This doesn't fully explain the underperformance of VFITX compared to a 3 year bond (which should have made 8 mo/12 mo * 2% = 1.3% in interest and lost around 0.7% on rising interest rates), a net gain of 0.6%.
So VFITX underperformed a three year bond by 1.2%. 0.13% of this is their management fee (0.2% * 8 mo/12 mo). I'm not able to explain the last 1% of difference.
However, in theory, a bond fund loses just as much value on an interest rate rise as the bonds it is holding lose. In my example above, the bond is worth ~$99 after the increase to a 3% rate, just like the fund. The only difference between the bond and the fund is the choice of when to liquidate or roll.
It's possible that VFITX got unlucky on the timing of their bond rolling (see cousin comment).
That said, I was very loose in my calculations. Without exact knowledge of their holdings and careful calculations, I'm not surprised that the numbers don't fully add up.
If you hold the bonds and/or VFITX instead, the interest pay out of the bonds and the distributions of VFITX should also come out equal (except not, the fund has the advantage that it can change its composition from buying/selling bonds, but also has the overhead of selecting and performing those transactions).
(In reference to your below comment, yes, fund != holding bonds. The fund is closer to you buying the bonds, but also buying/selling as bonds mature or you anticipate changes in rates)
But this is not the situation I described. My point is that buying the bonds in January would have been better than buying the bond fund in January.
If you look at mvilim's response to my comment you will see that the fund underperformed bonds between January and today.