good morning everyone and i wish you a good day of the italian republic i'd like to welcome in brand rams razoo my name is princess capacio i work for la stampa a daily i work on foreign policy especially on the middle east for many years and i'm also involved with immigration migration has had an impact on the traditional barriers between domestic and foreign policies become a whole sector also in view of the fear for terrorism imran rasool is a professor of economics at university college in london and he just actually does many more things and his lecture is about justice domestic discrimination well as a journalist i prepared a short introduction to the slides is going to present to us and then of course you're welcome to ask questions talking about justice justice and ethic discrimination is a monster topic a bit like matroska's it contains many other themes is just like a chinese box in order to introduce him ron rasul's work i've tried to use the definition given by the investigation commission on the hatred crimes at the joe cox committee in italy and they quoted a pyramid at the base of the pyramid we have the normalization of hostile language or hatred language towards the muslims and the at the top of the pyramid you have the crimes and discrimination and crimes uh proper and felony and then the sentences and the punishments that um in brothers who will get into more detail as regards the base of the pyramid that's where i worked um especially over the last few years starting from the middle east to the mediterranean migration so and so forth as regards italy but that applies to many other countries in the whole west i can say but talking about italy the the change occurred in september 2001 at that time the other started being identified in a more specific way up to then of course italy had experienced the manifestations of racism xenophobia the northern league is a party already existed but the fear and the diffidence was addressed towards the which is the migrants especially coming from north and africa i should ask that before 9 11 2001 italy was a young country in terms of immigration the number of immigrants were not many and actually up to the 70s italy was a country of immigration so italians were going abroad and starting in the mid 70s italy started receiving migrants from other countries so after 9 11 2001 what whatever used to be a lack of trust and suspicion for the strangers and foreigners turned into something like looking at islam and from that moment fear started spreading suspicion and control on the part of the law enforcement community and policing was focused on mosques and muslims actually became the others it doesn't mean foreigners because after 9 11 even attacks in london and madrid show that the other is not necessarily a foreigner can even be born in this country maybe a fellow citizen a resident but it's a culturally and resil religiously other different in terms of the the way they dress they pray and the way the customs in religious and cultural terms so the other becomes something different other than our civilization some all huntington the clash of civilizations let me quote from this over and above that there's something already present in our societies so the change of pace let me call it that way happened in that moment we perceived a threat against our civilization at the same time we felt we felt just legitimated to fight back so there was a gradual acceptance of a language which actually becomes racist and discriminatory and i i said acceptance because well before then well i i don't mean to say that suspicion was not there but there was a kind of prudence in a way in italy after the second world war on the ruins of racism and racial religious discrimination the shoha had sort of turned certain words and certain language into taboos and then the booming economy in europe was globalization the idea we all belonged to a global village and that of course led to an excessively politically correct language but we had to coexist and accept one another that was the model but the fear for islam threatening our civilization and asking for self-defense and fighting back actually authorized the use of a language that up to them was not tolerated and people were actually ashamed of or up to the actual races magazinophobia because confrontation with islam changed the rules of the game it was no longer about race or blood but it's apparently something more noble culture civilization religion and west the west at large and acceptance occurred at all levels because these themes were not only promoted by the right or the extreme right but in the wake of the perceived threat of islam they were taken over by the right at the left like in france of course islam is a religion threatening the conquest of the liberal left or just think when i worked on that for the first time i was really surprised but i thought it made sense imagine a pen in france and around her little by little a an aggregation developed of the french homosexual community in the wake of the fear um of islam i mean for islam to attack the french lai site the so the secularism the typical french secularism this was not just limited to europe it happened you know in the united states and brown roswell who will talk about it you might think of trump just banning some countries just because they were muslim not just not because they had ever attacked the united states in any way and we're coming to the second level of the pyramid switching from language to crimes hatred crimes and discrimination proper let me just to mention just to give an idea figures don't tell the whole story but they can give you an idea the data of the racial anti-discrimination or office with the oecd showed that in italy in 2016 out of 200 2 000 discrimination actions 65 percent where uh for racial discrimination in 2015 out of 155 hatred crimes or 369 were linked to xenophobia about 50 percent and then led to an increase in a number of court cases of the last three years appeal courts in italy so an increase of reports of crimes for rome 100 in bologna 157 milan over 130 in catania 14 baris seven and the south in spite of its contradictions and problems seems to be stronger this sense so coming to the recent years and the economic uh crunch crisis 2007 8 which actually made the middle class poorer and when you get poorer of course people start being afraid and feeling scared they have to look for a scapegoat and that escape court was already there because at that point it was no longer muslims the migrants were coming in but there was a widespread acceptance of scapegoating other people without having to justify to account for it as politically correct and from this viewpoint just think of italian news last year in 2017 and the beginning of the year support for the lower second generation immigrants in italy was about 60 percent italians were not against it at the end of last year after a long electoral campaign after until march the 4th the consensus had dropped to half of that just think of the italian electoral campaign that was played on that i'm not saying it's right or wrong but it's a fact it was actually played absolutely on fear for migrants and in the invasion of migrants while the previous well we're talking about 190 000 people coming to this country so in theory these numbers would wouldn't justify such a fear for an invasion especially in a situation where um official data shows that crimes in italy for all kinds have been declining for the last 20 years and we have reached the lowest point since the unification of italy and the real perceived context doesn't always match this data so 911 was a cornerstone really the way we get controlled we move uh with feel things and but the politically correct approach has been set aside and language has no limitations in calling other people enemies and different and then the credit crunch and economic crisis came along and we felt justified in behaving that way the media have actually been riding this wave and this perception of a correlation between between migration and the other the muslim the refugee the migrant potentially uh the author of terrorist attacks but in italy we had no such attacks so these people were actually identified low skills people taking jobs away from italians so a different kind of threat that was accepted by the media and the diffused by the media because of course that made the audience declining audiences on the radio and tv uh led to a situation where you try to attract audience back even with fake news if you remember the year before the weinstein's case was full of reports in hollywood of middle eastern actors that were fed up with being hired only to play the roles of kamikaze or terrorists and so on and so forth so let me come to my conclusions and then like i'll float in bran razor this of course leads to more reports more controls more law enforcement activities more policing that's a vicious circle that develops and we should actually count to the number of migrants in italian prisons because that turns into an electoral door opener in european jails foreigners are about 21 in italy 32 because in our law we criminalize irregular or illegal migrants even if they have committed no crime but regular legal migrants be they muslims or not are a very limited percentage accounting for only 10 percent of the foreign inmates and three percent of the total so so there are less of them in jail than italians but that's just numbers the perception is different and you tend to correlate the crime and the other as a person so it's very hard getting rid of this perception last point preventive incarceration is more easily applied to foreign migrants than italy and then of course taking us back to 9 11 suspicion what to do there's been a tightening of rules that the percentage of foreigners in preventive incarceration and total number of sentences is 28 percent versus uh 31 in total so preventive incarceration is more frequently applied to migrants of foreigners than each italian so that's um a sort of catch-22 situation there's a reason why we've come to this point and we've come to the slides that tim ron result is about to show us and it's the spirit of our times as well just think of the new government that's just been inaugurated the first theme tackled by the new uh home minister was reducing spending on for migrants and this is not just a result of an electoral campaign but it's actually the most popular thing to do and the least expensive actually so it's easier to cut these expenses an important area to be um tackled by politicians and the media as well is also fueled by the internet and the social media well the internet and social media beware they're just megaphones loud speakers of something already there and in the guts of people and it's worth thinking about thank you very much and i give the floor to imran rasool thank you francesca for um the very insightful introduction um and let me just start by saying it's a pleasure to be here and be able to present some of my work um today which is examining behavior in the criminal justice system in the united states so this is a country which is very much built on immigration and is probably one of the most culturally and ethically diverse countries on the planet but what i want to understand today and and to present to you is some work relating to behavior by very senior judges in the criminal justice system in the united states so before doing so i want to just put the us criminal justice system into context and just to give you some statistics about how this system delivers outcomes so the first statistic i wanted to present to you is just to show you that the united states incarcerates more individuals as a share of its population than any other country on the planet let me start by showing you the figures for a range of different oecd countries you can think of this as the as the share of the population um currently in in in prison or in some detention center so the us incarcerates more individuals um than any other country on the planet the us is obviously the top um country shown in the figure i've also highlighted italy towards the bottom of the figure which as a share of its uh population has about uh one-fifth the likelihood of somebody being um in prison than in the united states the the uk is is more similar to italy than than to the united states in fact that means that about a quarter of the people on the planet who are currently behind bars are in the united states whereas the u.s only has five percent of the world's population okay so this is a very very um extreme system on that dimension and of course i'm going to focus my discussion today on men the vast majority of prisoners in most countries are men what i want to focus in particular is how men from certain minorities are much more likely to come into contact with the criminal justice system than the native white population okay so although there are there are very extreme levels of uh incarceration they're actually very different for different um ethnic groups in the united states so for example we have some very shocking statistics by different uh ethnic group so currently about one in nine black men that's almost 10 of the of the population of black men aged between 20 and 34 are currently in prison we think about the share of men who expect to be imprisoned at some point in their adult lives that's one in three black men however it's been less recognized that hispanics are actually the fastest group being incarcerated in the us and currently in the most senior level of the criminal justice system that i'll be examining today about 40 percent of prisoners are of hispanic origin and so relative to white individuals hispanic men are four times as likely to go to federal prison at some point in their lives as white individuals what i would really like to try to understand in this research agenda is to understand what's driving these differences is this down to discrimination of different legal actors either the police judges prosecutors or is there some other explanation that is actually relevant for us to understand it might not all be a matter of discrimination obviously these huge differences between the way that different races and ethnic groups come into contact with the criminal justice system has huge impacts beyond just the uh the the impacts in in terms of prison so for example if individuals anticipate that they're going to be um incarcerated or be be come into contact with the with the police this has been shown to affect children's behavior say in schools or young adults incentives to acquire skills because they don't think they're necessarily going to be able to use those skills in the labor market these types of experiences with the criminal justice system have also been shown to have pretty devastating effects on marriage markets in amongst different minority groups where you have a huge absence of men of marriageable age or men who have young children so we can expect this to have impacts on families on children going forward into future generations and obviously for the individuals themselves these rates of incarceration will then have impacts in terms of their ability to come back into society to contribute say through their labor market um earnings and activities so there are a whole variety of reasons why we should care to understand what causes these big differences but what i want to also emphasize is the us criminal justice system is also incredibly severe in terms of the penalties that it um that it provides so of individuals who come up in front of judges 90 of them go to prison very few people are found not guilty or given a non-custodial sentence in terms of the average amount of time that individuals spend in prison this is for the federal criminal justice system this is the highest tier of the us system just below the supreme court on average for whites who go to prison they go to prison for about 41 months the average black male goes to prison for 83 months more than double and hispanic individuals go to prison for about 42 months quite comparable to whites however from an economic point of view what we'd really like to try to understand is if we compare like for like individuals if we imagine two individuals who've committed identically the same offence at the same moment in time they have the same quality of legal counsel they have the same age they have identical criminal histories they have the same marital status same number of dependents and they're being tried in a very similar court we compared two individuals who are identical in all of those dimensions the only difference is their ethnicity then we can say well what what are the differences the conditional differences and how the criminal justice system is treating them and that's what's shown in the figure in the bottom right hand corner so for two individuals that have identically the same uh characteristics so we're really making a like for like comparison then what we observe is that a black male when they go to prison they're sentenced for four months longer than a white individual and hispanic individuals are also sentenced between four and five months longer for identically the same offense and all those others like for like comparisons than a white individual so the key question that we're trying to understand is whether these differences for two otherwise identical individuals is driven by discrimination or it's driven by some other factor where the judges observe something about these individuals that might justify these differences so in order to understand whether those differences are driven by discrimination or something else we consider how judges change their behavior around 9 11. okay so we're going to think about 9 11 as being a very large shock to the u.s sort of cultural um set up it's an event that was going to be unanticipated and without doubt it's been documented by a range of human rights organizations and academics that 11 caused an increase in animosity that individuals felt towards muslims the figure there just shows you um what happened to hate crime towards muslims the types of crime that francesca uh mentioned in her introduction and we see a spike in 2001 that there was increased animosity increased aggression that we can measure statistically it doesn't just come from media reports this is as documented in fbi data so without doubt 9 11 in the immediate aftermath of 9 11 created animosity increased animosity towards muslims in the us what we want to try to understand in our research is whether that has impacts on non-muslim minorities uh within the federal criminal justice system in particular our sentencing outcomes for blacks and hispanics impacted by this increased animosity towards one other minority group okay so what we're trying to understand is whether animosity towards one minority group then impacts outcomes for other minority groups okay why might there be such interlinkages between how we as a society feel towards one particular minority group and how senior judges then treat other minority groups and here there are well-known theories from psychology to suggest that how we impact how how we think about one minority group can have such um spillover or interlinkage effects on how we then treat other minority groups francesca alluded to this in in her introduction when she contrasted immigrants and terrorists and how these two issues typically get conflated in public debate and in in the minds of individuals and so in psychology we have very well established theories one such theory is what's known as um contagious animosity that says if you feel more animosity towards one particular group then that can potentially cause you to have more uncontrolled or subconscious feelings of animosity towards other minority groups what we want to try to understand is when there's a spike in animosity just after 9 11 towards muslims whether that then impacts the way that very senior judges are then treating other minority groups who happen to be in the criminal justice system at the same moment in time okay and in particular we might think that hispanic defendants are more impacted by these interlinkages with muslims than for other minority groups for a whole variety of different explanations that i want to now come to so the first is uh political rhetoric here i've given a quote from a from a senior u.s politician just after um 9 11 which makes clear that in the minds of politicians in the minds of the media and potentially therefore in the minds of the general public there is a linkage between immigration policies and our vulnerability to terrorist activities this is despite the fact that all of the individuals involved in the 911 atrocity actually all entered the us uh legally this is a very stark quote but as francesca alluded to in a introduction these types of interlinkage between a fear of muslims a fear of terrorism is often very closely intertwined with language and debate around immigration i've just highlighted sort of the most recent sort of form in which this is taken which is sort of the proposals by donald trump to build his wall along the southern border so that's the first dimension in terms of political rhetoric the second dimension is in terms of policy so even before 9 11 in fact in the immediate aftermath of the oklahoma terrorist attacks various pieces of legislation were passed in the u.s which again increased the interlinkage between crimes related to immigration and crimes related to terrorism so that leads individuals who are involved in the criminal justice system from legal advisers judges and other key individuals to have in their mind in the background these important linkages between these somewhat different spheres of policy and here i've given another particular quote from the attorney general in the in the us at the time again making explicit this linkage but this process started actually well before 9 11 in response to the oklahoma terrorist attack in 1995. and the third key dimension that causes terrorists and immigration offenses to become interlinked is the creation of institutions okay so subsequent uh to this period what we observe is the creation of the department of homeland security which is the first time in us history that the same government agency is responsible for protecting the homeland and at the same time also dealing with immigration policy traditionally immigration policy was actually housed within the department of labor in the us but in the wake of terrorist attacks these these policies actually became implemented and housed within the same federal agency the department for homeland security and agreement again that strengthens the narrative between concerns over terrorist attacks and fears over over immigration and it's through those three dimensions of interlinkage that we might think that potentially individuals and judges may be influenced in how they then treat hispanic individuals in the wake of 9 11 which is very much a terrorist event to set the scene for for that potential possibility what i want to show you is again some statistics all from before 9 11 about sentiment that individuals felt towards different minority groups in the in the us this is taken from a nationally representative sample and collected by political scientists and these statistics are referred to as thermometer readings so in order to derive these statistics we take individuals in society and we ask them how much affinity do you feel towards certain groups towards whites towards hispanics blacks towards feminists towards environmentalists we have a whole sequence of different groups and then from that we can construct a measure which is known as our sort of thermometer reading which is how the average white individual in the us how much affinity or or um animosity do they feel towards other other groups so here i'm just going to show you those statistics all from before 9 11. these statistics are actually calculated in 2000. so you can see for whites everything is normalized by white so whites feel zero affinity towards themselves so in terms of the elderly they're the group who are most favored in terms of a minority group in the us in 2000 and then for the other groups you can see increasing degrees of animosity felt by whites towards these other minority groups this is something that surprised me but there's actually more animosity towards hispanics on the eve of 9 11 than there is towards blacks it's not as bad as the animosity felt towards gays and lesbians people on welfare or christian fundamentalists but in terms of differences between ethnic groups hispanics have more animosity they have lower thermometer readings than blacks and that this statistic is replicated going back to the 1970s actually so this is amongst the average population what we want to understand is are very senior and professional judges also influenced by these types of feeling of animosity around 9 11. okay for the average citizen we see that immediately after 9 11 there are start to be changes in people's perception so the left hand side figure shows you data from gallup before and after 9 11 showing you how sort of people's views about whether they think immigration should be restricted and if you look at the light blue line that shows the share of respondents of this gallup all over time that think that immigration should be decreased and around 9 11 the the the vertical black uh sorry the vertical dash line there you see a big jump up in the share of individuals who think that immigration should now be restricted in some way the right hand side figure shows you crimes of vandalism then committed against different ethnic groups in the us where the blue line shows you crimes committed against hispanic individuals and so we see a spike in that uh vulnerability to crimes of vandalism for example immediately after 9 11. so this all seems to suggest that animosity increased not just towards muslims but also towards immigrants and potentially hispanics immediately after 9 11. does this feed through into how people are treated in the criminal justice system so as i've mentioned we're going to study this question in the federal criminal justice system this is the highest tier of the american judiciary this is above the state levels it's only one level below the the supreme court this is a very high-stakes professional environment we have very senior judges who are appointed for life by the president to this tier of the judiciary at any at any one moment in time there are only around 700 federal judges who are operating that that's the degree of selection that we have of very high quality individuals and they're operating in a system in which sentencing guidelines are in place so the discretion that judges have is somewhat limited and should only be determined by the severity of the offence that's been committed and a defendant's criminal history so let me just describe a little bit the system that the judges are operating in so then we can see what how they might change their behavior so this is a an indication of the sentencing guideline so at any moment in time when i observe a case a judge has to be told what is the offence level of that individual and the offence is categorized from level one to level 43 there's a lot of variation in offense levels where level 43 are the most serious offenses level one are the least serious offenses and at the same time a defendant is characterized by their criminal history how frequently they've been in prison in the past how long they've been to prison for and again we have a system where we can define an individual at the time when they when they come up before the judge in terms of six different levels of criminal history where you can range from group one up to group six okay and so for any given combination if an individual has an offense level of severity 22 and their criminal history is group 4 then the judge is typically required to sentence them to prison for between 63 and 78 months the system doesn't allow judges much discretion and this system was put in place in the u s to exactly try to prevent discrimination from judges by saying you take this information as given the discretion that you have is where in this boundary between 63 and 78 months you're going to sentence an individual however judges do have one level of discretion which is to say i actually disagree with the offense level and the criminal history that's been presented to me by the legal teams and so a judge has the possibility to assign somebody to a slightly lower cell so in this case it may be the case that an individual instead of being assigned to 63 to 78 months the judge says actually the offence is not as severe their criminal history is not as severe and they should actually be assigned to 21 to 27 months that's the main decision that judges have in this system and this is what is known as the ability to downwards depart for white individuals in 17 of cases we observe downwards departures what i want to understand is how does that change for hispanics and for blacks just after 9 11. okay so what we observe in our administrative records we observe individuals from the point that are arrested by the police as they work through all different stages of the criminal justice system including the trial and the actual sentencing by the judge what we're going to do is take a set of defendants all of whom committed their offenses well before 9 11. all of the individuals committed their offences before 9 11 but then they're working through the timeline so we have a set of individuals whose sentencing by the judge occurs just before 9 11 and we're going to compare outcomes for those individuals who were sentenced just before 9 11 to another set of individuals who also committed their offenses before 9 11 but happened to be sentenced just after 9 11. their trial had already taken place their guilty verdict had already been given the only thing that needed to be determined by the judge was what was going to be their sentence okay so they're the two groups we'll be comparing and we want to see whether people are treated differently after 9 11 than before 9 11 and whether that is different for black individuals or for hispanics this is what we find remember the baseline level of downwards departures is 17 in 17 percent of cases judges for whites normally say we're going to downwards depart and give this individual a less severe sentence than than the than the prosecution have have asked for what we observe is for those decisions made just after 9 11 we find that blacks actually become one percentage point less likely to receive a downwards departure but for hispanics they're actually more than three uh three and a half percent less likely to receive a downwards departure so we start from a baseline of 17 that falls now and they're only 14 13 as likely to receive a downwards departure and these numbers are for like for like comparisons these are comparing a hispanic defendant and a white defendant who've committed exactly the same offence they should have been assigned with the same criminal history and the same offense severity they have the same age the same marital background the same number of children and they're being sentenced in the same court okay these two individuals are identical as far as we can observe the only difference between them is their ethnicity okay and despite being sentenced in a very similar time period after 9 11 we're seeing big reductions in the likelihood that these senior judges are granting downward departures to hispanic defendants okay this applies both to hispanic individuals who have u.s citizenship and also hispanic individuals who don't have u.s citizenship the numbers are three percent for hispanic citizens and more than four percent reductions and downward departures for non-citizens in terms of crimes we see the biggest reductions in the likelihood of downward departures amongst hispanic citizens to be for immigration offences but there's also a three percent reduction for hispanics who've committed other offenses which are not related to drugs but are often related to gun crime so the effects are not just concentrated amongst u.s citizen hispanics the effects are not concentrated just amongst immigration offences it's across the board and both for u s citizen hispanics and for non-us citizen hispanics when we translate these reductions in judges downwards departures into how much longer individuals then spend in prison we observe for hispanics before 9 11 they were spending about four months longer in prison than white individuals remember that was the like-for-like comparison i started out with after 9 11 they spend almost five months longer in prison okay so this is an additional month that they spend in prison for two otherwise identical individuals and this is costly this is costly to send people to prison for these types of uh uh difficult to explain uh reasons so we can think about the cost of this either from to society so we know how much it costs taxpayers how much it costs society to keep somebody in prison for one month longer that corresponds to um twenty nine thousand dollars to keep somebody in prison for an additional year so to keep them in prison for an additional month is about fifteen hundred dollars per defendant we have to spend that aggregates to a lot given that forty percent of all defendants are hispanics and twenty five percent of the world's prisoners are in the us in addition the individual will also bear some costs of going to prison for longer in terms of their future lives when they come out of prison the lack of opportunities that they might then have from having been in prison for longer and that's been estimated for an individual to be between thirty six thousand and forty thousand us dollars over the course of their lives this is hugely costly if you just try to translate this into monetary terms you can translate into into other terms as well in terms of lost opportunities for these individuals so it really makes a difference to these individuals it makes a huge difference to society as a whole and just to start to conclude the results are somewhat different across different parts of the united states so here what i've shown you is a map of the united states where each um square there each region corresponds to a different federal court so there are 90 federal courts in in the united states the darker areas are those areas that have the light the highest likelihood that hispanics are downwards departed and the shade the lighter areas are where hispanics are being dealt with more more harshly after 9 11. and when you try to understand the characteristics of these areas one key characteristic where hispanics are being treated less harshly after 9 11 is where there are more hispanic judges at this very senior level okay so that seems to suggest that the ethnicity of the judge also matters for what happened to defendants based on the defendant's ethnicity that again seems to suggest this may be acts of discrimination rather than any differences that can be justified from a legal or economic point of view let me just conclude so what we've observed is that 9 11 an event that was unanticipated and very much related to terrorist activities did have spillover effects for other minorities in the united states there are these types of contagious animosity where if we feel anger and increasing distrust towards some groups that causes us to change our behavior towards other minority groups what we're observing is that these types of potentially subconscious changes in individuals behavior apply to very senior judges they apply to highly skilled judges remember there are only 700 of these judges in the entire united states this is a very very select group of individuals and these types of biases almost you know also seem to apply at this very high level so if they apply at this level for certain they're likely to apply for other individuals they're in me they're in you these things will will happen um at all different tiers of the of the criminal justice system but the paper does suggest some policies that we can have to try to mitigate these types of implicit bias from an economic point of view we can think of three different dimensions from what can we do to try to mitigate these types of bias that lead to unfair unjust outcomes which is essentially you know eroding the basis on which any justice system is based the first might be to try to train judges to make individuals aware of the implicit biases that they have a second dimension might be to monitor judges for us to say that judges need to justify to a greater extent why they're making decisions related to downwards departure so potentially trying to monitor judges to a greater extent might have an impact on their behavior and the third dimension that economists think of is in terms of selection can we change the composition of the judiciary are the policies that we can put in place to encourage more individuals from minority backgrounds to potentially enter into these occupations and and to remain there to come back to the overall themes of the uh of the uh of the festival though there is an important notion here that what we're what we're documenting here are essentially implicit biases that individuals might have they may not be aware that they're changing their behavior in this way we can discover these biases statistically but individuals may not be conscious that they're making these decisions and so there's a larger question to be had among society and economists and lawyers have something to contribute to this which is the extent to which we might want to have automation in different parts of the criminal justice system we're already seeing this at the very first point where people come into contact with the criminal justice system so we've gone from what used to be science fiction in terms of uh robocop or predictive policing to the right-hand side of the figure where these types of algorithms where these types of predictive policing notions are actually being constructed and currently being used to inform the police where should they deploy their resources at any given moment in time often we observe very predictable patterns of crime when crime occurs where it occurs in relation to all kinds of characteristics of of a neighborhood fundamental question that we're then pushing towards in this paper and in this agenda is should we then try to automate other forms of decision making in the criminal justice system is it good to try to remove human biases or will it be the case that we introduce other forms of bias as we remove human interaction in these types of system and try to move towards more automated systems the the sentencing guidelines were a step in that direction they were explicitly designed to reduce the discretion that judges had and we see that that system actually didn't prevent discrimination taking place when judges are basing their decisions and somewhat emotional factors related to 9 11. so it's an open question but i think economists lawyers and the general public all have important things to decide and debate upon there will be costs and benefits of making those moves hopefully our research can be useful to think about um some of the some of the issues that at play there okay thank you thank you very much we can now open it to the floor the q a session is open now especially starting from the last point you raised a question with no answer basically so can automating policing contribute to correcting bias or can it have only a limiting distorted distorting impact because we're talking about automation and ai at a time in which probably western societies show behaviors that are very far away from your algorithm about predictability while you were talking i was thinking about the figures on imprisonment and costs that have to be born by keeping a person one month longer in jail for reasons due to an implicit bias and not the level of crime that was committed when talking about data on a migrant invasion like in 190 000 people and so on but in western societies in italy in the u.s in sociology there is only one theorem the thomas theorem stating if men perceive and define something and re as real the consequences of that will be real too and that's the very opposite of rationality so so i'd like to take some questions from you well otherwise i'll ask myself there's a question over there please use a microphone for translation thank you very much i have two short questions the first can you please tell us what the percentage of non-white judges is or non-unclear saxon judges in the federal courts and the second question is what's your opinion about the fact that our judiciary selected by public competition so they're not appointed they are not chosen by voters so do you think this difference has an impact because our system is generally considered as a more balanced system do you think this might be the cause of the consequences that you outlined i hope i've been clear enough we'll take a few more questions before you reply is that okay for you um i just wanted to ask um how the kind of like um uh if you know when a hispanic non-citizen has entered the country and been um you know and then they've got a sentencing and they've already committed a crime by entering the country so i was just wondering whether the fact that by the time they've already you know reached a judge they've already been guilty of of another crime does this have any impact or is there some you know if you've committed a crime before does that make the judge um more more discriminated or more harsher to you when you um sentence again thank you castle okay so um just to go through those in in sequence there's about 14 percent of judges in the us who are from non-white backgrounds and they tend to be concentrated in certain uh federal courts and it's in those federal courts that we find the discrimination against hispanics to actually be the least so that suggests that the ethnicity of judges also matters for how defendants are treated leading us to think this might be discrimination this is not just something that judges have been in general are practicing the question of how we select judges and you know clearly different countries have different mechanisms whether it's through public appointment lifetime appointments or um electoral appointees and so here we we know a little bit about how individuals behave the incentives that individuals have will be different who is willing to stand in those different systems and how their behavior might be impacted depending on what they think is going to happen in the future are they going to be reappointed are they are the voters are actually going to to reappoint them and we see this in the us in terms of regulators some regulators for some industries were electorally appointed some of them are appointed by um by by politicians directly the u.s system is designed to give politicians or absolutely no say in all voters no say in trying to discipline these federal judges they're appointed by presidents they have to be approved by congress but they're also appointed for life so they face no re-election or chances to actually be um to kicked out for bad decisions so this is a system in which compared to most other judges in the u.s these individuals are highly skilled highly trained we do find some additional differences beyond the ethnicity of the judge and driving outcomes such as whether they've been appointed by a democrat or by a republican and also the experience that they have of seeing other cases so clearly there are some factors that potentially we could manipulate amongst judges and see can we appoint them appoint them differentially the fundamental difficulty here is that i think what we're measuring here are essentially implicit biases that judges and individuals have when we have an event like 9 11 it triggers first and foremost an emotional response and that's very hard for us to try to suppress and perhaps if we do try to suppress that we might over compensate and lead to worse outcomes for other individuals so that's why it makes it a fundamentally difficult question to think about can we improve on this system or do we have to accept these differences if we have full automation no doubt we will have other errors that creep into the system because ultimately those algorithms will be designed by humans and they will feed in as has been documented so i think that there are important tradeoffs for us to think about um but what really surprised us is yeah we find such important levels of discrimination even amongst this very select group of professional judges in terms of how do the results vary depending on the criminal history of the of the defendant all of the individuals that we consider have been found guilty the only thing that is remains to be determined is the sentence that there that they're then given so here we're really comparing like with like of individuals who potentially for example could have been deported and that's true both before 9 11 that hasn't changed policy wise after 9 11. so these impacts that we're finding are really essentially for two clones one that happened to be sentenced before 9 11 one after 9 11 where they also differ in terms of the they were a pair of white clones versus a pair of hispanic clones so these impacts are really coming from something related to them being hispanic and being treated differently after 9 11 and it's not down to where they're being treated their criminal histories the level of the offense severity or anything else that determines where they end up in that guideline cell table well while you're thinking i have two questions myself the first question is more of a journalistic question i can say i comes to mind that you said that this contagious animosity after 9 11 this lack of trust towards the muslims in general had an impact on criminal on criminal justice more on the hispanic community than the black community which well in a way was more likely to have links with the muslim community being black and muslim so we have followed the black lives matter movement which was focusing on this very tall issue i i was reading today the family of a black guy that had been killed by police had received compensation of four dollars four dollars because uh the gun was not loaded and stood up so how comes there was no um movement called spanish or hispanic lives matter what do you think there's more of an acceptance more fear oh what this is my first question second question the phenomenon that you described this suspicion and lack of trust going from the muslims to other individuals after 9 11 and that actually happened in this country as well as i explained you focused on the united states but my impression is that it happened in italy too do you think this is something that the whole west has in common do you think it affects other countries too i'm talking about the west because the debate on discrimination in legal terms is more lively luckily okay so the issue of uh discrimination against blacks in the u.s criminal system is obviously one that uh is is highly relevant and one that gets a lot of attention in the u.s and outside clearly our statistics show that there are important differentials in how black defendants are being treated before 9 11 relative to whites what our results show is though that those levels of differences don't change because of 9 11. that's not to say that there aren't differences that there aren't important differences and they might be down to discrimination what our results show is that those don't change with 911 911 particularly triggers animosity towards muslims that then spills over onto hispanics rather than on to blacks in fact all of our results seem to suggest that there's absolutely no impact on black defendants at all as a result of 911 but an interesting question is what why does a movement like hispanic lives matter also not not exist or get the same levels of attention i think as hispanics you know become more prominent in u.s society on a whole range of different dimensions that that inevitably will be something that um that society will have to will have to confront and recognize that there are those differences i i showed you the statistics in terms of um the thermometer readings before 9 11. and it was surprising but the degrees of animosity felt towards hispanics were higher than felt towards blacks and that's been the case since the 1970s since the data first started to be collected so this is very much a deep rooted um animosity that's felt and this is actually something that huntington has actually described in his work as well you alluded to huntington in in the clash of civilizations he often refers to latinos as well as muslims as permanently being regarded as outsiders in in u.s society it's a different form of discrimination than potentially that that might be a against blacks is a different dimension of emotional response it's something that probably a social scientist we're still trying to understand what what triggers these different factors how do we perceive different minority groups and what are the linkages between these different minority groups there's a a range of different evidence collected by psychologists both in terms of lab environments and sort of in in real world environments where i could separate this room into into two groups and tell this group something about this group and this group something about this group and engender some kind of animosity between the two halves of the room i can just give you a false label and a very natural reaction for us is to identify more to the group that we've been assigned to than the group that we see ourselves in opposition to this is just a trait that we seem to have and has been documented by many different disciplines both in the natural sciences and the social sciences and i can engender this on a purely artificial characteristic i can call this group the left group this group the right group so it's hard to understand how can we try to offset these but they're there in in everybody and here what we're trying to do is use economic statistical techniques to try to uncover that to try to measure the impact of that and try to suggest some ways to potentially offset those those factors but i think it's very much the case that these types of behaviors therefore would exist in other societies and other criminal justice systems irrespective of how we try to select individuals either into the police into into prosecution into the judiciary there's always going to be these types of factors currently i'm working with the ministry of justice in the uk to try to replicate this analysis for the for the uk around terrorist events in in the uk and to try to understand whether something similar is happening there potentially against other groups be beyond muslims as well there's been other studies conducted in france for example that use very detailed and similar data to show that the sentences handed out by french judges are very much influenced by what's prominent in the media at the time so if there are very high profile cases in some particular part of the country about some particular type of crime that then impacts other defendants who've got nothing to do with that in terms of how judges behave potentially you could imagine those issues being magnified in a system where judges are being held to account by voters and by potential electoral arguments and in italy similar results have been shown in terms of how the criminal justice system impacts minority groups in terms of immigrants in in in particular that when for example it becomes easier for immigrants to obtain citizenship then you see criminality by immigrants plummeting that essentially it's not a choice it's something that they may well be forced to do and small changes in how the legal system treats individuals can have big impacts in terms of their behavior and that's what i alluded to at the start of the talk when i mentioned that the fact that individuals are going to come into contact with the criminal justice system at you know the most important time in their lives when they're young adults and just making so many important decisions has impacts not just on their lives after that but has impacts on their lives before that age when they're making choices or their parents are making choices about education and the skills that these individuals will have so these are important challenges not just for lawyers or for economists working on crime but for individuals who are working on education working on families working on labor markets essentially you know the whole of our lives we're going to be impacted by how we're treated by the system and when we have such big differences we can expect big differences and outcomes along all of those dimensions that are not unique to the us is there any other question yes there is one question from the united states so my question is the level of criminality judges everything is that just the volume is much higher at the state level than at the federal level and i would expect your statistics would be look even worse if you looked at the state level um so i guess my specific question is i can speculate to myself on why you restricted your study to the federal level maybe it's because that's where most of the terrorism related crimes exist but you ever have you thought about taking the the next step to the lower courts the state level because i think the results could possibly be much worse any other question important to to try to look at other actors in the in the judiciary importantly the state level as you mentioned as well as potentially um changes in behavior of the police in our current work we focus on the federal system essentially because that's where we expect the most professional judges to be present so we're thinking this is the worst place to try to find evidence of discrimination so let's start there and try to have the best case scenario and if we find something there that's important because of the severity of offenses here but also it probably will be indicative of exactly the intuition that you said that things are worse other parts of the of the justice system we look at the federal system though just because not just because of the severity but also because it's very algorithmic in terms of how long there's different stages of the system take so it's very easy to see individuals who come up just before 9 11 versus after 9 11. if we start looking at the state level there's many more margins of discretion in terms of how how speedily you put somebody through the system and that varies state by state so it becomes much more challenging statistically to show convincing like for like comparisons at that level because there's much more many more sources of discrimination we might try to get to that when we look at the uk system for example where one thing that you can do is speed people through the system and that may be another margin along which people are being discriminated against but that that's a much more complicated statistical problem to deal with as well that's why we've started here but my sense is that in other parts of the system whether we're thinking about the police or the state level or other countries as soon as you have more flexibility for agents to behave in this way things will potentially multiply along different margins here we've looked at judges who are essentially quite restricted they just have one decision that they can make on downwards departures and that's where we're able to focus our attention but i would love to be able to come back in a few years time and show additional results for other parts of the judiciary and that's going to help us understand both where it's worse but also maybe where it doesn't exist and are there important lessons that we can learn from from those parts the criminal justice system as well any other question well if not then i think we can draw to a close and thank you very very much mr imran razul and of course we thank you all for your attention and the questions you asked thank you you