Projects

Every month, predictions are generated using the CRISP model}. Currently, CRISP forecasts rebellion, insurgency, ethnic and religious violence, as well as domestic and international crises. CRISP is a suite of programs to aid CRISis Predictions. This report was created with CRISP package version 2013.12.16. The predictions identify onsets, as well as ongoing conflicts, and also provide estimates of the intensity of these events. To do so we construct (at least) three kinds of models for each event of interest. These models have a monthly time frame, and are applied to a population of 167 countries around the world. These include 1) a hierarchical model of whether an event of interest is present, 2) a count model of the number of incidences of conflict related to the event of interest, and 3) a split duration, event based model that generates an independent estimate of the onset of new events. This report highlights some of these predictions, made in November of 2013. Herein the actual models are very briefly discussed. Instead, focus is on predictions of continuations and onsets of these five classes of events

How do social networks among anti-government actors affect the decision of ruling authorities to challenge its opposition? Current literature focuses on the dyadic relationship between the government and potential challengers. We shift the focus toward exploring how network structures affect the strategic behavior of political actors. We derive and examine testable hypotheses using latent space analysis to infer actors' positions vis-a-vis each other in the network. Network structure is examined and used to test our hypotheses with data on conflicts in Thailand 1997-2010. We show the influential role of network stability in generating conflictual behavior.

The gravity model, long the empirical workhorse for modeling international trade, ignores network dependencies in bilateral trade data, instead assuming that dyadic trade is independent, conditional on a hierarchy of covariates over country, time, and dyad.   We argue that there are theoretical reasons as well as empirical reasons to expect network dependencies in international trade. Consequently standard gravity models are empirically inadequate.  We combine a gravity model specification with "latent space" networks to develop a dynamic mixture model for real-valued directed graphs.  The model incorporates network dependencies in both trade incidence and trade volumes at both levels simultaneously.  We estimate this model using bilateral trade data from 1990-2008.  The model substantially outperforms standard accounts in terms of both in- and out-of-sample predictive heuristics.  We illustrate the model's usefulness by tracking trading propensities between the USA and China.

Quantitative International Relations scholarship has focused on analysis of the so-called dyad. Few studies have given serious thought to the definition of a dyad, nor to the implications that follow from such a conceptualization. This piece argues that dyadic analysis is necessarily incomplete and even when rigorously pursued gives incomplete and at times incoherent pictures of the ebb and flow of interactions among actors in global politics and economics. Much of this myopia could be attributed in prior scholarship to the paucity of data, a defense no longer plausible.

For more than two decades, political scientist have created statistical models aimed at generating out-of-sample predictions of the popular vote in presidential elections. This exercise aims to develop the “best” model. Our approach is different. Rather than creating the best model or theory, instead we create an ensemble of prediction of the top ten models, and use that ensemble to produce a prediction of the current election, weighting each of the ten models by how accurate they have previously been. Our results produce a very close election, at least in terms of the popular vote, with the incumbent gaining only 50.3 % of the popular vote.

The separation plot package (current version at http://cran.r-project.org/) allows users to visually compare the fits of binary models, both in and out of sample. We present a visual method for assessing the predictive power of models with binary outcomes. This technique allows the analyst to evaluate model fit based upon the models’ ability to consistently match high-probability predictions to actual occurrences of the event of interest, and low-probability predictions to nonoccurrences of the event of interest. Unlike existing methods for assessing predictive power for logit and probit models such as Percent Correctly Predicted statistics, Brier scores, and the ROC plot, our “separation plot” has the advantage of producing a visual display that is informative and easy to explain to a general audience, while also remaining insensitive to the often arbitrary probability thresholds that are used to distinguish between predicted events and nonevents. We demonstrate the effectiveness of this technique in building predictive models in a number of different areas of political research.

Though weak states are associated with civil war, terrorism and other threats to
humanity, the social sciences provide scant insight into why states vary in their capacity to govern
across territory. This paper seeks to understand why states govern where they do in post-civil war
settings where leaders face stark geographic choices about extending state capacity across territory
in the face of resource constraints. We propose hypotheses derived from the distributive politics
literature and test them using satellite data in six countries (Burundi, Côte Ivoire, Kenya, Liberia,
Sierra Leone and Uganda). Contrary to several well-established theories, we find that state builders
do not reward core supporters or target swing districts. They do focus benefits on capital cities,
but this does not generalize to other urban settings. Instead, state leaders focus their efforts on
areas that have a history of violence.

We consider ensemble Bayesian model averaging (EBMA) in the context of small-n prediction tasks with high rates of missing component forecasts. With a large number of observations to calibrate ensembles and low rates of missing values for each component model, the standard approach to calibrating ensembles introduced by Raftery et al. (2005) performs well. However, data in the social sciences generally do not fulfill these requirements. The number of outcomes being predicted tend to be relatively small and missing predictions are neither random nor rare. In these circumstances, EBMA models may overweight components with low rates of missingness and those that that perform well on the limited calibration sample. This can seriously undermine the advantages of the ensemble approach to prediction. We demonstrate this problem and provide a solution that diminishes these undesirable outcomes by introducing a “wisdom of the crowds” parameter to the standard EBMA framework. We show that this solution improves predictive accuracy of EBMA forecasts in both political and economic applications.

Prediction is an important goal in the study of international conflict, but a large body of research has found that existing statistical models generally have disappointing predictive abilities. We show that most efforts build on models unlikely to be helpful for prediction. Many models essentially ignore the origins of conflict; studies look either at invariant structural features believed to affect the opportunities of conflict, or at factors that are believed to reduce the baseline risk of conflict, without attempting to identify the potential motivations and contentious issues over which conflicts typically arise. Researchers that have considered how contentious issues may motivate conflict and how these can be managed, using the Issues Correlates of War (ICOW) data, have not considered how these features may inform prediction. We assess the risk of dyadic interstate conflict based on the presence of specific contentious issues and conflict management events that may change the conflict potential of these contentious issues. We evaluate to what extent incorporating contentious issues and conflict management can help improve out-of-sample forecasting, as well as advance our understanding of conflict dynamics. Our results provide strong support for the idea that taking into account contentious issues can inform and improve out-of-sample forecasting.

GDELT and ICEWS are arguably the largest event data collections in social science at the moment. During their brief existence they have also been among the most influential data sets in terms of their impact on academic research and policy advice. Yet, we know little to date about how these two repositories of event data compare to each other. Given the nascent existence of both GDELT and ICEWS event data, it is interesting to compare these two repositories of event data. We undertake such a comparison for fighting in Syria, and for protest behavior in Egypt and Turkey, from 2011 to the present. You can view the visualizations here.

The gold-standard approaches to missing data imputation are complicated and computationally expensive. We present a principled solution to this situation, using Copula distributions from which missing data may be quickly drawn. We compare this approach to other imputation techniques and show that it performs at least as well as less efficient approaches. Our results demonstrate that most applied researchers can achieve great speed improvements implementing a Copula-based imputation approach, while still maintaining the performance of other approaches to multiple imputation.

This study investigates de facto states’ internal legitimacy—people’s confidence in the entity itself, the regime, and institutions.  Using original data from a 2010 survey in Abkhazia, we operationalize this using respondent perceptions of security, welfare, and democracy. Our findings suggest that internal legitimacy is shaped by the key Weberian state-building function of monopoly of the legitimate use of force, as well as these entities’ ability to fulfill other aspects of the social contract. 

For many, transnational capital is an important driving force of economic globalization. However, we know little about the political determinants for cross-border portfolio investments. Recent economic literature focuses upon information asymmetries. We move beyond this and intro- duce an explicitly political element into the study of international asset flows. Democratic institutions attract portfolio investments because they reduce the chances of predatory practices. Moreover, under real-world time constraints, investors are likely to use democracy as an important information short-cut for more credible property rights protection because the same underlying conditions (individual voice and rights, constraints on the executive, and rule of law) that make an established democracy also guarantee a credible property rights protection from the government. Applying a dynamic latent space model on the bilateral portfolio investment data from 2001 to 2005, we empirically examine the effects of important country-level characteristics of both net exporters and importers of portfolio investments. The empirical findings suggest that democracies are associated with higher levels of inward portfolio investments. We also find that portfolio investments are associated with business communities’ subjective estimate of property rights protection, but not with more comprehensive, index-based aggregate measures from international think tanks. This seems to suggest that investors do not have time to thoroughly study each country’s property rights system but rather rely on their subjective estimates, lending support to the democracy as information short-cut story in this paper. 

Developing political forecasting models not only increases the ability of political scientists to inform public policy decisions, but is also relevant for scientific advancement. As part of a larger project, a team at Duke University created a series of geographically informed statistical models for conflict prediction. The generated predictions have been highly accurate, with few false negative and positive categorizations. Predictions are made at the monthly level for six months periods into the future, taking into account the social-spatial context of each individual country. The model has a high degree of accuracy in reproducing historical data measured monthly over the past 10 years, and is approximately equally accurate in making  forecasts. This paper surveys the notion of forecasting and demonstrates the utility of creating forecasting models for predicting political conflicts in a diverse range of country settings. Apart from the benefit of making actual predictions, we argue that predictive heuristics are one gold standard of model development in the field of conflict studies and thatthe predictive heuristics shed light on an array of  important components of the political science literature on conflict dynamics.

Studies on civil conflict usually assume that state capacity is exogenous to fighting. However, the local level of state capacity and the onset of fighting are both strategic decisions and likely to be interdependent. We argue that the state imposes optimal tax rates across its territory to prevent groups outside of the government from challenging the state. We assume that peace-inducing tax rates are dependent on the varying local administrative and coercive power across a state's territory. Successful taxation of some localities broadens the state's ability to increase tax demands in other localities. As the state becomes stronger by taxing supportive localities, localities sympathetic to potential challengers of the state seek to preempt increasing demands, which leads to fighting. We test the empirical implications of our formal model by implementing a Seemingly Unrelated Regression approach to locally disaggregated data, supporting our claim that tax rate rate increases and fighting are strategic complements.

We extend ensemble Bayesian model averaging (EBMA) for application to binary outcomes and illustrate EBMA's ability to aid scholars in the social sciences to make more accurate forecasts of future events. In essence, EBMA makes improves prediction by pooling information from multiple forecast models to generate ensemble predictions similar to a weighted average of component forecasts. The weight assigned to each forecast is calibrated via its performance in some training period. The aim is not to choose some best model, but rather to incorporate the insights and knowledge implicit in various forecasting efforts via statistical postprocessing. After presenting the method, we show that EBMA increases the accuracy of out-of-sample forecasts relative to component models in three applied examples: predicting the occurrence of insurgencies around the Pacific Rim, forecasting vote shares in U.S. presidential elections, and predicting the votes of U.S. Supreme Court justices.

Every month, predictions are generated using the CRISP model}. Currently, CRISP forecasts rebellion, insurgency, ethnic and religious violence, as well as domestic and international crises. CRISP is a suite of programs to aid CRISis Predictions. This report was created with CRISP package version 2013.11.10. The predictions identify onsets, as well as ongoing conflicts, and also provide estimates of the intensity of these events. To do so we construct (at least) three kinds of models for each event of interest. These models have a monthly time frame, and are applied to a population of 167 countries around the world. These include 1) a hierarchical model of whether an event of interest is present, 2) a count model of the number of incidences of conflict related to the event of interest, and 3) a split duration, event based model that generates an independent estimate of the onset of new events. This report highlights some of these predictions, made in November of 2013. Herein the actual models are very briefly discussed. Instead, focus is on predictions of continuations and onsets of these five classes of events

What was the impact of the Eurozone economic crisis on interactions between political parties? Did the economic difficulties lead to an increase in conflict between them? Or did parties, in the face of an existential threat, come together to "rally around the flag" and interact in a more cooperative manner? In this paper, I use more than 110,000 machine coded news stories from 11 Eurozone countries between 2001 and 2011 and analyze them with latent network models. I show that the latter effect is true: The worse the economic situation, the more cooperative the relationships between parties. The effect is substantively large and robust for a variety of control variables, multiple model specifications, and an instrumental variable approach. The finding suggests that just like war or terrorist attacks, a severe economic crisis can lead to a suspension of the regular pattern of political interaction.

Reciprocity is a central concept in the international cooperation literature: strategies of conditional cooperation can create self-enforcing agreements. Another theme in this literature is that democracies follow international law more scrupulously than non-democracies. This study examines these claims in the context of the Laws of War. Using a model developed for dyadic data, we estimate reciprocity directly, which is an improvement over the standard techniques used in conflict research. We show that transgressions of the Laws of War are characterized by reciprocity, but the strength of reciprocity varies across issues, and over time, in interesting ways. Thus while states generally respond to violations, the incentives that create self enforcing agreements do not exist for all issues. We further show that standard estimation techniques detect differences between the behaviors of democracies and non-democracies that disappear once the dependencies that often exist in dyadic data are properly addressed.

Political transitions present a crucible for governance. Political systems are judged in large part by the nature of those transitions, and unfortunately in many states they manifest as coups by actors inside or outside the government who overthrow established rule. We examine the incidence of coups with a split-population duration model that  separates polities at risk from those that are not empirically, as well estimates how the chances of a coup evolve over the course of an established regime. Relying on monthly data about intra-governmental conflict, and information from the Polity project on changes in the character of political institutions, we are able to build a accurate model of coups d'etat that is theoretically grounded as well as useful for out-of-sample prediction.

Pages