Giovanni Fattore  is associate professor of health care and public management at Bocconi University (Policy Analysis and Public Management Department), Italy. Previously, he was research offi cer at the London School of Economics. His research focuses on health policy and public
management reforms. He holds a degree in economics from Bocconi University and did
p ostgraduate studies at the Harvard School of Public Health and the London School of Economics.
E-mail: giovanni.fattore@unibocconi.it
Hans F . W. Dubois  is a research offi cer at the European Foundation for the  I mprovement of Living and  W orking Condi-tions (Eurofound) in Dublin. Previously, he was assistant professor at Kozminski University in Warsaw and worked with the European Observatory for Health Systems and Policies in Madrid. He holds a doctorate in business administration and manage-ment from Bocconi University and studied economics and medical biology at the University of Amsterdam.
E-mail: Hans.Dubois@eurofound.europa.eu
Antonio Lapenta  is an economist and public management specialist. He holds a doctorate in health care management and economics from Magna Græcia University of Catanzaro, Italy, and has conducted training and research for the Public Management
D epartment at Bocconi University from 2004 to 2008. He served as strategic
a dvisor to the Bolivian Health Minister for the creation of a National Health System. He is currently an international consultant and advisor for government institutions.E-mail: Antonio.lapenta@yahoo
218 Public Administration Review • March | April 2012
Public Administration Review , Vol. 72, Iss. 2, pp. 218–227. © 2012 by The American Society for Public Administration. DOI: 10.111/j.1540-6210.
Giovanni Fattore
Bocconi University, Italy Hans F . W. Dubois
European Foundation for the Improvement of Living and Working Conditions, Ireland
Antonio Lapenta
Magna Græcia University of Catanzaro, Italy
New Public Management (NPM) recently has been
compared and contrasted with public governance (PG) to illustrate shifts in conceptions of public administrations and in reform agendas. Th  e authors develop measures to capture the relevance of NPM and PG in textual discourse and investigate the extent to which they have entered the political debate. Content analysis of electoral programs for the 2005 Italian regional elections reveals that even in this legalistic country, considerable attention was paid to both NPM and PG issues. An important explanatory variable in preference for NPM or PG is party ideology, highlighting often-ignored within-country dynamics. Furthermore, the authors show how a methodological approach adapted from mainly political science and business research can be exploited in the fi eld of public administration.
N
ew Public Management (NPM) and public governance (PG) are two core sets of ideas in the fi eld of public administration. In a
provocative article, tellingly titled “Th  e New Public Governance?” Stephen P . Osborne (2006) argues that NPM actually has been a transitory stage in the evolu-tion from traditional public administration to what he calls New Public Governance.
We investigate whether the theoretical interest in PG is expressed in the broader society, as refl ected by the political debate. We also explore whether reference to NPM and PG issues depends on candidates’ ideologi-cal roots. Th  e interest in NPM and PG tends to be generalized to the country level of analysis. Inclusion of a proxy for ideological background could reveal whether within-country diff erences and dynamics are of signifi cant magnitude. Furthermore, we make an attempt to develop specifi c measures for these two often referred to but rarely measured sets of issues.Th  e article is organized as follows: In the next  s ection, we discuss recent developments in the NPM/PG
debate and explore core ideas that fall under these two broad labels, setting the conceptual basis of the meas-ures presented and discussed in the following  s ections.
Next, we argue that content analysis of electoral
p rograms is an appropriate approach for our analysis. We discuss this in light of broader applications of content analysis to various types of texts in public administration research. Th  e 2005
regional elections in Italy are identifi ed as a relevant, critical case. We move on, carefully describing the procedure that we applied to measure reference to NPM and PG issues in these political documents. Descriptive statistics are presented and statistical analysis is applied to test for diff erences in prevalence. Th  e results are discussed and, fi nally, conclusions are drawn.The New Public Governance?
NPM and PG have received considerable academic attention over the last few decades, referred to as “the two grand narratives of public management reform” (Andresani and Ferlie 2006, 416). We present a brief literature review of both NPM and PG in turn and discuss how the second is gaining ground.
“New Public Management” is an umbrella term (e.g., Hood 1991). Scholars as well as professionals often have used the expression to refer to distinctive themes, styles, and patterns of public service manage-ment reforms of the last two decades (Barzelay 2001). While some see NPM and PG as two diff erent para-digms, we agree with those who see the two concepts as diff erent sets of ideas. NPM’s origins lie in a mar-riage of new institutional economics (public choice, transaction cost, and principle–agent theory) and scientifi c management–based business philosophies (Hood 1991). NPM’s principal focus is intraorganiza-tional processes and management, and it emphasizes the eco
nomy and effi  ciency of those service units in producing public services (Osborne 2006). Th  e litera-ture off ers a large number of investigations into the components of NPM (Ferlie et al. 1996; Hood 1991; Osborne and Gaebler 1992). Following Andresani and Ferlie (2006), NPM ideas stress a combination of empowered and entrepreneurial management rather than traditionally autonomous public sector professionals and administrators. It favors the use of
Measuring New Public Management and Governance in Political Debate
Measuring New Public Management and Governance in Political Debate 219
participation of representatives from all stakeholders (Bovaird and Loeffl  er 2002). Networks can include public and private (for profi t and nonprofi t) organizations and citizens. Existing quasi-legislative and quasi-judicial new governance processes provide ways to engage individual citizens, the public, and organized stakeholders in the work of government  (Bingham, Nabatchi, and O’Leary 2005). As opposed to NPM, the PG  a pproach overcomes the traditional dichotomy between policy and administration (Hansen and Ejersbo 2002). More generally, classical management tools require an evolu-tion and upgrade to be applied to networked settings, as public interest should be pursued by managing institutional networks in which governmental entities do not always have a pivotal role (
Bourgon 2007; Metcalfe and Richards 1990). Governance struc-tures can be seen as dynamic relationships that are infl uenced by the decisions and actions of public managers. Th  ese relationship struc-tures—through which public policy is made and implemented—are multiple and reciprocal: multiple in the sense that many diff erent interactions infl uence governance structures, and reciprocal in that public managers not only are subject to governance structures but also play important roles in creating, executing, and changing the dimensions of governance (Feldman and Khademian 2002).Th  e interest in NPM-inspired management tools has not evaporated and is likely to stay (Ferlie and Fitzgerald 2002), but it is being chal-lenged more than before. After two decades of public sector reforms inspired by NPM principles, the attention of scholars moved toward PG approaches. Th  e study of governance is growing in popularity (Van Kersbergen and Van Waarden 2004). Were this academic inter-est to refl ect a development in society true, we should be able to encounter references to PG issues in the political debate.
Now that the possibility of a new PG movement has been raised, it is important not to overstate its diff usion, by present-ing it as a global development, before investigating it outside the
A nglo-American/Commonwealth world. In England, moderniza-tion (Newman 2002) and/or governance (Newman 2001; Rhodes 2000) rhetoric indeed was present in key New Labour policy an
d “ideological” documents (Ferlie and Fitzgerald 2002). While there have been voices suggesting that PG issues might become more popular than NPM issues in continental Europe (Andresani and Ferlie 2006) and maybe elsewhere in the world, research takes an Anglo-American focus. To give some counterweight, we choose a strongly legalistic context. While observations are context specifi c, as always, if we are able to measure and assess the issues at stake in an “unfavorable environment,” they are likely to be observed elsewhere. In particular, our context is the 2005 regional elections in Italy, a country that widely has been recognized for its legislative, “Napoleonic” administra-tive tradition  (Capano 2003; Mussari 1997;  O ngaro 2006). In other words, we apply a critical case  s trategy (Patton 2002) by focus-ing on a context in which NPM and PG are
relatively unlikely to prevail.Furthermore, the NPM versus PG discus-sion tends to generalize at the country level. Within-country variance often is neglected. Because of their nature, NPM issues are expected to be relatively popular among quasi-market forces rather than planning. Furthermore, it stresses performance measurement, monitoring and management, and audit systems (Andresani and Ferlie 2006).
NPM reforms are typically top-down in nature, driven by a “reform-ist” central government trying to get the most out of large opera-tional agencies and functions. NPM aims to enhance government per
formance by promoting the “three Es” (economy, effi  ciency, and eff ectiveness) and by introducing management principles that often have been transferred from business. Th  e common prescription for governments is to downsize and outsource. In the popular notion of one of Osborne and Gaebler’s (1992) principles of entrepreneurial government, governments should seek to improve performance by “separating steering from rowing.” Governments should concentrate on steering (policy management) while reducing their involve-ment in rowing (operational management). Steering organizations set policy, deliver funds to operational bodies (public and private), and evaluate performance, but they seldom play an operational role themselves. NPM focuses on single government entities, and it promotes competition between government agencies and contracted providers. In NPM, the traditional policy–administration dichot-omy is restored, with a clear line between management and policy. Resources and responsibilities among individuals and organizations should be allocated accordingly (Metcalfe and Richards 1990).Similar to NPM, “public governance” also is an umbrella term, and there are many forms of governance (Kooiman 1993). Th  e theoreti-cal underpinnings of “new governance” ideals are constructed both from supposed virtues of markets and third sector alternatives to the state, as well as from the belief in the virtue of competition, choice, and multiagent collaboration (Considine and Lewis 2003). Th  e public administration/management/policy literature has highlighted in particular the network governance mo
del (Acevedo and Common 2006; Newman 2001; Pierre and Peters 2000; Rhodes 1997, 2000). Here, clients, suppliers, and producers are linked together as copro-ducers (Considine and Lewis 2003). When we use the term “public governance,” we refer to this form of governance. PG developed from diff erent academic disciplines, in particular organizational sociology and network theory (Osborne 2006). A common feature of the PG approach is that the focus of analysis shifts from single entities to include interorganizational networks. Second, emphasis is put on collaboration rather than competition.
PG implies an interaction between government and society—
between public, private, and voluntary actors—and attention
to a wider range of outcomes (Acevedo and Common 2006).
Th  e diff usion of administrative action, the
multiplication of administrative partners,
and the proliferation of political infl uence
outside government’s circles are at the heart
of PG. It requires improved skills in negotia-tion and coordination. While understanding the use of hierarchy and authority, public
administrators also must manage complex
networks, relying more on interpersonal and
interorganizational processes (Kettl 2002).
Th  e budget formulation, for example, was
seen as a top-down exercise in NPM, while with the PG  a pproach, it requires the active
Because of their nature, NPM issues are expected to be relatively popular among
stakeholders that favor smaller government. For PG issues,
the reverse might well apply, as it is more concerned with  e nhancing outcomes than with streamlining budgets.
220 Public Administration Review • March | April 2012
making a profi t. Public organizations lack this element and, from an
institutional perspective, have relatively more to gain from textual output. Th  us, it is not surprising that the public sector produces a relatively large textual output. Th  e nature of public sector accounta-bility provides us with another explanation for this abundance. Com-munication with citizens is expected to be more common than in the private sector, where the role of citizens as stakeholders is likely to be less pronounced. Content analysis uses such verbal and written com-munications as its principal data input. Logically, one would expect public sector researchers to gratefully exploit these (generally publicly available) textual resources and widely apply content analysis.We performed an explorative literature review to get a general  i mpression of the use of content analysis in public sector research. Th  e search strategy that we applied was equal to Duriau, Reger, and Pfarrer’s (2007), in that we used the same keywords: “content analysis” and “text analysis.” We searched 17 public management/administration/policy journals: Climate Policy, Governance, Interna-tional Journal of Public Administration, International Journal of Public Sector Management, International Public Management Journal, Journal of Policy Analysis and Management, Journal of Public Administration Research and Th  eory, Philosophy and Public Aff airs, Public Administra-tion and Development, Public Administration, Public Administration Review, P
ublic Policy Research, Public Management Review, Public Money and Management, Public Policy and Administration, Policy Studies Journal, and  Publius. Th  e time span of our search is from 1980 (or since the journal has been included in databases) to April 2009. We excluded two studies that did not perform a content analysis themselves but used the results of other researchers’ content analysis as secondary data (Bertelli 2006; McHenry 1986). Other studies applied content analysis in the form of systematically coding and analyzing transcribed interview data or questionnaire responses (e.g., Bryer 2009; Wright and Taylor 2005). Finally, in some cases. “content analysis” referred to a through literature review of academic articles or textbooks (e.g., Dunn 1988; Walsh et al. 1996). We excluded these studies, as our focus is on textual analysis of public sector documents.
T o our surprise, public sector studies using content analysis methods are scarce. Our literature search, which was similar in scope to that conducted by Duriau, Reger, and Pfarrer (2007) on private sector research, identifi ed only 22 relevant studies. Th  is amounts to about one-fourth (22 of 98) of the number of studies identifi ed by Duriau, Reger, and Pfarrer. While we arguably were somewhat stricter in terms of excluding literature reviews, the 30 studies that Duriau, Reger, and Pfarrer identify for one source category alone (annual reports) easily outnumber the total number we encountered. Th  is is in sharp contrast to the massive body of research that we were expecting to encounter, in a
fi eld so widely associated with textual output. In itself, this is a highly interesting observation that needs further research.Table  1 gives a brief overview of the public sector content analysis studies that we identifi ed. Data sources for these studies include texts as diverse as offi  cial public administration documents, collec-tive bargaining agreements, speeches, and submitted parliamentary questions. Th  e application of the method has been global, and most studies take a qualitative approach. Th  e relatively large number of studies published over the past fi ve years suggests that the method is gaining in popularity in public sector research.
s
takeholders that favor smaller government. For PG issues, the reverse might well apply, as it is more concerned with  e nhancing outcomes than with streamlining budgets. In addition, we can test whether there are diff erences between challenging and incumbent parties. It is expected that, all else being equal, incumbents have a better understanding of their administrations and NPM issues. Voters may easier relate public administration performance to incumbents than to challengers. Incumbents thus are more likely to be attacked on the current state of the public sector. Th  is would imply higher rates of both NPM and PG terms for the challenger, promising reform.
In conclusion, our research questions are as follows:
1. Is the theoretical interest in the public management/admin-istration literature on PG expressed in the broader society’s interests, as refl ected by political programs?
2. Is the theoretical interest in NPM expressed in the broader
society, even in a legalistic country such as Italy, as refl ected by political programs?
3. Does reference in political programs to NPM and/or PG
issues diff er depending on candidates’ ideological roots?4. Does reference in political programs to NPM and/or PG
issues diff er depending on candidates’ status, whether incumbent or challenger?Content Analysis of Electoral Programs
Content analysis is a research method that uses a set of categoriza-tion procedures to make valid and replicable inferences from textual data. Nevertheless, content analysis is more than a technique in that it embodies a theoretical perspective that seeks to assign a major role to communication. Content analysis was applied on a day-to-day basis during the First and Second World Wars to radio mes-sages and other textual material from the enemy. It was not until the late 1940s that it was appli
ed in science (Janowitz 1969). We were  u nable to fi nd any review of the use of content analysis in the fi eld of research on the administration of public entities. Conse-quently, we performed an explorative literature review ourselves and contrasted it with the fi ndings of Duriau, Reger, and Pfarrer’s (2007) thorough review of research on the administration of private e ntities.
Duriau, Reger, and Pfarrer (2007) present a meta-analysis of 98 publi-cations in the fi eld of business administration and management using content analysis methods. Th  eir literature search involves 15 promi-nent journals in the fi eld over the years 1980–2005. Th  ey search within these journals using the keywords “content analysis” and “text analysis.” Subsequently, the authors make a thorough comparison of these studies, with regard to issues such as the studies’ data source (mostly annual reports, 30.6 percent, and trade magazines, 18.4 percent), broad methodological approach (77.6 percent quantitative, 15.3 percent qualitative, 7.1 percent mixed methods), data analysis (84.7 percent report some form of frequency count), and reliability (62.2 percent use multiple raters and report interrater reliability).Before we focus on the particular data source that we used in this article (electoral programs), we review the use of the method more broadly in the fi eld. For private, competing organizations, an  i mportant factor in legitimizing their activities and their existence is
Table 1Content Analysis in 17 Public Sector Research Journals, 1980–2009
Author Year Publication Title Textual Source Country Context
Beierle and Konisky2000“Values, Confl ict, and Trust in Participatory
E nvironmental Planning”Case studies, project reports, etc., from cases
of environmental planning
U.S.
Douglas1987“Collective Bargaining and Public Sector Supervisors:
A Trend Toward Exclusion?”
Collective bargaining agreements U.S.
Jreisat 1988“Administrative Reform in Developing Countries: A
Comparative Perspective”Offi cial country statements submitted to a
conference
Iraq, Jordan, Morocco,
Saudi Arabia, Sudan,
Syria, North Yemen
Carnevale 1993“Federal Service 2000: Staff Training and Labor–
Management Cooperation”
Federal collective bargaining agreements U.S.
Blake et al. 1998“The Nature and Scope of State Government Ethics
Codes”
39 state governments’ codes of ethics U.S.
Ward and Spenneman2000“Meeting Local Needs? Case Study of a
C ommunication Project in the Pacifi c Islands”Final project document, two promotional
project documents
Fiji, Samoa, Vanuatu
Houston and Richardson2000“The Politics of Air Bag Safety: A Competition among
Problem Defi nitions”Offi cial record of one U.S. House and two
U.S. Senate hearings
U.S.
Hall 2002“Live Bureaucrats and Dead Public Servants: How
People in Government Are Discussed on the Floor
of the House”Floor speeches from the 103rd and 104th Congresses
U.S.
Rittberger and Richardson2003“Old Wine in New Bottles? The Commission and the
Use of Environmental Policy Instruments”
Environmental action programs European Union
Svensson and Wood2004“Codes of Ethics Best Practice in the Swedish Public
Sector: A PUBSEC-Scale”
27 public sector units’ codes of ethics Sweden
Cheung 2005“What’s in a Pamphlet? Shortfalls and Paradoxical
Flaws in Hong Kong’s Performance Pledges”
Performance pledges of public organizations Hong Kong
Higgins2005“Contemporary Public Library Provision in England:
A Content Analysis of the Highest and Lowest
S coring Inspection Reports”
Public library inspection reports U.K.
Erakovic and Wilson 2006“The Interaction of Market and Technology in
Radical Transformation: The Case of Telecom New
Zealand”
Annual reports New Zealand
Needham 2006“Policing with a Smile: Narratives Of Consumerism in
New Labour’s Criminal Justice Policy”Prime minister’s speeches, central
g overnment department papers (white and green papers and strategy documents),
local government corporate plans
U.K.
Schuh and Miller2006“Maybe Wilson Was Right: Espoused Values and Their
Relationship to Enacted Values”Agency mission statements, Speeches of Presidents George H. W. Bush, Bill Clinton,
and George W. Bush, plus a subject matter analysis of executive orders, and a survey
of senior executivesgovernance
U.S.
Penner, Blidook, and Soroka 2006“Legislative Priorities and Public Opinion:
R epresentation of Partisan Agendas in the
C anadian House of Commons”
Oral parliamentary questions Canada
Jewell and Bero2006“Public Participation And Claimsmaking:  E vidence
U tilization and Divergent Policy Frames in
C alifornia’s Ergonomics Rulemaking”Notice and comment submissions
to  C alifornia’s 1997 passage of an
e rgonomics standard
U.S.
Lim and Tang 2007“Urban E-Government Initiatives and Environmental
Decision Performance in Korea”
Government Web sites South Korea
Drumaux and Goethals 2007“Strategic Management: A Tool for Public
M anagement? An Overview of the Belgian
F ederal Experience”
Government strategic plans  Belgium
Marcuccio and Steccolini 2009“Patterns of Voluntary Extended Performance
R eporting in Italian Local Authorities”
Government extended performance reports Italy
Kapucu, Augustin, and Garayev 2009“Interstate Partnerships in Emergency Management:
Emergency Management Assistance Compact in
Response to Catastrophic Disasters”
News reports, government documents,
a fter-action reports from various
i nstitutions
U.S.
Kloot2009“Performance Measurement and Accountability
in an Australian Fire Service”Annual reports of 2006, for three fi refi ghting
services
Australia
As far as the analysis is concerned, compared to private sector
r esearch, our literature search revealed that few studies take a quanti-tative approach to content analysis. Exceptions in public administra-tion research include Marcuccio and Steccolini (2009) and Erakovic and Wilson (2006), who analyze frequencies; Beierle and Konisky (2000), who report correlations; and Hall (2002), who applies (logistic) regression. While quantitative approaches naturally have their limitations, they can provide important complementary analy-sis, as indicated by broad application beyond public sector  r esearch. Another trend that can be observed is that the textual sources refl ect the topics of interest at the time, from relatively greater use of bargaining minutes in the 1980s to the early 1990s, to ethical codes in the later 1990s and 2000s, and performance reports in the 2000s. Th e use of Web site data is likely to increase with the massive expan-sion of governmental Web sites.
Most important for our study, we were unable to fi nd any research in the fi eld that applies content analysis to electoral programs. One of the articles reviewed (Bertelli 2006) does make use of  s econ
dary data drawn from content analysis of electoral programs. It  m easures
Measuring New Public Management and Governance in Political Debate 221
222 Public Administration Review • March | April 2012
m
anner, and adopts the principles and laws establishing autonomy and decentralization. Implementation was delayed, and it was only in the mid-1970s that 15 regions were established and started to act. Regional parliaments have been elected by popular vote ever since. Over the past 30 years, regional governments have grown in size and scope in highly relevant policy domains such as health care, social services, housing, industrial development, professional education, and, more recently, policing and education.
Since 2000, regions have been governed by a directly elected governor who chairs the regional cabinet, which is approved by the regional parliament. In April 2005, citizens of 14 Italian regions
renewed regional councils and elected for the second time their governors. In these regions, accounting for about 80 percent of the Italian population, all but two outgoing governors ran for a sec
ond mandate and were challenged by only one principal candidate. Th  ese 28 incumbents and challengers constitute our sample.Methods Th  e fi rst step of the analysis is to document whether NPM and PG issues are present at all in the programs of each candidate. We collected the candidates’ political programs from their Web sites, checking their offi  cial Web site twice: fi ve and three weeks before election day. At the fi rst check, four incumbents’ programs were already available on the Internet compared to nine challengers’
programs. T wo weeks later, all candidates had their political programs available on their own Web sites, with the exception of two candidates who were running in diff erent regions. Most of the candidates presented a program that was intended as a formal docu-ment stating their aims, principles, and plans. Otherwise, in some cases, there were candi-dates who  u ploaded only a one-page docu-ment containing highlights of their political plans on their own Web site. In these cases, we considered these one-page documents  r elevant for the analysis. We converted the 26 electoral programs into Microsoft Word documents.Our next step was to read the programs and extract all sentences
that referred to any issue related to the management of government organizations or to public (government) governance. A “sentence” is defi ned as a unit of language having a subject, an object,
and a verb. Th  e end of each sentence is represented by a full stop. Th  e extracted sentences were collected in a grid including four columns: (1) NPM, (2) PG, (3) uncertain, and (4) NPM and PG. Next, we analyzed this classifi cation in order to reassign all of the sentences to NPM or PG. As a general rule, we classifi ed in the NPM category sentences that concerned primarily the internal perspective of the organization, while we privileged the other category when the focus was mainly on the relations between public organizations and its stakeholders. Some of the sentences were eliminated, as they were considered not to be substantially related to any of the two issues. Some others were put in one of the two columns if one of the two issues was promi-nent relative to the other. Sentences were attributed to NPM when they referred to internal operations and to effi  ciency criteria. Instead, they were referred to PG when they primarily concerned with inter-organizational relations and with decision-making processes.political party preferences toward quangos, drawing on a book by Budge et al. (2001) that maps policy preferences. Ever since Lasswell, Leites, and Fadner (1949) pioneered the content analysis of electoral programs, this has become popular in political science research and a common approach in this discipline. Political science even knows a “Manifesto Research Group” that created a coding book, with standard categories used to code party election  p rograms.
While electoral programs might not be the most obvious textual
input for public administration research, it is relevant for the fi eld
insofar as electoral programs refer to public administration issues. A similar argument can be made for several of the other textual sources that we found in the public administration literature (table  1), such
as prime minister’s speeches. We think that electoral programs can
be content-analyzed in a way to generate valuable data for pub-lic administration research as well. On the one hand, what-to-do
arguments in public management rely on rhetorical power (Hood
1998). On the other hand, political programs in particular provide
us with a view on whether the academic debate on certain public
administration issues refl ects interest in society. Electoral programs
are widely available and can be expected to refl ect a politician’s
position regarding issues better than a single speech. Th  ey also tend
to contain less anti-opponent rhetoric and more policy intentions
than verbal communications in the heat of upcoming elections.
Most importantly, political programs—at least to a certain extent— refl ect what citizens fi nd important at a certain
m oment in time. Politicians are unlikely to refer to such issues in their programs if voters are in fact totally ignorant of them. Candidates
choose the rhetorical campaign strategy that
they believe will give them the best chance of
winning the election, although there inevitably
will be a great deal of uncertainty associated
with this choice (Lau and Pomper 2002). So, when politicians refer to NPM or PG issues in their politi
cal programs, this refl ects the
expected values of their political target groups.
Furthermore, at least to some degree, electoral programs may well
communicate visions about the future of government actions that
may mobilize resources inside and outside public administration. As
far as politicians are true to their intentions (Pierson 2000), electoral programs thus provide us with a look into the potential future.
Context
Another advantage of our context is that Italian regional elections
held in 2005 contrasted two candidates who were part of either a
“center-left” or a “center-right” coalition. “Left” and “right” mean
very diff erent things in diff erent countries or even in diff erent
places within one country. In our Italian context, the label “left”
is  a ttached to candidates who are part of a coalition that gener-ally  f avors relatively large government, promotes, social services
and tends to be supported by labor unions. “Right,” in our Italian
context, refers to a coalition that is more pro-market and tends to
be supported by relatively conservative elements of the Catholic
Church.
Italian regions were granted a degree of autonomy in the 1948
c onstitution, which recognizes an
d promotes local autonomy, ensures that services ar
e provided in a maximally decentralized
[A]t least to some degree,
electoral programs may well
communicate visions about the future of government  a ctions that may mobilize resources inside and outside public
administration.
Measuring New Public Management and Governance in Political Debate 223
agreement. In a subsequent phase, sentences that were rated diff er-ently by the  a uthors were reviewed jointly once again, one by one. Slightly diff erent operationalizations of the categories or simple typing errors during the classifi cation process were the most com-mon causes of fi rst-stage disagreement. After the second stage, there still remained about 2 percent disagreement, which indicates some inherent ambiguity.
We then counted the number of words attributed to the two catego-ries reported in each document and calculated their relative frequen-cies (number of designated words over total number of words). Th  e computing procedure was performed electronically and double-checked. After aggregating these data, we compared the frequency of words between coalitions and between challengers and in
cumbents. Overall, the most-used roots refer to bureaucracy, participation, and network. In qualitative, contextual terms, the fi rst term generally is used to underscore the intention to move away from bureaucratic procedures and behaviors. Th  e other two mainly refer to the external relationships of public organizations.
To verify the reliability of our measure, all extracted sentences also were evaluated according to a diff erent approach. Th  e research team was asked to rank these sentences using a scale ranging from 0 to 10. Zero was assigned to extracted sentences focusing  e xclusively on NPM themes and 10 for sentences exclusively focusing on PG  i ssues. Intermediate values were expected to refl ect a mix of contents between the two themes. Mean scores were used to test diff erences between coalitions and between incumbents and challengers.Results
Candidate characteristics are presented in the fi rst four columns in table 3. Th  e table also presents the fi rst results of the analysis:
In order to help researchers classify sentences, we prepared a list of
words that have direct relevance for NPM and PG issues. We used the roots of these words as the element to be identifi ed in the docu-ments, as the root is the essential core of a word, from which all
potential words based on that root can be derived. Table 2 shows the relevant roots of the words that refer to public administration. Each root was used to identify the phrases that were potential candidates for the grid (NPM and PG categories), but the fi nal decision about the attribution to each category was made in accordance with the meaning they have in the programs.
In order to control and limit the errors inherent in the  s ubjective, judgmental process of classifying sentences, various measures were taken. Th  e task was fi rst performed by each author working  i ndependently. Discrepancy between the classifi cations performed by the authors was 10 percent. Interrater reliability, measured by Cohen’s Kappa, amounts to 0.798. Th  is suggests substantial
Note: Root in Italian, meaning in English.
Table 3 New Public Management and Public Governance in Electoral Programs Candidate Coalition*Incumbent/ Challenger**
Winner***
Total Document Length
(words)Words Contained in NPM Sentences % NPM Words Contained in PG Sentences % PG
% NPM + PG Abramo C-R I 031,186 1,456  4.67
399  1.28  5.95Antichi C-R C 011,188 739  6.61306  2.749.34Bocchino C-R C 01,035 14013.5316  1.5515.07Fitto
C-R I 017,385 369  2.12298  1.71  3.84Formigoni C-R I 115,545 994  6.39369  2.378.77alan C-R I 14,900 106  2.16298  6.088.24higo C-R I 013,449 1160.86317  2.36  3.22Laffranco C-R C 018,090 707  3.91655  3.627.53Latronico C-R C 07,963 390.49207  2.60  3.09Massi C-R C 0155 10  6.4500.00  6.45Monaco C-R C 02,003 103  5.1476  3.798.94Pace C-R I 02,951 119  4.03140  4.748.78Storace C-R I 04,471 86919.44103  2.3021.74Bassolino C-L I 1572 6511.3630  5.2416.61Bresso C-L C 115,973 720  4.511,4969.3713.87Burlando C-L C 13,565 00.00182  5.11  5.1
1Carraro C-L C 012,181 211  1.73851  6.998.72De Filippo C-L I 15,183 63812.3166712.8725.18Errani C-L I 115,562 1410.911,077  6.927.83Loiero C-L C 130,335 1,211  3.992,9129.6013.59Lorenzetti C-L I 115,670 400  2.551,3528.6311.18Marrazzo C-L C 148,758 553  1.132,028  4.16  5.29Martini C-L I 110,317 257  2.49715  6.939.42Sarfatti C-L C 021,912 552  2.522110.96  3.48Spacca C-L I 11,118 48  4.2911910.6414.94Vendola
C-L
C
1
16,662 676
4.06
574
3.44
7.50
*C-L = center-left, C-R = center-right. ** I = incumbent, C = challenger. *** 1 = yes, 0 = no.

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。