Estimating Vulnerability: Endogenous Modeling of Governance-Economy-Conflict

• In Event: Climate Change and Conflict

Abstract

Impacts from extreme climate and weather events depend fundamentally on the vulnerability of exposed societies. Vulnerability, in turn, is shaped largely by political and socioeconomic conditions. It is well known that political and socioeconomic factors are interdependent; most fragile and conflict-affected countries are poor whereas politically well-functioning and peaceful countries also tend to enjoy high average incomes. Put differently, poor economic performance threatens political stability whereas political instability hurts the economy, and both outcomes have adverse implications for societies’ capacity to cope with extreme climate events. In some contrast, quantitative models of economic growth, democracy, and civil conflict – as well as those used to quantify country-level vulnerabilities to climate change – typically control for prevailing socioeconomic and political contexts but they do not dynamically analyze this interdependent system, reducing the ability to produce good long-term forecasts. Here, we present probabilistic models of GDP, democracy, and armed conflict, built from state-of-the-art research on forecasting of these outcomes. The models are dynamically simulated over time to produce forecasts of each outcome, using the other two outcome variables as covariates. We evaluate the integrated forecasts over various time horizons and compare them to single (conventional) variable models, models where true outcomes are mixed in, and the out-of-sample observed outcomes. In a subsequent step, we investigate the ability of the endogenous system to forecast long-term trends in human cost of extreme climate events relative to the performance of projections derived from conventional models.

Authors

• Jonas Vestby, PRIO

• Håvard Hegre, Peace Research Institute Oslo

• Halvard Buhaug, Peace Research Institute Oslo (PRIO)