working papers

We study a version of a canonical model of attacks against political regimes where agents have an expressive utility for taking political stances that is scaled by the salience of political decision-making. Increases in political salience can have divergent effects on regime stability depending on costs of being on the losing side. When regimes have weak sanctioning mechanisms, middling levels of salience can pose the greatest threat, as regime supporters are insufficiently motivated to act on their preferences and regime opponents are sufficiently motivated to stop conforming. Our results speak to the phenomenon of charged debates about democracy by identifying conditions under which heightened interest in political decision-making can pose a threat to democracy in and of itself.

The R package CausalQueries can be used to make, update, and query causal models. Users provide a causal statement of the form X -> M <- Y; M <-> Y which is interpreted as a structural causal model over a collection of binary variables. CausalQueries can then (1) identify the set of principal strata—causal types—required to characterize all possible types of causal relations between nodes consistent with the causal statement (2) determine a set of parameters needed to characterize distributions over these types (3) update beliefs over the distribution of causal types, using a stan model and (4) pose a wide range of causal queries of the model, using either the prior distribution, the posterior distribution, or a user-specified candidate vector of parameters

What political and social features of states help explain the distribution of reported Covid-19 deaths? We survey existing works on (1) state capacity, (2) political institutions, (3) political priorities, and (4) social structures to identify national-level political and social characteristics that may help explain variation in the ability of societies to limit Covid-19 mortality. Accounting for a simple set of Lasso-chosen controls, we find that measures of interpersonal and institutional trust are persistently associated with reported Covid-19 deaths in theory-consistent directions. Beyond this, however, patterns are poorly predicted by existing theories, and by arguments in the popular press focused on populist governments, women-led governments, and pandemic preparedness. Expert predictions of mortality patterns associated with state capacity, democracy, and inequality, do no better than chance. Overall, our analysis highlights the challenges our discipline’s theories face in accounting for political responses to unanticipated, society-wide crises.

Process tracing is a strategy for inferring within-case causal effects from observable implications of causal processes. Bayesian nets, developed by computer scientists and used now in many disciplines, provide a natural framework for describing such processes, characterizing causal estimands, and assessing the value added of additional information for understanding different causal estimands. We describe how these tools can be used by scholars of process tracing to justify inference strategies with reference to lower level theories and to assess the probative value of new within-case information.

"I give an illustration of a simple problem in which a Bayesian researcher can choose between random assignment of a treatment or delegating assignment to an informed—but motivated—agent. In effect she compares between learning from an RCT or from an observational study. She evaluates designs according to an expected squared error criterion (ESE). I show that for a small problem (n “ 2) if she starts with a prior expectation of no treatment effect but believes that the agent is an advocate of treatment with probability q (and otherwise an opponent) then for all values of q she does at least as well delegating assignment as she does from an RCT and she does strictly better as long as q ≠0.5. For other priors on treatment effects, randomization can dominate delegation or be dominated by it. As n grows the expected errors from an RCT design fall but errors from delegated assignment do not. Although there is always some prior such that a delegated procedure beats randomized assignment, the converse is not true. For a given prior there may be no delegated procedure that trumps an RCT. With uniform priors for example the RCT dominates delegated assignment for all beliefs on agent motivations when n ≥4."

There is increasing support for the use of research registries in social sciences. One possible advantage of the use of a registry is that it would limit the scope for publication or analysis biases that result from selecting statistically significant results. However, to date, there is surprisingly little evidence for the claim that registration will reduce these biases. We look to historical data from medical publishing for evidence, comparing the distribution of p-values before and after the introduction of registration in prominent journals. We couple this analysis with a pre-analysis survey of medical experts and social scientists to assess their prior expectations of the impact of registration on medical publishing and to assess their perceptions on the specificity and sensitivity of our test of effects. Although there is evidence of publication bias in medical studies, our registered analyses uncovered no evidence that registration affected that bias, leading us to moderately downgrade our confidence in the curative effects of registration.

Identifying the conditions under which politicians are responsive to citizens’ needs and preferences is a central concern in the study of political economy. Does greater transparency improve political accountability? We use a simple model of political accountability to derive a set of hypotheses linking access to information to political behavior and provide results from a multi-level field experiment designed to test these hypotheses in the context of parliamentary behavior in Uganda. Between 2006 and 2011, working with a Ugandan partner, we developed a scorecard with detailed information on the behavior of Ugandan Members of Parliament (MPs), informed a randomly selected sample of MPs that the information would be disseminated in their constituencies, and provided voters with information about their MP’s performance through a variety of dissemination channels. Evidence from survey experiments indicate that Ugandan voters are strongly receptive to new information about the performance of their MPs. Evidence from the dissemination campaigns, however, provides no evidence that MPs respond to a higher level of transparency or that their prospects for reelection are threatened by it.

Final Report on the Effects of a Community Driven Reconstruction Program in Eastern Democratic Republic of Congo detailing null results.

We examine technology diffusion in rural Uganda, with a focus on the effectiveness of decentralized marketing to encourage the adoption of improved fuelwood cookstoves. Identifying the effects of a dissemination scheme is rendered difficult by the possibility of spillover effects—that areas that are not targeted with a new technology are nevertheless indirectly affected via targeting of their neighbors. To address this concern, we use a novel randomization scheme to allocate ambassadors to communities in a way that can allow for the identification of direct and indirect effects. We find evidence for positive direct and spillover effects on the knowledge of the stoves, but found little evidence of either direct or indirect effects on purchase decision. Our results are consistent with no impact of this scheme on takeup but we note that we cannot rule out the possibility of effects coupled with spillovers that do not decline appreciably over the ranges that we study.

In many contexts, treatment assignment probabilities differ across strata or are correlated with some observable third variables. Regression with covariate adjustment is often used to account for these features. It is known however that in the presence of heterogeneous treatment effects this approach does not yield unbiased estimates of average treatment effects. But it is not well known how estimates generated in this way diverge from unbiased estimates of average treatment effects. Here I show that biases can be large, even in large samples. However I also find conditions under which the usual approach provides interpretable estimates and I identify a monotonicity condition that ensures that least squares estimates lie between estimates of the average treatment effects for the treated and the average treatment effects for the controls. The monotonicity condition can be satisfied for example with Roy-type selection and is guaranteed in the two stratum case.

We use experimental methods to study coordination games analogous to voting games over fixed divisions of a pie. We find that players use the ethnic or gender identities of other players, rather than other payoff irrelevant features of a game’s description, to achieve coordination. Coordination along or across gender lines, we show, can be attributed in part to the presence of social norms rather than to the focal point properties of identity. Coordination along ethnic lines however is explained by focal point effects of ethnic categories and can not be accounted for either by the existence of other-regarding preferences between coethnics or to the existence of norms for cooperation within ethnic groups. This evidence suggests that the importance of identity in social interactions may can be explained without appealing to differential attitudinal characteristics of individuals or differences in norms across groups but more simply from the existence of categories that can be used by rational agents to structure their actions.

This report describes the initial findings of a survey of a representative sample of 1043 combatants from Sierra Leone’s civil war. It presents information on the demographic profile of the combatant population, their motivations for joining and incentives for staying within the different factions, and their attitudes about the disarmament, demobilization and reintegration (DDR) process.

A review of literature ca 2000 on economics and violent conflict.

I study a setting in which internally fragmented groups try to unilaterally alter the status quo, using violence or influence to pull it one way or another. A status quo is considered stable when the combined effects of the equilibrium strategies of different actors cancel each other out exactly. The–-costly–-equilibrium of the game has equivalence properties with solutions in institutionally rich environments; in these environments. The model provides answers to questions of the form: how do outcomes in institutionally weak environments compare to outcomes in institutionally rich environments? how does the aggregate level of violence vary with the coalitional, or perhaps ethnolinguistic structure of a polity? how does it vary as a function of the degree of homogeneity within organized groups and the heterogeneity across groups? or with the cost of violence? and how do political outcomes vary as the distribution of preferences.

Discussions of how to conceptualize salience in politics with connections to heresthetics