Bond Returns: Realistic expectations

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.

Bond Returns: Realistic expectations

Postby Kevin M » Wed Jul 24, 2013 9:06 pm

I've posted this chart in a couple of threads already, but I thought it deserved a post of its own. There's even been a post that linked to the presentation that included this chart, but the chart wasn't included in the post.

Image

The blue line is YTM for Barclays US Aggregate Bond Index, and the orange line is the return for the following 10 years.

And here again is a link to the Vanguard video in which this chart was presented: Ken Volpert on the current bond landscape

As already quoted in other threads, Ken Volpert, of Vanguard said:

So what this graph is really communicating is that the yield to maturity right now is a really good estimate of what you're going to get over the next 10 years in this investment. It's really been a very reliable indicator. So it gives you a very good approximation of what the return is going to be.


As of Dec 31, the time mentioned in the presentation, the YTM for this index, which is the TBM benchmark index, was 1.74%, and Ken said:

So investors looking forward should expect that that's the kind of return you're going to get: 1.7% to 1.8% return over the next 10 years.


Note that YTM for the aggregate bond index is a bit higher now, but we're still looking at an estimated annual return in the ballpark of around 2% for something like the Total Bond Market Index fund over the next 10 years. Of course these are nominal returns, not real returns.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby MitchC » Wed Jul 24, 2013 9:14 pm

That's about what I've been hearing / reading as well. May be a knee jerk (not BH-like) reaction, but I moved ALL of my TBM allocation in my 401k to our Stable Value fund currently earning a consistant (no-risk) 2% in late May.
MitchC
 
Posts: 63
Joined: Wed Jul 03, 2013 7:02 pm

Re: Bond Returns: Realistic expectations

Postby Noobvestor » Wed Jul 24, 2013 11:22 pm

Vanguard Long-Term Tax Exempt Admiral (7.3 year duration) has an SEC yield of 3.33% right now. For someone in the 28% bracket, that's a tax-equivalent yield of 4.63%. Intermediate-term tax exempt is 2.45% (tax-equivalent: 3.4%). Not bad, all things considered.
"In the absence of clarity, diversification is the only logical strategy" -= Larry Swedroe
User avatar
Noobvestor
 
Posts: 3927
Joined: Mon Aug 23, 2010 2:09 am

Re: Bond Returns: Realistic expectations

Postby Electron » Thu Jul 25, 2013 12:06 am

Common Sense on Mutual Funds provides data for the simple bond model. Projections appear to be higher than the actual result in three periods of rising interest rates.

Decade, Initial Yield, Decade Return, Error

1950s 2.1% -0.1% -2.2%

1960s 4.5% 1.4% -3.1%

1970s 6.9% 5.5% -1.4%

The model used data for Long Term U.S. Government Bonds. It's possible that the Total Bond Index will perform a little better as a result of higher yields in many bonds of a given maturity relative to Government bonds.
Electron
User avatar
Electron
 
Posts: 1040
Joined: Sat Mar 10, 2007 9:46 pm

Re: Bond Returns: Realistic expectations

Postby Simplegift » Thu Jul 25, 2013 2:37 am

Kevin M wrote:I've posted this chart in a couple of threads already, but I thought it deserved a post of its own.

Kevin, the chart makes a very convincing point, expressed simply. Thanks for re-posting it.

It would be interesting to see similar data for a longer time period, if anyone knows of such an analysis, especially the period prior to 1980 when inflation was much more volatile — though perhaps the results wouldn't be much different.

Your chart at least pushes back to 1979. The only previous chart I'd seen comparing forward returns to yield-to-maturity was this one, with five-year forward returns since 1990, but expressing your same point:

Image
Source: Above the Market
Cordially, Todd
User avatar
Simplegift
 
Posts: 1343
Joined: Tue Feb 08, 2011 4:45 pm
Location: Central Oregon

Re: Bond Returns: Realistic expectations

Postby Clive » Thu Jul 25, 2013 5:30 am

I've not read the other posts, nor any of the links, and maybe I'm being dense, but wont the YTM reflect what an investor would actually achieve when held to maturity - in advance? Fixed income does that, you know exactly how much you'll earn at the time of purchase when the intent is to hold to maturity.
Clive
 
Posts: 1156
Joined: Sat Jun 13, 2009 6:49 am

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 6:34 am

Yes, with all the hand wringing over the supposed impending bond armageddon, this simple fact seems to get lost in all the noise. Fixed income is, well, fixed. You pretty much know what you are going to get. For a fund that is the SEC yield. Yes, the outcome may differ from the expectation by a percent or two, but not much more. Note that you lock in the 5 to 10 year forward return for the shares on the day you buy them. If you are periodically investing over your earning years, your return is going to reflect the average rate during your investing lifetime.

The big problem now is not that rates may rise.
Its that they might not rise, and we are stuck with these low yields for the foreseeable future.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Call_Me_Op » Thu Jul 25, 2013 6:37 am

Let's hope for higher interest rates. :wink:
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein
Call_Me_Op
 
Posts: 4825
Joined: Mon Sep 07, 2009 3:57 pm
Location: Milky Way

Re: Bond Returns: Realistic expectations

Postby Clive » Thu Jul 25, 2013 7:30 am

If you are periodically investing over your earning years, your return is going to reflect the average rate during your investing lifetime


I was also thinking along those lines. A 10 year ladder, rolling maturing bonds into the 10 year at the time, with the projected compared to the actuals would produce a chart similar to those shown in this thread.

Let's hope for higher interest rates.


Be careful of what you wish for. 10% yields, 10% inflation, 25% tax on nominal income = -2.5% real

Hope for positive real yield spreads in excess of nominal taxation. Or deflation might be even nicer, 0% nominal, -2% deflation = no taxes and your money has 2% more purchase power.
Clive
 
Posts: 1156
Joined: Sat Jun 13, 2009 6:49 am

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 7:42 am

Its for sure that we all dream of rising real rates and lower taxes. Good luck.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Epsilon Delta » Thu Jul 25, 2013 1:25 pm

Kevin M wrote:Image

The blue line is YTM for Barclays US Aggregate Bond Index, and the orange line is the return for the following 10 years.


The Barclays US Aggregate Bond Index has a duration of just about 5 years. If its duration has been more or less constant the 5 year return should be a better fit, and the good fit for the 10 year return looks a just a little suspicious.
User avatar
Epsilon Delta
 
Posts: 3440
Joined: Thu Apr 28, 2011 8:00 pm

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 1:57 pm

Epsilon Delta: When you say suspicious, what exactly do you suspect? It is clear from the chart posted above by Simplegift that the 5-year returns do fit a bit more closely. The 10 year returns are nothing but a smoothed version of the 5 year returns.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Electron » Thu Jul 25, 2013 2:03 pm

I'll add a few additional comments.

The graph makes an excellent point, but note that it covers a very powerful 30 year bull market in bonds that will likely never be repeated. The ten year Treasury yield spanned a range from 15.84% to 1.43% between 1981 and 2012.

I'm not sure we lock in any particular return when we buy bond fund shares. One cannot hold a bond fund to maturity.

We partially lock in a certain dividend stream but the payout will change depending on changes in interest rates going forward. The SEC yield is applicable today only and it is relative to current NAV. The bond index holds bonds until they are one year from maturity and new bonds are added after they start trading.

Total return to any given future date will also depend on the sequence and timing of any future rate changes. Consider a case where rates are stable for several years followed by a large increase in rates. NAV could fall considerably in a short period of time. That would not be an issue for a longer term investor, but it could be for someone intending to cash out on a specific date.

One other issue is that the total bond market index is not a constant. The percentage of the index in Government debt has increased over time. There are Treasury auctions almost every week. Fortunately, this can be offset by new issues in other types of bonds.
Electron
User avatar
Electron
 
Posts: 1040
Joined: Sat Mar 10, 2007 9:46 pm

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 2:18 pm

Electron: I don't agree with you. You call it a 30-year bull market. Why? Total return on bond funds is falling the entire time (even though NAV's are going up, the falling rates offset that).

I think it is very clear that when you buy into a bond fund (equivalent to a rolling bond ladder) you will get approximately the SEC yield going forward in total return. They only impact of changing interest rates is how that total return is apportioned between NAV and dividend payments. A bond is a bond. Fixed income is fixed. The returns are known in advance. The payout is not dramatically affected by changes in interest rates going forward. The value of a bond will approach par as it approaches maturity (it is not necessary to actually hold it to maturity).

Just as over the last 30 years, falling rates have meant falling returns, rising rates will mean rising returns.

Yes, you don't want any money in an intermediate term bond fund that you will need in the next 5 years. That money should be in cash.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 2:27 pm

Clive wrote:I've not read the other posts, nor any of the links, and maybe I'm being dense, but wont the YTM reflect what an investor would actually achieve when held to maturity - in advance? Fixed income does that, you know exactly how much you'll earn at the time of purchase when the intent is to hold to maturity.

Not for a bond fund, and not exactly for an individual, non-zero-coupon bond. Let's look at individual bonds first.

For a non-zero-coupon bond, the YTM calculation assumes that interest is reinvested at the calculated YTM rate. This is unlikely to be the case since rates are unlikely to be constant over the term of the bond. So the the total return if held to maturity would be higher or lower than the YTM, depending on the rates at which the coupon payments are reinvested.

A callable bond could be called, so you may not have the choice to hold it to maturity. This is why yield to call or yield to worst are often published for bonds. So despite your intent to hold to maturity, you may not end up holding to maturity. If you buy at a premium or discount, that also will affect your return if the bond is called, since you may get paid less (or more) than you paid. Then you also probably are looking at reinvesting the proceeds at a lower rate.

For a zero coupon bond, the YTM does indicate exactly what you'll earn if held to maturity. This is because their are no coupons, so there is no reinvestment risk. Of course if the bond is callable, that could throw a wrench in the works.

For a bond fund, reinvestment risk also is a factor as it is with individual bonds (since a bond fund is just a collection of bonds), and if the fund owns callable bonds the potential of bonds being called is a factor, but there's also the issue that bond funds don't mature (at least typical bond funds), so there is no maturity to hold it to. Intermediate term bond funds probably don't hold bonds to maturity, as you can see if you look at the maturity distribution for such a fund, and it has been argued in another thread that this actually could result in the YTM (or SEC yield, which is similar for a bond fund) understating the expected return of the fund, at least based on the current shape of the yield curve and assuming that it won't change much (EDIT: this is more of a factor for something like Intermediate-Term Treasury fund, which does not hold any bonds in the 1-3 year maturity range, as opposed to TBM which holds 26% of its bonds in the 1-3 year maturity range).

Finally, just looking at the charts above shows that although YTM is a pretty good predictor of return, it's not precise. I believe John Bogle stated that there has been a 91% correlation between YTM and following 10-year returns for the aggregate bond market (as represented by total bond funds). Note the post above that showed a few periods where the prediction was off by a fair amount.

Kevin
Last edited by Kevin M on Thu Jul 25, 2013 2:52 pm, edited 1 time in total.
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 2:45 pm

billyt wrote:Electron: I don't agree with you. You call it a 30-year bull market. Why? Total return on bond funds is falling the entire time (even though NAV's are going up, the falling rates offset that).

If you define a bull market as a period of relatively high (nominal?) returns, then I would say someone who bought a 30-year bond 30 years ago has experienced a bond bull market. Pretty much the same for shorter-maturity bonds, but less dramatic. Whether held to maturity or sold before maturity, the returns were relatively high.

As Vanguard says, it is mathematically impossible for us to see anything close to this going forward from today. Even just a few years ago you were looking at 5% expected return, and indeed that's pretty much what you earned. Now we're looking at closer to 2%. Compared to the expected returns over the next 10 years, I'd say we have indeed experienced a bond bull market over the last 30 years.

Rather than saying falling rates offset rising NAV, I'd say rising NAV offset falling rates--at least until recently (and of course during other intermittent periods of rising rates).

Vanguard tells us that we should adjust our return expectations lower if we want to maintain a similar AA. Bonds will not provide the cushion they have in the past during stock downturns. So we either reset our return expectations, or increase our risk to increase the expected return of our portfolios. Important note: I am not suggesting that we do the latter!

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 2:53 pm

Kevin: The chart that you posted contradicts what you say. The period of high returns was in the 80's. It would be OK to say that there was a bull market in bonds 30 years ago. Returns have been falling ever since. Returns are controlled by the rates, not the NAV changes. NAV's went up, rates went down, returns went down. Rising NAV's did absolutely nothing to offset those rate changes.

Yes, forward returns now are lower than a few years ago. Forward returns a few years from now could be higher, if rates rise. If rates normalize from here, in 5 years this will all be a minor blip in the rear view mirror.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Electron » Thu Jul 25, 2013 3:29 pm

billyt wrote:Electron: I don't agree with you. You call it a 30-year bull market. Why? Total return on bond funds is falling the entire time (even though NAV's are going up, the falling rates offset that).

Falling interest rates define a bull market in bonds. I think you need to consider bonds purchased most of the way down in rates. That includes the 1980s, 1990s, and even since 2000.

Bond investors for many years saw two things: They locked in coupons higher than they could have in the following years. Secondly, their holdings appreciated in value above par. That resulted in unrealized capital gains. True, that premium would disappear if held to maturity.

However, I agree with you that returns for new bond investments fell over time, but they were typically higher than returns for new purchases.

If rates were to rise steadily for the next ten years, investors would see the opposite. They would have locked in lower coupons than currently available, and they would have unrealized capital losses for a time in their holdings. However, reinvestments would benefit from higher rates.

Maybe part of the puzzle is separating income and reinvested income from realized and unrealized capital gains and losses.

They say that income is the largest part of total return over time. However, both factors come into play.
Electron
User avatar
Electron
 
Posts: 1040
Joined: Sat Mar 10, 2007 9:46 pm

Re: Bond Returns: Realistic expectations

Postby ogd » Thu Jul 25, 2013 3:35 pm

Yup. It's only a bull market if you look at the NAV like you would for stocks. Bonds are not like stocks, I can't emphasize this enough.

If you are an intermediate bond fund investor, it's been a bit of a bear market in terms of returns. You just aren't getting as much as before, and the only way you will is if the interest rates go up.

Let's summarize the effects of the great 30-year bull market without complicated graphs:
Barclay's agg index returns for the last 30 years: 710%, 7.2% annualized (impressive indeed)
Returns if interest rates had stayed exactly the same, no bond capital gains whatsoever: 3018%, 12.15%

So the "great bond bull market" has robbed a bond investor (edit: who did not even invest new money, which would have made it worse) of 3/4 of their return. Yes, it was inflation and a number of other things that caused this. But you can't just look at a price graph and conclude that bonds are at all time highs and draw the same conclusions that you would for stocks. For stocks, the equivalent of the second scenario (no capital gains since 1983) would have been an unmitigated disaster. Something is very, very different about bonds over the long term.

Kevin M wrote:Bonds will not provide the cushion they have in the past during stock downturns.

Kevin, you are a very reasonable person, but I don't know what to make of this conclusion. If you mean that they're going to make a little less money than they did in 2008-2009, then probably so. But a cushion they will be. I would personally use the expression "handrail in a blizzard", from personal experience.
User avatar
ogd
 
Posts: 2592
Joined: Fri Jun 15, 2012 12:43 am

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 3:40 pm

I recognize that the financial media have repeated this endlessly, that falling rates mean rising prices, so the last 30 years were a bull market. If you look at a bond price chart, you say wow, prices have just gone up and up. What a bull market.
Well I am saying that is not a bull market; that is just bull. The orange line on the chart is total return, which includes, both NAV changes and interest. Total return has done nothing but fall for 30 years. The bull market was 30 years ago, which is when you had the opportunity to lock in some great returns.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 3:57 pm

This argument about whether or not we should call the last 30 years a bond bull market or not is kind of silly, and it's off topic (but it's hard to resist responding :wink:). How about we start another post to debate that :sharebeer.

The point of the chart, and this post, is that expected returns for bonds currently are very low compared to any time during the last 35 years or so. This is what the bond experts at Vanguard are telling us, not something I'm making up. I don't think we're disagreeing about this main point.

I'm most interested in using this to help educate folks who look at past bond returns and expect similar returns in the future, which I have seen more than once in other posts. Now I have a post I can just refer those folks to to help educate them.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 4:12 pm

We are certainly in agreement that rates are low. However, both 10 year treasury's and 10 year TIPS have gained more than 1% over the last year (Bloomberg). We are headed in the right direction. It is also encouraging that inflation remains low and both the 10 year TIP and the Vanguard TIPS fund have positive yields. 30 years ago interest rates were very high, but so was inflation, so to say rates are very low compared to the last 35 years is not the whole story. The bond market at the present time is not great, but on the whole it is not nearly as bad as people seem to want to make it.

If I can earn a positive real yield, I am satisfied.

If I could earn 2-3% real I am very happy.

We may yet get there.
Last edited by billyt on Thu Jul 25, 2013 4:20 pm, edited 1 time in total.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 4:19 pm

ogd wrote:
Kevin M wrote:Bonds will not provide the cushion they have in the past during stock downturns.

Kevin, you are a very reasonable person, but I don't know what to make of this conclusion. If you mean that they're going to make a little less money than they did in 2008-2009, then probably so. But a cushion they will be. I would personally use the expression "handrail in a blizzard", from personal experience.

I meant to say "the same cushion" or "as much of a cushion". I think it's a little more clear if you look at the entire statement I made rather than just the part you quoted. I did not make this up. I took it from this Vanguard Research paper.

Like I said (perhaps poorly) above, in this paper Vanguard points out that because of much lower yields, a higher bond allocation will be required to get the same protection from a major stock downturn, with correspondingly lower expected return, and a higher stock allocation with more downside risk will be required to get the same expected return as when bond yields were higher. From the conclusion of the research paper:

The diversification benefits of bonds in a stock/ bond portfolio will likely persist. This feature, more
than projected returns, justifies a strategic allocation to bonds. That said, lower projected returns from
bonds and their diminished ability to generate high offsetting returns have important implications for
downside risk and the asset allocation decision. If investors have a risk tolerance that is defined by
a maximum tolerable loss, then their asset allocation should become more conservative and their return
expectations must be lower. Conversely, if investors place a premium on generating higher returns, as
opposed to lowering downside risk, and as a result are reducing their bond exposure in favor of more
equities, they must be willing to tolerate more downside risk in their portfolios.


Hopefully this clarifies the point I was trying to make.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby billyt » Thu Jul 25, 2013 4:29 pm

Kevin: I do think the passage that you quoted from Vanguard is a fair assessment (except rates have risen since then). But I think the first two sentences are the most important: "The diversification benefits of bonds in a stock/bond portfolio will likely persist. This feature, more than projected returns, justifies a strategic allocation to bonds."
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby ogd » Thu Jul 25, 2013 4:38 pm

Kevin M wrote:The point of the chart, and this post, is that expected returns for bonds currently are very low compared to any time during the last 35 years or so. This is what the bond experts at Vanguard are telling us, not something I'm making up. I don't think we're disagreeing about this main point.

This might be a case of "glass half empty". *You* look at it and say, "forward looking 10 year returns are only 2% projected, low by historical terms". For others, this graph shows that it's 2% almost no matter what happens to interest rates, and interest rates going up would only help things. And that projected returns have moved up 1% lately.

Kevin M wrote:This argument about whether or not we should call the last 30 years a bond bull market or not is kind of silly, and it's off topic (but it's hard to resist responding ).

Fair enough, and I wasn't the one bringing it up. I just want to get rid of this silly notion than because bond prices are at historical highs, it's only downhill from here. This is a view that manages to take into account the specifics of interest rates ("can't go much lower than this"), while simultaneously ignoring any positive effects from rate increases.

Kevin M wrote:Like I said (perhaps poorly) above, in this paper Vanguard points out that because of much lower yields, a higher bond allocation will be required to get the same protection from a major stock downturn, with correspondingly lower expected return, and a higher stock allocation with more downside risk will be required to get the same expected return as when bond yields were higher.

So the Vanguard paragraph makes perfect sense, but the way I'm reading it it's not about cushion during the stock downturns -- it's what happens the rest of the time: the low projected returns force you to either accept a portfolio with lower returns, or one that is more exposed to future stock downturns.

During the dowturns themselves, you'll always wish you were 100% bonds. :oops:
User avatar
ogd
 
Posts: 2592
Joined: Fri Jun 15, 2012 12:43 am

Re: Bond Returns: Realistic expectations

Postby Epsilon Delta » Thu Jul 25, 2013 5:19 pm

billyt wrote:Epsilon Delta: When you say suspicious, what exactly do you suspect? It is clear from the chart posted above by Simplegift that the 5-year returns do fit a bit more closely. The 10 year returns are nothing but a smoothed version of the 5 year returns.

I suspect data mining, at least the soft data mining of stopping to look for more data when you find something that matches your thesis.
User avatar
Epsilon Delta
 
Posts: 3440
Joined: Thu Apr 28, 2011 8:00 pm

Re: Bond Returns: Realistic expectations

Postby dbr » Thu Jul 25, 2013 5:33 pm

ogd wrote:
Kevin M wrote:Like I said (perhaps poorly) above, in this paper Vanguard points out that because of much lower yields, a higher bond allocation will be required to get the same protection from a major stock downturn, with correspondingly lower expected return, and a higher stock allocation with more downside risk will be required to get the same expected return as when bond yields were higher.

So the Vanguard paragraph makes perfect sense, but the way I'm reading it it's not about cushion during the stock downturns -- it's what happens the rest of the time: the low projected returns force you to either accept a portfolio with lower returns, or one that is more exposed to future stock downturns.

During the dowturns themselves, you'll always wish you were 100% bonds. :oops:


Yes, somehow in this conversation risk and return are not being discussed explicitly as separate statistics. If stocks fall 50%, then a 50% allocation to bonds means the portfolio will fall by half that amount. That is the same now as at any other time. In a 25/75 portfolio, the loss is cut to a quarter of that or 12.5%. If we assume no correlation between stocks and bonds and respective SD's of 20% and 6%, then the portfolio SD would be 15%. If the bond SD were cut in half along with the return, the portfolio SD would go down 1% more to 14%.

If we assumed stocks would return 8% and bonds 4% so that a 50/50 portfolio returns 6% and now bonds return 2% we would need a 70/30 asset allocation to earn that same 6%. Otherwise we would have only 5%, not 6% Such a portfolio would now be exposed to a loss of 35% rather than 25% if stocks fell by 50%. The SD of the 70/30 portfolio would be 17%, which is not much of an increase in volatility.

A more relevant problem might be to calculate the SWR's for the different situations. It is plausible that lower bond returns would lower SWR some, although one would have to be pretty ambitious to extrapolate such a lower return persisting for a full thirty year retirement. How much of that could be recovered by investing a little more in stocks, I have no idea. We're all supposed to abandon bonds for SPIAs in retirement anyway, so it would not matter what bonds do as far as retirees are concerned.

If we would just do arithmetic we could stop the confusion created by using false, emotion laden analogies like protect and cushion. I note that this language originates at Vanguard and not on this thread.
dbr
 
Posts: 14347
Joined: Sun Mar 04, 2007 10:50 am

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 6:58 pm

dbr wrote:If stocks fall 50%, then a 50% allocation to bonds means the portfolio will fall by half that amount. That is the same now as at any other time.

Although this is OK as an approximation, I don't believe that it's consistent with the conclusion of the Vanguard paper:
Despite bonds’ importance for diversifying equity risk, today’s low bond yields mean that balanced
portfolios have more downside risk than in the past
.
(emphasis mine).

But they're only talking about a few percentage point difference:

With returns of 1.9% expected from bonds based on their current yields, these balanced
portfolios are likely to deliver returns ranging from −6.9% to −11.2% when equities decline −20%.
These results are between 2%−3% lower than what these portfolios would return (−3.6% to −9.1%) in
the same equity market sell-off if bonds’ expected performance matched the historical average yield of
7.3% (based on Barclays U.S. Aggregate Bond Index from January 1, 1976, through January 31, 2013).


In the case of a flight-to-quality scenario, especially for treasuries, the beneficial effect on portfolio returns of bonds starting at higher rates is more significant:

The flight-to-quality scenario is particularly relevant today because it highlights investors’
recent historical experience in balanced portfolios. For example, at the height of the technology bubble
in September 2000, bonds yielded just over 7% and provided a cumulative 22.8% return during the
ensuing bear market. And at the top of the stock market that preceded the global financial crisis in
2007, bonds yielded 5.3% and provided a cumulative 7.6% return through the bottom of the equity market
in March 2009. These return levels contributed meaningful downside protection to balanced portfolios
during both periods. Given these past experiences, investors may be disappointed by the level of
downside protection that bonds can offer when the next bear market occurs, given bonds’ current yield
of 1.9% (as of January 31, 2013).


Still, the differences are in the range of 5-7 percentage points (see preceding paragraph in the paper), so not huge, but as Vanguard says, "meaningful".

It's not about abandoning an allocation to fixed income, it's about having realistic expectations. IMO, it's also important not to overestimate the expected returns of bonds based on past performance (as I've seen people do in other threads) when comparing to alternatives like stable value funds and CDs; e.g., "Why should I invest in a stable value fund at 3% when I can earn 5% to 6% in a total bond fund?"

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 7:14 pm

Although the discussion of whether or not to call the past 30 years a bull market is off topic and not actionable, the actual statement that first mentioned the term in this thread might be worth exploring:

Electron wrote:The graph makes an excellent point, but note that it covers a very powerful 30 year bull market in bonds that will likely never be repeated. The ten year Treasury yield spanned a range from 15.84% to 1.43% between 1981 and 2012.

So a relevant question might be: Is YTM as a predictor of following 10-year return likely to be as good when rates have risen over the 10-year period as when they have fallen? I don't know the answer, but someone posted some periods where the prediction was the most off. Is there a positive or negative correlation there?

Of course even if there is a correlation, it really isn't actionable either, unless one wants to bet on the direction of interest rates over the next 10 years. I personally am hedging my bets by holding much of my fixed income in CDs, but still a significant chunk in higher-yielding bond funds (not TBM).

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby Clive » Thu Jul 25, 2013 8:41 pm

Kevin M wrote:Finally, just looking at the charts above shows that although YTM is a pretty good predictor of return, it's not precise. I believe John Bogle stated that there has been a 91% correlation between YTM and following 10-year returns for the aggregate bond market (as represented by total bond funds). Note the post above that showed a few periods where the prediction was off by a fair amount.

Seeing the actual comparing reasonably to the forward for the total aggregate isn't a surprise. But as you say its too much to expect 100% predictive accuracy.

A overall average for bonds might broadly compare. Deviations will of course occur, depending upon luck/timing etc. A junkier bond fo instance might pay 3% more than a respective treasury bond, but perhaps run with a 3% risk of a default occurring such that when defaults are accounted for the two broadly compare in total return. If however defaults are below average over a period of time the junkier bonds might relatively outperform. Over another period the reverse might hold and the junkier bonds might lag.

From Robert Shiller's data for one year and RLONG since 1871 (yearly granularity), the average yield from one year = 4.71% whilst RLONG = 4.65%. Whilst investors might expect to be rewarded more for lending for longer periods (buying longer dated bonds), the overall average is that there's little difference between lending for a few years or lending for longer periods (such as 30 years).

A relatively easy way to see that effect is to download Simba's backtest spreadsheet and use the data contained within that to compare various bond allocations since 1972 (earliest data sourced in that spreadsheet). Compare TBM with a 2 year T and LTT 50-50 barbell; Compare TBM with 100% 5 year Treasury; Compare 5 year T with 2 year T; Compare TBM with LTT ... etc and you'll see broadly similar overall results. There's some zigzagging around each other for some of those choices - but in the broad sense they're the same.

Assuming all bonds to be broadly similar overall, then the predictive power of any one, might also be a reasonable predictor for others - give or take.
Clive
 
Posts: 1156
Joined: Sat Jun 13, 2009 6:49 am

Re: Bond Returns: Realistic expectations

Postby dbr » Thu Jul 25, 2013 9:15 pm

Kevin M wrote:
dbr wrote:If stocks fall 50%, then a 50% allocation to bonds means the portfolio will fall by half that amount. That is the same now as at any other time.

Although this is OK as an approximation, I don't believe that it's consistent with the conclusion of the Vanguard paper:
Despite bonds’ importance for diversifying equity risk, today’s low bond yields mean that balanced
portfolios have more downside risk than in the past
.
(emphasis mine).
Kevin


They are not being explicit about whether they are talking about annual variability of returns due to the volatility of stocks or a secular change in expected return due to a decades long depression in bond returns. Most people think of downside risk as how much a portfolio could change over a short run in the context of annual returns in the stock market. The dilution of this risk by holding bonds is as valid as ever. When one looks at a trend in interest rates that has declined over a thirty year period and then one projects a future situation in which bond returns may be on the low side for a decade or two, that is an entirely different context for downside risk. They are not giving an explicit description of what they mean. That is why they need to discuss numbers and specifics instead of mouthing meaningless descriptions that are ambiguous. It is just poor writing, poor communication, and poor teaching.
dbr
 
Posts: 14347
Joined: Sun Mar 04, 2007 10:50 am

Re: Bond Returns: Realistic expectations

Postby Clive » Thu Jul 25, 2013 9:32 pm

dbr wrote:Yes, somehow in this conversation risk and return are not being discussed explicitly as separate statistics. If stocks fall 50%, then a 50% allocation to bonds means the portfolio will fall by half that amount. That is the same now as at any other time. In a 25/75 portfolio, the loss is cut to a quarter of that or 12.5%. If we assume no correlation between stocks and bonds and respective SD's of 20% and 6%, then the portfolio SD would be 15%. If the bond SD were cut in half along with the return, the portfolio SD would go down 1% more to 14%.

How are you calculating the standard deviations of the combined two assets?

http://www.zenwealth.com/BusinessFinanc ... lator.html indicates different figures to what you are suggesting
Clive
 
Posts: 1156
Joined: Sat Jun 13, 2009 6:49 am

Re: Bond Returns: Realistic expectations

Postby Kevin M » Thu Jul 25, 2013 10:38 pm

dbr wrote:They are not being explicit about whether they are talking about annual variability of returns due to the volatility of stocks or a secular change in expected return due to a decades long depression in bond returns.

Did you read the paper? They actually are explicit in the flight-to-quality scenario. They are looking at hypothetical 1-year returns; see the footnote in Figure 3. I'm just quoting summaries and conclusions, not the entire paper.

They are not as explicit as to what time period they are using in the first scenario (Figure 2), which doesn't assume any specific bond price increases due to flight-to-quality, but just the historical 7.3% bond return. I think given the context they also are looking at hypothetical 1-year returns here.

dbr wrote:That is why they need to discuss numbers and specifics
They do.
dbr wrote:instead of mouthing meaningless descriptions that are ambiguous.
To me the context seems pretty clear; they're trying to set realistic expectations for how bonds are likely to provide somewhat different offsetting returns in a bear market depending on the initial bond yields. Again, the differences aren't huge, and less so for higher equity allocations. All of this seems intuitive to me, but I think some folks may pay more attention to Vanguard's numerical analyses than to my intuition.

Maybe I shouldn't have introduced that paper in this thread, since it's not the main point of the post (other than having realistic expectations), and this discussion is getting deeper in the weeds than was my intention here.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby Electron » Fri Jul 26, 2013 12:01 am

Kevin M wrote:So a relevant question might be: Is YTM as a predictor of following 10-year return likely to be as good when rates have risen over the 10-year period as when they have fallen? I don't know the answer, but someone posted some periods where the prediction was the most off. Is there a positive or negative correlation there?

Here is the simple bond model data from Common Sense on Mutual Funds. Projections appear to be higher than the actual result in three periods of rising interest rates. The model used data for Long Term U.S. Government Bonds.

Decade, Initial Yield, Decade Return, Error

1927-36 3.5% 4.9% 1.4%
1930-39 3.4% 4.9% 1.5%
1940-49 2.3% 3.2% 0.9%
1950-59 2.1% -0.1% -2.2% *
1960-69 4.5% 1.4% -3.1% *
1970-79 6.9% 5.5% -1.4% *
1980-89 10.1% 12.6% 2.5%
1990-97 8.2% 9.9% 1.7%

Average 5.1% 5.3% 0.2%

* Period of Rising Rates

The starting yield in 1950 was only slightly higher than the current SEC yield on VBMFX at 1.89%. At the end of the decade the actual return was -0.1%.

The ten year projection today using the bond index would be 1.89%. The actual return could be anywhere from negative to somewhat higher than 1.89% based on the data in the table.
Electron
User avatar
Electron
 
Posts: 1040
Joined: Sat Mar 10, 2007 9:46 pm

Re: Bond Returns: Realistic expectations

Postby Kevin M » Fri Jul 26, 2013 12:48 am

Thanks Electron, that is interesting. According to that data, returns were lower than would be anticipated from the yield-predicts-return model during the decades in which rates increased, as you say. Also, it's interesting that there were three consecutive decades with returns that were lower than the model predicts, the first two decades quite a bit lower, and starting in 1950 when the yield was close to today's yield.

I do not want to promote fear, but this data doesn't seem to jibe with what some have been saying. I'm sure we'll get some informative replies. I'll just repeat what I've said elsewhere: rising rates may be good for bond owners, but we may have to be very patient.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4295
Joined: Mon Jun 29, 2009 4:24 pm

Re: Bond Returns: Realistic expectations

Postby Simplegift » Fri Jul 26, 2013 1:50 am

Finally ran across a chart below (courtesy of Mr. Bogle's Financial Market Research Center) that shows the long-term relationship between current bond yields and forward returns for the period starting in 1925. In this case, it's intermediate-term Treasury yields (red line) and 10-year forward returns (blue line):

Image
Source: Bogle Financial Market Research Center

It appears that prior to 1980 (a period of generally rising rates, post-1940), current yields tended to consistently underestimate forward returns a bit, sometimes significantly — but since 1980 (a period of generally falling rates), they've tracked each other more closely. Overall, as indicated on the chart, their correlation for the entire period was quite high, at 0.91.

In short, yield-to-maturity is not a perfect predictor of future bond returns, but it seems to be a fairly good one.
Cordially, Todd
User avatar
Simplegift
 
Posts: 1343
Joined: Tue Feb 08, 2011 4:45 pm
Location: Central Oregon

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 7:16 am

OK, so according to this chart, returns are mostly better than predicted during stable and rising rate scenarios. That is encouraging.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 7:20 am

Epsilon Delta: That's 3 different charts now showing the same thing. I think that eliminates some sort of conspiracy to torture the data until it tells you what you want to hear. If you wanted the chart to show some different result, how would you do it?
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Tom_T » Fri Jul 26, 2013 7:23 am

Another point about bond funds... if rates continue to rise, won't that make bond funds more attractive? The people who turned up their noses at 2% bonds might find them a lot more interesting at 5%.
Tom_T
 
Posts: 1393
Joined: Wed Aug 29, 2007 3:33 pm

Re: Bond Returns: Realistic expectations

Postby Ranger » Fri Jul 26, 2013 8:13 am

I have updated the chart posted by Simplegift till 2012 here
viewtopic.php?f=10&t=120317&p=1758708#p1758708

Here is another way to look at it.

Image

Red triangle series is the periods of "Rate increase". My definition of "Rate increase" is 2 consecutive 10 year rate increase compared to previous period. From my eye balling the data, when the rate increase happens from very low base, realized 10 year forward return will be slightly less.

Another observation from the data "Electron" posted is Bogle is using Long term rate US Govt. Bonds. I am not sure, why he choose Long term US data since TBM is intermediate term.
User avatar
Ranger
 
Posts: 446
Joined: Tue Feb 20, 2007 11:27 pm
Location: Wolverine Nation

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 8:31 am

That is very interesting, and I see what you mean. If that chart is right, it looks like the 'crossover' point is at about 4%. I am curious about the difference between your chart and the one that simplegift posted.
That chart clearly indicates that all through the period where rates were near 2%, the actual return is higher than the predicted return. What explains the difference between that chart and yours?
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Ranger » Fri Jul 26, 2013 8:47 am

One reason forward realized return is slightly less could be because of negative convexity of MBS has more effect at very low rates. This chart is better fit


Image



billyt wrote: I am curious about the difference between your chart and the one that simplegift posted.
That chart clearly indicates that all through the period where rates were near 2%, the actual return is higher than the predicted return. What explains the difference between that chart and yours?


I wondered about the same. Both are pretty similar during the periods of data after 1970, where Lehman Agg. Bond index data can be verified. I used the data posted in that thread. So period in question is the Ibbotson data. I am not sure about the data quality or one is using end of the year YTM other could be average YTM (or my calculation is off. I will take a look at it once again after end of trading day).
User avatar
Ranger
 
Posts: 446
Joined: Tue Feb 20, 2007 11:27 pm
Location: Wolverine Nation

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 9:04 am

Ranger, it looks like the plot that simple gift posted is not plotting the 10 year treasury rate (as in your figure), but the yield on the aggregate index as the 'prediction'. In simplegifts plot the predictive yields go below 2%. Yours do not. That is the main difference as far as I can see.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Simplegift » Fri Jul 26, 2013 9:51 am

billyt wrote:OK, so according to this chart, returns are mostly better than predicted during stable and rising rate scenarios. That is encouraging.

In trying to understand why the actual forward returns from a bond might be different from its yield-to-maturity, Investopedia seems to indicate that these will only be same under certain conditions:

    • That the bond is held to maturity;
    • That the coupons are reinvested (rather than spent);
    • That the coupons are reinvested at an interest rate equal to the yield-to-maturity.
Thus, if the coupons are being reinvested at higher rates than the original YTM, such as during the 1940-1980 period when rates were rising dramatically, then it makes sense that YTM would tend to underestimate the forward returns during these times. However, the opposite doesn't seem to be occurring during the period post-1980, when rates were falling — here YTM seems to predict forward returns fairly closely, both at 5 years and 10 years out.

Perhaps someone more knowledgeable about bonds can better explain these dynamics?
Cordially, Todd
User avatar
Simplegift
 
Posts: 1343
Joined: Tue Feb 08, 2011 4:45 pm
Location: Central Oregon

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 10:46 am

OK, in a falling rate environment, reinvested coupons are going to return a little less, but there is more incentive to hold onto the bonds longer (or alternatively, they will have a higher NAV because of the falling rates). Maybe these 2 effects balance each other out pretty well.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 10:51 am

Ranger: There is another thing I don't understand about the scatter plots you posted: If you draw a 1:1 through the data, you will see that most of the time the return on the intermediate term bond fund is more than the rate on the 10-year treasury. I'm not sure if that makes sense.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby ogd » Fri Jul 26, 2013 11:07 am

Electron wrote:The model used data for Long Term U.S. Government Bonds.

Decade, Initial Yield, Decade Return, Error

Kevin M wrote:I do not want to promote fear, but this data doesn't seem to jibe with what some have been saying. I'm sure we'll get some informative replies. I'll just repeat what I've said elsewhere: rising rates may be good for bond owners, but we may have to be very patient.

Or be mindful of durations.

Tom_T wrote:Another point about bond funds... if rates continue to rise, won't that make bond funds more attractive? The people who turned up their noses at 2% bonds might find them a lot more interesting at 5%.

Absolutely. If that 5% is 3% real, retirees will be all over them. Even Kevin here :)

Simplegift wrote:In trying to understand why the actual forward returns from a bond might be different from its yield-to-maturity, Investopedia seems to indicate that these will only be same under certain conditions:
• That the bond is held to maturity;

Remember that bonds move all the time towards par value, not at just the maturity point. Reinvesting a short or matured bond at higher yields is one avenue of taking advantage of them.

Simplegift wrote:• That the coupons are reinvested (rather than spent);

Clearly you must subtract any spending from the prediction as well, as well as from any alternative use of the money (like cash). Since coupon money and capital money are the same color (unlike in the ING commercial), it might be clearest to subtract spending from capital, or ignore it until after you've made total return projections.

Simplegift wrote:• That the coupons are reinvested at an interest rate equal to the yield-to-maturity.

This is the other way to take advantage of higher yields.

Prima facie, there is no huge obvious reason for the Ytm to mis-estimate returns more during rising rate periods than falling, and I only see them diverging greatly in the 80s. Perhaps the sheer magnitude of the coupons in those days ensured a faster growth because of reinvestments.
User avatar
ogd
 
Posts: 2592
Joined: Fri Jun 15, 2012 12:43 am

Re: Bond Returns: Realistic expectations

Postby billyt » Fri Jul 26, 2013 11:36 am

Looking at the chart that simplegift posted almost all of the misfit (returns are higher than predicted) occurs before 1981.
billyt
 
Posts: 632
Joined: Wed Aug 06, 2008 10:57 am

Re: Bond Returns: Realistic expectations

Postby Simplegift » Fri Jul 26, 2013 11:46 am

billyt wrote:Ranger: There is another thing I don't understand about the scatter plots you posted: If you draw a 1:1 through the data, you will see that most of the time the return on the intermediate term bond fund is more than the rate on the 10-year treasury. I'm not sure if that makes sense.

Is it possible that the two data series in Ranger's chart are using a different definition of yield?

If one uses current yield (coupon divided by market price) and the other uses yield-to-maturity (including interest-on-interest and capital gains or losses), then the results won't be exactly comparable. Just a thought.
Cordially, Todd
User avatar
Simplegift
 
Posts: 1343
Joined: Tue Feb 08, 2011 4:45 pm
Location: Central Oregon

Re: Bond Returns: Realistic expectations

Postby Epsilon Delta » Fri Jul 26, 2013 11:48 am

billyt wrote:Epsilon Delta: That's 3 different charts now showing the same thing. I think that eliminates some sort of conspiracy to torture the data until it tells you what you want to hear. If you wanted the chart to show some different result, how would you do it?

I believe that YTM is the best indicator of return for the duration of a bond fund. I am particularly suspicious of the data when it roughly confirms a preconceived notion. Here's how it should work:
  1. Formulate a hypothesis "YTM is the best indicator of return for the duration of a bond fund".
  2. Gather data, the time series of YTM and forward return for duration of the same portfolio.
  3. See if data is consistent with the hypothesis.

That did not happen in this case. There are several possibilities
  • The hypothesis was "The YTM on intermediate bonds predicts the return on long term bonds".
  • they gathered the data first and built the hypothesis round it.
  • The data they wanted was hard to gather and they settled for something else.

The first is a reason to ask why is that so, the others are reasons to regard the data as less convincing than a "properly conducted" experiment. And yes the fact that it's hard to perform proper experiments in finance does mean that any conclusions are less trust worthy than those of a hard sciences. That's not to denigrate economics, it's just that economics is difficult, and any conclusions are more tentative.
User avatar
Epsilon Delta
 
Posts: 3440
Joined: Thu Apr 28, 2011 8:00 pm

Next

Return to Investing - Theory, News & General

Who is online

Users browsing this forum: anakinskywalker, Bing [Bot], cfs, Day9, EyeDee, gvsucavie03, IlikeJackB, LittleD, ljmcin, Tanelorn and 64 guests