2024 Volume 41 Pages 53-73
This paper explores workplace-residence relationships at retail stores in interwar Tokyo (then known as Tokyo City) by examining the specific case of the Ginza-dōri shopping street. Through its discussion, the paper reaches five conclusions. First, the shift toward working and residing in separate locations (shokujū-bunri), instead of working and residing in the same place (shokujū-icchi), had begun prior to the Great Kantō Earthquake but underwent significant changes thereafter. Second, shokujū-bunri—including approaches where people made their dwellings in the suburbs—presented advantages from the standpoint of home life, given the poor living conditions, concerns about health and safety, and high house-rent levels in the area along Ginza-dōri. The third conclusion is that shokujū-bunri was also advantageous in terms of land and building use, as having fewer residents in workplaces gave businesses the space to expand their sales-floor areas and high sales per tsubo (a measure of area equivalent to roughly 3.3 m2) offset the high land prices at the time. On the other hand, the paper also finds that shokujū-icchi enabled stores to stay open for longer hours and had its own advantages on the employment side, evidenced by how employees apparently retained a stronger preference for live-in working arrangements than their store proprietors did. The fifth and final conclusion centers on conceptions of “family”: the “merchant family” ideal, which saw employees living in stores as a “fun,” “lively” arrangement, led many to opt for the shokujū-icchi approach.
Research on Japanese commerce and distribution has long focused on community development. Driving that interest is a demand for practical insights that could inform community-development policy and approaches to issues in the contemporary context, including an ongoing push to energize present-day shopping streets (Watanabe 2014; Yahagi, Kawano and Mitsuhashi 2017; Ishihara and Watanabe 2018). While this paper operates from a similar perspective, it also seeks to help pave the way for a new field—the history of retail-based community development, one might call it—by engaging in a historical, empirical analysis of workplace-residence relationships at retail stores.
Past scholarship on the postwar Japanese retail industry has highlighted the importance of workplace-residence relationships (Ishii 1996). As the model of the “modern family” took stronger root among merchant families, the study found, the shift toward smaller households and the shokujū-bunri approach resulted in conditions where merchant businesses found it increasingly difficult to stay open for long hours, find successors, and maintain deep, well-connected relationships with their surrounding communities. These types of discussions lend historical context to the decline of shopping streets in the mid-1980s and thereafter, but they simply suggest that shokujū-bunri gained constant momentum from the 1960s onward without making an empirical investigation of actual workplace-residence relationships. While my own past research has presented a different perspective, using postwar housing statistics and other resources to underscore how the shokujū-icchi approach persisted among individual proprietorships (Mitsuzono 2018), there has long been a need for more research that examines the actual conditions—including those during the prewar period—from a concrete, historical foundation and probes further into the factors shaping the changes that occurred.
The Ginza-dōri shopping street, which underwent the most significant shokujū-bunri shift in prewar Tokyo, has drawn significant attention in that context. Geographically speaking, the Ginza-dōri shopping street refers to the commercial sector of the “Ginza Hacchō” area (which took formal shape via a change of the town name in 1930) comprising stores on both sides of “Ginza-dōri” (the portion of Chūō-dōri in the Ginza area). An organization overseeing the shopping street, the “Kyōshin rengōkai,” formed in 1919 and eventually renamed itself the “Ginza-dōri rengōkai” (Ginza-dōri Association) (its current name) in 1930 (Ginza-dōri rengōkai 2019). Past research has already touched on the workplace-residence relationships on the Ginza-dōri shopping street. For example, a study in the field of urban and architectural history showed that while the brick buildings on the “Ginza Bricktown” of the early Meiji period were largely used for housing with store space, the Ginza-dōri shopping street of the early Shōwa period was home to more store-only establishments than combined housing-store establishments (Hatsuda 2004). Using these previous findings as a foundation, this paper aims to delve further into workplace-residence relationships by examining trends in land prices and rent, addressing factors relating to retail management, and incorporating the element of the home-living environment.
This paper comprises five parts. Part II looks at the ways in which contemporary observers theorized shokujū-bunri in the retail industry and the logical foundations that validated shokujū-bunri as a favorable approach. In Part III, I use historical sources to illuminate the workplace-residence relationships on the Ginza-dōri shopping street, examine the historical changes therein, and place the street in context with its contemporaries. Going deeper into the case study, Parts IV, V, and VI investigate ownership and use of land and buildings, employment in retail management, and living environments, housing expenses, conceptions of the family, and other elements of the home-living dimension. Taking the contextual position of the Ginza-dōri shopping street, the paper aims to apply its findings toward a better understanding of the various conditions that defined the contemporary workplace-residence relationships.1
As I discussed in another paper, the shokujū-bunri theory, which considered shokujū-bunri the optimal approach, began to emerge in the commercial sector around 1910 (Mitsuzono 2019). The theory centered on four points. First, it held that shokujū-icchi arrangements were prone to confusion about the line between private and professional matters because there was little to distinguish the two dimensions and businesses sometimes used their employees for household errands. The theory’s second main argument was that shokujū-bunri enabled proprietors to separate home life from work operations, thereby making it easier to draw the line between private matters and public concerns and concentrate better on work. The third point had to do with living environments: proponents contended that living in the suburbs was better for store employees, considering that places outside the city offered better sanitation, higher air quality, quieter conditions, and lower land prices. Turning the focus to expenses, the fourth point held that securing living spaces for employees and their own families in commercial districts with high land prices was economically unfeasible for proprietors, which made shokujū-bunri a better approach than shokujū-icchi on financial grounds. Scholars have already looked at the second and third points in detail; one study argued that marking a clear division between home and work life dovetailed with the context of the “modern family,” reinforcing the association of “family togetherness” and a “happy home” in people’s private lives (Muta 1996), while researchers looking at the formation of suburban residential areas have cited the perception of favorable conditions outside the city as a key factor (Katagi, Fujitani, and Kadono 2000).
In the 1920s, the shokujū-bunri theory began to appear in discussions of retail trade along similar lines. For example, prominent prewar commercial journalist and consultant Shimizu Masami visited the United States in 1920 and, drawing on what he saw there, made the case for Japan to adopt the type of home-store separation that was the norm Stateside (Shimizu 1924). Shimizu also emphasized, however, that decisions on whether to isolate home life from work life needed to account for the rise in expenses that separation could cause (Shimizu 1938). The next part examines the conditions on the Ginza-dōri shopping street, keeping the above points from the contemporary discourse in mind.
Before I begin my analysis, I should first outline the history of the Ginza area to provide the necessary contextual basis for the discussion to follow.
After a major fire devastated Ginza in 1872, work began on building Ginza Bricktown: an area with fireproof urban infrastructure (Hatsuda 2004; Fujimori 2004; Ginza-dōri rengōkai 2009; Matsuyama 2019). The brickwork lining Ginza-dōri was complete by 1873, and blocks of brick houses had gone up throughout the area by 1877. Roadwork expanded Ginza-dōri from a width of 8 ken2 (approximately 14.5 m) to 15 ken (around 27.3 m), and a new mode of transportation came to the area in 1882 when Tokyo Basha Tetsudō (Tokyo horse-drawn carriage railway) launched its Shinbashi–Nihonbashi operations and brought its service to Ginza-dōri. In the 1900s, an electrification project also ushered in streetcars that eventually became part of the Tokyo Municipal Electric Bureau’s streetcar network in 1911. The Ginza area first began to bustle with activity in the mid-Meiji period. Contemporary engravings featuring stores highlight the goings on: works dating from 1883 to 1885 show Ginza-area stores selling Western items and other imports, Japanese aesthetics coloring the local brick architecture, and proprietors doing on-shelf merchandising (Hatsuda 2004). In the field of architectural history, one study cited the area’s Westernization, street configuration, on-shelf merchandising, and show windows as elements that made Ginza-dōri a prototype of the modern shopping street (Fujimori 2004).
The Great Kantō Earthquake on September 1, 1923, would prove to be a pivotal juncture in the area’s development. The earthquake and ensuing fires consumed 43% of Tokyo’s city area and ravaged Kyōbashi ward, home to the Ginza area, to devastating effect: over 90% of Kyōbashi’s total area burned (Suzuki 2004). Among the places that the flames destroyed was Ginza Bricktown, where some of the brick buildings had been converted into wood structures that some suspect may have exacerbated the spread of the fires (Ginza roku-chō-me chōkai 1983).
As the recovery initiative began, the Kyōshin rengōkai (established in 1919), an official organization responsible for overseeing the Ginza-dōri shopping street, rejected a plan that would temporarily relocate businesses to a non-affected area. Instead, the group decided to have the businesses still in operation remain in their original locations (Noguchi 1997). The idea was to complete work on two-story barracks in mid-October 1923 and have all the businesses kick off their big year-end sales as soon as November. The organization’s members also agreed to use an essentially uniform style for their stores’ roof signs. By mid-November, roughly 60% of the previously existing stores were back up and running; year-end sales began on November 15. An article in the Yomiuri Shimbun dated December 7, 1923, reports that 210 of the 255 stores along Ginza-dōri had successfully resumed operations.
In the aftermath of the Great Kanto Earthquake also came the first department stores on Ginza-dōri, with Matsuzakaya opening in 1924, Matsuya in 1925, and Mitsukoshi in 1930. Meanwhile, the Tokyo government’s post-disaster reconstruction program placed Ginza in “Zone 20” for land readjustment. On March 4, 1930, the government officially redrew the boundaries of the area and formally renamed it “Ginza Hacchō.” Twenty days later, on March 24, 1930, the freshly christened area played host to the “Capital Reconstruction Festival”—and from there, Ginza began its development into the bustling “modern Ginza” it would become.
2. Historical sources and methodologyIn the process of elucidating the workplace-residence relationships on Ginza-dōri, I compared maps and guides listing the names and proprietors of business locations against the Tokyo section of the Nihon shinshiroku [Who’s who in Japan] to determine proprietor addresses. The scope of my investigation covered the business locations on both sides of Ginza-dōri in the Ginza Hacchō area, or what amounts to the store locations comprising the Ginza-dōri shopping street. The maps and guides meeting the conditions for my analysis were from three years: 1902, 1922, and 1940.
My source material for 1902 was a map of the Ginza area, published around the time in question (Hirata Yūbido 1902), that indicates business locations along with the corresponding proprietor names and business types. For 1922, I used a pamphlet issued by the Kyōshin rengōkai (Kyōshin rengōkai 1922). Designed as a Ginza-dōri guide, the document features illustrations of buildings along the street together with their corresponding business names, proprietor names, and lines of business. The historical sources for 1940 were a series of maps in the Chuo City Kyōbashi Library collection (Narukawa 1940a, 1940b, 1940c, 1940d) that specify the businesses in the locations and their respective proprietors.
Although the limitations of dealing with historical sources meant that I had access to just three years of information for my comparative analysis, the sources that I was able to secure provide a window on changes over a relatively extensive span—including a data point that immediately precedes the Great Kantō Earthquake (1923). Still, the primary sources present a variety of complications. First, the maps and guides for the three different points in time are all different in nature. I was also unable to corroborate information for any proprietors who lived outside Tokyo at the time. The Nihon shinshiroku has its own limitations, as well; the publication lists only high-income taxpayers, namely those who had high income levels or ran relatively large operations, so my analysis could not cover proprietors outside those strata. It would be logical to assume that the vast majority of retailers at the time were small- and medium-sized businesses that failed to meet the listing criteria for the Nihon shinshiroku, but I was still able to cover a good number, and I assume the location and character of Ginza-dōri played into that outcome. These factors are important to consider in interpreting the results.
3. How workplace-residence relationships changed over timeThe next section outlines the results of my investigation.
In each of the three years, there were approximately 250 proprietors that fell within the scope of the analysis. The percentages of those proprietors whose names appeared in the Nihon shinshiroku came to 57.3% for 1902, 63.0% for 1922, and 40.5%—a particularly low rate in terms of analytical coverage—for 1940. The limitations of the extant historical sources and my methodology certainly played a part in that relative lack of coverage.
I then checked the entries in the Nihon shinshiroku, found the proprietors listed under different addresses from their business locations in the maps and guides, and divided the number of those proprietors by the total number of listings for each year. The resulting percentages were 14.2% in 1902 and 19.6% in 1922, after which the proportion climbed to 49.0% in 1940. These proportions of proprietors living at separate addresses from their places of business (which I call “separate-address rates”) points to how the shift toward shokujū-bunri picked up substantial pace after the Great Kantō Earthquake (1923).
A look at the separate-address rates by business type, which I defined in the relatively broad terms of “merchandising industry,” “restaurant industry,” and “other,” brings the contours of the shift into sharper relief. While the “other” category had the highest separate-address rates for all three years, the rates for the “merchandising industry” and “restaurant industry” rose on similar trajectories but at different levels: the numbers for merchandisers went from 11.1% in 1902 and 15.0% in 1922 to 41.8% in 1940, while restaurants saw an increase from 20.0% in 1902 and 25.0% in 1922 to 68.4% in 1940. At each point in time, then, the restaurant industry was home to a larger percentage of proprietors with separate address listings than the merchandising industry. The shopping street’s composition also shifted: the number of merchandising establishments had dwindled by 1940 amid growth in the number of restaurants, a factor that may have also driven up the overall separate-address rate.
Examining the listings with separate addresses in detail, one notices signs of two distinct shifts. First, existing proprietors that had formerly adhered to the shokujū-icchi approach were moving their dwellings to different locations. Second, new proprietors were replacing old ones and setting up store-only establishments without living spaces. Overall, the proprietors’ separate dwellings tended to be in the Kyōbashi area and elsewhere in Tokyo (the old city area), although some were in adjacent suburban locales (new city areas). Data on income tax and business income tax amounts suggest that in 1940, the higher tax payments a proprietor was making, the more likely they lived and worked in separate places. The listings also show that of the stores whose proprietors lived off store premises, corporate-operated stores far outnumbered privately owned stores.
4. The Ginza-dōri shopping street in a broader contextIn the 1930s, two surveys examined the Ginza-dōri shopping street and other major shopping streets in Tokyo (Tokyo Chamber of Commerce and Industry 1936; Tokyo City 1937). The investigation by the Tokyo Chamber of Commerce and Industry (the “TCCI survey”), which looked at 17 shopping streets as of December 10, 1935, was part of a concurrent nationwide survey of shopping streets. The other (the “TC survey”), which was the work of Tokyo City, focused on 16 shopping streets as of May/June 1936 (but September–November 1936 for transportation-related items). While it covered some of the same streets as the TCCI survey, the TC survey was an independent project on the part of the Tokyo City government and delved deeper into facilities, transportation, management, and other specific areas. Both surveys concentrated primarily on retail stores in their corresponding shopping streets, but neither included department stores, uniform-price chain stores, or retail markets in its definition of “retail stores.”3 For the sake of clarity and consistency, then, this paper uses the term “retail stores” to denote regular retail stores other than the three types that the historical surveys excluded. The Ginza-dōri shopping street appeared in the TCCI survey and the TC survey, as well, with both investigations demarcating the street as the stores on both sides of Ginza-dōri in the Ginza Hacchō area. The numbers of retail stores on the Ginza-dōri shopping street differed slightly between the two surveys: 161 in the TCCI Survey but 171 in the TC survey.
The information in the TCCI Survey makes it possible to determine the numbers of retail stores in Tokyo’s shopping streets by purpose of use. According to the figures, store-only establishments represented an average of 16.2% of all the retail stores across the 17 shopping streets that the survey investigated. The ratio was markedly high on the Ginza-dōri shopping street, where store-only establishments accounted for 53.4% of the total—making the street the only one where store-only establishments were a majority of the retail picture. What these findings suggest is a connection between store space and shokujū-bunri. In most of the shopping streets in the 1930s, retail stores normally occupied housing with store space. The Ginza-dōri shopping street was an exception to that norm, with store-only establishments outnumbering those in the standard mold. Shokujū-bunri apparently commanded a stronger hold in shopping streets where the average retail store was larger.
Using the details in the TC survey, meanwhile, one can work out the numbers of retail stores by management organization. The data indicates that of all 171 stores on the Ginza-dōri shopping street as of 1936, individual proprietors accounted for 54.4% of the total (93 stores), kabushiki kaisha (corporations) 26.9% (46 stores), gōshi kaisha (limited partnerships) 15.2% (26 stores), and gōmei kaisha (general partnerships) 3.5% (6 stores). The breakdown is noticeably different for the other 16 shopping streets in the study, however. Of all the 2,329 stores in those areas, 84.8% (1,976 stores) operated under individual proprietors, 5.2% (122 stores) were kabushiki kaisha, 7.5% (175 stores) were gōshi kaisha, 2.2% (52 stores) were gōmei kaisha, and 0.2% (4 stores) were anonymous partnerships. Corporate management thus had a much larger presence on the Ginza-dōri shopping street than it did on the other streets, suggesting that store size was not Ginza-dōri’s only distinguishing characteristic—it also significantly differed from its counterparts in terms of management organization.
The city-specific numbers of retail stores by purpose of use in 1935 show that shokujū-icchi was still the predominant workplace-residence relationship on Japan’s shopping streets during the mid-1930s. The percentage of retail stores that doubled as housing averaged 91.3% nationwide (across 87 cities), an extremely high rate that left the shokujū-bunri approach a stark minority.
As the above findings indicate, Ginza-dōri was the Tokyo shopping street where the shift to shokujū-bunri was most prominent—and that quality certainly makes it an exceptional case. If one incorporates the postwar years into the overall picture, however, the case of the Ginza-dōri shopping street presents a valuable lens for assessing the historical conditions that have defined workplace-residence relationships, since shokujū-bunri did gain a solid footing on shopping streets over the long run. Examining the conditions on the Ginza-dōri shopping street actually helps delineate the logic behind certain proprietors’ decisions to opt for the shokujū-icchi approach. My discussion thus far has shown that the shokujū-bunri shift, which gained substantial momentum after the Great Kantō Earthquake, was particularly prevalent among proprietors that paid large amounts of taxes (income tax and business income tax) and establishments that had incorporated, suggesting that a proprietor’s income bracket, store size, and management organization all factored into the shokujū-bunri dynamic. Building on that foundation, the next sections examine the other conditions in play.
To understand the implications of land and buildings in the bigger picture, one first needs to understand the relationship between how retail sales methods changed over time and what store locations meant in that context. Retail operations in Japan had their origins in mobile (traveling) sales. Historically speaking, it was not too long ago that the market saw the emergence of the first permanent store locations, where zauri— “seated sales,” in which customers would take a seat in a shop, chat with a worker about what they wanted, and have the worker fetch the products accordingly—remained the standard sales method for many years. Considering the popularity of zauri, store locations themselves were extremely valuable because they afforded proprietors important space for holding business talks and storing products in back storerooms. In the late Meiji period, methods involving on-shelf merchandising and show windows began to spread (Hatsuda 1999; Takayanagi 1994). Store locations became actual sales floors, creating a direct correlation between store size (building size) and sales-floor area—a factor that had an increasingly strong bearing on sales volume. Under those conditions, retailers would decide how to use their land and buildings from a management standpoint that took sales per tsubo and a variety of factors into account.
In the mid-1930s, 95% of the retail stores on the Ginza-dōri shopping street merchandised their products on store shelves, and 90.6% of all store locations had show windows. A vast majority had thus already implemented on-shelf merchandising and show windows, suggesting that the close, intertwining relationship linking sales-floor area and sales volume had solidified. In terms of land and buildings, meanwhile, the Ginza-dōri area had another set of distinctive characteristics. According to one source, it was “rather difficult to secure the spacious plots of land that large-scale buildings required, as the commercial lots in Ginza had long been small and compartmentalized” (Ginza roku-chō-me chōkai 1983). Not only did Ginza’s tightly packed layout put physical limits on how far buildings could expand horizontally, but the cozy confines also restricted the extent to which vertical expansion could increase an establishment’s floor area. With Ginza’s unique set of circumstances in mind, let us now look at the contemporary land and building ownership structure for a more thorough understanding.
To get a clearer picture of land ownership along Ginza-dōri, I pulled together the names of the owners of lots abutting Ginza-dōri as of 1932. Overall, corporations had a relatively low profile in the ownership picture: they were not necessarily the holders of the area’s biggest lots, nor were they congregated at the top of the rankings (Uchiyama mokei seizu sha 1932). Of the 100 individual owners in the area, 40 were Ginza residents, and 28 listed their owned land as part of their residential addresses. In total, there were 100 owners of 135 lots, measuring a combined 17,953 tsubo in area, for an average of 132.98 tsubo per lot and 179.53 tsubo of owned land per landowner. For comparison, a survey of high-rise buildings by the Tokyo government found that the 2,211 buildings subject to the study had an average building area of 194.16 tsubo. Ginza’s reputation for “small and compartmentalized” layouts was an accurate one, then, as the area per lot along Ginza-dōri proved to be relatively small (Tokyo City 1935).
From the proprietor standpoint, meanwhile, there were several ownership categories of land and building ownership. Buildings were either owned or rented, and owned buildings fell into two subsets; some were on land that was also owned by the proprietor, while others were on land that was leased from a different party. That makes it possible to separate proprietors into three basic modes of land and building ownership: those who owned buildings on their own land (land-owning building owners), those who owned buildings on leased land (land-leasing building owners), and those who rented buildings (building renters). Building renters had no choice but to use existing buildings. Building owners could technically rebuild at their own discretion, but those with buildings on land that they leased from particularly powerful or unaccommodating landowners likely had little room to pursue their own reconstruction projects.
I decided to put the proprietors on Ginza-dōri into these three land- and building-ownership categories using the following method. First, I used historical records from the Ginza-dōri Association to identify store properties as of June 1934 (Ginza-dōri rengōkai 1980). From there, I drew on resources in the Tokyo Metropolitan Archives collection, dated December 1934 (Kaneko Collection, 167), to determine whether the properties were owned or rented. The last step involved looking at land records for 1932, which allowed me to identify the landowners among my set of proprietors.4 While my methodology does have its shortcomings, given that my materials came from different years and did not ensure complete coverage of proprietor names, I was able to identify a total of 244 properties. According to my findings, 9.4% (23 properties) belonged to land-owning building owners, 54.5% (133 properties) belonged to land-leasing building owners, and 36.1% (88 properties) belonged to building renters. These figures all merit some skepticism, of course, because of the inconsistencies I mentioned above and the fact that I had to place “unknown” properties into a category to complete the process (which I did by categorizing them as “rented” buildings). That said, the results paint a general picture of the relative proportions.
According to past research on the ownership structure of land and buildings in Tokyo, as many as 70% to 80% of properties in cities were rentals; that tendency was particularly strong in Tokyo, where building owners and landowners were often separate entities and there was substantial differentiation in the land-leasing and property-leasing businesses (Katō 1988; Katō 1990). A 1931 survey of housing in Tokyo looked at 113,719 housing properties and found that 4.8% belonged to land-owning building (house) owners, 21.1% to land-leasing building (house) owners, and 70.5% to renters. It also gathered data on “housing with store space” (52,947 buildings), determining that 4.2% belonged to land-owning building (house) owners, 28.2% to land-leasing building (house) owners, and 63.1% to renters (Tokyo City 1931).
The building-ownership rates among Ginza-dōri proprietors were thus higher than those in other areas of Tokyo, but the percentage of proprietors who owned both their land and their buildings still sat below 10% of the whole. From these conditions, one can infer that for many proprietors—even those on the Ginza-dōri shopping street—the idea of rebuilding their properties as they saw fit was likely quite a challenge. The legal restrictions that proprietors had to grapple with in the first place were another frustrating factor. Two legislative changes with a sizable impact were the Urban Building Act and “Barracks Order” in the aftermath of the Great Kantō Earthquake, which I discuss in the next section.
2. The Urban Building Act and the “Barracks Order”The Japanese government promulgated the Urban Building Act in April 1919, issued a corresponding enforcement order in September 1920, and issued the ordinance for enforcement in November 1920. The Act designated Ginza as a “commercial area,” a classification that came with its own sets of structure-specific building restrictions, and a “class-A fire-protection district.” In a class-A fire-protection district, every building needed to have fire-resistant structures for its exterior walls and roof, non-combustible materials for any protrusions, fire doors, and, if it was a large-scale building, fire-resistant internal structures (floors, pillars, and stairs).
These were the conditions in place when the Great Kantō Earthquake struck in September 1923 and devastated significant portions of Tokyo, including Ginza and the surrounding Kyōbashi ward. In response, the government passed the so-called “Barracks Order” (promulgated and enforced on September 15, 1923) to regulate buildings in areas that the disaster had affected (Tanaka 2006).
One point to note here is that the deadlines for both starting construction on the barracks and demolishing the barracks were extended. Originally, construction was supposed to begin no later than the end of February 1924; that deadline was pushed back for areas subject to land readjustment to the corresponding readjustment date (the date on notice was given for approval). The deadline for demolition, originally August 1928, was postponed on several occasions for different building classifications. One such category was “special buildings” (theaters, assembly halls, ryokan [Japanese-style inns], factories, etc.), whose demolition deadline went from August 1928 to August 31, 1933, and later to August 31, 1936. For buildings in class-A fire-protection districts, postponements pushed the demolition deadline to August 31, 1938, and ultimately to August 31, 1947. Buildings outside these classifications saw the demolition provisions change substantially, as well. For those with certain structural facilities, specifically those that either met the “Stipulations for Ordinary Structures” in the Chapter 3 of the Ordinance for Enforcement of the Urban Building Act or barracks that were modified to meet the requirements (“certified barracks”), the demolition requirement was lifted indefinitely. For those without the requisite facilities, demolition was to be completed by August 31, 1933. The reason for the deadline extensions was to help alleviate the financial burden on residents of earthquake-stricken areas—the economic strain of building in compliance with the Urban Building Act.
An earlier study argued that while the “Barracks Order” helped residents in affected areas secure housing quickly, the switchover from barracks to permanent buildings was sluggish; estimates suggest that fewer than half the barracks were converted into buildings that met the requirements of the Urban Building Act (Tanaka 2006). Numerous barracks remained unchanged in class-A fire-protection districts, even, owing to the steep costs that fire-resistant buildings entailed.
There were still many barracks left in Ginza in the second half of the 1930s, too. Relating to this point, the data on the 225 total buildings (including buildings besides retail stores) on the Ginza-dōri shopping street in 1936 by floor count (exterior) shows that there were 6 buildings with one floor, 157 buildings with two floors, 39 buildings with three floors, 7 buildings with four floors, 7 buildings with five floors, 4 buildings with six floors, 2 buildings with seven floors, and 3 buildings with eight floors. Two-floor buildings accounted for 69.8% of the whole, three-floor buildings for 17.3%, and buildings with four or more floors for 10.2%. Breaking down the buildings by architectural style, one finds that Western-style buildings had a dominant presence: 217 buildings (96.4%) were Western, with just 8 (3.6%) in the Japanese style. In terms of structure, 35 buildings (15.6%) were fireproof, 165 (73.7%) were semi-fireproof, and 25 buildings (11.1%) were wooden. The percentages for all the buildings in Tokyo’s 16 shopping streets (3,001 buildings), on the other hand, reveal some key differences: 7.2% of the buildings had one floor, 79.5% had two floors, 11.7% had three floors, and 1.6% had four or more floors; Western-style and Japanese-style buildings represented 61.7% and 38.3% of the whole, respectively; and 7.4% were fireproof, 49.7% were semi-fireproof, and 43.0% were wooden. Generally, then, vertical expansion and fireproofing efforts appear to have made more progress on the Ginza-dōri shopping street than they had in the area’s counterparts in Tokyo. For a commercial area with a class-A fire-protection designation, however, the Ginza-dōri shopping street was evidently making insufficient progress in rebuilding structures in compliance with the Urban Building Act.
Factoring in the earlier discussion of the land- and building-ownership structure at the time and the slow process of replacing barracks with permanent buildings, one could conjecture that many proprietors had begun to assume that the barracks would remain and thus shift their focus to how they could make effective use of them. That brings the discussion to the next point: building usage.
3. Building usageOf the 171 retail stores on the Ginza-dōri shopping street as of 1936, 139 (81.3%) were on the first floor, 28 (16.3%) were on the second, 3 were on the third, and 1 was on the fourth. The 37 non-retail establishments in the area had a much different breakdown, with 0 on the first floor, 23 (62.2%) on the second floor, 6 (16.2%) on the third floor, and 8 (21.6%) on the fourth floor. Nearly all the retail stores on Tokyo’s 16 shopping streets—2,257 of 2,329 (96.9%)—were on the first floor, too, so it would clearly appear that retail stores almost always occupied ground-floor locations. That played a vital role in shaping the notion of a shopping street, as store activity on the street level adds bustle to the environment, enables window shopping, and helps stores pull in “floating customers.” From that standpoint, business performance at a retailer on a shopping street must have depended heavily on whether the store was on the first floor of the building or not. The value of floor area, too, likely varied from floor to floor for those same reasons.
For a given store in a given building, expanding the sales-floor area generally entails changing how the store uses its available space. One viable option for a location in the shokujū-icchi mold would be to convert living space into selling space. However, repurposing in that way becomes less feasible when the living space is on the second floor; an attempt to expand the selling space from the first floor (where the vast majority of sales floors were) up to the second floor would run the risk of creating a disconnect in retail operations and create concerns about reductions in sales-floor efficiency (sales volume per unit of area).
A good way to explore this dynamic in our case is to examine the relationship between total building area and sales-floor area on the Ginza-dōri shopping street. First, there were 135 lots totaling 17,953 tsubo in area along Ginza-dōri. Banks, department stores, and post offices accounted for 12 buildings on 15 lots (2,909 tsubo). Removing those properties from the group, one arrives at a total of 120 lots covering 15,044 tsubo. The total number of buildings lining Ginza-dōri, meanwhile, was 225 (in 1936), or 213 if one excludes the 12 banks, department stores, and post offices. That makes the ground-floor area per building 70.6 tsubo (15,044 tsubo ÷ 213 buildings). Assuming a maximum building coverage ratio of 80% for commercial areas, one can work the building area per building out to roughly 56.5 tsubo (70.6 tsubo x 0.8). The average sales-floor area per retail store on the Ginza-dōri shopping street was 35.6 tsubo, which means that the average sales floor was smaller than the average building area. Of course, retail requires backrooms and other facilities in addition to actual sales-floor space, so a complete understanding of the relationship between building area and retail functionality needs to account for the number of tsubo for business use in general. Although the two surveys of Tokyo shopping street do not provide any figures on that point, a sample study of retail stores in Tokyo indicates that the number of tsubo for business use was generally as high as 1.7 times the sales-floor area (Tokyo Chamber of Commerce and Industry 1937). Combined, these figures put the average number of tsubo for business use per store at 60.5 tsubo.
The estimates thus suggest that an average retailer’s building area (56.5 tsubo) would have been big enough to hold the sales-floor area (35.6 tsubo) but not the business-use area (60.5 tsubo). In other words, a shop’s sales-floor area would fit comfortably on the first floor of the building, but the total area for business use would exceed the first-floor area—it would have stretched onto the second floor. However, even if many of the retail stores making up the Ginza-dōri shopping street operated out of two-story buildings, the data on building area suggests that proprietors still had room to add to their tsubo for business use—two floors of 56.5 tsubo (113 tsubo) were more than enough for 60.5 tsubo of business-use space—and further progress in the shift toward the shokujū-bunri approach would have enabled even more sales-floor expansion into the available space. To understand how retail proprietors might have viewed those prospects, one needs to factor in elements like land prices and rent, sales-floor efficiency, and rent burden.
4. Changes in land prices and rentFor information on land prices and rent, one can turn to the 1935 survey on shopping streets in Tokyo. The results suggest that the greater the ratio of store-only establishments on a shopping street, the higher the land prices along the street tended to be (with the Kaminarimon area of Asakusa an exception). However, the 1935 survey does not paint a clear picture of the conditions on the Ginza-dōri shopping street. The following section helps bring the area’s land prices and rent into better focus.
Pulling together a variety of fragmentary historical sources, I was able to trace the basic outlines of how land prices in Ginza changed over time. Land prices appear to have surged up to around 1930, after which the levels stabilized and the fluctuations steadied. According to an article in the Tokyo Asahi Shimbun on November 30, 1907, the land prices that year had hovered between 40 and 130 yen; around 1930, Andō Kōsei notes, the prices ranged from 4,000 to 6,000 yen (Andō 1931).
One variable that the available information provides details on is residential rent. The data shows that Ginza had the highest residential rent prices in Kyōbashi ward from 1905 to 1935, with the values rising during the period from around 1915 to roughly 1925 before proceeding to skyrocket from 1928 to 1929. Likely major factors behind these shifts were the post-earthquake recovery effort, which went into full swing in the mid-1920s, and progress on the land readjustment project. Rent prices then plateaued from 1929 to 1932 and went into a decline through 1935. The important point to note here is that residential rent in Ginza surpassed the average prices in Tokyo through the end of the 1920s by growing margins: whereas the Ginza prices more than doubled or tripled the Tokyo averages in the late 1920s, the gap from 1929 onward was as big as six or seven times.
Next, a historical source that I discussed earlier (from the Kaneko Collection)5 sheds light on rent along Ginza-dōri. The source, a survey of rent values along the street, details property lease–start dates, frontage, total floor space (in tsubo), house rent, leasehold charges and security deposits, past lease prices, leasing parties, and landlords. The survey covered 102 houses for rent among the 240 to 250 total buildings on Ginza-dōri (about 40% of the buildings), meaning that it probably encompassed a sizable share of the street’s rental properties at the time.
According to the figures in the survey, the average monthly rent per building was 902.5 yen—a value that drops to 506.32 yen when one excludes the two department stores from the calculation. The department stores—Mitsukoshi and Matsuzakaya—were in the 4-chō-me and 6-chō-me blocks, respectively, so those two blocks have extremely high average rent. Minus their department stores, however, the blocks’ average rent falls closer to the composite mean: 769 yen for 4-chō-me and 502.91 yen for 6-chō-me. Comparatively, though, per-property rent on the Ginza-dōri shopping street was markedly higher than the values for other major shopping streets. Taking the available numbers for total floor space per building, one can also find the average rent per tsubo: the average building floor space was 108.25 with the department stores and 62.02 tsubo without, so dividing the average monthly rent for the area (506.32 yen) by the average floor space (62.02 tsubo) produces an average rent per tsubo of 8.96 yen. Other valuables to consider include leasehold charges and security deposits. On the Ginza-dōri shopping street, there were 33 properties with leasehold charges and 32 with security deposits. That means that roughly one-third of the rental properties in the street were subject to either leasehold charges or security deposits, and many properties collected one or the other instead of both. Among the properties that did collect these fees, the per-building average leasehold charges were 13,642 yen, and the per-building average security deposit was around 2,942 yen. Leasehold charges were anywhere from 3.4 to 200 times the property’s rent, while security deposits ranged between 1.17 and 36 times the rent price. In light of that enormous variability and inconsistency, individual circumstances must have played a significant role in determining how much owners charged for these fees or if they collected them to begin with.
For comparison with Ginza, one can look at rent for rental houses in Tokyo’s “new city areas”—the five neighboring counties (gun) of Ebara, Toyotama, Kita-toshima, Minami-adachi, and Minami-katsushika. Survey results for April 1932 show that the average rent per tsubo in these areas was 1.45 yen (17.19 yen in average monthly rent divided by 11.82 tsubo in average floor space per house; Tokyo Prefecture 1932). The average rent per tsubo was thus 6.17 times higher in Ginza, a price differential similar to that separating the overall residential rent in Ginza and Tokyo as a whole. Without a doubt, the striking pace at which the Ginza-dōri shopping street’s land prices and rent figures were growing relative to its surroundings was one reason shokujū-bunri took such a firm hold in the area.
5. Sales-floor efficiency and rent burdenFor a retailer, though, the absolute value of monthly rent was just one piece of the puzzle; what also mattered was the burden of that rent expense in the context of store operations. High land and rent prices in a commercial area normally mean that the area is good for business, so the retailers in the pricey Ginza-dōri shopping street likely expected comparably high revenues. This section explores that dynamic in more detail, examining the elements of sales-floor efficiency and rent burden for area stores.
Before we focus on Ginza-dōri specifically, we should look at sales-floor efficiency levels at retail stores in shopping streets across Tokyo. The information in the TC survey (one of the two surveys on shopping streets that I mentioned earlier) makes it possible to gather data on yearly sales and operating expenses and, with supplementary data from the TCCI survey (the other survey), offers a basis for discussing sales-floor efficiency. The scope of my investigation here covered Ginza and six other shopping streets,6 the only seven for which I could obtain all the necessary figures. First of all, the data shows that the Ginza-dōri shopping street had much higher levels of sales-floor area per store, employees per store, and sales per store than its counterparts. In terms of business profits, too, the Ginza-dōri averages were noticeably higher than the rest. My calculations for sales-floor efficiency, or sales per tsubo of sales-floor area, put the Ginza-dōri shopping street’s average at 7,776 yen—a level that towered over the likes of Kōenji (1,260 yen) at the bottom of the ranking and Shinjuku (3,727 yen) in second place.
Data on operating expenses offers a window on land rent and property rent, but the figures do not include the portions of housing with store space that served as living quarters. Also absent from the data are specific details on properties where the proprietor owned either the land or the building. Since the only values to work with are averages for the entire area—including self-owned retail stores—I decided to calculate the property rent on the assumption that every property was a rental. Below, I refer to this property rent as “rentα.”
After performing my rentα calculations, I found that the retail stores on the Ginza-dōri shopping street once again had the highest value at 5,687 yen. The next-highest was Shinjuku at 2,228 yen, and Kōenji had the lowest at 480 yen. However, the Ginza-dōri shopping street had the lowest ratio of rentα to sales: 2.1%, which sat 1% to 2% under those of other shopping streets. Even when one calculates business profits on the hypothesis that all the properties were rentals (“business profitsα”), the Ginza-dōri shopping street still has the highest business profitsα per store at 11,881 yen. The retail stores on the Ginza-dōri shopping street did exceed the averages in other streets in terms of rentα per tsubo of sales-floor area, but they far outpaced their counterparts in sales per tsubo and also came out on top in business profits per tsubo (business profitsα). All together, the per-store data and per-tsubo data make it clear that the rent burden for retail stores on the Ginza-dōri shopping street was relatively light.
From a management perspective, retail stores on the Ginza-dōri shopping street thus had the wherewithal to expand their sales floors thanks to their high sales-floor efficiency levels, which would have allowed them to absorb the steep rent payments characterizing the area. If that had been the case, the percentage of store-only establishments on the Ginza-dōri shopping street—54.4%—might even seem low. The business profitsα per tsubo in the area came to 27.81 yen per month, higher than the average monthly rent in the five counties bordering the “old city” (17.19 yen). In simpler terms, the numbers suggest that expanding the sales floor by 1 tsubo would have given a proprietor enough profits to rent a house in the suburbs and still have money left over. Renting a living space in Ginza was a substantial burden to take on in the first place, as the residential rent there was six to seven times the Tokyo average and the rent for rental houses per tsubo was six times higher than that in the suburbs. One would thus think that even more proprietors on the Ginza-dōri shopping street would have repurposed their stores’ living quarters for business use and moved elsewhere, further pushing the momentum behind the shokujū-bunri shift. While the Ginza-dōri shopping street may have been the Tokyo street with the strongest presence of shokujū-bunri, a question remains: why did the approach not entrench itself even further in the community? In the next part, I examine the conditions in play to help elucidate the reasons.
Two key foreseeable advantages of shokujū-icchi on the employment front were that it could make it easier for stores to stay open longer, seeing as there was no need to account for commute times, and that it could allow for flexibility in defining work hours and break times (or business tasks and domestic tasks). Shimizu Masami took a critical stance on both of those points, treating them as problems in Japanese commerce, and contrasted them with the “efficiency” benefits that shokujū-bunri could unlock (Shimizu 1924). However, retail is obviously a type of business where a proprietor typically has to wait for customers to come; a store cannot generate revenue unless its doors are open. The retail world is also subject to the press and slack of business, fluctuations that occur even during open hours. It is hard to tell, therefore, how much weight Shimizu’s notion of “efficiency” carries in the retail context. In fact, some Ginza-area proprietors did see the latitude for longer business hours as an advantage of shokujū-icchi (Kimura 1954; Ginme-kai 1995).
The average opening and closing times on the Ginza-dōri shopping street in 1935 were 8:30 a.m. and 10:30 p.m. in the summer months (April–November) and 9 a.m. and 10:30 p.m. in the winter (December–March). Scholars have found that retail stores in Tokyo and other major cities tended to close around 11 p.m. (Fukumori 2003), so the Ginza-dōri shopping street appears not to have been far off its peers. In terms of business hours, most of the Ginza-dōri stores had 14-hour days, while slightly under 30% stayed open for 15 hours or more.
2. The establishment and impact of the Shop ActIn 1938, the Japanese government passed the Shop Act in an effort to protect commercial workers (Ishihara 2017). The law stipulated that establishments in the merchandising industry were to close by 10 p.m. (or, in special cases, postpone closing to 11 p.m. at the latest) and that employees were to have at least one day off per month, legal provisions that aimed to place checks on long working hours among commercial employees by regulating stores’ closing times. In addition, stores employing at least 50 individuals on a full-time basis were subject to labor regulations that prohibited those under the age of 16 (and girls) from working over 11 hours per day, required proprietors to give employees a 30-minute break when they worked more than 6 hours, give employees a 1-hour break if they worked over 10 hours, and provide at least two days off per month to workers under the age of 16 (and girls).
The available historical resources provide some details on the employment dynamic on the Ginza-dōri shopping street, albeit in a limited scope. I was able to find just 20 retailers that had at least 31 employees, which means that the number of establishments employing 50 or more full-time was likely negligible. On the point of business hours, however, the Shop Act did have an impact: since the provisions on closing time (10 p.m. as a general rule) applied to the entire merchandising industry, all shopping streets had to at least ensure they were in compliance. The Department Store Act, which had already entered force in 1937, required department stores to close no later than 7 p.m. from April to November and no later than 6 p.m. the rest of the year (Kōkai keiei shidō kyōkai 1983). When the Shop Act took effect the following year, then, retail stores must have felt a crunch as the closing-time regulations reduced the number of hours they could pull in floating customers after department stores shut down for the day.
According to an article in the Tokyo Asahi Shimbun on June 19, 1937, the Ginza-dōri Association was lobbying against the Shop Act in the final stretches leading up to the law’s establishment. The opposition ultimately proved fruitless, however. In the law, there was an exception allowing “amusement quarters” to stay open until 11 p.m. The point of contention thus became the question of whether Ginza-dōri would be designated as “amusement quarters,” and the Ginza-dōri Association pushed hard to secure the exception—but, again, to no avail (Miyashita 2019). In the end, stores on Ginza-dōri had to shutter at 10 p.m. and ended up seeing “foot traffic gradually dwindle” after closing (Kinoshita 1939).
As far as I can tell, the Shop Act’s closing-time regulations had no particular effect on the workplace-residence relationships on the Ginza-dōri shopping street. Having to close by 10 p.m. did not necessarily prevent proprietors from maintaining their long business hours, considering that the law did not include any provisions that prescribed opening times, so retailers did not feel compelled to make a change. Through at least the 1930s, the shokujū-icchi approach was apparently a crucial condition for keeping stores open for as long as possible. An interesting point to mention here is that 79.5% of the 171 retail stores on the Ginza-dōri shopping street in 1936 had live-in employees, exceeding the percentage of store-only establishments in the area. While more and more owners and their families may have been adopting the shokujū-bunri approach and living outside their stores, that did not mean that employees necessarily followed suit. As a store expanded its operations, the owner would generally spend less time minding the store and gradually assume a more hands-off, supervisory role. When one brings the workplace-residence relationships among employees into the picture, however, it is easy to see how long business hours and the shokujū-icchi approach shared close ties. In the logic of employment, the shokujū-icchi approach was obviously an important condition that made it easier for retailers to stay open longer.
On the other hand, individual descriptive primary sources suggest that one factor behind decisions to adopt the shokujū-bunri approach was the home-living environment. The Yamazaki Clothing Store is an illustrative example. The store’s owner relocated his residence to the suburban area of Ikebukuro, where he also established a boarding house for his employees to live in and commute from. According to a Yomiuri Shimbun article on the move (dated October 14, 1916), the move stemmed from several serious drawbacks of the living environment around Ginza-dōri: train noise, lack of sufficient sunlight, and concerns about children’s safety and health, to name a few.
An earlier study on changes in the character of stores and offices in Ginza addressed the living-environment issue (Maeda and Hatsuda 2006). Focusing on the Ginza Hacchō and Ginza Nishi areas, the authors found the numbers of facilities in various function-related categories at four points in time: 1902, a point prior to the Great Kantō earthquake, a point before land readjustment, and 1935. In the results, the study found that the Ginza area took on an increasingly “large-scale” character while the number of “small-scale” facilities (neighborhood-oriented facilities) waned. Take the estimated number of “facilities selling food products” to neighborhood residents, for example, which dropped from 91 (at the first point in time) to 56, 24, and finally 19 (at the last point in time). Still, the number never hit zero; other neighborhood-oriented facilities like the Konparu-yu public bath (sentō), which opened in Ginza 8-chō-me in 1863, and the famous Taimei Elementary School, remained. Home living was not necessarily impossible, therefore.
2. Housing expenses and conceptions of the familyFor a closer, more detailed picture of actual home living in the Ginza-dōri area, one can turn to the case of Yonedaya Clothing Store/Shibata Woolen Cloth Store. The next section uses this case to help illustrate transformations in the contemporary workplace-residence relationships (Shibata 1992).
Yonedaya Clothing Store, a clothing retailer originally founded by Shibata Kōnosuke in 1882, opened its Ginza-dōri location in 1888. Shibata Woolen Cloth Store, meanwhile, came about in 1903, when it launched operations as Yonedaya Clothing Store’s woolen-cloth division. The two conjoining establishments had separate entrances but connected at the back, allowing people to go back and forth.
Shibata Kōnosuke’s successor was Shibata Takeharu (born in 1888), who would go on to become president of Yonedaya Clothing Store when it became a kabushiki kaisha in 1919. Takeharu married in 1914 and lived on the second floor of the store until around 1920. As his family grew, he began to find the store in Ginza inconveniently small and decided to move to a house for rent in Takanawa before soon relocating to a house on the outskirts of Tokyo in Kamata. The family’s house in Kamata, located on a 2,000-tsubo plot of land that Kōnosuke had purchased during World War I, was completed in April 1923. Kōnosuke’s house was a traditional, one-story, Japanese-style house at 846 Ōaza Kita-kamata, Kamata-machi, Ebara-gun, while Takeharu’s abode comprised both a Western-style building and a Japanese-style home.
For my discussion, the point to focus on here is a choice that Takeharu’s younger brother, Sannosuke, made. Sannosuke was born in 1897 and lived a shokujū-icchi life at the store in Ginza, taking care of Shibata Woolen Cloth Store. When Sannosuke got married in 1925, he asked his wife where they should live. While they could have lived in a house like Sannosuke’s older brother Takeharu, Sannosuke’s wife, Takako, said it would be “convenient and fun” to live at the store with the employees (Shibata 1992). “Convenient” likely refers to the operational advantages of shokujū-icchi—but even more interesting here is the idea of living at the store with the employees also being “fun.” Takako came from a family that ran a rice store in Kanda. She had grown up in a lively environment, surrounded by employees, which is apparently why she chose to live at the Ginza store.
The values of “fun” and “liveliness” for Sannosuke and Takako had deep roots in a specific conception of the family, one that stood in sharp contrast with the “modern family” ideal of a private sphere consisting of a loving couple and their children. For the purposes of this discussion, I will use the phrase “merchant family” for an idea of the family where cohabitation with employees is “fun” and “lively.” In the case of Sannosuke and his wife, this merchant-family conception was a major motivation in actively opting for shokujū-icchi.
Later, in 1933, the Yonedaya Building went up, replacing the barracks on the site with a permanent structure. Sannosuke’s family continued to live there until around 1948. The building, made of steel-reinforced concrete, had a ground-floor area of 63.46 tsubo, a building area of 51.7 tsubo, five stories above ground and one below, central heating, and an elevator. On the basement level were a warehouse, assembly hall, employee canteen, and storeroom. The first floor housed Shibata Woolen Cloth Store (on the north side) and Yonedaya Clothing Store (on the south side), with the two connected at the back. On the second floor, there was a merchandise room (with merchandise tables and shelves), a sample and clerical room, and a storeroom. The third floor had a large, hall-type room and the employee sleeping quarters (26 iron triple bunk beds), along with a washroom. The fourth floor served as the workshop and sleeping quarters for craftspeople (26 triple bunk beds), while the fifth floor was where Sannosuke’s family made their home; there was also a meeting room, kitchen, and a “maid” room (normally for three, measuring 4.5 jō, or around 7 m2) on the fifth floor. The rooftop had a lawn area and a sandbox, and it also served as a place for doing laundry.
The average rent in Ginza 2-chō-me in 1934 was 8.45 yen per tsubo. For a building with an area of 51.7 tsubo (the Yonedaya Building), the rent would come out to 436.65 yen per month per floor. In comparison, a Tokyo City survey in 1931 showed that the average house in Tokyo had 3 to 4 rooms with 10.5 to 20 tatami mats, and the rent for most 4-room houses was between 20 to 50 yen (Tokyo City 1931). An average house in the outlying community of Kamata (Kamata-machi, Ebara-gun), meanwhile, had a total floor space of 12.76 tsubo and a monthly average rent of 19.03 yen in 1932—1.49 yen per tsubo, in other words. Considering that the average rent levels in Ginza 2-chō-me and Kamata at roughly the same point in time come out to 8.45 yen per tsubo and 1.49 yen per tsubo, respectively, Sannosuke’s housing expenses in Ginza were quite high. Living in the suburbs would have given the family not only more space but also a much smaller financial burden.
As my discussion has shown, the shift to the shokujū-bunri approach had already begun on the Ginza-dōri shopping street before the Great Kantō Earthquake but then grew dramatically after the disaster as existing proprietors continued to adopt the arrangement in greater numbers and new store-only establishments emerged onto the local scene. By the first half of the 1930s, store-only establishments outnumbered shokujū-icchi stores. With shokujū-bunri becoming a major force in the area, the Ginza-dōri shopping street was thus an exceptional case among the major shopping streets in Tokyo.
Trends in land prices and rent figures also demonstrate the uniqueness of the Ginza-dōri case. Land and rent were both comparatively costly in Ginza during the interwar years, due in large part to the impact of the Great Kantō Earthquake and subsequent land readjustment. As contemporary observers extolled the theory of shokujū-bunri, high land prices and rent figures no doubt gave the approach even more momentum. In explaining the exceptional progress of shokujū-bunri on the Ginza-dōri shopping street during the interwar period, then, one can easily cite high land prices and rent values, as well as the sizes of the businesses involved.
However, as of the mid-1930s, slightly under half of the street was still employing the shokujū-icchi approach. There were many aspects in which Ginza was not an ideal place to live—be it safety, quiet, cleanliness, or health—and there was a clear disparity in housing expenses between Ginza and surrounding areas. From a management perspective, retailers could have absorbed the high Ginza land and rent prices in their business operations, and the levels of sales-floor efficiency and rent burden suggest that they would still have had considerable room to expand their sales floors. Shokujū-bunri should have made even more progress than it did, considering the virtues it was supposed to have.
This paper has shown that there were elements in the logic of employment and the conception of the family that inherently relativized the theory of shokujū-bunri. In terms of employment, shokujū-icchi had significant benefits in facilitating long business hours because of how dependent the retail business was on actual store operations—a condition that the logic of efficiency central to shokujū-bunri theory was unable to overcome. This is evident in the difference in the workplace-residence relationships between store owners and employees, with many employees maintaining their live-in arrangements even when the store owners adopted shokujū-bunri. Conceptions of “family” represent another key point. There was a value system that interpreted cohabitation with employees as “fun” and “lively,” which I have referred to as the “merchant family” ideal. This type of conception had its roots in a different place from the modern image of the family as an intimate sphere comprising a couple and their children. At the heart of the “merchant family” was a certain idea of happiness; what the shokujū-bunri theory envisioned was not the same concept. That brand of employment logic and family ideals was surely not unique to the Ginza-dōri shopping street, likely characterizing contemporary retailers in general. The history of workplace-residence relationships at retail stores should not be understood as a unilinear history of shokujū-bunri, therefore, but rather as a process of change where a variety of elements intersected.
1 Due to space limitations, I have decided to condense my discussion to the most pertinent points and omit tables, graphs, quotations of historical sources, etc. as necessary.
2 1 ken is equivalent to approximately 1.8 meters.
3 A “department store” is a large establishment offering a far-reaching product lineup; a “uniform-price chain store” is a chain store where all the locations sell products at the same prices; and a “retail market” is a public or private market where customers can buy daily necessities. All are categorized separately from independently operated “retail stores.”
4 Ginza 1 chō-me yori itaru Ginza 8 chō-me kashiya yachin kenri-kin shiki-kin chōsho [Survey on rent of rental houses in Ginza 1-chō-me to 8-chō-me, December 1934], Kaneko Bunko [Kaneko Collection] 167, Tokyo Metropolitan Archives (hereinafter cited as the “Kaneko Collection”).
5 Kaneko Collection, 167.
6 The seven shopping districts were Ginza (Kyōbashi ward), Ningyōchō (Nihonbashi ward), Shinjuku (Yotsuya ward), Kagurazaka (Ushigome ward), and Ueno-hirokōji (Shitaya ward) in the “old city,” along with the “new city” area of Dōgenzaka (Shibuya ward) and Kōenji (Suginami ward).
