When Technology Raises Fertility The Labour-Service Loop and the Phase Shifts of Fertility in Time and Space

Japan's fertility fell in two steps, in 1974 and after 1990, while its population first spread out and then concentrated. This paper asks whether one object can carry all three. The object is the length of the chain that returns a unit of local spending as local income. Technology moves that chain the same way in both phases; whether that raises or lowers fertility depends on how much chain is left. With a strong loop, technology raised the local wage bill faster than it cheapened the substitute; with a weak one, it replaced the person. The sign flips near a gain of 0.35, and the loop has lost 50.9 per cent of its 1947 value against the 39.8 per cent needed, though the two ranges nearly touch at the edge.The framework is five equations, closed in form so that each link can be checked on its own. Six links are taken to Japanese data, each with its record: labour saving, the Tokyo loop gain, two waves of concentration, the employment rate across 47 prefectures, the price of an additional room, and fertility. Manufacturing employment fell by about two fifths in both the ten fastest-growing and the ten fastest-shrinking prefectures, so the pull came from services, and manufacturing was a buffer. An additional room costs 0.544 in elasticity, three times that of average rent, which closes the fertility gradient: −0.071 predicted against −0.067 observed. The congestion threshold is crossed by 45 per cent: 0.0425 against 0.0614. Technology has no sign; the loop has one.