Coproduction Of Government Services And The New Information Technology .

the “centralized local government public information centers thattake non-emergency service requests from citizens [enabling] 311systems to route requests to the appropriate department and followthrough on the fulfillment of service requests” (Ganapati 2011,430). The CRM logs all citizen-generated requests for services (suchas fixing a pothole, cleaning graffiti, etc.), as well as the work ordersgenerated to fix the problems. From this rich data source, we create maps of all service requests in the City of Boston and combinethat information with data about the neighborhoods where thoseproblems were reported. These maps and more detailed statisticalmodels allow us to develop an understanding of how a 311 systemmight affect citizens’ ability to participate in the coproduction ofpublic services.Yang and Sanjay Pandey, effective citizen participation is achievedwhen “government decisions and government–citizen relationshipscan be substantively improved” (2011, 889).The field of public administration has often encouraged citizenparticipation and has recognized a variety of different citizen rolesthrough which it might take place, including voter, customer, andactive citizen (Schachter and Yang 2012). Roberts (2008) outlinesseveral models of direct citizen participation, including coproduction,in which citizens adopt the role of volunteer or “coproducer.” Inthe coproduction model, administrators are responsible for facilitating the joint provision of service design and delivery with citizens(Roberts 2008, 21). Terry Cooper and Pradeep Chandra Kathiobserve that “[c]oproduction is joint provision of a public service bythe public agency as well as by the service consumers” (2005, 47),a view shared by Stephen Percy (1984) and by Sharp (2012), whoidentifies coproduction as “an alternate conception of citizen participation” (Sharp 1980, 109). In Citizen, Customer, Partner: Engagingthe Public in Public Management, John Clayton Thomas (2012)explains that coproduction has increased as government’s work hasshifted from products to services.The findings for Boston suggest that we have little cause for concernthat 311 systems may benefit one group over another based on race:the racial composition of a neighborhood provides little predictive power in assessing the number of service requests. However,the results give some indication that Hispanics may be left behindas 311 systems expand beyond the basic call center to the Internetand smartphone applications. The results demonstrate significantnegative relationships between the share of the Hispanic populaScholars have observed profound implications of the coproductiontion in a neighborhood and the likelihood of using the city’s servicemodel. Charles Levine argues that coproduction can strengthen therequest web portal. With respect to income, the findings show that“bridge” between citizens and government by revitalizing citizenthe poorer a neighborhood, the less likely it is to take advantageship (2008, 89–90), and Elinor Ostrom employs a similar metaof the 311 system. A notable exception is smartphone utilization.phor to describe coproduction’s potential for “breaching the greatThe greater the percentage of the two poorest income groups in adivide” between government and its citizens (1996, 1073). Basedneighborhood, the more service requests were made using Boston’son Ostrom’s observations and conceptualsmartphone application. This result sugframework, Taco Brandsen and Victor Pestoffgests that smartphones may have potential toThis result suggests that smartdescribe coproduction as “one way throughbridge the digital divide. We find additionalphones may have potential towhich synergy could occur between what aevidence that Boston’s smartphone applicationbridge the digital divide.government does and what citizens do” (2006,may help attract a large and rather transient496). Tony Bovaird asserts that “the copropopulation into the coproduction of governduction approach assumes that service users and their communitiesment services: young college-age individuals.can—and often should—be part of service planning and delivery”—a “revolutionary concept in public service” (2007, 846).This article is organized into five sections. The first section presentsthe grounding of coproduction in the literature of citizen particiIn addition to changing the relationship between citizens and theirpation and the emerging importance of the new communicationsgovernments, researchers find that coproduction can increase thetechnology for the coproduction model. The next section furthereffectiveness and efficiency of service delivery (Parks et al. 1981,explores the relevance of coproduction for 311 call center sys1001). One way in which improvements in service production cantems. The third section describes the data and methodology of thisresult is from citizens making requests for assistance or providinginquiry and is followed by empirical evaluation of the hypotheses.information to service agents, as noted by Gordon Whitaker (1980)The article concludes with a discussion of the findings and theirand others (Bovaird 2007; Brudney and England 1983; Thomasimplications for coproduction and the involvement of citizens in2012). Citizen service requests provide necessary information to311 systems.local governments concerning service needs and shortfalls (Whitaker1980, 243–44). Sharp observes that citizens exhibit “participatoryCitizen Participation and Coproductionbehavior” whenever they contact public officials, either to requestIn The Age of Direct Citizen Participation, Nancy Roberts definesa service or to lodge a complaint (2012, 103–4). Ostrom notesdirect citizen participation as “the process by which members of asociety share power with public officials in making substantive deci- that “if citizens do not report suspicious events rapidly to a policesions related to the community” (2008, 5). Elaine Sharp argues that department, there is little that department can do to reduce crimein an area or solve the crimes that occur” (1996, 1079). Percy citescitizen participation includes “any form of involvement in commuseveral examples of how citizens might participate in police services,nity affairs that has the potential to shape the allocation of publicsuch as “reporting crimes to police, providing information on crimiresources or the resolution of community issues” (2012, 102). Innal matters and suspects, and testifying in court” (1984, 432–33).contrast to indirect citizen participation (i.e., representation), directcitizen participation embraces increased cooperation between public By providing information to government officials, citizen coproduction facilitates better decision making and the ability to respond toadministrators and actively involved private citizens (Roberts 2008)citizen demands and concerns more effectively (Percy 1984, 437).and civic engagement (Sharp 2012, 102). According to Kaifeng688Public Administration Review September October 2013

Crime is not the only service domain in which citizens might makerequests for assistance or share information with public administrators and local governments. Indeed, Whitaker’s model recognizesthe possibility of active citizen participation across service domains(Brudney and England 1983, 60). Coproduction is observed in awide variety of policy areas, such as education, the environment,health care, and sanitation (Alford 2009; Ben-Ari 1990; Kiserand Percy 1980; Ostrom 1996; Sharp 1980). Educational successrequires the active engagement of student coproducers in learningand parent involvement in homework (Levine 2008, 83; Ostrom1996, 1079). Environmental programs engage citizens in recyclingand community cleanup campaigns (Kiser and Percy 1980, 1).Doctors must “rely on patients to behave in certain ways” to ensuresuccessful treatment, such as taking prescribed medication andparticipating in physical therapy (Alford 2009, 1). Coproductionin local sanitation services encourages citizens to “literally go halfway” in curbside trash pickup efforts (Sharp 1980, 111). Sharpargues that these “less spectacular forms of participation” are oftenforgotten in discussions of citizen participation but affect thecapability of government to improve service conditions (1980,112–13).From the inception of the coproduction model in the early 1980s,scholars have recognized the influence that technology and “technical feasibility” would have in setting the limits and potential ofcitizen coproduction activity (Parks et al. 1981, 1002; cf. Kiser andPercy 1980, 4). Sharp observes the need for public administrators toremain open to new forms of citizen coproduction and appropriategovernment responses (1980, 113). More recently, Bovaird concludes, “Though we cannot predict the outcomes of these complexadaptive coproduction processes, they clearly extend the opportunity space of available solutions for social problems” (2007, 857).Coproduction and Its Application to 311 Call CentersCentralized, nonemergency government 311 call centers originatedin Baltimore in 1996 as a response to an overload of the 911 systemwith nonemergency calls (Borins et al. 2007). Baltimore saw immediate and substantial positive results through lower crime rates andcost savings totaling 100 million in the first three years (Borins etal. 2007). Given Baltimore’s success, the Bill Clinton administration promoted 311 to improve government performance at thelocal level. Boston mayor Thomas M. Menino viewed 311 systemsas a way to achieve the core value of local government—“helpingpeople”—which could be accomplished by paying “attention tobasic quality of life issues such as filling potholes, removing graffiti, and ensuring that the city streets are clean, safe, and well-lit”(City of Boston Performance Management System 2011, 1). Sincethe Baltimore experience, a growing number of cities have expanded311 systems to encompass the Internet through dedicated municipalportals, social media, and smartphone applications.Research has not yet examined how 311 systems may have improvedcitizen interaction with government, however. Little is known regarding how 311 may have transformed citizens’ access to government oraffected the ability of different demographic groups to request andreceive services. Prior research on 311 systems is extremely limited(Borins et al. 2007; Mazerolle et al. 2005; Schultz 2003). Very fewstudies have investigated the use of advanced smartphone technologyin government services (Fioretti 2010; Floreddu and Cabiddu 2012;Ganapati 2011; Traunmüller 2011),3 and none that we are aware ofhas investigated the distributional consequences of these technologies.The questions that we seek to answer in this article pertain to thechanging role of citizens in the delivery of government servicesin the twenty-first century as a result of the introduction of newinformation technology. To understand theconsequences of 311 for public services, weinformation technolturn to the concept of coproduction.The new information technology has extendedThe newthe applicability of the coproduction modelogy has extended the applicabilin government service delivery. John Alford(2002) contends that recent technologicalCoproduction arose from a fiscal environity of the coproduction modeladvances may lead to an increased ability toment similar to the present era, with the localin government service delivery.perform coproduction activities. In an essaygovernment fiscal crisis of the 1970s “thaton applying technology to enhance citizenaffected provision of civic services” (Cooperengagement with local governments, William Barnes and Brianand Kathi 2005, 47). Levine asserts that limited public resourcesWilliams (2012) urge public administrators to embrace the newencourage coproduction in the delivery of services, a theme echoedinformation technology in response to a changing society and citiby Larry Kiser and Stephen Percy (1980, 1). But, as Nathan Glazerzenry. Albert Meijer claims that the Internet can facilitate improveobserves, coproduction is “more than a financial panacea for fisments in citizen contacts through social networking and onlinecally strapped governments” (quoted in Levine 2008, 83). To thesupport groups in the “networked coproduction of public services”contrary, Jeffrey Brudney and Robert England argue that coproduc(2011, 598). His results suggest that new media and online nettion’s main contribution is the “appreciation of the role that citizensworks can boost coproduction and information exchange betweencan and do play” in service delivery (1983, 62). Levine sees coprocitizens and their government (606).duction as a vehicle for “continuous day-to-day involvement ofindividuals and neighborhoods in government” (2008, 83). RichardDespite the growing interest of researchers and practitioners in theRich argues that for the average citizen, these “small increases in theapplications and consequences of the new information technolquality of the various municipal services delivered in his or her areaogy for the coproduction of government services, Meijer finds thatcan combine to improve significantly the quality of community life”“Scholarship on coproduction of public service repeatedly ignores(1981, 63). Similarly, Pestoff views coproduction as “an importantthe role of the new media” (2011, 598). Accordingly, the remaindermeans of enhancing both the quality and quantity of public servof this article will investigate how new information technology,ices” (2006, 507).such as 311 call centers, the Internet, and smartphones, affects theability of citizens to participate in the coproduction of public service Bovaird (2007) suggests that this increase in quality and quantitydelivery.of services occurs at very low cost to government. For example, theCoproduction of Government Services and the New Information Technology: Investigating the Distributional Biases689

economic consequences for cities using phone-based versus Internetor smartphone-based systems are substantial. The cost of operatorbased 311 systems range from 1.15 to 5.49 per call (Pew 2010),while the cost per reported case for the Internet/smartphone areestimated to be 80 percent to 90 percent less (Lagan 2011; 311:Charlotte and Mecklenburg County 2005). Because cost considerations will likely drive cities to invest even more in the Internet andsmartphone-based technologies, it is important to examine theirconsequences for the coproduction of government services.One possible implication concerns distribution: the potential advantages of coproduction for citizens and governments notwithstanding, scholars have found that participants in coproduction comeprimarily from wealthy communities. Thus, the apprehension arisesthat coproduction can perpetuate and worsen the disproportionatecontrol over community resources possessed by the more affluent(Bovaird 2007). Additionally, many citizens do not have the abilityto perform services that require specialized training, so that theycannot take advantage of coproduction (Pestoff 2006). The cost ofservices is not eliminated by coproduction, but instead, some of thecosts are transferred to the citizen. The efficiency and effectiveness ofcoproduction depend on the ability of citizens to perform their rolesin service delivery (Pestoff 2006).Scholars have examined the participants in coproduction anddisparities in involvement. According to one study, those engagingin coproduction activity tend to be older, female, and inactive inthe labor market; with the exception of older people, these samegroups are also more willing to engage in higher levels of coproduction (Löffler et al. 2008). Another study asserts that those withmore education and with higher-paying jobs are more likely to beinvolved in coproduction activities, including the improvement ofpublic services, the crafting of policy, and service in communityor neighborhood councils (Norris and McLean 2011). Althoughstudies may vary on specifics, most scholars agree that disadvantaged populations such as racial minorities, those with less formaleducation, and those in lower socioeconomic circumstances tendto participate less in coproduction activities, which will diminishthe benefits they can derive from the model (Barker 2010; Holmes2011). This research underlies our first hypothesis:Hypothesis 1: Historically disadvantaged groups are lesslikely to participate in coproduction through 311 systems.Mitlin (2008) shows that coproduction can be used to obtainpolitical influence and access to resources by creating opportunitiesfor citizen involvement in areas typically reserved for government.Scholars have noted the difficulty of engaging marginalized groupsin coproduction and public services (Barker 2010; Holmes 2011).The incorporation of information technology into coproductioncan further limit the groups that have the ability and access toparticipate. Older adults and those with low income, those who areunemployed, or those who have less formal education tend to sufferfrom digital exclusion (Nash 2011). One study shows that whitesare more likely than blacks or Latinos to participate in e-government (Bryer 2010). Another scholar identifies level of education,age, income, race, ethnicity, and gender as potential factors in digitalexclusion (Schradie 2011).Research has shown that minorities tend to have less access totechnological resources, limiting their ability to participate fully indigital coproduction activities (Nash 2011; Schradie 2011). Becausethese groups are often in greatest need of government services (Nash2011), their exclusion from technologically advanced coproductionmay be especially problematic. People from diverse backgrounds,then, may need supplementary support or facilitation to participate in coproduction (Needham 2009). Some communities areattempting to solve the problem of digital exclusion in coproductionby making computers and Internet access more widely availableat public libraries and community centers (Nabatchi and Mergel2010). Other cities provide handheld computers or digital camerasto citizens so that they can report on community conditions (Svaraand Denhardt 2010). Some governments promote Internet accessby funding programs to implement Internet connectivity in thehome (Hodgkinson 2011).Increasingly, governments are adopting new channels of information technology, such as telephone, smartphone, and the Internet,to broaden services and involve citizens (Gagnon et al. 2010).Technologies such as social networking and smartphones haverecently been used to help victims of natural disasters and otheremergencies, for instance, Google People Finder during the 2011tsunami and earthquake in Japan (Weintraub 2011). Informationand communication technology has enabled and initiated newpartnerships and coproduction relationships between citizens andgovernment (Alford 2002; Löffler 2011).Despite the growth in the use of new information technology bygovernment, little, if any, research has investigated the effects of theDespite the benefits of integrating new information technologyshift to new media on the provision of servinto coproduction, some citizens prefer moreices. For example, many cities have iPhone andtraditional ways of receiving services fromDespite the growth in the useAndroid-based applications but no Blackberryand interacting with government (Gagnonof new information technologyapplications for 311 reporting. A recentet al. 2010). These citizens may not have theNielsen survey found that African Americansexperience, training, or resources needed toby government, little, if any,are more likely than any other racial group touse the new information technology. Thisresearch has investigated the“digital divide” describes the gap betweeneffects of the shift to new media own a Blackberry (Kellogg 2011), so this preference for a particular technology could skewthe “technology haves and have-nots” causedon the provision of services.the usage and penetration of phone-basedby the lack of computer access, Internettechnology in government coproduction ofaccess, and technological expertise (Hortonservices. Research has not determined whether neighborhoods with2004,17). Dyson (2011) finds evidence “that marginalized groupsare the ones most adversely affected by the digital divide. Those who larger minority populations are taking advantage of 311 systems. Aare already disadvantaged by income or education are more likely to pattern of domination of 311 by certain groups will advantage thembut disadvantage others.be excluded from digital citizenship.”690Public Administration Review September October 2013

Scholars have long debated the relationship between the development of information and communications technologies (ICTs)and citizen participation. Using the Internet as a key technologyof ICTs, a large body of work addresses two perspectives on theeffects of ICTs on citizen participation: optimism and pessimism(Bimber 2001; DiMaggio et al. 2001; Katz and Rice 2002; Krueger2002; Norris 2010). The key argument of the optimistic positionis based on the mobilization hypothesis, which postulates that thedevelopment of ICTs will not only reduce the costs of informationand communication for citizen participation but also get citizens—particularly those who are inactive in the current system—moreinvolved and engaged in public life (Chen and Dimitrova 2006;Delli Carpini 2000; Krueger 2002; Norris 2001; Ward, Gibson, andLusoli 2003; Weber, Loumakis, and Bergman 2003). By contrast,the reinforcement hypothesis underpins the pessimistic perspective, which argues that technological innovations merely reinforceparticipation of those who are already informed and motivatedthrough traditional channels and thus exacerbate existing socialinequities between the information haves and have-nots (Best andKrueger 2005; Delli Carpini 2000; Kavanaugh 2002; Norris 2001).Even though one recent study finds both mobilizing and reinforcingeffects of the Internet on political participation (Nam 2012), themajority of empirical studies seem skeptical about the mobilizingeffects of the Internet on citizen participation, particularly on civicengagement and political participation, and find either a weak effecton political participation or little evidence to support a significantrelationship between the Internet and civic engagement (Bimber2001; Chen and Lee 2008; Delli Carpini 2000; Krueger 2002,2006; Park and Perry 2008).Although the new information technology may be facilitative forsome, it may pose another barrier to communicating with government for others. In this research, we examine the consequences ofexpanded technological channels for coproduction for disadvantaged groups. Our second hypothesis proposes:Hypothesis 2: As coproduction extends its application to theInternet and smartphones, the historical disparities will exacerbate the problems of unequal participation in coproduction.MethodologyThis section presents the data and methods used in the study totest the hypotheses concerning coproduction of services. We modelcitizen service requests as a function of sociodemographic variables,including race, income, and education; land use (zoning); and ageospatial dependence variable. We begin with the data sources.Data SourcesOur study relies on data drawn from the City of Boston’s CRMdatabase. The CRM tracks work orders generated by governmentemployees (e.g., by city inspectors) and citizens. The databaseincludes all information logged through the 311 system: information and service calls made to the Mayor’s Hotline, service requestsmade through the city’s Internet portal, and requests made throughthe smartphone application. The city limited the data available toone year, from March 1, 2010, to February 28, 2011. All citizengenerated requests that could not be mapped to their geographicorigin were removed (these items are almost exclusively generalinformation requests not tied to a specific location).In order to make inferences about the neighborhoods generatingcitizen service requests, we had to collapse the request data by theirlocation into census block groups, which allowed us to associatewith them demographic data derived from the U.S. Census Bureau’s2005–2009 American Community Survey. Because demographicinformation is not available for each service request, the associationwith the census data is the only way to incorporate this informationand examine the spatial or group implications for service delivery.Collapsing the data changes the unit of analysis from the individualservice request to the census block group. Consequently, we haveobservations from 537 block groups for statis

and the New Information Technology: Investigating the Distributional Biases Th is article examines whether the 311 system for requesting government services results in use throughout a jurisdiction or facilitates the "haves" gain-ing greater access and limiting opportunities for historically disadvantaged groups.