Why no foreign bond funds?
Why no foreign bond funds?
If one has significant allocation in foreign equities (e.g. 35% of equity allocation within tm total intern.) should not for a portion of the overall fixed income allocation, there also be a foreign bond/bond fund component (developed nations, high grade, unhedged)? Or more specifically, how do correlations match up among these four components of broad US and intern. equities, and U.S and intern. bonds?
Time for a decision for me nearing as defined benefit monies soon to be disbursed. Appreciate any feedback.
Bob
Time for a decision for me nearing as defined benefit monies soon to be disbursed. Appreciate any feedback.
Bob
Scotty, beam me up.
For me, it's a tax issue. Hold foreign bonds in a taxable account and you take a significant tax hit each year, but at least you get credit for the taxes those bonds paid to other countries. Hold them in tax-deferred and you lose the tax credit, but get long-term deferral.
Thus, I stick to domestic bonds (short term corp primarily) in tax-deferred and foreign equities in taxable.
Thus, I stick to domestic bonds (short term corp primarily) in tax-deferred and foreign equities in taxable.
A hedged portfolio of Global bonds has higher expected returns (due to diversification benefit of multiple country bond markets) and slightly lower risks that a similar portfolio of US Treasury bonds. With Global Bonds, you miss out on a bit of the "flight to quality" benefit of US Treasuries, but you aren't always going to rebalance if the "ftq" isn't enough to trigger a rebalancing action -- so part of that is just cosmetic.
It is unfortunate that there are no really good hedged low cost short term Bond funds that I am aware of other than DFA 2 and 5 Year global, but they are restricted. Best to just stick with some combo of Vanguard ST Investment Grade or ST Federal/Treasury and not sweat the few basis points you'll miss from global bond diversification.
SH
It is unfortunate that there are no really good hedged low cost short term Bond funds that I am aware of other than DFA 2 and 5 Year global, but they are restricted. Best to just stick with some combo of Vanguard ST Investment Grade or ST Federal/Treasury and not sweat the few basis points you'll miss from global bond diversification.
SH
You're so right: Payden Global Short Bond Fund Class R (PYGSX) is the only one I can find and it's 0.66 ER at Vanguard and its not NTF. Not a great option considering the expense.SmallHi wrote:It is unfortunate that there are no really good hedged low cost short term Bond funds that I am aware of other than DFA 2 and 5 Year global, but they are restricted. Best to just stick with some combo of Vanguard ST Investment Grade or ST Federal/Treasury and not sweat the few basis points you'll miss from global bond diversification.
Paul
foreign bond funds are controversial
There have been many posts on foreign bond funds on this forum. A number of studies advocate unhedged foreign bonds for diversification. Most of these studies however assume a small to modest allocation to foreign equities (e.g., 10-20% of overall portfolio). Swedroe and others argue that you are better off putting all your currency risk in foreign equities and buying hedged foreign bonds, if you can find them at the right price. Given that you can buy treasuries, agency bonds and TIPS for free, it doesn't take much of an ER to wipe out any diversification benefit. If you are with Vanguard (VBS) and have more than $25 K to invest, you can get access to PIMCO institutional foreign bond funds, which cost around 50 bps I believe.
Kevin
Kevin
-
- Posts: 48944
- Joined: Fri May 11, 2007 11:07 am
Although the best forecast of the returns of a bond is its current yield to maturity.SmallHi wrote:A hedged portfolio of Global bonds has higher expected returns (due to diversification benefit of multiple country bond markets) and slightly lower risks that a similar portfolio of US Treasury bonds.
SH
And for most developed markets, YTMs on 10 year bonds are actually lower than US treasury YTMs.
-
- Posts: 74
- Joined: Wed Oct 17, 2007 9:53 am
For non-dollar denominated bonds, the best forecast would require both the bond's YTM in its native currency and the appropriate forward exchange rate.Valuethinker wrote:Although the best forecast of the returns of a bond is its current yield to maturity.
And for most developed markets, YTMs on 10 year bonds are actually lower than US treasury YTMs.
-
- Posts: 48944
- Joined: Fri May 11, 2007 11:07 am
Thank you. Excellent point.old_dominion wrote:For non-dollar denominated bonds, the best forecast would require both the bond's YTM in its native currency and the appropriate forward exchange rate.Valuethinker wrote:Although the best forecast of the returns of a bond is its current yield to maturity.
And for most developed markets, YTMs on 10 year bonds are actually lower than US treasury YTMs.
However:
- in a currency hedged fund, isn't that already in the price (ie the NAV of the Fund)?
- in an unhedged currency fund, we can't be sure of what the currency outcome is?
(ie we are separating return into 2 components: the currency return and the interest return. So the interest return is roughly fixed by the YTM, and the currency return is separate?)
Part of my fixed income holding are in FNMIX. Performance this year has lagged but has been a good holding longer term.
'The fund normally invests at least 80% of assets in debt securities issued by companies and governments in emerging markets. It expects to emphasize investment in Latin America, and, to a lesser extent, Asia, Africa, and emerging European nations. The debt securities held by the fund may be below investment-grade; some securities may be in default.'
'The fund normally invests at least 80% of assets in debt securities issued by companies and governments in emerging markets. It expects to emphasize investment in Latin America, and, to a lesser extent, Asia, Africa, and emerging European nations. The debt securities held by the fund may be below investment-grade; some securities may be in default.'
Appreciate the replies.
Appreciate all the thoughtful feedback. However, please note, my thinking for diversification was not centered upon currency risk, but rather simply upon adding another non-correlating asset group (in this case within fixed income), with the same projected long term return as high grade domestic bonds. This assumes (and I surely do not know) that not all of the non-correlation is due to the currency risk, which if true would suggest a diversification benefit for including foreign bonds/funds unless offset by higher costs.
Bob
Bob
Scotty, beam me up.
EM bonds are mostly rated as junk. On top of that they are highly corelated with EM equities (~0.85). If there is any shakiness in the EM world equities can drop and the default risk on the bonds can rise.cdelena wrote:Part of my fixed income holding are in FNMIX. Performance this year has lagged but has been a good holding longer term.
'The fund normally invests at least 80% of assets in debt securities issued by companies and governments in emerging markets. It expects to emphasize investment in Latin America, and, to a lesser extent, Asia, Africa, and emerging European nations. The debt securities held by the fund may be below investment-grade; some securities may be in default.'
Paul
I have been reading about emerging market debt which, at the moment for us Canadians, mostly means closed-end funds offered by various institutions, although by the end of the year there will apparently be at least 2 ETFs tracking a Deutsch Bank index and a JP Morgan index, respectively.
My question is not about the merits of investing in emerging markets debt per se, but rather what would be the advantages/disadvantages of having the debt denominated in USD, or not? Its often mentioned in what I percieve to be a reassuring way that the bonds are denominated in USD, but does it actually influence risk?
My question is not about the merits of investing in emerging markets debt per se, but rather what would be the advantages/disadvantages of having the debt denominated in USD, or not? Its often mentioned in what I percieve to be a reassuring way that the bonds are denominated in USD, but does it actually influence risk?
-
- Posts: 48944
- Joined: Fri May 11, 2007 11:07 am
Interesting question. Brief thoughts.drejmd wrote:I have been reading about emerging market debt which, at the moment for us Canadians, mostly means closed-end funds offered by various institutions, although by the end of the year there will apparently be at least 2 ETFs tracking a Deutsch Bank index and a JP Morgan index, respectively.
My question is not about the merits of investing in emerging markets debt per se, but rather what would be the advantages/disadvantages of having the debt denominated in USD, or not? Its often mentioned in what I percieve to be a reassuring way that the bonds are denominated in USD, but does it actually influence risk?
Lower currency risk in principle. Also lower diversification.
However when countries default, they suspend US dollar debt payments (Argentina 2002, Ecuador now?)
Obviously US/Can$ a factor. Note many Canadian companies borrow in US dollars (the same principle at work).
Most of the risk in EM is credit risk, of the countries. Which you don't escape by being denominated in dollars. Actually the big money in EM debt the last few years has been investing in the *local* debt say paying in Brazilian Cruzeiros, which paid a much higher interest rate. (risk=> return)
lucky 7
Lucky 7,
There is a diversification benefit from hedged foreign bond exposure. Central banks are not always moving rates in the same direction. Nonetheless, I question whether this diversification is worth 50-65 bps in expenses. Have a look at the DFA website and review their foreign bond fund product for more insights.
Kevin
There is a diversification benefit from hedged foreign bond exposure. Central banks are not always moving rates in the same direction. Nonetheless, I question whether this diversification is worth 50-65 bps in expenses. Have a look at the DFA website and review their foreign bond fund product for more insights.
Kevin
Why no foreign bonds?
.
The global allocation of bonds suggests the efficient allocation is close to an even split between US and non-US obligations – however I currently exclude foreign bonds for the following reasons (last point not a primary reason but wanted to added it):
The table below shows portfolio returns when the listed bonds comprise 25% of a portfolio with the 75% equity allocation tilted to value and small cap stocks (0.2 size and 0.4 value loading). The equity portfolio is also split 50:50 US:Non-US but the value and size tilt is larger on the US side (my equtiy allocation benchmark).
Robert
.
Edited to add an unhedged foreign bond fund to the analysis.
.
The global allocation of bonds suggests the efficient allocation is close to an even split between US and non-US obligations – however I currently exclude foreign bonds for the following reasons (last point not a primary reason but wanted to added it):
- 1) Unwanted currency fluctuations:Currency fluctuations reflected in unhedged non-US bonds undermines the stabilizing role of fixed income in a portfolio. Over the last 15 year (time period of available data) Vanguard’s Intermediate US Treasury fund had about the same annualized return as the American Century International Bond fund (an unhedged foreign bond fund), 6.48% versus 6.40%, while the standard deviation of the unhedged foreign bond fund was almost 70% larger, 6.83% versus 11.44%. At a portfolio level this shows up as higher portfolio volatility but no compensation in higher return (unlike when volatility is added through equities). My preference is to take currency risk (and the associated volatility) on the equity side and leave fixed income for safety and stability.
2) Added costs for same expected return: Hedging currency risk adds complexity and raises costs, while long-term expected returns of US and Non-US hedged bonds are the same. Here is an extract from Swensen’s Unconventional Success - “Because monetary conditions differ from country to country, the two bonds [US and foreign] would likely promise different interest rates. An investor might expect that different interest rates and different economic conditions in different countries would lead to different investment results. If, however, the investor hedges each of the foreign bond’s cash flows by selling sufficient foreign currency in the forward markets to match the anticipated receipt of interest and principal payments, then the US dollar cash flows of the dollar-denominated bond match exactly the US dollar cash flows of the foreign-currency denominated bond hedged into US dollars.” [forward foreign exchange contracts derive from interest rate differentials]. There are no low cost foreign bond ETFs/mutual funds currently available.
3) Lower protection: Foreign bonds provide less downside risk protection than domestic bonds – the latter provides a clear and direct insurance against domestic financial crises and deflation. If a portfolio is tilted to US stocks (or has greater exposure to US value and small cap stocks) then domestic bonds (particularly US treasuries) provide the safest form of insurance against domestic deflation and financial crises. US treasuries also currently provide the greatest protection from global financial crises as the preferred safe haven in global flight to quality episodes.
4) No tax benefits: For those with taxable accounts – non-US bonds have no US tax benefits – such as the US state tax exemptions on US treasuries, and federal and state tax exemptions on most municipal bonds.
- 1. While the DFA 5 year global fund has a lower standard deviation than the Vanguard US intermediate treasury fund, and slightly lower returns – its characteristics over the 1992-2006 period were closely matched by a 55:45 US short:intermediate term bond index combination [see first below]. This suggests that it may have been the inclusion of corporate bonds in the DFA 5-yr global fund (40% at last check) that improved its mean-variance rather than (or at least as much as) inclusion of foreign bonds (following Swenson earlier comments in point 2 above).
2. Despite the Vanguard intermediate US treasuries fund having higher individual standard deviation than the DFA 5-yr global bond fund it had a greater mean-variance impact when included in a portfolio (point 3 above) [see second table below]. In addition, default risk in corporate bonds are positively correlated with equity returns (lowering their relative diversifcation impact).
3. The American Century International bond fund (an unhedged fund) had a standard deviation about double the other funds (its annual return ranged from –10.6% in 1999 to +24% in 1995). When added to a portfolio as half the bond allocation it raised portfolio volatility but not returns (see second table below).
Code: Select all
Individual characteristics, 1992-2006
Annualized Standard
Return Deviation
DFA 5 year Global 6.25 4.90
55:45 US Short:Intermediate bond index 6.24 4.91
Vanguard Intermediate Treasury 6.48 6.83
Unhedged foreign bond [BEGBX] 6.40 11.44
*The short and intermediate bond returns used are those of the underlying indices tracked the two Vanguard funds which extend back to 1992.
Code: Select all
Portfolio returns, 1992-2006
when the following are added
Annualized Standard
Return Deviation
Vanguard Intermediate Treasury 12.73 12.94
55:45 US Short:Intermediate Bond Index 12.71 13.17
DFA 5 year Global 12.71 13.23
50:50 Int. Trsy:Unhedged foreign bond [BEGBX] 12.73 13.48
*The short and intermediate bond returns used are those of the underlying indices tracked the two Vanguard funds which extend back to 1992.
.
Edited to add an unhedged foreign bond fund to the analysis.
.
Last edited by Robert T on Mon Oct 22, 2007 8:53 pm, edited 1 time in total.
Krelihan
Krelihan,
If I do not mind currency fluctuations which should cancel out over the long term, why not avoid the added expenses of hedging? The diversification benefit of adding this low/non-correlating asset class within my fixed income allocation is my primary focus. Appreciate any further thoughts, thanks.
Bob
If I do not mind currency fluctuations which should cancel out over the long term, why not avoid the added expenses of hedging? The diversification benefit of adding this low/non-correlating asset class within my fixed income allocation is my primary focus. Appreciate any further thoughts, thanks.
Bob
Scotty, beam me up.
Re: Why no foreign bonds?
A couple of thoughts as a follow up to Robert's very valid comments...
The efficiency of Intermediate Term bonds from the early 1990s through 2006 (and counting ) is unique to the entire last 40 or 50 years.
Since 1964, when the equity premium is negative on an annual basis, LT Bonds tend to underperform T-Bills by about (1.4%) annually. There is a small premium for extending maturities to about 2 years (+1.6% advantage for extending from 1 month up till 2 years), but not beyond that.
Also (and more concerning to me), when Value stocks underperform growth stocks, the term premium is negative by about 2% on average (but the premium from extending from 1 month up till 2 years is about +1.0%, but not beyond that).
From 1973-1991, a 75/25 portfolio that includes only 2YR Treasuries instead of Intermediate Treasuries for the bond component had a standard deviation that was 0.4 lower and an annual return that was +0.2% higher -- completely reversing the trend of the last few years.
A few thoughts on the struggles of Short Term Bonds:
Because 2YR T-Notes are a better inflation hedge (annual correlation with CPI of over +0.3 vs negative correlation for Int'd bonds and inflation), I would make the case that you shouldn't expect outstanding returns from shorter term bonds (2YR Treasury or DFA 5YR Global -- which is really a 2YR Fund) recently because inflation has been far from an issue.
Also, ST Bonds have been a very good HmL hedge, and, since 1992, US HmL has compounded at 6.4% a year, so this isn't really the type of period you'd expect to see shorter term nominal bonds do well anyway. (except the late 90s, as the tech bubble raged and HmL suffered, again, TERM turned negative and shorter term bonds provided a better hedge -- especially global short term bonds)
On DFA Global Bond Fund
I find a few things from the above research interesting:
a) it took 50% more TERM risk (with 50/50 STIndex/IntIndex) to approximate the risk/returns of DFA 5YR Global. Even though it didn't show up in the volatility, the 50/50 portfolio does have exposure to greater term risk -- it just didn't "show up" in the period as it would have in the 60s, 70s, or 80s.
b) I am just not sure how much of the DFGBX risk reduction came from taking on credit risk (and reaping some diversification benefits?). For one, going back to 1991, an index of hedged 1-5 YR Global Gov't bonds had less risk and slightly higher returns than a portfolio of only US 1-5YR Gov't bonds...so there is some evidence that the low correlation of global bond markets does lower the risk of a hedged global bond portfolio relative to domestic bond portfolio of the same duration and credit risk.
Although regressions are far from conclusive, DFGBX didn't act like a fund that took on a lot of credit risk (low DEFAULT sensitivity and standard errors.) And the credit it did hold was probably AAA rated or highest quality commercial paper.
Furthermore, DFGBX has gone many periods where it had very little if any corporate holdings. There were quite a few periods where DFGBX appeared to be over 80% US/Foreign bonds. Howerver, there have been periods of 50-70% corporate holdings (when the spreads were huge, like 2002).
All in all, I don't think Global bonds are crucial to anyone's success. And ST Global bonds, while far from disappointing the last few years, aren't going to be great in all periods. Usually, these periods will also coincide with significant HmL premiums, so its not that bad.
Int'd Treasuries do offer flight to quality hedges (usually this is cosmetic, however, as most FTQ scenerios don't actually trigger stock/bond rebalancing unless you have very tight bands), and will do well when inflation is less than expected, or very low in general.
Some may find the later preferable, while others will feel that, much like equities, it doesn't make sense to confine all fixed income holdings to just one country.
SH
The efficiency of Intermediate Term bonds from the early 1990s through 2006 (and counting ) is unique to the entire last 40 or 50 years.
Since 1964, when the equity premium is negative on an annual basis, LT Bonds tend to underperform T-Bills by about (1.4%) annually. There is a small premium for extending maturities to about 2 years (+1.6% advantage for extending from 1 month up till 2 years), but not beyond that.
Also (and more concerning to me), when Value stocks underperform growth stocks, the term premium is negative by about 2% on average (but the premium from extending from 1 month up till 2 years is about +1.0%, but not beyond that).
From 1973-1991, a 75/25 portfolio that includes only 2YR Treasuries instead of Intermediate Treasuries for the bond component had a standard deviation that was 0.4 lower and an annual return that was +0.2% higher -- completely reversing the trend of the last few years.
A few thoughts on the struggles of Short Term Bonds:
Because 2YR T-Notes are a better inflation hedge (annual correlation with CPI of over +0.3 vs negative correlation for Int'd bonds and inflation), I would make the case that you shouldn't expect outstanding returns from shorter term bonds (2YR Treasury or DFA 5YR Global -- which is really a 2YR Fund) recently because inflation has been far from an issue.
Also, ST Bonds have been a very good HmL hedge, and, since 1992, US HmL has compounded at 6.4% a year, so this isn't really the type of period you'd expect to see shorter term nominal bonds do well anyway. (except the late 90s, as the tech bubble raged and HmL suffered, again, TERM turned negative and shorter term bonds provided a better hedge -- especially global short term bonds)
On DFA Global Bond Fund
I find a few things from the above research interesting:
a) it took 50% more TERM risk (with 50/50 STIndex/IntIndex) to approximate the risk/returns of DFA 5YR Global. Even though it didn't show up in the volatility, the 50/50 portfolio does have exposure to greater term risk -- it just didn't "show up" in the period as it would have in the 60s, 70s, or 80s.
b) I am just not sure how much of the DFGBX risk reduction came from taking on credit risk (and reaping some diversification benefits?). For one, going back to 1991, an index of hedged 1-5 YR Global Gov't bonds had less risk and slightly higher returns than a portfolio of only US 1-5YR Gov't bonds...so there is some evidence that the low correlation of global bond markets does lower the risk of a hedged global bond portfolio relative to domestic bond portfolio of the same duration and credit risk.
Although regressions are far from conclusive, DFGBX didn't act like a fund that took on a lot of credit risk (low DEFAULT sensitivity and standard errors.) And the credit it did hold was probably AAA rated or highest quality commercial paper.
Furthermore, DFGBX has gone many periods where it had very little if any corporate holdings. There were quite a few periods where DFGBX appeared to be over 80% US/Foreign bonds. Howerver, there have been periods of 50-70% corporate holdings (when the spreads were huge, like 2002).
All in all, I don't think Global bonds are crucial to anyone's success. And ST Global bonds, while far from disappointing the last few years, aren't going to be great in all periods. Usually, these periods will also coincide with significant HmL premiums, so its not that bad.
Int'd Treasuries do offer flight to quality hedges (usually this is cosmetic, however, as most FTQ scenerios don't actually trigger stock/bond rebalancing unless you have very tight bands), and will do well when inflation is less than expected, or very low in general.
Some may find the later preferable, while others will feel that, much like equities, it doesn't make sense to confine all fixed income holdings to just one country.
SH
.
SH,
I agree – results often depend on time period analyzed.
Using the longest TERM data set I have from 1927-2005:
- when the average annual equity premium was negative the average annual term premium was +0.64%
- when the average annual value premium was negative the average annual term premium was +0.45%
- when the average annual returns of a US equity benchmark with 0.2 size and 0.4 value loading was negative the average annual term premium was +1.73%
Source of primary data: Fama-French Benchmark Factors, and Ibboston yearbook (to derive the term premium).
Robert
.
SH,
I agree – results often depend on time period analyzed.
Using the longest TERM data set I have from 1927-2005:
- when the average annual equity premium was negative the average annual term premium was +0.64%
- when the average annual value premium was negative the average annual term premium was +0.45%
- when the average annual returns of a US equity benchmark with 0.2 size and 0.4 value loading was negative the average annual term premium was +1.73%
Source of primary data: Fama-French Benchmark Factors, and Ibboston yearbook (to derive the term premium).
Robert
.
Robert,
All very good points. A couple of thoughts:
On the surface...
Secondly, because there is a very reliable "1YR to 2YR term premium" moving from 1 month to 1 or 2 years -- even when TERM is negative...I wonder if you could have picked up the entire TERM premium over the periods you mention with just 1 or 2 year bonds? For example, since 1963, even when TERM was negative, 1YR T-Notes and 2YR T-Notes still outpaced 1mo T-bills.
Looking closer at real world portfolios instead of abstract risk premiums, I find going back to 1927 that when the US Vector fund (DFVEX's simulated results are as close as I can muster to your 0.2/0.4 example) registered a negative year (20 out of 79 periods), TERM was negative -0.14%, and positive +2.65% when Vector had an annual gain. Furthermore, I find US Vector had an annual +0.2 correlation with TERM over this period.
(and I found similar return patters and correlations for the Core 2 strategy simulated results)
I really don't think the evidence is conclusive one way or another, to be honest. I will probably continue to advocate ST Nominal bonds with a heavy value tilt because I am comfortable with it, and based on my research I tend to lean that way...but I don't disagree with your position.
SH
All very good points. A couple of thoughts:
On the surface...
my first thought is...those aren't very big term premiums! :lol:- when the average annual equity premium was negative the average annual term premium was +0.64%
- when the average annual value premium was negative the average annual term premium was +0.45%
Secondly, because there is a very reliable "1YR to 2YR term premium" moving from 1 month to 1 or 2 years -- even when TERM is negative...I wonder if you could have picked up the entire TERM premium over the periods you mention with just 1 or 2 year bonds? For example, since 1963, even when TERM was negative, 1YR T-Notes and 2YR T-Notes still outpaced 1mo T-bills.
Looking closer at real world portfolios instead of abstract risk premiums, I find going back to 1927 that when the US Vector fund (DFVEX's simulated results are as close as I can muster to your 0.2/0.4 example) registered a negative year (20 out of 79 periods), TERM was negative -0.14%, and positive +2.65% when Vector had an annual gain. Furthermore, I find US Vector had an annual +0.2 correlation with TERM over this period.
(and I found similar return patters and correlations for the Core 2 strategy simulated results)
I really don't think the evidence is conclusive one way or another, to be honest. I will probably continue to advocate ST Nominal bonds with a heavy value tilt because I am comfortable with it, and based on my research I tend to lean that way...but I don't disagree with your position.
SH
Thinking more about the two periods we are referring to...one word comes to mind: inflation.
From 1927-1962, you had an inflation level of 1.5%, a much smaller HmL premium of +3.3%, and behavior from TERM that led to higher returns when HmL was negative than when it was positive.
From 1963-2006, you had a much higher level of inflation -- +4.4%, and a much higher level of HmL (+5.3%). During this period, term was strongly negative when HmL was negative. Term averaged -2.0% a year when HmL lost value, and averaged +4.1% when HmL was positive.
Now, if only my spreadsheet will tell me if the next 30 years will be like the former or the latter!
SH
From 1927-1962, you had an inflation level of 1.5%, a much smaller HmL premium of +3.3%, and behavior from TERM that led to higher returns when HmL was negative than when it was positive.
From 1963-2006, you had a much higher level of inflation -- +4.4%, and a much higher level of HmL (+5.3%). During this period, term was strongly negative when HmL was negative. Term averaged -2.0% a year when HmL lost value, and averaged +4.1% when HmL was positive.
Now, if only my spreadsheet will tell me if the next 30 years will be like the former or the latter!
SH
.
SH,
I agree that it's difficult to unpack to relationship between the equity premium, value premium, and term premium. The numbers are not as conclusive as other relationships/correlations. As you point out inflation impacts both value companies (in a positive way by reducing real cost of debt), and longer maturity debt (in an negative way by reducing the real income stream). Vice versa for deflation. However the relationship is complex and non-linear due to the economic cycle risk of value stocks. Larry’s book – The Successful Investor Today discusses this non-linearity in a section on the value premium, inflation, and portfolio risk and concludes as follows (pg. 238/9):
“If you are going to hold a high allocation to value stocks, you should also consider holding longer-term fixed income instrument. The recession/deflation/flight to quality risks of value stocks have negative correlation with the risk of longer-term fixed income instruments (which generally perform well in these environments. And since the evidence suggests that value stocks provide greater returns than growth stocks during inflationary periods, they offer greater protection than growth stocks against the inflation risk inherent in longer-term bonds. Thus holding a value tilt not only provides greater expected returns on your equity holdings, but it also allows you to take more interest-rate risk and thus potentially earn greater returns on your fixed income investments.”
Swensen also advocates longer-term bonds but mainly for protection in financial crises when most needed – “By holding portfolios of high-quality, long-term, non-callable instruments, investors emphasize the attributes of bonds that provide the greatest protection in times of financial crisis”. He discusses these attributed in some detail and some of this downside risk protection was highlighted in an earlier thread
Who knows what the future will hold – but as you say comfort level with investment choice is important. For me I am currently comfortable using intermediate term bonds for the two reasons provided above which are linked to the role I want fixed income to play in a portfolio.
Robert
.
SH,
I agree that it's difficult to unpack to relationship between the equity premium, value premium, and term premium. The numbers are not as conclusive as other relationships/correlations. As you point out inflation impacts both value companies (in a positive way by reducing real cost of debt), and longer maturity debt (in an negative way by reducing the real income stream). Vice versa for deflation. However the relationship is complex and non-linear due to the economic cycle risk of value stocks. Larry’s book – The Successful Investor Today discusses this non-linearity in a section on the value premium, inflation, and portfolio risk and concludes as follows (pg. 238/9):
“If you are going to hold a high allocation to value stocks, you should also consider holding longer-term fixed income instrument. The recession/deflation/flight to quality risks of value stocks have negative correlation with the risk of longer-term fixed income instruments (which generally perform well in these environments. And since the evidence suggests that value stocks provide greater returns than growth stocks during inflationary periods, they offer greater protection than growth stocks against the inflation risk inherent in longer-term bonds. Thus holding a value tilt not only provides greater expected returns on your equity holdings, but it also allows you to take more interest-rate risk and thus potentially earn greater returns on your fixed income investments.”
Swensen also advocates longer-term bonds but mainly for protection in financial crises when most needed – “By holding portfolios of high-quality, long-term, non-callable instruments, investors emphasize the attributes of bonds that provide the greatest protection in times of financial crisis”. He discusses these attributed in some detail and some of this downside risk protection was highlighted in an earlier thread
Who knows what the future will hold – but as you say comfort level with investment choice is important. For me I am currently comfortable using intermediate term bonds for the two reasons provided above which are linked to the role I want fixed income to play in a portfolio.
Robert
.
Robert,
I have given each of these thoughts/comments a lot of consideration lately, and I have come up with the following opinions. They may just be independent, or they may be factually incorrect, but its how I see it...
Theoretically, a 3YR TIP would have performed very well during this period had it existed. I fear the price declines on a 10YR TIP from rising interest rates, however, would have offset much of the inflation credit. So I am uncertain how helpful longer term TIPS would have been.
Yes, HmL and ST bonds will both perform poorly during deflationary periods...but a lot less worse on a "real" basis than on a "nominal basis". Also, you have to put deflation in perspective. Less than 3% of quarters over the last 40 years have been deflationary, and not one single year since the early 50s. I am fine not hedging deflation risk. Others may feel differently.
Treasuries do better during severe declines, giving you more to "rebalance with", but nothing is stopping you from a bit of dynamic rebalancing or overbalancing out of slightly less appreciated (non treasury) bond positions into stocks.
--I came away from this thinking that the flight to quality nature of treasuries is, for the most part cosmetic...it usually doesn't provide you any real opportunity to rebalance (given 20% or 25% bands) unless you do so daily like Yale. There is something to be said for a bond holding that "spikes" when stocks drop, but if you don't actually profit from it from rebalancing, it is more cosmetic and a "feel good" trait.
--I guess I could make the same arguement that ST bonds have a "feel good" trait in that they almost never fluctuate that much, so almost all of the volatility is on the equity side, where you are expecting to be paid. So volatility is directly in line with premium compensation, which is a nice feature.
Finally, over some periods of HmL decline, Int/LT bonds do very well in absolute and relative terms. Sometimes they do very poorly in absolute and relative terms. I would argue, outside of 1% interest rate scenerios, ST bonds almost always do reasonably well. Maybe not always as well as Int'd/LT bonds, but thats a relative thing. The risk of very poor absolute ST bond performance seems small. ST Bonds seem to me like a "safer, more dependable bet" given the limited (and sometimes conflicting) data we are trying to draw inferences from.
Again, just my thoughts. Some/many/most may not see it this way. And, as you may know, I don't own bonds anyway, so my thoughts/research are simply from a theoretical standpoint.
SH
I have given each of these thoughts/comments a lot of consideration lately, and I have come up with the following opinions. They may just be independent, or they may be factually incorrect, but its how I see it...
Yes, value stocks do well during inflationary times until the Fed starts cutting rates. In the late 70s/early 80s, Volker took a much tougher monetary stance on inflation via interest rates, and LT Bonds (and HmL) plummeted in unison. 78-80 was the 4th worst 36 month period for HmL on record. The best hedge during this period (which I only highlight because it seems like a plausable future scenerio) was ST bonds to offset the HmL premium decline.“If you are going to hold a high allocation to value stocks, you should also consider holding longer-term fixed income instrument. The recession/deflation/flight to quality risks of value stocks have negative correlation with the risk of longer-term fixed income instruments (which generally perform well in these environments. And since the evidence suggests that value stocks provide greater returns than growth stocks during inflationary periods, they offer greater protection than growth stocks against the inflation risk inherent in longer-term bonds. Thus holding a value tilt not only provides greater expected returns on your equity holdings, but it also allows you to take more interest-rate risk and thus potentially earn greater returns on your fixed income investments.”
Theoretically, a 3YR TIP would have performed very well during this period had it existed. I fear the price declines on a 10YR TIP from rising interest rates, however, would have offset much of the inflation credit. So I am uncertain how helpful longer term TIPS would have been.
Yes, HmL and ST bonds will both perform poorly during deflationary periods...but a lot less worse on a "real" basis than on a "nominal basis". Also, you have to put deflation in perspective. Less than 3% of quarters over the last 40 years have been deflationary, and not one single year since the early 50s. I am fine not hedging deflation risk. Others may feel differently.
Finally, I took a look not to long ago at all "flights to quality" dating back to the late 50s (on a month to month basis). I found only 6 opportunities in the last 5 decades that stocks and 5YR T-Notes decoupled by as much as 25% in just a month or two. Most rebalancing opportunities come from the long term, gradual outperformance of stocks v. bonds, or long drawn out declines from stocks and stable bond performance, not short "flight to quality bursts" from stocks out of bonds. So, most bond--> stock or stock--->bond rebalancing will occur over time as long as your stocks do reasonably well and your bonds are NOT of the very low quality, high equity correlation variety.Swensen also advocates longer-term bonds but mainly for protection in financial crises when most needed – “By holding portfolios of high-quality, long-term, non-callable instruments, investors emphasize the attributes of bonds that provide the greatest protection in times of financial crisis”. He discusses these attributed in some detail and some of this downside risk protection was highlighted in an earlier thread
Treasuries do better during severe declines, giving you more to "rebalance with", but nothing is stopping you from a bit of dynamic rebalancing or overbalancing out of slightly less appreciated (non treasury) bond positions into stocks.
--I came away from this thinking that the flight to quality nature of treasuries is, for the most part cosmetic...it usually doesn't provide you any real opportunity to rebalance (given 20% or 25% bands) unless you do so daily like Yale. There is something to be said for a bond holding that "spikes" when stocks drop, but if you don't actually profit from it from rebalancing, it is more cosmetic and a "feel good" trait.
--I guess I could make the same arguement that ST bonds have a "feel good" trait in that they almost never fluctuate that much, so almost all of the volatility is on the equity side, where you are expecting to be paid. So volatility is directly in line with premium compensation, which is a nice feature.
Finally, over some periods of HmL decline, Int/LT bonds do very well in absolute and relative terms. Sometimes they do very poorly in absolute and relative terms. I would argue, outside of 1% interest rate scenerios, ST bonds almost always do reasonably well. Maybe not always as well as Int'd/LT bonds, but thats a relative thing. The risk of very poor absolute ST bond performance seems small. ST Bonds seem to me like a "safer, more dependable bet" given the limited (and sometimes conflicting) data we are trying to draw inferences from.
Again, just my thoughts. Some/many/most may not see it this way. And, as you may know, I don't own bonds anyway, so my thoughts/research are simply from a theoretical standpoint.
SH
.
SH,
Thanks for the post. I haven’t looked at this in detail for a while but here are some of my earlier thoughts (drawing on earlier posts).
1. Bond impacts when Fed raises rates to curb inflation:
Value and term premiums have opposite signs in deflation and high inflation periods, but the same sign in periods of moderate to extremely high inflation (table below)
2. Treasuries versus corporates
Reasons I don't include corporate bonds:
IMO its important to look at a portfolio in total and not individual asset classes. ST bonds may be more stable as an individual investment but intermediate bonds have more often acted in opposite ways to equities (using my equity benchmark for comparisons), which has provided more stable portfolio returns. For illustration:
This seems consistent with Larry’s earlier insights quoted above.
Just my views and reasoning for choices made.
Robert
.
SH,
Thanks for the post. I haven’t looked at this in detail for a while but here are some of my earlier thoughts (drawing on earlier posts).
1. Bond impacts when Fed raises rates to curb inflation:
Value and term premiums have opposite signs in deflation and high inflation periods, but the same sign in periods of moderate to extremely high inflation (table below)
Code: Select all
1927-2005
Inflation range (%) >10 7-10 4-7 0-4 -ve
No. yrs 2 9 11 48 9
Av. Inflation rate 14 8.8 5.2 2.3 -3.8
Value premium (HmL) -3.7 13.2 7.5 6.0 -5.1
Term premium -7.7 -3.6 -0.8 4.2 4.1
Derived from Ken French website data and Ibbotson Yearbook
- Expected return on various investments:
Periods of deflation
Value stocks...............................................(+)
Longer term nominal bonds........................(-)
Inflation protected securities....................(flat)
Periods of modest inflation
Value stocks..............................................(+)
Longer term nominal bonds.......................(+)
Inflation protected securities.....................(+)
Very high inflation (and Fed tightens)
Value stocks..............................................(-)
Longer term nominal bonds......................(-)
Inflation protected securities....................(+)
2. Treasuries versus corporates
Reasons I don't include corporate bonds:
- i) Poor relative performance in flight to quality: Corporate bonds don't provide as much protection as treasuries when needed most - during financial crises/flight to quality episodes (as reflected in the earlier link). I prefer to take my default risk in equities (FF show default loadings on stocks, particularly small cap stocks, to be larger than for bonds). Adding additional default risk to fixed income just raises fixed income correlations with equities particularly if tilted to small cap value stocks.
ii) Limited upside, unlimited downside: Corporate bonds have limited upside (regular payment of interest and return of principal), but unlimited downside (worst being default without recovery). e.g. WorldCom “A” rated bond prices decline 88 percent in 6 month – no chance of recovery. There was a similar outcome for bondholders of the highly rated Penn Central Railroad. While infrequent it does and can happen – why take the risk?
iii) Heads you win, tails I lose characteristic (call provisions): If rates decline, the corporation wins by calling the bonds at a fixed price (investors lose out on the now high-coupon bond) but if rates rise investors are left holding the bonds (providing ‘cheap’ financing for the corporation).
IMO its important to look at a portfolio in total and not individual asset classes. ST bonds may be more stable as an individual investment but intermediate bonds have more often acted in opposite ways to equities (using my equity benchmark for comparisons), which has provided more stable portfolio returns. For illustration:
Code: Select all
1970-2006
Annualized Standard
Return Deviation Sharpe
75% equity/25% 2 yr T-Notes 14.00 13.28 0.659
73% equity/27% 5 yr T-Notes 14.00 13.09 0.667
Just my views and reasoning for choices made.
Robert
.