Vanguard GNMA Fund Versus Total Bond Market Index Fund

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Electron
Posts: 1658
Joined: Sat Mar 10, 2007 8:46 pm

Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Electron » Mon Oct 30, 2017 12:50 pm

Here are some current parameters for the Admiral Shares of Vanguard GNMA Fund (VFIJX) and Vanguard Total Bond Market Index Fund (VBTLX).

At present, the GNMA fund appears to offer higher yields with shorter duration and shorter average maturity. I'm wondering if the current comparison is typical or whether the market is attempting to discount future changes in interest rates and the effect on GNMA securities. I'd also be interested in comments on any special risks in GNMA Securities for the period ahead.

VFIJX vs VBTLX

SEC Yield: 2.80% vs 2.44%
Distribution Yield: 2.82% vs 2.51%
Yield to Maturity: 2.8% vs 2.5%
Average Coupon: 3.6% vs 3.0%

Average Maturity: 6.9 years vs 8.3 years
Average Duration: 4.3 years vs 6.1 years
Electron

User avatar
flamesabers
Posts: 1470
Joined: Fri Mar 03, 2017 12:05 pm
Location: Rochester, MN

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by flamesabers » Mon Oct 30, 2017 2:10 pm

When deciding what bond fund is most suited for you, I think what matters most is what your investment strategy is as opposed to how good a particular bond fund's performance has been recently. Are you looking to obtain broad exposure to the U.S. bond market? Or are you more interested with investing in a specific type of bond such as GNMAs?

lack_ey
Posts: 5715
Joined: Wed Nov 19, 2014 11:55 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by lack_ey » Mon Oct 30, 2017 3:03 pm

I haven't looked at this closely, but I would note that the actual index fund covering a similar category, Vanguard Mortgage-Backed Securities (VMBS, VMBSX), which if anything should be very slightly riskier because it includes Fannie and Freddie in addition to Ginnie mortgage bonds, has notably lower yields. For example, SEC yield of 2.28% on ER of 0.07%.

The actively managed fund is allocating riskier somehow that is not showing up in average duration, or there's something to how the yields are being measured. I see it has some repos and Fannie/Freddie on the margins, though the majority is the actual GNMAs.

User avatar
jhfenton
Posts: 2177
Joined: Sat Feb 07, 2015 11:17 am
Location: Ohio

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by jhfenton » Mon Oct 30, 2017 3:12 pm

lack_ey wrote:
Mon Oct 30, 2017 3:03 pm
I haven't looked at this closely, but I would note that the actual index fund covering a similar category, Vanguard Mortgage-Backed Securities (VMBS, VMBSX), which if anything should be very slightly riskier because it includes Fannie and Freddie in addition to Ginnie mortgage bonds, has notably lower yields. For example, SEC yield of 2.28% on ER of 0.07%.

The actively managed fund is allocating riskier somehow that is not showing up in average duration, or there's something to how the yields are being measured. I see it has some repos and Fannie/Freddie on the margins, though the majority is the actual GNMAs.
Those striking differences have been a mystery to me whenever I compare the two MBS funds, but I haven't taken the time to try to figure it out (because I don't own either).

User avatar
Taylor Larimore
Advisory Board
Posts: 26141
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Taylor Larimore » Mon Oct 30, 2017 3:24 pm

Electron wrote:
Mon Oct 30, 2017 12:50 pm
Here are some current parameters for the Admiral Shares of Vanguard GNMA Fund (VFIJX) and Vanguard Total Bond Market Index Fund (VBTLX).

At present, the GNMA fund appears to offer higher yields with shorter duration and shorter average maturity. I'm wondering if the current comparison is typical or whether the market is attempting to discount future changes in interest rates and the effect on GNMA securities. I'd also be interested in comments on any special risks in GNMA Securities for the period ahead.

VFIJX vs VBTLX

SEC Yield: 2.80% vs 2.44%
Distribution Yield: 2.82% vs 2.51%
Yield to Maturity: 2.8% vs 2.5%
Average Coupon: 3.6% vs 3.0%

Average Maturity: 6.9 years vs 8.3 years
Average Duration: 4.3 years vs 6.1 years
Electron:

Vanguard's diversified Total Bond Market Index Fund (VBTLX) holds 21% government mortgage-backed securities.

I suggest keeping investing simple with total market index funds.

Read my "Simplicity" link below.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

DaufuskieNate
Posts: 218
Joined: Wed May 28, 2014 11:53 am

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by DaufuskieNate » Mon Oct 30, 2017 3:52 pm

GNMA funds have negative convexity. This makes them higher risk to investors leading to the higher yields. You can learn more here:

https://vanguard.com/pdf/icrddc.pdf?2210045720

User avatar
Electron
Posts: 1658
Joined: Sat Mar 10, 2007 8:46 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Electron » Mon Oct 30, 2017 4:44 pm

Thanks all for the replies. The GNMA Fund looks attractive at present relative to the Bond Index and I was just trying to understand the reasons for the apparently favorable comparison. I also realize that the comparison will likely change over time. It's also possible that the fund exploited temporary inefficiencies in the market. The slow unwinding of the Fed's Balance Sheet which is expected to begin soon could also be a factor.

One factor in the comparison could be extension risk where duration in the GNMA fund will likely increase if rates rise. I should note that the duration of the Bond Index fund has been rising steadily in recent years as a result of new Treasury debt issuance and other factors.

I'll also mention that average credit quality as rated by Morningstar is AAA for the GNMA fund and AA for the Bond Index fund.

I do understand that the GNMA fund is less diversified than the Bond Index fund. It's good to know about the Vanguard MBS Index fund which I will monitor from time to time. Thanks also for the Vanguard information on Convexity.
Electron

alex_686
Posts: 2548
Joined: Mon Feb 09, 2015 2:39 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by alex_686 » Mon Oct 30, 2017 4:56 pm

DaufuskieNate wrote:
Mon Oct 30, 2017 3:52 pm
GNMA funds have negative convexity. This makes them higher risk to investors leading to the higher yields. You can learn more here:

https://vanguard.com/pdf/icrddc.pdf?2210045720
I am going to quibble a bit with this. First, the negative convexity does not mean higher risk - it means a different risk. Second, MBS can have negative convexity if prepayments are sufficiently high to break the supporting tranches. What is the chance of this happening? Will mortgage rates fall low enough that a unexpected number of people will refinance their mortgages? Most already have done so.

DaufuskieNate
Posts: 218
Joined: Wed May 28, 2014 11:53 am

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by DaufuskieNate » Mon Oct 30, 2017 5:42 pm

alex_686 wrote:
Mon Oct 30, 2017 4:56 pm
DaufuskieNate wrote:
Mon Oct 30, 2017 3:52 pm
GNMA funds have negative convexity. This makes them higher risk to investors leading to the higher yields. You can learn more here:

https://vanguard.com/pdf/icrddc.pdf?2210045720
I am going to quibble a bit with this. First, the negative convexity does not mean higher risk - it means a different risk. Second, MBS can have negative convexity if prepayments are sufficiently high to break the supporting tranches. What is the chance of this happening? Will mortgage rates fall low enough that a unexpected number of people will refinance their mortgages? Most already have done so.
Just making the point that all else equal, a bond with negative convexity involves a higher risk than the same duration and credit risk bond without negative convexity. I say this because the OP is comparing a GNMA fund with lower duration to Total Bond with higher duration. Negative convexity certainly explains some of the yield difference. And in the case of mortgages it cuts both ways. Mortgage rates can rise enough that an unexpected number of people hold onto their mortgages longer than expected. The duration of your fund increases at just the wrong time.

User avatar
Electron
Posts: 1658
Joined: Sat Mar 10, 2007 8:46 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Electron » Sat Nov 04, 2017 1:03 pm

lack_ey wrote:
Mon Oct 30, 2017 3:03 pm
I haven't looked at this closely, but I would note that the actual index fund covering a similar category, Vanguard Mortgage-Backed Securities (VMBS, VMBSX), which if anything should be very slightly riskier because it includes Fannie and Freddie in addition to Ginnie mortgage bonds, has notably lower yields. For example, SEC yield of 2.28% on ER of 0.07%.

The actively managed fund is allocating riskier somehow that is not showing up in average duration, or there's something to how the yields are being measured. I see it has some repos and Fannie/Freddie on the margins, though the majority is the actual GNMAs.
It's not clear if the modestly higher yields at times in the GNMA Fund result in higher returns compared with the Mortgage Backed Securities Index fund. The chart below shows the two funds tracking quite closely along with the Total Bond Market Index fund. However, different time periods show slightly different results. You can change the Zoom Control to 3Y, 1Y, YTD, 3M, and 1M.

http://quotes.morningstar.com/chart/fun ... A%5B%5D%7D

If I select the Performance Tab and add VMBSX, the GNMA fund shows a five year average annual return of 1.99% versus 1.89% for the Mortgage Backed Securities fund. The Total Bond Market Index fund VBTLX shows a five year return of 2.02%.

One possibility is that the GNMA fund picks up additional return through the use of Derivatives, Repurchase Agreements, and Mortgage Dollar Rolls.
Electron

aristotelian
Posts: 3033
Joined: Wed Jan 11, 2017 8:05 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by aristotelian » Sat Nov 04, 2017 3:32 pm

If you were to chase a little more return than total bond, I would suggest Vanguard Intermediate Bond (VBILX/BIV) which would be more diversified than the GNMA fund.

User avatar
Electron
Posts: 1658
Joined: Sat Mar 10, 2007 8:46 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Electron » Sat Nov 04, 2017 5:12 pm

aristotelian wrote:
Sat Nov 04, 2017 3:32 pm
If you were to chase a little more return than total bond, I would suggest Vanguard Intermediate Bond (VBILX/BIV) which would be more diversified than the GNMA fund.
Thanks for the suggestion on the Intermediate Term Bond Index fund.

This thread was primarily about an apparent anomaly between Vanguard GNMA and Vanguard Total Bond Market Index fund.

In other words, explain why the GNMA fund would have a higher SEC Yield, higher Distribution Yield, higher Yield to Maturity, Higher Credit Quality, shorter Average Maturity and shorter Average Duration. That comparison is not what you would normally expect.
Electron

jalbert
Posts: 2220
Joined: Fri Apr 10, 2015 12:29 am

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by jalbert » Sat Nov 04, 2017 5:53 pm

alex_686 wrote:
Mon Oct 30, 2017 4:56 pm
DaufuskieNate wrote:
Mon Oct 30, 2017 3:52 pm
GNMA funds have negative convexity. This makes them higher risk to investors leading to the higher yields. You can learn more here:

https://vanguard.com/pdf/icrddc.pdf?2210045720
I am going to quibble a bit with this. First, the negative convexity does not mean higher risk - it means a different risk. Second, MBS can have negative convexity if prepayments are sufficiently high to break the supporting tranches. What is the chance of this happening? Will mortgage rates fall low enough that a unexpected number of people will refinance their mortgages? Most already have done so.
There also is extension risk if home sales drop off precipitously, as they are also a source of prepayments.

Corporate bonds also can exhibit negative convexity. Unlike GNMAs, Fannie and Freddie MBS are not contractually backed by the US govt. Almost all of the MBS in TBM are not GNMAs.

GNMA spreads are around 90 basis points relative to intermediate treasuries. Historically, GNMAs total return has exceeded that of intermediate treasuries by about 35 bp/yr. The spread is thus likely indicative of increased risk of extension and/or prepayments at present.

The Vanguard GNMA fund (vfiix) has outperformed the Vanguard Total Bond Market index fund (vbmfx) with less volatility, measured over the full lifespan of the latter.
Last edited by jalbert on Mon Nov 06, 2017 1:27 pm, edited 1 time in total.
Risk is not a guarantor of return.

User avatar
Electron
Posts: 1658
Joined: Sat Mar 10, 2007 8:46 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Electron » Sun Nov 05, 2017 1:51 pm

jalbert wrote:
Sat Nov 04, 2017 5:53 pm
The Vanguard GNMA fund (vfiix) has outperformed the Vanguard Total Bond Market index fund (vbmfx) with less volatility, measured over the full lifespan of the latter.
Here is a Morningstar chart showing that comparison which goes back to 12-11-86. That time period represents nearly 31 years.

http://quotes.morningstar.com/chart/fun ... A%5B%5D%7D

The duration of the Total Bond Market Index Fund has risen quite a bit over that period and currently stands at 6.1 years. I don't know the exact duration in 1986 but it was 4.7 years in August 2007. I wonder if a broad bond index of relatively constant duration would be a better choice.
Electron

jalbert
Posts: 2220
Joined: Fri Apr 10, 2015 12:29 am

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by jalbert » Mon Nov 06, 2017 8:37 pm

The total bond index is great for low cost and the simplicity of holding a single fund. I don't think it is guaranteed to be the optimal bond subclass diversification, but I think Nobel Laureate Bill Sharpe would disagree with me.

GNMAs and maybe MBS in general are areas where active management seems to add net value. The TBM index fund holds a market index portfolio of agency MBS. Thus, I tend to prefer 80% VBILX/BIV (intermediate bond index) & 20% VFIIX/VFIJX (GNMA fund) to the TBM index fund, but the low cost and single fund portfolio simplicity of a TBM index fund is compelling as well.
Risk is not a guarantor of return.

not4me
Posts: 167
Joined: Thu May 25, 2017 3:08 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by not4me » Tue Nov 07, 2017 11:03 am

Electron wrote:
Sun Nov 05, 2017 1:51 pm

The duration of the Total Bond Market Index Fund has risen quite a bit over that period and currently stands at 6.1 years. I don't know the exact duration in 1986 but it was 4.7 years in August 2007. I wonder if a broad bond index of relatively constant duration would be a better choice.
Noting the change in duration over the past 10 year period brings up a point that I don't think is considered often enough. How often is it argued if one is re-investing distributions from a bond fund, then the future behavior over the upcoming period of time as measured by current duration can be predicted based on assumptions about interest rate changes. I think most realize that is an approximation, but I wonder how close it really is.

Isn't it is basically making the assumption that the issuers of bonds will continue to behave as they have in the either the recent past or previous times of similar rate changes? If the drivers behind the rate changes are different, I would assume the characteristics of the "market" (as defined by the index) may also change. As bond issuers change their behavior, might that not force the fund manager to sell bonds before they otherwise would?

I'm sure not smart enough to figure out how a "constant duration" index would perform in an environment where the "broader" market was expanding/contracting its duration.

User avatar
oldcomputerguy
Posts: 1999
Joined: Sun Nov 22, 2015 6:50 am
Location: In the middle of five acres of woods

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by oldcomputerguy » Tue Nov 07, 2017 11:40 am

Electron wrote:
Sat Nov 04, 2017 5:12 pm
This thread was primarily about an apparent anomaly between Vanguard GNMA and Vanguard Total Bond Market Index fund.

In other words, explain why the GNMA fund would have a higher SEC Yield, higher Distribution Yield, higher Yield to Maturity, Higher Credit Quality, shorter Average Maturity and shorter Average Duration. That comparison is not what you would normally expect.
The SEC yield is a figure created by the Securities and Exchange Commission meant to provide apples-to-apples comparison of short-term (30-day) performance. In this particular case, it may not be strictly apples-to-apples; Vanguard states that GNMA's SEC yield is based on prior 30 days' actual income, while Total Bond's SEC yield is based on its holdings' yield-to-maturity. So it may be a bit misleading simply to compare those numbers blindly.

For what it might be worth, I looked at a head-to-head comparison of Vanguard GNMA Fund Admiral Shares (VFIJX) and Vanguard Total Bond Market Index Admiral Shares (VBTLX), and noticed that Total Bond's YTD return is 3.14% versus 1.93% for GNMA. Total Bond also outperformed GNMA (slightly) in the 3- and 5-year timeframes, while GNMA outperformed in the 1- and 10-year timeframes, as well as since inception.
Anybody know why there's a 20-pound frozen turkey up in the light grid?

bt3
Posts: 3
Joined: Wed Dec 05, 2012 10:05 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by bt3 » Tue Nov 07, 2017 12:07 pm

Electron wrote:
Mon Oct 30, 2017 12:50 pm
Here are some current parameters for the Admiral Shares of Vanguard GNMA Fund (VFIJX) and Vanguard Total Bond Market Index Fund (VBTLX).

At present, the GNMA fund appears to offer higher yields with shorter duration and shorter average maturity. I'm wondering if the current comparison is typical or whether the market is attempting to discount future changes in interest rates and the effect on GNMA securities. I'd also be interested in comments on any special risks in GNMA Securities for the period ahead.

VFIJX vs VBTLX

SEC Yield: 2.80% vs 2.44%
Distribution Yield: 2.82% vs 2.51%
Yield to Maturity: 2.8% vs 2.5%
Average Coupon: 3.6% vs 3.0%

Average Maturity: 6.9 years vs 8.3 years
Average Duration: 4.3 years vs 6.1 years
Be very careful when comparing GNMA metrics vs. any non-mortgage bond fund.

If mortgage rates increase, prepayments are likely to decrease (all else equal) which will impact the attractiveness of the fund. Why? Refinancings should decline since a borrower's incentive to refi will decline & it's also likely that the propensity for people to move will decline as they'd end up with a higher mortgage rate in their new home. So in the world of increasing rates, the fund is likely to be stuck with lower coupon mortgages that have a longer duration than expected (either compared to today or compared to other funds at some point in the future). In English? The fund is likely to under-perform relative to other non-mortgage fixed-income funds. And if rates decrease from here, you're likely to have the flip-side: prepayments should increase (due to higher refinancings), durations will shorten & your coupon will drop. So the underlying mortgages left will also be less desirable. It's more complicated than this because it's never the case that all else is equal (and we're ignoring new home buyers) but it's a reasonable first approximation and it's the reason people talk about mortgages having negative convexity.

In the mortgage bond business, most people are using "option-adjusted" analytics to calculate the expected spread, duration, etc. averaged over a large number of interest-rate scenarios that are consistent with the spot interest-rate market (both in terms of level & volatility). Institutional investors spend a lot of time trying to figure out how homeowners will behave under various circumstances and adjust the mortgage bond cashflows accordingly. So they're better able to make an "apples-to-apples" comparison.

BTW, I'm not saying to avoid mortgage funds (I don't have a strong view) just that you need to be careful understanding the limitation of comparing these metrics.

jalbert
Posts: 2220
Joined: Fri Apr 10, 2015 12:29 am

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by jalbert » Tue Nov 07, 2017 2:03 pm

Be very careful when comparing GNMA metrics vs. any non-mortgage bond fund.

If mortgage rates increase, prepayments are likely to decrease (all else equal) which will impact the attractiveness of the fund. Why? Refinancings should decline since a borrower's incentive to refi will decline & it's also likely that the propensity for people to move will decline as they'd end up with a higher mortgage rate in their new home. So in the world of increasing rates, the fund is likely to be stuck with lower coupon mortgages that have a longer duration than expected (either compared to today or compared to other funds at some point in the future). In English? The fund is likely to under-perform relative to other non-mortgage fixed-income funds.
It is more complicated than that. Mortgage rates are not the only driver of home sales and, in the case of GNMAs, of refis. Home sales that do occur with lower rate mortgages are equivalent to a bond being called at par when market value is below par. The funds may then be reinvested at the higher rate. In the case of GNMAs, mortgage holders may refi at a higher interest rate to eliminate mortgage insurance wrapped into their payment. They may hold the FHA loan long enough for enough home appreciation to be at 80% loan-to-value to qualify for a conventional mortgage and come out ahead doing the refi at a higher rate.

MBS are also amortizing bonds, so each coupon payment includes return of principal that can be reinvested at a higher rate if rates rose. The speed at which rates rise determines how much return of capital occurs from amortization over the course of rates rising by any given amount. For instance, if rates rise by 50 basis points in a month, there is only 1 month of return of capital at par over the course of the 50 bp rise to offset extension effects, but if a 50 bp rise in rates occurs over a year, then there are 12 months of return of capital at par to offset extension effects. GNMAs also have a higher yield than treasuries as compensation for prepayment and extension risk. If rates rise more slowly, there is more accumulation of coupon interest as well over the course of the rate increase.

The behavior of GNMAs in a rising rate environment is quite complex. A good rule of thumb is that they will overperform similar duration treasuries if rates rise at a slow to moderate rate, but will underperform when rates rise more sharply.
Last edited by jalbert on Wed Nov 08, 2017 5:19 pm, edited 1 time in total.
Risk is not a guarantor of return.

User avatar
Electron
Posts: 1658
Joined: Sat Mar 10, 2007 8:46 pm

Re: Vanguard GNMA Fund Versus Total Bond Market Index Fund

Post by Electron » Tue Nov 07, 2017 6:40 pm

Thanks all for the additional replies.

not4me - You make an excellent point about duration. This forum often recommends that you limit the duration of a bond fund to when you may need the invested capital.

Here is an excellent reference in the Wiki. See the section on Duration and Point of Indifference.

/wiki/Bonds:_advanced_topics

In terms of the duration increasing in the US Aggregate Bond Index, I think it has been primarily new issuance of more long maturities by the Treasury and Corporations. The link below includes a Fact Sheet on the US Aggregate Bond Index and you can see how it operates. Maturity is one year and greater and the index is rebalanced monthly. New issues will be added to the index and other issues will fall out of the index. The Treasury has issued a lot of new paper in recent years as a result of the high budget deficits. Corporations have taken advantage of the low interest rates in recent years and also issued new paper. The relative distribution across all maturities would ultimately determine the duration of the portfolio.

https://bloombergindices.com/bloomb ... ethodology

oldcomputerguy - Thanks for pointing out the different SEC Yield formula used with the GNMA fund. I now see that information in footnote J under the Distribution tab. I think that explains a lot. There is undoubtedly a lot of complicated accounting in the GNMA fund especially when you include the derivatives, repurchase agreements, and mortgage dollar rolls.

https://personal.vanguard.com/us/funds/ ... =INT#tab=4

bt3 and jalbert - Thanks for the additional analysis and commentary which is a great help.
Electron

Post Reply