Given a minimal input POD document like this

=encoding utf8

=head1 NAME

Test document

the recently released pandoc 3.6.2, run with pandoc -f pod Minimal.pod (on Linux), outputs

<p>=head1 NAME</p>
<p>Test document</p>

instead of the expected

<h1>NAME</h1>
<p>Test document</p>

Placing a normal paragraph after the =encoding yields an expected result, i.e.

=encoding utf8

Dummy paragraph.

=head1 NAME

Test document

outputs

<p>Dummy paragraph.</p>
<h1>NAME</h1>
<p>Test document</p>

However, adding an extra blank line instead, as in

=encoding utf8


=head1 NAME

Test document

leads to a parsing error:

Error at "Minimal.pod" (line 4, column 1):
unexpected '\n'
expecting Pod paragraph

^

which does not seem reasonable.