I cannot just yet let the GEM Blog examination of Carlin and Soskice’s Macroeconomics (2015) go. It is very interesting to me how much rational-behavior generalized-exchange modeling would enhance that text, making it uniquely important. The advantage of the 2nd edition is its construction on Early Keynesian (EK) attentiveness to sluggish adjustment of nominal wages and prices that interact with adverse demand disturbances to generate involuntary job loss and reduced output. The EK model is transparently stabilization-relevant. The disadvantage is the C&S reversion to demand centricity shares the debilitating Early Keynesian problem of not being rooted in optimization and equilibrium, the two fundamental tenets of economic theory. This post looks at one more example of the mischief that results when model-building guidance from irrational assumptions of convenience replaces guidance from rational behavior.
In a section near the end of the text, entitled “sources of persistence in state-of-the-art NK models”, C&S (p. 610) write: “In the NK DSGE models found in central banks, many adjustments [i.e., irrational assumptions] are made to the model to allow it to capture the persistence of inflation and in the output response to shocks that characterize real economic data. The model which many central banks have used is a version of the Smets-Wouters (2003, 2007) model. This model incorporates features which are introduced to better match the sluggish adjustment of the economy such as:
- partial indexation of prices and wages, which introduces lagged inflation into the Phillips curve;
- habit formation in consumption and investment adjustment costs, which introduces a role for lagged output in the IS curve;
- lags in the Taylor rule, which introduces the lagged interest rate in the Taylor rule….”
Also from C&S: “These sources of persistence are necessary for NK DSGE models to fit the data. They considerably complicate the models and make it difficult to uncover the role played by micro-foundations that stem from the origins of the model in the RBC modeling tradition.”
The first source cited is no more than a particular restatement of the EK keystone assumption of wage catch-up to past inflation that motivated early Phillips curves. It blatantly violates the New Neoclassical Synthesis (NNS) prohibition of irrational assumptions of convenience. If positing “partial indexation” absent microfoundations is OK, EK stabilization-relevant modeling should be welcomed back, with apologies, into mainstream thinking. Related, but worse, the NK hypocrisy here continues to damage rational-behavior model-building guidance. (Recall the post two weeks ago.)
The second inertia source is more non-microfounded assumptions of convenience. C&S cite “the consumption function discussed at the end of Chapter 1, where a proportion of households use rule of thumb behaviour and base their consumption decisions on current income, whilst the rest are modelled as using the forward-looking permanent income hypothesis. Another frequently used rationalization is the idea of habit formation in consumption, which produces slow adjustment to income shocks.” It is instructive here that the GEM Project microfounds income as the principal driver of consumption and carefully constructs a framework for the analysis of habit-formation, both demonstrating the superiority of rational behavior in model-building guidance. Third, Taylor-rule lags is macro-speak for the obvious point that stabilization authorities experience recognition and implementation lags in responding to demand shocks. But logic and experience indicate that this is not a significant practical issue. Such lags have never generated problematic inflation momentum.
So far, the analysis has not been kind to mainstream NK theorists. Their supposed causes of macro inertia are little more than assumptions of convenience that embarrassingly contravene the New Neoclassical Synthesis and provide misleading model-building guidance. The especially hard fact is that the NK sources of inflation persistence cannot coexist with wage recontracting. They cannot rationally suppress that bedrock mechanism of neoclassical market-centric modeling.
Meanwhile, the generalized-exchange macro model, with its microfounded MWR, easily solves the nominal persistence problem. Its primary contribution here is the derivation of optimizing catch-up to past inflation in wage determination in workplaces restricted by costly, asymmetric employer-employee information. But, and this is important, not even microfounded wage-price persistence is able to rationally suppress wage recontracting. In a choice between losing one’s job or a wage reduction that does not violate opportunity costs, the latter is the worker’s only rational choice. The Keynesian reliance on lagged nominal wages/inflation to motivate recognizable macro cyclicality is, in GEM modeling, replaced by optimizing MWR.
In GEM Phillips curve analysis, the significant nominal-to-real action switches from lagged inflation to the careful specification of the role of unemployment, which is greatly influenced by the microfounding of MWR. Adequate analysis of the Phillips relation and related rational suppression of wage recontracting cannot be done in the NK market-centric general-equilibrium modeling. Next week’s post continues our look at the Phillips curve.
Blog Type: New Keynesians Saint Joseph, Michigan