An estimate of the underreporting of violent crimes against property applying stochastic frontier analysis to the state of Minas Gerais, Brazil

Abstract

The aim of this study was to determine the level of underreporting of violent crimes against property in the state of Minas Gerais, Brazil, using stochastic frontier analysis. Failure to report a crime to competent authorities has negative consequences for the effectiveness of public safety policies, because policy decisions regarding criminal matters are based on official crime statistics, which in turn are biased by underreporting. An awareness of the magnitude of underreporting should help policy makers to design more appropriate crime prevention strategies. The database used for the study defines armed robbery, robbery, and theft as violent crimes against property. The main results of the study show that, from 2004 to 2011, 32.7% of all violent crimes against property in the state were not reported. A robustness test was carried out on the estimates by employing the same technique to calculate the level of underreporting for the homicide rate, which is the least unreported crime.

Gendered mobility and violence in the São Paulo metro, Brazil

“…Solymosi et al, 2018). Although Sa˜o Paulo's metro authorities began a campaign to encourage users to report violence in 2015, and the number of reports has increased (see Ceccato and Paz, 2017), statistics are certainly still affected by underreporting (see Moreira et al, 2018). Moreover, as mentioned, the times that employees change shifts and move between stations are not included in our data.…”

Modelling underreported spatio-temporal crime events

“…This probability resulted from considering the costs of the crime and the benefit of informing it [ 17 ]. More recently, Chaudhuri et al [ 18 ], and Moreira et al [ 22 ] assumed crime as a linear function of demographic covariates and accounted for an additional term linked with inefficiency in the citizen’s report, i.e., underreporting. Although these last approaches may provide a short-time estimation of actual crime incidences, they require exogenous covariates, which may also vary in time, limiting their capabilities for underreporting crime estimation over time.…”

“…The main objective of this work was to study the identification over time of the “true” unknown crime incidence rates based on official reports of crime incidents and the “true” underreporting rates based on complementary information, particularly crime-related data acquired gradually over time. In contrast with previous works aimed to describe underreporting in long-time scales by exploiting victimization surveys [ 6 , 9 , 19 ], official crime data [ 17 , 18 , 22 ], or combining multiple crime data sources [ 27 ], this work aims to integrate additional incremental evidence about crime once is available, allowing to gain knowledge about the crime phenomena gradually, instead of forcing to wait for final integration. The proposed online estimation relies on a new crime underreporting combinatorial multi-armed bandit model [ 28 ] aimed to elicit the “true” average incidence rates and estimate the underreporting rates for different spatial units over time.…”