Conventional wisdom today follows Malthus in viewing the economic effects of population growth negatively. However Simon  has challenged this view, predicting that population growth will have a positive effect on per capita income, at least in the long run. The researches of Minami-Ono [1971, 72, 75] and later works of Kelley-Williamson  support this optimistic view. On the contrary, the works of Ogawa-Suits  and the earlier works of Kelley-Williamson  showed population to have large negative effects. In this paper an attempt is made to investigate these opposite views by using a general equilibrium growth accounting and some simulation techniques. Considering per capita income growth, it is clear that the direct contribution of population growth is negative while that of labor growth is positive. Less obvious are the positive indirect contribution of population and labor growth through the influences of population and labor on the rate of technical change in either sector. We measured the total contribution (direct contribution 4- indirect contribution) of population cum labor on per capita income by using a two sector growth accounting model. The calculated result shows that the total contribution of population cum labor growth tended to be negative in the decades 1880-1930 and positive in the decades 1930-70, with the exception of 1940-50. However, over the total period (1880-1970) population cum labor growth on average tended to make a positive contribution to per capita income growth. The research of Ogawa-Suits covers the period of 1885-1920, the one of Minami-Ono includes the period of 1930's which has a large positive population effect. This would be one of the reasons why we get these two opposite opinions for Japanese economic development. Another reason why Minami-Ono and Kelley-Williamson have such an optimistic view may come from the constant labor participation rates (i. e., they treat population and labor as identical).
We set up the hypothesis that the changes of mortality are progressing with the changes in socio-cultural phenomena. The grouping of socio-cultural indices by sex from the 1920's to 1975 in Japan is used as the explanatory variable (X), and the stationary mortality rate by sex in the above years is employed as the dependent variable (Y). We set up the working hypothesis that a multiple correlation exists between X and Y. We employ correlation and multiple regression analysis, principal component and factor analysis and path analysis to prove this hypothesis. In the structure of component variables to explain the variances in Y, rate of second level school attendance, physicians rate and urbanization rate for male or female have greater contributions than the other variables. But examining the partial correlation coefficients, the indices having the most striking influences on Y for male or female are found to be related to those of education, economy and urbanization. The indices relevant to industrialization and health have less influence than the above, and have differences in the weight of influences between male and female. Our conclusion, after testing the working hypothesis, is that the changes of mortality by sex are influenced by the changes of socio-cultural factors. Thus, we can regard the above hypothesis as largely proved by the data of Japan in the above years.