The Value of Work: The Gendered Outcomes of Organizational Wage Reforms

ABSTRACT

This study explores the influence of organizational practices on gender in/equality in a unique setting: the reformed Israeli kibbutz. The transition of the kibbutz from all members receiving an allowance to waged labor provides an opportunity to explore the impact of wage determination systems on gender pay inequality. The study uses a mixed-method approach: descriptive statistics of administrative data, in-depth interviews, and a focus group with kibbutz management. The findings demonstrate that a “market-based logic,” embedded with gendered preconceptions of “women’s work,” can seep into an organization through a shift in the wage determination system, and increase gender inequality. The findings also highlight the lack of organizational awareness about the gendered consequences of this shift, or even their uncritical acceptance of the “value hierarchy” as ordained by the market. The study thus supports the theory positing gender as a hidden but integral aspect of reward systems in contemporary labor markets.

HIGHLIGHTS

• The choice of a wage determination system is critical for gender inequality.

• “Women’s work” is valued higher by analytical job evaluation systems, which reduce gender biases.

• Market-based wage determinations introduce biases and increase organizational gender inequality.

• Management prefers market-based wage determinations due to competition and costs.

• The “free market” not only obscures gender hierarchies but also legitimizes them.

INTRODUCTION

What constitutes fair pay for a job? How do organizations evaluate and remunerate their workers, and how do such evaluations affect gender inequality within the organization and in the labor market? In this study, we trace the production and intensification of gender inequality in pay following the introduction of a job evaluation process in a unique setting – that of a reforming Israeli kibbutz.

The Israeli kibbutz is a shared community that was based on socialist, egalitarian ideas. Until relatively recently, these communities treated labor as a communal resource. All members received an equal allowance regardless of productivity, profession, job, or work effort. In the past few decades, however, due to a variety of socio-political, economic, and ideological reasons, many of these communities have reformed and commodified the labor of their members, moving from a “budget per membership” to a “wage per job” structure. This reform prompted the need to evaluate the economic value of each occupation and job in the kibbutz. It also introduced pay disparities not only between kibbutz members but also between men and women. Thus, it presents a unique opportunity to explore the organizational-level practices that produce gender inequality in pay, even in an organization that, at least until very recently, espoused equality as a main ideological tenet.

Gender pay inequality is caused by a myriad of interrelated factors. In addition to gender differences in attachment to the labor market and in wage-related human capital characteristics, common factors are gender bias in recruitment and promotion, occupational gender segregation, and the undervaluation of women’s work. These general factors are complementary. Together, they account for the gender pay gap in advanced labor markets, and within reformed kibbutzim (plural of kibbutz). Occupational gender segregation is a well-documented phenomenon in the kibbutz (Palgi 1994; Adar 1996; Palgi and Reinharz 2014; Fogiel-Bijaoui and Sharabi 2017) and a long-standing feature of the labor market at large (Charles and Grusky 2004; England, Levine, and Mishel 2020). In gender-segregated labor markets, pay differences between jobs and occupations are systemically linked to their gender composition, where male-dominated occupations tend to pay more on average than female-dominated occupations, creating vertical occupational segregation (England 1992; Blackburn, Brooks, and Jarman 2001; Cohen and Huffman 2003; Ridgeway 2011; Perales 2013).

Occupational segregation and pay differences between occupations are reproduced within organizations and through organizational practices (Acker 1990, 2006). The former, occupational segregation, occurs through the organization’s hiring and promotion practices that, together with the expectations attached to certain occupations, reinforce gender biases by directing men and women into different jobs and occupations (Acker 1990, 2006). The latter, pay differences between occupations and jobs, are determined either by pay negotiations, or by wage determination systems (Acker 1990, 2006).

Research has documented that the pay gap between men’s and women’s occupations is often driven by the inherent undervaluation of women’s work (England 1992; Chicha 2006). Given that the pay levels of occupations and jobs are often established by wage determination systems, the way in which the undervaluation of women’s work intertwines with wage determination systems is of great importance (see also Koskinen Sandberg [2017]).

The recent economic, social, and ideological transformation of the kibbutz provides a unique opportunity for tracing organizational practices, such as the use of different wage determination methods, which create, foster, or mitigate gender pay gaps. In this article, we analyze the gendered consequences of the implementation of a wage determination system in a case study of kibbutz during its transformation from a “budget per membership” to a “wage per job” structure. More specifically, during this transition process, our case study kibbutz used an analytical point-factor job evaluation system that during the implementation period was augmented and transformed into a market-based wage determination system. Following this process provided us with a window into the role of these systems in creating and entrenching gender pay in/equality.

Our findings underscore the effectiveness of the analytical job evaluation system in reducing gender inequality and shed light on the reasons behind its success in our case study kibbutz. They further demonstrate how during the implementation, a shift to a wage determination system that uses market pricing as a benchmark, introduced gender biases and thus intensified the gender pay gap within the organization. This shift occurred due to the inability of the organization’s management to resist the dominant market forces, and internal pressures to withdraw from the job evaluation system, which was perceived as costly and restrictive. Finally, our findings highlight the lack of awareness of the management and kibbutz members of the detrimental effect of their decision on gender inequality and even the acceptance of this inequality as driven by “market values.”

THE RESEARCH CONTEXT: ISRAELI KIBBUTZ COMMUNITIES

The kibbutz is a long-standing form of communal living founded on the values of the shared ownership of property, the de-commodification of labor, services, and goods, and the democratization of the community’s management (Palgi and Reinharz 2014; Abramitzky 2018). Organizationally, it can be classed as a “collectivist democratic organization” (Sagi 2000) – referring both to the collectivist form of ownership and sharing of resources, and to the democratic organization of the communities’ decision-making processes.

The ideology of the kibbutz was based on the concept of “from each according to one’s ability, to each according to one’s needs.” In practice, this meant that first, each member was expected to work to the fullest of his/her abilities; and second, regardless of their contribution, job, or occupation, all members had equal access to material and immaterial goods: food, housing, clothing, care, education, health services, and communal leisure activities and facilities. Consequently, one’s choice of occupation or job placement was not influenced by the prospective wage, but rather by the preferences and needs of the collective (Sagi 2000; Topel 2005; Abramitzky 2011). Another aspect of this ideology is that the kibbutz aspired to function as an autarky, that is, not to import or export labor. In practice, this aspiration was manifested by minimizing external (and therefore waged) labor, and by preventing members from working outside the kibbutz (Topel 2005; Palgi and Reinharz 2014).

Over the last three decades, a combination of forces, including ideological shifts in the country at large, and the increasingly precarious financial situation of many kibbutz communities, led most of these communities to reform their financial organization from “budget per membership” to “wage per job” (Palgi and Reinharz 2014; Abramitzky 2018). Y. Sagi (2000: 95) summarizes this transition: “If we can describe the reform in one sentence, it seems possible to say: from work as a value to the value of work.” In other words, work was transformed from an unmeasurable collective ideal into a measured and profit-oriented individual task. This transition led to a hitherto unknown phenomenon within kibbutzim – financial inequality between members. Currently, nearly all veteran communities have converted to the new financial system of the “reformed kibbutz” (Getz and Goldenberg 2022).¹

Another significant aspect that influenced the reformation of the kibbutz is the erosion of the ethos of a self-sufficient job market. The possibility of working outside the kibbutz had become viable even before the move to a differential wage system. However, at that stage, kibbutz members who worked outside of the kibbutz still transferred their earnings to the joint pooled income and received the same fixed “budget per membership” as other members (although one can say they did enjoy alternative benefits – such as more options for professional and personal development). In fact, in many less profitable kibbutzim, the salaries of members who worked outside the kibbutz became an important income source for the kibbutz economy (Palgi 2002). In parallel with this “exit” of members to the external market, kibbutzim began to hire nonmember workers in cases where a specific job required expertise unavailable within the kibbutz’s membership. These transitions gradually tied the kibbutz’s internal labor to the external labor market surrounding it.

The cooperative egalitarian ideology of the kibbutz did not eliminate gender inequality (Palgi 1994; Fogiel-Bijaoui and Sharabi 2017). In common with the external labor market, kibbutz employment tended to be gender-segregated. Women were over-represented in service and care jobs (Fogiel-Bijaoui 2009; Adar-Rotem 2013; Palgi and Reinharz 2014), while men were overrepresented in agriculture and manufacturing, as well as managerial positions (Zamir 1999). However, as a socialist cooperative influenced by the ideology of communism, the status and prestige of occupations in the kibbutz were founded on a different set of values than those that pertained to the surrounding labor market. The kibbutz accorded more prestige than the broader society to manual and agricultural labor as well as to the more “feminine” occupations of education and care (Adar 1982). Moreover, because remuneration was not tied to the job, the occupational hierarchy did not translate into a differentiated allocation of economic resources between individuals, or between men and women. However, this financial equality changed quite rapidly following the reform of the kibbutz.

WAGE DETERMINATION SYSTEMS

The kibbutz transition from “budget per membership” to “wage per job” created the need to attach a wage or salary to each job, a mission that required the use of wage determination tools (Gome-Lahav 2013). The complexity of wage reforms and the need to anticipate the social and financial implications of the reform led to the rise of new professions in kibbutzim:² legal advisors, financial experts, and specialized community advisors. Typically, reforming kibbutz societies use an external evaluator to estimate the appropriate wage for specific jobs, and to set the terms for recruitment negotiations and end-of-year job appraisals when necessary. The use of external wage evaluators introduced a sense of transparency, objectivity, rationality, and equality into the process of setting members’ wages – crucial for minimizing grudges, suspicions of favoritism, or discrimination in a sensitive process in close-knit communities.

Two wage determination systems were most commonly used in the preparation for a wage reform: a market-based wage determination system and a “point-factor” evaluation system. According to the Kibbutz National Movement’s Human Resources management, the former is the method most commonly used by kibbutzim. As the name implies, market-based wage determination systems seek to follow market prices. Thus, they are based on wage surveys, often conducted by private consultancy firms or assessors, estimating the earnings of the same or “similar” jobs in a defined market. The resulting estimate can then be further tailored geographically or revised based on the “productivity-related characteristics” of the individual employee filling the position.

This method thus assumes that it is the labor market that sets the real worth of a job through the process of supply and demand (Rynes and Milkovich 1986). However, identifying the appropriate market to survey and measuring the true wages of different jobs using such surveys is not as straightforward as it might seem (Rynes and Milkovich 1986; Sibbald 1993). Furthermore, basing wage outcomes on “productivity-related characteristics,” under the assumption that factors such as work experience are related to actual productivity, works against women, who, on average, are less attached to the market throughout their life course. Finally, and most importantly, proponents of comparable worth would claim that since gender biases, such as the undervaluation of women’s work, are inherent to labor market pricing, copying these wage levels will, by definition, weave these biases into the wage determination system of the organization (England 1992; Acker 2006).

In the case of kibbutzim, this method relies on wage surveys conducted by private assessors, in cooperation with the Kibbutz National Movement. When evaluating a specific job, kibbutz wage evaluators use these surveys to equate the wage level of an occupation in a specific kibbutz with the national average, or the average wage of this job in kibbutz communities of similar size and geographical location. The resulting figure is then personalized using the productivity-related characteristics of the employee, such as seniority and education. Thus, this system, in essence, aligns the organization's occupational hierarchy with the external job market standards.

The second more costly and time-consuming system used by kibbutzim is an analytical job evaluation system, usually termed in the literature a “point-factor” system (Baron and Newman 1990; England 1992; Kilgour 2008). John G. Kilgour (2008: 38) provides a general overview of the stages of such a point-factor system: (1) Listing all of the tasks performed by the job holder (analysis); (2) Combining the aggregate value of these tasks into a single score per job (evaluation); (3) Creating the pay scale using statistical methods (mechanics or architecture); and (4) Continuing the application of the evaluation method to new jobs, and monitoring the wage hierarchy in the organization (maintenance).

Although not originally designed to promote gender pay equity, analytical job evaluations carry the promise of side steeping or even exposing gender biases such as the undervaluation of women’s work (Koskinen Sandberg 2017). Indeed, both the European Commission (EC 2013) and the International Labour Organization (Chicha 2008) have recommended the use of these methods to achieve equal pay for “equal value” jobs. The assumption is that analytical job evaluations link the jobs’ actual tasks to their complexity and to the value these tasks have for an organization, using the same criteria across the board. The contention is that they do not use a benchmark presumably embedded with gender biases, nor do they take into account personal characteristics favoring men as part of the wage determination method.

Nevertheless, even these systems suffer from bias. For example, research has found that the criteria used in some job evaluations are gender biased because they deem certain skills, often held by men, as more important than other skills often held by women (England 1992). Additionally, in a real-world setting analytical job evaluation systems seem to be notoriously hard to fully implement (Acker 1989; Koskinen Sandberg 2017). Therefore, the wage determination system that is ultimately administered in an organization is often very different from the analytical job evaluation system officially established (Koskinen Sandberg 2017).

Before the actual implementation of the reform, our kibbutz case study evaluated all of the jobs in the organization using an analytical point-factor system developed by Noa HR^TM. This method used in-depth interviews with employees and their direct managers that meticulously detailed all of the tasks filled by the jobholders and the value that each of these tasks provides to the specific organization. It is important to note that the product of this interview is a “value based” job description. Thus, the score for each task can change from one organization to another based on its culture and values. A specialized algorithm then scored each of the listed tasks using comparable characteristics defined in the Dictionary of Occupational Titles (DOT). A weighted average of these scores was then calculated to generate a single measure of the complexity and value of each job to the organization. All jobs in the organization were codified on a single scale, enabling their comparison. Finally, in order to equate these scores with earnings outcomes, existing pay data of jobs in the organization were regressed against the evaluation scores. The predicted earnings generated by this regression model provided the final pay estimates. In our case study, since the kibbutz members did not have any earnings at this stage, the algorithm used the earnings of the kibbutz’s external workers in the regression model.

The implementation of “wage per job” in our case study kibbutz was a long and complex process and involved the two systems. After the data and methods section, we describe in detail the gradual transformation from one system to another in the specific context of our case study. We then turn to the gendered consequences of this transformation in the findings section.

DATA AND METHODS

We used a mixed-method approach, involving quantitative and qualitative data sources from a medium-sized kibbutz community (around 1,000 members) located in central Israel.³ The quantitative data include three sets of earnings data of all kibbutz members working in its various businesses. The first is the wage evaluation data produced by Noa HR^TM in 2005 (N = 132) after being commissioned by the kibbutz and prior to the actual application of a differential “wage per job” structure. The two other datasets, obtained with consent from the kibbutz’s accounting department, provide information about the actual earnings from 2010 (N = 172) and 2017 (N = 142) – the former being right after the transition to “wage per job” was completed and the latter being the most recent data available to us.

To provide a benchmark for our case study’s earnings estimates we compared them with estimates of the Israeli labor market surrounding the kibbutz. We based the comparison on a sub-sample of the households located in the central Israel district from the Israeli Expenditure Survey, administered by the Israeli Central Bureau of Statistics. We used the three-time points corresponding with the three data points of our kibbutz case study (2005, 2010, and 2017 with Ns of 3609, 3900, and 5457, respectively).

The qualitative data for our case study come from two main sources. First, we conducted twelve semi-structured in-depth interviews of about an hour and a half each on average with members of the kibbutz case study, recruited using a snowball sample. Nine of these interviews were with current and previous management representatives, who provided detailed and contextual information regarding the choices and justifications that underpinned the reform process. Three additional interviews were conducted with kibbutz members who were employed in the kibbutz at some point between 2010 and 2017. All of the interviews were recorded and transcribed verbatim. The interviewees were not compensated for their participation, and all signed an informed consent form. All names presented throughout are pseudonyms, and other identifying information has been anonymized.

The second qualitative source we use is a focus group we conducted with fifteen past and present members of the management committee of the kibbutz. The participants included two human resources personnel (the current officer and a predecessor who filled the role during the transformation period), managers of the service sector within the community, and members of the senior management team of the kibbutz’s commercial enterprises today. During the focus group, we presented our preliminary results regarding gender pay inequality in the kibbutz in the past and present to the participants. We recorded their responses to the findings and their perceptions regarding the causes and meanings of the processes we identified in the course of our research.

THE REFORM PROCESS IN OUR KIBBUTZ CASE STUDY

Our kibbutz case study was established before the founding of the State of Israel in 1948 and has already celebrated four generations of members. Its central location, along with its development of recreational and commerce facilities, and an industrial concern ensured its financial viability. The community began the process of transitioning to differential pay in the late 1990s, but the process was slow and gradual, taking many years to complete. Internal political and ideological disputes over the social and economic structure of the future kibbutz made it difficult to obtain the 60 percent approval of members required to reform its financial operations.

After voting for a transition in mid 2000s, the method initially chosen for determining the members’ wages was the Noa HR^TM point-factor job evaluation system, described above. This system offered the organization the possibility of creating a wage hierarchy based on the productive value of each job to the organization. This characteristic was appealing at the time when the kibbutz management was trying to achieve consensus for the reform. The evaluation process concluded in 2005 when each member employed directly by the kibbutz received an assessment of his/her projected earnings. However, although each employed member was informed of his/her projected earnings, the kibbutz never implemented the earnings established using the point-factor job evaluation system.

From 2005 to 2010, the period between the official decision to reform the kibbutz’s economic system and its final implementation, members still received a budget rather than a wage or salary. Nevertheless, during this period the management committee of the kibbutz tactically implemented several steps in order to gradually accustom members to the concept of pay for labor and consequently, to the concept of economic inequality between members. Additionally, the management committee’s goal was to motivate members to maximize their earnings potential, in preparation for the full implementation of a “wage per job” system.

Aharon, a senior member of the management committee, recalled this period: “We started with remuneration for weekends and for overtime. All sorts of ploys. That wasn’t so bad, and people accepted it as is. And slowly, slowly, we moved towards the market.” Shlomo, another high-ranking official of the kibbutz at the time, added: “Each change that connected the job value to a member’s earning was another ‘push’ towards the market, so to speak.” Both managers were referring to changes in the financial perceptions and ideology of the members, a consequence of the partial adoption of a wage system. According to Sagi (2000), mechanisms of this nature not only create an incentive structure that encourages members to maximize their wages but also change the ideology attached to work – turning it into a measurable, personalized endeavor, rather than a collective mission.

Another noticeable and important change that occurred during this period was the adjustment of the wage determination system itself, making it incrementally more compatible with market prices. In other words, the management and community gradually shifted away from the analytical point-factor job evaluation system toward a market-based system. As part of the transition process kibbutz workers were entitled to ask for adjustments to their potential earnings. The adjustment process is described in the “Renewal Model Guidebook”⁴ as follows:

Towards the end of each calendar year, an update to the wages of a member who works in one of the kibbutz’s branches will be conducted upon the member’s request and/or the request of the branch manager … an external wage evaluator will meet with the member and his direct supervisor in order to assess the request and any change in the job’s characteristics, including assessing the ability of the branch to accommodate any increase in wages. In updating the wages, the evaluator will consider criteria such as “market value,” professionalism, skills, education, conduct, and the evaluation of the direct manager.

Thus, by 2010, when the kibbutz eventually finalized the transition to “wage per job,” the resulting wage hierarchy was a mixture of the original estimations generated by the point-factor system and reevaluations based on the market value of the jobs. The kibbutz has continued to adjust wages and salaries using the market-based system ever since. Unfortunately, we do not have data regarding the identity of the workers who requested such an update, nor the resulting changes to their earnings. However, as noted above, we do have data for three distinct time points: the wage evaluations produced by the point-factor system of Noa HR^TM in 2005, the actual wages in 2010, after completing the transition and adjusting many of the wages following the workers’ requests, and finally, in 2017, seven years after the implementation and continuous wage adjustments. Since we know that pay adjustments between 2005 and 2017 were tied to the market value and that the vast majority of members did not change jobs, we can conclude that differences between the periods can largely be attributed to the adoption of a market-based wage determination system.⁵

THE KIBBUTZ REFORM PROCESS AND GENDER PAY INEQUALITY

Table 1 displays the median and mean pay, by gender, at each of the stages described above (2005, 2010, and 2017). The left-hand side of the table presents the median and mean monthly pay (in Israeli new shekels) of working kibbutz members (that is, excluding nonmembers working in the kibbutz or members working outside the kibbutz), and the right-hand side of the table presents the same results for the surrounding market of central Israel (the kibbutz’s location). The significance tests for differences between the means and medians within each year are based on regressions that appear in Appendix Tables A1 and A2. Figures 1, 2, and 3 complement these central tendency measures by presenting the full earnings distributions, by gender, for each time point, within and outside the kibbutz (panel (a) and (b), respectively). To measure the overall divergence between men and women’s earnings distributions depicted in each one of the panels of Figures 1–3, we employ the Total Variation Distance (TVD – for more details see Jann [2021]). By comparing the relative density of the distributions, the TVD denotes the similarity (or dissimilarity) between the distributions by quantifying the proportion of data that would have to be redistributed in one distribution in order for the two distributions to completely overlap (Jann 2021: 892).⁶ These results are presented in the bottom row of each year in Table 1.

Table 1.

Mean and median earnings in the Kibbutz compared to the surrounding market, by gender and year

		Case study Kibbutz			Central Israel market			Kibbutz/Central Israel
Year	Gender	Median	Mean	N	Median	Mean	N	Median	Mean
2005	Women	6,757	6,694	68	5,624	7,159	1,852	1.20**	0.94
	Men	7,165	9,361	64	8,596	11,685	1,757	0.83*	0.80**
	Ratio	0.94	0.72**	132	0.65**	0.61**	3,609	1.44**	1.17*
	TVD	0.132*			0.242**			0.55
2010	Women	6,150	7,183	82	5,875	7,555	2,014	1.05	0.95
	Men	8,336	9,821	90	8,587	11,867	1,886	0.97	0.83**
	Ratio	0.74**	0.73**	172	0.68**	0.64**	3,900	1.08	1.15*
	TVD	0.188**			0.235**			0.80
2017	Women	6,617	7,422	73	7,234	9,231	2,775	0.91	0.80**
	Men	9,793	12,167	69	10,634	13,818	2,682	0.92	0.88
	Ratio	0.68**	0.61**	142	0.68**	0.67**	5,457	0.99	0.91
	TVD	0.279**			0.219**			1.27

Notes: All figures are adjusted for inflation (2010 as base); Estimates regrading central Israel are derived from the Israeli Bureau for Statistics Expanse survey of the corresponding year; TVD – Total Variation distance. **, * denote significance at the 5 and 10 percent levels, respectively (see Appendix Tables A1 and A2).

Figure 1.

Wage distribution (percentage of each category from total) in the kibbutz (a) central Israel (b) in 2005

Figure 2.

Wage distribution (percentage of each category from total) in the kibbutz (a) and central Israel (b) in 2010

Figure 3.

Wage distribution (percentage of each category from total) in the kibbutz (a) and central Israel (b) in 2017

To recall, the 2005 data depict the expected mean/median earnings, based on job evaluations produced by the Noa HR^TM point-factor evaluation system, preceding the actual implementation of the wage reform. In contrast, the values in 2010 and 2017 are derived from administrative data of actual earnings.

The first period: wage determination using the analytical job evaluation system

As the 2005 estimates in Table 1 show, had the point-factor system been implemented as is, men's and women's median wages would have been close to parity – an exceptionally equal ratio of 0.94 (the gap is not significant). The ratio between women’s and men’s mean pay is much less equal (0.72, p < 0.01), but it is still much closer to equality than the corresponding ratio in the surrounding market (0.61). As Figure 1 illustrates, according to the wage evaluations made in 2005, there were more men than women at the higher end of the distribution – explaining the difference between the mean (0.72) and median ratios (0.94). However, most of the kibbutz’s employees were located in the middle of the distribution with the peak of both genders aligning – indicative of relative equality in pay between the workers in general, and between genders specifically.

This relative equality not only accords with the egalitarian ideology of the kibbutz institution but also suggests that the job evaluation system, based on task performance scores, resulted in a narrow gender pay gap. Moreover, since occupational gender segregation is prevalent in kibbutzim no less than in the Israeli labor market as a whole (Fogiel-Bijaoui 2009; Adar-Rotem 2013; Palgi and Reinharz 2014), these findings imply that the evaluation given to “women’s jobs” under this evaluation system might be higher than in the market.

A comparison with the labor market surrounding the kibbutz strengthens these conclusions. The wage distribution for men and women in the surrounding market is much more dispersed, right-skewed, and has larger gender gaps. Indeed, while in the kibbutz the dis/similarity between men and women’s distribution (TVD) is measured at 0.132 (p < 0.05); meaning that around 13.2 percent of women or men would have to be redistributed to attain similarity, this figure is almost twice as large in the surrounding market (0.242 – p < 0.01). Additionally, in the surrounding market, the gender earnings ratios were much less equal in 2005, not only between the means (0.61) but also between the medians (0.65). These results indicate that the gender gap was not merely due to the unequal representation of men and women at the top end of the earnings distribution (as in the kibbutz) but across the entire distribution.

Likewise, it is also apparent that while the average evaluated wage of women in the kibbutz in 2005 was very similar to that of women outside the kibbutz (a ratio of 0.94), their median wage was significantly higher (a ratio of 1.2 – p < 0.05), with the peak of women in the kibbutz closer to the middle (around 6,000–8,000_ILS monthly) rather than the bottom (around 4,000–6,000_ILS monthly) of the distribution, as was the case in the surrounding market. This was not the case for the evaluated wages of men in the kibbutz, who earned only about 80 percent of the average and median salary of the men outside it (p < 0.05). Taken together, we note that the gender pay gap in both means and medians was significantly lower in the kibbutz, compared to the surrounding market (by 1859 and 2534 NIS, respectively – p < 0.05).

It seems, then, that under the point-factor system, the wage distribution patterns for both men and women working in the kibbutz would have been quite egalitarian, both within and between the gender groups. Compared to the surrounding market, both wage distribution patterns were much more concentrated around their medians. Women would have benefitted from relatively low representation at the bottom of the distribution, and men would have had a relatively low representation at the higher ranks of the distribution.

Several factors may explain the relative success of this system in mitigating gender inequality and side-stepping gender biases. The first is the focus of the Noa HR^TM evaluation system on the job itself, meaning the actual tasks performed and the skills required to perform them, and on the products created by these tasks, rather than on the qualifications and characteristics of the employees filling the jobs. An evaluation based on the job’s actual tasks brought to light the complexity of the jobs often filled by women on kibbutzim and elsewhere (England and Folbre 1999). Second, all of the employees’ jobs were analyzed using the same analytical system with the same criteria. This eliminated manipulations of managers and power relations between employers and employees that often shape earnings outcomes in organizations in favor of men, even when they officially have a job evaluation system (Ridgeway 2009; Koskinen Sandberg 2017). Third, the higher value that the kibbutz, at the time, gave to care, education, and service jobs relative to other organizations is expressed in the evaluation outcomes (Adar 1982). As noted above, the job evaluation system decomposed each job into its components and then attached a value to each component that reflected its worth to the specific organization.

Despite the successes of this job evaluation system in preserving relative equality, kibbutz officials also noted significant problems with the implementation of the job evaluation system that eventually drove them away from it. Aharon, a senior member of the management committee at the time, described one such problem:

In [our community], in my opinion, we shouldn’t give up on what [another manager] called “the romance.” There is no reason in the world that women should have less than men, and, although challenging, if possible, we should match the wages of the service branches to the “masculine” earnings standard. It is hard because we used to say, “for every wage proposed there should be money in the bank to cover it,” okay? How much are you willing to raise [community] taxes, so that childcare workers will be paid more? When we worked with the [Noa HR] system, the child minder, the kindergarten teacher were at the highest wage levels, because they were doing extremely complex tasks. Educating, shaping souls and minds … And we had a problem. Because at the end of the day, you had to adjust wages to the market, because otherwise … So, on the one hand, as I said earlier, we gave in [to the market] … now, in hindsight, we also kept the gap at bay, so it wouldn’t get out of hand.

This citation supports the claim that the point-factor system better captured the complexity of “women’s work” such as education and childcare tasks (England and Folbre 1999), and thus gave them higher monetary rewards. However, it also demonstrates the inherent tension between the socialist ideology and market forces. The higher rank in the hierarchy of “women’s work” corresponds better with the kibbutz’s values. However, at the same time, the quote reveals a certain lure toward the power of the market, and the inability of the kibbutz’s management to resist it, even when it contradicted their underlying ideology. The management felt pressured to align their pay scale with the surrounding market. On one hand, they needed to pay higher wages to members with lucrative occupational opportunities outside of the kibbutz to prevent a human capital flight from the kibbutz (“brain drain”). On the other hand, they could not afford to pay more to members without many alternative occupational possibilities, such as women care workers, and stay within the budget without raising the community’s taxes.

A second reason given for moving away from the point factor system is the cost and complexity involved in using it. Alon, a male HR manager at the time, said:

The system of “Noa HR” is a great system in my opinion. But it is very hard to upkeep. Once, twice … It is so detailed and transparent. It annoys managers immensely because wage setting is no longer exclusively their realm of authority … it has objective, scientific aspects to it. And, again, you need to upkeep it … It was a valiant effort on our part, but it eventually gave way. It was more worthwhile to move forward and not to invest more time and money in it. To swiftly give the worker [the pay increase] and be done with the process. You stop with this nuisance that takes so much [resources].

Alon’s perception is that an analytical job evaluation system might be a better tool for achieving unbiased wage outcomes that reflect a job’s true value to the organization. However, the continuous upkeep of the elaborate job evaluation system demands a lot of the organization’s finite resources. Moreover, it frustrated the managers because it limited their authority to control pay and use it as a managerial tool. Therefore, its full implementation as an ongoing wage determination system failed and, in general, seemed inoperable.

The second and third periods: shifting to a market-based wage determination system

In 2010, after the kibbutz finally implemented the “wage per job” reform, the picture changed. The first and most evident change was the relatively large increase in men’s median earnings, combined with a (smaller) decrease in women’s median earnings. Consequently, the median gender pay gap increased and became significant (p < 0.01), and the ratio decreased from 0.94 in 2005 to 0.74 in 2010.

It is important to note that, during these five years, the gender earnings gap in Israel actually narrowed slightly (Dagan-Buzaglo, Hasson, and Ophir 2014), as also indicated by the comparison between the results from 2005 and 2010 in Table 1 for employees in central Israel. Thus, the increase in the gender pay gap in the kibbutz was not only substantial but also went against general societal trends. The increase in the median gender pay gap within the kibbutz on the one hand, and its small decrease outside of the kibbutz on the other hand, lowered the differences between the two markets in the median gender pay gap substantially, rendering it no longer significant (p > 0.05).

Figure 2 shows that the earnings distribution in the kibbutz became less egalitarian for both genders. A few women on the kibbutz shifted toward higher wages, but the percentage of women at lower pay levels increased substantially, pulling their median wages closer to women in the surrounding market (the median ratio of 1.05 is still in favor of women in the kibbutz but no longer significant). The distribution of men shows a similar, but even more pronounced process. Thus, while both gender distributions became more similar to the distribution pattern in the surrounding labor market, the process was much more evident among men. Whereas quite a few men moved down from the 2005 value, many others moved up toward the right end of the distribution, skewing the distribution and increasing both the mean and the median. This is expressed by an increase in TVD from 13.2 percent to 18.8 percent, indicating that men’s and women’s distributions grow further apart. As accounts of various kibbutz members and officials attested, these growing similarities to the gendered distributions of the surrounding market are most possibly explained by the ongoing use of a market-based wage determination system during this period.

Even though the gendered distributions in 2010 were much less egalitarian and similar than those produced by the initial job evaluation process in 2005, they were still more similar in terms of TVD and means ratio than the corresponding gender distributions in the surrounding market. However, the market-based adjustment process described above continued from 2010 until 2017, seven years after the implementation of the reform. The most substantial change was the rise in gender pay gaps within the kibbutz, making it even more compatible with the surrounding market. By 2017, the gap between women’s and men’s median and mean earnings had widened further to a ratio of 0.68 and 0.61 (p < 0.01), respectively – almost identical or even lower than estimates of central Israel (0.68 and 0.67, respectively; see Table 1). In fact, by 2017, the gender pay gap within the kibbutz, as indicated by both means and medians (4744 and 3176, respectively), became statistically indistinguishable from the gender pay gap in the surrounding market (4587 and 3401, respectively). Additionally, by 2017 the mean pay gap among women (1808 NIS) favored women working in the surrounding market (p < 0.01), and even median gaps among women turned in favor of the outside market (by 616 NIS a month), although the gap is not significant. Examining men’s results, we find that men working within the kibbutz moved further toward higher earnings, thus increasing their mean pay level substantially and making it quite similar to that of men working outside of the kibbutz. Their median pay also increased substantially, but not as substantially as that of men working outside the kibbutz.

Comparing Figure 3 to Figures 1 and 2 indicates that the process of increasing inequality both within the two gender distributions and between them continued and intensified. In fact, by 2017 the gendered distributions within the kibbutz became less similar than the corresponding distribution in the surrounding market (TVD = 0.279 in the kibbutz, compared with 0.219 in the surrounding market). The larger dissimilarity between the gendered distributions within the kibbutz can be explained by the fact that the earnings distribution of women within the kibbutz remained relatively homogenous, with less women placed at the highest ranks of the distribution compared to the surrounding market, while the distribution of men changed more significantly and became more similar to that of men outside the kibbutz. This greater convergence among men implies that men sought to realize their earnings opportunities in the surrounding market and adjusted their pay accordingly. As the distributions in Figures 1–3 and the differences between the means and the medians in Table 1 indicate, this adjustment to market wages is especially evident in higher wage levels.

The adjustments of earnings to market prices exacerbated earnings inequality between kibbutz members within a rather short time span. However, the resulting earnings hierarchy raised very little concern among kibbutz members. Most of them appear to have accepted the “market authority” and possibly even welcomed it. The following two quotes, the first from a retired man who worked for decades as a senior manager in the kibbutz industry, and the second from a woman working in the service sector, demonstrate this point:

Life here is life without equality. I don’t believe in equality at all. That’s just how it is. But my life physically improved. It improved because, at the end of the day, I earned more money. Fact. I earned more money. And I had more options. And slowly, slowly, one progresses.

I don’t judge those who earn more. I think those jobs demand a lot of responsibility, and [attract] a lot of ungratefulness. A lot of criticism. And the price one must pay in order to earn that money is high. If I had the opportunity to be in that position, if I had to weigh the personal cost against the money [he is earning], I would not want it.

The two interviewees demonstrate how members in very different positions internalized, and even justified, the earnings hierarchy and the inequality produced by the reform process. Our interviewees stated that the introduction of inequality, in effect, improved their economic status or their freedom of choice. Even when their own earnings were relatively low, they saw the reform as increasing their personal choices and rewarding individuals on the basis of their achievements and life choices.

Furthermore, while overall earnings inequality as a whole was perceived as inevitable, it was not necessarily seen as influencing gender inequality. When we presented our preliminary findings to a focus group of members, they were surprised to realize that the increase in overall earnings inequality was also linked to an increase in gender inequality. Confronted with this information, some naively acknowledged this finding. However, others knowingly subordinated the resulting gender inequality to the logic of “value” as seen through the lens of the labor market. The following conversation, which took place in the focus group, exemplifies this point:

Participant: I don’t get it. You said that there is no difference between the earnings of individual men and women in the kibbutz working in the same job.

Researcher: Right. The gap is at the level of the occupations.

Participant: What you are researching is like, sort of, jobs of “comparable worth.” So the question is really about the value we give jobs. That is, for what and not for whom.

Researcher: Right.

Participant: And what we see here is, allegedly, an increasing entry of the labor market to the internal one. That in the market we reward masculine jobs more than feminine jobs.

Researcher: Right.

Participant: So the issue is what position we compensate and not whom.

Researcher: Yes, right.

Participant: Cool. So the question here is not gendered at all; it is a question of values. What we value and not whom we value.

The speaker in this dialogue was a member of the kibbutz‘s committee that was responsible for the process of setting the members’ earnings. In a few short paragraphs, he summarized a very prevalent gender-blind point of view, according to which gender inequality in earnings is attributed to and driven by the “market value” of occupations, rather than by gender biases. Before concluding the meeting, he repeated this understanding, insisting that this is not a gendered question but rather a “value” question. When we suggested in response that, “market values are influenced by gender power relations,” it seemed that this suggestion was intriguing to some group members but baffling for others.

To sum up, a comparison of the three pairs of wage distributions of our case study and the surrounding market over time reveals that the most evident process is a convergence of the two distributions, and the consequent widening of the gender pay gap in our case study. Although time trends can potentially be affected by other occurrences that might impact the gender pay gap during the same time, we should take into account that the widening of the gender pay gap within the kibbutz contrasts with the relative stability (or even small decrease) in the gender pay gap in the surrounding market. This rapid process is not entirely surprising, given the shift toward a market-based determination system, which in itself was seen as inevitable given the growing affinity between the kibbutz’s market and the one surrounding it. What is surprising, however, is that while some somewhat lamented the loss of equality as an essential value of the community, overall, the kibbutz’s members swiftly internalized the unequal consequences of the change, and even sought to justify them. Furthermore, the community’s members did not anticipate, nor did they completely realize, the consequential increase in the gender pay gap. Once confronted with this reality their response was layered and complex. They acknowledged the resulting gender gap as unbecoming and contrary to their ideology. Notwithstanding, they also maintained that these gaps are inevitable due to the unchallengeable power of the market, forcing them to align with it. Finally, some even seemed to justify the resulting gender earnings hierarchy as driven by the unbiased “value” of the occupations.

DISCUSSION: THE IMPLICATIONS OF OUR FINDINGS FOR GENDER IN/EQUALITY

This study explores the way in which organizational practices – in this case, the choice of a wage determination system and the occupational hierarchy it produces – create and reproduce gender inequality in the organizational setting. Job evaluation systems are methods used to align a specific wage hierarchy to an organization’s aims, in order to attach higher wages to jobs with a greater degree of responsibility and that produce more value for the organization (Schwab and Wichern 1983; England 1992; Steinberg 1992). Ideally, such systems should be neutral with regard to irrelevant personal characteristics such as gender or race (England and Kilbourne 1991; Kilgour 2008). However, empirical research has demonstrated that they are often either inherently imbued with gendered biases, based on the lower value attached to “women’s work,” skills and professions, or become tainted with such biases during an erroneous implementation process (Baron and Newman 1990; Steinberg 1990; Koskinen Sandberg 2017).

In our case study, we compared the earnings outcomes of a specific job evaluation system to those generated after switching to a market-based wage determination system, as deployed in a kibbutz society undergoing widespread socioeconomic reforms. We found that in the context of a society that initially sought to promote economic equality between members and certain highly valued “feminine” positions, the job evaluation system produced an earnings distribution characterized by relative gender equality. It yielded a median wage projection for kibbutz women that was 20 percent higher than the median wage of women working in the surrounding labor market and much closer to the median wage of men within the kibbutz (a ratio of 0.94, relative to a ratio of 0.65 outside the kibbutz).

We can propose three explanations for this result. First, the outcomes of the job evaluation system – as opposed to the wage determination system relying on market pricing – are based on the actual tasks done. Thus, it captured the complexity of care jobs and other “women’s work.” As such, it reduced biases related to the undervaluation of these jobs (England and Folbre 1999; Koskinen Sandberg 2017). Second, its output ranked all jobs in the organization strictly according to the same analytical criterion – the benefit of the job to the organization – without being affected by power relations in the organization such as pressure by employees’ groups or adherence to senior or middle management requests. Third, the job evaluation system incorporated in its apparatus the higher regard the kibbutz’s culture at the time held for “women’s work” relative to the surrounding society (Adar 1982).

However, when market-based determinations were added to the wage determination system, these “feminine” jobs lost their status, and therefore their relative wage parity with “masculine” jobs. The kibbutz management justified the transition to a market-based wage determination system as driven by necessity, resulting from a combination of two factors. The first was economic forces that affected the kibbutz’s labor market following its intensifying intertwining with the surrounding market and the occupational possibilities it offered to (mostly men) members of the now-reformed kibbutz. Second, the cost of the job evaluation system – which created tensions among the managers because it limited their tools for supervising personnel and was inherently complex and expensive to maintain – was too burdensome.

Our findings highlight several important points. First, they show that the earnings distribution of men within the kibbutz changed dramatically following the shift in the wage determination system, becoming more similar to the earnings distribution of men outside the kibbutz, and less similar to the earnings distribution of women in the kibbutz. The earnings distribution generated by the job evaluation system was relatively homogenous, with a large percent of working men in the kibbutz concentrated around the median. After only five years of earnings adjustments using a market-based system, the distribution became much more dispersed. By 2017, the percentage of men in the right tail of the distribution doubled. In our case study kibbutz, as in other kibbutzim (Orchan, Palgi, and Adar 2016) and society at large, most of the highly paid managerial positions were occupied by men. As their earnings became more and more aligned with the pay levels of such positions in the surrounding market, they rose relative to the rest of the distribution, and so did gender inequality. Within the vast literature exploring the mechanisms that create gender pay inequality (for example, Blau and Kahn [2017]), our findings underscore the importance of firm or organizational-level practices and decisions (Acker 2006), which are often not even perceived as related to gender pay inequality by the social actors themselves.

Second, our findings demonstrate how the “invisible hand” of the market generates and reproduces gender inequality in earnings. Specifically, they show how the undervaluation of “women’s work,” prevalent in the “free market,” seeps even into an equality-espousing organization, and its unequal consequences (when acknowledged) are seen as inevitable. In this regard, it provides additional support to the well-documented feminist argument that gender biases are inherent in contemporary labor markets, even though these biases are frequently not evident (Acker 1990, 2006; Koskinen Sandberg 2017; Mandel 2018). Furthermore, because these gendered biases are often hidden, they are often unacknowledged. We show that even though the transition to market-based wage determinations almost immediately increased the gender pay inequality in the kibbutz (the median gender ratios widened systematically, from 0.94 in 2005 to 0.74 in 2010 and to 0.68 in 2017), the kibbutz’s management and its members generally did not acknowledge these outcomes. While the overall increase in earnings inequality was foreseen, and even embraced, the rise in the earnings gap between men and women came as a surprise. Moreover, when we confronted kibbutz members with this finding, some interpreted it as “a value question” rather than “a gendered question.” This blindness to the connection between the gender composition of an occupation and its socially constructed value encapsulates the essence of the undervaluation bias and its complexity.

The third and related point is that kibbutz members came to regard the “market authority” as an unavoidable reference point, in effect justifying even the emerging gender earnings gaps. Our findings show that while the kibbutz members still conceptually endorsed a “civic equality” style of justification historically championed by Israeli kibbutzim, in order to justify the emerging inequality between members, many came to utilize and accept “free market” justifications as superseding “romantic” notions of civic equality (see Boltanski and Thévenot [2006]). This change was apparent in the initial stage when members voted for the transition to a “wage per labor” system as a whole, then later when they moved to a market-based determination system, and then when some blindly assumed that the value of an occupation is driven by unbiased economic laws. Thus, although the growth in the gender pay gap was remarkable and unanticipated, the “free market” justification legitimated their implications and was even used to justify them when explicitly criticized. Our conclusion, therefore, is that the market, as a hierarchy-setting mechanism and a source of justification, not only obscures gender hierarchies but also legitimizes them by characterizing these hierarchies as “value driven,” and thus justified in a profit-oriented economic system.

Earnings inequality has been at the center of feminist struggles for more than half a century, a fight that has evolved in stages: first in legislation against pay discrimination based on individual characteristics such as gender, race, and sexuality; and second in legislation against pay discrimination among jobs with “comparable worth” (O’Reilly et al. 2015). In Israel, legislation regarding the former point has been in place since 1964 (Equal Pay for Male and Female Workers Act, 1964), and enhanced to include “comparable worth” in 1996. However, the latter posed a significant challenge to the legal system, as there is no objective criteria to assess the value of occupations and to determine conclusively the worth of one profession in comparison to another (Hartmann, Roos, and Treiman 1985; England and Kilbourne 1991; England 1992). Since the evaluation of the worth of a job or occupation is a complex and often elusive task, discrimination based on structural mechanisms continues to pose a great challenge to courts in Israel as in other countries (Renan-Barzilay 2017).

The bleaker conclusion that emerges from our findings is that a separation from the surrounding market is nearly impossible. Even in a democratically managed, worker-owned cooperative, defined by its overt egalitarian ideology, initial aspirations of equality and fairness were eroded in the face of the “free market.” Our findings therefore demonstrate the domination of market forces, and the inability of organizations to resist them. With respect to gender, they highlight the unexpected and obscured gendered consequences of the most prevalent wage determination method – market-based pricing.

We therefore hope that organizations will not only recognize these implications but also act explicitly to mitigate gender bias, by fully implementing analytical job evaluation systems that have the potential to reduce gender inequality. In a labor market riddled with gender bias, organizations are required to methodically and purposefully formulate mechanisms capable of acknowledging that female-dominated occupations are invariably penalized and to invest resources to minimize it.

Biographies

Hadas Mandel is Professor at Tel Aviv University, and the Principal Investigator of the ERC-funded (Horizon 2020) ‘Struct.vs.Individ’ project. Her published work focuses on the intersection between gender, class, race, and social policies, and the complex and seemingly paradoxical implications of structural and institutional aspects on women’s economic attainments.

Amit Lazarus is a PhD candidate in the Department of Sociology and Anthropology at Tel Aviv University. His main research interests are structural mechanisms of inequality in a variety of socioeconomic resources and among various social groups. His current research examines gender inequality in unpaid work and labor outcomes.

Adi Moreno is a gender studies lecturer at the Tel-Aviv-Yafo academic college. Her research interests involve gender inequality, sexuality, and evolving forms of familial living. Adi received her PhD in 2016 from the University of Manchester. Her research project followed the development of new parenting norms among the gay community in Israel. Her current project addresses the social and economic changes in the reformed kibbutz.

APPENDIX.

Table A1.

Regression of gender and working place on monthly pay

	OLS on monthly pay			Median quantile on monthly pay
Variables	2005	2010	2017	2005	2010	2017
Constant	6,694.23**	7,183.14**	7,422.31**	6,754.66**	6,214.83**	6,617.11**
Men [Women]	2,667.03**	2,637.83**	4,744.41**	437.47	2,193.76**	3,176.11**
Surrounding market [Kibbutz]	464.58	372.27	1,808.65**	−1,130.54**	−339.83	616.58
Men*Surrounding market	1,859.49*	1,673.65*	−157.75	2,534.14**	518.24	224.44
N	3,741	4,072	5,599	3,741	4,072	5,599

Table A2.

Regression of gender and working place on monthly pay.

	OLS on monthly earnings			Median quantile on monthly earnings
Variables	2005	2010	2017	2005	2010	2017
Constant	11,685.32**	11,866.89**	13,817.62**	8,595.73**	8,587.00**	10,634.23**
Women [Men]	−4,526.52**	−4,311.48**	−4,586.65**	−2,971.62**	−2,712.00**	−3,400.55**
Kibbutz [Surrounding market]	−2,324.06**	−2,045.92**	−1,650.90	−1,403.60*	−178.42	−841.01
Women*kibbutz	1,859.49*	1,673.65*	−157.75	2,534.14**	518.24	224.44
N	3,741	4,072	5,599	3,741	4,072	5,599

Notes: **, * denote significance at the 5 and 10 percent levels, respectively. The differences between Appendix Tables A1 and A2 are only the reference categories of gender and work place used. Toggling between them enables to test all possible differences.

Funding Statement

This work was supported by H2020 European Research Council under the European Union's Horizon 2020 research and innovation program [Grant Number 724351] and the Israel Science Foundation (ISF) [Grant Number 1111/22].