From the DBI documentation:

When all the data has been fetched from a SELECT statement, the driver
+ should automatically call finish for you. So you should not normally
+ need to call it explicitly except when you know that you've not fetc
+hed all the data from a statement handle. The most common example is 
+when you only want to fetch one row, but in that case the selectrow_*
+ methods may be better anyway. Adding calls to finish after each fetc
+h loop is a common mistake, don't do it, it can mask genuine problems
+ like uncaught fetch errors.
[download]

And also says:

Calling finish resets the /Active attribute for the statement. It may 
+also make some statement handle attributes (such as NAME and TYPE) un
+available if they have not already been accessed (and thus cached). 

The finish method does not affect the transaction status of the databa
+se connection. It has nothing to do with transactions. It's mostly an
+ internal "housekeeping" method that is rarely needed. See also /disc
+onnect and the /Active attribute.
[download]

Update: The DBI docs (same link) also says for the execute method:

If execute() is called on a statement handle that's still active ($sth
+->{Active} is true) then it should effectively call finish() to tidy 
+up the previous execution results before starting this new execution.
[download]