The best books on social and economic collapse are from Tainter and Quigley. They are the best since they both provide systemic models of collapse that can apply across all historical instances. For example, Tainter sees civilization as problem solving system that collapses when it begins to succumb to diseconomies of complexity. Quigley sees civilization as a patchwork of wealth/solution creation engines that fail when they become self-serving institutions (organizations that perpetuate the problem they were built to solve) that fight renewal. Better yet, the models they propose are not mutually exclusive. They compliment each other, Tainter's model anticipates increasing turbulence (the crisis of 2008) and Quigley's anticipates systemic predation (like the financial oligarchy).
However, as good as these theoretical models are, they miss one critical insight. It occurred to me while reading these works in conjunction with Elinor Ostrom's work on self-managed resource commons (everything from fisheries to irrigation systems). The common finding is that every system is unique at the local level. No two irrigation systems are the same. No two fisheries are the same, etc. Further, the problems they represent are all rapidly changing, in that what's required to effectively manage them is divergent.
This implies that complex societies, over time, develop a yawning mismatch between the local and the general. More formally:
The need for evolutionary advances at the local level will always outstrip the pace of evolutionary change at the center. When the mismatch grows too large, the entire system collapses.
I suspect this applies not only to governance (politics, bureaucracy and legal systems) but also to markets, in that the centralization of markets works against the development of local wealth (scale requirements, predatory behavior, etc.).
Of course, this is also very similar to the problem I saw in John Boyd's decision making models (see "Open Decision Making" for more). It also implies that a workable solution to this problem (as well as the drivers of collapse outlined by Tainter and Quigley) is to return sovereignty and autonomy to the local, to the extent possible, to enable massively parallel decision making. The only way I can envision this working is through the establishment of resilient communities (aka, a semi-autonomous polis) that is economically self-sufficient, able to protect itself, and highly democratic.
NOTE: Wow, thinking deeply today. Hope you enjoy.