Recall that in the growth models of Harrod-Domar, Kaldor-Robinson, Solow-Swan and Cass-Koopmans, we argued, explicitly or implicitly, that technical change was “ exogenous”. In Schumpeter's version, this was not true: we had "swarms" of inventors arising under particular conditions. The Smithian and Ricardian models also featured technical changes resulting from profit squeezes or, in Smith's particular case, due to prior technical conditions. Allyn A. Young (1928) had argued for the resurrection of the Smithian concept in terms of increasing returns to scale: the division of labor induces growth which allows for further division of labor and therefore even faster growth. The idea that technological change is induced by prior economic conditions can be termed “endogenous growth theory”. The need for a theory of technical change was there: according to some rather famous calculations by Solow (1957), 87.5% of the growth in production in the United States between 1909 and 1949 can only be attributed to technological progress. . Therefore, the so-called “Solow residual” – the term g(A) in the growth equation given above, is enormous. One early reaction was to argue that by reducing much of this influence to pure capital improvements, capital intensity appears to play a more important role than imagined in these 1957 calculations. – Solow goes on to argue, for example, that increased capital – intensive investment embodies new machines and new ideas as well as increased learning for even greater economic progress (Solow, 1960). However, Nicholas Kaldor was actually the first post-war theorist to consider endogenous technical change. In a series of articles, including a famous one from 1962 with JA Mirrlees, Kaldor postulated the existence of a “technical progress” function. that per capita income was indeed an increasing function of per capita investment. Thus, “learning” was seen as a function of the rate of increase in investment. However, Kaldor considers productivity increases to have a concave character (i.e., increases in labor productivity decline as the investment rate increases). Of course, this proposition does not satisfy Solow's insistence on constant returns. asdsadasdasdaK.J. Arrow (1962) considers that the level of the “learning” coefficient is a function of cumulative investment (that is to say, past gross investment). Unlike Kaldor, Arrow sought to associate the learning function not with the growth rate of investment but rather with the absolute level of knowledge already accumulated. Because Arrow claims that the new machines are improved, more productive versions of existing ones, the investment not only induces growth in labor productivity on existing capital (as Kaldor claims), but it would improve also labor productivity on all subsequent machines. manufactured in the economy.