Who wasn’t invited?
Last week Federal Reserve Chairman Warsh announced the external co-heads of his five advisory task forces on communications, the Fed’s balance sheet, data analysis, productivity and jobs, and the Fed’s inflation framework. The general assessment was that the intellectually and politically diverse group of well respected experts signalled genuine institutional reform rather than political theater.1[1] That is consistent with Chairman Warsh’s philosophy of “epistemic humility.” I will provide a deeper dive into what I expect from the five task forces in Part II of my Warsh cycle series at Thematic Markets. But here I’d like to focus on the powerful signal sent by who was not invited to the party. The absence of a notable thought leader from any particular task force may simply be idiosyncratic. But wholesale exclusion of any adherent of what is arguably macroeconomics’ dominant intellectual paradigm, the New Keynesian Synthesis (NKS), is a powerful statement by Mr. Warsh that he rejects its guidance for central banking. That has important implications for the conduct of Fed policy.
Nominal rigidities
If you aren’t an economist, a bit of background will be helpful to understand the significance of NKS to modern central banking and the implications of its rejectionion by the Warsh Fed. I’m simplifying a lot, but in the 1970s and ‘80s economists divided into two broad camps: Keynesians who believed that monetary policy can induce economic fluctuations due to “sticky prices” that force adjustment onto real variables like employment and output, and Real Business Cycle (RBC) economists who thought that primarily changes in preferences and technology explain business cycles. Keynesians’ models better fit the macroeconomic data but lacked internal consistency and, in RBC economists’ eyes, “microfoundations” that grounded macroeconomic outcomes to rational choices by individuals. Keynesians in turn criticized RBC models for lacking a role for monetary policy.
New kids on the block
The Keynesians began to experiment with adding sticky prices to RBC models and by the 1990s generated a class of models that became known as the New Keynesian Synthesis that seemed to bridge economists’ divide. The models were mathematically complex — e.g. see the NKS equations in the cover art of one of my early critiques of their abuse in monetary policy — but were theoretically consistent at the level of individuals like RBC models while generating aggregate behavior that seemed to match the real economy like the old Keynesian models.
Look at my new toys!
More excitingly (for economists) NKS models offered a powerful new set of tools for economic analysis and formulation of monetary policy. For instance, because expectations drive behaviors in the new models, NKS implied that central banks could affect the economy through credible communication. This provided a more rigorous basis for inflation targeting and adherence to transparent rules (rather than policy discretion) to anchor expectations, and forward guidance to manipulate expectations. NKS also generated new metrics to guide policy: the difference between actual and potential output (the output gap) and the neutral interest rate (r*, “r-star”) that keeps the output growth and inflation steady at potential and target, respectively. By adjusting both interest rates relative to r* and expectations through forward guidance an optimizing central bank could keep the economy running at a Goldilocks “just right” temperature with no output gap and inflation anchored at its target.
The eye of Sauron
It’s easy to see why economists would fall in love with the NKS framework. It provides a mathematically elegant — economists love math! — description of the world that claims to be predictive and makes economists into all-powerful managers of our expectations using variables that they construct (the output gap and neutral interest rate). Unsurprisingly, NKS scholars came to dominate both the academy and central banking and their insights became the foundations of what many have considered modern monetary policy in the last two decades. The proof of that is that many lay people are familiar with terms like “forward guidance,” “output gap,” and “r-star” used by central bankers and the financial press but unfamiliar with non-NKS economic argot like “Ricardian equivalence” or “Pareto efficient.”
Signs of trouble
While some of us questioned whether any central banker, even Janet Yellen, possessed the omniscience required to “optimally control” monetary policy as NKS suggested one could,2[2] signs of trouble began to emerge with the Global Financial Crisis (GFC). The crisis itself brought attention to the fact that the NKS largely ignored banks, credit, liquidity constraints, and balance sheets of both households and firms. More generally, it became increasingly apparent that the real economy has many rigidities that affect behavior and can propagate changes in monetary policy than just the cost of updating restaurant menus: financial frictions between borrowers and lenders, information asymmetries between buyers and sellers, and differences in risk tolerance between individuals. Ironically, the GFC caused most central bankers to double down on the NKS framework as interest rates hit the zero lower bound because it left them with forward guidance as one of the few remaining tools.
Representative of everything? Or nothing?
One of the biggest problems the GFC exposed was the NKS’s reliance on so-called “representative agent models,” i.e. a simplifying assumption that everyone in the model is exactly the same, working in one-person firms, paying themselves identical wages, and consuming the same as everyone else. Thus, the NKS model predicts that when interest rates fall, everyone consumes more because saving earns less. But in reality, many retirees consume less because their interest income falls, those living hand-to-mouth can’t change their consumption, and debtors (like mortgaged home owners) use their interest savings to raise consumption. A new class of NKS models “Heterogeneous Agent New Keynesian” (HANK) models with multiple “agents” arose to address these problems, creating “insights” into new channels of monetary transmission, but they still leave many of the financial and inequality frictions that plagued original NKS models.3[3]
About that fit…
Another problem with NKS models — even the newly improved HANKs — is that their claimed fit to the data is very subjective. Because the models are either simulated or require the estimation of unobserved variables — e.g. the output gap and neutral interest rates — their fit requires a lot of assumptions from the estimating economist. Change the assumptions and the fit worsens or even disappears entirely.4[4] Yet, it is these models that central banks have been basing policy upon.
The crux of the problem
This gets to the crux of the problem with the NKS. It isn’t that this vein of research hasn’t generated useful insights: expectations do matter for economic outcomes, hence my view that Being is believing drives inflation; and potential GDP growth and neutral real interest rates are valid concepts for economic analysis, hence the impact of Localization on both. But the Illusory omnipotence that central bankers have derived from NKS zealotry over the last two decades has led to repeated policy errors. Potential growth, neutral interest rates and even expectations are unobserved and can only be inferred through an array of signals from the economy that no model is complex enough to reliably capture. Yet over the last couple of decades central banks have made policy as though their model-estimated values were real observations. This is the real problem with the NKS revolution in Economics and central banking.
My bête noire slain?
Long-time readers of my research know that abuse of NKS models has been a consistent source of my criticism of central bank policy for over a decade. Whether it is the deeply flawed, NKS-derived Laubach-Williams model of neutral interest rates, estimates of potential growth that appear detached from reality, or confusing bond market forecasts for consumers’ inflation expectations, the over reliance of the Fed and other central banks on NKS models for policy analysis has generated the worst inflation in more than 40 years, not only in peak magnitude but in duration. Hence, Chairman Warsh’s embrace of epistemic humility and distancing from NKS is a welcome sign of a return to orthodox central banking that will hopefully return the US (and perhaps other major economies) to price stability in the near future.
The first step to recovery
That said, admitting the problem is only the first step. It’s all well and good to end reliance on constructed models and data, and to cease questionably effective, credibility-reducing practices like forward guidance. But something has to replace it. It also is important not to throw the baby out with the bathwater. As noted, NKS has provided useful economic insights even if it failed as a framework for policy (in my view). Chairman Warsh seems to understand this as he, too, uses its conceptual terms in his speeches, despite ignoring NKS in his advisory body selection. Let’s hope the consensus impression that the announced task forces appear to signal real reform rather than performative theater and are staffed with Economics’ all-star team is correct, because the Fed needs it after two decades of folly.
Comments are available to paid subscribers only.