Hello Pratikashokraut,
The performance relies on a number of aspects: the complexity of the trigger logic and the number of rows affected, indices.. there are many things that can flip this one one or the other so it’s pretty hard to come up with general statistics

One advantage of the approach in the article, though, is that the delete into is atomic; either BOTH the delete and insert get executed or none. When the tigger fails there is a change that the insert is not rolled back i.e: you insert without deletion. This can be solved by wrapping the trigger in an explicit transaction (with BEGIN, TRANSACTION, COMMIT and ROLLBACK statements)