1.2 Regionalization Evaluation Model

It is not recommended that an evaluation model of security policy for the State of Mato Grosso use municipalities as the unit of observation. The 1988 Federal Constitution gives state governments the power on this particular policy. While the recent debate on public security policy and the public policy developed by the very SENASP confer a large role to municipalities, the decentralization of public safety is not yet sufficiently established in the State of Mato Grosso for them to be subject to a comprehensive evaluation. Including the municipalities in the evaluation model would restrict the evaluation to a very limited number of cases, with a loss of comparability.

Thus, the evaluation model consists of 16 observation units, now called regionalization. Regionalization is based on three criteria:

(i) the evaluation refers to the policy of the State Government;

(ii) the State Civil Police is organized into regional, which implies that the regional police authorities respond by regional policy performance;

(iii) the population size, with the goal of standardizing the data analysis in the range of 100,000 inhabitants, in order to avoid statistical distortions associated with the scale. Since the raw performance indicators selected tend to have fairly low numbers of cases in the numerator, the denominator’s small values would tend to produce large fluctuations in exchange rates between the two years analyzed. For this reason, regional Civil Judicial Police with a population of less than 100,000 inhabitants were grouped into a single unit of analysis.

Regionalization as a priority was adopted by the Civil Police of the State, since this is the law enforcement agency responsible for statewide criminal justice system. The civil police is the national security agency with assignments of investigative and judicial skills. It's under its scrutiny to assess whether the reported phenomenon will be investigated or not. It is from its investigation that the fact enters the formal justice system. For this reason, the data of the civil police feeds most of the data systems of the public security in Brazil from all the states.

This way, the production (collection and systematization) of the raw data has been a central criterion for regionalization in the evaluation model. The criterion of quantitative population of the region was applied to the regionalization adopted by the Civil police (PCJ). As noted, for statistical reasons, this second criterion is essential for calculating the rate of observed phenomena. Thus, the evaluation model considers the range of 100,000 inhabitants for both aggregate regional population with less than this limit as to disaggregate units with a population higher than this number, so that all units of analysis present a population around (or higher to) 100,000 inhabitants.

Implementation of scale of 100,000 inhabitants on the map of regionalization of civil police produced four different regions of the original matrix, i.e., the one adopted by the civil police, as can be seen on the map below:
 

Map 1

Map of regionalization of the evaluation model

mapa1.jpg

Source: own

The five-region variety of the original matrix corresponds to (i) the breakdown of the cities of Sinop, Cuiabá, Rondonópolis and Várzea Grande and their respective regional Civil police, because these counties alone have more than 100,000 inhabitants and (ii) aggregation of the Regional of Barra do Garças and Alto Araguaia, as these regions each have less than 100,000 inhabitants.

Thus, the region of Sinop has been disaggregated in Sinop (district) and Sinop. The City of Sinop now constitutes a single region. The other municipalities follow aggregated composing the regional Sinop, which excludes the City of Sinop.

We proceed similarly with the cities of Cuiabá, Rondonópolis and Várzea Grande. The City of Cuiabá became a separate regional municipality of its own surroundings. Its denomination is Cuiabá (municipality). The City of Várzea Grande, also broken down, has become a regional for proceeding the analysis for the model, whose denomination follows the same pattern as before, i.e., Várzea Grande (city). The same procedure was adopted to Rondonopolis. The surrounding municipalities keep constituting their respective regions, but this excludes the regional county seat.

The regions Barra do Garças and Alto Araguaia, both with populations less than 100,000 inhabitants, were added to achieve the scale-adopted population. The regional municipalities and their correspondents may be accessed at the Methodological Annex.
 

Support
fapesp
cepid
usp
cebrap