Assessing the performance of freedom of information

Assessing the performance of freedom of information
Robert Hazella and Ben Worthy
a Constitution Unit at University College London, UK

Available online 3 August 2010.


How well has the UK FOI Act worked in practice now that it has been in force for 4 years? This article discusses how to measure the performance of FOI regimes. It presents the evidence on the performance of FOI in the UK measured against comparative data from Australia, New Zealand, Canada, and Ireland, countries with access to information legislation and similar political systems. On a range of measures, the UK Act is found to have performed reasonably well, but it also suffers from problems common to all FOI regimes. The article concludes with some observations on what makes for a successful FOI regime, and how to measure it.
Keywords: Freedom of Information; FOI; Performance

1. Introduction: the global reach of FOI

In the past two decades freedom of information legislation has moved from being a legislative “luxury” enjoyed by a few advanced democracies to becoming an accepted part of the democratic landscape. Around ninety countries now have access to information regimes in place and another fifty have legislation pending ([Banisar, 2006], [Mendel, 2008] and [Vleugels, 2009, September]). From India to South Africa and Mexico to China, states of varying degrees of development, size, and political persuasion have embraced openness and FOI. There is increasing legal recognition from national courts and from international bodies such as the European Court of Human Rights and the Inter-American Court of Human Rights that FOI is a human right ([Mendel, 2008] and [, 2009]). Alongside the rapid spread of FOI, there has been growing interest in trying to measure its impact and effectiveness. Many FOI laws are “paper” laws, passed in response to domestic or international pressures for transparency and good governance for “symbolic purposes,” with little or no implementation machinery (Relly & Sabharwal, 2009). International donors and civil society organizations have begun to develop performance measures in an attempt to evaluate the effectiveness of the new laws. These evaluation attempts are still in their infancy. Broadly they encompass four different kinds.
First, there are the attempts to compile aggregate indicators of good governance, which include indicators of transparency and accountability. Examples are provided by the World Bank Institute ([Kaufmann et al., 1999] and [Relly & Sabharwal, 2009]), and by indices compiled by international NGOs such as Freedom House, the Centre for Public Integrity, and Transparency International. Some studies also attempt to examine factors that can help determine transparency levels or perceptions of transparency, such as access to information legislation, telecommunications access, or press freedom ([Islam, 2006] and [Relly & Sabharwal, 2009]). Government departments in the lead on FOI, such as the Justice Department in the US, the Ministry of Finance in Ireland, and the Ministry of Justice in the UK, also collect and publish statistics on FOI on their respective websites. The Obama administration has ordered that each department and agency create a new FOI annual report detailing request numbers and also delays (EOP, 2009). All the above indicators are at too high a level of aggregation to provide any useful data on the effectiveness of FOI.
Second, there are cross-country comparative surveys conducted by journalists. These tend to focus on only one group of FOI requesters, the media (e.g., Lidberg, 2002), and some of the studies are concerned as much with freedom of the press as they are with FOI. Third, there are studies which consider the use of standardized FOI requests in different countries and compare the quality of the responses. These are methodologically more rigorous, but complicated and expensive to organize; only one such study has been conducted, funded by the Soros Open Society Justice Initiative (Open Society Justice Initiative 2006). Fourth, there have been attempts to analyze the impact of FOI in a particular country, often with a view to reforming the Act. White's 2007 work on New Zealand and the large scale analysis of the Indian Right to Information legislation are two good examples of this emerging trend ([White, 2007] and RTI Assessment and Analysis Group (RaaG) and National Campaign for People's Rightto Information(NCPRI), 2009). Others have begun to examine the impact of FOI on particular parts of the world, with Darch and Underwood's excellent study examining the particular issues around FOI in the developing world (Darch & Underwood, 2010). The difficulty of measuring the performance of FOI laws was the subject of a conference in 2008 organized by the Carter Centre in Atlanta. The Centre's conference paper opens by asking how we should define “success” for a FOI regime (Horsley, 2008).
The difficulties of how to assess FOI feed into a wider problem of performance measurement. Julnes and Holzer (2001) point out that there exist two distinct approaches: rational assessment of technical measurements, versus the influence of the political environment. The “rational/technocratic” approach sees evaluation as a “technological task.” In relation to FOI, this would mean a measurement of different “technical” aspects of the FOI process, such as the number of requests or the resources used. This is not to say that technical measurement is wholly objective or permanent. Legislation can be adjusted or interpreted subjectively. For example, the statutory time limit has been lengthened and shortened in different regimes. Even within regimes, different departments will have different criteria regarding what constitutes a request or what constitutes simply a question. The following analysis needs to be read with this in mind.
However, as Julnes and Holzer (2001) observe, such procedures “do not operate in a vacuum”; all reforms operate within a “political context that may weaken or bolster” the operation of the reform initiative (Julnes & Holzer, 2001, p. 696). For example, although measurements of technical aspects are often assessed relative to a goal, “the goals for objectives for a program are always vague [as] only the lofty goal evades challenge” (Julnes & Holzer, 2001, p. 696). Similarly, as seen in a number of FOI regimes, adoption of FOI may be symbolic, and agencies may be given signals to obey through “symbolic action” without actually implementing the Act (Julnes & Holzer, 2001, p. 696). The political approach acknowledges that a “number of important factors are largely beyond the control of managers, including support among elected officials and the interests of the public and the media” (Moynihan & Pandey, 2004, p. 431).
In this preliminary article we have chosen to focus on the performance data which governments themselves collect on the effectiveness of their FOI laws. Our study will seek to measure the technical aspects of operation, in terms of quantifiable measures across FOI regimes. It is important to keep in mind that FOI operates within a political context that can, and does, profoundly influence the performance of different regimes. Disentangling the two aspects is problematic as, for example, reforms generated by rising or falling political support can have a decisive impact, as seen in Ireland. We have sought to highlight where and how the external factors have a significant impact upon the performance measures and to address the complex interplay between the two.

2. History of FOI in Australia, Canada, Ireland, New Zealand, and the UK

We have chosen as our comparator countries Australia, Canada, Ireland, and New Zealand. We have done so for three reasons. First, they all have Westminster parliamentary systems, with recognizably similar political, legal, and bureaucratic cultures. Second, the operation of their FOI laws is reasonably well documented, and measured in terms of basic performance data. Third, they introduced FOI at broadly the same time. Australia, Canada and New Zealand passed Acts in 1982 in the ‘second wave’ of FOI reforms after the US FOI Act of 1966, and Ireland passed an FOI Act in 1997 as part of the third wave, slightly ahead of the UK in 2000.
The performance measures collected and published by their governments focus on: use of the Act and the volume of requests; the success and failure rate of requests; and performance of the appeal system. In no case do these provide absolute measures to assess FOI; at best, they offer a range of different proxy measures of good performance. A positive combination of the above factors, such as high levels of awareness and use, high rates of successful requests resulting in disclosure, and a strong appeals process potentially locks FOI into a positive cycle of use, learning, and improvement, in which the request process and appeal system improve and the exemptions are clarified through interpretation. Such a finding would be a sign of an Act performing well. Conversely, if FOI is not used or the appeal system is weak FOI may become locked into a negative cycle of disuse, neglect and stagnation.
FOI came into operation in the UK in 2005, so it is still in its early days, but it does well to be aware of the risk of possible future decline (for some possible scenarios, see Glover and Holsen, 2008). FOI laws can be launched with initial enthusiasm, but then undergo revisions to restrict the operation of the Act when politicians start to feel the pain, or simply suffer from bureaucratic neglect when starved of resources. After observing the development of FOI in the Australian states, Zifcak and Snell developed a four-stage typology characterizing the life of an FOI regime: initial “optimism,” increasing “pessimism,” giving way to “revisionism” designed to alter the FOI law, normally to limit its scope or performance, and then later a return to the “fundamentals” of FOI (Snell, 2001, p. 343). In each case, in line with the literature outlined above, the performance of FOI has been influenced by external events in the political environment, in particular the government's attitude towards the costs and benefits of FOI.
The case of the Australian Federal FOI Act of 1982 provides a useful illustration. Initial optimism and strong support gave way after 3 years to a series of revisions, including an increase in fees which deterred requesters, and after 5 years to gradual neglect of FOI by governments damaged politically by FOI requests ([Hazell, 1989] and [Terrill, 1998]). A review by the Australian Law Reform Commission (ALRC) in the mid 1990s highlighted three problems: conflict, lack of co-operation with the spirit of the Act, and lack of an FOI champion (ALRC, 1995). In 2008 the new Rudd government, enthusiastic about FOI in opposition, announced plans to revive FOI by creating an Information Commissioner, to publish more information proactively, and to extend the scope of the act (Australian Department of Prime Minister and Cabinet, 2009).
Ireland, after a honeymoon period comprised of explicit government support and positive assessment by the Information Commissioner, experienced similar pessimism and revision (McDonagh, 2006). After 5 years the government introduced fees, which led to a reduction in request numbers, lengthened the time period before Cabinet papers could be released, and introduced greater protections for the decision-making process (Irish Information Commissioner, 2008). The introduction of application fees, charging 15 Euros per non-personal request, cut the number of requests by almost 50% (McDonagh, 2006). The Information Commissioner describes acceptance of FOI as “uneven” with some public bodies “reluctant” and others suffering “FOI fatigue” due to resource constraints (Irish Information Commissioner, 2008).
New Zealand's Official Information Act is widely regarded as a model of how progressive access to an information regime should work. The Act's success in its early years was assisted by the strong support of the Prime Minister (Hazell, 1989). New Zealand's “revisionism” in 1987 went against the grain of the typology, as it increased the scope of the Act and limited the way in which a ministerial veto (to prevent release) could be used. The Act has succeeded in its aim of gradually extending the boundaries, with Cabinet papers and advice to ministers being regularly published. But even in New Zealand, requests from the media or opposition parties still cause friction and political pain (White, 2007).
In Canada the Act was implemented with diligence, although the public uptake was slow ([Gillis, 1998] and Access to Information Review Task Force [AIRTF], 2002). The Canadian FOI regime was hampered by a negative reaction following a series of early controversies resulting from FOI requests, one involving the Prime Minister himself ([Gillis, 1998] and [Hazell, 1989]). This led to a combination of resistance, decreased resources, and informal systems designed to limit the impact of requests (Roberts, 2006). The Act is also increasingly hampered by the fact under section 6 of the Act requests have to be made in writing and cannot be made electronically (MOJ, 1985).
The above regimes will be measured against the UK FOI Act of 2000, which came into force in 2005 and is the most recent Act of the compared regimes. The UK Act had a peculiar gestation, having begun as a ‘radical’ white paper which was then translated into a more realistic draft Bill, and then an Act ([Hazell, 1998] and Public Administration Select Committee [PASC], 1999). The Act was passed in 2000 but not enforced until 2005. Pessimism set in early in the UK: implementation was delayed until the latest possible date allowed by the statute. But since then it has been implemented diligently and used effectively, as is shown by the performance data set out below.

3. Performance data on the operation of FOI

Essentially, governments collect data to answer five questions:
• How much is the Act used? How many FOI requests are there?
• How many FOI requests are granted?
• How many FOI requests are refused, and for what reasons?
• How many refusals are taken to appeal?
• How many appeals are successful?
The data they collect and publish are the main quantitative data available about the performance of FOI. These data provide an important addition to the qualitative assessments of how different FOI regimes are working, which inevitably contain a strong subjective element. But the reliability of the quantitative data should not be overstated. In all FOI regimes there are problems of defining what counts as a FOI request, and the figures of usage almost certainly understate the real volume of requests. The figures record those requests which the government has decided to treat as formal FOI requests. These are likely to be the more difficult requests; many easy requests, granted informally, do not get counted. Nevertheless such data can and is used to measure FOI. Piotrowski points to the example of certain federal departments in the US that “formally or informally” seek to assess and measure FOI performance, showing that, although the measurements are not “common” across government, they are “attainable” (Piotrowski, 2007, p. 51). Indicators used include time taken to process requests, number of appeals undertaken, and reduction of backlog; though these are not necessarily the most appropriate (Piotrowski, 2007).
There are also real difficulties in comparing the figures between different countries. These difficulties typically arise for four reasons. First, there are differences of jurisdictional and geographical coverage: the jurisdiction of the federal governments in Australia and Canada is more limited than that of the governments in Ireland or New Zealand, which are unitary states. Second, there are differences between the laws: for example, there are those countries which initially included access to personal files within their FOI regime (e.g. Australia), and those which had a separate Privacy Act (e.g. Canada). Differences also exist in terms of the type of appeals system (whether using a commissioner, an ombudsman, or tribunal) and how the Ministerial veto can be deployed. Third, there are differences of coverage in terms of the number of agencies subject to FOI: the UK has exceptionally wide coverage, with an estimated 100,000 public bodies being subject to the Act all at once, whereas Ireland implemented FOI over the course of a number of years. Finally, FOI in Ireland and the UK took place within a very different context than it did in Australia, Canada, and New Zealand.
The latter group of countries legislated before the information revolution or the spread of computer and information and communications technology (ICT) had begun to take hold. Britain and Ireland enacted FOI within the context of an information revolution that has made government more open and information easier to use, store, access, and distribute, which has led to very different ideas regarding citizens’ rights in relation to it. This may have affected use levels and interest in the second wave of FOI regimes. All of the figures in the subsequent tables should be read subject to these caveats about the different coverage of each country's respective FOI regime. But, even allowing for these differences, there are striking differences in some respects, and strong similarities in others.

4. Requests: numbers and patterns of use

The first indicator of a healthy FOI regime may be the number of requests, though this may be dependent on who is making the requests and for what information. Heavy use by business or the media may produce a very different type of FOI regime than one driven by requests from the public. If the Act is well publicized by government and the media, public awareness of the Act should be high, enabling a high rate of use. In practice only a tiny proportion of the population make FOI requests; but in other advanced democracies, which have recently introduced FOI, the rate of use of FOI is even lower than in our comparator countries. Compare, for example, the volume of requests in Switzerland (which introduced FOI in 2007). In the first 2 years they experienced twenty times fewer requests than in the UK.1
Of course, Switzerland is a federal government with strong regional authorities which may be the subject of requests instead. But even so, with twenty times fewer requests, the much lower numbers in Switzerland is striking, pointing towards the impact of wider political context and culture, notably Switzerland's more open political system and consensual style of democracy, in shaping how FOI legislation performs (see Holsen and Pasquier, 2009).
The volume of requests in the early years of FOI in Australia, Canada, Ireland, and the UK is shown in Table 1. (New Zealand is not included in this table because they collect no data on the number of requests.)
Table 1. Number of requests in Ireland, Canada, Australia, Switzerland, and to UK central and local government in the first three years of operation of FOI as a percentage of the population.

Year 1
Year 2
Year 3
Average per year as per 1000 of population

Total Percent Total Percent Total Percent
UK central government 38,108 0.063 33,688 0.056 32,978 0.056 0.6
UK local government 60,000 0.100 72,000 0.120 81,000 0.135 1.2
Ireland 3,731 0.100 11,531 0.310 13,705 0.370 2.6
Canada 1,508 0.005 2,228 0.008 3,607 0.014 0.1
Australia 19,227 0.137 32,956 0.235 36,512 0.260 2.1
Switzerland 249 0.003 221 0.003 0.03
(Sources: [Constitution Unit, 2006], [Constitution Unit, 2007], Constitution Unit, 2008 Constitution Unit, Freedom of Information Act 2000: the third year. The experience of local authorities in England in 2007, Author, London (2008).[Constitution Unit, 2008], [Irish Information Commissioner, 2000], [Irish Information Commissioner, 2001], [Hazell, 1987], [MOJ, 2008b] and Federal Data Protection and Information Commissioner [FDPIC], 2008).

These figures illustrate how small the requester community is and how FOI requesters are, by their numbers, a select group. Even under the unlikely assumption that each request is made by a different individual, in all countries requesters comprise only one or two per thousand of the population. The UK figures are slightly lower than those of Australia and Ireland, despite the fact that the Irish Act was restricted initially to few public bodies (McDonagh, 2006). The volume of requests is much lower in Canada, but many requests there were made under the Privacy Act passed in the same year (Hazell, 1987).

5. How many FOI requests are granted?

Another way by which we can measure the performance of the Act is through the proportion of successful requests (see Table 2).

Table 2. Proportion of requests as a percentage granted and refused in Australia, Canada, Ireland, and the UK in the first three years of FOI.

Full release
Part release
Fully withheld
Year 1 2 3 1 2 3 1 2 3
Australia 62 67 25 28 13 7 5
Canada (ATI) 39 42 37 20 25 29 10 9 7
Canada (Privacy) 62 59 69 18 24 18 4 3 1
Ireland 42 52 18 19 18 17
UK 66 62 63 13 15 13 18 19 20
(Sources: [Attorney General of Australia, 1987], [Hazell, 1987], [Irish Information Commissioner, 2000], Ministry of Finance [Ireland], 1999, [Missen, 1984, July 17] and [UK Information Commissioner, 2008b]).

Here the figures do start to become more directly comparable. The main difficulty arises when drawing comparisons between those regimes which include access to personal files within the FOI regime (Australia and Ireland), and those which enable access to personal files under separate privacy or data protection legislation (Canada and UK). With this in mind, it needs to be said that significant proportions of requesters are obtaining the information they seek.
The proportion of FOI requests which are granted in the UK is higher than at a similar point in the development of the Irish and Canadian Acts, significantly so in the Canadian case. The Australian figures are high because 75% of their requests were from individuals seeking access to their personal files, where access is more easily granted (Hazell, 1987). The figures for the Canadian Privacy Act show the difference in success rates between requests under FOI and Privacy legislation. The other main difference between the UK and the other countries is that generally disclosure rates increase over time and non-disclosure decreases, whereas in the UK the reverse seems to be the case. This may be due to initial rates starting out far higher and then stabilizing to a more typical number as experience and case law bring clarity to the way in which the Act works; or it could also be the case that as volume increases, authorities are more prepared to apply the cost limit.

6. How many FOI requests are subject to delay beyond their statutory deadline?

Delay is a problem common to all FOI regimes. One of the most challenging tasks is meeting the statutory deadline, which in the case of the UK is 20 working days. Both central government and local government have improved timeliness in responding to FOI requests on the first year, which is to be expected.
Table 3 shows roughly one in five requests delayed in central government beyond 20 days, and one in ten by other agencies, excluding requests that are not being pursued or need to have the time extended (as is possible under the UK Act). Though there is little data on delays in the other commonwealth regimes, the UK appears to be doing better than Australia during the same period. By comparison, 3 to 4 years into the life of the Australian Act, just over half of requests were responded to within the time limit of 30 days, with a further one in five processed within 45 days (Hazell, 1987).

Table 3. Percentage of responses within 20-day deadline for UK Central Government not lapsed or on hold.

Year 1 Year 2 Year 3
Departments of state 70 79 79
Total 77 84 85
(Source: Ministry of Justice, 2008b).

7. How many refusals are taken to appeal?

A further performance measure lies in the appeal system. FOI regimes generally adopt one of four different models: the courts, tribunals, the ombudsman, or a specialist commissioner. The UK is unusual in that it has a two-stage appeal system, in the first stage the Information Commissioner makes a decision regarding the appeal, and in the second stage, an Information Tribunal reviews the appeal and the Information Commissioner's decision. By contrast, Ireland has an Information Commissioner; New Zealand and Australia both use the generalist Ombudsman. Australia has two parallel appeal bodies, the Ombudsman and the Administrative Appeals Tribunal (AAT), although they have recently committed to creating an independent Information Commissioner, as described above.
The Information Commissioner or appeal body can serve to rectify problems with the Act, establish precedents and provide guidance. An effective appeal body–one that polices the Act to ensure compliance and rectifies mistakes and promotes openness–is a key part of an Act performing well. By contrast, a weak appeal body–one that lacks resources, time, influence, or power–can have a negative effect upon the overall success of the Act. This has proven to be the case in Australia, where the lack of a full champion and focal point for FOI contributed to the gradual deterioration of the Act's performance (ALRC, 1995).
Two questions about the appeal process can assist in measuring the performance of the Act. What proportion of requests is taken to appeal? And what proportion of appeal rulings upholds the original decision?

7.1. How many requests are taken to appeal?

The first question can act as a “proxy” satisfaction index for the system: if few requests are appealed this may indicate that requesters are satisfied, though it could be also seen as a measure of confidence in the appeal system. It is difficult to know why a requester does or does not take the case to appeal and in any system the number of requesters using the system is very small, particularly when compared with the percentages of those receiving the information they seek (see above).
Few requests, as a proportion of the total number of requests received, have been taken to the UK Information Commissioner since 2005. This compares favorably with the number of requests taken to appeal in both Ireland and Canada and is level with the proportion in Australia (see Table 4).

Table 4. Proportion of requests taken to external appeal as percentage of overall requests in the first three years of FOI.
Jurisdiction Year 1 Year 2 Year 3
UK 1 1 1
Ireland 5 3
Canada 9 8 9
Australia 1 1 1
(Sources: [Attorney General of Australia, 1987], [Irish Information Commissioner, 2000], [UK Information Commissioner, 2008b] and [Hazell, 1987]).

7.2. How many appeals are successful?

The second question examines the results of the appeal system for requests made to the central/federal government (see Table 5). The UK Commissioner started by upholding decisions of central authorities in around 70% of all appeals, but in years 2 and 3 this proportion fell to around 60%. This is the opposite of the Irish experience, where the Information Commissioner's decisions varied less over time and confirmed decisions more. Although the figures for Australia are not available, Hazell (1987) estimated that the ombudsman varied the decision in around 16% of cases and the Administrative Appeals Tribunal in around 20%, though in the latter case many requests were dropped at an early stage.

Table 5. Number of appeals and percentage of outcomes in the first three years of FOI.

No. of complaints
Dismissed as ineligible
Decision confirmed
Decision varied
Year 1 2 3 1 2 3 1 2 3 1 2 3
Canada 162 228 369 0 0 0
Ireland 57 141 237 18 7 6 21 38 39 21 14 18
New Zealand 318 575 506 19 2 2
UK 127 384 222 72 60 62 29 39 38
Australia 273 476 378
(Sources: [Irish Information Commissioner, 2000], [Hazell, 1987], [MOJ, 2008a] and [UK Information Commissioner, 2008b]).

However, the decisions of the Commissioner should not be measured only in quantitative terms, as certain decisions can have great influence or resonance. Two controversial and high profile decisions are demonstrative of the UK Commissioner's influence. The first case involved the Cabinet minutes relating to the legal decision to go to war in Iraq. Although Cabinet minutes are potentially exempt under several provisions in the Act–though subject to a public interest test–the Information Commissioner took the view that the subject of the Iraq war was of such public interest that the information should be released (UK Information Commissioner, 2008a). The case was then appealed to the Information Tribunal, which, in a majority decision made “not without difficulty,” agreed with the Commissioner that the circumstances of the decision “created very powerful public interest reasons why disclosure was in the public interest” (UK Information Tribunal, 2009, p. 3). The release was then subject to the first use of the veto (see below).
The second case concerned the expenses of Members of Parliament, which have been a controversial subject since the passage of the Act. In anticipation of the Act, the House of Commons released details of the MP's expenses. However, FOI requests sought further breakdowns. The MPs’ anxiety over the issue was demonstrated by an abortive Private Member's Bill, designed to remove Parliament from the scope of the Act and a series of attempts to amend the Act or restrict the level of detail in expenses disclosures. The Information Commissioner ruled in favor of disclosing further information in a series of cases (Winetrobe, 2008). Ultimately, the release (initially by leak then through FOI) led to political turmoil and the resignation of a number of MPs, including the Speaker and a number of ministers. Both cases demonstrated the Information Commissioner's willingness to take strong positions on releases and, indeed, both decisions had far reaching consequences. The Iraq case led to the first use of the veto and the decision on the MPs' expenses contributed to the ongoing exposure that, eventually, culminated in a series of leaks that damaged all parties in the spring and summer of 2009. Interestingly, the Irish Information Commissioner issued a similar decision notice relating to the expenses of Irish members of Parliament in the early years of the Irish Act (Irish Information Commissioner, 1999). Ireland has suffered a more drawn out expenses scandal which led to the resignation of the speaker of the Dail in October 2009 (Sunday Tribune “How O'Donoghue fell on his sword,” 2009). The Scottish Parliament had a rather different experience, releasing all MSP's expenses following a disclosure that led to the resignation of the leader of the Scottish Conservatives, David McLetchie ([BBC, 2005a] and [BBC, 2005b]).
It will take more precise measures and more time to see the impact of such cases. However, such cases may raise awareness of FOI and encourage others to use it. So, for example, the MP's expenses stories in the UK led to requests to the BBC, police, and local government regarding the use of expenses and may also have motivated use in other regimes. By contrast, the issues around the use of the veto may lead to an exclusion ruling for Cabinet documents (see below).

8. Delay at the UK Information Commissioner

One of the key problems that have hindered the work of the Information Commissioner's office has been delay. Although there has been an improvement in the number of cases closed per year since 2005, the Commissioner is still experiencing significant problems with a backlog of cases. This is shown in Table 6, though the figures only take account of outcomes known at the end of the year and do not include, for example, the appeals referred towards the end of the year.

Table 6. Delay at the UK Information Commissioner in the first three years of the operation of FOI.

Year 1 Year 2 Year 3 Average
Cases referred 101 301 186 196
Outcomes known by end year 15 68 51 45
Known outcomes as % of cases referred 15 23 27 22
(Source: UK Information Commissioner, 2008b).
Note. This information is based only upon outcomes known by the end of the year.

Delay in the appeal system may deter requesters from appealing and act as a brake upon the expansion and development of the Act through the appeals process. Similar problems were experienced by the Irish Commissioner in the early years of the Act, who admitted in 2001 that “significant delays” of 12 months or more were “becoming the norm [in] many instances” (Irish Information Commissioner, 200, p. 12).

9. Use of the executive veto

There is one other performance measure which is not normally published by governments, which is indicative of their respect for the appeal process and the uncomfortable decisions which they occasionally must deliver. This is the use of the executive veto, the provision for which is found in the FOI legislation in Australia, Ireland, New Zealand, and the UK. The exercise of the veto is intended to be used only in exceptional circumstances to pre-empt or overturn appeal rulings. We have included it as a performance measure on two grounds: because it is a measure of (1) how the regime is working in terms of confidence in the appeals process and (2) the sensitivity of information asked for. Each exercise of the veto amounts to a vote of no confidence in the appellate authority by the government.
The veto within the UK FOI Act, contained in section 53, is to be used rare circumstances, when ministers believe the Information Commissioner has seriously misjudged where the balance of the public interest lies (Ministry of Justice, 2008a). Ministers exercising the veto must have reasonable grounds to form the opinion, inform the requester, and Parliament, of the reasons. Following the New Zealand model, the veto is a collective decision of the Cabinet and a Minister is required to consult Cabinet colleagues before exercising it (Ministry of Justice, 2008a).
The mechanism is broadly similar in other jurisdictions. The Australian veto, for example, covers Cabinet documents, national security, executive council documents, and internal working documents. It is engaged before, rather than after, an appeal decision—though its scope has been limited by the Rudd government (Campaign for Freedom of Information, 2001). The New Zealand veto covers all classes of exempt information, and the Irish veto covers law enforcement, confidential information, security, defense, and information related to Northern Ireland (Campaign for Freedom of Information, 2001). The information in Table 7 below illustrates the use of the veto in these countries for the 4 years for which information is available.

Table 7. Use of the Executive Veto in FOI regimes.
Jurisdiction Number of times used in first four years
Australia 48
New Zealand 14
Ireland 2
UK 1
(Source: Campaign for Freedom of Information, 2001).

All the countries have made use of the veto, though Australia has made use of it many more times than New Zealand, Ireland, or the UK. The relatively heavy use by Australia may be due to the fact that it can be exercised by a single Minister and thus can be used more easily.
In the UK, the first use of the veto came after 4 years of FOI. In February 2009, Justice Minister Jack Straw issued a veto to prevent the release of the Cabinet minutes on the Iraq war. Disclosure would damage the doctrine of Cabinet responsibility, and the exceptional nature of the Iraq case made it more, not less, necessary to maintain confidentiality. The government argued that “exceptional cases create an exceptional need for confidence in Cabinet confidentiality to be strong” (Straw, 2009, p. 4). To reveal the details of Cabinet discussions could, in the future, lead to Ministers being less frank in their opinions for fear of their statements becoming public knowledge (Straw, 2009). A second veto was used over discussion relating to Scottish devolution with similar reasons expressed.
Press reaction to the veto, where reported, was critical. Christopher Ames accused the Justice Minister and the Prime Minister of trying to “strangle the life” out of the Act to save the government from embarrassment (Guardian 2009 “The Act that Jack Wrecked,” 2009). An opinion piece in the Times accused Mr. Straw, who was present at the Cabinet meeting which was the subject of the request, of breaking a key aspect of the British constitution which states that no-one should be allowed to judge their own case (Times 2009 “Iraq Cabinet Minutes,” 2009). To put this into context, the use of the veto in the UK is the lowest, equal with Ireland.

10. Government and FOI

A key element of the political context that can profoundly affect how FOI works is the attitude of the government, which is central to the success of any reform (Moynihan and Pandey, 2004). As the current Australian Ombudsman pointed out:
One of the truisms of FOI law reform is that it is brought in on a wave of government enthusiasm against public service scepticism [sic] yet the public service then learns to live with it at about the same pace that government antagonism grows (McMillan, 2002, p. 10).
And, as Prime Minister Gordon Brown pointed out, and as the MPs’ expenses scandal in the UK has shown, FOI legislation “can be inconvenient, at times frustrating and indeed embarrassing for governments” (Brown, 2007, para. 62). Writing in the late 1980s, Hazell observed that “the Australian and Canadian governments find it difficult to conceal their dislike of the legislation” (Hazell, 1989, p. 202). This dislike has led to a neglect which, combined with conflict and a lack of central leadership, has seriously hindered the two countries’ performance over the long term. McDonagh claims that in Ireland the “disenchantment” in public sector bodies was also reflected at a “political level” in government (McDonagh, 2006, p. 2). Moreover, the reforms of 2003 “polarized” opinions between parties in Ireland and the debate over the Act since has been “divisive and acrimonious” (Irish Information Commissioner, 2008, p. 13).
In the UK, as in the other countries analyzed in this study, the initial implementation was well led and organized. The UK suffered from the (self-inflicted) disadvantage that the Act was applied to 100,000 public bodies at once, rather than being gradually phased in, as was the case elsewhere (James, 2006). However, the government's attitude seemed to harden in 2006, when it undertook a review of the cost of FOI, and considered the option of revising the charging regime and capping multiple requests (Frontier Economics, 2006). This review, coming at the same time as the Maclean Bill, which sought to exclude Parliament from FOI, appeared to the media to be part of a concerted attack on FOI.
However, the proposed tightening of the charging regime was dropped in October 2007 and the Maclean Bill failed to find a sponsor in the Lords. Moreover, the then Prime Minister Gordon Brown spoke in favor of FOI, arguing that there is more we can do to change the culture and make the workings of government more open [...] [P]ublic information does not belong to government. It belongs to the public on whose behalf government is conducted. Whenever possible that should be the guiding principle behind the implementation of our FOI Act (Brown, 2007, para. 63).
Brown followed this by setting up two reviews. One review considered the possible expansion of the Act to incorporate private bodies carrying out government work, which occurred in other FOI regimes including Australia. The second review considered the reform of the 30-year rule governing release of public records by the National Archives (Brown, 2007). The latter review reported in 2009 and recommended that the 30-year rule be reduced to 15 years (Dacre, 2009). The former reported in the summer of 2009, proposing a limited expansion of the Act to four other public bodies (Ministry of Justice, 2009). The significance of the Prime Minister both explicitly endorsing the FOI Act and advocating its expansion, has symbolic value and may, as in the case of New Zealand, send a powerful message that the Act must be taken seriously.
Overall, the government attitude remains uncertain, presenting contrasting messages that properly reflect the differing feelings and views of Ministers and officials across public bodies. The failure of the two attempts to amend FOI and Brown's subsequent endorsement of the Act indicate that the UK government is now more supportive of FOI than many of their counterparts in other FOI regimes at a comparable point. That said, the veto use and the drawn out controversy over MPs' expenses shows that there is doubt and tension surrounding the Act. Moreover, Brown's support for FOI was tempered by the possibility of creating an absolute exemption for Cabinet documents and Royal matters that would exclude the two areas completely from FOI (Brown, 2009).
The failure of the two attempts to restrict FOI, followed by Brown's endorsement of the Act, suggests that the UK government is officially more supportive of FOI, and more supportive than some other governments after 4 years of operation. But political support for FOI is fragile, and can change from one Prime Minister to another. The UK may yet go through a period of pessimism and revisionism about FOI, or see public expenditure cuts leading to reductions in FOI staffing levels, with consequent reductions in performance.

11. Conclusion

This brings us back to Julnes and Holzner's (2001) distinction between “rational or technocratic” assessment as existing against evaluation within a wider political context. This article has been primarily a technocratic evaluation in its assessment of how far we can measure FOI performance with numbers. The numbers do matter. They show:
• How much a FOI law is used (volume of requests).
• How long requesters must wait for a response (delays).
• How much information is disclosed (proportion of requests resulting in full or partial disclosure).
• How many requesters appeal.
• How many appeals are upheld.
Based on these technocratic criteria, the UK is found to be in the middle of the range compared with the comparator countries. It has reasonably high levels of requests; slightly higher rates of total disclosure, but also high rates of total refusal; but low rates of appeal, suggesting disappointed requesters are not so dissatisfied that they want to appeal. The UK has also been the most restrained in the use of the veto.
Problems of delay between request and response, the cumbersome nature of the appeal system, and the difficulties involved in effecting a culture change towards openness are common to all FOI regimes. Every FOI regime experiences similar problems to these, many of which stem from the inherent contradiction in FOI that seeks to create “a legal framework based on reasonableness [...] operating in an unreasonable environment,” again highlighting the political context within which FOI operates (White, 2007, p. 295). Other valuable data, which has only been collected in a few FOI regimes, could include estimates of resource costs for requests or details of the types of requester, be they private individuals, businesses, or members of the media.
But the numbers only take us so far. Governments seeking to improve their performance will not achieve that by focusing on the numbers alone. This is where the political context is so important. Above all, an effective FOI regime requires strong government commitment and political will. Officials cannot do it on their own. Given strong political support, it is much easier to put other supportive factors in place: a strong lead department, with authority across government; central support and training (often removed after the early years); an effective appeals mechanism and related clear case law; an effective fees regime, which helps to control demand and reduce administrative costs (though what constitutes a “balanced” regime is difficult to determine).
Based on these wider factors, the New Zealand FOI regime probably fares best, given its progressive openness and high level of political and official support, sustained by a wider pluralistic political culture. The UK follows New Zealand, with reasonably high rates of disclosure, a strong Information Commissioner, single use of the veto, and some explicit political support. Third is Ireland and fourth Australia, both of which, despite high levels of use and disclosure, suffer from a high level of appeals, a lack of political support and consequent restrictive reform. Canada comes last as it has continually suffered from a combination of low use, low political support and a weak Information Commissioner since its inception.
The statistics on FOI can only tell us so much. They give us little insight into whether FOI has met the objectives set by those who introduced the Act. Nor does the information tell us whether FOI has borne out any of the fears expressed in the UK and elsewhere that it could erode the traditional “pillars” of the Whitehall system of government, such as ministerial or collective responsibility.
FOI is shaped by a whole range of factors, from the mechanisms within the Act to the action of the key groups who drive FOI: the requesters, the media and the government. Moreover, as the article argues, FOI is shaped by its political environment. The state of government and media relations, the nature of the media, levels of political support and levels of public trust can all influence how FOI operates. These wider issues are ones we are examining in a pair of research projects, which seek to assess the extent to which FOI can achieve its original objectives. We hope to report the results of that in future articles.

12. Notes

Estimated population in first year of Act: Switzerland 7.5 million; UK 60 million, Ireland 3.7 million; Australia 14 million; Canada 25 million.


Access to Information Review Task Force [AIRTF], 2002 Access to Information Review Task Force [AIRTF], Access to information: Making it work for Canadians, Supply and Services Canada, Ottawa, Canada (2002).
Attorney General of Australia, 1987 Attorney General of Australia, Freedom of Information Act Annual Report 1986-87, Australian Government Publishing Services, Canberra, Australia (1987).
Australian Law Reform Commission [ALRC], 1995 Australian Law Reform Commission [ALRC], Open government: A review of the Federal Freedom of Information Act 1982, ALRC 77, Australian Government Publishing Services, Canberra, Australia (1995).
Australian Department of Prime Minister and Cabinet, 2009 Australian Department of Prime Minister and Cabinet, Freedom of information (FOI) reform: Companion guide, Australian Government, Canberra, Australia (2009).
Banisar, 2006 D. Banisar, Privacy International: freedom of information around the world 2006: a global survey of access to government information laws (2006) Retrieved June 29, 2009, from
BBC, 2005a BBC, MSP expenses published in detail: officials at the Scottish Parliament have published unprecedented details of MSPs' expenses claims (2005, December 13) Retrieved from
BBC, 2005b BBC, McLetchie resigns as Tory Leader (2005, October 31) Retrieved June 26, 2009, from
Brown, 2009 G. Brown, Statement on Constitutional Renewal (2009) Retrieved June 26, 2009 from
Brown, 2007 G. Brown, Speech on Liberty (2007) Retrieved June 26, 2009, from
Campaign for Freedom of Information [CFOI], 2001 Campaign for Freedom of Information [CFOI], The Ministerial Veto Overseas (2001) Retrieved May 24, 2010 from
Constitution Unit, 2008 Constitution Unit, Freedom of Information Act 2000: the third year. The experience of local authorities in England in 2007, Author, London (2008).
Constitution Unit, 2007 Constitution Unit, Freedom of Information Act 2000: the second year. The experience of local authorities in England, Wales, and Northern Ireland in 2006, Author, London (2007).
Constitution Unit, 2006 Constitution Unit, Freedom of Information Act 2000: the first year. The experience of local authorities in England in 2005, Author, London (2006).
Dacre, 2009 P. Dacre, Review of the Thirty Year Rule, TSO, London (2009) [Chair].
Darch & Underwood, 2010 C. Darch and P. Underwood, Freedom of Information in the Developing World: the citizen, the state and models of openness, Chandos, Oxford (2010).
Executive Office of the President [EOP], 2009 Executive Office of the President [EOP], Open Government Directive (2009) Retrieved on February 1, 2010, from
Federal Data Protection and Information Commissioner [FDPIC], 2008 Federal Data Protection and Information Commissioner [FDPIC], Fifteenth Activity Report 2007–2008, FDPIC, Bern, Switzerland (2008)., 2009, European Court of Human Rights: Right to information essential to free expression (2009) Retrieved February 12, 2010 from
Frontier Economics, 2006 Frontier Economics, Independent review of the Freedom of Information Act, Author, London (2006).
Gillis, 1998 Robert P. Gillis, Freedom of information and open government in Canada. In: A. McDonald and G. Terrill, Editors, open government, Macmillan, London (1998), pp. 143–167.
Glover & Holsen, 2008 M. Glover and S. Holsen, Downward slope? FOI and access to government information. In: R. Hazell, Editor, Constitutional futures revisited, Palgrave, Basingstoke, UK (2008), pp. 178–196.
Guardian, 2009, February 24 Guardian, The act that Jack wrecked (2009, February 24) Retrieved from
Hazell, 1998 R. Hazell, Briefing from the Constitution Unit 19: commentary on the Freedom of Information white paper, The Constitution Unit, London (1998).
Hazell, 1989 R. Hazell, Freedom of Information in Australia, Canada, and New Zealand, Public Administration 67 (1) (1989), pp. 189–210. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (1)
Hazell, 1987 R. Hazell, Freedom of Information in Australia, Canada, and New Zealand, Cabinet Office, London (1987).
Holsen & Pasquier, 2009 S. Holsen and M. Pasquier, The Swiss Federal Law on Transparency: Much ado about nothing?, Paper given to 'The Future of Governance in Europe and the U.S.', George Washington University, Washington, D.C., June 11-13 (2009).
Horsley, 2008, February 27-29 J. Horsley, Working group five: Indicators and measurement of ATI, Paper presented at the International Conference on the Right to Public Information, Atlanta (2008, February 27-29).
Irish Information Commissioner, 2008 Irish Information Commissioner, Freedom of information: The first decade, OIC, Dublin (2008).
Irish Information Commissioner, 2001 Irish Information Commissioner, Annual Report, OIC, Dublin (2001).
Irish Information Commissioner, 2000 Irish Information Commissioner, Freedom of information: Annual report, OIC, Dublin (2000).
Irish Information Commissioner, 1999 Irish Information Commissioner, Case 99168—Mr. Richard Oakley, The Sunday Tribune newspaper and the Office of the Houses of the Oireachtas (1999).
Islam, 2006 R. Islam, Does more transparency go along with better governance?, Economics and Politics 18 (2) (2006), pp. 121–167. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (9)
James, 2006 S. James, The potential benefits of freedom of information. In: R.A. Chapman and M. Hunt, Editors, Open government in a theoretical and practical context, Ashgate, Abingdon, UK (2006), pp. 17–33.
Julnes & Holzer, 2001 P. Julnes and M. Holzer, Promoting the utilization of performance measures in public organizations: an empirical study of factors affecting adoption and implementation, Public Administration Review 61 (6) (2001), pp. 693–708.
Kaufmann et al., 1999 D. Kaufmann, A. Kraay and P. Zoido, Governance Matters (World Bank Policy Research Working Paper No. 2196), World Bank, Washington, D.C (1999).
Lidberg, 2002 Lidberg, J. (2002). ‘For Your Information’The Impact of Freedom of Information Legislation on Journalism Practice and Content in Western Australia and Sweden. Unpublished master's thesis, Murdoch University, Perth, Australia.
McDonagh, 2006, June 8 M. McDonagh, FOI: the Irish experience. Paper presented at the Conference on Access and Privacy at the University of Alberta, Edmunton, Canada (2006, June 8).
McMillan, 2002, March 4 J. McMillan, Twenty years of open government-what have we learnt? Inaugural Professorial address to the Australian National University (2002, March 4).
Mendel, 2008 T. Mendel, Freedom of information: a comparative legal survey (2nd ed.), UNESCO, Paris (2008).
Ministry of Finance [Ireland], 1999 Ministry of Finance [Ireland], (6) of the Freedom of Information Act [p. 52823] (1999) Retrieved June 5, 2009, from
Ministry of Justice [MOJ], 2009 Ministry of Justice [MOJ], Freedom of Information Act 2000: designation of additional public authorities: response to consultation (2009) Retrieved May 23, 2010, from
MOJ, 2008a MOJ, Exemptions Guidance, About Exemptions. London, Author (2008).
MOJ, 2008b MOJ, Freedom of information: statistics on implementation in central government 2007, Author, London (2008).
MOJ, 1985 MOJ (Canada). Access to Information Act (R.S., 1985, c. A-1). Retrieved May 23, 2010, from
Moynihan & Pandey, 2004 D.P. Moynihan and S.K. Pandey, Testing how management matters in an era of government by performance management, Journal of Public Administration Research and Theory 15 (3) (2004), pp. 421–439.
Missen, 1984, July 17 A. Missen, Freedom of information—the Australian experience, Speech to the Campaign for Freedom of Information at the House of Commons, Westminster, UK (1984, July 17).
Open Society Justice Initiative, 2006 Open Society Justice Initiative, Transparency and silence: a survey of access to information laws and practices in fourteen countries, Open Society Justice Initiative, New York (2006).
Piotrowski, 2007 S. Piotrowski, Transparency in the path of administrative reform, State University of New York Press, New York (2007).
Public Administration Select Committee [PASC], 1999 Public Administration Select Committee [PASC], HC 570-I, HMSO, London (1999).
RTI Assessment and Analysis Group (RaaG) and National Campaign for People's Rightto Information(NCPRI), 2009 RTI Assessment and Analysis Group (RaaG) and National Campaign for People's Right to Information(NCPRI), Safeguarding the right to Information: Report of the People's RTI Assessment 2008 (Executive Summary) (2009) Retrieved May 25, 2010, from
Relly & Sabharwal, 2009 J.E. Relly and M. Sabharwal, Perceptions of transparency of government policy making: a cross-national study, Government Information Quarterly 26 (2) (2009), pp. 148–157. Article | PDF (296 K) | View Record in Scopus | Cited By in Scopus (9)
Roberts, 2006 A. Roberts, Dashed expectations: governmental adaptation to transparency rules. In: C. Hood and D. Heald, Editors, Transparency: the key to better governance?, Oxford University Press, Oxford (2006), pp. 107–125.
Snell, 2001 R. Snell, FOI: the experience of Australian states: an epiphany?, Federal Law Review 29 (2001), pp. 343–358.
Straw, 2009 Jack Straw, Exercise of the executive override under section 53 of the Freedom of Information Act 2000: Statement of reasons (2009) Retrieved 25 May 2010 from
Sunday Tribune, 2009, October 11 Sunday Tribune, How O'Donoghue fell on his sword (2009, October 11) Retrieved from
Terrill, 1998 G. Terrill, The Rise and Decline of FOI in Australia. In: A. McDonald and G. Terrill, Editors, Open Government, Macmillan, London (1998), pp. 89–117.
The Times, 2009, February 24 The Times, Iraq Cabinet Minutes: ‘Jack Straw should not be his own judge’ (2009, February 24) Retrieved February 24, 2009 from
UK Information Commissioner, 2008a UK Information Commissioner, Freedom of Information and Privacy—the Regulatory Role of the Information Commissioner. Speech at the Centre for Regulated Industries, 9/2/2008 (2008).
UK Information Commissioner, 2008b UK Information Commissioner, FOI Three Years On, ICO, Wilmslow, UK (2008).
UK Information Tribunal, 2009 UK Information Tribunal. (2009). Information Tribunal Appeal Number: EA/2008/0024 and EA/2008/0029.
Vleugels, 2009, September R. Vleugels, Overview of all 90 FOIA countries and territories. Fringe Special (2009, September).
White, 2007 N. White, Free and frank: making the official information act work better, IPS, Wellington (2007).
Winetrobe, 2008, July 26-27 B.K. Winetrobe, Precedent vs. Principle: The convergence of the UK and Scottish Parliaments. Conference of Parliamentary Scholars, Wroxton College, Oxfordshire, UK (2008, July 26–27).

Post a Comment