Brad Delong and I agree on at least one thing: Donald Luskin is mighty dumb (see Brad here and me here). Now Slate's Mike Kinsley has ventured very near to Luskin territory in this piece.
He gives figures for inflation, unemployment and real per capita income growth from 1959 - 2007 which show better average performance under Democratic presidents than Republican presidents and uses this information to conclude that Democrats are better for the economy than Republicans.
Now people, you know I like gridlock, but I am not a Republican, and I believe that it is quite possible that Democrats are better for the economy than Republicans, but comparing sample averages does not provide any relevant information about who is better.
One reason is the potential of reverse causation. Suppose for example, that when the economy is going bad, voters turn to the Republican party to "fix" the problem, and when the economy is booming voters turn to Democrats to advance a non-economic agenda. If this is true then it's the economy is causing the party of the president, and it's not the party of the president causing the economy.
A second reason is the lack of a counterfactual. What we really need to know is if all circumstances were exactly the same except for the party of the president, would the outcomes differ, and if so, in what way. One way to impose the counterfactual would be to use a multiple regression and control for all the other relevant external factors that affect economic performance.
A third reason is that much responsibility for economic policymaking in the US resides with the Federal Reserve, which is supposedly independent from partisan politics, and even if they are not, presidents do not automatically get to pick a new FOMC when they take office.
Please don't read this as an apology for the Republicans. Read it as a plea for numeracy among our chattering class.