A nice article in the economist:
1. Statistical Artifact of Goodhart’s law - statistical (causal) relationships collapse once exploited by policy makers
2.Inflation Expectations- Inflation Expectations are reacting more slowly to economic information
3.Philips Curve is non linear. Unemployment needs to fall even more before inflation gets triggered.
What about a 4th possibility, related to a liquidity trap.
Rates are so low and alternative investments are all relatively low yielding. As such, businesses see little yield difference across investment opportunities. Instead, they use risk to rank order investments. As such and in very low rate environments, businesses would rather invest in lower risk investments (like cash / low yielding securities) than higher risk people investments. Also, businesses likely see capex associated with labor automation substitution as lower risk than people investments. This is because of the tax environment is favorable to capex and the complicated legal environment associated with hiring and firing people.
Another good Phillips curve article, that beyond the “typical” reasons for the Philips curve flattening, are other issues like 1) the typical measures of unemployment ( u-3) don’t get at the available employment slack that exists and is not counted. 2) the stickiness if wages in down turns. There is a culture of maintaining wage levels even as demand dampens. Thus making the timing of wage changes out of sync with demand changes that drive inflation.
I think, like many economic theories, the dichotomy logical fallacy is at play, meaning the simple relationship is not so simple. The Philips curve is relevant but is a unique case, not a rule of thumb. It is instructional but should only be considered in context of the market it is being applied.