That’s a comment by Steven Kopits.
Figure 1: M2 in billions, seasonally adjusted, on log scale (blue). Source: Federal Reserve via FRED.
- Reject unit root null (intercept) at 10% using ADF: No,
- Reject unit root null (intercept, trend) at 10% using ADF: No
- Reject unit root null (intercept) at 10% using DF-Elliott-Rothenberg-Stock: No
- Reject unit root null (intercept, trend) at 10% using DF-Elliot-Rothenberg-Stock: No
- Reject mean stationary null at 1% using KPSS: Yes
- Reject trend stationary null at 1% using KPSS: Yes
If there was a debate of this sort, I missed the conclusion that Mr. Kopits refers to (as well as the general consensus of the economics profession).
Maybe he meant velocity (M2):
Figure 2: Velocity for M2. Source: Federal Reserve, BEA via FRED, and author’s calculations.
- Reject unit root null (intercept) at 10% using ADF: No,
- Reject unit root null (intercept, trend) at 10% using ADF: No
- Reject unit root null (intercept) at 10% using DF-Elliott-Rothenberg-Stock: No
- Reject unit root null (intercept, trend) at 10% using DF-Elliot-Rothenberg-Stock: No
- Reject mean stationary null at 1% using KPSS: Yes
- Reject trend stationary null at 1% using KPSS: Yes
