The way that Jira handles sub-tasks is inconsistent with standard project management methodology. A sub-task, by definition, is a fractional component of a parent task. In other words, a parent task can be broken down into multiple smaller steps, each of which is a sub-task. Assuming there is an complete listing of sub-tasks, once all of the sub-tasks are completed, the parent task is also complete. So that if a parent task is estimated to take 10 hours to complete, it can be broken down into, for instance, five two-hour sub-tasks. The time required to complete the sub-tasks does not add to the total time of the parent task; rather, the time for the parent task is already inclusive of the sub-task time.
Jira, however, treats sub-tasks as "related tasks," which is a completely different concept. When a sub-task is created and a time estimate entered, that time estimate adds to the total time of the parent task, which then results in inflated project time estimates and an inability to effectively track and report on the remaining work because time is getting double counted – once in the sub-task and once in the parent task. The only workaround for this situation is to enter a zero time estimate for the sub-task and to log work against the sub-task in the parent task, which does solve the problem but removes the ability to track estimate accuracy against the sub-tasks within Jira.
To resolve this situation, Jira would be improved by not adding time estimate for a sub-task to the total estimate for a parent task unless the total estimated time for all of the sub-tasks for a given parent task exceeded the original estimate for the parent task. Logging time against a sub-task should then reduce the total remaining time for the sub-task as well as the parent task.