Bogleheads

[SmallHi]

I find, sadly, that there are some common and persistent misconceptions about Short Term Corporate Bonds as a stand alone investment and within the context of a balanced portfolio.

I will keep this thread brief and try to avoid a lot of charts and numbers, instead sticking with facts. If there is any disagreement or disbelief, certainly #s can be provided.

A few myths/realities:

MYTH: Its not worth taking credit risk in the bond markets.

REALITY: Credit risk works very consistently on the short end of the curve (1-3 years), and has produced annualized premiums of about 0.8% over short term treasuries dating back to the 1970s. In only 4 of the last 30 years have ST Treasuries outpaced ST Corporate bonds. 0.8% may not sound like a lot, but many investors have used the 0.7% annual return of 5YR Treasuries over 1YR Treasuries since 1964 as justification to extend out on the curve.

...Longer term credit risk is a more dicey proposition.


MYTH: corporate bonds do not mix well with a balanced stock portfolio -- relative to treasuries

REALITY: In fact, 60/40 Global Balanced Equity/ST Corporate portfolios had about 0.2% higher returns annually for a given amount of volatility relative to Equity/ST Treasury portfolios.


MYTH: on average, Treasuries are a better hedge against equity risk than corporate bonds

REALITY: On the short end, high quality corporate bonds have actually hedged equity risk better than treasuries. In the 8 years since 1978 that TSM underperformed T-bills, ST Corporates averaged almost 0.6% a year higher returns


MYTH: Treasuries offer better defense during periods of general economic decline

REALITY: According to the NBER, there have been 4 recessions since the late 70s (80, 82, 90, and 01) spanning 38 months. On average, 1-3YR Corporate bonds returned +1.28% during these months, whereas 1-3YR Treasuries only returned +1.19%. Obviously these are economically different, but not statistically different...both of these returns, however, are statistically different from the +0.7% and +0.6% returns for ST Corporates and ST Treasuries during all periods since 1977. But the ST Corporate/Treasury spread does not change whatsoever depending on the current stat of the business cycle.


MYTH: on average, if you take value or size risk in the equity markets, you should avoid corporate bonds, as treasuries are a better hedge against negative factor periods

REALITY: There is no evidence that Corporates underperform treasuries when value or small is out of favor. Actually, on average, corporates have beaten treasures by 0.3% to 0.5% a year on average with large beats small and growth beats value. Corporates a good bet whether you hold neutral or tilted portfolios


MYTH: Its prudent to avoid credit risk and instead opt for term risk in a balanced portfolio -- using Int'd Term Treasuries instead of ST Corporate bonds

REALITY: While Int'd Treasuries perform modestly better during equity downturns, a Stock/ST Corporate + Stock/Int'd Treasury portfolio (adjusted for volatility) have had identical returns since 1977.

a) this period is obviously biased against ST bonds, as interest rates have fallen dramatically over this period. Its unlikely the two portfolios would look the same if interest rates were likely to rise over a prolonged periods...such as the 1964-1981 period.

b) downside losses are only one "risk". Inflation is another risk especially pressing for income oriented retirees. Int'd Treasuries expose oneself to a rather large amount of short term inflation risk, on that many investors would prefer to avoid relative to an extra loss of 1% to 2% every so often.


MYTH: If short term corporates make sense, then even lower quality junk bonds (BBB and below) work even better because of even higher potential returns

REALITY: The "credit play" really only works at the short end of the yield curve, and breaks down at lower credit levels. Junk bonds have higher correlations with stocks, size, and value firms, and don't diversify you as well as higher quality issues.

To summarize, I think short term corporate bonds (not short term treasuries or intermediate term treasuries) should represent the core fixed income holding for most fixed income only or balanced investors, but more isn't better in this case -- as lower quality bonds (junk) aren't worth the effort...and, luckily, Vanguard and iShares have two excellent strategies to capitalize on this opinion. The Vanguard ST Investment Grade fund and the iShares 1-3 Year Credit bond index.

sh

edited to add interesting bit about corporate/treasury returns and economic recessions that I don't believe anyone is aware of

[gbs]

Would be hard to debate you without having the returns

Anybody care to post them?

Regards, gbs

[larryswedroe]

SH
Your conclusions are exactly what the literature shows--the ONLY place that credit risk has been rewarded is on the very short end of the curve. My guess is that is the call risk is then minimized

And of course you get the freebee of the tax issue if you hold in tax advantaged accounts.


BTW-my books cite the studies that have show that. I believe I have posted the data here more than once.

[Robert T]

.
SH,

Just to focus on one point:

[docneil88]

[SmallHi]

Robert,

I chose to use the ML 1-3 Year Indexes because it was the longest data I could find on a credit and treasury bond index that wasn't long term, and seems to correspond too Bernstein's thoughts (from EF '97):

[SmallHi]

Neil,

I think the nature of longer term credit risk (say corporate bonds with 10 years or more till maturity) have a very different risk profile than, say, 1-3YR or 1-5YR corporate bonds.

For one, I assume short term corporate financing will be more securely collateralized, and prospects are certainly easier to forcast over short periods.

Finally, the price movements of longer term corporate bonds (regardless of how long an investor plans to hold them) behaves very differently from short term corporate bonds, both relative to treasuries, and within a balanced portfolio. Consider this:

From 1978 to 2007, LT Corporate Bonds beat LT Gov't Bonds 50% of the time. On average, during these 15 years where there was a long term default deficit (LT Gov beat LT Corp), 1-3YR Corporate Bonds beat 1-3YR Treasuries by 0.5% per year. There is very little correlation between LT credit risk and ST credit risk -- of the 15 negative LT DEF years, ST corps only trailed ST treasuries in 3 of these years.

sh

[docneil88]

[Willy]

The most typical recommendation I see on this great forum for the fixed income portion of one's portfolio is ST treasuries or Intermediate treasuries, mixed with TIPS. Larry, if SmallHi's analysis on this thread is correct, can you please explain why you and many others recommend ST treasuries instead of ST corporate?

Thanks to both SmallHi and Larry for your consistently terrific and helpful input!

[TheEternalVortex]

Thank you for this excellent post. It was very informative.

[Trev H]

Interesting and thanks SH for the Post and Robert and Larry, others for the contributions.

Could someone please post the raw data, annual returns for the 1-3 yr investment grade corp bonds.

Something simple like by year...

5.2
1.4
6.3
4.8

So I could yank em and put them in my backtesting spreadsheet.

Or if you could provide a link to a site where they could be viewed or downloaded, that would be great.

I would specifically be interested in the years 1970-2007 but I will replace the later years with returns from Vanguards ST Corp Fund.

Simba may also want to include them in his excellent spreadsheet.

Thanks

Trev H

[Robert T]

.
Sh,

We just seem to reach different conclusions on the evidence “that Corporates underperform Treasuries when value or small is out of favor”. You seem to suggest there is no evidence that this happened – I suggest there is…specifically the data of actual Vanguard fund returns (or their underlying benchmarks) over the last 25 years, shows that in all 6 years that SV had a negative annual return Vanguard Short-term Corporate underperformed Vanguard Short-term Treasury (or its underlying benchmark). [this was a reality in actual returns of the funds available to Vanguard investors].

[Les]

[financialguy]

Trev,

If you want to plug in the data for ST Investment Grade, here are the returns since 1985 you can copy and paste in the "Data_85_07" tab of the Simba spreadsheet:

14.90
11.40
4.50
6.90
11.40
9.20
13.10
7.20
7.07
-0.08
12.74
4.79
6.95
6.57
3.30
8.17
8.14
5.22
4.20
2.11
2.20
4.99
5.86

Returns from before 10 years ago are taken from http://www.fundadvice.com/feht..../0206.html

[Kenster1]

For those who are unsure either way -- you can always ride the Short-Term Bond Index Fund which is about...
65% Treasury, 35% Corporate.
_________________
SURGEON GENERAL'S WARNING: Any overconfidence in your ability, willingness and need to take risk may be hazardous to your health.

[Trev H]

Thanks financial guy, I do have a copy of Simba's sheet but what I was really interested in was the 1970-1984 timeframe yearly returns for ST Corp (or a decent benchmark index).

I usually just backtest wtih IT Treasury or Total Bond, or perhaps include the SynTIPS. Would be nice to have ST Corp since I have heard Larry in the past say they would be a reasonable choice if you wanted more risk/return from the bond side of your portfolio.

Perhaps SH will be able to list them later on.

Trev H

[Robert T]

.
FWIW - this is what I get using Vanguard Corporate bond data from 1983-2007 (Data provided at end of post):

-ST Corporates produced an annualized premium of 0.26% over Short Term Treasuries with similar standard deviation. i.e. since 1983 an all bond investors would have earned a higher ‘risk-adjusted’ return in ST Corporates than in ST Treasuries. This is consistent with the Ilmanen finding in ‘Which Risks Have been Rewarded Paper?”. (note this just looks at bonds in isolation - I get a different result when looking at a stock:bond portfolio).

- Short-Term Treasuries outperformed ST Corporates in 7 of the last 25 years. In all but one of the years that ST Treasuries outperformed ST Corporates, SV returns were negative (IMO this is largely consistent with theory – flight to quality, default risk correlates).

- When added to a global small cap and value tilted portfolio, ST Corporates added small incremental returns relative to treasuries (0.08% and 0.04% for a 60:40 and 75:25 portfolio respectively), but at the cost of higher volatility. (Sharpe Ratios are higher when ST Treasuries are added to the portfolio than when ST corporates are added - the differences, however, are fairly small).

- In my view Corporate bonds have higher downside risk than Treasuries (greater negative skewness) as reflected in 2007. While this looks like an outlier in the historical data, as Taleb indicates - in financial markets outliers typically dominate the mean, and it just takes a few outliers to have a significant effect. And as above IMO Corporate bonds are more susceptible to these outliers than Treasuries. Just my opinion.

[larryswedroe]

few quick thoughts
I have no real problem with ST very high quality corporate though really only benefit from them in tax advantaged accounts--you get the freebee of the tax differential.
But for me since I can buy Treasuries or government agencies directly with no manager needed but corporate bonds I would want diversification so you need a fund, the advantages tend to disappear.
And the risks show up at the wrong time. (as Robert showed)

But, again, no problem with very high grade corporate at short end. Go longer then you likely have the call risks. Would only buy bullets. But one other point, at least with very high grade corporates that are callable you don't the I lose/I don't win problem that you have with junk

Junk gets called either from falling rates OR improved credt. If already have AAA or AA hard to have the later happen, so the risk is only the former.

BTW DFA will shortly introduce two new investment grade bond funds, 1-5 year and 5-10 year. It will hold only investment grade (but different from their other funds that hold only the highest end) and only either non callables or what are called make whole callables.

Hope that helps

[SmallHi]

Hi gang,

Just a quick addition. I will add the 1977 to 2007 returns for 1-3 Year Corporate Bonds and Treasury bonds (both indexes can be purchased through iShares for 0.2%, fyi). My personal preference is to draw conclusions based on the index returns (especially given the availability of low cost options for direct investment today).

[financialguy]

I would love to see returns of ST Corporates vs. ST Treasury Bills/Notes during a black swan event like the Great Depression. I say this because conceptually one would think that owning Treasury debt would give a better risk-adjusted return during times like that, since the US government can fund its obligations through the power to tax. Theoretically, a primary reason for owning government obligations is to protect against black swans.

[dumbmoney]

[tc101]

[tc101]

[SmallHi]

[Robert T]

.
Sh,

Time will tell...(the Vanguard live data seem consistent with theory IMO - perhaps this will change in the future or not with live index fund data)...

[financialguy]

[pastafarian]

[Les]

Treasuries did better then equities between 7/80 to 6/82, but there was 18.1% inflation during that period. Taxes and inflation were killers then.

[Les]

[larryswedroe]

SH
Just to add the two new funds will not use a shifting maturity approach, as trading costs would be too high, not as liquid as treasuries and agencies.

With investment grade risk especially at shorter end there is little default risk and you don't have much of the equity risk that you get in junk bonds and remember they will also not buy callables which creates problems (unless make wholes)

As you would expect they do have good data on this. Let me add they are suggesting it for people that have high equity allocations already and will only look at the portfolio as a whole and not the components.
The portfolio's (if high equity allocation) SD will be dominated by the SD of the equities so little added risk (at least in terms of SD). And will be highly diversified with hundreds of names (as you would expect).

FYI-It will be called I think an extended investment grade fund. Isolating A/Baa. which is about 2/3 of the investment grade universe.
73-3/08 returns and SD
AAA 8.19/5
AA 8.28/5.08
A 8.39/5.09
Baa 8.88/5.32
That data is the Lehman Intermediate Credit Index

73-07 Interesting DATA
75/25 equity and AAA and then 75/25 with Baa: Equity DFA balanced strategy
13.43/11.18 and 13.57/11.22
now 25/75
10.08/5.75 with AAA and 10.63/8.46 with Baa

So seems to add bit of value if have high equity allocation but with high fixed income allocation you get incremental return but much more volatility, so probably not appropriate.

And just a note, in my book Wise Investing Made Simple there is a table of recommended funds and one of them is the Vanguard Short Term Bond Fund which includes the full spectrum of investment grade bonds.

[pastafarian]

[gbs]

X - STDEV
Y - Return
Blue curve - arithmetic efficient frontier
Red curve - geometric efficient frontier

Short term treasuries vs Short Term corporate 1977 - 2007.


Basically same risk return if more than 65% TSM.

Intermediate term treasuries vs Short Term corporate 1977 - 2007.

Intermediate treasuries a tad better after 50% equity.

TIPS vs Short Term corporate 1977 - 2007.

TIPS - well ahead.

Treasuries/ TIPS don't look bad at all.

Will run against other portfolios but I will need the portfolio returns from 1977 to 2007.

Regards, gbs

[matt]

Wow. I didn't even know most of those were myths. Shows what I know, I guess. But then again, I don't pay too much attention to decisions that will result in 5 bps of performance difference. If I have to think hard, crunch a bunch of numbers, and make choices about such things, I would like them to pay off a little better.

[Random Musings]

At current rates, the almost 200 basis point spread between ST Treasury Fund and ST Inv Grade makes the Inv Grade a compelling argument.

Placed a portion in my tax-deferred after the final Fed cut - but also still have some treasuries.

GBS: what was the "spread" between ST Treasuries and ST Corp at the starting point of your 1997-2007 analysis? If this spread was more "normal" (or even lower) compared to other times during that time period you were looking at (and perhaps that isn't even the case), could you run the efficient frontiers starting at time periods when the spreads look more like today's?

Regards,

RM

[gbs]

[james22]

[dumbmoney]

Here's an earlier thread on this topic with some more data:

http://www.diehards.org/forum/....p?p=196961

Given that treasuries have advantages in taxes, liquidity, safety, etc., one would expect corporate bonds to have done much better. So it's odd to frame this as a love affair with treasuries, when it seems that treasuries have been undervalued. Maybe the love affair is with yield. Bond fund managers have incentives to inject equity risk into their funds (it made Bill Gross a billionaire).

[Robert T]

.
The above discussion IMO raises the question (at least to me) of: Can credit risk be effectively indexed?

As Vanguard is regarded as the leader in market index funds – How did they do?

[Sunny Sarkar]

[stratton]

I started a new thread for this at http://www.bogleheads.org/foru....707#218707

Paul

[nisiprius]

Does it matter much?


_________________
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

[Taylor Larimore]

Hi Bogleheads:

I will attempt to answer nisiprius's question: "Does it matter much?"

The bond market is EXTREMELY efficient. Thousands of highly paid and experienced bond managers are sitting at computers day and night looking for bonds with any small advantage in risk or return. We can be almost certain that any higher expected return from bonds will be offset by higher expected risk. It can't be otherwise, or most investors would soon crowd into the superior bond fund.

ANY of of Vanguard good quality short- or intermediate-term bond funds should do the job of providing income and safety for a portfolio. We can narrow our choices by looking at bond yields--the higher the yield, the more the risk (small as it is).

So, to answer nisiprius's question: It doesn't matter much.

Best wishes.
Taylor

[SmallHi]

Robert,

[SmallHi]

Robert,

[larryswedroe]

If I had to choose only one the answer is very clear--based on the academic papers, TIPS should dominate the fixed income portfolios.

Beyond that, for investors with high equity allocations then taking small amount of credit risk, still sticking with investment grade is likely to get you best risk/return (Sharpe Ratio), at least based on the data. But if have low equity allocation then staying only with Treasuries and AAA/AA is likely to get best result.

So like with most things the answer depends.

Personally I like the safety of the Treasuries/agencies/highest investment grade, as don't want that risk to show up at the wrong time, which it has tendency to do--as Robert has nicely pointed out. But then again I fit the investor profile of low equity allocation.

[SmallHi]

Larry,

As I understand it, the idea of keeping bonds short and high quality (at any equity/fixed split) was more about saving risk so that you can amplify it on the equity side (by tilting more to small/value if you so desire).

Let me give you an example (I am sure you are aware of, but more for the benefit of others):

[alec]

[larryswedroe]

SH
Exactly, the rewards of taking risk on equity side have been better and certainly more tax efficient. That is the reason for my recommendations.

As another example for everyone's benefit you might rerun your numbers to come up with P1 and P3 with not the same SD (with P3 being higher return) but with the same return and see how much lower the SD is. Have to play around to find the answer but not hard. Good exercise if you have the technology.

[Robert T]

.
Sh,

On 6/81 – 5/82: The Lehman Brothers 1-3 yr Government Index return was 16.96%, not sure why there is more than 3% difference with the ML 1-3 yr Treasury for the same time period?

Black Swan’s and outliers: Your point is well taken, although I think there are two things being discussed, both raised by Taleb. The first is Black Swan – “events with small probabilities but significant repercussions” (he highlights wars, weather related events, bank collapses…), the second is ‘outliers’ which he indicates in his book often dominate the means in finance (e.g. 10 days accounted for about half the S&P returns over the 10 year period to mid 2005; 3% of stocks accounted for 60% of SV excess returns on average from 1927-2005; 33 stocks accounted for 75% of the S&P 500 return in 1998, …etc…). His suggestion (as I understand) is to recognize this phenomenon when making decisions (try to reduce sampling error risk).

On Vanguard tracking error: Maybe Vanguard Manager’s goofed, I don’t know. But perhaps the margin for ‘goofing’ is higher with corporate bonds than with Treasuries. I have tried to find other ST corporate index funds for comparison but did not find any. Will this ‘tracking error’ happen in future? I don’t know – but it just takes one year to significantly effect mean differences and with corporates the risks seem higher than for treasuries. Will the 107 of 550 bond sub-sample of the iShares 1-3 Credit fund be immune to similar sampling error? Time will tell… (I'll just stick with my real and nominal treasury allocation and get credit related risk in equities).

On another point on sampling error in TM equity funds: IMO this is a real challenge as the sub-sample used to improve tax efficiency may raise sampling error risk (missing those stocks which make the positive difference, or including those stocks which contribute to a negative difference). DFA’s TM funds in 2002 come to mind (e.g. DFA TM Mktwide value fund return = -27.2%, compared to DFA large value = -14.9%). Not saying its DFA’s fault – just that its difficult to do and trade-offs have to be made, and sometimes (given the nature of outliers often dominating the mean) they come back to bite. IMO this is a risk of TM funds (don’t get me wrong I still think TM value funds are excellent for taxable accounts but just that we should recognize the sample error risks are slightly higher).

Robert
.

[larryswedroe]

Robert
FYI- the big difference you note in the TM value and the large value of DFA was a purely bad luck random anomaly. Having said that the one year average difference they estimated might be as much as 6% but after 10 years was very close to zero. This was in presentation before they even launched the funds.

That year here is what occurred. The regular LV fund was not getting any cash flows as all the new money was going into the TM fund and even some money coming out of the fund to move to the TM fund. Now there were stocks (e.g.,telecom) that were dropping into the buy range for LV but DFA had no cash flow to buy---and they won't sell something just to buy something else (another advantage over pure index funds). On the other hand, the TM value funds had plenty of new cash and they were buying these new value stocks (today they would not because they added the negative momentum screen after that). So it was really that the LV fund OUTPERFORMED its benchmark that year. As comparison the Vanguard Value index fund fell 20.9%. The DFA LV fund got lucky and did not buy the falling telecom stocks. Random bad luck for the TM fund and random good luck for the regular fund. That is the price of not looking like an index. You have to accept random tracking error.

Now imagine if DFA were available to the public--and it had to explain those differences. That is one of the reasons they don't want to deal with the public directly.