The selection of a file for a tax audit is based on the level of risk of tax revenue loss. Therefore, files with a higher level of risk are targeted. Selections are made using risk models based on several factors, including:

• the sector of economic activity;
• the type of business;
• the types of returns and rebate applications filed;
• the types of income and expenses;
• the amounts at stake;
• the business's or the individual's tax profile; and 
• the results of previous audits.

When selecting files using risk models, we establish links between the data available to us. We also use information from other sources, including:

• information obtained while processing other files;
• information shared with other public bodies; and 
• information gathered from reports of failure to meet fiscal obligations.

At the end of the selection process, the auditor's superior assigns files. The auditor does not choose the files he or she will be auditing.

A follow-up using several criteria enabling us to validate the efficiency of the selection process is then done for every selected file. The criteria themselves are regularly examined in order to confirm their relevance and their level of effectiveness.