Latest updates - 15 March 2024
15 March 2024 - this guidance was updated to include the following points:
- Which exemptions are prejudice-based?
Section 38 – Health and Safety has been removed from the list to reflect the fact the terms ‘endanger’ and ‘prejudice’ are different. Please see our section 38 guidance for more details.
- When should we consider the prejudice?
The time at which the prejudice test should be considered and the fact that this is the same as the time at which the public test should be considered has been clarified.
- What should we do if we want to neither confirm nor deny we hold the information?
The example used how the prejudice test may apply in these situations has been updated
About this detailed guidance
This guidance discusses the prejudice test under the Freedom of Information Act (FOIA) in detail and is written for public authorities to use. Read it if you have questions not answered in the Guide to FOIA, or if you need guidance to help you consider the prejudice test in practice.
In detail
- Types of exemptions
- Which exemption are the prejudice-based exemptions?
- What does ‘prejudice’ mean in FOIA?
- What is the prejudice test?
- When should we consider prejudice?
- What should we do if we want to neither confirm nor deny we hold information?
- What should we do if the requested information is misleading?
- The public interest test
Types of exemptions
The exemptions in Part II of FOIA are “absolute” or “qualified”. If an absolute exemption applies, you do not have to release the information. If the exemption is qualified, you must weigh the public interest in maintaining the exemption against the public interest in disclosure.
You can further divide the qualified exemptions into ‘class-based’ and ‘prejudice-based’ exemptions. This guidance discusses the prejudice-based exemptions.
Which exemptions are prejudice-based?
The prejudice-based exemptions are:
- Section 26 – Defence
- Section 27(1) – International relations
- Section 28 – Relations within the UK
- Section 29 – The economy
- Section 31 – Law enforcement
- Section 33 – Audit functions
- Section 36 – Public affairs
- Section 43(2) – Commercial interests
We categorise these exemptions as prejudice-based because of their wording. Most of them use the word ‘prejudice’.
Example
Section 43(2):
“Information is exempt information if its disclosure under this Act would or would be likely to prejudice the commercial interests of any person (including the public authority holding it).”
What does ‘prejudice’ mean in FOIA?
In legal terms, the word ‘prejudice’ is commonly understood to mean harm. To say that disclosure would or would be likely to prejudice the interests specified in the exemption implies that it would (or would be likely to) harm those interests.
Section 36(2)(b) provides an exemption where, in the opinion of the qualified person, disclosing the information would or would be likely to “inhibit” the provision of advice or the exchange of views.
The term “inhibit” does not have exactly the same meaning as “prejudice”. Nevertheless, in our view, the approach to adopt when deciding whether there is a likelihood that the provision of advice or the exchange of views will be inhibited, is the same as that when applying the prejudice test.
However, in section 38 - health and safety, the word “endanger” is used rather than “prejudice” and we consider the test for engaging this exemption to be different. For more detail, please see our guidance on section 38.
What is the prejudice test?
We have based our approach to the prejudice test on the Information Tribunal in Christopher Martin Hogan and Oxford City Council v the Information Commissioner (EA/2005/0026 and 0030), (EA/2005/0030, 17 October 2006) (‘Hogan’). This involves the following three steps:
1. Identify the “applicable interests” within the relevant exemption.
2. Identify the “nature of the prejudice”. This means you need to show that:
o the prejudice claimed is “real, actual or of substance”; and
o there is a “causal link” between disclosing the information and the prejudice claimed.
3. Decide on the “likelihood of the occurrence of prejudice”.
Step 1: identify the applicable interests
You need to show that the prejudice you envision affects the particular interest that the exemption is designed to protect. Arguments about prejudice to any other interests will not engage the exemption. So, for example, if you are claiming the exemption under section 29(1)(a) (economic interests of the United Kingdom), any arguments about damage to the UK’s interests abroad that are not clearly economic are not relevant. Those arguments are more likely to engage section 27 (international relations).
Where the exemption has subsections about different interests, the prejudice will relate to the specific subsection(s) that you want to engage. For example, section 31 distinguishes between a variety of law enforcement interests, including the prevention and detection of crime under subsection (a) and the operation of immigration controls under subsection (e).
Step 2: identify the nature of the prejudice
As the Hogan Tribunal explained, this step involves two parts.
First, the prejudice that you have envisaged should be real, actual or of substance.
This means that disclosing the information will at least be capable of harming the interest in some way, ie have a damaging or detrimental effect on it. If disclosure would have either trivial or insignificant consequences, then there is no prejudice. This does not mean, though, that the prejudice must be particularly severe or unavoidable. There may be a situation where disclosure could cause harm, for example to commercial interests. However, you can mitigate the effect of the disclosure, perhaps by publishing clarification to provide context. Where you can mitigate the severity of the prejudice, the exemption might not be engaged. Alternatively, we could still accept that the exemption is engaged but then consider the effect of these mitigating actions as a factor in the public interest test.
Second, there should be what the Hogan Tribunal called a “causal link” between the disclosure and the prejudice claimed. You need to show how disclosing the specific information requested would or would be likely to lead to the prejudice described in the relevant exemption.
Example
The Information Tribunal case of Pauline Reith v Information Commissioner and London Borough of Hammersmith and Fulham (EA/2006/0058) concerned a request for the council’s policy on towing illegally parked vehicles. The council confirmed that they targeted towing operations on certain types of illegal parking. In effect, they therefore partially disclosed the policy but withheld their more detailed criteria under section 31 (law enforcement). The relevant subsection was section 31(1)(g): prejudice to the exercise of the functions specified in 31(2)(c). This is to do with “regulatory action in pursuance of any enactment”. The Tribunal found (in paragraphs 30-40) that the council had not demonstrated that disclosing this information would be likely to cause any prejudice to their parking enforcement functions.
The council believed, given their experience, that disclosing their detailed criteria for towing vehicles would encourage people to park illegally. However, they had no actual evidence to support this.
The ICO had argued that people who park illegally would know, based on the detailed criteria, where they could park and only receive a fine, rather than being towed away. The Tribunal however found no evidence that there were people who would take this risk, given that they would receive a substantial fine in any case. If there were, they could already make an ‘educated guess’ as to their chances of being towed from the partial information released by the council. Furthermore, publishing the full criteria might actually reduce illegal parking if people realised that they are more likely to be towed than they had thought.
The Tribunal therefore did not accept that there was a causal link between the disclosure and the prejudice envisaged and so the exemption was not engaged.
Although there should be a causal link, the prejudice test concerns something that may happen in the future, if the information were disclosed. Therefore, it is not usually possible to provide concrete proof that the prejudice would or would be likely to result. As the above example shows, there should be more than just an assertion or belief that disclosure would lead to prejudice. There should be a logical connection between disclosing the information and the envisioned prejudice to engage the exemption.
Example
In Colin P England v London Borough of Bexley and the Information Commissioner (EA/2006/0060 and 0066) the Information Tribunal said that providing actual evidence of the link between disclosing a list of empty properties and prejudice to the prevention of crime “… is a speculative task, and as all parties have accepted there is no evidence of exactly what would happen on disclosure, it is necessary to extrapolate from the evidence available to come to the conclusion about what is likely.”
Establishing the causal link means that the prejudice claimed is at least possible, ie there are circumstances in which it could arise. The next step in engaging the exemption is to consider how likely the prejudice is to occur.
Step 3: decide how likely the prejudice is
The prejudice-based exemptions use the phrase “would or would be likely to” prejudice (or inhibit or endanger). If you cannot show that the prejudice would or would be likely to occur, then the exemption is not engaged.
The causal link discussed in step 2 shows the circumstances, or the chain of events, that could lead to prejudice. It may be possible to show that prejudice would occur even if those circumstances would only occur once or affect only one person or situation. However, the more frequently those circumstances arise, the more likely the prejudice is to occur. So, while the chances of prejudice occurring in any one case may be low, if the number of cases in which it might arise is high then it may be possible to say that prejudice would or would be likely to arise.
In establishing whether prejudice would or would be likely to occur, you should therefore consider:
- the range of circumstances in which prejudice could occur (for example, whether it would affect certain types of people or situations);
- how frequently the opportunity for the prejudice arises (ie how likely it is for these circumstances to arise); and
- how certain it is that the prejudice results in those circumstances.
The terms ‘would’ and ‘would be likely’ have separate and distinct meanings in this context.
Would prejudice
The Tribunal in Hogan said at paragraph 33:
“there are two possible limbs on which a prejudice-based exemption might be engaged. Firstly, the occurrence of prejudice to the specified interest is more probable than not, and secondly there is a real and significant risk of prejudice, even if it cannot be said that the occurrence of prejudice is more probable than not.”
The first limb is about ‘would’ and the second is about ‘would be likely’. ‘Would’ therefore means ‘more probable than not’. There should be more than a 50% chance of the disclosure causing the prejudice, even though it is not absolutely certain that it would do so.
If you claim that prejudice would occur, you need to establish that either:
- the chain of events is so convincing that prejudice is clearly more probable than not to arise. This could be the case even if prejudice would occur on only one occasion or affect only one person or situation; or
- given the potential for prejudice to arise in certain circumstances, and the frequency with which such circumstances arise (ie the number of people, cases or situations in which the prejudice would occur) the likelihood of prejudice is more probable than not.
Would be likely to prejudice
‘Would be likely’ refers to a lower level of probability than ‘would’, but one which is still significant. This interpretation is based on the judgment of Mr Justice Munby in R (on the application of Lord) v Secretary of State for the Home Office [2003] EWHC 2073 (Admin) (a Data Protection Act case) who said:
“Likely connotes a degree of probability that there is a very significant and weighty chance of prejudice to the identified public interests. The degree of risk must be such that there ‘may very well’ be prejudice to those interests, even if the risk falls short of being more probable than not.” (paragraph 100)
The Information Tribunal relied on this interpretation in John Connor Press Associates v Information Commissioner (EA/2005/0005):
“We interpret the expression “likely to prejudice” as meaning that the chance of prejudice being suffered should be more than a hypothetical or remote possibility; there must have been a real and significant risk.”
On the basis of these judgments, ‘would be likely’ means that there must be more than a hypothetical or remote possibility of prejudice occurring. There must be a real and significant risk of prejudice, even though the probability of prejudice occurring is less than 50%.
If you claim that prejudice would be likely to occur you need to establish that:
- there is a plausible causal link between disclosing the information in question and the argued prejudice;
- there is a real possibility that the circumstances giving rise to prejudice would occur, ie the causal link should not be purely hypothetical; and
- the opportunity for prejudice to arise is not so limited that the chance of prejudice is in fact remote.
Stating the level of likelihood
If you withhold information under a prejudice-based exemption, you will make a choice between ‘would’ or ‘would be likely to’ and state this in your refusal notice.
If the ICO is considering a complaint where you have not specified the likelihood of prejudice, we will apply the test for ‘would be likely’ unless it is clear that your position is ‘would’ prejudice. This is in line with the approach of the Information Tribunal in Ian Edward McIntyre v the Information Commissioner and the Ministry of Defence (EA/2007/0068):
“We consider that … in the absence of designation as to level of prejudice that the lower threshold of prejudice applies, unless there is other clear evidence that it should be at the higher level.”
The “clear evidence” would include the language you use. For example, we may take references to the consequences of disclosure, rather than the possible consequences of disclosure, as evidence that you meant ‘would’ rather than ‘would be likely to’.
This is our approach in exceptional circumstances. Situations in which you do not specify the level of prejudice should not arise. You should always be able to state whether you mean ‘would’ or ‘would be likely to’.
Cases may arise in which you say that prejudice would occur, but the ICO does not accept that you have demonstrated this. Instead, the ICO considers that the exemption is only engaged on the basis that the prejudice would be likely to occur. In such cases we will proceed on the basis of ‘would be likely’. You should note that this approach does not apply if you are claiming section 36(2). We explain this in our guidance on Prejudice to the effective conduct of public affairs.
Establishing the appropriate level of likelihood is also important because it has an effect on the balance of the public interest test. For an explanation of this, see our separate guidance on the public interest test.
When should we consider prejudice?
In carrying out the prejudice test you should consider the circumstances that exist within the time you respond to the request in accordance with the statutory timeframes for compliance. For most public authorities, that is within the 20 working day limit.
This means that you can take account of any change in circumstances between the date you receive the request and the date you respond to it, providing you respond within the 20 working days.
The time for considering the prejudice test is the same as that for the public interest test. The time for considering the public interest test was established by the Upper Tribunal in Montague v Information Commissioner and Department for International Trade [2022] UKUT 104 (AAC). That case clarified that the correct time was the time at which the public authority responds to the request, or the statutory time for compliance, normally 20 working days, whichever is earlier.
Following the Montague case, the First-tier Tribunal then confirmed, in Centre for Animals and Social Justice v Information Commissioner and Department for Environment, Food and Rural Affairs EA/2022/0317 (15 August 2023), that the same approach should be applied to assessing the prejudice test as the Upper Tribunal had taken to the public interest test in Montague.
If you are a public authority covered by one of the Freedom of Information (Time for Compliance with Request) Regulations, you should apply the prejudice test by considering the circumstances that exist within the statutory time for compliance established by the relevant regulations. For example, if you are a school and due to school holidays, you are not required to respond to the request until the sixtieth working day following its receipt, the time for considering the prejudice test will be the time at which you respond, or the 60 working day limit, whichever is earlier. See Time for compliance under the Freedom of Information Act (section 10) for more details.
What should we do if we want to neither confirm nor deny we hold information?
You normally have a duty under section 1(1)(a) of FOIA to confirm whether you hold the requested information, even if you are going to withhold the information itself. This is the ‘duty to confirm or deny’. However, each of the prejudice-based exemptions contains a subsection to the effect that the duty to confirm or deny does not arise where even to confirm or deny whether you hold the information would (or would be likely to) itself prejudice the interest protected by the exemption.
Example
In ICO decision notice IC-108046-X9B0 the requester had asked the Financial Conduct Authority (FCA) for information about a complaint. The FCA relied on section 31(3) of FOIA to neither confirm nor deny it held any information relevant to the request because doing so would be likely to prejudice its regulatory functions under section 31(2)(b) and 31(2)(c). The FCA argued that if it confirmed or denied it held any relevant information it would, in effect, be confirming whether or not the body referred to in the request had been subject to an investigation. The FCA considered that confirming or denying the existence of such an investigation could tip off other bodies or individuals in similar positions of its regulatory interest in a particular issue. This could lead those parties to take steps designed to frustrate the regulatory process.
The Commissioner accepted that confirming or denying whether FCA held the requested information would be likely to result in the prejudicial effects to the FCA’s purposes, described at sections 31(2)(b) and 31(2)(c) of FOIA. Section 31(3) was therefore engaged.
You should therefore take the decision to neither confirm nor deny whether you hold information in relation to a particular prejudice-based exemption. Furthermore, the decision neither to confirm nor deny is itself subject to the public interest test. For further information, see our guidance on the public interest test and on When to refuse to confirm or deny holding information.
What should we do if the requested information is misleading?
You might argue that the harm the exemption is designed to protect against either exists or is increased because the requested information is misleading or could be misunderstood.
Our view is that it is generally possible to avoid this perceived difficulty by putting the disclosure into context.
Example
In ICO Decision Notice FS50130316 the complainant had requested information from the Food Standards Agency (FSA) about a review into the safe cooking time and temperature for burgers. The FSA argued that section 43 applied; the company had expressed the view that disclosing the information would be misleading and could prejudice its commercial interests. It maintained that the information could be taken out of context, leading to misunderstanding and misrepresentation, which in turn could result in a loss of trust and damage to its reputation.
The FSA acknowledged that the information could have been released with an accompanying explanatory statement setting it in context. However, it took the view that this would not have removed the risk of harm, particularly as selected media bodies would have been likely to ignore such clarification in the interest of providing sensationalist and misleading headlines and reports concerning the company. The Commissioner rejected these arguments, deciding that section 43 was not engaged: “The Commissioner is generally reluctant to accept arguments for withholding information based on the contention that disclosure might result in the information being misunderstood or that certain parts of the media might seek to misrepresent the information in order to provide sensationalised news stories. His view is that it is always possible to offset the potential for this to happen by issuing an accompanying statement placing the information in context.”
However such arguments may be relevant in a small number of cases where strong and persuasive arguments are presented which are specifically tied to the exemption claimed. The Commissioner accepts that in some cases it might not be possible to provide the necessary explanation or context, and consequently not possible to mitigate the argued prejudice effectively. An example of this might be where the only person that could provide the necessary explanation is no longer employed by the public authority.
Public interest test
As noted, the prejudice-based exemptions are also qualified exemptions. Once you have established that the exemption is engaged because the test of prejudice is met, the next stage is to consider the public interest. You must consider whether the public interest in maintaining the exemption, and so in withholding the information, outweighs the public interest in disclosing it. We explain this in more detail in our guidance on the public interest test.
Further Reading