I often discuss institutions and transaction costs here, but it was only recently that I read a very useful paper [PDF] by Oliver Williamson, who won the Nobel Prize* for his work on transaction costs.
I was pleased to see -- contrary to what I'd been told -- that Williamson explains institutions quite clearly. The following diagram appears on page 597:
In it, Williamson makes a useful clarification of the different types of institutions. At the top level (L1) are institutions such as culture that last for hundreds to thousands of years. Then there are formal rules (e.g., a constitution) that will be constrained by culture but last for decades or more. Next (L3) comes governance in how the game is run.** These are various laws and administrative rules that are constrained by informal culture and norms and formal constitutional bounds. Finally, we get to L4, the ongoing process of allocating resources (water, labor, capital, time, etc.) in accordance with out individual goals.
This schema helps clarify the ideas that we discuss here (and in my book) for reforming L3 governance or affecting L4 actions, even while taking formal rules and informal norms into account. That's why, e.g., I propose all-in-auctions as a way of reallocating water (given rights in L3) or redistributing rights (given norms in L1 and constitutional definitions in L2) when it comes to a human right to water.
Bottom Line: Institutions matter, and it's useful to know which institutions you want to change/respect. Read this paper.
* Yes, it's not a real Nobel. Whatever.
** Also see this review of The Calculus of Consent (on constitutions).