We think, and perhaps Sachs disagrees, a framework that says there are 17 factors, each of them hugely important is no framework at all. The power of a framework comes from its ability to focus on the most important elements at the exclusion of the rest and in so doing in providing a way of thinking about these elements, how they function, how they have come about, and how they change. For us those elements were related to institutions and politics, and we have focused on them.I think A&R are mostly wrong about this. Simple models of the world certainly have an important place in social science. After all, the simple models of neoclassical economics explain quite a bit (though not nearly all) of the human behavior we observe; it gives us a lot of bang for the buck, so to speak. But simple models take us only so far. As Elinor Ostrom continually reminded us, the world is a complex place, and it is populated by complex human actors. To understand how humans interact with one another (not just as governors and citizens) and with the ecological systems they inhabit, scholars need to embrace complexity, not reject it in favor of simple (and simplistic) nostrums. See, for example, Lin's 2005 book Understanding Institutional Diversity, her Nobel Prize address (here) and her more recent articles, e.g., here, and here, elaborating a Social-Ecological Systems Framework comprised of approximately four dozen variables.
If A&R are right that a framework with 17 variables cannot be a framework, then everything Lin Ostrom thought she was doing since the publication of her landmark study, Governing the Commons (1990), was a fool's errand. Heck, she didn't even understand what constitutes a proper analytical "framework." But Lin was no fool, and she had a far richer and better grounded understanding of analytical frameworks and models than A&R display. Indeed, Lin's approach to solving collective action problems to improve material welfare of individuals and the social and ecological communities they inhabit is not only more ambitious and hopeful than A&R's simple dichotomy of "extractive" and "inclusive" institutions (which is redolent of recent political distinctions between "producers" and "moochers"); it is also more productive.
A&R's simple model hardly advances the ball beyond work done over the last generation by Douglass North and other political-economic historians. As far as I can tell, it basically relabels what North, Wallis, and Weingast refer to as "natural states" v. "open-access societies," with a bulk of additional stories to support the simple dichotomous model. If anything, Acemoglu and Robinson's approach holds even less hope for fundamental institutional change in "extractive" societies than North's approach. Indeed, reading Acemoglu and Robinson, one might easily conclude that the emergence of "inclusive institutions" is mostly a matter of historical luck or accident, rather than the conscious decisions of human actors interacting in various constitutional, collective-action, and operational arenas. In their "framework," the UK and France are both "inclusive" societies, and little if any basis exists for assessing (or preferring) their very different roads to that end. As for "extractive" societies, it seems A&R have little to offer but the sage advice to become "inclusive" by extending and protecting property rights, freeing markets (with reliable contract enforcement), and establishing the rule of law. What is novel about any of that? And yet, A&R don't seem to appreciate the complexities involved in doing any of those things. As Lin showed in Governing the Commons (and many other works), establishing a property regime is one thing; establishing one that actually works in the local circumstances is quite another.
A focus on a wider array of lower-level variables can at least indicate areas where marginal improvements might be politically feasible (via collective action). And marginal improvements in one area might facilitate further improvements in other areas, leading over time to larger-scale institutional change. Moreover, with a wider variety of variables, it is possible to construct a wider variety of specific models (under the framework) that can be tested in experiments and in the field, and to code data more precisely for purposes of more precise quantitative analysis, which would be a valuable addition to the thickly descriptive approach, albeit one focused exclusively on higher state structures, A&R take in their book.