Risk Complexity and Policy System Resilience: Effect of Crises, Discord and Political Agency to Policy Outcomes

This paper provides an assessment of key policy mechanisms affected by systemic resilience. The aim is to determine risk levels during contested policies where deliberation between political actors and other stakeholders shapes an agenda and provides the basis for building consensus. The complexity of this system entails a number of assumptions including that, actor’s autonomy of action varies in policy systems, reflecting their power; power is also manifest through influence networks; policy domains are typically interdependent; and external shocks can alter system properties (i.e. affect the terms of reference for the interaction between actors). From the perspective of individual agents, their resource investment in a policy system depends on the salience of an issue to their policy agenda. From the perspective of a system, the different phases of the policy process/cycle determine the level of activity, while the level of agitation is contingent to the proximity to key stages in the policy process, such as agenda setting or a decision/vote. The dearth of critical information in extant longitudinal datasets of policy making, (which typically map 2-3 stages in the policy process), provide limited opportunities in the analysis of these inherent complex policy systems. This study employs existing datasets as a baseline, and compares them with simulated deliberations of multistage synthetic data, during agenda setting-consultation-coalition building. The aim is to test a range of political risks, including system polarization, sensitivity to conflict during negotiations, system disintegration, and variance from ex ante preferences of policy actors (i.e. risk in the predictability of outcomes ex ante). The model employed is a hybrid network model, combined with an agent-based model. Modelling includes variables for issue salience, clustering level, network properties (reciprocity, transitivity, etc), tie maintenance costs, and agent power among others. Extreme events are also modelled such as external shocks, while agents influence one another via a DeGroot process. Also modelled is the presence of political entrepreneurs, defined here as exceptional agents who can optimise their network position depending on circumstance. Initial results indicate resilient political systems have significant and negative relationship to risks from policy shifts, and to shocks from conflict and clustering. Resilience has a positive relationship to entropy risks. Furthermore, the presence of political entrepreneurs is significantly affecting the probability of achieving policy consensus. This study confirms the sensitivity of political outcomes to policy system robustness and resilience.