The *what* and *how* aspects are discussed during sprint planning sessions. Our problem is that while estimating the PBIs after they are accepted by the dev. team, it follows the practise of estimating them so that the development time is almost double i.e. a task that can be completed in 8 working hours is estimated to 13 or 15 hours. As a result product increment suffers. The team argues it prefers to err on the side of caution i.e. it does not want to overcommit so that the sprint remains incomplete at the end of the iteration. However, this is extending the release data and stakeholders are not happy about this. What should be done?
1) Should the PO estimate the PBIs?
2) Should the team be "forced" to estimate more realistically?