[imtringued]

The author talks about how accounting identifies are weird gotchas but the problem here isn't that there is a weird gotcha, the problem with MV = PT is that pretty much all variables are unknown except the general price level. Nobody knows what the real quantity of money is, nobody knows what the velocity is since it would require marking individual dollars and counting how many times they change hands.

[notahacker]

The real quantity of money is pretty well known (notwithstanding disputes over which monetary aggregate is the "correct" measure of inflation, which made monetarist monetary aggregate targeting impractical in practice). PT is basically GDP, and obviously we also track the P component so transaction volumes can be inferred

The problem with the monetarist version of MV = PT isn't that we can't measure the variables, it's that we have measured the variables and that makes it clear that the monetarist assumption that the residual V is fairly stable in the long run and with respect to monetary policy change is clearly incorrect.

[js8]

Forgetting the money velocity is very apparent for example in discussions about basic income, where people who claim it will cause inflation forget that BI redistribution is a big change of V.