He also has a good discussion about the implications of the words we use to talk about policy.
What I love about Bob's crusade against the LIHTC is that he really gets into issues of policy design. A lot of people see a problem - say, with low income housing - and assume that just putting money into things related to that problem will solve it. Sometimes you have to go a little deeper and think about the implications of taking a supply side approach or a demand side approach, thinking about eligibility rules, thinking about implementation, etc.
This is a pretty straightforward "a demand-side approach will work better than a supply-side approach argument".
You see this all the time in the science and engineering workforce discussion too. Yes, there are good reasons for public investments in science. But the problems are largely on the demand side and not the supply side. For some reason, some people still seem intent on throwing money at the problem without thinking of policy design and end up supporting the most obvious policy intervention (in the case of science policy, it's also often a supply-side solution).
There's nothing special about the demand side of course. Bob takes a demand side approach on low income housing policy, I take a demand side approach on science workforce policy, and I also think on the demand side when we talk about recessions here too. That's just a coincidence - there's a lot of potentially good policy on the supply side too.