SSHLaw Founder and Partner Bjorn Harsanyi on Generalized Risk and Gang Violence Presented at CARL in 2014

Generalized Risk and Gang Violence

The protection afforded to a claimant under paragraph 97(1)(b) of IRPA must be personal and not one faced generally by others in the country. The million dollar question in the last number of years is what is meant by personal risk, and how have the courts defined it in recent years? In the past few years the jurisprudence has become muddied to say the least, with often competing lines of interpretation on when a risk becomes personalized when an individual is faced with gang related violence. Decisions of the RPD emboldened by certain decisions from the Federal Court have repeatedly denied otherwise well founded claims for protection because the risk to them is generalized.

This is particularly the case for applicants from countries in Central America, where crime and gang violence is rampant. In recent jurisprudence, the courts have clearly grappled with where the line in the sand should be drawn in these cases. For example, is a shopkeeper who flees Guatemala because he refuses to pay extortion money simply a victim of crime for whom s. 97 does not apply? What about the young man who flees Honduras rather than be recruited by a gang? These are some of the fact scenarios that courts have dealt with in recent years.

Personalized Risk vs ‘Shared’

With the release of Prophete[1] in 2008, the court clarified that a risk faced by a claimant was not personal if that risk was one faced by others in the country generally. In that case, the Haitian applicant argued that he should be protected based on his perceived wealth as a businessman, which made him a target of gang violence.

The Court disagreed, stating that the risk of harm due to gang violence was a type of criminality experienced by all Haitians. While a specific number of individuals may have been targeted more frequently because of their wealth, all Haitians were at risk of becoming the victims of this type of gang violence:

[18]           The difficulty in analyzing personalized risk in situations of generalized human rights violations, civil war, and failed states lies in determining the dividing line between a risk that is “personalized” and one that is “general”. Under these circumstances, the Court may be faced with an applicant who has been targeted in the past and who may be targeted in the future but whose risk situation is similar to a segment of the larger population. Thus, the Court is faced with an individual who may have a personalized risk, but one that is shared by many other individuals.

After Prophete, many applicants fleeing crime, civil war and gang violence were thus excluded from protection. The required threshold was virtually unattainable; most claimants who were victims of extortion or targets of gang violence lost their claims even if they were being personally threatened. Some of the decisions of the RPD were simply nonsensical; a person was specifically and directly targeted by a violent gang, but the risk to them was deemed not to be personal.   More emphasis was placed on general country conditions, the genesis of the risk and crime levels rather than an individualized inquiry.  It appeared that any claimant who was targeted by criminal groups in countries where crime is rampant (no matter how personal the risk) was found to face a generalized risk.  Section 97 was for all intents and purposes being stripped of all meaning.  In my view, this was never the intent of Prophete and was a distortion of the finding of the Court of Appeal.  The genesis of the risk became paramount  (i.e. the reason for the initial targeting) and whatever intervening events occurred thereafter was found to be irrelevant.  Many very good cases from the RPD were being rejected on this basis and it became a cookie cutter approach to rejecting large numbers of otherwise well founded refugee claims.  Ironically, the more dangerous and widespread the gang the more likely the case was to be rejected on generalized risk.  To be frank, a disturbing and depressing time for anyone practising refugee law with claimants from Central America.


The Turning Point

Although the jurisprudence around s. 97 continued to develop post-Prophete, it became clear that the court was beginning to view such a narrow interpretation of s. 97 as problematic. Justice de Montigny’s decision in Pineda[2] was often cited and used as anchor decision in judicial reviews of negative RPD decisions citing generalized risk in for those fleeing gang violence. In that case, the claimant had been targeted by the Maras Salvatruchas gang in El Salvador, however, the RPD found, unsurprisingly, that he was subject to the same risks as those faced by the general population. In what became a heavily relied upon quote in the case law to follow, Justice de Montigny stated:

[15]      Under these circumstances, the RPD’s finding is patently unreasonable. It cannot be accepted, by implication at least, that the applicant had been threatened by a well-organized gang that was terrorizing the entire country, according to the documentary evidence, and in the same breath surmise that this same applicant would not be exposed to a personal risk if he were to return to El Salvador. It could very well be that the Maras Salvatruchas recruit from the general population; the fact remains that Mr. Pineda, if his testimony is to be believed, had been specifically targeted and was subjected to repeated threats and attacks. On that basis, he was subjected to a greater risk than the risk faced by the population in general.


If Pineda was the anchor case, a 2011 decision from Justice Zinn really provided new hope to gang related generalized risk claims. In Guerrero[3], Justice Zinn first set out the ‘test’ – essentially the elements of a claim – under s 97:


[25] Subparagraph 97(1)(b)(ii) of the Act defines a person in need of protection as“a person in Canada whose removal to their country or countries of nationality… would subject them personally to a risk to their life or a risk of cruel and unusual treatment or punishment if the risk would be faced by the person in every part of that country and is not faced generally by other individuals in or from that country.”


[26] Parsing this provision, it is evident that if a claimant is to be found to be a person in need of protection, then it must be found that:


  1. The claimant is in Canada;


  1. The claimant would be personally subjected to a risk to their life or to cruel and unusual treatment or punishment if returned to their country of nationality;


  1. The claimant would face that personal risk in every part of their country; and


  1. The personal risk the claimant faces “is not faced generally by other individuals in or from that country.”


All four of these elements must be found if the person is to meet the statutory definition of a person in need of protection; it is only such persons who are permitted to remain in Canada.


In that case, the Guatemalan applicant refused to deliver drugs for a notorious gang. He fled to Canada after armed gunmen shot at his house with automatic weapons, killing his grandmother in his presence. The RPD found that the applicant was a “particular” target of the drug trafficking gang, however the risk to him was still generalized given the prevalence of gangs in the region and crime in the country. In overturning the decision, Justice Zinn found that the applicant was not only at risk of being recruited into the gang like other young men, but his life was personally at risk when he became a target after refusing to comply with the gangs’ demands.

Justice Zinn expressly rejected the idea that an applicant was automatically excluded from protection under s. 97 simply because he faced a risk of violence at the hands of a criminal gang in a country where gang violence is prevalent. He also observed that a finding of personal risk was essential before determining whether it was a risk faced generally by others in the country.

[27]           The majority of cases turn on whether or not the last condition has been satisfied, that is, whether the risk faced by the claimant is a risk faced generally by others in the country.  I pause to observe that regrettably too many decisions of the RPD and of this Court use imprecise language in this regard.  No doubt I too have been guilty of this.  Specifically, many decisions state or imply that a generalized risk is not a personal risk.  What is usually meant is that the claimant’s risk is one faced generally by others and thus the claimant does not meet the requirements of the Act.  It is not meant that the claimant has no personal risk.  It is important that a decision-maker finds that a claimant has a personal risk because if there is no personal risk to the claimant, then there is no need to do any further analysis of the claim; there is simply no risk.  It is only after finding that there is a personal risk that a decision-maker must continue to consider whether that risk is one faced generally by the population.


In Portillo[4], the court, following the line of reasoning in Guerrero, found the two ideas – that a person could be personally targeted but still face a generalized risk – could not exist simultaneously. In other words, a finding that an applicant faces a personal risk because a violent gang will continue to pursue him cannot be reconciled with a determination that the same applicant faces a generalized risk, simply because the risk arises from criminal gang activity.  One would assume such a finding was common sense but it took a long time for the Court to make this determination.

Diverging Authorities

Despite some very strong statements by the court in this regard, the jurisprudence in this area appears to diverge into two lines of authority. There are several instances where the court has reached opposite conclusions on very similar sets of facts. In Portillo the court stated:

[37]           In recent years, this Court has been called upon to review a multitude of RPD decisions interpreting the notion of generalized risk enshrined in paragraph 97(1)(b) of IRPA. To a large extent, these decisions turn on their own particular facts, including the reasoning from the RPD that the Court was called upon to review. The cases may be divided into two groups.


The court listed several cases where it had overturned RPD decisions where the claimant had been personally targeted for violence by one of the criminal gangs operating in Central or South America (see e.g. Pineda (2012); Lovato v Canada (Minister of Citizenship and Immigration), 2012 FC 143 (CanLII), 2012 FC 143 at para 7, [2012] FCJ No 149 (Rennie) [Lovato]; Guerrero v Canada (Minister of Citizenship and Immigration. In contrast, several decisions reached the opposite conclusion, where the Court upheld the RPD’s decisions in situations where gangs made threats of future harm to the claimants but the threats were found to be insufficient to place the claimant at any greater risk than others in the country (see e.g. Rodriguez v Canada (Minister of Citizenship and Immigration)2012 FC 11 (CanLII), 2012 FC 11, [2012] FCJ No 6 (Russell); Rajo v Canada (Minister of Citizenship and Immigration)2011 FC 1058 (CanLII), 2011 FC 1058, [2011] FCJ No 1277 (Kelen); Chavez Fraire v Canada (Minister of Citizenship and Immigration)2011 FC 763 (CanLII), 2011 FC 763, [2011] FCJ No 967 (Zinn)

Despite some differing lines of interpretation, it does appear that the court is making an effort to reach a common ground. Some Justices have even referred to the idea that competing lines of authorities have emerged as more imagined than real[5], however it is clear that the court has become far more comfortable in finding that the RPD frequently mischaracterizes the risk faced by claimants due to criminality.  It is I would assume a troubling admission for the Court to make that they are essentially split into two camps – those that think gang related violence can form the basis of refugee protection under Section 97 and those who do not.  I would suggest this schism was rooted in ideology and the law was vague enough that decisions could be justified either way.  Rather than acknowledge the opposing lines of authority it was simply chalked up to the individual facts of the case. But as the cases began to mount on both sides some common ground had to be found.

A Common Approach

In the most recent case from the Federal Court in this area, Justice Russell not only provides an excellent summary of the law but moves the court further towards a common ground. Correa[6] is a clear illustration that the court is finally prepared to accept that someone whose life is personally at risk due to violence and criminality can be a person in need of protection:

[46] While a full consensus has yet to emerge, I think that there is now a preponderance of authority from this Court that personal targeting, at least in many instances, distinguishes an individualized risk from a generalized risk, resulting in protection under s. 97(1)(b). Since “personal targeting” is not a precise term, and each case has its own unique facts, it may still be the case that “in some cases, personal targeting can ground protection, and in some it cannot” (Rodriguez, above, quoted with approval in Pineda v. Canada (Minister of Citizenship and Immigration), 2012 FC 1543 (CanLII), 2012 FC 1543 [Pineda (2012)]. However, in my view there is an emerging consensus that it is not permissible to dismiss personal targeting as “merely an extension of,” “implicit in” or “consequential harm resulting from” a generalized risk. That is the main error committed by the RPD in this case, and it makes the Decision unreasonable.

In that regard, there are some common principles that can be gleaned from these cases. As a starting point, an applicant fleeing gang violence is no longer barred from protection under s. 97 simply because gang violence is rampant in the country of origin and affects large swathes of the population. The RPD must conduct an individualized inquiry, and there are several factors to consider.

First, the initial reason for the risk is largely irrelevant to the assessment. As stated in Portillo, above, the fact that the risk to an applicant arises from criminal activity does not in itself foreclose the possibility of protection under section 97. In Guerrero, Justice Zinn further stated: “When one conflates the reason for the risk with the risk itself, one fails to properly conduct the individualized inquiry of the claim that is essential to a proper s. 97 analysis and determination.”[7]

In other words, the question is not whether the risk to a claimant is created by criminal activity, but instead whether the claimant would be subjected personally to a risk to his or her life and whether that risk is one not faced generally by other individuals in or from that country. In Contreras[8], the court also overturned a decision of the RPD where the member focused on the reason for the risk as opposed to the risk itself:

[19]     Where the Member erred was in conflating the current risk with the original reason for that risk. In her reasons for decision, the Member stated that “A key matter in this analysis is the initial reason for the targeting” (para 15). However, the fact that the risk had initially arisen from criminal activity which might have threatened any small business operator was not relevant to the personalized situation which had eventually developed and which was the basis for the s 97 claim: Camargo Vivero v Canada (MCI), 2012 FC 138 (CanLII), 2012 FC 138 at para 11. Section 97 must not be interpreted in a manner that strips it of any content. The issue is not what first created the risk, but whether it is a risk faced by the claimant personally and not faced generally by other individuals in the country: Vaquerano Lovato v Canada (MCI), 2012 FC 143 (CanLII), 2012 FC 143 at paras 13-14.


In this regard, a discussion of risk must distinguish between s. 96 and 97. The claimants in the cases above are mainly young males from Central America who have been targeted by gangs. In these examples however, the applicants cannot claim protection because they are young males at risk of being recruited by gangs. This is too generalized. Rather, the claimant needs protection because he is being personally targeted for a specific action, refusing to join a gang, for example. A shopkeeper in Colombia who refused to pay extortion money could succeed under s. 97 not because he is a business owner with perceived wealth, but because he failed to pay extortion money and is now being threatened. In Correa Justice Russell elaborates:

[56]           The problem with this approach lies in assigning too much importance to the initial reasons for the threat. In doing so, it seems to improperly import elements of the s. 96 test into the s. 97 context. Under s. 96, the reason one is targeted is at the heart of the analysis, because of the requirement to establish a nexus to Convention grounds of protection. Under s. 97, by contrast, it has very little relevance at all. Someone may be initially targeted for extortion because he/she is a shopkeeper, but that is irrelevant to the risk faced now and in the future except to the extent that it provides clues to the nature and extent of the threat objectively considered. It does not matter what personal characteristic of the victim prompted the perpetrator to target them (e.g. youth, perceived wealth or ownership of a business) or what motivates the perpetrator to target anyone in the first place (e.g. increasing wealth through extortion or acquiring “drug mules” through forced recruitment).

Ultimately, the distinction can be made by asking a straightforward question: is the claimant at risk because of who he is (his profile) or is he at risk for what he has done (or failed to do) with the former likely falling within the realm of generalized risk and the latter being personal.

Second, the risk itself can evolve. If it is irrelevant why a claimant was at risk in the first place, it follows that risk is not a static concept. A person might initially be targeted because they are perceived as wealthy, but the situation may evolve after they report the threats to the police and become direct targets as a result.[9] A risk assessment is a contextual and common sense assessment rather than a rigid or quantitative exercise.[10]

Third, the proximity, or degree, of the risk must be assessed in addition to the broader nature of the risk. In this sense, the analysis is highly fact driven. The RPD must look at the specific, unique circumstances of the claimant, and not simply whether or not the risk is general in nature. To do otherwise would strip s. 97 of its meaning entirely:

[61]           It seems inescapable that, if s. 97(1)(b) is to be given any meaning (for anyone, not just victims of criminal gangs), proximity to the risk (or as some have put it, the degree of the risk), must be considered in addition to its “nature” broadly defined. That is why it is problematic to speak of personal targeting as simply “consequential harm” or an “escalation of” an initial generalized risk (contra Baires Sanchez, above, at paras 21 and 27).

In the words of Justice Shore: “The risks of those standing in the same vicinity as the gunman cannot be considered the same as the risks of those standing directly in front of him.”[11]

Finally, in assessing proximity and degree, what has also emerged in the case law is that a risk to a ‘sub-group’ of people – as opposed to all citizens generally – can be contemplated by s. 97. The cases found that if you could be part of a large enough sub group then the risk could still be generalized even though it does not apply to the entire population (mental gymnastics at its finest and a gross overcomplication of a simple concept). This then raises a host of questions such as how large that sub-group must be for a risk to be considered as faced generally by others, or how do we define the sub-group and what are its parameters. Justice Russell advises using a common sense approach to this question, looking at whether the sub-group is of a size that the risk could be considered widespread or prevalent.[12] In Munoz[13] Justice Campbell elaborates:

[9]               Defining the characteristics of the sub-group is particularly important. In the present case the following features of a potential sub-group were established by the Applicant: shopkeepers who are extorted; and who have not met extortion demands; and who have suffered severe bodily harm and death threats; and who continue to receive death threats communicated through family members, together with death threats directed against family members. It was for the RPD to determine whether persons who have had the same experience as the Applicant form a sub-group of a character and size sufficient to make the Applicant’s risk a generalized risk.


[10]           In the sub-group determination just described, a point might be found where a personalized as opposed to a generalized risk exists. In the present case, on the evidence produced by the Applicant, the turning point might be found where the Applicant’s conduct of not paying money in response to extortion demands drew particular attention to him personally, or perhaps when the extortion threats became actualized in extreme violence and escalating death threats against the Applicant and his family. It was for the RPD to determine whether these, or any other turning point existed.



The determination of the size of the sub-group necessitates an individualized inquiry and will turn on the facts of the case.


It seems while we have not come full circle in the case law it is very close to being just that. Prior to Prophete the issue of generalized risk when an individual was specifically and personally targeted by gangs was somewhat of a non issue. This then changed dramatically for the worse and Section 97 was severely restricted for a period of time and many good refugee cases were lost (even when the risk to the individuals life was acknowledged by the Board).  With the most recent cases the tide appears to have turned back and gang related cases are back on the table.

While a clear consensus remains elusive, thankfully there is now an abundance of authority from the Federal Court to decisively state that individual, personal targeting distinguishes personalized from generalized risk.  The wolves have been beaten back for another day. In these cases, a claimant fleeing gang violence can no longer be merely dismissed as a victim of crime. The key for any practitioner in this area when conducting a refugee file relating to gang related or other criminal violence in countries where this is prevalent is to make the risk as personal as it can be.  The more unique and personal the facts are the less likely a rejection can justified on the basis of the risk being generalized. The more intervening acts that occur from the initial reason for targeting the more personal the risk will become.

Please know this paper is not meant to be comprehensive and is simply snapshot of my view of the evolution of this case law.  I hope you found it useful.


Bjorn Harsanyi LL.B Partner

[1] Prophète v. Canada (Citizenship and Immigration), 2008 FC 331

[2] Pineda v Canada (Minister of Citizenship and Immigration), 2007 FC 365

[3] Corrado Guerrero v Canada (Minister of Citizenship and Immigration), 2011 FC 1210

[4] Portillo v. Canada (Citizenship and Immigration), 2012 FC 678

[5] See Coreas Contreras v. Canada (Citizenship and Immigration), 2013 FC 510

[6] Correa v. Canada (Citizenship and Immigration), 2014 FC 252

[7] Guerrero, para 29

[8] Contreras, supra


[9] See Martinez De La Cruz v Canada (Minister of Citizenship and Immigration),  2013 FC 1068

[10] Correa para 84

[11] Balcorta Olvera v. Canada (Citizenship and Immigration), 2012 FC 1048 para 41

[12] Correa, para 8, quoting De Parada

[13] Munoz v. Canada (Citoyenneté et Immigration), 2012 FC 716