Section 31 is a prejudice-based exemption. This means you can only rely on it where disclosing the information (or confirming or denying that you hold the information) could cause harm. To demonstrate the harm, you must satisfy a prejudice test. You can find general information in our guidance on the prejudice test.
When you are applying section 31, you should answer three questions:
- Which law enforcement interest(s), protected by section 31, could be harmed by the disclosure?
- Is the harm you have identified real, actual or of substance and is there a causal link between disclosure and that harm?
- What is the likelihood of that harm actually occurring: would it occur, or is it only likely to occur?
The more likely the harm, the greater weight it will carry when you consider the public interest. In this context, the term “would prejudice” means that it has to be more probable than not that the harm would occur. “Would be likely to prejudice” is a lower test: even if the risk of harm occurring is less than 50 per cent, it must still be a real and significant risk.
If we receive a complaint about your reliance on section 31, we will expect you to demonstrate that you have met each part of the prejudice test.
The mosaic effect
The prejudice test is not limited to the harm that the requested information could cause on its own. You can take account of any harm likely to arise if someone pieced together the requested information with other information to form a broader picture. This is commonly known as the ‘mosaic effect’ and is explained in more detail in our guidance on information in the public domain.
Complying with one request can make it more difficult to refuse requests for similar information in the future. You are therefore entitled to consider any harm that could be caused by combining the requested information with the information you could subsequently be required to provide, if the current request was complied with.
Example
In decision notice IC-209255-G8H5, the Commissioner considered a request to Transport for London (TfL). The requester asked about the number of penalty charge notices issued and revenue raised through the enforcement of a particular yellow box junction.
TfL successfully argued that, whilst the information in question was not particularly significant on its own, the complainant had also requested similar data for five other locations. Combining all this data would start to show broader traffic enforcement patterns across London – enabling motorists to avoid detection. TfL could therefore engage section 31(1)(b).
The mosaic effect is relevant to all the parts of section 31.