
In (1966) and (1967), I introduced an endogenous theory of technological progress in which there were a (then new) nonconventional factor of production, “technological knowledge” A, and a (then new) sector R, inventive activity. Constant returns to scale were assumed to prevail with respect to conventional factors of production, but increasing returns prevailed in all factors taken together. Hence purely competitive provision of inventive activity is not possible. Also because of increasing returns, the growth process is historydependent, permitting both explosive growth and contractionary growth. I put forward in (1973) the first growth model in which inventive activity depends on the prevailing industrial organization. The presentation of this 1973 IO paper is more Marshallian than I would have preferred; I used the best tools available to me at the time. Before the “new growth theory”, I was pursuing endogenous growth.

In 1971, I introduced the formal generalequilibrium analysis of Paul Samuelson’s overlappinggenerations (OG) model. I showed that the restrictions on market participation caused by births and deaths are not the cause of the “Samuelson’s friction.” The open horizon is the sole source of nonoptimality in Samuelson’s model. Yves Balasko and I provided the equilibrium and welfare analyses of monetary (and nonmonetary) OG economies. Yves and I studied shortrunParetooptimal (SRPO) allocations. SRPO is a weakening of the PO criterion: PO allocations are SRPO, but not all SRPO allocations are PO. A “transversality” condition characterizes the SRPO allocations that are indeed PO. The two welfare theorems for OG economies are valid if the PO criterion is replaced by SRPO criterion. There are at least two innovations in the Samuelson model. One, of course, is the demography. The other — equally important — is that taxes and transfers are denominated in money. Yves and I introduced bonafide taxtransfer policies — policies that allow money to have positive value. Contrary to Ricardo, retirement of the public debt is not necessary for bonafidelity in OG economies: there can be rational bubbles. These bubbles might even be stochastic. Indeed, the SSE literature began with rational bubble money with pricelevel volatility. See my Malinvaud lecture (1977).

Dave Cass and I invented the sunspot equilibrium (SSE) concept. “Sunspots” is shorthand for purely extrinsic uncertainty — uncertainty that does not affect the fundamentals but may affect the outcomes. There are two interpretations: (1) SSE are the limit of the equilibria as the volatility of extrinsic uncertainty become small. In this context, SSE provide a rationalexpectations explanation of “excess market volatility” — the excess of the randomness in outcomes over the randomness in the fundamentals. This interpretation is what motivates the use of the word “sunspots”. Realworld (i.e., Jevons) sunspots do affect the fundamentals, but not very much. (2) Economies — like games — are social institutions in which agents must base their beliefs and actions on the beliefs and actions of others. Just as in game theory, there can be stochastic solutions (SSE), even in economies with nonstochastic fundamentals.