The use of a hash in this application is positively dangerous unless the elements of the  @change array (which become the keys of the hash) can be guaranteed to be unique. E.g.:

>perl -wMstrict -le
"my @change         = (1.15, -0.1, 5.4, 1.03, 1.15, -0.241);
 my @numberofshares = (9999,  400, 200,  300,  500,    240);
 my %hash;
 @hash{@change} = @numberofshares;
 printf qq{%d keys in hash (oops...) \n}, scalar keys %hash;
 my $sum = 0;
 while (my($k, $v) = each %hash) {
   $sum += $v * $k;
   }
 print qq{sum == $sum (?)};
"
5 keys in hash (oops...)
sum == 1866.16 (?)
[download]

Update: The algorithm can be expressed more concisely with some list-processing functions:

>perl -wMstrict -le
"use List::Util qw(sum);
 use List::MoreUtils qw(pairwise);
 use vars qw($a $b);
 my @change         = (1.15, -0.1, 5.4, 1.03, -0.241);
 my @numberofshares = ( 100,  400, 200,  300,    240);
 my $sum = sum pairwise { $a * $b } @change, @numberofshares;
 print qq{sum == $sum};
"
sum == 1406.16
[download]

I think it could be expressed yet more concisely in Perl 6, but my 6-fu is not that great. Would it be something like:
    my $sum +=<<*<< @change, @numberofshares;