Anyone who has drawn a demand curve has watched the invisible hand at work. Pick one—a price or a quantity of a good—and the curve traces out the other. Standard New Keynesian models, the workhorse framework for modern analysis of business cycles and macroeconomic policy, follow this logic: firms choose prices, while an invisible hand makes them produce whatever quantity consumers demand at those prices. That shortcut simplifies analysis by ensuring that markets clear, but paints a rather unrealistic and unsatisfying picture of how businesses make choices. Firms in the standard model have no direct control over their own production decisions, which can even force them to supply goods that they fully know would generate profit losses.
In this paper, the authors take a step toward closing that logical gap. They build a general equilibrium model in which firms optimally choose both prices and production before demand and productivity are known. The framework combines the rational expectations and optimal pricing of New Keynesian economics with the disequilibrium that the Old Keynesian tradition identified as a core mechanism for economic fluctuations. Crucially, the fact that markets may be rationed or slack is not assumed: instead, it emerges from firms’ optimal decisions given the information that they have.
In their model, two principles govern how much is sold. First, exchange is voluntary: no buyer or seller can be forced to trade more than they want. Second, exchange is efficient: no mutually beneficial sale goes unmade. Together, these principles imply that sales equal the smaller of what the firm produced and what customers demand at the posted price. Markets therefore end up rationed, when demand exceeds production; slack, when production exceeds demand; or, only in a knife-edge case, exactly balanced.
The authors then embed this firm problem in an otherwise familiar macroeconomic model. When one product sells out, households redirect spending toward products that remain available, so conditions in one market change effective demand in others. Tracking all combinations of slack and rationing has long made multi-market disequilibrium models unwieldy. Here, uncertainty and differentiated goods generate a smooth distribution of market states, allowing the authors to characterize those spillovers analytically and describe their macroeconomic implications. This yields several new insights about how shocks propagate through the economy:
- A firm’s markup is no longer pinned down solely by the elasticity of demand. In conventional models with market clearing, the equalization of marginal revenues and marginal costs of production leads to a tight connection between markups and the elasticity of demand. Here, an extra unit of production earns revenue only if the firm would otherwise sell out, while its cost is paid in every state. As a result, markups are governed by the probability of rationing.
- Uncertainty changes firms’ choices at first order. Because marginal revenue jumps when a market switches between slack and rationed, even a small increase in volatility raises prices and reduces planned hiring.
- Cross-market demand spillovers can run in the opposite direction from the standard New Keynesian logic. Producing more in a rationed market lets households spend more there, leaving less to spend in slack markets. Expansion by one firm can therefore reduce demand elsewhere, overturning the usual presumption that firms’ production decisions are strategic complements.
- An economy-wide increase in uncertainty acts like a stagflationary cost-push shock. Firms raise prices, some customers are rationed while goods elsewhere go unsold, consumption falls, and measured labor productivity declines even though physical technology has not changed. Employment moves little. In the nested New Keynesian benchmark, this uncertainty shock produces none of these effects.
- An unexpected monetary expansion raises consumption by absorbing product-market slack. Households use the additional spending power to buy goods that were already produced but had gone unsold. This mechanism does not require firms to instantaneously expand production to meet demand, as standard New Keynesian models do; consumption can rise quickly while production and employment respond only modestly.
- Monetary expansion therefore raises measured labor productivity rather than lowering it through price dispersion. Sales and consumption rise faster than labor input because the economy uses existing capacity more fully. This helps the model account for evidence that spending responds to monetary shocks much more quickly than employment—a pattern the simple New Keynesian benchmark struggles to generate.
Methodologically, the authors’ analysis shows how the over- and under-supply of goods —long understood to be crucial to economic fluctuations by the Old Keynesian tradition —can be combined with various features of modern macroeconomic models including rational expectations and rational, payoff-maximizing choices. This synthesis of “New Old Keynesian” ideas yields new insights about how shocks affect the macroeconomy, while remaining tractable enough to be embedded in even larger and more complex business-cycle models.