Homicide Declines, 600-2060 AD: a Generalising Framework

Homicide Declines, 600-2060 AD: A Generalising Framework

Manuel Eisner Institute of Criminology University of Cambridge 16 Theories of the Crime Decline

1. Incarceration 2. Better policing (Sherman, Zimring) 3. Demographic trends (Blumstein) 4. Abortion 5. Lead/environmental neurotoxins (Nevin) 6. Unemployment, consumer sentiment, 7. Inflation (Rosenfeld) 8. Self-control (Eisner/Pinker) 9. Psycho-pharmaceuticals (Finkelhor) 10. End of crack/cocaine epidemic 11. State legitimacy (Roth/LaFree) 12. Divorce – family stability 13. More effective public health/crime prevention policy (Finkelhor) 14. Changing norms and values 15. Dissipation of effects of cultural revolution (Pinker) 16. Technical Crime Control (Clarke et al.)

Methodological Approaches

1. National Case Studies – One period, one country: – Typical answers: Compstat; drugs/crack epidemic, abortion, incapacitation

2. International Trend Analyses – Several places, on period – Typical answers: anomie (Messner/Rosenfeld), loss of legitimacy (LaFree), Postmodernity (Young).

3. Historical Cross-cultural Generalising Perspective – All violence declines, at all times, in all places.

Is there a limited number of universal mechanisms behind all major declines of interpersonal violence in all societies at all times? i.e.

How do societies become less conflict-ridden and homicidal places?

A Homicide Scale Violence as Politics 1000 > 200 per 100,000 • Unrelated men in public space Iraq after Invasion • Organized and functional • > 100 Civil War South African Townships Private justice & revenge • Violent enterpreneurs

Ciudad Juarez, Mexico • Fights over goods and territories

100

10-100 Violent Societies

10 1-10 Semi-Pacified Societies Murders per 100,000 pop

< 0.3 per 100,000 • Large % domestic, female vict. Japan in 1970-90s • Disorganized and disfunctional < 1 Pacified England, 1920s • Individual pathologies Societies Denmark, 1950s • Marginal groups

0.1 Violence as Pathology Buss, D. M. and T. K. Shackelford (1997). "Human aggression in The Evolutionary Perspective evolutionary psychological perspective." Clinical Psychology Review 17: 605-619.

What purposes does violence amongst men serve? -> adaptionist, strategic, functional view informed by evolutionary theory (Archer, Pellegrini, Pinker, Buss/Shackelford, Tremblay)

• Proactive instrumental violence Violence as a means of gaining access to scarce resources, particularly goods, power/status, sex – Examples: raids, robberies, private taxation, piracy, staged fights, attacks on same-sex competitors.

• Reactive retaliatory violence Violence as a means of defending against aggressors, producing protection, maintaining norms of reciprocity – Examples: vendetta, revenge killing, vigilantism.

Under what circumstances does killing (a group member) becomes a strategically less attractive option?

Some hypotheses:

– A high availability of law reduces the chances that people perceive the need to use violence as a means of self-help (retaliation and revenge)

– A high legitimacy of social order increases the chances that people believe they can fulfil desires in cooperation with others

– A high level of protection from risks increases the chances of investing into building up durable relationships

– A cultivation of self-restraint and manners reduces the chances for friction resulting in violence

Decline 1 Elite Violence, 600-1800 AD Eisner, Manuel (2011). The Decline of Regicide in an Evolutionary Perspective, British Journal of Criminology, 2011.

Why? • The involvement of elites in intra-societal violence is not a constant. • Theorists such as Elias, Cooney, Black expect a decline in elite violence.

What? • All monarchs of all major dynasties across all of Europe from 600 to 1800 AD (N = 1820). • Coded for cause of death and further particulars.

Trend in Regicide Rates, 600-1800 AD 50-Year Periods

3000 Note:

Average Regicide Rate 2500 1060 per 100,000 (ruler-years)

2000 General Population Today 1 per 100,000 (person years) 1500

Rate perRate 100,000 1000 R² = 0.4598

500

0 649 699 749 749 849 899 949 999 ------1149 1199 1049 1099 1249 1299 1349 1399 1449 1499 1549 1599 1649 1699 1749 1799 ------600 650 700 750 800 850 900 950 1100 1150 1000 1050 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 Regicide Rates by Region, 600-1800 AD

4500 East: Byzantium, 4000 Ottoman Empire, Russia, Bulgaria

3500

North 3000 Scandinavian Monarchies, England, Scotland 2500 All East 2000 North West&South West&South All others 1500 Rate per 100,000 MonarchYears

1000

500

0 7th 8th 9th 10th 11th 12th 13th 14th 15th 16th 17th 18th Why did Regicide Decline?

• Fewer internal rivals with sufficient military power to challenge royal authority.

• Decreasing probability of successful acquisition of power through usurpation/murder.

• Power transition becomes more law-bound (e.g. inheritance laws).

• Change in elites’ conduct of life (“from warriors to courtiers”).

• Higher stability and legitimacy of state institutions.

Decline 2 Homicide in Europe, 1300-2000

• Eisner, M. (2003) Long-Term Historical Trends in Violent Crime, Crime and Justice.

• Johnson, E and E. Monkkonen (eds.) (1996). The Civilization of Crime; Violence in Town and Country since the Middle Ages. Urbana: Illinois University Press.

• Spierenburg, P. (2008) A History of Murder. Cambridge: Polity Press.

Homicide Rates across Europe, 1300-2000

50 England Netherlands 40 Scandinavia 30 Germany Italy 20 Homicide perRate 100,000 10

0 1300 1400 1500 1600 1700 1800 1900 2000

See Eisner (2003). Long-Term Historical Trends in Violent Crime, Crime and Justice Why did Homicide Decline?

• Increasingly stable state monopoly of power

– Elias (1978 [1939]) The civilizing process.

• Growing protection decreases necessity of violence as self-help

– Black (1983). Crime as Social Control.

• Interdependence and social complexity increase benefits of self-control

– Elias (1978 [1939]) The civilizing process.

• Disciplining institutions promote impulse control

– Krieken (1989). Violence, Self-Discipline and Modernity: Beyond the Civilizing Process

• Growing legitimacy and integration of state

Regicide and Homicide

250

200

England 150 Netherlands Scandinavia Germany Rate per 100000 100 Italy Regicide

0 600 800 1000 1200 1400 1600 1800 2000 Any way to test a theory empirically?

How about this hypothesis:

The decline was a result of regular frontal-lobe training in self- control by means of reading books? Book Production in Europe 600

500

Great Britain

400 Ireland Belgium Netherlands 300 Germany Switzerland Italy Books per 1000 per 1000 Books population 200 Spain Sweden

100

0 1454 - 1500 1501 - 1550 1551 - 1600 1601 - 1650 1651 - 1700 1701 - 1750 1751 - 1800 Source: Baten, J., & Van Zanden, J. L. (2008). Book production and the onset of modern economic growth. Journal of Economic Growth, 13(3), 217-235. The Books-Murder Connection, 1500-1800 45

20 Murder

Homicides per per population 100,000 Homicides 10

0 0 100 200 300 400 500 600 Books per 1000 Population

Note: All available pairs of 50-year estimates of book production and homicide rates, 1500-1800 N = 46.

… this looks like an elasticity

The Books-Murder Connection, 1500-1800 100

y = 28.804x-0.363

R² = 0.4464

10 Murder Power (Murder) Homicides per per population 100,000 Homicides

1 0.1 1 10 100 1000 Books per 1000 Population

Note: All available pairs of 50-year estimates of book production and homicide rates, 1500-1800 N = 46.

Decline 3 Homicide in the USA, 1600-1800 Roth, Roland (2009). American Homicide. Cambridge (Mass): Belknap.

Empirical Observation: Parallel to the development in Europe, Homicide rates among unrelated adults in New England, New Netherlands, Virginia and Maryland fell from over 100 per 100,000 to about 1-5 per 100,000 between the early 1600s and around 1790.

The Core Roth Hypothesis

• Both in the USA and in Europe there are four correlates of declining homicide amongst unrelated male adults. 1. The belief that government is stable and that its legal and judicial institutions are unbiased and will redress wrongs and protect lives and property. 2. A feeling of trust in government and the officials who run it, and a belief in their legitimacy. 3. Patriotism, empathy, and fellow feeling arising from racial, religious, or political solidarity. 4. The belief that the social hierarchy is legitimate, that one’s position in society is satisfactory and that one can command the respect of others without resorting to violence.

Decline 4 The Victorian Success Story • National series available in many countries • Victorian decline of homicide much studied by early generation of criminologists (Boschi, Durkheim, Ferri).

The Victorian ideal of the The bad drunk husband family Homicide Trends, 1840-2005

10 Austria Spain Italy 8 Switzerland Belgium France 6 Germany Netherlands Denmark 4 Norway Sweden Ireland England and Wales 2 Scotland

0 1840 1860 1880 1900 1920 1940 1960 1980 2000 Explaining the Victorian Success Story

• Cultivating character – Wiener, Martin (2004,) Men of Blood. Cambridge: Cambridge University Press

• Social/national/political Integration

• Rise of control bureaucracies – Emsley, Clive (2007) Crime, Police and Penal Policy: European Experiences, 1750-1940.

• Temperance, self-control movements – Carter-Wood, J. (2004), Shadow of our Refinement.

• Elimination of collectivist honour (i.e. self-help) cultures. – Durkheim, Emile (1957 [1899] Professional Ethics and Civic Morals – Gallant, Decline 5 The Homicide Decline Since the Early 1990s

• Are some of the mechanisms associated with declining homicide also present during the past 20 years? – Is it broad rather than a locally limited phenomenon? – Is it significant? – Does is share characteristics of larger and longer declines?

• Homicide data for 22 Western countries from 1950 to 2008, – Based on WHO and LaFree Dataset

Mean (1950-2006) = 100 for each series; Scoret = homicide ratet*100/mean

250 Austria Belgium Denmark Finland France 200 Germany Hungary Ireland Netherlands Norway 150 Poland Portugal Spain New Zealand USA 100 Canada Australia Sweden Italy Switzerland 50 England and Wales Scotland Mean-1SDEV Mean Mean+1SDEV 0 1940 1950 1960 1970 1980 1990 2000 2010 2020 Mean (1950-2006) = 100 for each series; Scoret = homicide ratet*100/mean

250 Austria Belgium Denmark Finland 200 France Germany Hungary Ireland Netherlands 150 Norway Poland Portugal Spain 100 New Zealand USA Canada Australia Sweden 50 Italy Switzerland England and Wales Scotland 0 1940 1950 1960 1970 1980 1990 2000 2010 2020 Timing of Peaks and Troughs , 1950-2008

% % Trough Peak Trough Trough Peak Trough Increase T-P Decrease P-T Austria 1963 1983 2008 0.94 1.71 0.58 83 -66 Belgium 1963 1999 2007 0.46 1.31 0.77 186 -41 Denmark 1960 1992 2005 0.46 1.31 0.77 186 -41 Finland 1965 1993 2007 1.95 3.22 2.10 65 -35

France 1952 1984 2007 0.52 1.29 0.66 146 -49 Hungary 1964 1992 2006 1.67 3.95 1.81 137 -54 Netherlands 1955 1996 2008 0.28 1.27 0.87 348 -31 Norway 1954 1987 2008 0.32 1.41 0.73 335 -48 Poland 1965 1992 2008 0.91 3.04 1.32 234 -57 Spain 1963 1984 2008 0.10 1.12 0.75 1055 -33 New Zealand 1957 1991 2001 0.55 2.07 1.42 276 -31 USA 1957 1992 2008 4.70 10.20 5.33 117 -48 Australia 1950 1989 2006 1.00 2.07 1.27 107 -39 Sweden 1961 1990 2007 0.58 1.42 0.93 145 -34 Italy 1969 1990 2007 1.03 2.47 0.89 139 -64 Switzerland 1969 1992 2007 0.68 1.47 0.70 117 -53 England and Wales 1959 2002 2009 0.59 1.75 1.38 199 -22 The Homicide Decline Since the Early 1990s

• Similar periodicity across Western countries Mean Trough: 1959 Mean Peak: 1992 Mean Trough: 2007

• TSFA (time-series factor analysis) from 1950 to 2006: – About 50 % of the variation across all countries shared by one underlying factor.

• Similarity holds across countries with very different national trajectories in – Policing – History of substance abuse – Unemployment – Imprisonment – Left/right political tradition

Decline 6 Looking into the future

• Understanding = controlling USA mechanisms = predicting outcomes. 14.00

12.00

• We know that liberal democratic 10.00 capitalist countries have homicide rates of about .3-.4 per 8.00 100,000. USA 6.00

• How do we get there in the next 4.00 30 years? 2.00

0.00 1940 1960 1980 2000 2020 2040 2060 To conclude: some problems

• How many cases are we looking at? – 22 cases with 22 explanations or - 1 trajectory with 22 (or more) variations

• At what level of disaggregation do we need to look for explanatory variables?

• What explains the commonalities in – Timing of trough, peak, and decline?

• What are those commonalities and are they similar to previous declines? – Increasing political stability (the ‘victory’ of the capitalist/democratic model of society) – End of ‘self-fulfilment’, laissez faire parenting ideals

Some final thoughts

• Necessity of theory – A theory that doesn’t contradict consolidated developmental, situational, biological, etc. knowledge is not a disadvantage.

• Contextual Meaning of Predictive Indicators – All macro-level indicators have multiple meanings. Moreover their meaning can change over time and between contexts.

• Limitations of Quantitative Models – In most cases the conditions for rigorous causal tests (e.g. TS analyses of Granger causality) are not given • Lack of precision of outcome and predictor measures, lots of noise • Multiple meanings of indicators • Unknown ‘transfer function’ between cause and effect. • No data to test micro-level ‘thread needle’ of macro-level hypotheses