The Economics and Demography of Balanced Growth

During the period from the Meiji Restoration until World War I productivity in agriculture grew, as did productivity in the nonagricultural sector, and while growth in manufacturing outstripped growth in farming and fishing, differentials in growth were not large. At the time of the Meiji Restoration a common labor pool moved back and forth on a regular basis between farming and by-employment, including small-scale light industry, and the forces of supply and demand tended to equalize earnings and marginal labor productivity in the agricultural and nonagricultural sectors. Because differential labor productivity growth after the 1870s did not unduly favor the nonagricultural sector and because the vast majority of the newly created industrial jobs were in textiles and food processing, which made extensive use of short-term female labor that did not require long periods of training, employers had little incentive to offer premium wages to keep workers who by dint of job experience were especially skilled. As a result, marginal productivity of labor in agriculture set a wage floor for manufacturing and there was almost no divergence in earnings between manufacturing and agriculture.^[1]

One of the reasons growth was balanced was because productivity of farmworkers increased as population quality improved. Consider the figures in panel A of table 16. While the number of farm families remained virtually unchanged throughout the 1880-1920 period, the number of workers per farm family gradually declined, mainly because the number of female workers per farm household was dropping. For example, the ratio of female farm household workers to male farm household workers, always less than 1.00 throughout the period, declines somewhat over the four decades. Now one reason why farm households could increasingly afford to release workers for dekasegi employment in factories is apparent from the figures on workdays per worker and on labor productivity: both days worked and productivity per workday go up. As each worker became potentially more efficient, the household was better able to "lend" some of its members to industry without much sacrifice in output. With the steady improvement in

population quality and with the lengthening in number of years mandated under the compulsory education requirement, the physical and mental capacity of those who entered into agricultural pursuits was enhanced. Moreover, as can be seen from the figures on fertilizer and machinery inputs, the worker also benefited from more fixed (machinery) and variable (fertilizer) capital. In this context it is useful to keep in mind that increasing the application of fertilizer not only improves the inherent quality of the soil but also, by curtailing the amount of time devoted to weeding, reduces labor requirements (see Brandt 1993). Indeed as farm children, who often did the weeding, were withdrawn from regular work to attend school, presumably the demand for fertilizers increased. In short, the secular improvement in population quality commenced during the late Tokugawa period and continued into the Meiji period as traditional best-practice techniques (the so-called rono gijutsu , technology of veteran farmers) diffused throughout the country under the active promotion of the government. This diffusion raised per capita agricultural output, and in its wake per capita gross nutrition also improved. And the effect of these trends was to free up workers for manufacturing, thereby encouraging the continuance of a balanced growth process begun during the late Tokugawa period.

Hence under conditions of balanced growth the agricultural sector and the legacy of techniques built up in the more productive fiefs made a major contribution to economic growth both in terms of generating output growth and in terms of elastically freeing up labor for manufacturing activities. Some indicators of the quickening pace of growth in nonagricultural activities and of its implications for urbanization and income growth can be seen in panel B of table 16. Note that the proportion of the population living in the six big cities where industrial production tended to be concentrated—five of the six big cities were either in the Tokyo (formerly Edo) or Osaka region or lay along the Tokkaido where by-employments had most vigorously flourished during the late Tokugawa period—grows rapidly and the proportion in primary industry declines. With the shift out of agriculture into the manufacturing sector, which enjoyed labor productivity growth rates somewhat in excess of those experienced by agriculture, real income and consumption per capita grew slowly but steadily (see panel B of table 16). It is important to keep in mind that this growth in manufacturing production during the 1880-1920 era was mainly restricted to light industry and that in light industry female workers predominated. Consider the figures on percentage of hours in cotton spinning supplied

by females. As can be seen from panel B over 75 percent of the hours were supplied by females and this proportion actually grew during the period 1890-1900. Moreover, wages for the factory girls actually fell short of wages paid females in farming; being a seasonal activity, farming could not provide the regularity of work offered by many textile enterprises.^[2]

What about male workers for whom—as panel B of table 16 demonstrates—manufacturing wages usually exceeded agricultural wages, albeit not by a large margin? Why does the wage pattern vary between the sexes? The critical differences were the skill levels required and the nature of the labor contract. Male employment was largely concentrated in government-managed military arsenals, mines, and a few heavy industrial facilities (some of which were later sold off to the private sector) and in a relatively small number of privately operated heavy industrial concerns that manufactured ships and the like. During the Meiji period when labor capable of employing the techniques necessary for using Western machinery was in short supply, a typical pattern was for these factories to contract out work with labor bosses who brought in with them to the shop floor their crews of subordinate workers. The labor boss trained these subordinate workers and allocated wages to them from the overall payment negotiated between him (it was always a male) and the firm. In fact, the labor boss system was a carryover from the Tokugawa period and is one more example of balkanized entitlement insurance as the labor boss "took care" of the basic subsistence needs of his subordinates in good times or bad. Fearing that the labor bosses could and would use their oligopoly position to extract excessively high rents out of the enterprise, corporations attempted to establish direct control over their workers, replacing the free-floating labor boss with the co-opted labor boss or internally promoted foreman whose fate was more closely linked to the company. Increasingly, as heavy industry grew and established direct managerial control over the shop floor, special status within companies was accorded professional workers graduating from the higher educational system and co-opted labor bosses. These favored workers were called shain (literally company workers), and as an elite they earned high wages, wages the enterprises tied to seniority and age in order to reduce turnover. Corporate paternalism in Japan initially developed around this elite and was associated with wage payments that, in comparison with agriculture, were generous^[3]

The vast majority of male industrial workers, however, were not given shain status. Until the slowdown in growth of the nascent heavy

industrial sector, coupled with an expansion in the number of graduates from industrially oriented vocational schools during the 1920s, spelled the end of excess demand conditions for skilled labor, most highly trained blue-collar workers moved about with considerable frequency. Some moved with labor bosses and some moved on their own initiative. Consider, for instance, figures for the Shibaura and Ishikawajima engineering works in 1902 (Gordon 1985: 35):

Not only did blue-collar workers leave plants with frequency, employers often complained of poor work habits and lax work discipline among their ranks and as a result often resorted to firing these workers who were denied shain status and were known as koin , which implies hired worker/outsider status. For these workers wage levels were not especially generous, although the more skilled they were, the greater the premium. At the time the typical heavy industrial enterprise saw no incentive in offering a generous seniority-based wage package as a vehicle for eliciting loyalty and effort.

As it had during the late Tokugawa period, female dekasegi labor continued to be the backbone of the industrial labor force. In other words, despite the great advances being made in importing and adapting American and European technology in Japan, continuity in the labor market remained strong in the sense that high turnover and the short-term labor contract prevailed. What was new was the fact that large plants operating with central power sources like steam engines or with electricity began to supplant the small rural workshop of late Tokugawa. And what was new within the large confines of textile plants or in the depths of coal mines was the danger of industrial contamination and exposure to airborne infections that had not been prevalent in the small shops of the late Tokugawa period. In short, as the large enterprise employing the bulk of its workers on a short-term basis increased its share of the total labor force, the proportion of employers with an active incentive to protect the health and physical well-being of their workers declined. And at the same time the health of the industrial work environment deteriorated. Slowly but steadily the health-

enhancing nature of the Japanese market was breaking down. And this was taking place in an environment in which the government was reluctant to intervene for reasons to which we shall now turn.