98718 RTI Right to Information Working Paper Designing Right to Information Laws for Effective Implementation By Toby Mendel January 30, 2015 Right to Information Series World Bank Victoria L. Lemieux, General Editor Stephanie E. Trapnell, Working Paper Series Editor strapnell@worldbank.org Working Paper No. 3 The Right to Information Series brings forward current and ongoing research on issues related to transparency and the right to information. It aims to provide a range of information on policy, practice, experience, and frontier issues related to public sector openness and transparency, including the underlying functions and outcomes of open government efforts. 1 Right to Information Series The Right to Information Working Papers disseminate the findings of works in progress to encourage the exchange of ideas about issues related to transparency and right to information. Findings are released quickly in order to stimulate discussion and further refine the conclusions of research. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Project Leader and Right to Information Series General Editor: Victoria Lemieux, vlemieux@worldbank.org Working Papers Series Editor: Stephanie E. Trapnell, strapnell@worldbank.org Titles in the Right to Information Series Right to Information: Case Studies on Implementation (2014), edited by Stephanie E. Trapnell Right to Information: Requests and Appeals Data in RTI Systems (2014), by Jesse Worker with Carole Excell Right to Information: Recent spread of RTI legislation (2014), by Toby Mendel Right to Information: Identifying Drivers of Effectiveness in Implementation (2014), by Stephanie E. Trapnell and Victoria L. Lemieux Disclaimer The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Executive Directors of the International Bank for Reconstruction and Development/The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. © 2015 The World Bank, 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000, Internet: www.worldbank.org 2 Designing Right to Information Laws for Effective Implementation Right to Information Series Working Paper No. 3 January 30, 2015 ALL COMMENTS WELCOME strapnell@worldbank.org vlemieux@worldbank.org Contents Abstract ................................................................................................................................................. 5 Introduction .......................................................................................................................................... 6 Direct Impacts ....................................................................................................................................... 6 Central Support Bodies ....................................................................................................................... 11 Bureaucratic Integration and Fit ......................................................................................................... 19 Proactive Disclosure ............................................................................................................................ 28 Conclusion ........................................................................................................................................... 29 3 Acknowledgments The author is deeply grateful to the following individuals for their contributions to this report: Perrine Canavaggio, archival expert, Helen Darbishire, Executive Director, Access Info Europe, Mukelani Dimba, Executive Director of Open Democracy Advice Centre, Maurice Frankel, Director of the UK Campaign for Freedom of Information, Shailesh Ghandi, former member of the Indian Central Information Commission, David Goldberg, right to information expert, Susan Healy, Venkatesh Nayak, Coordinator - Access to Information Programme, Commonwealth Human Rights Initiative, Anne Thurston, Director of the International Records Management Trust, Dirk Voorhoof, Professor, University of Ghent. The generous support of the Nordic Trust Fund is acknowledged as having made this study possible. 4 Abstract Over 100 countries have now adopted laws that give individuals a right to access information held by public authorities, now recognised globally as an international human right. An ongoing challenge has been the significant policy-practice gap in many countries caused by a failure to implement these laws properly. Despite this, relatively little research has been done on how to design RTI laws so as to facilitate better implementation. This paper addresses this research gap, looking at different areas where better or more careful legal design might reduce the burden on public authorities and others tasked with implementing RTI laws while also ensuring strong respect for this human right. The first area is ‘Direct Impacts’ or issues where there is a direct causal relationship between the rules and how they are supposed to be implemented, such as definitions, the regime of exceptions and the procedures for processing requests. In these areas, the main recommendations are to focus on the need for clarity and limits on administrative discretion while promoting strong overall standards. The second area – Central Support Bodies – looks at the roles accorded in the law to different central bodies, including both internal support bodies (nodal agencies) and external support bodies (independent administrative oversight bodies and courts). The paper analyses the three main roles undertaken by these bodies – dealing with complaints, supporting supply-side or internal implementation and public awareness-raising and support – and assesses various institutional design and role allocation options. It recognises that, ultimately, the ability of these bodies to discharge the three roles is in many cases more dependent on resources than anything else, but it also points to a number of design features which can help to optimize efficiency while facilitating access. The third area is Bureaucratic Integration and Fit, which looks at how legal design can better be adapted to ensure fluid integration of RTI obligations into internal bureaucratic systems. This part of the paper devotes significant attention to issues relating to the information officer function within public authorities, or the individuals/units that bear primary responsibility for many RTI implementation tasks. Issues addressed here include structural features such as how the information officer function is formally designated and its relationship with the rest of the public authority, and incentives and sanctions. This part of the paper also addresses a number of the systems that are needed to meet RTI obligations, such as records management, training and reporting. The fourth and final area – proactive disclosure – looks at different legal design options relating to this important RTI obligation. The paper ends by concluding that careful thought is needed when adopting or amending RTI legislation to ensure that the law is designed in a way that will facilitate, as far as possible, ease of implementation. This cannot, however, come at the expense of strong standards of respect for this international human right, and striking an appropriate balance between these two competing demands will remain a challenge not only for legal design but for implementation efforts more generally. 5 Introduction This paper looks at the relationship between the design of a law which aims to give individuals a right to access information held by public authorities, i.e. a right to information (RTI) law, and the successful implementation of that law. The legal framework involves both laws and subordinate legislation, such as regulations, which complement the law and are easier to amend, with the result that there is likely to be a more dynamic relationship between the design of regulations and implementation challenges. There is also, of course, the question of how laws are interpreted by the courts, as well as other players, such as oversight bodies, which can impact significantly on implementation of the law. A key issue for this paper is the fact that there is, at least in many countries, a law-implementation or policy-practice gap in the sense that implementation of the RTI law is significantly sub-optimal.1 No law is perfectly implemented, but the gap between the standards of the formal rules and what actually happens is often quite significant for RTI laws. In some settings where observance of the rule of law is low, RTI laws are almost entirely ignored and/or certain key provisions in them are routinely ignored. This sort of radical policy-practice gap makes it difficult to discuss sensibly the relationship between legal design features and implementation, which is the focus of this paper. The paper therefore focuses on contexts where there is a reasonable expectation or an established record of medium to better practice in terms of implementation. A key focus is to discuss ways to reduce the policy-practice gap through more carefully tailored legal design. Direct Impacts There are a number of areas where there is a relatively direct causal relationship between provisions in an RTI law and its implementation, in the sense that the rules essentially define the nature of implementation (subject to the issue of a policy-practice gap noted above and, of course, the question of how legal provisions are interpreted). Examples of this are the scope of the law (in terms of coverage), the procedures for making and processing requests for information and the regime of exceptions. Another commonality among these areas is that, although there is notable variance in the detailed rules in different laws, the structural approach taken tends not to vary considerably. Thus, virtually all RTI laws include a definition of the types of public authorities that fall within their ambit (i.e. that are subject to information disclosure obligations), although the definition itself varies. In terms of the scope of the law, key issues are the range of public authorities it covers, who is authorized to make requests for information (which may be everyone or may be limited to citizens) and, albeit to a lesser extent, the scope of information covered. In relation to these issues, the law defines the scope of application of the rules, and implementation essentially follows. If the parliament is excluded from the scope of the law, then it is simply not subject to information disclosure obligations. Similarly, if only citizens are allowed to make requests, that defines the scope of the law in this respect (i.e. requests from non-citizens will not be entertained). It may be noted that recognition of RTI as a human right under international law has important normative implications in this area. Under international law, human rights obligations, including the 1 This paper does not treat questions of how RTI laws are interpreted as part of the policy-practice gap (i.e. interpretation is treated as part of the policy framework rather than as implementation practice) although this is of course extremely important. 6 right to freedom of expression and hence RTI, which it encompasses, extend to all State actors. In other words, the scope of public authorities covered by an RTI law should be defined broadly, for example to include all three branches of government, as well as the full range of players that undertake public functions. The same conclusions flow from practical considerations given that the main practical drivers for RTI – such as participation and accountability – apply to public authorities defined broadly. Different RTI laws take different approaches on this issue which may have implications for implementation. In most countries, the scope of the law in terms of public authorities is defined in generic terms and it is then a matter of interpretation as to whether or not a specific authority is covered. The definition is often broad but the fact that it needs to be interpreted provides an opportunity for at least some types of public authorities to claim that they are not covered. This can create conflicts around implementation of the law and can also be abused. In some countries, however, such as Canada, Scotland and the United Kingdom, the scope of public authorities is entirely or largely defined through a list of specific authorities included in the law (usually in a schedule or annex) and the responsible minister has fairly wide discretionary powers to add public authorities to the list provided for in the law.2 The advantages of this system are that it removes any doubt or debate about which public authorities are included in or excluded from the scope of the law. However, it can also lead to a situation where the scope of the law gradually reduces as public functions are contracted out, which has become more pronounced in recent years due to the growth in public- private partnerships and a tendency to privatize government functions. This has been a problem in Scotland, for example, where a recent press release by the Information Commissioner noted: “ Rights have been gradually lost over the last 10 years as the responsibility for public service delivery is passed to third parties.”3 This approach also imposes a degree of administrative burden on the responsible ministry, given the rapid changes in terms of which public authorities exist and should be subject to the RTI law. For example, a report by the United Kingdom Ministry of Justice in 2011, just six years after the law had come into force, detailed nine orders adding bodies and six removing them.4 In some RTI laws, ‘everyone’ has the right to make requests while in others only citizens5 or sometimes citizens and residents6 have this right. In addition to the principled point made above, that everyone should benefit from human rights, these limitations impose essentially unnecessary administrative burdens on both public authorities and information applicants. The former are required to verify citizenship or residence. A superficial check of this can be done, for example by requiring applicants to submit documentary proof of citizenship or residence. However, public authorities are not well placed 2 See section 77(2) of the Access to Information Act of Canada, sections 3-6 of the Freedom of Information (Scotland) Act 2002 and sections 4 and 5 of the Freedom of Information Act 2000 of the United Kingdom. 3 Press Release issued on 15 January 2015. Available at: http://www.itspublicknowledge.info/home/News/20150119.aspx. 4 Ministry of Justice, Memorandum to the Justice Select Committee Post-Legislative Assessment of the Freedom of Information Act 2000, December 2011. Available at: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/217339/post-legislative-assessment- of-the-foi-act.pdf. Some stakeholders feel that the system fails adequately to cover all bodies undertaking public functions in the United Kingdom. For example, a wiki on proposed changes to the UK Act states that there are “many bodies with substantial public responsibilities which are not currently subject to the FOIA” and calls for these to be added. Available at: http://foiwiki.com/foiwiki/index.php/Changes_we_would_like_made_to_FOI_law#Add_more_public_authorities. 5 Some examples of this include Argentina, Bangladesh, China, India, Indonesia, Nepal, Pakistan, the Philippines, Poland, Russia, Thailand and Uganda. 6 Some examples of this include Canada, Israel, Lithuania, New Zealand and Zimbabwe. 7 to verify the accuracy of such documents, while some individuals, particularly those who are more marginalized, may lack the requisite documentation. This also imposes an additional step (i.e. an additional barrier) in the process for both applicants and public authorities, and, in extreme cases, can lead to discrimination against marginalized groups that make it difficult for them to exercise their right of access to information In some cases, interests such as national security are posited as justifications for these sorts of limitations, but this makes little sense in the modern world where citizens can hardly be expected to pose no risk to security and where a determined foreigner could almost always find a citizen to make a request on his or her behalf. In any case, sensitive national security information is not subject to disclosure under RTI laws. Another justification is that states should not have to bear the burden of satisfying requests by foreigners. It is unclear how significant a burden this is, although anecdotal evidence suggests that it is small. More significantly, it may be assumed that, in a majority of cases, foreigners’ requests support national public interests, for example if they are making a request for purposes of research or investment in the country. Many laws simply define the scope of information covered as being all recorded information held by any public authority, regardless of the form in which that information is held. This is both simple and broad. It also limits the administrative burden on public authorities since, when a request is received, the process can be initiated by simply identifying any information they hold which is responsive to that request. In some cases, laws limit the scope of information covered for example through qualifications such as “serving official purposes”,7 being “used for official purposes”8 or having been “created or obtained in the exercise of [official] functions”.9 These sorts of qualifications impose an unnecessary additional obligation on public authorities, namely to assess whether the information meets the prescribed condition. This is also problematical because these conditions are almost inherently vague, meaning that the determination of whether or not they are met is likely to be subjective, and this also opens up an additional potential ground for appeal. There is also no particular operational reason for imposing limitations like this. A few countries even impose the substantive condition on information applicants of having a specific interest in the information they are requesting.10 This runs against the international principle that applicants should not be required to provide reasons for making requests for information, as well as the idea that one should not have to provide a justification for exercising a human right. A second area where provisions in RTI laws tend to define the nature of implementation rather directly is the procedures for making and processing requests for information. Typically, these rules describe what must be included in a request for information and how such a request may be lodged with public authorities (including whether it needs to be on a specific form and how the request may be communicated to the public authority). In terms of making requests, these rules also typically impose an obligation on public authorities to provide assistance to applicants who need it and to provide receipts acknowledging requests. The rules on processing requests normally establish timelines within which this must be done, and set out what fees may be charged. They may also describe how to deal with requests 7 Section 2 of the German Federal Act Governing Access to Information held by the Federal Government. 8 Section 2(i) of the Pakistan Freedom of Information Ordinance 2002. 9 Article 2 of China’s Ordinance on Openness of Government Information. 10 See, for example, Article 8 of Uzbekistan’s Law on Principles of and Guarantees for Freedom of Information. 8 for information which is not held by the public authority, as well as the ways in which access should be provided (such as via inspection, or a physical or electronic copy). Although these rules tend to be rather technical in nature and, they can have a profound impact on the success or otherwise of the system from the standpoint of how they encourage or limit accessibility. First, the manner in which requests for information may be lodged is likely to affect the way these requests are processed. It will always be important to retain non-digital forms of access (i.e. providing physical, paper copies of documents). At the same time, digital, and especially Internet-based, requests have some obvious advantages from the perspective of both applicants and public authorities (assuming both are digitally enabled). Such requests are much easier to make (this can be done without one ever leaving one’s desk) and they are also easier to process and manage. In Mexico, for example, the implementation of digital forms of making requests has had a very positive effect on increasing access to the right to information among remote rural populations, often among the poorest.11 This is especially true where strong technological platforms have been designed to this end, but it is even true where such tools have not been put in place. In some cases, laws specifically indicate that requests may be made electronically, while in other cases this is achieved instead via policy tools and practical arrangements. It is not clear whether it makes any difference in practice whether these arrangements are set out in law or dealt with administratively; perhaps not, since what is important here is for public authorities to develop the systems needed in practice to receive digital requests. Assistance obligations tend to be cast in terms of requirements to provide “reasonable assistance”12 or “necessary assistance”.13 In practice, this probably leaves the matter largely up to the discretion of public authorities both because what is reasonable is highly subjective and depends on the level of resources which have been allocated to this issue (i.e. how much time the specialized information officer(s) has to do this) and because of barriers to appealing against a perceived failure to provide adequate assistance. The legal rules regarding timelines can also impact in important ways on implementation. The most common approach is to provide for a time limit for responding to requests – typically ranging from around ten to thirty days – and then allow for it to be extended, as needed in the context of more complex requests, typically for a similar length of time. If the conditions for extending time limits are too broad or vague, officials are likely to abuse their discretion to delay the provision of information. In a number of countries, however, time limits can be extended essentially indefinitely,14 and this has often resulted in lengthy delays in responding to requests.15 11 Silvana Fumega, El Uso de Las Technologias de Informacion y Communicacion para La Implementation de Leyes de Acceso a La Informacion Publica (2014: Santiago, Chile, Consejo para la Transparencia). 12 See, for example, Article 17 of Guyana’s Access to Information Act, Articles 5(3) and 6(1) of the Indian Right to Information Act, and Article 19 of the South African Promotion of Access to Information Act. 13 See, for example, Article 38 of the Serbian Law on Free Access to Information of Public Importance. 14 This is the case, for example, in Canada, Thailand and the United States. 15 In a case from Canada, Information Commissioner v. Minister of National Defence , 3 March 2014, 2014 FC 205, the Minister of Defence claimed it needed 1110 days beyond the original 30 days to process a request, and the Federal Court held that no refusal to provide information takes place before this deadline passes, thereby limiting the appeal opportunities of the applicant. Case available at: http://decisions.fct-cf.gc.ca/fc- cf/decisions/en/item/67139/index.do. In the United States, historic delays in processing requests have been well documented and the adoption of the Open Government Act in 2007 was aimed at part in addressing those problems. See: http://www2.archivists.org/statements/saa-statement-on-delays-by-federal-government-agencies-in-responding- to-foia-requests. 9 A similar situation pertains to fees. In many laws, fees are limited to the costs of reproducing (photocopying) and sending (mailing) the information.16 In such cases, applicants do not have to cover the human resource costs associated with finding, assessing and processing the information, all of which depend at least to some extent on the degree of internal efficiency in relation to records management and assessing and applying exceptions. In some cases, these fees are set centrally, so as to avoid a patchwork of fees across public authorities. In other countries, however, public authorities are allowed to charge requesters for the costs of processing requests. Many aspects of this – such as how much time is spent assessing the applicability of exceptions – are quite discretionary in nature, and this can lead to very high charges being levied on applicants.17 In most cases, RTI laws establish their own regimes of exceptions, while a few countries simply refer to exceptions provided for in other laws. The issue of the relationship between the regimes of exceptions in RTI laws and secrecy provisions in other laws is dealt with below. Otherwise, regimes of exceptions tend to follow a roughly similar pattern, which involves describing legitimate aims which need to be protected against disclosure of information, imposing tests, often referred to as harm tests, for when information may be withheld to protect the legitimate aims (i.e. when disclosure of the information would harm the aim), and providing for a public interest override which requires disclosure even in the context of a risk of harm to a legitimate interest where this would serve broader public goals (such as exposing corruption or the truth about human rights abuses). Both the harm test and the public interest override flow directly from the human rights status of RTI, as conditions on restrictions on this right. There is a standard list of exceptions in most RTI laws, along with some ‘special’ ones in some laws. For the most part, the regimes of exceptions in RTI laws are fairly sophisticated legal constructs which lie at the heart of the RTI system. They often pose challenges for officials, and even specialized information officers, given their legal complexity and the need for specialized expertise to apply them. However, exceptions almost inevitably provide officials with a measure of discretion in terms of application, which is often exercised against disclosure.18 This discretion derives in part from the undefined nature of some exceptions; for example, exceptions to protect internal processes are often drafted in a vague manner while some concepts, such as national security, are inherently difficult to define. But discretion here also flows from challenges in applying the harm test, which effectively provides space for officials to claim harm even when the risk of it materializing is very remote. Finally, the public interest override is almost inherently subjective, once again effectively affording officials discretion in applying the law. Discretion in this area is obviously problematic, but there are limited tools in terms of legal drafting to address this. Attempts can be made to ensure that exceptions are drafted in as clear a manner as possible and, in particular that they clearly identify the interest to be protected. Many laws refer to vague notions when carving out an exception for internal processes instead of clarifying exactly what 16 This is the case, for example, in Estonia pursuant to Article 4 of the Public Information Act, in El Salvador, pursuant to Article 4 of the Ley de Acceso a la Informacion Publica, in Jamaica pursuant to section 12 of the Access to Information Act and in Slovakia pursuant to section 21 of the Act No. 211/2000 Coll. on Free Access to Information and on amendments of certain acts. 17 In the National Freedom of Information Audit 2009-10, conducted by the Canadian Newspaper Association, a request to the Ministry of Transportation of the province of British Columbia for the number of cell phones and how much they cost generated an estimate of $98,603 to satisfy the request, while most other provinces released the information for free (see page 31). Available at: http://www.newspaperscanada.ca/system/files/CNA%20FOI%20Audit%202010%20efinal.pdf. 18 This is reflected, for example, in the high rate of cases in many countries where oversight bodies overturn refusals to provide access to information. 10 interest is being protected, and then this can be relied upon to refuse access. Better practice is to define clearly the legitimate interests to be protected here, such as the free and frank exchange of advice within government and the success of policies (which might be thwarted by premature disclosure).19 In other cases, for example for national security and privacy, laws can provide non-exclusive lists of the types of interests involved, as a means of clarifying the scope of these exceptions. However, this can also be problematic because such concepts are often highly context dependent, so that an item on the list may justify a refusal to provide information in one context but not in another one.20 It may be better to leave this sort of elaboration to policy or guidance documents rather than including it in a binding legal document. Public interest overrides are central to the balancing that takes place between imperatives for openness and reasons for secrecy, but they can be difficult for officials to apply and they also provide for a significant degree of administrative discretion in application. Some laws simply include a general rule that information should be disclosed whenever this is in the overall public interest, despite the fact that this would pose a risk of harm to a protected interest.21 This is formally a strong test, but it provides very little guidance to officials. In Colombia, a general public interest override along these lines is accompanied by absolute overrides in relation to information that exposes human rights abuses or crimes against humanity.22 In other cases, the law provides a list of the types of public interests that might justify overriding the exceptions. In South Africa, for example, exceptions may be overridden where disclosure of the information might expose illegal acts or risks to public safety or the environment.23 This has the benefit of helping to clarify the scope and nature of the public interest override, but the disadvantage of being limited to the overriding public interests listed. Perhaps the best approach is exemplified by Bosnia and Herzegovina, where the law combines a general public interest override with a non-exclusive list of examples of when this might be engaged.24 Central Support Bodies Broadly speaking, central support bodies in different countries perform three main roles to assist in the proper implementation of RTI laws. The first is to review complaints about the way in which requests for information have been treated. This always covers cases where access to information has been refused, 19 Section 44(1)(b) of the South African Promotion of Access to Information Act provides a good example of a narrowly drafted exception in this area, exempting advice where disclosure of that advice would either inhibit the candid provision of such advice or frustrate the success of a policy through premature disclosure. The Indian Right to Information Act does not include any exception for internal documents, although it does protect Cabinet documents (see section 8(1)(i)). In Mexico, internal documents are only protected until such time as a decision is made, after which they must be published. See Article 14(VI) of the Federal Transparency and Access to Public Government Information Law. 20 For example, troop movements are normally considered highly sensitive during wartime, but at the outset of the Iraq war, great efforts were made to publicise the fact that the ships with troops from the United States were leaving, so as to give Saddam Hussein a last chance to make concessions. 21 See, for example, Article 20(b) of the Maldivian Right to Information Act and Article 4.8(c) of the Liberian Freedom of Information Act. 22 See Articles 21 and 28 of Law 1712 of 2014. See also Article 14 of the Mexican Federal Transparency and Access to Public Government Information Law. 23 See Article 46 of the South African Promotion of Access to Information Act. 24 See Article 5 of the Law on Freedom of Access to Information for Bosnia and Herzegovina. 11 but many laws also allow for complaints regarding other problems in the way requests have been processed (such as a breach of the timelines or charging excessive fees). In some cases, the law even provides for non-request based complaints about how the law is being implemented (such as whether proactive publication targets are being met). In some countries, external complaints may be lodged only after a request for an internal review of the problem by the same public authority has been made. The manner in which external complaints are addressed may include mediation, quasi-judicial decision- making and judicial decisions. The second role of central support bodies is to provide advisory and other support services to public authorities tasked with implementing the law. This can cover a range of activities including training, the development of central systems, such as tracking platforms for requests, the development of model tools, such as a model annual report or model protocol for processing requests, the provision of general advice, such as on how to interpret exceptions, and the provision of advice in response requests from individual public authorities. The third role is to undertake public promotional and awareness-raising activities, and other outward- looking support roles (such as providing advice to requesters). In some cases, this role is shared between one or more central support bodies and individual public authorities (and, on an informal basis, civil society). This role can cover an enormous range of specific activities, including public education, public events, advertising (including through the public offices of public authorities), training for potential requesters and so on. Cross-cutting this, there are three main types of central support bodies. In some countries, central internal bodies (i.e. bodies which form part of the executive), often referred to as ‘nodal bodies’, play a central support role for RTI implementation. Such bodies tend to focus on the second main role noted above and may engage in some or all of the activities outlined there. Nodal bodies do not normally engage in the processing of complaints, the first main role undertaken by central bodies, although they may engage in some external awareness-raising (i.e. the third role). In most countries, some such body, usually a ministry, at least carries out the minimal role of being responsible for adopting regulations under the law and perhaps also of having a default role in terms of law reform whenever this is proposed. Although these might technically be considered nodal bodies, the concept as used here refers to bodies that play a somewhat more active role in terms of supporting RTI implementation, including by helping public authorities meet their obligations under the law. A nodal body support role may be allocated to an existing body, such as the Ministry of Justice in Canada or the Department of Personnel and Training in India, or to a new body, such as the Pilot Committee for Access to Information in Tunisia. In most countries, these arrangements are not specifically provided for in the RTI law, but are set up administratively afterwards, although laws often do allocate legal custody over the law to a ministry (which is not, of itself, a nodal role, as indicated above). Some laws have references to wider roles that may or must be undertaken by government or a ministry. Thus, in the Indian law, the government may, subject to financial constraints, undertake public education efforts and assist in the training of officials, and it is required to produce a guide for the public on how to use the Act.25 In Slovenia, the responsible minister is required to inform the public about how to use the law, provide counseling to public authorities and undertake other promotional and developmental tasks.26 25 Article 26 of the Indian Right to Information Act. 26 Article 32 of the Access to Public Information Act. 12 In the end, and regardless of what is specifically stipulated in the law, the extent to which central nodal support takes place is probably more a function of the amount of staff and funding allocated to this role than of the law. At the same time, referring to this role in the law can help to fix both the scope of the function (i.e. what the body is expected to do) and the locus of this responsibility (i.e. which body is expected to do it). Second, many RTI laws establish an external administrative oversight body, such as an information commission, or allocate RTI oversight functions to a pre-existing body, such as a human rights commission or ombudsman. These bodies are ‘external’ in the sense that, although they are public authorities, they do not form part of the executive branch of government. In almost all cases, these bodies are tasked with reviewing complaints about the way requests have been processed (and, as noted above, potentially other implementation failures). Some are limited to this role but in many cases these bodies also often have a public awareness-raising role and it is not uncommon for them also to provide internal support to public authorities in implementing the law (i.e. they may undertake tasks relating to all of the three main central roles noted above). For the sake of completeness, it should be noted that a range of other actors often also play a role vis-à-vis public awareness-raising, such as civil society actors and, in some countries, education authorities.27 A number of legal design issues come up in relation to these bodies which can have a profound impact on their ability to promote the successful implementation of RTI laws. A primary function of most of these bodies is to perform the first main central role, dealing with complaints. It is clear that these bodies can only undertake this role effectively if they are independent of the public authorities whose actions they are reviewing. This is true of any adjudicative role, almost by definition, and is important not only to ensure that decisions are in practice made independently but also that the body has the trust of the public it serves (i.e. it is seen to be independent). The extent to which the RTI law promotes independence can have a profound impact on the actual independence or otherwise of these oversight bodies. For example, in Japan and Thailand the oversight bodies – the Information Disclosure Review Board and the Official Information Board, respectively – are effectively under the direction of the Prime Minister and so lack the requisite independence to carry out their functions.28 Some examples of how members are appointed to more independent administrative oversight bodies are India, where the President appoints the members who are nominated by a committee consisting of the Prime Minister, the Leader of Opposition and a Minister29; Mexico, where the President appoints the members (but this is subject to veto by the Senate or the Permanent Commission, a body that reviews senior appointments in the civil service)30; and Indonesia, where members are nominated by parliament and appointed by the President.31 In many countries there are also other conditions that promote the independence of members. For example, in Indonesia, members must be persons of integrity with relevant expertise. In India, the salaries of members are linked to the salaries of members of Election Commissioners. In Mexico, members must have “performed outstandingly” in a relevant field and not have been a senior civil 27 Pursuant to Articles 43 and 44 of the Nicaraguan Ley de Acceso a la Informacion Publica both school and higher educational authorities must include RTI among the subjects that they teach. 28 See Article 21 of the Japanese Law Concerning Access to Information Held by Administrative Organs and Article 27 of the Thai Official Information Act. 29 Section 12(3) of the Right to Information Act. 30 Article 34 of the Federal Transparency and Access to Public Government Information Law. 31 Article 31 of the Indonesian Public Information Disclosure Act. 13 servant or elected official. The precise provisions that are required to ensure independence will depend on the country; the point is that independence is highly dependent on the rules set out in the law. A second issue is whether or not the RTI law creates a dedicated oversight body for information (which in many countries also address data protection and/or privacy issues, given that they are closely related fields) or allocates oversight functions to an existing body, normally an ombudsman or a human rights commission. This may affect the question of how independent the body is (in principle, both ombudsmen and human rights commissions are supposed to be independent)32 as well as many of the issues that are addressed below. Although few if any comparative studies have been done on this issue, there is strong anecdotal evidence that allocating RTI oversight and promotional roles to pre-existing bodies is less robust in terms of promoting RTI than creating a dedicated body for this purpose. Resources are central to the effectiveness of this function and, although there is no technical reason why equivalent supplementary resources should not be allocated to a pre-existing body as to a new, dedicated body, in practice the former approach often results in only limited additional funding being allocated. Information functions, and especially information complaints, are very different from many of the other areas in which these bodies work, particularly for ombudsmen, and this can make it difficult for a pre-existing body to adapt well to performing an information function. Furthermore, layering an information role on top of a pre- existing role usually takes place within the framework of the existing set of powers held by the body, instead of it having powers which are more carefully to the specific needs of the information role (see the discussion below on this). Third, the formal powers these bodies have in relation to the investigation of complaints can vary quite significantly. In some countries, administrative oversight bodies have extensive powers in this area, akin to the powers of courts. These may include the power to require public authorities to produce information, including confidential information, for their review, to order witnesses to appear and to testify before them, including potentially very senior officials, and to conduct on-site investigations of public authorities. In Mexico, for example, the oversight body has the power to review any and all information,33 while both the Indonesian and Serbian Information Commissions have that power and also the power to compel witnesses to appear before them and to provide evidence under oath.34 The Canadian law allocates very wide powers to the Information Commissioner, including to summon and hear witnesses, to compel the production of documents and to enter any premises for purposes of conducting an investigation.35 Experience suggests that there is often a gap between the formal powers held by these bodies and the extent to which they can exercise them. It can be very controversial, for example, for an information commission to try to compel very senior officials, and especially ministers, to testify before it, even if it formally has the power to do so. Similarly, it is not always easy to get information from intelligence 32 On the former, see the International Ombudsman Association (IOA) Standards of Practice , available at: http://www.ombudsassociation.org/IOA_Main/media/SiteFiles/IOA_Standards_of_Practice_Oct09.pdf. On the latter, see the Principles relating to the Status of National Institutions (The Paris Principles), adopted by UN General Assembly Resolution 48/134 of 20 December 1993, available at: http://www.ohchr.org/EN/ProfessionalInterest/Pages/StatusOfNationalInstitutions.aspx. 33 Article 17 of the Federal Transparency and Access to Public Government Information Law. 34 Article 27 of the Indonesian Public Information Disclosure Act and Article 26 of the Serbian Law on Free Access to Information of Public Importance. 35 Article 36 of the Access to Information Act. 14 bodies, even if they are formally required to provide it. In addition, the ability to conduct proper investigations into complaints, which are often very numerous, is frequently as much a question of resources as of formal powers. Several recently established information commissions – such as in Yemen and the province of Punjab in Pakistan – have not yet received any resources despite being created two and one year ago, respectively, while information commissions in India lack the resources to investigate in any depth the tens of thousands of complaints many receive on an annual basis. A fourth issue concerns the powers oversight bodies have in relation to the remedies they may impose if they decide that there has been a breach of the law. There are two key sub-issues here. In many countries, oversight bodies are limited to providing remedies to individual complainants, such as access to the information they are seeking, a reduction in the amount of fees required to be paid for accessing information or the provision of information in an alternative format. These are all relatively uncontroversial remedies which go straight to the matter of implementing the RTI law in accordance with its provisions. It may be noted that, even where the oversight body has access to the information sought, which is often the case, it is not normally supposed to provide that information directly to applicants; this should instead be done by the public authority itself. This can be quite important in terms of enforcement of remedies (see below). In some cases, oversight bodies also have the power to require public authorities to provide financial compensation to information applicants where they have suffered financial loss through a failure to provide information,36 although anecdotal evidence suggests that this happens relatively rarely in practice. In a small number of countries, however, oversight bodies also have the power to require public authorities to take more structural measures to address systematic failures to implement the RTI law properly, such as to provide more training to officials or to put in place more robust systems for processing requests.37 In theory, this is a very important means of addressing structural problems in implementing an RTI law, since these sorts of issues would not normally be addressed through the individual complaints mechanisms which are designed to address failures in the proper processing of specific requests. However, further research is needed to assess how effective such mechanisms are in practice. This likely involves a number of complex administrative and relational issues, such as the ability of the oversight body to follow up on whether its remedies have been implemented (which is largely a question of resources, often very much in short supply at oversight bodies due to the high volume of appeals), as well as its effective power to coerce resistant public authorities (which will often be the very same authorities where structural changes are needed in the first place) to take the desired action. A second sub-issue regarding remedies is the extent to which the decisions of the oversight body are binding. There are three main approaches here. In many countries, oversight bodies undertake quasi- judicial processes following which they issue formally binding orders, for example for public authorities to disclose information to an applicant. In some countries, oversight bodies go through a quasi-judicial process, with both parties arguing their case. In other countries, they may simply issue findings and make recommendations, which the concerned public authority may ignore without formal consequences. In other countries, oversight bodies along the lines of ombudsmen only have the power to engage in mediations, i.e. processes that are designed to bring the parties to agreement on a solution. These may or may not result in a formal decision and, where they do, the decision is not 36 See, for example, section 19(8)(b) of the Indian Right to Information Act. 37 See, for example, section 19(8)(a) of the Indian Right to Information Act and Article 37(v) of the Mexican Federal Transparency and Access to Public Government Information Law. The Sierra Leonean Commission can impose fines on public authorities. See section 45(1) of the Right to Access Information Act, 2013. 15 binding. Many bodies with quasi-judicial powers attempt to mediate disputes as a first step, following which they engage in an adjudicative process. There is an extensive body of literature on which approach is preferable and, in particular, whether or not oversight bodies should have binding order powers.38 While the views expressed in the literature do vary, there is a clear bias towards binding order powers. Perhaps most significant are the statistics reflected in a survey of 53 oversight bodies, of which 33 had binding order powers and 20 did not.39 Of those with binding powers, 55% indicated that their orders were always complied with and another 30% indicated that a significant majority were complied with. In stark contrast, none of those with recommendatory powers reported full compliance with their decisions and only 45% indicated that a significant majority were complied with.40 In other words, decisions were complied with in at least a significant majority of cases 85% of the time for bodies with binding order powers and only 45% of the time, or about one-half as often, for non-binding order bodies. Beyond these significant statistics lie a rather complex range of factors and considerations that, furthermore, vary considerably from one country to another. For example, although the survey is not broken down by type of country, there is some evidence that compliance in developed countries with non-binding decisions is significantly higher than compliance with these decisions in developing country contexts.41 It has been argued that binding decisions tend to require more due process, although in fact this is heavily dependent on the legal framework, and some jurisdictions have developed extensive procedures even for non-binding decision-making. It has also been suggested that non-binding decision- making processes are less likely to lead to adversarial relations between oversight bodies and public authorities, although contrary arguments can also be made and more work is needed to establish whether or not this is the case. In a number of jurisdictions, oversight bodies have reported problems with enforcement even of legally binding decisions. Although a provision in the RTI law stating that decisions are binding formally renders them so, this is in practice quite a different matter from actually enforcing those decisions. In a number of countries, oversight bodies have complained that they lack any practical means of actually enforcing their otherwise formally binding decisions, in the sense of not being able to take any steps to enforce compliance if public authorities simply ignore their orders. Some laws, on the other hand, do put in place mechanisms for enforcing the decisions of oversight bodies. There is a fairly readymade solution for this in Common Law countries, namely by allowing oversight bodies to register their decisions with courts, upon which a failure to respect them becomes a contempt of court, with the attendant mechanisms for enforcing court decisions coming into play. 38 See, for example, Laura Neuman, Enforcement Models: Content and Context (2009: Washington, World Bank) and Laverne Jacobs, “Building on the Ombudsman: Polyjuralism and the Impact of Dispute Resoluti on in the Canadian Access to Information Context” in Laverne Jacobs & Sasha Baglay, eds., The Nature of Inquisitorial Processes in Administrative Regimes: Global Perspectives (2013: Farnham, Surrey, Ashgate). 39 It should be noted that this was not a representative sample of oversight bodies from around the world. For example, about one-half of the total came from developed jurisdictions, although these represent only about one- quarter of all countries with RTI laws, and nearly one-quarter (12) came from just two countries (including sub- national bodies), namely Australia and Canada. 40 Centre for Freedom of Information, In the Experience of Information Commissioners: The Information Commissioners’ International Exchange Network Survey 2014, November 2014, table 3.2. Available at: http://www.centrefoi.org.uk/edocs/pdfs/centrfoi_survey_november_2014.pdf. 41 In discussions with the author, the Canadian Information Commissioner has indicated a high degree of compliance with her rulings while oversight bodies in Pakistani identified very low levels of compliance with their decisions. 16 Fifth, while in some countries external oversight bodies are restricted essentially to a complaints and/or appeals role, along with a mandate to report on implementation,42 in many other countries oversight bodies also play a role in terms of both internal support (i.e. to public authorities) and external awareness-raising. This raises a number of issues vis-à-vis implementation, including, importantly, resources, since the extent to which an oversight body can discharge these functions is ultimately always bounded by available resources, essentially regardless of what the law says. Another important issue here is the relationship between the external oversight body and any central nodal body. It terms of resources, it is significant that in the 2014 survey of oversight bodies, 41% felt they were receiving sufficient or wholly sufficient funding for their work, despite a natural tendency among administrative bodies to want more resources, a sharp increase from 2013, when only 15% voted in these categories.43 It is unclear whether the significant change was based on differences in the oversight bodies participating in the survey or changes in terms of funding. Furthermore, the survey was not broken down according to the scope of activities of these bodies (i.e. whether they were just complaints bodies or had a wider support role). The way resources are allocated can have a very important impact on the independence of these bodies. Where resources are controlled by a ministry, for example, this can undercut independence while giving a budget approval role to parliament tends to help insulate these bodies from political interference. In some countries, neither an external oversight body nor a nodal body plays a robust role in terms of either internal support or external awareness-raising, and this clearly leaves an important gap in terms of implementation.44 It is also a significant inefficiency, at least as far as internal support goes. A central body can develop many of the necessary implementation tools and systems for RTI that individual public authorities can then simply adapt to their particular circumstances. This will not only save considerable time and effort (as public authorities are relieved of the task of reinventing the wheel for each of their obligations) but it will also help promote consistency and quality across the public sector. This avoids a patchwork of solutions across government, which can give information applicants a feeling that government decisions are capricious or unfair. Imagine, for example, being charged significantly different rates for photocopies at two different public authorities. And in some cases – for example in the areas of training and tools for tracking requests – there are even more significant inherent advantages of centralization. In other countries, both an external oversight body and a nodal body play these roles, while in yet other countries either one or the other does. There are potential advantages to having nodal (i.e. internal) bodies conduct at least some of the internal support roles, particularly those relating to systems that need to be integrated into internal administrative arrangements. Depending on where they sit within the bureaucracy, nodal bodies may have binding powers over public authorities, which external oversight bodies generally do not have (apart from in relation to complaints), although this is not unknown. External bodies in some countries are able, for example, to require public authorities to use certain forms for requests, meet certain minimum proactive publication standards, establish a website within a certain period of time, or adopt binding fee schedules for public authorities.45 42 This is the case, for example, in Canada and India. 43 See Centre for Freedom of Information, note 40, table 1.2. 44 This is the case, for example, in Pakistan. 45 For example, in Antigua the oversight body plays a role in setting standards regarding the proactive disclosure of information and records management (see sections 11 and 12(3) of the Freedom of Information Act), in the Maldives the oversight body undertakes the same functions (see sections 38 and 39(c) of the Right to Information 17 Even where nodal agencies do not have formally binding powers, they can still use the fact that they are internal bodies to play a powerful coordinating role on these issues, something an external body is less able to achieve. Specifically, they are more likely to enjoy the trust and confidence of public authorities than a body which, when it is not advising them, is potentially making adverse decisions against them (as is the case with most external oversight bodies). Furthermore, nodal agencies, by virtue of the fact that they are government bodies, are more likely to be aware of and sensitive to the particular constraints facing public authorities, as well as how to integrate new systems into existing administrative arrangements. This means they will be better positioned to propose and design solutions that are more carefully tailored to the particular circumstances of the bureaucracy, i.e. which are more appropriate and effective. At the same time, in some countries, such as Mexico, a powerful external oversight body with a broad legal mandate has played a very significant internal support role. Furthermore, certain internal roles, such as training, can be done perfectly well by external oversight bodies, which often have the advantage of having better relations with civil society, a group that can also play a very significant role in facilitating effective implementation. There can also be advantages to having oversight bodies play a role – either by leading directly or at least by being consulted – in relation to regulatory activities, such as setting fees or elaborating on rules relating to exceptions, among other things, because they will often be seen as being a more neutral player in this respect than a ministry or other internal nodal body. The matter is rather different with respect to the issue of public education, where there is less risk of overlap and duplication and where it is more a question of letting a thousand flowers bloom (i.e. more is generally better). In this case, both internal and external central bodies, as well as individual public authorities and a range of other players (civil society, the media, education bodies) all have a role to play. For example, public bodies with extensive physical interfaces with the public – such as the education and health sectors and the police – have a readymade way of raising public awareness, even just by exhibiting posters about RTI at their public offices. Central bodies – whether internal or external – can conduct media campaigns and develop promotional materials, while civil society groups can collaborate with external oversight bodies on attention getting activities such as award ceremonies to celebrate good and expose poor performers. Finally, in most countries, the courts ultimately play a final role in adjudicating disputes under the RTI law, most obviously in relation to complaints, often as an appeal from a decision by an administrative oversight body. Courts can also potentially adjudicate other disputes, for example in relation to the enforcement of (formally binding) decisions of external administrative oversight bodies.46 Many RTI laws say little or nothing about this role, because it falls within the scope of the inherent legal oversight role allocated to courts on a constitutional basis. In some countries, the law refers generally to the ability of the courts to hear RTI appeals, either directly47 or from a decision of the administrative oversight body.48 Act) and, in Mexico, the oversight body sets proactive publication and records management standards, and also approves extensions of the period of classification beyond 12 years and prescribes the form for making applications for information (see, respectively, Articles 7, 16, 15 and 32(VIII) of the Federal Transparency and Access to Public Government Information Law). 46 Some RTI laws do allocate additional roles to the courts. For example, in India, the President or Governor can refer a case to the Supreme Court asking it to determine whether or not a member of the Information Commission ought to be removed from office. See Articles 14 and 17 of the Indian Right to Information Act. 47 See, for example, sections 37 and 38 of the Ugandan Access to Information Act. 18 In a few countries, such as India, the law seeks to limit the oversight role of the courts,49 although this has not generally been successful, in the sense that the courts have still asserted their inherent constitutional power to review administrative actions.50 In a few countries that do not have an administrative level of appeal, such as Uruguay, the RTI law provides for special procedures for RTI appeals before the courts, essentially as a means to mitigate the difficulty of bringing a case to the courts. These include shorter timelines for processing such cases and also the right of the applicant to benefit from legal aid.51 In this way, the law seeks to address the two main shortcomings of court actions in this area – namely that they take too long and cost too much for most applicants. While this cannot entirely replace the need for an administrative level of appeal (not to mention the other roles that such administrative oversight bodies play), it is at least a way of making the courts far more accessible to information applicants. Bureaucratic Integration and Fit One of the most complex issues for any RTI system is to try to ensure that there is effective integration of the system into the bureaucracy which, in turn, depends at least in part on the legal design of the system and how well it fits or meshes with the way the bureaucracy actually works. There are a number of overlapping sets of issues involved here. An overriding issue is the extent to which there is internal support within both the bureaucracy and the government (i.e. including political players) for implementation of the RTI law. Where senior bureaucrats within a particular public authority, and/or the minister responsible for that authority, are essentially hostile to RTI, this can create significant barriers to effective implementation of the RTI law. Mixed messages can be provided to civil servants which formally signal support for what is ultimately a legal obligation while at the same time indicating in more subtle ways that it is not a priority or, worse, that administrative resistance is desired. Studies have shown that there is considerable power within bureaucracies to frustrate proper implementation of an RTI law, without formally breaching the legal rules (or at least not exposing oneself to effective legal liability).52 This is not necessarily something that is directly impacted by the content of an RTI law, perhaps unless the law contains provisions that are quite unreasonable from an administrative point of view (which is rare). However, bureaucratic buy-in can be affected by the way the law is developed. Where officials 48 See, for example, Article 4(4) of the Indonesian Public Information Disclosure Act. 49 See Article 23 of the Indian Right to Information Act. 50 Indeed, at one point the Supreme Court of India even required judges to sit on the panels of the Information Commissions that were deciding cases, although this was later withdrawn. See Namit Sharma v. Union of India, Writ Petition (Civil) No. 210 of 2012, 13 September 2012 and Union of India v. Namit Sharma, Review Petition [C] No. 2309 of 2012 in Writ Petition [C] No. 210 of 2012 with Review Petition [C] No. 2675 of 2012 in Writ Petition [C] No. 210 of 2012. 51 See Articles 25-30 of Ley Nº 18.381 Derecho de Acceso a la Informacion Publica (Law Nº 18.381 on the Right of Access to Public Information). 52 See, for example, Roberts, Alasdair (2002): “Administrative Discretion and the Access to Information Act: An internal law on open government?” 45 Canadian Public Administration, 175, available at: http://www.aroberts.us/documents/journal/cpa_hrdc_02.pdf and Rick Snell (2005): “Usin g Comparative Studies to Improve Freedom of Information Analysis: Insights from Australia, Canada and New Zealand”, presented at 6th National and 2nd International Congress on the Right to Information, Mexico 8-11 Nov. 2005 (Mexico City: National University of Mexico), available at: http://ricksnell.com.au/Articles/Snell_Mexico_2005.pdf. 19 were consulted as part of the process of developing the law and feel they had effective input into its design, this can help reduce opposition to the law once it comes into force. On the other hand, where officials were excluded from the process and feel that the law is being forced on them, this can generate feelings of hostility and opposition. Consulting with officials when the law is being developed can also help ensure that it is more appropriately tailored to civil service realities and the constraints within which officials operate. This is not to suggest that robust standards regarding the right to information should be watered down to suit officials but, rather, that consulting with officials can identify ways to improve legal design which facilitate implementation without sacrificing quality and that these should be taken advantage of where possible. A first set of issues relates to the extent to which RTI rules and responsibilities are fully integrated into wider formal bureaucratic systems. Implementing RTI is much like rolling out any major project in government, and it is not likely to be successful if it is not integrated into major planning processes relating to budgeting, human resource allocation and so on. It is also important that the range of rules and policies which bind staff at public authorities are aligned with, or at least do not contradict, the RTI rules (the issue of aligning other laws is addressed below), so as to protect officials against being bound by contradictory sets of instructions. The precise rules and policies which are relevant here will depend on the context. In some cases, employment contracts, which may have been designed some time ago, impose broad duties of secrecy on civil servants. In some cases, internal codes of conduct may include rules on secrecy and there may well be other internal organizational rules or systems that do the same. RTI should also ideally be incorporated into internal incentive systems, at the very least for dedicated information officers, whose job descriptions would presumably include a significant focus on RTI responsibilities. In addressing this issue, care needs to be taken to ensure that the other incentive systems do not promote potentially conflicting objectives. For example, formal duties of loyalty to the civil service, rather than to serve the public, may be interpreted as suggesting that civil servants should protect the service against embarrassing disclosures. There are also less formal ways of incentivizing support for RTI, such as prizes or other forms of recognition for strong performance in this area. Equally important are disincentives for openness, which can seriously undermine proper implementation of an RTI law and which may need to be countered by positive incentives. RTI laws rarely, if ever, directly address these sorts of rules and systems. However, clarity in the law about the fact that it imposes direct obligations on public authorities can be important in signaling indirectly that there is a need to adapt internal systems to facilitate RTI implementation. A second set of issues revolves around how to design the system of information officers, those individuals who are specifically tasked with implementing the RTI law, which are provided for in most RTI laws. There are three main sub-issues here. The first is how this function is designed structurally, especially for larger public authorities, as well as authorities which have different offices in different locations around the country or even different types of offices (such as a ministry of health, which may have a central headquarters, provincial administrative branches, hospitals and rural clinics, and even be responsible for individual doctors). In many countries, the law simply provides for the appointment of an information officer. In other countries, however, more sophisticated design features associated with this function are built directly into the RTI law. In Mexico, for example, the law requires each public authority to appoint both a liaison 20 section and an information committee (which includes the head of the liaison section as a member). The former is responsible for such activities as processing requests, proactive publication, assisting information applicants and training, while the latter is more of a policy body that generally coordinates and supervises, adopts relevant procedures and policies, and oversees classification and records management.53 It is at least implied in the Law that the liaison section should be a group of people, as necessary, but otherwise the law does not indicate how it should be structured. In India, every public authority must designate as many information officers “as may be necessary to provide information” to those who request it, and processing requests and providing assistance are the main responsibilities of these officers. Furthermore, assistant information officers must be designated at “each sub-divisional level or other sub-district level”, provided that requests are only dealt with by information officers and, where a request is received by an assistant, five days is added onto the time limit for responding to requests, presumably to accommodate for the additional time needed to transfer the request to the centre.54 In El Salvador, each public authority is required to have an access to information unit, which shall be designed to reflect the characteristics of the authority. Auxiliary units should be established as necessary taking into account the structure and size of the authority and the oversight body may recommend the establishment of additional units based on the same criteria.55 In several other countries, the idea of appointing as many information officers as are necessary has been incorporated into the law.56 It is reasonably clear that public authorities that receive very large volumes of requests will need to appoint more than one information officer. Furthermore, as the Indian example demonstrates, some system should be put in place for facilitating the receipt and processing of requests from ‘remote’ locations (i.e. locations other than the headquarters of a public authority). This is particularly important, but also challenging, in lower capacity settings where access to the Internet and even literacy rates may be low. While this sort of detail does not necessarily need to be set out in the law, and doing so may even introduce unhelpful rigidities, it might be useful for the law to at least set clear standards about the sort of service that is expected from the information officer, so as to ensure consistency and minimum quality standards across the public service. A second sub-issue is the scope of responsibility of information officers. As noted above, in Mexico the various bodies (liaison sections and information committees) are responsible for requests, proactive publication, training, classification and records management. In Indonesia, the law refers generally to the responsibility of information officers to develop a “fast, easy and reasonable” information service, without defining what that means.57 However, a legally binding regulation adopted by the Central Information Commission stipulates in more detail the responsibilities of information officers, including for records management, proactive publication, preparing a list of information held by the public authority, processing requests, including by deciding on the applicability of exceptions, public educational activities and serving as the secretariat for the processing of internal appeals.58 A similarly detailed list of tasks for the information officer is established by the implementing circular adopted under the Tunisian law, which includes processing requests for information (along with some detail as to what this involves), preparing an action plan for implementation activities, preparing a guide for the 53 Articles 28 and 29 of the Mexican Federal Transparency and Access to Public Government Information Law. 54 Section 5(1)-(3) of the Indian Right to Information Act. 55 See Article 48 of the Ley de Acceso a la Informacion Publica. 56 See, for example, section 17(1) of the South African Promotion of Access to Information Act. 57 Article 13 of the Indonesian Public Information Disclosure Act. 58 Information Commission Regulation Number 1 of Year 2010 About Public Information Services Standards, 30 April 2010. 21 public and leading on the reporting that public authorities are required to do under the law.59 In India, in contrast, the formal scope of responsibilities of information officers under the law, as noted above, is limited to processing requests, as is the case in many countries, while in others, the law only includes even more general language as to the tasks of information officers.60 It seems clear that it is helpful for the law to set out in reasonable detail the main responsibilities of information officers, if only for the sake of clarity. There is another reason, however, why this is important. If no one is identified as having lead responsibility for key tasks such as annual reporting on implementation or planning and monitoring, it is all too easy for these tasks to fall by the wayside, as it were. For example, in its 2012-13 Annual Report the South African Human Rights Commission states: “Since the passage of the PAIA, compliance by public bodies with regard to the submission of [the annual] 32 reports has been low.”61 At the very least, identifying responsibilities in the legal framework avoids a responsibility gap for these key implementation functions. Beyond the legal framework, it is useful for the position of information officer to be a formal designation and to be accompanied by terms of reference or at least integrated into the job descriptions of the responsible officials. This can be important as a planning and responsibility tool, and it can also help ensure institutional recognition of the fact that certain individuals have been designated as information officers. Otherwise, this task can all too easily be simply added to the existing responsibilities of individuals, which is both unfair to them and also likely to result in low levels of effort being made to implement the rules. Systems for formally designating information officers, along with clear terms of reference, are in place in some countries but, at the same time, anecdotal evidence suggests that in many countries individuals are designated as information officers without any formal process at all, and without any formal definition of their tasks. A third, closely related, sub-issue is how the information officer function is integrated into the bureaucracy. In Tunisia, for example, the implementing circular for the RTI law stipulates that information officers must be at least of rank A1, a fairly senior grade within their bureaucracy.62 An important issue is whether this function is seen as a proper career track position within the bureaucracy – for example with minimum training or expertise requirements and the possibility of advancement – a place to put staff who are not advancing in their careers, or something in between. This is in most contexts a combination of formal recognition of this role within the organizational structure and less formal issues, such as the time allocated to the position and the status actually accorded to it. A third and rather thorny set of bureaucratic integration issues, with particular implications for the information officer function, is how the system of imposing sanctions and disciplinary punishment works. Sanctions are generally considered to be an important means of signaling not only the importance of respect for an RTI law but also that obstruction will not be tolerated, and as a counterweight to the often rather powerful ‘culture of secrecy’ forces and pressures that augur against openness. In line with this, most RTI laws provide for criminal responsibility where individuals, or sometimes only information officers, willfully obstruct access in contravention of the law. Sanctions pursuant to these provisions may only be applied by the courts, after a trial conducted in accordance 59 Circulaire d'application du décret-loi 2011-41, 5 May 2012, Clause III(2). 60 Article 36(a) of the Maldivian Right to Information Act provides, simply, that the information officer shall have the “responsibility to provide information to the public requesting for the information under this Act”. 61 The Promotion of Access to Information Act Annual Report 2012-2013, p 33. 62 Circulaire d'application du décret-loi 2011-41, 5 May 2012, Clause III(2)(a). 22 with the strict due process standards that must be observed before criminal penalties may legitimately be meted out (which may be fines and/or imprisonment). In part due to this, and perhaps in part also due to a sense that, as inappropriate as obstruction of access is, it rarely rises to the level of criminal behavior, such sanctions are in practice applied only exceedingly rarely, and never in many countries.63 In a number of other countries, the law provides generally that obstruction of access or breach of its provisions may lead to disciplinary sanctions as provided for in internal disciplinary systems.64 It is unclear how effective such provisions are, but anecdotal evidence suggests that they are of limited usefulness and that disciplinary sanctions of this sort are in practice rarely if ever applied in many countries. This may to some extent flow from the fact that there are limited internal bureaucratic incentives to apply such disciplinary sanctions. Put differently, if sanctions are intended to address the bureaucratic culture of secrecy, it may not make sense to locate responsibility for applying those sanctions within the bureaucracy. Another approach, best exemplified in India,65 is where the oversight body has the power to impose penalties directly on officials for various breaches of the rules, including an unreasonable failure to provide information (and, in that case, the Information Commissions also have the power to report individuals for disciplinary breaches).66 This system of sanctions is perhaps the only one which has in practice been applied more frequently, and it is often claimed to have played an important role in fostering positive implementation of the law in India. A major recent study stated: “The provision to allow for imposition of penalties under the RTI Act is widely seen as the clause that gives the law its teeth.”67 In Panama, the law also allows individuals who have been affected by a denial of access to information to bring a civil suit against the responsible official.68 It is, of course, very important that any system of sanctions is appropriately aligned with the allocation of responsibilities and powers. In other words, it is only appropriate to impose sanctions where the individual (or entity) in question actually bears effective responsibility in the sense that he or she wields sufficient power to decide individually on the action that attracts the sanction. It is not fair to impose sanctions on an information officer, for example, where the failure to provide the information was due to other officials refusing to cooperate with him or her or where administrative arrangements severely hamper access, for example because records management is very weak. The Indian law does something to mitigate this problem by requiring all officials to cooperate with the information officer.69 63 See Thomas M. Susman Ashwini Jayaratnam David C. Snowden Michael Vasquez, Enforcing the Public’s Right to Government Information: Can Sanctions Against Officials for Nondisclosure Work?, available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2295466. 64 This is the case, for example, in Tunisia (see décret-loi n°41 daté du 26 Mai 2011 relatif à l’accès aux documents administratifs des organismes publics tel que modifié par et complété par le décret-loi n°2011-54 du 11 Juin 2011, Article 20). See also Article 21(1) of the Romanian Law on Free Access to Public Information. 65 But also found in Bangladesh (section 27 of the Right to Information Act), the Maldives (section 71(a) of the Right to Information Act) and Nepal (section 32 of the Right to Information Act). 66 Section 20 of the Indian Right to Information Act. 67 RTI Assessment and Advocacy Group (Raag) and Samya-Centre for Equity Studies (2014), Peoples’ Monitoring of the RTI Regime in India, 2011-13, p. 118. The study noted that, between January 2012 and November 2013 3,870 penalties had been imposed by information commissions. The study is available at: http://www.rti- assessment.com/raag---ces--rti-study-2011-13.html. 68 See Asamblea Legislativa Ley N° 6 Que dicta normas para la transparencia en la gestión pública, establece la acción de Hábeas Data y dicta otras disposiciones, section 21. 69 Sections 5(4) and (5) of the Indian Right to Information Act. 23 It seems likely that systems of sanctions that are never applied, even over time, are unlikely to achieve whatever ends they were designed for. Although a deterrent effect might apply for a while simply due to the existence of sanctions, over time it seems reasonable to assume that this effect would wane and eventually largely disappear. The difficulty of applying sanctions in many countries also highlights the need for positive measures to support implementation. A shortcoming of individual sanctions is that they focus on individual failures to provide information but, as alluded to above, often the problem is more systemic within public authorities. It is both unfair and ineffective to punish individuals for problems that are essentially rooted within the overall organizational culture and systems. As noted above, some countries grant oversight bodies the power to require public authorities to undertake structural reforms, while yet others provide for sanctions to be imposed on public authorities which systematically fail to implement their RTI obligations,70 but these measures are relatively rare. The complex issue of the relationship of the RTI law to other laws and, in particular, secrecy provisions in other laws also affect the way sanctions work. This is a particularly thorny issue at the present time, given heightened concerns about national security in many jurisdictions. In many cases, formal legal coherence is relatively easily achieved by stipulating either that the RTI law dominates in case of conflict or that it does not. However, this simple means of resolving legal conflicts does not provide a practical solution for officials, due to the complexity of the way legal systems work. In assessing legal dominance, for example, courts are likely to take into account a number of factors, including the date of the law (so that even if the RTI law explicitly provides for its own dominance, this may not be credited vis-à-vis a later law on the basis that parliament clearly wanted the later law to prevail) and the degree of specialization of the law or provision (in general, laws that are more specific are deemed to dominate more general laws). But what constitutes specialization is itself a complex matter, for example in a conflict between secrecy and openness provisions in laws that specialize, respectively, in national security and transparency. Officials simply cannot be expected to understand and interpret these legal complexities. They should certainly not be penalized for interpretations that are ultimately deemed to be wrong (i.e. for releasing information that is later deemed to be exempt). One solution to this problem, which is found in a quite a few RTI laws, is to provide protection for officials as long as they acted in good faith and/or reasonably when disclosing information.71 In practice, given the array of factors that militate against disclosure – including historic practices of secrecy and prevailing cultures of secrecy, as well as the fact that incentives tend to be stacked up against disclosure – the release of exempt information can be expected to be fairly rare, although of course there will inevitably be some instances of this. Absent protection for good faith disclosures, however, and where there are sanctions for obstruction of access, officials will find themselves in the unenviable position of being compelled, at risk of sanction, to disclose information under the RTI law while potentially risking often even more serious sanctions for wrongful disclosures. On balance, the benefits of providing such protection clearly outweigh the risks of wrongful disclosures. 70 See, for example, section 8(5) of the Nigerian Freedom of Information Act which provides for fines to be imposed on public authorities for wrongful denial of access. Section 45(1)(iv) of the Sierra Leone Right to Access Information Act grants the oversight body the power to impose fines on public authorities for breach of the law. 71 For example, section 39(1) of the Ethiopian Proclamation to Provide for Freedom of the Mass Media and Access to Information provides: “No person is criminally or civilly liable for anything done in good faith in the exercise or performance of any power or duty under this part of the proclamation.” 24 A fourth issue concerns the annual reporting obligations on public authorities and often also on central authorities that are found in many RTI laws. The content of these reports normally covers what has been done by the public authority to implement the law, including in terms of the processing of requests. This provides an invaluable and essentially irreplaceable flow of information about how the RTI law is functioning, which is important to address shortcomings and other problems, and generally to understand what is happening in this area. There are different models for this. In some cases, every public authority is required to produce a dedicated report on RTI implementation, while in others they are only required to include a section on RTI in their general annual reports. A third model is to place an obligation on public authorities to cooperate with the central body as required to help it produce its own report.72 As with many of the issues addressed in this paper, very little comparative research has been done in this area.73 However, the third model seems to include an inherent design flaw since, absent a general requirement to report, public authorities are unlikely to collect the requisite information on a systematic basis in the first place, and are thus unlikely to be able to provide this information upon request (i.e. from the central reporting body). Otherwise, there would not appear to be a major difference in terms of functionality between the other two models. A dedicated report would probably tend to be more in-depth and comprehensive, but the delivery rate of included reports might be better. In terms of the former, better practice, reflected in many RTI laws, is to provide a detailed list of the types of information that should be included in such reports, whether this is part of a larger annual report or a dedicated report on RTI.74 As noted above, this is an area where performance by public authorities in many countries falls short of the legal requirements, with reporting often only being done by a small percentage of all public authorities. One possible reason for this is that many laws fail to allocate specific responsibility for the preparation of such reports, which would presumably normally fall to the information officer (see the discussion on the tasks of such officers above). Obviously such shortfalls seriously undermine the potential benefits of reporting, since conclusions about how the system is working will be skewed (i.e. because one could expect generally better performing public authorities to be more heavily represented among the public authorities which did file annual reports). In some countries, the RTI law allocates the task of central reporting to the independent oversight body while in others this is done by a nodal or other body. It is not clear whether this makes any difference, although reporting done by an independent body might be more credible and the recommendations might be expected to be more robust. 72 See, for example, section 40(2) of the Freedom of Information Act of Trinidad and Tobago and section 30(5) of the Right to Information Act of Bangladesh. 73 See, however, Requests and Appeals Data in Right to Information Systems: Use and Compliance Report for South Africa, Jordan, Thailand, the United States, United Kingdom, India, Mexico, and Brazil (July 2014), part of the World Bank’s Right to Information Working Paper Series. Available at: http://siteresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/Resources/285741- 1343934891414/8787489-1344020463266/8788935- 1399321576201/Requests_and_Appeals_RTI_Working_Paper.pdf. 74 Such lists are found, for example, in section 5.3 of the Liberian Freedom of Information Act, Article 43 of the Serbian Law on Free Access to Information of Public Importance and section 11 of the Yemeni Law Regarding The Right of Access to Information. 25 A fifth area where it is important for RTI systems to be designed carefully so as to fit as well as possible into pre-existing bureaucratic systems is in the area of records management. Good records management is, for fairly obvious reasons, central to the successful implementation of RTI laws; if public authorities need to spend a lot of time locating information which is responsive to requests or, worse yet, cannot locate the information at all (even though they hold it or should hold it), that will obviously create barriers to successful implementation and will also increase the cost of operating the system. Despite this, records management provisions in many RTI laws are very underdeveloped and simply consist of general language about how public authorities are under an obligation to manage their records well, so as to facilitate implementation of RTI obligations.75 There are several reasons why provisions along these lines are unlikely to be effective. First, they fail to set any standards for records management, and so lack any yardstick against which the performance of a particular public authority might be measured. Second, they place responsibility for developing systems of records management on each individual public authority, despite the fact that this is a specialized area where taking advantage of a body with expertise in this area, such as an archival body or RTI oversight body, would reap significant advantages (in terms of efficiency, better quality standards and consistency across the public service). Third, and related to the first, this sort of approach fails to build in any enforcement system both because it lacks any clear minimum standards to enforce against and because it fails to provide for any specific mechanism of enforcement (see below). A natural way to address this problem is to give a central body the power to set binding records management standards and to establish some system to enforce those standards. In some countries, the rules on records management found in the RTI law are supplemented by provisions in dedicated archival legislation. However, in many cases such legislation, instead of addressing how records should be created and managed during the active life of the record at the originating public authority, focuses more on the retention of records (i.e. by original authorities after a pre-set time limit), the preservation of records (by the archival body) and the process of selecting records for destruction. An alternative, and potentially far more powerful, approach, found in countries such as the United Kingdom, is to allocate specific responsibility to a central actor to set minimum records management standards. In the United Kingdom, for example, this is done by the Lord Chancellor, who is also responsible for The National Archives, which in practice prepares these standards for formal adoption by the Lord Chancellor. Interestingly, enforcement of these standards rests largely with the Information Commissioner’s Office rather than The National Archives. This may be due to the fact that the National Archives has limited capacity for enforcement resting, as it does, within the executive branch of government, while the Information Commissioner’s Office sits outside of the executive branch. However, even so, enforcement is relatively weak in nature, and consists essentially of providing guidance, albeit of a public nature, rather than more coercive measures. This also allows for the strengthening of records management standards over time, inasmuch as the rules indicate that the Lord Chancellor “shall issue, and may from time to time revise, a code of practice” on records management.76 This approach allows for improvements and adjustments as records management systems change, for 75 A typical example of this sort of provision is Article 22(5) of the Croatian Act on the Right of Access to Information, which states: “The information officer shall undertake all the actions and measures necessary for keeping the catalogue in an orderly manner, for which he will be directly responsible to the head of the public authority.” 76 See sections 46-48 of the Freedom of Information Act of the United Kingdom. 26 example in response to technological developments, or as records management capacity increases, which is especially important in lower capacity settings. A sixth issue is how the delivery of training for officials is addressed in the RTI law. This varies considerably. In many countries, training is described in the RTI law simply as an activity that falls within the mandate of the independent oversight body, or perhaps an internal nodal body, which may then undertake it to the extent that its resources permit.77 In other countries, each public authority is formally responsible for ensuring that its staff receives appropriate training,78 even though the fulfillment of this role often lies beyond the immediate capacity of that body (i.e. in terms of expertise). In yet other countries, the matter is not addressed at all in the RTI law. Ideally, information officers should receive fairly specialized training on the right to information whereas it is sufficient to provide other officials with far less intensive training, although they do need at least some training in this area. Providing both initial and ongoing training – to both information officers and all officials – is a huge task that requires some prioritization, especially during the early stages of implementation of a new RTI law. It is unclear how effective it is to formally impose training obligations on individual public authorities in an RTI law. Anecdotal evidence suggests that this has little impact and that what really drives training, as is the case in several other areas, is resources and support from central bodies, which in this case may include civil society organizations, perhaps with the benefit of funding from the international community. On the other hand, it is probably quite important for the law to include training within the mandate of at least one or another central body (and more tends to be better here), so as to make it clear that they are allowed to work in this area. Otherwise, it can be expected that adequate funds to support this work will not be forthcoming. A common theme with many of the issues addressed in this part of the paper is that they are hard to enforce or lack effective enforcement frameworks. This is a difficult challenge in many countries because it is not clear who might play an oversight role here or how these rules might be enforced. In some countries, NGOs and/or oversight bodies have come up with interesting encouragement and public exposure (carrot and stick) ways of promoting better practices, such as public events where they give out awards to good performers and expose poor performance. In some countries, as noted above, administrative oversight bodies have the power to order public authorities to take general remedial action in some of the areas discussed above – including reporting, records management and training – but only where such problems are exposed in the context of an individual complaint. An even smaller number of RTI laws, as also noted above, provide for sanctions to be imposed on public authorities per se for serious failures to implement the law. One possible approach to explore would be to allocate more of a role to the external oversight body to undertake suo moto monitoring of public authorities so as to identify problems in these areas, along with the power to order appropriate remedial action where needed. Such an approach needs to be 77 See, for example, Section 5(2)(g) of the Liberian Freedom of Information Act and section 83(3)(e) of the South African Promotion of Access to Information Act. In Mexico, both the oversight body and the information officer units are given this role, but this does not appear to be obligatory for either. See Articles 37(XII) and 28(VI), respectively, of the Federal Transparency and Access to Public Government Information Law. 78 See, for example, section 15(5) of the Irish Freedom of Information Act, Article 43 of the Nicaraguan Ley de Acceso a la Informacion Publica and Article 42 of the Serbia Law on Free Access to Information of Public Importance. 27 designed carefully, taking into account the challenge many oversight bodies already face simply in securing compliance with their (binding) orders in the context of individual complaints. An alternative might be to put in place a public reporting and response system, whereby administrative oversight bodies have a mandate to make public recommendations to public authorities regarding systematic remedial needs and those public authorities are then required either to implement those recommendations or to explain (publicly) why they are not. Proactive Disclosure Proactive disclosure is a key part of any modern RTI law as well as of any system of public transparency. If anything, proactive disclosure has become even more important in recent years, with the massive growth in access to the Internet globally and other technological developments. Although the very early RTI laws did not include proactive publication regimes, these have been a feature of most laws for some time now.79 Most laws provide for a list of categories of information that must be published by all public authorities, while some laws provide for special proactive disclosure obligations for certain types of public authorities.80 In some jurisdictions, open data initiatives have effectively started to take over from proactive disclosure provisions in RTI laws. Anecdotal evidence suggests that proactive disclosure and/or open data initiatives are more effective when undertaken within the legal and institutional framework established by an RTI law. Otherwise, there is a risk that open data initiatives leave a lot of discretion to officials to determine what will be released with little public accountability for how these choices are made or their outcomes. At the same time, there are almost conflicting issues with legal regimes for proactive disclosure, specifically as these are set out in RTI laws. In many developing countries, public authorities struggle to meet the sometimes quite extensive proactive publication obligations placed on them by RTI laws. This can lead to situations where many or even most public authorities are in breach of their RTI obligations right from the beginning, undermining the law and implementation more generally.81 On the other hand, in higher-capacity settings, the proactive disclosure obligations in an RTI law quickly become outdated (or, for older laws, became outdated a long time ago) as public authorities move towards broad e-government and open data programs and very extensive proactive disclosure practices. One way to address potentially both of these challenges, which has been tried in a few countries, is to replace or supplement a fixed list of categories of information subject to proactive disclosure with a more flexible system for levering up proactive disclosure obligations over time. These systems usually allocate some role to an independent administrative or other oversight body, such as an information commission or nodal body. This might be to set progressively higher standards over time (i.e. to increase proactive publication requirements) or to require each public authority to put forward its own proposals 79 For example, rules on this were included in all three of the laws adopted in 1982 in Australia (see sections 7A to 9 of the Freedom of Information Act), Canada (see section 5 of the Access to Information Act) and New Zealand (section 21 of the Official Information Act). 80 See, for example, Articles 15 and 16 of the Indonesian Public Information Disclosure Act, which provides for special proactive publication by, respectively, political parties and NGOs. 81 For an assessment of this problem in the Indian context, see Peoples’ Monitoring of the RTI Regime in India, 2011-13, note 67, pp. 81-6. 28 as to what it will publish, subject to this being approved by the oversight body.82 Another potential means to address this problem is to revert to a more sectoral approach towards proactive disclosure whereby, instead of setting rules centrally for all public bodies, proactive rules are set by sector. Another possibility would be to set strong proactive disclosure goals in the legislation, but give public authorities a more substantial period of time (say between three and seven years) to reach these goals, perhaps being required to meet interim targets in between. There are disadvantages associated with all of these approaches. Giving oversight bodies’ important functions in this area can burden them and designing systems around individual public authorities may result in some simply not participating in the system. At the same time, broad non-compliance by public authorities with their legal obligations is obviously also a serious problem. More research in this area is needed, perhaps along with modeling of different options to see what works best. Another key issue with most proactive disclosure systems is that they lack strong enforcement mechanisms, in common with a number of other features of RTI laws such as training, reporting and even appointing information officers. In most RTI laws, enforcement revolves largely around complaints based on the processing of individual requests, which does not extend to proactive disclosure. At the same time, there are potential problems with granting a broad right of complaint to anyone who feels that proactive disclosure practices do not fully comply with the legal rules. As noted above, in some countries oversight bodies have the power to require public authorities to take remedial measures to improve their structural compliance with the RTI law and these may extend to proactive disclosure. This can at least create some pressure on public authorities to comply with their proactive publication obligations. Conclusion There has been a massive increase in the number of RTI laws globally, with the rate of adoption of these laws picking up from around 1997 to an average of just over four per year. This flurry of legislative activity has been accompanied by an attendant growth in civil society and academic attention devoted to this issue. A nearly universal observation has been that, if it is often difficult to get (strong) RTI laws adopted, it is significantly more difficult to ensure proper implementation of these laws. This has led to an all-too large policy-practice gap in many countries whereby legal provisions on RTI are simply not implemented properly. Despite the growth in the attention being given to RTI in its different aspects, and the widespread recognition of a policy-practice gap, very few if any studies have looked at the issue of how the legal design of these laws could be adapted to facilitate better implementation and thereby reduce the policy-practice gap. This paper aims to start to fill that research lacuna by looking at a number of areas where legal design could be improved – in the sense of reducing the barriers to implementation without sacrificing in any way the right of access. It includes a particular focus on the institutional impact of legal design, especially how the rules impact on the individuals and bodies tasked with implementing the laws. The paper includes a large number of recommendations – both explicit and implicit – and also points to a number of areas where more research is needed. These will not be repeated here, but it is possible to 82 Such a system is in place in the United Kingdom. See sections 19 and 20 of the Freedom of Information Act. 29 identify some of the general ways in which the recommendations seek to improve implementation of RTI laws. One is by reducing administrative discretion. Discretion can lead to differential application by different public authorities, undermining public confidence and generating unmet expectations, and it can also provide opportunities for abuse of that discretion to prevent disclosure of information. Discretion can also impose a burden on officials, as they struggle to define what is needed in any given situation. Examples of this are calling for clear definitions of exceptions, not limiting the right to citizens (so that officials do not need to verify this), providing a list of non-exclusive examples of what sorts of public interests might override exceptions and not qualifying the definition of information in ways that would require officials to consider whether the qualification has been met. Other recommendations seek to promote approaches to institutional design, understood broadly, that facilitate better implementation of the RTI law. The benefits here include efficiency, more robust protection of the right, fairness (for example to information officers) and fostering consistency of application across different public authorities, among others. Many of these recommendations relate to the structure, power and roles of the independent administrative oversight body, as well as of the nodal body and the information officer function. Several recommendations relate to the way implementation systems are designed. These include systems for annual reporting, records management, training and proactive disclosure. In most cases, there is no one-size-fits-all solution to these systems, but experience with different approaches has helped to identify at least some of the pros and cons of each. Separate from institutional design, but closely related, are recommendations that seek to ensure that implementation of the RTI law is integrated well with other, pre-existing systems. This covers, among other things, the way in which the RTI rules and systems are integrated into bureaucratic planning and regulatory systems, including in relation to sanctions, as well as the relationship with secrecy provisions in other laws. Overall, the paper points to a large number of areas where careful thought is needed when designing RTI laws to make sure that they are able to give full effect to this human right while not placing an unduly large burden of the obligation bearers, namely public authorities. The paper also identifies a number of measures beyond legal design, but flowing from legal RTI obligations, which can serve the same overarching goal. Where RTI laws are well designed, taking into account local legal, administrative and institutional arrangements, the size of the policy-practice gap that is almost universally observed in this area should shrink. It is hoped that the recommendations in this paper will help serve that end. About the Author Toby Mendel is the Executive Director of the Centre for Law and Democracy. 30