Chapter 5
Enterprise, Community, and Human Growth in an Era Unbalanced Economic Growth, 1920-1940

A Changing Market Environment

After World War I, heavy industry, which hitherto had constituted a minuscule sector submerged by a sea of agricultural and light industrial enterprises, emerged as the driving force, the engine, of Japanese economic growth. As a result the era of balanced growth ended. And a new epoch characterized by wide and widening differentials in wages and labor productivity between the heavy industrial sector and the rest of the economy commenced.

The new market environment brought down once and for all the curtain on the system of Tokugawa entitlements. Why? Recall that Tokugawa entitlements had supplemented a set of market institutions that had automatically built in—through incentives—a considerable amount of health protection for employees. Indeed, in the case of the majority of workers in Tokugawa Japan the "employer" and the "employee" were really inseparable since the enterprise for which the latter worked was the family into which he or she was born. The entitlement system was therefore oriented to providing insurance for households that were also enterprises so that they could continue to stay in operation. Since most enterprises engaged in rice cultivation and rice was ultimately the basis for taxes and for stipends for the samurai, entitlements were focused around rice. Entitlements basically involved transfers of rice and were aimed at dealing with short-run fluctuations in supply and demand at the local level, with market price serving as a barometer for

this supply/demand balance or lack of balance. Most ikki protests, most conflicts between landlord and tenant, ultimately had their origins in local conflicts over the distribution of rice. Entitlements were balkanized and therefore conflicts over them were also local.

But the trends underlying and leading up to the new market environment made a mockery of the Tokugawa system of entitlements. First of all, with the growing breakdown within the labor market of the industrial sector of enhancement of population quality, the assumption that matters could be left to enterprises was no longer credible. And even more important was the decline in the proportion of the labor force working in agriculture, and within the farming sector itself the decline in farming income relative to total income. As long as domestic rice production expanded rapidly enough to satisfy a population that, increasingly, was engaged in pursuits other than producing rice, the government allowed the farmer to enjoy the benefits of a healthy expansion in demand in the form of favorable—favorable to farmers, that is—movements in terms of trade of agricultural products vis-à-vis manufactured products. But as best-practice Tokugawa agricultural techniques diffused throughout the countryside, growth in agricultural productivity began to falter. It became apparent to a government concentrating its focus on the building up of production of iron and steel, chemicals, and transportation equipment that relying on domestic sources for foodstuffs might hamper the speed with which the nation could turn toward nonagricultural pursuits. The Rice Riots of 1918 heralded the end of an agricultural policy aimed at solving the problem of food supply for urban industrial regions through technological progress in domestic agriculture. The government began to aggressively promote the importation of foodstuffs, especially rice, from its colonies, Korea and Taiwan. In short, the central government adopted a policy that inevitably depressed either the level or growth of domestic rice prices relative to manufacturing output price levels and growth rates.

The political tilt toward industrial districts created havoc with the traditional concept of entitlements for two very important reasons. It made clear that the government was now sufficiently satisfied with domestic conditions in farming—of the ability of agricultural villages to avert subsistence food crises—that it could turn the principal focus of its food policy away from promoting domestic productivity to guaranteeing to the growing legion of industrial workers that prices for basic food staples would remain fairly stable relative to nominal wages. In effect, the government was trying to solve the growing population quality

problem caused by the gradual breakdown in health-enhancing labor market institutions within industry with its rice import program. In weighing the needs of the industrial sector against those of the agricultural sector the central authorities came down on the side of industry, effectively transforming the rice price question from an agricultural into an industrial entitlement issue. Moreover, partial justification for this policy twist was the fact that agricultural crises occurred with diminishing frequency and ferocity. And a corollary of this was the fact that tenants benefited less and less from the entitlement insurance traditionally offered by landlords since the probability of harvest failure was steadily in decline. Thus the two mainstays of Tokugawa entitlement insurance, which had been falling into disuse even during the last decades of the balanced growth era, now ceased to exist altogether. The result was rural rebellion—limited perhaps, but rebellion nevertheless—as the rural sector, especially its poorer segments, began to aggressively voice its discontent with the attention shown to the industrial sector and the almost exclusive attention now being devoted to building up a new system of entitlements to cope with the social problems of manufacturing. The increasing policy of according wage premiums to male workers in heavy industry served to further fan the flames of rural discontent.

The wrenching changes experienced in the structure of labor markets can be seen in table 19. First consider wage differentials in panel C of that table. The premium enjoyed by male industrial workers both over male workers in agriculture and over female industrial employees sharply and decisively increases during the interwar period, especially during the 1930s. In response to the changes in relative wage ratios for males and females in industry and in agriculture, which in turn reflects the surge in demand for males in heavy industry, farm families began to change the sexual composition of their labor forces. This is apparent in panel A of the table: the number of male workers declines with some fluctuation until the mid-1930s and then precipitously drops thereafter as the number of female workers increases, most dramatically after the mid-1930s. Female workers, who constituted the backbone of the low-wage/low-skill light industrial labor force—and increasingly this was the case as a glance at the percentages of cotton spinning hours worked by females confirms—were in a sector of manufacturing that was declining in importance relative to high-wage/high-skill heavy industry, which was becoming the new growth engine. The expansion of heavy industry as represented by machinery production and the contraction in textile production is clearly seen in the bottom right-hand portion of

panel C. In short, a new labor market structure was emerging and an old labor market that had much in common with that of late feudal Japan was dying.

But drastic structural change was hardly limited to labor markets. It was also apparent in product and in capital and land markets. Consider the sharp fluctuations in the terms of trade between agriculture and manufacturing: during the late 1920s the rice price and the price level of a general bundle of foodstuffs sinks; and then it soars during the late 1930s. The uncertainty of getting an acceptable harvest had now been replaced by the uncertainty of the income one could secure from that harvest. Among the industrial products declining in price relative to agricultural prices is the output of the chemical fertilizer industry. Note from panel A of table 19 that the prices of nitrogen-based fertilizer fall relative to land prices while wages rise—with some fluctuation—relative to land. A rise in the cost to farming families of labor relative to land brought on by a vigorous expansion in demand for labor within industrial production should encourage households to substitute factors for labor that are becoming, in relative terms, comparatively cheap. And trends of this sort are evident from panel A: fertilizer and machinery inputs are increased (the changing terms of trade encourage this after the early 1930s) and labor input in terms of number of workers is decreased. Because of a secular trend toward improvement in population quality, adult workers are better able to work and hence workdays per worker increases somewhat, but labor productivity increases even faster, each hour of work garnering more real output because of the quality improvements in the labor input and because labor now works with more fertilizer and with more machinery. In short, interwar agriculture was by no means incapable of productivity growth.

But while it was not incapable of generating growth, the returns to investments in agriculture were now, in the new market environment, far less potentially remunerative than those that one might secure from industrial investments. The rapid expansion of the six big cities evident from. panel D is testimony to the feverish activity centered around new heavy industrial enterprises in the major metropolitan conurbations. This is one reason wealthy landlord families—the very households that pioneered the diffusion of best-practice techniques and the construction of social overhead capital in the countryside like roads and drainage ditches—now turned away from rural pursuits and began to invest funds and time and energy in the newly developing industrial sector. Hence in the new market environment landlords had little, if anything,

to offer since they were no longer spearheading the drive for improvements in the countryside; they were not even particularly interested in the problem of how to redress the growing imbalance in entitlements as they saw their fate now tied increasingly to investments in urban industrial enterprises, and moreover their entitlement insurance was hardly needed. Is it surprising, therefore, that their tenants now turned against them?

Continuity and Change in Regional Differentials in Population Quality

Looming over the striking restructuring of Japan's labor market was a growing military crisis; indeed, the military crisis overshadowed and conditioned the economic transformation and played an important role in shaping the proposals for a centralized entitlement program to replace the defunct Tokugawa scheme. The deepening military crisis that saw Japan expand its de facto occupation of the mainland of Asia from its colonial base in Korea to control over Manchuria and portions of China and that brought it step by step into conflict with its great Pacific rival, the United States, conditioned the economic transformation because it sped up the demand for heavy industrial output like chemicals and transportation equipment, for instance, battleships and armed vehicles and airplanes. And it conditioned the economic transformation because it increasingly drew the government into a imperialist strategy that presumed that the colonized areas of Asia would serve as suppliers of foodstuffs like rice and raw materials for the industrial heartland in Japan. And it conditioned the demand for a centralized program of entitlements because the growth in military activity increased the demand for healthy soldiers; that is, it increased the demand for population quality. Hence those sectors of the rural economy concerned about the growing market-driven disparities between urban income and opportunities and rural income and opportunities found a listening ear among the militarists who increasingly directed national affairs as Japan entered the 1930s. And insofar as it was perceived that population quality depended on household investments, which in turn depended on the market opportunities and the entitlements available to households, the problem of redesigning entitlements appeared all the more critical to decision makers in Tokyo.

With the military needs of the central government in mind, let us consider geographic differentials in population quality as measured in

terms of height and weight on military recruitment examinations. After all, these were the figures that the armed forces assiduously collected in order to see which districts were supplying the best potential soldiers and which districts the worst. Had the military planners in the mid-1930s drawn a prefectural map of Japan and shaded in more darkly those districts with the higher proportions of short—stunted—recruits, they would have found the pattern displayed in map 3. As can be seen from a comparison of this map with map 2, many features of the 1918 pattern continued to persist in the interwar period. For instance, the regions with the smallest proportions of short examinees continued to be Hokkaido and the Kinki district, including Osaka, Hyogo, and Kyoto and the great cities of Osaka, Kobe, and Kyoto. But the availability of prefectural data for 1934 allows us to pick out some patterns that either did not exist in 1918 or are not evident in 1918 because of aggregation. For example, Tokyo prefecture, which consists of the ward areas of the city of Tokyo proper and its immediate semiurbanized and semirural suburbs, stands out as producing few short examinees within a northeastern region that, relative to southwestern Japan, tended to have a large proportion of short individuals as it had in 1918. Nagasaki prefecture on the southern island of Kyushu, which includes the city of Nagasaki, off of which is Dejima, the island occupied by the Dutch during the Tokugawa period, is also a supplier of mainly nonshort examinees. Now the map is primarily of interest because it gives us useful information about the nature of, and particularly the perpetuation of earlier, geographic differentials in population quality. For instance, it makes clear the superiority of the major urbanized industrial zones relative to the hinterland, especially the northeastern hinterland. But this map is also of interest because it shows us the nature of the social problem confronting Japan's military leaders: How were the areas lagging in population quality to be brought up to the standards set by the high-income industrial districts?

And it should have been clear to the military leaders that relying on the market alone would not help close the gap. Consider table 20. This table gives means and maxima and minima for groups of prefectures classified by proportion of the total labor force in primary industry averaged for the years 1920, 1930, and 1940 (PPI). Group A is the most agricultural and group D is the least agricultural. As can be seen, not only are the per capita income levels far higher in group D than in group A, but absolute increases in per capita income are greater in groups C and D than they are in groups A and B. Moreover, there is a large

Map 3.
Percentage of Males Examined for Military Service Who
Are Short (150 cm or Less) in Each Prefecture of Japan, 1934

disparity in the per capita index of public health and medicine (PHMEDI) for 1927—with group D prefectures far more favored than those in groups A and B—and the disparity also seems to grow larger over time (between 1927 and 1935). Again dependence on primary industry is positively associated with high levels of child/youth labor services. So disparities in all three of the factors that in chapter 2 we found were important in shaping trends in population quality—income (hence gross nutritional intake), public health and medicine, and child/youth labor services—are correlated with each other in the cross section. And as has already been discussed in the context of map 3, and as can be further seen in the right-hand columns of panel A and panel B of table 20, differentials in population quality mimic these differentials in net nutritional intake. Moreover, gains in height, weight, and BMI appear—not dramatically, but slightly—to be greater in the districts with higher levels of population quality. In short, the Japanese military authorities were faced with a serious problem of how to eliminate differentials in population quality between advanced industrial districts and the rural hinterland, especially in the Northeast. And the market was not eliminating these differentials. Indeed, because of the strong economies due to transportation costs, heavy industry was concentrating more and more in the great urban conurbations around Tokyo-Yokohama and Osaka-Kobe-Kyoto; hence if nothing was done in the policy arena, disparities in height and weight for military recruits were likely to increase, not decrease.

For these reasons Tokyo was becoming concerned about the lack of an adequate entitlement program in the countryside. And for this reason alone it is not surprising that the ministries in Tokyo began to consider or experiment with price supports for rice, a two-tier system of governmental purchase of rice under which marginal tenant farmers would receive higher prices than other farmers, a wholesale land reform that would do away with absentee landlordism, and some form of insurance for farm households administered through local rural cooperatives. But the interest of the central government in moving in the direction of erecting a new set of centrally administered and organized entitlements to replace the balkanized system inherited from the Tokugawa period is not simply due to long-run military concerns. It is also attributable to something far more immediate and pressing: the danger of wholesale rebellion in the countryside. With greater salience than any other single sequence of events, the outbreak of rural rebellion during the interwar period demonstrates the great importance of the voicing of

demand for entitlements in conditioning the historical development of population quality within twentieth-century Japan.

Tenancy, Paddy Production, Population Quality, and Agricultural Labor Productivity

Rural revolt in interwar Japan centered around confrontations, at times violent, between tenants and landlords. The institution of tenancy was already well established by the latter part of the Tokugawa period, especially in the commercially advanced Kinki district. However, it was not until 1871 that the government began to issue deeds guaranteeing ownership to the individuals assuming the responsibility for paying taxes on the land. And it was not until 1872 that the formal Tokugawa ban on land sales was actually abolished. In I873 the Land Tax Revision was promulgated, converting taxes from payments in rice to payments stipulated in monetary terms and based on the estimated value of land. Despite the conversion of taxes to monetized payments, however, under the typical arrangement between landlords and tenants rents continued to be paid in terms of a stipulated volume of rice or in terms of a proportion of the rice crop on the land worked under tenancy.

How prevalent was tenancy? Y. Nishida (1986) estimates that in 1874 30.6 percent of cultivable land was in tenancy and that the percentage rose thereafter, reaching 40 percent in 1887 and 45.4 percent in 1912. A similar-story is told by the following figures on the distribution of farm households according to landownership status for 1883-1884 and 1912 (PO/PT stands for part-owner/part-tenant):

In short, over the course of the balanced growth era tenancy grew. Perhaps this growth was due to the fact that the land tax was now collected in a monetized form, and therefore during the Matsukata Deflation of the 1880s when rice prices dropped, many marginal farm households were forced to sell their land. Or perhaps it was due to the fact that reclaiming land for cultivation was an attractive investment opportunity

during the era of balanced growth and therefore enterprising merchants and risk-taking farmers were encouraged to invest in the development of land that they could then rent out to tenants. In any case, because so much cultivated land was farmed under tenancy, by the time Japan entered into the era of unbalanced growth the potential for a major political struggle between landlords and tenants was well developed.

Beginning in the early Meiji period tenants here and there had taken collective action potentially directed at landlords: they formed tenant unions.^[1] The first tenant union was established in 1875, and the number gradually grew after that. However, after over forty years of organizing, in 1917, the total number of unions did not even reach two hundred. But then, all of sudden, the number of unions and their activity proliferated. As can be seen in panel B of table 19, landlord-tenant disputes increased at a fever pitch during the 1920s. Over the course of the interwar period both the quantity and the character of the disputes changed. During the early 1920s tenants appear to have been the main initiators of disputes as the vast number centered around demands for rent reductions. However, during the later 1920s and the 1930s landlords took the offensive in many areas and the proportion of disputes involving continuation of tenancy or compensation due to termination of a landlord-tenant contract soared. Hence it is clear that landlord-tenant conflict became especially pronounced precisely at the time when the balkanized Tokugawa entitlement system was collapsing and a centralized replacement emanating from Tokyo had not yet filled the vacuum. It is my view that the outbreak of rebellion in the countryside was not merely due to conflict between tenant and landlord but rather was the voicing of a demand on the part of villagers for a new system of entitlements to replace the defunct Tokugawa system. Because landlords had once been a source for traditional entitlement insurance, they often bore the brunt of the attack on the old system. But the rebellion in the countryside was only in part a rebellion designed to initiate land reform. It was only part of a much broader demand, namely, demand for a system of entitlements consistent with the realities of a new market environment.

The rural village, especially the poorer stratum of the village represented by the marginal owner and the smaller tenant, was being left behind: the new market was leaving it behind; the government was leaving it behind. That was the fear that fanned rural discontent. With the abolition of feudal classes, upward social mobility became a reality for many poorer but ambitious farmers. In the balanced growth era of the

Meiji period agriculture offered opportunities comparable to those offered by industry; hence the low-income hardworking tenant household, by dint of hard work and a detailed knowledge of farming techniques or by dint of the assistance of a landlord family interested in increasing its own rental income in tandem with its tenant's income, could advance economically and socially. But in the new market environment favoring industry—especially heavy industry—the small farmer and especially the small tenant farmer watched with growing anxiety as the relative agricultural wages and returns on investment fell.

The key point to keep in mind in understanding why rural rebellion exploded during the interwar period is that it was triggered by a sense that the government was abandoning the rural sector to the whims of an increasingly hostile and unpredictable market in which less prosperous farmers expected to experience relative losses of income and status. The decision to commence wholesale imports of rice from the colonies was a symbol of the apparent lack of interest by the central authorities in building up a new system of entitlements to replace those that in the late Tokugawa period had stabilized local rice markets in times of dearth. It is important to stress that relative income and status were at the heart of the matter, not absolute levels of income. For instance, consider the figures in table 21 on household membership and sex composition, relative income per household worker and relative agricultural income per household worker, and growth rates in real income per household worker and growth rates in real expenditure per household member. As can be seen from the figures that compare levels and growth rates for large and small owner households, large and small part-owner/part-tenant households, and large and small tenant households, income was growing in real terms throughout the 1930s for all six groups.^[2] But the gaps between the groups were also slightly growing in most, but not all, of the measures looked at here. In short, even with the terms of trade moving in a favorable direction, rural households, especially marginal tenant households, had a growing sense of insecurity: the tenant household did not enjoy security of land ownership and to this insecurity was added the growing insecurity of the rice market, which was now being affected by conditions of production in distant colonies like Korea and Taiwan.

Central to the concerns over a deterioration in relative incomes was a concern about the long-run—by long run I mean a period of several generations—economic and social status of the ie. For "feedback" of income into expenditures on education, food, and direct health maintenance,

which, by improving population quality, had the potential to enhance future income-generating potential for the members of the family, was important. This is the message of panel B of table 21, where I report on time series regressions for the 1930s of two forms. The first set of regressions gives for each type of farm households estimates from log-log regressions for the elasticities of per-household member expenditures on population quality-enhancing items—on food and drink, on education and health, and on the total of the two—with respect to real income per household worker (where the symbol RYPW here refers to real income adjusted for price changes, not to relative income as it does in panel A of table 21). As can be seen, the elasticities appear to be high. Moreover, as can be seen from the bottom part of panel B where I report on a second set of regressions for each type of farm household in which, controlling for the sex ratio, labor productivity in a given year is regressed against per household member expenditures on food, education, and health and per work hour real expenditure on fertilizer in the previous year, the more a household put into building up its population quality and its fertilizer input, the better off it was likely to be in future years. In short, because of the feedback effect of investments in population quality on future labor productivity, any increase in income disparities was feared by rural households. And it must be remembered that marginal farmers were concerned about two types of income disparities: the gap between urban and rural households, which was growing; and the gap between large owner households and marginal producers, which was not shrinking. And since large owner households enjoyed incomes sufficient to allow them to educate their children in higher echelons of the educational system that opened the doors to employment opportunities in heavy industry, they could escape the "squeeze" arising from the new market environment. But in general tenant households could not expect to escape this squeeze and therefore they sought redress in the form of new government entitlement programs.

A key part of this argument rests on the notion that in relative terms tenant households were suffering a deterioration in population quality. That is to say, even though they were enjoying improvements due to the strong secular trend, their relative position was deteriorating. In regard to this point consider the findings reported in table 22. In panel A I report on cross-sectional regressions using logarithms of all variables for 1927, 1930, and 1935 (the observations used in both panel A and panel B regressions are for the thirty prefectures of Japan whose average level of proportion of labor force in primary industry for 1920, 1930, and

1940 exceeds or is equal to 55 percent). I regress various measures of height for military recruitment examinees (i.e., percentage tall, percentage small, and average height) and the body mass index against the index of public health and medicine per capita, average real income per capita, and the percentages of cultivable acres in the prefecture (multiplied by the PPI variable to adjust for the degree to which the prefectural population was actually engaged in farming) farmed in tenancy, and the percentage of the cultivable acres in the prefecture (again multiplied by PPI) devoted to paddy rice production. As can be seen, tenancy is associated with lower levels of population quality; and in the regressions the elasticities on the tenancy variable are considerable. Moreover, as can be seen from panel B, tenancy was associated with slightly diminished gains in heights between 1926 and 1937 (and with slightly elevated gains in weights and BMIs). In short, lower levels of income were associated with lower levels of net nutritional intake and lower levels of population quality.^[3]

It is also possible to use the estimates in panel B of table 22 to say something about long-run coevolution in rural Japan. Analysis of skeletal remains suggests that the population of Japan was on average taller before the Tokugawa period than during the Tokugawa period. G. Honda and T. Shay (1994: 13) give the following estimates on mean heights (in cm) for males and females in different periods of Japanese history (it goes without saying that the sample sizes on which the estimates are based are very small):

Could the diffusion of rice cultivation during the Tokugawa period have led to a diminution of height due to some long-run coevolutionary adjustment, say, to diet or to working in paddy rice fields? The estimates in table 22—albeit for the twentieth century and, therefore, perhaps of limited relevance for the preindustrial period—do not lend credence to this thesis.

To summarize: after the government opened up Japan to rice imports following the Rice Riots of 1918, marginal farmers and tenant farmers, concerned about a potential decline in relative income and greater instability in income-generating opportunities, brought on by the governmental tilt toward the industrial sector and by the new market conditions associated with unbalanced economic growth, suffered growing anxiety and frustration. This pent-up frustration exploded during the interwar period. And as had the ikki of the late Tokugawa period, it sent a signal to the central authorities in Tokyo to address the plight of the agricultural sector through the creation of a new entitlement system to replace the now-defunct Tokugawa entitlement system. In demanding changes the farmers were expressing concern about more than their own lives. They were also expressing concerns about their children and the long-run viability of their ie lines because of the feedback of labor productivity onto future population quality and future labor productivity that existed in rural Japan. Moreover, the population quality of districts characterized by high levels of tenancy was declining relative to that in the rest of rural Japan. And, since rural districts were declining in relative terms compared to industrial districts, the population quality of areas with high levels of tenancy was declining all the more relative to nonagricultural Japan.

Industrial Regions, Industrial Occupations, and Anthropometric Measures

In the face of the rural crisis, the government continued to pursue reforms aimed at dealing with the breakdown in health-enhancing labor market institutions within the industrial sector. A variety of factors explain the government's intense concentration on promoting reforms in the industrial sector: the pressure of the International Labor Organization, which Japan joined; the momentum gained by the reform-minded liberal wing of the bureaucracy from implementation of the Factory Act in 1916; the growing interest of employers and professional managers in health-enhancing programs because of their potential payoff in terms of improved worker morale and productivity, which reduced resistance to reforms in the business community; and the spread of socialist ideas among a small but growing labor movement in industrial districts.^[4] The Health Insurance Law passed the Diet in 1922, although it was not actually implemented until 1931, the same year a law providing aid to injured workers was passed. And in 1916 the Noshomusho (Ministry of

Agriculture and Commerce) issued directives listing a group of occupational diseases to which special attention was to be paid by factory and mine inspectors (however, silicosis for miners was not included in the original list and was not placed on the list under 1930). Perhaps the most important single factor in this trend was the decline in resistance to reforms in the business community, which seems to have emanated from two sources: from the evidence generated from a variety of industrial health studies that demonstrated that healthier factory workers were more productive; and from the fact that as employees successfully brought the labor boss system to an end, they found themselves increasingly devoting resources to the training of their blue-collar workers, and as their investment in these workers rose so did their interest in keeping them through a host of paternalistic programs.^[5] In short, the new market environment was encouraging a positive attitude on the part of the business community—especially large enterprises—to the creation of a new system of health-enhancing entitlement programs, especially those programs that linked entitlements to labor contracts, giving workers incentives to work hard and to remain with the company.

There was a convergence between market forces and governmental (and military) interest in enhancing population quality that was making the task of building up a new system of health-enhancing entitlements far easier in urban industrial districts than in agricultural regions. That the children of urban households tended to do less physical work than those in villages—the Factory Law of 1911 banned child labor—and that per capita expenditures on public health and medicine were far higher in urban areas than in less densely settled villages were certainly factors. That these environmental factors were playing a role can be seen from table 23, which reports on regressions for the sixteen nonagricultural prefectures of Japan. The dependent variables are, respectively, gains in average height, gains in average weight, and gains in the body mass index for recipients of military recruitment examinations between 1926 and 1937. In short, urban employers benefited from the environments they operated in.

But the improvement of conditions in industrial zones was not simply due to a better environment. As factory owners became more aware of the potential benefits to productivity of reducing pollution on shop floors and as factory inspectors pointed out substandard levels of ventilation and grime, and so forth, and assessed fines, industrial work conditions improved. An excellent test of the hypothesis that factory improvements were important is afforded by a comparison between the

relative height and weight of female textile workers around 1910 with the relative height, weight, and BMI of textile factory workers during the late 1930s. By "relative," I mean relative to students. Consider the figures in panel C of table 24 which we can usefully compare to those in table 18. In making the comparison it is important to keep in mind that female textile workers tended to be drawn from tenant farm households, that is, from households whose relative levels of population quality were in decline. And yet as can be seen from table 25 by simply taking arithmetic differences between the figures for the interwar period and for 1910, the gap in human growth measures between female students and factory girls was decisively closing. More nutritious meals in company cafeterias, regular periods of rest, improved ventilation, and more attentive medical staffs all made a contribution. By late interwar Japan, the large-scale factory—that covered by the Factory Act—was no longer an impediment to the convergence in population quality differentials as it had been during the era of balanced economic growth.

The other tabular material in table 24 is also of interest. The figures in panel A show that the BMI tends to be smallest at the extremes of the height spectrum and that it is largest somewhere in the middle of that spectrum. And the figures in panels C and D show that food intake was greatest for groups whose physical activity was the most demanding; but it also shows that in spite of the greater gross nutritional intake of those doing heavy physical work, heights and weights tended to be greater for those consuming less food and subjecting their bodies to lower levels of physical exertion. In short, these data underline the importance of demands placed on nutrient intake.

The Emergence of Government Welfare Policy in Interwar Japan

During the interwar period, as unbalanced economic growth transformed labor, capital, and product markets, the government found itself increasingly drawn into the problem of building a new system of entitlements to replace the old balkanized Tokugawa system fallen into disuse. Political pressure from workers and especially from tenant unions and marginal farmers played a fundamental role in bringing the government's attention to this issue. And by the 1920s the liberal wing of the national bureaucracy was committed to introducing a program of reform that was inspired by that already introduced in many Western European countries and in the United States. Had the military crisis not

grown over the course of the interwar period, that program, or at least key parts of it, might have been instituted. As it was, a new centralized health-enhancing entitlement system did emerge: the key features of that program were the Factory Act, health and disability insurance for industrial workers, and, in 1938, the creation of the National Health Insurance Law which aimed at creating, with central government assistance, a system of health insurance societies in local communities, especially in agricultural districts.

But the military crisis grew worse. And as it did so, the military found itself caught on the horns of a terrible dilemma. Military spending diverted scarce tax resources and bureaucratic attention and energy away from population quality-enhancing social welfare programs. And yet the military, perhaps more than any other central government agency, was concerned about the perpetuation and even widening of differentials in height and weight and body mass index.

After Japan's surrender, freed from the burden of making excessive military expenditures and from the ideological trammels of fascism imposed by the conservative wing of the bureaucracy (many members of which were purged after 1945), the liberal wing of the bureaucracy was able to decisively complete the program of reform envisioned in the 1920s and 1930s. With bewildering speed the government passed a tough labor standards law governing work conditions and labor contracts, sweeping land reform that did away with absentee landlordism, price supports for rice that guaranteed a minimum standard of living for marginal farm households comparable to that enjoyed by blue-collar households, and a comprehensive national health insurance law. By the early 1960s Japan had joined the ranks of the advanced industrial nations in terms of its system of national entitlement programs.

But passage of this sweeping program of reform, realizing the dreams of the liberal bureaucrats of the 1920s, had to wait until the late 1940s, the 1950s, and the early 1960s. It was hardly uppermost in the minds of the Japanese bureaucrats—it was certainly not uppermost in the minds of the military planners—during the late summer of 1941 as Japan plunged into full-scale war against its great Pacific rival, against the very country that had forcibly opened it up to international commerce during the 1850s, the United States.