"If they enter any data, those data also need to be taken into account in this report. " Pls read on PG's MVCC architecture. In SQLServer, unless you enabled its bad implementation of Snapshot isolation, you can't achieve the same. So it makes sense there. In PG it is easy to ensure that your report gets a point in time consistent view of the data. Trying to mimic one database in another is not a smart way.
