Can VMD improve the estimate of the muon g-2 ?

We show that a VMD based theoretical input allows for a significantly improved accuracy for the hadronic vacuum polarization of the photon which contributes to the theoretical estimate of the muon g-2. We also show that the only experimental piece of information in the $\tau$ decay which cannot be accounted for is the accepted value for ${\rm Br}(\tau \ra \pi \pi \nu_\tau)$, while the spectum lineshape is in agreement with expectations from $e^+ e^-$ annihilations.