If your file is small enough (no more than several hundred meg and guaranteed not to grow beyond that size), slurping and searching the entire file may be useful. Code:

use warnings;
use strict;

my $doc = do { local $/;  <DATA>; };  # slurp entire file
# print "<<<$doc>>> \n";  # FOR DEBUG

my $rx_account = qr{ (?<! \d) \d{12}       (?! \d) }xms;
my $rx_amount  = qr{ (?<! \d) \d+ [.] \d\d (?! \d) }xms;

my ($two_lines, $account, $amount) = $doc =~ m{
    ( ^ ACCOUNT \s+ NUMBER \s+ ($rx_account) .*?
      ^ YOUR \s+ LOAN \s+ PAYMENT .*? ($rx_amount)
    )
    }xms;

print "two lines: <<<$two_lines>>> \n\n";
print "account '$account'  amount '$amount' \n";


__DATA__
ACCOUNT NUMBER    000111111111
YOUR LOAN PAYMENT FOR THE YEAR WILL BE 00.00
   OF WHICH 00.00  WILL BE FOR PRINCIPAL AND INTEREST,         00.00
   WILL GO ESCROW ACCOUNT, AND .00  WILL BE FOR
   DISCRETIONARY ITEMS THAT YOU
   CHOSE TO BE INCLUDED WITH YOUR LOAN PAYMENT.
THE EFFECTIVE DATE OF YOUR NEW SCHEDULED PAYMENT IS 00/00/00.
[download]

Output:

c:\@Work\Perl\monks\crusty_collins>perl match_multiline_1.pl
two lines: <<<ACCOUNT NUMBER    000111111111
YOUR LOAN PAYMENT FOR THE YEAR WILL BE 00.00>>>

account '000111111111'  amount '00.00'
[download]

Updates:

1. See also File::Slurp and friends.
2. What you have shown in the OP is not what I would think of as a "paragraph" match, but rather a multi-line match. It's difficult to see from the given data just what a paragraph would be.