Backup job sizes always include any additional folder and file meta-data - for example, the size of the ACLs on all leaf objects (folders and files) plus any fully supported extra file-system extensions and characteristics (e.g. NTFS can "note" MP3 bit rates), so related meta-data will always be included in the total length of the "byte stream" which forms a backup. AFAIK, the size of this restored meta-data is not reported by a restore job. So, what this means in practice is that... in most use cases the size of backup will not match the size of a restore job.
If your goal is to prove restore capability, then your test should be... backup folder structure "A", and restore to folder structure "B", and then use your operating systems' differencing / reporting / listing / scripting tools to compare "A" to "B".
