In Stata, it's possible to fit SEM models with longitudinal data, making syntax much shorter (implicitly constraining coefficients to be identical across time of course). Making this possible in lavaan would be great!
For example, instead of
dem2 ~ dem1
dem3 ~ dem2
dem4 ~ dem3
dem5 ~ dem4
...

we can write:
dem ~ lag(dem)

which is much easier and less error-prone, especially if there are many time-periods.