The Choice consists of an interdisciplinary group of scholars focusing on how institutions influence allocations, shape our decision-making, and organize how we relate to one another politically and economically. Within any domain, be it health or education, law or the economy, a society must choose among institutional forms to distribute resources, assign power, or otherwise structure the way that we interact with one another and pursue our interests. Those institutional choices matter for their specific domains and also more broadly.
Our inquiry focuses on five core questions related to those choices:
How societies choose and design individual institutions
why do we have markets for cars and democracy for judges?
(2) How technological advances and demographic changes influence these choices
the rise of markets to structure employer-employee relationships; algorithms for criminal sentencing
(3) How nested and linked institutions produce robust outcomes
eg, the functionality of boards of directors atop hierarchies and federalism as a marketplace for citizens' ideas
(4) How institutional ensembles contribute to civic capacity
how communities build trust and order
(5) How civic capacity limits or enables institutional choices and forms
why development projects fail in some countries; how norms can constrain or enable innovation
As a starting point, we assume that societies choose between five pure institutional types to make allocations and decisions:
Market: exchange of goods, services, information, and knowledge by assigning property rights and using a price mechanism
Hierarchy: an organizational structure that assigns roles, quantities, incentives, and responsibilities, and resolve disputes through procedures and authority relations
Democracy: deliberation and aggregation of opinion through consensus or a voting mechanism
Community/Network: self-organized behaviors and connections in which coordination and cooperation are enforced through norms
Algorithm: rule-based aggregation or allocation based on information extracted from the environment revealed information from participants
These pure types have well-defined attributes. Theory, empirical research, and case studies provide guidance as to when to choose each type of institution. If we care about equality and general welfare or if a decision impacts everyone, we often choose a democratic institution - we vote on defense spending and some laws. When allocating private goods that have no externalities, we care less about inequality and wish to unleash innovation, so we choose markets. Thus, we have markets for food.
Markets do not work for all such tasks. If a task requires coordination among uncertain tasks or frequent adaptations, such as building an airplane, we choose a hierarchy instead. Hierarchies facilitate coordination but can have substantial transaction costs. If those costs of setting up a formal institution outweigh the benefits, such is the case with time allocation within many organizations and definitely within families, we choose to self-organize. Finally, advances in technology now allow people to send information to a central source where an algorithm makes an allocation. Algorithms have long been used for allocating medical residencies but are now used for assigning routes to cars and buses.
Below are some attributes and features of environments in which each institution thrives (adapted from Stinchcombe 1985).
resources, goods, and services can be evaluated in monetary terms or another medium of exchange
competition among buyers and sellers: important exchanges and actions involve multiple sellers and buyers
access to law (or coercion) equal across actors
long-lived firms to guarantee efficiency
decisions determined by efficiency not welfare or justice
reliable information about goods delivered later or at long distances (standardization)
incentives for innovation
common knowledge and agreement on opportunity, responsibility, e.g. common defense, schooling
participation equal within boundaries (citizenship)
applies to task where agreement needed: national policies
open discussion to improve actions and decisions
good at finding best decision if there is one
authority to implement decisions
elected leadership
written constitution
activities that require adjustments or that make specifying performance in advance is difficult
resolution of conflicts through authority relations
establish standard operating procedures
incentives to build organization specific skills and relationships
no formal roles
coordinated and cooperative behaviors emerge
norms and punishments reinforce behaviors
connections and networks adapt to the task
individuals send information within protocols
one shot revelation - no interaction
decisions made by rules applied to received information
algorithm can also generate reputations
By thinking in terms of these pure types and their uses, we necessarily err given the messiness and complexity of the real world. Actual institutions do not map neatly into types. And a domain may seem ideally suited for one type of institution, yet we may choose another. Thus, empirical reality reveals double muddle.
First, impure types abound. A hierarchical organization may be run by market principles or make decisions democratically. A mostly self-organized community may include authority relations. Those authority relations may be emergent, that is they may not be written down in an organizational chart or constitution but they may be just as real in practice. Further, algorithms may rely on implicit or explicit markets or democracies. An algorithm could sell to the highest bidder or decide by a majority vote. We consider something to be an algorithm if individuals send a single message, piece or vector of information, and a rule then makes a decision or allocation. An online deliberative democracy or an online market such as eBay would not satisfy that definition.
Second, the boundaries between where we should expect to see each type of institution prove far more complicated and nuanced than theory would predict. That is especially true for the case of markets and hierarchies (Stinchcombe 1985). Activities that one might think would be far better accomplished by a hierarchy, might be instead be handled through a sophisticated system of networks.
Institutional analysis traditionally focuses on allocative outcomes: are they efficient, equitable, and sustainable? We choose one institution over another because we think it more appropriate for that domain. As noted above, we use markets to allocate cars, democracies to decide on tax rates, hierarchies to build airplanes, algorithms to assign work rooms, and communities of people to self-organize where they go to dinner.
Institutions influence the behaviors, beliefs, norms, information, knowledge (explicit and tacit) and networks. Evaluating institutional types in isolation ignores spillovers between institutions and culture. These spillover effects imply that a collection of institutions, each suited to its own domain, could produce a selfish, disconnected, uninformed population with few of the tacit skills and social norms necessary to flourish.
Thus, the choice over institutions requires an ensemble approach that transforms the choice among institutions from a sequence of isolated decisions – should we have school vouchers, government provided health care, markets for carbon, an algorithm to assign access to roadways - to a multidimensional problem. In our stylized framing, a society chooses among five institutional types: market (M), democracy (D), hierarchy (H), self-organized community (S) or algorithm (A). We refer to this as an institutional ensemble.
The success of a society's institutions to enable flourishing depends in part on that society's cultural capacity. However, as already noted, that civic capacity will depend upon the institutional ensemble. That interplay - institutional performance depends on civic capacity, while civic capacity depends on institutions - lies at the core of our inquiry. The interplay implies that making an optimal choice of institution in each domain could result in a suboptimal ensemble of institutional choices if those one shot "optimal" choices do not build sufficient civic capacity.
By civic capacity we mean the information and knowledge, behavioral repertoires and norms, belief systems, and social networks that characterize a population or society. Markets promote specialized knowledge acquisition, strategic, often self-interested behavior, beliefs in rationality and the self-interest of others, and symbiotic network connections. Democracy promotes general knowledge, norms of reciprocity, beliefs in collective interest, and dense social networks. Hierarchy promotes within firm specialized knowledge, rule following, trust, and networks organized by position. Self organized communities promote common knowledge of processes, high degrees of trust, beliefs in equality, and rich social networks. Algorithms encourage self reflection, self interest, and a trust in explicit rules, but may not build new social connections.
Civic capacity spillovers could, for example, mean that each type of institution creates promoting spillovers for its own type (Dolan and Galizzi 2015). We represent these promoting spillovers with arrows from each institutional type pointing back to itself. Promoting spillovers can occur across institutional types. The figure below shows promoting spillovers from democracies to markets and self- organized communities. Mutually promoting spillovers between markets and algorithms exist in the form of markets creating information that algorithms can exploit (Fourcade and Healy 2017). There also likely exist promoting spillovers between self organized communities and democracies in the form of prosocial norms. Spillovers can also be purging. They can hinder the performance of other institutions. Democracies, by creating a norm of equality may make hierarchies, which limit voice and rely on authority relations less effective. The figure show this negative spillover with a grey arrow.
The institutional DNA metaphor implies a string of individual institutions. Within many domains, multiple institutions may be nested. This nesting allows the strengths of the various types of institutions to make up for weaknesses in he other types. As shown in the figure below, a hierarchical organization may have a board of directors (a democracy) that provides general guidance. Within the hierarchy, algorithms may be used for staffing and data storage, markets may be used for non shared staff allocation and internal pricing of some resources, and conference room and vacation scheduling may be allowed to self organize.
Taking the wide view and thinking of a society as comprised of an ensemble of nested institutions offers an opportunity. Institutional choices and designs can increase civic capacity. Acemoglu and Robinson (2019) argue that there exists a narrow corridor that strikes a balance between state power and individual freedom within which societies can grow civic capacity.
Readings:
Acemoglu and Robison 2019. The Narrow Corridor: States, Societies, and the Fate of Liberty. Penguin Random House, New York.
Bednar, Jenna. 2019. Podcast. "States As Laboratories for Policy Experimentation: In Conversation with Jenna Bednar." Centre for the Study of Governance and Society.
Bednar, Jenna and Scott E. Page. 2018. "When Order Affects Performance: Culture, Behavioral Spillovers, and Institutional Path Dependence." American Political Science Review 112(1):82-98.
Davis, Jerry. 2016. The Vanishing American Corporation: Navagating the Hazards of A New Economy
Dolan, Paul and Matteo Galizzi 2015. ``Like ripples on a pond: Behavioral spillovers and their implications for research and policy'' Journal of Economic Psychology Volume 47, Pages 1-16
Marion Fourcade, Kieran Healy, 2017. ``Seeing Like a mMarket", Socio-Economic Review, Volume 15, Issue 1, JanuaryPages 9–29, https://doi.org/10.1093/ser/mww033
Greany and Rob Higham, 2018 Hierarchy, Markets and Networks Analysing the ‘self-improving school-led system’ agenda in England and the implications for schools, UCL Press. London.
Malone, Thomas Superminds: The Surprising Power of People and Computers Thinking Together
Stinchcome, Arthur. 1985 "Contracts as Hierarchical Documents" in Organization Theory and Project Management (A.L. Stinchcombe and C.A. Heimer eds). Norwegian University Press and Oxford University Press. Oslo and London.