Matias Vernengo has a very interesting post on the long-run and the short-run in economics. As he says, the long-run and short-run are just thought experiments (he calls them “methodological tools”). So, the long-run is an imagined period when all adjustments had taken place with which the economy is pregnant with at any given time, while the short-run is the “imperfect” period of gestation where these adjustments are taking place.
Vernengo then correctly points out that when economists actually designate periods of time — 3 months, 6 years and so on — to these periods they get into an awful mess. I totally agree. But I would go one further: the long-period is not just a thought experiment; in actuality is a metaphysical idea. Joan Robinson, who took the same position I think, used to in her later writings call it “the mysterious long-period”.
Not only does the long-period not exist and will never exist, but it is not even a real concept and any application of it to real historical time will always result in nonsensical statements. This is because historical time is non-ergodic and so as forces of adjustment move through time they cause changes that alter the course of the trajectory of the economy. In conclusion, talking about the long-period is really just a metaphysical thought experiment and the moment it can be seen to have any influence in a discussion of any real-world situation it can be known with confidence that the discussion has gone completely awry.
Fixing the Economists 
The long run can be understood in terms of probability theory as non-chronological. Probability holds in the long run although there may considerable variation in the short run, which is what gamblers’ bet on when gambling against the house, which always comes out ahead “in the long run,” which is at any point in time given the number of gamblers. Someone’s breaking the bank rarely if ever happens. If the house has deep enough pockets, it will easily recover from any individual loss and its take from gambling will always be very close to the odds in its favor.
Say “the long run” in coin flips is 1000 flips. There is no difference in the meaning of “in the long run” by flipping the same coin 1000 times or 1000 coins simultaneously. Historical time is not relevant.
This is clearly not the case over historical time in a non-ergodic system like a modern economy subject to reflexivity and changing historical circumstances that alter the conditions.
This is crucial to the neoliberal myth of return to equilibrium at optimum potential “in the long run,” meaning over an unspecified period historical time, owing to operations of market forces iaw neoclassical assumptions. It’s misunderstanding or misrepresenting the meaning of terms. Moreover the idealized model is non-representational owing to its construction, e.g., assumption of atomism and ergodicity, adopted to make modeling tractable in spite of uncertainty.
Billyblog had a somewhat relevant blog some years ago. Money neutrality – another ideological contrivance by the conservatives, where he quotes a bit from
Fernando Carvalho’s paper on the meaning of long and short run: ‘Alternative Analyses of Short and Long Run in Post Keynesian Economics’ (also reprinted in Alternatives to Economic Orthodoxy: A Reader in Political Economy; I gave a free link there to that book for most of the paper) See also his quotes from Kalecki.
As an aside, I am all for metaphysics. Everything interesting is metaphysics. Nobody can or ever has thought without metaphysics. The problem is that the logic and metaphysics of neoclassical/mainstream economics is impoverished, and really really bad.
Tom:Someone’s breaking the bank rarely if ever happens. But it does happen. It’s a traditional way for casinos to change hands.