cancel
Showing results for 
Search instead for 
Did you mean: 
     
Highlighted
Valued Contributor

Fixed Income OEF Correlations

Just some musings over a morning coffee ...

Looking at various bond mutual fund rolling correlations this morning, there's some interesting data.  All charts cover the past five years using daily price data.

1. Vanguard High Yield VWEHX correlations to Treasuries VFITX oscillate around the zero line and recently turned negative:

ScreenHunter 608.png

2.  Investment Grade bond funds like PIGIX have provided true diversification from equities, excepting during the late 2016 "taper tantrum" when both stocks & bond sold off:

ScreenHunter 612.png

3.  HY correlations to the S&P 500 are positive and have been increasing over the past year:

ScreenHunter 609.png

4. Some multi-sector bond funds like PTIAX have been moving with strong correlation to Treasuries ...

ScreenHunter 610.png

4. ... while other multi-sector funds like PIMIX have moved independent of Treasuries, depending on active management strategy decisions ...

ScreenHunter 611.png

5. ... and other multi-sector funds like LSBDX have largely had weak correlation to Treasuries:

ScreenHunter 614.png

Is this information useful for portfolio construction?  I draw a few tentative conclusions:

  • High Yield bond funds provide only limited diversification to equities.
  • Investment grade bond funds usually (but not always) offer good diversification to equities.
  • Multi-sector bonds are all different.  We should look beyond their historical CAGR and volatility performance to understand their contribution to portfolio diversification.  For example, PTIAX has behaved similar to Treasury bond funds, while PIMIX seems to go anywhere and LSBDX is typically risk-on.

FWIW.

N.

20 Replies
Highlighted
Participant ○○○

Re: Fixed Income OEF Correlations

Thanks, N., Cheers, rm

0 Kudos
Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations

Good data. Notable point is that correlations vary over time. Often a single value is used.

YBB
0 Kudos
Highlighted
Follower ○○

Re: Fixed Income OEF Correlations

The PTIAX-VFITX correlation is interesting.  I had previously noticed the almost-0 correlation of PTIAX with total stock, one of the reasons I own it. 

0 Kudos
Highlighted
Contributor ○○

Re: Fixed Income OEF Correlations


@GaryTheSnail wrote:

The PTIAX-VFITX correlation is interesting.  I had previously noticed the almost-0 correlation of PTIAX with total stock, one of the reasons I own it. 


I checked the correlation of PTIAX to the S&P 500 on several different occasions.
The correlation was minimal each time.
Here's the most recent info from the parent firm's website.

PDF

PTIAX.JPG

0 Kudos
Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations

"High Yield bond funds provide only limited diversification to equities."

High yield bond funds have performed better than bond market index funds whenever used in a retirement portfolio.  VFINX/VWEHX/VBMFX link

  • VFINX/VWEHX performed better than VFINX/VBMFX in terms of both TR and income produced
  • VWEHX/VBMFX produced more income than VFINX/VBMFX and had a better risk adjusted return (higher Sharpe Ratio) than VFINX/VBMFX (although the TR was lower.)

Rather than thinking about high yield in terms of diversification, think about high yield in terms of equity replacement.

0 Kudos
Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations


@ElLobo wrote:

"High Yield bond funds provide only limited diversification to equities."

High yield bond funds have performed better than bond market index funds whenever used in a retirement portfolio.  VFINX/VWEHX/VBMFX link

  • VFINX/VWEHX performed better than VFINX/VBMFX in terms of both TR and income produced
  • VWEHX/VBMFX produced more income than VFINX/VBMFX and had a better risk adjusted return (higher Sharpe Ratio) than VFINX/VBMFX (although the TR was lower.)

Rather than thinking about high yield in terms of diversification, think about high yield in terms of equity replacement.


Yes, I think of Fixed Income CEFs like this; PCI, for example.  Though it's mostly been Pimco's Ivascyn who's been shooting out the lights. 

However, I'm not yet convinced that "buy & hold" exposure to High Yield OEFs make sense for me.  My tests suggest that a "barbell" approach of Equity + High Quality OEFs works better ... Wellesley, for example.

N.

0 Kudos
Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations

"Yes, I think of Fixed Income CEFs like this; PCI, for example. Though it's mostly been Pimco's Ivascyn who's been shooting out the lights."

As do I, at this point in my retirement!  I barbell PCI and SDYL, both leveraged funds, with a correlation coefficient of 0.45.  The coefficient for VFINX/VWEHX is 0.6.  SDYL/PCI has returned twice that of VFINX/VWEHX at twice the risk, due to leverage.  Mega ditto for the amount of income produced.

VFINX/VWEHX, SDYL/PCI, VWINX link.

0 Kudos
Highlighted
Participant ○

Re: Fixed Income OEF Correlations

Thanks to all the contributors of this forum. I will try to move back to investment grade bond funds may be by end of October. I am in bond funds like JMUTX which have slightly more than 50% in BBB and lower. 

SRT

0 Kudos
Highlighted
Participant ○○○

Re: Fixed Income OEF Correlations

Interesting.  Good job with this, N.

0 Kudos
Highlighted
Contributor ○○○

Re: Fixed Income OEF Correlations

For years, many bond oef investors have considered junk bonds and aggressive multisector bond funds as equity-like bond funds.  The high correlation of these type bond funds to equities, is not a great revelation to us as both equities and these type bond funds are just categorized as "higher risk" investments.  As a result, we look at various other bond categories, and bond funds, for an opportunity to lower risk.  Funds like PTIAX are less risky than many other multisector bond funds like LSBDX, but PTIAX is heavily influenced by its limited portfolio of risky mortgages and longer duration Munis.  You have other multisector bond funds like VCFIX/VCFAX which will give you less risk than PTIAX but with more total return, so the question then becomes why would you hold PTIAX if you can get a lower risk multisector bond fund with better total return.  Then there is another bond category that Norbert did not even consider that is becoming a popular multisector bond option--nontraditional bond funds.  Like true M* Multisector bond funds, nontraditional bond funds are not carbon copies of each other, with some being lower risk and solid total return options than many of their multisector bond fund siblings.  Within the nontraditional bond category, you can choose a fund like ISIAX, which closely resembles lower risk multisector bond funds, or you can look at lower risk nontraditional bond funds like SEMMX, PMZIX, and MWCIX as providing even more ballast, low correlation with equities and equity like bond funds.  For me, and I think many other bond oef investors, it is simply a matter of what risk level of bond oefs do you want in your portfolio, either for ballast roles correlating to equities or risky equity like bond funds.  As Norbert chooses to do, you can use the asset correlation measures, from the Portfolio Visualizer to help see statistically and graphically how to select these various funds for "ballast" roles.

Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations

@dtconroe : ".....either for ballast roles correlating to equities or risky equity like bond funds....."

I think you meant "ballast roles inversely correlating to equities", but I got your meaning.

YBB
0 Kudos
Highlighted
Contributor ○○○

Re: Fixed Income OEF Correlations

Having thrown the ballast issue out there, what correlation do people feel comfortable with? Currently I am at about .41 and would be at about .43 if I add a fund like JMSIX. Or I could add a fund like DODIX and drop to about ,36. I have an open slot for my 401k rollover. These numbers are through PV Backtest.

0 Kudos
Highlighted
Contributor ○

Re: Fixed Income OEF Correlations


@Gary1952 wrote:

Having thrown the ballast issue out there, what correlation do people feel comfortable with? Currently I am at about .41 and would be at about .43 if I add a fund like JMSIX. Or I could add a fund like DODIX and drop to about ,36. I have an open slot for my 401k rollover. These numbers are through PV Backtest.


I don't have a direct answer for you, but it occurs to me that there would be too little difference in your particular case to prevent you going in either direction.  I would tend to say: Do whatever feels right to YOU!

0 Kudos
Highlighted
Contributor ○○○

Re: Fixed Income OEF Correlations

In my case yes. 0.4 has a lot of risk still. But I am curious about others. What do you like for correlation? Need some risk to get better yield/income I think.

IOFAX is another super low correlation MS fund not mentioned.


@racqueteer wrote:

@Gary1952 wrote:

Having thrown the ballast issue out there, what correlation do people feel comfortable with? Currently I am at about .41 and would be at about .43 if I add a fund like JMSIX. Or I could add a fund like DODIX and drop to about ,36. I have an open slot for my 401k rollover. These numbers are through PV Backtest.


I don't have a direct answer for you, but it occurs to me that there would be too little difference in your particular case to prevent you going in either direction.  I would tend to say: Do whatever feels right to YOU!


 

0 Kudos
Highlighted
Contributor ○

Re: Fixed Income OEF Correlations


@Gary1952 wrote:

In my case yes. 0.4 has a lot of risk still. But I am curious about others. What do you like for correlation? Need some risk to get better yield/income I think.


@racqueteer wrote:

@Gary1952 wrote:

Having thrown the ballast issue out there, what correlation do people feel comfortable with? Currently I am at about .41 and would be at about .43 if I add a fund like JMSIX. Or I could add a fund like DODIX and drop to about ,36. I have an open slot for my 401k rollover. These numbers are through PV Backtest.


I don't have a direct answer for you, but it occurs to me that there would be too little difference in your particular case to prevent you going in either direction.  I would tend to say: Do whatever feels right to YOU!



I personally think you're right; we're being forced to take on some risk to make any gains.  I'm not as concerned about correlations, because they can be either up OR down through coincidental factors.  The old saw: correlation is not causation.  I'm being careful here in the equity market, the long end of bond duration, and volatility generally.  Most of my 'riskiness' is in PCI and PDI, where I think the need for income works in their favor, and their holdings seem likely to retain value.  I own PDIIX and PIGIX, but I'm keeping a close eye on them; as they've been faltering a bit of late.

0 Kudos
Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations


@dtconroe wrote:

For years, many bond oef investors have considered junk bonds and aggressive multisector bond funds as equity-like bond funds.  The high correlation of these type bond funds to equities, is not a great revelation to us as both equities and these type bond funds are just categorized as "higher risk" investments.  As a result, we look at various other bond categories, and bond funds, for an opportunity to lower risk.  Funds like PTIAX are less risky than many other multisector bond funds like LSBDX, but PTIAX is heavily influenced by its limited portfolio of risky mortgages and longer duration Munis.  You have other multisector bond funds like VCFIX/VCFAX which will give you less risk than PTIAX but with more total return, so the question then becomes why would you hold PTIAX if you can get a lower risk multisector bond fund with better total return.  Then there is another bond category that Norbert did not even consider that is becoming a popular multisector bond option--nontraditional bond funds.  Like true M* Multisector bond funds, nontraditional bond funds are not carbon copies of each other, with some being lower risk and solid total return options than many of their multisector bond fund siblings.  Within the nontraditional bond category, you can choose a fund like ISIAX, which closely resembles lower risk multisector bond funds, or you can look at lower risk nontraditional bond funds like SEMMX, PMZIX, and MWCIX as providing even more ballast, low correlation with equities and equity like bond funds.  For me, and I think many other bond oef investors, it is simply a matter of what risk level of bond oefs do you want in your portfolio, either for ballast roles correlating to equities or risky equity like bond funds.  As Norbert chooses to do, you can use the asset correlation measures, from the Portfolio Visualizer to help see statistically and graphically how to select these various funds for "ballast" roles.


Traditionally, it was stocks for growth, bonds for income and to reduce overall portfolio volatility.  Bond funds, in turn, were based on bond market indices (VBMFX), which are higher quality intermediate term duration.  Ifn you wanted/needed higher yield income, you either went down in quality (VWEHX) or out in term (VWESX/VLTCX).  Other bond characteristics were implemented by 'tilting' towards higher quality (treasuries), lower volatility (shorter term), inflation (TIPs), positive/negative convexity (MBS based funds), domestic/foreign/EM debt, active versus passive indexing, and so forth.

The biggest change I see coming into play these days is leverage, unheard of a decade ago in any great detail, in turn implemented with CEFs, ETFs, and ETNs.  As a result, ifever I want to compare an OEF with a CEF/ETF/ETN, I first see if it uses leverage and, if so, mentally divide the risk/return/yield characteristics by 2.  Not rocket science, just a state of mind.

In terms of the OP, a quick look at the correlation coefficients for a leveraged fund, SDYL, compared to the unleveraged product, SDY, to both the stock market, SPY, and the bond market, BND, shows little difference in the correlation coefficients as a result of using leverage.  Here is the link.

Here is the table from that link:

TickerSPYSDYSDYLBND
SPY10.910.91-0.1
SDY0.9110.980.01
SDYL0.910.981-0.02
BND-0.10.01-0.021

 

(What seems to be different these days from days of olde is that inflation doesn't seem to enter much into discussions anymore!)

0 Kudos
Highlighted
Contributor ○○○

Re: Fixed Income OEF Correlations

Gary: "In my case yes. 0.4 has a lot of risk still. But I am curious about others. What do you like for correlation? Need some risk to get better yield/income I think.

IOFAX is another super low correlation MS fund not mentioned."

I would be cautious about generalizing about risk and correlation statistics, without fully understanding the specific assets of funds you are comparing.  IOFAX has been the subject of many threads, is focused on a single category of high risk mortgages, and has the "potential" to change dramatically if that asset class deteriorates.

0 Kudos
Highlighted
Participant ○○○

Re: Fixed Income OEF Correlations

A little off topic, but I haven’t seen much on IOFAX/IOFIX in a while. On their website, if you scroll down, they have two buttons -

  1. Download Fact Sheet - which has details of their asset allocation
  2. Download Fund Presentation – which if you go to Slide 23, lower right hand table, shows an interesting statistic:

Legacy RMBS Collateral Historical Performance

90+Day Delinquencies (DQ): 15.3% Today (data as of 6/30/2019)

http://alphacentricfunds.com/products/income-opportunities-fund/

0 Kudos
Highlighted
Valued Contributor

Re: Fixed Income OEF Correlations

Good thread!!!

As always, what should I do based on the info in this thread?

Norbert

  • High Yield bond funds provide only limited diversification to equities.
  • Investment grade bond funds usually (but not always) offer good diversification to equities.
  • Multi-sector bonds are all different.  We should look beyond their historical CAGR and volatility performance to understand their contribution to portfolio diversification.  For example, PTIAX has behaved similar to Treasury bond funds, while PIMIX seems to go anywhere and LSBDX is typically risk-on.

FD: based on the above.  No need to use HY.  Usually, IG offer good diversification and why we should use them. MS funds depend on what funds you own which means you better know what you own.

The above is very common to what you find in most articles and research.  They mostly discuss stock(SPY) + IG bonds because these are "simpler" indexes with more consistent correlation and results.

What should you do with the info next week?  after all, I always want to know what to do with my portfolio.  If you want to be a B&H (boglehead) for decades to come you can just hold indexes and can expect to get market performance with expected volatility and risk.  That is a very good way that worked for decades and very simple to accomplish.

The minute you step away from "simple" indexes correlation loses its effectiveness. 

But...mmm...there is always a but, since 2000 I selected my own unique way.  How to accomplish similar performance with better risk attributes(SD, Max Draw, Sharpe, Sortino) or higher performance with similar risk attributes or the best, better performance + better risk attributes.  That is not a simple or guaranteed task but if you don't try you won't get it.

The following samples are proof of this. 

10 year performance of PIMIX vs VFITX (link) - one of the easiest way to show what a genius manager can do.

PortfolioCAGRStdevBest YearWorst YearMax. DrawdownSharpe RatioSortino Ratio
PIMIX9.09% 3.58%22.17%0.58%-3.30% 2.35.7
VFITX(IG)3.22% 3.76%9.80%-3.09%-4.49% 0.741.25

 

Performance(link) of PRWCX vs VFINX since 01/2000

PortfolioCAGRStdevBest YearWorst YearMax. DrawdownSharpe RatioSortino Ratio
PRWCX10.39% 10.27%33.05%-27.17%-36.61% 0.861.3
VFINX5.55% 14.56%32.18%-37.02%-50.97% 0.330.47

 

USMV vs VFINX is another case (link)

So, are you participating in the risk/reward challenge?  :-)

Announcements

Morningstar is here to help you respond to the Coronavirus crisis.