SUNGEO: Sub-National Geospatial Data Archive
Kollman, Kenneth (PI) and Yuri M. Zhukov (co-PI). “SUNGEO: Sub-National Geospatial Data Archive.” Data infrastructure, 2022.
Abstract | R package (CRAN) | R package (GitHub)
Research on political, social, and economic behavior and phenomena increasingly depends on combining multiple distinct sources of sub-national data, which are often collected at disparate spatial scales and units of analysis. Using different methods of linking data, not to mention different data sources, can affect inferences. Yet analysts’ critical decisions on both dimensions are often ad hoc and driven mainly by the idiosyncratic needs and constraints of a particular project. These practices reduce the reliability, transparency, and replicability of empirical research.
The Sub-National Geospatial Data Archive (SUNGEO) will relieve bottlenecks in research by integrating multiple sources of sub-national data in a common data repository at multiple, customizable spatio-temporal scales, and developing a suite of methods for data processing and analysis. This infrastructure includes three main components. First is a user-friendly web interface, where researchers can select among many pre-loaded variables [e.g. elections, violent events, weather, land use, public health outcomes], choose levels and methods of spatio-temporal (dis)aggregation, interpolation and integration, and easily construct their own sub-national datasets. Second is an open-source software package, in the R statistical programming language, that processes user-supplied data, merges them with pre-loaded geo-referenced data, and produces a more customizable output based on user needs and specifications. Third is an archiving tool, which allows users to contribute original data to the repository.
National Science Foundation RIDIR Grant (SES-1925693).
Integrating Data Across Misaligned Spatial Units
Zhukov, Yuri M., Jason S. Byers, Marty Davidson, Kenneth Kollman. “Integrating Data Across Misaligned Spatial Units.” Working paper, 2022.
Abstract | PDF
Theoretical units of interest often do not align with the spatial units at which existing data are available. This problem is pervasive in social science, particularly in sub-national empirical research that routinely requires integrating data across incompatible geographic units (e.g. administrative areas, electoral constituencies, postal codes). Overcoming this challenge requires researchers to not only align the scale of empirical and theoretical units, but also to diagnose the costs and consequences of alternative spatial transformation methods. We propose a framework for addressing such change-of-support problems in social science. We investigate the relative performance of 12 spatial transformation methods, including overlays, interpolation, kriging and other model-based approaches, using election data and Monte Carlo simulations. We show that both the accuracy of transformed values and the estimation of regression coefficients depend on the relative scale of source and destination units (i.e. aggregation, disaggregation, hybrid) and their degree of nesting (i.e. whether source units fall completely and neatly inside destination units). We propose simple nonparametric measures of relative scale and nesting, and validation procedures to assess the quality of the transformed geospatial data. Additionally, we introduce new infrastructure and open-source software to elucidate the consequences of these choices, and to make transformation options more accessible, customizable, and intuitive.
Fighting for Tyranny: How State Repression Shapes Military Performance
Rozenas, Arturas, Roya Talibova, and Yuri M. Zhukov. “Fighting for Tyranny: How State Repression Shapes Military Performance.” Working paper, 2022.
Abstract | PDF
Utilizing over 100 million declassified Red Army personnel records from World War II, we study how state repression shapes soldiers’ motivation to exert effort in fighting. Exploiting three complementary identification strategies, we find that soldiers from places with higher levels of pre-war repression under Stalin’s rule were more likely to fight until death and less likely to shirk on their duties, but they also received fewer decorations for personal bravery. The coercive incentives created by repression appear to have induced obedience at the expense of initiative and increased the human costs of war.
Repression Works (just not in moderation)
Zhukov, Yuri M. “Repression Works (just not in moderation).” Working paper, 2022.
Abstract | PDF
Why does government violence deter political challengers in one context, but inflame them in the next? This paper argues that repression increases opposition activity at low and moderate levels, but decreases it in the extreme. There is a threshold level of violence, where the opposition becomes unable to recruit new members, and the rebellion unravels — even if the government is responsible for more civilian suffering overall. I show this result theoretically, with a mathematical model of coercion and popular support, and empirically, with micro-level data from Chechnya and a meta-analysis of sub-national conflict dynamics in 145 countries. The data suggest that such a threshold exists, but the level of violence needed to reach it varies. Many governments, thankfully, are unable or unwilling to go that far. I explore conditions under which this threshold may be higher or lower, and highlight a fundamental trade-off between reducing government violence and preserving civil liberties.
A Near-Real Time Data Analysis of Russia’s 2022 Invasion of Ukraine
Zhukov, Yuri M. “A Near-Real Time Data Analysis of Russia’s 2022 Invasion of Ukraine.” Working paper, 2022.
Abstract | PDF
Who did what to whom, when, and where? Event data can be an invaluable resource in establishing the “hard facts” of warfare. Policymakers, activists, journalists and scientists often rely on such data to track the dynamics of armed conflict as it unfolds. Yet this information is not always easy to acquire, and it can be surprisingly perishable — if event reports are not preserved in a timely and systematic way, reconstructing a sequence of events becomes very challenging. We will examines this problem in the case of Russia’s 2022 invasion of Ukraine, using VIINA (Violent Incident Information from News Articles) — a near-real time territorial control and violent event tracking system, which scrapes and parses news reports from Ukrainian and Russian media, georeferences them, and classifies them into standard event categories (e.g. firefight, tank battle, artillery shelling) through machine learning. We will provide an overview of the project, and illustrate potential applications of these data for policy and social science. These applications include investigations of changes in territorial control, and variation across Russian and Ukrainian media sources in the relative location, intensity and attribution of reported violent events.
Are Competitive Elections Good for Your Health? Evidence from the 1918 Flu and Covid-19
Walden, Jacob, and Yuri M. Zhukov. “Are Competitive Elections Good for Your Health? Evidence from the 1918 Flu and Covid-19.” Working paper, 2021.
Abstract | PDF
Do more electorally vulnerable government officials respond to public health emergencies more aggressively than officials in less competitive seats? Using novel data on local government responses to the 1918 influenza A (H1N1) “Spanish Flu” and 2020 Covid-19 pandemics in the United States, we study how the competitiveness of federal, state and local elections shapes the policy choices of incumbents. We find that, in 1918, vulnerable incumbents enacted more and longer nonpharmaceutical interventions (e.g. quarantines, closures), enforcing them more aggressively than in less-competitive jurisdictions. Their constituents subsequently experienced fewer influenza-related deaths and lower overall excess mortality. In 2020, more competitive constituencies similarly experienced lower rates of Covid-19 infection and death, but they implemented fewer nonpharmaceutical interventions and relied more on pharmaceutical measures. This policy substitution was feasible in part due to political geography: more competitive localities became more suburban, and more conducive to social distancing in the absence of government mandates.
Fratricidal Coercion in Modern War
Lyall, Jason, and Yuri M. Zhukov. “Fratricidal Coercion in Modern War.” Working paper, 2021.
Abstract | PDF
Does the threat or use of violence against one’s own soldiers make them more willing to perform their duties in battle? Existing theories largely dismiss this kind of fratricidal coercion as ineffective or obsolete, suggesting that positive inducements like ideology, material rewards, and primary group bonds drive soldiers’ behavior. We argue instead that fratricidal coercion can improve soldier compliance, reducing wartime desertions, missing in action, premature surrender, and other forms of indiscipline. Yet it also places soldiers at greater risk of physical harm, and potentially impedes an army’s ability to inflict costs on enemy forces. To test our claims, we use a three-pronged empirical strategy that draws on (1) a monthly panel dataset of 609 Soviet Rifle Divisions in 1941–45, built from 34 million personnel files; (2) a close-range paired comparison of two Rifle Divisions selected via matching; and (3) 526 land battles (1939–2011) to assess the cross-national generalizability of these micro-level findings. Fratricidal coercion improves soldier compliance across all of these samples, but at the cost of higher casualties. These findings highlight the need to bring coercion back into our theories of combat motivation and military effectiveness.
Political Regime Type and Warfare: Evidence from 600 Years of European History
Blank, Meredith, Mark Dincecco and Yuri M. Zhukov. “Political Regime Type and Warfare: Evidence from 600 Years of European History.” Working paper, 2017.
Abstract | PDF
This paper presents new evidence that, historically, the relationship between political regime type and warfare was different than it is today. Using a novel database of interstate conflict in Europe between 1200 and 1800, we perform the first quantitative analysis of domestic political institutions and warfare across the pre-modern era. We find that early parliamentary regimes – the institutional predecessors of modern democracies – were disproportionately more likely to experience armed conflict than their absolutist counterparts. Our empirical strategy makes use of two complementary approaches: a standard dyadic analysis of conflict initiation, and a dynamic network analysis that accounts for interdependence between dyads. These analyses show that early parliamentary regimes fought in significantly more wars than absolutist monarchies, both against one another and overall. Such regimes, we argue, had a relatively large capacity to make war, but, unlike modern democracies, not enough institutional constraints to prevent it.
How Selective Reporting Shapes Inferences about Conflict
Zhukov, Yuri M. and Matthew A. Baum. “How Selective Reporting Shapes Inferences about Conflict.” Working paper, 2017.
Abstract | PDF
By systematically under- or over-reporting violence by different actors, media organizations convey potentially contradictory information about how a conflict is likely to unfold, and whether outside intervention is necessary to stop it. These reporting biases affect not only statistical inference, but also public knowledge and policy preferences. Using new event data on the ongoing armed conflict in Eastern Ukraine, we perform parallel analyses of data from
Ukrainian, rebel, Russian and third party sources. We show that actor-specific reporting bias can yield estimates with vastly different implications for conflict resolution: Ukrainian sources predict frequent unilateral escalation by rebels, pro-Russian rebel sources predict unilateral escalation by government troops, while outside sources predict that transgressions by either side should be quite rare. Experimental evidence suggests that news consumers tend to support intervention against whichever side is shown to be committing the violence. We argue that these kinds of reporting biases can potentially make conflicts more difficult to resolve — hardening attitudes against negotiated settlement, and in favor of military action.