The temporal misalignment might not be so bad. Residential dwelling construction and public infrastructure should both lag the actual tax revenue.

I also thought this was bad until I gave it some thought recently. I'm not saying that I'm out of that camp completely because I think that the approach in some cases leads to overspending and may increase the overall cost in real, not just nominal, terms.

However, bonding out projects does spread the expenditure in time, more directly charging the future users for infrastructure they use. Sure you could recoup the costs through user fees, but then you're effectively borrowing money from the current tax base, so "bonds" by another name.