Modernization:
A Mexican Perspective on
Democratization and Violent
Crime
Student: Mikki Sistermans
E-mail: mikki.sistermans@student.uva.nl
Study: Political Science (specialisation: International Relations)
Stud.nr: 10183736
Date: Haarlem, 15-7-2015 (rewritten version)
Bachelor project: Armed Conflicts, Political Actors, and Statebuilding
Professor: Dr. H. Matthee
1
Table of contents
Introduction 2
Democracy and Crime:
A Multilevel Analysis of Homicide Trends in Forty-Four Countries, 1950-2000 3
Theoretical background 6 Conceptualization 8 Methodology 10 Hypothesised mechanism 13 Analysis 15 Discussion 20 References 21
2
Introduction
In general one assumes that a strong democracy is associated with low levels of violent crime. Statistics showing us that most of the strong and well established democracies in the world do indeed have relatively low murder rates seem to support this idea. But why is it that democracies seem to have these relatively low violent crime numbers? Less democratic countries can have equally strong or even stronger police forces than fully democratic countries. They can also have equally strong or even stronger economies. Yet somehow increased democratization generally seems to lead to a decrease of violent crime as LaFree and Tseloni point out in Democracy and Crime: A Multilevel
Analysis of Homicide Trends in Forty-Four Countries, 1950-2000 (2006). This thesis offers an
exploration of the modernization perspective offered by LaFree and Tseloni by applying it to Mexico. A country known mostly for its war on drugs and the violence this brings forth, Mexico has
surprisingly seen mostly declining violent crime statistics until then newly elected president Calderon declared his ‘war on drugs’ in 2007. Looking at the two administrations prior to Calderon’s this thesis will try to shed some light on the modernization perspective which suggests that while in a
transitional phase to democracy a country will see increased violent crime, but these numbers will drop as soon as the democratic institutions have become embedded in the country’s political and social structures. Such a transitional phase is exactly what happened in Mexico between the 1990s and early 2000s, especially during the Ernesto Zedillo and Vicente Fox administrations. By applying theory-testing process tracing this thesis will try to find empirical evidence for a question coming forth out of the statistical claim that democratization of a country will lower its violent crime
numbers: How has increased democratization led to a decrease in violent crime in Mexico during the Zedillo and Fox administration between 1994 and 2006?
3
Democracy and Crime: A Multilevel Analysis of Homicide Trends in Forty-Four
Countries, 1950-2000
This thesis will be using LaFree and Tseloni’s (2006) article as main theoretical basis for research. Therefore, a thorough understanding of the article is required. This chapter will serve as an analysis of the theories and claims it makes.
LaFree and Tseloni (2006) start of by explaining that the urge to write their article came forth out of a lack of existing literature at the time. They claim to be the first to take two major trends and combine them in scientific research: A big increase in democratization between 1975 and 1995 and a
substantial rise in violent crime during the same time period. This correlation leads for reason to believe both variables might influence each other (LaFree and Tseloni, 2006) and trying to explain this possible relation is the main aim of the article. In order to examine this possible relation, three compelling theoretical frameworks are brought forward: Norbert Elias’s civilization theory (1939, 1978), conflict theory from William Chambliss (1976) and Robert Bohm (1982) and Emile Durkheim’s modernization theory (1947, 1950). Though not specifically mentioned the research question used by LaFree and Tseloni (2006) is likely closely related to the following: Can the global increase in
democratization since 1975 help explain the global increase in violent crime in the same period?
The civilization perspective
To answer this question LaFree and Tseloni (2006) develop four hypotheses, based on existing theories, which they will test using an econometric analysis. The first hypothesis is based on Norbert Elias’s civilization theory (1939) and argues that increased democratization will lead to a decline of violent crime. LaFree and Tseloni state that: “Although Elias’s civilization perspective does not directly reference the role of democratization in reducing crime, his arguments are nevertheless centered on a set of West European countries that eventually became strong democracies.”(2006: 28). An elaborate explanation tells us this is based on a prediction from Elias (1939, 1978) that modern states will claim a monopoly on violence and make punishment for violent crime committed against a person a public rather than a personal matter. The second prediction made by Elias (1939, 1978) is that “Increasingly complex social configurations such that advancing individual self-interest less often required the use of violence.” (LaFree and Tseloni, 2006: 29). LaFree and Tseloni (2006) support this with more recent research from Gurr (1981), Osterberg (1996), Spierenburg (1996) and Eisner (2001). However, there’s also evidence that conflicts with Norbert Elias: Gurr, Grabosky and Hula (1977), LaFree and Drass (2002) and Eisner (2001) all conclude that violent crime has seen an increase after World War II as opposed to a decline since the Middle Ages until before WWII. LaFree and Tseloni still justify the use of the civilization perspective hypothesis with the following statement: “However, for our purposes, such conclusions must be tempered by the fact that no study to date has specifically examined connections between levels of democratization and violent crime rates for a large sample of countries.” (LaFree and Tseloni, 2006: 29).
The conflict perspective
The next hypothesis introduced by Lafree and Tseloni (2006) states the opposite of the civilization perspective: An increase in democratization will lead to an increase in violent crime. This claim is made based on work from Chambliss (1976) and Bohm (1982) and states that increased
democratization should be closely associated with a capitalist market economy as used by most modern democracies. Increased economic inequality, unemployment and social misery encourage
4 greed and selfishness and propagate aggression. This development exists not just within, but also between countries as the gap between highly industrialized Western democracies and a much larger group of economically dependent countries grows (LaFree and Tseloni, 2006). The article then proceeds with providing a few more authors who reinforce the conflict perspective hypothesis: Neuman and Berger (1988), Wallerstein (1979), Hopkins and Wallerstein (1981), Chirot and Hall (1982), Spitzer (1975) and Applebaum (1978). The research that showed rising violent crime number since World War II can also be seen as evidence of the conflict perspective and LaFree and Tseloni (2006) end with more research, mainly in budding democracies in Eastern Europe and Latin America, backing up the hypothesis.
The modernization perspective
The third hypothesis is based on Durkheim’s (1947) assessment of the transition from traditional to modern society and more recent research showing evidence for Durkheim’s claim. It states that violent crime will rise when a country transitions from an autocracy to a democracy due to modern values and norms clashing with the existing autocratic ones. “The emergence of new roles not yet fully institutionalized and integrated into society make [sic] normative guidelines ambiguous and weaken traditional support mechanisms.” (LaFree and Tseloni, 2006: 31), according to research from Merton (1938), Messner and Rosenfeld (1997), Davies (1962), Smelser (1962), Tilly, Tilly and Tilly (1975), Useem (1985), Lodhi and Tilly (1973), Cloward and Ohlin (1960) and Agnew (1992). The article then continues with the claim that most research done along the lines of the modernization
perspective may have been wrong. According to the article most research has found that there was no significant link between economic development and crime rates or that this link was negative (LaFree and Tseloni, 2006). Yet LaFree and Drass (2002) claim the idea of modernization is inherently longitudinal, while most of the research has not been longitudinal. Based on this LaFree and Drass (2002) conclude that research that is not longitudinal does not properly show the full relation between economic development and violent crime. LaFree and Drass (2002) also found evidence in their longitudinal study of 34 nations between 1995 and 1998 that supports the modernization theory (LaFree and Tseloni, 2006). The biggest difference with the civilization and conflict
perspectives lies in the fact that the modernization theory only links an increase of violent crime with the transitional phase of a new democracy. Once a country has established and strengthened its democratic institutions crime should start to decline. However, LaFree and Tseloni (2006) mention only limited evidence to support this theory with related research from Montinola and Jackman (2002) and Sung (2004).
The fourth hypothesis is a null hypothesis, claiming no connection between democratization and violent crime (LaFree and Tseloni, 2006). Figure 1 visualizes the four different hypotheses as proposed by LaFree and Tseloni (2006).
5 Figure 1 A visual display of the link between democratization and crime in the four hypotheses.
For their data LaFree and Tseloni (2006) use statistics for the WHO on violent crime that measure homicide rates per one hundred thousand citizens. They did so for forty-four countries between 1950 and 2000. The variation between the amounts of data per country was only due to availability of the data. For democracy LaFree and Tseloni (2006) used the scale developed by Gurr et al. for the Polity project (1974, 1990, 1995, 2003) which divides countries on a 21 point scale and rates them from autocracy, -10 points, to a full democracy, 10 points (LaFree and Tseloni, 2006). The article also explains use of several control variables and the use of a hierarchical linear model approach (LaFree and Tseloni, 2006). For the purpose of this thesis the actual analysis doesn’t offer any extra necessary understanding hence we will continue with the conclusion of the article.
Conclusion
From the results LaFree and Tseloni (2006) conclude that just the modernization perspective hypothesis holds up. First, they argue that homicide numbers in full democracies were not significantly lower than in autocracies, which disproves the civilization perspective. Second, they argue that homicide numbers in full democracies weren’t significantly higher than in autocracies, disproving the conflict perspective. Third, they argue that homicide numbers in countries
transitioning towards democracy were significantly higher than other categories, thus supporting the modernization perspective. Last they argue that homicide numbers between autocracies and full democracies did not significantly differ over time which indirectly supports the conflict perspective. Based on these findings they conclude that the modernization perspective is the most fitting (LaFree and Tseloni, 2006).
6
Theoretical background
To gain a better understanding of what LaFree and Tseloni (2006) measure in their research it is important to take a look at the scale they use and the research from Ted Gurr (1974) they use for their democracy scale. This chapter will also take a more extensive look at the modernization perspective and the authors that are mentioned by LaFree and Tseloni in relation to the
modernization perspective. Where necessary it will also relate the theories to the case of Mexico.
Ted Robert Gurr’s Democracy Scale
The democracy scale used by LaFree and Tseloni (2006) was first developed by Ted Robert Gurr in
Persistence and Change in Political Systems, 1800-1971 (1974) and has been widely accepted as one
of the best datasets on polity (Center for Systemic Peace, 2014). Gurr develops his scale under the assumption that “The most important feature of any political system […] is its pattern of authority relations.” (Gurr, 1974: 1484). From this he derives that a ‘stable’ political system is one whose patterns of authority do not change much over a longer period of time, relatively to other political systems. The most durable systems are those that are able to gradually adapt over time as a result of and to counter internal and environmental pressures (Gurr, 1974).
LaFree and Tseloni (2006) derive their scale from this article. For this thesis their adaption will be used since it is their modernization perspective we are looking to test. It is also a more recent article and thus likely adapted from a more fully developed democracy scale as Gurr et al. (1974, 1990, 1995, 2003) have continued to work on the democracy scale. This scale recognizes four
categories including: “(1) the competitiveness of political participation, (2) the openness of executive recruitment, (3) the competitiveness of executive recruitment, and (4) constraints on the chief executive.” (LaFree and Tseloni, 2006: 35). For each of these categories countries earn points depending on how well they do (LaFree and Tseloni, 2006).
Since the assumption is that Mexico has improved on democratization it is less important to know the autocracy side of the scale. But because it makes sense to have at least a basic
understanding of the full scale it is worth noting that “the autocracy measure is based on values on the opposite side of the scales that make up the democracy scale.” (LaFree and Tseloni, 2006: 36). Finally the total score is found by subtracting the amount of points gained for autocracy from the points gained for democracy. This will produce a number ranging from -10 (autocracy) to +10 (full democracy) on a 21-point scale (LaFree and Tseloni, 2006).
Heung-En Sung, Montinola and Jackman
Besides an understanding of the democracy scale it is also necessary to have a closer look at the reasons why crime will decrease once a democracy has established itself. LaFree and Tseloni (2006) mention there has not been a great amount of research on the matter, but the authors they do mention help to gain some insight on the subject.
First of all it has to be said that the article used in this thesis is not the exact same article as used by LaFree and Tseloni (2006) in their research. Unfortunately Sung’s work from 2004 was no longer available, but based on the conclusions LaFree and Tseloni draw on Sung’s work and the conclusions that can be drawn from Sung’s 2006 article it is safe to say that both articles offer roughly the same insights on the relation between democratization and violent crime. Sung (2006) argues in his study on democratization and criminal justice systems that countries that experience an improvement in democracy are very likely to increase the strength in personnel of courts, police and prisons. Sung (2006) explains that the main difference between authoritarian and democratic
7 regimes lies in its focus: authoritarian regimes tend to focus on social control and efficiency, locking up as many lawbreakers as efficiently as possible to ensure criminals get punished; democratic regimes are more focused on following the correct protocols to ensure that when a criminal is convicted he is truly guilty and has received a punishment that matches the severity of the crime. Using data from the United Nations Surveys on Crime Trends and the Operations of Criminal Justice Systems, Sung (2006) finds that staff of court, police and prison organizations is larger in democracies compared to autocracies. This contradicts with the hypothesis he used, which predicted democracies to have relatively smaller police forces than autocracies. Sung notes that a “plausible explanation for the smaller size of authoritarian police forces is that the domestic surveillance of political figures and the suppression of opposition groups are usually conducted not by the regular police but by the state security or by secret police organizations.” (Sung, 2006: 326-327), while democracies are more focused on having a police force that protects and assists the public (Sung, 2006). This does not directly explain the decline of violent crimes in democracies compared to autocracies. However, it does suggest that democratic authorities are more focused on protecting citizens, whereas
authoritarian regimes are more focused on social control of the public and control of political actors. Gabriella Montinola and Robert Jackman (2002) have conducted a cross-country study on sources of corruption. One of their hypotheses predicts that “more competitive political structures inhibit corruption” (Montinola and Jackman, 2002: 153), with the structures mentioned implying democracies. Contrary to this hypothesis Montinola and Jackman (2002) find that politicians in democracies run a greater risk at being replaced giving them more incentive to try to cheat the system, for example by accepting bribes. However, they also find that a country chooses democracy as its system and enters the transformation stage it will likely see an increase in corruption compared to autocratic regimes. But once a country passes the transitional phase and moves into a phase of stabilized democracy corruption numbers drop significantly (Montinola and Jackman, 2002). It must be stressed that this research is based on actors with a political function. Even though this also is not a direct answer to the question why violent crime is likely to drop due to democratization, it does imply politicians in democracies to be less interested in bribery and more focused on doing their work to ensure a longer stay in office. This can include reducing violent crime, especially in countries with fairly high homicide numbers, as one can imagine that a politician who succeeds in reducing murders gains popularity among citizens.
8
Conceptualization
This chapter will explain some of the most recurring concepts in this thesis. Whenever one of these concepts is used it will be based on the elaboration offered in this chapter.
Crime
One of the variables around which LaFree and Tseloni (2006) based their research in Democracy and
Crime: A Multilevel Analysis of Homicide Trends in Forty-Four Countries, 1950-2000 is crime. Although
the article does not offer any definition on crime it seems clear that the focus is on violent crime involving the death of at least one individual. Reason to believe this can be found in the use of homicide statistics, collected by the World Health Organization (WHO), which is used to conduct the research (LaFree and Tseloni, 2006). Whenever crime or a variation on the concept of crime is mentioned in this thesis it will be in reference to this paragraph unless specifically mentioned. In line with LaFree and Tseloni (2006) any homicide statistics used will also be based on evidence collected by the WHO.
Democracy (scale) and Autocracy (scale)
Democratization is the second main variable used by LaFree and Tseloni (2006) and this is measured on a scale which ranges from (full) democracy to autocracy. The democratic part of the scale has already been explained in the theoretical background and that is the explanation that will be used in this thesis. However, Gurr et al. (1974, 1990, 1995 and 2003) developed not one, but two sides for their scale including the aforementioned democracy part and an autocratic part. That latter part will be explained here. “For the most part, the autocracy measure is based on values on the opposite side of the scales that make up the democracy scale.” (LaFree and Tseloni, 2006: 35). An autocracy with suppressed competitiveness of political participation receives 2 points and if it has restricted participation it will receive 1 point; countries with closed recruitment of executives are awarded 1 point; if chief executives are determined by hereditary succession or designation 2 points are given; for slight to moderate constraints on the chief executive a country receives 1 point, 2 points for an intermediate category and 3 points when there are no constraints on the chief executive. “The final 2 points on the autocracy scale are based on evaluations of the extent to which countries have binding rules on when, whether, and how political preferences are expressed.” (LaFree and Tseloni, 2006: 36). This means that 2 points are awarded to countries where competitive groups, issues, and types of conventional participation are regularly excluded from the political process. “Countries that oscillate between parties that favor [sic] their own group members in central allocations and restrict competing group members receive 1 point.” (LaFree and Tseloni, 2006: 35). After a country has received its points on both sides of the scale the final score is obtained by subtracting the autocracy measure from the democracy measure. This produces a 21-point scale that ranges from -10 for autocracies to +10 for full democracies (LaFree and Tseloni, 2006).
Democratization
Democratization is the process of a country becoming a better democracy based on the scale
explained above from Gurr et al. (1974, 1990, 1995 and 2003). Thus whenever a country gains a point on this scale this is considered democratization.
A direct example of democratization in Mexico is the 1996 reform of the countries electoral system which resulted in a big step forward in its process of democratization. These reforms
9 300 of the 500 seats which translates to 60%; a maximum level of over-representation; lowered requisites for registration of new political parties; greater levels of fairness in the distribution of radio and television broadcasting time; the right to clarify information provided by the media if said
information is distorted; the incorporation of an electoral court into the judiciary with maximum authority (except for actions of unconstitutionality) (Ace Project, 2015). Besides these changes Zedillo also reformed the Federal Electoral Institute (FEI) which has the duty to guarantee the fairness of elections on federal level. These were some of the changes made to the FEI: a complete split from the highest executive power; voting within management of the FEI is held entirely among citizen advisors; several national matters regarding political parties came under exclusive control of the FEI (i.e. funding, access to media), incorporation into the judiciary and a permanent nature as opposed to existence only during election periods (Ace Project, 2015).
10
Methodology
Process tracing
Process tracing is a method to analyse how a specific cause let to a certain outcome: how did A lead to B? B is the given outcome while we think that A is the cause, thus we are testing the cause while the outcome stays the same. (Punton and Welle, 2015a). Process tracing uses probability tests to assess the strength of the evidence and gives a researcher the tools to find out how and why a certain effect occurred and evaluate its impact (Punton and Welle, 2015b). Punton and Welle’s
Applying Process Tracing In Five Steps (2015a)1 will be used as the main guideline towards creating a theoretical framework for process tracing. Once the standard framework is explained it will be adapted to fit the needs of this thesis, while trying to stay as close to the original literature as possible.
Three different types
Punton and Welle (2015a) recognize three different types of process tracing: Theory-testing process tracing, theory-building process tracing and explaining-outcome process tracing. These three different categories all have different assumptions about the specific cause leading to a certain outcome. The first type of process tracing, theory-testing process tracing, assumes the following conditions: (1) we know the cause and the outcome. (2) We think there is a causal link between the cause and outcome. (3) We think we know why the cause led to the outcome (Punton and Welle, 2015a). This means we will be testing the causal link between the cause and outcome to see if it holds up. The second type of process tracing, theory-building process tracing, is divided in two different versions: the first assumes that we know what the cause and outcome are and that we think there is a link between the two. The second assumes we know the outcome, but we do not know the cause. In both versions of theory-building process tracing there is no knowledge on how the cause led to the outcome (Punton and Welle, 2015a). The aim of this is to find out the cause and what lead to the given outcome. The third type of process tracing, explaining-outcome process tracing, assumes that we know the outcome, but not the cause and that we want to explain what exactly led to the outcome “in order to craft a ‘minimally sufficient’ explanation” (Punton and Welle, 2015a: 2). Thus, explaining-outcome process tracing aims at getting the fullest understanding of why a certain outcome occurred.
Theory-testing process tracing
To find out if LaFree and Tseloni’s (2006) modernization perspective applies to Mexico it seems most reasonable to use theory-testing process tracing, since we know the outcome and we think we know what led to this outcome. The modernization perspective provides the proof to assume there is a causal link between increased democratization and declining violent crime numbers. This paragraph will further explore the concept of theory-testing process tracing to gain a better understanding.
Punton and Welle use a five step approach to process tracing starting with “Developing a hypothesised causal mechanism.” (Punton and Welle, 2015a: 4). A causal mechanism is described as the explanation of what it is exactly that makes A lead to B, where A is the cause and B is the
outcome. A mechanism consists of at least two components: an entity (or actor) and an activity. Logically the entity carries out the activity. A mechanism that includes only one entity and one
1 This article is based on Beach, D. and Pedersen R. B. (2013) Process-Tracing Methods: Foundations and
11 activity is the most basic form, but a mechanism can also include many more entities undertaking various activities. The most important aspect is to recognize that an entity can only be part of the mechanism if its activity is caused by the activity from another entity within the mechanism and adds to the outcome (Punton and Welle, 2015a). When developing a hypothesised causal mechanism Punton and Welle (2015a) note that during the development every part of the mechanism can be framed as a hypothesis which can then be tested.
The second step is to operationalize the hypothetical causal mechanism developed in step one. This means finding a way to make a part in the mechanism recognizable through certain characteristics that can be observed (Punton and Welle, 2015a). To do so Punton and Welle (2015a) mention four different types of evidence: (1) account evidence is actual empirical content
(interviews, observations etc.). (2) Trace evidence is proof of the mere existence of a part of the hypothesised mechanism. (3) Pattern evidence is based on statistical patterns and probabilities. (4) Sequence events evidence is based on a sequence in which we expect events to happen in case (a part of) the hypothesised mechanism exists. Operationalization also requires proof of causation between all parts of the mechanism, forcing the researcher to think about alternative parts in the mechanism and exclude unlikely parts (Punton and Welle, 2015a).
Step three involves the collecting of evidence. Considering how crucial it is to gather good evidence it is important to make sure evidence is reliable and valid (Punton and Welle, 2015a).
During step four the inferential weight of the evidence is assessed. Punton and Welle (2015a) mention four different tests: the ‘straw-in-the-wind’ test, the ‘hoop’ test, the ‘smoking gun’ test and ‘doubly decisive’ test. The first test is the straw-in-the-wind test. Straw-in-the-winds include hints that certain events might have taken place, but they do not offer firm conclusion. A straw-in-the-wind test is neither necessary nor affirmative of a hypothesis, but it is also the easiest and can still provide valuable information. For example if there are multiple pieces of evidence for a hypothesis that pass a straw-in-the-wind test it improves the likelihood of the hypothesis being probable (Collier, 2011).
The ‘hoop’ test gives us a slightly better idea of the probability of a hypothesis. The basic idea behind it is that a hypothesis must pass the test in order to still stand, if it fails the test, the
hypothesis should be eliminated. It must ‘jump through the hoop’ (Collier, 2011). A hoop test can eliminate a hypothesis, but it cannot confirm one.
The third test is the ‘smoking-gun’ test and can be taken quite literally. As Collier notes: “The metaphor of a ‘smoking gun’ conveys the idea that a suspect who is caught holding a smoking gun is presumed guilty.” (Collier, 2011: 827). However likely the probability of the hypothesis being true, this does not actually mean the hypothesis is true. It mainly offers very strong support for a hypothesis, but to fail a smoking-gun test does not mean a hypothesis isn’t correct (Collier, 2011).
The last test offered is the most decisive one and is fittingly named the ‘doubly decisive’ test. When evidence passes this test it not only strongly supports the hypothesis, but it also denounces other hypotheses. The doubly decisive test builds upon the combination of multiple of the
aforementioned tests and can only be carried out when one has found most of the pieces of the puzzle (Collier, 2011). Through elimination of other hypotheses in the end only one hypothesis should remain hence “when the investigator has eliminated all plausible alternatives, the remaining scenario must be the correct one.” (Collier, 2011: 827). A doubly decisive test is very difficult to attain, but if a hypothesis passes this test one can assume with strong certainty that the hypothesis is incorrect.
12 the evidence and a conclusion on the strength of the hypothesis. They point out that because all the parts of the mechanism should be necessary to make it work the mechanism itself is only as strong as its weakest part. If for example one part only passes a straw-in-the-wind test there is still a big possibility that this part is not causal, which means the mechanism does not work as described. However this does not exclude the hypothesis from being true, it merely questions the
trustworthiness of the evidence. Thus it is important that as much research as possible is put into finding as many evidences as possible so as many tests as possible can be conducted. With every piece of evidence that passes a test, the likeliness of the hypothesized mechanism being correct increases.
13
Hypothesised Mechanism
The first step in exploring the case is to hypothesize a mechanism that can offer an explanation as to how an increase in democracy leads to a decline in violent crime. Before doing that however, it is important to note that his research will focus on the political aspect of the modernization
perspective offered by LaFree and Tseloni (2006). The reason behind this is simply a matter of practicality since I do not have the time or the resources to fully explore all of the Mexican society to uncover every single part of the mechanism. The choice for the political aspect seemed the most logical one given the fact that the democracy scale used by LaFree and Tseloni (2006) mainly focuses on the political aspects of a country. A change for the better or worse on the democracy scale is very likely to be caused by decisions made by political actors. Taking this into account the hypothesized mechanism will consist out of mainly political entities carrying out activities that influence the outcome through political ways. Since LaFree and Tseloni (2006) do not offer very concrete literature on how increased democratization leads to a decline in violent crime the hypothesized mechanism will be very abstract and will focus on the already existing literature about democracy and violent crime in Mexico. The choice for a more abstract hypothesised mechanism over a more complicated and in-depth version implicates a few things. First of all the conclusions drawn from this mechanism about the specific case of Mexico will likely not be as strong as conclusions that can be drawn from a more advanced hypothesised mechanism. This means the validity of this research might be lower compared to other research conducted for similar topics in Mexico. However, the aim of this research is ultimately to test the modernization perspective and by using an abstract hypothesised mechanism that is more generalizable it might be easier to translate the mechanism and its results to other cases. Regardless of the results of this research it would likely take more cases than just Mexico to reach a thorough verdict on the modernization perspective.
Using the analysis of the modernization perspective offered by LaFree and Tseloni all the basic elements for the hypothesised mechanism can be derived. Starting with democratization the first part of the mechanism are politicians who create new institutions and laws and/or abolish old institutions and laws. This leads to power vacuums and clashes between old and new institutions and laws, it also leads to the second part of the mechanism where individuals with ill intent will look to exploit these power vacuums. As a result violent crime will increase in the third part of the
mechanism. The fourth part sees the aforementioned politicians respond to the increase in violent crime by further strengthening their new democratic institutions and laws. As a result the new institutions and laws on one side of the scale and the individuals of ill intent on the other side will balance each other out, leading to equilibrium in the fifth part. Over time this leads to the sixth part where the new institutions and laws become accepted throughout society. This increase in legitimacy will help the institutions and laws to do their jobs, including the prevention of violent crime, which is the seventh and final part of the mechanism. The result is a decrease in violent crime.
Mexico
The choice for the Ernesto Zedillo administration in Mexico (December 1st 1994 – November 30th 2000) and the Vicente Fox administration following it (December 1st 2000 – November 30th 2006) comes forth out of a multitude of reasons. The country qualifies as a critical case, which means that if the theory is right this case should prove it (Bryman, 2008). During Zedillo’s administration Mexico’s score on the democracy scale improved from a score of 0 to a score of +6 which happens to be a threshold for democracy on said scale. The biggest incline took place between 1995 and 1996 as
14 Mexico registered a +4 score in 1996 (Center for Systemic Peace, 2014). This suggests that during the Zedillo administration Mexico completed its transition to democracy. Homicide statistics show a steady decline in murder rate from 1995 until 2000 going from 16.9 to 10.7 per 100,000 citizens, with the biggest drop happening in the last two years of the Zedillo administration: 14.1 in 1998, 12.4 in 1999 and aforementioned 10.7 in 2000 (UN Data, 2015). The Fox administration then registers an increase on the democracy scale from +6 to +8 thus implying a consolidation of democracy in Mexico (Center for Systemic Peace, 2014) The homicide rates for the Fox administration show a
consolidation of the murder rate around 10 per 100,000 citizens (UN Data, 2015). Combining the democracy score and the homicide rate one can see a connection with the modernization perspective: First of all there is a big improvement on the democracy score in 1996 with smaller improvements towards a democracy threshold in the following years up to 2000 and the
consolidation of democracy up to 2006. Second, there is a more significant drop in homicide rate in the later years of the Zedillo administration suggesting that after some sort of a stabilization of democracy the violent crime decreased more drastically than during the earlier years.
15
Analysis
This chapter will assess the found evidence through the four tests offered by process tracing literature as described earlier in this thesis. First we will shortly2 assess the relevant authors LaFree and Tseloni (2006) use to support their theory. By doing so we can gain an understanding about the validity of their arguments and create a starting point from where new authors will be added with supporting evidence. It is important to keep in mind that the authors LaFree and Tseloni (2006) mention (Montinola and Jackman, Sung) are not focused on Mexico which limits the validity of the articles for the case of Mexico. After evaluating their literature the focus will shift to articles written specifically about Mexico.
Heung-En Sung
Democracy and Criminal Justice in Cross-National Perspective: From Crime Control to Due Process
(2006) is not the exact article used by LaFree and Tseloni (2006) since this was no longer available. However, the article used (Democracy and organized crime activities: Evidence from 59 countries (2004)) and the 2006 article written by Sung seem to be very complementary to each other in what they offer on democratization and violent crime, hence they will be treated as equal. Sung (2006) claims that democracies are more focused on protecting their citizens with their police force while autocracies mainly aim for social control. Sung (2006) qualifies as trace evidence and passes the straw-in-the-wind test for step four of the hypothesised mechanism as politicians in democracies will look to further improve democratic institutions and laws for the protection of their citizens.
Montinola and Jackman
Montinola and Jackman (2002) argue that once a country chooses to convert to a democracy it will likely see an increase in corruption during the transitional period. This might be due to the newly created power vacuums in step two of the hypothesised mechanism. However, it is very hard to link Montinola and Jackman’s evidence to step three and violent crime in general, thus the article provides merely trace evidence that might just pass a straw-in-the-wind test. Because the article does not address violent crime or Mexico it is very hard to classify this as useful within the analysis.
Andrés Villarreal
Andrés Villarreal (2002) offers the following conclusion: “The results provide considerable support for the hypothesis that in societies characterized by patronage relations, an increase in electoral competition will result in higher rates of homicidal violence.” (Villarreal, 2002: 494). Villarreal comes to this conclusion based on his statistical research conducted among voting results of 1,800 Mexican municipalities and violent crime statistics between 1987 and 1999. His research is specifically aimed at the impact from political competition and factionalization on informal social and political
networks. He expects these informal social and political networks, specifically hierarchical patron-client relationships, to be stronger and more important in rural areas. This results in the choice for municipalities as they are the smallest political and administrative units in the Mexican political system (Villarreal, 2002).
At first glance this speaks against the assumption that increased democratization should lower violent crime; however, after putting the article in context it offers a different view. The main
2 A more thorough analysis of the articles of Montinola and Jackman (2002) and Heung-En Sung (2006) can be found on page 7.
16 reason Villarreal’s conclusion should be seen as support instead of contradictory is the time frame when this research was conducted. For his research Villarreal uses data between 1987 and 1999 which, according to the research conducted by the Center for Systemic Peace (2014) is the
transitional phase for Mexico. An increase in violent crime is what is to be expected for this period because Villarreal’s research is based on the time period that corresponds with the first phase of the modernization perspective: the phase where new laws and institutions are introduced and clash with old and existing ones. It is important to notice that Villarreal mentions his conclusion only holds up for rural areas where he mentioned patron-client relationships to be strongest. This means that Villarreal’s research does not represent all of Mexican society and only holds up for small, rural societies. Villarreal’s conclusion is also contradictory with the UN data which show a steady decline in violent crime starting in 1995, but this is likely because Villarreal focuses on rural areas whereas the UN data account for all of Mexico. A possible explanation could be that the patron-client relations in rural Mexico are indeed very strong and are more difficult to remove; as a result it could have taken longer for the rural parts of Mexico to consolidate democracy. This means step four of the
hypothesised mechanism might have been delayed in rural areas compared to other parts of Mexico. Villarreal also offers a very interesting notion on a case specific condition by mentioning patronage relations. Villarreal shows that these patronage relationships are a factor by testing them in his models. He finds that the relationship between increased democracy and increased homicide statistics due to patronage relations was only found in rural areas even after controlling for a multitude of standard correlates of violent crime. He suggests that this effect occurs due to a disturbance in patronage networks. It must also be noted that Villarreal does not offer a way to research the underlying mechanism of this specific effect (Villarreal, 2002). Regardless of the absence of a specific mechanism it does seem that the statistics offered by Villarreal give an interesting insight in how the collision of existing social frameworks with the arrival of a new, more democratic, social framework lead to increased violent crime. It would have been interesting to see how the homicide statistics in these specific rural communities hold up against a well-established democracy once the transition has been completed.
This article provides pattern evidence for the existence of a power vacuum created by a clash between democratization in the form of more open political competition on the one hand and pre-existing patronage relationships representing the old power structures in Mexican society. Villarreal’s statistical research can qualify for both the straw-in-the-wind and the hoop tests for the second and third part of the hypothesised mechanism. However, because it is based on statistics it cannot pass the other two tests as Villarreal does not show actual empirical evidence of causation. Villarreal notion that his conclusions only hold up for rural areas of Mexico means that the evidence might not be translatable to all of Mexico. Final conclusion on Villarreal would have to be that he provides proof for part two and three of the hypothesised mechanism, but that this proof has limited validity for the entire country of Mexico.
Michel Estefan Gutierrez
In line with LaFree and Tseloni (2006), research conducted by Gutierrez also shows a link between democratization and a decrease in violent crime. The research done by Gutierrez differs from LaFree and Tseloni’s mainly because Gutierrez’ research focuses mainly on Mexico. Using different datasets for homicides (including five control variables for other causes of murder) and democracy he still ends up with roughly the same findings as LaFree and Tseloni as he concludes that “political competition is associated with lower homicide rates.” (Gutierrez, 2011: 61) and “the more people
17 vote in an election, the less likely it is that homicides will increase.” (Gutierrez, 2011: 61). Using multiple regression models Gutierrez shows that a higher degree of vote concentration in the winning party leads to an increase in murders and that a larger number of voters in an election is linked to a smaller chance of an increase in homicides (Gutierrez, 2011).
It is interesting to note that Gutierrez builds on the models used by aforementioned Andrés Villarreal, but with the difference that Gutierrez uses the time period of 1990 until 2008 and that he measures democratization based on voting participation and political competition in five elections for members of the Mexican House of Representatives (Gutierrez, 2011). Thus the main difference between Villarreal and Gutierrez is that the latter focuses on a national level, while the former focused on local level. The two main variables for democratization used by Gutierrez are: the first is fraction of votes received by the winning party, where a lower fraction of votes should imply more political competition. The second is voter participation. A higher percentage of votes from registered voters imply a higher participation in elections. High voter turnouts also imply more legitimacy for elected politicians and more legitimacy is said to generate more respect for social norms from these politicians (Gutierrez, 2011).
Unfortunately Gutierrez does not show the differences in fractions of votes received by winners and voter participation for the five different elections he used in his research. This means there is no possibility to see how these statistics would have corresponded with the homicide statistics and democracy statistics provided by the Center for Systemic Peace (2014) and the UN (2015). However, Gutierrez does prove that during the 1990-2008 time period political competition in the form of a lower percentage of votes for the winning party corresponds with less violent crime. And also that for the same time period higher voter participation corresponds with a lowered possibility of an increase in violent crime.
Even though Gutierrez’ research shows correlations between democratization and lowered violent crime it is less easy to fit this within the hypothesised mechanism and the time frame used in this thesis. The time frame used by Gutierrez is longer (1990-2008) than the one used in this thesis (December 1994 – November 2006) which means that it is difficult to rate the validity of this research because the statistics cannot be connected with shorter time periods in Mexican history and the modernization perspective necessary for this thesis. His research also does not seem to provide any real evidence for any of the steps in the hypothesised mechanism as his statistics only show
correlation, but he does not connect this to any actual causation evidence.
Aguirre and Herrera
Evidence what can go wrong during the power vacuum is offered by Aguirre and Herrera (2013) as they argue that democratic changes have only been made at a federal level while state and local level have not changed as much. Because of this power structures have shifted from a national to a more localized level but otherwise have not changed, suggesting that power is still held by small elites, specifically governors, instead of being more democratically allocated. As a result Mexican citizens still suffer from violent crime and the danger of being killed is still very high.
Aguirre and Herrera base their research on a case study of the state of Michoacan in Mexico. With the use of “the authors’ own experience, documented sources, and participative investigation of diverse events related to the article’s theme” (Aguirre and Herrera, 2013: 224) they test the hypothesis that organized crime and drug trafficking have flourished due to weak political institutions (Aguirre and Herrera, 2013). They focus on the creation of the illicit organization called
18 structure of local cells who embed themselves in the local community through economic, social, political and religious ties. Because of their interest in local society members of
‘LaFamiliaMichoacana’ are different from most criminal organizations and more accepted among local society. By doing so this particular criminal organization has created a situation of coexistence with the local governments and has even replaced local governments in some of their functions, creating a parallel government (Aguirre and Herrera, 2013).
The article then proceeds by stating the process of democratization on federal level has led to a loss in effectiveness and because of that legitimacy, particularly in the areas of public safety and justice administration. This is the power vacuum in which an organization like ‘LaFamiliaMichoacana’ emerged. Aguirre and Herrera argue that local representatives are “elected through top-rank
negotiations of the political parties” (Aguirre and Herrera, 2013: 231) which implies that
democratization has not yet found its way to Michoacan. Furthermore the authors note that the state of Michoacan has a history of social movement struggles and because of this history organizations like ‘LaFamiliaMichoacana’ earned some sort of local legitimacy just by existing (Aguirre and Herrera, 2013).
Even though their research and knowledge on the state of Michoacan seems extensive, Aguirre and Herrera fail when they try to translate their state level findings to a federal level. They do not mention any data on violent crime like the data provided by the UN (2015) which clearly shows a steady decline in violent crime in the last 15 years, yet they seem to blame the federal government and their failure to democratize for the violence in the state of Michoacan. This article therefor provides empirical evidence that in Michoacan politicians have thus far failed to proceed to step four of the hypothesised mechanism (responding to increased violence) and the evidence seems strong enough to pass the straw-in-the-wind test, the hoop test and the smoking-gun test. But it only proves this point for the state of Michoacan and not for the country of Mexico as a whole. This does seem to indicate that democratization does not happen at the same speed throughout the whole country as ‘LaFamiliaMichoacana’ wasn’t founded until 2006. It implicates Michoacan was only at step 2 of the hypothesised mechanism and thus seemed to be far behind the general level of democratization in Mexico.
Snyder and Duran-Martinez
Does illegality breed violence? Drug trafficking and state-sponsored protection rackets (2009) offers
an insight on the relationship between the breakdown of an existing political and social order to create a new one. In the case of Mexico, this translates to the democratization of the country as argued earlier in this thesis. Snyder and Duran-Martinez (2009) begin in the mid-1980s when the Mexican Attorney General’s Office issued reforms that increased the rotation of police officers and civilian personnel in an attempt to fight corruption. In 1996 three Assistant Attorney General’s Offices were created to decentralize the power and all of these offices held non-contiguous jurisdictions spread out over Mexico. As a result of these changes the Mexican drug cartels now faced a problem related to their spatial organization. This led to heightened uncertainty about their operations because they did not know who to bribe and an increased need to find their own means for protection of these operations. Indeed Sneyder and Duran-Martinez (2009) note that as soon as 1997 one of the biggest cartels had created its own paramilitary force. They then proceed by assuming: “These paramilitary groups likely contributed to higher levels of drug-related violence. Moreover, the uncertainty […] led to conflicts among criminal organizations over market share.” (Sneyder and Duran-Martinez, 2009: 265).
19 This article shows how a simple decision from Mexico’s attorney general to try and fight corruption inside its office and become more transparent has led to the creation of a paramilitary force by a drug cartel. The article qualifies as trace evidence and passes the straw-in-the-wind and hoop tests as Sneyder and Duran-Martinez describe how the power vacuum created by the attorney general leads to exploitation by one of the cartels and an increase in violent crime.
In a short summary of the hypothesised mechanism and the assessed literature it can be said that Snyder and Duran-Martinez offer evidence for part one, two and three as they show how changes in an existing institution led to a power vacuum and an increase in violent crime. Evidence for part two and three is also provided by Villarreal as he provides evidence of increased violent crime as a result of increased democratization because of clashes between existing social structures and new
democratic reforms in rural areas. Some evidence is then offered on part four by Sung as he explains that democracies will look to improve the safety of their citizens. Evidence for parts five until seven could not be found in the literature.
20
Discussion
Using Democracy and Crime: A Multilevel Analysis of Homicide Trends in Forty-Four Countries,
1950-2000 (2006) as a starting point the aim of this study was to test the modernization perspective
proposed by said article and explore the case of Mexico between 1994 and 2006 in relation to the modernization perspective. The article by LaFree and Tseloni was thoroughly analysed together with the theories they based their modernization perspective on and this created the main theoretical framework for this thesis. Using theory-testing process tracing methodology this thesis has tried to examine the literature written about democratization and violent crime in Mexico, specifically between the mid-1990s and the early 2000s. The case of Mexico’s transitional phase towards increased democratization during this showed that there does indeed seem to be a link in the statistics between democratization and an increase in violent crime during a transitional phase and a decrease of violent crime once a democracy has been well established as has been shown by
statistics from the Center for Systemic Peace and the UN. Using a seven part hypothesised mechanism this thesis has tried to find empirical evidence for the existence of this statistical link, only to find that evidence is limited. Looking at each individual part of the mechanism it must first of all be noted that parts five, six and seven are completely missing. This means a balance between crime and democratization and later on an acceptance of the new institutions and laws, leading to a decrease in violent crime were nowhere to be found in the assessed articles. For all the other parts some evidence has been found, but it seems that for each of these parts the word ‘part’ is very fitting as all of the authors mentioned in this thesis seem to have found their own part of Mexico or their own part of Mexican history to fit the needs of their research. Unfortunately this means that a complete picture of the process of democratization and a decrease in violent crime during the Zedillo and Fox administrations was not found. This also means that it is very difficult to answer the question if and how increased democratization has led to a decline in violent crime numbers during the Zedillo and Fox administrations.
However, there is still something that can be said about the modernization perspective and Mexico as evidence was found for democratization and power vacuums following the start of democratization. It seems that the statistics seem to favour the 1994-2006 period as Mexico’s transitional phase from autocracy towards democracy. But the empirical evidence found in the literature shows that democratization and the events following democratization seem to have many different starting points on Mexico’s timeline. Where Sneyder and Duran-Martinez mention a period as early as the mid-1980s
, there’s also Villarreal and Aguirre and Herrera who imply the existence of a power vacuum as a result of democratization in the late 90s in rural Mexico and even in 2006 in the state of Michoacan. This could imply that the process described by the modernization perspective simply takes Mexico much longer than the statistics on democracy and violent crime seem to indicate. Mexico might not have been as far in the process of democratization as expected, before being brutally interrupted by a war on drugs.
21
References
Written sources
Agnew, R. (1992) Foundation for a general strain theory of crime and delinquency. Criminology 30(1): 47-87
Aguirre, J. and Herrera, H. A. (2013) Institutional weakness and organized crime in Mexico: the case of Michoacán. Trends in Organized Crime 16(2): 221-238
Applebaum, R. P. (1978) Marx's theory ofthe falling rate of profit. American Sociological Review 43(1): 67-80
Beach, D. and Pedersen R. B. (2013) Process-Tracing Methods: Foundations and Guidelines. Ann Arbor MI: University of Michigan Press
Bohm, R. M. (1982) Radical criminology: An explication. Criminology 19(1): 565-89 Bryman, A. (2008) Social Research Methods. Oxford: Oxford University Press
Chambliss, W. J. (1976) The state and criminal law. In Whose law? What order? ed. W. J. Chambliss and M. Mankoff. New York: Wiley
Chirot, D. and Hall, T. D. (1982) World system theory. Annual Review of Sociology 8(1): 81-106 Cloward, R. A. and Ohlin, L. E. (1960) Delinquency and opportunity: A theory of delinquent gangs. New York: Free Press
Collier, D. (2011) Understanding Process Tracing. Polticical Science and Politics. 44(4): 823-30 Davies, J. C. (1962) Toward a theory of revolution. American Sociological Review 27(1): 5-19 Durkheim, E. (1947) The division of labor in society. Translated by George Simpson. New York: Free Press
Durkheim, E. (1950) Suicide: A study in sociology. Translated by George Simpson. New York: Free Press
Eisner, M. (2001) Modernization, self-control, and lethal violence: The long-term dynamics of European homicide rates in theoretical perspective. British Journal of Criminology 41(4): 618-38 Elias, N. (1939 and 1978) The civilising process. 2 vols. Translated by Edmund Jephcott. New York: Urizen Books
Gurr, T. R. (1974) Persistence and change in political systems, 1800-1971. American Political Science
Review 68(4): 1482-1504
Gurr, T. R. (1981) Historical trends in violent crime: A critical review of the evidence. Crime and
Justice: An Annual Review of Research 3(1): 295-350
Gurr, T. R., Grabosky, P. N. and Hula, R. C. (1977) The politics of crime and conflict: A comparative
22 Gurr, T. R., Jaggers, K. and Moore, W. (1990) The transformation of the Western state: The growth of democracy, autocracy, and state power since 1800. Studies in Comparative International
Development 25(1): 73-108
Gutierrez, M. E. (2011) Violence, In Mexico? Homicide in a Democratizing Society. Berkeley, CA: University of Berkeley
Hopkins, T. K. and Wallerstein, I. (1981) Structural transformations of the world economy. In
Dynamics of world development. ed. R. Rubinson. Beverly Hills, CA: Sage
Jaggers, K. and Gurr T. R. (1995) Tracking democracy's third wave with the Polity data. Journal of
Peace Research 32(4): 469-82
LaFree, G. (1998) Losing Legitimacy: Street Crime and the Decline of Social Institutions in America. Boulder, CO: Westview
LaFree, G. and Drass, D. A. (2002) Counting crime booms among nations: Evidence for homicide victimization rates, 1956 to 1998. Criminology 40(4): 769-99
LaFree, G. and Tseloni, A. (2006) Democracy and Crime: A Multilevel Analysis of Homicide Trends in Forty-Four Countries, 1950-2000. Annals of the American Academy of Political and Social Science 605(1): 26-49
Lodhi, A. Q. and Tilly C. (1973) Urbanization, crime and collective violence in nineteenth century France. American Journal of Sociology 79(2): 296-318
Marshall, M. G. and Jaggers K. (2003) Polity IV Project: Political Regime Characteristics and
Transitions, 1800-1999. Dataset users manual. College Park: University of Maryland, Center for
International Development and Conflict Management
Merton, R. K. (1938) Social structure and anomie. American Sociological Review 3(5): 672-82 Messner, S. F. and 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 Montinola, G. R. and Jackman, R. W. (2002) Sources of corruption: A cross-country study. British
Journal of Political Science 32(1): 147-70
Neuman, W. L. and Berger, R. J. (1988) Competing perspectives on cross-national crime: An evaluation of theory and evidence. Sociological Quarterly 29(2): 281-313
Osterberg, E. (1996) Criminality, social control, and the early modern state: Evidence and interpretations in Scandinavian historiography. In The civilization of crime: Violence in town and
country since the middle ages. ed. E. A. Johnson and E. H. Monkkonen. Urbana: University of Illinois
Press
Smelser, N. (1962) Theory of collective behavior. New York: Free Press
Snyder, R. and Duran-Martinez, A. (2009) Does illegality breed violence? Drug trafficking and state-sponsored protection rackets. Crime Law and Social Change 52(3): 253-273
23 Spierenburg, P. (1996) Long-term trends in homicide: Theoretical reflections and Dutch evidence, fifteenth to twentieth centuries. In The civilization of crime: Violence in town and country since the
middle ages. ed. E. A. Johnson and E. H. Monkkonen. Urbana: University of Illinois Press
Spitzer, S. (1975) Toward a Marxian theory of deviance. Social Problems 22(5): 638-51
Sung, H-E. (2004) Democracy and organized crime activities: Evidence from 59 countries. Security
Journal 17(4): 21-34
Sung, H-E (2006) Democracy and Criminal Justice in Cross-National Perspective: From Crime Control to Due Process. Annals of the American Academy of Political and Social Science 605(1): 311-337 Tilly, C., Tilly, L. and Tilly, R. (1975) The rebellious century, 1830-1930. Cambridge, MA: Harvard University Press
Useem, B. (1985) Disorganization and the New Mexico prison riot of 1980. American Sociological
Review 50(5): 677-88
Villarreal, A. (2002) Political Competition and Violence in Mexico: Hierarchical Social Control in Local Patronage Structures. American Sociological Review 67(4): 477-498
Wallerstein, I. (1979) The capitalist world economy. Cambridge: Cambridge University Press
Electronic sources
Ace Project (2015) Legislative Framework. [online] Available at:
<http://aceproject.org/main/english/lf/lfy_mx.htm> [Last accessed 17 June 2015] Center for Systemic Peace (2014) The Polity Project. [online] Available at: < http://www.systemicpeace.org/polityproject.html> [Last accessed 11 June 2015]
LaFree, G. and Tseloni, A. (2006) Democracy and Crime: A Multilevel Analysis of Homicide Trends in
Forty-Four Countries, 1950-2000 [image online] Available at: <
http://www.jstor.org.proxy.uba.uva.nl:2048/stable/pdf/25097798.pdf?acceptTC=true> [Last accessed 23 June 2015]
Punton, M. and Welle, K. (2015a) Annex Applying Process Tracing In Five Steps. In Centre For
Development Impact Practice Paper 10. [PDF] Institute of Development Studies. Available at:
<http://opendocs.ids.ac.uk/opendocs/bitstream/handle/123456789/5997/CDIPracticePaper_10_Ann ex.pdf?sequence=2> [Last accessed 9 June 2015]
Punton, M. and Welle, K. (2015b) Straws-in-the-wind, Hoops and Smoking Guns: What can Process Tracing Offer to Impact Evaluation? Centre For Development Impact Practice Paper 10. [PDF] Institute of Development Studies. Available at:
<http://opendocs.ids.ac.uk/opendocs/bitstream/handle/123456789/5997/CDIPracticePaper_10.pdf;j sessionid=BB71EC667173AEAD7361195D3CD10FCD?sequence=1> [Last accessed 10 June 2015] UN Data (2015) Intentional homicide, number and rate per 100,000 population. [online] Available at: <https://data.un.org/Data.aspx?d=UNODC&f=tableCode%3A1> [Last accessed 17 June 2015]
