Cities and Violence: An Empirical Analysis of the Case of Costa Rica

DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 1-35 h ps://doi.o g/10.1590/dados.2021.64.1.225 ARTIGOS ORIGINAIS Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica G ego io Gimenez1 Liubo Tkache a2 Ka a ína S i ko á3 Bea iz Ba ado4 1P o esso o Economics and Business a he Uni e si y o Za agoza. Za agoza, Spain. E-mail: g egim@uniza .es. ORCID: h ps://o cid.o g/0000-0003-3702-4017 2P o esso o Psychology a he S a e Uni e si y o S . Pe e sbu g. S . Pe e sbu g, Russia. E-mail: [email p o ec ed]; ORCID: h ps://o cid.o g/0000-0002-9822-1914 3PhD in In e na ional Rela ions a he Facul y o Social Sciences o Cha les Uni e si y. P ague, Czech Republic. E-mail: ka a ina.s i ko a@ s .cuni.cz; ORCID: h ps://o cid.o g/0000-0002-4839-7313 4PhD s uden in Economics a he Facul y o Economics and Business, Uni e si y o Za agoza. Za agoza, Spain. E-mail: [email p o ec ed]; ORCID: h p://o cid.o g/0000-0003-2033-801X Da abase: h ps://doi.o g/10.7910/DVN/3556RW INTRODUCTION The wo ld’s u ban popula ion is booming; by 2050, six and a hal billion people will be li ing in u ban a eas (Wo ld Bank, 2011a). On he one hand ci ies a e signi ican popula ion, in as uc u e and econ- omy hubs, as well as cen e s o poli ical powe ; on he o he hand, acco ding o a ime-p o ed assump ion p o ided by he social sciences, he g ow h o ci ies ends o lead o dis up ion and c ime (Shaw and McKay, 1942). The ocus o his s udy on he ci y le el is in line wi h he g adual ans o ma ion o go e nance and he g owing impo ance o u ban poli ical economies (Sassen, 2006). In a public policy domain, he u ban lens is use ul o add essing si e-speci ic issues – especially in he ield o secu i y (UN-Habi a , 2012; Jai man and Gue e o, 2015). C ime a es end o e eal geog aphic pa e ns, wi h highe concen a- ion in u ban a eas (Eck and Weisbu d, 2015; Johnson e al., 2007; Cu - man, Ande sen and B an ingham, 2015; Gill, Woodi ch and Weisbu d, 2017). Unde s anding he deg ee o which u ban concen a ion a ec s DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 2-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica c ime inciden s is a undamen al issue o be e plan c ime p e en ion s a egies and educe iolence. Focusing on he u ban dimension o iolen c ime, he main aim o his s udy is o examine he ela ionship be ween he deg ee o u ban concen a ion and he numbe o se ious o enses, pa icula ly homicides. Mos exis ing esea ch is ocused on de eloped coun ies 1 and o en has me hodological p oblems, mos ly due o he use o da abases wi h poo da a quali y. This p oblem is especially acu e in de eloping coun ies, whe e c ime s a is ics a e usually agmen ed, inconsis en , and agg ega ed only o he mos mac o le els. The lack o in o ma ion and he weak na ional s a is ic sys ems a e an impo an challenge o conduc ing igo ous esea ch. I is impo an o emphasize ha ou con ibu ion is no only in e ms o p o iding addi ional e idence, bu also in e ms o he ype o con ex ha we analyze. As Ajzenman and Jai man (2016:9) ema k, p o iding empi ical e idence om de eloping coun ies is c ucial, gi en ha “de eloping and de eloped coun ies a e di e en in many dimensions, and also because c ime le els end o be much highe in he de eloping wo ld and, in pa icula , in La in Ame ica”. Speci ically in La in Ame ica, one o he main cha ac e is ics o he phenomenon o c ime and iolence is he deg ee o geog aphic con- cen a ion. U ban homicide a es a e much highe han he a e age homicide a e, and almos hal o all homicides a e concen a ed in 10% o municipali ies (IDB, 2017). Violence in ci ies is no only homicidal and includes all ypes o c imes. Among he main causes associa ed wi h he concen a ion o c imes in u ban a eas, Al a ado and Mug- gah (2018) highligh hose linked o he exis ence o la ge poo and pe iphe al neighbou hoods, he p esence o diso de ly u baniza ion and you h unemploymen . Rega ding ou case s udy, Cos a Rica, Vilal a Pe domo, Cas illo and To es (2016) highligh a se ies o speci ici ies in he link be ween u baniza ion and delinquency. They poin ou ha he Cos a Rican popula ion is la gely concen a ed a ound so-called G an Á ea Me o- poli ana (GAM). They ema k ha , acco ding o he Ins i u o Nacional de Es adís ica y Censos (INEC), in 2011 his a ea occupied 4% o he na ional e i o y and concen a ed 53% o he popula ion. Mos o he coun y’s c imes ake place in his a ea, which is cha ac e ized by highe unemploymen , bu also highe educa ion and income le els. Highe educa ion is ela ed o highe ea nings, which means ha he e DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 3-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado a e mo e po en ial ic ims o c imes agains p ope y. The au ho s also indica e ha , wi hin he GAM, San Jose is conside ed he mos impo an zone o he coun y o he cocaine ade. Consequen ly, he GAM as a whole, and San Jose in pa icula , p esen he g ea es challenges o public policies, especially in ela ion o u ban de elop- men and public sa e y. Ou in-dep h econome ic analysis o Cos a Rica – a coun y o he Global Sou h wi h signi ican u baniza ion and inequali y a es – uses highly disagg ega ed socioeconomic da a. I ocuses on a wide ange o ypes o c imes, in combina ion wi h econome ic echniques which es o he p esence o endogeinei y, which cons i u es a signi ican empi ical con ibu ion o he exis ing li e a u e. The a icle is o ganized as ollows. The nex sec ion p esen s a li e a u e e iew in he ield o u ban iolence in La in Ame ica and Cos a Rica. I s main objec i e is o p o ide a gene al con ex o ci izen insecu i y in he egion and, pa icula ly, he coun y, be o e p esen ing he case s udy in he hi d sec ion. This empi ical pa de elops an econome ic model in o de o demons a e he connec ion be ween u ban concen- a ion and c ime, pa icula ly homicides. URBAN CRIME: CONTRIBUTING FACTORS IN LATIN AMERICA AND COSTA RICA Exis ing esea ch on he ela ion be ween ci ies and secu i y con i ms ha dense u ban a eas gene ally end o be mo e iolen han small owns o u al a eas. This is due o he e y na u e o ci ies as poin s o concen a ion o popula ion and social in e ac ions, including a - ious ypes o con lic s. The link be ween ci ies and c ime has been widely esea ched in he ield o c iminology since he ea ly indings o he Chicago school (Shaw and McKay, 1942). As poin ed ou by Glaese and Sace do e, h ee ac o s a e especially ele an o u ban c ime: “highe pecunia y bene i s, highe chance o a oiding a es and a concen a ion o c ime-p one indi iduals in ci ies” (1999:227). In ecen yea s, he connec ion be ween he u ban en i onmen and c ime has been explo ed using a a ie y o me hodological app oaches and empi ical examples om ci ies all a ound he wo ld (Cecca o, 2012). Among he cha ac e is ics which a ec c ime a es in la ge ci ies, i is cus oma y o alloca e social inequali y and po e y (Cozens, 2007), especially aking in o accoun i s spa ial s a i ica ion wi hin neighbo - DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 4-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica hoods (G ai , Glad el e and Ma hews, 2014); social and acial seg e- ga ion (Jü gens and Gnad, 2002); acial inequali y (Beall, C ankshaw and Pa nell, 2004); he p esence o “ho spo s” o c ime in di e en sec o s o he ci y (Goldsmi h and McGui e, 2000; CAF, 2014; Weis- bu d and Am am, 2014; Weisbu d, 2015; Gill, Woodi ch and Weisbu d, 2017); he exis ence o zones o su icien socioeconomic dep i a ion (Ma ínez J ., Rosen eld and Ma es, 2008; Sho and Yang, 2012); and he con inually inc easing social dispa i ies accompanied by spa ial agmen a ions wi hin he ci y (Coy and Pöhle , 2002; Rodge s, 2004). Th oughou he las decades, heo e ical and empi ical s udies ha e ocused on geog aphic, social and economic aspec s o c ime in ci ies 2 . Taking in o accoun globaliza ion and ci ies’ pe manen expansion, Osca Newman de eloped he heo y o “de ensible space”, which ocuses on c ime p e en ion h ough en i onmen al design (New- man, 1996). A ecen body o esea ch in his ield, e e ed o as “new u banism”, has ound i s way o p ac ical implemen a ion in ci ies, al hough i s c i ics ha e poin ed o a lack o e idence o ac ual c ime educ ion (Cozens, 2011). Policies based on u ban o i ica ions and hei implica ions in e ms o social in eg a ion ha e been c i ically examined (Landman, 2012). Tempo al and seasonal e ec s on iolen c ime in ci ies ha e also been explo ed (Cecca o, 2005; Azjenman and Jai man, 2016). B idging heo y wi h me hods, spa ial and empo al s a is ical analysis o c iminal beha io is o pa icula ele ance o he s udy o iolence in ci ies (Ra cli e, 2010). The e o e, a wide- anging ield o geog aph- ical c iminology has elied on s a is ical me hods o s udy u ban isk, mos ly in de eloped coun ies. Acco ding o Be nasco and El e s (2010:700), his ield o inqui y has b ough subs an ial inno a ions – ei he me hodological o empi ical – o he s udy o iolen c ime. In his con ex , ou a icle aims o de elop a solid s a is ical amewo k o s udy c ime in he La in Ame ican u ban con ex , which equi es he clea iden i ica ion o ac o s which con ibu e o delinquency in ci ies. These ac o s, analyzed below, will be used as con ol a iables in he model de eloped in he subsequen sec ion. Cu en le els o u ban iolence in he Global Sou h a e conside ed a se ious “ h ea o public heal h” (Cecca o, 2014). Thus, acco ding o Be gman (2018), in spi e o apid economic expansion and signi ican imp o emen s in social bene i s, La in Ame ican coun ies ha e su - DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 5-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado e ed unp eceden ed c ime ise and ins abili y, comp omising public sa e y. He highligh s he accessibili y o illici inancial lows, which is so a ac i e o many people – including o icials –, as he main cause. While he de ini ion, na u e, and ex en a y be ween and wi hin coun ies, La in Ame ica is one o he wo ld’s mos iolen egions, wi h iolence p ima ily occu ing in ci ies (Human Secu i y in Ci - ies, 2007; UNODC, 2014; Lappi-Seppala and Leh i, 2014; Jai man and Gue e o, 2015). CAF’s (2014) s udy o he concen a ion o c imes in Suc e (Venezuela) and ou Colombian ci ies (Ba anquilla, Bogo a, Cali and Medellin) e ealed ha 50% o homicides ook place in 1.6% o he blocks o he ci ies, and 50% o he s and obbe ies ook place in app oxima ely 7% o he blocks o he ci ies du ing 2011/2012. This concen a ion has been highligh ed in o he ci ies om B azil, Colombia, Mexico, U uguay, and Venezuela. Using geolocalized da a, Ajzenman and Jai man (2016) ha e ound ha 50% o c imes a e con- cen a ed in 3% o 7.5% o s ee segmen s, and 25% o c imes a e concen a ed in 0.5% o 2.9% o s ee segmen s. Acco ding o Ga i ia and Pagés (2002:190), he isk o becoming a ic im o iolen c ime inc eases subs an ially in La in Ame ican ci ies wi h mo e han a mil- lion esiden s. E en in coun ies whe e mos o he popula ion is u al, c ime ends o concen a e in ci ies. In Gua emala, 40% o homicides occu ed in he capi al ci y in 2006, which houses only 20% o he popula ion (Ma u e and Ga cía, 2007 apud Wo ld Bank, 2011b). Al hough he causes o u ban c ime can depend on a wide a ay o explana o y ac o s, we ocus on hose ela ed o “demog aphic” o “social” condi ions which p o ide an en i onmen in which c iminal ac i i ies a e mo e a ac i e. Rega ding demog aphic condi ions, a popula ion’s age s uc u e is ypically ela ed o c ime a es. Acco ding o adi ional li e a u e om c iminology, bo h pe pe a o s and ic- ims o c iminal ac i i ies a e, in he majo i y o cases, young. Being an immig an has also been ound posi i ely ela ed wi h c ime beha iou (UN-Habi a , 2007:66-71; Wo ld Bank, 2011b). E idence seems o sug- ges ha his may ha e been an impo an ac o in some La in Ame - ican coun ies (Bide man, Mello and Schneide , 2009). Inc eases in popula ion, u baniza ion and densi y ha e also been iden i ied as explana o y ac o s. Di e en s udies ha e showed ha agglome - a ion o popula ion is posi i ely co ela ed wi h c ime (Glaese and Sace do e, 1999). DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 6-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica Among socioeconomic condi ions, inequali y and po e y ha e been iden i ied as impo an explana o y ac o s o c ime a es. These ac- o s, clea ly isible and inhe en o he La in Ame ican egion, become especially appa en in dense u ban a eas. Unequal access o employ- men , educa ion, heal h, and basic physical in as uc u e, along wi h he u ban poo s’ li ing condi ions, a e closely ela ed o he incidence o c ime and iolence, especially a ec ing he poo es (Rodge s, 2004). Mo eo e , neighbou hoods wi h a high concen a ion o economic disad an ages end o ha e less social capi al and collec i e e icacy, uelling he in ensi ica ion o iolence and c ime (Sampson, Rauden- bush and Ea ls, 1997; Sampson and Raudenbush, 2001). S udies ha e suppo ed his ela ionship, inding meaning ul co ela ions be ween inequali y and se e al measu es o c ime a es o di e en La in Ame - ican coun ies (Fajnzylbe , Lede man and Loayza, 2002; Bou guignon, Nunez and Sanchez, 2003; Soa es, 2004). Fu he mo e, a ange o o he d i e s o u ban iolence ha e been iden i ied, including public sensi- i i y o c ime, small-a ms a ailabili y, gang p esence, na co ics ade, d ugs and alcohol use, ine icien c iminal jus ice sys ems, low quali y u ban in as uc u e planning, ha e also been iden i ied as explana o y ac o s o u ban iolence (Fay, 2005; UN-Habi a , 2007; B iceño-León, Villa eces and Concha-Eas man, 2008; Jü e sonke, Muggah and Rod- ge s, , 2009; Soa es and Na i omi, 2010; Muggah, 2012; PNUD, 2013). In he case o Cos a Rica, iolence ac o s a e compa ible wi h hose ound by simila epo s on La in Ame ica. The Minis e io de Salud (2004) includes po e y, inequali y, lack o educa ion, u baniza ion and o e c owding among isk ac o s. Thus, wi hin he g ea e me - opoli an a ea o he capi al, which con inues o expand, isks a e e en highe . Acco ding o he epo (Minis e io de Salud, 2004:38), his phenomenon di ec ly con ibu es o “s uc u al oppo uni ies o homicides”. A he same ime, p oblems de i ed om he social, eco- nomic and demog aphic si ua ion s ongly a ec esiden s’ men al and physical heal h, which can also lead o inc eased a es o iolen c ime. Acco ding o ano he s udy, by he In e -Ame ican De elopmen Bank (2013), he mos signi ican a iable wi h espec o homicide in Cos a Rica is inequali y – which is he key ac o in he econome ic model p esen ed in ou empi ical sec ion. Despi e he ele ance o a iables such as unemploymen , po e y and you h popula ion men ioned abo e, his epo does no conside hese ac o s as con ibu ing o inc eased homicide a es in Cos a Rica (IDB, 2013:81). Taking spa ial DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 7-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado loca ion in o accoun , app oxima ely hal o homicides occu in places which could be conside ed as ypical o an u ban en i onmen , such as s ee s, sidewalks and public pa ks (p. 84). In sum, he e e ences summa ized in his sec ion demons a e a high deg ee o con inui y in e ms o con ex and speci ic a iables ela ed o iolen c imes in La in Ame ica – and pa icula ly in Cos a Rica. The ollowing sec ion analyzes a conside able olume o highly dis- agg ega ed da a ega ding domes ic socioeconomic ac o s which, based on he concep s e e ed o in he p esen sec ion, a e widely belie ed o con ibu e o iolen c ime. Con olling o his ex ensi e se o explana o y a iables, he esul s o he analysis suppo he conclusion ha he e is a ac o , inhe en o ci ies and independen om mos s udied ac o s, which leads o inc eased homicide a es in u ban en i onmen s. This empi ical wo k ela es o academic and policy-o ien ed s udies ci ed abo e, which explo e he link be ween u ban concen a ion, phe- nomena inhe en o u ban socie ies and inc eased le els o iolence. As shown below, he e is an empi ically obus connec ion be ween u ban concen a ion and iolence when con olling o a b oad se o he causal a iables. Case s udy The e a e a numbe o easons why Cos a Rica is sui able o ou empi - ical analysis. Fi s , Cos a Rica is loca ed in La in Ame ica, a egion whe e he issue o iolence has singula impo ance in bo h social and economic e ms. In addi ion, i sha es ce ain endemic cha ac e is ics wi h o he coun ies in he egion, such as he p esence o ansna ional o ganized c ime, p onounced inequali y and pola iza ion in de elop - men be ween u al and u ban a eas. The coun y has expe ienced an ala ming inc ease in c ime a es, especially in he a e o in en ional homicides (UNODC, 2019) and in he amoun o cocaine con isca ion (Unidad de In o mación y Es adís ica Nacional sob e D ogas, 2018). In 2005, he numbe o homicides pe 100,000 popula ion was 7.9 (see Figu e 1). Be ween 2005 and 2017, homicides inc eased by 75% o each 13.8 – a huge inc ease in a egion whe e mos coun ies’ igu es ha e been dec easing 3 . The Wo ld Heal h O ganiza ion conside s a es abo e 10 as an epidemic. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 8-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica Figu e 1 E olu ion o he homicide a e in Cos a Rica (2005-2017) Sou ce: made by he au ho s, based on da a om UNODC (2019). EMPIRICAL ANALYSIS OF CRIME AND URBAN DEVELOPMENT Me hodology The baseline model o he esea ch can be exp essed syn he ically as: Vi = α + βUi + δXi + εi (1) Violence (V i ) in each dis ic i is de e mined by he deg ee o u ban de elopmen (Ui) and a ec o Xi ha con ains a b oad se o explan- a o y a iables (con ol a iables). The s uc u al pa ame e β de e - mines he e ec o u ban de elopmen on iolence. The model is ini- ially es ima ed wi h he homicide a e pe 10,000 inhabi an s as a dependen a iable4. Al hough i is di icul o ind an accu a e indica o o iolence, he homicide a e is he mos widely used in he empi ical li e a u e. Tha happens because he legal de ini ion o homicide ends o be mo e homogenous ac oss coun ies han ha o o he c iminal o enses, hus acili a ing compa ison. In addi ion, i s measu emen is mo e accu a e, because o he o enses, especially misdemeanou s, a e o en no epo ed o he police o a ious easons, such as he cos s and hassles inhe en o he p esen a ion o complain s and he lack o us DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 9-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado in he local police o ce and/o he judicial sys em. Subsequen ly, he es ima ion esul s a e subjec ed o a sensi i i y analysis in which he homicide a e is eplaced by a b oad spec um o c iminal a iables. As explained in sec ion 2, which analysed he ac o s behind c ime a es, he model used can be conside ed a s anda d model in he li - e a u e on iolence, and he choice o a iables included in he ec o X i ollows exis ing empi ical esea ch, as jus i ied in he concep ual amewo k. These a iables a e: numbe o police o ice s deployed pe 10,000 inhabi an s, popula ion densi y, annual popula ion g ow h, p opo ion o men o women, p opo ion o immig an s, Gini index o income inequali y, po e y index o unsa is ied basic needs, unem- ploymen a e and a e age schooling yea s o he popula ion o e 15. Despi e he common usage o he Gini index as he main measu e o inequali y, we ha e conside ed i app op ia e o include, in addi ion o his a iable, a po e y index based on unsa is ied basic needs. The Gini index collec s in o ma ion abou how weal h is dis ibu ed in an economy, bu i does no p o ide an accu a e idea o he popula ion li ing in si ua ions o exclusion. The Gini index may show an equal dis- ibu ion o income e en in egions wi h a high le el o po e y. Gi en ha inequali y and po e y a e clea ly di e en concep s and ha bo h a iables may a ec he a es o iolence, hey a e bo h inco po a ed in o he model. All a iable de ini ions, sou ces o da a, s a is ics and co ela ions a e in Appendix A. Two me hodological conside a ions ega ding he baseline es ima ion a e necessa y. Fi s , due o da a cha ac e is ics, he a iance o he dis u bance is e y likely no cons an ac oss obse a ions, leading o a loss o e iciency o he es ima o . The esul o he B eusch-Pa- gan/Cook-Weisbe g es o he e oscedas ici y, x 2 (1) = 70.86, leads us o ejec he null hypo hesis o cons an a iance. The e o e, obus a iance es ima es and consis en s anda d e o s a e epo ed in he es ima ions. Second, he inclusion o he police o ce as an explana - o y a iable leads us o conside he p esence o simul anei y p ob- lems be ween his a iable and he homicide a e. In o he wo ds, an inc ease in he police o ce may lead o a educ ion in he numbe o o enses, bu , in u n, he c ime a e will condi ion he numbe o police o ice s deployed. This would imply double causali y, which could lead o endogenei y conce ns in he es ima ion, esul ing in E(εi Xi) ≠ 0. I his we e he case, he use o ins umen al a iables could p o ide unbiased and consis en es ima o s, which would be DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 16-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica Table 3 Baseline es ima es o Cos a Rica c ime model wi h u ban de elopmen (dependen a iables: di e en c ime ca ego ies) Mugging Robbe y o i ems in ehicles Robbe y wi h io- lence om people Robbe y wi hou iolence om people Robbe y (non esiden ial wi hou iolence) Robbe y ( esi- den ial wi hou iolence "The (non esiden- ial wi h iolence) The (non esiden ial wi hou iolence) The ( esiden- ial wi h iolence) The ( e- siden ial wi hou iolence) Vehicle he (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Police o ce 0,142 -0,049 -0,001 0,285 0,031 -0,051 0,256 0,062 0,032 -0,017 0,169 (0,733) (0,220) (0,023) (0,967) (0,282) (0,092) (0,222) (0,066) (0,206) (0,028) (0,143) Popula ion densi y -0,410 -0,188 -0,064 -3,029 -1,088 -0,097 -0,674* 0,041 -0,618** -0,029 0,444 (2,647) (0,418) (0,043) (3,176) (0,965) (0,140) (0,378) (0,186) (0,256) (0,032) (0,316) Annual popula- ion g ow h -565,550** -59,241 -6,361 -519,727* -166,673* -7,776 -63,821 -40,378** 76,674 -4,182 -52,704* (254,349) (52,337) (5,751) (300,879) (92,669) (21,040) (42,467) (17,982) (49,101) (3,949) (27,059) Men -609,736 -19,736 4,062 -567,582 -208,863 15,659 27,580 -38,922 52,885 -5,465 -16,540 (423,510) (65,284) (13,106) (511,936) (156,094) (27,635) (68,221) (28,520) (57,868) (6,997) (43,778) Immig an s 488,444* 196,305** 8,940 576,292* 197,143* 47,617* 60,220 32,877* 38,843 6,181 69,495** (272,399) (78,841) (5,929) (325,361) (103,479) (28,305) (47,028) (19,644) (45,696) (4,981) (33,208) Gini index -71,229 41,291** 4,307* 8,426 3,344 36,091*** 45,958*** -9,007 79,816*** -0,082 -10,007 (81,842) (19,206) (2,207) (99,063) (30,363) (7,952) (17,093) (5,909) (17,241) (1,896) (11,471) DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 17-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado Table 3 Baseline es ima es o Cos a Rica c ime model wi h u ban de elopmen (dependen a iables: di e en c ime ca ego ies) (con .) Mugging Robbe y o i ems in ehicles Robbe y wi h io- lence om people Robbe y wi hou iolence om people Robbe y (non esiden ial wi hou iolence) Robbe y ( esi- den ial wi hou iolence "The (non esiden- ial wi h iolence) The (non esiden ial wi hou iolence) The ( esiden- ial wi h iolence) The ( e- siden ial wi hou iolence) Vehicle he (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Po e y index (unsa is ied basic needs) 77,485** 18,280*** 0,060 64,530 25,782* 3,565 18,115** 8,687*** -2,835 4,050*** 4,496 (36,811) (6,766) (1,310) (43,630) (13,591) (3,066) (7,514) (2,816) (7,690) (0,999) (5,724) Unemploymen a e 214,724** 7,029 14,132* 116,661 62,186* 4,860 94,073** 25,691*** 43,536 18,931** 35,016 (93,576) (36,533) (7,281) (96,346) (33,121) (16,979) (38,096) (9,616) (49,536) (7,830) (21,714) A e age schooling yea s +15 10,602** 5,176*** 0,036 9,672 3,751* 0,484 1,921** 1,488*** 1,182 0,607*** 2,503*** (5,362) (1,193) (0,110) (6,377) (1,982) (0,431) (0,961) (0,407) (0,916) (0,106) (0,817) DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 18-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica Table 3 Baseline es ima es o Cos a Rica c ime model wi h u ban de elopmen (dependen a iables: di e en c ime ca ego ies) (con .) Mugging Robbe y o i ems in ehicles Robbe y wi h io- lence om people Robbe y wi hou iolence om people Robbe y (non esiden ial wi hou iolence) Robbe y ( esi- den ial wi hou iolence "The (non esiden- ial wi h iolence) The (non esiden ial wi hou iolence) The ( esiden- ial wi h iolence) The ( e- siden ial wi hou iolence) Vehicle he (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) U ban ca ego y 1. <20 pe cen iles 3,622 1,051 0,362 4,442 1,491 1,151 3,287** 0,334 2,320 0,022 1,697** 2. 20-40 pe cen iles (2,227) (1,517) (0,360) (3,314) (0,964) (0,832) (1,599) (0,272) (2,264) (0,208) (0,780) 0,295 -1,966 -0,269* -3,493 0,340 0,893 2,579 -0,077 0,657 0,304 -0,088 3. 40-60 pe cen iles (3,299) (1,384) (0,151) (3,692) (1,376) (0,711) (1,657) (0,391) (1,973) (0,313) (0,811) 0,023 -0,629 -0,101 0,714 -0,252 2,467** 1,626 -0,183 0,722 -0,122 -0,003 4. 60-80 pe cen iles (5,608) (2,006) (0,231) (6,780) (2,116) (0,974) (1,823) (0,467) (2,069) (0,257) (1,084) 0,923 0,680 -0,167 -2,110 -0,465 1,477 -1,263 -0,016 -1,887 0,080 1,263 5. >80 pe cen iles (5,706) (2,701) (0,248) (6,896) (2,352) (1,114) (1,963) (0,523) (2,555) (0,256) (1,180) Cons an 226,485 -49,651 -3,822 188,139 67,744 -23,942 -45,879 9,189 -54,143 -2,740 -5,270 (198,176) (34,852) (6,312) (240,791) (73,129) (14,783) (36,065) (13,529) (33,445) (3,999) (21,545) Numbe o dis ic s 459 459 459 459 459 459 459 459 459 459 459 R2 0,23 0,34 0,08 0,14 0,18 0,16 0,13 0,33 0,16 0,23 0,35 Adjus ed R2 0,21 0,32 0,05 0,11 0,16 0,13 0,11 0,31 0,14 0,21 0,33 No es: *** p<0.01, ** p<0.05, * p<0.1. Hube -Whi e es ima ed s anda d e o s a e in pa en heses below coe icien s. Sou ce: made by he au ho s, based on he a iables explained in he Table A.1 o he appendix. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 19-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado As o why u ban de elopmen is an explana o y ac o o homicides bu no o o he c imes, i mus be poin ed ou ha he co ela ion be ween homicides and o he c iminal ypologies is ela i ely low: be ween 0.11 and 0.45 (Table 4). Tha is, he e is no clea associa ion be ween homicide and o he c imes, since some homicides occu du ing o he c ime e en s – such as a ape, obbe y o bu gla y – and equen ly in ol e adul o ju enile gang iolence, bu a signi ican numbe o homicides ha e hei o igin in o he ac o s – he in lu- ence o alcohol o na co ics, disag eemen s abou money, a gumen s be ween couples o jealousy a e ypical causes – and a e commi ed by some o he ic im’s in ima e con ac s – iends, boy iends, gi l iends, spouses o ex-spouses. Table 4 Co ela ions be ween homicide a es and o he c ime ca ego ies Mugging Robbe y o i ems in ehicles Robbe y wi h io- lence om people Robbe y wi hou iolence om people Robbe y (non esiden- ial wi hou iolence) Robbe y ( esiden ial wi hou iolence) The (non esiden ial wi h iolence) The (non esiden ial wi hou iolence) The ( esiden ial wi h iolence) The ( esiden ial wi hou iolence) Vehicle he 0,369 0,298 0,115 0,32 0,335 0,29 0,257 0,4 0,152 0,448 0,318 Sou ce: made by he au ho s, based on he a iables explained in he Table A.1 o he appendix. One should also no e ha he co ela ion is s onge when mo e se ious c imes – a acks o esiden ial (0.45) and non- esiden ial buildings (0.4) o mugging (0.37) – a e compu ed and weake in he case o misde- meanou s. This lack o associa ion could be caused by measu emen e o s, o special impo ance in less se ious o enses because, as no ed be o e, hese may no be accoun ed o due o he incon eniences in ol ed in epo ing c imes and a lack o us in police o ice s and/ o he judicial sys em. These measu emen e o s would no allow us o es ima e he ue impac o u ban de elopmen on hese c imes. This is why he empi ical li e a u e o en uses homicide a es as dependen a iables, since hei de ini ion and measu emen a e mo e accu a e. DISCUSSION I is well known ha as u baniza ion has pu esou ces and se ices unde g ea p essu e. In de eloping coun ies, he u ban poo who a e able o ind wo k a e o en bound o spend hei li es a ec ed by DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 20-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica insecu i y and insu icien ly paid jobs. Ru al mig an s as well as immi- g an s o en ha e no o he choice han o se le in slum a eas o owns and ci ies, whe e hey expe ience ex eme po e y (Ha o -Ta el, 2010) while bu dened wi h social seg ega ion, isola ion, spa ial agmen a- ion and g owing ulne abili y (G ai , Glad el e and Ma hews, 2014). Ad e si y is no limi ed o socioeconomic condi ions – his s udy has ocused on he spa ial dimension o insecu i y, ied o he physical en i onmen o he mos ulne able u ban dwelle s. The phenomena men ioned abo e lead o an inc ease in c ime a es in ci ies. The esul s o he p esen s udy alida e his hypo hesis, using highly disagg ega ed da a o ace he occu ence o iolen beha iou in all 473 dis ic s o Cos a Rica. Con olling o a iables usually con- nec ed o c ime inciden s, he co ela ion be ween u baniza ion and homicide is signi ican and obus . The quali y o he da a se used o his s udy, in addi ion o i s speci ic ocus on he ole o u baniza- ion in explaining homicide a es, cons i u e a majo con ibu ion o he li e a u e in he ield. In sum, esul s highligh he signi icance o u ban de elopmen in explaining c ime, measu ed by homicide a es. We do no ind his ela ion in he case o o he o enses. I is clea ha , gi en he cu en u baniza ion a e ends in La in Ame ica, he s udy o he ela ion be ween u ban en i onmen s and iolen c ime is o g ea ele ance. The case o Cos a Rica can hus be conside ed an example o wha happens in La in Ame ica, one o he mos u banized egions in he wo ld, and i s u baniza ion a es, which o e each 80% and con inue o g ow. The a iables adi ionally measu ed in ela ion o – and con- side ed as con ibu ing o – iolen c ime in La in Ame ica con inue o be ele an in local o egional scales. This ein o ces he p essing need o add ess he nega i e e ec s o u baniza ion on ci ies and ci - izens h oughou La in Ame ica and he impo ance o sus ainable u baniza ion in designing de elopmen and public sa e y policies. Ou indings a e in line wi h exis ing li e a u e and simila s udies ocused on he egion. Bu inic and Mo ison (2000) ind ha iolence in La in Ame ica is co ela ed wi h ci y size. The au ho s poin ou ha agglome a ion can in ensi y an i-social beha iou , educing social cohesion and, hence, inc easing c ime a es. The u baniza ion p o- cess has also been posi i ely co ela ed wi h iolence. Ga i ia and Velez (2001) examined Colombian ci ies and ound ha u ban g ow h DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 21-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado is posi i ely co ela ed wi h c ime. In a con ex o apid u baniza- ion, he inabili y o many ci ies o suppo he inc easing demand o u ban in as uc u e and public se ices nega i ely a ec s pub- lic sa e y, allowing diso de ly u baniza ion p ocesses, which in u n a ec s c imes a es. Ri e a (2015) examined he sou ces o c ime in 19 La in Ame ican coun ies o e he pe iod 1980-2010. The au ho inds ha posi i e ends in u ban popula ion a e co ela ed wi h inc eases in coun ies’ mu de a es. As he au ho explains, oppo uni ies o c ime end o be plen i ul in u ban a eas, since social in e ac ion is mo e equen and he isk o punishmen is lowe han in u al a eas. Ca dia (2000) shows ha he apid p ocess o u baniza ion mani es ed in he headlong g ow h o ci ies in B azil is one o he main ac o s o he inc ease in iolen c imes. Acco dingly, Chon (2011) e eals ha high homicide a es o La in Ame ican na ions a e due o high po e y and income inequali y le els, which a e also di ec consequences o ampan u ban g ow h and lack o esou ces. CONCLUSIONS Violence in La in Ame ica cons i u es a conside able public heal h p oblem and has huge social cos s. Unde s anding causes and impac s o iolence has become a op p io i y in o de o design iolence-p e- en ion p og ams and o imp o e li ing s anda ds. The main aim o his s udy was o p o ide a con ex -speci ic unde s anding o iolence, paying pa icula a en ion o c ime in u ban a eas. We conside his app oach especially use ul in La in Ame ica, a egion no o ious o i s high homicide a es, whe e o mal and in o mal u baniza ion is a he p onounced. The empi ical analysis o ou case s udy, using highly disagg ega ed da a on 473 Cos a Rican dis ic s, e eals ha : 1. The deg ee o u ban de elopmen plays a key ole in explain- ing homicide a es, once we ha e con olled o a wide ange o explana o y a iables. 2. This e ec is p og essi e: he g ea e he u ban concen a ion, he g ea e he inc ease in homicide a es. 3. This causal ela ionship be ween iolence and u ban concen a ion is no obse ed in o enses o he han homicide. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 22-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica As o his s udy’s limi a ions, wo main issues a e men ioned in he analysis. Fi s is he omission o da a abou ansna ional o ganized c ime. Al hough his phenomenon is e y signi ican in he geog aph- ical egion o La in Ame ica, including his kind o da a would o e - each he scope o his case s udy, a he conce ned wi h quan i a i e da a on he domes ic le el only. Second, he s udy does no inco po a e da a abou he numbe o legalized weapons despi e i s ele ance o homicides. The main easons o his omission a e he una ailabili y o eliable da a in his espec and he ac ha egis e ing a weapon in a ce ain dis ic does no limi i s use o his pa icula geog aphical uni . Fu he mo e, i is likely ha a signi ican po ion o i ea ms a e held illegally, which would u he educe he eliabili y o such da a. Using only incomple e da a on he possession o legal i ea ms would a he dis o he model. Fu he esea ch in his a ea migh seek o apply his model and his analy ical me hodology o o he coun ies in Cen al Ame ica, as new da abases wi h c ime and socioeconomic a iables a his le el o disagg ega ion become a ailable. Resea che s ypically ace g ea di icul ies in accessing ich and comple e da abases in his egion o he wo ld. A he same ime, “sma policing” which ies o alloca e exis ing esou ces mo e wisely, aking in o accoun a ailable in o ma- ion, has become inc easingly impo an in he igh agains c ime. This app oach elies on an in ensi e use o s a is ical da a and high-qual- i y c iminal analysis, which help o iden i y a eas wi h la ge c ime concen a ion. In his con ex , he esul s o he p esen s udy show he impo ance o wo king wi h such da a, and con ibu e o a wide unde s anding o causes and consequences o u ban iolence. (Recei ed on Decembe 13, 2016) (Resubmi ed on June 4, 2019) (Accep ed on Oc obe 15, 2019) NOTES 1. Fo some excep ions, see Caldei a (2001) and Cecca o (2005). 2. See o example Cecca o (2014). 3. El Sal ado , Jamaica, Hondu as, B azil and Gua emala eached he highes homicide a es du ing 2017 wi h 60, 55.7, 42.8, 29.7, 26.1 homicides pe 100,000 espec i ely. 4. Homicide a es a e usually epo ed o e e y 100,000 o 10,000 inhabi an s, which is also common in c iminology. We chose his la e in o de o make he igu es o he DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 23-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado es ima ed coe icien s o he eg essions mo e easily comp ehensible ( hey a e isible wi h 3 decimals). I we wo ked wi h homicide da a in e ms o 100,000 inhabi an s, he coe icien s o he a iables in he ables would be close o 0.000. Howe e , ollowing e iewe s’ sugges ion, we ha e p o ided s a is ical igu es in e ms o homicides pe 100,000 inhabi an s. 5. Au ho s like Luallen (2006) ha e highligh ed his kind o ela ion wi h empi ical esea ch o U.S. schools, concluding ha loss o school days (due o eache s ikes) signi ican- ly a ec s you h iolence in u ban a eas, al hough he impac is i ele an in u al and subu ban a eas,. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 24-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica REFERENCES AJZENMAN, Nicolas; JAITMAN, Lau a. (2016), C ime concen a ion and ho spo dynamics in La in Ame ica. Washing on, In e -Ame ican De elopmen Bank. Re ie ed om h ps:// publica ions.iadb.o g/bi s eam/handle/11319/7702/C ime-Concen a ion-and-Ho - -Spo -Dynamics-in La inAme ica.pd ?sequence=1. ALVARADO, Na halie; MUGGAH, Robe . (2018), “C imen y iolencia: un obs áculo pa a el desa ollo de las ciudades de Amé ica La ina y el Ca ibe”. IDB Wo king Pape IDB- DP-644, ol. 2. BEALL, Jo; CRANKSHAW, Owen; PARNELL, Susan. (2014), Uni ing a di ided ci y: go e nance and social exclusion in Johannesbu g. London: Rou ledge. BERGMAN, Ma celo. (2018), Mo e money, mo e c ime: p ospe i y and ising c ime in La in Ame ica. Ox o d Uni e si y P ess. BERNASCO, Win; HENK, El e s. (2010), “S a is ical analysis o spa ial c ime da a”. In: A. R. Pique o; D. Weisbu d (eds.), Handbook o quan i a i e c iminology. New Yo k: Sp inge , pp. 699-724. BIDERMAN, Ci o; MELLO, João M. P. de; SCHNEIDER, Alexand e. (2009), “D y laws and homicides: e idence om he São Paulo me opoli an a ea”. The Economic Jou nal, ol. 120, no 543, pp. 157-182. BOURGUIGNON, F ancois; NUNEZ, Jai o; SANCHEZ, Fabio. (2003), “A s uc u al mo- del o c ime and inequali y in Colombia”. Jou nal o he Eu opean Economic Associa ion, ol. 1, no 2-3, pp. 440-449. BRICEÑO-LEON, Robe o; VILLAVECES, And és; CONCHA-EASTMAN, Albe o. (2008), “Unde s anding he une en dis ibu ion o he incidence o homicide in La in Ame ica”. In e na ional Jou nal o Epidemiology, ol. 37, no 4, pp. 75-757. BUVINIC, May a; MORRISON, And ew R. (2000), “Li ing in a mo e iolen wo ld”. Fo eign Policy, no 118, pp. 58-72. CALDEIRA, Te esa P. R. (2001), Ci y o walls. C ime, seg ega ion and ci izenship in São Paulo. Oakland, Uni e si y o Cali o nia P ess. CARDIA, Nancy. (2000), U ban iolence in São Paulo. Washing on: Wood ow Wilson In e - na ional Cen e o Schola s. CARMONA VILLALOBOS, Keylo ; RAMOS ESQUIVEL, Ma io A.; SÁNCHEZ MATA- RRITA, Fe nando J. (2005), Dimensión espacial de la pob eza, dis ibución del Ing eso y pola ización social en Cos a Rica. PhD Thesis, Uni e sidad de Cos a Rica, San José. CECCATO, Vania. (2005), “Homicides in São Paulo, B azil: assessing spa ial- empo al and wea he a ia ions”. Jou nal o En i onmen al Psychology, ol. 25, no 3, pp. 307-321. CECCATO, Vania (ed.). (2012), The u ban ab ic o c ime and ea . Do d ech : Sp inge Ne- he lands. CECCATO, Vania. (2014), “The geog aphic, socioeconomic and cul u al de e minan s o iolence”. In: P. D. Donnelly; C. L. Wa d (eds.), Ox o d ex book o iolence p e en ion: epidemiology, e idence, and policy. Ox o d: Ox o d Uni e si y P ess, pp. 77-86. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 25-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado CHON, Don Soo. (2011), “Con ibu ing ac o s o high homicide a e in La in Ame ica: A c i ical es o Neapoli an’s egional subcul u e o iolence hesis”. Jou nal o Family Violence, ol. 26, no 4, p. 299. COY, Ma in; POHLER, Ma in. (2002), “Ga ed communi ies in La in Ame ican megaci ies: case s udies in B azil and A gen ina”. En i onmen and Planning B: Planning and Design, ol. 29, no 3, pp. 355-370. COZENS, Paul M. (2007), “Public Heal h and he po en ial bene i s o c ime p e en- ion h ough en i onmen al design”. New Sou h Wales Public Heal h Bulle in, ol. 18, no 11-12, pp. 232-237. _____. (2011), “U ban planning and en i onmen al c iminology: Towa ds a new pe spec i e o sa e ci ies”. Planning P ac ice and Resea ch, ol. 26, no 4, pp. 481-508. CURMAN, And ea S. N.; ANDERSEN, Ma in A.; BRANTINGHAM, Paul J. (2015), “C ime and place: a longi udinal examina ion o s ee segmen pa e ns in Vancou e , BC”. Jou nal o Quan i a i e C iminology, ol. 31, no 1, pp. 127-147. ECK, John E.; WEISBURD, Da id. (2015), “C ime places in c ime heo y”. C ime and Place: C ime P e en ion S udies, ol. 4, pp. 1-33. FAJNZYLBER, Pablo; LEDERMAN, Daniel; LOAYZA, No man. (2002), “Inequali y and iolen c ime”. The Jou nal o Law and Economics, ol. 45, no 1, pp. 1-39. FAY, Ma ianne (ed.). (2005), The u ban poo in La in Ame ica. Washing on: The Wo ld Bank. GAVIRIA, Alejand o; VELEZ, Ca los Edua do. (2001), Who bea s he bu den o c ime in Co- lombia?. Re ie ed om h ps://ss n.com/abs ac =256746. GAVIRIA, Alejand o; PAGÉS, Ca men. (2002), “Pa e ns o c ime ic imiza ion in La in Ame ican ci ies”. Jou nal o De elopmen Economics, ol. 67, no 1, pp. 181-203. GILL, Cha lo e; WOODITCH, Alese; WEISBURD, Da id. (2017), “Tes ing he ‘Law o C i- me a Place’ in a subu ban se ing: Implica ions o esea ch and p ac ice”. Jou nal o Quan i a i e C iminology, ol. 33, no 3, pp. 519-545. GLAESER, Edwa d L.; SACERDOTE, B uce. (1999), “Why is he e mo e c ime in ci ies?”. Jou nal o Poli ical Economy, ol. 107, no S6, pp. 225-258. GOLDSMITH, Vic o ; McGUIRE, Philip G. (2000), Analysing c ime pa e ns: on ie s o p ac ice. London: Sage Publica ions. GRAIF, Co ina; GLADFELTER, And ew S.; MATTHEWS, S ephen A. (2014), “U ban po e y and neighbo hood e ec s on c ime: Inco po a ing spa ial and ne wo k pe spec i es”. Social Compass, ol. 8, no 9, pp. 1140-1155. HARROFF-TAVEL, Ma ion. (2010), “Violence and humani a ian ac ion in u ban a eas: new challenges, new app oaches. Humani a ian deba e: Law, policy, ac ion”. U ban Violence. In e na ional Re iew o he Red C oss, ol. 92, no 878, pp. 329-351. HUMAN SECURITY IN CITIES. (2007), Human Secu i y o an U ban Cen u y. Local Chal- lenges, Global Pe spec i es. Re ie ed om h ps://docs.unocha.o g/si es/dms/HSU/ human_secu i y_ o _an_u ban_cen u y%20Sou h%20Ame ica%20%281%29.pd . DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 32-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica Table A.1 Lis o a iables, sou ces and summa y s a is ics (con .) Va iable De ini ion Sou ce Yea Mean S.d. U ban popula ion Pe cen age o he popula- ion o a dis ic li ing in an u ban a ea. U ban a eas we e de ined a p io i by Ins i u o Nacional de Es adis ica y Censos wi h physical and unc ional c i e ia, aking in o accoun angible i ems such as clea ly de ined quad an s, s ee s, sidewalks, u ban se ices (ga bage collec- ion, s ee ligh ing) and economic ac i i ies. The adminis a i e cen es o each coun y o dis ic pe o med geog aphic delinea ion. INEC and IGN 2011 0.523 0.377 Vehicle he O ense consis ing o he seizu e o ano he ’s ehicle wi h he in en ion o p o i - ing, employing o ce, io- lence o in imida ion. Pe e e y 10,000 inhabi an s. OIJ and INEC A e age alue 2010- 2013 7.531 8.801 Sou ces: CRS: Ca mona; Ramos and Sánchez (2005); IGN: Ins i u o Geog á ico Nacional; INEC: Ins i u o Nacional de Es adís ica y Censos; MSP: Minis e io de Segu idad Pública; OIJ: O ganismo de In es igación Judicial. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 33-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado Table A.2 Co ela ion ma ix Homicide a e Police o ce Popula ion densi y Annual popula ion g ow h Men Immi- g an s Gini index Po e y index Unemploy- men a e A e age schooling yea s +15 U ban popula ion Homicide a e 1.000 Police o ce -0.028 1.000 Popula ion densi y 0.280 0.156 1.000 Annual popula ion g ow h -0.097 0.041 -0.175 1.000 Men -0.124 -0.143 -0.542 -0.008 1.000 Immig an s 0.120 -0.022 0.046 0.185 -0.011 1.000 Gini index 0.122 -0.017 -0.076 0.216 -0.162 0.101 1.000 Po e y index 0.089 -0.145 -0.421 -0.074 0.678 0.005 0.022 1.000 Unemploy- men a e 0.181 0.006 0.116 0.034 -0.125 -0.073 0.224 -0.011 1.000 A e age schooling yea s +15 0.085 0.204 0.554 0.076 -0.776 0.149 0.174 -0.824 0.026 1.000 U ban popu- la ion 0.200 0.242 0.579 0.087 -0.674 0.093 0.069 -0.673 0.138 0.760 1.000 Sou ce: made by he au ho s, based on he a iables explained in he Table A.1 o he appendix. DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 34-35 Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica RESUMO Cidades e Violência: Uma Análise Empí ica do Caso de Cos a Rica Es e a igo oca no e ei o da u banização no que ange a c imes iolen os – pa icu- la men e homicídios – na Cos a Rica. Apesa de a iolência se um g a e p oblema em oda a Amé ica La ina, poucos es udos empí icos conduzidos nessa á ea usam bancos de dados socioeconômicos e de c iminalidade de al a qualidade, com al o ní el de desag egação geog á ica. Nes e a igo, emp egamos dados de odos os 473 dis i os de Cos a Rica en e 2010 e 2013. Desen ol emos um modelo que le a em conside ação p oblemas de endogeneidade e usa con as es de p edições linea es ma ginais. Concluímos que o g au de concen ação u bana desempenha um papel de e minan e na explicação de axas de homicídio, man endo cons an es as demais a iá eis. Esse e ei o é p og essi o: quan o maio a concen ação u bana, maio es são as axas de homicídio. Essa elação causal não é obse ada em ou os c imes. Pala as-cha e: c ime; iolência; u bano; cidades; Amé ica La ina; Cos a Rica ABSTRACT Ci ies and Violence: An Empi ical Analysis o he Case o Cos a Rica This a icle ocuses on he e ec o u baniza ion on iolen c ime – pa icula ly homicide in Cos a Rica. Al hough iolence is a majo p oblem h oughou La in Ame ica, ew empi ical s udies ca ied ou in he a ea use high-quali y socio- economic and c ime da abases wi h a high le el o geog aphical disagg ega ion. In his a icle, we employ da a om all 473 dis ic s o Cos a Rica be ween 2010 and 2013. We de elop a model which akes in o accoun endogenei y p oblems and uses con as s o ma ginal linea p edic ions. We conclude ha he deg ee o u ban concen a ion plays a key ole in explaining homicide a es, o he hings being equal. This e ec is p og essi e: he g ea e he u ban concen a ion, he g ea e he inc ease in homicide a es. This causal ela ionship is no obse ed in o enses o he han homicide. Keywo ds: c ime; iolence; u ban; ci ies; La in Ame ica; Cos a Rica DADOS, Rio de Janei o, ol.64 (1): e20190127, 2021 35-35 G ego io Gimenez, Liubo Tkache a, Ka a ína S i ko á and Bea iz Ba ado RÉSUMÉ Villes e Violence: Une Analyse Empi ique du Cas du Cos a Rica Ce a icle se concen e su l’e e de l’u banisa ion su la c iminali é iolen e – en pa iculie l’homicide au Cos a Rica. Bien que la iolence soi un p oblème majeu dans ou e l’Amé ique La ine, peu d’é udes empi iques menées dans la égion u ilisen des bases de données socio-économiques e c iminelles de hau e quali é a ec un ni eau éle é de désag éga ion géog aphique. Dans ce a icle, on u ilise les données des 473 dis ic s du Cos a Rica en e 2010 e 2013. On dé eloppe un modèle qui p end en comp e les p oblèmes d’endogénéi é e u ilise des con as es de p édic ions linéai es ma ginales. On conclu que le deg é de concen a ion u baine joue un ôle clé dans l’explica ion des aux d’homicides, ou es choses é an égales pa ailleu s. Ce e e es p og essi : plus la concen a ion u baine es éle ée, plus l’augmen a ion des aux d’homicides es impo an e. Ce e ela ion de cause à e e n’es pas obse ée dans les in ac ions au es que l’homicide. Mo s-clés: c ime; iolence u baine; illes; Amé ique La ine; Cos a Rica RESUMEN Ciudades y Violencia: Un Análisis Empí ico del Caso de Cos a Rica Es e a ículo se cen a en analiza el e ec o de la u banización en los deli os io- len os – pa icula men e el homicidio en Cos a Rica. Aunque la iolencia es un p oblema gene alizado a lo la go de Amé ica La ina, pocos es udios ealizados en es a á ea usan bases de da os socioeconómicas y de c iminalís ica con un al o ni el de seg egación geog á ica. En es e a ículo, u ilizamos da os de los 473 dis i os de Cos a Rica en e 2010 y 2013, pa a lo cual desa ollamos un modelo que ome en cuen a los p oblemas de endogeneidad y que u ilice con as es de p edicciones linea es ma ginales. Concluimos que el g ado de concen ación u bana juega un papel undamen al pa a explica las asas de homicidio, en igualdad de o as condiciones. Es e e ec o es p og esi o: cuan o mayo es la concen ación u bana, mayo es el aumen o de las asas de homicidio. Es a elación de causalidad no se obse a en deli os dis in os al homicidio. Palab as cla e: c imen; iolencia; u bano; ciudades; Amé ica La ina; Cos a Rica