Galí on the State of New Keynesianism, Part II

Print/Save PDF

Last week’s post looked at, and found wanting, Jordi Galí’s 2018 optimistic assessment of the state of New Keynesian macroeconomics. My negative judgment was not influenced by his downplaying of the great shame of NK macro modeling, which I planned to treat separately in this post. The disgrace, of course,  is NK theory’s uselessness in helping stabilization authorities understand and effectively confront collapsing nominal demand that was ominously organizing itself late in 2008. Galí eventually gets to that failure, about which he also optimistic

Galí offers his own instability model, which he believes is indicative of important progress on the great shame that is being made within the NK market-centric general-equilibrium framework. From late in his survey: “in Galí (2017a), I explore the possibility of fluctuations driven by stochastic bubbles in a New Keynesian model with overlapping generations. Stochastic bubbles grow at a rate above the long-term growth of the economy, generating a boom in output and employment. But these bubbles may collapse at any time with some (exogenously given) probability, pulling down aggregate demand and output when they do. Despite the highly stylized nature of the model, the implied equilibrium appears consistent with the pattern of asset price booms followed by sudden busts (and the induced recession) that has characterized historical financial crises.”

To me, Galí’s reimagining real-side Austrian business cycles is as unconvincing as every other attempt to use them to describe highly specialized economies. Simply positing “highly stylized” (read arbitrary) bubbles, he jumps through hoops to avoid central roles for meaningful wage rigidity and involuntary job loss. He badly serves stabilization policymakers by not clearly establishing the criticality of aggressive interventions to halt and reverse collapsing nominal demand. He is occupied with tiny issues while ignoring the overwhelming need to promote changes that preserve and strengthen the central bank’s capacity to adequately manage total spending. That NK theorists miss the main point in stabilization analysis provides room for foolish Congressional actions to prohibit remedial actions that actually worked in the Great Recession.

The GEM Project’s model of extreme instability, informed by the generalization of rational exchange and microfounded MWR, provides an instructive contrast to Galí. First, it assigns nominal demand disturbances a central role, separating them into two classes.  Stationary demand disturbances (SDD) are inherently self-correcting, helped along by automatic stabilizers augmented by central-bank “lean-against-the-wind” intervention. Nonstationary demand disturbances  (NDD) feature contracting total spending that overwhelms automatic stabilizers and orthodox central-bank intervention. If not contravened, extreme instability induces rapidly accumulating forced job and income loss, collapsing output and profits, impaired lending by and funding for financial institutions, gathering price deflation, widespread debt default, wealth destruction, and cataclysmic depression.

Second, it draws upon Roger Farmer’s (2010b) revival of the intuitive idea that confidence exerts a “separate, independent” influence on spending behavior. His specific contribution is that investor/lender confidence is rooted in a positive feedback relationship with mark-to-market financial asset prices and, as a result, is effectively calibrated as an increasing, nonlinear function of those prices. In short, Farmer posits asset prices to be an effective proxy for the varied determinants of investor/lender confidence that have been identified in the literature.

A GEM-enhanced version of Farmer’s feedback mechanism, especially relevant in NDD circumstances, is:

₡(t)=ƒ(V(t-1),ΔV(t-1)) such that Δ₡/ΔV>0, Δ₡/Δ(ΔV)>0, and

V(t)=ƒ(P(t), ₡(t), Ƈ(t)) such that ΔV/ΔP>0, ΔV/Δ₡>0, 0≤ Ƈ ≤1,

where ₡ is investor/lender confidence, Ѵ represents composite asset prices, Δ denotes change, P represents discounted rational expectations of future profits (subject to a stable probability distribution), and Ƈ measures notional investor/lender perceptions of the credibility of stabilization authorities’ trend real-side objective. P is rooted in neoclassical determinants that fundamentally drive asset pricing, such as the capital stock, labor force, and state of technology. Ƈ is the contribution of the GEM Project, mediating the relative influence of P and in the determination of V and providing a critical demand link to stabilization policy. (See below.)

Third, GEM theory assigns a critical role to Nancy Stokey’s (2009) thesis that, as investors/lenders become less certain about macro prospects, simple inaction becomes increasingly rational. Many buyers of financial assets respond to uncertainty inherent in a nontrivial possibility of unchecked NDD by moving to the sidelines, waiting for the emergence of a credible floor under falling prices. (A maxim of veteran traders is not to try catching a falling knife.) The GEM Project enriches Stokey’s analysis by motivating rational inaction with breakdowns in the credibility of stabilization authorities’ trend full-employment objective. Asset buyers’ moving to the sidelines provides low-risk opportunities for short-sellers, imposing outsized downward pressures on asset pricing. Moreover, if and when stabilization policymakers’ credibility is restored or the NDD threat otherwise dissipates, the rebound in asset prices will be relatively rapid.

Stabilization authorities’ trend real-side credibility, once damaged, introduces uncertainty into the valuation process. When credibility is high, investor/lenders’ expectations of the macro future are governed by stable probabilities; and asset pricing is driven by expectations of profit fundamentals. As credibility erodes, investors become uncertain; and asset-price dynamics are increasingly driven by investor confidence. A linear version of the feedback system is:

V(t)= Ƈ (t)P(t)+(1- Ƈ (t))[a1 Ƈ (t-1)+a2ΔV(t-1)], such that a1+a2=1.

The GEM Project easily extends CV feedback from asset pricing to total nominal spending, the focus of its NDD analysis. In the enriched framework, investment outlays are already described, determined by P, , and Ƈ, with policymaker credibility mediating the relative influence of economic fundamentals and confidence. In addition, consumption spending is influenced by household wealth and is, therefore, already sensitive to mark-to-market asset prices. Finally, in episodes of contracting aggregate demand, incomes throughout the economy are destabilized by powerful multipliers unleashed by contracting employment and income. Putting the pieces together, broad macro uncertainty produces a second, more extensive destabilizing feedback between the interacting confidence/credibility measures ( and Ƈ) and aggregate demand.

In contrast to Galí’s arbitrary bubbles, the GEM theory of extreme instability (constructed on Farmer confidence, Stokey rational inaction, and credibility of the Fed’s trend full-employment objective) features endogenous demand disturbances propagating rational financial shocks, broad asset-price collapse nowhere near aligned with actual resource endowments, the translation of contracting total nominal spending to millions of involuntary job losers and evidence-consistent contractions of output and employment, and the identification of remedial policies recognizably consistent with policies that actually worked in 2008-09. The GEM Project rejects as a foolish pipedream the NK reliance on our regulatory capacity to prevent future financial shocks.

Blog Type: Wonkish Saint Joseph, Michigan

Write a Comment

Your email address will not be published.