David Swensen--Is he recommending long term bonds?

[NeedWisdom]

I'm in the process of reading Unconventional Success and was just wondering this question.

Also, could someone explain to me the difference between long and short and intermediate bonds. Why would one choose one over the other? I've heard inflation/deflation mentioned as reasons--I'm a bit confused about that. Why would the duration change how inflation impacts them?

Thank you all.

[CyberBob]

See this thread.
Swensen was emailed directly about that question and he responded that your bond duration should match that of the overall Treasury market (5.07 years at the time).

Bob

[jwillis77373]

There are a lot of good people here how can answer about Bonds better than I can.

But here's my point of view as a new Bond holder.

Bonds are basically loans that make you interest and hopefully return your principal when the deal is done.

They have ratings based on who you make the loan to.. the Government "Treasuries" are considered a safe bet that you will get your principal back.

The safer the loanee.. generally the lower the interest you can expect to make on the loan.

But there are lots of type of borrowers.

There is the Government, Corporations, City Governments (municipalities), and people who need money to buy a new home (Mortgagers).

Each has a credit rating and that kind of sets the interest rate you can ask for when making the loan.

Its also influenced by market factors, if there are a lot of people offering to loan money the interest rate can drop.

The length of the loan until they give back your prinicpal can also effect the interest rate. Generally the longer the loan the more interest you can ask for because you are risking locking up your money and passing up potentially better offers down the road. Shorter term loans are usually for less money.

But not always, when this flips they say the yield curve is inverted. Its a bad thing and generally preceeds things like recessions. It more or less says people are thinking in the future people will need to borrow less because the economy will be slowing down... so long term interest rates fall below short term and its "inverted".

Some people like to buy and sell Bonds direct with the borrowers, others prefer to buy and sell shares in a Bond fund. Dealing direct means you get all the interest and do all the paperwork. Dealing with a Bond fund means you pay a fee for someone else to handle the details.

Generally the interest earned from Bonds is taxed like income (there are exceptions) in the year earned so most people like to only hold them in retirement or non-taxable accounts.

There is so much to learn about Bonds I could never cover all the corner conditions and exceptions.. but those are the basics I think.

My recommedation is to read read read, and hold off on investing in Bonds until you feel somewhat educated in the topic. Or seek a professional who can help you learn about Bonds before investing in them.

In a strange sense if your young, Bonds are high risk, since they earn low interest they may not keep up with inflation and you just can't afford not to keep up with inflation.

- later

[craigr]

Why would the duration change how inflation impacts them?

Bonds are inversely related to interest rate changes. For instance when interest rates go up, existing bond prices go down and vice versa.

The reasoning is this: If I own a bond paying 5% a year and then the issuer comes out with a bond paying 6% a year my bond is worth less to an investor. I need to lower the price on my bond to make it attractive to new buyers who could simply buy the higher paying 6% bond.

Likewise, if the interest rate falls the existing bond price goes up. If the interest rate were to drop to 4% my 5% yielding bonds are worth more money because they pay higher interest. Therefore they are worth more money to buyers and they'll pay a higher price.

Now, let's look at how inflation/deflation plays into this.

Under an inflationary scenario the market interest rates are going up. Sometimes they are going way up. In the 1970's interest rates were in the mid-teens. If I purchased long-term bonds in 1970 yielding, say, 6% and bonds in 1980 were yielding 12%+ then my bonds are worth a lot less money. After all, if I can get a bond paying 12% why would I pay you the same amount of money for one paying only 6%? Inflation is nasty business and each dollar you receive is worth less so people demand higher interest rates to make up for the fall in the dollar.

However under deflation the scenario reverses. If I have bonds paying 6% and interest rates plunge (as they did in the 1930's to 1% or so) then my bonds are worth a lot more money. Under deflation, each dollar is worth more so a higher interest rate is paying you in more valuable dollars.

Short-term bonds are less affected by interest rate changes because they are coming to maturity so much faster than long-term bonds. If interest rates spike my short-term bonds may be maturing in a year or so and the principal will be returned. The spike in interest rate won't make them move much.

However long-term bonds are heavily affected by interest rate moves. If I have a bond that is maturing in 25-30 years then a move in interest rates will make it swing wildly up or down because they mature so far out in the future. Long-term bonds are more volatile because the market has to anticipate interest rate changes much further out.

Hope that makes sense.

Now as to holding short/long term bonds. Many people think you should just hold intermediate term bonds as they traditionally represented the "sweet spot" on risk/reward.

However there are good arguments for holding short and long term bonds to hedge against potential interest rate changes going either direction.

[NeedWisdom]

See this thread.
Swensen was emailed directly about that question and he responded that your bond duration should match that of the overall Treasury market (5.07 years at the time).

Bob

Um, ok. How do you do that with bond funds? Which Vanguard Treasury fund would best fit that description?

[SmallHi]

Intermediate Term Treasury. (Admiral shares are tough to beat!)

If you are going that route, I image TIPS would also make sense.

Or, you may just wanna stick with ST Treasury or ST Corporate which have very low risk and volatility, and structure your portfolio so that all of the volatility occurs in the asset classes with meaninful long term return premiums (Global Value and Small).

SH

[kimchee]

See this thread.
Swensen was emailed directly about that question and he responded that your bond duration should match that of the overall Treasury market (5.07 years at the time).

Bob

I asked this question before when rich pointed out this same post. Can rich or CyberBob clarify (or anyone else):

1) Whether the "market" referred to was the overall bond market or the overall Treasury market?
2) How does one determine the current duration of the market (from question 1)?

Thanks!

kimchee

[Robert T]

.
Swensen seems to be using intermediate treasuries in the Yale Endowment (certainly closer to intermdiate than long-term). This is from the last time I checked:

Code: Select all

Last 10 years annualized returns to June 30, 2006 

Yale endowment fixed income returns                        6.6 
Endowment benchmark [LB Treasury Index]                    6.2 

Vanguard Intermediate Treasury                             6.2 
Vanguard Long Term Treasury                                7.4

Source: http://www.yale.edu/investment....ent_06.pdf and Vanguard Semi-Annual Reports

Code: Select all

Here are some choices of intermediate treasuries

                   Benchmark                   Duration   Expense 
Vanguard [VFITX}   LB 5-10 US Treasury Index     5.2        0.26*
State Street [ITE] LB 1-10 US Treasury Index     3.5        0.13
IShares [IEI]      LB 3-7  US Treasury Index     4.0        0.15 

*Admiral shares have an expense ratio of 0.10
.

Robert
.

[rich]

I asked this question before when rich pointed out this same post. Can rich or CyberBob clarify (or anyone else):

1) Whether the "market" referred to was the overall bond market or the overall Treasury market?
2) How does one determine the current duration of the market (from question 1)?

Thanks!

kimchee

Hi Kimchee:

In Unconventional Success Swensen recommends US Treasuries only. In his "generic" portfolio he recommends 1.) 15% US TIPS and 2.) 15% nominal US Treasuries allocated by market duration. For the later, and per the email from Swensen, he recommended a barbell approach with a portion short term treasuries and a portion long term treasuries.

[kimchee]

A closer rereading of the post from Swenson indicates he is trying to set the duration based on the Treasury market and not the entire bond market. However, what is unanswered still is my second question. How does one obtain this basic piece of data:

>Durations as of August 31, 2006:
>Lehman US Treasury Index - 5.07

for, say, October 26, 2007? It seems basic but I can't find it by googling.

Robert, look at the comments by Tony G in the original post. It seems to me that a proper mix of ST and LT Treasury funds to achieve the market duration is what Swenson recommends.

kimchee

Thanks for the response rich, I didn't see your reply before I posted mine. Can you point to source for the duration data?

[rich]

Thanks for the response rich, I didn't see your reply before I posted mine. Can you point to source for the duration data?

I'm sorry but I'm just not sure where to find up to date information. Perhaps someone else knows?

[Robert T]

.
kimchee,

I saw the earlier note from Swensen but just wanted to check whether the Yale endowment fixed income was also aligned to the US treasury market index (as he suggested for individual investors) and not to a longer-term US treasury index. The comparsion shown above suggests that it is similar.

The Lehman Brothers website has estimates of the duration of several bond indices. See first 'numbers' column in this link - you will find the US treasury index in the list (for its composition see the factsheet). Interestingly the duration is listed as being the same as it was on August 31, 2006 in Swensen's note: 5.07.

Swensen suggested that "The easiest way to match the market would be to own some of the 1-3 Treasury (70%) and some of the 20+ Treasury (30%) so the weighted average duration matches the market's 5.07 years".

While the overall market duration has not changed the duration of the underlying funds has changed slightly: IShares 1-3 Treasury (SHY) was 1.67 it is how 1.66, and the IShares 20+ Treasury (TLT) was 13.00 and is now 13.52 so a 70:30 combination would now give a weighted average duration of 5.22 years. The combination would need to be adjusted slightly to get an exact match.

IMO the key element of fixed income performance against the US Treasury market index is to closely match the duration of the market index (which can be done several ways). The current duration of the Vanguard Intermediate Treasury fund comes fairly close at 5.2 years and as in the earlier post above the last 10 year returns of the fund to Dec 2006 matched almost exatly the returns of the US Treasury market index at 6.2%. As long as the Vanguard fund duration stays fairly close to the US Treasury Index I would expect returns to be similar. Swensen's suggestion is more precise but I prefer the simplicity of one fund rather than two.

Robert
.

[Valuethinker]

.

Swensen suggested that "The easiest way to match the market would be to own some of the 1-3 Treasury (70%) and some of the 20+ Treasury (30%) so the weighted average duration matches the market's 5.07 years".

.

When the Treasury stopped issuing greater than 10 year bonds, it led to a drop in the yields for very long term bonds. There are natural buyers for such bonds (insurance companies issuing annuities, and pension funds who make use of annuities eg for members with frozen benefits) with durations of 20 years +, and the Treasury cut off supply.

So 10 year plus bonds came to have lower yields than 10 year bonds, rather than higher, as theory would predict (longer maturity, more inflation risk).

I believe the Treasury is again issuing long term bonds, but long duration is still a relatively difficult thing to buy in the market*, and with increasing longevity, this poses a real problem for pension funds and insurers (who have very long term liabilities to pensioners and annuitants).

Accordingly I think the intermediate term Treasury bond (up to 10 years) is more likely to be 'fairly' priced than the long term bond. There is less risk of overvaluation due to supply/demand factors.

Even with large government deficits, the world is afflicted by a surplus of savings (one of the reason long term interest rates are so low despite rising concerns re inflation).

This effect may be very small, however.

* because corporate credit risk can change overnight, as we found with Worldcom, or any number of leveraged buyouts of blue chip companies, a 30 year AAA or AA corporate bond has much higher credit risk than a similar term government bond.

[larmewar]

FDIC/NCUA insured CD's with yields above 5% are still available. The risk would seem to be the same as Treasuries. Has Swensen taken a position on substituting CDs for some of the Treasuries?

Lar

[kimchee]

Robert T,

Sorry for the belated response. Thanks for the link, that was exactly what I was looking for. I will consider the single fund route, there does not seem to be too much of a difference although I'm not sure if that has been/will always be true.

kimchee

[Valuethinker]

FDIC/NCUA insured CD's with yields above 5% are still available. The risk would seem to be the same as Treasuries. Has Swensen taken a position on substituting CDs for some of the Treasuries?

Lar

Not that I know of.

But yes, the risk is (fractionally) more than Treasuries. Not enough to matter.

There is a tax position thing (exemption of interest from US treasuries for US state taxes?) but otherwise, I think the 2 are pretty much equivalent (within FDIC limits).

[larryswedroe]

Keep this in mind

Swensen I don't think recommends LT Bonds in isolation. But as part of the portfolio--the OVERALL portfolio. He likes the deflationary hedge risk they provide. But he also recommends including CCF in the portfolio--perhaps it is the idea of how the two mix together--LT bonds and CCF--they hedge each other's risks. This is the way I look at it and manage my own portfolio.

As I have mentioned the allocation to CCF allows me to comfortably extend the duration of my munis a bit longer than I would otherwise as the two risks hedge each other and I pick up some incremental yield along the way. And the incremental yield should be part of the "total return to CCF" as otherwise I would not be able to take that risk.

[Robert T]

.
Larry,

I have not read anywhere that Swensen recommends CCFs, I think he prefers investments in oil reserves, at least for the Yale Endowment – but understand your reasoning on the LT bond:CCF mix.

Here are extracts from various Yale endowment annual reports

On fixed income: “Fixed income plays an important role in the Endowment by providing a diversifying hedge against financial accidents or periods of unanticipated deflation. To achieve this hedge, the Endowment invests primarily in high-quality instruments backed by the full faith and credit of the U.S. government. Yale’s portfolio emphasizes non-callable securities, ensuring that the Endowment receives the hoped-for protection in periods of declining interest rates.” “The Lehman Brothers U.S. Treasury Index serves as the portfolio benchmark.”

“Sensible investors focus on the diversifying characteristics of long-term government bonds, holding only the amount necessary to protect portfolios against financial trauma. If portfolios include the minimum allocation necessary to provide insurance against catastrophe, investors free up assets to diversify into alternative asset classes, achieving volatility reduction without sacrificing return. A low allocation to high-quality fixed income reduces the costs associated with holding bonds during normal circumstances and periods of unanticipated inflation, the environments in which fixed income positions tend to impair portfolio performance. Tailoring the bond portfolio to emphasize fixed income’s essential diversifying characteristics increases expected benefits in time of crisis, while reducing the long-term costs of holding bonds.”

On real assets in general (real estate, oil and gas, and timber): “Real assets investments provide claims on future streams of inflation-sensitive income, supplying protection against unanticipated inflation and playing an important diversifying role in the portfolio. Real assets provide relative stability to the Endowment during periods of public market turmoil, at the price of an inability to keep pace during bull markets.”

“Yale prefers real assets investments that generate a current cash yield, whether from property rents, reserve production, or sustainable timber harvests. The presence of a substantial cash yield makes the total return on investment less sensitive to the length of the holding period and reduces valuation risk.”

On oil and gas in particular: “In the oil and gas and timber arenas, price changes in the underlying commodity strongly influence investment returns. Unfortunately, macroeconomic and political factors drive commodity prices, making them extremely difficult, if not impossible, to forecast. Rather than depend on uncertain future price increases, Yale’s natural resource investments must meet return targets in flat price environments. If commodity prices rise, Yale’s natural resource portfolio will generate handsome performance even as other parts of the Endowment suffer from the higher costs of basic materials and energy.”

“In the oil and gas portfolio, Yale emphasizes the low-risk purchase of high-quality proven reserves. In finding managers that evaluate and operate assets more efficiently than large oil and gas companies, Yale generates substantial returns without depending on higher-risk exploration strategies. A portion of the energy portfolio is allocated to private investments in which investment managers take meaningful stakes in energy exploration, production, or service companies, and attempt to influence the management and growth of the companies.”

Robert

PS:kimchee - pleased the link was useful.
.

[richard]

Swenson has different recommendations for individuals and for institutions. He has explicitly said that individuals should not follow his institutional recommendations.

Larry is right - he recommends LT treasuries as a deflationary "hedge" as part of a total portfolio.

I've never seen him recommend CCF for individuals and would like to see some evidence on this. I don't remember it from his book.

[grok87]

.
Larry,

I have not read anywhere that Swensen recommends CCFs, I think he prefers investments in oil reserves, at least for the Yale Endowment – but understand your reasoning on the LT bond:CCF mix.

Here are extracts from various Yale endowment annual reports

On fixed income: “Fixed income plays an important role in the Endowment by providing a diversifying hedge against financial accidents or periods of unanticipated deflation. To achieve this hedge, the Endowment invests primarily in high-quality instruments backed by the full faith and credit of the U.S. government. Yale’s portfolio emphasizes non-callable securities, ensuring that the Endowment receives the hoped-for protection in periods of declining interest rates.” “The Lehman Brothers U.S. Treasury Index serves as the portfolio benchmark.”

“Sensible investors focus on the diversifying characteristics of long-term government bonds, holding only the amount necessary to protect portfolios against financial trauma. If portfolios include the minimum allocation necessary to provide insurance against catastrophe, investors free up assets to diversify into alternative asset classes, achieving volatility reduction without sacrificing return. A low allocation to high-quality fixed income reduces the costs associated with holding bonds during normal circumstances and periods of unanticipated inflation, the environments in which fixed income positions tend to impair portfolio performance. Tailoring the bond portfolio to emphasize fixed income’s essential diversifying characteristics increases expected benefits in time of crisis, while reducing the long-term costs of holding bonds.”

On real assets in general (real estate, oil and gas, and timber): “Real assets investments provide claims on future streams of inflation-sensitive income, supplying protection against unanticipated inflation and playing an important diversifying role in the portfolio. Real assets provide relative stability to the Endowment during periods of public market turmoil, at the price of an inability to keep pace during bull markets.”

“Yale prefers real assets investments that generate a current cash yield, whether from property rents, reserve production, or sustainable timber harvests. The presence of a substantial cash yield makes the total return on investment less sensitive to the length of the holding period and reduces valuation risk.”

On oil and gas in particular: “In the oil and gas and timber arenas, price changes in the underlying commodity strongly influence investment returns. Unfortunately, macroeconomic and political factors drive commodity prices, making them extremely difficult, if not impossible, to forecast. Rather than depend on uncertain future price increases, Yale’s natural resource investments must meet return targets in flat price environments. If commodity prices rise, Yale’s natural resource portfolio will generate handsome performance even as other parts of the Endowment suffer from the higher costs of basic materials and energy.”

“In the oil and gas portfolio, Yale emphasizes the low-risk purchase of high-quality proven reserves. In finding managers that evaluate and operate assets more efficiently than large oil and gas companies, Yale generates substantial returns without depending on higher-risk exploration strategies. A portion of the energy portfolio is allocated to private investments in which investment managers take meaningful stakes in energy exploration, production, or service companies, and attempt to influence the management and growth of the companies.”

Robert

PS:kimchee - pleased the link was useful.
.

Robert,
Thanks for the above- very useful.
FWIW I haven't been able to bring myself to stash 20% of my portfolio in Real Estate as Swensen recommends. My current plan is 8% TIAA Real Estate, 4% MSUSX (morgan stanley us reit fund- only reit option in my 401k) and 4% PCRIX. I consider this a 16% allocation to "Real Assets".

cheers
grok

[larryswedroe]

Re Swensen--stand corrected should have said at least based on my memory anyway that he recommends allocation to commodities (real assets anyway) as hedge against inflation type risks.

[Robert T]

.
What I also find interesting about the Yale Endowment is:

“Yale’s portfolio typically favors value and small-capitalization stocks.”

“Emerging markets, with their rapidly growing economies, are particularly intriguing, causing Yale to target one-half of its foreign portfolio to developing countries.”

[both are extracts from the annual reports]

Yale's domestic and foreign equity investments will likely perform worse than a ‘market allocation’ in financial crises/flight to quality (given its value, small. EM tilt), which Swensen seems to protect against with ‘intermediate’ (US treasury index) rather than short-term treasuries (just my take).

Grok,

I agree, 20% is a high number for my tastes – but can understand why real assets are often included. FWIW – I don’t hold a separate allocation of real assets – just focus on the allocation to four risk factors that have historically been rewarded with premium returns and have had relatively low correlation (market, ‘value’, small, and term).

Robert
.

[SmallHi]

Robert,

Any thoughts on why DEF (default) hasn't carried a reliable return premium over the last 80 years?

SH

[Robert T]

.

Any thoughts on why DEF (default) hasn't carried a reliable return premium over the last 80 years?

I don’t have a good answer – but here are some thoughts.

You are right that default risk – measured as the returns on long-term corporate bonds minus long-term treasury bonds has not carried a reliable premium. It was about 0.4 percent from 1926-2005 and close to zero from 1972-2005 according to Ibbotson data.

Code: Select all

                                     Geometric   Arithmetic    Standard
                                       Mean         Mean      Deviation  
1926-2005 
Long-term corporate bonds              5.9           6.2           8.5
Long-term government bonds             5.5           5.8           9.2
LT Corporate-LT government (DEF)       0.4           0.4           

1972-2005
Long-term corporate bonds              9.0           9.5          10.9
Long-term government bonds             8.9           9.5          11.7
LT Corproate-LT government (DEF)       0.1           0.0

Source: Ibboston Yearbook

But surely if capital markets are functioning relatively efficiently then small default premiums should reflect relatively low risk of corporate bonds versus government bonds? Is this true? Possibly yes. Corporate defaults were higher prior to 1972 hence the higher premiums and if we consider the data below, one average, longer-term corporate bonds are, on average, higher quality than short term bonds.

Code: Select all

Data from 1974-1994

Company             Av. Company Size          Average Compnay      
Moody Rating        [Market cap $bn]           bond duration
                 
Aaa                      16,132                    7.13
Aa                        7,211                    6.85
A                         5,435                    6.61
Baa                       2,546                    6.26
Ba                        1,326                    5.60
B                           441                    5.13      
Caa                          96                    3.78
Ca                          120                    4.05

Source: Book-to-Market Equity, Size, and the Segementation of the Stock and Bond Markets – Roberto Gutierrez

Okay. If shorter term corporates have, on average, lower quality, have the default premiums been higher? Yes – if we consider the data below the ‘default premium’ on 1-3 year bonds was 0.64% while on longer term bonds it was 0.01 percent.

Code: Select all

Data from 1985-2002

1-3 yr treasuries                   7.30
1-3 yr AAA/AA Corp                  7.94
Difference (default premium)        0.64

7-10 yr treasuries                  9.59
7-10 yr AAA/AA Corp                 9.60
Difference (default premium)        0.01

Source: Which risks have been best rewarded? Ilmanen, Byrne, Gunasekera, Minikin.

But shouldn’t bonds with the same ratings have the same risks for a given duration? IMO may be not. If on average the companies issuing 1-3yr corporates are smaller than those issuing 7-10 yr corporates, perhaps there are size risks the market has priced into the short term default premiums that are not pick-up by the ratings given by the rating agencies.

Perhaps some gaps in the argument – but just my take on an answer.

Robert
.

[SmallHi]

Interesting thoughts, Robert.

I haven't read the papers you linked to, and certainly need to.

Interesting that, even on the very short end (1YR), DFAs One Year Fixed fund hasn't picked up much of the default premium after accounting for expenses:

1983-2007

DFA 1YR Fixed = +6.15%
1YR T-Note = +6.05%

If we assume weighted average expenses of 0.25% for the entire period (guessing here), it looks as though only about 1/3% annually can be attributed to default risk. Of course, that "alpha" could just as easily been due to variable maturity mangement, as, over this period, there was no default premium on longer term bonds (-0.65% annually). Using Lehman's intermediate indexes, OTOH, we see about +0.8% annually of default premium.

appreciate the thoughts.

SH

[Robert T]

.
Another study tries to explain the difference in between the yield to maturity on a zero-coupon corporate bond (corporate spot rate) and the yield to maturity on a zero-coupon government bond of the same maturity (government spot rate).

For a 10 A-rated corporate bond – the difference in the ‘corporate spot rate’ and ‘government spot rate’ was explained as follows (assuming that I have interpreted the results correctly):

Code: Select all

Expected loss from default                               17.8%
Less favorable tax treatment than government bonds       38.1%
Factors that explain equity risk premiums                37.5%
Unexplained                                               6.6%
Total difference                                        100.0%

Source: Explaining the Rate Spread on Corporate Bonds. Elton, Gruber, Agrawal, and Mann. JoF. 2001

Robert
.

[SmallHi]

If on average the companies issuing 1-3yr corporates are smaller than those issuing 7-10 yr corporates, perhaps there are size risks the market has priced into the short term default premiums that are not pick-up by the ratings given by the rating agencies.

It could be, but, based on a very limited sample of 20 years, I don't find that 1-3YR Corporate Index performes worse than a 1-3YR Treasury index during periods of:

a) negative market premiums
b) negative HmL or SmB premiums
c) negative SV periods

...either quarterly or annually since 1986.

This is consistent with the research that indicates moving out from 1month t-bills to 1-3 year bonds and taking on AAA/AA credit risk is one of the largest risk adjusted premiums available, and the presistence of this premium is much more dependable than most other dimensions we target.

Longer term default risk does, however, have a noticeably positive correlation with SmB, so its something to consider.

(ps -- I had read the Ilamanen paper, mistakenly said I didn't...but haven't read the Gutierrez study)

Are you aware of any studies that discuss the appropriatness of longer term bonds in a balanced portfolio (especially a S/V tilted portfolio -- obviously stretching it there)?

Thanks

SH

[lucky7]

When Farrell lists the Yale University Lazy Portfolio, it included Vanguard Long Term US Treasury. But don't recall Swensen specified in his book, "Unconventional Success". Was curious about the same issue, so thanks for the posting.

Bob