have happened in the past, although it will be a short time period.

Long bonds only might make sense in a high equity allocation, but I doubt that is equally true with access to the G Fund.

20 bp per year is probably not realistic any time soon.

User avatar
Topic Author
Ketawa
Posts: 2192
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by Ketawa » Sat Oct 26, 2019 10:15 am

michaelsieg wrote:
Fri Oct 25, 2019 10:38 pm
I also think that the term risk you are considering is not worth the additional yield at this time. Just about a year ago, it would have been worth considering, with intermediates yielding more than 3%. Obviously I have no idea which way interest rates will go but there have been fairly large movements over the past 2 years and I think anything is possible.
Intermediates were over 3%, but so was the G Fund. I doubt that there is ever a time when taking additional term risk is compensated in yield if one has access to the G Fund.

2018-01 2.375%
2018-02 2.750%
2018-03 2.875%
2018-04 2.750%
2018-05 2.875%
2018-06 2.875%
2018-07 2.875%
2018-08 3.000%
2018-09 2.875%
2018-10 3.000%
2018-11 3.125%
2018-12 3.000%
michaelsieg wrote:
Fri Oct 25, 2019 10:38 pm
Quick question, what are the prepaid cards earning 5%? - sorry for my ignorance.
I have 5 flavors of NetSpend cards limited to $1K each: NetSpend, ACE Elite, Brink's, Western Union, HEB. I have 1 Insight card limited to $5K. Those all make 5%. I have DCU savings account limited to $1K with 6.17% APY. All told, about 4% of my portfolio. The NetSpend cards used to have a limit of $5K, and I have had Mango, UnionPlus , and additional Insight cards over the years.

User avatar
Topic Author
Ketawa
Posts: 2192
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by Ketawa » Sat Oct 26, 2019 1:21 pm

I ran some numbers through Portfolio Visualizer using a variety of portfolios. I can't link to them because they use my imported G Fund data series. I picked the longest time periods I could find where interest rates were slightly rising.

Longest time period available for my actual portfolio, which is limited by the time QEELX has been in existence.

30 year rate slightly rising
01/30/15 = 2.25%
07/31/19 = 2.53%

Image


A generic tilted portfolio using DFA funds with longer track records, removing QSPIX from the allocation.

30 year rate slightly rising
12/31/08 = 2.69%
04/30/19 = 2.93%

Image

And finally a non-tilted portfolio using only VT, same time period.

30 year rate slightly rising
12/31/08 = 2.69%
04/30/19 = 2.93%

Image

Worth the effort? Maybe. Over these time periods, the benefit was negligible to 15 bp with slightly rising interest rates. This was from extending duration from 0 to 5 years. Presumably, this was almost solely due to a correlation benefit.

User avatar
Topic Author
Ketawa
Posts: 2192
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by Ketawa » Sat Oct 26, 2019 5:06 pm

Thinking about this some more, I will also consider holding individual long-term TIPS or PIMCO 15+ Year US TIPS Index Fund (LTPZ). TIPS would eliminate the risk of unexpected inflation.

I can purchase individual TIPS at auction or buy/sell them on the secondary market without a commission at Fidelity, save on the expense ratio of a fund, and go ultra-long for efficiency in achieving a potential duration target in combination with the G Fund. However, they might cost extra in taxes due to the loss of capital gains efficiency from an ETF and ongoing taxes for the inflation adjustment.

User avatar
Topic Author
Ketawa
Posts: 2192
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Extend Duration from Zero (G Fund) to Intermediate Term (G Fund/LT Treasury ETF)?

Post by Ketawa » Tue Oct 29, 2019 7:44 pm

Putting to paper a few more of my thoughts after some analysis:

Slim pickings for long term TIPS funds. LTPZ has a duration of 21.5 years and holds the 10 TIPS maturing in 20+ years. If I wanted to use TIPS, it would be better to hold them individually to save on the ER and more efficiently hit a target average duration while maximizing the G Fund free lunch. TIPS maturing in 2049 have a duration of 25.6 years.

However, it doesn't seem like high equity portfolios have benefited from allocations to long term TIPS like they have with long term nominal Treasuries. Throughout their history, TIPS (VIPSX or LTPZ), have had a zero to slightly positive correlation with equities, while long-term nominal Treasuries (VUSTX) have had a slightly negative correlation.

Looking specifically at 2008:
- The real yield curve rose overall in 2008, while the nominal yield curve fell. What was the reason? It seems like the market was pricing in lower expectations for future inflation by the end of the year.
- For some time, the real yield curve was dramatically inverted, with rates over 4% real on 5-7 year maturities for a few days in November, and 3% for 20 year maturities. What was the reason? I've read that Lehman had to dump all their TIPS and this depressed the market. I'm skeptical of that view; it seems possible that the market was pricing in deflation in the intermediate term, with a return to normal inflation levels after some time. On 11/26/08, the real yield on 5 year TIPS was 4.24%. The nominal yield on 5 year Treasuries was 2.01%. Assuming a mostly rational market, the break-even rate was 2.23% deflation.
- After the fact, when inflation expectations returned to normal, it ended up as a great buying opportunity for people with an allocation to TIPS who wanted to lock in real returns, but it wasn't great for anyone with an allocation to long-term TIPS hoping to rebalance.

It seems that TIPS might not be the best asset to use for a goal to benefit from negative correlations. While nominals are exposed to unexpected inflation, changes in inflation expectations seem to have been positively correlated with equities in the past. Fears of deflation, or falling expectations for inflation, have been associated with falling equities. Of course, this may not have been true in the 1970s.

I also do not have much exposure to risks from unexpected inflation. I will have an inflation-adjusted pension from a military retirement starting in as soon as 10 years. My military income, while not guaranteed to rise in lockstep with inflation, is supposed to rise in conjunction with private sector wages, unless legislation specifically changes the adjustment. In the 70s, military pay increased closely with inflation. For example, the increase from 1979 to 1980 was 11.7%.

If I want to extend duration to hopefully benefit from correlations with equities, nominal Treasuries seem like a better fit.

I looked at zeros, as they would allow me to simplify my taxable portfolio by not having to deal with actual interest payments, as well as maximize the use of a G Fund and Treasury barbell since zeros have a duration equal to their maturity. However, spreads were over 50 bp for zeros when I looked today! I'll have to check again, but I'd rather have the simplicity of a fund like EDV if holding individual zeros means losing out on over 5 years of a fund's expense ratio (at least 35 bp higher spread, 7 bp cost for a fund like EDV).

At this point, I'm somewhat leaning towards EDV as the ideal vehicle, but unsure if I'll go through with the change. There's a part of me that thinks it would be "fun" to do. There's also a side benefit that EDV is more efficient than equities in my tax-adjusted asset allocation calculation; effective tax rate of 10% for EDV vs 18% for VT, which effectively means I have a larger tax-adjusted portfolio. Regardless, it probably won't make much of a difference if the duration of my fixed income holdings increases to 5-10 years, representing 3-6% of my portfolio.

Post Reply