If you don't care about volatility, are long-term bonds better than total bond?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
jimkinny
Posts: 1271
Joined: Sun Mar 14, 2010 1:51 pm

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by jimkinny » Fri Jul 10, 2015 9:38 am

Thanks every one and a special thanks to doc, Kevin M and ogd for sharing your time and knowledge.
Jim

dyangu
Posts: 187
Joined: Wed Nov 03, 2010 4:25 pm

Re: If you don't care about volatility, is long-term bonds better than total bond?

Post by dyangu » Fri Jul 10, 2015 9:57 am

Browser wrote:Well, there is the theory that you can get the same portfolio effect with a smaller allocation to long bonds than to intermediate term bonds, so your bond allocation doesn't have to be as large. I barbelled between CDs and long term treasuries.
I do something like this as well. I have limited tax shelter room and I don't want to waste any on short term bonds, especially when CD yield is higher.

User avatar
Doc
Posts: 8309
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Doc » Fri Jul 10, 2015 10:13 am

dyangu wrote:
Browser wrote:Well, there is the theory that you can get the same portfolio effect with a smaller allocation to long bonds than to intermediate term bonds, so your bond allocation doesn't have to be as large. I barbelled between CDs and long term treasuries.
I do something like this as well. I have limited tax shelter room and I don't want to waste any on short term bonds, especially when CD yield is higher.
Currently short term Treasuries are more tax efficient than an S&P 500 fund for many even most of us because of their near zero return. Even in the 50% tax bracket 50% of zero equals zero - no loss to taxes.

There are a lot of threads recently that show that the conventional wisdom of bonds in tax advantaged may not be valid in the current low interest rate environment.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.

hafius500
Posts: 217
Joined: Sat Sep 10, 2011 4:31 pm

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by hafius500 » Fri Jul 10, 2015 2:12 pm

GMO Quarterly Letter 1 Q 2015 - Ben Imker - Breaking Out Of Bonadage
At current yields, the utility of long-term government bonds in most investment portfolios is questionable at best. To our minds, any investors who are not required to own longterm government bonds in their portfolios should warmly consider getting rid of them...
...So the odds that investors are buying 30-year bonds because they have a higher expected return than rolling shorter-term bonds seems pretty unlikely. Now, how about the first scenario – that investors own the bonds because of their exceptional hedging properties?...
...Admittedly, it is a tough ask to an investment committee to replace their holdings in “safe” long-term government bonds with a short in those same bonds. But no one should be under the illusion that long-term bonds trading at these yields are truly safe investments. In all probability they are safe against a loss of value under deflation, but in just about any other circumstance their expected real returns are somewhere between approximately zero to strongly negative
prior username: hafis50

User avatar
Kevin M
Posts: 9779
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Kevin M » Fri Jul 10, 2015 4:36 pm

Larry says that his firm's bond strategy (and other investment strategies) is based on academic research, not opinions. ogd's opinion is that he disagrees with Larry, because sometimes what Larry says the research supports doesn't happen. I haven't seen the academic research, but based on Larry's demonstrated deep and broad knowledge about the academic research, I suspect he could back up what he says.

As Larry often says, we should not confuse strategy with outcome, since a sound strategy based on solid research and reasoning won't always produce positive outcomes.

Nevertheless, I'm skeptical about basing investment strategies too heavily on empirical research, since the results tend to be heavily period dependent, and the future may not resemble the past. I personally do not play the yield curve game, nor do I use "the Larry portfolio", which also is based on empirical research.

At any rate, I think a more interesting, related topic that deserves more attention in the context of this thread is the negative estimated term premium of longer-term bonds. From the Bernanke blog post on the term premium topic (linked above):
To explain the behavior of longer-term rates, it helps to decompose the yield on any particular bond, such as a Treasury bond issued by the US government, into three components: expected inflation, expectations about the future path of real short-term interest rates, and a term premium. At present, all three components are helping to keep longer-term interest rates low.

<snip>

The focus of this post, though, is on the behavior of term premiums—the third component of bond yields. Briefly, a term premium is the extra return that lenders demand to hold a longer-term bond instead of investing in a series of short-term securities (a new one-year security each year, for example). Typically, long-term yields are higher than short-term yields, implying that term premiums are usually positive (investors require extra compensation to hold longer-term bonds instead of short-term securities).

<snip> ... term premiums have recently been zero or even slightly negative.
So if the term premium is zero or negative, investors are not getting paid, or are actually paying, to take term risk. This statement is kind of an oxymoron though, since a zero risk premium implies that the perception is that there is no risk; i.e., there is no perceived term risk in holding longer maturity bonds. This is pretty much what Bernanke goes on to explain in the blog post.

So, I have the following questions. On the one hand, if you believe that longer-term bonds are not risky, why not own them, since you are getting paid for expected inflation and the expected movement of short-term rates? On the other hand, if you are skeptical that there is no risk, why own them unless you are getting paid a term risk premium?

I personally tend more toward the second view, and so don't hold many long-term bonds, but do hedge my bets a bit by holding a relatively small portion of my portfolio in long-term bond funds (not Treasuries though).

Of course ex-post reality does not always comply with ex-ante expectations, as evidenced by the YTD performance of the Vanguard Long-term and Intermediate-Term Treasury funds:

Long-term: -4.62%
Intermediate-term: +1.15%

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

jimkinny
Posts: 1271
Joined: Sun Mar 14, 2010 1:51 pm

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by jimkinny » Sat Jul 11, 2015 6:49 am

I have read that demand by pension funds has increased for two reasons: lowering risk due to lack of need for risk due to equity increases over last several years. % funding of pension obligations has increased. Also some regulation change. I am not sure what this but seems that I read in hazy past pesnion funds had to meet some funding guideline but were given a temporary delay due to 2008/2009 crash.

If pension fund obligations in nominal dollars this makes since because the fund does not care about inflation and therefore no inflation risk.

my links and quotes do not seem to work with new computer but will try
[urlhttp://www.bloomberg.com/news/articles/2014-05 ... here-s-why][/url]

What I find astonishing is that I can buy a brokered 10 year CD yielding just lightly less than a 30 year Treasury. 10 year CDs yields were in range of 3.2-3.3 several weeks ago and by buying some secondary market CDs of term 9-<10 year, close to that.

User avatar
ogd
Posts: 4854
Joined: Thu Jun 14, 2012 11:43 pm

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by ogd » Sat Jul 11, 2015 11:24 am

Kevin M wrote:Larry says that his firm's bond strategy (and other investment strategies) is based on academic research, not opinions. ogd's opinion is that he disagrees with Larry, because sometimes what Larry says the research supports doesn't happen. I haven't seen the academic research, but based on Larry's demonstrated deep and broad knowledge about the academic research, I suspect he could back up what he says.
I agree with Larry when he says long bonds work well with high equity portfolios, without paying heed to yields at all. I disagree with him when he (elsewhere) says you should demand 20bp per year. Unlike the former, I haven't actually seen a proper study supporting the latter.

Another rule of thumb is (or used to be) "avoid California munis". Neither of us two pays attention to that one, I believe. I think the rules of thumb are the more speculative part of the Larry worldview, or as he says himself about the first one, "marginal".
Kevin M wrote:So if the term premium is zero or negative, investors are not getting paid, or are actually paying, to take term risk. This statement is kind of an oxymoron though, since a zero risk premium implies that the perception is that there is no risk; i.e., there is no perceived term risk in holding longer maturity bonds. This is pretty much what Bernanke goes on to explain in the blog post.

So, I have the following questions. On the one hand, if you believe that longer-term bonds are not risky, why not own them, since you are getting paid for expected inflation and the expected movement of short-term rates? On the other hand, if you are skeptical that there is no risk, why own them unless you are getting paid a term risk premium?
The conundrum / oxymoron is much improved if you add the hedging aspect into the picture. So instead of "the perception is that there is no risk" you might say, "for the average player, the hedging value of long bonds balances the term risk, or exceeds it" (Bernanke too acknowledges the hedging, as I mentioned).

This leaves the door wide open to differing assesments of "hedging value" for different players. Like other things depending on aspects extraneous to the asset being discussed (e.g. tax bracket for munis), this can lead each of us to differing or opposite conclusions. I might say "well I have a high equity allocation, so for me the hedging value is higher than average", or perhaps, the thing that gives me pause about long bonds, "a lot of players in this field have long-term nominal obligations, driving the average assessment uncomfortably high for me", or even "long bonds currently work well with CDs and/or separating high and low yields between tax-advantaged and taxable, both things more advantageous to me vs the market at large".

So I think it's necessarily more complicated than merely checking how much yield you get for duration, just like "it's more complicated" than the OP's proposal that if horizon is long enough, get long bonds.

User avatar
Doc
Posts: 8309
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Doc » Sat Jul 11, 2015 11:57 am

ogd wrote:Unlike the former, I haven't actually seen a proper study supporting the latter.
Apparently the original research on the concept comes from Eugene F. Fama and Robert R Bliss "The information in Long-Maturity Forward Rates" September 1987 edition of the American Economic Review. The 20 bps apparently comes from DFA.
Swedroe/Hemple in 'Winning Bond Strategy' March 2006 wrote: ... which has successfully used this approach since 1983... DF imposes an arbitrary rule that a longer maturity must provide at least twenty basis point per annum in higher returns ...
I know this is a stretch in references especially since you don't tend to agree with Swedroe but it is what it is.

Found a paper from Dimensional that may be the source. I haven't read it. http://www.ifaarchive.com/media/images/ ... income.pdf
Investors can increase their risk-adjusted returns with an alternative approach
developed by Professor Eugene Fama of the University of Chicago. This variable
maturity strategy shifts the maturities of the portfolio as yield curve changes create
the possibility for lower risk, higher expected return outcomes.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.

User avatar
Kevin M
Posts: 9779
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Kevin M » Sat Jul 11, 2015 8:26 pm

jimkinny wrote:I have read that demand by pension funds has increased for two reasons: lowering risk due to lack of need for risk due to equity increases over last several years. % funding of pension obligations has increased. Also some regulation change. I am not sure what this but seems that I read in hazy past pesnion funds had to meet some funding guideline but were given a temporary delay due to 2008/2009 crash.

If pension fund obligations in nominal dollars this makes since because the fund does not care about inflation and therefore no inflation risk.
Bernanke also discusses this type of effect in Why are interest rates so low, part 4: Term premiums | Brookings Institution:
Ben Bernanke wrote:Changes in regulation and market practices also affect the demand for safe, liquid assets, such as Treasury securities, lowering their term premiums. For example, new regulations require banks to hold ample liquidity and securities dealers to post more collateral in derivatives transactions. Insurance companies and pension funds also face rules that effectively require them to hold significant amounts of safe, longer-term bonds. This mandated demand seems likely to put downward pressure on longer-term yields for the foreseeable future.
He doesn't use the term, but this is an example of what is referred to as the segmented market hypothesis in term structure of interest rates (yield curve) theory. The idea is that rates over specific maturity segments are determined by supply and demand within each segment, the demand side of the segmentation being determined by factors like those mentioned by Bernanke.

The more dominant theories are the expectations hypothesis and liquidity preference preference hypothesis. The former refers to the belief that long-term rates are determined by the expectations for future short term rates. The latter basically says that investors demand a term premium for holding longer maturity bonds due to greater uncertainty. As Bernanke points out, expectations are that short term rates will remain low, and there seems to be less uncertainty about long-term bond returns than in previous times due largely to low inflation expectations. So both dominant theories support the case for relatively low long-term bond rates.

I tend to agree with Bernanke's presentation, that pretty much says that all three theories are at play in determining the yield curve. However, according to one of my investment textbooks, the segmented market hypothesis has "meager empirical support", but "enjoys wide acceptance among market practitioners, and according to another one "This view of the markets is not common today".

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

jimkinny
Posts: 1271
Joined: Sun Mar 14, 2010 1:51 pm

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by jimkinny » Mon Jul 13, 2015 11:18 am

In part 4 of Bermake's blog post re interest rates he shows a graph, fig. 3, plotting 10 year Treasury yield, term premium and risk-neutral yield over time.

The estimated risk-neutral yield is about 2-2.5% above the actual term premium as of most current data points.

I understand that Bernake discusses matters outside of domestic issues like QE by the ECB and unease as possible factors to explain the lack/low levels of a term premium currently as well as domestic issues.

It will be very interesting to see how this plays out.

That missing 2.5% risk-neutral yield seems pretty large when consider we now have a 2.4% yield on the 10 year. If I am interpreting the graph correctly, that is what the estimate risk-neutral would have predicted, an additional 2.5% yield to the current 2.4, which is a doubling of yield. I am not qualified to judge the reliability of the risk-neutral yield of course, but it seems a bit out of whack.

It is also interesting that Larry Swedroe discusses how his standard approach is a 10 year ladder, regardless of current rates and that rule of thumb of 0.2 increments for adding term risk, that Doc traced back to 1983 or so. So, at least based on the 2014 article, he does't fiddle much except in some cases going longer when the 0.2 premium is reached, but not always.

I have learned some things, thanks every one.

Jim

User avatar
Kevin M
Posts: 9779
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Kevin M » Mon Jul 13, 2015 4:01 pm

jimkinny wrote:In part 4 of Bermake's blog post re interest rates he shows a graph, fig. 3, plotting 10 year Treasury yield, term premium and risk-neutral yield over time.

The estimated risk-neutral yield is about 2-2.5% above the actual term premium as of most current data points.
This is just another way of saying that the term premium has been close to zero lately. As Bernanke explains:

Risk Neutral Yield = Yield - Term Premium
Or, RNY = Y - TP

Which we can rearrange to be:

Y = RNY + TP

Looking at either equation, we see that if TP = 0, Y = RNY. In other words, the yield is just the risk neutral yield if investors don't demand a term premium.
jimkinny wrote:That missing 2.5% risk-neutral yield seems pretty large when consider we now have a 2.4% yield on the 10 year. If I am interpreting the graph correctly, that is what the estimate risk-neutral would have predicted, an additional 2.5% yield to the current 2.4, which is a doubling of yield. I am not qualified to judge the reliability of the risk-neutral yield of course, but it seems a bit out of whack.
I think this is confused. The risk neutral yield is not "missing"; it is just Y + TP, and if TP = 0, Y = RNY. RNY does not predict anything--at least not from what we read in the blog post.

It's important to keep in mind that the term premium cannot be observed, but can only be estimated, so there is some fuzziness in all of this. If one estimates TP to be 0, then by definition RNY = Y, and using your 2.4% rate, Y = RNY = 2.4%. If were to estimate the term premium at say 0.5%, then RNY = Y - TP = 2.4% - 0.5% = 1.9%.

There is no "doubling of the yield" unless one assumes that the term premium will jump from 0% to 2.4% (and nothing else changed; i.e., no change in expected inflation or expected short term rates). The second graph in the blog post shows the estimated term premium since 1961, and we can see it has varied a lot.

Note that even though the estimated term premium increased from about 0% in the mid-2000s to well over 2% in the late 2000s, the yield actually decreased! Using Bernanke's decomposition of yield into inflation expectations, short-term rate expectations, and term premium, the implication is that inflation expectations and short-term rate expectations declined more than the term premium increased during this period.

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

jimkinny
Posts: 1271
Joined: Sun Mar 14, 2010 1:51 pm

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by jimkinny » Tue Jul 14, 2015 8:56 am

Kevin M wrote:
jimkinny wrote:In part 4 of Bermake's blog post re interest rates he shows a graph, fig. 3, plotting 10 year Treasury yield, term premium and risk-neutral yield over time.

The estimated risk-neutral yield is about 2-2.5% above the actual term premium as of most current data points.
This is just another way of saying that the term premium has been close to zero lately. As Bernanke explains:

Risk Neutral Yield = Yield - Term Premium
Or, RNY = Y - TP

Which we can rearrange to be:

Y = RNY + TP

Looking at either equation, we see that if TP = 0, Y = RNY. In other words, the yield is just the risk neutral yield if investors don't demand a term premium.
jimkinny wrote:That missing 2.5% risk-neutral yield seems pretty large when consider we now have a 2.4% yield on the 10 year. If I am interpreting the graph correctly, that is what the estimate risk-neutral would have predicted, an additional 2.5% yield to the current 2.4, which is a doubling of yield. I am not qualified to judge the reliability of the risk-neutral yield of course, but it seems a bit out of whack.
I think this is confused. The risk neutral yield is not "missing"; it is just Y + TP, and if TP = 0, Y = RNY. RNY does not predict anything--at least not from what we read in the blog post.

It's important to keep in mind that the term premium cannot be observed, but can only be estimated, so there is some fuzziness in all of this. If one estimates TP to be 0, then by definition RNY = Y, and using your 2.4% rate, Y = RNY = 2.4%. If were to estimate the term premium at say 0.5%, then RNY = Y - TP = 2.4% - 0.5% = 1.9%.

There is no "doubling of the yield" unless one assumes that the term premium will jump from 0% to 2.4% (and nothing else changed; i.e., no change in expected inflation or expected short term rates). The second graph in the blog post shows the estimated term premium since 1961, and we can see it has varied a lot.

Note that even though the estimated term premium increased from about 0% in the mid-2000s to well over 2% in the late 2000s, the yield actually decreased! Using Bernanke's decomposition of yield into inflation expectations, short-term rate expectations, and term premium, the implication is that inflation expectations and short-term rate expectations declined more than the term premium increased during this period.

Kevin
Yes, I was confused. I was looking at the graph as missing risk neutral as opposed to zero or near zero term premium. The simple formula makes it very clear.

The decomposition of rates is actually quite informative. It provides a way of thinking about the yield that had been invisible to me.

thanks, jim

User avatar
Kevin M
Posts: 9779
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Kevin M » Wed Jul 15, 2015 2:52 pm

jimkinny wrote: The decomposition of rates is actually quite informative. It provides a way of thinking about the yield that had been invisible to me.
Yes, it is a good way to think about it. Bernanke's decomposition of longer-term Treasury rates is: "expected inflation, expectations about the future path of real short-term interest rates, and a term premium". Another way to decompose it is simply expectations about the future path of nominal short-term interest rates, and a term premium, with the understanding that inflation expectations are built into all nominal rates. This simpler, two-component approach is basically a combination of the expectations hypothesis (long-term nominal rates reflect future path of nominal short-term nominal rates) and the liquidity preference hypothesis (long-term nominal rates reflect current short-term nominal rates plus a term premium).

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Re: If you don't care about volatility, is long-term bonds better than total bond?

Post by rca1824 » Fri Jul 17, 2015 7:34 am

nisiprius wrote:Most most people do care about volatility.
Warren Buffet doesn't. He focuses on earnings, not price. Price only matters during the initial buy.

Of course Buffet would be 100% equities making the bond discussion moot.
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB

User avatar
Kevin M
Posts: 9779
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: If you don't care about volatility, is long-term bonds better than total bond?

Post by Kevin M » Fri Jul 17, 2015 9:29 pm

rca1824 wrote:
nisiprius wrote:Most most people do care about volatility.
Of course Buffet would be 100% equities making the bond discussion moot.
Ah, not true! Buffet (Berkshire Hathaway) also holds cash (short-term Treasuries)--the shortest-term bonds of all. And for the inheritance he will leave his wife, 90% S&P 500 and 10% short-term government bonds (www.berkshirehathaway.com/letters/2013ltr.pdf.

Buffet has made it very clear that he thinks bonds (anything other than the shortest-term) are not currently a good value.
If I had an easy way, and a non-risk way, of shorting a whole lot of 20- or 30-year bonds, I’d do it,” Buffett said Monday on CNBC. “But that’s not my game, and it can’t be done in the kind of quantity that would make sense for us. But I think that bonds are very overvalued, I’ll put it that way.
Source: Buffett Says He’d Short 30-Year Bond If Could Be Done Easily - Bloomberg Business

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

User avatar
Rx 4 investing
Posts: 735
Joined: Sat Apr 25, 2015 11:03 am

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Rx 4 investing » Fri Jul 17, 2015 11:57 pm

Image

" The Bureau of Labor Statistics released the June CPI data this morning (7-17-15) .

--The year-over-year unadjusted Headline CPI came in at 0.12% (rounded to 0.1%), up from -0.04% (rounded to 0.0%) the previous month.

--Year-over-year Core CPI (ex Food and Energy) came in at 1.76% (rounded to 1.8%), up fractionally from the previous month's 1.72% (rounded to 1.7%)." This still continues to run below the Fed's target of 2%.

Source: http://www.tradingeconomics.com/united- ... lation-cpi

Image

The $USD is strengthening. Has anybody seen the recent trend in oil, commodities, and gold? The world's largest consumer of commodities, China, is trying to halt their economic slide with monetary stimulus. I don't understand all the fuss about bonds in this dis-inflationary environment.

The 30 year T-Bond closed the week at 3.08%, and the 10 year T-Bond at 2.34%. They seem appropriately priced to me given inflation expectations.
“Everyone is a disciplined, long-term investor until the market goes down.” – Steve Forbes

User avatar
Kevin M
Posts: 9779
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: If you don't care about volatility, are long-term bonds better than total bond?

Post by Kevin M » Sat Jul 18, 2015 1:40 am

Rx 4 investing wrote: The 30 year T-Bond closed the week at 3.08%, and the 10 year T-Bond at 2.34%. They seem appropriately priced to me given inflation expectations.
Warren Buffett wrote:If I had an easy way, and a non-risk way, of shorting a whole lot of 20- or 30-year bonds, I’d do it <snip> I think that bonds are very overvalued.
Reasonable minds can differ.

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

Post Reply