Why does vacuum have a nonzero characteristic impedance towards electromagnetic radiation? Intrinsic impedance given by $\eta=\sqrt{j\omega\mu / (\sigma +j \omega \epsilon)}$. It gives slope of transformation of $\mathbf E$ to $\mathbf H$ and vice versa. Here $\eta$ is complex. And in this expression real part is the cause of attenuation and imaginary part is the cause of phase shift.
In case of free space since $\sigma = 0$, we have $\eta = \sqrt{j\omega\mu / j\omega\epsilon} = \sqrt{\mu / \epsilon}$, which is real. This suggests presence of resistive part in intrinsic impedance which means there should be attenuation. Also curiosity is how free space can offer resistance and however, the expression for electric field in plane wave $\mathbf E = E_0 \exp(wt-\beta z)$ where $\beta =2 {\pi}/{\lambda} $ and $\lambda $: wavelength
suggests constant electric field. How can we reconcile the real impedance of space with the expression for electric field, which has no attenuation?