The Access to Information Act adopted by Parliament in 1982 is intended to give Canadians a qualified right to documents held within federal ministries and agencies. The arguments in favour of the right to informa- tion are compelling, but the task of enforcing that right has proved difficult.
In fact, the ATIA is suffering a crisis of legitimacy with- in the federal government. Many officials and ministers no longer respect the law. They believe it is prone to abuse and imposes disproportionate administrative costs, and that journalists and politicians twist the information they receive to create a distorted view of federal ministries. They do not like the way in which non-governmental organiza- tions, empowered with inside information, are able to dis- rupt the process of planning and implementing policy.
This crisis of legitimacy has deepened since the election of the Chrétien government in 1993. Over the last eight years, the government has found many ways of chopping away at the ATIA. It has transferred many functions””and billions of dollars””to new quasi-governmental organiza- tions that are not subject to the law. It has litigated aggres- sively to exclude ministers and their offices from the law. In the Anti-terrorism Act adopted in November 2001, the gov- ernment also acquired an unprecedented new power to sus- pend access rights in the name of national security.
These are the obvious signs of the government’s ani- mosity toward the ATIA. Access rights can also be defeated in more subtle ways. For example, the explosion of com- plaints about delays that followed budget cuts to ATIA administration in 1994 taxed the capacity of the Office of the Information Commissioner for several years.
In May 2001, Ontario’s Information Commissioner, Ann Cavoukian, described another form of subversion in the administration of that province’s Freedom of Information and Protection of Privacy Act””the practice of giving special atten- tion to politically sensitive requests:
Although we have not been provided with details or copies of any policy documents, we have learned through our work in mediating and adjudicating provin- cial appeals that certain access requests that are deter- mined to be ”œcontentious” are subject to different response and administrative procedures. This ”œcon- tentious issues management” process is managed by Cabinet Office. Our understanding of the process is sketchy, and ministry Freedom of Information and Privacy Co-coordinators are extremely reluctant to provide us with details; however, as we understand it, the process generally operates as follows: if an access request is made by certain individu- als or groups (i.e., media, public interest groups, politicians), and/or the request con- cerns a topic that is high profile, politically sensitive or current, ministry Freedom of Information and Privacy Co-coordinators must follow the contentious issues proce- dures. Once designated into this category, the process requires the immediate notifica- tion of the Minister and Deputy Minister, along with the preparation of issue notes, briefing materials, etc. Cabinet Office is often involved in this process.
The Commissioner observed that there was no provision in law for special treatment of polit- ically sensitive requests. On the contrary, the law presumes that requests will be handled without regard to the identity or motivation of the requester. Nevertheless, internal practices appeared to have an important impact, produc- ing longer response times and less generous deci- sions on disclosure. The Commissioner worried that this had become a ”œsystemic problem” in the administration of Ontario’s freedom of informa- tion law.
There is evidence that federal departments and agencies handle politically sensitive requests in similar fashion. A report on ATIA administration completed for Treasury Board Secretariat in February 2002 observes that in some depart- ments, senior officials are given weekly updates about the content of incoming requests so they can identify areas of sensitivity for close moni- toring. Similarly, some ATIA officers routinely prepare impact statements that alert senior offi- cials and communications staff to the anticipated consequences of disclosing information.
The case management software used by most federal departments, known as ”œATIPflow,” appears to support such practices. It allows ATIA officers to categorize incoming requests by type of requester””including distinct categories for the media and political parties””and by the political sensitivity of the request. The version of ATIPflow used by Human Resources Development Canada allows ATIA officers to tag incoming requests as routine or sensitive, or as requests about which it is necessary to ”œAdvise [the] Minister.”
If it had no impact on the processing of ATIA requests, this kind of tracking might not be objectionable. However, there is evidence that federal institutions do treat sensitive requests dif- ferently, typically by slowing down the process of responding to requests. In May 2000, Federal Information Commissioner John Reid concluded that recurring delays in replying to ATIA requests submitted to the Department of National Defence had been caused by the desire to manage ”œpolitical considerations, including the commu- nication needs of the minister.” Requests from the media seemed most likely to be singled out for special treatment.
The problem did not seem to be confined to National Defence. In the same report, Reid criti- cized the government’s response to controversies that arose in early 2000 over the policies of Human Resources Development Canada. HRDC had scored well on compliance ”œreport cards” issued by the Information Commissioner in May 1999. But its performance slipped badly in 2000. The Information Commissioner attributed the change to the controversy over allegations of mismanagement in HRDC job creation programs that erupted in January 2000.
The imbroglio led to a surge in ATIA requests to HRDC that may have overwhelmed its ATIA staff. But Commissioner Reid detected another reason for pervasive delays in responding to requests:
[G]overnment couldn’t help but let its reflex- ive need to ”œcontrol” the story take prece- dence over the legal rights of access requesters to obtain timely responses. Ministers wanted to be out front of any access request””making a clean breast of any bad news before it hit the street and, when it did, being armed with an action plan. Of course, the only way to accomplish this strategy was to buy time””to slow down or postpone the release of these requested audit reports on grants and contributions programs.
Many journalists share Reid’s suspicions about this sort of governmental obduracy. A recent study of journalists’ views on the ATIA reported that the most common criticism was unjustifiable delays in responding to requests:
Many interviewees also reported that when information was eventually obtained, some- times after an appeal, it was unclear why it had been withheld or delayed in the first place. … Indeed, it seemed to some interviewees that the 30-day limit, and the delays beyond it, might constitute bureaucratic strategies intended to ”œkill” stories by letting them drift into irrelevance and frustrating requesters.
Complaints about the differential treatment given to politically sensitive requests are serious, but the available evidence is not always compelling. Too often it is based on hearsay, anecdotes, or impressions gleaned from the investigation of a small proportion of access requests. Is there any better evidence to support these claims?
In fact, there is. The federal access-to-information system can be criticized on many grounds, but on one dimension it is clearly unparalleled. Federal departments and agencies have deployed better software systems for track- ing information requests than any other jurisdic- tion in the world. One system””the Coordination of Access to Information Requests System (CAIRS)””is a government-wide program that contains data about new requests received by all major federal institutions. A new version of CAIRS was installed in August 2001, at a cost of $166,000. The federal government also spent sev- eral hundred thousand dollars over the last five years installing ATIPflow, which, as mentioned, is used to manage the ATIA workload in major departments.
These software programs capture a large amount of data that can be used to provide a detailed view of the internal operations of the ATIA system. Much of this data can be obtained from federal departments by making a request under the Access to Information Act itself. The fol- lowing analysis draws on data collected with the ATIPflow database maintained by Human Resources Development Canada. It includes data for 2,120 ATIA requests completed by HRDC between January 1, 1999 and December 31, 2001. The analysis does not include data for requests that were transferred, abandoned or handled informally, or whose final disposition was unclear.
One important dimension of institutional performance is the time taken to complete ATIA requests. Analysis of ATIPflow data in Table 1 shows that processing time within HRDC did, in fact, vary according to the profession of the requester or the sensitivity of the request.
Table 1 provides rough evidence of differen- tial treatment. It shows that in every year the average number of days required to process requests was greater for requests from the media or political parties than it was for other requesters””in 2000 and 2001, almost a complete month longer. Of course, these averages could be misleading. For example, it could be that media or party requests were larger, or sought distinct kinds of information that took longer to process. A more careful analysis of processing time is obvi- ously needed.
I performed just such a statistical analysis, using the techniques of multiple regression, and the results are shown in Table 2. My analysis took into account several considerations that could lengthen the time required to process requests. One obvious factor was the sudden spike in work- load after January 2000. A second was the vol- ume of records that relate to the request: many requests produce no relevant records at all, while others cover a large number of records. A third factor was the type of information requested. Some types of information might take longer to process than others. This analysis took into account three important types: commercial infor- mation, which typically requires consultation with the business that provided the information to the government; Cabinet records, which require consultation with the Privy Council Office; and information about internal delibera- tions. The analysis also took into account whether a request was identified by HRDC as coming from the media or a political party. Finally, the analysis took into account whether HRDC had used ATIPflow’s features to tag the request as ”œsensitive”””although because HRDC stopped using this feature of ATIPflow after early 2000, this last step had limited usefulness.
The results reported in Table 2 reveal some of the key determinants of processing time within HRDC between 1999 and 2001. The entry for each attribute represents the extra number of days required for a request with that attribute, compared to the number of days for the default case. Processing time spiked sharply in 2000”” though whether this was a result of increased workload or of the increased recalcitrance sus- pected by the Information Commissioner, it is not possible to say. The volume of records requested also proved influential: if no records were requested, processing time dropped by about 25 days, while it increased by 21 days for large requests. Requests for commercial informa- tion took about eight days longer; for deliberative processes, about 14 days longer; and for Cabinet records, about two months longer.
Stay in the know with veteran reporter Kathryn May. Sign up for routine and out-of-the-ordinary news about the public service with The Functionary, our new newsletter.
Other findings are much more problematic. For example, the analysis still suggests that requests identified as coming from political par- ties or the media were completed more slowly. They took at least three weeks longer to complete than requests from other sources. And requests tagged as sensitive took longer still, with another 14 days added to processing time. Because HRDC abandoned this categorization after 1999, this last result should be treated with caution. On the other hand, it does suggest that judgments about the sensitivity of a request were not based exclu- sively on the identity of the requester. Other con- siderations also came into play.
The ATIA generally requires that institutions reply to requests for information within 30 days. However, they are permitted to extend this deadline if they believe that more time is needed to search for records, engage in consultations, or notify non-governmental organizations affected by a request for information. A ”œdeemed refusal” occurs when the time taken to complete a request exceeds the original 30-day limit or any exten- sion of that limit.
For the last several years, the Office of the Information Commissioner has used deemed- refusal rates as a rough measure of departments’ ability to comply with ATIA requirements. It is not a demanding test, because departments have broad discretion in deciding both whether to take extensions and how long extensions should be. Nevertheless, a ”œdeemed refusal” constitutes a clear violation of ATIA requirements. (On the other hand, departments argue that deemed refusal rates may overstate non-compliance when they may have made a partial release of docu- ments before the deadline.)
As Table 3 shows, there is rough evidence that media and political party requests were more likely to become deemed refusals. Again, howev- er, a more careful analysis is required. This fol- low-up analysis of ATIPflow data examined the factors that influenced the chances a request would become a deemed refusal. All of the vari- ables included in the previous analysis were also included here, and with similar results. The analysis found a clear connection between the profession of the requester, the perceived sensi- tivity of the request and the chances of a request giving rise to a deemed refusal.
Some implications from this analysis are shown in Table 4. The largest impact on the chances of deemed refusal is the change in work- load after 1999. The second largest is whether there are records relevant to the request. As expected, that relationship is positive: if no records are requested the chance of a deemed refusal falls. The next most important influences on the chance of a deemed refusal are whether the request comes from a political party or the media””if it does, the chance of refusal rises sig- nificantly””or has been tagged as sensitive. These factors prove more important than the type of information requested.
Table 5 provides a more concrete illustration of how some of these factors influence the chances of a deemed refusal. The table shows the estimated chances of a deemed refusal for a request for information about internal delibera- tive processes for each of the three years. (Because this table is limited to requests for information about deliberative processes, the numbers are quite distinct from those in Table 3.) The analysis suggests that requests from the media or political parties always ran a higher risk of deemed refusal””about 14 per cent higher in 2000, and 20 per cent higher in 2001.
Difficulties with data collected in the ATIPflow software make it hard to judge whether the same pattern of differential treat- ment emerges with respect to other aspects of ATIA request processing.
For example, one critical aspect of ATIA administration is the estimation by officials of the fee that must be paid for processing larger or more complex requests. There is good evidence that large fee estimates will cause requesters to narrow or abandon their requests. Unfortunately, the apparently erratic nature of recordkeeping on initial fee estimates precludes precise analysis of differential treatment in the exercise of discretion over fee estimates.
There is limited evidence of differential treatment in other areas. For example, the data sug- gest that HRDC was more likely to exercise its right to extend the statutory response time on the grounds of unreasonable interference with its operations if a request was submitted by a politi- cal party or the media. However, there was no evidence of a comparable effect with regard to the use of other time-extension provisions.
Similarly, there is limited evidence of varia- tion in disclosure practices. There is a significant drop in the proportion of reviewed records that are released in response to requests sent by polit- ical parties, even when other factors are taken into account. There was no comparable effect for media requests or sensitive requests. The average number of statutory exemptions invoked in response to requests from parties or the media was also significantly higher than the average for other requests. However, because it does not con- trol for variations in the kinds of information sought by each group of requesters this finding must be treated cautiously. Proper controls of that sort would require more extensive research.
As the statistical investigation has shown, on one dimension””completion time””there is good evidence to suggest that there were two standards for administration of ATIA requests within HRDC between 1999 and 2001. Requests that were identified by HRDC as coming from the media or political parties or that were tagged as sensitive took substantially longer to process, even after other considerations””the type of request, the size of the request, HRDC’s overall workload””are taken into account. In fact, the chances that processing time would exceed statutory deadlines””an unam- biguous violation of ATIA requirements””rose signifi- cantly for media and party requests, as well as for requests tagged as sensitive.
Completion time is a crucial dimension of ATIA administration. As the feder- al government’s 1977 White Paper on access to informa- tion observed: ”œthe essence of the so-called ”˜freedom of information’ idea is not sim- ply access to government documents, but timely access.” Timely access is obviously of critical importance to media and party requesters. University professors might not be unduly harmed by an added delay of three or four weeks. However, this sort of delay could make a world of difference in everyday politics, where agendas shift quickly and attention spans are short. The value of information that is with- held for any appreciable time drops sharply.
This might explain why an unusually large proportion of requests received by HRDC in the first half of 2000 were later abandoned by requesters. In 1999””that is, before HRDC became the subject of controversy””requesters abandoned just five per cent of all requests. The abandonment rate spiked to 17 per cent in the first quarter of 2000, and to 32 per cent in the second quarter.
It is not surprising to find that differential treatment is most obvious in processing time. This is the area in which the enforcement mecha- nisms within the ATIA are weakest. In fact, it would not be much of an exaggeration to say that there is no effective remedy for problems of delay under the ATIA. The Office of the Information Commissioner (OIC), staggering under a burgeon- ing workload, is incapable of responding quickly to complaints about delay. In 1999-2001, the OIC needed more than three months to resolve a delay complaint (Table 6). And even if the OIC finds the complaint justified, there is little it can do to undo the harm done by delay. It can hardly reverse time or the political agenda so that requested information regains its relevance.
This breakdown in enforcement is not attrib- utable to mismanagement within the OIC. On the contrary, it is a flaw inherent in the law itself. The enforcement strategy built into the law, which aims to resolve individual complaints about non- compliance, simply does not work in cases of delay. Recent OIC studies have shown that a large majority of individuals who could make justified complaints about delay choose not to, perhaps because they know there is no effective remedy. For example, the OIC’s own statistics suggested that Health Canada had a relatively good record in compliance: only six per cent of requests filed in 1998 resulted in a complaint to the OIC about delay or abuse of exemptions. But a closer investi- gation by the OIC showed that in almost 40 per cent of cases, individuals would have been justi- fied in making a complaint about delays alone.
It may be just as well that most cases of non- compliance go unreported. If all these individuals did make complaints, the OIC’s caseload would skyrocket and its investigative procedures would break down completely.
There are better ways to deal with problems of delay and differential treatment. One is with publicity. Even under existing law, the OIC could use its influence to require federal institutions to produce more detailed and informative statisti- cal reports on the processing of ATIA requests. The OIC could also audit institutions to ensure that data-collection and reporting procedures are adequate and accurate. These annual reports could include a description of processing trends for different categories of requesters, in the style of the analysis presented earlier. This kind of reporting would impose a small burden on departments. At the same time, publicity would generate pressure for better and more even- handed compliance.
The OIC also needs better tools for dealing with departments with a persistent habit of non- compliance. In 1999, the OIC attempted to use its existing investigative powers””including its power to subpoena senior public servants””in an effort to draw attention to pervasive problems of delay. This strategy did produce some improve- ments in overall performance. However, the use of these investigative powers also alienated some senior officials, who apparently felt that the OIC had treated administrative problems as though they were almost criminal matters. There is there- fore pressure within the government to introduce new statutory checks on the investigative power of the OIC.
The problem is simply one of policy design. The OIC needs instruments that are better suited to the task of eliminating broad patterns of non- compliance. In my view, the best approach would be to give the OIC explicit authority to require that non-compliant departments publish plans for improving performance. Authority to impose sanctions that go beyond that simply may not be necessary. Departments would become publicly accountable for deficient performance and for any failure to take appropriate remedial action.
These are workable, inexpensive remedies to problems of delay and differential treatment. They would help improve public understanding of the ATIA system, reduce conflict between the OIC and senior officials””and bolster Canadians’ right to information.
A longer version of this paper with more detailed discussion of methodology can be downloaded from http://www.campbellinstitute.org. This research was supported by a grant from the Canadian Newspaper Association. The author wishes to thank Professors Vernon Greene and Stuart Bretschneider for their assistance with methodological questions.