BONUS: Noah Smith on the State of Macroeconomics

Episode Summary

Executive Summary: Noah Smith argues macroeconomics remains “in its infancy”: DSGE/New Keynesian models improved theory but often rest on weak microfoundations and limited, noisy data. He favors building knowledge from empirical micro studies, accepting ugly but testable models, and using a “fox” approach at the Fed. He also criticizes MMT and other “nutty” theories as unfalsifiable, while distinguishing them from merely bad but testable ideas like price controls.

Main Topics: The limits of DSGE and New Keynesian macro (Priority: 5/5): Smith traces the rise of RBC/DSGE and New Keynesian models, noting their dominance after the 1980s and their post-crisis updates for financial frictions and the zero lower bound, but argues they still fail to provide trustworthy structural explanations. Weak microfoundations and empirical mismatch (Priority: 5/5): He contends that core microeconomic building blocks used in macro models, especially Euler equations and Calvo-style pricing, do not fit actual micro data well enough to support strong claims about structural policy effects. A bottom-up, empirical research agenda (Priority: 5/5): Smith advocates studying consumption, investment, labor supply, debt, and firm behavior through quasi-experimental micro studies to accumulate smaller facts that can later be integrated into broader macro models. Model uncertainty, general equilibrium, and limited data (Priority: 4/5): While acknowledging general equilibrium concerns, Smith says micro evidence is still valuable because it provides conditional behavioral responses that can later be checked across different macro regimes and settings. How policymakers should think: be a fox (Priority: 4/5): For central banking, Smith recommends a ‘fox’ approach inspired by Tetlock: combine many models, forecasts, and informal judgment rather than relying on one grand theory. Critique of MMT and other non-falsifiable theories (Priority: 5/5): He argues that MMT and praxeology are not just wrong but gibberish because they evade clear predictions and formal testability, unlike mainstream macro, which may be imperfect but remains testable. Distinguishing bad ideas from nutty ideas (Priority: 3/5): Smith draws a line between ideas like price controls, which are testable and sometimes grounded in history, and theories like MMT that lack explicit models and never admit failure.

Key Arguments: Macro since the 1980s has seen a methodological revolution, but the new mathematical sophistication did not produce reliable microfoundations or robust policy guidance. Prescott’s RBC work mattered more for championing general equilibrium and mathematical formalism than for accurately explaining recessions. Consumption Euler equations, a core ingredient of many models, often fail in micro data and may even show the opposite sign from model predictions. Heterogeneous-agent New Keynesian (HANK) models are a useful exercise, but their complexity, convergence issues, and lack of rigorous existence proofs limit confidence in their outputs. Because macro data are sparse, serially correlated, and highly interdependent, top-down validation of large structural models is inherently weak. Young macroeconomists are making progress by studying narrow causal questions—like UI extensions, debt forgiveness, and firm responses to recessions—rather than trying to solve the entire economy at once. The Fed should use multiple models and sources of information, then apply judgment; no single framework is sufficient. MMT is criticized as unfalsifiable and vague because it avoids clear written theory and evades accountability for failed predictions. Neo-Fisherian ideas are cited as an example of a testable but generally wrong theory, distinct from incoherent theories like MMT. Price controls are bad policy in Smith’s view, but they are not ‘nutty’ because they are at least testable and historically grounded. Economic science should aim for falsifiable, directional predictions rather than claims of complete certainty or universal truth.

Data Points: Decades of macro data: A few decades - Smith emphasizes how little aggregate time-series evidence exists for validating large structural macro models. Macro theory explosion: 1980s - He dates the major DSGE/RBC methodological revolution to the 1980s. Zero lower bound: 0% lower bound - Mentioned as a key post-Great Recession update to New Keynesian models. Inflation in Turkey: Spiraling / hyperinflation - Used as an example of policy failure when a government keeps interest rates low amid rising inflation. Model ranking concept: Foxes vs. hedgehogs - Tetlock-inspired framework Smith uses to describe better forecasting and policymaking behavior.

Pivotal Quotes: "Macroeconomics is in its infancy." — Noah Smith: His core thesis about the state of the field and why humility is warranted. "The macro is in its infancy." — Noah Smith: Reiterated as a summary of the need to build up evidence from smaller empirical studies. "If you don't know how micro agents work... can you trust a macro model that seems to fit aggregate data?" — Noah Smith: His Lucas-critique-based objection to relying on aggregate fit without credible microfoundations.

Implications: Listeners should expect macro policy debates to remain uncertain and pluralistic. For the Fed and researchers, the path forward is more micro evidence, more model competition, and less confidence in grand theories—especially those that cannot be tested or falsified.