Crime Concentration and Hot Spot Dynamics: An Examination of Homicides in Santa Fe, Argentina

doi:10.21203/rs.3.rs-3112743/v1

Download PDF

Research Article

Crime Concentration and Hot Spot Dynamics: An Examination of Homicides in Santa Fe, Argentina

https://doi.org/10.21203/rs.3.rs-3112743/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

This study aimed to facilitate a general understanding of the spatial concentration of crime in Latin America. Thus, the spatial concentration of homicides committed in the Argentine city of Santa Fe between 2001 and 2020 was examined. In addition, this study examined the temporal stability of the concentration and trajectories of the identified hot spots. The results were contrasted with the “law of crime concentration” proposed in a previous study and with the results of other studies on crime concentration conducted in the region of Latin America. The findings suggested that homicides committed in the city of Santa Fe were spatially concentrated and that their concentration levels had remained stable over the years. Although two-thirds of the microunits did not experience homicides during the analysis period (2001–2020), a disproportionate number of homicides occurred in a small group of microunits. The study findings will prove useful in decision-making regarding place-based crime prevention strategies.

Criminology

hot spots

spatial concentration

temporal stability

persistence in the trajectory

homicides

The study of where crime occurs can be traced back to the 19th century. However, it was during the last quarter of the 20th century and the early 21st century that studies on places and crime were established as a highly dynamic area in the field of criminology. Among these studies, those focusing on the concentration of crime in small geographic spaces have attracted attention. Currently, there is significant evidence that shows that crime “is extremely concentrated in small places; this concentration is stable across time and crime hot spots are not bad neighborhoods but rather are a few small crime pockets that are driving the crime rates for large areas” (Weisburd et al., 2016, p. 34).

Although the concentration of crime has been extensively studied in the US and other developed countries (Weisburd, 2015; Lee et al., 2017; Eck et al., 2017; Braga et al., 2017), minimal empirical research has been conducted in Latin America. The few publications in this regard are quite recent and are typically based on data corresponding to nonextensive temporal arches. A better understanding of the spatial concentration of crime in Latin America would enable managers and political decision-makers to apply place-based preventive strategies. Thus, this study tested the “law of crime concentration” proposed by Weisburd (2015), which predicts that crime is concentrated in street segments within certain bandwidths, in relation to homicides committed in the Argentine city of Santa Fe between January 1, 2001, and December 31, 2020. In addition, it seeks to determine if the concentration of crimes within the expected bandwidth is stable over time. Finally, the study attempted to understand if the temporal stability of the concentration could be explained by a small number of micro-places that remained hot over time or if despite the temporal stability of the concentration, this concentration is the result of the entire city, in this case, the city of Santa Fe, i.e., if these micro-places varied in their trajectories and did not remain hot over time.

2.1. The spatial concentration of crime

In the late 1980s, Sherman et al. took a critical step toward understanding crime concentration (Weisburd et al., 2016). Although they were not the only researchers to identify strong concentrations of crime at the microgeographic level, they were the first to recognize the criminological importance of such findings and called for a new area of study—the criminology of place. This occurred when they measured the concentration of crime in the city of Minneapolis (Minnesota) and observed that 50% of all service calls to the police made over 2 years had come from 3% of the addresses.

In 1995, Eck and Weisburd edited a collection dedicated to the study of crime at the micro-space level and coined the expression, crime places, giving new impetus to an emerging branch of space criminology.

Weisburd et al. (2004) conducted a trajectory study in Seattle (Washington) and analyzed more than 1.4 million incident reports from the Seattle Police Department. They observed that approximately 5% of street segments represented 50% of all criminal incidents. They also discovered that all crime incidents were in the range of 48–53% of street segments. The study period was 14 years (1989–2002).

Weisburd (2015) formulated the “law of crime concentration.” For a defined measure of crime in a specific microgeographic unit, the concentration of crime will fall within a narrow bandwidth of percentages for a certain cumulative proportion of crimes. Weisburd (2015) argued that whatever the variability, the general conclusion is that there is a narrow bandwidth of crime concentration in places, which suggests a law of crime concentration in cities. For a crime concentration of 50%, the bandwidth was approximately 4% of street segments (between 2.1 and 6%). For a crime concentration of 25%, the bandwidth was less than 1.5% of street segments (between 0.4 and 1.6%).

Empirical studies strongly support the law of crime concentration in micro-places. For instance, Weisburd (2015) showed that the law of crime concentration operates in large and small cities, and Ridner (2019) showed that the law also operates in medium-sized cities.

Lee et al. (2017) reviewed studies on the US, UK, Israel, and Turkey, which showed the expected concentrations, although crime in the US appeared to be more concentrated.

Gill et al. (2017) considered that although most studies had focused on urban areas, there were indications that the law of crime concentration also applied to suburban areas.

The law of crime concentration has been tested in Europe. Hardyns et al. (2018) studied two major Belgian cities with a 9-year dataset (2004 to 2012), using grid cells of 200 m × 200 m. They observed that 25% of crime was concentrated according to crime type and, according to the city, between 0.44 and 1.86% of the micro-places, with 50% of crime concentrated between 1.61 and 5.37% of the places. In addition, they observed that residential burglaries tended to be more dispersed than other types of crimes. Hardyns et al. (2018) concluded that significant variations and patterns specific to each crime type can be hidden during the analysis of crime indices.

Using street segments as spatial analysis units, Favarin (2018) studied burglaries and robberies registered between 2007 and 2013 in Milan (Italy). The burglary and robbery concentrations were also determined. On average, 4.0% and 1.6% of the street segments in Milan accounted for 50% of the total burglaries and robberies, respectively. By aggregating all the incidents that occurred between 2007 and 2013, 8.2% and 4.0% of the street segments accounted for 50% and 50% of the total burglaries and robberies, respectively. These overall percentages were higher than the concentration levels experienced each year. This is because burglaries and robberies did not always occur in the same street segments over the years. In addition, robberies had a higher crime concentration than burglaries over the years. Thus, Favarin (2018) argued that the analysis must be specific to properly tailor prevention policies.

The law of crime concentration has also been tested in Asia, where Amemiya and Ohyama (2019) studied property crimes in two Japanese cities and observed concentrations with variations according to crime type.

In Latin America, the use of geospatial crime analysis is generally related to the study of the effects of certain policies or strategies—typically police—deployed over roughly broad territories.

However, the study of crime concentration in micro-places, particularly Weisburd's law, remains relatively unexplored, despite a few recent publications, such as those by Mejía et al. (2015), Jaitman and Ajzenman (2016), Chainey et al. (2019), and Chainey and Muggah (2020).

Table 1 presents the results of recent studies conducted in Latin America. These investigations showed that cumulative proportions of 25% and 50% of crimes were committed in a low percentage of the microunits of analysis. This low percentage was within the bandwidths predicted by Weisburd (2015).

Table 1

Spatial concentration of crimes in relevant studies
	Spatial unit of analysis		Percentage of microunits
	Cities studied	Spatial units	25% of crimes	50% of crimes
Weisburd (2015)	Cincinnati, Seattle, Tel Aviv, New York, Sacramento, Brooklyn Park, Redlands, and Ventura	Street segment with an average length of 144 m	Between 0.4 and 1.6%	Between 2.1 and 6.0%
Mejía et al. (2015)	Barranquilla, Bogotá, Cali, and Medellín	Street segment	0.8% (0.3%*)	3.6% (0.9%*)
Jaitman and Ajzenman (2016)	Belo Horizonte, Bogotá, Montevideo, Sucre, and Zapopan	Street segment	Between 0.5 and 2.9%	Between 3 and 7.5%
Chainey et al. (2019)	37 cities in Latin America	Street segment with an average length of 139 m	0.8%	2.5% (1.4%*)
Chainey and Muggah (2020)	Rio de Janeiro	Square cell of 150 m × 150 m (22500 m²)	0.3%	1.0%
Note: Values that specifically pertain to homicide case studies are indicated with an asterisk.

2.2. The temporal stability of the concentration of crime

Weisburd et al. (2016) stated that despite fluctuations in crime in cities over time, the concentration of crime remained within a relatively narrow range, although the number of annual criminal incidents was volatile within and among cities. Weisburd (2015) examined the temporal stability of the concentration of crime in four cities (Tel Aviv, Seattle, Brooklyn Park, and New York) and observed that the concentration of crime remained stable over time. In Seattle, over 7 years, the bandwidths of cumulative proportions of 50 and 25% of crime ranged from 4.6 to 5.8% and 0.9 to 1.2% of the street segments, respectively. Similarly, in New York, the bandwidths ranged from 4.7 to 6% and 1.1 to 1.5% for 50 and 25% of crimes, respectively, over 9 years. In Brooklyn Park, the concentration was high, but the bandwidth was small, ranging between 1.5 and 2.6% and 0.3 and 0.5% of street segments for 50 and 25% of crimes, respectively, over a 14-year period. Tel Aviv followed the general pattern of stability; however, the variation over time was slightly high. The bandwidths of 50 and 25% of the cumulative crime ratio ranged from 3.9 (1990) to 6.5% (2003) and 0.8 to 1.8%, respectively.

2.3. Persistence in the trajectory of the hot spots

According to Eck et al. (2005), although there is no common definition of a crime hot spot, it is generally understood as an area in which the number of criminal incidents or disorders is above average or an area where people have a higher-than-average risk of victimization. According to Chainey (2021), a hot spot is an area with a high concentration of crime in relation to the crime distribution throughout the study area.

As described by Weisburd et al. (2016), hot spots are micro-places (i.e., street segments, a conjunction of street segments, or a group of housing units).

From this perspective, the stability of crime concentration does not necessarily imply that hot spots remain stable over time. It is possible that the law on crime concentration applies over time and that the specific places where crime is concentrated change yearly. However, Weisburd et al. (2012) suggested that this is not the case. They examined a 16-year dataset (1989–2004) from 24000 Seattle street segments each with its “trajectory” or crime pattern over time and determined that a 22-group model was optimal for understanding crime data in street segments. To simplify the description, Weisburd et al. (2012) divided the trajectory groups into eight developmental patterns representing the main levels of crime and crime trends observed (high decreasing, low decreasing, high increasing, low increasing, moderate, stable, crime-free, and chronic). Chronic crime patterns represent the most serious crime hot spots in cities.

Weisburd et al. (2012) observed that 1% of the street segments followed this pattern, indicating that there were 247 street segments in the city of Seattle with extremely chronic levels of crime. These 247 street segments accounted for 22% of crime incidents in Seattle between 1989 and 2004. Thus, the findings of Weisburd et al. (2012) regarding the stability of crime concentration, particularly the stability of crime hot spots, showed that crime is tightly connected to place.

2.4. Homicides in the city of Santa Fe: the provincial, national, and regional contexts

At the global level, the United Nations Office on Drugs and Crime [UNODC] (2019) stated that the homicide rate had been slowly declining for more than 2 decades, from a peak of 7.4 per 100000 inhabitants in 1993 to 6.1 per 100000 inhabitants in 2017. However, homicide rates have been consistently high in the Americas over the past 3 decades. Between 1990 and 2016, the average homicide rate in this region remained between 14.5 and 16.7 per 100000 inhabitants, approximately two or three times the world average, before increasing to 17.2 in 2017, the highest rate since 1990.

A significant proportion of the deaths caused by intentional homicides at the global level in 2017 (37%) was registered on the American continent. Central and South Americas, with 25.9 and 24.2 homicides per 100000 inhabitants, respectively, were the subregions with the highest average homicide rates in 2017, rendering Latin America and the Caribbean the most violent region in the world with a homicide rate four times higher than the global average.

The UNODC (2019) maintained that South American countries can be divided into two categories in relation to homicides. The first group comprises countries with persistently high homicide rates, such as Brazil, Colombia, and Venezuela. The second group comprises countries with homicide rates that are lower than those of the first group but still above or around the world average; Argentina belongs to the second group.

In Argentina, the distribution of violence is inhomogeneous but is concentrated in certain subnational jurisdictions (provinces) that exhibit high concentration levels and persistence. As reported by the Ministry of Security of the Nation of Argentina (2019a, 2019b, and 2020), the province of Santa Fe had the highest homicide rate per 100000 inhabitants in Argentina in 9 of the 11 years in 2010–2020. In the other 2 years (2012 and 2014), it registered the second-highest homicide rate after Chubut, a province with extremely low population density.

According to the Public Prosecutor's Office of the Province of Santa Fe (2017, 2019, 2020, 2021), in 2020, 8 out of 10 of the homicides that occurred in the province of Santa Fe occurred in the departments of La Capital and Rosario (81.5%). This relationship (8 out of 10) had previously been manifested (80.1% in 2019, 83.2% in 2018, 80.7% in 2017, 80.9% in 2016, 80.3% in 2015, and 88.3% in 2014).

The Department of La Capital, whose main population nucleus is the city of Santa Fe, which is also the capital city of the province, had the highest homicide rate per 100000 inhabitants from 2002 to 2019, except for 2013. In 2013 and 2020, the La Capital Department ranked second after the Rosario Department (21.4 against 20.2 in 2013 and 16.4 against 15.7 in 2020).

2.5. Aims, objectives, and hypotheses

The current study was aimed at improving the understanding of the spatial concentration of crime in Latin America, particularly Argentina, to facilitate the creation and implementation of data- and place-based crime prevention strategies.

This study had three objectives. First, this study tested the “law of crime concentration” proposed by Weisburd (2015), which predicts that crime is concentrated in microgeographic units within certain bandwidths, in the city of Santa Fe between January 1, 2001, and December 31, 2020. Second, the study determined if the crime concentration within the expected bandwidth remained stable over time. Third, the study examined if there was persistence in those micro-places that can be characterized as hot spots, i.e., if the stability of the crime concentration was due to the persistence of crime in the same places over time.

Thus, the following hypotheses were tested: 1) In Santa Fe, the places where homicides are committed are spatially concentrated. Therefore, there is a specific spatial relationship, concentration, among the places of occurrence of homicides, such that these places are not dispersed or randomly distributed in space. 2) In Santa Fe, the spatial concentration of the places where homicides are committed remains stable over time. Noteworthily, despite certain fluctuations within a limited range, the concentration remains stable over the years. 3) In Santa Fe, the stability of the spatial concentration of the places where homicides are committed is due to a small number of spatial microunits in which homicides are persistently committed. This implies that the concentration that manifests yearly and is maintained is not due to particular phenomena that occur in different parts of the city but is due to a small set of places where homicides are persistently committed.

3.1. The city of Santa Fe

The city of Santa Fe, the capital of the homonymous province, is located in the center-east of Argentina, north of Buenos Aires, separated by approximately 400 km in a straight line (Fig. 1).

Santa Fe is the main urban nucleus of the La Capital Department. According to the projections of the Provincial Institute of Statistics and Censuses (IPEC) for 2021, the estimated population of the La Capital Department is 588016. Of these, approximately 426145 belong to the city of Santa Fe.

Santa Fe is a city of colonial origin; its location in the territory showed little variation until the second half of the 19th century. According to Peretti and Gómez (2013), the urban area, which has hardly moved away from the main square for 3 centuries, began to grow north at the end of the 19th century, when the city became a railway port epicenter in the agro-export economy. In this regard, this paper argues that the increase in population and the expansion of the city of Santa Fe during the 20th century were closely related to the role and functions that the city and its regional environment experienced at different economic stages through which Argentina passed. Between 1895 and 1930, Santa Fe exercised its role as a city port, consolidating itself as the epicenter of the railway port in the framework of the agro-export model. After 1930, the economic changes that occurred generated new urban transformations. Among these changes, the first informal settlements emerged west of the city, mainly on low-lying land with low real-estate value. In recent decades, the western area lying on the Salado River has recorded an accelerated population growth compared with the rest of the municipality. This western area comprises a significant young population, contrary to the aged population that characterizes other neighborhoods in the city. This area also experiences the highest levels of poverty.

3.2. Data

The Public Prosecutor's Office of the Province of Santa Fe publishes official crime reports annually. As of 2020, these publications were coauthored by the provincial Ministry of Security. For certain years, the data were obtained at the city level; for other years, the data were obtained at the department level, a spatial (and political-administrative) unit larger than cities.

Crime data are not georeferenced. Thus, crime incidents that occurred in the city of Santa Fe from January 1, 2001, to December 31, 2020, were geocoded using open-source data, primarily newspaper reports and accounts. Open-source data are commonly used to identify and collect information on crime incidents, such as homicides (Parkin & Gruenewald, 2017), mass murders (Capellan & Gomez, 2018), and terrorism (Freilich et al., 2014). Here, the open-source data search successfully identified and georeferenced more than 97% of the incidents.

3.3. Unit of analysis

The present study used street segments as a unit of analysis, similar to studies, such as those by Weisburd (2015) and those recently conducted in Latin America. Street segments in the city of Santa Fe were obtained from the microsite of urban plants at the IPEC of the province of Santa Fe. After exploring the data, it was observed that certain lines did not properly correspond to a street segment understood as “both sides of the street between two intersections” (Weisburd et al., 2012, p. 23). Thus, prior to the analysis, the data were cleaned by removing lines that did not represent a street. It was also observed that the original layer contained short lines. Therefore, all lines with lengths of less than 30 m were removed. The clean file contained 12526 street segments. Despite the removal of a significant number of lines, the average length of the segments remained at approximately 103 m (with more precision, 103.45 and 102.75 m for the original and purified layers, respectively).

This specification led to the number of incidents (homicides) that were considerably lower than those of microunits (street segments). This can lead to incorrect conclusions since the crimes will have necessarily occurred in a few microunits (Bernasco & Steenbeek, 2017; Hardyns et al., 2018; Andresen, 2020; Chainey, 2021). To obtain robust results, this study employed a hexagonal cell with an area of 42436 m² as the second unit of analysis. The area of the hexagonal cell was obtained by multiplying the average length of the street segments by itself and then multiplying the obtained value by four. Therefore, it is an area equivalent to that of four average blocks. The hexagonal shape was considered to better fit the shape of the land of the city, achieving a more visually harmonious representation. The tessellation process¹yielded 2059 units. The ratio of the number of incidences and spatial units significantly reduced the problems regarding street segments.

3.4. Analysis plan

To explore the temporal stability of the spatial concentration of homicides in the city of Santa Fe, the data were binned into 5-year ranges, beginning with 2001–2005 and ending with 2016–2020. This was performed to mitigate the asymmetry between the places and incidents (a few homicides and many street segments/hexagonal cells) that occurred when the homicide data were disaggregated by year. The annual number of incidents was very low. Assuming that the maximum possible dispersion was given (i.e., that each incident occurred in a different spatial unit), the results would still be within the bandwidth predicted by the law of crime concentration (Weisburd, 2015).

To examine the persistence of hot spots, two strategies were employed; both strategies used 2059 hexagonal cells of 42436 m². The first strategy involved analyzing the cells using two properties: a) the total number of homicides committed within a cell and b) the number of years in which homicides were committed.

The second strategy involved analyzing the temporal variation of the micro-cells identified as statistically significant hot spots. For every 5 years, local Moran’s I was used to identify the microcells that fell into hot spots (high-high areas). The stability of the hot spots was examined across four observation windows.

^[1] The tessellation was conducted using the street layer as the extension parameter. Afterward, the street layer was used again, selecting, by location, those hexagons that intersected with it. This parameter was preferable because it was possible to cover only those effectively urbanized areas with polygons, leaving out large green spaces, water bodies, and wetlands. Thus, oversizing the final number of cells, which would have led to an overestimation of the concentration, was avoided.

4.1. Spatial concentration of homicides committed between January 1, 2001, and December 31, 2020

After analyzing the homicides committed in the city of Santa Fe from January 1, 2001, to December 31, 2020, where street segments of 103 m in average length were used as spatial units of analysis, it was observed that cumulative proportions of 25 and 50% of the crimes were committed in 1.3 and 3.9%, respectively.

If the maximum dispersion had been reached, i.e., if each homicide had been committed in a different street segment, the values would have been 3.3 and 6.6%, respectively.

Oppositely, considering only the street segments in which at least one homicide was committed, it was observed that cumulative proportions of 25 and 50% of the cases occurred in 12.6 and 39.3% of the street segments, respectively.

Replicating the analysis of the spatial concentration of homicides using hexagonal micro-cells with an area of 42436 m² as spatial units of analysis, it was observed that cumulative proportions of 25 and 50% of the crimes were committed in 2.6 and 7.3% of the micro-cells, respectively.

If the maximum dispersion had been reached, i.e., if each homicide had been committed in a different microcell, the values would have been 20.1 and 40.0%, respectively.

Considering only the micro-cells in which at least one homicide was committed, it was observed that cumulative proportions of 25 and 50% of the crimes occurred in 7.8 and 21.6% of the micro-cells, respectively.

4.2. Temporal stability

At the street segment level and considering the investigated year ranges, it was observed that for cumulative proportions of 25 and 50% of homicides, the spatial concentration ranged between 0.4 and 0.6% and 1.1 and 1.6% of the street segments, respectively.

Considering only the street segments in which at least one homicide was committed, it was observed that for cumulative proportions of 25 and 50% of homicides, the spatial concentration ranged between 14.8 and 20.0% and 43.8 and 50.1% of the street segments, respectively (Fig. 2).

Replicating the preceding analysis using hexagonal micro-cells with an area of 42436 m² as spatial units of analysis, it was observed that for cumulative proportions of 25 and 50% of homicides, the spatial concentration ranged between 1.4 and 2.2% and 3.8 and 4.9% of microcells, respectively (Fig. 3).

Considering only the micro-cells in which at least one homicide was committed, it was observed that for cumulative proportions of 25 and 50% of the homicides, the spatial concentration ranged between 9.6 and 14.6% and 28.7 and 34.5% of the micro-cells, respectively.

4.3. Persistence in the trajectory

The persistence in the trajectory of the hot spots was analyzed using 2059 hexagonal micro-cells with an area of 42436 m2. The findings showed that 66.29% of the micro-cells did not experience homicides during the 20 years (Fig. 4). Furthermore, 32.25, 1.36, and 0.10% of the microcells experienced homicides in 1–5, 6–10, and 11–15 years, respectively. Finally, no microcell experienced homicides in more than 15 years.

A complementary analysis following the same analytical strategy involved concentrating on 694 hexagonal microcells in which at least one homicide was committed. These microcells were characterized according to two variables: the number of homicides committed and the number of years in which homicides were committed within these micro-cells.

For the first variable (number of homicides), the data range was 1–18 with a mean of 2.35 and a standard deviation of 2.03. For the second variable (years with homicides), the data range was 1–13, with a mean of 2.09 and a standard deviation of 1.59.

The microcells with a count above the mean in both variables added up to 181 (Fig. 5); thus, they represented 26.1% of the microcells in which at least one homicide was committed and 8.8% of the total microcells. In addition, 910 homicides were committed in these 181 micro-cells, representing 55.4% of the total homicides.

The microcells with a high count in both variables (between 6 and 18 homicides and between 5 and 13 years) added up to only 40; thus, they represented 5.8% of the microcells in which at least one homicide was committed and 1.9% of the total microcells. In addition, 319 homicides were committed in these 40 micro-cells, representing 19.4% of the total homicides.

The second strategy involved analyzing the temporal variation of the microcells identified as statistically significant hot spots, to understand the trajectory of specific hot spots and if there was persistence in terms of the hot spot areas of the city.

First, this study determined if a statistically significant spatial autocorrelation existed for the dataset (1643 homicides distributed in 2059 hexagonal micro-cells) corresponding to the 20-year period. In this regard, the calculation of the global Moran's I returned the following values: Moran’s index of 0.442364, a Z-score of 31.641528, and a p-value of 0.000000.

Subsequently, the global Moran's I was calculated for year ranges. The results indicated that the statistically significant spatial autocorrelation was stable over time (Table 2). In all the cases, the Z-scores were several standard deviations above the mean, ranging from 11.214490 (2016–2020) to 21.161639 (2010–2014).

Table 2

Statistical analysis of the concentration of homicides by year ranges in hexagonal cells with an area of 42.436 m² (global Moran's I)
Range	Moran’s index	Z-score	p-value
2001–2005	0.228282	16.377622	0.000000
2002–2006	0.216566	15.523989	0.000000
2003–2007	0.207945	14.910397	0.000000
2004–2008	0.196088	14.063264	0.000000
2005–2009	0.213189	15.288688	0.000000
2006–2010	0.247693	17.744057	0.000000
2007–2011	0.268145	19.210153	0.000000
2008–2012	0.256289	18.356700	0.000000
2009–2013	0.270630	19.364233	0.000000
2010–2014	0.295680	21.161639	0.000000
2011–2015	0.247063	17.688447	0.000000
2012–2016	0.260484	18.648418	0.000000
2013–2017	0.244327	17.509304	0.000000
2014–2018	0.241040	17.272397	0.000000
2015–2019	0.174420	12.535764	0.000000
2016–2020	0.155698	11.214490	0.000000
Note: the search was performed with "square inverse distance," and the neighborhood search threshold was set using ArcGIS PRO to 221.3835 m.

After verifying the existence of a statistically significant spatial concentration and the temporal stability of the statistically significant spatial concentration, the statistically significant hot micro-cells were located in space and mapped through the local Moran's I calculation for four 5-year ranges (2001–2005, 2006–2010, 2011–2015, and 2016–2020). These ranges comprised the investigated 20-year period.

Figure 6 shows that the spatial autocorrelation (local Moran's I) was concentrated in the microcells found west of the city, as well as in the northeast and the neighborhood of Alto Verde (located in the southeast area of the city, separated from the rest of the city by a watercourse).

Although it can be concluded that there was persistence in terms of the areas of the city where hot spots were dominant, only five micro-cells remained hot throughout the four 5-year ranges (indicated in yellow in the maps shown in Fig. 6).

The five micro-cells that remained hot according to the local Moran's I calculation (Alto-Alto grouping) represented 0.7% of the micro-cells in which at least one homicide was committed and 0.2% of the total micro-cells. Furthermore, 45 homicides were committed in these five micro-cells, representing 2.7% of the total homicides.

Considering street segments as spatial units of analysis, the results obtained were within the range proposed by Weisburd (2015). The results also corresponded to those obtained by Jaitman and Ajzenman (2016).

Considering the hexagonal cells with an area of 42436 m² as units of analysis, the values were still consistent with those obtained by Jaitman and Ajzenman (2016). However, they represented a lower concentration than expected, according to Weisburd (2015). Thus, by expanding the size of the spatial unit of analysis, thereby reducing the number of units into which the study area was divided (in this case, the city of Santa Fe), a concentration “reduction” will be achieved because a fixed number of incidents is being divided by a smaller number of (spatially larger) units.

Although the present results were consistent with those obtained by Weisburd (2015) and Jaitman and Ajzenman (2016), they did not appear to correlate with those of other studies conducted in the region such as those by Mejía et al. (2015), Chainey et al. (2019), and Chainey and Muggah (2020). These studies generally observed higher spatial concentrations of crime, and these concentrations were even higher for homicides. The disparities could be explained by shorter year ranges being used in these studies.

Mejía et al. (2015) analyzed the data on incidents that occurred over 2 years (2012 and 2013), whereas Chainey et al. (2019) analyzed the incidents of 1 year (2017). Chainey and Muggah (2020) also analyzed the incidents of 1 year (2016).

When crime concentration was calculated in 5-year intervals for the city of Santa Fe, the results were consistent with those obtained by Mejía et al. (2015), Chainey et al. (2019), and Chainey and Muggah (2020). For cumulative proportions of 25 and 50% of homicides, the spatial concentration ranged between 0.4 and 0.6% and 1.1 and 1.6% of street segments, respectively.

These results were consistent with those specifically observed for homicide crimes in previous studies. In this sense, Mejía et al. (2015) and Chainey and Muggah (2020) reported 0.3% of street segments for 25% of homicides, and Mejía et al. (2015), Chainey and Muggah ( 2020), and Chainey et al. (2019) reported 0.9, 1.0, and 1.4% of street segments for 50% of homicides, respectively.

Regarding stability, the present results corresponded to those of Weisburd (2015) and Jaitman and Ajzenman (2016).

Regarding persistence, Weisburd et al. (2012) discovered that most street segments remained relatively stable over 16 years. This corresponds to the coupling of crime to place, as referred to by Weisburd et al. (2016).

Here, the micro-cells with a count above the average in the variables (number of homicides and number of years with homicides) represented 8.8% of the total micro-cells. In these micro-cells, the number of homicides committed represented 55.4% of the total homicides. Among the microcells with a count above average for both variables, those with a high count for both variables (between 6 and 18 homicides and between 5 and 13 years) represented 1.9% of all microcells, and 19.4% of all the homicides (n = 40) were committed in these micro-cells (Fig. 5).

Finally, 66.29% of the microcells never experienced homicides. These results were consistent with those of Weisburd et al. (2004) on the trajectory of crime in Seattle (Washington). They examined more than 1.4 million incident reports from the Seattle Police Department for 14 years (1989–2002) and discovered that approximately half of the street segments did not experience any crime.

This study improved the understanding of the spatial concentration of crime in Latin America, particularly in Argentina. The findings suggest that homicides committed in the city of Santa Fe from 2001 to 2020 were spatially concentrated and that concentration levels remained stable over the years. Although two-thirds of the microunits never experienced homicides over the 20 years, a disproportionate number of homicides occurred in a small group of microunits, and this occurred in many different years.

The findings confirmed the law of crime concentration proposed by Weisburd (2015) and confirmed that this law applied to Latin America, even in an extensive timeframe. However, the study encountered certain limitations, for example, only 97% of the homicides were georeferenced; it is possible, albeit unlikely, that the remaining 3%, if included, could change the results. Additionally, newspaper reports may not provide accurate location data for all the incidents. Even small disparities in location could lead to certain incidents being included in certain microcells and not others.

Consistent with previous studies, the concentration was reported using the percentage of spatial units where 25% and 50% of the homicides were committed. However, the large difference between the number of spatial units (when street segments are used) and criminal incidents can pose problems (Bernasco & Steenbeek, 2017). An attempt was made to partially correct this situation by reporting the percentage of spatial units in which 25% and 50% of the homicides were committed, using only the units in which at least one homicide was committed as the denominator. Furthermore, since there were few incidents in relation to the total number of spatial units, the analysis of the temporal stability could only be performed using microcells, not street segments. Finally, connecting chronic hotspots to environmental features that create opportunities for crime to concentrate there remains a challenge. Therefore, although this study contributes to the determination of where homicides occur, the question regarding what happens where the homicides occur remains unanswered.

Weisburd et al. (2016) argued that the concentration of crime in one place in a city provides a solid basis for the concentration of crime prevention in that place. For the city of Santa Fe, the results of the present study revealed interesting starting points. For example, five microcells were observed to remain statistically significant hot spots over the 20 years. The study findings are expected to aid decision-making regarding place-based crime prevention strategies.

The author has no competing interests to declare that are relevant to the content of this article.

Amemiya, M. & Ohyama, T. (2019). Toward a test of the “Law of Crime Concentration” in Japanese cities: a geographical crime analysis in Tokyo and Osaka. Crime Sci, 8, 11. https://doi.org/10.1186/s40163-019-0106-z
Andresen, M. (2020). Environmental Criminology. Evolution, Theory, and Practice. Second Edition. New York: Routledge.
Bernasco, W. & Steenbeek, W. (2017). More Places than Crimes: Implications for Evaluating the Law of Crime Concentration at Place. J Quant Criminol, 33, 451–467. https://doi.org/10.1007/s10940-016-9324-7
Capellan, J. & Gomez, S. (2018). Change and stability in offender, behaviours, and incident‐level characteristics of mass public shootings in the United States, 1984–2015. Journal of Investigative Psychology and Offender Profiling, 15(1), 51–72. https://doi.org/10.1002/jip.1491
Chainey, S. (2021). Understanding Crime. Analyzing the Geography of Crime. Redlands: Esri Press.
Chainey, S. & Monteiro, J. (2019). The dispersion of crime concentration during a period of crime increase. Secur J, 32, 324–341. https://doi.org/10.1057/s41284-019-00165-x
Chainey, S., Pezzuchi, G., Guerrero Rojas, N. et al. (2019). Crime concentration at micro-places in Latin America. Crime Sci, 8, 5. https://doi.org/10.1186/s40163-019-0100-5
Chainey, S. & Muggah, R. (2020). Homicide concentration and retaliatory homicide near repeats: An examination in a Latin American urban setting. The Police Journal. https://doi.org/10.1177/0032258X20980503
Eck, J., Chainey, S., Cameron, J. et al. (2005). Mapping crime. Understanding hot spots. Washington DC: National Institute of Justice. https://www.ojp.gov/pdffiles1/nij/209393.pdf
Favarin, S. (2018). This must be the place (to commit a crime). Testing the law of crime concentration in Milan, Italy. European Journal of Criminology, 15(6), 702–729. https://doi.org/10.1177/1477370818757700
Freilich, J., Chermak, S., Belli, R., Gruenewald, J., & Parkin, W. (2014). Introducing the United States extremist crime database (ECDB). Terrorism and Political Violence, 26(2), 372–384. https://www.jstor.org/stable/26544642
Gill, C., Wooditch, A., & Weisburd, D. (2017). Testing the “Law of Crime Concentration at Place” in a Suburban Setting: Implications for Research and Practice. Journal of Quantitative Criminology, 33, 519–545. https://doi.org/10.1007/s10940-016-9304-y
Hardyns, W., Snaphaan, T., & Pauwels, L. (2018). Crime concentrations and micro places: An empirical test of the “law of crime concentration at places” in Belgium. Australian and New Zealand Journal of Criminology, 52(3): 390–410. https://doi.org/10.1177/0004865818807243
Jaitman, L. & Ajzenman, N. (2016). Crime concentration and hot spot dynamics in Latin America. BID. https://publications.iadb.org/publications/english/document/Crime-Concentration-and-Hot-Spot-Dynamics-in-Latin-America.pdf
Lee, Y., Eck, J., O, S. et al. (2017). How concentrated is crime at places? A systematic review from 1970 to 2015. Crime Sci, 6, 6. https://doi.org/10.1186/s40163-017-0069-x
Malleson, N., Steenbeek, W., & Andresen, M. (2019). Identifying the appropriate spatial resolution for the analysis of crime patterns. PloS one, 14(6). https://doi.org/10.1371/journal.pone.0218324
Mejía, D., Ortega, D., & Ortiz, K. (2015). Un análisis de la criminalidad urbana en Colombia. Documento de trabajo N° 2015/06. CAF. https://scioteca.caf.com/bitstream/handle/123456789/810/Criminalidad%20urbana%20en%20Colombia%20-%20Enero%202015.pdf?sequence=1&isAllowed=y
Ministerio de Seguridad de la Nación. (2020). Estadísticas Criminales República Argentina – 2019. https://estadisticascriminales.minseg.gob.ar/reports/Informe%20Nacional%20Estadisticas%20Criminales%202019.pdf
-- (2019a). Estadísticas Criminales en la República Argentina – Año 2018. https://estadisticascriminales.minseg.gob.ar/reports/Informe%20Nacional%20Estadisticas%20Criminales%202018_V10.pdf
-- (2019b). Informe Estadísticas Criminales. Homicidios Dolosos. Serie histórica – 2001 al 2018. https://estadisticascriminales.minseg.gob.ar/reports/Informe%20serie%20historica%20HOMICIDIOS%20DOLOSOS%202001%20al%202018.pdf
Ministerio Público de la Acusación [MPA]. (2017). Informe sobre homicidios. Departamento La Capital. Ciudad de Santa Fe. 2016. https://mpa.santafe.gov.ar/static/files/fa4fe162bd01b7cce27e7cc83a1ab34c
Ministerio Público de la Acusación / Observatorio de Seguridad Pública [MPA/OSP]. (2021). Reporte anual. Homicidios. Provincia de Santa Fe. Año 2020. https://www.santafe.gob.ar/ms/osp/informes/informe-anual-sobre-homicidios-en-provincia-de-santa-fe-2020/
Municipalidad de la Ciudad de Santa Fe / Bolsa de Comercio de Santa Fe [MCSF/BCSF]. 2018. Santa Fe. Cómo vamos 2017. https://www.bcsf.com.ar/ces/downloads.php?file=U0ZDVl8yMDE3LnBkZg%3D%3D
United Nations Office on Drugs and Crime [UNODC]. (2019). Global study on homicide. https://www.unodc.org/unodc/en/data-and-analysis/global-study-on-homicide.html
Parkin, W. & Gruenewald, J. (2017). Open-source data and the study of homicide. Journal of interpersonal violence, 32(18), 2693–2723. https://doi.org/10.1177/0886260515596145
Peretti G. & Gómez N. (2013). Territorio, población y expansión de la ciudad de Santa Fe durante el siglo XX. In F. Platino (coord.), Convivir con el Río (pp. 11–13). Santa Fe: MCSF. https://issuu.com/santafeciudad/docs/fasciculo13_convivirconelrio_enbaja
Ridner, H. (2019). The Law of Crime Concentration in Midsized Cities: A Spatial Analysis". Masters Theses & Specialist Projects. Paper 3122. https://digitalcommons.wku.edu/cgi/viewcontent.cgi?article=4139&context=theses
Trento, J. (January 01, 2020). Hubo 102 homicidios en el departamento La Capital en 2019: 87 en la ciudad de Santa Fe. Diario Uno Santa Fe. https://www.unosantafe.com.ar/policiales/hubo-102-homicidios-el-departamento-la-capital-2019-87-la-ciudad-santa-fe-n2553846.html
-- (January 02, 2019). Los asesinatos crecieron un 33 por ciento en 2018: hubo 88 en el departamento La Capital. Diario Uno Santa Fe. Recuperado de https://www.unosantafe.com.ar/policiales/los-asesinatos-crecieron-un-33-ciento-2018-hubo-88-el-departamento-la-capital-n2160938.html
Weisburd, D. (2015). The law of crime concentration and the criminology of place. Criminology, 53, 133–157. https://doi.org/10.1111/1745-9125.12070
Weisburd, D., Eck, J., Braga, A. et al. (2016). Place Matters: Criminology for the twenty-first century. New York: Cambridge University Press.
Weisburd, D., Groff, E., & Yang, S. (2012). The criminology of place: street segments and our understanding of the crime problem. New York: Oxford University Press.
Weisburd, D., Bushway, S., Lum, C. et al. (2004). Trajectories of Crime at Places: A Longitudinal Study of Street Segments in the City of Seattle. Criminology, 42(2), 283–322. http://dx.doi.org/10.1111/j.1745-9125.2004.tb00521.x

Download PDF

Version 1

posted

You are reading this latest preprint version

Crime Concentration and Hot Spot Dynamics: An Examination of Homicides in Santa Fe, Argentina

Status:

Version 1

Abstract

Figures

1. INTRODUCTION

2. THEORETICAL FRAMEWORK

2.1. The spatial concentration of crime

2.2. The temporal stability of the concentration of crime

2.3. Persistence in the trajectory of the hot spots

2.4. Homicides in the city of Santa Fe: the provincial, national, and regional contexts

2.5. Aims, objectives, and hypotheses

3. METHODOLOGY

3.1. The city of Santa Fe

3.2. Data

3.3. Unit of analysis

3.4. Analysis plan

4. RESULTS

4.1. Spatial concentration of homicides committed between January 1, 2001, and December 31, 2020

4.2. Temporal stability

4.3. Persistence in the trajectory

5. DISCUSSION

6. CONCLUSIONS

DECLARATION

REFERENCES

Status:

Version 1