I just finished Skanda Amarnath’s and Alex William’s piece on the Phillips Curve and maximum unemployment.1 There is a lot to like in the piece, especially the dismissal of the Phillips Curve, also known as the worst idea to come out of American Keynesianism other than hydraulic modelling.
Some Critiques
Before I get into the meat of this post, I want to discuss what I didn’t like about the article.
My main issue with the article really is only about their assertion2 that the failure of the labour utilization rate to rise leads to recessions or anti-recessionary actions. It is a mere causation due to capacity labour utilization being endogenous to the strength of nominal output growth, both on the real output and stabilization policy side.
I also take issue with their stance on interest rates. Interest rates must be looked at through a Wicksellian3 lens—period. Understanding both how the spread between the natural and policy rate and Fisher Effect play into rates is crucial.4
I also find their discussion of the Great Inflation to be more or less false. The price level increases in the 1970s were the result of expansionary policy. It is impossible for capacity constraints to affect input-output linkages in the way they did, whereas the hot potato effect and the Fed’s overshoot5 can certainly explain the crisis (and consistently with theory too).
The Two Drivers of the Labour Market
Their piece is really about the relationship between the labour market and inflation. To summarize, inflation is born of excess demand over supply constraints, while limits to labour market tightening are born of capacity constraints and a lack of policy accommodation.
In my mind, we can divide up the driving forces of labour markets into two types: real and nominal.
Real driving forces in labour markets are ultimately about the need and ability for sectors to take on workers and the desires and constraints worker have imposed while bargaining. Things like a microchip shortage limit the ability of certain sectors to hire, while things like unions and occupational licensing restrict labour supply. It is important to note the relation between labour supply and demand and real wage acceleration. Ceteris paribus, restrictions to labour supply increase real wage acceleration. However, on a grander scale, they decrease real wage acceleration because firms are less survivable. That’s why paying attention to occupational licensing, trade unions restricting the entry of potential workers,6 and capacity constraints that hurt firms’ bottom lines all serve to drive down labour market tightness.
Nominal forces are merely stabilization policy. Both monetary policy driving up income expectations and easing credit conditions and fiscal policy adding new capacity and spending to the economy count here. Labour markets will tighten when the economy approaches equilibrium, as that is when competition is most perfect. Firms with more spending capacity can also allow wages to rise more, and workers with more potential fallbacks drive a harder bargain. The inverse (when policy is contractionary) is true as well.
The Two Types of Inflation
Let us now evaluate the opposite side of the Phillips Curve: inflation. Inflation can be one of two types: transitory and permanent.
Transitory inflation happens when there is a disequilibrium in the economy that manifests as a price increase. It is typically, but not always, a cost-push phenomenon. When the disequilibrium is resolved, the inflation ends.
On the other hand, permanent inflation is embedded into the equilibrium structure. This can only be a demand-pull phenomenon. It is sourced entirely from an overwrought gain in nominal income relative to capacity, or the nominal income gap. It can only be resolved by tightening policy or re-spurring real growth.
Demographic changes matter for inflation insofar as they govern the dynamics of real output capacity.7 Paying attention to the ability for demand to come or offset from the rest of the world is also important (never mind the additional effects on interest rates).
Declining Capital Stock and Frictions
My favorite assertion in the Employ America piece was that targeting conditions from the pre-recession is appropriate:
Demand-side policies should be actively addressing the fallout of painful recessions that inflict obvious cyclical dislocation. It is well within reason to aim for a labor market that proved feasible just 14 months ago.
Historically, though, commentators and policymakers alike have often responded to long, slow recoveries by revising their estimates of “maximum employment” downwards. In this narrative — rather than demonstrating a failure to manage the business cycle — persistent unemployment following a recession instead indicates some “structural” mismatch preventing the economy from returning to pre-recessionary strength. If job loss can be reframed as “structural” rather than “cyclical,” it becomes the responsibility of workers to “find new skills” and “learn to code,” rather than the responsibility of policymakers to support workers through an economic downturn. Consider Charles Plosser’s claim in 2008 that unemployment would remain elevated simply because, “you can’t change a carpenter into a nurse.” This is not much different from the obsession with “skill-biased technological change” that gripped macroeconomists as the recovery dragged on past the two year mark. Unfortunately, this “structural” narrative can easily lead to premature policy tightening that stifles a nascent recovery.
I’ve felt this way for a while now. I see no reason why the potential of physical capital, human capital, and factor productivity should necessarily decline after a recession. In fact, it’s vague if they decline in general. The only things that always decline are the financial sector and pure labour hours. However, those can recover quickly. In fact, I partially credit the shadow banking sector with driving the faster US recovery from the Great Recession compared to Europe which lacks such a sector.
It is further necessary to note the way frictions adjust the labour recovery rate. Mostly in the hiring process, vagaries can obscure the true rate of recovery in the labour market.
End the Phillips Curve
Milton Friedman recognized it. Emi Nakamura recognizes it. It’s time for the Phillips Curve (as Keynesians know it) to be kissed goodbye. Unemployment does not drive inflation.
Instead, there is a delicious web of factors influencing the relationships of what can be called the Monetarist non-Phillips not-really-Curve. One, inflation can ease labour market conditions, meaning a lower unemployment rate. That’s a standard relationship. However, it is more accurate to say that nominal output growth drives both. Depending on the shape of real factors, the curve ceases to be, as inflation is held down while employment is pushed up.
Those real factors are twofold: standard union and capacity constraints, and the shape of markets as influenced by demand policy. You can perhaps denote the natural rate of unemployment as the rate on which policy will be ineffective unless said policy changes the long run factors (such as via communication), or the point at which the curve must shift. In effect, don’t look at the curve. Look at what NGDP is doing and how EPOP/NPOP and inflation expectations respond.
Amarnath, Skanda, and Alex Williams. Beyond The Phillips Curve: A Dynamic Approach To Communicating Assessments of “Maximum Employment”. Employ America Medium Article, June 17th, 2021. https://employamerica.medium.com/beyond-the-phillips-curve-a-dynamic-approach-to-communicating-assessments-of-maximum-employment-c3eff48b2fcf.
I will note that it is ambiguous whether they believe in a correlation or causation here.
Sumner, Scott. A Critique of Interest Rate–Oriented Monetary Economics. Working Paper, Mercatus Center at George Mason University, November 23rd, 2020. https://www.mercatus.org/system/files/sumner-critique-interest-rate-mercatus-working-paper-v1.pdf.
I shall write about this another day.
I have a forthcoming paper on this subject.
I have the American Medical Association in mind here.
I mention this later.
In the U.S. Golden Era in Capitalism, 2/3 was financed by velocity (by putting savings back to work, by the thrifts/nonbanks, backstopped by the FSLIC, NCUA etc.). A dollar of savings (income held beyond the income period in which received), is more potent than a dollar of the money stock. R-gDp averaged 5.9% during 1950-1966 (in spite of the 3 recessions).
In the U.S. Golden Era in Capitalism, 2/3 was financed by velocity (by putting monetary savings back to work, by the thrifts/nonbanks, backstopped by the FSLIC, NCUA etc.). A dollar of savings (income held beyond the income period in which received), is more potent than a dollar of the money stock. R-gDp averaged 5.9% during 1950-1966 (in spite of the 3 recessions).