This is independent of use utf8;
On my Linux system, I don't have this problem, on Windows I could solve it by switching to Code page 65000 (!)

see https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/chcp and/or CPAN module Win32::Codepage

(!) - which is the code page for UTF-7…

Update: I hadn't taken the -C flag into account. With -CS, I also get wrong encoding on Linux. So haj's answer is it.

Update 2:with chcp 65001, it also works with -CS on Windows