In Complexity and the Art of Public Policy, David Colander and Roland Kupers introduce the “complexity frame” of policy-making, and argue for bottom-up policy-making like nudging individuals to make decisions more in line with the goals of the public choosers.
In Part 1, “The Complexity Frame for Policy,” Colander and Kupers reduce the old policy debate to one of methodology. Social scientists, they say, have been raised on a strict diet of equilibrium theory and market failure. Those social scientists who accept the market failure hypothesis use equilibrium theory wielded by well-meaning apolitical interventionists as a panacea to all mixed-economic ills. Those social scientists who do not accept the market failure hypothesis use equilibrium theory wielded by marginalist freedom-lovers as a panacea to all mixed-economic ills. Colander and Kupers demarcate their position in this starkly painted landscape as strictly not one (control) or the other (market anarchism), for methodological and practical reasons.
The complexity frame is neutral regarding the level of government, for it doesn’t make assumptions about the nature of government. It does assume, like my professor Richard Wagner likes to say, that “Government is with us, and will probably always be with us.” The nature of government will likely change as a result of the implications of doing policy in the complexity frame. How it will change as an institution – how it is formed, what incentives must propel public choosers compared to now – is expressed somewhat in future parts of the book, but never satisfactorily, in my view. What new things government should do to change social outcomes are elaborated on with greater detail.
The complexity frame as set forth in the book is similar in every way to how social systems have been expressed as complex adaptive systems over the past 20 years especially. Equilibrium theory is described as an inappropriate tool with which to investigate market behavior for many reasons, not the least of which is that equilibrium theory is a static theory in a reality of constant change. The main takeaway is that trends, like the statistics we associate with macroeconomic variables, are generated from underlying, bottom-up processes. As such, the authors claim, policy measures should strive to be as bottom-up as possible, to subvert or mimic natural social processes like shifts in norms.
In Part 2, “Exploring the Foundations,” Colander and Kupers work hard to separate their approach to doing economics from more traditional approaches. For the most part, their methodological case is solid. Where they come in weak is in detailing the goals that bottom-up policy should strive towards. They shower praise upon J. S. Mill, Smith, Pigou, Marshall, and Keynes, and signal disapproval of Milton Friedman, George Stigler, Leonard Read, and Don Boudreaux. Their choice of heroes and anti-heroes colors the book with an ideological tint that I believe obscures their overall message of bottom-up change.
Influencing norms is the vehicle for the authors’ bottom-up policy change (“norms” appears 100 times in the book). The authors suggest that individual decision-making isn’t based on utility and other kinds calculations, rather, experimental and behavioral economics shows that a good deal of decision-making uses heuristics and shortcuts like adhering to prevailing norms. Changing how people make decisions, therefore, is how Colander and Kupers want to affect real and lasting change in the emergent patterns of the social system. Changing norms as a way of changing emergent patterns of social action is not unlike the famous Smithian point about how the chess master moves chess pieces without a thought to whether the internal laws of motion of the chess pieces are in line with the movements forced upon them; in Smith’s view, government works best when these two movements align.
Further, according to the authors, policy effectiveness can’t be measured simply as a function of what they call “material welfare” (p. 87). “Social welfare” is itself an important goal. Sadly, they fail to express how social welfare is supposed to be measured as a part of policy effectiveness accounting, and how the determination of social welfare itself might be colored by the ideology of the authors and, more importantly, the policy-makers. The power of using “material” measurements as a metric for social welfare is their disassociation with any particular subjective measure of good. The authors do not adequately find a replacement metric that isn’t heavily subjective.
The new “policy patterns” suggested by the authors are a laundry list of systems theoretical phenomena: nonlinearity (p. 113), emergence (p. 116), multiple equilibria (p. 117), path dependence and lock-in (p. 118), phase transitions (p. 121), diversity (p. 121), power laws (p. 123), networks (p. 125), and agent-based modeling (p. 127). Multiple equilibria and lock-in are the focus of their ensuing policy suggestions. The authors also rely heavily on game theoretical constructions, as a kind of panacea to traditional utility maximization equilibrium theorizing. The problem with this being is that in general, finding a Nash equilibrium is an NP-hard problem. So while the authors in many places indicate they are aware of the mathematical challenges facing complex system theorizing, they aren’t consistent in their modeling and analysis advice.
In Chapter 9 of Part 2, the authors bring forward a scheme of “nudging” individual behaviors and outcomes. Norms are the primary target of nudges, the authors describe, but nudging is really about manipulating decision-making heuristics. As described above, people don’t always or perhaps even usually utilize maximization calculations when making decisions. Often they choose based on some set of heuristics. The power of policy is to exploit heuristics to steer people to decisions deemed overall “better” for society by some metric of social welfare.
If you’re mystified as how these nudges as such are dissimilar from the kind of top-down control governments have always tried to enact, so am I. The heart of top-down control isn’t in the method prescribed to actualize goals, it’s in the determination by a governing group of public choosers that some goals are better than others, without sufficient proof that these goals are calculably better (as the problem is fundamentally subjective and thus incalculable).
In Parts 3 and 4, Colander and Kupers suggest that while governance structures are created from the bottom-up, these bottom-up structures are somehow free of the apparent behavioral flaws of the individuals whose decisions led to the emergence of these governance structures. To suggest that one (the private individual) is driven by self-gain while the other (the public chooser) is driven by social welfare ignores how most governance structures have traditionally arose. It also ignores the numerous non-governance organizations that arose to promote social welfare throughout history.
The authors introduce the idea of a “for-benefit corporation,” which is pretty much like an old-fashioned fraternal organization that is blessed and nudged into existence by favorable (top-down) policies. They do not examine why fraternal and other charitable organizations vanished in the first place, instead leaning on their implicit assumption throughout the book of an almost Pigovian-like market failure theory of private lock-in due to suboptimal norms and individual decision-making heuristics.
More egregious than their academic oversights and ideological biases, however, is their waving-away of economic truths that are true regardless of the complexity of the social system. The primary truth ignored is the recognition that resources arrogated to government have alternative uses in the private sphere. The authors list many examples of nudging policy in which they explicitly encourage a trial-and-error method of blank checks written to bureaucrats for jolting social systems out of states those same bureaucrats deem as “locked-in” to some suboptimal “basin of attraction” (sounds like Pigovian welfare theory, doesn’t it?). How officials are able to calculate which methods will jolt society in which direction, and what the more optimal basin of attraction is, remains a mystery unexplained in this text.
I saw little indication of new theory in the book. Most of what I saw was an attempt to re-package Pigovian social welfare theory and utilitarianism in complex systems theoretical trappings. It ignores what I see is clearly the biggest hurdle in policy-making in a complexity regime: how government as we know it is premised on a group of people with the power to arrest and tax being able to dream up goals that will make everyone else better off than in that group’s absence.
I enjoyed the introduction of complexity ideas to the realm of public policy-making. I agreed that the top-down “control” method is a naive view of how policy changes (or should change) outcomes.
My main problem with the text is that although it rejects the ideas of equilibrium theory and Pigovian welfare economics in a traditional sense it fully embraces them in a modern sense, without an acknowledgment of why rejection of one does not imply the rejection of both. For example, some of the most-used terms in the book are: 1) lock-in (used 45 times), 2) nudge/nudging (56 times), 3) basin (of attraction) (19 times). These ideas are not qualitatively unlike 1) price/wage/variable “stickiness,” 2) short-term intervention to jolt a system out of a suboptimal sticky point or long-term intervention to prevent from ever getting in the suboptimal sticky place, 3) multiple equilibria, some of which are associated with suboptimal states where everyone could be better off if they were somehow jolted towards a different point on whatever plane of relationships is being considered.
I do not think it is useful to recast complexity-aware social welfare theory as such a close cousin to social welfare theory’s first iteration. It discards in large part the power and insight of complex systems theory, and does not challenge the policy-making status quo nearly as heartily as it deserves.