Control and Causation
Suppose I am driving down a long road. I tell my passengers that I intend to keep the car in the center of the lane. Suddenly, there is a curve in the road. I turn the steering wheel and the car remains in the center of the lane. One of my passengers says “If it hadn’t been for that curve, you would have driven into a ditch on the side of the road”. To quote Group Captain Lionel Mandrake, “Well, if you’ll excuse me saying so, sir, that would be, to my way of thinking, rather… well, rather an odd way of looking at it.” It is technically true that given the steering wheel’s movements, without the curve the car would have driven off the road. But that’s ignoring the intentionality of the driver.
If one variable is being held steady by a purposeful actor, there will be no correlation with any other variable. You can’t get a correlation out of a straight line. Furthermore, if a variable was once following a random walk and then becomes a straight line, that seems to me evidence for purposeful control.
Update: “A bet is a tax on bullshit.” – Alex Tabarrok
I am willing to bet anyone $100 that NGDP will continue to follow this trend for the next 2 years. Email me if you think NGDP will deviate from the trend above and we’ll work out the details, which I am willing to post on my blog.