Built on the Hands of Women: Data, Automation, and Gender in West Germany's Financial Industry

Abstract

This article examines the history of data entry from the 1950s to the 1970s, when office automation required transforming information into computer-legible data. Often women's work, data entry was an essential step in electronic computing—one that automation proponents did not anticipate and historians have largely overlooked. Two West German cases, Allianz insurance company and Sparkassen savings banks, encountered issues when they computerized their offices. Allianz, pioneers of computerization in West Germany, unsuccessfully sought a technological solution for the data entry problem. In contrast, Sparkassen pursued a sociotechnical solution, delaying the introduction of computers until they developed technologies that circumvented the need for data entry in the long run. Before that, they employed women for data entry on a part-time basis. These cases show that routine manual data entry was—and continues to be—a crucial part of computerizing business operations.

Introduction

Debates about (big) data often revolve around issues of data management, preservation, and analysis, as well as privacy and security. Office automation, however, faced a different problem in its first three decades: data entry. Data accrued in offices in various forms—on time cards, invoices, letters, insurance policies, and account statements, and first had to be made computer-legible before a computer could process it at lightning speed.¹ Women performed most data entry, keying data onto punch cards. This article uncovers the work of these "unsung heroines" of automation, as one computer scientist called them: the countless data entry typists without whose tedious labor computer automation would have been impossible—and whose work has been left out of our histories of computing and technology.²

Manual data entry was a precursor to today's routine information labor, from Amazon's "Mechanical Turks" performing microjobs through online platforms to the human content moderators for social media sites like Facebook.³ While high-tech companies and technology proponents lead us to believe that automation and artificial intelligence technologies relieve workers from mundane tasks, data entry reveals the long history of human labor being moderated by and inscribed into electronic technologies. Manual labor has been interwoven with electronic information technologies for decades; it is not just a temporary fix until technology catches up with ever more complex tasks. This examination of changes in work processes provides a more complete picture of computerization by showing that it created more rather than less work, what kind of work, and for whom.

The emergence of electronic computing technologies accompanied the midcentury transition in the West from industrial to postindustrial societies, with more employees in the service sector than in industrial manufacturing. On shop floors, assumptions about the inevitability of technological progress influenced reactions to technological replacement, although automation raised conflicts of control between (male) workers and managers.⁴ In offices, however, automation technologies arrived after white, young, and unmarried or widowed women had become a widely accepted mainstay in lower-level routine positions within the white-collar workforce.⁵ Bridging the gap between historical studies of women's office work in the early twentieth century and sociological studies on the impacts of personal computers and word processing software on women's work in the 1980s, this article focuses on women's midcentury data entry work. It investigates shifting relations among technology, work, gender, and race at a time when, spurred by the women's and civil rights movements, married women stayed in the workforce and Black women moved into office work.

This article builds on the work of computer historians who have shown that, unlike many other technological areas, computing was initially open to women but turned masculine over time. Nathan Ensmenger and Thomas Haigh integrated labor issues into the study of computing by revealing the efforts of (mostly male) computer programmers and systems analysts to professionalize their field in the 1950s and 1960s.⁶ Following Jennifer Light's classic article on the ENIAC (Electronic Numerical Integrator and Computer—the first electronic, programmable, general-purpose digital computer, which was completed in 1945) women, leading edited volumes by Joanne Cohoon and William Aspray and by Thomas Misa brought gender questions to the field, investigating the historical reasons for women's dwindling participation in computing.⁷ More recently, Ensmenger, Mar Hicks, and Janet Abbate integrated labor and gender approaches by studying gender in the American and British programming profession. While Ensmenger and Hicks argue that professionalization and government labor classifications made computing a domain for men, Abbate reveals the niches, such as software service companies, that women carved out for themselves.⁸ This article looks at the work of (mostly women) keypunch operators. Doubly disadvantaged as women and as presumably unskilled employees, keypunch operators need to be integrated in the standard narrative of computer history.

Two cases of office automation highlight the situation in the West German financial industry: Allianz insurance company and Sparkassen savings banks. Both were financial companies with data processing at the core of their business operations. In addition, both operated in the capital-deprived postwar German economy and applied conservative family policies that shaped women's participation in the workforce by enforcing traditional gender roles. However, both approached their data entry problems in different ways. Allianz acquired the first IBM computer in Germany in 1956 and soon identified data entry as the bottleneck in computerization. To solve this problem, the company optimistically tried to find automatic scanners, but a technological solution remained elusive. Eventually, Allianz reorganized its data entry work to alleviate some of the pressures, turning to sociotechnical measures. By contrast, Sparkassen savings banks delayed the introduction of computers until the late 1960s. In the meantime, they developed paperless processes for money transfers that bypassed data entry altogether and hired a newly emerging interim part-time workforce of women for immediate data entry needs. Sparkassen thus pursued an interim sociotechnical solution before implementing technologies that eased data entry. It is still debatable whether differences in company size, geography, and local labor markets explain the dissimilar approaches to data entry. However, the two cases show that there was no single or "West German" solution to the data entry issue. Political, social, and economic factors—for example, ideas about capital investments, labor costs, and what kind of work, when, and how long was acceptable for women—shaped the organization of data entry work.

The Unexpected Data Entry Crisis

Automation proponents like consultant and writer John Diebold promised that automation technology would take over repetitive work, end "the subordination of the worker to the machine," and release the worker "for work permitting development of his inherent human capacities." Diebold suggested that those performing simple repetitive tasks could be trained to perform more engaging maintenance and repair jobs.⁹ Less known automation proponent Hunt Brown, whose prolific publications include loose-leaf handbooks on office automation, stated bluntly that automation would replace "drudgery, not people" and create more jobs for more employees who would "not be doing the same dull, drab, routine work" but would "have more interesting assignments."¹⁰ However, a closer look reveals the opposite: rather than relieving employees from routine work, the introduction of computers to offices created more of the same.

As the office managers who installed computers in mid-twentieth-century companies soon realized, information accrued in offices in various forms. One office manager noted in 1956 that office automation differed from factory automation because offices did not control what he called "raw materials." While factory managers could control in what standardized form materials arrived or automatically record information from certain instruments or processes, data arrived in offices in numerous formats, including letters, forms, notices, applications, claims, phone messages, and personal communications during meetings. These pieces of data differed both in content and in shape and size, some handwritten and others typed, and many on diverse types of paper. As the office manager optimistically noted, information needed to be "normed" for processing by a computer.¹¹ Just like standardized raw materials in factory automation, office information could be processed automatically once normed—or mostly punched on a punch card for reading into an electronic computer.

For the companies introducing computers, data entry—then commonly called keypunch operations—soon became critical. Ensmenger and others argue that the scarcity of qualified computer programmers caused the late 1960s "software crisis."¹² The industry was, however, also facing a scarcity of qualified keypunch operators, a situation that historians have largely ignored despite contemporaries' increasing concerns. Data entry was costly, requiring salary payments for large numbers of keypunch operators, machine rental for the equipment, and often scarce office space for the operations. Data entry was also a time-consuming process that delayed the fast processing of information by computers. Most importantly, data entry threatened the accuracy of computer operations since wrongly punched digits rippled through entire computer calculations.

While some office managers urgently sought psychological tests and other means to identify suitable keypunch operators, prominent computer scientist Carl Hammer tried to calm the heated rhetoric. He argued that the perception of data entry error rates was relative. He showed that if you counted the cards with mistakes in relation to the total number of cards, the error rate was as high as 17.5 percent for an exemplary job; however, if, for the same job, you counted the number of incorrect digits in relation to all the positions the operator entered, the error rate was as low as 0.3 percent. Hammer urged focusing more on "(wo)man-machine interfaces," including how changes in keyboard design or in the input document's layout affected error rates. Over the course of his article, Hammer developed sympathy for the operator, who turned from an "unknown keypunch operator" into "our girl"—now viewed more emphatically but designated (and denigrated) by her gender.¹³

Indeed, as Hammer's formulations implied, primarily women took on keypunch operator jobs. Operating a keypunch physically resembled typists' work. Historians of gender and office work have shown how an initially gender-neutral technology, the typewriter, became feminized when office work was divided in the late nineteenth century, relegating the new workforce of women to routine positions in the emerging typing pools.¹⁴ Like typing and other routine office work, keypunch operation was a highly monotonous task and easy to control. Supervisors could measure keypunch operators' output in inches of cards punched, and even if supervisors did not keep exact production counts, they could easily assess a keypunch operator's work by her output. To reduce the number of errors, cards were mechanically checked, providing another easy assessment of a keypunch operator's work quality. In other words, keypunch operations bore the typical characteristics of women's low-paid, routine office work.

Women keypunch operators' working conditions were often dreadful. The noisy keypunch machines created an unpleasant work environment where operators needed to concentrate over long stretches of time in order to work quickly and accurately. The constant noise and pressure turned the women operators into "nervous wrecks." Supervisors reported that "if you happen to speak to an operator while she is working, she will jump a mile," and "if you just tap one of them on the shoulder when she is working, she'll fly through the ceiling." Many women dreaded their keypunch assignments, feeling they were "working for the machine," and retaliated with high turnover and absentee rates.¹⁵ Computer automation thus did not eliminate routine tasks but created increasing volumes of physically taxing and repetitious work for women keypunch operators.

Casting Women as the Guardians of Family Values in Postwar West Germany

To better understand keypunch operations in Germany, we first need to look at the role of women and their work in the postwar years. In the young West German republic, a public debate on the "crisis of the family" ensued, partly funneled by demographic shifts that raised questions about men's and women's proper roles at work and at home. More men than women were among the millions of military or civilian fatalities during World War II or had been taken as prisoners of war. The population now had 7.3 million more women than men, with the most imbalanced gender ratio in the marriageable age group, between 20 and 40 years old. In 1946, there were 167 women aged 20–30 for every 100 men in the same age group, and 151 women aged 30–40 for every 100 men of that age. In 1950, single women headed one-third of all households.¹⁶ Many of these women had been the sole providers for their families during the war. They continued as such in the postwar years, coping with severe housing and food shortages—in 1946, Germany had eight million dwellings for fourteen million households, as well as food rations of about 1,146 calories per day per person in the British occupation zone. Germans worried about the "abundance of women" (Frauenüberschuss) that seemed to prevent women of marriageable and childbearing age from finding partners and, some feared, heightened the threat of promiscuity and adultery.¹⁷

In addition, men returning from war often had problems reintegrating into their families' lives, with the excesses of the Nazi regime having discredited male authority and women often unwilling to relegate their newfound independence. Divorce rates soared from 8.9 divorces per 10,000 persons in 1939 to 11.2 in 1946, peaking at 18.8 in 1948.¹⁸ Many Germans longed for the return to what they considered a normal family life of a wage-earning father, a homemaking mother, and two or three children. Similar trends occurred in the United States, where women returned to their roles as housewives and mothers in the postwar years.¹⁹

Representatives of all political parties in Germany passed legislation that promised to promote a return to normalcy and protect conservative family values. While the 1949 German Basic Law—the German constitution—for the first time recognized equal rights for women, it also granted special protection of marriage, motherhood, and families, thus tying women to their domestic roles. The ensuing debate on how to adapt the German Civil Law to women's new constitutional rights aimed to reconcile these contradictory values. Social, Liberal, and Christian Democrats agreed that husbands and wives carried equal responsibility for the financial security of the family, and they equated women's nonwage housework to men's wage-earning work. The legislation eventually passed in 1957 and encoded the "housewife marriage" (Hausfrauenehe). Women no longer needed their husband's approval to work, giving them, at least in principle, the right to engage in gainful employment; at the same time, women remained tied to their primary duties as housewives and mothers.²⁰ Other legislation similarly emphasized women's role as housewives.

Given rising Cold War tensions, the strong emphasis on the nuclear family in West German legislation can only be properly understood against the backdrop of family policies in East Germany. The East German government urged women to work outside the home, both to ease growing labor shortages and due to the political view that gender equality could only be achieved through integration in the workforce—a person's position at work determined their social standing. While historians point out that glass ceilings remained for women at work and that women's emancipation often stopped short of relations within families, women did receive full equal rights in 1949 in East Germany, and the 1950 Law for the Protection of Mothers and Children created social services for working mothers, expanded women's opportunities to work, and mandated equal wages and affirmative action programs.²¹ Conversely, in West Germany, the notion of the "family as a refuge" from outside influences (Fluchtburg Familie) became popular in the 1950s. It promoted the idea that families provided security against the intrusions of (Communist or National Socialist) totalitarian regimes and the degrading forces of individualism, materialism, and secularism.

Young Unmarried Women Doing Routine Office Work—before Computing

In the 1950s, West German corporations initially operated under conditions of economic scarcity. Office buildings had been destroyed during air raids on city centers, and many corporations focused first on rebuilding their office operations.²² In the cash-strapped postwar economy, companies lacked capital for expensive office machinery. Therefore, many companies—including Sparkassen and Allianz—relied on young, unmarried women for their routine office operations, continuing longer trends of eschewing mechanization for women's manual labor in data processing.²³ However, they followed wider social ideas about women's proper role.

Sparkassen occupied a unique position in the German banking sector, different from cooperative and large private banks such as Deutsche Bank and Dresdner Bank. Chartered by cities or counties and geographically limited to their charter municipalities, Sparkassen operated as institutions of public law on a not-for-profit basis. They were dubbed the "banks of everyone" because they primarily offered personal banking services.²⁴ These were high-volume, low-value transactions, suitable for rationalization to reduce costs, and in the 1920s, Sparkassen had acquired some calculating, accounting, and bookkeeping machines. After World War II, Sparkassen were charged with special short-term tasks, such as conducting the currency reform, settling bonds, and compensating savings account holders displaced from Eastern Europe, because, unlike large private banks, they were not discredited for having financed the Nazi regime.²⁵ These tasks were high-volume operations that required reorganization and temporary solutions, and Sparkassen relied on the division of (women's) manual labor to pursue them cost-efficiently.²⁶ But within a few years, Sparkassen again worked on rationalizing their operations.

The newly recreated national Sparkassen association supported these rationalization efforts by conducting and publishing comparisons between Sparkassen branches to give managers benchmarks and best-practice models. It recommended that staffing "all appropriate positions (50%)" with women and others with trained workers, mostly men, helped optimize performance rather than increase costs. The association propagated a gendered division of labor, reporting that Sparkassen separated "mostly mechanical work" and "continuously recurring schematic work"—performed by semiskilled "low-paid female" employees—from work that required "thinking, planning and decision making" and allowed for "more autonomy and responsibility and produced more satisfaction and pride in work," performed by professionally trained employees (mostly men). By 1955, the association noted with satisfaction that the increasing employment of "primarily young and especially female" employees allowed Sparkassen to reduce personnel costs in relation to overall assets despite salary increases, overtly advocating a gendered employment policy.²⁷

Throughout the 1950s, Sparkassen relied heavily on mostly young women to process routine transactions. As the overall workforce increased sevenfold in only eight years, from under 5,000 in 1950 to about 35,000 in 1958, the number of women employees also increased proportionally, from 29 percent in 1950 to 39 percent in 1955 to 48 percent a decade later.²⁸ Young women far outnumbered older women, and more than two-thirds of them were under thirty in 1958. Most of these women worked only for a limited time when they were young, whereas their male colleagues—evenly distributed over all age cohorts—pursued lifelong careers. Tightly organizing this young, low-paid women's work allowed Sparkassen to reduce operating costs in the 1950s.²⁹

Like Sparkassen, Allianz also employed young women for routine office work. Founded in 1890, Allianz was the largest insurance company in postwar Germany, with five million insurance contracts in 1950 more than doubling that over the following decade to thirteen million.³⁰ The company had rationalized its office operations in the late 1920s, decentralizing office tasks formerly all handled at company headquarters. In 1954, the company established a new business department, Betriebswirtschaftliche Abteilung (BWA), to coordinate, simplify, and plan business operations across Allianz's local branch offices, as well as oversee punch card operations. Heinz-Leo Müller-Lutz, a young, college-educated business administrator, headed the new department. With a colleague, he traveled to the United States to study insurance technology, and he used the U.S. model for his rationalization program at Allianz. Addressing women's routine work, Müller-Lutz's department introduced a new work- and space-saving filing system, acquired over 3,500 dictation machines, created a typing pool, and implemented other workspace and workflow improvements that almost doubled typists' performance from an average of 25,000 machine strokes in 1953 to almost 40,000 in 1963.³¹

The BWA department also aimed to increase women keypunch operators' productivity even before installing a computer. Keypunch work required varied skills, including good hand-eye coordination and consistent concentration, attention to detail, and the ability to learn and work under pressure. BWA aimed to quickly and easily identify suitable keypunch operators, for example, through psychological testing or analysis of physical attributes, rather than finding out which women were good at the work only months after hiring them.³² In addition, the department tried to improve keypunch operators' performance throughout the day. Müller-Lutz had observed that American companies gave their office staff additional midmorning and midafternoon breaks because studies showed that regular breaks every two hours improved performance over a longer period. Müller-Lutz and his team therefore suggested similar ten-minute breaks for keypunch operators. They addressed concerns that such extra breaks went beyond contract agreements with the labor unions and would set an undesirable precedent for other departments, arguing that the special strains on keypunch operators, who had to adapt to the rhythm of the machines, legitimated special treatment. In addition, Müller-Lutz suggested that Allianz provide a light lunch for staff—not the heavy, fatty food typically eaten in Germany—to avoid reduced performance on account of indigestion, and he advocated offering stimulating drinks such as tea, coffee, or Coca-Cola.³³

At both Sparkassen and Allianz, young women performed routine work for low wages, while the men managing them aimed to rationalize their work and further reduce costs. Employing women might at first glance seem to violate the conservative family values dominant in the young West German republic, yet a closer look shows that it did so in socially acceptable ways. Companies like Sparkassen and Allianz did not generally employ women of all ages, including housewives and mothers; they hired young women for several years before releasing them to take on their family duties. Companies thus followed the socially acceptable two-phase model of women working from the end of their schooling (if they had any education during the war years) until their marriage or motherhood. These short-term employees were ideal for monotonous tasks. Unable to pursue permanent careers and with a high built-in turnover rate, the young women presented a cheap workforce that could be relegated to routine work for several years.

How Data Entry Eluded Allianz's Quest for a Technological Solution

Allianz was the first German company to install an electronic computer, an IBM 650, in 1956. In doing so, Allianz closely followed the U.S. model, and it continually looked across the Atlantic for a technological solution to its data entry problem. Like automation proponents Diebold and Brown, Müller-Lutz and his colleagues expected that fewer keypunch operators would be needed once Allianz introduced computers. A few months after the first computer arrived, Müller-Lutz was already thinking ahead to the next computer generation. In a report on automation technology, he wrote that

the goal of automation was not to methodologically exclude as many persons from the production process or even from the whole economic life as possible. Rather, automating repetitive manual and now also mental tasks will free humans from the monotony of work. The number of unskilled and trained laborers will go down, and the number of qualified workers and technicians will go up.³⁴

Concretely, he expected computers to replace women keypunch operators, as well as clerks operating tabulating machines and maintaining the punch card files. Regarding public debates about technological unemployment from computer automation, he assured that all these operators and clerks would be able to transfer to other tasks without problems since they would be released slowly over a longer time period.

However, within years, Müller-Lutz and his colleagues identified keypunch operations as a major hurdle to office automation. The more tasks they transferred to the computer, the more information had to be made computer legible for processing. Indeed, when Müller-Lutz claimed that his team's work helped keep the size of the punch card department stable at about 350 employees, this claim concealed a demographic shift in that department: within a decade of Allianz acquiring its first computer, the proportion of keypunch operators in the punch card department had grown from under 50 percent to almost 70 percent in 1966.³⁵ The number of keypunch operators at Allianz had increased by about sixty throughout the decade, or roughly six per year. While Allianz managers had been confident in how to rationalize these keypunch operators' work back in 1955, over the following decade, they became increasingly exasperated over keypunch work.

Müller-Lutz continually complained that keypunch operations were costly because they required manual labor, needed space for over two hundred keypunch operators' work stations by the mid-1960s, and were prone to errors.³⁶ In the early 1960s, the labor aspect had become the BWA department's primary concern. BWA managers complained repeatedly that the fluctuation of keypunch operators was higher than in comparable positions, and that they had a higher absentee rate of about 15 percent.³⁷ The managers claimed that other companies had introduced punch card and computing equipment, increasing the demand for keypunch operators to such a degree that there were more open positions than candidates. The overall employment level in postwar Germany aggravated the situation. Due to the postwar economic boom (Wirtschaftswunder), Germany's unemployment rate dropped to 2 percent in 1960 and then hovered under 1 percent for the next six years. On average, only 66,000 women in West Germany were registered as seeking employment, reaching a low of 41,604 in 1965.³⁸ Unable to fill the keypunch positions, not to mention adequately control keypunch operators' work, BWA managers turned to technical solutions. What had caught Müller-Lutz's attention during his visit to the United States in 1957 were what he called "reading machines" that read or scanned information. He described two prototypes: the widely publicized ERMA (electronic recording machine—accounting) check reader, and an experimental reader for automatic mail sorting. Müller-Lutz estimated that it would take two to three years until the trial results were available.³⁹ By the early 1960s, BWA was pursuing reading technology in earnest, cooperating with IBM to test whether the latest IBM optical character reader was suitable for Allianz's operations.⁴⁰ Unfortunately, based on these experiments and the associated cost calculations, BWA recognized a dilemma: the volume of Allianz's operations was too low to make the use of an IBM reader economically viable.

IBM's automatic reader offered limited technical capabilities in that it could only read up to seventy numbers in a maximum of two rows per page. Outgoing money orders seemed suitable for an automatic reader because they came in large volumes, were typed, and contained only numerical data (the IBM 1418 was a numerical reader, whereas the slightly later IBM 1428 could also read letters). Unlike its usual cost-saving measures through rationalization and automation, BWA argued that introducing a reader would make sense even if the company just broke even, given the acute lack of keypunch operators.⁴¹ However, cost calculations showed that, to break even, for the machine rent of DM 11,000 per month, the reader would have to replace seven keypunch operators earning an average of DM 1,530 a month and together punching about 10,500 cards a day. Unfortunately, Allianz received only about 2,000 money orders a day—not a high enough volume to operate an IBM reader economically, even with the large volume of outgoing money orders.

Still, BWA set out to analyze which forms were transferred to what kind of punch cards and which forms an automatic reader could read. This analysis revealed that reorganizing the keypunch work would enable a 25 percent improvement in performance. This reorganization allowed BWA to continue its keypunch operations for two years without needing to employ more staff.⁴² BWA focused on reorganization until suitable reading technology became available.

In its ongoing commitment to addressing the data entry problem, BWA reorganized its working groups in 1964 so that the electronic experts who had devised Allianz's computer strategy were now working alongside punch card managers in a newly formed Working Group for Integrated Data Processing (Arbeitsgruppe für Integrierte Datenverarbeitung, AGID), recognizing that data input and output were central to automatic data processing.⁴³ BWA continued to collaborate with IBM on data entry solutions; it also collected information on other companies' experiences with automatic reading machines, discussed forming a reading machine pool with other companies for cost-effectiveness, and even considered switching from IBM for better data entry solutions—notable considering Allianz's close cooperation with IBM.⁴⁴

Data entry remained an important and intractable challenge, as comparison with other BWA activities reveals. In the 1960s, BWA successfully completed other complex projects, such as implementing the customized Electronic Integrated Allianz-System, transferring payroll to the computer, installing microfilm technology, and improving its typing service and performance statistics. Other problems withstood easy solutions, and BWA stopped working on them. For example, teleprocessing technology was proven technologically feasible in the United States, but in Germany, the postal system had a monopoly over data lines, and its high prices and low quality rendered teleprocessing uneconomical.⁴⁵ While BWA resolved or temporarily ceased pursuing these issues, the data entry problem—exacerbated by tense conditions in the German labor market—remained a primary concern.

In 1965, Müller-Lutz projected that economically feasible reading machines would be available between 1970 and 1972.⁴⁶ However, in 1974, Allianz ended up installing new data entry terminals. At these terminals, women—now called data typists rather than keypunch operators—entered data directly onto magnetic disks. The new terminals reduced the noise from the click-clacking mechanical punching machines, promised a 20 percent productivity increase through easier work, and reduced costs because magnetic disks could be used multiple times, ending the need to purchase and store punch cards. BWA estimated that the new data entry terminals would release twenty keypunch operators and help save between DM 750,000 and 950,000 in staff and material costs.⁴⁷ However, it was still manual operators, mostly women, who entered the data for the computer to process. Even in the early 1970s, BWA still could not come up with a technological solution for automatically entering data in computer-legible form, which it had been seeking for fifteen years; therefore, it instead relegated itself to sociotechnical fixes through reorganizing and streamlining data entry operations.

Sparkassen's Part-Time Women and Data Storage Exchange

Unlike Allianz, Sparkassen delayed the introduction of electronic computers for a decade. Throughout the 1950s, Sparkassen focused on work process reorganization rather than capital-intensive technologies in their pursuit of rationalization. The national Sparkassen association strongly cautioned avoiding the mistakes of the 1920s and 1930s, when banks had invested in various technologies that prevented cooperation within the banking system and soon became obsolete.⁴⁸ In 1958, the association noted with satisfaction that savings banks showed "appropriate restraint" toward costly automation technology in order to avoid investment mistakes.⁴⁹ Indeed, Sparkassen only switched to computing technology in the late 1960s, but it did so rapidly. In 1966, only 21 percent of Sparkassen banks used computers, and of them, half—mostly smaller branches—cooperated in computing centers. In just two years, more than 60 percent of all savings banks operated electronic computers—200 banks independently and 350 as part of fifty-nine computer center cooperatives.⁵⁰

While delaying the introduction of computers, Sparkassen urgently worked on reducing the need for data entry in the long term. They adopted types of money transfers for many transactions that avoided printing paper records and reading them again. For the remaining financial transactions, Sparkassen developed an optical reading method that promised to be cheaper and more flexible than other automatic readers on the market. In other words, Sparkassen focused on long-term solutions for the data entry problem by creating the organizational environment for these solutions and relying on an interim workforce for immediate data entry needs.

In the 1960s, Sparkassen focused on the transition to cashless salary payments, tying up their organizational and financial resources and delaying computerization. Until then, German employees usually received their pay at the end of the week in a cash envelope. Partly motivated by the transition to electronic office machines, more and more German companies shifted to cashless salary payments in the 1960s. Employees in America had received weekly paychecks since the 1920s, but German institutions opted against check payments because there were no legal consequences for writing uncovered checks and checks were not considered convenient or safe. Bank transfers had become the preferred type of cashless payment since the early 1900s, and Sparkassen also favored them for paying wages. This meant that thousands of salaried employees had to open checking accounts to receive their wages. Given Sparkassen's traditional role of providing banking services to underserved private customers, the majority of Germans opened accounts with a Sparkasse bank. Sparkassen expanded their branch network to serve this rapidly increasing customer base, almost tripling the number of checking accounts in under a decade, from seven million in 1961 to seventeen million private and corporate accounts in 1970.⁵¹ This dominance in private checking accounts positioned Sparkassen to shape cashless payment modalities.

Together with other banking associations, the Sparkassen association pushed for three types of cashless payments: bank transfers, standing orders, and debit notes—possibly because of their automation advantages. Bank transfers allow clients to transfer individual cashless payments from their bank account to a recipient's; standing orders are used for recurring payments of the same amount, such as monthly rent; and debit notes are for recurring payments of varying amounts, such as utility bills. Unlike checks, which are handed to recipients and then returned and routed through the banking system individually, these three types of cashless payments share an important feature: they do not require actions by three parties (the issuer, the recipient, and the bank or banks), but only two parties (the bank(s) and either the recipient or the issuer). This meant that transfers with common features—for example, all salary payments by one company, or all bills from a utility company—could be processed together in bulk transactions between the banks and client companies, without involving individual bank customers.

Sparkassen envisioned that these bulk transactions would be processed by exchanging electronic storage mediums between banks and large corporations or utility companies, thus eliminating the need for paper and data entry altogether. It is unclear whether and when Sparkassen pursued bank transfers for their data entry advantages. While the national association of private banks remarked as early as 1958 that bank transfers were "especially adequate for automatic processing because of the constantly increasing, massive volume of identical business transactions which have to be sorted according to specific aspects," the Sparkassen association initially talked only about the convenience of bank transfers for corporate and private customers.⁵² Not until the late 1960s did the national Sparkassen association acknowledge that bank transfers were "more rational" for banks and their corporate clients because they suited electronic data processing.⁵³ Still, the German banking associations had been planning the exchange of data storage media as an automation method as early as 1964.⁵⁴ Corporate customers would no longer deliver their information on paper, submitting instead data storage media to their banks, whose computers would read them directly, eliminating the need for data entry.

The remaining individual transactions were similar to check transactions in that they required manual data entry. The need to handle check payments individually had already exerted such pressure on the American banking system in the 1950s that within a year, U.S. banks agreed on a common standard for printing and automatically reading the routing and account numbers on checks through magnetic ink character recognition.⁵⁵ In Germany, the number of checks had also increased in the 1950s, but the pressure was nowhere near what it was in the United States. Sparkassen opposed magnetic ink character recognition and rejected the technological solution pioneered in America. They opted for optical character recognition, which was cheaper than magnetic ink character recognition because it did not require special ink; it also promised to recognize other information besides the preprinted routing and account numbers. However, optical character recognition was a more complicated technology that, Sparkassen knew, would take longer to develop.

To achieve their long-term planning for bulk-processing paperless monetary transactions and optical character recognition for automatic data entry, Sparkassen relied on a newly emerging interim workforce to solve their immediate data entry needs: part-time working women. Part-time employment promised to meet the demands of many social groups. It presumably met consumer demands, allowing housewives, alongside their housework and childcare, to earn additional income for consumer goods such as a television set, a couch, or a car. It provided banks with a willing yet low-paid workforce for drudgework without career opportunities and alleviated any worries about women suffering fatigue over a full workday. And for politicians, part-time work seemed to conveniently link conservative family values with addressing the national shortage of white-collar workers.

The demographic composition of the Sparkassen workforce in the mid- to late 1960s underlines this shift. In 1966, the percentage of women employees under age thirty dropped slightly from 75.5 to 75.1 percent, indicating that the age of the women working there was rising.⁵⁶ In addition, Sparkassen hired more part-time employees. The ratio of part-time employees among the new female hires grew over the next three years from one in ten to almost one in three: about 200 out of 2,000 in 1967; 500 out of 2,000 in 1968; over 1,000 out of almost 5,000 in 1969; and over 1,500 out of 4,800 in 1970.⁵⁷ Sparkassen thus hired increasingly more part-time employees while simultaneously shifting to electronic data processing technologies.

Tapping into the female workforce, Sparkassen took advantage of widely changing attitudes toward women's work in Germany. Despite the overwhelming emphasis on family values in postwar West Germany, more and more women joined the workforce in the 1950s. The ratio of women between fifteen and fifty-nine years old in salaried employment rose steadily, from 47 percent in 1950 to 52 percent in 1955, and then to 54 percent in 1960, when it leveled off for the next decade.⁵⁸ In addition, women's employment shifted from the domestic and agricultural sectors—considered more appropriate for women's disposition and their future roles as wives and mothers—to the service sector. With one in three working women joining the service sector in the 1950s—up from 5 percent in 1907—women's employment moved to the center of public attention.⁵⁹ This shift, as well as the expanding cohort of working women, may have fueled public concern about women's work. But as women's employment rates show, while the image of the full-time homemaking wife and mother may have been an ideal that public opinion demanded, it was not the reality even in the 1950s.

Indeed, studies show that German women coming of age in the postwar decade made career-oriented choices, eschewing routine positions with higher immediate payments for apprenticeship training with little compensation, often as office clerks or other typical women's occupations such as salesclerks, childcare providers, and hairstylists. In addition, more women alternated family and salaried work throughout their lives, not following the idealized two-phase model, often against their husband's will; the ability to return to their learned occupations eased the negotiations with spouses.⁶⁰ A late 1950s sociological study showed that many working women expressed shame about neglecting their families.⁶¹ Likely exacerbating these feelings were the public debates on "latchkey kids" (Schlüsselkinder), the supposedly neglected children of working mothers who had a key to let themselves into their homes after school and remained without adult supervision until their mothers returned from work. Consequently, working women felt out of step with the mainstream and were conflicted about trying to meet the demands of both work and family.

Mounting labor shortages in Germany in the late 1950s and early 1960s changed attitudes. Germans embraced part-time work. When unemployment levels dropped in the 1950s and German rearmament exacerbated the labor shortage, the German government recruited Italian and Turkish workers in the manufacturing and heavy industries. The immigrant workforce was perceived as lacking the linguistic requirements for clerical and other white-collar positions. The silent labor reserve of women, however, could fill these jobs on a part-time basis without abandoning their family obligations. Part-time employment tripled, rising from 7 percent in 1960 to 19 percent a decade later.⁶²

This newly emerging group of women working part-time perfectly met the needs of the data processing sector. Trained often in nonclerical fields, and therefore seen as unskilled, they were low-paid employees without opportunities for career advancement. However, "fresh and relaxed" each day, they achieved relatively high performance levels at monotonous jobs like keypunch operations or telephone switching, and even "if the warm-up period doubles when a single position is filled by two women, on the flipside, the fatigue time is also reduced, resulting in greater performance and less absenteeism."⁶³ Part-time office employment skyrocketed between 1964 and 1966, as small and medium-sized companies introduced punch card machines. In the words of one German historian, "the Hollerith operator, who entered the data into the computer's punch card system, became the most popular unskilled transitional occupation for female office workers in the 1960s."⁶⁴ Computer automation only increased the need for such data entry personnel, further accelerating the growth of the keypunch operator workforce.

Sparkassen relied fully on middle-aged women working part-time for their data entry needs in the late 1960s and early 1970s, once they had introduced computers but could not yet use paperless exchanges of storage mediums or optical character recognition technology. In these years, Sparkassen were increasingly concerned that more and more middle-aged women were taking on low-paid part-time positions. From 1970 to 1972, Sparkassen hired more than twice as many women as men, and in 1973 they even hired almost exclusively women, about 2,200, compared to only about 200 men. A quarter to a third of the newly hired women worked part-time, and the ratio of women employees under 20 years old dropped in favor of women aged 20–40.⁶⁵

By 1972, Sparkassen had devised a standard for data sets, was exchanging the first magnetic tapes with corporate customers, and optically read paper forms. Within a year, Sparkassen performed 17 percent of all their cashless payments through magnetic tape exchange and automatically read 300,000 checks and debit notes, and the proportion of their paperless transactions increased rapidly to 45 percent by 1979.⁶⁶ That year was the first time the total number of paperbound transactions was lower than the previous year. Cashless transactions almost doubled over the 1970s, while the workforce only increased by a third.⁶⁷ Sparkassen expanded the exchange of storage media to also include magnetic disks and magnetic cassette tapes. This allowed them to also exchange data with small-sized companies that issued hundreds of financial transactions at a time but had smaller computer installations without magnetic tape units. Sparkassen also began working with service centers that provided payroll and other financial services for companies.⁶⁸ By 1985, Sparkassen performed 58 percent of their cashless transactions through data storage exchange, and this rate hovered around 60 percent for the rest of the decade.⁶⁹

When Sparkassen introduced automation solutions in the early 1970s, the number of women they employed part-time plummeted. For the first time in years, the Sparkassen workforce remained stable, with only 495 new employees hired in 1974. Sparkassen focused in earnest on increasing their employees' qualifications, and the composition of the workforce changed radically.⁷⁰ New work processes contributed to these shifts, as well as growing unemployment in the 1970s due to the oil crisis and baby boomer cohorts entering the job market. Responding to calls from the federal government for corporations to alleviate youth unemployment, Sparkassen expanded their apprenticeships. By the late 1970s, they provided apprenticeships for their own benefit, not just to promise a job. In addition, thanks to the higher education system opening up to disadvantaged groups such as women and children from rural and working-class families, Sparkassen could choose from a larger applicant pool. They also took on as apprentices increasing numbers of high school graduates who had completed thirteen years of schooling, not the usual ten years, and passed the college-entry exam, Abitur. While almost one in three Sparkassen employees had no training in 1970, this ratio fell to only one in ten by 1987, indicating a radical shift in the composition of the Sparkassen workforce.⁷¹

Conclusion

Sparkassen and Allianz were among many companies in Germany creating the rapidly increasing need for keypunch operators in the 1960s. Wherever corporations installed computers, their managers soon faced a growing need for data entry. While Allianz urgently but unsuccessfully sought a technical solution for its data entry bottleneck, Sparkassen expressed concern but not the same urgency. Several factors contributed to this difference. Allianz operated large branch offices in major cities such as Munich, Stuttgart, Cologne, and Hamburg, needing comparably high numbers of keypunch operators in urban labor markets, where it competed with other large companies for this coveted workforce. Many Sparkassen were smaller organizations in rural areas where women possibly had fewer employment options, particularly the more desirable white-collar positions in the service sector. As suitable readers remained elusive for Allianz, a technological solution was impossible. Sparkassen, by contrast, relied less on keypunch operators after a short transitional period, once their long-term solutions were in place. Sparkassen, Allianz, and other corporations inadvertently created growing demand for keypunch operators when they installed electronic computers in their offices.

This demand emerged as attitudes toward women's roles and women's work were changing in West Germany. Having emphasized women's roles as housewives and mothers in the 1950s, a decade later, Germans accepted women's work, particularly part-time, promising that women could integrate salaried employment with family duties. However, not all keypunch operations were part-time, and not all were short-term. Keypunch operations—albeit low-paid and routine—may have appeared suitable and even desirable positions for mothers since they offered work in a white-collar environment at respectable institutions. These positions may have been instrumental in drawing the silent reserve of women into the workforce during a time of labor scarcity in Germany.

This article has shown German corporations' responses to the emerging need for keypunch operators, yet it leaves open many questions about the part-time data entry "ladies," as keypunch operators were sometimes called.⁷² Initial research shows that women chose to do data entry work at Sparkassen to gain social recognition by working in the public service sector, and they stayed there for the sociability of working outside the home—not for the additional income, as historians have suggested.⁷³ Some women held jobs in data entry until retirement, and they worked several full days a month, relying on the invisible work of mothers or in-laws for childcare on work days—not just a few hours in the morning that allowed easy integration of work and childcare, as the public rhetoric on part-time work suggested.⁷⁴ Further research is needed to identify the data entry workforce and its social composition and reasons for accepting and continuing data entry work.

Finally, the comparison between Allianz and Sparkassen raises the question of how other industries, as well as companies around the globe, responded to and resolved the data entry issue in their organizations. Demographic shifts in the Sparkassen workforce show that it increasingly included women in their twenties and thirties—who presumably had responsibilities as housewives and mothers—not the young girls under age twenty that they had employed for routine tasks in the 1950s. Some studies suggest that U.S. corporations likewise hired women for mechanical positions in the 1960s, when their data entry needs expanded.⁷⁵ But spurred by the American civil rights movement, Black men and women entered office work, including routine information work in telephone switchboards and the postal service, and possibly also in data entry.⁷⁶ Further research is needed to understand how computer automation played out in different countries and related to local images and expectations of the workforce, including factors of class and race. International comparisons reveal that local expectations about women and their roles shaped the data entry workforce in postwar West Germany. Companies elsewhere did not necessarily apply the same solutions to data entry problems when they installed electronic computers. However, they all had to create human-computer interfaces suited to their local labor, social, and political conditions.