Rationality Report Cards: Assessing the Economic Rationality of Large Language Models