Introduction

In recent decades, Kenya has witnessed a remarkable surge in female labor participation and employment rates, driven by a confluence of multifaceted factors. Demographic and cultural shifts have paved the way for greater acceptance and empowerment of women in the workforce, while the diversification and evolution of Kenya's economy have created new employment opportunities. Moreover, the promulgation and implementation of the 2010 constitution has played a pivotal role in promoting gender equality and supporting initiatives aimed at enhancing women's economic participation. Parallel to this, the realization of educational gender parity has equipped women with the necessary skills and qualifications to compete in the labor market.

The rise in female employment rates has been accompanied by constant changes in the country’s occupational and industrial structures. There has been a transition in the labor market from predominantly formal employment to a prevalent informal sector (Kenya National Bureau of Statistics [KNBS] Economic Survey, 2023; Abdiaziz & Kiiru, 2021; UN Women, 2023). The total employment outside small-scale agriculture and pastoral activities in Kenya increased from 18.3 to 19.1 million from 2021 to 2022 with a total employment rate of 65.3% for individuals aged 15–64. However, the employment rate for women (60.3%) is lower than that for men (70.4%). Looking at the occupational and industrial structures, women have a higher representation as professionals (10.9% women, 8.4% men), technical and associate professionals (15.5% women, 10% men), service and sales workers (22.3% women, 13% men), and elementary occupations (34.2% women, 32.9% men). Conversely, men dominate plant and machine operator and craft and related worker roles (KNBS Economic Survey, 2023; UN Women, 2023). These employment trends mask significant gender disparities within specific sectors and occupations which help to explain the prevalence of gender wage inequality in the Kenyan labor market despite the advancement in women's economic participation. The present study attempts to examine these phenomena and contribute to the literature by examining the role of occupational and industrial segregation in explaining the gender wage gap in Kenya. An understanding of the factors driving male-female wage differentials is relevant since it helps direct public policies and evaluate if previous efforts to reduce gender wage inequalities have been rewarded.

The literature has sought to explain the well-known discrepancy between the average wages of men and women with a focus on the issue of unequal pay for equal work (Blau & Khan, 2017; Firpo et al., 2009; Oaxaca, 2007; Machado & Mata, 2005). However, there is a growing body of evidence that gender differences in occupational and industrial distribution comprise an equally important source of the aggregate wage differentials (Orraca et al., 2016; Ismail et al., 2017; Goy & Johnes, 2012; Demoussis et al., 2010; Khitarishvili et al., 2018). Although the literature recognizes the importance of these drivers of the gender wage gap, most studies focus primarily on the former. As a result, we do not yet have reasonable measures of the relative size of the wage gap resulting from occupational segregation, nor can we identify the effects of discrimination that remain after accounting for characteristics that help to explain both wages and occupational distribution. This paper attempts to fill in these gaps by decomposing the male-female wage differential into occupational and wage components in the Kenyan context.

Most studies that have examined gender wage differentials in Kenya (Kabubo-Mariara, 2003; Agesa et al., 2009, 2013; Omanyo, 2021; Abdiaziz & Kiiru, 2021, UN women, 2023) have done so under the framework developed by Oaxaca (1973) and Blinder (1973) and its variants (Firpo et al., 2009; 2018), focusing on the issue of wage discrimination or unequal pay for equal work. For the “index number” problem in Oaxaca-Blinder type decompositions, the discrimination effect is calculated as the unexplained difference between two reduced-form wage equations estimated separately for men and women. However, this conventional approach does not account for occupational distribution between men and women and assumes that occupational choices between men and women are exogenously determined. Consequently, if occupational choice is subject to labor market discrimination, then these conventional approaches would be inappropriate; in other words, the OB approach fails to adequately separate the effects of wage/job discrimination from occupational segregation (Liu et al., 2004). To address this, Brown, Moon, and Zoloth (1980) proposed an alternative framework treating occupational choice as endogenously determined and decomposing of the gender wage gap into within-occupation and between-occupation differences. The BMZ framework further decomposes the within- and between-occupation components into effects attributable to differences in observable productivity characteristics and differences in estimated coefficients. Specifically, the between-occupation effects provide a direct estimation of the impact of occupational segregation on the gender wage gap (Brown et al., 1980; Orraca et al., 2016; Ismail et al., 2017; Goy & Johnes, 2012; Teo, 2003; Demoussis et al., 2010; Khitarishvili et al., 2018).

This paper conducts a novel examination of the role of occupational segregation in explaining the gender wage gap in Kenya, using data from the Kenya Continuous Household Survey (KCHS-2021). Given the limited research on this topic in Kenya, international comparisons are made to gauge the relative magnitude. This study distinguishes itself in three ways. First, it constructs a dissimilarity index to objectively measure the differences in occupational and sectoral structures between males and females. Second, it treats individuals’ occupational/industrial attainment as an endogenous variable using observable characteristics of males and females to determine occupation selection.  Thus, we simulate an occupational distribution for female workers, assuming they face the same occupational attainment structure as men. Third, it investigates the separate effects of within- and between-occupation differences by adopting Brown et al.’s (1980) decomposition technique. To account for the potential bias introduced when occupation is incorrectly included and address sample selection bias of self-selectivity, a reduced-form multinomial model is employed to estimate the predicted occupational distributions, considering that securing a job may not be an individual’s only priority. Therefore, this study addresses the following research questions.

1. Do the gender differences in occupational and industrial structures drive the wage disadvantage experiences by women in Kenya?

2. What is the extent of divergence or differences in occupational, industrial, and sectoral structures between men and women in Kenya?

Literature Review

The literature on the gender wage differentials in Kenya is limited but growing. Studies have proved that women earn less than men, and the magnitude of the gender wage gap fluctuates depending on the period of study and the coverage of the data source used. Maina (2021) conducted a study in Kenya on the impact of occupational segregation on the gender wage gap. The study, employing data from the 2019 Quarterly Labor Force Survey, found a substantial gender wage gap, with male workers earning approximately 58.8% more than their female counterparts. Moreover, the study assessed segregation across occupations using the Duncan index and found that 42.73% of women would need to switch jobs for full integration to occur.

Abdiaziz and Kiiru (2021) conducted a study in Kenya using data from the 2013 World Bank Skills Towards Employability and Productivity Survey to examine gender pay gaps across industries. The study analyzed the data using the Mincerian earnings framework and Oaxaca–Blinder decomposition. They found that men’s wages were 27.2% higher than women’s in the commerce and trade sector. In the services sector, men earned 28.5% more than women and, in the manufacturing, and construction sector, men earned 23.1% more than women. In the agriculture, fisheries, and mining sectors, 57.9% of the wage difference was attributed to human capital characteristics.

Chakraborty (2020) examined gender pay differentials in India across the private and public sectors using data from the Periodic Labor Force Survey for 2018–2019 by employing the Oaxaca–Blinder decomposition approach and the Brown–Moon–Zoloth technique for robustness checks and comparison to identify factors contributing to the gender wage gaps in both sectors. The findings confirmed significant gender pay differentials resulting from occupational discrimination. Specifically, the study revealed that rural women faced more significant wage gaps across occupations than urban women. Furthermore, the study argued that eliminating occupational discrimination from the labor force could reduce average wage differentials by 57% in rural and 67% in urban areas.

Ismail et al. (2017) conducted a study in Malaysia on gender occupational segregation and the wage gap. They used data from the Malaysian Working Households 2011 survey and identified that women’s participation in the labor market has increased. However, occupational segregation and wage disparities between men and women persist. The study employed the wage decomposition model proposed by Brown et al. (1980) to analyze the factors contributing to gender-related wage differentials. The findings indicate that within-occupation differences account for the most significant portion of the wage gap between men and women. Furthermore, the study underscores the significance of wage discrimination within occupations as contributing to the gender wage gap.

Orraca et al. (2016) conducted a study in Mexico on the role of occupational segregation in explaining the gender wage gap. Using census data and applying the Brown et al. (1980) decomposition technique, they determined that the wage differentials between men and women increased between 2000 and 2010. In both years, within-occupation wage differentials largely contributed to widening the gender wage gap, while between-occupation wage differentials had the opposite effect. Notably, the within-occupation wage differentials were largely driven by the unexplained component, implying that differences in the average returns to productivity-related characteristics within occupations are the main contributors to the gender wage gap. Notably, despite the notable differences in the occupational structure of male and female workers, occupational segregation did not exacerbate the gender wage gap. Women do not encounter significant barriers to accessing high-paying occupations.

To sum up, empirical studies investigating gender wage disparities present varied findings regarding the influence of occupational segregation on the gender wage gap. Some studies find that occupation segregation contributes to widening the gender wage gap (Herrera et al., 2019; Chakraborty, 2020), while others suggest that there is minimal or no effect (Orraca et al., 2016). In the case of Kenya, limited research exists on gender wage differentials that consider occupational disparities. To the best of our knowledge, only one study in Kenya (Abdiaziz & Kiiru, 2021) estimates interindustry wage differences without detailed decomposition while not explicitly examining occupation-specific wage functions in relation to gender wage differentials. Therefore, further investigation is necessary for a better understanding of the issue of gender occupational wage disparities in Kenya.

 

Methodology

The conventional decomposition of wage disparities typically focuses on wage discrimination (Oaxaca, 1973; Blinder, 1973; Nuemark, 1988; Cotton, 1988; Oaxaca & Ransom, 1994; Machado and Mata, 2005; Firpo et al., 2009), overlooking the variations in occupational choices between men and women (Brown et al., 1980; Meng, 1998; Orraca et al., 2016). Consequently, these approaches assume that the factors determining wages and occupational selection are the same. However, if additional factors, such as discriminatory barriers to entry, influence occupational status, these approaches may underestimate the discrimination component (Brown et al., 1980). To address this limitation, Brown et al. (1980), hereafter BMZ decomposition, expanded the model and incorporated the distinction between wage differences across-occupation (industries) and within-occupation (industries) into the analysis of wage differentials. The BMZ decomposition can be expressed as follows:

​

​

​

​

​

​

​

​

​

​

​

​

Where  represent the natural logarithm of the hourly wage for an individual  employed in occupation  which denotes the total number of occupational groups being considered,  represent the individual’s characteristics,  is a vector of wage coefficients specific to occupation j to be estimated. Moreover,  and  represent the average natural logarithm of male and female wages, respectively, hence, gross logarithmic gender hourly earnings differential ( and  and indicate the proportions of male and female workers employed in occupation category j.   represents the hypothetical occupational distribution structure for women under the assumption that they face the same occupational distribution as men. This predicted occupational distribution for female employees is generated from female characteristics using male occupational attainment as the non-discriminatory norm. Hence, the difference between the occupational distribution of actual males and predicted females is simply the non-discriminatory differences arising from male and female productivity-related characteristics (Sung et al., 2001). The computation of non-discriminatory occupational distribution for female employees requires a model of occupational attainment to be estimated. Brown et al. (1980) suggested a reduced form multinomial logit model to capture how various variables influence the probability of an employee i working in occupation j. We follow this approach. Goy and Johnes (2012) point out that gender differences in characteristics alone are insufficient in explaining why women tend to be concentrated in low-paid occupations. To address this issue, we employ a methodology (Heckman, 1979) that considers sample selection bias when estimating wage regressions for both men and women to generate the inverse Mills ratio .

In accordance with Brown et al. (1980), the first component on the right-hand side of Equation (1) is referred to as the within-explained (WE) component and it represents wage differences resulting from gender disparities in average characteristics within-occupations. The second component is the within-unexplained (WU) which captures the wage effect resulting from vertical or hierarchical segregation within-occupations. It captures unexplained differences in the occupational achievement structure or the unobserved factors that contribute to the under- or overrepresentation of certain groups in higher-paying positions within specific occupations (Salardi, 2013). The third component is known as the between-explained (BE) and it accounts for wage differentials arising from variations in participation shares between-occupations. The last component is the between-unexplained (BU) term represents the wage effect of horizontal or occupational segregation between-occupations, providing insights into the equality of access between men and women to different occupations (Liu et al., 2004). The ‘unexplained’ term refers to wage differentials that cannot be accounted for based on productivity endowments and is commonly interpreted as a measure of labor market discrimination.

Data Description.

This study uses data from the 2021 Kenya Continuous Household Survey (KCHS) collected by the KNBS. The KCHS-2021 survey covered 17,042 households and 68,677 individuals within the households. Our sample is restricted to workers aged 15–65, which is considered the working age range in Kenya in accordance with the Employment Act of 2007 (Omanyo, 2021). Following these restrictions, we have a final sample of 6,653 waged employees, including 4,210 male employees and 2,443 female employees, aged between 15 and 65 years. This study’s decomposition of the gender wage gap takes the term “wages” to refer to gross income from waged employment, covering wages, salaries, and other earnings, including allowances, received in the past month. Table A1 (see Appendices) presents the main variables applied in the analysis while Table A2 describes the nine occupational groups which are defined according to Kenya Standard Classification of Occupations (KeSCO-2022, State Department for Labor and Skills Development).

Results and Discussions

The Duncan index of dissimilarity shows that approximately 8.9% of women would need to change jobs across employment sectors to obtain the same sectoral distribution as men. Furthermore, approximately 37% and 30% of women would need to change occupations and industries of work to have identical occupational and industrial distribution as men. Notably, the dissimilarity index for occupational segregation in Kenya has decreased over time. Maina (2021) reported a value of 0.4273 based on the 2019 Quarterly Labor Force Survey.

​

Table 1: Dissimilarity Index for KeSCO-2022 Occupations, Industrial classification, and Sectors of employment.

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

The first stage of the Brown et al. (1980) decomposition requires the estimation of K occupation specific wage regressions for male and female workers as defined in Eqns. (1) and a multinomial logit occupational attainment model. While we do not present the first-stage estimations due to space constraints, our focus is on the wage decomposition results; these results are available upon request.

​

The Brown-Moon-Zoloth decomposition.

Table 2 and Figure 1 present the findings of the BMZ wage decomposition between men and women disaggregated by occupational segregation and place of residence. The results demonstrate that aggregating data for all the occupations categories, a mean wage differential of 0.1144 log points exist between male and female employees. That is, female employees earn approximately 87.9%[1] of men’s monthly earnings, indicating a gender wage gap of 12.1% in the Kenyan labor market. Notably, the gender wage differential has declined since earlier studies were conducted in Kenya. Previous studies reported mean log wage differentials of 0.2 and 0.6 log points in the public and private sectors, respectively (Omanyo, 2021), suggesting a positive trend toward reducing the gender wage gap in Kenya, potentially attributed to government policies to close the wage gap.

Table 2: BMZ wage decomposition between men and women by occupational segregation.

 

​

​

​

​

​

​

​

​

​

​

​

​​

​

​

​

​

​

​

​

​

​

​

​

Several notable points arise from the BMZ decomposition results. First, the gender wage gap is primarily driven by gender differences within-occupation. The sum of WE and WU, which equals 0.206 log points, or 106.6%, indicates that the gender differences within-occupation account for the entirety of the observed gender wage differentials. However, the sum of BE and BU equals −0.0972 log points, accounting for (−) 85% of the wage differentials, while only 5.2% of the differential is due to sample selection bias.

Nonetheless, analyzing these components of the total wage differential shows substantial heterogeneity in their effects. Explicitly, the WE component (due to productivity-related characteristics) decreases to 0.0487, accounting for 42.6% of the total observed wage differentials. The result indicates that nearly half of the wage gap is due to gender differences in endowments within occupations that positively contribute to widening the wage gap. This implies that based on their average productivity-related characteristics within occupations, men have better endowments than women, ceteris paribus, and this gender difference increases the wage gap.

Figure 1: BMZ wage decomposition between men and women by occupational segregation.

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​​

​

​

The WU component stands out prominently, rising to 0.157 log points and accounting for approximately 138% of the total wage differentials, all else being equal. This finding indicates that the largest share of the observed gender wage gap is primarily due to the unexplained component of the within-occupation, suggesting that the wage gap is largely driven by vertical segregation, that is, by gender differences in returns to average productivity-related characteristics within each occupation. Consequently, this signals that based on the returns to their average characteristics within each occupation, the size of the actual gender wage gap would have been more extensive than what is observed. This finding is consistent with Ismail et al. (2017) and Orraca et al., (2016), who found that a larger portion of the gender earnings differential in Malaysia and Mexico was attributed to unexplained factors within occupations.

Second, the BE term is −0.141 log points, accounting for (−) 123% of the observed wage gap, while the BU term is equivalent to 0.0438 log points or 38% of the total wage gap. Notably, we observed that the effect of the BE component is greater than the BU component; that is, the BE component is not offset by the BU effect. This finding indicates that on average, women are allocated in better-renumerated positions than men across occupations. Although the BU component is positive, indicating a positive effect, that is, increasing the gender wage gap, surprisingly, this effect is offset by the BE component; thus, women encounter fewer barriers to entry into occupations. If women had the same occupational attainment structure and choice as men, their shift to higher-paying occupations would completely curb the gender wage gap, implying that occupational segregation partially favors women in the Kenyan labor market. However, we caution against completely ruling out horizontal occupational segregation in the Kenyan labor market since the BU component is positive, suggesting the existence of barriers to entry into higher-paying occupations for women. 

Third, the sum of WE and BE terms (total explained), capturing gender differences in characteristics within and between occupations, equaled −0.0923 log points, suggesting that according to their average characteristics, women should have had higher earnings than men. In other words, women possess better labor market endowments, and based on these average endowments, the gender wage gap would significantly reduce by 81%. However, the sum of the WU and BU, which captures differences in coefficients to endowments, equals 0.2008 log points or 176% of the observed gender wage gap, signaling that the size of the gender wage gap should have been larger than what is observed, according to the returns on their average characteristics. Nonetheless, when separating the two unexplained components, a different pattern emerges: the WU component is equivalent to 0.157 log points, while the BU term is 0.0438 log points. Notably, the effect of BU is offset by the BE component, while the WE component does not offset the WU term. Thus, we can conclude that the returns to average characteristics, i.e., unexplained factors associated with discriminatory practices (e.g., vertical segregation) largely drive the wage differential in Kenya within each occupation. Moreover, women somehow encounter fewer entry barriers across professions. The higher contribution of the total unexplained portion may suggest less effective legislative controls in the Kenyan labor market to curb discriminatory practices in the pre- and postentry labor market.

The gender wage gap varies significantly between urban and rural localities, being higher in rural areas, at 0.246 log points, compared with 0.0273 log points in urban areas. In metropolitan regions, the combined effect of the WE and BE components accounts for −0.037 log points or −136% of the gender wage gap. This suggests that based on their average characteristics, women in urban areas have better endowments, leading to a decline in the wage gap. However, the WU and BU components total 0.0614 log points or 225% of the wage gap, indicating that the size of the wage gap would be larger based on the returns on their average characteristics. Moreover, the combined WE and BE components account for −0.753 log points or −306% of the wage gap in rural areas, indicating that most of the wage gap in rural areas would be eliminated due to differences in average characteristics within and between occupations. Conversely, the WU and BU components total 0.991 log points or 403% of the gender wage gap, suggesting that the size of the wage gap would have been even more prominent based on the returns to their average characteristics.

The results indicate that the effect of “explained factors” (WE and BE) in reducing the gender wage gap is more pronounced in rural areas, while the impact of “unexplained factors” (WU and BU) is less significant in urban areas. Notably, the urban BU component is −0.0456 log points, indicating that on average, women in urban areas are allocated to better-remunerated positions than men. Subsequently, the positive WU component in rural and urban regions suggests that the wage differential is largely driven by vertical segregation.

Finally, the selection term differential, which is attributed to sample selection bias, is 0.0059 log points (5.2%), 0.0029 log points (10.6%), and 0.008 log points (3.3%) for the entire sample, urban, and rural localities, respectively. The selection correction terms are favorable, indicating the wage offer gap is slightly smaller than the total wage gap due to the selectivity bias in occupation choice. The findings regarding the contribution of sample selection bias in this study appear to be more realistic than those of Goy and Johnes (2012) and Ismail et al. (2017), who reported that sample selection bias accounted for 63.1% and 35.8% of the total wage differential, respectively.

Conclusions

The literature on the gender wage gap in Kenya typically attributes the wage gap to gender differences in productivity-related characteristics and returns to endowments. However, wage determination may be altered by a sorting mechanism that slots workers into various occupations according to their skills and gender. Moreover, occupational choices are not exogenous but rather endogenously determined. Employing the Brown et al. (1980) decomposition technique, this study examined the role of occupational segregation in explaining the gender pay gap in Kenya. The analysis is based on the 2021 KCHS data, comprising 17,042 households and 64,677 individuals.

The findings are remarkable and novel. The results show that approximately 8.9%, 37%, and 30% of women would need to change jobs across sectors of employment, occupations, and industry of work, respectively, to achieve the identical sectoral, occupational, and industrial distribution as men. Based on decomposition results, men on average have higher log monthly earnings than women, and the pay gap is significantly greater in rural regions than urban localities. Notably, since within-occupation pay differentials are largely driven by the unexplained component, we conclude that the gender page gap is primarily a product of differences in average returns to endowments within occupations. Moreover, while the results reveal that male workers’ occupational and industrial structures differ considerably from those of their female counterparts, occupational segregation does increase the gender page gap since women appear to experience barriers to high-paying occupations.

Based on these results, we recommend that for the pre-labor market entry policies, the monitoring mechanism could be applied to prevent gender discrimination in hiring, working to the earnings disadvantage of female workers being recruited for similar occupations as men. In all establishments, affirmative action laws should be implemented, requiring human resource departments to prepare regular reports on the hiring process, detailing the job positions by gender, and rewarding firms and enterprises that do not discriminate hiring. Legal compliance with the affirmative action labor law should be among the criteria for business license renewal and eligibility for tax exemption incentives by county and national governments in Kenya.

References

Abdiaziz, A. and Kiiru, J., 2021. Inter-Industry Gender Wage Differentials in Kenya. Gender Statistics for Evidence-Based Policies, p.10.

Agesa, R.U., Agesa, J. and Dabalen, A., 2013. Sources of the persistent gender wage gap along the unconditional earnings distribution: Findings from Kenya. Oxford Development Studies, 41(1), pp.76-103.

Blau, F.D. and Kahn, L.M., 2017. The gender wage gap: Extent, trends, and explanations. Journal of economic literature, 55(3), pp.789-865.

Blinder, A.S. (1973). Wage Discrimination: Reduced Form and Structural Estimates. The Journal of Human Resources, 8 (4), pp. 436–455.

Brown, R., S. Moon, M. and Zoloth, B.S., 1980. Incorporating occupational attainment in studies of male-female earnings differentials. Journal of Human Resources, 5(1), pp.3-28.

Chakraborty, S. (2020). Gender Wage Differential in Public and Private Sectors in India. The Indian Journal of Labour Economics, 63(3), pp.765-780.

Demoussis, M., Giannakopoulos, N. and Zografakis, S., 2010. Native–immigrant wage differentials and occupational segregation in the Greek labour market. Applied Economics, 42(8), pp.1015-1027.

Firpo, S., Fortin, N.M. and Lemieux, T., 2009a. Unconditional quantile regressions. Econometrica, 77(3), pp.953-973.

Goy, S.C. and Johnes, G., 2012. Revisiting the impact of occupational segregation on the gender earnings gap in Malaysia. Jurnal Ekonomi Malaysia, 46(1), pp.13-25.

Heckman, J.J., 1979. Sample selection bias as a specification error. Econometrica: Journal of the econometric society, pp.153-161.

Ismail, R., Farhadi, M. and Wye, C.K., 2017. Occupational segregation and gender wage differentials: evidence from Malaysia. Asian Economic Journal, 31(4), pp.381-401.

Kabubo-Mariara, J., 2003. Wage determination and the gender wage gap in Kenya: Any evidence of gender discrimination? AERC Research paper No. 132.

Kenya National Bureau of Statistics (KNBS), 2023. Economic Survey 2023. Nairobi: Government Printer. https://www.knbs.or.ke/economic-survey-2023/

Khitarishvili, T., Rodriguez Chamussy, L. and Sinha, N., 2018. Occupational segregation and declining gender wage gap: The case of Georgia. World Bank Policy Research Working Paper, (8583).

Maina, J.N., 2021. The Effects of Occupational Segregation on Gender Wage Gap in Kenya (Doctoral dissertation, University of Nairobi).

Machado, J.A. and Mata, J., 2005. Counterfactual decomposition of changes in wage distributions using quantile regression. Journal of applied Econometrics, 20(4), pp.445-465.

Oaxaca, R. L. & Ransom, M. 1994. On Discrimination and the Decomposition of Wage Differentials. Journal of Econometrics, 61(1), pp.5-21.

Oaxaca, R., 1973. Male-female wage differentials in urban labor markets. International Economic Review, 14(3), pp.693-709.

Omanyo, D., 2021. Gender Wage Differentials between Public and Private Sectors in Kenya. Kenya Institute for Public Research and Analysis. KIPPRA Discussion Paper No.255

Orraca, P., Cabrera, F.J. and Iriarte, G., 2016. The gender wage gap and occupational segregation in the Mexican labour market. Econoquantum, 13(1), pp.51-72.

UN Women, 2023. Why Women Earn Less: Gender Pay Gap and Labour-Market Inequalities in Kenya. https://africa.unwomen.org/sites/default/files/2023/un_women_kenya_gender_pay_gap_report.pdf

​

​

​​

​