Buffering from Labor Costs
The last chapter argued that men are more likely to be hired into occupations in which employers are significantly buffered from having to economize on labor costs. In order to understand the rationale for this prediction, we need to consider the interrelation between patriarchy and the economic needs of employers.
Maintaining patriarchy is a secondary utility for employers. A secondary utility is a goal that an actor would like to attain but that can be overridden by more important concerns. For most employers, their primary goal is assuring the economic viability and profitability of their firms. For managers, secondary utilities might be the preservation of executive expense accounts, upgrading the food served in the company cafeteria, or providing corporate sponsorship for the arts. These goals are not essential to the economic well-being of the firm, although they may have mildly positive consequences for organizational functioning. As long as the company grows at a healthy rate, few people would question the allocation of funds for these peripheral projects. If these side activities came into serious conflict with the survival of the firm, however, most managers would eliminate the budget line for these items and concentrate their resources on more immediate needs.
Discrimination is a secondary utility for management when it does not provide an immediate economic return. There are certainly occasions when a policy of not hiring women is profitable, as would be suggested by many of the arguments in the introduction. Nevertheless, it is clear that many employers are openly sexist and would simply prefer not to work with women. There are several forms such overt sexism can take.
Some managers have moral objections to hiring women. They sincerely believe that women should remain in the home, and they are reluctant to participate in arrangements that they perceive as having deleterious effects on the family. Other employers are concerned with preserving male employment; they don’t wish to hire women when men have no jobs. In both cases, the manager in question is purchasing a public good; he is using corporate resources to maintain patriarchy in the rest of society. This is a secondary expenditure comparable to any other form of corporate philanthropy, albeit for a somewhat more questionable goal.
Other employers believe that women are inferior employees. As the examples below will show, managers in the Great Western Railway and the Post Office certainly perceived women as being sickly, physically frail, and incapable of performing tasks requiring intellectual sophistication or training.1 While many of these statements can be dismissed as rationalizations, it would be hard to deny that the image many employers had of women was extremely negative and that these stereotypes exist to the present day. If women were in fact inferior employees, then the secondary utility argument would not hold; not hiring women would be consistent with a bona fide need to staff the organization with the most productive employees available. However, since women and men are equally productive in virtually every setting, these beliefs are the product of imperfect information. When managers consistently operate on the basis of faulty information, this suggests that their data collection and research has been less than authoritative and complete. Thus, the persistence of sexist beliefs among managers often reflects the secondary utility of not wanting to invest resources in obtaining accurate data about male and female capabilities.
A third form of sexism is the unwillingness by males to interact professionally with females. Rosabeth Moss Kanter has argued compellingly that managers prefer to surround themselves with colleagues similar to themselves. Employers prefer social interactions that are routine and unstrained; they attempt to achieve this by filling their environment with workers whose background is similar to their own. This provides a comparability of personal style that eliminates awkward “cross-cultural” contacts (Kanter 1977). This concern with homophily is an obvious secondary utility.
There is a last secondary utility that is germane to the exclusion of women that is not explicitly sexist in nature. Some employers might have resisted feminization because this would have produced administratively inconvenient changes in what was an otherwise satisfactory situation. The transition from an all-male to a mixed labor force does entail a number of minor adjustments. Rest facilities will have to be reallocated, or possibly redesigned. New sets of pay and benefit schedules will have to be developed. Offices and personnel may have to be reshuffled, especially if the women are expected to be segregated. None of these obstacles are especially time-consuming or difficult to overcome. However, managers are human, which means they are lazy. Delays in feminization may have been the product of procrastination and a concern with avoiding the nuisances associated with any type of major personnel change.
Firms can tolerate policies that are not in their best interests so long as the costs of such policies are reasonably contained. The best analysis of such suboptimalities comes from the work of the decision theorists. The decision theorists argued that firms do not attempt to maximize profits but merely to satisfice them. The rationale for this is that managerial energy and attention are not unlimited. Employers tend to work on a small set of problems at a time and turn their attention to other matters only as needed. This sequential problem-solving implies that many organizational matters are relatively neglected until they deteriorate into major crises: these crises produce dramatic interventions and responses by management, which in turn distract attention from yet another subset of problems. The result is that those matters that are central to organizational growth are often perceived as being at unsatisfactory levels and are constantly upgraded, while more secondary matters are administered in a relatively inefficient and negligent fashion (Simon 1957; Cyert and March 1963).
The cost of discriminating against women is a labor force that is overly expensive. Such a policy can be tolerated if labor costs are a relatively unimportant factor in determining organizational survival. There are a number of situations in which employers are buffered from the need to minimize labor costs. An important example is when firms are capital-intense. Although such firms may hire a large number of workers, their expenses are relatively small compared to their expenditures on raw material and technology. Such firms can afford to hire males if the extra labor costs are absorbed by savings in the management of the capital budget. No such subsidies are available for the labor-intense firm. Most of the working budget is committed to paying for personnel; the slightest upward drift in wages will have an immediate impact on overall operating costs. Therefore labor-intense firms should take advantage of every form of staff economy available, which implies an increased willingness to employ female labor.
This hypothesis has an obvious implication for the two firms in this study, the Great Western Railway and the General Post Office. The Great Western Railway was extremely capital-intensive, with its heavy investments in locomotives, rolling stock, permanent way, stations, and both the machinery and the raw materials used in the locomotive factory. The Post Office was far more labor-intensive. Mail delivery was virtually unmechanized; the telephone and telegraph services required surprisingly small amounts of physical plant relative to the armies of signallers and operators and required virtually no expenditure on raw materials. As would be expected, the labor force of the Great Western Railway was overwhelmingly male, while the General Post Office was heavily female.
Employers may also choose to hire men for certain occupations if these occupations make up a relatively small percentage of the total labor force. Just as high labor costs can be subsidized by efficient capital management, high labor costs for a minority of workers can be subsidized by efficient management of the majority. This principle has an important implication for clerical work. Clerical work should be more likely to be female in firms in which office work makes up a relatively large proportion of the total labor force. There should be a larger proportion of males in the clerical force of firms whose overall labor forces are blue collar. The Great Western Railway and General Post Office fit this pattern as well. The Great Western Railway hired mostly manual workers to run the trains, maintain the lines, and staff the factories. Clerical workers made up a far greater proportion of the General Post Office, which hired massive numbers of sorters, telegraphists, and telephone operators as well as pure clerks. The clerical force of the Great Western Railway was male while the clerical force of the General Post Office was female.
Another type of buffering occurs when organizational success depends less on the conservation of costs than on other criteria of performance. Most firms cannot allow their total overhead to become excessive; however, there are cases when objective criteria of economic efficiency become irrelevant to organizational survival and functioning. An important example is an organization run by patronage. The purpose of patronage is to provide lucrative jobs for supporters of the current political regime. Because of this, salaries are expected to be generous. Political hiring not only makes the savings from hiring women irrelevant, but it may disqualify women from employment totally if the administration decides to concentrate its rewards among its male followers. In general, the Post Office was not a patronage organization; however, there were particular offices and particular periods in which politics did interfere with the hiring process. An internal examination of the use of women within the Post Office clearly shows the impact of patronage in reducing female employment.
Four types of evidence will be presented to support the buffering model that has been presented here. First, the model will be used to account for variance within each organization in the use of female clerks. Longitudinally and cross-sectionally, men should appear in the least clerical-labor-intense departments and in those offices subject to political hiring. Second, the Post Office will be shown to be quantitatively more clerical-labor-intense than was the Great Western Railway. This difference in cost structure will be shown to outweigh the competing pressure from the small sector of the Post Office that was politically buffered. Third, it will be shown that managers sought to exclude women from their offices. This theory places the prime responsibility for excluding women upon management. If the historical record showed managers doings everything within their power to feminize, but being prevented by some sort of environmental pressure, then the buffering model would be incorrect. The arguments presented here claim that managers chose not to hire women due to their own personal preferences for a male labor force. Thus, there must be clear evidence of exclusionary sentiments among employers.2 Fourth, non-clerical-labor-intense offices will be shown to be relatively inefficient in the handling of clerical matters not related to hiring. Most economic theories are based on the maximizing profitability and minimizing cost. This model follows decision theory in arguing that managers merely “satisfice” these utilities; they only reduce costs to levels that will avoid a crisis in profitability. Materialists should always exercise caution in using models that explicitly posit some level of non-rationality. What the analyst views as being inherently irrational may upon subsequent examination be discovered to have a sound economic base. If the hiring of men in non-clerical-labor-intense offices is the product of inattention to economic detail, this inattention should be manifest in a large number of areas besides hiring decisions. Record-keeping, office procedures, and the organization of the division of labor should all show signs of carelessness, neglect, and objectively low levels of operating efficiency.
In general, women were extremely underrepresented in offices that were not clerical-labor-intensive. Very few women worked in engineering, storehouse, or warehouse offices. Each of these administered a large number of factory workers and warehousemen. Women did not sort mail, because the same officers that administered mail-sorting also administered the mail carriers, a large blue-collar labor force. Women rarely did the clerical work in railway depots. The stationmasters in charge of the booking clerks also controlled the operating staff of the railway stations, an enormous force of freight loaders, train drivers, signalmen, and rail switchmen. The clerks were a very small proportion of the station staff.
Women appeared in very large numbers in the departments whose specialty was paperwork. Among the heavily female offices were financial offices, such as accounting offices, the Money Order Department, and the Postal Savings Bank. The labor forces in paper-collecting offices such as record registries and the Returned Letter Office were all primarily female.
Within the Post Office, an illustrative contrast can be made between, on one hand, the Telegraph and Telephone Departments and, on the other, the Engineering Department. All three departments were involved in the general business of telecommunications. Both the telephone and telegraph services required extensive quantities of both capital and labor: a nationwide system of transmitting equipment and wiring with which to send messages, and a large force of signallers or operators with which to work the system. The Post Office bifurcated the administration of these systems, with responsibility for physical plant entrusted to the engineers and labor management entrusted to the telephone and telegraph departments. The engineers predominantly used male clerks. The telephone and telegraph clerks were overwhelmingly female.
This impressionistic account can be supported by more rigorous quantitative tests. Data for the analysis of the Great Western Railway come from their staff censuses. As we saw in Chapter 1, before World War I so few women were hired that there was no variance between departments in their sex composition. After World War I, departmental sex ratios were virtually constant. Therefore, the analysis considers only one panel: 1933. Departments were subdivided into the headquarters, the divisional offices, and the staff working in depots. Each of these subsections was treated as a separate unit.
There are two measures used for clerical labor intensity. The first is the ratio of clerks to total staff within each department. This is a literal measure of the degree to which clerks dominate the total budget for labor. As expected, this ratio varies inversely with capital costs: the departments with obvious raw material and physical plant expenses have very low clerk/staff ratios.
The second measure classifies the observed unit into office buildings or depots. The major concentrations of the blue-collar staff on the railway were found in depots. Even though the clerical staff in the depots could be large, it was always dwarfed by the operating staff. Most depot managers would have thus concerned themselves largely with blue-collar matters. The headquarters and divisional offices contained the clerical-labor-intense enclaves of the department, the central bookkeeping offices being an example. Thus we would expect to find lower percentages of female clerks in the depots, despite the clerical labor intensity of the whole department.
Table 2.1 shows that both measures are effective predictors of the percentage of clerks that are female. The zero-order correlations with the dependent variable are .43 for the percentage of clerical workers and –.46 for being a depot. When combined, the two variables explain almost 30 percent of the variance after adjusting for the small sample size. Both coefficients are significant at the .03 level. Overall, then, the clerical-labor-intensity model seems to be supported by the railway data.
The Post Office data come from the Postal Establishment Books. The sample consists of those years for which there were women hired by the Post Office: 1876 to 1936. The postal analysis uses two measures of buffering from labor costs. The first is the clerk/staff ratio that was used for the railway analysis. The second is the degree to which hiring was done on an economic rather than a political basis.
Before 1870, most civil service positions, including those in the Post Office, were filled by patronage. Members of Parliament and other notables would receive the right to nominate various individuals for government posts. Since government jobs were designed to be political rewards as well as positions of responsibility, salaries were set at fairly high levels, usually substantially above prevailing market rates. For comparable reasons, job holders were also given lifetime job security. It should be no surprise that, before 1870, the Post Office never made any attempt to feminize.
The patronage era came to a partial end in 1870, when a major program of civil service reform was instituted. There had been extensive parliamentary agitation for the abolition of patronage throughout the 1850’s and 1860’s, sponsored by such reformers as Sir Stafford Northcote and Sir Charles Trevelyan. In 1870, a set of Orders in Council were passed that completely overhauled recruitment procedures for civil service jobs; the nomination system was abandoned in favor of a system of nationwide, open, competitive examinations (Cohen 1941; Richards 1963; Gladden 1967). Subsequently, new hiring was based on the relatively economic criteria of merit, except in the case of offices involved in the sorting of mail. Here patronage continued, albeit at reduced levels, until its demise in the 1890’s (Hanham 1960; Richards 1963).
Regressions of Selected Variables on the Percentage of Female to Total Clerks in The Great Western Railway, 1933
Note: N = 30. The Top figure is the unstandardized regression coefficient. The lower figure is the significance level.
It is significant that, the very year that the Orders in Council were passed, the Post Office began to hire women. Furthermore, the period of the most rapid feminization of the Post Office was the 1870’s. Some of this was purely coincidental. In 1868, the Post Office absorbed the private telegraph companies. The introduction of female telegraphists in 1870 was in response to a shortage of trained male telegraph operators (Scudamore 1871). In 1870, the Accountant General’s Office began hiring women and in six years had become nearly 15 percent female. In 1872, the Returned Letter Department introduced women, and thereafter it feminized so dramatically that merely four years later it was 48 percent female. In 1875, the Savings Bank introduced women, and in 1876 a new Postal Order Division was created that was virtually all-female. These were all departments that were vigorously growing in both the 1860’s and the 1870’s. Before the Orders in Council, new positions had been filled with male political appointees, but after 1870 new openings were filled with women.
There was a second form of political hiring that affected women’s opportunities. This was the use of government positions to support the unemployment policies of the welfare state. In the early twentieth century, the British government was advocating National Unemployment Exchanges and preferential hiring for veterans. In each case, the government decided to act as a model employer and itself reserve a substantial number of positions for both registered exchange users and veterans. Nearly all of the reserved openings were for postmen and sorters, including both permanent positions and the seasonal positions used for the Christmas rush. These political considerations undoubtedly restricted women’s access to employment as sorters. Throughout the entire 1857–1940 period, sorting remained an all-male occupation. In 1912, the Huddersfield postmaster experimented with using women to assist with sorting during the Christmas pressure. He reported that the women performed their duties completely capably and at costs substantially below those associated with temporary male labor. His superiors ordered him to dismiss the women immediately. They acknowledged that using women was cost-efficient, but it would undermine the official policy of giving these positions to male labour exchange candidates (Christmas Pressure 1912).
In 1921, the Post Office systematically considered the future of female employment within the department. A substantial number of temporary women had been hired to replace men serving in the armed forces. The government had to develop a systematic policy as to whether such temporary women would be allowed to staff any of the permanent positions that had previously been held by men. Among the most important areas that had been feminized during the war had been mail sorting. The central administration avoided making any commitments about women in general. Women were to be used in situations where female employment was “appropriate.” However, the relevant committees did make one consistent policy recommendation: that it was important to give some preferential consideration to the claims of ex-servicemen. Subsequently, the Post Office dismissed all of the women who had worked as mail sorters during the war. This created a manpower shortage that was partially filled by exams that were open to veterans only. The remaining vacancies were filled with an open all-male exam similar to those normally used for sorting recruitment (Whitley Council 1919; Ex-servicemen’s Employment 1921). The national campaign to provide for ex-servicemen undoubtedly contributed to the de-feminization of post-war sorting. By removing economic criteria from hiring from some positions not only was a special enclave created for male hires but the general disregard for cost efficiency was extended to other positions in the department. The non-veterans’ positions could have been filled with women. However, the administration exploited the laxity encouraged by veterans’ hires to indulge their uneconomic pre-war male preference.
The measure of economic hiring is an adjusted estimated percentage of jobs that are recruited by open competitive hiring. Before the Orders in Council, this percentage is assumed to be zero. For all offices that were created after the Orders in Council, the score is 100 percent. For other offices, this figure is the percentage of all employees that were hired after 1870. This figure is adjusted for sorting offices to reflect those positions whose sex composition was determined by welfare state considerations.3
One control variable was added, namely the percentage of non-entry-level jobs in each department. Most writers on occupational sex-typing argue persuasively that women are confined to jobs with low pay and status. As Chapter 4 will show, in both the Post Office and the Great Western Railway, women were disproportionately represented in entry-level jobs, for reasons that are different from those discussed here. The percentage of high-status jobs was measured by the percentage of positions attainable only through promotion.4
The analysis is presented for each panel separately. Such a presentation is somewhat unwieldly when compared with pooled equations, but it possesses a number of statistical advantages. It is absolutely free from autocorrelated error and allows for the most explicit delineation of period effects. This removes most of the time-based disturbances that distort pooled ordinary least squares analyses of cross-sectional time-series data.5
There are two types of findings presented in Table 2.2: those that are essentially the same in all periods and those that change over time. The first general finding is that the model does a good job in explaining the variance within each panel in the use of women clerks. The adjusted R2 is greater than .30 in every period, and in most of the early twentieth century it is over .50. Second, both the control variable, percentage of non-entry jobs, and the substantive variable, percentage of non-political hiring, perform as expected. In most cases, the zero-order correlations with the dependent variable are in excess of .40, and the regression coefficients are highly significant.
Regressions of Selected Variables on the Percentage of Female to Total Clerks in the British Post Office by Date
Note: The top figure is the unstandardized regression coefficient. The bottom figure is the significance level.
The behavior of clerical labor intensity is somewhat more complex. The size and significance of the coefficients for this variable increase monotonically over time. In the nineteenth century, clerical labor intensity is a weak predictor of sex composition, the raw correlations being only in the twenties, with the regression coefficients being insignificant. By 1911, however, the effect of clerical labor intensity has doubled, with the regression coefficients being both significant and strong.6
The steady increase in the predictive power of clerical labor intensity reflects the slow implementation of cost efficiency in the Post Office. In the nineteenth century, there was no tidy relationship between office economics and sex composition because sex composition was determined in part by pre-existing policies of patronage. Every office contained a number of male employees hired before 1870 who were doing jobs that might otherwise have been done by women. The number of these holdovers declined steadily throughout the nineteenth century until they disappeared completely in 1911. The strength of the clerical labor-intensity coefficient directly reflects the decline of this confounding factor; as the hiring policies of the Post Office increasingly came to resemble those of the private sector, clerical labor intensity increasingly affected sex-typing.
These equations show a consistent correlation between clerical labor intensity and sex-typing. However, there has been little evidence presented to rule out alternative interpretations of this data. Women were relatively underrepresented in sorting offices, engineering offices, factory offices, and depots. There could be some other distinctive quality about these offices that may have contributed to excluding women.
The most plausible alternative explanations of the clerical-labor-intensity correlation involve the presence of physical labor, manual laborers, or unusual skills. It could be that the non-clerical-labor-intense offices involved some sort of dirty or physical work. Women might not work in factories or depots because such clerks might be expected to do non-clerical tasks as part of their duties. Second, working in non-clerical-labor-intense offices might involve social contact with blue-collar workers. Males employers might have kept women from such positions out of a paternalistic sense of gentility and class prejudice. Finally, it could be that non-clerical-labor-intensive offices involved some sort of special skills. If the work was highly technical or involved unusually high proficiency in math or science, women may have been educationally unqualified to take these positions. While each of these theories sounds plausible initially, they are inconsistent with the historical data on actual job content.
Few of the clerical jobs in these organizations actually involved physical, dirty labor. There were exceptions to this generality, such as shipping clerks, who occasionally had to handle their own packages, and inventory clerks, who often examined merchandise in unsavory surroundings. However, the majority of clerks in blue-collar settings actually did work that was clean and sedentary. Shipping clerks were a trivial proportion of the clerks in the analysis. Most goods and stores divisions used blue-collar workers to do much of the physical inspecting. The number of clerks who would have had such double duties is small. Depot clerks rarely got involved in the physical work of the station; they remained behind their windows, selling tickets and keeping accounts. Even mail sorters did not carry their own mail; the Post Office kept a special force of porters and bagmen to do the moving of heavy mailbags. Thus, in any of these offices, the physical demands of the jobs would have been extremely modest.
These findings are not likely to be the spurious result of managers’ unwillingness to allow women clerks to mingle with blue-collar labor. In the non-clerical-labor-intensive offices, most of the work was done in offices physically removed from the shop floor. In the Chief Mechanical Engineer’s Department, only 26 percent of all clerks actually had jobs requiring contact with manual laborers (Staff Charts 1933). Depot clerks would have worked in the booking windows and the accounts offices, while the physical work of handling freight and trains was done in the yards and on the loading quays. Of the workers in non-clerical-labor-intense offices, only sorters would have had sustained interactions with manual laborers. This is because they worked in mailrooms shared with porters and postmen.
Furthermore, despite the enormous amount of paternalistic concern many Victorian employers showed for their female employees, it is not clear that managers particularly objected to women’s working in proximity to blue-collar workers. There was substantial concern about shielding women from unsavory members of the general public. When the management of the London Post Office was determining what sex of telegraphist to put in what branch, they explicitly took into account the quality of the neighborhood; women were explicitly kept out of the Bethnal Green and Crystal Palace branch offices. Both offices were in slum areas with a lot of street crime (London Branches 1914). However, nobody seems to have cared about lower-class employees. Most London Post Offices were sex-segregated. If women worked in a branch, then that branch was usually all female. However, the physical labor, such as hauling mailbags, was always entrusted to men. Thus the female offices would have had a blue-collar male staff that not only interacted with women but would have taken all of its orders from women.
These findings are probably not a spurious result of unmeasured variations in skill or human capital requirements. In the postal equations, differences in status have already been taken into account by the control variable, percentage of non-entry jobs, an imperfect but nevertheless valid measure of human capital stocks. Furthermore, although the non-clerical-labor-intense offices had distinctive types of workers doing tasks dissimilar from those performed elsewhere, the characteristic tasks of these offices were not particularly challenging. Mail-sorting was done in non-clerical-labor-intense offices. Mail-sorting, however, is quite simple, requiring merely good hand-to-eye coordination and an ability to read addresses. Temporary assistants hired during the Christmas season were able to perform quite competently after receiving only minimal training. Depot work was probably simple as well. Although the content of work cannot be measured directly, depots had the highest concentration of entry-level jobs of any type of office in the GWR. This suggests either that depot work was comparatively unskilled or that depot jobs were not highly valued in the railway status system. Either way, it would be unlikely that depot jobs were reserved for men because of their reputation for difficulty.
The only case in which unusual technical skills would have been required would have been in early engineering offices. There was often little differentiation in such offices between the actual engineering force and the clerical establishment; engineering trainees served their apprenticeship doing clerical work, while their superiors often took on some of the clerical work themselves (Engineering Clerks 1904,1919). Early Postal Establishment Books made no semantic distinction at all between engineers and clerks; both were called “engineering clerks” (GPO Establishment Books 1857–1900). Thus some of the workers nominally called “clerks” would have been doing work of an advanced technical nature.
While technical skill somewhat distorts the equations in Table 2.2, eliminating scientific workers from the analysis has a negligible impact on the overall findings. Later twentieth-century Post Office data permits the division of engineering workers into two categories: those who did work we would now consider to be engineering and those whose duties were purely clerical. Eliminating the technicians from the data does not significantly increase the percentage of engineering clerks who were female. In 1921, only 6 percent of the pure clerks in engineering offices were female. Including the scientific personnel would only lower this figure to 3 percent. Both figures were among the lowest of any department in the Post Office. Running the regressions in Table 2.2 with either figure produces equations that are virtually identical.7
The robustness of the analysis can be explained by the high proportion of purely clerical positions within technical offices. The engineering offices did more than just design production processes. They also supervised the routine administration of large factories, docks, and communications networks. As such, they had to do many of the clerical tasks associated with any business: maintaining personnel files, paying wages, ordering raw materials, keeping records of stock, constructing budgets, keeping accounts, and corresponding with the outside world. Many of these tasks are fairly standardized and use the same procedures in both scientific and commercial settings. Thus, while the technical nature of engineering work may help explain why technical offices were female, this can not comprise the whole explanation. Clerical labor intensity helps to explain the use of men for tasks that do not have distinctive human capital requirements.
The Post Office feminized before the Great Western Railway because the Post Office was more clerical-labor-intensive. The Post Office did suffer from the effects of patronage and political hiring. It can be shown, however, that the enormous gross differences between the two organizations in overall cost structure outweighed the distortions that came from the public sector.
The Great Western Railway had a much lower ratio of clerks to total staff than did the Post Office. Although clerk/staff ratios are only available for the GWR for 1860,1919, and 1933, it is remarkable that three independent sources working over a span of seventy years produced exactly the same clerk/staff ratio: 11 percent exactly (Railway Report 1860; GWR Staff Census 1919; GWR Staff Census 1933). Estimates of the postal clerk/staff ratio are approximately 50 percent. The ratio was 57 percent in 1876,49 percent in 1896, and 52 percent in 1921. By these estimates, the Post Office was roughly five times as clerical-labor-intensive as the railway, a differential that was both substantial and constant over time.
However, the clerk/staff ratio is only one component of clerical labor intensity. Clerical labor intensity is the product of the clerk/staff ratio and overall labor intensity, the ratio of labor costs to total costs. Although the intra-organizational analysis did not incorporate labor intensity, the data being unavailable at the departmental level, this variable can be included in an inter-firm comparison. In 1906, 41.5 percent of the Great Western Railway’s operating budget went towards paying labor. The comparable figure for the Post Office was 61.6 percent. The ratio of labor to total costs in the operating budget slightly underestimates levels of capital expenditure, because large-scale capital acquisition was usually financed out of a separate budget. Furthermore, the purchase of telephone lines and equipment was not included in the operating budget. However, including these items only lowers the postal labor/cost ratio to 59.0 percent, a figure virtually identical to the original estimate. The entire cost of a year of phone construction, in the period of the most rapid growth of the network, was less than 10 percent of the wage bill for mail delivery alone. Thus both measures of clerical labor intensity strongly favor the Post Office.
Economic efficiency in the Post Office was somewhat impaired by political hiring. The effect of this dropped dramatically after 1870, however, and thereafter political hiring was concentrated in a small subpopulation of offices. By 1890, half of the pre-Orders-in-Council employees had retired, and they had completely disappeared by 1911. Thus, by the twentieth century, 92 percent of all office hires were economically based. In a large minority of offices, those that had been created after the Orders of Council, political hiring had never had any impact at all. Among these were the telegraph offices, the telephone offices, the engineering department, and the factory departments. In the nineteenth century, patronage would have had an impact on postal hiring. However, the burden of the railway’s massive capital and blue-collar labor obligations probably provided a far more substantial buffer against clerical costs than did the limited degree of postal political hiring.
Managerial Opposition to Clerical Feminization
The previous analysis was predicated on the assumption that the primary responsibility for excluding women lies with management. Male managers are assumed to prefer an all-male labor force, but are frustrated by economic factors that compel feminization. It is easy to imagine an alternative model in which managers prefer an all-female labor force, but are constrained by forces that compel de-feminization. This might be the case if employers face strong sexist unions or vigorously enforced female labor laws. An employer seeking to use women to reduce costs might find his plans frustrated by strike action or by the fear of prosecution. One strategy for addressing these counter theories is to test them directly.8 Another is to examine the actual behavior of managers considering feminization, to determine if they are supporting or opposing the use of women. Both the GPO and the GWR left extensive archival records of the policy decisions concerning their female employees. Within these files is a rich assortment of letters, reports, and memoranda in which supervisors openly discussed their opinions of women as workers. Using these letters, we can reconstruct the managerial ideologies of the period and determine whether employers facilitated or limited women’s economic options.
In general, managers bitterly opposed the introduction of female workers. The overt objections took two forms: the claim that women were physically and mentally incapable of doing a particular kind of work, and the claim that feminization would pose prohibitively complex administrative difficulties. The vigor with which these arguments were pursued suggests that managerial resistance was sincere. These objections were fairly thin, however, in that they were based on gross misperceptions of feminine incapacity or on exaggerations of problems that could be overcome with very minor levels of inconvenience.
A typical case of the posing of thin objections and their subsequent easy resolution occurred in London Postal Sorting during World War I. Late in 1914, the Women’s Institute had written to the Postmaster General suggesting that since the military was quickly absorbing the supply of available male labor, women should be used as temporary replacements of men on leave from sorting offices. The letter was referred to the head of the London Postal Service, who considered the idea entirely unworkable. First, according to the Controller, major structural alterations would be required to provide restrooms for the women. Second, women could not work side by side with men, but would require completely separate offices. This would require fundamental alterations to the design of the main building. Third, sorting mail was too arduous for women. Men traditionally sorted while standing up; women would require chairs. Since the sorting boxes were at unusual heights, standard-issue chairs would be inadequate, and the Post Office would have to invest in specially ordered equipment. Furthermore, women could not be expected to work at night, a problem since a substantial percentage of mail-sorting had to be done at four or five in the morning. Given the enormous costs in construction and supplies, and the difficulties of scheduling inherent with an inflexible workforce, it would be far more economical to raise male wages and attract a larger pool of job applicants. The Postmaster agreed with the substance of the analysis, and the female sorter project was abandoned.9
A mere five months later, the situation had altered. The labor shortage had become so intense that the Post Office could not find suitable males at any price, making the use of women unavoidable. Once the commitment to use women had been made, the supposedly insuperable obstacles of a few months before were easily overcome. Women were given their own floor at the sorting station, providing them with separate offices and lavatories at no construction cost whatsoever. Women were supplied with standard-issue chairs, which they rarely used, preferring to sort standing up like the men. Women were also quite willing to cover the late night shifts. The experiment in feminizing was so successful that the Post Office used women in increasingly large numbers throughout the war years (WWI Sorters 1915).
A similar story can be told about the upper management of the Great Western Railway. In 1876, a mere six years after the introduction of women to the Post Office, the board of directors of the GWR suggested an inquiry into the viability of using female clerks. The initial response from the department heads was overwhelmingly negative. Virtually every administrator raised the same objections, the non-availibility of lavatories and the impropriety of women’s working side by side with men. Most of the reporting offices were located in the firm’s headquarters at Paddington station, a very large multistory complex, making it hard to believe that these difficulties could not have been resolved with a modest amount of shuffling.
The board felt the accommodation argument was persuasive and ruled out the use of women in already-existing facilities. However, the argument was hard to justify in the case of offices currently undergoing construction for expansion. In these cases, the structural features necessary for the segregation of women could be obtained at nominal expense. Therefore, the board approved the use of women in a facility undergoing complete renovation, the goods department in Birmingham. However, before the plan could be implemented, managers from all over the system started submitting remodeling requests using as their rationale a desire to incorporate women workers. The same individuals who had opposed feminization were willing to accept women if this was the cost of having their offices redecorated. The board feared that granting these requests would set an expensive precedent and, rather than embark on a firm-wide construction project, canceled the plans for feminization. This suggests an extremely weak commitment to feminization by the board. Neither the option of mixing men and women nor the option of dividing pre-existing offices between the sexes was ever considered as an alternative to remodeling. On the basis of this very superficial analysis of the use of women clerks, the issue died entirely for twenty-nine years.
In 1905, the question of women clerks was renewed, after the General Manager of the company had taken a tour of American railways. At the time, the United States was considered to be in the vanguard of progressive railway administration, and the General Manager had noted that the Americans used female clerks. He took a poll of department heads, and as was the case in 1876, the response was overwhelmingly negative. The traditional complaints about sex-mixing and bathrooms re-emerged. These were buttressed with a variety of individual objections. The Relief Cashier argued that women could not take a job involving travel and thus could not go to out stations to perform relief duties. The Rates and Taxation Office, which had a wide variety of clerical and accounting positions, declared at the outset that women would only be suitable for messenger positions and then provided a long discussion of why messenger work is inherently better performed by men.
This time the General Manager was in favor of feminization, so he used his position to override the objections of his subordinates. However, he did acknowledge the legitimacy of the accommodation arguments. Therefore, he ordered the hiring of women clerks, but made remodeling funds generally available throughout the system to provide for segregated offices and facilities. As before, women were first placed in an office that for exogenous reasons was undergoing physical renovation, in this case the clerical section of Paddington Goods Station. The remodeling stricture severely impeded the rate of feminization, since the delays in implementing and completing construction could often be substantial. Therefore, by World War I, only a small number of female clerks had been hired, and these were concentrated in a select set of remodeled offices (GWR Women 1906).
In the Great Western Railway, a strong central executive had to force the introduction of women over the initial objections of upper and middle management. A similar set of incidents occured in the Postal Savings Bank. The debate over women in the Savings Bank began in 1874, by which time several other clerical-labor-intense departments, such as the telegraph offices, the Receiver and Accountant General’s Office Department, and the Returned Letter Department, had already feminized. The Secretary of the Post Office observed that the Savings Bank was still all male, and suggested to the Controller of the Bank that he consider hiring some women clerks. The Controller’s response was extremely negative. His main objection was a lack of “suitable accommodation” for women. Subsequent correspondence clarified what he meant by “suitable accommodation.” Not only would women need to have separate lavatories and offices, but they would require separate lunchrooms and kitchens as well. The use of women would thus require a complete duplication of Post Office facilities.
The Secretary of the Post Office was unimpressed with these arguments. He ruled by fiat that the next sets of staff increases for the Savings Bank would consist entirely of female clerks, with the Controller left free to use these women as he saw fit. Because the Bank was very severely understaffed, the Controller could not afford to turn down the help. After several futile pleas for additional men, rather than women, he was forced to take in the new female clerks.
However, the battle was not over. The Controller next sought to demonstrate that women could not do bookkeeping. Since bookkeeping comprised the main activity of the Bank, barring women from the ledgers would guarantee that some fraction of future replacements would have to be male. Therefore, the Controller put all of his women to work addressing envelopes and copying receipts. Receipts were printed forms in which one filled in the name of a depositor with the date and amount of the deposit, as simple a job as could be found in the office. He then commissioned a set of pseudo-scientific tests to demonstrate that women lacked the mental capacity to keep accounts. Several women sorters were borrowed from the Returned Letter Department. These were basically filing clerks, who were hired with comparatively little skill or education and whose work did not involve any computational tasks. They were then given a one-morning crash course in calculating compound interest by hand. This involves some fairly complex arithmetic and requires practice to obtain proficiency. That very afternoon, the women were given several trial passbooks and told to balance the accounts and calculate the interest. About two-thirds of the women did very poorly at this exercise. The remaining third were given a second test with more difficult passbooks, and they too failed the exam. The Controller then sent the results of both tests to the Secretary of the Post Office as clear evidence of women’s lack of quantitative skills. The Secretary criticized the unfairness of the makeshift exams and sent an explicit order to put the women on ledger duties.
The Controller then attempted to demonstrate that women were physically incapable of doing bookkeeping because they would be unable to lift or handle the heavy ledgers. This sounds outrageous to the modern reader, but would have been easier to justify in Victorian times. Before modern filing systems were invented, accountants kept related records together by writing them on the same page of a ledger. If the account was complex, this could require a very large physical document. The Savings Bank kept some record books that weighed over twenty-five pounds and stood over four feet high. The Controller argued his case successfully before the Chief Medical Officer, who in turn filed a report claiming that allowing women to lift ledgers would seriously impair their health. This argument the Secretary found convincing, but he decided that, rather than get rid of the women, he would have the ledgers redesigned so that workers of both sexes would find them more portable. The Medical Officer was then told to produce a signed statement giving full approval of women’s doing ledger work, and the Controller was ordered to provide a comparable signed statement announcing his intention to put women on bookkeeping duties. The entire debate, from the Secretary’s first suggestion to the receipt of the final signed concessions, had taken over a year and four months (Bank Introduction 1875).
Several points are illustrated by these stories. Managerial opposition to clerical feminization was quite widespread. This vigorous resistance to feminization was not merely confined to non-clerical-labor-intensive offices. It was endemic to managers as a whole, even in such clerical-labor-intense enclaves as the GWR Accounting Office and the GPO Savings Bank. In each of these cases, women were only introduced after the exertion of pressure from outside forces, usually central management. Cost structure did not cause local managers to change their preferences concerning women so much as it introduced pressures from organizational gatekeepers not to discriminate and instead to emphasize cost efficiency. The effect of cost structure can be clearly seen in the differing policies of the 1906 GWR General Manager and the Secretary of the Post Office in 1874. The GWR General Manager exerted significantly less pressure than his counterpart for the implementation of feminization. He never advocated the reshuffling of offices to attain segregation; he never redesigned normal work procedures to facilitate women’s employment; he permitted the preliminary remodelings to occur at a very slow rate, significantly delaying the introduction of women to many offices. In distinct contrast to the Postal Secretary, he never accelerated feminization by banning male hires in particular offices. The greater capital intensity of the GWR decreased the urgency of achieving clerical cost controls, producing a far less intense commitment to implementing feminization over managerial resistance.
The substance of the objections the department heads raised is also important. In many cases, the administrators were clearly trying to avoid having to make adjustments in their normal procedures that would have been required by the addition of new types of employees. An example of such an objection is the binder reform, which, though precipitated by the introduction of women to the Savings Bank, was nevertheless a long-overdue rationalization of record-keeping. However, not all of the resistance can be attributed to bureaucratic inertia alone; some represents sexist ideology and a genuine preference for male colleagues. Both in the railway and in the Post Office, employers insisted that women be segregated from men. Segregation came to serve latent functions in the Post Office; however, the initial justification of this policy in both organizations was the normative impropriety of men’s and women’s having informal contact in an office setting. That such a factor dominated the discussion of an issue that had enormous implications for the lowering of operating costs suggests that Victorian sex-role ideology still played an important part in managerial decision-making. This can be seen as well in the continual references to feminine frailty in the debates over women’s suitability as employees.
There is also evidence that men really wanted male offices and were willing to work quite hard to get them. The sixteen-month campaign of the Controller of the Savings Bank is an excellent case in point. A more dramatic example can be found on the Irish railways. When the Post Office nationalized the telegraphs, railway telegraphs were explicitly excluded. Subsequently, special arrangements were made to allow them to remain in private hands. In most cases, this merely formalized the railways’ existing control over their own lines. In Ireland, however, a large number of telegraph lines on railway property had been built and operated by the Post Office. In these cases, the telegraphs were to be transferred to the appropriate private railways. When the Post Office operated these telegraphs, they were operated by women. After the transfer, the railway companies fired all of the women and hired men to do the signaling (Telegraph Report 1876).10 This expulsion is remarkable for several reasons. First, telegraph operators took a long time to train, and the nation was suffering from a shortage of telegraph workers. Dismissing a fully functional operating crew and replacing them with raw recruits would have been an extremely substantial upheaval. Managers could not have been avoiding administrative complexities, for in this case clearly the simplest solution would have been to maintain the status quo. The railways voluntarily gave themselves an enormous organizational problem as the price of a male labor force. Second, ignorance about female capacities is not likely to have been a factor. Female telegraphists were now well established in both England and Ireland. Railway management would have had ample opportunity to see how well women worked when the Post Office was running telegraphs on the railways’ own lines. This suggests a solid commitment by management to a male labor force, even when presented with good information on objective criteria favoring female employment.
Variations in Attention to Office Management
The main argument of this chapter is consistent with the earlier claims that patriarchy is in contradiction to the needs of both individual and collective capital. Discrimination is not wholly dysfunctional for capitalism; the GWR and the GPO used sex discrimination to solve problems involving their internal labor markets, as subsequent analyses will demonstrate. When employers chose not to use cheap female labor, however, their reasons were often cultural rather than economic, and the profitability of their offices often suffered as a result. The offices that chose not to save money on clerical labor took other liberties with office management, ignoring other economies in the administration of the clerical work itself. The fact that the most discriminatory offices were particularly inefficient lends support to the proposition that discrimination can run counter to the logic of capitalist rationality. There is unlikely to be an ulterior materialistic logic to the exclusion of women, if discrimination is correlated with inferior economic performance.11
The Post Office Engineering Department, which used a predominantly male clerical force, was notorious for its disorganized and inefficient office management. The paperwork fell into such overwhelming disarray that two separate formal inquiries had to be made into the department’s clerical procedures. In 1904, investigators reported that the administrative work of the office was in extensive arrears. There was widespread confusion as to who was processing what materials and what actions the department had taken previously. The committee diagnosed the problem as largely self-induced. First, the engineering department had never made any attempt to create a division of labor among the clerks. There had not even been a clear division of labor between technical and clerical duties. Paying the staff and keeping accounts sometimes devolved upon technicians hired to design telephone systems. Second, the filing system had completely broke down. Part of this was a simple backlog in returning documents to the files; however, this was compounded by a system of file organization that was archaic and inefficient, resulting in very difficult document retrieval. Third, there had been little attempt to provide an overview of just what records were being produced. Many of the assignments that were clogging the office workflow were duplicates of tasks that had already been done elsewhere. No attempt had been made to eliminate redundant paperwork. Finally, the committee noted, the department was severely understaffed.
This last consideration was beyond the control of the engineers, staffing levels being determined exogenously by the Post Office Secretary. But these personnel shortages only intensified the problems; the basic inefficiencies of the system had been created by the engineers themselves. The relationship between the engineering department’s capital intensity and its poor administrative practices was probably not coincidental. When the Post Office Secretary sought experts in office management who could rationalize the engineering department’s procedures, he turned to the heads of the clerical-labor-intense departments. The filing system was assigned to officials from the Registry and the Returned Letter Office. The other procedures not handled by the Secretary’s own staff were assigned to the Receiver and Accountant General’s office (Engineering Clerks 1904). These were three of the most clerical-labor-intense departments in the Post Office, so it is significant that they were viewed as models of efficiency in data management.
Despite the objective shortcomings of their administration, the engineers were not personally criticized for their inattentiveness. The explanation of this is consistent with the clerical-labor-intensity model. In a post-crisis report of 1919, the Engineering Work Committee wrote:
The position of the Engineering Department is peculiar in this respect that while the clerical work is of course ancillary to the engineering work and must to some extent be under the control of the principal engineering officer, it has at the same time so special a character that it is impracticable without loss of efficiency to require of a technical officer more than a general knowledge of the clerical procedures which he controls. . . . It is imperative that technical matters should make the first and by far the greatest demand upon his time and that any change of procedure which may tend to encroach upon the time which he gives to engineering questions is in our opinion strongly to be deprecated [Engineering Clerks 1919].
The Post Office thus explicitly encouraged its engineers to concentrate on matters other than the administration of clerical work. The presence of capital responsibilities allowed upper management to tolerate some inattention to office management per se. This toleration can be exaggerated, since the mere presence of a commission on clerical work in engineering suggests a certain amount of overall concern. The Post Office at various times did authorize high-level accountants and office managers to concentrate exclusively on the administration of the engineers’ clerical labor. However, these officials were always subordinate to the Chief Engineer, who retained the ultimate control over staffing decisions. Engineering clerical policy continued to be treated as a secondary matter.
In this particular case, lax office management produced a crisis of serious proportions. Most administrative inefficiencies do not result in such dire consequences. The Great Western Railway seems to have handled its office work with dispatch and efficiency. Data collected in the 1870’s by both the Railway Clearing House and the GWR itself suggest that the error rates in freight returns and accounts were typical of those of other railways (Staff Statistics 1870; Northern Recruitment 1904). Taylorite investigations of clerical performance in the 1920’s reported adequate levels of overall efficiency (Clerical Work Committee 1916-28; Goods Stations 1922). Nevertheless, there is suggestive evidence that the railway did not devote its full attention to reducing clerical costs.
In 1879, a national recession severely reduced the GWR’s traffic and revenues. The railway could not respond to the recession by reducing services and thus found itself under great pressure to reduce the costs of its fixed operations. The GWR introduced a program of labor austerity that cut pay rates for newly entering employees and imposed a stretch-out on existing staff. Most manual workers were paid by the day, with the hours of labor left up to the discretion of management. The GWR began imposing eighteen- and twenty-hour days upon most of the operating staff, compounded with consecutive shifts merely two or three hours apart. Workers could be on duty continuously for several days with only a couple of hours allotted for sleep. These abuses were so blatant and widespread that the eight-hour day became the central issue around which railway unionism crystalized (Bagwell 1963).
The Great Western Railway left written plans on the precise means by which these economies were to be obtained. They list pay cuts for several types of employees including guards, brakemen, switchmen, signalmen, porters, and police.12 Clerks are conspicuously absent from the list. There was no discussion at any time of achieving cost reductions in the offices. A subsequent analysis reported the economies in total staff that had been made by each department. All of the staff cuts had been made in departments that hired predominantly blue-collar labor. Traffic, goods, signalling, and engineering all reported significant savings. In the offices that hired only white-collar labor, no staff reductions or pay reductions were made at all (Staff Expenses 1879). Such a policy could be justified economically in terms of the greater savings available from the numerically predominant blue-collar force. However, the presence of such a source of savings did buffer the clerical force from cost-cutting and allowed for relatively lax office management.
Male office workers were thus associated with inefficiency in clerical administration. However, this was not the consideration affecting the use of men as opposed to women. The postal equations show that, net of variables associated with buffering models, the status and promotion prospects associated with an office had a critical impact on determining the sex-typing of its jobs. The next two chapters will explore the relationship between women and low-status jobs: Chapter 3 will discuss whether clerical feminization can be explained by deskilling; Chapter 4 will consider the relation between sex, turnover, and patterns of promotion within firms.
1. Citing all the cases I have found of managerial criticisms of women workers would be tedious and uninformative. The reader seeking particularly flagrant examples of belief in feminine frailty will find plenty of material in the writings of British civil service medical advisers. A good published source for these is the testimony given by medical personnel at the 1914 McDonagh Commission hearings. The 1881 Telegraph Controller’s Report on the use of women for telegraph work contains an extensive discussion by local management of the physiological incapacity of women to work long distance lines (Telegraph Substitution 1880; McDonagh 1912-13).
2. Note that, for the theory to apply, the amount of sexism need not vary across departments in any systematic way. The claim is not that clerical-labor-intense employers are intrinsically less opposed to women, but rather that such employers are organizationally less capable of acting on their non-economic preferences.
3. The percentage of post-1870 hires was estimated by calculating the number of employees working in 1870 and then estimating retirements per year based on Post Office actuarial data (War Cabinet 1919). The average male postal career was approximately forty years in length. In each year from 1871 to 1910, one fortieth of the number of workers in 1870 were estimated to have retired. Once the number of survivors was calculated, it was possible to calculate the percentage of staff not in this survivor pool. In sorting offices, all postmen were coded as having been hired through political channels. Other occupations were treated normally. This overestimates the political hiring of postmen, but underestimates the political hiring of sorters, providing a roughly accurate overall score.
4. This control was also included in the preliminary analyses of the GWR. As Chapter 4 shows, on the railway men were ten times more likely than women to hold non-entry-level jobs. However, this control does not predict office sex composition in this case. Due to the high percentage of male clerks overall, low-status offices contain large numbers of low-status male clerks along with the women. This weakens the ecological relation between status levels and gender.
5. The analysis begins in 1876. The Establishment Books generally reflect hiring practices two years before the observed date. This is because there was bureaucratic delay in formalizing and recording staffing levels. Typically, when new employees would be hired, they would not be recorded in the Establishment Books until the details of their long-term salary scales were determined. Thus the first observation point in the data that contains women is 1876, rather than 1871.
6. An exception to this trend is 1936. There are severe missing-data problems with this panel, producing an N of only 13.
7. The analysis in Table 2.2 includes technical personnel to allow comparability between the equations for different dates. Engineers can only be excluded from engineering offices for selected dates in the twentieth century. Combining technical and clerical personnel was the only way to create a standard sample that could be used for every panel in the analysis.
8. See Chapters 5 and 6 for explicit discussions of the roles of law and labor.
9. Note the use of this opportunity by the Postal Controller to put in a plug for higher wages. In general, postal administrators were always advocating better conditions for their staff, exploiting any issue to refer to the need for higher pay. Since budgetary control was centralized outside of the individual departments, it cost the department heads nothing to make these requests, which facilitated good relations between management and labor.
10. Fortunately, in the transfer the women were legally guaranteed employment, so the Post Office had to hire them back at considerable expense. Because the rail telegraphs had been a substantial proportion of postal employment, the Dublin office, which absorbed the women, was the only telegraph department in the country with a labor surplus.
11. In this section of the study, the term “office management” will be used in a very narrow sense. Office management here refers to the administration of data-processing and paperwork within the office. It includes accounting, filing, and the processing of correspondence, but not the administration of substantive policy. This usage is consistent with that of the literature on office administration.
12. Engine drivers and firemen were excluded from wage cuts, although not from hours abuse. Engineers and firemen were the craftsmen of the railroad, being highly skilled and, more significantly, highly organized.