Maximum Bond Upside

[MachineGhost]

[sophie]

I must say it's going to be REALLY difficult to buy long bonds when I go to invest next month's PP contribution.

FWIW I followed your recent suggestion when my wife made her Roth contribution for 2015... just closed my eyes and bought some TLT, some IAU and some S&P 500 shares. Don't even know the purchase prices but those TLT shares have gone up.

That's absolutely what I'm going to do too, difficult or not! Remember that discussion about what we'll do about long bonds when the yield goes down to 1 or 1.5%? Not there yet but somehow I think we might be breaking 2% sometime soon.

[barrett]

I must say it's going to be REALLY difficult to buy long bonds when I go to invest next month's PP contribution.

FWIW I followed your recent suggestion when my wife made her Roth contribution for 2015... just closed my eyes and bought some TLT, some IAU and some S&P 500 shares. Don't even know the purchase prices but those TLT shares have gone up.

That's absolutely what I'm going to do too, difficult or not! Remember that discussion about what we'll do about long bonds when the yield goes down to 1 or 1.5%? Not there yet but somehow I think we might be breaking 2% sometime soon.

Not only do I remember it but I think it's going to be a big topic on here going forward. Just got a report a couple days ago from one of the Europeans who uses German Bunds. They are the best performing asset in the European PP so far this year (with gold a close second).

Not trying to hijack but what I'd love to understand better is gold's performance over the last few months. The bond performance makes sense to me. Are we all convinced that negative real interest rates are good for gold?

[clacy]

Swiss 30 year bond yield now below zero.

It's hard to believe that people would lend money to governments at a negative yield beyond a few months, let alone for 30 years.

[Kbg]

I'm not sure it's good, but it certainly inverts the normal relationship and makes it not as yucky to hold in terms of potential lost income.

Random question I just thought of...in the US what would be the tax treatment of negative interest treasurys?

[barrett]

Random question I just thought of...in the US what would be the tax treatment of negative interest treasurys?

Oooo! That's a tasty one!. Offsetting losses for everyone when it comes time to file taxes?

[Lang]

There are now just two Swiss government bonds left that still have a positive yield. One is a 2058 (41 years) bond that yields 0.02% and the other is a 2064 (almost 50 years) bond that yields 0.06%.

Also, German Bunds...

[craigr]

It's also possible the market expects Brexit to pass and then the entire EU might unravel. So investors are looking to park money where they think their money would be the safest if the Euro goes away. That could be bonds, cash, and gold.

[barrett]

The German 30-year bond closed today yielding .381%. The Swiss 30-year bond closed yielding -.033. That yield turned negative today. In Japan the yield is .13%.

Might soon be time to revisit that conversation of if & when to get out of 30-year bonds here in The US. So far the lesson from Europe seems to be to stick with 30-year bonds as part of the PP. Not sure I could do it though.

Thoughts?

[sophie]

The US Treasury 30 year bond's yield of 2.28% (as of today) looks really good compared to these numbers. The fact that there is still large-scale purchasing of 30-year bonds with negative yields is proof, to me, that there is no magic resistance at even 1% yields on these bonds.

Difficult it may be, but I plan to forge ahead and buy those bonds. I'm due to buy some next month. The major question to me is, should I buy them in taxable, or save taxable space for gold bullion purchases? I've been figuring to buy physical gold, and convert an equal amount of paper gold in retirement accounts to the desired asset (e.g. bonds). However, if/when 30 year yields increase, I wouldn't be able to tax-loss harvest bonds if I did that.

[Lang]

40 and 50 year bond yields of Switzerland are now both at 0.04%. Ignoring possible changes in price, this means that the cumulative yield of the 50 year bond is just about 2%. Two percent for 50 years of sitting on a bond.

[MachineGhost]

40 and 50 year bond yields of Switzerland are now both at 0.04%. Ignoring possible changes in price, this means that the cumulative yield of the 50 year bond is just about 2%. Two percent for 50 years of sitting on a bond.

I don't think anyone actually believes they will ride out the duration. They're just parked until something better comes along. But man, 50 years at that rate has a duration of about 49.5 years. That's almost half your capital wiped out on a 1% move in yields! That's just insane.

It seems to me what we're getting is essentially a Zero Coupon the closer the yield goes to 0%. Once the yield goes negative, the duration continues to increase! So with that in mind, what someone needs to do is research is what the ideal duration is for use in the PP.

[MachineGhost]

I did a little tentative research.

Current 30-year duration is 1.79x higher that what was necessary to cover Black Monday's -20.5% single day crash and which took about two weeks to cover. That should have alarm bells going off like eating a five alarm chili in a whorehouse.

And lazy calculated equity duration right before Black Monday was about 35 vs about 48.5 now.

[Lang]

Actually, Switzerland's 50 year bond has a 2% annual coupon rate - the bond was issued two years ago, before the NIRP craze. So its duration is "just" 36.5 years.

Anyway, the yield on that bond just went down to zero percent. That's it, guys. We're finished.

[MachineGhost]

Actually, Switzerland's 50 year bond has a 2% annual coupon rate - the bond was issued two years ago, before the NIRP craze. So its duration is "just" 36.5 years.

Anyway, the yield on that bond just went down to zero percent. That's it, guys. We're finished.

I get about 37.5 years, but who's quibbling? At such low a YTM you definitely don't see much of a convexity effect. And 0% YTM breaks my convexity calculator.

So just adjust your duration exposure downards. 25% into 50-year duration is bloody stupid.

[buddtholomew]

Does % decline and duration work in the same fashion when yields are negative (ignoring duration increases)?
In other words, does a -1% decline earn the investor approximately 20% in returns if the 20-year treasury was purchased at a yield of 0%?

[MachineGhost]

Does % decline and duration work in the same fashion when yields are negative (ignoring duration increases)?
In other words, does a -1% decline earn the investor approximately 20% in returns if the 20-year treasury was purchased at a yield of 0%?

Yup, at least as far as I can calculate it. You're just paying interest instead of earning it, but it's still a loss that lengthens the effective duration. Technically, it would be a loss if held to maturity because you don't have enough time to recoup the negative yields. Sort of like bitcoin mining through difficulty increases.

So what I did was record the duration right before Black Monday and adjust it upwards so that the gain on the bonds before they topped out two weeks later covered exactly that one day -20.5% loss. And that's about 12.5 years of duration. So taking VUSTX for example, you would only allocate 18.50% to it instead of 25% to get that duration. Unless anyone has some other brilliant ideas, this is what I'm capping it at.

OTOH, as I mentioned elsewhere, you could use 6.5 years duration at the 1981 peak, allocating 9.6% to bonds. That's the absolute worst case scenario... until it's surpassed. But I don't think that has enough juice to protect equity much. It would have made only about 10% for that 1.45% yield decline after Black Monday.

I suspect it is wrong to look at the overall fixed income duration of the PP by including T-Bills. A Frankenstein intermediate bond doesn't have the flexibility to move against equity declines as a right barbell does. But maybe it is all academic and I'm willing to be persuaded otherwise.

[Lang]

The 50Y Swiss bond keeps on rising! The yield is now at -0.02%. Whoever bought that bond two years ago is up almost a 100% now!

[buddtholomew]

Does % decline and duration work in the same fashion when yields are negative (ignoring duration increases)?
In other words, does a -1% decline earn the investor approximately 20% in returns if the 20-year treasury was purchased at a yield of 0%?

Yup, at least as far as I can calculate it. You're just paying interest instead of earning it, but it's still a loss that lengthens the effective duration. Technically, it would be a loss if held to maturity because you don't have enough time to recoup the negative yields. Sort of like bitcoin mining through difficulty increases.

So what I did was record the duration right before Black Monday and adjust it upwards so that the gain on the bonds before they topped out two weeks later covered exactly that one day -20.5% loss. And that's about 12.5 years of duration. So taking VUSTX for example, you would only allocate 18.50% to it instead of 25% to get that duration. Unless anyone has some other brilliant ideas, this is what I'm capping it at.

OTOH, as I mentioned elsewhere, you could use 6.5 years duration at the 1981 peak, allocating 9.6% to bonds. That's the absolute worst case scenario... until it's surpassed. But I don't think that has enough juice to protect equity much. It would have made only about 10% for that 1.45% yield decline after Black Monday.

I suspect it is wrong to look at the overall fixed income duration of the PP by including T-Bills. A Frankenstein intermediate bond doesn't have the flexibility to move against equity declines as a right barbell does. But maybe it is all academic and I'm willing to be persuaded otherwise.

Thanks MG. I'm basically deer in the headlights for now and not adjusting any fixed income duration by selling LTT's.
Still maintaining 5.6 years and obviously including cash in the calculation.

[MachineGhost]

Thanks MG. I'm basically deer in the headlights for now and not adjusting any fixed income duration by selling LTT's.
Still maintaining 5.6 years and obviously including cash in the calculation.

I did some backtesting and I can't see any relation between stock duration and bond duration. But going too low in bond duration will certainly offset the ability to recoup stock losses. It's also complicated by the fact there's no such thing as a constant duration bond and I'm not sure if bond funds keep it fixed (doubt it).

A 30 year + 1 year = 11.5 year overall duration which is rather high.
A 20 year + 1 year = 8.79 year overall duration which is rather moderate.
A 10 year + 1 year = 5.13 year overall duration which is rather low.

I still don't have an answer but the lower the duration the worse it is for offsetting equity declines. 2008 needed a lot of help and 10 year + 1 year just didn't hack it.

So I don't think there's any duration magic to be had here. Disgusted!

[Lang]

Entire stock of Swiss bond yields turns negative:

The topsy-turve world of government bond markets has just hit a new milestone.

As of today, Switzerland’s entire stock of government debt is now trading at negative yields – with the last to go a bond that does not mature for almost 50 years.

Yields on the Swiss 2064 paper dropped below zero for the first time on Friday, meaning investors are willing to pay more to hold the bond than they will get back in interest and repayment of principal, writes Elaine Moore.

Around the world, government bond yields have fallen to new lows this week as investors anticipate years of ultra-low interest rates.

The yield on benchmark, 10-year UK gilts has reached a fresh low of 0.78 per cent following the UK’s vote to withdraw from the EU and Bank of England governor Mark Carney’s announcement on Thursday that monetary easing was possible in the summer. Swiss 10-year bonds are trading at minus 0.66 per cent.

https://next.ft.com/content/1205b735-e7 ... e96dec4fe3

[buddtholomew]

Thanks MG. I'm basically deer in the headlights for now and not adjusting any fixed income duration by selling LTT's.
Still maintaining 5.6 years and obviously including cash in the calculation.

I did some backtesting and I can't see any relation between stock duration and bond duration. But going too low in bond duration will certainly offset the ability to recoup stock losses. It's also complicated by the fact there's no such thing as a constant duration bond and I'm not sure if bond funds keep it fixed (doubt it).

A 30 year + 1 year = 11.5 year overall duration which is rather high.
A 20 year + 1 year = 8.79 year overall duration which is rather moderate.
A 10 year + 1 year = 5.13 year overall duration which is rather low.

I still don't have an answer but the lower the duration the worse it is for offsetting equity declines. 2008 needed a lot of help and 10 year + 1 year just didn't hack it.

So I don't think there's any duration magic to be had here. Disgusted!

I combine TLT (currently 17 years duration) with highest term CD to create my +/- 5.6 year duration. TLT and SPY are both 25% each so I maintain the PP balance. Love the extra cash even if it does not provide optimal results. I don't underestimate ability to sleep well.

One thing I was thinking of doing is replacing a portion of TLT for EDV and placing the difference in cash.
That way I have less absolute $ in LTT's but still have the duration to hedge losses in equities.

[iwealth]

One thing I was thinking of doing is replacing a portion of TLT for EDV and placing the difference in cash.
That way I have less absolute $ in LTT's but still have the duration to hedge losses in equities.

It's a good plan I think. Only problem is that EDV has lousy liquidity and you'll probably pay a solid 0.1-0.2% more than they are worth due to the bid/ask spread.

[buddtholomew]

It's a good plan I think. Only problem is that EDV has lousy liquidity and you'll probably pay a solid 0.1-0.2% more than they are worth due to the bid/ask spread.

That's unfortunate.
Funny a lack of liquidity in a treasury ETF

[MachineGhost]

EDV has phantom income taxation issues also.

Thinking ahead, we should just assume rates are going to go to 0%. In that case, I don't think we should be buying past 20-year T-Bonds right now. I don't know how bond funds will be reflected, but it seems like it would favor VUSTX at 16.9 duration vs TLT or TLO at 18 duration. I'm willing to pay an extra .05% to .10% just to have autoinvest capability.

This bond crap is so stressful!