Weighted average maturity vs. duration

[Chuckanut]

I am looking at a particular bond fund.

Its weighted average maturity is 4.9 years.

Its duration is 1.32 years.

If I understand duration correctly, this fund is not very sensitive to interest rate changes despite an average maturity that is almost 4 times longer. A 2% increase in interest rates would cause this fund to lose about 2.6% of its value. Correct?

 

[pb4uski]

That is what it would imply. If interest rates increase 1% that the fund's investments would decline only 1.32%.

Typically the difference between weighted average maturity and duration is not so extreme. My guess is that the fund is using derivative to hedge interest rate risk in some manner. Is there anything unusual about the fund that might explain why the two metrics are so different?

 

[Chuckanut]

It's Fidelity's GNMA fund - FGMNX

Strategy
Normally investing at least 80% of assets in Ginnie Mae's and repurchase agreements for Ginnie Maes Investing in other U.S. Government securities and instruments related to U.S. Government securities. Engaging in transactions that have a leveraging effect on the fund.

 

[pb4uski]

Interesting. I just looked at Vanguard's GNMA fund and its weighted average maturity is 5.0 years and duration is 3.1 years and their MBS fund has a weighted average maturity of 5.0 years and duration of 2.8 years. They must be doing something to cause the more significant difference between weighted average maturity and duration.

You might look a those metrics for a number of other GNMA funds and if FGMNX is an outlier start reading or asking questions as to why before investing.

 

[mathjak107]

weighted maturity and duration do not coincide because the duration value is always changing for many reasons while maturity stays close to the same.

aside from credit risk changing many times a fund may sell a longer term bond and replace it with a shorter term lower paying one. it may not alter such a big mix of bonds as far as average maturity but it can effect duration.

 

[brewer12345]

Duration figures on funds that hold MBS (like any GNMA fund) are rarely worth the ink they are printed with. These are model-driven estimates that are making lots of assumptions about how mortgage borrowers will react to future interest rates, states of the economy, etc. To paraphrase, models are like xxxxs: everyone has one and most of them stink.


Google "negative convexity" for more fun educational material.

 

[OrcasIslandBound]

I just looked up FGMNX on Morningstar, found its duration to be 2.85 years.

Sent from my Nexus 4 using Early Retirement Forum mobile app

 

[target2019]

Some disparity between M* site (12/31/2014) and Fidelity site (1/31/2015)...

 

[pb4uski]

Perhaps Fidelity changed assumptions between 12/31/14 and 1/31/15.....

 

[audreyh1]

Duration figures on funds that hold MBS (like any GNMA fund) are rarely worth the ink they are printed with. These are model-driven estimates that are making lots of assumptions about how mortgage borrowers will react to future interest rates, states of the economy, etc. To paraphrase, models are like xxxxs: everyone has one and most of them stink.


Google "negative convexity" for more fun educational material.

Good to hear from Brewer, as it's been a while! It's not just the negative convexity, but also investors tend to dump these kind of bonds during interest rate spikes, compounding the problem.

I have learned not to trust the statistics on FGMNX as such a mortgage loan type fund simply cannot be modeled or understood using the standard bond fund characteristics.

Duration looks nice and low, so it appears as it has little interest rate sensitivity, but have the little taper tantrum of 2013 and all of a sudden duration increases markedly. I think it popped up over 5 years at one point, so fast it made your head spin. Too bad you can't get history of things like a fund's average duration and see how it really behaves in different interest rate environments!

This fund behaved very badly in 2013 compared to normal bond funds. It recovered spectacularly in 2014, but beware.

 

[OrcasIslandBound]

I can't help but wonder about the accuracy of fidelity's duration calculation. This is from their website. You can add up the durations above 3 years and below 3 years and the weighted sum suggests a duration of at least 3 years. Still, this fund looks quite good to me despite its 10 % drop in 2013.

Sent from my Nexus 4 using Early Retirement Forum mobile app

 

[audreyh1]

I don't think it's an accuracy problem. The problem is the model. I think you simply can't use the same characteristics to model a GNMA or any mortgage type (MBS) fund compared to other bond funds.

 

[Chuckanut]

Wow! Lots to chew on. Grazie a tutti!

 

[OrcasIslandBound]

I suspect that a ginny Mae fund has a different sensitivity to interest rates due to the portfolio of mortgage monies. Mortgages tend to get refinanced when interest rates drop. But they don't when rates rise. This is a bit like call options being exercised for callable bonds, I guess.

Any way, the effect is that this fund may act like a long term duration when rates rise, (hence the 10% drop in 2013) but when rates fall, the increase in the fund value is much more subdued. You probably will still get the benefit, but it will take a couple years.

Sent from my Nexus 4 using Early Retirement Forum mobile app

 

[Big_Hitter]

weighted maturity and duration do not coincide because the duration value is always changing for many reasons while maturity stays close to the same.

^ this - duration depends on the interest rate used to measure it

 

[Lsbcal]

...
Duration looks nice and low, so it appears as it has little interest rate sensitivity, but have the little taper tantrum of 2013 and all of a sudden duration increases markedly. I think it popped up over 5 years at one point, so fast it made your head spin. Too bad you can't get history of things like a fund's average duration and see how it really behaves in different interest rate environments!

This fund behaved very badly in 2013 compared to normal bond funds. It recovered spectacularly in 2014, but beware.

Good points. Why are bonds so darned complicated for what is suppose to be a safe investment alternative to equities but with lower expected returns?

Anyway, if I wanted to analyze this I'd look at the Yahoo historical price data for the fund (use adjusted price). Throw it in a spreadsheet and look at the volatility versus, say, the same price data for the Vanguard Intermediate Treasury fund.

Or M* has a fund chart compare tool.

 

[robertf57]

weighted maturity and duration do not coincide because the duration value is always changing for many reasons while maturity stays close to the same.

Although I am no expert, this is really because the models for "effective duration" (there are different formulas of duration varying in complexity with effective duration the more complicated duration) with mortgage instruments that can be retired at will or held to maturity will yield different durations as interest rates change. Think of a fund of 3.5% mortgages in an environment that quickly rises to a 4.5% prevailing rate. I would assume the "effective duration" of that fund is going to lengthen significantly while the technical weighted average maturity will be unchanged.

 

[audreyh1]

Good points. Why are bonds so darned complicated for what is suppose to be a safe investment alternative to equities but with lower expected returns?

Anyway, if I wanted to analyze this I'd look at the Yahoo historical price data for the fund (use adjusted price). Throw it in a spreadsheet and look at the volatility versus, say, the same price data for the Vanguard Intermediate Treasury fund.

Or M* has a fund chart compare tool.

It's GNMAs and MBS type bonds that are so complicated. The other types - not so much.

 

[audreyh1]

Although I am no expert, this is really because the models for "effective duration" (there are different formulas of duration varying in complexity with effective duration the more complicated duration) with mortgage instruments that can be retired at will or held to maturity will yield different durations as interest rates change. Think of a fund of 3.5% mortgages in an environment that quickly rises to a 4.5% prevailing rate. I would assume the "effective duration" of that fund is going to lengthen significantly while the technical weighted average maturity will be unchanged.

Yes, I observed this exact thing realtime in 2013 with FGMNX. Couldn't believe how fast that duration leapt up!