Orin Kerr has a post up at Volokh titled "What if Lower Court Judges Weren't Bound by Supreme Court Precedent? He makes the sensible point that reasonable people are sure to disagree on interpretative issues, and that if lower court judges were left free to decide constitutional cases according to their ideas of what the true Constitution says, that there would be more uncertainty in the law. I agree, of course, that uncertainty in the law is very bad. However, getting the substantive law wrong is an as or more important problem. All of the reasons why Orin points to as to why lower court judges are likely to disagree with each other also apply to Supreme Court judges. The choice is between lower courts' making judgment calls about what a confusing constitutional provision and some of them getting it wrong, meaning that some parts of the country live under the wrong rule, or one high court wrestling with a difficult constitutional provision and possibly imposing the wrong rule on the entire country. If some parts of the country live under the wrong interpretation and others under the right one, it's more likely in time that the parts of the country living under the wrong rule will see the superior reasoning of the better interpretation. Whereas if the entire country is living under a bad rule that only the Supreme Court can overturn, it will take much longer for the right rule to win out.
I should note also that I don't think there's much reason to think that the Supreme Court is more likely to get difficult interpretative questions right than the lower courts. The most talented jurists of a particular generation often don't make it onto the Court. Presidents don't want to take a chance on radicals or on gifted lower court judges who don't fit neatly into a particular ideological camp (see, e.g., Richard Posner). Or they want to appeal to particular demographic groups (see, e.g., the controversy over whether Sonia Sotomayor was really the best person Obama could have picked for that slot.)
Leaving that aside, wisdom of crowd effects might also be relevant here. A group of several hundred lower court judges trying to come to the right conclusion about difficult interpretative questions might be more likely to come to the right conclusion than a group of just nine Supreme Court judges.
Will robots really boost productivity all that much?
30 minutes ago