The general method involved may be very simply stated. In cases where the equilibrium values of our variables can be regarded as the solutions of an … - Paul Samuelson

Well, I will say this. And this is the main thing to remember. Macroeconomics -- even with all of our computers and with all of our information -- is not an exact science and is incapable of being an exact science. It can be better or it can be worse, but there isn't guaranteed predictability in these matters.

To the student of economic history the preponderant truth is that technical change has since 1750 tended to raise market clearing real wage rates. This property of the Age After Newton is hard to understand and explicate if you believe that sterile congealed-dead-labor is embodied in machines almost infinitely substitutable for live labor; equally confusing to you will be the truth that inventions which are labor saving may at the same time be wage raising! The doctrines of equated rates of surplus value moved Marxians backward from square one in the understanding of the laws of motion of the capitalistic system or the system of the Mixed Economy.

I reproach myself for a gross error. But I would reproach myself more if I had persisted in an error after observations revealed it clearly to be that. I made a deal of money in the late 1940s on the bull side, ignoring Satchel Paige’s advice to Lot’s wife, “Never look back.” Rather I would advocate Samuelson’s Law: “Always look back. You may learn something from your residuals. Usually one’s forecasts are not so good as one remembers them; the difference may be instructive.” The dictum “If you must forecast, forecast often,” is neither a joke nor a confession of impotence. It is a recognition of the primacy of brute fact over pretty theory. That part of the future that cannot be related to the present’s past is precisely what science cannot hope to capture. Fortunately, there is plenty of work for science to do, plenty of scientific tasks not yet done.