Economic Model Predictive Control

Economic model predictive control (economic MPC) is the application of model predictive control principles to economic systems — markets, supply chains, resource allocation, and macroeconomic policy. Rather than tracking a pre-specified reference trajectory, economic MPC optimizes a direct economic objective: profit, welfare, cost minimization, or risk-adjusted return, subject to dynamic constraints on inventories, capital, labor, and regulatory limits.

The shift from tracking to direct economic optimization changes the mathematical structure of the problem. Classical MPC seeks to minimize deviation from a setpoint; economic MPC seeks to maximize a flow of value. The steady-state behavior is no longer a fixed point but an optimal operating regime, and the controller's task is to drive the system to this regime while satisfying transient constraints. This requires new stability concepts — notably dissipativity and turnpike properties — that guarantee the system does not get trapped in suboptimal cycles.

Economic MPC has been deployed in chemical process control, energy market bidding, and inventory management. Its most consequential application is in macroeconomic policy, where dynamic stochastic general equilibrium models function as implicit MPC systems with very long horizons. The Federal Reserve's interest rate decisions, for instance, can be interpreted as solving an economic MPC problem with a multi-decade horizon and highly uncertain dynamics. The objective function is not disclosed; the optimization is performed by human committees rather than algorithms. But the structure is the same.

Economic MPC reveals that all policy is control, and all control is optimization under constraint. The question is not whether to use MPC in economic governance, but whether to make the objective function explicit and contestable — or to leave it implicit and therefore beyond democratic scrutiny.