The authors treat macroeconomic models as composed of large numbers of micro-units or agents of several types, and explicitly discuss stochastic dynamic and combinatorial aspects of interactions among them. In mainstream macroeconomics sound microfoundations for macroeconomics has meant incorporating sophisticated intertemporal optimization by representative agents into models. Optimal growth theory, once meant to be normative, is now taught as a descriptive theory in mainstream macroeconomic courses. In neoclassical equilibria flexible prices led the economy to the state of full employment and marginal productivities are all equated. Professors Aoki and Yoshikawa contrariwise show that such equilibria are not possible in economies with a large number of agents of heterogeneous types. The authors treat equilibria as statistical distributions and not as fixed points. They employ a set of statistical dynamical tools via continuous-time Markov chains, and statistical distributions of fractions of agents by types available in the new literature of combinatorial stochastic processes, to reconstruct macroeconomic models.