Estimation of Gender Wage Differentials using Oaxaca Decomposition Technique

Loyola University Chicago Loyola ecoons Topics in Middle Eastern and North Arican Econoies Quinlan School o Business 5-1-2014 Estiation o Gender Wage Dierentials using Oaxaca Decoposition Technique Marwa Biltagy Cairo University Recoended Citation Biltagy, Marwa, "Estiation o Gender Wage Dierentials using Oaxaca Decoposition Technique" (2014). Topics in Middle Eastern and North Arican Econoies. 188. http://ecoons.luc.edu/eea/188 This Article is brought to you or ree and open access by the Quinlan School o Business at Loyola ecoons. It has been accepted or inclusion in Topics in Middle Eastern and North Arican Econoies by an authorized adinistrator o Loyola ecoons. For ore inoration, please contact ecoons@luc.edu. This work is licensed under a Creative Coons Attribution-Noncoercial-No Derivative Works 3.0 License. 2014 the authors

ESTIMATION OF GENDER WAGE DIFFERENTIALS IN EGYPT USING OAXACA DECOMPOSITION TECHNIQUE Marwa Biltagy Cairo University, Egypt bilarwa@hotail.co bilarwa@eps.edu.eg ABSTRACT This paper ocuses on estiating wage dierences between ales and eales in Egypt to understand the deterinants o the gender wage gap and control o this variation. The ethodology o this paper is based on studying and analyzing the topic o wage dierentials between ales and eales by using the Oaxaca decoposition technique. The data used in the analysis is obtained ro the Egypt Labor Market Panel Survey 2006 (ELMPS 2006), which was presented by Central Agency or Public Mobilization and Statistics (CAPMAS) in cooperation with Econoic Research Foru (ERF). The indings o this paper help deepen the understanding o the wage gap between ales and eales in Egypt by deterining dierences in wages due to real variations in characteristics between both genders, or exaple, education and experience and dierences due to discriination against woen in addition to deterining dierences due to selectivity bias. It is estiated that, the wage gap between ales and eales is 25% and the results ascertain that the overall gap is attributed to discriination against woen. KEYWORDS: Incoe Inequalities- Gender wage inequality- Wage dierentials- Oaxaca decoposition technique- Egypt. JEL CLASSIFICATION: J31, J38, J71 17

1. INTRODUCTION Wage inequality between ales and eales is a topic that has received a great deal o attention in the econoic literature. The issue o dierence in incoe between ales and eales is o great iportance because it aects a very large nuber o people. This iplies that, the gender wage gap aects woen and their children and uture generations as well. I the deterinants o the gender wage gap can be ound, policy could be ipleented to reduce the incoe disparities. In recent decades, ale-eale earnings dierentials have been studied in any developed and developing countries around the world, including Egypt. The existence o such dierentials is usually associated with discriination against woen in the labor arket. It has been asserted that discriination has iportant econoic, political and social consequences that call or corrective actions. Feales jobs do pay less than ales jobs, even ater accounting or observable dierences in worker and job characteristics; a considerable raction o the gender wage gap reains unexplained. An unexplained gender wage gap has oten been explained as evidence o labor arket discriination. However, it could relect additional unobserved or uneasured dierences in worker and job characteristics between ales and eales (Wood et al., 1993; Blau and Kahn, 2000). The act that eales are ore likely to choose occupations that oer ore lexibility and that do not require continual investents in skills unique to a ir, or occupations where skills do not depreciate signiicantly because o career interruptions helps explain the selection hypothesis. The higher concentration o eales in these jobs would then explain why eale-doinated occupations pay lower wages than ale-doinated ones (Görlich and de Grip, 2008; ILO, 2010). This paper ocuses on estiating wage dierences between ales and eales in Egypt to understand the deterinants o the gender wage gap and control o this variation. This study helps deepen the understanding o wage gap between ales and eales in Egypt by deterining dierences in wages due to real variations in characteristics between both genders and other dierences due to discriination against woen. The study attepts to answer one iportant question, i.e. what are the ain deterinants o gender wage dierentials in Egypt? The ethodology o this paper is based on studying and analyzing the topic o wage dierentials between ales and eales by using the Oaxaca-Blinder decoposition technique and exaining the selection bias eect using the Neuan-Oaxaca decoposition technique. This paper consists o ive ain sections, including the introduction and conclusion. The second section surveys the wage syste in Egypt and its structure, the trends and proile o the Egyptian labor arket. The third section presents the data description and the characteristics o ales and eales saples. The ourth part proposes the epirical odel, its estiation and the results o the regression analysis o ales and eales saples. 18

2. BACKGROUND 2.1 THE WAGES SYSTEM IN EGYPT: MAJOR CHARACTERISTICS A wage can be deined as a onetary reward that is paid to an eployee or the services provided by hi. There are a nuber o principles o a coprehensive wage policy. For exaple, the wage policy should achieve the national objective o econoic growth with social justice; reduce the relative poverty o the working class; proote eployent, productivity and capital oration; reove ibalances aong sectors and wage dierentials and ensure rising real wages that are consistent with the capacity o the industry and the national econoy. The wage structure depends on any actors, such as, wage settleents, the labor arket situation, and the nature and size o dierent institutions. In addition, the wage structure consists o certain grades and scales; each scale has a iniu and a axiu liit and the actual pay in a grade depends on the length o service and the perorance o the eployee. The wage structure in Egypt generally consists o basic wage, bonus and other incentives. A rational wage policy ust eectively ix a iniu and axiu wage strategy. This eans that, there is a need to provide a wage which would be suitable to ulill basic needs. On the other hand, the need or a wage ceiling is essential in order to check the upward inlationary trend o the wages. This iplies that, the wage structure ust contribute to price stability. The wages syste in Egypt has any probles and has been criticized because o its coplexity and inequality. For exaple, the basic salary on which the bonus and the pension payents ater retireent are calculated represents approxiately 20% o the total wage. Furtherore, the syste o wages suers ro inequities between the public and private sector. It has been asserted that, the iniu wage or workers is dierent in both public and private sectors. Eployees who doing siilar jobs whether in the public sector or private one oten earn very dierent wages depending on the inistry (Abdelhaid and El Baradei, 2010). Many eployees have deanded a iniu wage policy to be settled by the governent, and this request increased heavily especially ater the 25 th o January Egyptian revolution in 2011. According to the International Labor Convention (1970), iniu wages policies should be established and regularly updated within the rule o law. The National Council or Wages was established by Law 12 o 2003 as the entity which is responsible or drawing up the wages policy in Egypt and ensuring that wages are suitable to the cost o living and are adequate to cover rising prices. There are any dierent laws and regulations controlling the wages syste in Egypt. A revision o all laws and legislations that aect the wages syste in Egypt is an iportant issue that should be taken into consideration. Another characteristic o the coplexity o the Egyptian wages syste is the series o special bonuses, annual increases and incentive structures. Beore the 25 th o January Egyptian revolution in 2011, there was a correlation between the presidential election year and the increase in the percentage o special bonuses that should be added to the basic wage. However, there is an increased level o dissatisaction, especially aong eployees in the public sector, with the wage level, which lacks the international standards o iniu wages. Egyptian Law 19

guarantees the iniu wage or all governent and non-governent eployees, in addition to yearly increents to salary ranging usually between 10-15%. In Egypt, the iniu wage policy should take into consideration the dierences between sectors and geographical areas. Moreover, this policy should be odiied to be copatible with the level o productivity and the increased inlation rate as well. At least one quarter o Egyptians live in harsh poverty, spending no ore than LE3444 ($500) each year. In July 2012, a iniu wage o LE700 ($100) was set in the public sector or the irst tie since 1980s. Fro January 2014, Egyptian public sector workers will have a iniu onthly wage o LE1200 ($171). Egypt s annual urban inlation rate rose to 10.1% in the 12 onths prior to Septeber 2013, up ro 9.7% recorded in year-to-august (CAPMAS, Egypt in Figures, 2013). The inlationary pressure resulting ro a iniu wage increase can be contained i the governent ollows a sound policy to enhance production. Then again, Egypt s heavy reliance on iports or basic necessities akes it unlikely that raising the iniu wage would result in an increase in local production. Egypt spent around LE33 billion ($4.8 billion) on iports in the irst eight onths o 2013. Moreover, Egypt s oicial uneployent rate reached 13.3% o the country s labor orce in the second quarter o 2013, with 74% o job seekers aged between 15 and 29 years old (CAPMAS, Egypt in Figures, 2013). In suary, it can be said that reoring the wage syste in Egypt requires an overall plan and a civil service reor that would satisy the objectives o eployer and eployees as well as lead to an eicient and responsive governent syste. It can be said that, parallel political reor allowing or greater levels o accountability and transparency in governent operations and the staged ipleentation o reors, are iportant issues to achieve the target o reoring the wage syste in Egypt. 2.2 THE EGYPTIAN LABOR MARKET: TRENDS AND PROFILE The ain coponents o the labor arket in Egypt are a large public sector, a growing private sector, and an inoral sector. For a long tie, the governent guaranteed eployent associated with lietie job security and several other beneits such as public health insurance or every secondary and postsecondary graduate. These characteristics have changed in the context o the econoic reor and structural adjustent progra initiated in 1991. The governent has increased the waiting period or governent appointents. As a result, the public sector work orce growth rate declined and the private sector growth rate increased due to the privatization process. Moreover, pay and work conditions have changed, resulting in changes in the labor arket structure (Assaad, 1997). Despite signiicant progress in eale labor orce participation over past years, gender dierences reain in productivity and earnings across dierent sectors and jobs. In spite o lower earnings and productivity, woen are not worse workers than en. Instead, the gender dierences in labor productivity and earnings are ainly the result o dierences in the econoic activities o en and woen, in addition to gender dierences in huan capital and in the returns to worker and job characteristics. In act, ales and eales jobs dier greatly, 20

across sectors, industries or types o irs. Woen all over the world appear to be concentrated in low-productivity jobs. They work in sall ars and run sall irs. They are overrepresented aong unpaid aily workers and in the inoral sector. Woen are ore likely to work in jobs that oer lexible working arrangeents, so that they can cobine work with care responsibilities, or instance, part-tie or inoral jobs. These jobs oten pay lower wages than ull-tie and oral jobs. Soeties, they rise to positions o power in the labor arket (World Developent Report, 2012). Over the past years, woen have joined the labor arket in increasing nubers, partially closing the gender participation gap. Between 1980 and 2009, the global rate o eale labor orce participation rose ro 50.2 % to 51.8 %, while the ale rate ell ro 82.0 % to 77.7 %. It can be said that eale labor orce participation is lowest in the Middle East and North Arica (26 %) and South Asia (35 %) and highest in East Asia and Paciic (64 %) and Sub-Saharan Arica (61%). The changes in eale participation rates over the past 25 years can be attributed to the eects o econoic developent and rising levels o education aong woen, which increases the deand or eale labor (World Bank, 2011). In Egypt, it is a act that arriage has usually been associated with a decline in eale labor orce participation. Soe woen rejoin the labor orce several years ater arriage. The exit rates or woen working in inoral irs or household enterprises are higher than the exit rate o governent eployees. Exit rates are deterined ainly by arriage, but the eect o arriage on exit aong inoral sector eployees is considerably higher than aong governent eployees. Moreover, increases in the age o arriage and declines in the percentage o ertility are likely to have contributed to higher participation rates. To su up, the ipact o econoic developent and changes in the levels o education and aily oration on eale labor orce participation varies across individuals and regions and depends on institutions and individual preerences (World Bank, 2010). The ajority o working urban woen held governent jobs, and in rural areas, the governent and household enterprises reported ore than 70 % o eale eployent. Private sector irs accounted or less than a quarter o eale eployent in urban Egypt in 2006. Their share in rural eale eployent was lower (approxiately 8%). Work in the public sector is ore copatible with woen s needs since it oers shorter hours, ore access to childcare and greater assistance or aternity leave. In 2006, the proportion o workers who reported having been at work during their last pregnancy was signiicantly higher in the public sector. It has been asserted that 86% o public sector workers who had a baby while working were given paid aternity leave o at least six weeks, in contrast to only 47% o those working in the oral private sector. Furtherore, the percentage o working woen aged 15 29 years who coplain o long working hours is signiicantly higher in the private sector (50%) than in the public sector (32%) (World Bank, 2010). Table (1) deonstrates eployent growth ro 1988 to 2011, based on the data o ELMPS 2012. Eployent has increased ro 14,508 thousand in 1998 to 19,987 thousand in 2006 and to 22,487 in 2011. Over the sae period, job creation has changed, and it reached 1,488 thousand in 2011. Siilarly, job exit luctuated ro 1998 to 2003 and then increased gradually ro 2003 to 2011. The net job growth has declined ro 4.6% in 1998 to 1.1% in 2011; this 21

could be explained by the growing nuber o job exits. Additionally, working age population growth has decreased over the sae period and has reached 2% in 2011. Table (1): Eployent Growth, 1998-2011 Year Eployent (Thousands) Job Creation (Thousands) Job Exit (Thousands) Net Job Growth (Thousands) Net Job Growth (%) Working Age pop. Growth (%) 1998 14,508 1,251 599 652 4.6 4.3 1999 15,125 1,216 611 605 4.1 4.4 2000 15,915 1,580 801 779 5.0 4.0 2001 16,678 1,449 709 740 4.5 3.7 2002 17,371 1,472 774 698 4.1 4.2 2003 17,910 1,270 678 592 3.3 3.7 2004 18,621 1,425 760 665 3.6 3.0 2005 19,393 1,631 901 730 3.8 3.2 2006 19,987 1,515 948 567 2.9 2.7 2007 20,484 1,386 903 483 2.4 2.5 2008 21,063 1,616 1,074 542 2.6 2.4 2009 21,695 1,754 1,137 617 2.9 2.4 2010 22,218 1,697 1,328 369 1.7 2.1 2011 22,487 1,488 1,245 243 1.1 2.0 Source: Assaad and Krat, 2013. The labor orce in Egypt has increased ro 19.3 illion in 2001 to 27 illion in 2012, however the eale labor orce has only increased ro 4.1 illion in 2001 to 6.1 illion in 2012 (see Figure 1). In contrast, the ale labor orce participation rate has been increasing over this period; in act, the ale labor orce has increased ro 15.2 illion in 2001 to 20.9 illion in 2012. In ters o uneployent, eales have continued to ace uch higher rates over the whole period. The eale uneployent rate has been approxiately 3 to 4 ties higher than that o ales. 22

Figure (1): Annual Estiates o Labor Force by Sex (2001-2012) (1) All nubers in Y axis are in hundreds. (2) Males Feales Source: CAPMAS, Annual Statistical Book, 2013. Figure (2) illustrates that, the ale uneployent rate as reported by the Labor Force Saple Surveys (LFSSs) conducted by the Central Agency or Public Mobilization and Statistics (CAPMAS) nearly doubled ro 2010 to 2011 (i.e. ro 4.9% to 8.9%). This could be attributed to the econoic slowdown caused by the 25 th o January Egyptian revolution. The ale uneployent rate then increased urther in 2012 to 9.3%. On the contrary, the eale uneployent rate was not aected heavily by the revolution; it increased ro 22.57% in 2010 to 22.73% in 2011 and reached 24.1% in 2012. The overall uneployent rate increased ro 9% in 2010 to 12.7% in 2012. 23

(%) Topics in Middle Eastern and Arican Econoies Figure (2): Annual Estiates o Uneployent Rate by Sex (2001-2012) Annual Estiates o Uneployent Rate by Sex (2001-2012) 30 25 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 Years Males Feales Source: CAPMAS, Annual Statistical Book, 2013. There have been substantial changes in the coposition o eployent over tie in Egypt. The public sector eployed 25% o the workorce in 2006 and 26% in 2012. Eployent in public enterprises continued to decrease, ro 7% o eployent in 1998 to 5% in 2006 and 4% in 2012. Moreover, oral private regular wage eployent has increased ro 8% in 1998 to 9% in 2006 and 11% in 2012, while inoral private regular wage eployent has decreased ro 17% in 2006 to 15% in 2012. Coparing between the Egypt Labor Market Panel Survey 2006 (ELMPS 2006) and the Egypt Labor Market Panel Survey 2012 (ELMPS 2012), it has been asserted that, the largest change was the substantial increase in irregular wage work. While 8% o the eployed were irregular wage workers in 2006, this had increased to 17% o the eployed in 2012. Regarding ales and eales, there has been very little decrease in the public sector s share o ale and eale eployent over the 1998 to 2012 period. A sall increase in oral private regular work has occurred over tie aong ales and eales. The inoral private regular wage work decreased aong ales over the 2006 to 2012 period. Siilarly, the inoral private regular wage work declined aong eales, ro 9% in 2006 to 7% in 2012. Furtherore, ales experienced a signiicant increase in irregular wage work, ro 9% in 2006 to 20% in 2012, while a sall nuber o eales are engaged in irregular wage work over the 2006 to 2012 period (Assaad and Krat, 2013). 24

2.3 RELATED LITERATURE REVIEW There are a large nuber o previous studies o gender wage dierentials. Shaban et al. (1993) and Assaad (1997) estiated joint odels o sector choice and wage deterination in the public and private sector using 1987 and 1988 household-level labor orce saple survey data. Assaad (1997) attepted to quantiy the value o non-wage beneits in the public sector. Kanellopoulos and Mavroaras (1999) presented an epirical study o the developent o labor arket participation and wage dierentials between ales and eales in Greece between 1988 and 1994. This paper used recent survey data generated by the National Statistical Service. The authors used selectivity corrected earnings estiations. The results show that the adverse treatent o eale labor arket participation is the largest identiiable reason why the wage gap is there and why it increased between 1988 and 1994. The study ound that the observed lower eale relative pay can be priarily attributed to the actors which deterine paid eployent participation. The participation process was ound to be highly discriinatory in avor o ales. Assaad (1999) copares the earnings o workers in and out o public enterprise, while taking account o dierences in non-wage beneits and non-rando sector selection. The author relates workers' losses to observable characteristics such as seniority, age, educational attainent, and gender, and evaluates how well alternative redundancy pay orulas typically used in severance progras atch copensation payents to these estiated losses. The results o this study show that woen ore than en tend to ace strong barriers to entry into wage jobs in the private sector and thus have poorer earning prospects. El-Haidi and Said (2005) studied the changes in the distribution o returns to education and gender wage in the Egypt and Morocco arket using joint odels o educational choice and wage deterination. Their epirical analysis is based on the 1988 and 1998 Egypt Labor Force Saple Surveys and The Morocco Living Standard Measureent Studies o 1990/1991 and 1998/1999. In Egypt, the ale public sector wage preiu declined ro 7% in 1991 to 3% in 1998; whereas the eale one reained alost the sae at 16-17 %. Overall, wage inequality by education and gender appears to have declined substantially in Egypt during that decade o pursuing econoic liberalization policies. By contrast, all changes in public sector preius and unexplained wage gaps in Morocco appear to be in the opposite direction. Male preius in the public sector increased ro 33% in 1991 to 58% in 1999 and so did eale preius, which draatically juped ro 14% in 1991 to 81% in 1999. In general, the nineties appear to be a decade o increasing wage inequality by gender and education in Morocco. Schagans and Stelcnery (2006) re-exained the gender decoposition o wages in the presence o selection bias. They derived the appropriate saple selection corrections, based on a reduced or odel or the joint participation decisions o both couples. The inluence that husbands participation decision has on the eale participation decision also highlights the iportance o using data on both spouses or the analysis o the gender wage gap. The authors analyzed the gender earnings dierential using Canadian census data. They ound that adding additional controls or ield o study did not signiicantly aect the decoposition analysis, but there was 25

soe reduction in discriination when additional controls or occupation and industry were added. Haas (2006) studied the dierence in incoe between en and woen, the greatest equity is ound in Switzerland, which has a ale-to-eale wage ratio o 1.11, eaning that en ake approxiately 1.11 ties as uch as woen. The greatest inequality is seen in Egypt, with a ratio o 3.84, iplying that en ake alost quadruple the wages o woen. This study exained the relationship between the gender wage gap and the degree o econoic developent o a country as easured by the gross doestic product per capita. This study used the United Nations Huan Developent Index as a ore coprehensive easure o developent. It also analyzed the relationship o educational attainent and general wage inequality to the size o the gender wage gap. The results suggested that the decrease in gender wage inequality is not expected to be seen until countries reach developent levels close to 0.80 on the Huan Developent Index. Herrera and Badr (2011) suggested that the returns to education and experience in Egypt increase with ir size and are larger in the oral sector. Given the act that eales are overrepresented aong inoral workers, this would translate into eale jobs paying less than ale jobs. Abdelhaid and El Baradei (2010) identiied what needs to be done to reor the pay syste or governent eployees in Egypt through proposing a set o policy solutions and strategies. The authors propose a syste or pay adjustent and strategies to resolve the proble. For exaple, securing the needed extra unding or increasing governent eployees pay, right sizing the governent civil service, enhancing transparency, reducing wage discrepancies, reoring the iniu wage policy and establishing a better link between pay and perorance. Moreover, the study ephasizes an urgent need or capacity building in governent sta, as uch as or salary and wage revision. Yasin et al. (2010) concentrated on the gender eployent positions and wage dierentials in Pakistan. They analyzed the deterinants o gender wage discriination in Pakistan using descriptive and regression analysis based on the cross-sectional data o the Pakistan labor orce survey. They concluded that illiteracy, poor, and low levels o education as well as low vocational, technical, and proessional copetence are iportant eatures o the labor arket participants in Pakistan. The results o this epirical analysis showed that dissiilarity in the attainent o jobs is a rearkable phenoenon between ales and eales. It is also proved that, soe socio-econoic and cultural constraints hinder the participation o eales. Moreover, the results conired that woen were not dierent in their productivity ro en and without discriination woen could earn ore copared to en in soe cases. 3. DATA, METHODOLOGY AND ANALYSIS The data used in the analysis is obtained ro ELMPS 06, which was presented by CAPMAS in cooperation with Econoic Research Foru (ERF). The questionnaire or the ELMPS 06 is coposed o three ajor sections; the irst section proposes the household questionnaire adinistered to the head o household that contains inoration on basic deographic 26

characteristics o household ebers. The second section presents the individual questionnaire adinistered to the individual containing inoration on parental background, detailed education histories, detailed eployent characteristics, job characteristics and earnings. The third section discusses the incoe sources o the household. The size o the saple o ales is 5107 individuals and the size o eales saple is 1465 individuals. The saples contain waged workers whose ages range ro 15 to 64 years. Those individuals answer all the questions needed or the estiation o the Mincerian equation and basic earnings unctions and the equation that deterines the wage dierentials between ales and eales. The epirical raework ollows Mincer s estiation o the siple schooling odel, which relates earnings to work experience and years o schooling. The dependent variable in the log earnings regression is the log o a respondent's total wage. The ethodology o this paper is based on analyzing wage dierentials between ales and eales by using Oaxaca and Blinder decoposition technique. The Oaxaca/Blinder decoposition technique explains whether dierences in wages between ales and eales are due to variations in characteristics between the or alternatively due to discriination. Tables (2) and (3) present the variables that are used in the estiation o the epirical odel: 1- Age: The age o ales in the saple ranges ro 15 to 64 years old, while or the eale saple it varies ro 15 to 62 years. In addition, the ean value o age is 35.86 years or ales and 36.34 years or eales. 2- W: The wage o ales ranges ro 90 to 20340 pounds per onth, while it ranges ro 85 to 15000 pounds per onth or eales. Moreover, the ean value o wages is 709.5 and 536.4 Egyptian pounds per onth or the saples o ales and eales, respectively. It is clear that, the wages o ales per onth are greater than eales and this is the wage gap, which will be decoposed. 3- lnw: The ean value o the log wages is 6.2 and 5.95 or ales and eales, respectively. 4- S: The nuber o years o schooling (S) ranges ro 0 to 20 years or both genders. Tables (2) and (3) show that, the ean value o S is greater or eales (13.5 years) than ales (10.9 years). 5- T: There is a dierence in the ean value o the nuber o years o experience (T) between the saples o ales and eales. While the ean value o T is 19.9 years or ales, it is only 17.3 years or eales. 27

Table (2): List o the Variables used in the Model (Saple o Males) Variable No. o Mean Std. Dev. Min. Max. Observations Age 5107 35.86 11.28 15 64 W 5107 709.5 1247 90 20340 lnw 5107 6.2.7015 4.5 9.9 S 4675 10.9 5.046 0 20 SS 5107 132.3 94.27 0 400 experience 5107 19.9 12.89 0 58 expexp 5107 561.1 637.7 0 3364 Urban/rural 5107.6015.4896 0 1 Marital Status 5107.7235.4473 0 1 Sector o ep. 4373.5657.4957 0 1 Source: Author s calculations based on ELMPS 2006. 28

Table (3): List o the Variables used in the Analysis (Saple o Feales) Variable No. o Mean Std. Dev. Min. Max. Observations Age 1465 36.34 10.8 15 62 W 1465 536.4 839.5 85 15000 lnw 1465 5.95.715 4.4 9.6 S 1412 13.49 3.14 0 20 SS 1465 185.1 76.5 0 400 experience 1465 17.34 11.42 0 54 expexp 1465 431.2 461.3 0 2916 Urban/rural 1465.7727.4192 0 1 Marital Status 1465.661.4736 0 1 Sector o ep. 1450.761.4268 0 1 Source: Author s calculations based on ELMPS 2006. THE CHARACTERISTICS OF THE MALE AND FEMALE SAMPLES Tables (2) and (3) illustrate that, the average earnings are 709.5 and 536.4 pounds per onth or ales and eales, respectively. In addition, the average age is 35.86 years or ales and 36.34 years or eales, the average nuber o years o schooling is 10.9 years or ales and 13.5 years or eales, the average aily size is 6 individuals or both genders and the average nuber o years o experience is 19.9 and 17.3 years or ales and eales, respectively. Moreover, the average level o schooling o an individual s parents is low or both genders; however, the eales parents are ore educated. It is ound that, 60% and 77% o the ale and eale saples, respectively, live in urban areas. In addition, the data shows that, 72.4% and 66.1% o the saples o ales and eales, respectively, are arried. 29

It is concluded that, 56.6% and 76.1% o the ale and eale saples, respectively, are eployed in governent and public enterprises. The average nuber o working days is 6 days per week or both genders and the average nuber o working hours is 8.6 and 7.5 hours per day or ales and eales, respectively. 4. EMPIRICAL MODEL, ESTIMATION AND RESULTS This section attepts to provide an answer to the question: What are the deterinants o the wage gap between ales and eales in the Egyptian labor arket? In order to do that, two ethods o decoposition will be used. They rely on estiating log earnings unctions based on Huan Capital Theory or each gender separately. The standard Oaxaca-Blinder procedure is used to estiate the extent to which the overall wage gap between ales and eales can be explained by dierences in observed huan capital characteristics such as education and experience (Oaxaca 1973; Blinder 1973). Then, the Neuan-Oaxaca decoposition technique is applied to correct or the selectivity bias eect. 4.1 THE EMPIRICAL MODEL The earnings unction is, ln W B B S B T B T u. (1) j 0 1 j 2 j 3 2 j j Where W represents the onthly wage o an individual, S relects his/her years o schooling. The variable S = 0 or illiterates, S = 6 or priary education, S = 9 or preparatory education, S = 12 or secondary education, S = 14 or above interediate education, S = 16 or university education and S = 20 or post graduate studies. Furtherore, T j represent the nuber o years o experience or individual j. It is assued that this unction exhibits positive but diinishing arginal returns to experience. The nuber o years o experience (T j ) is calculated by using a siple rule, that is, T j = A j S j 6, (2) where A is the age o an individual and S is the nuber o years o schooling. The earnings unctions or ales and eales are, ln W b b S b T b T u (3) i 0 1 i 2 i 2 3 i i ln W B B S B T B T u (4) i 0 1 i 2 i 3 2 i i The total dierence in wages or both genders can be expressed by, lnw lnw lnw. (5) The Oaxaca-Blinder decoposition equation is, 30

lnw lnw lnw lnw X B ( X X X B X B X B ) B ( B B ) X (6) Where the irst ter eales and the second ter The selectivity-corrected wage equation is, lnw ( X X ) B reers to dierences in characteristics between ales and (B B ) X captures the discriination eect. lnw ( X X ) B ( B B ) X ( ). (7) Where the ter ) denotes the selectivity bias eect. ( 4.2 ESTIMATION AND RESULTS RESULTS OF WAGE FUNCTIONS ESTIMATION Tables (4) and (5) present the results o estiation o the wage unctions or ales and eales. The results show that, the private rate o return to education is 3.9% or ales and 8.3% or eales. It is clear ro the results that, there is a positive relationship between the nuber o years o experience and the earnings or both genders. In addition, the wages o ales and eales will increase at a decreasing rate as the nuber o years o experience increases, because b 3 is negative in tables (4) and (5). Moreover, the rate o return to the nuber o years o experience is 1.91% or ales and 3.56% or eales. In other words, each additional year o experience will increase the wages by 1.91% and 3.56% or the saples o ales and eales, respectively. Table (4): The Regression Results o Wage Function (Males Saple) t P> lny Coe. Std. Err. t S.0380302.0021382 17.79 0.000 experience.0294937.0027469 10.74 0.000 expexp -.0002667.0000597-4.47 0.000 cons 5.382155.0382285 140.79 0.000 Source: Author s calculations based on ELMPS 2006. 31

Table (5): The Regression Results o Wage Function (Feales Saple) lny Coe. Std. Err. t t P> S.0801928.0052082 15.40 0.000 experience.0440133.004959 8.88 0.000 expexp -.0002621.0001319-1.99 0.047 cons 4.251109.0833718 50.99 0.000 Source: Author s calculations based on ELMPS 2006. OAXACA-BLINDER GENDER WAGE GAP DECOMPOSITION RESULTS As entioned above, the irst ethod o decoposition is siple and it depends on the Oaxaca (1973) and Blinder (1973) decoposition technique. By subtracting the earnings unctions by parts, it is possible to decopose the earnings gap into two dierent coponents. The irst coponent is the part o the gap that can be attributed to the dierences in ean huan capital characteristics o the two groups and this coponent is called dierences in characteristics. The second coponent is the part o the gap that can be attributed to dierences in the estiated paraeters o the earnings unctions o ales and eales, this part is called discriination. This ethod o decoposition can be illustrated by equation (6). The results o the estiation o gender wage dierentials using the Oaxaca-Blinder decoposition technique are illustrated in table (6). The ean value o real onthly wages aounted to 709.5 Egyptian pounds or ales and 536.4 Egyptian pounds or eales in 2006 (The wage gap is: 709.5-536.4 = 173.1 Egyptian pounds). The characteristics o the saple iply that the average nuber o years o schooling S is greater or eales (13.5 years) than ales (10.9 years). Given the characteristics o the eales saple, the estiation o the odel iplies that, without discriination against woen, their onthly wages should equal 749.8 Egyptian pounds. This eans that, due to discriination; eales are receiving 213.4 Egyptian pounds less in ters o their real onthly wages. The value o discriination represents 39.8% o the ean value o real onthly wages they are actually receiving. 32

Table (6): Oaxaca-Blinder Wage Decoposition Results Wage Decoposition coponents Males Feales Mean value o real onthly wages 709.5 536.4 Overall wage gap 173.1 Endowents eect -40.3 Discriination eect 213.4 Wage without discriination eect (1) All values in the table are in Egyptian pounds. Source: Author s calculations based on ELMPS 2006. 749.8 Figure (3) reports the results o the Oaxaca-Blinder wage decoposition. The overall gap is attributed to discriination. The data ascertains that ales have higher potential years o experience and eales have ore years o education that ade the better endowed in ters o overall huan capital characteristics. It is estiated that, the wage gap is 25%, where lnw lnw = 0.25, the endowents or characteristics eect is -5.8, where ( X X ) B = - 0.058 and the discriination eect is 30.8%, where (B B ) X = 0.308. The discriination eect includes dierences in coeicients in addition to dierences in the constant ter. 33

(%) Topics in Middle Eastern and Arican Econoies Figure (3): Oaxaca- Blinder Wage Decoposition Technique 35 30 25 20 Oaxaca- Blinder Wage Decoposition Technique 15 10 5 0-5 -10 Total Gap Characteristics Discriination Wage Decoposition Gap Source: Author s calculations based on ELMPS 2006. NEUMAN-OAXACA GENDER WAGE GAP DECOMPOSITION RESULTS The second ethod o decoposition is ore coplicated and it depends on wage decoposition with selectivity-corrected wage equation adopted by Neuan and Oaxaca (2004). The estiates o ales and eales wage equations adopted Heckan s two step estiates with correction or selection bias. The act that soe people have chosen to work as eployees in the public/private sector could be a rando choice. However, probles arise when those eployees have soe coon characteristics that are responsible or their particular choice. In order to account or the selection bias eect on the earnings unction, the two-stage selectivity correction ethod is used (Heckan, 1979). This ethod can be illustrated by equation (8), in which: lnw lnw ( X X ) B ( B B ) X ( ) (8) where ln W is the ean value o log wage, X is the ean vector o wage deterining variables i.e. the nuber o years o education and the nuber o years o experience, B is vector o coeicients i.e. the estiated returns to the wage deterinants, is an estiate o ρσ u and is an estiate o the ean Inverse Mills Ratio (IMR). The irst two ters in this equation are the 34

discriination and huan capital coponents. Moreover, the last ter signiies the selectivity bias eect. According to this ethod, a selection equation using the two-stage Heckan selection odel o the probability o eployent in the private sector is estiated or ales and eales separately [see tables (7) and (8)]. The explanatory variables are the nuber o years o education, age, duy variable or residence in urban or rural areas and arital status. Then, the Inverse Mill s Ratio (IMR) is calculated and the earnings unctions are re-estiated separately or ales and eales and subtracted by parts to get the new decoposition o the gender earnings gap that includes a third coponent, corresponding to selectivity. Table (7): Heckan Selection Model -- Two-Step Estiates (Saple o Males) ln W Coeicient Std. Err. z P> z [95% Con. Interval] S 0.0460321 0.0093177 4.94 0.000 0.0277697 0.0642944 experience 0.0494474 0.0068097 7.26 0.000 0.0361006 0.0627942 expexp -0.0006405 0.0001008-6.36 0.000-0.000838-0.000443 _cons 5.141805 0.1002343 51.30 0.000 4.945349 5.33826 Sector o eployent S -0.0728771 0.0047995-15.18 0.000-0.0822839-0.0634703 age -0.0646856 0.0027429-23.58 0.000-0.0700616-0.0593095 urban/rural 0.3587335 0.0477924 7.51 0.000 0.2650622 0.4524048 arital -0.2777026 0.05833-4.76 0.000-0.3920274-0.1633778 status _cons 2.935747 0.1086344 27.02 0.000 2.722828 3.148667 ills labda 0.030174 0.1164952 0.26 0.796-0.1981523 0.2585004 rho 0.04754 siga 0.63467632 labda 0.03017405 0.1164952 Source: Author s calculations based on ELMPS 2006. 35

Table (8): Heckan Selection Model -- Two-Step Estiates (Saple o Feales) ln W Coeicient Std. Err. z P> z [95% Con. Interval] S 0.0903815 0.0089856 10.06 0.000 0.0727701 0.107993 experience 0.039431 0.0074317 5.31 0.000 0.0248651 0.0539969 expexp -0.0001053 0.0001609-0.65 0.513-0.0004207 0.0002101 _cons 4.109094 0.1979487 20.76 0.000 3.721122 4.497067 Sector o eployent S 0.1050601 0.0140692 7.47 0.000 0.077485 0.1326352 age 0.0636172 0.0051356 12.39 0.000 0.0535517 0.0736828 urban/rural -0.4610893 0.1149472-4.01 0.000-0.6863817-0.2357969 arital status 0.8008961 0.0936082 8.56 0.000 0.6174273 0.9843648 _cons -2.892989 0.2618105-11.05 0.000-3.406128-2.37985 ills labda 0.0351295 0.1045332 0.34 0.737-0.1697518 0.2400108 rho 0.06306 siga 0.55709517 labda 0.03512947 0.1045332 Source: Author s calculations based on ELMPS 2006. The results o the estiation o gender wage dierentials using Neuan-Oaxaca wage decoposition technique are displayed in table (9). The estiation o the odel iplies that, without discriination against woen, their onthly wages should be equal to 797.5 Egyptian pounds. This eans that, due to discriination; eales are receiving 261.1 Egyptian pounds less in ters o their real onthly wages. Furtherore, the value o selectivity bias eect is -34.8 Egyptian pounds; this value is sall but it reduces the gender wage gap. 36

Table (9): Neuan-Oaxaca Wage Decoposition Results Wage Decoposition coponents Males Feales Mean value o real onthly wages 709.5 536.4 Overall wage gap 173.1 Endowents eect -53.2 Selectivity eect -34.8 Discriination eect 261.1 Wage without discriination eect Source: Author s calculations based on ELMPS 2006. All values in the table are in Egyptian pounds. 797.5 Figure (4) deonstrates the results o the Neuan-Oaxaca wage decoposition. It is estiated that, the endowents or characteristics eect is -7.7%, where ( X X ) B = -0.077, the selectivity eect is -5%, where ( ) = -0.05 and the discriination eect is 37.7%, where (B B ) X = 0.377. The results illustrated in igure (4) take into account the issue o selectivity in the private/public sector jobs, using the Neuan and Oaxaca (2004) ethod o decoposition. It appears that the eect o selectivity is sall, but it tends to reduce the gender gap. For exaple, without selectivity in 2006 the earning gap as a proportion o the average ale wage (LE709.5) would be 5% higher (i.e. LE745). Again, the observed wage gap is attributed totally to the discriination eect, while dierences in huan capital characteristics tend to decrease the earnings gap between the two genders. 37

(%) Topics in Middle Eastern and Arican Econoies Figure (4): Neuan-Oaxaca Wage Decoposition Technique Neuan-Oaxaca Wage Decoposition Technique 40 35 30 25 20 15 10 5 0-5 -10 Total Gap Characteristics Discriination Wage Decoposition Gap Selectivity Source: Author s calculations based on ELMPS 2006. 5. CONCLUSION This paper concentrates on wage gap dierentials in Egypt in order to help governent and decision-akers involved in the design o the wage policy control or the incoe inequalities and distortion o wages between ales and eales in the Egyptian labor arket. The indings o this paper help in understanding the wage gap between ales and eales in Egypt by deterining dierences in wages due to real variations in characteristics between both genders and other dierences due to discriination against woen. This study uses ELMPS (2006) to estiate the size o the gender wage gap in Egypt. The data shows that the real onthly wages are higher or ales than eales. The Oaxaca-Blinder decoposition o the gender wage gap deonstrates that the gap is due to the discriination against woen in the labor arket, not to low levels o huan capital characteristics (i.e. lower education levels or less experience). Using the Neuan and Oaxaca (2004) ethod o decoposition, it appears that the eect o selectivity is sall, but it tends to reduce the gender gap. For exaple, without selectivity in 2006 the earning gap as a proportion o the average ale wage would be 5% higher. Again, the observed wage gap is attributed totally to the discriination eect, while dierences in huan capital characteristics tend to decrease 38

the earnings gap between the two genders. To decrease this wage gap, labor standards need to be prooted, especially in the private sector. There is a need or stronger versions o antidiscriination legislation, such as an equal pay rule and a regulation which orbids discriination in job ranks, pay scales and the criteria or entrance into the labor arket. Actually, article 35 o the uniied labor law o 2003 prohibits wage discriination based on gender, but this law should be ore strongly enorced. The ain recoendation o this paper is that several costs, which ake the eployent o eales ore expensive to the eployer, should be decreased in order to give ales and eales an equal opportunity in the labor arket. Given the encouraging environent (e.g. beneits o contracts, paid vacations or child care and sick leave and edical insurance) or woen in the governent and publicly owned enterprises copared to the private sector and the lower levels o discriination in that sector, it is expected that the burden o privatization and civil service downsizing ay negatively aect the already low levels o participation rates o eales, unless eorts are ade to reduce gender discriination in the private sector. At present, changes in wage policy are driven by a philosophy that its poorly with how labor deand and labor supply are changing. For uture policy, the key challenge is to work with institutional arrangeents and do the best that is possible to capture the beneits o both coordination and lexibility. This will provide a basis or generating labor arket standards appropriate to ales and eales as well. Finally, there is an essential need or wage policy reor, based on linking wages to productivity. 39

REFERENCES Abdelhaid, D. and L. El Baradei. (2010). Reoring the Pay Syste or Governent Eployees in Egypt. International Public Manageent Review, Vol. 11 (3). Assaad, R. (1997). The Eects o Public Sector Hiring and Copensation Policies on the Egyptian Labor Market. The World Bank Econoic Review, Vol. 11(1). Assaad, R. (1999). Matching Severance Payents with Worker Losses in the Egyptian Public Sector. World Bank Econoic Review,Oxord Journals, Vol. 13(1). Assaad, R. and C. Krat. (2013). The Structure and Evolution o Eployent in Egypt: 1998-2012. Econoic Research Foru, Working Paper No. 805. Becker, G., (1971). The Econoics o Discriination. Chicago: University o Chicago Press. Blinder, A. (1973). Wage Discriination: Reduced For and Structural Eleents. Journal o Huan Resources, Vol. 8(4). Blau, F. and L. Kahn. (2000). Gender Dierences in Pay. Journal o Econoic Perspectives, Vol. 14 (4). CAPMAS. Central Agency or Public Mobilization and Statistics. (2013). Egypt in Figures, Arab Republic o Egypt. CAPMAS. Central Agency or Public Mobilization and Statistics. (2013). Annual Statistical Book, Labor Force, Arab Republic o Egypt. Econoic Research Foru and Central Agency or Public Monetization and Statistics (2006). Egypt Labor Market Panel Survey. Location: http://www.erdataportal.co/index.php/catalog/27 Econoic Research Foru and Central Agency or Public Monetization and Statistics (2012). Egypt Labor Market Panel Survey. Location: http://www.erdataportal.co/index.php/catalog/45 El-Haidi, F. and M. Said. (2005). Wage Inequality by Education and Gender in MENA: Contrasting the Egyptian and Moroccan Experiences in the 1990s. 12 th Annual Conerence o the Econoic Research Foru (ERF), Deceber 19, 2005. Görlich, D. and A. de Grip. (2008). Huan Capital Depreciation during Hoe Tie. Oxord Econoic Papers, Vol. 61. Haas, S. (2006). Econoic Developent and the Gender Wage Gap. Illinois Wesleyan University, Departent o Econoics, Honors Projects, Paper 1. Heckan, J. (1979). Saple Selection Bias as a Speciication Error. Econoetrica, Vol. 47. 40

ILO. International Labor Organization. (2010). Woen in Labor Markets: Measuring Progress and Identiying Challenges. ILO: Geneva. Herrera, S. and K. Badr. (2011). Why Does Productivity o Education Vary across Individuals in Egypt? Fir Size, Gender and Access to Technology as Sources o Heterogeneity in Returns to Education. Policy Research Working paper Series 5740, World Bank, Washington, DC. International Labor Organization (1970). Miniu Wage Fixing Convention. Geneva, 54th ILC session, No. 170 Kanellopoulos, C. and K. Mavroaras. (1999). Male-Feale Labor Market Participation and Wage Dierentials in Greece. The Centre o Planning and Econoic Research, No. 70. Mincer, J. (1958). Investent in Huan Capital and Personal Incoe Distribution. Journal o Political Econoy, Vol. 66. Mincer J. (1974). Schooling, experience and earnings, Colubia University Press, New York. Neuan, S. and R. Oaxaca. (2004). Wage Decopositions with Selectivity-Corrected Wage Equations: A Methodological Note. Journal o Econoic Inequality, Vol. 2. Oaxaca, R. (1973). Male-Feale Wage Dierentials in Urban Labor Markets. International Econoic Review, Vol.14 (3). Schagans, M. and M. Stelcnery. (2006). Selectivity and the Gender Wage Gap Decoposition in the Presence o a Joint Decision Process. The Suntory Centre, Suntory and Toyota International Centres or Econoics and Related Disciplines, London School o Econoics and Political Science, Discussion paper No. EM/2006/513. Shaban, R., R. Assaad and S. Al-Qudsi. (1993). Labor Markets in Arab Countries: A Survey. Paper presented to the First Annual Conerence on Developent Econoics, Initiative to Encourage Econoic Research in the Middle East and North Arica, Cairo, June 4-6, 1993. Wood, R., M. Corcoran and P. Courant. (1993). Pay Dierences aong the Highly Paid: The Male-Feale Earnings Gap in Lawyers Salaries. Journal o Labor Econoics Vol. 11 (3). World Developent Report. (2012). Gender Dierences in Eployent and Why They Matter. Chapter 5. World Bank. (2010). Arab Republic o Egypt: Gender Assessent 2010: Narrowing the Gap, Iproving labor Market Opportunities or Woen in Egypt. World Bank, Washington, DC. World Bank. (2011). Woen Business and the Law: Measuring Legal Gender Parity or Entrepreneurs and Workers in 128 Econoies. World Bank, Washington, DC. 41

Yasin, G., I. Chaudhry and S. Azal. (2010). The Deterinants o Gender Wage Discriination in Pakistan: Econoetric Evidence ro Punjab Province. Asian Social Science, Vol. 6 (11). 42