A General Theory of Macroeconomics: Path Laws, Endogenous Amplification, and Constraint Geometry

This paper states a general theory of macroeconomics whose primitive object is the law of an economic path rather than an equilibrium point perturbed by exogenous shocks. The theory rests on three laws --- paths are the primitives; fluctuations are generated by endogenous amplification; feasibility geometry disciplines dynamics --- articulated through nine postulates, one master system, and a body of theorems. Equilibrium is relocated, not discarded: it is a fixed point of the path-law map , rational expectations is its linear--quadratic special case, and the classical theory as a whole is recovered as a restrictive corner of parameter space (no feedback, fast mean reversion, convex constraints, reversible paths, thin tails). The proved core comprises: (1-ρ)⁻¹ amplification laws, including a stationary Volterra energy identity with an exact pre-window correction; a crisis trichotomy (amplification, endogenous cycles, constraint collision); a self-organized-criticality theorem in which profit gradients consume the spectral gap; an exact Pareto law for the crisis multiplier X=(1-ρc)⁻¹, with tail index α=h₀/c --- the hazard-to-drift ratio --- and a stated regular-variation condition under which realised severities inherit Pareto tails; an identification theorem establishing the L\'evy area as a detector of directed feedback that is exactly null under reversible common-factor structure, with the correlated-factor confound stated explicitly, together with a proof that univariate second moments carry no information about endogeneity; a hysteresis theorem replacing the I(0)/I(1) dichotomy with a kernel-exponent continuum; a convexification-error bound for nonconvex feasibility; the expected signature as a moment hierarchy capped by the depth barrier in crisis-tail regimes; a thinning microfoundation of the crisis hazard; area-estimator asymptotics; and a welfare theory in which the expected loss rate is U-shaped in the tail index with interior optimum α^=β+√β²+β, so that both hyper-fragile and hyper-rigid economies are dominated inside the stated reset-cost objective. Sectoral chapters develop money and inflation, labor and hysteresis, growth and the trend--cycle question, and the open economy inside the same architecture. The theory is falsifiable, and its failure modes are stated in advance.