This is the 5th in this year’s job market series.
As economies develop, average firm size rises. Large firms differ from smaller firms in many ways, but in particular, they use managers more intensively. Because large firms account for a larger share of employment in richer economies, those economies also use managers more intensively overall. What explains these patterns? One explanation is that growing demand for goods or technological progress makes it profitable for firms to operate at a larger scale, increasing their demand for managers to support firm growth. But causality may also run in the opposite direction: a larger supply of managers may itself drive firm growth and economic development.
In my job market paper, I use Japan’s forced repatriation after World War II as a natural experiment to examine how a larger supply of managers affects firm size and economic development. Many repatriates had worked as managers running Japanese organizations overseas. I find that inflows of managers increased firm size, helped firms expand into other districts, and shifted employment from agriculture and self-employment toward wage work. To quantify the aggregate effects, I build a model that accounts for firms’ geographic expansion. I find that the increase in manager supply raised aggregate output per worker by about 3%.
Japan’s postwar repatriation provides a natural experiment
To estimate the effects of greater manager supply, I compare Japanese districts that received larger and smaller inflows of workers with managerial experience. There are 1,113 districts in total. The empirical challenge is to isolate changes in quantities that stem from supply shifters rather than the demand-side mechanism.
Japan’s forced post-WWII repatriation provides a natural experiment to address this challenge. In 1940, the agriculture sector still accounted for 44% of Japan’s employment. At Japan’s surrender on August 15, 1945, more than three million Japanese civilians lived overseas, compared with about 70 million people in Japan. After the empire’s collapse, Allied repatriation policies forced them to return to Japan. The Allied occupation authorities limited repatriates to ¥1,000 per person, making repatriation primarily a labor supply shock. Figure 1 shows repatriates’ origins and numbers.
Historical map provided by the author and intended for illustrative purposes only and does not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.
I address the endogeneity issue by using predetermined family-registry locations as an instrument for realized repatriate inflows. These records identify households’ legal domiciles at Japan’s surrender, often in hometowns or ancestral places where they remained registered after moving abroad. My instrument follows the logic of immigration-network instruments, which use earlier settlement patterns of immigrants from the same country (Card, 2001). Family-registry locations predict repatriate inflows because people tend to return to places with family or historical ties.
It is possible that districts where more repatriates had their family registries were already on a different development path before the war. To check, I look at how employment shares in agriculture, manufacturing, and services changed from 1920 to 1940. These trends look similar across districts that received high and low inflows of managerial repatriates (Figure 2).
Repatriation brought a large pool of workers with experience relevant to management. Japan’s territorial expansion led to the establishment of Japanese firms and government institutions overseas, creating demand for people to run them. Within these organizations, Japanese employees were concentrated in managerial and supervisory positions. For example, managers at the South Manchuria Railway, one of Japan’s largest firms at the time, assigned employees to tasks and coordinated with other departments—familiar managerial responsibilities today. More than 60% of repatriate household heads had worked in occupations related to management, compared with about 14% of employed men in Japan proper in 1940.
Despite this occupational difference, repatriates and workers who had remained in Japan had similar schooling and family backgrounds. Average schooling was 9.3 and 9.0 years, respectively, and about 46% in each group came from farming families. These patterns suggest that repatriation disproportionately increased the supply of managers rather than a generic increase in the supply of skilled labor.
Do managers matter for firm growth and structural transformation?
To test whether greater manager supply helps firms expand, I examine changes in establishment numbers, average size, and sales per worker in services across 1,113 districts between 1936 and 1960. I use the 1936 tax records for prewar establishment counts and sales and the 1940 Population Census for employment. For 1960, I get all three from the 1960 Census of Commerce. I find that greater manager supply increased establishment size and productivity. An interquartile range (IQR) increase in managerial inflows led to 11% more establishments, 4% more workers per establishment, and 7% higher sales per worker by 1960. These patterns are consistent with firms finding it easier to hire managers to supervise larger workforces and coordinate more establishments. Managers may also help firms adopt better practices that raise productivity.
To test whether greater manager supply contributes to structural transformation, I examine changes in sectoral employment shares over 1940–1960 and wage- and self-employment shares over 1940–1970 across the same 1,113 districts. An IQR increase in managerial inflows reduced agriculture’s employment share by 2.7 percentage points, shifting employment mainly toward services. It reduced self-employment’s share by 0.7 percentage points and increased wage employment’s share by 2.0 percentage points. These shifts are consistent with expanding firms drawing workers from agriculture and self-employment into wage jobs, contributing to structural transformation. Figure 2 shows the sectoral effects over 1920–1970.