Why is this necessary? Are the queries more performant?

It's necessary because aggregate functions are run for each group separately.
So if there's GROUP BY, COUNT(*) will count a number of grouped rows in each group (i.e. COUNT(*) will be calculated for each returned row).
But we need to return the count of returned rows.

https://www.w3schools.com/sql/trysql.asp?filename=trysql_editor
Try here the following queries:

SELECT Country, COUNT(*) FROM Customers GROUP BY Country;

vs

 SELECT COUNT(*) FROM (SELECT Country FROM Customers GROUP BY Country);

The first query returns COUNT(*) for each country, the second query returns COUNT(*) of countries.
The first query returns 21 rows, the second query return one row with value 21.

Makes sense 👍🏾