xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 002021449
007 cr mnu|||uuuuu
008 090729s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002545
Trent, Carol L. S..
State dominance and political corruption :
b testing the efficacy of an alternate configuration of institutional-anomie theory cross-nationally
h [electronic resource] /
by Carol L. S. Trent.
[Tampa, Fla] :
University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 145 pages.
Thesis (M.A.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
ABSTRACT: Extant assessments of Messner and Rosenfeld's institutional-anomie theory (1994) have generally supported the thesis that, in social collectives where the economy dominates, non-economic institutions (i.e. the family, education, polity) are rendered feeble, unable to exert their normative controls. The cultural values of these societies place primacy on "making it" (monetary success), while at the same time placing a much weaker emphasis on the licit means of achieving these goals. The resultant state is one of anomie, conducive to crime. Messner and Rosenfeld have extended their argument stating that it is not economic dominance per se that contributes to high crime rates, but any tip in the institutional balance of power. The current study examines one of these configurations which hypothesizes that, in nation-states where the state dominates other institutions, the dominant cultural orientation is one of moral cynicism, conducive to corruption-prone behaviors. Using macro-level data, the current study assesses the efficacy of this alternate configuration of institutional-anomie theory as a predictor of corruption cross-nationally. Using a sample of 125 nations, state dominance is positively related to corruption. The effects of the state were both mediated and moderated by economic strength, measured as levels of industrialization.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Co-advisor: Michael J. Lynch, Ph.D.
Co-advisor: John K. Cochran, Ph.D.
t USF Electronic Theses and Dissertations.
State Dominance and Political Corruption: Testing the Efficacy of an Alternate Configuration of Institutional Anomie Theory Cross Nationally by Carol L. S. Trent A thesis submitted in partial fulfillment of the requirements for t he degree of Master of Arts Department of Criminology College of Arts and Sciences University of South Florida Co Major Professor: Michael J. Lynch Ph.D. Co Major Professor: John K. Cochran Ph.D. Shayne E. Jones, Ph.D. Date of Approval: June 20, 2008 Keywords: criminology, macro level theory, political crime, comparative, cross cultural Copyright 2008 Carol L. S. Tr ent
Dedication For my dad, who always believed in me. I wi sh that you could have seen thi s.
Ac knowledgements This thesis would not have been possible without the guidance of my co major professors Dr. Michael Lynch and Dr. John Cochran, and my committee member Dr. Shayne Jones. Thank you Dr. Lynch for sharing your wisdom and genius, and for, perhaps more importantly, reducing my panic level, helping me clear my head, and getting me to just finish the darn thing. Thank you Dr. Cochran for accepting nothi ng less than the best and not letting me get away Jones for reading the various drafts of this document with a critical eye ; your comments and assistance were invaluable. Finally, I would like to thank my husband Bobby for ed. Now we can leave
iv Table of Contents List of Tables ................................ ................................ ................................ ....................... vi List of Figures ................................ ................................ ................................ ..................... v i i Abstract ................................ ................................ ................................ ............................. viii Chapter One Introduction ................................ ................................ ................................ ... 1 Chapter Two Corruption ................................ ................................ ................................ ..... 9 Schools of Thought and Empiric al Evidence ................................ ......................... 13 Chapter Three Theoretical Perspective ................................ ................................ ............ 21 Durkheim ................................ ................................ ................................ ............... 21 Merton ................................ ................................ ................................ .................. 25 ................................ ....... 28 Institutional Anomie Theory and Cultural Dynamics ................................ ............ 36 Empirical Tests of Institutional Anomie Theory ................................ ................... 40 Alternate Con figurations of Institutional Imbalance ................................ ............ 54 Markets, Morality, and Crime ................................ ................................ ............... 57 Chapter Four Data and Methods ................................ ................................ ...................... 60 Design Strategy and Research Purpose Revisited ................................ ................ 62 Methodological Considerations in Cross National Research ............................... 62 Sample ................................ ................................ ................................ ................... 64 Dependent Variable ................................ ................................ .............................. 66 Independent Variables ................................ ................................ .......................... 72 ................................ ..... 72 State Dominance ................................ ................................ ....................... 77 Economy ................................ ................................ ................................ .... 86 Family ................................ ................................ ................................ ....... 88 Education ................................ ................................ ................................ .. 89 Control Variables ................................ ................................ ................................ ... 90 Analytic Strategy ................................ ................................ ................................ ... 9 3 Chapter Five Results ................................ ................................ ................................ .......... 94 Multivariate Models ................................ ................................ .............................. 97 Mediation ................................ ................................ ................................ .............. 99
v Moderat ion ................................ ................................ ................................ ......... 100 Cultural Indicators ................................ ................................ ............................... 105 Chapter Six Discussion ................................ ................................ ................................ .... 110 Limitations ................................ ................................ ................................ ........... 115 Conclusion s ................................ ................................ ................................ ......... 118 References ................................ ................................ ................................ ...................... 119 Appendix ................................ ................................ ................................ ......................... 142
vi List of Tables Table 1: Karklins Typology of Post Communist Corruption ................................ ............. 12 Table 2: Crime Types by Dominant Institution and Cultural C onditions ......................... 55 Table 3: Negative Binomial Regression Estimates for Bribery Rates ............................... 68 Table 4: Principal Components Analysis for Eight Dimensions of Polyarchy ................... 82 Table 5: Principal Components Analysis for Five Dimensions of Industrialization .......... 88 Table 6: Descriptive Statistics and Selected Zero Order Correlations ............................. 95 Table 7: Corruption Perceptions (TI CPI) Regressed on IAT Indicators ........................... 98 Table 8: Corruption Perceptions (TI CPI) Regressed on IAT Indicators, Moderation Models ................................ ................................ ................................ ................................ ......... 102 Table 9: Corruption Perc eptions (TI CPI) Regressed on IAT Structural and Cultural Indicators ................................ ................................ ................................ ........................ 106
vii List of Figures Figure 1: Corruption Perceptions Worldwide ................................ ................................ .... 2 Figure 2: State Dominance Corruption Functional Form ................................ ................ 83 Figure 3: State Dominance Corruption, Reverse Inverse Tr ansformation ...................... 85 Figure 4: State Dominance Corruption, Log Transformed ................................ .............. 86 Figure 5: Corruption Perceptions and Industrialization Scores ................................ ..... 104
v iii State Dominance and Political Corruption: Testing the Efficacy of an Alternate Configuration of Institutional Anomie Theory Cross Nationally Carol L. S. Trent ABSTRACT anomie theory (1994) have generally supported the thesis that, in social collectives where the economy dominates, non economic instit utions (i.e. the family, education, polity) are rendered feeble, unable to exert their normative controls. The cultural values of these societies weaker emphasis on th e licit means of achieving these goals. The resultant state is one of anomie, conducive to crime. Messner and Rosenfeld have extended their argument stating that it is not economic dominance per se that contributes to high crime rates, but any tip in the institutional balance of power The current study examines one of these configurations which hypothesizes that, in nation states where the state dominates other institutions, the dominant cultural orientation is one of moral cynicism conducive to corruption prone behaviors. Using macro level data, the current study assesses the efficacy of this alternate configuration of institutional anomie theory as a predictor o f corruption cross nationally. Using a sample of 125 nations, state dominance is positively related to corruption. The effects of the state were both mediated and moderated by economic strength, measured as levels of industrialization
1 Chapter One Introduction Over the past two decades, interest has risen sharply in the international community on the topic of political corruption and its damaging impact on societies, especially those in the devel oping and transitioning world. International financial institutions such as the International Monetary Fund (IMF) and the World Bank have corruption policy reforms (Bro wn and Cloke, 2004; Elliott, 1997), and corruption has emerged as a global issue in bodies such as the Organization for Economic Co Operation and Development (OECD), the Organization of American States, and the United Nations General Assembly (Fitzsimons, 2002). Anti corruption civil society organizations such as Transparency International have been calling for greater accountability and transparency on the part of nations. These organizations may look to academic research to guide policy and reform, howe ver; the dissemination of empirical corruption studies has, to date, been scant. Figure 1 (below) illustrates the pervasiveness of corruption worldwide in 2006. Published by Transparency International, the Corruption Perceptions Worldmap is based upon t from within and outside each nation. The United Nations estimates that bribery has
2 bec Perceptions Worldmap by the red to dark red or wine colored nations (those whose corruption perceptions score is 4.0 or below), corruption is endemic on every continent, with the possible exemption of North America and Oceania. The countries most rife with corruption are largely the developing nations of Africa, Latin America and Southeast Asia, as well as the transitional post Soviet nations of Eastern Europe. Figure 1: Corr uption Perceptions Worldwide (2006) Reprinted from Corruption Perceptions Worldmap. Copyright 2007 Transparency International: The Global Coalition Against Corruption. Used with permission. 1 According to a 2002 survey, Transparency International re ports that 96% of respondents in Pakistan who have had contact with the courts have encountered corrupt 1 Transparency International (TI) is the global civil society organization leading the fight against corruption. Through more than 90 chapters worldwide and an international secretariat in Berlin, Ge rmany, TI raises awareness of the damaging effects of corruption and works with partners in government, business and civil society to develop and implement effective measures to tackle it. For more information, visit http://www.transparency.org.
3 practices (Rodriguez and Ehrichs, 2007). Susan Rose Ackerman notes that in Guinea continuous demands for bribes are a feature of virtually every busine ss deal, while of that agency signed control of that department over to a private Swiss firm (1997, p. 32). Estimates suggest that annually up to five percent of the glob al economy is lost to corruption (Karklins, 2005). What is corruption? Scholarly researchers and public policy makers have struggled with the ambiguity of a definition of corruption and its myriad forms, frequencies, and consequences (Johnson and Sharma, 2004). Definitional issues will be discussed at length in Chapter 2. For the purposes of this study, the succinct definition given by Transparency International, the misuse of public power for private gain at the expense of the public good will be used (www.transparency.org). This definition limits the scope of the study to the political realm, yet is broad enough to allow for comparative study across a range of nations and cultures. The very nature of corruption makes this crime inherently difficult to study and, partially due to this, scholarly research on the topic has been largely underdeveloped. While this theoretical and analytical gap has been shrinking in the past decade, criminology lags behind disciplines such as political science, economics anthropology, and public administration in its treatment of this problem. A possible explanation as to why criminologists seem to ignore this phenomenon, aside from the intrinsic sub rosa character of corrupt acts, is that criminology in general, and A merican criminology is
4 criminologists working in the arena of state and political crime have noted that, from 2 allocated a mere 3% of their 575 articles to topics involving crimes of the powerful, with only seven of these addressing either political or state crime. The remaining space was devoted to lower class crime and/or its control (Lynch and Michalowski, 2006 p. 194). Although in his 1989 American Society of Criminology (ASC) presidential address, William J. Chambliss called for the all of the works on state crime in the la st thirty years have included calls (and sometimes a desperate plea) for more criminological attention to illegal actions of states and Matthews, and Miller, 2001, p. 174) Definitional disputes and measurement issues have in large hindered the study of this phenomenon (Jain, 1998, p. 4). As corruption is an international concern, research pertaining to this phenomenon must be comparative. However, cross national/compar ative criminology, including the study of state crime and other crimes of the powerful has been relegated to a small sub field in orthodox criminology. Globalization has increased the need for scholars to understand their ever changing world as a whole. This is not new. Starting as early as 1889, anthropologists sought to construct cross cultural theory, and, political science followed suit (Bennett, 1980). Criminology m ust also look to 2 These the American Society of Criminology ( Criminology ) and the Academy of Criminal Justice Sciences ( Justice Quarterly ), as well as the British Journal of Criminology.
5 comparative study for both theory testing and theory generation, and for the universal while avoiding the pitfalls of grand theories that are a cultural and a historical (Lynch and Groves, 1995; Mills 1959/2000). Furthermore, cross national testing of theory country based 4) that is narrow in scope, its theories applicabl 253.), toward a more encompassing approach. In addition to adding to the knowledge base, there are practical applications for comparative research. Gl obalization has made corrupt practices a transnational crime much like human trafficking, drug smuggling, industrial and technological espionage, and dissemination of child pornography, among others. Transnational crime, by its very nature, requires a coo rdinated and cooperative response in order to determine best practices, analyze how criminal organizations operate, and guide public policy (Bennett and Lynch, 1990; Howard, Newman, and Pridemore, 2000, Williams, 1999). With advances in transportation, co mmunication, and other technologies, economic crimes such as fraud, money laundering, identity theft, and corruption also extend beyond the realm of the political limits of the nation state (Howard et. al., 2000). Criminology must also broaden its horizon s. Similar to other white collar offenses, (price fixing, for example), individual acts of corruption often victimize persons without their direct knowledge Victims also may fail to report solicitations for bribes out of necessity or fear, due to threat s of extortion.
6 At the societal level, indirect violations of human rights and democracy through corrupt acts may lack a clear offender and victim, but constitute state crimes against the citizenry as a whole, and are thus worthy of criminological investi gation (Green and Ward, 2004). Corruption weakens the legitimacy of governments and undermines the confidence of citizens. It also distorts the distribution of benefits, exacerbating existing ly to flow to the few and the well connected while costs are extracted from society at large ultimately, from the See also Girling, 1997; Fitzsimons, 2002; Lynch and Michalowski, 2006; Rose Ackerman, 1997). As Zimring and Johnson suggest: corruption is one category of crime where the strong will prey on the weak and where the net effect of many acts of corruption may be regressive rather than redistributive of income. In many, if not most, settings where corruption flourishes, the offense pattern produces greater rather than lesser concentrations of wealth among advantaged populations (2005, p. 798) The effects of corrupt practices on the disadvantaged range from further degraded living condition s among already impoverished citizens to even death. Rose Ackerman (1997) cites several case studies that provide examples of the harms caused by corruption. For example, one irrigation district in India indicated that 20 50% of government provided funds were wasted due to corruption and misuse and, in Korea, a shoddily built department store collapsed, killing several people building inspectors had allegedly accepted bribes from substandard contractors (1997, p. 32). What is
7 more, corrupt practices are associated with other criminal activities including drug trafficking, organized crime, money laundering, and illegal money transfers, the latter of which have been suggested to support international terrorist organizations (Gambetta, 2002; Jain, 1998; Mora n, 2001; United Nations, 2004). The current study seeks to add to this small but growing body of literature on corruption, a topic worthy of study due to its pervasiveness, global nature, and perhaps most importantly, its impact on humanity as a cause of suffering and despair. As noted by Jain (1998), the prevalence and potential harms produced by corrupt acts has, as of yet, far exceeded the scholarly that this phenomenon has received The current study seeks to aid in the remedy of this omission The purpose of this research is to transport Anomie national comparative analysis (Karstedt, 2001, p. 292). In contrast t o the conventional interpretation of IAT, where economic dominance weakens the normative controls of other institutions leading to high levels of instrumental crime, here the institutional balance of power is tipped in favor of the state, which predicts h igh levels of crimes of power such as bribery, clientelism, maladministration, and abuses of authority. Under this configuration, rather than a normative breakdown leading to the anomic state, state dominance leads to cultural patterns of moral cynicism a nd withdrawal in the citizenry, which Messner and Rosenfeld suggest will enable corruption. This thesis is organized as follows: First, a working definition of corruption will be established and correlates and schools of thought discussed. An overview of the
8 findings from extant corruption literature will be provided, with an emphasis on macro level, cross national research. Chapter 3 provides the theoretical background of institutional an omie theory, a review of literature, and introduces the alternate configuration of institutional anomie to be tested in the current study Chapter 4 will state the hypotheses to be tested, and provide an in depth discussion of the data and methods used. Chapter 5 will present univariate (descriptive), bivariate, and mu ltivariate results. An examination of the assumptions of ordinary least squares will be presented, with alternate models proposed as a from of sensitivity analysis. Chapter 6 will discuss the findings, acknowledge the limitations of the current study, an d present calls for future research.
9 Chapter Two Corruption What is corruption? Scholarly researchers and public policy makers have struggled with the ambiguity of defining it and its myriad forms, frequencies, and consequences (Gambetta, 2002; John son and Sharma, 2004, United Nations, 2004). The literature provides multiple definitions, yet none are clear cut or definitive. According 2001, p. 28). Legalistic, relativistic, and descriptive definitions have been suggested, but a clear consensus has yet to be reached The concept of The use of legal definitions a is always Western expectat Rooted in the concept of the rule of law, legal definitions of corruption often go hand in Cultural relativi sm, however, suggests that these rule and norms are not universal, ledgling democracy, or as a holdover from a traditional p ast based upon
10 norms of kinship (Saj, 2003, p. 187) In some countries, rather than corrupt activities being proscribed, it is considered a moral duty for a state agent to act for the benefit of fa mily and/or friends (Amundsen, 1999) legitimate and legal varies from country to country and time to time, the definition of Alth ough corruption may occur in private enterprise, as well as at the intersection of the state and corporations where the state initiates or facilitates illegal or injurious actions perpetrated by corporations, generally, when we speak of corruption, we are concerned with administrative or political malfeasance that which occurs in the public sphere (i.e. governments, the state apparatus). Nye (1967) supplies one of the most popular definitions of public corruption. In his seminal book Bribes Nye defines c orruption as a: behavior which deviates from the normal duties of a public role because of the private regarding (family, close private clique) pecuniary or status gains; or violates rules against the exercise of certain types of private regarding influen ce. This includes behaviors as bribery (use of rewards to pervert the judgment of a person in a position of trust); nepotism (bestowal of patronage by reason of ascriptive relationship rather than merit); and misappropriation (illegal appropriation of pub lic resources for private regarding use) (quoted in Kotkin, 2002, f.n. 25). Robinson (1998) further distinguishes between incidental (individual), institutional, and systemic (society wide) corruption, while Philp (1997) notes that
11 definitions are mostly c entered on public office, public interest, and the market. perverted) state In order to facilitate comparative research, Johnson and Sharma (2004) s uggest that researchers define corruption by example. Few states have laws that expressly define a category of illicit acts as political corruption, rather laws are codified that condemn crimes such as bribery, extortion, or electoral misconduct (Philp, 1 997). International bodies such as the United Nations and the International Chamber of corruption that exist now, but also enabling states to deal with other forms that may emer developed a typology of practices in post communist nations that illustrates the range of activities that fall under the rubric of corrupt acts (see Table 1).
12 T able communist corruption Low level administrative Bribery of public officials to bend rules Deliberate overregulation, obfuscation, Using licensing and inspection powers for extortion Self serving asset stripping by offi cials Diverting public resources for civil servant spoils Mismanagement and profiteering from public resources Profiteering from privatization Malpractice in public procurement pt networks De facto takeover of public institutions for private business or criminal activity Forming collusive networks to limit political competition Undermining free elections through slush funds, hidden advertising, etc. Misuse of legislative p ower Corruption of the judicial process Misuse of auditing, investigatory, and oversight powers Using kompromat (propaganda) for political blackmail and coercion Corruption of and in the media (Adapted from Karklins, 2002, p. 24) As illustra criminal offense of bribery. In addition to the examples of malfeasance listed corrupt administrations often operate in collusion with other criminal enterprises such as drug traffic king, money laundering, organized crime, the underground economy, and illegal money transfers (Gambetta, 2002; Jain, 1998; Moran, 2001). Although definitional issues are important, they remain unresolved in the literature, and lie beyond the scope of the p urpose of the current analysis. For the purposes of this research, a conventional, albeit imperfect, core definition of political corruption supplied by Transparency International will be used the exercise of public power for private gain (www.transparenc y.org) This definition spans the corruption level administrative acts
13 such as bribing an officer of the law to rent seeking (public officials illegally charging for services after creating artif state capture (Johnson and Sharma, 2004; Karklins, 2002; 2005). The latter may be exemplified by cases such as Zaire under Mobutu, the Philippines under Marcos, and Nicaragua under Samosa ( Jain, 1998, p. 23). Compared to the relativistic stance, where corruption is placed within respective social contexts, the broad definition adopted in the present study allows for cross national comparison across a range of cultures What we gain is gene ralizability, and the ability to compare the same behaviors across space and time, even where legalistic definitions may prove inapplicable ( s ee Gambetta, 2002, p. 51). Finally, this definition is consistent with the current consensus found in the literat ure (Goudie and Stasavage, 1998; Kaufmann, Kraay and Mastruzzi, 2003; 2007a; 2007b; Montinola and Jackman, 2002; Sandholtz and Koetzle, 2000; Svensson, 2005; Treisman, 2000; Xin and Rudel, 2004) as well as with the analytical purpose of the current study. Schools of thought and empirical evidence Ackerman, 1978, p. 3). The definitional discrepancies discussed above lead Jain (1998) to have yet developed a vocabulary, leave [ sic ] alone a theory, that can explain the Although scholarly research pertaining to this phen omenon exists prior to the m ost notably, Leff, 1964; Nye, 1967; Ro se Ackerman, 1978; Scott, 1972 ), a
14 majority of the cross national, cross cultural comparative research has appeared only in the past two decades T he bulk of these studies emerg ed f rom the disciplines of political science, public policy, and economics and were largely driven by anti corruption agendas (Kotkin and Saj, 2002) The principal barrier to comparative studies of corruption has been measurement problems, such as the lack of valid measures across nations (Ades and DiTella, 1999; Husted, 1999; Knack, 2006; Mauro, 1997; Rose Ackerman, 1997; Svensson, 2005; Xin and Rudel, 2004), and therefore, large n empirical literature is greatly outnumbered by theoretical literature and ca se studies (Husted, 1999; Treisman, 2000). With the advent of subjective cross national corruption indices, such as those published by Business International (BI), the World Bank Group, and Transparency International a number of macro level studies across nations have been made available in recent years. Early theories of corruption assumed that corruption and bribery are universally immoral and ethically suspect (Kotkin and Saj, 2002; Nye, 1967; Saj, 2003). Under the 19 th century idea of the rule of l and less developed nations (Kotkin and Saj, p. 25). Under a moralistic approach, definitions focus on behaviors, c ertain motive or traits of an agent, (or individual), that leads to corrupt acts and practices. These suspect behaviors may include the integrity, or even to their depra
15 theories of moral degradation. Cultural relativist thought explains corruption by societal emphases on gift giving, kinship loyalties, and persona l ties rather than an adherence to the rule of law (Kotkin and Saj 2002; Jain, 1998; Montinola and Jackman, 2002). Based upon cultural norms and rooted in the respective social context, what may appear in a purely legalistic sense as corruption, can be seen as mere cultural differences rather than social harms, making a core definition of the practice impossible. Few quantitative studies include measures of cultural norms, due to the difficultly of obtaining valid and reliable measures of culture, espec ially at the macro level. Additionally, norm based several correlates of corrupti on have appeared in the literature that are consistent, on their face, with a cultural relativist school of thought. Ethno linguistic fractionalization -or the probability that two persons drawn from a population at random will not belong to the same g roup -has been positively related to corruption (Mauro, 1995, Morris, 2004; Shliefer and Vishny, 1993; Treisman 2000) Research indicat es that corruption is lower in culturally homogeneous populations, as bureaucrats are likely to favor members of thei r own ethnic group (Mauro, 1995, p. 693) Eastern Orthodoxy, Islam) present less challenge to office holders, while, in the case of Islam, for example, the division between church and state may be blurred (p. 403 See also Paldam, 2001). Protestantism, on the other hand, with its values of egalitarianism
16 and individualism has been found to reduce corruption by state institutions in its congregates, which aids in exposing cases of corrupt behavior (2000, p. 421) Lastly, regional groupings of nati ons high on corruption rank order corruption have been observed in African, Latin American and post Soviet nations (Montinola and Jackman, 2002; Moran, 2001; Xin and Rudel, 2004), although this may be confounded by factors such as slow development, bloated states sectors, a history of colonization, or other economic or political factors. Emerging in part from the cultural relativist viewpoint, the neo functional app roach to the study of corruption introduces efficiency enhancing models of corruption ( Montinola and Jackman, 2002, p. 148 ). Under the neo functionalist theoretical framework, corruption is a functional necessity, as governments, especially burgeoning d emocracies, are unable to provide basic services through legitimate means ( s ee Leff, 1964). Much like the relativist argument, neo functionalism disputes the application of Western norms, ideals, and legal culture to developing and transitional societies, suggesting that corrupt behaviors are inevitable by products of modernization (Scott, 1972). According to this school of thought, corruption is seen as beneficial to The neo (Montinola and Jackman, 2002, p. 148). Empirical research has refuted this viewpoint.
17 Maur deters economic growth by lowering the investment rate (1995, p. 695). Subsequent studies confirm that the net effect of corruption on economic development is negative (Mauro, 1998; Rose Ackerman, 1999; T anzi, 199 7 ; Tanzi and Davoodi, 1997). Several studies have examined the relationship between democratization and neoliberal free market principles and corruption. Public choice theory (and what is roughly its micro level counterpart principal agent theor y) suggests that corruption will be reduced by economic deregulation and by political competition via the election process. Neoliberal policies of market competition (versus government intervention and regulation), measured as openness to trade, decreases corruption (Ades and DiTella, 1997; 2000; Gerring and Thacker, 2005; Sandholtz and Koetzle, 2000), while accountability, transparency, openness, predictability, and rule of law have been linked to a reduction in corrupt activ ities within social collectives (Goudie and Stasavage, 1998). Transparency, accountability, and a strong civil society (such as a free press) further diminish opportunity structures conducive to corruption, for example monopolistic government services (M ontinola and Jackman, 2002; Rose Ackerman, 1978, 1999). The stylized fact that poor nations tend to be more bureaucracies government size is related to corruption (Mauro, 1995, p. 706). In the literature, r esults on this point have been mixed Montinola and Jackman (2002) and Husted (1999) found that the size of government, measured as the government share of the Gross Domestic Product (GDP), actually reduced corruption, while LaPalombara
18 (1994) reports the opposite. The relationship between democratization and corruption, the theoretical foundation that most informs the current study has also been mixed in the literature. For example, a linear, negative relationship be tween political democratization and corruption has been found (Goldsmith, 1999; Sandholtz and Koetzle, 2000; Xin and Rudel, 2004), while other researchers have found the inverse (Scott, 1972; Moran, 2001) Treisman (2000) found no significant relationship between corruption and present democratization, but rather, that the association is dependent upon the length of uninterrupted democratic rule, with twenty years being the minimum before any significant effect emerges These mixed findings may be due to the functional form that the democracy corruption relationship may take. democracy freedom of group opposition, political rights, and legislative effectiveness found that, du ring the post 1974 third wave of democratization ( See Huntington, 1991), no relationship emerged between democratization and corruption. Only when a non linear quadratic transformation was applied to the democracy measure did a significant association sur face Sung (2004), on the other hand, suggests that the democracy corruption relationship takes on a cubic form. As nation states move toward more open, predictable, and transparent policy making, structured under the rule of law, corruption rates drop precipitously, but then appear to resurge. This may indicate a non recursive relationship, where the causal arrow then flows back from corruption to democracy levels, dependent upon initial conditions, as in the case of
19 post soviet Russia Clearly, such findings warrant further investigation of the democracy corruption fun ctional form. Additional socio structural correlates of corruption found in recent literature include structural elemen ts of the polity. For example, larger governments become more difficult to manage, providing more opportunity for corrupt acts due to the large amounts of impersonal dealings ( Montinola and Jackman, 2002; Scott, 1972), as do federalist states (Goldsmith, 1999; Treisman, 2000). Natural resource endowments, for example the oil rich Middle Eastern states or ore rich former Soviet nations provide sources of rents, as these commodities can be sold at a price that greatly exceeds the cost of extraction (Ades a nd DiTella, 1999; Fish, 2005; Fish, 2005; Lambsdorff, 1999; Leite and Weidmann, 2002; Mauro, 1997; Robbins, 2000 ; Treisman, 2000). Economic traits such as overall wealth per capita has been found to decrease corruption (Goldsmith, 1999; Xin and Rudel, 20 04), while relative and/or absolute deprivation can be both a cause and consequence of corrupt administrations (Goudie and Stasavage, 1998; Lambsdorf, 1999) Finally, common law systems and a legal culture inherited by British colonization (Sandholtz and Koetzle, 2000; Treisman, 2000) reduces corruption through the emphasis on precedent and procedure. It is noteworthy that a bulk of corruption literature appears in the disciplines of political science, public policy, and economics. Although corruption, as a crime of the powerful, influences and impedes economic development, weakens public trust and undermines political legitimacy, diverts spending designed to help the poor and reduce inequality, and erodes social justice (Morris, 2004; Zimring and Johnson, 20 05), this
20 subject has been largely ignored by criminologists for theory testing and theory building. The proposed thesis seeks to fill this gap in the literature by testing a general theory of crime that, through an alternate articulation, proposes an exp lanation for corruption anomie theory.
21 Chapter Three Theoretical Perspectives Durkheim under normal conditions, and Bachman, 2001, p. 141, emphasis in original) In motivational theories such as anomie, we must account for rule breaking The anomic paradigm in criminology can trace its theoretical roots to the work of French sociologist Emile Durkheim. Durkheim first introdu ced the concept of anomie which typically refers to a societal condition where social norms are unclear or non existent, in his doctoral dissertation The Division of Labor in Society (hereafter Division ). In Division Durkheim suggested that anomie was d of the division of labor, brought about by the progression from the mechanical solidarity of simple societies to the more complex organic solidarity common in industrialized societies (1984, p. 301). In simple societies, collect ive moral beliefs (generally centered on religion) yield societies become more complex, labor becomes specialized and differences between individuals become more ap functionally differentiated parts must ultimately depend on mutual consensus
22 ess either does not exist or is not related to Conditions of the and 108). There is a large body of literature, including critiques, empirical evaluations, and reviewed here T wo represent a gen eral overview of cross national findings based upon advantage of the natural experiment that is the Russian Federation. ross national level (N = 33) using population and urbanization measures to capture the moral density of a society and energy consumption as a proxy for industrialization A division of labor measure developed by Gibbs and Martin (1962) captured the evenne ss of the distribution of individuals into various industrial categories; greater diversification indicated greater interdependence in a given society. Lastly, anomie was operationalized using Feierabend and of systematic frustration computed by subtracting the sum of the mean standard score of the following four variables : g ross n ational p roduct per capita (GDP) and the number of radios, newspapers, and telephones per capita model failed to support anomie variable was removed, bivariate analysis showed an expected, moderate, and
23 positive relationship between population, moral density, a diversified division of labor, and both property and total crime rates across nations (1978, p. 665 666). Messner (1982) argued that development will have no effect on homicide rates cross by structural changes of an egalitarian nature which permit the b with equality (measured as 1 Gini coefficient of income inequality) inversely related to homicide rates, consistent with the hy pothesis that egalitarian trends promote organic solidarity However, population has the opposite effect suggesting that rapid in and of itself (p. 236 ). Recent research has been able to take adv antage of a unique natural experiment -the events that have occurred in Russia after the collapse of the Soviet Pridemore (2005 a ) found that rapid socio economic change i ncreased homicide rates across 78 sub national Russian regions (out of a total of 89 regions) Additionally, their research indicates that rapid political change (measured as the vote for opposition parties versus ballots for the Communist Party of the Russian Federation or the Libera l Democratic Party of Russia mbling Communist or Soviet Rule ; ( Pridemore and Kim, 2005b p. 88 ) exerts a strong, positive influence on the change in post 1991 property crime rates This relationship holds after controlling for socioeconomic indicators. The researchers interpret these findings as indicative of a short
24 [I]t is fair to argue that long standing Russian cultural tradition s included strong bonds based on collective sentiments, and the Soviet era resulted in ideologically rooted and exaggerated sentiments about the collective. Such cherished beliefs, according to Durkheim, are bound to elicit heightened passions when threat ened (p. 99). Using interrupted time series analysis with the 1991 fall of the Soviet Union as post Soviet collapse homicide rates are not due to confounding influences (Pride more, Chamlin, and Cochran, 2007). This lends further support for change in a collective yields an increase in deviant behavior, as individual desires replace the collective, and the society struggles to adapt to new, a nd perhaps, unclear norms. In Division Durkheim argued that societal discordance is due to inadequate procedural rules. With 1897 publication o n suicide rates in France, anomie takes on a different meaning. In Suicide, Durkheim posited that the systematic and patterned nature of rates of suicide across social collectives suggested a socio structural, rather than individual, explanation for this behavior. Durkheim developed a typology of suicides egoistic, altruistic, fatalistic, and anomic -the latter being the this moderating ro (Durkheim, 1997, p. 252) Durkheim discovered that anomic suicides fluctuate with economic and political conditions, especially sudden change, when new rules for
25 interaction have not yet form ed (Willis, 1982). Most interesting is that Durkheim found anomic suicide rates to increase not only during economic depressions, as may be expected, but also during periods of economic boom. According to Durkheim, booms and busts yield the same result, Hence, the causative factor is not the material circumstances themselves, but the instability, which they introduce into social life. In circumstan ce of social are placed under stain, and tend to lose their hold (1979, p. 50). In sum, anomie, as developed in Suicide of inadequate mora l norms to guide and control the actions of people and groups in the crime relationship as normal, even beneficial in a society, as it strengthens the conscience collect ive: Crime therefore draws honest consciousnesses together, concentrating them. We only have to observe what happens, particularly in a small town, when some scandal involving morality has just taken place. People stop each other in the street, call upon one another, meet in their customary places to talk about what has happened. A common indignation is expressed (1984, p. 58). Merton If for Durkheim, deviance prevents anomie, for Robert K. Merton, anomie produces deviance. With the 1938 publication o
26 Merton, argued that the variance in the rates of crime across social collectives could be Similar to Durkheim, Merton explains differences in crime rates across social aggregates in terms of the fundamental properties of social systems. According to Merton, societies, such as the United States, suffer from unusually high rates of crime due to the strong, universalistic emphasis placed upon monetary success goals, coupled with a weak emphasis on legitimate institutionalized means for achieving these goals. This configuration and its resultant state of malintegration leads to an For Merton, a healthy society is one in which goals and means are in equilibrium An anomic society is characterized by mal integration. In the Mertonian explanation of goals, but rather the means used to obtain these goals that become problematic (Paternoster & Bachman, 2001, e is, then, based upon a disjunction between the goals that individuals set for themselves, (or rather, that society prescribes that they set for themselves), and the availability of legitimate means, or opportunities that ar e afforded to them (1938; 1968) level anomie, exists in the literature, although several studies have included variables that tap into constructs of inequality/deprivation consistent with the basic premise of anomie theory. For example, Blau and Blau (1982), using data for 125 metropolitan areas in the United
27 substantially raises its rate of study, conceptualize anomie disruption (percent of population that is divorced/separated). Th explanations such as the Southern culture of violence thesis (p 122). Recently, Baumer and Gustafson (2007), developed a precise causal model of anomie theory (discussed below) using macro level data from the General Social Survey (GSS). In their interpretation, instrumental crime is expressed as a function of cultural structure (societal emphasis on pursuing monetary success goals, and the emphasis of the collective on pursuing these goals through licit mean s) interacting with social structure (limited access to opportunities (means) of obtaining such goals) (p. 621). version of the GSS, the researchers measured commitment to monet ary success as agreement with the statement : W eak commitment to legitimate means was measure u s ing agreement with the statement: (Baumer and Gustafson, 2007 p limited
28 was captured by the ratio of the total numbe r of persons employed/seeking employment to the number of jobs available in the local labo r market area A dditional variation in opportunities was also measured ( low educational and economic attainment and educational and income inequality ) I n stitutional anomie theory (IAT) measures introduced into the study included the strength of the following institutions described in Mes education, the family, and the polity; two other socio structural variables were included to represent two additional institutions religion and social capital. Net the effects of appropriate controls, the researchers found sign ificant interactions between high commitment to monetary success and low commitment to obtaining this success through legitimate means, although limited and unequal opportunities failed to moderate this effect as 655), as well as calls for further integration of the theories of the anomie paradigm. anomie theory In their contemporary elaboration of anomie theory in Crime and the American Dream framework, atte mpting to explain the structural aspect of the theory beyond social relationships that ensure the basic survival requirements of the society are fulfille d, i.e. on social institutions,
29 institutions, as an explanation for the cultural orientations that favor criminal behavior (Messner, 2003, pp. 4 5, emphasis in original). The result is an institutional anom ie theory of crime ( hereafter IAT). commitment to the goal of material success, to be pursued by everyone in society, Rosenfeld, four core values underlie this defining characteristic of American culture: (1) achievement orientation tes with a failure individualism where people are universalism where all members of society are expected to strive for the same goal, regardless of the avail ability of the means to fetishism of money monetary success), coupled with a much weaker emphasis on the licit means of achieving these goals (Messner, 2003, p. 3). Using the United States as the archetypal case, Messner and Rosenfeld posit that a climate of competition, ind ividualistic achievement orientation, and pressures to due, in part, to the consistent emphasis and reinforcement of the monetary success oriented culture (through ou tlets such as mass media and advertising), while, at the
30 same time, equal weight is not given to the legitimate channels that allow these objectives to be realized (mainly education and employment opportunities). This malintegration of goals and means is of social organization; the consequence of this disjunction is anomie. Under this condition, a breakdown occurs in the normative social controls that promote conformity, as individuals are unable to attai n culturally prescribed success goals when the opportunities to do so are beyond their grasp. Anomic societies are criminogenic 58). This cultural ethos does not exist in a vacuum; it is expressed in the social structure of a given collective tructural arrangements that encourage and and Rosenfeld expand anomie theory beyond mere cultural analysis and stratification. Institutional anomie theory examines the i nterplay between four primary institutions in light of the collective culture of a society. Messner and Rosenfeld focus on four elemental institutions found in social collectives. The economy functions around the production of goods, trade, and distribut ion, providing the basic requirements for sustenance: food, clothing, shelter. The polity is responsible for the safety and security and distributes power to attain col family is
31 other institutional domains Education provides additional socialization skills, preparing children for future occupations, as well as expanding the knowledge base of the citizenry (Messner and Rosenfeld, 2007, p. 72.) These institutions are to some degree interdependent and coordin ated, held in balance by the values of the collective populace. For example, the economy is dependent upon produced by the educational system, which is, in turn, dependent upon the early socialization of students by their respec tive families, or institution s are influenced by the culture of the collective, while, at the same time, these mately, where culture economic institutions, those weakened institutions lose their ability to control the populace (2006, p. 130 131), resulting in a state of normlessness, or anomie (Messner, 2003; Messner and Rosenfeld, rst of these mechanisms is devaluation of non economic institutions. For example, education is regarded as a means to earn a degree to obtain employment or to learn a marketable skill, rather than knowledge for its own sake. Likewise, is valued over the homemaker or
32 caregiver and politics is left to the career politician rather than the average citizen (Messner and Rosenfeld, 2007, p. 76 79). A second way that the economy dominates is through accommodatio n For example, in the United States, long work schedules and overtime is valued, while family and parental leave is given, but often without pay. few workers must dominance is through penetration economic institutions. Education becom business. Since the original 1994 publication of Crime and the American Dream, Messner and Rosenfel d have elaborated upon th e idea of institutional anomie theory to suggest s ee also Rosenfeld, 2004). The first type, exemplified above, of a strong free market e conomy as dominant, with non economic institutions weakened or subsumed by the strength of limited to, offenses of theft, burglary, and homicide. A second arrangeme nt occurs group loyalty and civil rights violations. A third configuration may occur when th e state dominates.
33 ssner and Rosenfeld, 2004, p. 96). This latter articulation is the focus of the proposed thesis. anomie theory states that dominance of the economy coupled with weakened non economic normative con trols through the three processes described above, along with an exaggerated anomic state, as manifested by the disproportionately high crime rates in the United States ( p. 84). They do, however, acknowledge that other play in the dominance of other institutional entities, resulting in the rise of different types of crime. Thus, it is institutional imbalance per se rather than strictly eco nomic dominance, that IAT proposes leads to crime. The literature, however, has narrowly focused on this configuration and on street crimes, conventionally defined. This serious limitation robs IAT of full analysis of its scope and generalizability. In f act, to date few empirical tests of institutional anomie theory have been published Messner and Rosenfeld suggest that this is due to the high level of [ T ] ranslating the abstract theoretical claims of IAT into specific empirical propositions is thus a challenging, i 9). This ambiguity of operationalization and measurement calls into question whether IAT can be subjected to falsification (Chamlin
34 and Cochran, 2007). Furthermore, the theory is not amenable to direct testing due to the nec essity of employing indirect measures of key theoretical constructs. This drawback is not unique to IAT, rather, it is common amongst the majority of macro level theories, especially when cast at a high level of aggregation. Chamlin and Cochran note that it is possible to evaluate macro make predictions about the relationships between structural predictors and rates of 2007, p. 42). They do so in their initial evaluation of IAT by the inclusion of cross product terms to evaluate conditional relationships between the economy and non economic institutions T (to be discussed in detail shortly) support their hypotheses that institutional anomie uniquely predicts that the effect of economic conditions on profit motivated 2 007, p. 42, emphasis added). On a cross national level, support for or against institutional anomie appears to depend on the nations included in the sample. Messner and Rosenfeld state that developed societies of the world prompt the quest ion of why the rates of serious crime are so surprisingly high in one emphasis added). In their critical evaluation of IAT, Chamlin and Cochran (2007) question this restriction, citing that there are no inherent pr actical or theoretical reasons for excluding less developed nations. Indeed, their analysis of mean homicide and robbery rates for a developmentally diverse sample of nation stated showed that
35 against a sub sample of Western European countries. Messner and Rosenfeld counter that, although nations such as the Russian Federation have exceeded the United States tive about crime in the American proclivity for firearms. Nonetheless, the greatest empirical support for the theory arises from tests of nation states at similar levels of industrial advancement, Additional criticisms of IAT include: (1) the possibility that its scope may be limited to instrumental crimes (Chamlin and Cochran, 1995); (2) the c oncept of and across nations; (3) IAT fails to incorporate in its analysis variables central to other macro level theories of crime, instead relying on a universal shared va lue system based upon economic roles, risking possible spurious relationships (Jensen, 2002, p. 55 56); and (4) 002, p. 739). This latter criticism suggests articulating IAT through a multilevel lens ( s ee Baumer, 2007 for a multilevel model of Mertonian anomie theory that may be able to inform a possible MLM interpretation of IAT). The bulk of extant research has sought to use institutional anomie theory to assess the effects of socio structural and institutional dynamics on crime rates. Only a limited body of research has looked at the value orientations resulting from the
36 cs at work (Cao, 2004; Chamlin and Cochran, resulted in incomplete, partial tests of IAT due, in part to the aforementioned difficulty in operationalizing and measuring the dominant cultural orientation of a social collective. This synthesis of ideas is necessary in order to lend support to, or dispel the Institutional Anomie Theory and Cultural Dynamics In Crime and the American Dream, Messner and Rosenfeld ground their theory on a basic premise of Mertonian anomie theory that the exaggerated emphasis placed on pecuniary success, countered with only a weak emphasis on the importance of legitimate means to achieve these success goals leads to a state of anomie. Not only is forth as an explanation of why the rate of violent crime in the United States consistently exceeds that of other, similarly situated industrialized nations (p. 12, 25). To date, few empirical studies have examined the effects of the conflict between cultural values emphasizing material success and the available means to achieve these ends. Four studies that have addressed this thesis of IAT have turned to comparative, cross cultural analysis. Jensen (2002) and Cao (2004) examined the comparative cultural dynamics o f materialism using items from the World Values Survey. This instrument taps into individual opinions on family, work, religion, and success goals. Jensen assessed the premise of American exceptionalism as an explanation for the
37 higher homicide rate. He found no support for the notion that the United States is inherently more criminogenic nor for the decommodification hypothesis as suggested by Messner and Rosenfeld (1997) Jensen suggests that diversity, and conflict as opposed to economic national level (p. 69). culture is more anomic than other societi es, a proposal extended in the work of Messner and Rosenfeld. Cao used the summed score of six items from the World Values Survey as a proxy operationalization of anomie (justification of: claiming government benefits without entitlement, avoiding fare on public transport, cheating on taxes, knowingly purchasing stolen goods, accepting a bribe, and failing to report accidental damage to a parked vehicle). Similar to Jensen (2002), Cao found no support for American exceptionalism in the data, con trolling for social and demographic covariates. Chamlin and Cochran (2007) utilized two items from the World Values Survey, terialistic values. Additionally, their sample included developing and transitional nation states that were omitted from prior cross the USA are substantial, they are not did not rank highest on the World Values Survey measures included, inconsistent with countries, we find no evidence in support
38 that American culture places an unparalleled emphasis on the acquisition of goods and anomie appears only to hold for advanced industrialized nations. In the final cross Hoegl (2004) applied basic tenets of IAT to managerial ethical reasoning at the cross national level, again using World Values Survey data. Using hierarchal linear modeling for indivi dual and aggregate level data on 3450 managers from 28 nation states, the predictors of managerial willingness to justify ethically suspect behavior. This study provide s the first possible confirmatory link between the dynamics of IAT at the level of the social aggregate and that of the individual. Unfortunately, criminological research has not yet attempted to replicate these findings on instrumental crime, perhaps due to biases within the discipline in general, and anomie theories in particular, to keep micro and macro level articulations of theory distinct (Baumer, 2007, c.f. Agnew, 1987; Bernard, 1987. See also Messner, 1988). In a study that integrated both institutional and cultural dynamics consistent with institutional analytic scope, examining institutional and cultural values at the micro level. While institutional anomie theory was in troduced as a possible explanation for serious crime, student cheating. She rationalized this provides the student a type o
39 letter, of institutional anomie theory. mate rialistic values of the American Dream and have a higher likelihood of cheating than non U.S. born students. Additionally, those students involved in the economy (working outside of school) will be more success oriented and more likely to cheat. Finally, students who are involved in non economic pursuits (family, education, polity) will be less likely to engage in cheating, as these institutions will moderate the effects of the economically dominant society (2006, p. 636 637). Support for institutiona l anomie was mixed in this study. As predicted by theory, students who adhered to the cultural values of the American Dream were more likely to cheat, although this was not universal for U.S. born students. Moreover, those students who worked outside of school (actively participating in the economy) reported significantly lower instances of cheating. This was interpreted by the researcher as contrary to institutional anomie theory. Finally, neither interaction effects (moderation) nor mediation by non e conomic institutions emerged as significant for this sample N = 162), although this N is actually larger than the majority of studies of IAT. (p. 649). Due to study limitations and methodological issues in this study, it is not clear then at this time if IAT is applicable at the micro level or as an explanation of minor forms of deviance. Indeed, nothing in the literature suggested that IAT was intended to explain individual variation or minor deviant behav iors.
40 willingness of communities to commit scarce resources to the aid and comfort of the cultural aspect of institutional Cochran hypothesize that communities that place greater emphasis on volunteerism and char itable contributions enjoy lower rates of both violent and property crime (p. 209 211). Social altruism was measured as the ratio of United Way contributions to aggregate income in 354 United States cities The researchers controlled for both absolute a nd relative economic inequality, urbanism and opportunity measures, and demographic variables (1997, p. 211 institutional violent a members to respect and engage in behaviors that promote the welfare of others enjoy 221). Empirical Tests of Institutional Anomie Theory In an analysis of the relationship between social institutions and crime rates, Chamlin and Cochran (1995) published the first empirical test of institutional anomie theory using the 50 United States as the unit of analysis. They noted that a comprehen sive test of both core tenets of the theory culture and social structure would be difficult, if not premature at the time of their research (p. 415). Measures of
41 cultural dynamics for the United States were unavailable and, therefore, their research focus ed on the structural effects of economic and non economic institutions on four profit oriented crimes: robbery, burglary, larceny, and auto theft, measured as aggregate crime rates for 1980. While Crime and the American Dream (2007) laid a foundation for a socio structural analysis of crime rates, the authors do not provide detailed operationalizations of these constructs, nor do direct measures of these theoretical assumptions exist (Piquero and Piquero, 1998). Chamlin and Cochran (1995) operationalized percentage of families below the poverty line as a measure of absolute economic deprivation. The strength of non economic institutions (as buffers against the anomic effect of economic deprivation) was operationalized by three measures: ratio of yearly divorces to marriages per 1,000 (family disruption), adjusted church membership rate per 1,000 (strength of re ligious organizations), and percent of persons of voting age who cast ballots in the 1980 congressional elections (strength of the polity). Recall that these institutions are interdependent and that institutional anomie hypothesizes that a strong free mar ket economy coupled with the weakened ability of the non economic social sector to fill a normative role leads to the anomic state (Messner and Rosenfeld, 2007). To capture this interaction the researchers added three product terms: poverty*family, povert y*religion and poverty*polity. Chamlin and Cochran found that the additive effect of the independent variables
42 explained 96% of the variation in profit oriented crime across the 50 states. With the introduction of each product term into the regres sion model, the explained variance effect is a key contribution and assessed in all subsequent tests of institutional anomie rs that the interplay between economic and other social institutions determines the level of anomie within a 1995, p. 423). In their second empirical test of institutional anomie theory, Messner and Rosenfeld (199 7) examined the eff ects of decommodification on homicide rates cross nationally. They hypothesized that, similar to the restraining effect on the market by non stems (1990). Messner and Rosenfeld apply Esping (Esping Anderson, 1990, p. 22), to cross national crime rates.. Messner and Rosenfeld (1997, p. 1395), and applied this notion of deco mmodification to institutional anomie theory. These social welfare policies provide a bulwark to harsh economic dominance that, they believe, contributes to an anomic state and, thereby, crime. As Esping Anders o only contains dat a for 18 capitalist nations, Messner and Rosenfeld developed a proxy measure of decommodification so that a
43 larger sample of nations could be analyzed ( N = 45 maximum). Based upon International Labor Office data, this proxy measure is comprised of three indicators of national expenditures: average annual welfare benefits per household, social security expenditures as percent of gross domestic product, and percent of expenditures nt variable was lethal violence, measured as multi year homicide rate averages. The study supported this precept of institutional anomie theory, such that, across various model specifications (exclusion of Syria as a possible outlier and a smaller sub sam ple of 39 nations with complete data across all variables), nations exerting greater political restraint of the market through social welfare programs and model e xplains between 32.6% and 48% of the variance in homicides cross nationally (p. 1404), while advising that future tests of their theory should include non economic social institutions such as family, religion, and education (p. 1408). While Messner and R osenfeld provided the first cross national test of a facet of institutional anomie theory, Piquero and Piquero (1998) provide the first test using both varying operationalizations of the strength of two non economic institutions, along with assessing the e ffects of all four institutions identified in Crime and the American Dream on both violent and property crime rates. As a form of sensitivity analysis, Piquero and Piquero (1998) performed separate multivariate analyses, first operationalizing polity as percentage of population receiving public aid (a decommodification type proxy measure) and second as percent of citizens
44 voting for in the 1988 presidential election (as a measure of public involvement in government). The researchers represented the stre ngth of education using three variables: proportion of population enrolled in college, percentage of high school average annual pay (p. 69 70). Family disruption was mea sured by the percentage of single parent households; percent of population below the poverty level represented the economic dominance construct. Three interaction terms were introduced to the economic institutions (p. 72). Regression analyses of the varying specifications yield mixed support for institutional anomie theory, with the alternative specifications of polity and education failing to reach statistical significance in one model, and co ntradictory results found in other must be undertaken with extreme caution and respect for alternative est is extremely foremost challenges of empirically testing institutional anomie theory stems from employs concepts that are highly abstract and amenable to alternative interpretations. Savolainen (2000) examined institutional anomie theory cross nationally usi ng
45 of his findings on a different sample of nations ( N = maximum of 45) with partially different measures (p. 1027). The most significant contribution of this parall el sample is the inclusion of seven post Soviet/ Warsaw Pact nations in transition to market economies. Savolainen (1997) hypothesis that decommodification of labor tames the harsh anomic state caused by the free market economy tested the effects of both the proxy decommodification index and welfare expenditures on homicide rates disaggregated by gender. Overall, the findings for the male and female homicide rate were substantively the same. Consi stent with institutional the strength of non econom ic normative controls. Thus, Savolainen found support for the moderating effects of both decommodification (explained variance of 32% to 51%) and welfare spending (explained variance 38% 76%) on homicide rates (p. 1032 1034). Point estimates for the co nditional effects of inequality were calculated for Finland (a welfare state) and Mexico (wh ich scored 12.9 on the centered measure of welfare spending in these data). Consistent with IAT, the point estimates for Mexico equaled 4.5 (male homicide rate) and 4.91 (female homicide rate), while the predicted values for Finland were 5.34 and 6.54, respectively. The latter seems counterintuitive, suggesting that income inequality may actually lower the level of homicide, but Savolainen notes that this is co nsistent with the maximin principle theorized in the work
46 the well acceptable, insofar th at it stimulates hard work, innovation, and economic productivity in general, and, by way of tax revenues, improves the situation of the poorest segment 1036). Pratt and Godsey (2003) reported results similar to Savo lainen using cross national sampling and homicide rates as the dependent variable. This study integrated concept of decommodification derived from institutional anom ie theory. Using percentage of gross domestic product allocated to health care expenditures as a measure of social support, and inequality measured as the ratio of the median incomes of the richest to the poorest 20% of citizens, the researchers found tha t the main effects of these key covariates remain statistically significant when controlling for the effects of .359, p < 0.05), indicating that the criminogenic effects of inequality are reduced by high l evels of public social support (Pratt and Godsey, 2003, p. 626). This moderating effect is consistent with the predictions of both theories. Moreover, this relationship holds under different sampling methods and different modeling conditions to test for spurious relationships. Pratt and support are capable of producing a concomitant reduction in crime rates even in the absence of a social and economic revolution, and th at higher levels of social support can In a similar vein to the preceding examinations of the effects of
47 decommodification on homicide rates cross nationally (Messner and Rose nfeld, 1997; Pratt and Godsey, 2003; Savolainen, 2000), Batton and Jensen (2002) examined the effects of decommodification on homicide rates sub nationally using United States data. Using time series analysis, this study analyzed homicide rates for 1900 t o 1997 and the effects of a time series decommodification proxy measure, controlling for inflation and value of the U.S. dollar (p. 16). This study is unique among the empirical tests of institutional anomie theory in that it examines historical trends an of control the state exercises over the economy and the extent to which it attempts to While it has been noted that tests of institutional anomie theory should assess the relationship between economic and non economic institutions cross nationally Batton and Jensen rejoin, might facilitate a more in depth discussion of the decommodification homicide relationship by allowing f or comparisons between rural and urban areas, which often experience shifts in market forces differently because of employment in difference segments of the labor force (2002, p. 14). Additionally, changes in the relations between capital and labor over time should be related to changes in the crime rate over time both across and within social collectives. the effects of decommodification on homicide rates over temporal varia tion. For the
48 period 1900 to 1945, their results generally support institutional anomie theory; however, these results do not carry over to the latter period 1946 1997, which they manifests in the post WWII period (p. 28 29). To date this study has not been replicated in other social aggregates or other historical periods and remains the only time series treatment of IAT. This is a serious limitation in the evaluation of this theory since Messner and Rosenfeld acknowledge the value of Several other empirical tests of institutional anomie theory have focused on within nation analysis. Hannon and De fronzo (1998) examined public assistance as a counties. Their analyses explain 71% of the variation in overall crime rates, 68% of property crime, and 63% of violent crime (p. 387). The researchers interpret these high level of explained variance as both support for traditional anomie/strain theories, such criminogenic frustratio anomie theory. Welfare allows people to maintain a standard of living free of the market, thus lowering levels of anomie and allowing social institutions to provide their normative functions. Stucky (2003) focused on on theory the polity -through an integration of institutional anomie and social representation, responsiveness of the ci ty government to organized groups within the
49 1115). Direct and conditional effects were found, such that district based political representation and cities with Africa n American mayors reduced crime, while the effects of poverty and family disruption were moderated by the strength of mayor council type governmental structures. Stucky interprets this as both support for social disorganization theory and Messner and Rose found in institutional anomie theory. A recent study by Cancino, Varano, Schafer, and Enriquez (2007) continued this vein of theoretical integration by combining tenets of social disorganization theory with IAT in the context of the systemic network thesis. Their study focused on community level characteristics in a predominantly Latino urban area on the outcome variables of both violent and property crime rates. Cancino et al. found support for both social di Maume and Lee (2003) assessed the relationship between social institutions and crime and a sub nati onal level, with counties treated as cases. The researchers examined both instrumental (profit motivated) and expressive homicides, hypothesizing that institutional anomie will better explain the former (p. 1145). It is noteworthy that this is the first anomie theory to look all four institutions identified by Messner and Rosenfeld economy, polity, family, and education; along with religious adherence (as first suggested by Chamlin and Cochran, 1995), and political market restraint (Messner and
50 Rosenfeld, 1997). Inequality (Gini coefficient) was used to measure economic dominance; polity is operationalized as voter participation, family disintegration by divorce rates. Educational exp enditures and adherence to religious denominations community variation in support for non economic was added to the model. While prior r esearch suggests that non economic institutions a divergence, suggesting another possible causal link mediation (p. 1147). that for societies where the economy and penetrate non economic institutions ( 2007, p. 76 83). In an attempt to integrate institutional institutions are restricted by the economic base in the degree to which they can shift imbalance [is] a logical result of a capitalist ec rect impact on how other institutions function and are able to effectively control criminal non economic institutions, undermined by the dominance of the economy, direct ly affect, or partially mediate the effect of the economy on, crime rates (p. 1149). Where prior tests of IAT have found moderating effects, whereby the effects of
51 the non economic institutions on crime rates is dependent upon the level of economic domina effects of the non economic institutions on the economic crime relationship. Under causal chai n is in effect. The introduction of the non economic institution variables into the negative binomial regression model reduces the criminogenic effects of the economic dominance proxy (Gini coefficient). Only one measure, the decommodification proxy, exh ibits both mediation and moderation. Lastly, the researchers found their model to particularly salient for profit motivated homicides; the inclusion of the non economic measures reduced the direct effects of economic dominance on homicide rates by 43.2%. Following prior sub national studies, Kim and Pridemore (2005a), using regional data, applied institutional anomie to transitional Russia. They sought to explain the ee market and cross anomie theo ry and consistent with the Durkheimian version of anomie theory, Kim and Pridemore suggest that we economic] institutions on any association between social change and crime move toward capitalism, coupled with the likelihood that its citizens have begun to at
52 rising crime rates. Kim and Pridemore operationalized institutional anomie variables similar to prior research (Chamlin and Cochran, 1995; Maume and Lee, 2003), creating an economic strength factor, along with family stability, educational strength, and political involvement. Using negative binomial regression techniques, homici de rates were regressed on poverty, social institutions, and socioeconomic change variables. Following what has become convention in IAT studies, interaction terms were also included in the model. The researchers found support for the Durkheimian social deregulation model as suggested by Bernberg (2002); regions experiencing the worst effects of rapid socioeconomic change also reported the highest homicide rates (Kim and Pridemore, 2005, p. 1391). They did not, however, find support for the hypot hesis that non economic institutions moderate the effects of the dominant economy, with none of the interaction terms reaching statistical significance. This may suggest that institutional ential conditioning effects of social institutions simply may be overwhelmed because the researchers tested IAT to account for serious property crime in Russia (Kim and Prid emore, 2005b), but again found support for a Durkheimian interpretation of anomie, rather than IAT. The study that most closely informs the current proposal is Schoepfer and anomie theory on white
53 an illegal act or series of illegal acts committed by nonphysical means and by concealment or guile to obtain money or property, to avoid th e payment or loss of money or property, or to obtain bu siness or personal advantage, the researchers hypothesized that white organizational gain and therefore are well suited to be studied within the IAT anomie theory suggests a cultural ethos whereby it is always possible to obtain more material wealth; the researchers suggest that this speaks to the applicability of institutional motivations niform Crime Report data, they looked at a single offense previous research that puts forward the opinion that this par offenders has much in common with street offenders (p. 231). While generally supportive of institutional anomie theory, Schoepfer and Piquero found the effects of unemployment to be significant predictors of embezzlement, but in t he opposite direction than that hypothesized by IAT. This is intuitive. The v e ry nature of the offense studied, embezzlement, presupposes that the offender be employed. The researchers caution that when examining white collar offenses, economic variabl es such as unemployment should be handled as opportunity controls rather than economic disadvantage measures (p. 233). Overall, despite limitations such as measurement issues of both the independent and dependent va riables, coupled with the previously discussed difficulty in operationalizing IAT
54 measure in general, this study fills a gap in the literature similar to the study proposed herein. Alternate c onfiguration s of i nstitutional i mbalance The overarching trend in the aforementioned research indicates that economic dominance, coupled with enfeebled non economic institutions incapable of exerting their normative controls leads to anomie, which, in turn, results in high crime rates across social collectives. Messn institutional anomie theory is that institutional imbalance per se and not simply high crime rates. As previously t displays these confi gurations. Table 2: Crime types by dominant institution and cultural conditions Dominant institution Cultural condition Predicted crime Economy Anomie Instrumental crime Civil (i.e. religion, ethnic, kinship) State Moral cynicism/withdrawal Corruption In societies with strong civil institutions, such as kinship systems, ethnic or hypermoralism 156, emphasis in original). Under this configuration, societal members develop a strong sense of obligation to those with and the victim of, what Messner and Rosenfel
55 or may not be condemned by codified law -repressions of personal freedoms, hate crimes, violations of human rights, vigilantism, ethnic cleansing, and the like. This institutional arrangement and the resultant cul defense of the moral order itself, albeit understood in narrow and highly particularistic Rosenfeld suggests this articulation of IAT may provide an explanation for terrorism: Terrorism is also nouris hed under the dominance of the primordial institutions in a world in which contrasting institutional arrangements have been on the ascendance for two centuries...religious and ethnic solidarity, real and imagined, figures predominantly as a protest ideal a gainst the rationalism and universalism of the modern world. (2004, p. 25) The third possible configuration is that by which the institutional balance of power tips in favor of the state. Under this arrangement, the emphasis is on the dominance may manifest in ways similar to those discussed earlier, namely devaluation of non political institutions, accommodations to the state by other institutional bodies, and penetration of political or i deological norms into these entities. A historical archetype of this state dominant society may be exemplified by the Soviet Union. Under this state dominant arrangement, government control is infused into d, 1997). During the totalitarian rule
56 asunder or subordinated to the needs of the state. The traditional family was devalued Any cohabita tion could be considered a family, abortions were available on demand, and divorces were obtained simply by requesting one, sometimes by the low birth rate began to cause a signif icant reduction in military conscripts that media propaganda began to extol marriage as a duty to the state. Abortions were outlawed, and divorce made prohibitively expensive. The Russian Orthodox Church embracing ide underground secret services. Muslims in the Soviet Union were forced to accommodate (under threat of arrest) to the demands of t he state, phasing out sharia Penetration of Soviet ideology was apparent in all realms of social life, especially in the schools. Soviet schools taught both political and technical curriculums in order to party Other historic and contemporary examples of state dominance include: state control of the birth rate (China), economic statism (Ita lian fascism), single party of Syria), and nationalized education (Cuba), while other nation states may also be thought to lie somewhere on an autocracy democracy continuum. an ever expanding role in regulating everyday life, the opportunities for the exercise of personal agency are diminished...a sense of direct responsibility for the well being of others accordingly
57 ch as altruism and empathy serving crimes of power (Rosenfeld, 2004, p. 25). What eac h of these institutional configurations lacks is the balance that the and Rosenf eld, 2004, p. 156). The influence of strong families and educational institutions as a buffer against state dominance is intuitive; each provides an escape and a refuge from the repression of a dominant government. In addition, both provide a socializati on function. These mechanisms are similar to the normative influences of these institutions in an economically dominated society, as described in the conventional reading of IAT. The possible moralizing effect of the third institution the market, is less obvious and warrants further discussion. Markets, Morality and Crime The topic of markets and morality can be framed under two opposing viewpoints described by Hirschman (1992, cited in Rosenfeld and Messner, 1997). Under what Hirschman terms the self destruction thesis capitalist markets erode (Rosenfeld and Messner, 1997, p. 209). Th is thesis, found in the extant tests of IAT where a dominant economy is seen as criminogenic, is theoretically informed by
58 Marxist and radical accounts of capitalism. Under this arrangement, traditional social the replacement of substitute values save for efficiency and profit (p. 209). In this economic system, individuals become egoistic. According to Bonger, this society is ividuals by weakening the bond that unites becomes the organizing (Curr ie, 1991, p. 255 ) that is the basis for the conventional reading of IAT as an explanation of instrumental crime. For the proposed study, however; it is necessary to consider, an opposing viewpoint, what Hirschman terms the doux commerce thesis Based upon liberal political and economic theory of the enlightenment, and found in the writings of David Hume and Adam Smith, under the doux commerce principle the market functions as a configuration: market involvement promotes personal attributes such as thrift, industriousness, honesty, and reliability, behaviors that are required to sustain the central organizing principle of markets: reciprocity. (Rosenfeld and Messner, 1997, p. 20 8) Historian E.P. Thompson (1963, cited in Karstedt and Farrall, 2006) similarly justice, roles and rules, and shared notions of acceptable behaviours [ sic. ], profits and ent
59 individualists with a strong self interest in preserving mutually beneficial, cooper ative, anc[ing] the 209).
60 Chapter Four Data and Methods is that institutional imbalance per se and not simply domin p. 155, emphasis in original) is responsible for high crime rates. As previously discussed, (Messner and Rosenfeld, 2004, p. 96; See Table 2, this thesis). Messner and Rosenfeld propose that, in societies where the state dominates the other institutions, high levels of corruption can be expected, leading to the first hypothesis: H 1 : The level of state dominance (defined herein as the absence of popular control over the state) is positively related to corruption cross nationally. In their initial partial test of IAT, Chamlin and Cochran (1995) posed that institutional anomie theory hypothesizes t hat a strong free market economy, coupled with the weakened ability of the non economic social sector to fill a normative role, leads to an interplay between economic and other social institutions determines the level of Thus, a conditional relationship is expected, such that: H 2 : The effects of non polity institutions will moderate the criminogen ic effects of
61 state dominance. This moderation hypothesis has been supported in the literature (Hannon and DeFronzo, 1998; Messner and Rosenfeld, 1997; Piquero and Piquero, 1998; Savolainen, 2000, Schoepfer and Piquero, 2006; Stucky, 2003); however, in the ir sub national test of IAT, Maume and Lee (2003) suggest another possible causal link mediation. According y will continue to have a direct impact on how other institutions function and are able to effectively control criminal mediation hypothesis, which has also has garnered prelim inary support at the cross national level (Bjerregaard and Cochran, forthcoming, Trent, 2007). This leads to a third hypothesis: H 3 : The effects of non polity institutions will mediate the criminogenic effects of state dominance. Finally, although often theory contains a cultural element in addition to its structural component. In the alternate, state dominant configuration of IAT as tested in the current study, the dominant cultural ethos of the state dominated society is not one of anomie, but rather one of moral cynicism (Rosenfeld, 2004). Under this arrangement, individuals eschew personal responsibility and are characterized by marked interpersonal distrust. At the aggregate level, this cultural o rientation is conducive to high levels of corruption. Therefore,
62 H 4 : Nations whose cultural value systems exhibit high levels of moral cynicism will have higher levels of corruption. Design Strategy and Research Purpose Revisited The dominant strategy of both data collection and data analysis is quantitative toward the end of empirically testing hypotheses derived from an alternate configuration of institutional anomie theory. Here the purpose is to confirm (or refute) this alternate configuration of M relationships between its theoretical constructs and corruption cross nationally. Specifically, what is the relationship between state dominance and corruption at the cross national level of analysis? Do the effects of the economy, the family, and/or education attenuate that relationship, or do they interact? Is there a relationship between the cultural value system of a nation, measured as an aggregate response to an opinion survey, and the level of corr uption in that nation? Methodological considerations in cross national research Extant research in comparative criminology falls into three general n quantitative analyse s (Howard et. al., 2000, Neapolitan, 1999). The first type of comparative research is an in depth case study, generally macro level, of one nation. While critics have claimed that this type of research is not comparative, a methodologically sound case st udy both increases the knowledge base of the discipline (Bennett, 1997; Howard et. al., 2000; Flyvberg, 2006) and provides a potential for future replication with positive, negative, and/or non conforming cases (Brady and Collier,
63 2004). A majority of pub lished criminological research on corruption falls into this single nation category (e.g. Brovkin, 2003; Green, 2005; Lu and Gunnison, 2003; Markovskaya, Pridemore, and Nakajima, 2003; Rodgers, 2006). Parallel studies such as topical comparison (e.g. Neap olitan, 1994; 1999) provide an opportunity to compare and contrast crime rates and/or criminal justice systems across nations. As pointed out by Howard et. al. (2000), this methodology allows for comparisons that are more meaningful, overcoming many of th e difficulties resulting from the lack of comprehensive macro social data. The researcher is able to account for and better explain differences in police reporting, local historical and cultural differences, and political and economic structures, although generalizability of results diminishes. Criminologists have used this methodology to examine corruption victimization in Latin America (Seligson, 2006), corruption in post communist countries (Karklins, 2002; 2005; Karstedt, 2003; Saj, 2003), and civil versus common law responses to corruption (De Sousa, 2002). Multinational quantitative methods use the nation as the unit of analysis, examining theoretically derived correlates of crime such as income inequality, social disorganization measures, or modern ization, usually through multivariate regression analysis. The advantage of this method is its ability to test hypotheses, as well as to identify patterns and trends across different societies (Howard, et. al, 2000; Neapolitan, 1999). A majority of studi es use homicide rates as the dependent variable, and with good reason; compared to other offenses, the availability and relative validity and reliability of the data is greatest for homicide (e.g. Chamlin and Cochran, 2005; Gartner,
64 1990; LaFree, 1999; LaF ree and Drass, 2002; Messner, 1982; Messner and Rosenfeld, 2007; Neapolitan, 1994, 1999; Pampel and Gartner, 1995; Savolainen, 2000). Others have looked at other violent and/or property crimes (e.g. Fajnzylber, Lederman and Loayza, 2002; Krohn, 1978; Kroh n, and Wellford, 1977; Jensen, 2002; Messner and Rosenfeld, 2007; Neapolitan, 2003; Stack, 1984). While growing in the political science, corruption cross nationally are li mited in the criminological literature (for exceptions see Sung, 2004; Xin and Rudel, 2004), or restricted to descriptive analyses. As the purpose of the proposed study is to test hypotheses at the cross national level, variable n on analysis is the most appropriate research strategy. This method is not without limitations. The operationalization of highly abstract theoretical constructs reduces these social processes simply to a collection of (Goldthorpe, 1997, p. 2). This simplification tends to ignore the causal processes and the historical, cultural, social, and economic antecedents and trends that contribute to a This drawback is noted; however, as this study is exploratory in nature, multivariate analysis retaining the largest number of cases is most suitable to the task of evaluating theory. Sample The appropriate level of analysis for the proposed study is t hat of the nation state. The universe is all legally sovereign and independent nation states. For quantitative analysis, ideally, a random sample would be drawn from the population of
65 approximately 193 3 entities recognized by international bodies; howeve r, practical concerns influence the sample used the in the final analysis. Perhaps the bane of cross national and macro social research remains the limited availability of reliable comparative data across theoretical constructs (Bennett, 1980; Bennett and Lynch, 1990; Boyle, 2000; Howard et. al., 2000; Karstedt, 2001; Krohn, 1978; Krohn and Wellford, 1997; Marenin, 1997; Neapolitan, 1997, 1999, 2003). The researcher often cannot efficiently and feasibly collect all of the desired data due to cost, languag e barriers, and difficultly and resistance from the national stakeholders. Additionally, these data are rarely compiled systematically within respective nation states. This necessitates the use of secondary data. Following Bennett and Lynch (1997), the sample for the quantitative analysis in the proposed study will be limited to nations available in existing data sets with complete data across all key variables. Extant empirical assessments of institutional anomie theory, where the level of analysis i s the nation state, have used samples ranging from 33 to 84. Cases were included because either they had complete data across all variables or imputation methods were utilized to maximize the amount of nations available for analysis (Chamlin and Cochran, 2006, 2007; Jensen, 2002; Messner and Rosenfeld, 1997; Pratt and Godsey, 2003; Savolainen, 2000). In the present study, data are obtained primarily 3 Based in part upon the list of countries and territories assigned a three digit International Standard ISO 3166 1 Code for the representation of names of countries and their subdivisions as used by the Statistics Division of the United Nations (see http ://unstats.un.org/unsd/methods/m49/m49alpha.htm), nations recognized by the United States Department of State (see http://www.state.gov/s/inr/rls/4250.htm ), and The World Factbook (see https://www.cia.gov/library/publications/the world factbook/index.html)
66 from the Shared Global Indicators Cross National Database (Norris, 2008), which compiles data on over 700 i ndicators from several published sources for 191 nations. Full variable descriptions and sources of data appear in Appendix A. Due to missing values, the current study includes a maximum of 125 nations. Dependent variable In the study of crimes of the powerful, there are innate practical limitations in the choice of definitions that can be applied and the data that can be used. The hidden nature of corruption makes this crime inherently difficult to study (Karklins, 2002, Xin and Rudel, 2004; Zi mring and Johnson, 2005). Initially, the United Nations Surveys on Crime Trends and the Operations of Criminal Justice Systems officially recorded cross national corruption and bribery rates appeared promising 4 (Burnham, 1998) Ten waves of annual data ha ve been collected by the United Nations Crime and Justice Information Network (UNCJIN) since 1975. Data for these surveys are collected via voluntary surveys sent to both member and non member nation states. Each survey year contains official recorded in stances of crime disaggregated by offense. Relevant to the current study, this dataset includes bribery rates, defined as: Bribery and/or corruption may be understood to mean requesting and/or accepting material or personal benefits, or the promise thereo f, in connection with the performance of a public function for an action that may or may not be a violation of law and/or promising as well as giving material or personal benefits 4 In a white paper published by the U.S. Department of Justice, Barnett (2002) states that bribery is the most reported and cleared white collar offense using NIBRS; at the cross national level, however; this indicator appears quite unreliable for comp arative analysis.
67 to a public officer in exchange for a requested favor. (United Nations, n.d ., p. 5 6) Following Bennett and Lynch (1990), it is instructive to note that, while the use of official crime rates may be deceptive in evaluating the exact amount of crime in any given nation, it has been found reliable in determining and comparing tren ds across social collectives. We should be able, then, to assess the efficacy of institutional anomie theory cross nationally by analyzing these trends and their relationship to a bribery rates and to draw reasonable conclusions. Preliminary analyses u sing the United Nations dataset proved both instructive and disappointing. The sixth (1995 1997) and seventh waves (1998 2000) of the United Nations Survey were used; to control for annual fluctuations, the average bribery rate was computed from the annua l recorded data. Only twenty nine of the ninety two total respondent nations had complete data across all IAT indicators, an attrition rate of almost 68%. As crime rates are rare events and the dependent variable was overdispersed, effects were estimated using Poisson based negative binomial regression analysis (Osgood, 2000). When regressed on the cultural and institutional variables identified by institutional anomie theory (to be discussed shortly), the direct effects model as a whole was not signific df ) = 9.43(5), p > 0.05). Individual coefficients and incidence rates ratios (IRR) are shown in Table 3.
68 Table 3: Negative binomial regression estimates for bribery rates, cross national data ( N = 29), 1995 2000 Variable Coef. S.E. I.R.R. S.E. State dominance .34 .37 .71 .27 Industrialization 1.04 .61 .35 .21 Family disruption .04 .04 1.0 .04 Education .35 .42 1.4 .60 Moral cynicism .90 .39* .41 .15* Pseudo R square .069 Log likelihood 63.44 9.43(5) p < 0.05 Only one predictor reaches statistical significance in this model. The cultural condition of moral cynicism is positively related to bribery rates, as suggested by theory. This should be approached with caution; since the model itself i s not significan t, this relationship is likely significant purely by chance (McClendon, 2002, p. 172). The reliability and validity of cross national crime statistics is often suspect (Vigderhous, 1978). This is due to problems at the level of data collection, rather th an methodological concerns (Bennett and Lynch, 1990; Reichel, 1999). Based upon this preliminary analysis, United Nations data are inappropriate for the present study, as the measure of official bribery rates cross nationally provides little confidence in the results they simply do not allow the researcher to adequately answer the research questions and test the alternative configuration of IAT. Furthermore, it is unclear if we are measuring incidences of corrupt acts, the quality of law enforcement, cour ts and prosecutors, and/or anti corruption policies. whereby many acts of official malfeasance go unreported and undiscovered. Only some crimes are reported to the police who, in turn, may only record a portion of incidents.
69 Furthermore, when looking at the offense of bribery, the police themselves are often the offender. At the level of grand or systemic corruption, a single, identifiable victim may not even exist. Erro rs may be aggravated by (1) definitional discrepancies, or the legal definition of bribery in differing jurisdictions (Huang and Wellford, 1989); (2) differing levels of reporting crime to the police or different traditions of policing. For example, in na tions where the state is dominant, residents may have little or no confidence in the authorities and/or fear retribution by the very nature of the offense (for example, bribery) being studied (Marenin, 1997; Rosenfeld, 2004). Unfortunately, these nations are the most interesting cases. Finally, (3) different socio economic and political contexts could lead to different levels of reporting and recording (i.e., poorer nations may not have the available resources to adequately and systematically collect vali An alternate to official incidence rates exists for offenses of corruption and bribery corruption perceptions in dices. Both the World Bank Group (Kaufmann, Kraay, and Mastruzzi, 2003) and Transparency International (TI) provide indicators of corruption perceptions cross nationally based upon multi source surveys of international organizations, think tanks, citizens non governmental organizations (NGOs), and business risk analysts from both within and outside individual nation states. overall climate of corruption in a given nation a s perceived by those stakeholders who
7 0 Jackman, 2002, p. 156). Both the World Bank Worldwide Governance Indicator (WGI) Control of Corruption measure and Transparency Int ernational Corruption Perceptions Index (TI CPI) are considered valid and reliable indicators of a nation corruption, measured as perceived levels (Kaufmann, Kraay and Mastruzzi, 2007; Sandholtz and Koetzle, 2000; Treisman, 2000; Xin and R udel, 2004). As the level of corruption varies widely not only across, but within nations, it has been suggested that measuring the true magnitude of corruption, especially in large, diverse countries Ackerman, 1 997, p. 31). Furthermore, as suggested by Meny (1996), due to the clandestine nature of corruption, frequency perception and feeling 310, emphasis added). Thus, the use of corruption perception measures, such as those published by the World Bank and Transparency International, are likely the best measure to gauge the level of corruption in a nation for large n comparat ive, cross national hypothesis testing. Although the World Bank Corruption Control measure and Transparency International Corruption Perceptions Index are compiled from different sources, the two are highly correlated ( r = .98) and thus the latter is cho sen for the current study based upon the nations ( N = 180 maximum) and years recorded. Transparency International is a non partisan, global civil society organization geared toward measuring corruption cross nationally and propagating the anti corruption Corruption Perceptions Index (TI CPI) is a composite index of survey data, the first
71 systematic attempt to measure corruption cross nationally using the same metric for each nation. The most recent CPI (Lambsdorff, 2007) incl udes 14 sources, originating from 12 independent institutions, and covers 180 nations. For a country to be included, it must feature at least three polls. Poll inclusion is contingent upon the following: (1) the source must include a ranking of nations f or comparative purposes, and (2) the poll must not include other issues such as political instability or decentralization. Overall correlation between the respective sources is, on average, 0.77 indicating that the individual assessments do not differ sub stantively ( f or full methodology of the CPI, please consult Lambsdorff, 1998; 2007). Transparency International has been criticized on its methodology for its use of a bootstrapping approach when the number of sources for a nation is small, which can re sult in spuriously small amounts of variance and exaggerated estimates (Kaufmann, Kraay, and Mastruzzi, 2003). Saj has been critical of indices measuring corruption (Sa Additionally, Saj argues that perception indices label a whole society, creating the Criticisms notwithstanding, the TI CPI has been used in prior studies when the purpose is comparative research (Goldsmith, 1999; Karstedt, 2003; 2007; Sung, 2004;
72 purpose, it remains the most appropriate index. The TI CPI provides a large number of observations from nations at varying stages of development ac ross several years, necessary for testing the hypotheses at hand. For the present study, the dependent CPI score, reverse coded so that higher scores indicate possibility of high profile scandals skewing results, annual scores are averaged for the years 2000 2005. Independent variables Among the greatest challenges to empirical tests of institutional anomie theory, is the difficulty in operationalizing its highly abstract theoretical claims (Messner, 2003; Messner and Rosenfeld, 2004; 2006). As IAT is cast at a high level of abstraction, the use interpretations. Such concepts can be rich in meaning, but they defy easy IAT is sensitive to the operationalization of its key variables (Piquero and Piquero, 1998). This challenge is magnified in cr oss [must be employed] to ensure that the variables reflect the same concepts in [different] from existing tests of IA T and cross national corruption studies (when possible). In the economic dominant/anomic configuration of IAT, Messner and Rosenfeld created in societies where, within the culture of the collective, there is malintegration between the
73 emphasis on the pursuit of goals and the available normative means to achieve these goals: d the prospect of accumulating wealth and on the other, they are largely denied effective opportunities to an achievement orientation, indiv idualism, universalism, and a 7). Operationalizing and quantifying this cultural ethos has proven most difficult, and, as a result, it is often omitted from the bulk of tests of institutional anomie theory. Sever al researchers have, however, used theoretical framework ( See Cao, 2004; Chamlin and Cochran, 2007; Cullen, Parboteeah, and Hoegl, 2004; Jensen, 2002). In the present test of the state dominant configuration of IAT, the dominant cultural ethos is not a malintegration between goals and means, but rather between state regulation at the structural level and personal agency and direct responsibility in the citizenry. Rath er than a manifestation of anomie, in the current study the researcher must operationalize and quantify moral cynicism at the aggregate level. First this concept must be defined. Based upon differing dominant institutions and resultant cultural condition s, Messner and Rosenfeld suggest that the varying forms of institutional imbalance are possible; each believed to produce diverse offense types (2001; 2004; Rosenfeld, 2004; Refer to Table 2, this thesis). The authors, however, do not provide any guidanc e on the
74 Referring back to the theoretical roots of anomie Durkheim claims that moral feelings toward others are strongest in social groups whe n individual members are bound by attachment and commitment to, and involvement in the group (1961). In his feed the trust, responsibility, and sense of morality on which rest interpersonal trust, and social group membership are the hallmark of well functioning societies, then, as Hearn notes, malfunctioning societies are those marked by distrust, In his ecological theory of deviant places, Rodney Stark proposes that densely populated neighbo trust amon gst members of the collective. According to Braithwaite, effective social control relies on communitarian interdependencies. Once these bonds between
75 vacuum is created t hat will attract the most brutal, repressive, and intrusive of police The writings of Durkheim, Hearn, and Braithwaite suggest that the direction of causality is from moral cynicism to a repressive polity, while Jacobs (2002) notes that discussion of the alternative configurations of IAT, Messner and Rosenfeld speculate that, in state dominant societies, societal members are characterized by moral citizenry, suggesting a concomitant relationship between moral cynicism and state repression (Rosenfeld, 2004, p. 25). While the direction of causality and poss ible feedback loops cannot be estimated in the current cross sectional study, from these speculated relationships, an operational definition of moral cynicism can be formed. For this study, the dominant cultural orientation of moral cynicism is understood as an aggregate of atomized, distrustful individuals, withdrawn from the collective anti communitarian, and anti altruistic, and devoid of personal civic responsibility. In the tradition of prior research that has attempted to operationalize and meas ure culture at the aggregate level (Cao, 2004; Chamlin and Cochran, 2007; Cullen, Parboteeah, and Hoegl, 2004; Jensen, 2002), the current study uses the World Values Survey, an instrument that endeavors to capture individual sociological, cultural, and pol itical values and opinions cross nationally. The World Values Survey (originally the European Values Study, cited a such), has been conducted in four waves since 1981 and consists of about 250 questions yielding more than 800 variables that capture indivi dual
76 sociological, cultural, and political values of residents of a maximum of 90 nations. An average of 1300 face to face interviews is conducted in each country by a network of social scientists (full methodology is available at www.worldvaluessurvey.or g). For the present study the 2000 World Values Survey is used; aggregated responses ( see Liska, distrust (Messner and Rosenfeld, 2001, p. 155). The first indicator is t he total number of voluntary organizations (including church or religious, sport or recreational, art, music or educational, environmental organization, professional association, and/or charitable organization) in which individuals report membership. The mean value was calculated for each nation and signifie d views on personal responsibility a nd agency based upon the aggregated mean response to the following: government should take more responsibility to ensure that everyone is provided d take more responsibility to provide for Additionally, Stark posits that high levels of moral cynicism will be accompanied by correspondingly high levels of interpersonal distrust Rooted in social capital theory (Putnam, 2000), interperson al distrust has been linked to corruption in a sample of Latin American (Morris, 2004) and post Soviet nation states (Fish, 2003). For the current
77 study, the lack of social trust is captured by the response to the following question, aggregated to the mea n level for each country: values indicate distrust) (European Values Study Group, 2004). In order to preserve degrees of freedom, factor analysis is used in order to reduce these three measures into a single variable using principal components analyses (PCA). The voluntary organization and personal responsibility and agency variables load on the same dimension (factor loading = 0.858; Eigenvalue = 1.47), and explain 73.7% percent of the variance. This measure also exhibits a high degree of internal consistency coded (*( 1)) so that higher values indicate higher lev els of moral cynicism The measure of interpersonal distrust does not appear to tap into the same latent construct as the previous two measures and will be entered into the regression analysis separately State Dominance While Crime and the American Dre am focused on an institutional arrangement dominated by the economy, Messner and Rosenfeld also envision another possible orruption and related forms of manipulation study, in order to conceptualize state dominance, an appropriate definition of its antithesis is first supplied. This st polyarchy or
78 pluralism (rather than democracy, which Dahl regards as an unachievable ideal). According to Dahl, a polyarchy consists of [C]ontrol over government decisions about policy is constitutionally vested in elec ted officials; (2) elected officials are chosen in frequently and fairly conducted elections in which coercion is comparatively all adults have the right to run for elective office right to express themselves without the danger of severe punishment on political matters broadly defined; (6) citizens have the right to seek out alternative sources of information. Moreover, alternative sources of information exist right to form relatively independent associations or organizations, including independent political parties and interest groups (1982, p. 11). State dominance is thus operationalized by first constructin g a multi dimensional, multi source polyarchy scale and then reversing to a one to five scale of state dominance. This variable then is essentially the lack As the concept of polyarchy is multi dime nsional; multiple measures are used to tap into this complex construct. The first set of variables is from the Governance Indicators Index published by the World Bank Group (Kaufmann, Kraay, and Mastruzzi, 2003; 2007). This perception based, or subjectiv e, set of aggregate indicators has been collected annually since its inception in 1996; the latest survey in 2007 is based upon
79 311 variables from 33 different sources and covers 212 nations. Survey respondents include NGOs and commercial risk analysts, m ultilateral development agencies, and internal and external observers. The survey employs an Unobserved Component Model (UCM) to aggregate responses into six dimensions: voice and accountability, political stability and absence of violence, government eff ectiveness, regulatory quality, rule of law, and control of corruption (Kaufmann et. al., 2007). voice and accountability regulator y quality, and rule of law The first captures the extent to which citizens are able to select their leaders and participate in the political process. This measure falls in line basic tenets of public choice/principal agent theory, such that, in closed polities, principals lack the ability to monitor the agents, providing an opportunity structure conducive to corrupt practices (Jain, 1998; Rose Ackerman, 1978). market unfrie ndly policies such as price controls Low levels of regulatory quality are indicative of policies of economic statism or state interference in the free market. for hones length or objective and unbiased, relationships between government officials and private sector i See also Ades and DiTella, 1997; 2000; Gerring and Thacker, 2005; Sandholtz and Koetzle, 2000).
80 predictable rules form the basis fo 2003, p. 4). Multi year averages spanning five years of annual data (2000 2005) on each governance indicator are used in the current study. Two measures published by Freedom House (2004) are employed i n the current study. The Freedom House Civil Liberties and Political Rights Index is comprised of a 1 to 7 scale (free to not free) and is based upon the United Nations 1948 Universal Declaration of Human Rights. The social activist school of thought sug gests that the absence of civil liberties, such as freedom of expression, associational and organizational rights, and personal autonomy, greatly increases political corruption (Kaufmann, 1998). The Freedom of the Press index is a 1 to 100 scale that rat es nations based upon the legal environment for reporting and political pressures on journalists. Both indices are based upon in depth nation summaries compiled by a team of regional experts and scholars; historical background and current events are also included in the overview of the country. One hundred and ninety two nation states and 14 territories are included in the final index. In the current study, both indices are reverse coded for interpretability, so that higher scores equal greater freedom. Finally, two additional composite indices of democracy are used. The Vanhanen Index (VI) covers the period 1810 through 1998. The VI is an objective index that
81 combines two basic dimensions of democracy competition (% seats occupied by dominant party) and participation (electoral turnout) (2000). These measures are aggregated using the following formula: C(ompetition)*P(articipation)/100 y. Competitive government reduces corruption since party turnover minimizes opportunities, while a larger selectorate yields greater transparency amongst the politicians (Montinola and Jackman, 2002, p. 153 154) The Cheibub standardized scale is bas fold classification of regime type (monarchic dictatorship, military dictatorship, civilian dictatorship, presidential democracy, mixed democracy, parliamentary democracy). This scale covers the years 1946 2002; higher sco res designate greater popular rule (2004). Rose Ackerman (2001) has found that corruption is correlated with the constitutional structure of a given country, ranking parliamentary systems as best at the avoidance of corruption and party centered president ial (presidential democracy)as worst. Nations under dictatorships are often predatory kleptocracies, where economies become wealth building mechanisms for those in power, at the expense of the populace (Jain, 1998). In order to preserve degrees of freedom the polity measures described above are factored into a single polyarchy variable using principal components analysis (PCA). The resultant values are then multiplied by ( 1) in order to remove negative scores for interpretability and, finally, one to fi ve scale is created. This scale indicates the level of
82 state dominance, or the lack of popular rule, in a given nation state. Factor loadings, Eigenvalues, and percentage of variance explained are shown in Table 4. Table 4: Principal components analysis for eight dimensions of polyarchy Variable Factor loadings Eigenvalues % of variance Voice and accountability .982 7.19 79.87 Regulatory quality .894 Rule of law .846 Political rights .931 Civil liberties .950 Freedom of the press .930 Vanh anen Index .648 Regime type (Cheibub) .781 Extant research has suggested that the democracy corruption relationship is U (quadratic) or S (cubic) shaped (Montinola and Jackman, 2002; Sung, 2004). As the dominance is, in essence, the inverse of polyarchy (or democracy), it is instructive to assess the functional form that the state dominance corruption relationship takes in the current data. TI CPI scores were plotted against the state dominance scale, and the resultant distribution followed an s shaped, or cubic, pattern as evidenced by the blue fit line (R 2 = 0.819) in Figure 2 ( below ).
83 While hierarchal polynomial regression is the preferred method of estimating the parameters for nonlinear y x relat ionships, the transformation of the non linear x to higher order polynomials results in regression coefficients that become difficult to interpret (Fox, 1991). In the current study, these interpretability issues would be compounded, since the key independ ent variable, state dominance, is a factor score derived from six components, each of which capture a dimension of polyarchy. Finally, as the moderation hypothesis of the current study suggests a non additive relationship, the use of higher order polynomi als would further muddle any clear substantive interpretation of the regression parameters. As the research purpose of the current study is not to assess the functional form of the state dominance corruption Figure 2 : State dominance corruption functional form
84 relationship, but rather to test hypotheses der ived from an alternate configuration of a criminological theory, an alternate strategy is warranted. linear associations is the use of power transformations of x lea st squares regression. The shape and direction of the curvature produced by the non linear x y relationship determines which direction one moves on the power ladder (see Tukey, 1977). Based upon Figure 2 (above), the first segment of the S curve is a con cave U shape, as state dominance and corruption both initially increase. Next, the curve appears to somewhat level out, before again increasing. As the overall curve appears monotonic, the power transformation for the concave U shape is either the natura l log (ln x ) or the negative inverse of x ( 1/ x ) (Fox, 1991). As shown below (Figure 3), the negative inverse ( 1/x) transformation yields the closest approximation to linearity (again, indicated by the blue fit line), with the R squared value of the linea r equation increasing from 58% to 79%. Figure 4 shows the log transformation of x which only increases the linear fit to approximately 71%.
85 Figure 3: State dominance corruption, reverse inverse transformation
86 Figure 4: State dominance corruption, log transformed Economy The role of the economy is to provide for basic human subsistence necessary for humans to adapt to their environment (for example, food, clothing, shelter). A standard strategy in tests of IAT has been to operationalize and measure the strength of the economy based upon an indicator of absolute (Chamlin and Cochran, 1995; Piquero and Piquero, 1998) or relative deprivation (Maume and Lee, 2003; Messner and Rosenfeld, 1997; Pratt and Godsey, 2003; Savolainen, 2000), or unemployment ra tes (Schoepfer and Piquero, 2006). The preferred method uses relative deprivation, typically the Gini coefficient, as this indicator includes reference groups (Passas, 1997), although recent research by Pridemore (2008) suggests that the deprivation crime relationship becomes
87 null when a proxy for poverty (infant mortality) is entered into the equation This finding suggests that reference groups may not be as influential as previously believed. For the study of corruption, the use of inequality as an ind icator of the economy is problematic. One of the most damaging consequences of corruption is the accumulation of wealth to the (already) privileged at the expense of the poor. Gupta, Davoodi, and Alonso Terme (1998) found that, across 37 nations, corrupt ion and the Gini coefficient were positively related, net the effects of other exogenous variables. Additionally, this impact remains significant when controlling for GDP. Other research has questioned the direction of causality of this relationship, sug gesting that inequality and poverty is likely both a cause and a consequence of corrupt governments (Husted, 1999; Lambsdorff, 1999). The cross sectional nature of the current study precludes testing this last hypothesis. Consequently, the proposed study marks a departure from the bulk of prior research by conceptualizing the strength of the economy by the level of industrial development within a nation. Empirical studies suggest that industrialization is a valid indicator of the dominant economic system within a country (Bonger, 1969; Cullen, Parboteeah, and Hoegl, 2004; Esping Anderson, 1990; Krohn, 1978), and prior research has linked industrialization, measured as government subsidies to manufacturing as a percentage of the Gross Domestic Product (GDP) to reduced levels of corruption (Ades and DiTella, 1997). In the current study, industrialization is operationalized via the following proxy variables: commercial energy use (oil equivalent) per capita (kilograms) 1980 2002,
88 electric consumption per capi ta (kilowatt hours) 1980 2001, and the average percentage of population living in urban areas (1990 1998). Commercial energy use measures levels of resources used in the industrial production of goods, as does electric consumption per capita. Urban popul ation is a consequence of industrialization, as citizens flock to city centers for work, citizens who also contribute to an increase the Gross Domestic Product (GDP) (p pp), averaged for the years covered by the current study (2000 2005). Research has shown that poorer countries are, in general, more corrupt than wealthier counties (Andvig et. al, 2000; Mauro, 1995), as they lack the resources to effectively fight corrup tion. Finally, the economic distribution by sector (percent non agrarian) measures the distribution of the domestic product and labor force involved in both the industrial and service sectors. PCA results are shown in Table 5. Table 5: Principal componen ts analysis for five dimensions of industrialization Variable Factor loadings Eigenvalues % of variance % of GDP non agriculture .769 2.79 69.84 Commercial energy use .864 Elec. consumption .838 Urban population (%) .868 GDP (ppp) .789 0. 65 Family Functioning similar to the conventional social bond in micro level control theory
89 time, and for the socialization, nurturance, and care of societal 2003, p. 5). Additionally, the family provides a buffer from the stresses of the other institutional domains (Messner and Rosenfeld, 2007). Empirical tests of IAT have operationalized this concept as family disruption using divorce rat es (Batton and Jensen, 2002; Chamlin and Cochran, 1995; Jensen, 2002; Maume and Lee, 2003; Schoepfer and Piquero, 2006), or the percentage of single parent families (Kim and Pridemore, 2005; Piquero and Piquero, 1998). Divorce rates at the cross national level are limited to a small number of nations making this variable unviable for the current study. Instead, a measure of female economic activity was used to measure family disruption. The current study employs a measure from the United Nations Human Development Report (2003) that reports the percentage of women over the age of fifteen who work outside the home. Admittedly, when compared to divorce and single parenthood, female employment is certainly not as disruptive to the traditional nuclear famil y, yet it may suggest that children are, at least for a period of time during the day, left alone in the home or under the supervision of older family members or non familial child care workers. Education The role of education, similar to the family, is s education has been operationalized as enrolled in college full time (Kim and Pridemore, 2005; Piquero and Piquero, 1998), comparat ive teacher salaries (Piquero and Piquero,
90 1998), percentage of high school dropouts (Piquero and Piquero, 1998; Schoepfer and Piquero, 2006), and/or school expenditures (Maume and Lee, 2003). One study by Mauro (1997) found a significant negative relatio nship between public education expenditures and corruption. The proposed study follows this convention using a factor variable comprised of the combined gross enrollment ratio for primary, secondary, and tertiary schools (2004), and the mean adult litera cy rate (age 15 and above, 2000 2002). Both measures load on a single factor (factor loading = 0.819; Eigenvalue = 1.34), this measure is only constructed from two c omponents. Control variables Known correlates of corruption are included in the regression analysis to control for possible spurious relationships. First, a population heterogeneity factor variable is created, comprised of ethnic fractionalization (Ales ina et. el., 2002), and the Gini coefficient (United Nations, 2004). Studies suggest that ethnic divisions impact the quality of institutions; indeed, ethnic conflict has been associated with poor economic performance and policy and political instability, as well as higher levels of corruption within nations (Alesina et. al, 2002, Mauro, 1995, Morris, 2004; Treisman 2000). The fractionalization variable is computed as one minus the Herfindahl index of ethnolinguistic group shares, and signifies the proba bility that two randomly selected persons from a population will belong to different groups. The theoretical maximum is reached when each individual belongs to a different group (or the value of 1) (Alesina et. al 2002, p. 156). For the current study, th e combined linguistic and racial/ethnic
91 variable from the Norris (2008) dataset is used. The second dimension of population heterogeneity is economic disparity, herein defined as income inequality, as captured by the Gini coefficient. Cited as both a ca use p. 4. See also Goudie and Stasavage, 1998). Based upon the Lorenz curve, the Gini coefficient is defined as a ratio with values between 0 and 1, where 0 denotes perfect equality (everyone earns the same amount) and 1 denotes perfect inequality (one individual holds all of the income) (Sen, 1973). Gini coefficient values from the U nited Nations Human Development Report (2004) are used in this study. Principal components analysis (PCA) indicates that these measures load on a single factor (factor loading = 0.816; Eigenvalue = 1.33), explaining 66.6% of the variance. While the Cro number of components. Regression analysis was performed used an un collapsed version of this measure with no substantive change in the results. The literature has also linked government size to levels of corruption within nation states (Husted, 1999, Morris, 2004; Treisman, 2000; Xin and Rud el, 2004). Larger bureaucracies tend to increase corruption opportunities by the sheer number of officials who hold a monopoly over resources needed by citizens (Husted, 1999). In particular, less developed and developing countries tend to have over expa nded large state sectors when compared to nations in the industrialized world (Theobald, 1990).
92 38). The size of government in the current study is measured as central government expenditures are a percentage of GDP. Five year averages covering the years included in this study (2000 2005) are estimated. To control for historical effects, a dummy v ariable is created to indicate if the strong, negative relationship between British heritage and current levels of perceived corruption a reduction of nearly two poin ts on a ten point corruption index, while no Spanish/Portuguese colonization and corruption (p. 418 419 S ee also Sandholtz and Koetzle, 2000). Treisman credits this negative British heritage corruption relationship to Finally, em pirical analysis has found the Protestant religious ethos to reduce 40. See also Lipset and Lenz, 1999; Treisman, 2000). In the current study, Data for both the former British colony variable and predominantly Protestant are from the Norris dataset (2008).
93 Analytic strategy After a brief examination of the descriptive (univariate) and bivariate statistics, multivariate analysis will be estimated using ordinary least squares regression (OLS). Baseline levels of the strength and direction of corrup tion perceptions regressed on state dominance, net the effects of the control variables, will be estimated. The possible causal link of mediation will be tested by examining the direct effects of all variables in the model. Should the inclusion of the no n polity variables reduce the magnitude of the effects of state dominance, this shall indicate partial mediation; full mediation would occur if state dominance were to become non significant or change its algebraic sign (Baron and Kenny, 1986). In order t o assess the possible moderating effect of the non polity variables, a series of models will be run with the inclusion of cross product interaction terms (e.g. state*economy, state*family, etc.). Diagnostic tests will be performed to determine if any seri ous violations to the assumptions of OLS exist (Fox, 1991). Results will be discussed along with directions for future research and
94 Chapter Five Results The following analysis was conducted in order test the fo llowing four hypotheses (as described in Chapter 4): H 1 : The level of state dominance (defined herein as the absence of popular control over the state) is positively related to corruption cross nationally. H 2 : The effects of non polity institutions will mo derate the criminogenic effects of state dominance. H 3 : The effects of non polity institutions will mediate the criminogenic effects of state dominance. H 4 : Nations whose cultural value systems exhibit high levels of moral cynicism will have higher levels of corruption. Descriptive statistics appear in Table 6 (below); the final column lists zero order moment correlation coefficient ( r )) for the dependent and key independent variables. For the nations included in this study, the average level of corruption is four, based upon the one to ten Transparency International Corruption Perceptions Index (TI CPI). Finland ranks lowest in corruption, while the TI CPI ranks Bangladesh as the most corrupt.
95 Table 6: Descriptive statist ics and selected zero order correlations Variable Mean S.D. Min. Max. r** Dependent variable Corruption 4.03 2.09 9.50 1.22 Independent variables State dominance 2.67 1 1 4.62 Industrialization .02 .99 2.01 3 Family disr uption 53 14.26 29.20 82.60 .13 Education 0 1 .2.25 1.40 .22 Moral cynicism .08 .93 1.79 1.08 .11 Interpersonal distrust 3 1 0 5 Controls Population het. .05 1 1.95 2 Government size 16.05 6.21 5 29.9 Prev. British colony 0* Protestant 0* Dichotomous variable, statistic is mode ** Zero order correlation between theoretical variables and corruption (TI CPI). 5 The average level of state dominance in the nations sampled is 2.67 (based upon the non tra nsformed, one to five state dominance scale). The highest levels of dominance appear to cluster in African (Sudan = 4.62, Zimbabwe = 4.51, Cameroon = 4.04, Guinea = 4.02, Togo = 4.01) and post Soviet nations (Uzbekistan = 4.53, Belarus = 4.40, Tajikistan = 3.96). The lowest levels of state dominance are found in Nordic social democracies, with Denmark, Sweden, and Norway each scoring a one. Zero order correlations between state dominance and corruption perceptions suggest a fairly robust relationship at the bivariate level ( r = 0.78) strength of the economy (operationalized as the level of industrialization within a nation) is highest in the Organization for Economic Co O peration and Development 5 All p values are based upon a one ta iled test.
96 (OECD) nations (e.g. Iceland, Norway, Canada, Sweden, and the United States) and lowest in sub Saharan Africa (e.g. Ethiopia, Tanzania, Togo, Cameroon, and Sudan). Industrialization exhibits a strong, negative relationship with co rruption perceptions ( r = 0.82). In the nations sampled, a little over half of females age fifteen and up work outside the home (53%) economic activity is lowest in Syria (29.2%) and highest in Moza mbique (82.6%). Educational strength, measured as a factor score comprised of literacy rates and gross enrollment, is highest in Spain and lowest in Ethiopia. At the bivariate level, neither the family disruption proxy, nor the education measure, is sign ificantly related to corruption perceptions. Of the two cultural measures included in the analysis, moral cynicism (operationalized as the aggregate number of voluntary organizations to which citizens belong and scores on a personal responsibility and age ncy scale) is lowest in the United States ( 1.79), and highest in former Warsaw Pact and Soviet states (Ukraine = 1.10; Croatia = 0.95, Belarus = 0.87, Lithuania = 0.76, Latvia = 0.70, Hungary = 0.69). This measure does not exhibit a significant relation ship with the corruption perceptions measure. The second cultural measure, interpersonal distrust, is lowest in the Nordic states (Finland, Denmark, Sweden and Norway = 0) and highest in two Latin American nations (Peru and Brazil = 5). Interpersonal dis trust is positively associated with cross national corruption perceptions, and this relationship is fairly strong ( r = 0.65) The control variables indicate that, in this sample, population heterogeneity is lowest in Japan, and highest in Syria, with the d istribution following a normal curve. The
97 average level of central government expenditures as a percentage of GDP (government size proxy) is 16.05%. Only about one forth (24.7%) of the nations sampled are predominantly Protestant, and 33.2% were former c olonies of Great Britain. As two of the structural IAT indicators, as well as the cultural interpersonal distrust measure, are significantly related to corruption at the bivariate level, multivariate analysis is warranted in order to disentangle the uniqu e effects of each when controlling for the other variables in the model. First, the effects of state dominance and the control measures on corruption perceptions will be estimated in order to provide a baseline analysis for comparison. Next, a model that includes the other, non polity, institutions identified by IAT will be added to assess whether these variables mediate the relationship between state dominance and corruption. Then, a series of models will be run, each including one of the following cros s product terms state dominance industrialization, state dominance education, and state dominance family disruption. These interaction terms will be used to test for moderation effects between the component variables. Next, the full IAT model identified in the mediation model will be analyzed with the addition of the two cultural measures moral cynicism and interpersonal distrust, to see if the introduction of these cultural factors produces an increase in the overall variance explained, as determined by a significant change in the coefficient of determination ( R 2 ). Lastly, in order to assess the efficacy of the model, the results of regression diagnostic testing will be discussed. All models are estimated in SPSS using ordinary least squares (OLS) regre ssion. Multivariate models
98 Table 7 presents the regression results for the direct effects of state dominance on corruption perceptions, controlling for population structure, government size, and known historical correlates of corruption. As a whole, th is model is statistically significant, Table 7: Corruption perceptions (TI CPI) regressed on IAT indicators Baseline model Mediation model Variable b t b t State dominance ( 1/x) 9.37 .87 14.33** 7.08 .65 9.76** Industrialization .76 .31 4 .72** Education .12 .05 1.08 Family disruption .01 .06 1.43 State X Industrialization State X Education State X Family disruption Population heterogeneity .11 .05 .92 .01 .00 .06 Government size .04 .10 2.00 ** .02 .03 .76 Fmr. British colony .38 .07 1.64* .45 .08 2.07** Protestant .04 .01 .17 .21 .04 .80 F statistic (df) 94.7 (5)** 82.1(8)** Adjusted R 2 .80 .86 2 .06** N = 121 N = 109 p < 0.10; ** p < 0.05 (F = 94.7, p < 0.05), explaining 80% of the variation in corruption perceptions across the nations sampled. In order to introduce linearity to the state dominance corruption relationship, the negat ive inverse ( 1/ x ) of the state dominance scale was substituted for the five point state dominance scale in this, and all subsequent, all estimations. Net the effects of the control variables, state dominance is significant and positively related to corru ption perceptions, such that a one unit increase in state dominance yields a 9.37 increase in corruption perceptions. Moreover, this rel
99 0.87). Mediation As discussed in Chapter Three, Messner and Rosenfeld argue that, for societies devalue, accommodate, and penetrate non econ omic institutions (2007, p. 76 83). the economy will continue to have a direct impact on how o ther institutions function and are able to a second causal link where the strength of non economic institutions, undermined by the dominance of the economy, directly a ffect, or partially mediate the effect of the economy on, crime rates (p. 1149). Extrapolating this finding to the current study, the researcher expects that the effects of the non polity institutions will mediate the corruption prone tendencies of domina nt states. In order to test for mediation effects, the non polity variables (industrialization, education, and family disruption) were added to the baseline model. As shown in the second panel of Table 7 (above), the addition of these mediator variables significantly increase the explained variance by 6% (F change = 10.43, p < 0.05). Should the state dominance variable become non significant or reverse its algebraic sign due to the inclusion of the non polity variables, this would signify full mediation (Baron and Kenney, 1986). As indicated by the significant t value and positive unstandardized beta for state dominance, this is obviously not the case in the present data, however; the
100 unstandardized beta is reduced 24.47% when the additional institutiona l variables are added to the model (b = 7.08); this suggests partial mediation. Furthermore, this attenuation is significant (p < .01, two tailed test) based upon the statistical test for the equality of regression coefficients developed by Paternoster et al (1998). z = b 1 b 2 1 2 + SE b 2 2 z = 2 + (.527) 2 z = 2.34 Only one of the non polity institutions is exerting this mediating effect. Industrialization (proxy variable for economic strength within a nation state) was negatively related to corruption perceptions such that a one unit increase in industrialization results in a 0.76 reduction on the TI argument that the economy acts as an institution of normative control, as first su ggested by liberal political and economic theories of the enlightenment, later echoed commerce thesis (1992. See Chapter Three, this thesis). Consistent with prior bivariate findings ( See Table 6), neither education nor the family dis ruption measures produces significant direct effects on corruption. Moderation using non additive modeling. In the current study three product terms (state industrialization, state education, and state family) were added to the full socio structural model to test for moderating effects. In creating these interactions the variables were mean centered prior to transformation in or der to
101 eliminate multicollinearity between the component variables and their product terms (McClendon, 1994). Regression results appear in Table 8 (below).
102 Table 8: Corruption perceptions (TI CPI) regressed on IAT indicators, moderation models Full IA T structural model Moderation model 1 Moderation model 2 Moderation model 3 Variable b t b t b t b t State dominance ( 1/x) 7.08 .65 9.76** 6.36 .59 8.67** 6.67 .62 7.77** 7.08 .65 9.27** Industrialization .76 .31 4.72** .10 .04 .31 .81 .33 4.68** .76 .31 4.51** Education .12 .05 1.08 .02 .01 .19 .39 .14 1.17 .12 .05 1 .06 Family disruption .01 .06 1.43 .01 .07 1.90* .01 .05 1.23 .01 .06 .60 State X Indust. 1.70 .41 3.15** State X Education .67 .11 .86 State X Family 0 .00 .10 Population heterogeneity .01 .00 .06 .01 .01 .10 .00 .0 0 .04 .01 .00 .06 Government size .02 .03 .76 .02 .03 .74 .01 .03 .61 .02 .03 .75 Fmr. British colony .45 .08 2.07** .55 .10 2.6** .45 .08 2 04** .45 .08 2.04* Protestant .21 .04 .80 .08 .02 .32 .14 .02 .52 .21 .04 76 F statistic (df) 82.1(8)** 80.6(9)** 72.9(9)** 72.2 (9)** Adjusted R 2 .86 .87 .86 .86 2 .01** 0 0 N = 109 p < 0.10 ; ** p < 0.05
103 Panel 1 shows the regression estimates for the full structural IAT model in order to provide a baseline for model comparison. Moderation model 1 introduces the state industrialization interaction term. At the model level, the inclusion of this product term significantly increases the explained variance by one percentage point (F change = 9.90, p < 0.05) confirming the necessity of the non additive model. The significant t value for the state industrialization product term suggests that the effects of these institutions do interact (b = 1.70). In other words, the effect on state dominance on corr uption attached to the coefficient indicates that, contrary to the doux commerce thesis, higher levels of industrialization increase corruption perceptions when the lev el of dominance by the state is high. This finding may be indicative of the fact that, for autocracies and dictatorships, economy has been nationalized by the state (Garland, 1997). Examining a scatterplot of TI CPI scores against industrialization (see Figure 5, below) confirms that corruption perceptions are relatively high, even in the presence of high levels of industrialization in the oil producing rentier states of Sau di Arabia, Kuwait, United Arab Emirates, Bahrain, and Qatar. Indeed, political scientists have advanced rentier state theory to explain authoritarianism and resistance to democratization in the petroleum rich states of the Middle East (Smith, 2004), natio been positively linked to rent seeking activities and corruption in the literature (e.g.
104 Ades and DiTella, 1999; Fish 2005; Lambsdorff, 1999; Leite and Weidmann, 2002; Robbins, 2000, Treisman, 2000). Figure 5: Corruption perceptions and industrialization scores As shown in moderation models two and three (Table 8, above), the inclusion of the state education and sta te family interaction terms do not increase the overall predictive power of the model, with both explaining approximately 86% of the variation in corruption measures, the same as the direct effects baseline model. Furthermore the F change statistic fails to reach significance in either model two (F change = .73, p > 0.05) or model three (F change = .00, p > 0.05).
105 Cultural indicators Although generally omitted in empirical tests of IAT due to lack of data, Messner cultural, as well as structural, component. In the alternate configuration of IAT tested in the present paper, the dominant cultural ethos is not that of the anomic state, but rather one of moral cynicism, defined herein as a lack of personal agency and responsibly on the part of the citizenry and interpersonal distrust. As these measures, gleaned from aggregate responses to the World Values Survey, are not available for a number of the nations included in the prior analyses, a separate analysis was perf ormed ( N = 50). Panel one of Table 9 provides the benchmark regression of the TI CPI regressed on IAT structural indicators and appropriate controls. Panel 2 includes the two cultural indicators as discussed in Chapter 4 moral cynicism and interpersonal distrust.
106 Table 9: Corruption perceptions (TI CPI regressed on IAT structural and cultural indicators Baseline model Culture Variable b t b t State dominance ( 1/x) 7.08 .65 9.76** 5.80 .54 5.05** Industrialization .76 .31 4.72** .86 .28 3.50** Education .12 .05 1.08 .05 .02 .25 Family disruption .01 .06 1.43 .02 .01 1.50 Moral cynicism .18 .05 .75 Interpersonal distrus t 3.85 .20 2.83** Population heterogeneity .01 .00 .06 .01 .01 .06 Government size .02 .03 .76 .00 .00 .03 Fmr. British colony .45 .08 2.07** .34 .05 .79 Protestant .21 .04 .80 .45 .07 .84 F statistic (df) 82.1(8)** 34.4( 10)** Adjusted R 2 .86 .87 2 .0 1 ** N = 121 N = 50 ** p < 0.05 At the model level, the inclusion of the two cultural indicators increases the model fit by one percentage point over the structural model; this increase is significant (F change = 4.856, p < 0.05 ). As in the bivariate analysis, moral cynicism is not significantly related to corruption perceptions in these data. Interpersonal distrust does exhibit a moderate an alysis constant. For every one point increase in the interpersonal distrust scale, there is a corresponding 3.85 increase in the TI CPI. This is consistent with the alternate reading of IAT, as well as with corruption research that has linked this dimens ion of social capital with corruption at the cross national level (Uslaner, 2004). Although Uslaner (2004) claims that the direction of causation is from social trust to corruption, there is a likelihood of a non recursive relationship, or feedback loops, such that a lack
107 of social trust increases corruption, which than increases interpersonal distrust. Unfortunately, the cross sectional nature of the current study precludes testing this hypothesis. Finally, there is also a possibility of mediation and/or moderation of the institutional variables by the inclusion of the social tru st measures. At this time, however; th ese possibilit ies have not been fully theorized. Regression diagnostics and sensitively analysis The use of ordinary least squares regression analysis requires that certain [This] may produce m current study, several transformations of the data were performed in order to introduce linearity (transformation of the state dominance variable) and deal with a non additive model specificat ion (introduction of appropriate interaction terms). In macro level research studies, multicollinearity often poses problems amongst highly correlated theoretical constructs. When multicollinearity is present, parameter estimates may be biased and unreli able, standard errors become inflated and Type II errors become likely (Fox, 1991). In the present study, several component variables were combined using principal components analysis (PCA) in order to reduce the highly correlated dimensions into a single factor score. Furthermore, in the creation of the interaction terms for the moderation models, variables were mean centered prior to estimation as per convention (Jaccard, Turrisi, and Wan, 1990). Although at the bivariate level, state dominance and ind ustrialization were highly correlated ( r = 0.72),
108 an examination of the variance inflation factors (VIF) indicated that none reached the dominance measure in the mediati on (full structural) model (Fox, 1991) 6 Finally, both the state dominance and industrialization measure reaches statistical significance in each of the models tested, thus, it appears that the data are free from serious collinearity. Examination of the dependent and independent variable residuals did not indicate heterogeneity in the error variance, another issue often prevalent in macro level research. Although one of the variables, moral cynicism, did appear to exhibit a W constant error variance in this measure. Spatial autocorrelation was also not problematic in the present data, as the Durbin Watson statistic for each of the models estimated approximately two. A final test of the regression assumptions indicated the presence of two distance values greater than one. An examination of the DFbeta scores for the individual independent variable show ed that these two nations were outliers on the state dominance, education, family disruption, moral cynicism, and interpersonal 6 The state dominance variance inflation factor did reach 4.76 in moderation model 1, which includes the state dominance industrialization interaction. Furthermore, that interaction product term had a VIF of 12.50. This is unavoidable to the inherent perfect correlation of the state dominance component variable and the high correlation between state dominance and industrialization in the bivariate findings. Both reach statistical significant in the model, indicating that high collinea rity is not problematic, although the exact coefficients should be interpreted with caution.
109 distrust measures. Additionally Nigeria was an outlier on industrialization, population heterogeneity, and government size indi cators. To test the sensitivity of the models indicated in the present study, each was rerun with Nigeria and the Czech Republic variable removed. Regression analysis with these outlier nations removed showed no substantive difference in either the cult ural or two of the moderation models. However, in the mediation (full socio structural IAT model), the omission of the Czech Republic and Nigeria increases the model fit, such that the model explains 90% of the variation in corruption perceptions cross na tionally. Additionally, in the revised mediation model the education measure reaches statistical significance (b = .54), p < 0.05), although this industrialization interaction) and moderation model three (state family disrupti on interaction) are substantively unchanged, however; in moderation model two (state education interaction) the algebraic direction of the state education interaction reverses direction (b = 0.57), but the effect size is weak and fails to reach statistica l .09, p > 0.05). This finding indicates that IAT is not only sensitive to the measures included in analysis (Piquero and Piquero, 1998), but also the nations included in analysis as first suggested by Trent (2007). As this is the firs t study of this alternate configuration of the institutional balance of power, as well as the first study to include over 100 nations at varying stages of development, further research is warranted.
110 Chapter Six Discussion The purpose of this study was to test the efficacy of an alternate configuration of anomie theory. Prior research has supported of IAT at the cross national level ( Messner and Rosenfeld, 1997; Savolainen, 2000 ), when economic institutions, which in turn, lose their ability to exert normative controls over a populace (Messner and Rosenfeld, 2006, p. 130 131). Support for IAT has also been found sub nationall y at various levels of aggregation ( Chamlin and Cochran, 1995; Maume and Lee, 2003; Piquero and Piquero, 1998) All but one empirical test of IAT has focused on street crime ( s ee Schoepfer and Piquero, 2006, for an analysis of white collar crimes across a United States sub na tional sample) ; the current study expands upon this small body of literature to include other institutional configurations, while focusing on less studied crimes of power. A ccording to Messner and Rosenfeld, it is not merely economic d ominance that per se (2001, p. 155, emphasis in original). The authors describe other, analytically distinct configurations of institutional imbalance that predict different typ es of crime. One of these, a dominant state, is hypothesized to be related to high levels of corruption within a social collective (Messner and Rosenfeld, 2001; 2004; Rosenfeld, 2004). This study is
111 the first to test this alternate configuration across a relatively large sample of nations at varying levels of development Additionally, t his study provides a complete test of this alternate configuration of IAT by developing an operationalization and measure for t Messner and Rosenfeld posit as the To date, f ew empirical tests of IAT have included measures of culture dynamics, an integral facet o f IAT, and with limited support (Cao, 2004; Chamlin and Cochran, 1997; 2007; Cullen, Parboteeah, and Hoegl, 2004; Jensen, 2002; Muftic, 2006). The current study confirms the institutional bala nce of power. As hypothesized, the level of state dominance (defined and measured as the absence of popular control over the state), is positively related to corruption cross nationally, with the highest levels of state dominance clustered in several Afri can and post Soviet nations, countries which also rank high on CPI). As suggested by the literature (Montinola and Jackman, 2002; Sung, 2004), a curvilinear relationship was found between state dominance and corruption, one that takes a cubic functional form. In order to introduce linearity to the state dominance corruption relationship, the negative inverse ( 1/ x ) of state dominance was used in the multivariate regression models. The direct e ffects of state dominance on corruption, controlling for known correlates of corruption was very strong, explaining 80% of the variation in corruption across the nations included in the sample.
112 According to Messner and Rosenfeld, social institutions are to some degree interdependent and coordinated. It is hypothesized that the non polity institutions will mediate and/or moderate the corruption prone tendencies of the dominant state. In order to test for mediation, the non polity variables (economy (indust rialization), education, and family disruption) were added to the baseline model. The addition of these variables increased the coefficient of determination for the model to 86%. One of the non polity variables, industrialization (proxy for economic stre ngth), was negatively related to corruption cross nationally, and, additionally, attenuated the corruption producing effects of the dominant state. This finding is consistent with Maume and Lee s mediation hypothesis ( 2003) t that the economy acts as an institution of normative control. The other institutions described in IAT (education and family disruption) failed to relate significantly to corruption at the cross national level, consistent with bivariate findings in the present study. A secon d possible causal link moderation, was tested through the introduction of interaction terms. T he state dominance industrialization interaction was significantly related to corruption cross nationally indicating that a non additive model was appropriate. In other words, the effect of state domination on corruption is dependent upon levels of industrialization. Higher levels of industrialization actually increases corruption when the level of state dominance is high. While this may appear to contradict th e earlier mediation finding, this is consistent with prior research on corruption in dictatorships and rentier state theory, and it indicates that the mediation
113 and moderation arguments are not mutually exclusive. Finally, unlike many prior tests of IAT, t he prese n t study included two measures of cultural conditions gleaned from the World Values Survey moral cynicism and interpersonal distrust. While the former (measured as a factor score of volunteerism and personal responsibility and agency) was unrelated to corruption at the cross national level, interpersonal distrust was positively related to corruption. Although the significance in a smaller sub sample of nations ( N = 50), or under le ss than ideal circumstances. This finding is consistent with the small body of literature linking social trust (or social capital ( s ee Putnam, 2000)) with lowers instances of corrupt activities (Bjornskov, 2004; Uslaner, 2004; Zak and Knack, 2001). This may be an indirect link, as it bureaucracies and higher levels of development, both correlates of reduced corruption (Harris, 2007, p. 2). The inverse has also been supported suggesting that societies with high levels of social capital may also be characterized by clientelism or familism, which reduces the ability for in group members to cooperate with outsiders (Harris, 2007; Lipset and Lenz, 1999). The positive association between distrust and corruption in the current study warrants further investigation. Regression diagnostics identified two nations who proved to be outliers on nearly every independent variable Nigeria and the Czech Republic. The removal of these two nat ions increas ed the fit of the mediation model, such that the coefficient of determination reached 90% Furthermore, this change caused the direct effects of
114 education to become significantly, positively related to corruption, an unexpected finding. Limit ed research suggests that corruption nations will have lower levels of education, as resources are diverted and distorted benefitting political organizations rather than the populace (Mauro, 1997). Although higher levels of education have been found to increase white collar criminal behaviors, the inverse has been found when dealing with street crimes (Schoepfer and Piquero, 2006 ). This seemingly incongruent findings warrants further investigation as the effects of education on corruption levels cross nationally has largely been ignored. Additionally, as the pres ence of only two outliers significantly impacts the results, future research should continue to examine the applicability of IAT to a such large array of nations. As found by Chamlin and Cochran (2007) and echoed in Trent (2007), the efficacy of IAT seems to be dependent upon the levels of development of the nations included in the analysis. When state dominance is plotted against corruption ( s ee Chapter 4), it appears that only a few highly industrialized, generally core OECD nations appear in the low co rruption, low state dominance range. Similarly, only a handful of nations are high corruption, high state dominance, and most are politically unstable, low development, peripheral nations. Replication of the current study, sampling f ro m may provide more interesting results as to the role that that various characteristics of the state may play in these cases. IAT alone may not be able to explain these findings, suggesting integration with other theories such as dependency and world syste ms, quantifiable through political economy measures
115 (Wallerstein, 1974) 7 Limitations Several limitations exist in the current study, each providing avenues for future research. At the macro level, direct measures of theoretical constructs are unavailable. Proxy variables must be used. The current study used a factor score that measured a he economy within a collective. This measure strayed from the standard measures of economy found in macro level analysis relative or absolute deprivation, as the common variables representing these constructs (GINI coefficient and poverty levels, respectively) present problems with endogeneity and causal order. Prior research suggests that these measures are both a cause and consequence of corruption ( s ee Lambsdorff, 1999 ) Following Cullen et. al. ( 2004 ) who draw on Bonger ( 1969 ), the current study operationalized the strength of the economy as levels of industrialization, yet in the globalized world, the relevance of this measure may be called into question Many industrialized nations, includ ing the United States, have moved from production to service economies as globalization shrinks o u r world. Research suggests that the advent of large m ulti national conglomerates actually leads to competition that promotes corruption in developing nations (Goudie and Stasavage 1998 p. 119), while Saj (2003) puts forth globalization as an explanation of recent increases in corruption cross nationally, e specially instances of grand corruption Future research should address measures that tap into the globalized economy. These may include measures of 7 The author would like to thank Michael Lynch for pointing out this interesting link.
116 transnational trade, labor migration, capital investment, and technology, using indices such as the KOF I ndex of Globalization (Dreher, 2006). Still, the interesting results of the current study, such that industrialization acts as both a mediator and a moderator to the corruption prone effects of state dominance, should not be discarded. Rather, further re search should build upon this foundation when determining how to best measure the economy both within and outside the IAT framework. Future studies of IAT must continue to find applicable measures for the other social institutions described in Messner and In particular the family disruption measure, usually captured as the ratio of marriages to divorces is unavailable for a large number of nations, thus the effects of divorce on corruption remains unknown. The current study used a mea sure of female economic activity to capture family disruption, but this measure was not related to corruption rates In the absence of other measures, we cannot determine what role, if any, the family institutions plays in the study of corruption Unfort unately, due to barriers such as cost, feasibility, and efficiency, researches are often limited to the use of secondary data not systematically collected to specifically test theory Lacking direct and sound cross national measures, the effects of this a nd other institutions such as religion will remain largely hypothetical. Furthermore, the unavailability of sound measures precluded testing this configuration of IAT against any rival theories at this time. Only as macro level measures become more devel oped and refined, will critical theory testing become possible at the cross national level. Finally, the use of a cross sectional dataset limits the current findings to
117 correlates, rather than causes of corruption. It is common practice to use multi year averages to control for economic and political fluctuations, however; in order to disentangle causal order, it is imperative to test theories cross nationally using longitudinal, time series data Examination of nations over time would also better assist in determining if the non political institutions mediate or moderate the corruption prone effects of state dominance by examining the effects of these institutions consecutively at equal intervals and identifying turning points Future research should begin to use longitudinal data as this becomes available. Finally, it is troublesome that the moralistic approach to the study of corruption corruption legislation and agenda rather than comprehension of the phenomenon (Kotkin and Saj, 2002). It is imperative that we first begin to understand corruption, a behavior that lacks even a definitional consensus, before we use empirical findings to inform public policy. Simply pu t, at this time, not enough is known. Conclusion In sum, in the twelve years since the publication of Crime and the American Dream, researchers have found institutional anomie theory to provide a sophisticated expansion of the Mertonian tradition of anomie theory as a possible explanation of the variance in crime rates across various macro social units of analysis b ut the authors do not limit their theory to instrumental crime, nor to cultural conditions of anomie. As shown in the current study, alternate specification s of the theory show empirical promise in predicting other forms of crime, or crime equivalents, under other cultural
118 conditions At present, too little has been done in both furthering theories of the anomie paradigm, and in studying other forms of crime and deviance beyond conventional street crime. Further research will assist in better specifying theoretical and causal assumptions, and, if sustained, institutional anomie theory may provide guidance to instruct public policies and the allocation of national assets in combating not only crime, conventionally defined, but also a wide range of serious offenses.
119 References Ades, A. & DiTella, R. (1997). National champions and corruption: Some unpleasant interventionist arithmetic. E conomic Journal, 107 (443), 1023 1042. Ades, A. & DiTella, R. (1999). Rents, competition, and corruption. The American Economic Review, 89 (4), 982 993. Agnew, R. (1987). On testing structural strain theories. Journal of Research in Crime and Delinquency, 24 (4), 281 286. Alesina, A. et. al. (2003). Fractionalization. Journal of Economic Growth, 8 155 194. Amendsen, I. (1999). Political corruption: An introduction to the issues. Unpublished manuscript. Retrieved November 19, 2007 from http://www.cmi.no/publications/publication/?1040=political corruption Andvig, J. et. al. ( 2000). Research on corruption: A policy oriented survey Bergen, Norway: Chr. Michelsen Institute (CMI) and Norwegian Institute of International Affairs (NUPI). Bardham, P. (1997). Corruption and development: A review of issues. Journal of Economic Literature, 35, 1020 1046. Baron, R.M. & Kenny, D.A. (1986). The moderator mediator variable distinction in so cial psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51 (6), 1173 1182.
120 Batton, C. & Jensen, G. (2002). Decommodification and homicide rates in the 20 th century United States. Homi cide Studies, 6 (1), 6 38. Theoretical Criminology, 11 (1), 63 93. Baumer, E.P. & Gustafson, R. (2007). Social organization and instrumental crime: Assessing the empirical validity of classic and contemporary anomie theories. Criminology, 45 (3), 617 663. Beblawi, H. & Luciani, G. (1987). Introduction. In H. Beblawi & G. Luciani (Eds.), The rentier state, (pp. 1 21), London: Croom Helm. Bennett, A. (1997). Lost in translatio n: Big (N) misrepresentation of case study research. [Electronic version]. Paper presented at the 38 th annual convention of the International Studies Association. Retrieved February 23, 2007 from http://www.georgetown.edu/faculty/bennetta/bign.htm. Bennet t, R.R. (1980). Constructing cross cultural theories in criminology: Application of the generative approach. Criminology, 18 (2), 252 268. Bennett, R.R. & Lynch, J.P. (1990). Does a difference make a difference? Comparing cross national crime indicators. Cr iminology, 28 (1), 153 181. Bernard, T.J. (1987). Testing structural strain theories. Journal of Research in Crime and Delinquency, 24 (4), 262 280. Bernberg, J.G. (2002). Anomie, social change and crime: A theoretical examination of institutional anomie the ory. British Journal of Criminology, 42, 729 742. Berry, W.D. & Feldman, S. (1985). Multiple regression in practice. Beverly Hills, CA: Sage.
121 Bjerregaard, B., and Cochran, J.K. Does the Strength of Other Social Institutions Mediate or Moderate the Effect s of the Economy on the Rate of Crime: A Cross National Test of Institutional Anomie Theory. Western Criminology Review Forthcoming. Bjornskov, C. (2004). How does social trust affect economic growth. Working paper 06 02, Department of Economics, Aarhus S chool of Business. Blalock, H.M., Jr. (1967) Toward a theory of minority group relations New York: Capricorn. Blau J. R. & Blau, P.M. (1982). The cost of inequality: Metropolitan structure and violent crime. American Sociological Review, 47 (1), 114 129. Bollen, K.A. (1993). Liberal democracy: Validity and method factors in cross national measures. American Journal of Political Science, 37 (4), 1207 1230. Bonger, W.A. (1969). Criminality and economic conditions. Boston: Little, Brown. (Original work publish ed 1916). Boyle, E.H. (2000). Is law rule? Using political frames to explain cross national variation in legal activity. Social Forces, 78 (4), 1195 1226. Brady, H.E. & Collier, D. (2004). Rethinking social inquiry: Diverse tools, shared standards. Lanham, MD: Kogan Page. Braithwaite, J. (1989). Crime, shame, and reintegration. Cambridge: Cambridge University Press. Brovkin, V.N. (2003). Corruption in 20 th century Russia. Crime, Law and Social Change, 40, 195 230.
122 Brown, E. & Cloke, J. (2004). Neoliberal ref orm, governance and corruption in the South: Assessing the international anti corruption crusade. Antipode, 36 (2), 272 294. Burnham, R.W. (1998, November). Analysis of the United Nations data set on crime trends and the operations of criminal justice sys tem s: Part I, Final Report." NCJ 1 83461, Washington, DC: United States Department of Justice, National Institute of Justice. Inter university Consortium for Political and Social Research [distributor], 1999. Retrieved January 3, 2007 from http://webapp. icpsr.umich.edu/cocoon/NACJD STUDY/02513.xml Cancino, J.M., Varano, S.P., Schafer, J.A. & Enriquez, R. (2007). An ecological assessment of property and violent crime rates across a Latino urban landscape: The role of social disorganization and institutiona l anomie theory. Western Criminology Review, 8 (1), 69 87. national data. International Journal of Comparative and Applied Criminal Justice, 28, 17 31. Chamlin, M.B. & Coc anomie theory: A partial test. Criminology, 33 (3), 411 429. Chamlin, M.B. & Cochran, J.K. (1997). Social altruism and crime. Criminology, 33 (2), 203 228. Chamlin, M.B. & Cochran, J.K. (200 5). Ascribed economic inequality and homicide among modern societies. Homicide Studies, 9 (1), 3 29.
123 Chamlin, M.B. & Cochran, J.K. (2007). An evaluation of the assumptions that underlie institutional anomie theory. Theoretical Criminology, 11 (1), 39 61. Ch eibub, J.A. & Gandhi, J. (2004). A six fold classification of democracies and dictatorships. Paper presented the 2004 meeting of American Political Science Association. Cullen, F.T. (1994). Social support as an organization concept for criminology: Preside ntial address to the academy of criminal justice sciences. Justice Quarterly, 11, 527 559. Cullen, J.B., Parboteeah, K.P. & Hoegl, M. (2004). Cross national differences in l anomie theory. Academy of Management Journal, 47 (3), 411 421. Currie, E. (1991). Crime in the market society. Dissent. (Spring), 251 259. Dahl, R. A. (1982). Dilemmas of pluralist democracy. New Haven, CT: Yale University Press. De Sousa, L. (2002). Hard responses to corruption: Penal standards and the repression of corruption in Britain, France and Portugal. Crime, Law and Social Change, 38, 267 294. Dreher, A. (2006). Does globalization affect growth? Evidence from a new Index of Globalization. Applied Economics, 38 (10), 1091 1110. Durkheim, E. (1961). Moral education. New York: Free Press. Durkheim, E. (1984). The division of labor in society. New York: The Free Press. (Original work published 1893).
124 Durkheim, E. (1997). Suicide. New York: The Free Pres s. (Original work published 1897). Elliott, K.A. (1997). Introduction. In K.A. Elliot (Ed.) Corruption and the global economy, (pp. 1 5), Washington, DC: Institute for International Economics. European Values Study Group and World Values Survey Association (2004). European and world values survey four wave integrated data file, 1981 2004 v.20060423, 2006. Aggregate File Producers: Anlisis Sociolgicos Econmicos y Polticos (ASEP) and JD Systems (JDS), Madrid, Spain/Tilburg University, Tilburg, The Nethe rlands. Data Files Suppliers: Analisis Sociologicos Economicos y Politicos (ASEP) and JD Systems (JDS), Madrid, Spain/Tillburg University, Tillburg, The Netherlands/ Zentralarchiv fur Empirische Sozialforschung (ZA), Cologne, Germany:) Aggregate File Distr ibutors: Anlisis Sociolgicos Econmicos y Polticos (ASEP) and JD Systems (JDS), Madrid, Spain/Tillburg University, Tilburg, The Netherlands/Zentralarchiv fur Empirische Sozialforschung (ZA) Cologne, Germany. Esping Anderson, G. (1990). The three worlds of welfare capitalism Princeton, NJ: Princeton University Press. Fajnzylber, P., Lederman, D. & Loayza, N. (2002). What causes violent crime? European Economic Review, 46, 1323 1357. Fish, M.S. (2005). Democracy derailed in Russia: The failure of open pol itics. New York: Cambridge University Press. Fitzsimons, P. (2002). The politics of corruption in the 21 st century. [Electronic version]. Globalization, 2 Retrieved October 10, 2007 from http://globalization.icaap.org/
125 Flyvberg, B. (2006). Five misunders tandings about case study research. Qualitative Inquiry, 12 (2), 219 245. Fox, J. (1991). Regression diagnostics Thousand Oaks, CA: Sage. Freedom House (2004). Freedom in the World. Washington, DC: Freedom House. Retrieved January 16, 2008 from http://www. freedomhouse.org/template.cfm?page=15. Gambetta, D. (2002). Corruption: An analytical map. In S. Kotkin & A. Sajo (Eds.), Political (pp. 33 56), New York: CEU Press. the problem of crime: Foucault, criminology, sociology. Theoretical Criminology, 1 (2), 173 214. Gartner, R. (1990). The victims of homicide: A temporal and cross national comparison. American Sociological Review, 55 (1), 92 106. Gerring, J. & Thacker, S.C. (2005). Do Neoliberal policies deter political corruption? International Organization, 59 (Winter), 233 254. Gibbs, J.P. & Martin, W.T. (1962). Urbanization, technology, and the division of labor: International patterns. American Sociological Review, 27 (Oc tober), 667 677. Giddens, A. (1979). Emile Durkheim. The Viking Press: New York. Girling, J.(1997). Corruption, capitalism, and democracy. New York: Routledge. Goldsmith, A.A. (1999). Slapping the grasping hand: Correlates of political corruption in emergi ng markets. American Journal of Economics and Sociology, 58 (4), 866 883. Goldthorpe, J.H. (1997). Current issues in comparative macrosociology: A debate on methodological issues. Comparative Social Research, 16, 1 26.
126 Goudie, A.W. & Stasavage, D. (1998). A framework for the analysis of corruption. Crime, Law and Social Change, 29, 113 159. Green, P. (2005). Disaster by design: Corruption, construction, catastrophe. British Journal of Criminology, 45 (4), 528 546. Green, P. & Ward, T. (2004). State crime: Gov ernments, violence and corruption. London: Pluto Press. Gupta, S., Davoodi, H. & Alonso Terme, R. (1998, May). Does corruption affect income inequality and poverty? International Monetary Fund Working Paper, 98/76. Hannon, L. & Defronzo, J. (1998). The tr uly disadvantaged, public assistance, and crime. Social Problems, 45 (3), 383 392. Harris, D. (2007). Bonding social capital and corruption: A cross national empirical analysis. Unpublished manuscript. Department of Land Economy: University of Cambridge. H earn, F. (1997). Moral order and social disorder: The American search for civil society. New York: Aldine de Gruyter. Hellman, J.S. (1998). Winners take all: The politics of partial reform in postcommunist transitions. World Politics 50 (2), 203 234. Hirsch i, T. (1969). Causes of delinquency. Berkley: University of California Press. Hirschman, A.O. (1992). Rival views of market society and other recent essays. Cambridge: Harvard University Press. Hosking, G. (1992). The first socialist society: A history of the Soviet Union from within. Cambridge: Harvard University Press.
127 Howard, G.J., Newman, G., & Pridemore, W.A. (2000). Theory, method, and data in comparative criminology. In D. Duffee, D. McDowall, B. Ostrom, R.D. Crutchfield, S.D. Mastrofski, & L.G. Maza rolle (Eds.), Measurement and analysis of crime and justice (pp. 139 212). Washington DC: National Institute of Justice. Huang, W.S.W. & Wellford, C.F. (1989). Assessing indicators of crime among international crime data series. Criminal Justice Policy re view, 3 (28), 28 48. Huntington, S.P. (1991). The third wave: Democratization in the late twentieth century. Norman: University of Oklahoma Press. Husted, B. (1999). Wealth, culture, and corruption. Journal of International Business Studies, 30 (2), 339 360. Jaccard, J., Turrisi, R. & Wan, C.k. (1990). Interaction effects in multiple regression. Thousand Oaks, CA: Sage. Jacobs, J.B. (2002). Dilemmas of corruption control. In S. Kotkin & A. Sajo (Eds.), Political (pp. 81 90), New York: CEU Press. Jain, A.K. (1998). Economics of corruption. Norwell: Kluwer Academic. Jain, A.K. (2001). The political economy of corruption, London: Routledge. Jensen, G. (2002). Institutional anomie and societal variations in crime: A critical appraisal. International Journal of Sociology and Social Policy, 22 (7/8), 45 74. Johnson, R.A. & Sharma, S. (2004). About corruption. In R.A. Johnson (Ed.) The struggle against corruption: A comparative study (pp. 1 19). New York: Palgrave MacMi llan.
128 Johnston, M. (2001). The definitions debate: Old conflicts in new guises. In A.K. Jain (Ed.), The political economy of corruption, (pp. 11 29). London: Routledge. Johnston, M. (2005). Syndromes of corruption: Wealth, power, and democracy. Cambridge, UK: Cambridge University Press. Karklins, R. (2002). Typology of post communist corruption. Problems of Post Communism, 49 (4), 22 32. Karklins, R. (2005). The system made me do it: Corruption in post communist societies. Armonk, New York: M.E. Sharpe. Kars tedt, S. (2001). Comparing cultures, comparing crime: Challenges, prospects and problems for global criminology. Crime, Law, and Social Change, 36 285 308. Karstedt, S. (2007). The strength of institutions: Corruption cultural values, and institutional re gimes. Paper presented at the 2007 annual meeting of The American Society of Criminology. Atlanta, GA. Karstedt, S. & Farrall, S. (2006). The moral economy of everyday crime British Journal of Criminology, 46, 1011 1036. Kaufmann, D. (1998). Research on c orruption: Critical empirical issues. In A.K. Jain (Ed.) Economics of corruption, (pp. 129 176), Norwell: Kluwer Academic. Kaufmann, D. Kraay, A. & Mastruzzi, M. (2003). Governance matters III: Governance indicators for 1996 2002. Washington, D.C.: The Wor ld Bank. Kaufmann, D. Kraay, A. & Mastruzzi, M. (2007a). Governance matters VI: Governance indicators for 1996 2006 Washington D.C.: The World Bank.
129 Kaufmann, D. Kraay, A. & Mastruzzi, M. (2007b). Measuring corruption: Myths and realities. In D. Rodriguez & L. Ehrichs (Eds.) Global corruption report 2007: Corruption in judicial systems, (pp. 318 323). Cambridge, UK: Cambridge University Press. Kauzlarich, D. Matthews, R.A. & Miller, W.J. (2001). Toward a victimology of state crime. Critical Criminology, 10 (3), 173 194. Kim, S.W. & Pridemore, W.A. (2005a). Poverty, socioeconomic change, institutional anomie, and homicide. Social Science Quarterly, 86 (Supplement), 1377 1398. Kim, S.W. & Pridemore, W.A. (2005b). Social change, institutional anomie and serious property crime in transitional Russia. British Journal of Criminology, 45, 81 97. Knack, S. (2006). Measuring corruption in eastern Europe and central Asia: A critique of the cross country indicators. World Bank Policy Research Working Paper 3968. Washingt on, DC: World Bank. Kotkin, S. (2002). Liberalism, geopolitics, social justice. In S. Kotkin & A. Sajo (Eds.), (pp. 387 412), New York: CEU Press. Kotkin, S. & Saj, A. (2002). Political corruption New York: CEU Press. Krohn, M.D. (1978). A Durkheimian analysis of international crime rates. Social Forces, 57 (2), 654 670.
130 Krohn, M. & Wellford, C.F. (1977). A static and dynamic analysis of crime and the primary d imensions of nations. International Journal of Criminology and Penology, 5, 1 16. LaFree, G. (1999). A summary and review of cross national comparative studies of homicide. In M.D. Smith & M.A. Zahn (Eds.) Homicide: A sourcebook of social research. (pp. 12 4 148) Thousand Oaks, CA: Sage. LaFree, G. & Drass, K.A. (2002). Counting crime booms across nations: Evidence for homicide victimization rates, 1956 to 1998. Criminology, 40 (4), 769 800. Lambsdorff, J.G. (1998). Corruption in comparative perception. In A. K. Jain (Ed.) Economics of corruption, (pp. 81 110), Norwell: Kluwer Academic. Lambsdorff, J.G. (1999, November). Corruption in empirical research A review. Transparency International Working Paper. Lambsdorff, J.G. (2007). The methodology of the corrupti on perceptions index 2007 Passau, Germany: Transparency International. Lapalombara, J. (1994). Structural and institutional aspects of corruption. Social Research, 61 (2), 325 350. Leff, N. H. (1964). Economic development through bureaucratic corruption. A merican Behavioral Scientist, 8, 8 14. Leite, C. & Weidmann, J. (2002). Does mother nature corrupt? Natural resources, corruption, and economic growth, In G.T. Abed & S. Gupta (Eds.) Governance, corruption, and economic performance, (pp. 159 196), Washing ton, D.C.: International Monetary Fund.
131 Lipset, S.M. & Lenz, G.S. (1999). Corruption, culture, and markets. Arlington, VA: George Mason University. Liska, A.E. (1990). The significance of aggregate dependent variables and contextual independent variables f or linking macro and micro theories. Social Psychology Quarterly, 53 (4), 292 301. Liu, J. (2007). Developing comparative criminology and the case of China. International Journal of Offender Therapy and Comparative Criminology, 51 (1), 3 8. Lu, H. & Gunnison E. (2003). Power, corruption, and the legal process in China. International Criminal Justice Review, 13, 28 49. Lynch, M.J. & Groves, W.B. (1995). In defense of comparative criminology: A critique of general theory and the rational man. In F. Adler & W. S. Laufer (Eds.), The legacy of anomie theory (pp. 367 392), (Vol. 6), New Brunswick: Transaction Publishers. Lynch, M.J. & Michalowski, R. (2006). Primer in radical criminology: Critical perspectives on crime, power and identity. New York: Criminal Justic e Press. Marenin, O. (1997). Victimization surveys and the accuracy and reliability of official crime data in developing countries. Journal of Criminal Justice, 25 (6), 463 475. Markovskaya, A., Pridemore, W.A. & Nakajima, C. (2003). Laws without teeth: An overview of the problems associated with corruption in Ukraine. Crime, Law, and Social Change, 39, 193 213. Maume, M.O. & Lee, M.R. (2003). Social institutions and violence: A sub national test of institutional anomie theory. Criminology, 41 (4), 1137 1172
132 Mauro, P. (1995). Corruption and growth. Quarterly Journal of Economics, 110 (3), p. 681 712. Mauro, P. (1997). The effects of corruption on growth, investment, and government expenditures: A cross country analysis. Institute for International Economics. Retrieved February 15, 2008 from http://www.iie.com. Mauro, P. (1998). Corruption and the composition of government expenditure. Journal of Public Economics, 69 263 279. McClendon, M.J. (2002). Multiple regression and causal analysis. Long Grove, IL: Wave land Press. International Social Science Journal, 48 (3), 309 320. Merton, R.K. (1938). Social structure and anomie. American Sociological Review, 3, 672 682. Merton, R.K. (19 68). Social Theory and Social Structure. New York: Free Press. Messner, S.F. (1980). Income inequality and murder rates: Some cross national findings. Comparative Social Research, 3, 185 198. Messner, S. (1982). Societal development, social equality, and h omicide: A cross national test of a Durkheimian model. Social Forces, 61 (1), 225 240. Deviant Behavior, 91 (1), 33 53.
133 Messner, S.F. (2003). Sozialstruktur und anomie: An insti tutional anomie theory of crime: Continuities and elaborations in the study of social structure and anomie. Cologne Journal of Sociology and Social Psychology, 43, 93 109. (Translation from German provided by author, pagination differs from print journal) Messner, S.F. & Rosenfeld, R. (1997). Political restraint of the market and levels of criminal homicide: A cross national application of institutional anomie theory. Social Forces, 75 (4), 1393 1416. Messner, S.F. & Rosenfeld, R. (2001). An institutional anomie theory of crime. In R. Paternoster & R. Bachman (Eds.), Explaining criminals and crime (pp. 151 160), Los Angeles: Roxbury. McCord (Eds.). Beyond empiricism: Ins titutions and intentions in the study of crime (pp. 83 105), New Brunswick, NJ: Transaction Publishers. Messner, S.F. & Rosenfeld, R. (2006). The Present and future of institutional anomie theory. In F.T. Cullen, J.P. Wright, & K.R. Blevins (Eds.). Taking stock: The status of criminological theory (pp. 127 148), New Brunswick, NJ: Transaction Publishers. Messner, S.F. & Rosenfeld, R. (2007). Crime and the American Dream. Belmont, CA: Thomson Wadsworth. (Original work published 1994). Mills, C.W. (2000). Th e sociological imagination. New York: Oxford University Press. (Original work published 1959).
134 Montinola, G.R. & Jackman, R.W. (2002). Sources of corruption: A cross country study. British Journal of Political Science, 32, 147 170. Moran, J. (2001). Democr atic transitions and forms of corruption. Crime, Law and Social Change, 36, 379 393. Morris, S.D. (2004). Corruption in Latin America: An empirical overview. SECOLAS Annals, 36, 74 92. connecting culture, institutions, and deviance. International Journal of Offender Therapy and Comparative Criminology, 50 (6), 630 653. Neapolitan, J.L. (1994). Cross natio nal variation in homicides: The case of Latin America. International Criminal Justice Review, 4, 4 22. Neapolitan, J.L. (1997). Cross national crime: A research review and sourcebook. Westport, CT: Greenwood Press. Neapolitan, J.L. (1999). A comparative an alysis of nations with low and high levels of violent crime. Journal of Criminal Justice, 27 (3), 259 274. Neapolitan, J.L. (2003). Explaining variation in crime victimization across nations and within nations. International Criminal Justice Review, 13, 76 89. Norris, P. (2008). Shared global indicators cross national database [SPSS]. Retrieved January 25, 2008 from http://ksghome.harvard.edu/~pnorris/Data/Data.htm. Nye, J.S. (1967). Corruption and political development: A cost benefit analysis. The America n Political Science Review, 61 (2), 417 427.
135 The Sociological Quarterly, 6 (1), 37 44. Osgood, D.W. (2000). Poisson based regression analysis of aggregate crime rates. Journal of Quantitative Criminology 16 (1), 21 43. Paldam, M. (2001). Corruption and religion: Adding to the economic model. Kyklos, 54 (2 3), 383 413. Pampel, F.C. & Gartner, R. (1995). Age structure, socio political institutions, and national homicide rates. European Sociological Review, 1 1 (3), 243 260. Passas, N. (1997). Anomie, reference groups, and relative deprivation. In N. Passas & R. Agnew (Eds.), The future of anomie theory, ( pp. 63 94), Boston: Northeastern University Press. Passas, N. & Agnew, R. (1997). The future of anomie theor y. Boston: Northeastern University Press. Paternoster, R. & Bachman, R. (2001). Explaining criminals and crime. Los Angeles: Roxbury. Paternoster, R., Mazerolle, P. & Piquero, A. (1998). Using the correct statistical test for the equality of regression coe fficients. Criminology, 36 (4), 859 866. Philp, M. (1997). Defining political corruption. Political Studies, XLV, 436 462. Philp, M. (2002). Political corruption, democratization, and reform. In S. Kotkin & A. Sajo (Eds.), Political corruption in transition (pp. 57 80), New York: CEU Press.
136 Piquero, A. and Piquero, N.L. (1998). On testing institutional anomie theory with varying specifications. Studies on Crime and Crime Prevention, 7 (1), 61 84. Pratt, T.C. & Godsey, T.W. (2003). So cial support, inequality, and homicide: A cross national test of an integrated theoretical model. Criminology, 41 (3), 611 643. Pridemore, W. A. (2008). A methodological addition to the cross national empirical literature on social structure and homicide: A first test of the poverty homicide thesis. Criminology, 46 (1), 133 154. Pridemore, W.A. & Kim, S.W. (2006). Democratization and political change as threats to collective sentiments: Testing Durkheim in Russia. The Annals of the American Academy of Politic al and Social Science, 605 (1), 82 103. Pridemore, W.A., Chamlin, M.B. & Cochran, J.K. (2007). An interrupted time series Justice Quarterly, 24 (2), 271 290. Putnam, R.D. (2000). Bowling alone: The collapse and revival of American community. New York: Simon and Schuster. Rawls, J. (1972). A theory of justice. Oxford: Oxford University Press. Reichel, P.L. (1999). Comparative criminal justice systems: A topical approach. Upper Saddle River, NJ: Prentice Hall. Robbins, P. (2000). The rotten institution: Corruption in natural resource management. Political Geography, 19 (4), 423 443. Robinson, M. (1998). Corruption and development: An introduction. In M. Robinson (Ed.) Corruption and development, (pp. 1 14), Lond on: Frank Cass.
137 Rodgers, E.J.(2006). Conflict diamonds: Certification and corruption A case study of Sierra Leone. Journal of Financial Crime, 13 (3), 267 276. Rodriguez, D. & Ehrichs, L. (2007). Global corruption report 2007: Corruption in judicial systems Cambridge, UK: Cambridge University Press. Rose Ackerman, S. (1978). Corruption: A study in political economy. New York: Academic Press. Rose Ackerman, S. (1997). The political economy of corruption. In K.A. Elliott (Ed.), Corruption and the global econo my, (pp. 31 60), Washington, DC: Institute for International Economics. Rose Ackerman, S. (1999). Corruption and government: Causes, consequences, and reform, Cambridge: Cambridge University Press. Rose Ackerman, S. (2001). Political corruption and democra tic structures. In A.K. Jain (Ed.), The political economy of corruption, (pp. 35 62). London: Routledge. Rosenfeld, R. (2004). Terrorism and criminology. Sociology of Crime, Law, and Deviance, 5, 19 32. Rosenfeld, R. & Messner, S.F. (1997). Markets, morali ty, and an institutional anomie theory of crime. In N. Passas & R. Agnew (Eds.), The Future of Anomie Theory, ( pp. 205 224), Boston: Northeastern University Press. Sajo, A. (2002). Clientalism and extortion: Corruption in transition. In S. Kotkin & A. Sajo (Eds.), (pp. 1 22), New York: CEU Press.
138 Saj, A. (2003). From corruption to extortion: Conceptualization of post communist corruption. Crime, Law, and Social Change, 40 171 194. Sandholtz, W. & Koetzle, W. (2000). Accounting for corruption: Economic structure, democracy, and trade. International Studies Quarterly, 44, 31 50. Savolainen, J. (2000). Inequality, welfare state, and homicide: Further support for the institutional anomie theory. Crimin ology, 38 (4), 1021 1042. Schoepfer, A. & Piquero, N.L. (2006). Exploring white collar crime and the American dream: A partial test of institutional anomie theory. Journal of Criminal Justice, 34, 227 235. Scott, J.C. (1972). Comparative political corruptio n. Englewood Cliffs, New Jersey: Prentice Hall. Seligson, M.A. (2006). The measurement and impact of corruption victimization: Survey evidence from Latin America. World Development, 34 (2), 381 404. Sen, A. K. (1997). On economic inequality, Oxford: Clarend on Press. Shleifer, A. & Vishny, R. (1993). Corruption. Quarterly Journal of Economics, 108 599 617. Sims, B.A. (1997). Crime, punishment, and the American dream: Toward a Marxist integration. Journal of Research in Crime and Delinquency, 34 (1), 5 24. Smi th, B. (2004_. Oil wealth and regime survivial in the developing world, 1960 1999. American Journal of Political Science, 48 (2), 232 246. Stack, S. (1984). Income inequality and property crime: A cross national analysis of relative deprivation theory. Crim inology, 22, 229 257.
139 Stark, R. (1987). Deviant places: A theory of the ecology of crime. Criminology, 25 (4), 893 909. Stucky, T.D. (2003). Local politics and violent crime in U.S. cities. Criminology, 41 (4), 1101 1135. Sung, H. (2004). Democracy and polit ical corruption: A cross national comparison. Crime, Law and Social Change, 41, 179 194. Svensson. J. (2005). Eight questions about corruption. Journal of Economic Perspectives, 19 (3), 19 42. Tanzi, V. (1998). Corruption around the world: Causes, consequen ces, scope, and cures. IMF Staff Papers, 45 (4), 559 594. Tanzi, V. & Davoodi, H.R. (1997). Corruption, public investment and growth. IMF Working Paper Series WP/97/139, Washington, DC: International Monetary Fund. Tanzi, V. & Davoodi, H. (2001). Corruption growth and public finances. In A.K. Jain (Ed.), The political economy of corruption, (pp. 89 110). London: Routledge. Theobald, R. (1990). Corruption, development and underdevelopment. Durham: Duke University Press. Thompson, E.P. (1963) The makings of t he English working class. London: Penguin. Treisman, D. (2000). The causes of corruption: A cross national study. Journal of Public Economics, 76, 399 457.
140 Trent. C.L.S. (2007). Examining the variability of economic crime rates cross nationally: A partial test of institutional anomie theory. Paper presented at The American Society of Criminology annual meeting, Atlanta, GA. Tukey, J.W. (1977). Explanatory data analysis. Reading, MA: Addison Wesley. United Nations (1948). Universal declaration of human right s, 10 December 1948. [Electronic version]. New York: United Nations. Retrieved June 28, 2007 from http://www.un.org/Overview/rights.html United Nations (2004). Global Compact Guidance document: Implem entation of the 10 th principle against corruption [Electronic version]. New York: United Nations. Retrieved March 15, 2007 from http://www.unglobalcompact.org/index.html. United Nations (n.d.) Questionnaire for the seventh United Nations survey of crime trends and operations of criminal justice systems, covering the periods 1998 2000. Vienna: United Nations. Retrieved February 5, 2008 from http://www.unodc.org/pdf/crime/seventh_s urvey/InstrumentE.pdf United Nations (Various years). Human development report. New York: United Nations. Retrieved January 18, 2008 from http://hdr.undp.org/en/humandev/hdi/ Uslaner, E.M. (2004). Trus t and Corruption. In J.G. Lambsdorf, M. Taube, & M. Schramm (Eds.), Corruption and the new institutional economics. London: Routledge. Vanhanen, T. (2000). A new dataset for measuring democracy, 1810 1998. Journal of Peace Research, 37 (2), 251 265. Vigderh ous, G. (1978). Methodological problems confronting cross cultural criminological research using official data. Human Relations, 31 (3), 229 247.
141 Wallerstein, I. M. (1974). The modern world system: Capitalist agriculture and the origins of the European worl d economy in the sixteenth century. New York: Academic Press. White, H. (1980). A heteroskedasticity consistent covariance matrix and a direct test for heteroskedasticity Econometrica, 48, 817 838. Williams, P. (1999). Emerging issues: Transnational crime and control. In G. Newman (ed.) Global report on crime and justice New York: Oxford University Press. Sociological Inquiry, 52 (2), 103 113 Xin, X. & Rudel, T.K. (2004). The context for political corruption: A cross national analysis. Social Science Quarterly, 85 (2), 294 309. Zak, P. & Knack, S. (2001). Trust and growth. The Economic Journal, 111, 295 321. Zimring, F.E. & Johnson, D.T. (2005). On the comparative study of corruption. Brit ish Journal of Criminology, 45 (6), 793 80 9.
143 APPENDIX A: DATA SOURCES Variable Description Source Corruption perceptions 1 to 10 scale (Higher scores indicate higher levels of perceived corruption); 2001 2005 avg. (annual) Transparency Internat ional Corruption Perceptions Index www.ti.org Moral cynicism Cynical orientation Voluntary organization mean and government responsibility factor score; 2000 World Values Survey International Network of Social Scientists World Values Survey (2000) www.worldvaluessurvey.org Interpersonal distrust Social trust measure; 2000 World Values Survey International Network of Social Scientists World Values Survey (2000) www.worldvaluessurvey.org State dominance Voice and accountability Extent to wh ich citizens may participate in government, freedom of speech/association; 2000 2005 avg. (annual) World Bank Group Governance Indicators www.worldbank.org/wbi/governance/ Regulatory quality Regulations that promote and permit private sector developm ent; 2000 2005 avg. (annual) World Bank Group Governance Indicators www.worldbank.org/wbi/governance/ Rule of law Quality and predictability of law enforcement and judiciary, extent to which agents have confidence in rules of society ; 2000 2005 avg. (annual) World Bank Group Governance Indicators www.worldbank.org/wbi/governance/ Civil liberties and political rights Freedom of expression, associational and organizational rights, personal autonomy; electoral process, political pluralism, 2000 Fr eedom House Freedom Around the World 2005 www.freedomhouse.org Freedom of the press Legal environment, political pressures, economic factors, access to information; 2004 Freedom House Freedom Around the Worl d 2005 www.freedomhouse.org Dimensions of democracy (Vanhanen) Competition (% of seats held by dominant party) and electoral turnout; 1810 1998 T. Vanhanen (2000) Journal of Peace Research, 37 (2), 251 265
144 APPENDIX A: DATA SOURCES (cont.) State dominance (cont.) Regime type Monarchic dictatorship, military dictatorship, civilian dictatorship, presidential dem ocracy, mixed democracy, parliamentary democracy; 2000 J.A. Cheibub. & J. Gandhi, J. (2004). fold classification of democracies and Paper presented the 2004 meeting of American Political Science Association. Industrialization facto r score GDP per capita (PPP) Final value of goods and services within a nation adjusted for cost of living and inflation; 2000 2005 avg. (semi annual) World Bank Group World Development Indicators 2007 www.worldbank.org Commercial energy use Oi l equivalent per capita; 2002 2005 avg. (semi annual) World Bank Group World Development Indicators 2007 www.worldbank.org Electrical consumption Per capita in kilowatt hours; 2001 United Nations Human Development Report 2003 www.undp.org Ur ban population Population living in urban areas (defined by within nation census data) as % of total pop.; 2002 2005 avg. (annual) World Bank Group World Development Indicators 2007 www.worldbank.org Distribution by sector Economic distribution by s ector; percent non agrarian (combined industry and service); various years United States Central Intelligence Agency CIA World Factbook 2007 https://www.cia.gov/library Family disruption Female economic activity (non domestic labor) rate (% ages 15 and a bove); 2002 United Nations Human Development Repo rt, 2003 www.undp.org Education factor Enrollment Combined gross enrollment (primary, secondary, and tertiary schools); 2001 2004 avg. (annual) United Nations Human Development Report, 2003;2007 www.undp.org Literacy rate Adult literacy rate (ages 15 and above); 2000 2002 (annual) United Nations Human Development Report, 2003 www.undp.org
145 APPENDIX A: DATA SOURCES (cont.) Control variables Population hetero geneity Economic disparity Gini index of income equality; 2004 United Nations Human Development Report, 2004 www.undp.org Ethno linguistic fractionalization Ethnic fractionalization (combined linguistic and racial); 2002 A. Alesina et. al (2002) Journal of Economic Growth, 8, 155 194 Former British colony Dichotomous variable (0= no; 1 = yes) United States Central Intelligence Agency CIA World Factbook 2007 https://www.cia.gov/library Protestant Dichotomous variable (0=no; 1= yes) United States Central Intelligence Agency CIA World Factbook 2007 https://www.cia.gov/library Size of government Central government expenditures (as % of GDP); 2000 2005 mean value (annual) World Bank Group Wo rld Development Indicators 2007 www.worldbank.org