The United States has experienced economic downturns, or recessions, roughly every 5-10 years since they began being tracked in the late 18th century. 1785, 1789, 1796, 1802, 1807, 1812, 1815, 1822, all the way to 1937, 1945, 1949, 1953, 1958, 1960, 1969, 1973, 1980, 1981, 1990, 2000, and most recently 2007.

What I'm confused about is, assuming the economy falls on some balanced distribution around the mean, there should be equally as many economic upturns, booms, or manias if we want to borrow a term from psychology, as there are downturns, recessions, and depressions.

Yet as far as I can tell, this isn't the case. The economy goes from neutral to negative and back, with an occasional foray into positive, showing a clear bias for negative periods over positive ones.

Why is this?

  • 2
    $\begingroup$ Do you think the USA economy today is larger or smaller than it was in 1785? Why do you think "the economy falls on some balanced distribution around the mean"? What is the evidence for " a clear bias for negative periods over positive ones."? $\endgroup$ – 410 gone Dec 20 '19 at 7:26
  • 1
    $\begingroup$ What is your definition of an upturn? $\endgroup$ – Kenny LJ Dec 21 '19 at 3:35

As others pointed your statement about upturns and downturns is not right. An empirical evidence clearly shows that there are short sharp recession followed by long periods of moderate expansion (see Romers Advanced Macroeconomics handbook) and also see the picture below. It is the Fred series for US growth where you can see recessions (downturns) highlighted in grey color and you can clearly see they are very small time periods compared to periods of expansions (upturns), although it plots just US, you would get the same story for most countries around the globe:

enter image description here

Furthermore, normal distribution is not necessary best distribution to describe the rate of growth as it has very thin tales. With normal distribution Great Depression or Great Recession would be 6-sigma events but they are clearly not. Also, the growth rate exhibits dependence so normal distribution could be at the best assumed for the error. However, even if you assume normal distribution as first approximation it would be not correct to assume it has mean 0. The mean of such distribution would center around the natural rate of growth (for an advanced economy about 2%). I made a simulation in R for you of how that would look like with assuming just pure normal distribution with center at 2% and then also as some autoregressive process.

As you see below, the normal distribution is bad description of an growth rate evolution, but you still get the same story as from real life data if you center it at 2% which is about the natural rate of growth for advanced economies, few recessions most expansions.

enter image description here

However, a better description would be some auto-regressive model (Here I use AR$(1), \mu=2, \rho=0.9$). Below you see that this could actually be some rate of some hypothetical country, again you get the same story, few recessions a lot of expansions (in the picture below 75.3% of time periods happened to be expansion periods).

enter image description here

So to sum up:

  1. your description of economy does not match real life data. Its actually the opposite. You get more expansions than recessions.
  2. Why is this happening? Because even with symmetric distribution the growth rate is not necessary centered at 0. It is centered most likely at its long run natural rate - which for modern advanced economies is positive (empirically around 2%).
| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.