Commonalities are striking. Commonalities here include questions about how coordination among concurrent processes and multiple agents in distributed systems is possible. Common questions with potentially cross-pollinating answers aside, lots of important phenomena in socioeconomic systems and in biological systems (e.g. mitosis in frogs) look strikingly similar. A monostable system in biology looks like a single-equilibrium supply-demand curve in economics. A bistable biological system looks like a coordination game in game theory: there are two equilibria with some questions about what it takes for the system to switch states. The distinction between steadiness and stability also matters in both domains: it is possible for a biological system, for example, to be in an unstable steady state instead of a stable steady state. Local perturbations sometimes leave a system in its current state but other times result in what one might call a phase transition -- a change to another static or even dynamic equilibrium and thus a global change in state, with coordination communicated and global change propagated via local interactions. Plus there are lots of other interesting similarities. This is more than a striking metaphor or colorful analogy. Rather, even though cells aren't people and people aren't cells, this points to what biological and socioeconomic (and likely also physical) systems have in common: they are complex, modular, and have dynamics that are sometimes linear but oftentimes non-linear (or at important times even if infrequently).
Differences are also important. The fact that people are cognitive agents (i.e. have cognition and agency) is an important difference in our understanding of socioeconomic systems, as well as in guiding us in our efforts to continue to improve the condition of humankind and in our attempts to nudge our own behavior in a direction that is better socially but yet does not cross the line into the morally objectionable or even the morally impermissible. On the one hand, we are rational animal; on the other hand, for efficiency's sake, our cognitive system takes shortcuts which, under unfavorable conditions, result in cognitive biases that make us predictably irrational (to borrow Dan Ariely's catchy turn of phrase). Moreover, we often mean and understand different things and form very different expectations even though we use the same exact words. Sometimes a difference in meaning is not serious but other times it runs deep and stems from our not having thought about some important distinction (like the difference between a relative reduction in risk and an absolute one), from us having divergent perspectives (like that of a reasonable physician vs. a reasonable patient -- a concept of a reasonable person here is insufficient), and sometimes even from our having very different ideas about the good and the right (e.g. divergent conceptions of justice).
In my work, I have strived to pay attention to both: what makes socioeconomic systems similar to biological and other systems and what also makes us different and unique. Below you will find my work divided into two categories: processes and patterns. What I am ultimately interested in, of course, is the interplay between the two.
Risk behavior: conceptual analysis of moral hazards, the human propensity to increase exposure to risk when insured. Of course propensities do not always obtain. But when moral hazards do obtain, they can be found in insurance markets, be it private insurance or social insurance, narrowly or broadly construed. I argue that, contrary to what has been argued in the past, moral hazards are not morally neutral even if they are not inherently unethical. That's because moral hazards determine rather directly the allocation of benefits and burdens and the allocation of benefits and burdens is not morally neutral as it always requires moral justification (hence the field of distributive justice). On the philosophical end, this argument reminds us that "not inherently unethical" does not entail "morally neutral". On the behavioral economics end, this analysis brushes up against interesting issues at the intersection of behavioral economics and economics of information. (Braynen, 2013, Cambridge University Press) related topic: risk communication
Opportunities as information rather than information about opportunities. (I recommend skipping right to the Positive Thesis portion of this dissertation introduction.) This is the part of my research which, in my analysis of the socioeconomic system, reaches beyond what economists call "price signaling", as does the next piece below. (Braynen 2012, dissertation.)
The internal dynamics of social structures and their effect on the distribution of mental states, motivational and epistemic (e.g. beliefs about their place in the distribution of talents), which in turn affects the payoffs the agents get as a result from their actions. This is a proof of concept that bad outcomes in the form of unmet expectations as returns on investment of talent can obtain without any wrongdoing in the form of deception or lying purely. Specifically, this is about outcomes that result from nonpecuniary signals that the configuration of a social structure emits to the agents who make up that structure. It is also a proof of concept that good outcomes can arise without any explicit right-doing the same exact way. This is similar to how price signaling might work, except that instead of price signaling the mechanism in this case is the internal dynamics of social structures and its role in "self-organizing" behavior. (This is a dissertation chapter from Braynen (2012), not currently posted.)
Lighthouse: creating an alternative to rankings and, also, giving users control over their own frames and other cognitive biases. This was was partially developed as a response to the book Nudge (2008), as well as to the use of framing effects, anchoring, and other psychological tricks in marketing, sales and negotiation. (Built by Braynen with Hutchins.)
Genotype or Phenotype? The conflation of two concepts in evolutionary agent-based modeling. The two distinct concepts of genotype and phenotype are often conflated by Neo-Darwinian evolution theory, as well as by agent-based models that use evolutionary principles derived from that theory (e.g. in evolutionary game-theoretic models). In evolutionary game-theoretic models, the concept of genotype is analogous to an agent's strategy while the concept of phenotype is analogous to an agent's behavior. Strategy and behavior, however, are often treated as equivalent concepts. Using a spatialized prisoner's dilemma, we show this conflation is a non-trivial simplification that may lead to drastically different outcomes. Additionally, we discuss implications for using the evolutionary paradigm to model market interactions.
Reducing Prejudice: A Spatialized Game-Theoretic Model for the Contact Hypothesis. Software for the simulations can be found here (Java and C/C++/OpenGL) and here (NetLogo).
Visual Distribuenda: using divergent conceptions of distributive justice to benchmark distributions of goods and bads. This was developed as a pedagogical tool and could in theory also be used to explore and benchmark real data. (Built by Braynen with Hutchins.)
Deviating from equal distributions: using KL-Divergence to measure the distance between an actual distribution of goods or responsibilities (not probabilities!) on the one hand and a relevant benchmark, e.g. the ideal distribution, on the other. The philosophical argument is a response to the so-called "leveling-down objection" against egalitarianism: if the ideal distribution is indeed an equal distribution, that does not mean that any equal distribution is better (e.g. less unjust) than any unequal distribution. This work could easily be extended to goods and responsibilities that obtain with some probability, although that is not necessary for the main argument. (Co-authored by Christiano and Braynen).
Game-theoretic robustness: a graphic volumetric measure. Game-theoretic robustness is a concern that, for us, arose in agent-based models that (a) use evolutionary game theory as their transformation function and which, as a result, (b) exhibit complex systems behavior and so have tipping points. Our graphic measure allows to measure and also provides a visualization for how robust the result is. If, for example, there are lots of tipping points and the simulation's results are highly sensitive to initial conditions (and, for instance, we do not know which of these model assumtions are more likely than others), then the result is not very robust at all. But if, at the other extreme, there are no tipping points at all, then the result is optimally robust. We show you a new way of thinking about the cases in-between. (Built by Braynen, co-authored with Grim et al. and published in Synthese)
Our mission is two-fold.
Firstly, it is to enhance our understanding of the various ways in which the knowledge on which people base their plans is communicated to them in epistemically-diverse and dynamic adaptive socioeconomic systems.
Secondly, it is to test and capture our theoretical insights by creating prototypes which improve the quality of decision making and the quality of informed choice without increasing the unit cost of information, thereby empowering people to miscommunicate less and coordinate more.