The COI-CG1 Working Party wishes to commend to all members of the Association the following ‘COI Judicial Checklist’ (see below). Although we hope this checklist and accompanying Explanatory Memorandum (see paras. 1-64) will be of general interest, its primary aim is to furnish a guide to judges2 in cases where they face having to assess Country of Origin Information (COI) in the context of deciding asylum or asylum-related appeals.

The Checklist is the result of 18 months of deliberations involving the efforts of a considerable number of people with knowledge in this area. The following are current members of the COI-CG Working Party: Hugo Storey (Rapporteur, UK), Bostjan Zalar (Deputy Rapporteur, Slovenia), Graham Davies (UK), Bernard Dawson (UK), Nigel Osborne (UK), John Barnes (UK), Dallal Stevens (UK), Anna Bengtsson (Sweden), Patrick Hurley (Ireland), Rory McCabe (Ireland), Vaclac Novotny (Czech Republic), Manoj Kumar Sinha (India), James Simeon (Canada), and Hannah Lily (Assistant to the Rapporteur, UK). The following are those who attended the June 2006 London Roundtable, which was devoted to debate on earlier versions: Mark Ockelton (Senior Immigration Judge and Deputy President, Asylum and Immigration Tribunal, UK (Chair)), Oldrich Andrysek (Department of International Protection, UNHCR), Chris Attwood (Country of Origin Information Service, Home Office, UK), John Barnes (Former Senior Immigration Judge, UK), Chantal Bostock (Legal and Research Unit, Asylum and Immigration Tribunal, UK), John Bouwman (Judge, Holland), Eamonn Cahill (Judge, Refugee Appeals Tribunal, Ireland), Jane Coker (Immigration Judge, UK), Heaven Crawley (Senior Lecturer, Swansea University, UK), Steve Crawshaw (Human Rights Watch, UK), Alice Edwards (Amnesty International), Mark van Elzakker (Immigration Service, Holland), Jonathan Ensor (Immigration Advisory Service, UK), Professor Anthony Good (Edinburgh University), Mark Henderson (Barrister and representative of Immigration Law Practitioners Association, UK), Catriona Jarvis (Senior Immigration Judge, UK), Andrew Jordan (Senior Immigration Judge, UK), Hannah Lily (IARLJ Working Party Assistant and British Refugee Council, UK), Nigel Osborne (Immigration Judge, UK), Ilkka Pere (Justice, Supreme Administrative Court, Finland), Professor Terence Ranger (St Antony's College, Oxford University), John Ryan (Judge, Refugee Appeals Tribunal, Ireland), Hugo Storey (Senior Immigration Judge, UK), Nick Swift (Advisory Panel on Country Information (Secretary), UK), Mark Symes (Barrister, UK), Patrice Wellesley-Cole (Immigration Judge, UK) and Bostjan Zalar (Judge, Slovenia). The Working Party wishes to pay particular thanks to Allan Mackey (immediate Past President of the IARLJ) who co-wrote the original version of the paper and presented it to the November 2005 IARLJ European Chapter Budapest Conference, Alice Edwards of Amnesty International whose paper, ‘Amnesty International's Comments on Hugo Storey & Allan Mackey, “In Search of Judicial Criteria for Assessing Country of Origin Information”’, was also presented to the same Budapest Conference and Barbara Svec of ACCORD who wrote a specific commentary on a revised draft of the Storey/Mackey paper, ‘ACCORD Comments on Hugo Storey & Allan Mackey, “In Search of Judicial Criteria for Assessing Country of Origin Information”’, for the June 2006 London Roundtable.

A particular debt is also owed to Hannah Lily, Bostjan Zalar, Andrew Grubb and Andrew Jordan for their assistance with the final stages of revision, albeit ultimate responsibility for any shortcomings is mine. Thanks are also due to Geoffrey Care, Bernard Dawson and John Barnes who contributed their ideas at various stages.

Hugo Storey, (Rapporteur) on behalf of the COI-CG Working Party

October 2006

COI Judicial Checklist

When assessing Country of Origin Information (COI) in the context of deciding asylum or asylum-related cases judges may find the following 9 questions useful:

Relevance and adequacy of the Information

  • i) How relevant is the COI to the case in hand?

  • ii) Does the COI source adequately cover the relevant issue(s)?

  • iii) How current or temporally relevant is the COI?

Source of the Information

  • iv) Is the COI material satisfactorily sourced?

  • v) Is the COI based on publicly available and accessible sources?

  • vi) Has the COI been prepared on an empirical basis using sound methodology?

Nature / Type of the Information

  • vii) Does the COI exhibit impartiality and independence?

  • viii) Is the COI balanced and not overly selective?

Prior Judicial Scrutiny

  • ix) Has there been judicial scrutiny by other national courts of the COI in question?

COI Judicial Checklist: Explanatory Memorandum

  • 1. In the course of dealing with asylum appeals, judges3 will depend to a great extent for their ability to make sound judgments on having before them up-to-date and reliable country background information or ‘Country of Origin Information’ (COI).4 The probative value of an asylum seeker's evidence has to be evaluated in the light of what is known about the conditions in the country of origin.5 The demands on the judge are huge. Sometimes within a very short period he6 may be called on to decide cases of claimants from several different countries. He may be expected to decide at one moment on whether an asylum seeker is a member of a sub-clan of a minority clan based in Mogadishu, Somalia, and also to determine whether that clan is without effective protection. At another moment he may be asked to assess whether a member of the former communist government of Afghanistan would be at risk from the current Northern-Alliance-based regime. He may have to decide whether a Chaldean Christian from Northern Iraq would be at risk from Muslim extremists. In a rapidly changing world he may need to decide whether a Tamil member of the LTTE from Northern Sri Lanka would today face a risk of persecutory harm from the authorities in the light of renewed clashes between government troops and LTTE militias. Faced with diverse cases and shifting political scenarios, judges desperately need accurate and reliable information in order to determine justly who is in need of international protection.

  • 2. COI is evidence the judge should take into account. It is a crucial aid. But it will rarely be determinative. How much it will help the judge determine the individual case will vary depending among other factors on the extent to which the claimant's case is based on personal characteristics or circumstances which he shares with others similarly situated. COI may not be relevant to the same degree in every case.7

  • 3. For a judge making findings on country conditions is not an end in itself: indeed it is not his function to pass judgment on the human rights performance of other countries.8 He is only required to make a finding on a particular case. Nevertheless, within that context sometimes general findings as to country conditions must of necessity be made.

  • 4. Conversely, it is not an end in itself for most bodies that produce COI to assist refugee decision-makers: usually their aim is to provide an analysis for general circulation of a country's human rights performance or some related aspects. That has perhaps the advantage from the point of view of the judge that it cannot be suggested the COI has been ‘tailored’ for use in supporting asylum appeals.

  • 5. In recent years, a number of states who are signatory to the Refugee Convention have written in to their national law specific provisions as to how decision-makers (including judicial decision-makers) are to approach assessment of a person's asylum claim.9 There has also been a major regional initiative within the European Union (EU) designed to harmonise national approaches in this and other respects. From 9 October 2006, all EU Member States except Denmark are bound by the provisions of (and should have transposed into national law) the ‘Qualifications Directive’, that is, Council Directive 2004/83/EC on minimum standards for the qualification and status of third country nationals or stateless persons as refugees or as persons who otherwise need international protection and the content of the protection granted. Article 4 of this Directive deals with assessment of facts and circumstances relating to a claim for international protection. Article 4(3) states:

    ‘The assessment of an application for international protection is to be carried out on an individual basis and includes taking into account: … ’

  • 6. Five matters are then mentioned. The first specifies:

    ‘(a) all relevant facts as they relate to the country of origin at the time of taking a decision on the application; including laws and regulations of the country of origin and the manner in which they are applied’.

  • 7. This provision highlights the importance of COI to all refugee decision-makers.

  • 8. Background country materials or COI (Country of Origin Information) will derive from diverse sources, including reference works (maps, encyclopaedia, yearbooks), reports or papers by international bodies (for example, UNHCR, UN Human Rights Committee), international NGOs (for example, Amnesty International reports, Human Rights Watch reports, International Crisis Group (ICG) reports), national bodies (for example, the U S State Department Reports, the Danish Immigration Service reports, the United Kingdom Country of Origin Reports (COIR),10 news and media clippings and databases, legal materials (laws, jurisprudence, etc) and cross-checking of other refugee claims.11 Reports can be generic (for example, US State Department reports), event or group specific (for example, reports from trials, minority profiles) or claimant specific (for example, embassy checks). There are a number of databases which are specific to asylum-related work: for example, UNHCR's Refworld and ACCORD.12

  • 9. Practices vary as to how COI comes to be placed before judges in asylum and asylum-related cases. Adversarial systems often depend on the parties submitting such materials. Judges in inquisitorial systems may obtain COI by their own initiative, usually with the help of dedicated staff/research units/trained documentalists.13 Other systems mix the two approaches and are sometimes able in important cases to hold a preliminary hearing at which the parties are notified of relevant country materials known to the judge(s) and are asked to cover them in their submissions.

  • 10. Another source of COI comes in the form of reports written by country experts who are typically academics, researchers or journalists with considerable experience in the field.

  • 11. Despite the fact that judges are not country experts, they are often faced with having to evaluate country materials in order to make findings, where relevant, on general country conditions, for example, on whether draft evaders in Eritrea are a risk category or whether ordinary Christian converts are at risk on return to Iran. The judicial focus is always on the individual case, but individual cases can sometimes involve generally occurring facts.14 Although he must at all times avoid stereotyping,15 the judge may sometimes have to make a finding on what is generally the case in respect of one or more specific ‘risk categories’.

  • 12. The question arises, by reference to what criteria should judges evaluate background country materials?

  • 13. In approaching this question we must seek to build on the very considerable work which has been done, particularly over the past 15 years on developing reliable COI databases. UNHCR together with many other bodies have been in the forefront of efforts to develop proper systems and criteria for COI.16 UNHCR sees scope for considerably enhanced international cooperation in the field of COI, particularly at the regional level and is actively co-operating with the European Commission on a number of COI initiatives.17 Major country report-writing bodies both at governmental level (for example, the US State Department reports) and at NGO level (for example, Amnesty International) have developed their own methodologies for compiling and evaluating COI. But there are particular features of the judicial decision-making role which require us to develop and identify our own criteria. Below we offer a nine-point COI ‘judicial checklist’ which lists in the form of questions, a number of (non-exhaustive), criteria which reflect current best international judicial practice adopted when assessing how much weight can be attached to a particular COI source or reference. There then follows an explanation for each inclusion. It will be obvious that some of these criteria overlap. No single criterion should be treated as decisive. They are grouped under four main sub-headings. Whilst the ordering given is not to be seen as fixed, it is intended to reflect the usual order in which questions relating to the evaluation of COI will normally be raised.

Part I

Relevance and adequacy of the Information

i) How relevant is the COI to the case in hand?

  • 14. Relevancy is an obvious criterion; for the judicial decision-maker the primary concern is with information that is legally relevant in the sense of helping to answer case-related questions.

  • 15. Obviously there is little value in background materials that do not bear on the principal country issues that have to be determined. As trite an observation as this may sound, it is remarkable how often judicial decision-makers find nothing in background country materials directly on the point about country conditions with which they have to grapple. That does not mean that COI found by the judge to be of no or little relevance is not extremely salient in other cases or in other contexts. Relevance of the material is a judgement about the case, rather than the COI.

  • 16. Generally speaking preference will be given to reports whose content relates to asylum-related issues, for example, which deals with human rights violations and the situation of minorities and displaced persons. The pioneering Evian Report 1990 identified as a key criterion: ‘Scope – the main scope of the database would be material describing the human rights situation in countries from where there are refugees coming or likely to come’.

ii) Does the COI source adequately cover the relevant issue(s)?

  • 17. One obvious criterion for evaluating the worth of certain types of COI sources is whether or not they give a full or adequate treatment of relevant country conditions/issues. If, for example, there is an issue about the fairness of a country's judicial system, then it is obviously important that the judge should be able to learn from the evidence before him about all relevant factors, relating, for example, to the national justice system.

  • 18. Given the duty on a judge normally to consider a person's asylum claim in the context of the evidence relating to conditions in the country of origin as a whole, considerable value may be placed on reports that furnish both a detailed overview of conditions in a particular country and particulars about relevant groups and categories (for example, the position of different ethnic minorities or of vulnerable categories). Thus within the EU judges dealing with cases from Somalia have increasingly begun to have regard to periodic Joint reports drawn up by officials from several EU countries who have conducted a fact-finding mission.18 The 2004 Joint report contains sections dealing in detail with diverse aspects of Somali affairs: its history, political institutions, legal system, clan structure, the position of vulnerable categories etc.

  • 19. However, the extent to which COI that is both general and particular is required will vary from case to case and over time.

  • 20. Comprehensiveness will obviously not be an appropriate feature to expect of sources that only seek to deal with a specific incident or situation, for example, a press cutting describing recent arrests of dissidents. But it will be appropriate for reports which purport to give a detailed overview of the general country situation or to deal fully with specific issues. However, just because a report which purports to be comprehensive does not mention a particular event or fact does not necessarily mean it did not happen/is not true.19

iii) How current or temporally relevant is the COI presented?

  • 21. Most national refugee determination systems require (or allow in certain circumstances for) the judicial decision-maker to decide the issue of whether someone is a refugee or is at risk of human rights violations if returned according to the up-to-date situation.20 What is normally being assessed is ‘future risk’ by reference to the prevailing circumstances as at the date of hearing. This requirement is not an easy one for judges to apply, since the reports placed before them will by definition be dealing with events that by then are past. But in order to maintain the integrity of the decision-making it is vital, when our national legislation requires us to assess current risk,21 that we make our assessments in the light of the latest evidence and that we avoid reliance on obsolete or out-of-date COI. That can be a tall order in some cases, since even some very well-established country reports, when examined closely, can be seen to rely on sources that are no longer recent. The 2004 UNHCR COI Report highlights problems of this type.22

  • 22. It is largely because of the importance of basing decision on current information that particular value is often attached to reports which are produced on a regular or periodic basis. UNHCR Position Papers, the US State Department reports, Amnesty International reports and Human Rights Watch reports are produced annually, the latter two bodies sometimes producing additional interim or periodic reports. In the UK the Home Office Country of Origin Services reports (COIR) reports (formerly CIPU reports) on a number of countries (currently 20) are produced bi-annually in April and October. Sometimes it may be important to know about events from reliable media sources only a day or two old (for example, if there has just been a coup).

  • 23. Of course, COI can also be vitally relevant in testing or establishing matters relating to historical aspects of the appellant's experiences. As Alice Edwards put it in her paper to the November 2005 IARLJ Budapest Conference:

    ‘While “future risk” of persecution is a key question in any asylum determination, it is almost always necessary to review the individual's past experience and past practices in order to determine the likelihood of harm in the future. An individual who fled in 2000 due to serious abuses at the hands of government officials arising out of their political activities should have this information taken into account. It would produce a distorted picture of his or her claim if a decision-maker only considered the practices of the government in 2005. Historical evidence and patterns of behaviour and practices are important indicators of potential future risks.’

  • 24. Having to decide questions about current risk categories by reference to COI which is not up-to-date may not be an easy situation for judicial decision-makers in some countries, since their legal system can still require an answer on the basis of whatever evidence that is before the judge. However, as a general rule judicial decision-makers will try in such cases to avoid anything which could be taken as country guidance for other cases.

Part II

Sources of the Information

(iv) Is the COI material satisfactorily sourced?

  • 25. Depending on the context sourcing may be about accurate referencing (for example, footnoting) or about corroborating statements or reports.

  • 26. Attribution where possible increases judicial confidence in a report. A report which simply sets out its account and conclusions without making clear from where or from whom it has obtained its own information can rarely be given credence. Judges may well regard such reports as being of uncertain or unknown provenance. On the other hand, judges have to be aware that sometimes sources are anxious not to be identified.

  • 27. In a world in which there are often vested interests in how a country's human rights performance is presented, judges are understandably wary of COI or reports which depend wholly or mainly on just one or two sources. For this reason they tend to place more reliance on reports which are multi-sourced and demonstrate cross-referencing or corroboration for what they describe.23 Where there is more than one source for any particular observation contained in a country report the judge may be able to consider that that observation has been corroborated. Sometimes a judge may be able to seek corroboration in the fact that there is more than one report confirming the same point.

  • 28. The independent research unit within the Immigration and Refugee Board of Canada (IRB) employs what it refers to as a ‘Triple “C” Methodology’: compare, contrast and corroborate.24 This captures very well the need for COI from which one can see that its contents are the result of cross-checking.

  • 29. In certain cases, for example, reports which purport to be definitive on a particular issue, it may be appropriate to expect them to annex all the background materials on which they have relied, so that readers can know precisely the data on which their principal conclusions were based.25

  • 30. Much will depend on the quality of the sources cited. Judges will be wary of too ready acceptance of accounts based on obscure, unrepresentative or inaccessible sources. In Ireland, it is seen as a helpful rule of thumb for judicial decision-makers to corroborate information by taking examples from at least three strata in a ‘hierarchy’ starting with (1) intergovernmental sources, then governmental sources and international NGOs, (2) then international news reports, national NGOs, national news, then local governmental sources, local news, then (3) ordinary witnesses. Whether or not one agrees with the notion of a hierarchy – perhaps better would be the notion of perspectives from different vantage-points – recourse to different types of sources as indicated would appear to be useful.

  • 31. What the judge needs to be assured of is whether the COI is accurate, but he can only do that by reference to multi-sourced information.26 Otherwise there is no proper basis of comparison for deciding whether information given is accurate. At the same time it may be important on occasions to make allowances for the fact that the source has tried to give vital information quickly without knowing the full story, so that the outside world will begin to take an interest. Sometimes having one source will be better than none.

  • 32. It may be that on occasions information will emerge that is not or cannot easily be corroborated, yet which may be said to be highly indicative of the real situation.27 Clearly judges must always be astute to the possible value of all kinds of sources, but it remains that they are obliged to decide cases in accordance with the evidence, not hunches or inspired guesses.

  • 33. The judge also needs to assess accuracy within the context of the facts of the individual appeal.

  • 34. When considering accuracy it will always be important to keep a sense of proportion. A source may be found to contain several errors but not necessarily ones which undermine the reliability of the rest of the report. In this regard it may be necessary to consider how well-established the source is, and whether, over time, it seeks to correct and remedy inaccuracies in later reports.28

  • 35. Because we do not live in an ideal world where all COI meets rigorous standards, it is inevitable that to some degree judges will tend to attach weight to materials that have achieved an international reputation and are frequently-used: for example, reports of the UN Human Rights Committee, UNHCR Position Papers, US State Department reports, Human Rights Watch reports, Amnesty International reports.29 They will do so in part because of their need for digested information: even in inquisitorial systems, judges do not have the time to go hunting for uncollated/unassimilated country information or conducting their own statistical analyses. The rationale for considering reputation is that such sources have earned respect from many quarters for having been shown to provide a relatively reliable picture of country conditions over a significant period of time.30 The reputation may attach to the organisation or body producing the report and/or to the report itself.31

  • 36. Judges are aware, however, that even reputable sources are criticised from time to time and that it may be necessary on occasions to examine whether such criticisms are valid in relation to a particular issue and/or whether those writing the reports have acted to improve the standards of their reports.32 We have also to be aware of new bodies in the field with emerging reputations as providing reliable country data, not necessarily known to the judicial decision-maker.33

  • 37. Furthermore, reliance on a source because it has an established reputation may not always assist, for example, when two well-established sources adopt opposite or conflicting views or where an eminent expert disputes for cogent reasons what is said in an established source.

  • 38. For these reasons, although considering the reputation of a source may be justified on pragmatic grounds, it is not itself a criterion going to the merits of the COI directly.

v) Is the COI based on publicly available and accessible sources?

  • 39. The pioneering 1990 Evian Report identified as a basic criterion:

    ‘Public Material – the database would contain only public material, including non-conventional and unpublished material provided it is from a named and traceable source’.

  • 40. This criterion remains of enduring importance.34 The Report closely related it to the requirement of access to databases containing only public material. Part of the thinking behind the requirement that material be public is that it should be clear to the asylum-seeker what evidence is available and where it can be found and that he should be able to make use of it in support of his asylum claim and/or appeal. This helps achieve an ‘equality of arms’ between the decision-maker and the claimant. A further factor here is user-friendliness: qualities such as appropriate formatting, divisions into appropriate headings and clear tables of contents will assist here.

  • 41. Obviously there will from time to time be a need to consider confidential data, for example, testimonies of human rights researchers in a country of origin who cannot disclose their identities directly without placing themselves at risk,35 reports whose authors are bound by professional ethics not to disclose the identity of a particular source. Whilst this may raise difficulties about the accuracy of the informant's material, the weight to be attached to the information may be greater if the reason for anonymity is explained or if it possible to assume that the publisher of the report is an organisation of sufficient probity to ensure the source will have been checked insofar as it is possible to do so. But, subject to exceptions of this kind, COI may only be viewed as generally reliable if it is in the public domain and transparent as to its authorship.

vi) Has the COI been prepared on an empirical basis using sound methodology?

  • 42. Just as judges are not country experts, neither are they social scientists. Nevertheless, they will naturally attach more weight to sources that demonstrate in transparent fashion a sound empirical basis for their principal findings. There is a premium on objectively verifiable facts. Sometimes even methods of obtaining statistical information will need to be scrutinised. It will ordinarily be apt to ask of a document, two particular questions, ‘How does the source know what it says it knows?’ and ‘To what extent is it based on opinion and to what extent is it based on observable or established facts’?

  • 43. One aspect here is to what extent a source is based on reports from persons ‘on the ground’ in a particular country. One of the reasons why UNHCR Position Papers are often accorded considerable weight is because it is known that in relation to many countries UNHCR relies for its evaluation, not only on background sources, but also on reports from UNHCR staff that are posted in the particular country concerned.36

  • 44. Credit is also seen to accrue to reports identifying in explicit fashion what their own data-gathering methods and processes are. For example, the Preface to the US State Department reports37 for 2004 stated that:

    ‘Throughout the year, our embassies collect the data contained in it through their contacts with human rights organisations, public advocates for victims, and others fighting for human freedom in every country and every region in the world. Investigating and verifying the information requires additional contacts, particularly with governmental authorities. Such inquiries reinforce the high priority we place on raising the profile of human rights in our bilateral relationships and putting governments on notice that we take such matters seriously. Compiling the data into a single, unified document allows us to gauge the progress that is being made. The public release of the Country Reports sharpens our ability to publicise violations and advocate on behalf of victims. And submission of the reports to the Congress caps our year-round sharing of information and collaboration on strategies and programs to remedy human rights abuses – and puts us on the path to future progress’.38

  • 45. The wording of this Preface has been criticised for disclosing a US foreign policy bias39 and one can certainly see that it does contain several value-judgments. Equally it does this in a way which lays bare the principal focus of its concerns and it is arguable that transparent statements of this kind permit the reader to take account of any bias that results. But it should not be ruled out that in particular instances, despite reports being transparent in this way, their stated agenda or value judgments may get in the way of objectivity, for example, by being too heavily influenced by that country's foreign policy concerns. In relation to US State Department reports, for example, it could possibly be argued, especially in relation to countries in which the US is presently involved in the internal affairs of a country (for example, Afghanistan, Iraq) that its reports lacked independence. Having said that, the clear primacy US State Department reports place on the monitoring and gauging of the human rights performance of particular countries (by reference to international human rights norms) may be thought to render such reports of particular assistance to judges. That is because by and large judges, when determining whether a state of affairs is persecutory under the Refugee Convention or contrary to international human rights guarantees, likewise seek to base their decisions on internationally accepted standards as enshrined in public international law. Much the same can be said of reports by international NGOs such as Amnesty International which clearly pursue a number of political goals (for example, trying to shift world opinion against capital punishment) but try as far as possible to assess country conditions by reference to methods of analysis40 and evaluations based on international human rights norms.

  • 46. Another aspect has to do with methodology. It may not be easy to place great reliance on a source which states, without giving any relevant background facts and figures, that there are ‘reports’ or ‘incidents’ or ‘cases’ of detainees being tortured in custody. Obvious questions arise in respect of such statements. How many cases? In which prisons (all or just some)? Involving what type of prisoners (political/ordinary)? If a report gives specific figures of persons reported to have suffered human rights abuses in detention, they will generally carry more weight if they include relevant comparators: for example, what is the prison population in the relevant country? If a report refers to certain human rights abuses being widespread or routine or frequent, but elsewhere indicates small numbers of persons are affected, that will tend to detract from the weight such evidence may be given. Questions of scale and frequency can be vital in assessing risk. In the UK, for example, in Harari [2003] EWCA Civ 807, the Court of Appeal has held that for prison conditions in general in a particular country to be considered as giving rise to a ‘real risk’ of persecution or treatment contrary to Art 3 of the ECHR, there has to be shown ‘a consistent pattern of gross, frequent or mass violations of fundamental human rights’.41 On the other hand, judicial decision-makers must always be astute to real constraints that may affect data-gathering in certain countries, for example, the authorities might deliberately prevent journalists or others from learning anything about certain detention centres, or official statements may significantly downplay the real numbers of detainees involved, etc.

  • 47. The excellent reputation of particular sources (whether they be governmental, such as, the US State Department Reports, or non-governmental, for example, Amnesty International and Human Rights Watch) should not deter the judicial decision-maker from scrutinising their methodology and data-gathering research methods as much as that of any other source. Nor can it be automatically assumed that because past reports from a particular body have normally been of a high standard that the specific report before the judge presently measures up to the norm.

Part III

Nature/Type of the Information

vii) Does the COI exhibit impartiality and independence?

  • 48. For credence to be placed on COI it is essential for the judicial decision-maker to be satisfied that it is not partisan or affected by bias. Although this is an elusive criterion to state with any precision, it is clearly a very important one. It is elusive because of the recognition that there is no such thing as ‘value-free’ assessment of country conditions. Arguably every report adopts a particular vantage point. As can be seen from their Preface,42 US State Department reports are an example. However, it remains that perceptible bias or partisanship or having an ‘axe to grind’ may be seen as reducing the value of a particular report.

  • 49. To this end judges need always to pose a number of critical questions of any source so as to evaluate its purpose, scope and authority.43 It may add value to a report that it is known to emanate from an independent source, for example, a report prepared by a reputable research body dedicated to compiling reliable data for use by international agencies.

  • 50. Nevertheless judges should be cautious of being too judgmental about such matters. For example, it may be that the only recognised country expert on a particular country is an émigré who has aligned himself (or herself) to a particular political group in exile. One of the reasons why he may have come to be regarded as an expert is that he has ‘frontline’, on–the-ground knowledge of recent events. If a report from such a person nevertheless exemplifies an objective and balanced treatment of relevant issues, it may be given as much (if not sometimes more) weight as if it came from an academic body or source with no apparent political colouring.

  • 51. In respect of reports from governmental agencies, or from joint government fact-finding missions, it may be necessary to consider whether there is any governmental bias. Factors of some importance are the extent to which the agency or agencies in question can be said to be shielded from political pressures by having a separate budget coupled with administrative independence from the decision-making authority.44 A further safeguard may be an independent monitoring body able to check on the quality and accuracy of ongoing reports.45

  • 52. The language and tone in which COI reports are written is also of some importance. Reports that frequently resort to hyperbole or employ emotive terminology, or that contain rhetorical and prejudicial phrases, risk not being taken seriously.46

  • 53. In respect of country experts it is important to establish what material has been provided to that expert (other than that relating to the claimant's individual history). Has he referred to the most recent COI? If he takes a different view as to risk than that taken by established sources such as UNHCR, what are his reasons for doing so? Has he taken an empirical approach to the evidence? Do the facts he identifies logically support the inferences he draws from them? Does he provide sources for his various statements? Is he bringing direct knowledge of relevant political events or political actors to bear or is he simply relying on (and making inferences from) very much the same body of evidence which is before the judicial decision-maker? Has he noted evidence or opinions which are contrary to his? What are his credentials?

  • 54. Judicial experience of country expert reports may count for a lot here. For example, if judicial decision-makers see over time that a particular expert constantly seeks to paint a worse (sometimes rosier) picture than do other recognised sources, this may lead to a conclusion that the expert has lost the right to be considered impartial and has become an advocate. As was said by Collins J in the UK Tribunal case of Slimani:47

    ‘In all cases, we have to distil the facts from the various reports and documents. Bodies responsible for producing reports may have their own agenda and sources are not always reliable. People will sometimes believe what they want to believe and, aware of that, those with axes to grind may feed willing recipients. Many reports do their best to be objective. Often and inevitably they will recount what is said to have happened to individuals. They will select the incidents they wish to highlight. Such incidents may be wholly accurately reported, but not always. This means that there will almost always be differences of emphasis in various reports and sometimes contradictions. It is always helpful to know what sources have been used, but that may be impossible since, for obvious reasons, sources are frequently anxious not to be identified. We are well aware of criticisms that can be and have been levelled at some reports and are able to evaluate all the material which is put before us in this way’.

  • 55. It is particularly important to assess the impartiality of so called ‘expert witnesses’. If their evidence is sound academically, demonstrably objective and the expert is not acting as an advocate for the applicant's case, strong weight can, and should rightly, often be given to such reports.48

  • 56. Such country experts are not usually legally trained. Nor can they be expected to have a firm understanding of the skills or concepts judicial decision-makers have to deploy when making a credibility assessment. They may not even know that their reports will end up being used in a judicial process. Matters relating to standards and burdens of proof must be matters for the judge. Consequently what country experts describe as a serious risk or danger cannot be taken as determinative of that question. This does not mean their reports are to be given no weight, or to be treated as devalued or irrelevant simply because they are unaware of the precise legal criteria.

  • 57. Even when country expert reports fail to exhibit all the characteristics of a good report, and so only limited weight can be attached to them, that does not necessarily mean they are to be entirely discarded. Such ‘expert’ reports are still part of the totality of the applicant's case which the judge has to evaluate and then apply the correct legal principles to before reaching his own conclusions.

  • 58. The independence of experts must also be considered. It may be relevant in certain cases, for example, to consider whether an expert who derives a significant level of income from preparing country reports for claimants can be regarded as independent.

viii) Is the COI balanced and not overly selective?

  • 59. Closely allied to the impartiality and independence criteria is that of non-selectivity.49 The judicial decision-maker expects a report to present a balanced account noting items of evidence that go one way and the other.50 COI which was found, for example, to ignore consistently or overlook reports of acts of impunity by police and security forces would be deeply suspect. Conversely, a report which highlighted human rights abuses exclusively, without noting evident and significant improvements in a government's human rights record, would be received with scepticism. What judges want to learn is the real picture. However, a report is not necessarily lacking in balance simply because it comes down on one side of the argument about conditions in a particular country: the balancing required here is only to take account of all relevant considerations for and against.

Part IV

Prior Judicial Scrutiny

ix) Has there been judicial scrutiny by other national courts of the COI in question?

  • 60. It is widely recognised by those involved in data-gathering in the asylum field that sources should cover case law emanating from different courts and tribunals.51 That is a valid requirement for at least two reasons. A judge's decision on a particular case may sometimes necessitate a forensic analysis of, and conclusions about, conflicting sources of evidence. Such analyses and conclusions may be of value to all. Secondly, much of the skill of judicial decision-makers in dealing with COI consists in correlating what it says about risk and dangers for particular categories with the legal concepts arising under the Refugee Convention and international human rights treaties. For example, a country report or expert may state that the risk to a particular category is ‘serious’ or ‘real’ etc. But whether such assertions are accepted as demonstrating a ‘well founded fear of being persecuted’ under Art 1A(2) of the Refugee Convention or substantial grounds for believing that there is a ‘real risk’ of treatment contrary to basic international human rights is a matter for judges to decide in particular cases. Thus the judicial decision-maker may have before him a UNHCR Position Paper which frames its evaluation of risk categories more broadly than is justified under the terms of the Refugee Convention (or even under international human rights law). It may, for example, base itself on a concept of international protection which embraces, for example, humanitarian categories such as persons fleeing from the ordinary incidents of civil war or famine.52

  • 61. For this reason judicial decision-makers benefit from sight of decisions reached in different countries. They are aware that just as refugee law judges pursue a single universal or autonomous meaning of key concepts under the Refugee Convention, so they should strive to reach common views on the same or broadly similar country data.

  • 62. We would accept, however, that reliance should only be placed on decisions from judges in other countries in limited circumstances and subject to careful review. There a number of reasons for this. Country conditions are mutable and in any event the primary focus must always be on the individual claimant's particular circumstances. It will sometimes be difficult to know the status of a decision from another jurisdiction (whether, for example, there has been a further appeal reversing the case). It may be unclear whether the court of tribunal in question has employed different standards of proof or different legal principles. However, at least within the EU, this difficulty will greatly reduce as a result of the partial harmonisation of standards brought into effect by the Refugee Qualification Directive as from 9 October 2006.

  • 63. A further difficulty is that, as the theme of our paper highlights, we are a long way away from a stage where we can be confident that judges always have to hand COI meeting all of the standards we have identified.


  • 64. The above Judicial Checklist and Explanatory Memorandum are the product of considerable discussion and exchange, involving judges as well as those active in the refugee law and policy field. Whilst they seek to reflect the views of judges generally – that is, to furnish a specifically judicial perspective – it must not be thought that they necessarily achieve that; they are only a work in progress. The COI-CG Working Party will endeavour to keep them under review and from time to time post revised versions on the IARLJ website taking into account the latest developments.

The Country of Origin-Country Guidance Working Party.
The term ‘judges’ or ‘refugee law judges’ is used here to cover all types and levels of judicial or quasi-judicial decision-makers regardless of whether they deal with asylum or asylum-related cases regularly or only occasionally.
COI has been defined as ‘[a]ny information that should help to answer questions about the situation in the country of nationality or former habitual residence of a person seeking asylum or another form of international protection’. See Barbara Svec of the Austrian Centre for Country of Origin and Asylum Research and Documentation (ACCORD), Vienna, in presentation to the IARLJ November 2005 Budapest Conference.
1979 UNHCR Handbook para 42: ‘… The applicant`s statements cannot, however, be considered in the abstract, and must be viewed in the context of the relevant background situation. A knowledge of conditions in the applicant's country of origin - while not a primary objective - is an important element in assessing the applicant's credibility’.
‘He’ is used throughout to cover both the masculine and the feminine gender.
See paper by Alice Edwards, op.cit.: ‘AI also reiterates that country of origin information alone cannot foresee the range or types of abuses that a particular individual may suffer in a given context and so cannot be relied upon to the same degree in every case’.
1979 UNHCR Handbook, para 42.
See, e.g., s.8 of the Immigration and Asylum Act (Treatment of claimants, etc) Act 2004 (UK).
Formerly CIPU (Country Information and Policy Unit) reports. CIPU was formerly part of the Home Office Asylum and Appeals Policy Directorate, but in May 2005 was moved to the government's Research Development and Statistics (RDS) section. Reports produced by this section are now called Country of Origin Reports (COIR).
See ‘Country of Origin Information: Towards Enhanced International Cooperation’, UNHCR Feb. 2004 (hereafter ‘2004 UNHCR COI Report’), para 13(iii).
Austrian Centre for Country of Origin and Asylum Research and Documentation. For a helpful list, see Elisa Mason, ‘Guide to Country Research for Refugee Status Determination’, Jan. 2002, at para 38 gives a useful list of asylum and refugee resources.
In Canada the Immigration and Refugee Board (IRB) has a research programme that makes available current, public and reliable information to all parties in the refugee protection determination system.
See UK case of Manzeke [1997] Imm AR 524 (Lord Woolf): ‘It will be beneficial to the general administration of asylum appeals for Special Adjudicators to have the benefit of the views of a Tribunal in other cases on the general situation in a particular part of the world, as long as that situation has not changed in the meantime. Consistency in the treatment of asylum-seekers is important in so far as objective considerations, not directly affected by the circumstances of the individual asylum-seeker, are involved’. See further 2004 UNHCR COI Report para 9: ‘The information needed to assess a claim for asylum is both general and case-specific’.
See High Court of Australia case, Applicant NABD of 2002, Case Ref.: [2005] HCA 29 S70/2004.
See ‘Country of Origin Information: Towards Enhanced International Cooperation’, UNHCR Feb. 2004 (hereafter ‘2004 UNHCR COI Report’). The European Council on Refugees and Exiles (ECRE) in the 4th paper in its Way Forward series entitled ‘Towards Fair and Efficient Asylum Systems in Europe’ suggests as one of the areas of cooperation for EU Member States: sharing of existing country of origin information and coordinated use of joint fact-finding missions.
See 2004 UNHCR COI Report, para 7ff.
For example, the joint British, Danish and Dutch fact-finding Mission (17-24 Sept. 2000); the joint British and Danish fact-finding mission to Nairobi, Kenya, and Baidoa and Belet Wayne, Somalia (20 May to 1 June 2002), ‘Report on political, security and human rights developments in southern and central Somalia, including South West State of Somalia and Puntland State of Somalia’; the joint Danish, Finnish, Norwegian and British Fact finding mission to Nairobi, Kenya (7-12 Jan. 2004), published 17 Mar. 2004, entitled ‘Human Rights and Security in Central and Southern Somalia’.
In this regard it must not be overlooked that bodies involved in the production of COI are often working under pressure and may be under-resourced.
In systems which confine assessment to an error of law or judicial review approach, it may be that all that can be examined is whether the evaluation made by the original decision-maker was within the range of reasonable responses, i.e., not perverse or irrational. However, where a material error of law is found, some countries then allow at that stage for the appeal to be considered on its merits, in the light of the latest country information: see, e.g., the position in the UK of the Asylum and Immigration Tribunal as analysed by the Court of Appeal in R (Iran) [2005] EWCA Civ 982.
In the US it is apparently risk at the date of the application.
Para 19: ‘One general problem is that certain types of information age quickly and lose relevance when country situations can change rapidly. Collections, unless regularly up-dated, become retrospective rather than forward-looking. Another widely recognised problem is “round-tripping” when secondary sources begin to cite each other’.
To similar effect the 2004 UNHCR COI Report states at para 24: ‘Experience shows that a coherent body of information requires multiple sources and that no particular source can generally be ruled out.’
We are grateful to the IRB for its presentation to the IARLJ November 2005 Budapest Conference in which this point, among others, was explained.
In the UK it is now routine that decisions by the Asylum and Immigration Tribunal (AIT) which are designated as ‘country guidance’ cases, contain an appendix listing all the sources considered: see AIT website under ‘Country Guidance’.
Care must always be taken to ascertain whether sources are genuinely different and are not in fact based on the same primary source: this is the well-known problem of information ‘round-tripping’.
Alice Edwards, op.cit. p 5.
Alice Edwards in her paper for the IARLJ November 2005 Budapest Conference stated: ‘For AI (Amnesty International), accuracy means that researchers always seek to verify or corroborate information; that information is gathered from different sources, wherever possible; all sides of the story are to be pursued; testimonies are to be collected from different witnesses; and the information must be carefully distinguished (e.g., rumours versus allegations versus confirmed reports). AI analyses the information, identifies patterns, and chooses its language carefully, to avoid misleading or inaccurate reports.’
In a UK Court of Appeal case, R v Special Adjudicator, ex p K (FC3 1999/5888/4, 4 Aug. 1999) Amnesty International was recognised as ‘a responsible, important and well-informed body’ and judges were exhorted to ‘always give consideration to their reports’.
This is similar to UNHCR criteria: see 2004 UNHCR COI Report para 19: ‘Given finite resources and the need to enhance productivity, preference is naturally given to information and/or assessments already “digested” (evaluated from a reputable source (another government, an intergovernmental agency, or an NGO).’
See Alice Edwards, op.cit. p.3.
See, e.g., critique of US State Department reports by Lawyers Committee for Human Rights, 30 Apr. 2003, ‘A Review of the State Department Country Reports on Human Rights Practices`, before the Committee on International Relations, Subcommittee on International Terrorism, Non-Proliferation and Human Rights; ‘Critique of State Department's Human Rights reports’, by Human Rights Watch (4 Apr. 2003); Gramatikov v INS, 128 F.3d 620 (7th Cir.1997).; Kasvari v INS, 400 F 2d 675, 677 n.1 (9th Cir 1968). In Gramatikov it was said: ‘[T]here is perennial concern that the State Department softpedals human rights violations by countries that the United States wants to have good relations with’. In a recent judgement of the European Court of Human Rights in the case of Said v The Netherlands (Application no. 2345/02) 5 July 2005, Judge Loucaides in a Separate Opinion disagreed with the opinion of the majority who had viewed the US State Department report as a reliable source of information on the human rights situation in Eritrea: ‘They are not prepared by an independent and impartial institution but by a purely political government agency, which promotes and expresses the foreign policy of the United States. Therefore, they cannot by definition be relied on as a neutral and impartial exposition of the facts mentioned therein. There is always an element of suspicion that such Reports are influenced by political expediency based on US foreign policy with reference to the situation in the country concerned and that they serve a political agenda. … Therefore I do not see how any judgment of the European Court of Human Rights can rely in any way or to any extent on any US Department of State Country Reports on Human Rights Practices in respect of any country’. We are indebted for some of these references to the IAS publication, ‘Country guideline cases: benign and practical?’, Colin Yeo (ed.), Jan. 2005, Immigration Advisory Service (IAS), London.
Alice Edwards, op.cit: ‘It is also important to be aware of a judge's or a jurisdiction's own limitations in knowing “the field” or knowing what organisations exist and the types of work they are doing. Sometimes smaller or national organisations may not be known to the judge or decision-maker, but may be well-known outside of judicial circles as having a very solid reputation’.
See 2004 UNHCR COI Report and Elisa Mason, ‘Guide to Country Research for Refugee Status Determination’, Jan. 2002,
Alice Edwards, op.cit: ‘… for security reasons and personal safety reasons of both the source and the [Amnesty International] staff member, the sources relied upon in the report may not be named. AI is an organisation dedicated to researching human rights violations, commonly involving governments that do not live up to their international obligations. AI has a responsibility to its sources not to disclose their names where appropriate, but this does not and should not detract from the truth or accuracy of the information contained in a given report’.
See 2004 UNHCR COI Report Annex 1: ‘Information systems within UNHCR para 4: UNHCR papers are a result of a collaborative effort between the Regional Bureaux concerned and the Department of International Protection (DIP). This means that as a rule information is not only corroborated but also incorporates comments from experienced staff and up-to-date assessment directly from the field’. However, courts and tribunals have not always found it possible to accept the evaluation of risk categories contained in UNHCR Position Papers: see below n. 52.
These reports are prepared pursuant to ss.116 (d) and 502(b) of the Foreign Assistance Act of 1961 (FAA) as amended and Section 504 of the Trade Act of 1974 as amended. This legislation requires the Secretary of State to report to the Speaker of the House of Representatives and the Committee on Foreign Affairs of the Senate.
Country Reports on Human Rights Practices 2003, US Department of State (25 Feb. 2004), Preface. We have taken this quotation from the IAS publication ‘Country Guideline cases: benign and practical?’, above n. 32.
See above n.32.
Alice Edwards in her paper to the November 2005 IARLJ Budapest Conference paper giving Amnesty International's views stated that: ‘AI carries out on-site or field missions to many parts of the world, so the majority of the reports include first-hand knowledge and experience. AI spends considerable efforts in building networks with regional, national, local and community organisations, professional bodies, associations such as trade unions, academics, and individuals. Prime responsibility for global monitoring of the human rights situation rests with the International Secretariat with offices in 10 countries (London, New York, Geneva, Hong Kong, Kampala, Senegal, Moscow, Costa Rica, Beirut and Paris). AI also has national representation through Amnesty International Sections/structures in 75 countries. … research reports are prepared according to internal research policy that endorses four main principles, namely: accuracy, impartiality, respect for confidentiality and collaborative approaches.’ Her report highlights, however, that in certain instances the methodology used varies with the type of report and so the reader must check what is said about methodology in the report in question: see p. 7.
See further, Batayav [2003] EWCA Civ 1489. The ‘consistent pattern …’ terminology is borrowed from Article 3 of the 1983 UN Convention Against Torture.
See above para 46.
These are similar to those used by bona fide researchers: see Eliza Mason, ‘Guide to Country Research for Refugee Status Determination’ op.cit. who at D suggests the following questions: ‘Who has produced the information and why? Answer this question by asking additional questions: “If it is an NGO, what is its philosophy? If an international organisation, what is its mandate? If a newspaper, what are its politics? If a government, what is its record in the area of human rights and the rule of law? If a report by a UN Rapporteur, who wrote it and under what restrictions? How independent or impartial is the producer?” Essentially “objectivity” can be established by learning something about the organisation itself, i.e., where its funding comes from, who makes the management decisions and does she have anything to gain or lose by the outcome of a case etc? …’
See 2004 UNHCR COI Report para 49. The 1990 Evian report identified as a key criterion: ‘Control body – control over the content of the database should be left in the hands of a relatively independent centre with a professional information staff, responsive to the needs of the users’. In the UK the Advisory Panel on Country Information (APCI) is an independent body established under the Nationality, Asylum and Immigration Act 2002, ‘to consider and make recommendations to the Secretary of State about the content of country information’. For further background, see Andrew Jordan, paper for November 2005 IARLJ Budapest Conference, ‘Country Information: The United Kingdom and the Search for Objectivity’.
On the UK experience, see paper for IARLJ, November 2005 Budapest conference, by Andrew Jordan, copy on IARLJ website.
See Elisa Mason, ‘Guide to Country Research for Refugee Status Determination’, op.cit.para 41: ‘What is the tone of the report? What kind of language and definitions does it employ? Given the nature of the subject-matter of many human rights reports, it is understandable that a bitter tone might resonate throughout the text. However, reputable human rights organisations are normally careful about overstating a case, and will attempt to characterize abuses according to defined categories without resorting to superlatives and angry verbiage’.
SSHD v S (01/TH/00632) 1 May 2001, para 19.
See Report from Expert Evidence Working Party, paper by John Barnes, in The Asylum Process and the Rule of Law, IARLJ Netherlands, 2006 (Manak Publications) pp.263-293.
Meaning here failure to mention all relevant facts.
See 2004 UNHCR Report on COI para 5: ‘By comparing and contrasting information from a variety of different sources, decision-makers are assisted in forming an unbiased picture of prevailing conditions in countries of concern’.
Elisa Mason, ‘Guide to Country Research for Refugee Status Determination’ op.cit. p.2: ‘Typical categories of sources include international instruments … national legislation in the country of origin … case law (decisions from administrative and judicial bodies which granted asylum or other forms of protection to an individual with a claim similar to the one you are researching) … guidelines … etc issued by UNHCR and other international bodies as well as governmental agencies … human rights reports … news reports and newswire services … background materials … experts’.
See UK case of NM (Lone women-Ashraf  ) CG Somalia [2005] UKIAT 00076 and its comments as follows:

‘This is illustrated by UNHCR position papers, such as the January 2004 one dealing with Somalia. In Somalia UNHCR has responsibility for voluntary repatriation programmes, currently confined to northern Somalia, and has evident consequential concerns referred to in paragraph 3 of this report about “over-stretched absorption capacity” even in the relatively stable northern part of Somalia. Reasons of this kind lead UNHCR to discourage signatory states from going ahead with enforced returns of rejected asylum seekers. However, the only issue arising on statutory appeals on asylum or asylum-related grounds before Adjudicators and the Tribunal is whether the claimant is a refugee and if so, whether to return a person to Somalia would breach the Geneva Convention or constitute treatment contrary to Article 3 ECHR or any other Article, where engaged. The question of absorption problems that might flow from any United Kingdom government decision to enforce returns in numbers is not of itself the basis for showing that return would breach either Convention.’

The Tribunal went on to say:

‘111.The UNHCR, in such circumstances and they arise very frequently, is pursuing what it sees as its wider remit in respect of humanitarian and related practical considerations for the return of people, particularly on a large scale. This is a common problem where the country of refuge borders the country of past persecution or strife. What it has to say about the practical problems on the ground will be important where it has staff on the ground or familiar with the conditions which a returnee would face.

112. But the assessment of whether someone can be returned in those circumstances is one which has to be treated with real care, if it is sought to apply it to non Refugee Convention international obligations, especially ECHR. The measure which the UNHCR uses is unclear; indeed, realistically, it may be using no particular measure. Instead, it is using its own language to convey its own sense of the severity of the problem, the degree of risk faced and the quality of the evidence which it has to underpin its assessment. It is often guarded and cautious rather than assertive because of the frailties of its knowledge and the variability of the circumstances.

113. This is not to advocate an unduly nuanced reading of its material, let alone an unduly legalistic reading. It is to require that the material be read for what it actually conveys about the level of risk, of what treatment and of what severity and with what certainty as to the available evidence. But there may be times when a lack of information or evidence permits or requires inferences to be drawn as to its significance, which is for the decision-maker to draw. There is often other relevant material as well.

114. UNHCR's language is not framed by reference to the ECHR and to the high threshold of Article 3 as elaborated in the jurisprudence of the Strasbourg Court and of the United Kingdom. That is not a criticism – it is not an expert legal adviser to the United Kingdom courts and couches its papers in its own language. So its more general humanitarian assessments of international protection needs should to be read with care, so as to avoid giving them an authority in relation to the United Kingdom's obligations under the ECHR which they do not claim. They may give part of the picture, but the language and threshold of their assessments show that the UNHCR quite often adopts a standard which is not that of the United Kingdom's ECHR obligations.

115. UNHCR papers are often not the only ones which Adjudicators or the Tribunal has to consider. Other organisations may have first-hand sources and differ from UNHCR; experts may bring a further perspective. A considered UNHCR paper is therefore entitled to weight but may well not be decisive.’