03 LA Riots Uprising

8/12/2019 03 LA Riots Uprising

1/17

Immigration, Race, and Riot: The 1992 Los Angeles UprisingAuthor(s): Albert Bergesen and Max HermanSource: American Sociological Review, Vol. 63, No. 1 (Feb., 1998), pp. 39-54Published by: American Sociological AssociationStable URL: http://www.jstor.org/stable/2657476.

Accessed: 18/09/2013 17:26

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

American Sociological Associationis collaborating with JSTOR to digitize, preserve and extend access to

American Sociological Review.

http://www.jstor.org

This content downloaded from 68.181.10.34 on Wed, 18 Sep 2013 17:26:11 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=asahttp://www.jstor.org/stable/2657476?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2657476?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=asa


2/17

IMMIGRATION, RACE, AND RIOT:THE 1992 LOS ANGELES UPRISING*Albert Bergesen Max HermanUniversity f Arizona University f Arizona

We test the hypothesis that the 1992 Los Angeles race riot represents back-lash violence in response to recent Latino and Asian immigration into Afri-can American neighborhoods. Wepropose a variant of ethnic competitiontheory that links residential ethnic succession withpropensities for riot vio-lence. Wedepartfrom previous research on riots by comparing census tractsrather than cities, and we find that, controlling for economic conditions andracial/ethnic composition, there is a significant association between ethnicsuccession in neighborhoods (Latino and Asian in-migration and black out-migration) and riot violence.t 3:00 P.M. on April 29, 1992, the ac-quittal verdict of four white Los An-geles police officers charged with the beat-ing of the African American motorist RodneyKing was announced by a jury in Simi Val-ley, California (see Jacobs 1996). Almost im-mediately, crowds gathered in South CentralLos Angeles to protest the verdict. By 4:15

P.M. there were reports of looting at the inter-section of Florence Boulevard andNormandie Avenue, and by 5:45 there werereports of motorists being assaulted. At 8:15P.M. the first fatality was reported, and in thenext six days of rioting 2,383 people wereinjured, 8,000 were arrested, 51 were killed,and over 700 businesses were burned. Prop-erty damage was estimated at over 1 billiondollars (Webster 1992:23). The hardest hitareas were South Central Los Angeles andKoreatown. It was the largest outbreak of ra-cial violence since the riots of the 1960s.Previous explanations of the 1992 riot havebeen based on interpretations of employ-ment, poverty, and immigration statistics(Baldassare 1994; Chang and Leong 1994;

Gooding-Williams 1993; Hazen 1992;Morrison and Lowry 1994; Oliver, Johnson,and Farrell 1993), but these studies do notsystematically test hypotheses purporting toexplain the riot's underlying dynamics. Inthis paper we formally test a variant of eth-nic competition theory (Olzak 1992) as anexplanatory model of the L.A. riot. Previousempirical research has primarily used city-level aggregations of economic, political,and demographic data (Lieberson andSilverman 1965; Myers 1997; Olzak andShanahan 1996; Olzak, Shanahan, andMcEneaney 1996; Spilerman 1970, 1971,1976). While such studies have made impor-tant contributions to understanding city-leveldifferences in the frequency and severity ofrioting, this high level of aggregation maynot fully capture the ecological dynamics ofgroup contact thathave been hypothesized tounderlie the L.A. riot (Davis 1993a, 1993b;Miles 1992; Oliver, Johnson, and Farrell1993; Olzak et al. 1996). For example, re-search has shown that people tend to engagein riot activity close to where they live(Capeci and Wilkerson 1991; Porter andDunn 1984; Webster 1992:134), making theestimation of economic and demographiccharacteristics of local neighborhoods cen-tral to any explanation of why and where ri-oting takes place. Therefore, we use censustracts rather than cities as our basic unit ofanalysis. Although the census tract is not aperfect approximationof the social constructof "neighborhood," it should provide a rea-sonable approximation of local conditions.

*Direct all correspondence to Albert Bergesen,Department of Sociology, University of Arizona,Tucson, Arizona 85721 ([email protected]).An earlier version of this paper was presented atthe 1994 annual meeting of the American Socio-logical Association in Los Angeles. For assis-tance and comments on early drafts the authorsthank Stanley Lieberson, Michael Hout, LenHochberg, Calvin Morrill, Doug McAdam,Michael Sobel, Paula England, Jim McElvany,ASR reviewers, and the ASR Editor.

American Sociological Review, 1998, Vol. 63 (February:39-54) 39

http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/17

40 AMERICAN SOCIOLOGICAL REVIEWETHNIC COMPETITION THEORYOlzak and her colleagues have advanced atheory explaining race riots that emphasizescompetition among ethnic groups overboundaries and identities ratherthan empha-sizing simple economic conditions. General-izing ecological ideas of Barth (1969) onniche overlap and resultant conflictual com-petition, they generate the counterintuitivehypothesis that desegregation heightenscompetition among ethnic groups, which inturn can result in ethnic violence that in-cludes riots (Olzak 1992; Olzak andShanahan 1996; Olzak et al. 1996). "In con-trast to models emphasizing that segregationof ethnic and racial populations causes eth-nic conflict, competition argumentshold thatdesegregation causes ethnic conflict. A corehypothesis . . . is that desegregation of labormarkets intensifies ethnic competition,which in turn raises the rate of ethnic collec-tive action." (Olzak 1992:3). Efforts by adominant residential ethnic group to excludein-migrants are also considered part of thecompetition process (Olzak et. al. 1996).A recent analysis of riots from 1960through 1993 that includes the 1992 L.A. riotproposes to "examine a new hypothesis-ur-ban settings that experience increases in in-terracial contact (defined as increasing resi-dential exposure of Blacks and Whites) willexperience more unrest. . . " (Olzak et al.1996:591). A backlash by whites againstblacks migrating into their neighborhoodsappears to explain the racial violence of theearly twentieth century, such as the 1917East St. Louis, 1919 Chicago, and 1921Tulsa riots (Grimshaw 1960; Janowitz 1969;Lieberson 1980; Massey and Denton 1993;Wilson 1978). But increasing residential ex-posure of blacks and whites was not an issuein the 1992 Los Angeles riot, the 1980 Mi-ami riot (Bergesen 1982; Portes and Stepick1993), or in any of the riots that broke outduring the 1960s.Competition models may be misspecifiedwhen African Americans are considered theonly minority population. Although AfricanAmericans are minorities at the national leveland in many cities, they are not always mi-norities in the local neighborhood level. Intheoretical models assuming society-wideprocesses, specifying blacks as minorities is

the correct model specification. But forBarthianethnic competition models in whichthe hypothesized dynamics operate at theneighborhood or census tract level, the cor-rect specification-at least since the 1960s inmost major U.S. cities-will have AfricanAmericans as the inner-city majority.This assumption is particularly importantfor ethnic competition theory, as its centralproposition hypothesizes exclusionary vio-lence by dominant ethnic groups in reactionto the intrusion of in-migrating minorities. Inthis regard, South Central Los Angeles, theepicenter of the riot, is 60.4 percent black,and a correctly specified competition modelshould treat African Americans as the major-ity, not the minority, population. With Afri-can Americans now correctly identified asthe inner-city majority, Mexican, CentralAmerican, and Koreanimmigrants constitutethe relevant in-migrant minority. In such acorrectly specified competition model, thehypothesis that "desegregation following ini-tially high levels of segregation sparksracialtensions that lead to rioting" (Olzak et. al.1996:592) fits the racial situation in Los An-geles. Note, however, that it isn't the deseg-regation of white neighborhoods, but the de-segregation of black neighborhoods that ishypothesized to be associated with riot vio-lence. The great demographic shift in LosAngeles during the decade preceding the1992 riot involved an out-migration of blacksand an in-migration of Latinos/Asians. Thisproduced a classic ethnic succession process,or what could be called the "desegregation"of black neighborhoods in South CentralLosAngeles. In sum, desegregation may lead toethnic violence, but the ethnic group whoseneighborhood is being desegregated must becorrectly identified: In 1992 it was the blackneighborhoods of South Central LosAngeles.Although economic competition suppos-edly results from desegregation ( Olzak1992; Olzak and Shanahan 1966; Olzak et.al. 1966), there is little evidence of economiccompetition among blacks, Latinos, and Ko-reans that could produce exclusionary vio-lence such as a race riot. Espenshade andGoodis (1985), for instance, report thatvariation in the numberof immigrantshad noeffect on unemployment rates for blacksacross 247 SMSAs, and Bean, Lindsay, and



4/17

THE LOS ANGELES RIOT 41Taylor (1988) find only small effects of un-documented Mexican immigration on the av-erage wage of black, white, and nativeLatinos in a study of 47 southwestern metro-politan areas. A study of relations betweenblacks and new immigrants concludes: "Thecase for displacement would have to rest ona finding that immigrants pushed African-Americans out of jobs that the latter onceheld; evidence of that sort, however, is hardto find" (Waldinger 1996:314). Several stud-ies have examined the effects of immigrationon wages and unemployment for blacks(Borjas 1990; Fix and Passel 1994; Lewis1994; Simon 1989). A review of these stud-ies concludes ". . . that variation in the num-ber of immigrants produced little effect oneither unemployment or wages of the nativepopulation as a whole" (Heer 1996:187).With regard to the question of black/Latinoniche overlap and economic competition forjobs, Heer concludes, "[N]one of the studiesreviewed so far provide evidence of harm toblacks in California due to high levels of im-migration from Mexico" (Heer 1996:187).Similarly, Wilson concludes that there is not"the kind of interracial competition and con-flict that has traditionally plagued the labormarket ... [such that] theories which associ-ate labor-marketconflicts with racial antago-nisms have little application to the presentperiod of modern industrial race relations"(Wilson 1978:16).HYPER-ETHNIC SUCCESSIONTHEORYWe propose an ethnic competition modelsimilar to Olzak's. Our formulation differsfrom Olzak's version, however, by acknowl-edging African Americans as the majoritypopulation when the neighborhood is the rel-evant level of analysis. We also propose thatthe key element is not so much labor marketcompetition, but the arrivalof new racial/eth-nic immigrants into residential areas with adifferent racial/ethnic majority alreadyestab-lished. In Los Angeles, this means the rapidin-migration of Latinos and Asians in the1980s into neighborhoods that were pre-dominantly African American. In such a situ-ation, we hypothesize defensive backlashviolence from African Americans in thoseneighborhoods in which the rate of in-migra-

tion by nonblack minorities is increasing.Our model differs from Olzak's in that wepredict riot violence from rates of residentialin-migration rather than from competitionfor housing or jobs. We argue that housingand job competition are secondary processessuch that by the time a neighborhood has be-come desegregated enough to foster politicaland economic competition, the high rates ofin-migration already will have had their ef-fect.What could be called hyper-ethnic succes-sion, then, represents demographic situationsin which a rapid in-migration of new ethnicgroups alters the ethnic composition and pro-duces resentment that fuels defensive back-lash violence by a residential area's domi-nant racial/ethnic group. This hypothesis re-sembles those proposed in studies that findan association between prejudice and reac-tionary collective violence by groups dis-placed through rapid social change(Bergesen 1977, 1980; Hofstadter 1955;Olzak 1992; Tilly 1969, 1978) and the resis-tance of a dominant group to rising minori-ties (Blalock 1967; Bobo and Hutchings1996; Bobo and Zubrinsky 1996; Quillian1995; Vanneman and Pettigrew 1972).A clear example of collective violence inreaction to such hyper-ethnic succession isthe 1900-1920 wave of race riots thatJanowitz (1969:393) called the "communalriot,"which represented"aninterracialclash,an ecologically based struggle at the bound-aries of the expanding black neighborhoods."

During the years immediately preceding itsrace riot each city experienced large increasesin Negro population primarily because of theinflux from the South. Between 1910 and 1917the East St. Louis community grew from nearly6,000 to perhaps as many as 13,000. In Chi-cago there were nearly 110,000 Negroes in1920 compared to 44,000 a decade earlier. De-troit in 1940 had about 160,000 Negro resi-dents, but three years later there were an esti-mated 220,100 Negroes. (Rudwick, as quotedin Wilson 1978:75)The racial/ethnic groups engaged in suchcommunal violence have changed since the1900-1920 period, but we argue that the un-derlying sociological process remains thesame. During the 1980s, immigration was atits highest since 1900-1910 as some 8.6 mil-lion immigrants entered the United States



5/17

42 AMERICAN SOCIOLOGICAL REVIEW(Webster 1992). In Los Angeles in the 1980sthe relevant wave of immigrants came largelyfrom Mexico, Central America, and Asia, andit infringed upon black residential areas. Asof the 1990 census, there were more foreign-born residents in Los Angeles County than inany other SMSA (Heer 1996:185), and theethnic composition of the city was rapidlychanging. In 1980, Los Angeles was 28 per-cent Latino and48 percent white, but by 1990it was 40 percent Latino and only 37 percentwhite. Furthermore, during this same periodthe African American share of the city'spopulation declined from 17 to 13 percent,and the Asian share grew from 7 to 13 per-cent (Webster 1992:36). The ethnic transfor-mation of Los Angeles was underway, and asMorrison and Lowry (1994) note, "[T]he pat-tern of ethnic succession in this part of LosAngeles County reveals ... potential under-lying tensions between groups.... The blackcommunity of South Central Los Angelesmay regard the incursion of Hispanics as'threatening'" (p. 32). While this link be-tween ethnic succession andriot violence hasbeen suggested elsewhere (Baldassare 1994;Miles 1992; Oliver, Johnson, and Farrell1993) it has not yet been formally tested.TESTING HYPER-ETHNICSUCCESSION THEORYTo test the hypothesis that hyper-ethnic suc-cession leads to collective violence we gatherdata at the census-tract level on riot violence,ethnic population change, and economic con-ditions (to control for the competing hypoth-esis that poverty and economic deprivationare the causal factors). Our population con-sists of all census tracts in Los AngelesCounty. Along with citywide analyses, wemore closely examine South Central Los An-geles, the contiguous city of Compton, andthe region that encompasses the areas ofKoreatown and Pico Union.1 Together theseareas account for 80 percent of the fatalities.

Dependent VariableIndicators of riot violence can include ar-rests, injuries, property damage, and fatali-ties. Each has specific biases. For example,many arrestswere for curfew violations, andas such these arrests may include personswho were not actual riot participants. It alsohas been suggested "that police and federalagencies used the time of civil unrest toround up and deport people who they be-lieved to be undocumented (American CivilLiberties Union 1992:1), which may bias thepercentage of Latinos arrested(Pastor 1993).Injured persons may choose not go to thehospital, or if the nearby hospital is full theymay go to another hospital. Thus, injuriesmay be underreportedand subject to selec-tion bias. Property damage is also subject toselection bias, and estimates of damage costsmay be inflated for insurance purposes.2We focus primarily on riot fatalities,which as a measure of riot violence also hasits limitations. Fatalities are relatively rare.Because most riot-related deaths in Los An-geles resulted from gunshot wounds, theirperpetrators are usually unknown (Webster1992). Yet for our purposes fatality data areless biased than other indicators: Althoughthe number of riot deaths is small, their num-ber is fairly accurate, specification of theneighborhood is likely to be accurate, and

I Within the city of Los Angeles these areas aredefined as follows: South Central is bounded onthe north by the Santa Monica Freeway, on thewest by Van Ness Boulevard, on the south by Im-perial Highway, and on the east by Central Av-enue. Compton includes the city of Compton andparts of Lynwood and South Gate. Koreatown isbounded on the north by Sunset Boulevard, on the

west by Crenshaw Boulevard, on the south by theSanta Monica Freeway, and on the east by HooverAvenue. Pico Union is bounded on the north bySunset Boulevard, on the west by Hoover Avenue,on the south by the Santa Monica Freeway, andon the east by the Harbor Freeway. Because SouthCentralL.A./Compton and Koreatown/Pico Unionare geographically contiguous and share demo-graphic similarities, they are combined in some ofthe analyses. South Central L.A. andCompton are60.4 and 54.2 percent black, respectively, whereasKoreatown and Pico Union are only 15.2 and 6.1percentblack. Koreatown and Pico Union are 19.7and 16.1 percent Asian, respectively, and SouthCentral and Compton are only 2.9 and 2.3 percentAsian, respectively.2 In a separate analysis, we plotted addressesof reported property damage on a Los Angelescensus tract map shaded for racial/ethnic compo-sition (see Figure 3). We acknowledge the assis-tance of Mike Ridland of Environmental SystemsResources Institute for providing data on the lo-cations of reported property damage.



6/17

THE LOS ANGELES RIOT 43Table 1. Selected Population Characteristics of Different Regions of Los Angeles County, 1990

RegionSouth OtherLos Angeles Central Los Angeles

Characteristics County Los Angeles Koreatown Pico Union Compton County AreasPercentblack 11.7 60.4 15.2 6.1 54.2 7.7Percent foreign-born 30.8 28.4 58.6 63.6 26.0 29.1Percent Hispanic 34.1 35.0 49.0 65.8 41.2 32.5Percent Asian 10.6 2.9 19.7 16.1 2.3 10.7Percent white 58.6 10.9 36.6 33.1 12.9 64.1Percent unemployed 7.6 13.4 10.7 11.0 15.9 6.8Medianhousehold income 38,109 20,766 22,060 17,368 24,683 40,551(in 1990 dollars)Numberof riot fatalities 51 15 10 6 10 10

Source: U.S. Bureau of the Census (1990).

there is less bias in how data were gathered(less, for example, than police bias in arrest-ing Latinos). We decided on riot fatalities asour measure because they are a clear and un-ambiguous indicator of serious racial vio-lence. They also meet an important require-ment for this ecological research-a fairlyunambiguous and unbiased specification ofwhere the violence occurred (i.e., the neareststreet intersection where a body was found).We employ several statistical techniques, in-cluding Poisson regression, designed tomodel rare events.Thus the dependent variable in this analy-sis is the presence or absence of a riot fatal-ity in a Los Angeles census tract. The near-est street intersection for each of the 51 re-ported riot fatalities was identified from po-lice reports (Webster 1992, app. 9-3:5; LosAngeles Times, May 4, 1992, p. A8). Usingthis information, we located these intersec-tions on a map of census tracts for the LosAngeles- Long Beach SMSA. Census tractswere then coded according to whether theyhad a riot fatality (yes = 1; no = 0).3 The ma-jority of fatalities occurred at street intersec-tions that were also the defining border oftwo or more census tracts. Because in these

cases we could not be sure in which tract thefatality occurred, we assigned a 1 to eachtract whose boundary was also the street in-tersection where a riot fatality occurred.Independent VariablesThe ethnic transition process has two parts:an established residential group is movingout and an immigrant group is moving in.Our empirical analysis to test our theoreticalmodel required estimating the rates of Latinoand Asian in-migration and African Ameri-can out-migration. To do this we looked atthe changing population of African Ameri-cans, Latinos, Asians, and the general censuscategory of "foreign-born," for each censustract during the decade prior to the riot(1980-1990). For each tract we also com-piled data on its racial/ethnic composition,the percent and proportion of the labor forceunemployed, and the median household in-come in both 1980 and 1990 (U.S. Bureau ofthe Census 1980, 1990). Table 1 presentssummary statistics for the neighborhoods in-cluded in the analysis.

AnalysisAs a preliminary exploration of the ethnicsuccession hypothesis, we employ a differ-ence-of-means test (t-test) that comparestracts with and without a fatality on each ofour independent variables. We expect thattracts with a fatality will have higher rates of

3 There are 1,637 census tracts in Los AngelesCounty, and 92 were coded as having a fatality.Of the 129 tracts in South Central L.A./Compton,44 were coded as having a fatality, and of 99tracts in Koreatown/Pico Union, 24 were codedas having a fatality.



7/17

44 AMERICAN SOCIOLOGICAL REVIEW

a} SX~~~~~~~~~~ofX5~~~~~~~~~~~~X X .'.'-Sfi m , g i : E7YX~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~....10~~~~~~~~~~~~~~~~~~~~~~~~~.. ........

Miles~~~~~~~~~~......| * Riot fatality, t992 g Latino population > 50%, 1990 g Black population > 50%,1990... ...................Figure 1. Riot Fatalities and Ethnicl~~~~~~~~~~~~~~~~~~~~~~~acialomposition Los Angeles County Census Tracts,1990~~~~~.........................

black out-migration and Latino and Asian in-migration then do tracts without a fatality.Next, we perform a logistic regression analy-sis on the likelihood of riot fatalities occur-ring in a tract while controlling for the ef-fects of changes in ethnic in- and out-migra-tion, economic deprivation, and relative sizeof the host and immigrantracial/ethnic popu-lations.4RESULTSWe begin with a visual representation of ar-eas of ethnic succession and riot fatalities.Figure 1 plots the street locations of riot fa-talities and census tracts that are at least 50percent African American or at least 50 per-cent Latino.5 The lighter shaded tracts on the

right-hand side of Figure 1 represent moreheavily Latino neighborhoods, such as EastLos Angeles; the darker shaded tracts on theleft-hand side represent more heavily AfricanAmerican neighborhoods. Most of the dotsrepresenting riot fatalities cluster along theborder between the more heavily AfricanAmerican and Latino census tracts.Figure 2 presents another perspective. Ifwe imagine a Venn diagram of two circles,with one being heavily Latino and the otherheavily African American, the overlap areawill be the zone that has a high proportion ofboth Latinos and African Americans. Thiscould be considered a "contact zone," orwhat has been referred to in earlier riots as a"contested area" (Grimshaw 1960). Figure 2plots fatalities on a map where the shadedcensus tracts contain at least 40 percentAfri-can Americans and at least 40 percentLatinos. This "contact zone" has the largestBecause riot deaths are rare events, we alsoran our logistic models with Poisson regression.Results were comparable to those of the logisticregression analysis.5 Asian Americans account for only 3.4 percent of the population in South Central Los Angeles,so their presence is not indicated on these maps.



8/17

THE LOS ANGELES RIOT 45

Mile

N*Hft I \ '~~~~~~~~~~~.1''..1.-......,_

Fiur 2. Rio Faaite in Ara of:':EtncR ia Ovra: Los Angle Cont Cesu Tat,{ 1990O

\-{ _ - Lz,,,,,,,,~~~~~~~~~~~........W I e } z~~~~~~~~~~~~~~~~~~~~~~~~.......

.....~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ii.; .. ....g~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~... ....._.. . . . .M les.. .. ...* Riotatality,992 2:1 Latinopopulation> 40%, 1990, and black population.40%,.1990

Figure2. Riotatalities n Areasof Ethnic/RacialOverlap:LosAngelesCountyCensusTracts,990.......

cluster of riot fatalities: The rest of the fa-talities are scattered around Los Angeles, ex-cept for a smaller cluster in the Koreatown/Pico Union area (just north of the shadedarea in Figure 2).Finally, Figure 3 displays the relationshipbetween mixed ethnic composition and prop-erty damage in the 1992 L.A. riot. Theshaded areas again represent tracts that areat least 40 percent African American and 40percent Latino-the zone where the twopopulations overlap. The dots represent ad-dresses of buildings that sustained propertydamage.6 Figure 3 shows that the preponder-ance of property damage occurred in theshaded overlap zone. Riot fatalities and prop-erty damage, then, were not concentrated inareas that were overwhelmingly African

American. This suggests that rioting is not asimple function of segregation; rather, vio-lence occurs in areas of mixed racial/ethniccomposition.Difference-of-Means TestThe hyper-ethnic succession hypothesis ar-gues that backlash violence is more likely tooccur in areas where rates of in- and out-mi-gration are high. To examine this, we statis-tically compare rates of in- and out-migra-tion for those census tracts that had a fatalitywith those that did not have a fatality. If thehyper-ethnic succession hypothesis is sup-ported, those tracts with fatalities shouldhave significantly higher rates of in- and out-migration. Table 2 presents data that supportthis hypothesis. For Los Angeles censustracts with fatalities, the mean increase inLatino and Asian population between 1980and 1990 is significantly greater than the in-crease for tracts without fatalities. Similarly,Los Angeles County tracts with fatalities

6Reported by the Civil Disorder Damage Sur-vey and compiled by the City of Los Angeles De-partmentof Building and Safety, revised October23, 1992. Data file obtained from Mike Ridlandof Environmental Systems Resources Institute,Inc.



9/17

46 AMERICAN SOCIOLOGICAL REVIEW

4P ~ ~ N

0 S:

8 0 8 16Miles

|* Property amage, 1992 HLatino opulation 40%,1990,and blackpopulation 40%,1990|Figure 3. Reported Riot Property Damage in Areas of Ethnic/Racial Overlap: Los Angeles CountyCensus Tracts, 1990

have a significantly higher rate of black out-migration in this period than do tracts withno riot fatalities. In sum, tracts with riot fa-talities have higher rates of ethnic successionthan do tracts with no fatalities.Examining the two clusters of violenceshown in Figure 2 emphasizes the same sig-nificant relationship of higher rates ofLatino in-migration and black out-migrationfor the South Central Los Angeles/Comptontracts with fatalities compared to tracts inthis area with no fatalities. Rates of Asianin-migration into South Central L.A./Compton, though, are not significantly asso-ciated with fatalities, perhaps because thearea is only 1 percent Asian-there was notenough Asian in-migration to generate back-lash violence. For tracts harboring the sec-ond cluster of fatalities, Koreatown/PicoUnion (which has a much smaller blackpopulation than South Central L.A./Compton), tracts with fatalities do not differsignificantly from tracts with no fatalities in

terms of rates of Latino and Asian in-migra-tion or African American out-migration.Logistic Regression AnalysisThe previous analysis was conducted with-out control variables. The significant differ-ences observed between tracts with and with-out fatalities could disappear after we con-trol for the relative sizes of the black, Latino,and Asian populations and for the level ofeconomic deprivation. Table 3 presents re-sults of a logistic regression analysis inwhich we enter variables for unemploymentand median household income along with theethnic composition variables. In this analy-sis, the census category "foreign-born"serves as a general indicator of immigrantpopulation. For Los Angeles County (col-umns 1 and 2) Table 3 shows a significantrelationship between fatalities and the out-migration of blacks and the in-migration offoreign-born, while controlling for changes



10/17

THE LOS ANGELES RIOT 47Table 2. Demographic Characteristics of Census Tracts with and without a Riot Fatality, by Region:Los Angeles County, 1980 and 1990

Los Angeles South CentralL.A./ Koreatown/County Compton Pico Union OtherAreasTracts Tracts Tracts Tracts Tracts Tracts Tracts Tractswith a with No with a with No with a with No with a with NoCharacteristic Fatality Fatality Fatality Fatality Fatality Fatality Fatality Fatality

Percent Change, 1980-1990Black -11.7 ttt -1.2 -21.1 tt -13.6 -2.0 -3.1 -4.1 .0Foreign-born 13.7 ttt 8.8 16.5 tt 11.8 11.7 10.3 10.6 8.7Latino/Hispanic 17.8 ttt 7.5 23.9 ttt 17.3 13.2 11.5 11.2 6.7Asian 1.9 t 4.8 -.1 .4 3.4 3.5 5.2 4.0White -4.6 -10.5 .5 -2.1 -13.7 -9.6 -4.8 t -11.0

Mean Percent, 1990Black 31.5 ttt 10.5 55.7 58.2 6.6 14.0 12.1 7.7Foreign-born 44.2 ttt 30.0 30.7 t 25.4 58.2 66.7 46.6 ttt 28.8Latino/Hispanic 53.6 ttt 33.0 43.9 t 35.0 52.8 60.0 64.7 ttt 31.9Asian 8.3 t 10.7 1.1 t 3.4 21.2 17.6 8.5 10.7White 26.0 ttt 60.5 10.2 12.9 36.5 35.1 44.6 ttt 64.5

Meanhousehold 21,800 ttt 39,080 19,958 24,134 19,343 20,722 26,527 ttt 40,791income (in 1990dollars)Percent unemployed 13.6 ttt 7.3 16.6 t 13.6 10.4 10.9 11.1 ttt 6.7Numberof census 92 1,501 44 85 24 75 24 1,397tracts

Sources: U.S. Bureauof the Census (1980, 1990).tp < .05 ttp < .01 tttp < .001 (t-tests for difference between means)

in income, unemployment, and the relativesizes of the black and foreign-born popula-tions. This important finding runs counter toSpilerman's (1970, 1971) well-known find-ing of an association between the size of theblack population and the frequency/severityof rioting. Although Spilerman comparedcit-ies, we compare census tracts. Both, how-ever, are defined areas in which riot violenceoccurred. More important, in the 1960s riotsthe principal correlate of racial violence wasthe size of the black population, but in this1992 riot the correlate is a declining blackpopulation, even when a number of controlvariables are added to the equation.7 This

suggests a re-thinking of the generalizabilityof the 1960s model of racial violence.We now conduct a logistic regressionanalysis for the epicenter of the riot, SouthCentral L.A./Compton. Table 3 (columns 3and 4) shows that the succession variables-change in proportion black and change inproportion foreign-born-remain significantafter adding the control variables for income,unemployment, and the relative sizes of theblack and foreign-born populations. In SouthCentral L.A./Compton, the control variablesfor economic deprivation and relative sizesof the black and foreign-born populations areno longer significant. These findings suggestthat the ethnic transition hypothesis isstrongly supported in the neighborhoods of7 Blalock's (1967) power/threat hypothesis sug-gests a curvilinear relationship between the sizeof the minority population and violence. In sepa-rate analyses, we found no curvilinear effect for either the percent black or percent Latino and thepresence of riot fatalities.



11/17

48 AMERICAN SOCIOLOGICAL REVIEWTable 3. Logistic Regression of Riot Fatalities on Change in Racial/Ethnic Composition and EconomicCharacteristics, by Region: Los Angeles County 1990 Census Tracts

Los Angeles South CentralL.A./ Koreatown/County Compton Pico UnionIndependentVariable (1) (2) (3) (4) (5) (6)Racial/Ethnic Composition

Change in proportion -3.904** -6.617** -7.797black, 1980-1990 (1.031) (2.200) (9.108)Changein proportion 2.625* 7.122* 3.036foreign-born, 1980-1990 (1.325) (3.186) (3.655)Proportion black, 1990 .948 .454 -9.384(.505) (1.068) (5.409)Proportion oreign-born,1990 1.927*** .692 5.171*(.674) (1.760) (2.439)

Economic CharacteristicsChangein proportion -.862 -.601 .893 1.365 -6.268 -8.886unemployed, 1980-1990 (2.534) (2.709) (3.445) (3.495) (6.881) (6.625)Change in median household -.150*** -.170*** .004 .002 .010 .024income, 1980-1990 (.024) (.025) (.049) (.049) (.025) (.029)(in $1,000s)

Constant 1.075** -1.542** -2.290* -2.029* -.140 -5.029*(.351) (.493) (1.066) (.862) (.621) (2.098)Chi-square 140.207***122.905*** 12.217* 9.975* 10.183* 9.806**Degrees of freedom 4 4 4 4 4 4Number of census tracts 1,591 1,591 129 129 97 97

Note: Numbers n parenthesesare standard rrors.Sources: U.S. Bureauof the Census (1980, 1990).aEstimates and standard rrorsshown for changein median householdincome aremultipliedtimes 1,000.p < .05 Hip< .01 *** < .001 (two-tailed tests)

South Central Los Angeles and the city ofCompton. For the Koreatown/Pico Unionarea(columns 5 and 6) results show that onlyproportionforeign-born is significantly asso-ciated with fatalities.Rates of Racial/Ethnic ChangeThe hypothesis linking hyper-ethnic succes-sion with riot violence suggests it is not justthe changing proportion of racial/ethnicgroups, but the rate of change in relativegroup sizes that triggers racial violence.Table 4 enters variables for the rate of changein the proportion of black and foreign-bornresidents from 1980 to 1990 along with thecontrol variables. For Los Angeles Countycolumn 1 shows a statistically significantnegative association between the rate of

change in proportionblack and riot fatalities:The more rapid the decline in the blackpopulation, the more likely are fatalities.Again, this is the opposite of Spilerman's(1970, 1971) finding for the 1960s riots of asignificant association between the size ofthe black population and riot violence. Col-umn 2 of Table 4 shows a significantly posi-tive association in Los Angeles County be-tween rate of change in the foreign-bornpopulation and fatalities: The more rapid thein-migration the more likely are fatalities.Both of these findings remain statisticallysignificant when control variables for eco-nomic deprivation and the relative size of theblack and foreign-born populations are in theequation.We repeatedthis analysis for South CentralL.A./Compton (columns 3 and 4). The rate of



12/17

THE LOS ANGELES RIOT 49Table 4. Logistic Regression of Riot Fatalities on Rate of Change in Racial/Ethnic Composition andEconomic Characteristics, by Region: Los Angeles County 1990 Census Tracts

Los Angeles South CentralL.A./ Koreatown/County Compton Pico UnionIndependentVariable (1) (2) (3) (4) (5) (6)Racial/Ethnic Composition

Rate of change in proportion -7.066** -8.547** -.699Black, 1980-1990 (1.680) (2.78) (.565)Rateof change in proportion .358** .167* -.126foreign-born,1980-1990 (.102) ( 100) (.976)Proportion black, 1990 .971 - 3.751* -6.476(.490) (1.74) (3.559)Proportion oreign-born,1990 3.208** 3.321 4.832(.741) (1.546) (2.573)

Economic CharacteristicsChangein proportion -.812 -1.116 .971 .164 -.594 -8.940unemployed, 1980-1990 (2.526) (2.756) (3.465) (3.439) (6.821) (6.666)Change in median household -.150** -. 150*** -.014 -.006* -.010 -.031income, 1980-1990 (.024) (.026) (.052) (.048) (.036) (.045)(in $1,OOOs)

Constant -2.081* -2.294** -5.248** -1.995* -.301 -4.236*(.699) (.569) (1.899) (.867) (.492) (2.087)Chi-square 144.075 135.509** 14.709** 8.469 11.201* 9.083Degrees of freedom 4 4 4 4 4 4Number of census tracts 1,591 1,591 129 129 97 97

Note: Numbers n parenthesesare standarderrors.Sources: U.S. Bureau of the Census (1980, 1990).a Estimates and standarderrors shown for rate of change in median household income are multipliedtimes 1,000.* < .05 < .01 (two-tailed tests)

decline of the black population and the rateof increase in the foreign-born population issignificantly associated with fatalities, con-trolling for economic deprivation and therelative sizes of the black and foreign-bornpopulations. These relationships are not sig-nificant for Koreatown/Pico Union.These results support our hypothesis link-ing changing rates of in- and out-migrationwith fatalities in Los Angeles. The analysisfor the Koreatown/Pico Union neighbor-hoods, though, did not show the same asso-ciation. Hayes-Bautista, Hayes-Bautista, andSchink (1993) and Davis (1993a, 1993b)suggest that while the rioting in South Cen-tral L.A./Compton is associated with rates ofethnic succession, the rioting in Koreatown/Pico Union may representopportunistic loot-

ing that occurred after the riot had begun inSouth Central L.A. and then spread north-ward. The marginally significant associationbetween percent foreign-born and fatalitiesin Koreatown/Pico Union may indicate thepresence of poor Salvadoran immigrantswho, along with others, engaged in extensivelooting once the riot began (Morrison andLowry 1994:29). In this regard, the largestgroup of immigrants into Koreatown/PicoUnion are ethnically similar to the residen-tially dominant ethnic group (i.e., Latino im-migrants are moving into neighborhoods al-ready dominated by Latinos). This may pro-duce a conflict between new and establishedimmigrant groups of the same ethnicity(Miles 1992), thus, along with black/LatinoKorean tensions, there also may be tensions



13/17

50 AMERICAN SOCIOLOGICAL REVIEWTable 5. Logistic Regression of Riot Fatalities on Interethnic Contact, Racial/Ethnic Composition,and Economic Characteristics, by Region: Los Angeles County 1990 Census Tracts

Los Angeles South CentralL.A./ Koreatown/County Compton Pico UnionIndependentVariable (1) (2) (3) (4) (5) (6)Racial/EthnicComposition

Interethniccontact measure 7.936* 8.003** 14.506 15.782** 7.817 1.775(3.173) (2.724) (7.572) (5.677) (11.487) (6.175)Proportionblack, 1990 3.981** 4.764** 5.622 5.905* -2.673 -1.230(1.143) (.955) (7.443) (3.338) (3.281) (3.182)Proportion Mexican, 1990b 2.298 - 4.747 - -3.732(1.252) (8.314) (4.289)Proportion non-Mexican 7.288** 10.352 4.523Hispanic, 1990C (1.649) (7.490) (3.508)Proportion Asian, 1990 3.284* -10.406 .853(1.589) (15.706) (2.819)Proportion oreign-born, 1990 6.775** 8.822* 3.419(1.501) (5.141) (3.263)

Economic CharacteristicsProportionunemployed, 1990 4.243 4.317 3.103 4.237 -7.746 -7.934(3.036) (3.016) (4.475) (4.662) (9.357) (8.745)Medianhousehold -.02 -.009 .02 .015 -.004 .006income, 1990 (.02) (.02) (.05) (.049) (.035) (.033)(in $1,000s) d

Constant -6.277** -7.436** -9.307 -9.878* -1.514 -2.648(1.434) (1.468) (7.695) (4.654) (2.292) (2.590)-2 log-likelihood 216.773** 215.101** 23.599** 22.456** 12.066 6.383Degrees of freedom 7 5 7 5 7 5Number of tracts 1,589 1,589 123 123 97 97

Note: Numbersin parenthesesarestandarderrors.Sources: U.S. Bureau of the Census (1980, 1990).a Black, Mexican, otherHispanic,Asian.b Spanish-speakingpersonsof Mexican heritage.c Spanish-speakingpersonsnot of Mexican heritage.d Estimates and standarderrorsshown for medianhousehold income aremultipliedtimes 1,000.


14/17

THE LOS ANGELES RIOT 51panics (largely Central Americans), andAsians.Table 5 examines Blalock's hypothesis ofethnic conflict when ethnic groups shareequally in the census tract population. Thereis no significant association between fatali-ties and the interethnic contact variable fortracts in Koreatown/Pico Union. The unem-ployment and household income variablesalso do not have a significant effect. Thus,both the economic deprivation and the intra-ethnic succession hypotheses are not sup-ported in this region of Los Angeles.The riot violence in Koreatown/PicoUnionmay represent a spillover or diffusion of thecollective violence (Pitcher, Hamblin, andMiller 1978) that began the previous day inSouth Central L.A. Normandie and VermontAvenues flow directly north from South Cen-tral L.A. into Koreatown, and the riot movedup these corridors. Some data suggest thatmany of those rioting in Koreatown werefrom South CentralL.A. (Webster 1992, app.8-15 to 8-17). For example, arrest recordsshow, that only 37 percent of those arrestedin the police bureau that includes Koreatownactually lived there and that 35 percent actu-ally lived in the south and central police bu-reaus. In contrast, among those arrested inthe south and central police bureaus, two-thirds were residents of the area and only 5percent were from the police bureau that in-cludes Koreatown. Thus it appears that riot-ers from South Central L.A. had moved upVermont and Normandie Avenues intoKoreatown and were arrested there, whereasmost of those arrestedin the south and cen-tral bureaus resided there.DISCUSSIONWe have shown that for Los Angeles Countyas a whole, changing percentages of blacksand immigrants in the decade prior to the1992 Los Angeles riot are significantly asso-ciated with riot fatalities, controlling for therelative sizes of these groups and for mea-sures of economic deprivation.This relation-ship also holds for the epicenter of the riot-South Central Los Angeles and the city ofCompton-but not for the neighborhoods ofKoreatown and Pico Union. These findingsadd support to the growing literature linkingethnic group boundaries, identities, and com-

petition to racial prejudice and conflict(Olzak 1992; Olzak et al. 1996; Quillian1995). However, these results also suggestsome rethinking of traditionally held notionsabout the underlying causes of racial vio-lence.To begin with, the conceptual model ofU.S. race relations-built on a black/whitefulcrum-is incapable of handling the multi-cultural vectors of ethnic tension that wehave shown to be associated with racial/eth-nic violence. We second the call to developexplanatory models that move beyond the"prevailing bipolar black/white model ofrace relations" (Omi and Winant 1994:153)toward a multivalent model of race that rec-ognizes the Technicolorrcomplexity of eth-nic and class conflict" (Davis, 1993b:34). AsWaldinger (1996) notes, "[N]ow ... that im-migration has returnedAmerica's biggest cit-ies to the diverse, multiethnic worlds theyonce were, analysis in black and white willno longer do" (pp. 300-301). Thus, variantsof ethnic competition theory that specifyonly blacks as minorities when they are of-ten majorities at the neighborhood level, con-tinue to misspecify the dynamics of U.S. ra-cial violence. African Americans react likeany other residential majority to a rapid in-flux of new immigrants. Given rapid in-mi-gration, whether during the period 1900-1920 or 1980-1990, a defensive reactionamong the host residential majority may oc-cur, whether that host group be white orblack. Backlash violence is not limited to onegroup or one circumstance. This is not to saythat resistance by whites to black in-migra-tion has ended-residential segregation con-tinues to be high (Massey and Denton 1993).But Latino or Asian in-migration into resi-dential neighborhoods occupied by blackscan generate a backlash.Second, and more generally, two character-istics of the 1992 L.A. riot suggest that werethink the cause of racial/ethnic violence:the seeming reappearance of what Janowitz(1969) called the "communal riot," and ourfindings (which contradict Spilerman's1960s results) that it is areas with a relativedecline in black population that have riotviolence. The "1960s model" of race riots aspolitical protest against economic conditionsand a struggle to gain political power maynot be the current model. The riots of the



15/17


16/17

THE LOS ANGELES RIOT 53Rodney King/Reading Urban Riots. New York:Routledge.Grimshaw, Allen D. 1960. "Urban Racial Vio-lence in the United States: Changing Ecologi-cal Considerations." American Journal of So-ciology 64:109-19.Hayes-Bautista, David E., Maria Hayes-Bautista,and Werner 0. Schink. 1993. "Latinos and the1992 Los Angeles Riots: A Behavioral Sci-ences Perspective." Hispanic Journal of Be-havioral Science 15:427-48.Hazen, Don, ed. 1992. Inside the L.A. Riots. NewYork: Institute for Alternative Journalism.Heer, David. 1996. Immigration in America's Fu-ture. Boulder, CO: Westview.Hofstadter, Richard. 1955. The Age of Reform.New York: Random House.Jacobs, Ronald N. 1996. "Civil Society and Cri-sis: Culture, Discourse, and the Rodney KingBeating." American Journal of Sociology 5:1238-72.Janowitz, Morris. 1969. "Patterns of CollectiveRacial Violence." Pp. 393-422 in Violence inAmerica, edited by H. D. Graham and T. R.Gurr.New York: New American Library.Lewis, Staughton Y. 1994. "Impacts of U.S. Na-tives' Employment and Earnings: A Summaryof the Evidence." Pp. 169-215 in A Stone'sThrow from Ellis Island: Economic Implica-tions of Immigration to New Jersey, edited byT. J. Espenshade. Lanham, MD: UniversityPress of America.Lieberson, Stanley. 1980. A Piece of the Pie:Blacks and WhiteImmigrationsince 1880. Ber-keley, CA: University of California Press.Lieberson, Stanley and A. R. Silverman. 1965."The Precipitants and Underlying Conditionsof Race Riots." American Sociological Review30:887-98.McAdam, Doug. 1982. Political Process and theDevelopment of Black Insurgency. Chicago, IL:University of Chicago Press.Massey, Douglas S. and Nancy A. Denton. 1993.AmericanApartheid: Segregation and the Mak-ing of the Underclass. Cambridge, MA:HarvardUniversity Press.Miles, Jack. 1992. "Blacks vs. Browns: Immigra-tion and the New American Dilemma." TheAt-lantic 270:41-68.Morrison, Peter A. and Ira S. Lowry. 1994. "ARiot of Color: The Demographic Setting." Pp.19-46 in The Los Angeles Riots: Lessons forthe Urban Future, edited by M. Baldassare.San Francisco, CA: Westview.Myers, Daniel J. 1997. "Racial Rioting in the1960s: An Event History Analysis of LocalConditions." American Sociological Review62:94-112.Oliver, Melvin L., J. H. Johnson, Jr., and W. C.Farrell, Jr. 1993. "Anatomy of a Rebellion: A

Political-Economic Analysis." Pp. 117-41 inReading Rodney King/Reading Urban Upris-ing, edited by R. Gooding-Williams. NewYork: Routledge.Olzak, Susan. 1992. The Dynamics of EthnicCompetition and Conflict. Stanford, CA:Stanford University Press.Olzak, Susan and Suzanne Shanahan. 1996."Deprivation and Race Riots: An Extension ofSpilerman's Analysis." Social Forces 74:931-61.Olzak, Susan, Suzanne Shanahan, and ElizabethH. McEneaney. 1996. "Poverty, Segregation,and Race Riots, 1960-1993." American Socio-logical Review 61:590-613.Omi, Michael and Howard Winant. 1994. RacialFormation in the United States From the 1960sto the 1990s. 2d ed. New York: Routledge.Pastor, Manuel, Jr. 1993. "Latinos and the LosAngeles Uprising: The Economic Context."Claremont, CA: Tomas Rivera Center, Occi-dental College.Pitcher, Brian L., Robert L. Hamblin, and JerryL. L. Miller. 1978. "The Diffusion of Collec-tive Violence." American Sociological Review43:23-35.Porter, Bruce and Marvin Dunn. 1984. The Mi-ami Riot of 1980: Crossing the Bounds. Lex-ington, MA: Lexington Books.Portes, Alejandro and Alex Stepick. 1993. City onthe Edge: The Transformation of Miami. Ber-keley, CA: University of California Press.Quillian, Lincoln. 1995. "Prejudiceas a Responseto Perceived GroupThreat:Population Compo-sition and Anti-Immigrantand Racial Prejudicein Europe." American Sociological Review 60:586-611.Simon, Julian. 1989. The Economic Conse-quences of Immigration. Cambridge, MA: Ba-sil Blackwell.

Spilerman, Seymour. 1970. "The Causes of Ra-cial Disturbances: A Comparison of Alterna-tive Explanations." American Sociological Re-view 35:627-49.. 1971. "The Causes of Racial Distur-bances: Tests of an Explanation." AmericanSociological Review 36:427-42.. 1976. "Structural Characteristics of Cit-ies and the Severity of Racial Disorders."American Sociological Review 41:771-93.Tilly, Charles. 1969. "Collective Violence in Eu-ropean Perspective." Pp. 4-42 in Violence inAmerica, edited by H. Graham and T. Gurr.New York: Signet Books..1978. From Mobilization to Revolution.Reading, MA: Addison Wesley.U.S. Bureau of the Census. 1990. U.S. Census ofPopulation and Housing: STF3A, Los AngelesCounty [CD-ROM]. Washington, DC: U.S. Bu-reau of the Census.



17/17

54 AMERICAN SOCIOLOGICAL REVIEW. 1980. U.S. Census of Population andHousing: STF3A, Los Angeles County[MRDF]. Washington, DC: U.S. Bureau of theCensus [producer]. Ann Arbor, MI: Interna-tional Consortium for Political and Social Re-

search [distributor].Vanneman, Reeve D. and Thomas Pettigrew.1972. "Race and Relative Deprivation in theUrban United States." Race 13:461-86.Waldinger, Roger. 1996. Still the Promised City?African Americans and New Immigrants in

Postindustrial New York. Cambridge, MA:Harvard University Press.Webster, William H. 1992. "The City in Crisis:A Report by the Special Advisor to the Boardof Police Commissioners on the Civil Disorderin Los Angeles." Los Angeles, CA: SpecialAdvisor Study [601 South Figueroa Street,Suite 3425, Los Angeles, CA 90017].Wilson, William Julius. 1978. TheDeclining Sig-nificance of Race. Chicago, IL: University ofChicago Press.

03 LA Riots Uprising

Documents

Transcript of 03 LA Riots Uprising