Sketches of an evolutionary constitutional economics

Gerald Frug has coined a nice term: The architecture of governance. He uses it to describe the vast array of service delivery arrangements at the local government level; the way that authority is allocated to a variety of different kinds of institutions – city neighborhood associations, city governments, private enterprise, and the like. See him speaking about this here.


The term architecture has also been applied to organisations; most notably by Carliss Baldwin and Kim Clark in their work on the power of modularity in organisation design. Their work has spawned an entire sub-genre of study on organisational structure and performance.

I am interested in meshing these two approaches into what might be termed a kind of ‘evolutionary constitutional economics’. In content, this is more closely aligned to Frug’s work; but in practice it would look a lot more like the Baldwin and Clark approach.

To get things started here are some passages from Frug that I have shamelessly edited/rearranged/co-opted to introduce the architecture metaphor:

The most serious design problem facing the world’s large social organisations, such as governments and nations, is the design of their governance system. Without an effective governance regime, no idea for institutional innovation and improvement can be implemented with any degree of success or confidence.

I suggest that we think of the construction of governance systems as a form of architecture. By governance, I am referring to the rules that determine who has power to solve any of the problems engendered by political and social life. I’m not talking about the content of any particular solution or any particular problem. Thinking about the governance system requires us to ask: who has the power to determine what the solutions to political problems are? And, equally importantly, who has the power to frame the rules that govern political and social life?

This brings us to architecture. If, at the most basic level, architecture deals with the design and fabrication of structures, we can think of governance as a particular kind of structure. The governance structure – the organisation of decision-making – is designed, and it can be designed well or badly. Of course, governance systems do not have an architect in the conventional sense of the term. But the people who design governance systems can be analogised to architects. Who are these designers?

Initially, they were the constitutional framers of any given nation. Today, depending on the country, they are either state or provincial legislators or members of the national legislature. These legislators are the people – circumscribed, to be sure, like architects, by others in the society, but nevertheless powerful – who determine what the governance structure looks like. To be sure, the upkeep and retrofitting of the governance structure is in the hands of different individuals than the original designers. But the same is true of architects – new people come in to redo and update what their predecessors have built.

As architects know, when designing a building, it is not easy to ensure that it can be retrofitted – many buildings seem to defy being re-used for different purposes. The only option, it seems, is to tear them down. Retrofitting is harder still for governance mechanisms. Many of our current government institutions were designed decades – even centuries – ago. Everyone realises that there are problems with how they work now. And no one thinks that a radical change in the whole system could happen all at once. The problem is that we do not have a way for re-designing the structure to be a regular, routine part of governance. So, instead of trying to redesign it, we create a series of additions to an otherwise unquestioned structure. At times it seems the only escape from this cycle is to begin to think about the architecture of governance and, then, step-by-step, government-by-government, work on redesigning it.

There is, however, another option: evolvable constitutions.

The “how to be a constitutional architect” conundrum dovetails with the questions of why constitutions matter and whether they really do evolve; whether they can or should be rationally designed or allowed to adapt in a process of spontaneous ordering. Most scholars think they have an angle on constitutional design – that is, as constitutional architects – that is rational because they frame the problem as axiomatic. They position themselves as constitutional architects and endeavour to draw up blueprints that are fitting for today, and forever. This makes constitutional analysis an exercise in logical reasoning (Locke, Kant, Mill, Rawls, etc.). The master social planer is the master architect; and essentially this amounts to an exercise in constitutional monument building.

Conversely, the very idea that constitutions really evolve can be found in Hayek (1960) (and elaborated in 1973abc). This line of argument is opposed to constructivist-rationalist design and advocates the decentralised use of dispersed knowledge and spontaneous ordering. Of course, it has evolutionary foundations, but a neglected contribution to Hayek’s research program is that no one has yet thought evolutionary mathematically about the problem. That is, no one has taken the computational-combinatorial view of constitutional design seriously. Most theorising starts with the ideal type, or is a justification of the scholars own vision of what the right constitution should be – and yes this is basically axiomatical. If, however, we proceed from a true appreciation of the overwhelming size and complexity of the ‘constitutional design space’, we might have profoundly less optimism in our ability to  rationally construct a fitting architecture of governance.

I propose something of a synthesis of the

  • Beinhocker-Kauffman-Dennett model of evolutionary search over fitness landscapes
  • Nelson-Winter-Dosi model of social technologies and paradigms
  • Dopfer-Foster-Potts rule-based micro-meso-macro framework

as the analytical basis for an evolutionary constitutional economics. This would proceed from the “library of Babel/Mendel/Lego” evolutionary parable as outlined in Eric Beinhocker’s The Origin of Wealth. Beinhocker calls the “business plans” equivalent to Babel/Mendel/Lego, an economic design space, the “library of Smith”. I suppose I might call the “governance plans” equivalent, an institutional design space, the “library of Hayek”. The central idea is that there is a mind-bogglingly large and complex institutional design space, which is a space of all possible rules (connecting people and resources in governments, organisations, society, etc.).

So the way I would like to re-pose the question is “how do you navigate the library of Hayek?” This is a (meta) constitutional question, about the correct amount of “unbundle-ability” of a state or constitution or institutional system, not necessarily the correct final configuration you arrive at. I think the “constitutional evolvability” question is also the “navigation” question, and is ultimately about also keeping an eye on the practical consequences of such thinking, which is about how to we get from here (bundled, layered) to there (unbundled, decomposed).

Then what is a constitution in relation to this library of Hayek – this institutional design space? A constitution dictates specifically some rules that must be and are followed (like the architecture of government), but it also leaves open space for individuals to originate and retain their own rules. Who: politicians, entrepreneurs, and citizens.

So the constitution actually cordons off parts of the massive institutional design space and opens spaces up for action – it reduces complexity and makes things slightly more tractable. The constitution says, “along these design dimensions the allele value must be this” when it dictates just how authority relations in the architecture of the state must be organised. Other times the constitution says, “along this design dimension the allele value can be between x and y”; i.e., it doesn’t dictate what a government policy is or what a company or person must do, but what they can do.

Do constitutions evolve? Indeed they do, in the sense that they change. Of course, this is a trivial insight; so perhaps a more pertinent question is: how do constitutions evolve? Moreover: can constitutions code for their own evolvability? Richard E. Wagner has argued that since constitutional designers are never in the mythical Rawls-Buchanan pre-society time we don’t have the option of changing constitutions prior to action but rather they come to be through action – and hence are spontaneously ordered. This is much they same as the trivial observation above that constitutions do actually evolve in the sense that they change and adapt; but importantly he also recognises the realities of ‘genuine’ complexity when it comes to constitutional design (perhaps a sort of parallel of Knightian uncertainty); and the fact that constitutions can only be ‘constitutive’, i.e., dictate a space of actions/rules and not a final allocation of actions/rules.

Constitutions cordon off the institutional space; the question is: what dimensions get cordoned off and over what allele values, i.e., how do you actually go about designing the constitution? Evolution or rational design/logical reasoning? When one attempts to design the blueprint for the architecture of governance, wouldn’t it makes sense to think about retrofit, renovation, and redesign?

We need not necessarily relinquish control of design completely to evolution (i.e., in a kind of second-order evolution where the rules of the game themselves evolve more freely than in piecemeal constitutional change). Rather, one might suggest that instead of letting loose with evolution at the constitutional level, we just free up some of the dimensions that have hitherto been constrained by constitutions; in this sense constitutional design emerges in a bottom-up fashion as a result of the actions of lower-level political entrepreneurs and experiments.

While there has been little formal thought ‘evolutionary-mathematically-combinatorially’ about constitutional design, fortunately, it is now possible to create a theoretical framework for doing this using new language and techniques of complexity theory that were previously unavailable to constitutional theorists. Such a research program would represent an exciting development in constitutional theorising: toward ‘evolutionary constitutional economics’.

In doing so we reiterate the importance of a constitution as a complexity-reducing device, while acknowledging that we need not reduce complexity simply by rigidly restricting constitutional design dimensions (i.e., drawing up the blueprints, enshrining them as an unquestioned structure, and sticking with them until the architecture requires major overhaul). Rather, we might take the logical reasoning aspect (burden, responsibility) and transfers it from a set of omniscient constitution framers or central planners (architects) to a swarm of social entrepreneurs and dispersed citizenry (deductive tinkerers).

Coming at this from an evolutionary economics perspective; this might involve a synthesis of the rule-based micro-meso-macro framework (Dopfer, Potts) and the NK fitness landscape model (Kauffman et al) for describing ‘institutional design spaces’ and studying different systems of governance. This research program opens up new ways of looking at some of the classic constitutional/governance questions, re: clubs, commons, federalism, FOCJ, etc. The ultimate goal is to compare the performance of relatively rigid constitutional structures with an ‘evolvable’ architecture. This would be explicitly seeking to develop an evolutionary methodology for constitutional economic analysis, and a return to Hayek’s theory of constitutional evolution.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s