Is okay to use one instrumental variable for 2SLS multiple regression?