6

Exclusion by Organized Labor

One of the most fundamental issues in the area of occupational sex-typing is assessing the role of organized male labor in reducing economic opportunities for women. Male workers have often been vociferous opponents of equal opportunity in employment. Men have barred women from union membership when such membership was needed for access to employment. Unions have agitated for protectionist legislation excluding women from various “dangerous” occupations. They have gone on strike to protest the introduction of female co-workers. Besides these formal mechanisms, there have been informal actions such as the harassment within the workplace of new female employees. Opposition to women workers has not been universal, but it has been quite common. In the American case, Alice Kessler-Harris has documented the existence of anti-feminization campaigns among printers, iron workers, cotton spinners, and subway conductors, among others (Kessler-Harris 1982).

In most cases, these actions can be explained in terms of the traditional union tactic of raising wages by restricting the labor supply. Organized labor has attempted to create an artificial scarcity of workers by invoking a large number of criteria that might disqualify new entrants to employment. There have been explicit limits on the number of apprentices; these have been accompanied by restrictive definitions of who can qualify for apprenticeship. Excluding female workers from employment can effectively cut the labor pool in half. It can be plausibly argued that such a strategy can be counterproductive. Exclusionary campaigns have often resulted in the creation of hostile pools of scab labor. Women and blacks have been quite willing historically to cross the picket lines of sexist or racist white male unions. Nevertheless, most male workers have not conceived the issues in these terms. When men have dropped their opposition to female employment, it has usually only been in the face of an inevitable feminization against which opposition would have been futile.¹

For the student of sex-typing, the question of what causes worker resistance to feminization is only of secondary interest. The fundamental question is whether such campaigns have an effect on the percentage of women in an occupation. There are two primary positions on this issue. The first is that organized labor can exert a significant influence on management in determining occupational sex-type. No one would say that labor determines occupational sex-type unilaterally. However, sustained and militant opposition to female workers might induce management not to use female workers in settings that otherwise would have been feminized.

The opposite position is that sex-typing is determined unilaterally by management. This is the assumption that intrinisically exists in any mode of feminization based purely on “economic” (in this case managerial) utilities. Thus, the clerical labor intensity theory and turnover theories in general all assume a process of determining sex type that gives no role to the sentiments of employees.

It seems likely that, if organized labor has an impact on gender-related matters at all, this impact is confined to a relatively limited number of industrial settings. Male workers can only ban women where they have the labor strength to control entry into their trade effectively. This represents a fairly high level of worker control. Historically, most workers have faced real challenges in maintaining unions against the resistance of employers. Many have not been able even to obtain union recognition. Others have obtained meaningful wage concessions, but have not been able to restrict management’s right to make decisions concerning the labor process unilaterally. Only a small elite of workers consisting largely of traditional crafts workers such as printers or construction workers, have obtained significant control over the labor process. These workers have been only occasionally able to control entry into employment. Thus, if strong workers will be more successful than weak workers in controlling entry, the empirical question of interest is where on the continuum of organized labor strength workers obtain the ability to influence sex-typing.

The logic of inquiry can be more easily understood by considering two extreme findings. Imagine that one studied an incredibly powerful set of workers and found that these workers were incapable of having any impact on feminization. Such an example would strongly suggest that workers in general have little imput into sex-typing decisions. At the other extreme, imagine one studied a completely weak and unorganized occupation. If one found that these workers made a significant contribution to managerial decisions concerning sex-typing, one could conclude that probably all workers exert such an independent influence. Neither of these scenarios is likely to occur in real life. In the absence of such pre-emptory case studies, an appropriate research strategy would be to do a series of studies of occupations with varying degrees of labor power, seeking a cutting point between effective and ineffective exclusionists. By finding precisely how much organizational strength is required for a role in sex-type determination, one can discover what percentage of the labor force could reasonably be expected to possess such resources.

The present research is a first step in the execution of this grander design. The Great Western Railway and the General Post Office allow for a strong test of whether male clerical workers had any input into determining the rate of feminization.

Clerical work represents a case of weak unionization and unmobilized labor power. Only slightly over 10 percent of all private sector clerks were unionized by 1975, despite a renaissance in white-collar organizing (Freeman and Medoff 1979). Those clerical unions that have existed have generally not been particularly militant. Few of them ever strike. The successful unions rarely achieve more than union security and wage concessions. If male clerical workers were able to pose significant obstacles to feminization, this would suggest that a very large percentage of the labor force would be capable of winning such victories. If male resistance to feminization were ineffective, it would leave open the possibility that better-organized workers might yet have a meaningful impact. Considering clerical work thus represents a test of the role of labor strength that is biased towards finding no effect.

Within clerical work, however, the General Post Office and the Great Western Railway represent the cases most likely to generate a significant labor effect. These two organizations represented the vanguard of clerical organizing. Before World War II, the civil service unions and the Railway Clerks Association were two out of the three clerical unions in Britain that had ever obtained any concessions from management (Lockwood 1958). The density of clerical union membership in the Post Office and in the railways was the highest among all industries in Great Britain. The Post Office had been continuously organized from the 1870’s; the GWR had been organized from the turn of the century. Both sets of unions had been involved in major strikes, and had engaged in successful parliamentary agitation. On the minus side, it must be noted that the independent economic power of these organizations was questionable; they lost every strike they engaged in. Nevertheless, the presence of actual walkouts, long campaigns of workplace agitation, the successful use of political tools, and the maintenance of consistently high levels of union membership put these unions far in the forefront of worker organization and militancy. If the postal and rail clerks could not slow down the flow of feminization, it is unlikely that other clerks would have had more impact.

Clerical-Industrial Conflict within the Post Office and the Railways

In order to understand the evidence and interpretations provided below, it is helpful to consider first the overall histories of labor relations in these two firms. The railway situation was substantially simpler, so it will be considered first.

Labor Relations in the Railways

In the nineteenth century, the clerical force of the railways was wholly non-unionized. Labor relations worked in a fashion similar to that described by David Lockwood in The Blackcoated Worker (1958). Office supervisors exercised personal control over their subordinates. Standardized pay rates and job descriptions did not exist. Remuneration was determined largely by the department head based on the individual characteristics of the worker. The absence of any organized counterforce to managerial autonomy forced employees to get ahead solely through obedience, loyalty, and the development of personal ties with their employers.

The Railway Clerks Association (RCA) was formed in 1897 as an industry-wide union. In its early days, if two railways served the same city, unionists from both companies were in the same local. The Great Western Railway lagged slightly behind the other companies in the early recruitment drives. Bristol and South Wales had a strong chapter from the beginning, and Paddington later became very important. However, Paddington was one of the last major London stations to open its own local branch (RCA Annual Reports 1903-28; Walkden 1928).

In the early stages, the RCA largely confined itself to political tactics. It affiliated with the Trades Union Congress Political Committee, and early on successfully sponsored a small number of parliamentary candidates. This allowed the RCA to win significant concessions by strategically delaying important railway legislation. The most important of these victories was the winning of a government inquest into railway pension funds that resulted in the payment of substantial compensatory damages by management. This tactic was also used to prevent the harassement of local RCA organizers (Walkden 1928).

The first economic concessions were won in 1911. Traditionally, railway clerks worked six days a week. However, supervisors could routinely demand an extra day on Sunday, which was unpaid. In 1908, the RCA petitioned the GWR to provide payment for Sunday labor. Indirect negotiations on the issue lasted for three years, with the railway finally granting remuneration to all clerks who worked at least six hours on Sunday. The victory, however, turned out to be largely symbolic. When the railway implemented the policy, the Sunday shifts were all made five and one-half hours long, allowing very few clerks to qualify for the supplemental payments (GWR Petitions 1914).

The next concessions, however, were substantial. In 1912, after extensive petitioning, the RCA won an upgrading of entry-level pay. This was followed in the same year by amelioration of the other workers’ salaries. The RCA was only partially responsible for these improvements. The Great Western Railway for some time had been suffering from difficulties in recruiting junior staff. One problem had been that starting salaries had been allowed to slip to levels uncompetitive with those of other employers. Thus, it is not insignificant that entry-level pay was increased first and senior pay was raised subsequently. Nevertheless, the RCA probably played an instrumental role in accelerating the pay raise and seeing that it was generalized (Salaried Recruitment 1912).

World War I produced major gains for the RCA. Due to the fact that railways played a vital role in the defense effort, the government semi-nationalized the industry. The railway companies were guaranteed a high fixed level of profit in return for providing uninterrupted service, eliminating any financial incentive on the part of the companies to conserve on costs. In return, the government put pressure on management to avoid strike-related shutdowns in an economy in which labor was generally scarce (Pratt 1921; Aldcroft 1968). The companies were thus strongly motivated to grant the railway unions whatever they wanted, and the unions took immediate advantage of this leniency. Wages were raised no less than eight times during this period. Although inflation absorbed much of this improvement, railway employees made substantial gains (GWR Petitions 1914; Colours 1920; Pratt 1921).

After World War I, the government rescinded control of the railways and began forcibly to amalgamate many private lines. During the amalgamation period, labor relations on the railroad were completely reorganized and formalized. The British government had developed a plan for the overall reduction of industrial conflict. Every industry in the British economy was to adopt industry-wide collective bargaining between employer confederations and trade unions in forums known as Whitley councils. Disputes not resolvable by mutual consent were to be submitted to public arbitration. The British political elite believed that union recognition and open collective bargaining were in the public interest (Phelps Brown 1983). During the amalgamation period, all aspects of railroad organization were reconstituted with an active participation by the government. Railway labor relations incorporated many of the institutions of the Whitley Council movement (Bagwell 1963; Aldcroft 1968). Thus the Railway Clerks Association along with the blue-collar unions participated in industry-wide collective bargaining. Salary scales were standardized across the major railways using a uniform job classification scheme (Grand Classification 1920). In the first formally negotiated pay settlement, a substantial portion of the wartime temporary increases was incorporated into permanent wages (Contract Talks 1919-20; Bagwell 1963).

The RCA’s role in obtaining these benefits was mixed. The union was directly responsible for its own inclusion in the railway negotiating machinery. The original plans had called for collective bargaining only with the blue-collar unions. The RCA threatened a strike if they did not receive comparable treatment. The railways could not tolerate a major union recognition strike during the sensitive amalgamation proceedings, particularly when public policy militated for union recognition, so management conceded the RCA procedural demands just before the strike deadline (Lockwood 1958).

Once bargaining rights were granted, however, management proved to be unyielding on other issues. The first collectively negotiated wage agreement provided few new benefits for the clerks. Some of the “temporary” war bonus payments were incorporated into regular salaries, a fact that merely reflected the inflation of British currency during the war. However, the union had little impact in determining the amount and form of these concessions. The pay package that was put into the final contract was virtually identical to management’s initial offer, while none of the RCA negotiating planks appeared in the final settlement, even in diluted form (Contract Talks 1919-20; RCA Contract 1925). In later years, the pay issue remained deadlocked. The union made several attempts to increase wages, while management made several attempts to reduce them. Both types of demands were repeatedly rejected in arbitration (Railway Clerk 1907-37; Walkden 1928).

The early post-war period shows some ability by the RCA to at least maintain a defensive strength against hostile managerial initiatives, but the balance of power swung decidedly in managments’ favor in 1926. The RCA joined the blue-collar railway unions in supporting the General Strike of 1926. The railway clerks stayed out for the duration of the strike. When the strike collapsed, labor power on the railways was effectively crushed. The strike leaders were suspended and punitively transferred. Pay for all railway employees was dramatically reduced, and it was cut several times more in the following years (General Strike 1926; Labor Report 1930-33).

Thus, despite the bargaining privileges that the RCA enjoyed, labor control was very uneven. The RCA won major benefits in 1912 and during the war. It was able to prevent management attempts to reduce wages in the 1920’s. However, its ability to obtain wage increases on its own was very weak after World War I, and after 1926 its power was negligible.

Labor Relations in the Post Office

The labor relations in the Post Office were somewhat more complex. The Post Office workers were represented in a wide variety of single-occupation associations. The identities of these unions changed over time as varying organizations underwent mergers. However, the distinctive character of postal industrial relations existed largely apart from the reorganizations of the unions themselves.

The lead sector of postal unionism was the telegraph force, who were the earliest group to organize. In the mid-nineteenth century, postmen had engaged in ad hoc organization, and clerks throughout the civil service had lobbied in the 1850’s for the right to vote (Swift 1900). However, the telegraph clerks were the only workers with a visible and commanding strike threat. Most of the work in the Post Office was either semi-skilled or involved little demand for timely performance: delivering mail was simple manual labor; sorting mail required low levels of training; most of the bookkeeping in the GPO was not done under pressing deadlines. Substitute labor could be obtained from elsewhere in the civil service or the private sector. The telegraphists, however, were irreplaceable. Telegraphy was a skilled occupation requiring long training. Merely learning the rudiments of using Morse code took a full year. It took years of practice before a telegraphist could work at high speed. The use of untrained scab labor to operate the telegraphs would have resulted in a significant deterioration of the quality of service. Furthermore, the telegraph system was indispensible to the functioning of British business. The financial community would have been unlikely to tolerate a telegraph strike that produced delays and inaccuracies in stock market reports or impeded the transmission of urgent commercial news. The Post Office was thus under strong pressure to avoid a protracted stoppage of telegraph service, and the signalling force was quite willing to exploit this vulnerability.

The second most militant workers after the telegraphists were the sorters and postmen. These workers were not particularly skilled, but they were employed in sufficient numbers to provide a modestly visible strike threat. Any particular postman on strike was easily replaceable, but replacing every postman on strike even in a small city would have made extraordinary demands on the local labor market.

The postal clerks were the third largest category of unions. In many cases their participation in industrial conflict was nominal. The postal clerks had some firm-specific skills, but in many respects their labor resembled tasks being done by other financial clerks. In the event of a strike, they could have easily been replaced with workers from the temporary clerical market. Clerks were rarely employed in large numbers except in London, and in London the clerks were divided among a multiplicity of offices scattered among many buildings. It would have been extremely difficult to organize a clerical strike that would have been sufficiently large to challenge management’s ability to use scab labor.

The postal unions did not rely exclusively on the strike in pressing their demands. They made extensive use of political tactics and vigorously pursued their claims in Parliament. They monitored the votes of Members of Parliament on bills germane to postal workers and rewarded those members who sought to raise civil service salaries with electoral support from a large bloc of single-issue voters. Most civil service employees were concentrated in London. Because of this, only the London representatives needed to pay attention to the clerical constituency. However, telegraphists, postmen, and sorters were distributed throughout England. There was not a seat in Parliament that could not be affected by the “postal bloc.” This political threat bolstered the claims of the unionists and gave the postal workers power far beyond what one would expect from their economic position.

Postal voting power should not be overestimated. Postal unionists commanded respect but not obedience. They were but one of many interest groups within a politician’s district. Many politicians chose to seek support from more conservative sources, openly resisting postal union claims. Still others temporized by making concessions to the unions that were purely symbolic. Many parliamentarians would pay off their debts to the postal workers by delaying the postal budget appropriations in committee, making long speechs deploring the postman’s conditions but rarely putting forward any resolutions or making amendments to the budget. Despite the show, most postal bills passed in their original form. By the twentieth century, the postal unions finally learned how to press their sponsors for more concrete forms of political assistance, but then the Whitley Council movement neutralized many of their legislative victories. Lobbying posed great potential as a tactic for agitation; nevertheless, the returns from this activity were often disappointing.

In the late 1860’s and early 1870’s, labor relations within the Post Office were undergoing a fundamental transformation, due to two changes. The first was the enfranchisement of government employees in 1868. The second was the nationalization of the telegraphs and the subsequent absorption of the signallers. The enfranchisement of the civil service gave birth to the strong stream of labor-related political organizing that would characterize the postal unions. The precise reason why Parliament gave government employees the vote is unknown. The civil service had lobbied for their electoral rights and may have attracted help from sympathetic democratic allies. Enfranchisement may have also been related to the civil service reform movement of the late 1860’s. The original rationale for denying government workers the vote was that, under a patronage regime, politicians might improperly reward supporters with jobs. As patronage declined, the threat of machine politics faded and the need for disenfranchising government employees was reduced. Whatever the reasons for the change, the result was the creation of a new actor in the national political arena.

The incorporation of the telegraphists into the Post Office meant the introduction of a well-organized set of artisanal laborers. Before incorporation, there had been occasional public meetings by postmen protesting wages and conditions, but these meetings had generally been ineffectual; there had never been a strike or any labor-induced pay increase (Swift 1900; Hall 1902). The telegraphists posed a much more severe labor relations problem. They had been formally organized under the private companies and had actually had a small strike in the 1860’s. In 1871, soon after absorption, they organized the first strike in postal history, when the telegraphists of several northern English, Scottish, and Irish offices went out on a one-week stoppage. The Post Office had never dealt with such a situation before, but managment was able to obtain a short-term victory. By firing six of the union leaders, management precipitated the strike before the union was ready. The defensive strike that followed was largely confined to a minority of urban stations. Management then imported replacements from non-striking districts and broke the workers’ resistance. The strikers returned, receiving no immediate concessions, and the leaders were punitively transferred (Telegraph Strike 1872). However, management’s success was dependent on the partial support the strike had received; a more complete shutdown would have had more devastating consequences.

In 1873, the Post Office raised the telegraphists’ wages. A wage increase had been promised back in 1868; nevertheless, the amount of the increase was non-trivial. When the telegraph increase was finally granted, this spurred a parallel movement among the sorters, who supported their wage demands by calling what was to be a much-imitated tactic, a no-overtime strike. Management traditionally had had the right to demand unpaid overtime from the men. The sorters refused to work more than seven hours a day. The 1873 strike was foiled completely, however; six men were suspended, and no concessions were made (Swift 1900).

The sorters then turned to political agitation. In defiance of postal regulations, they held two large public meetings to portray their plight. Public meetings were usually demonstrations of political strength. They were chaired by Members of Parliament sympathetic to the postal workers, and they often received favorable attention from the press. Such meetings were good devices for obtaining public action. This time the results were dramatic. The Queen called for an investigation into the sorters’ conditions. The tribunal found in favor of the sorters, and pay rates were raised dramatically.

The sorters attempted to follow up on their victory the next year, in 1874. They put out an memorial requesting further increases over the 1873 increase. However, this time they met with substantially less success. Five men were dismissed merely for delivering the memorial, and the request was rejected out of hand. The sorters did get a debate in Parliament to discuss their claims, but their advocates made no formal motions. Nothing concrete emerged from the hearing (Swift 1900).

After this setback, the postal workers temporarily concentrated on obtaining new parliamentary forums for their grievances. The late 1870’s were characterized by a long campaign to obtain the right to testify at a major inquest into the overall operation of the civil service. Despite an extensive lobbying campaign, the workers were denied access.

In 1881, the postal unions won another major wage increase. This drive was spearheaded by the telegraphists and supported by the sorters and postmen. The early states of the campaign consisted of collecting memorials from postal workers all over Britain. An earlier set of petitions in 1880 had been turned down. The new memorials were to be supported by a two-pronged offensive. Pro-union MP’s were to voice support of the proposal during the debate on the postal budget. At the same time as the debate, the London telegraphists were to go on a no-overtime strike. The Post Office capitulated before the double threat. A new telegraph pay scheme was announced just before the no-overtime strike and the debates were to have occurred.

The central role of the telegraphists in this activity should not be underestimated. Most of the parliamentary speeches protesting low levels of postal pay focused primarily on the telegraphists. It was the telegraphists who threatened the no-overtime strike, so when the original postal revisions were announced, only the pay of the telegraphists was raised. Other employees did receive some improvements but only much later. Although participation in the 1880-81 campaign was general, placating the telegraphists seems to have been the prime concern of both parliament and management (Swift 1900; Hall 1902).

The events of 1889 and 1890 paralleled those of 1880 and 1881. The postal unions pressed to participate in another other civil service commission. This time, too, they were denied. This was followed by a general pay increase, won by the combination of a London telegraphists’ no-overtime strike threat and political support in Parliament. As before, a pay increase for the telegraphers was announced just before the no-overtime strike was to be instituted and immediately before a pro-union parliamentary debate (Swift 1900).

In 1888 there had been a petition by sorters and postmen for higher pay. This memorial had been denied, but subsequentially the Postmaster General had agreed to an inquiry by postal management into sorters’ and postmen’s wages. In 1890, the in-house review concluded that no adjustments in wages were necessary. This decision provoked great discontent among the staff. The independent agitation by the telegraphists was diminishing the postmaster’s ability to quell this second wave of discontent, and he agreed to reconsider the commission’s findings.

While this reconsideration was pending, the sorters and postmen suffered a few blows. The postmen decided to threaten a strike, and held a large rally before the event to articulate their claims. Several postmen who attended this rally were dismissed. Four sorters publicly offered support to a potential postmen’s strike. These workers too were dismissed. The unions then took defensive actions protesting the harassment of the militants. After achieving no results in this agitation, the postmen finally called a strike. The results of the strike were dismal. Fewer than 600 postmen supported the strike, most of these from a single district of London. The strike was crushed in one day, resulting in the suspension of 435 men (Swift 1900).

Similar tactics of harassment began to be used on the telegraphists. During 1889, in the initial periods of the telegraph campaign, several Cardiff telegraphists published some protest telegrams in the local newspaper. They were punitively transferred. Several meetings were held in sympathy for the “Cardiff martyrs” but no restitution was ever made.

In 1890, on Jubilee Day, the head of the Central Telegraph Station assembled the signallers and gave a customary patriotic speech. After finishing his presentation he called for three cheers for Queen Victoria, which were given. He then requested three cheers for the Postmaster General. The Postmaster General received three groans.

Most modern managers would dismiss such a display as a joke. The Post Office took the incident very seriously. Management circulated a formal disclaimer among the staff to sign. While some employees signed it, most did not. Management then called each of the union leaders and gave them a choice: sign a formal apology with a vow of future obedience or face immediate dismissal. The union leaders had to undergo the humiliating ritual of applying to their supervisors for pardon. These setbacks were obviously less dramatic than those suffered by the sorters and postmen, but they do show that even the most powerful postal workers faced some risk in engaging in union activity (Swift 1900; Hall 1902).

The only incident of a strike among the clerical staff took place in 1891. The second division clerks in the savings banks went on a no-overtime strike a few months after their general pay revisions were announced. This was a gender-related strike. The issues involved will be discussed in greater detail below; for present purposes it is enough to say that the strike collapsed after one day. No concessions were made, and 250 male clerks were suspended (Swift 1900; Hall 1902).

Between 1890 and 1914, the strategy of the postal workers changed. The unions continued to seek inquests into their wages. However, the 1889 sorters’ pay inquest had shown that, when management was allowed to run these inquests internally, the investigations would turn into nothing but an elaborate stalling device. The unions began to press for inquests that would be staffed by sympathetic referees. Since the workers had more leverage on members of parliament than on any other group, the overall stategy became one of seeking a parliamentary wage investigation. Over the course of this twenty-five-year period, the unions won a series of different investigations into wages. The composition of each committee became closer and closer to the union ideal; finally in 1906 and 1914 they obtained two all-parliamentary commissions (Humphreys 1958).

In 1895, the postal unions obtained a wage commission staffed by representatives of management throughout the civil service. This represented an improvement over 1889, when the commissioners had been drawn exclusively from the upper levels of the Post Office. Nevertheless, this new committee not surprisingly showed little sympathy to the postal workers’ claims. After three years of investigation, they wrote a final report recommending negligible increases in overall wage levels.

The events of 1890 repeated themselves. The London telegraphists threatened a no-overtime strike, demanding higher increases. Parliamentary allies of the postal unions negotiated a postponement of the strike pending the working out of a superior arrangment. After some parliamentary debates and a second impromptu inquiry made by the postmaster himself, postal wages were raised by a generous amount (Swift 1900).

Agitation for an all-MP commission continued. In 1898, the unions polled every candidate for a parliamentary seat in Britain, asking if they would explicitly endorse an all-MP commission. Nearly every candidate promised future support. Circulating to such a survey by civil servants was technically illegal, however, and the newly elected Parliament refused to pass a bill legalizing the practice. In 1899, a resolution pressing for an an all-MP commission was defeated. Nevertheless, the postal union pressure continued to be intense. From the late 1880’s up through 1906, every single vote on a postal budget had been delayed pending a forced debate on postal worker wages and conditions. Amendments were added to the postal budget moving the impounding of £100 from the Postmaster General’s salary if postal wages were not improved. (Such amendments usually got withdrawn before the final reading). Nearly thirty questions a year were being put on the floor of the House to the Postmaster General demanding written explanation of why postal conditions should not be improved. There clearly was a small minority of parliamentarians such as Albert Rollins from Islington who vigorously supported the union’s position. The mainstream was also having trouble resisting the union’s appeals (Swift 1900).

In 1903, the Postmaster General noted that several MP’s had come to him seeking some buffer from the persistant lobbying of postal unionists. To alleviate the pressure, he nominated a new commission. This time the committee was made up of businessmen in the private sector (Humphreys 1958). The 1903 Bradford Commission surprised everyone by reporting in favor of a major across-the-board pay raise. The Postmaster General refused to implement the report. For two years there were parliamentary attempts to force compliance with the Bradford recommendations. Some wage adjustments were made, but the arbitrators’ recommendations were generally ignored (Hobhouse 1906).

In 1906, the postal unionists finally triumphed. As a device for resolving the dispute over the Bradford Commission, Parliament agreed to a full parliamentary investigation into wages. The investigation was a slow process, taking over three years. The results of the Hobhouse Commission were mixed. Provincial workers received a wage increase that pushed them towards parity with their better-paid London counterparts. The rest of the postal claims were denied (Hobhouse 1906). The Post Office delayed the implementation of the plan so that it was not until after 1910 that the postal workers received the benefits of the inquest begun in 1906. Nevertheless, the majority of employees received real wage benefits from the endeavor.

The next move was to press for a second parliamentary commission to benefit the workers left unsatisfied. In 1914, the Holt Commission was established with a composition similar to that of the Hobhouse panel. World War I prevented the Holt Commission from issuing their report. It also marked the close of an era of parliamentary influence by civil servants.

The government took action after the war to insulate the legislature from postal union interference by the establishment of Whitley Councils. The Whitley Council movement was the same movement that helped bring about collective bargaining for the railway clerks: the attempt by the British government to institute industry-wide collective bargaining throughout the nation as a whole. Since the government was expending considerable resources in attempting to persuade private industrialists to adopt the system, the state was under moral pressure to use Whitley Councils to take care of its own labor relations problems. Thus, throughout the civil service, an elaborate machinery of grievance committees and consultation boards was set up to deal collectively with all manner of disputes concerning personnel policy (Humphreys 1958; Wigham 1980).

The effect of permanent dispute machinery on civil service negotiating strength was mixed. The postal workers no longer had to fight a major political battle every time they wanted their wages reconsidered. At the same time, however, they lost the ability to staff such commissions with referees of their own choosing. In fact, their political organization was effectively neutralized. Parliamentarians with mixed sympathies could now argue that the postal workers had a valid channel for expressing their grievances and that parliamentary intervention was not only superfluous but inappropriate. The amount of postal parliamentary agitation dropped dramatically after 1920. While the success rate of such tactics had always been mixed, even in the best of times, after 1920 the unions won no major parliamentary victories.

At the same time, the economic bargaining strength of the telegraphists took a major downturn. This was brought on by the introduction of the telephone and the telex. The telephone had been around since the beginning of the twentieth century. However, only after World War I did it achieve fundamental penetration into routine patterns of business communication. As the telegraph system lost its monopoly on the transmission of urgent commercial information, the national ability to withstand a telegraph strike was significantly increased.

The telex allowed for the significant de-skilling of the signalling labor force. The scarcity of telegraphists had centered around the long training times required to be able to transmit quickly in Morse code. One not only had to learn the code itself but had to develop high levels of dexterity and accuracy in the physical tapping out of the signals. Furthermore, the instruments were unreliable and often required continual adjustment and maintenance by the operator.

The telex eliminated the need for most of these skills. Telegraph messages could now be sent out on a keyboard resembling a typewriter, with the coding being done by machine. Since the physical skills of typing at a keyboard were well distributed throughout the female labor force, most telex operators were fully trained at the time of their recruitment. The machines were reliable and did not require continual maintenance between transmissions. The result was that the number of replacements to striking telegraphists was limited only by the number of available telex devices (Telegraph 1927). This produced a quiet period in postal industrial relations. Wages stagnated for sorters and postmen. They declined for telegraphists. For some reason, clerical wages increased in the late 1920’s, although this was accompanied neither by parliamentary agitation nor by a strike threat (GPO Establishment Book 1918-36; Humphreys 1958).

These clerical union histories should make one point reasonably clear. The bargaining strength of these unions was always problematic. The Railway Clerks Association achieved a few successes in the 1910’s and 1920’s. Its strength, however, was somewhat dependent on the patronage of the railwaymen’s union; even with this asset there were many periods, such as after 1926, when the union suffered from complete powerlessness. The telegraphists had a consistent record of winning pay advances, but even they suffered from frustrations in Parliament, a failed strike, and the final undermining of their bargaining power. The independent power of the other postal workers was even more modest. Thus, in wage negotiations, the issue of central concern to these unions, the winning of concessions was difficult and subject to failure. On matters of gender, the ability of these unions to enforce their will was substantially more tenuous.

The Efficacy of Gender-Related Union Campaigns

Neither the RCA nor the postal unions were neutral to the sex issue. The postal unions aggressively sought to limit clerical feminization. The Railway Clerks were mixed in their sex demands. The male rank and file was anti-feminist. The leadership favored equal opportunity for women. Thus, the official goals of the Railway Clerks Association was to increase employment for women clerks. However, very few of these campaigns, whether for or against women, could be said to have had any success. In the short term, the unions were able to stall or to accelerate a variety of reforms. However, ultimately the occupational sex type of most disputed jobs became that which was preferred by management.

The sex-related campaign that would have seemed to have had a high prospect for success was the Postal Telegraph Clerks’ campaign of 1910-14. This was an attempt to forestall the feminization of London branch telegraph offices. As London gained in size and population, the demand for mail and telegraph service increased. The London Postal Service was thus required to open a large number of new branch offices. Very little sorting was done at these offices, this function being performed at large centralized sorting depots. The primary functions of the branch post offices were to provide window service and to serve as a local base for the transmission and receipt of telegrams. These offices when opened were usually sex-segregated. They were all male, all female, or occasionally female by day and male by night.

At the turn of the century, the London Postal Service committed itself to a policy of feminizing these branch offices. In order to minimize disruption, the Post Office arranged to leave all pre-existing offices with their sex composition intact. However, all new offices were to be all female. The policy was kept secret, but the demographic trend soon became clear to the rank and file. (An exception was made for exceptionally busy offices, such as that at Charing Cross, and those in rough or unsavory neighborhoods, such as Bethnal Green or Crystal Palace.)

In 1910, the Postal Telegraph Clerks Association began to memorialize management about the transition. The male telegraphists’ concern was based on substantive economic considerations. The opening of each office created a series of new supervisory positions. Single men were supervised by men, and women were supervised by women, the new offices represented opportunities for promotion. The men were thus attempting to reserve for themselves a share of the new supply of superior jobs. The PCTA demands were two-fold. First, they wanted some of the branches to be reserved to men. Second, in redress for the new feminization, they wanted offices that were female by day and male by night to be made all male. By sending an old female staff to a new office, the telegraphists could have ensured that the newly created supervisorship at the old office would be a male position (London Branches 1914).

At the time, the Postal Telegraph Clerks had a very reasonable bargaining threat. Their parliamentary influence was at a maximum. The market monopoly the telegraphists had on Morse signalling was still intact. Nevertheless, the telegraphists were unable to extract any meaningful concessions from management. Every new office opened from the inception of the program until World War I was staffed by women. Between 1901 and 1911, the percentage of females in the district signalling offices had gone up 9 percent. In the years between 1911 and 1914, it went up another 7 percent. In the three main years of the protest, as many women had come in as had been introduced in the previous ten years (GPO Establishment Books 1901–14).

In compensation, the Post Office did eliminate some of the mixed-sex branch offices. However, this proved to be a nominal rather than a substantive improvement. Management did a cost-efficiency study of the split offices and discovered a wide variety of practical problems. The transfer of keys and work instructions between shifts was inefficient and error-prone. The night male staff often had to do day duties at other offices, resulting in inefficient between-shift shuttles. Postal management decided to eliminate the split-sex offices, but in a way that would incur no extra expense to the department. Single-sex offices were created by shuffling pre-existing personnel. The final plan did not result in a single new male hire or promotion. Technically the union’s demands were met, but in a form that violated the basic intentions of the campaign (London Branches 1914).

Following World War I, the development of the teletype induced a further feminization of telegraphy. The telegraphists (now merged with the sorters and postmen to make the Union of Post Office Workers, or UPOW) attempted to block this feminization. Although they failed in their ultimate goal, they were able to delay the introduction of teletyping. The rules of the new Whitley procedure required that management consult with the unions before implementing such a change. Feminization was supposed to occur under a set of guidelines jointly agreed upon by union and management, with these guidelines being expected to include compensation to the unions for whatever damage might be suffered by them in the process of the transition. The UPOW simply refused to accept any of the proposed guidelines or compensations. Their position was that nothing could reasonably buffer them from the harm that they would suffer by a major wave of feminization. The conciliation process allowed for the operation of multiple levels of grievance machinery. By fully exploiting these, the UPOW was able to tie up negotiations for four years.

In 1927, the conciliation process was brought to a close with no significant concessions having been made by either side. Management then implemented a policy of hiring two women for every three new hires. The postal unions resorted to their traditional tactics, memorializing and political pressure. Several deputations were sent to top management, both by the UPOW and by various supervisory unions. Sympathetic MP’s took their legal prerogative of demanding formal written responses from governmental officials to questions from the floor. At one point, the Postmaster General had to demonstrate to the house that increasing the proposition of female telegraphists would not jeopardize the nation’s signalling capacities in the event of war. Such harassment tactics had only a symbolic impact. A one-sentence response was considered a legal and acceptable answer to a parliamentary question. The efficacy of the UPOW campaign can be measured by its objective results. Between 1926 and 1931, the percentage of females in the Central Telegraph Station went from 47 percent to 61 percent. This was the fastest rate of feminization since 1870 (GPO Establishment Books 1926-31; Telegraph 1927).

The sorters rarely had to make any explicit gender claims due to the all-male composition of their labor force. However, in World War I, it became necessary to introduce female substitutes for men who had left for military service. The Fawcett Association, the major sorter’s union, vehemently protested this temporary labor. The union feared that the use of female temporary sorters might create a precedent whereby sorting might become feminized after the war. To resist this threat, union leadership actually offered to exchange extension of the normal workday in return for the preservation of an all-male force. Negotiations over the sex of temporary substitutes occurred several times during the war, re-emerging every time the use of temporaries was extended. In no case was the union able to prevent or even delay the use of women substitutes. They were not even able to obtain a trade-off benefit (WWI Sorters 1915).

The Savings Bank clerks were the only workers driven to striking over gender issues. The militancy of their tactics is remarkable in the light of their modest degree of bargaining power. Their numbers were too small to allow for a meaningful voting threat. Their skills were common to a wide variety of private clerks and were nearly duplicated by the staff of the Postal Accountant General’s Office and the Money Order Department. A post office-wide coalition of financial clerks might have had some strike power. A financial office acting on its own account had nearly none.

The Savings Bank no-overtime strike occurred in the aftermath of the 1890 salary revision. The Savings Bank reacted to the general increase in pay rates by increasing the rate of feminization. The management announced that all future vacancies were to be filled exclusively by women. This produced widespread protest among the male staff of the bank. As in the case of the telegraphists, one issue was the loss of male supervisory positions. However, there was an additional issue concerning the distribution of nightwork. Women were not expected to work overtime. The Savings Bank was perpetually understaffed and required substantial inputs of evening labor in order to handle peaks in demand. The declining proportion of male employees ensured that the night duties would desolve upon an increasingly smaller percentage of the labor force. The male clerks naturally resisted this substantial increase in their hours of labor.

The Savings Bank clerks circulated two memorials, held a public meeting, and lobbied for letters of public support in newspapers. In 1891, they decided to imitate the London telegraphists and threaten a no-overtime strike. Management refused to concede, forcing the clerks to implement their walkout. The strike was an immediate disaster, collapsing after one day with over 240 clerks being suspended. Feminization continued at a brisk rate, and the financial clerks never struck again (Swift 1900; Hall 1902).

The Railway Clerks were equally incapable of exerting an independent influence on gender hiring decisions. This assessment, however, becomes complex when one considers the conflicting goals of the organization. The Railway Clerks Association was characterized by an anti-feminist rank and file dominated by a pro-feminist leadership. The Executive Committee of the RCA considered that the feminization of clerical work was inevitable in the long run. Short-term opposition to female employment was only likely to create a body of female workers outside the union who could effectively serve management as scabs. By 1914, the Executive Committee was explicitly telling local unionists not to resist the extension of women’s employment. The official strategy was to work towards the amelioration of both men’s and women’s conditions by actively working to organize and support female clerks (Bevan Letter 1914).

This policy was not well accepted by the rank and file. The majority of local union activists probably favored an aggressive policy of anti-feminism. Such a claim can be validated by a reading of the union newspaper. Local columnists repeatedly inveighed against the dangers of female employment. The Bristol and South Wales correspondent wrote several stories a year noting how female employment led to favoritism among employers and overwork among the men required to make up for female deficiencies. Satirical poems appeared, attacking the suffragetting and slow work that the authors associated with women clerks. The editors usually balanced these with feminist comments or with centrally written counterpoems urging solidarity among the sexes. During World War I, the newspaper attempted to balance the male commentary by hiring a female clerk to write a column on women’s issues. After World War I the editors moved to eliminate the flow of hostile commentary. They closed down the offending regional columns, but still felt pressure from the rank and file to discuss the problem of women clerks. In 1920, the editors permitted the publication of a one-time forum on sex equality on the railways. Views on both sides of the issue were published, with the editors adding favorable comments on the feminist letters and unfavorable comments on the sexist letters. Subsequently, the newspaper declared the issue closed and forbade any further letters or communications on gender (Railway Clerk 1907–37).

The conflict between the rank and file and the executive became particularly pointed at annual conventions. At the conventions it was customary to vote on a number of planks representing the next years’ union program. The Executive Committee usually sponsored a plank supporting equal pay for equal work. This plank was a sanitized expression of the board’s feminist sentiments. If challenged, such a position could be defended to exclusionists as protective of male privilege. Since women were “inherently less productive” than men, equal wages would cause employers to prefer efficient male labor.

The Executive Committee, however, was often seriously embarrassed by male local representatives who wanted the supposed exclusionist goals of the policy explicitly stated in a formal motion. In 1914, 1915, and 1916, against Executive Committee recommendations, hostile amendments were moved to the equal pay plank that specified female labor as being by definition two-thirds that of labor performed by men. In 1914, the amendment passed. In 1915 and 1916, the amendment was only defeated after extended debate. In 1917, the Executive Committee solved its problem. Amendments to convention resolutions were explicitly forbidden. Subsequently, the equal pay platform in its feminist version was passed without problems. Thus, the egalitarian policy of the RCA was only achieved by the centralization of power by the Executive Committee, requiring the censoring of dissent and the brokering of union conventions. These policies silenced but did not eliminate mass exclusionist sentiment. In 1946, twenty-five years after the final censoring of public anti-feminist complaints, the Executive Committee at an annual convention put forward a plank condemning discrimination against married women. The convention rejected the plank overwhelmingly (Railway Clerk 1907-37; Convention 1946).

The split had a predictable impact on the efficacy of union gender policy. It was very difficult for exclusionists to implement any campaign of opposition to female employment. In the entire period from 1900 to 1940, the RCA only made one demand for the restriction of female employment, a request that women be barred from all-night employment. Since these jobs were very few and were traditionally all male anyway, the impact of this demand was slight. Exclusionists were thus forced to rely on informal wildcat tactics if they wished to press any claims upon management independently. The ability of the union to obtain benefits was problematic even when it made use of its full range of organizational resources. The impact of a small, resourceless, faction would have been much less. This is particularly true when one considers that the dissidents involved would merely have memorialized, rather than struck. Thus the rank-and-file members of the railway clerical force would have had little impact on the hiring policies of the Great Western Railway.

Ironically, it would seem that the leadership was equally ineffective at increasing opportunities for women. The Executive Committee only put forward two sex-related negotiating claims during the 1900–1940 period. Both were made during the 1920 contract negotiations. The first demand was for equal pay for men and women. This was rejected by management in the first day of bargaining on women’s salaries and was never raised by the union again. The second was for the protection of women’s jobs during the return of male workers after World War I. During World War I, the GWR had hired a large number of temporary women to replace clerks who had joined the military. After the war, the RCA wanted the railways to keep the temporary women on as permanent employees. They, however, also wanted the returning men to be allowed to resume their jobs with no loss of benefits. The union plan for achieving this seemingly contradictory program was to put the women on short time for the indefinite future. Some vacancies had already been created by former clerks who would not be returning. Others would be created over the long run by natural attrition. The RCA hoped to avoid the mass dismissal of women by having the railway overstaff for a few years. In this way the temporary women could be integrated into permanent staff as vacancies arose. The railway did not explicitly reject this demand. However, they postponed discussion of the issue until the rest of the 1920 contract details were negotiated. In the meantime, they dismissed most of the temporary women. The mass firings made the RCA request academic. At the end of negotiations, the subject was dropped by mutual agreement (Contract Talks 1919–20).²

The anecdotal evidence presented above suggests that unions played a negligible role in determining sex type. Consistently, management stonewalled union sex demands, denied them outright, or granted them in a form that deprived the concession of substantive meaning. In some cases, this can be explained by union weakness, as was the case with the telegraphists of 1926. Even relatively strong unions, however, like the telegraphists of 1910, were unable to obtain significant concessions. The resources of these unions were exhausted in the struggles over wages and salaries, leaving them with little capital with which to pursue auxiliary demands.

Quantitative Tests of the Role of Informal Exclusion

Although this discussion thus far has emphasized formal mobilizations against women, informal resistance may have been equally important. Many males when confronted with the introduction of a set of unwanted women have acted on an individual level, rather than working through their unions. By complaining to their supervisors, sexually harassing the women, and not cooperating with or training the new employees, men can create a deterioration of productivity and office climate that is so severe that employers will remove the offending women and restore the pre-existing order. Most of the opposition to women clerks in the Great Western Railway probably took this form; the feminist tendencies of RCA leadership would have deprived sexist railway clerks of any formal union channels for acting on exclusionary demands. Such informal mobilizations could have occurred in the Post Office as well, where anti-feminist campaigns were stalled or ineffectual. Thus, to consider fully the impact of resistance by male workers to feminization, it is necessary to measure the impact of such clandestine exclusionary campaigns.

There are obvious methodological problems involved with estimating the effects of informal campaigns. Such activity tends to be lost to the historical record, with both management and workers denying that such a process occurred. Even when such activities are acknowledged, one can not be sure that the absence of a record implies an absence of resistance. Thus, it is impossible to correlate direct measures of the amount of informal opposition to female employment with measures of the subsequent use of women.

If one wishes to study clandestine exclusions, there is a second strategy available. One can study the distribution of worker militancy and union power and observe whether this correlates with the sex composition of offices. If work settings with well-organized, powerful men had few women, while settings with unorganized, marginal men were heavily feminized, a plausible interpretation might be that the strong workers were excluding women. The finding of no relation between levels of worker power and the use of women would suggest that sex-typing is determined entirely by managerial discretion. In general, the data overwhelmingly support the low union influence hypothesis. There is a very poor correlation between changes in sex type and changes in worker bargaining strength.

Data from the Great Western Railway will be considered first. Table 6.1 shows the ratio of female to total clerical employment for the Great Western Railway for several dates. It also shows comparable figures for the Chief Mechanical Engineer’s Office, a major department within the GWR that will be used in subsequent analyses. The critical contrast implicit in the table is the sex type before and after 1926, the date of the General Strike. Before the strike, the Railway Clerks Association was relatively strong; afterwards, its bargaining power was completely undercut. If worker resistance had any impact on the percentage of women hired, then feminization should have increased after 1926. No such pattern emerges in the data. Both the Great Western Railway and the Chief Mechanical Engineer’s Office show little change in sex from 1923 to the mid-1930’s. The years around 1926 show changes in occupational sex-type of less than a percentage point.

The departments with a high level of labor organization were not particularly prone to be less female. The most rigorous support of this claim comes from an internal analysis of the Chief Mechanical Engineer’s Department. Sketchy, but nevertheless suggestive, data are available for the GWR as a whole. Ideally one would like to correlate departmental sex ratios with an explicit measure of union activity. Union membership would be an acceptable measure. Participation in the 1926 General Strike would be equally valid. None of these data is available in systematic form for departmental comparisons.

An indirect measure of labor strength is the level of wages. Raw wages per se are a poor measure of organized labor strength. The impact of union activity is confounded with skill, status, and age. The effect of some of these exogenous considerations can be neutralized if one contrasts pay in two periods in which the rate of change of pay would have been particularly subject to the effects of union agitation. Not all of the noise is eliminated by this method, but the impact of such factors is considerably reduced.

In 1920, the first industry-wide clerical pay contract was negotiated. This entailed the widespread standardization of wages throughout the railways. Before 1920, clerks were paid on a series of local-specific scales. Pay rosters of this early era show that a given department could have an enormous number of maximum salaries for differing types of work (Staff Statistics 1870; Salaried Recruitment 1912; GWR Petitions 1914). The RCA contract had but seven grades of male salaries. This meant a necessary assimilation of the old grades into the new.

This transformation had obvious implications for the advancement and maximum salaries of the clerks in question. These individuals had a substantial interest in pressing for the best classification possible. Probably most of these decisions were made by managerial fiat. However, there was a neutral appeals process for the correction of unfavorable classifications and clerks could make informal complaints to their immediate supervisors. The extent and distribution of such activity are, of course, unknown. Nevertheless, such individual complaints are the precise informal mechanism one would wish to analyze for its impact on female employment. One can thus analyze wages before and after the 1920 contract negotiations. The offices where there was substantial wage inflation are likely to be those where employees had some impact on the classification procedure.

The following regression should be interpreted cautiously. The best estimates of departmental wages before and after the contract negotiations come from observations in 1919 and 1933. The 1933 figures were affected by wage changes that occurred after the contract negotiation. Some of these wage changes reflect subsequent informal bargaining and thus measure latent worker strengths. Other influences include changes in the labor market and the working of seniority, which would not have been affected by the internal labor relations of the railway. Thus 1919-33 salary differentials are an imperfect measure of clerical bargaining strength, although they represent the best data that are available at a departmental level for the GWR as a whole.

Table 6.2 regresses the percentage of females within departments in 1933 with the male wage change from 1919 to 1933. A second equation includes clerical labor intensity and past sex composition as controls. The small N reflects missing data due to the non-comparability of some departments between the two dates.

The bivariate equation shows almost no zero-order relation between sex composition and wage change. This persists when the control variables are included in the equation. The coefficient of the wage variable is negative, which is consistent with an exclusionary model. However, the overall effect is both small and statistically insignificant.

The above analysis is flawed by the indirectness of the measure of male labor strength. More direct measures are available for offices within the Chief Mechanical Engineer’s Office. The Chief Mechanical Engineer’s Office was located in the west of England in a company town called Swindon. The GWR built all of its own locomotives and rolling stock, as well as rolling a substantial proportion of its rails, and all of these factory operations were concentrated in Swindon and supervised by the Chief Mechanical Engineer. Furthermore, the department functioned as a central accounts station for the western branch of the railway.

Not surprisingly, the CME encompassed a wide variety of work situations. There were factory offices, where clerks worked in close proximity to workers on the shop floor. There were technical offices, where clerks worked in laboratories next to technicians. There were purely administrative offices. Some of these involved secretaries working as personal assistants to teams of managers. Some were large concentrations of clerical proletariats doing semi-mechanical data entry and processing. The shop-floor and technical offices tended to be all male. The purely administrative offices varied from nearly all male to all female. Thus within the CME one could find examples of virtually every kind of clerical work organization imaginable.

Data exist for each of these offices on the percentage of clerks who joined the General Strike of 1926. These rates are available separately for men and for women. The male strike rate is a good indicator of the degree of commitment of male clerks to the Railway Clerks Association. The strike rate represents the extent to which clerks were willing to defy management in order to show their concrete support for labor. The data were compiled from non-attendance records kept by management and are available for each of the 101 offices the CME maintained in Swindon.

Table 6.3 shows the bivariate relationship between male militancy and sex composition. The equation provides modest support for an exclusionist interpretation. The male strike rate is negatively correlated with the percentage of female clerks. The coefficient is close to but not quite significant at the .05 level. The equation only predicts 3 percent of the variance, which is an unimpressive performance. Nevertheless, the correlation could be interpreted as supporting a hypothesis of a weak but real relationship between male militancy and sex-type.

However, the bivariate relation very much overstates the role of male exclusionism. This is because male radicalism was coincidently associated with clerical labor intensity. The primary centers of strike support were the shop-floor offices. The major organizers of the General Strike were the blue-collar unionists; the shop-floor clerks would have been swept up in the militant struggles typical of the railway manual labor force. The “office” clerks would have been less exposed to class conflict, their only mobilizing agent being the somewhat more staid Railway Clerks Association.

A second feature of the shop-floor offices is that they were decidedly less clerical-labor-intensive than the administrative offices. The shop-floor clerks were part of work units that were truly non-clerical. They were not on the standard clerical job classification scheme, which effectively meant they could not be promoted into the mainstream clerical hierarchy. They were supervised by blue-collar foremen and superintendents. These foremen would have been primarily responsible for the effective management of the shops as a whole and thus have been primarily concerned with technological issues and the control of the blue-collar labor force. Reducing clerical costs would have been a lower priority. The bivariate relation between militancy and sex composition thus could have been the spurious result of clerical labor intensity.

Informal support for this assertion can be found by considering the technical offices. These are offices whose functions and personnel were primarily scientific. Among them are drafting offices, material inspection sections, and laboratories. The ratio of clericals to scientific personnel in these offices was very small, usually under 10 percent. However, the militancy of these offices closely resembled that of the administrative offices. In the average technical offices, 37.1 percent of the male clerks struck. The comparable figure for the administrative office was 35.1 percent.

The clerks in these offices thus resemble the shop clerks in clerical labor intensity, but resemble the administrative clerks in labor strength. If an employee discrimination model is correct, then the administrative and technical offices should have comparable sex distributions. If a clerical-labor-intensity model is correct, the technical offices should resemble the shops. Table 6.4, which gives the mean percentage of females of the various types of office in the CME, shows that the technical offices were far more similar to the shops than they were to the administrative offices. Both the labs and the shops had clerical forces that were 99 percent male. This is consistent with a clerical-labor-intensity model, but not a union-exclusion model, for there was little exlusionary pressure from the clerks in the labs.

Table 6.5 shows the result of including clerical labor intensity as a control in the bivariate equation in Table 6.3. Actual clerk/staff ratios are not available at this low level of aggregation. Administrative status is used as a proxy for clerical labor intensity. An office is considered to be administrative if it was not technical and was not located on the shop floor.

Including clerical labor intensity diminishes the effect of the strike rate to insignificance. The coefficient shrinks to one-tenth of its former size, representing a negligible impact in absolute terms. Administrative status, however, is highly significant. Overall, the two variables explain nearly half the variance in clerical sex type within the CME, with clerical labor intensity carrying by far the brunt of the explanatory power.

The overall sex ratios of the Chief Mechanical Engineer’s Department did not generally respond to changes in the power of organized labor. Table 6.1 shows that, like the GWR, the sex ratio of the CME was generally fairly constant over time. The decline of RCA strength that came with the collapse of the 1926 General Strike did not significantly change the patterns of sex preference in hiring. Thus, both longitudinally and cross-sectionally, in both the CME and on the Great Western Railway as a whole, variations in the bargaining strength of male workers do not significantly explain variations in the use of women.

A similar pattern emerges when one considers data from the Post Office. If employees with labor power successfully excluded women from employment, then telegraph offices should have been less female than the rest of the Post Office. The signallers’ position as a labor vanguard should have given them superior capacity to restrict entry into employment. Table 6.6 shows the average percentage of women in telegraph and non-telegraph offices at four critical dates in postal labor history: 1881, the year after the telegraphists won their first retroactive pay demand, the Fawcett revision; 1891, the period of the Raikes revision, when the telegraphists obtained a Post Office-wide wage increase; 1906, the date of the Hobhouse Commission, the highwater mark of postal parliamentary agitation; and 1926, a low point in signalling labor strength, the period of the transfer to telex and telephone communications.

The findings in Table 6.6 are precisely the opposite of those one would expect from a labor exclusion model. In each of the three periods of telegraph labor strength—1881, 1891, and 1906—there were more women in telegraph offices than in the less well-organized sectors. The one date in which telegraph offices had fewer women relative to other offices was when their strength was vitiated.

These counterintuitive findings can be explained. As shall be seen in the next chapter, management feminized the telegraphs service very heavily in the early years. Part of this policy was due to a scarcity of skilled labor, requiring the training of new labor sources. Part of it was also due to a misconceived strategy of labor control through gender segmentation. For whatever reasons, the telegraph offices started with substantial complements of women. This presented labor with a fait accompli that was very difficult to reverse.

The increase in the percentage of females in non-telegraph offices is due both to general feminization and to compositional changes in the Post Office. Most offices feminized between 1870 and 1926. Another contributory factor was that the Post Office continued to absorb new functions that were staffed by women. The high percentage of females in 1926 is partly due to the addition of several new telephone offices that had large staffs of women operators.

Changes in labor strength within the telegraph service seems to have very modest explanatory power in accounting for changes in telegraphic sex types over time. The figures on the left side of Table 6.6 could be used to construct a labor-exclusion argument. The percentage of females declines during the years of labor strength and increases during the years of labor weakness. However, this seeming trend is somewhat misleading; the figures are heavily influenced by several very small offices within the sample. There were several elite telegraph offices that were reserved for male workers. This includes Central Telegraph Control, the International Cable Office, and the Intelligence Office, which transmitted news and race results. These offices were filled with lucrative jobs for workers who were at the apex of their careers. They were non-entry-level; they were well-paying; and the actual number of men involved was very small. Most of these offices employed fewer than twenty clerks, while the main offices each employed over a thousand. The elite offices would have been all male regardless of their union policies simply because they contained no entry-level positions. The critical question for a employee-discrimination theory is whether unions were capable of excluding women from mainstream jobs that employers might have been interested in feminizing. Thus, the cases of interest in the telegraph example are the large, centralized, urban signalling stations with their high percentage of entry-level positions. The middle row of Table 6.6 shows the percentage of females in the non-elite telegraph offices: the Central Telegraph signalling galleries, the London East Central Telegraph section, the London District Telegraph offices, and the telegraph departments of Dublin and Edinburgh. Note that when these offices are considered, the weakness of a labor-exclusion model becomes even more apparent. During the years of telegraph union strength, the telegraph offices were more female than their counterparts; during the years of union weakness they were less female than their counterparts. Furthermore, during the years of union strength, they were unable to secure any lasting trend towards de-feminization. Sex-typing was virtually constant during the 1881–1906 period, and whatever minor fluctuations occurred were feminizations just as often as de-feminizations. Although the percentage of women did increase when union strength became weakened, it is not clear which was cause and which was effect. It was the very introduction of female typists in the first place that undercut the monopoly power of the Morse signallers. Thus the mainstream telegraph workers seem to have been unable to control entry into their profession in periods either of strength or of weakness; the only areas that remained all male were offices in which management had little economic incentive to feminize.

Another test of the role of labor power is available. One can estimate the correlation between union activity and sex-typing for a large sample of departments of different types, controlling for exogenous factors that could be obscuring a relation between male militancy and the exclusion of women. No sophisticated measures of union membership or strike participation exist at the department level. Thus, the best available measure of union activity is a simple dummy variable with 1 representing activity and 0 representing passivity. The dummy was coded in such a way as to recreate the historical account given in the earlier section. In addition, organizational histories of postal unions such as B. V. Humphreys’ history of civil service labor relations (1958) were used to find the dates of origin of various early unions. In most cases, postal unions once founded never became inactive. Thus, once some form of organizing activity had been located within an office, it is assumed that such activity continued at later dates.

The N for this analysis is smaller than that presented for other postal equations. This is for two reasons. First, in some cases, such as the Edinburgh offices, the historical record is very thin. It thus becomes impossible to distinguish between a true absence of unionization and gaps in the survival of archival material. Provincial activity where known is included in the sample, but passive provincial cases must be treated as missing data. London, however, is sufficiently well studied to allow for the confident coding of some offices as quiet. Second, after World War I all civil servants came under the coverage of the Whitley Councils. Inter-departmental variance in formal participation in collective grievance machinery virtually disappeared. Thus, interdepartmental tests of the role of union activity are only meaningful up through 1914.

Table 6.7 shows the bivariate regression of sex-typing on union activity by year. Overall, the performance of union activity is very poor. In none of the cases did the regression coefficient attain statistical significance, and in only one year was male militancy capable of explaining as much as 10 percent of the variance. Furthermore, in that year, the sign of the coefficient is in the wrong direction.

Table 6.8 shows regressions of the percentage of females on union activity and the three control variables from the core postal model of Chapter 2. The performance of militancy is as weak here as it was in the bivariate equations. In most of the equations, the coefficient is in the wrong direction, and in no case is the variable statistically significant.³ Longitudinal analyses of these data also fail to uncover a meaningful labor effect. Thus, overall, one would have to conclude from the Post Office materials that cost structure and synthetic turn-

over exerted a far more important influence on sex-typing than either formal or informal campaigns of union exclusion.

Conclusion

Overall, neither the anecdotal or the quantitative evidence support a strong impact of employee discrimination on workplace sex composition. It would seem that managerial responsibility was primarily responsible for the exclusion of women from all-male environments. This is consistent with the anecdotal evidence given in Chapter 2 and with the general good fit of the earlier statistical model.

These findings imply that the attitudes of male co-workers probably had little to do with the determination of sex type in clerical work. These firms represent an unusual high point in clerical union-organizing and sophistication. If these clerks could not make an impact on sex-typing, it is hard to imagine what office workers would have had a better chance.

However, such a finding does not rule out the role of unions in influencing sex issues in the workplace. Many blue-collar workers, particularly those in craft occupations, have had much more control over the labor process than did the unions discussed here. The present findings, in fact, suggest a plausible labor-related explanation that clerical work feminized because of the absence of a body of male workers capable of putting forth effective resistance. Women’s concentration in white-collar occupations is due to the prevailing weakness of labor organization in this sector. The present research can not parse between this explanation and one based purely on managerial utilities. The degree to which blue-collar managers are frustrated feminizers can only be determined by direct study of manual work settings.

This discussion does not exhaust the possible links between class conflict and sex. It is also conceivable that women can be used to reduce class conflict. Sex-based hiring may be determined by the dynamics of the segmentation of the labor force and by managerial reaction to labor’s offensives. The next chapter considers the role of industrial-conflict models that are not based on the assumption of employees’ determining hiring policies.

Notes

1. An example of such concession to the inevitable is the change in the policy of male cigar-makers. Initially vehemently opposed to women workers, they became quite conciliatory to women after feminization had become a fait accompli (Kessler-Harris 1982).

2. The dismissals severely weakened overall RCA negotiating strengths. The RCA had a benefit program that provided substantial relief to any railway clerk who was involuntarily dismissed. Before World War I, this has been a gratuity, since forced resignation was extremely exceptional on the railways. During World War I, the Executive Committee extended this benefit to female temporary employees (over the substantial protest of many local activists). At the end of the war, the RCA had to pay benefits to an enormous number of departing women. In 1919 alone, the RCA paid out over one quarter of its entire assets in unemployment compensation (RCA Annual Reports 1918–20).

3. It should be noted that clerical labor intensity does not fare much better in these equations. To some extent, this is an artifact of the reduced sample; on an N of 18 it is difficult for four variables to be statistically significant simultaneously. However, even the zero-order relation between clerical labor intensity and sex type is weak in these instances. Because the cases for which information on union activity is available is only a subset of the total population of offices, it is thus possible that union militancy might have shown more of an effect in a more complete database.