Can the Introduction of a Minimum Wage in FYR Macedonia Decrease the Gender Wage Gap?

Transcription

1 Publc Dsclosure Authorzed Publc Dsclosure Authorzed Publc Dsclosure Authorzed Publc Dsclosure Authorzed Pol c y Re s e a rc h Wo r k n g Pa p e r 4795 Can the Introducton of a Mnmum Wage n FYR Macedona Decrease the Gender Wage Gap? The World Bank Europe and Central Asa Regon uman Development Sector Unt December 2008 Dego F. Angel-Urdnola WPS4795

2 Polcy Research Workng Paper 4795 Abstract Ths paper reles on a smple framework to understand the gender wage gap n Macedona, and smulates how the gender wage gap would behave after the ntroducton of a mnmum wage. Frst, t presents a new albet smple decomposton of the wage gap nto three factors: () a wage level factor, whch measures the extent to whch the gender gap s drven by dfferences n wage levels among low-sklled workers of opposte sex; () a sklls endowment factor, whch quantfes the extent to whch the gender wage gap s drven by the dfference n the share of hgh-sklled workers by gender; and () returns to educaton, whch measures the extent to whch the gender gap s drven by dfferences by gender n returns to educaton. Second, the paper presents smple set of smulatons that ndcate that the ntroducton of a mnmum wage n Macedona could contrbute to decrease the gender wage gap by up to 23 percent. Nevertheless, n order to sgnfcantly mprove the wage gap, a rather hgh mnmum wage may be requred, whch may contrbute to reductons n employment. Ths paper a product of the uman Development Sector Unt, Europe and Central Asa Regon s part of a larger effort n the department to lead polcy makng n Labor Market Polcy and Regulaton n transton countres. Polcy Research Workng Papers are also posted on the Web at The author may be contacted at dangelurdnola@worldbank.org. The Polcy Research Workng Paper Seres dssemnates the fndngs of work n progress to encourage the exchange of deas about development ssues. An obectve of the seres s to get the fndngs out quckly, even f the presentatons are less than fully polshed. The papers carry the names of the authors and should be cted accordngly. The fndngs, nterpretatons, and conclusons expressed n ths paper are entrely those of the authors. They do not necessarly represent the vews of the Internatonal Bank for Reconstructon and Development/World Bank and ts afflated organzatons, or those of the Executve Drectors of the World Bank or the governments they represent. Produced by the Research Support Team

3 Can the Introducton of a Mnmum Wage n FYR Macedona Decrease the Gender Wage Gap? Dego F. Angel-Urdnola * December 2008 JEL classfcaton: J38, J23, J71, I32 Key words: Mnmum wages, Gender Gap, Wage Dfferentals, Macedona. Correspondng Author: Dego F. Angel-Urdnola Street, NW Washngton, DC Mal Stop E-mal: dangelurdnola@worldbank.org * The author acknowledges support from grant TF from the Gender Acton Plan at the World Bank, as well as comments and suggestons from Vctor Macas, Arvo Kuddo, Jan J. Rutkowsk, Jane Armtage, and Gordon Betcherman. The vews expressed here are those of the author and need not reflect those of the World Bank, ts Executve Drectors or the countres they represent.

4 1. Introducton Ths paper reles on smple framework to understand the gender wage gap n Macedona and then smulates how the gender wage gap would behave after the ntroducton of a mnmum wage. A recent labor market assessment conducted n the country by the World Bank (Angel-Urdnola and Macas, 2008) ndcates that female labor force partcpaton n Macedona (at 49%) s one of the lowest n the ECA regon (and sgnfcantly lower than that of males, at 75%). The authors fnd that low labor force partcpaton for women s manly explaned by very low partcpaton rates among women wth low levels of educaton. Whle there are cultural values affectng the choce of low-educated women to engage n domestc producton, especally among women of Albanan orgn, low market wages, hgh reservaton wages, and harsh employment condtons seem to be mportant excluson determnants nfluencng uneducated women to stay at home (CRPM, 2008). Accordng to the law (Artcle 108 n the labor relatons law of 2005), employers shall be oblged to pay equal salary to employees for equal work wth equal responsbltes at the poston, regardless of ther gender. owever, estmates n ths paper ndcate that a large wage gap exst between men and women that s not necessarly explaned by labor segmentaton (whereby women enter sectors offerng lower-pay) or by dfferences n returns to educaton by gender, but more lkely by dscrmnaton (whereby men n smlar sectors, wth smlar educaton, and dong smlar obs earn hgher wages than ther female counterparts). Angel-Urdnola and Macas (2008) fnd that low labor moblty, especally among low-sklled women, strengthens the sense of local (and non-convergent) labor markets whereby dfferences n employment outcomes across regons are qute large. Nevertheless, workers especally women seem not to move from worse to better performng regons n order to seek better ob opportuntes. These combned factors are generally prevalent n labor markets where frms have monopsonstc power. The mplcaton of ths s that frms may be payng women workers below ther margnal product of labor, whch causes the supply of labor to be below that n a compettve settng (a feature that may be affectng women dsproportonally). Indeed, n some regons n Macedona, a large share of workng women s employed by a few large textle companes. These women clam havng to accept obs wth very precarous condtons (n terms n pay, safety, and workng hours) due to lack of alternatve employment opportuntes (CRPM, 2008). If ndeed low-educated workers n Macedona, especally women, face a monopsonstc market; standard economc theory would ndcate that the ntroducton of a mnmum wage above the monopolstc wage but below the compettve market wage would lkely ncrease employment wthout causng addtonal unemployment (Aaronson et al., 2008; Kaas and Madden, 2008; Josh and Pac, 1998). Whle the ntroducton of a mnmum wage n the prvate sector has been a polcy already consdered n Macedona, t has not been ntroduced. 1 Ths s because mnmum wage polcy s heavly nfluenced 1 Artcle 107 n the labor relatons law specfes that an employee s salary for carryng out full-tme work may not be lower than the mnmum salary determned by law and collectve agreement. owever, mnmum wages are not enforced n the prvate sector and are currently set unlaterally by the Mnstry of Labor and Socal Protecton as benchmark to determne the salary grd for publc servants. 2

5 by poltcal factors, such as aggressve lobbng by the textle ndustres (whch strve to reman compettve n the global textle market) and fears that the level of the mnmum wage, once ntroduced, would tend to ncrease rapdly to the pont that t may create negatve effects n employment growth (Angel-Urdnola, 2008; Brown et al., 1982; orrgan and Mncy, 1993; Machn and Mann, 1994; and Card and Ashenfelter, 1999). Furthermore, the ntroducton of a mnmum wage n Macedona, beng t the country wth the hghest unemployment n Europe (at 36%), seems a rsky move; especally at a tme when the government s movng towards makng ts labor market more flexble by reducng labor costs and tax wedges (Lebfrtz, 2008). On the other hand, Macedona s the process to accesson to the EU and thus looks to comply wth the Lsbon Agenda, whch targets female labor force partcpaton rates of 60% for EU-member countres. As such, the ntroducton of a mnmum wage as a tool to ncrease low wages and thus female labor supply especally among low-sklled women consttutes a polcy opton. Ths paper has two man contrbutons. Frst, t presents a new albet smple framework that decomposes the wage gap nto three factors: () a wage level factor, that measures the extent to whch the gender gap s drven by dfferences n wage levels among low-sklled workers of opposte gender; () a sklls endowment factor, that quantfes the extent to whch the gender wage gap s drven because the share of hghsklled workers dffers by gender; and () a returns to educaton factor, that measures the extent to whch the gender gap s drven by dfferences n returns to educaton by gender. Second, the paper presents smple set of smulatons that suggest that the ntroducton of a mnmum wage could contrbute to a decrease n the gender wage gap of up to 23%. The ntuton of ths result s very smple: low-sklled women earn much less than lowsklled men, despte rather low segmentaton and hgher returns to educaton. A smple Oaxaca decomposton ndcates that most of the wage gap among low-educated workers s due to unexplaned factors (generally assocated wth dscrmnaton; among other unobservable factors). The ntroducton of a mnmum wage wll lkely be more bndng for low-sklled/low-pay female workers and thus would contrbute to ncrease ther wages relatve to those of males. owever, these results should be treated wth care. In partcular, results n ths paper ndcate that n order to obtan a reducton n the wage gap that s statstcally sgnfcant, a rather hgh level of the mnmum wage (close the medan wage) may be needed. A hgh mnmum wage (set above the market-clearng prce of labor) could lead employers to move back along ther demand curves, causng a reducton n employment. The paper s structured as follows. Secton 2 descrbes the data used and presents smple descrptve statstcs to quantfy labor market segregaton and dscrmnaton. Secton 3 presents a smple decomposton, whch serves as a descrptve tool to analyze the gender wage gap. Secton 4 dscusses the results of the decomposton and quantfes (through smulatons) the effects of ntroducng a mnmum wage n the wage gap. A bref concluson follows. 3

6 2. Data and Descrptve Statstcs Ths paper uses data for year 2006 from the natonal Labor Force Survey (LFS) conducted by the Macedonan State Statstcal Offce (SSO). The Survey ncludes a rotatonal panel, whereby households are ntervewed more than one tme durng a year. The samplng frame, based on the 2002 Census, s stratfed, rotatonal, two stage random and natonally representatve. The data contans basc nformaton on demographcs, educaton, and labor market outcomes for ndvduals 15 years old and plus. Our sample ncludes approxmately 6,536 workng men and 4,445 employees between 15 and 64 years old. Due to the nature of the study, unpad and self-employed workers are not ncluded n the sample. About 77% of the sample s made of low-sklled workers, as defned as those wth less than complete tertary educaton. Most workers n the sample work n manufacturng (28%), wholesale (12%), publc admnstratons (10%), and educaton/health/socal servces (16%); whch consttute the largest ndustres provdng employment n Macedona. More detaled descrptve statstcs by gender, age-group, regon, educaton, and ndustry of employment are presented n Table 1. An nterestng feature of the labor market n Macedona s the exstence of a large wage gap between men and women. Fgure 1 plots the cumulatve densty functon (CDF) of hourly wages for both men and women. The fgure llustrates that the men s CDF stochastcally domnates that of women, suggestng that at all ponts of the wage dstrbuton males earn hgher wage rates than ther female counterparts. On average, uncondtonal estmates ndcate than men earn wages that are 25% hgher than women. 2 Table 2 provdes descrptve statstcs on populaton shares and wage rates by gender and educaton for all employees. An nterestng result s that gender wage dfferentals between men and women are large, especally among low-sklled workers n the prvate sector. A smple Oaxaca decomposton (Oaxaca, 1973) ndcates that only 17.4% of the gender gap among low-sklled workers s explaned by dfferences n endowments between men and women whle the remanng 82.6% of the gap s unexplaned; whch s generally attrbuted to the exstence of dscrmnaton (and/or other unobservable factors) n the labor market. Large gender wage gaps among less sklled workers could be explaned by labor market segregaton (Becker, 1971; Bergman, 1974; Johnson and Stafford, 1998) to the extent that women are more lkely to work n low-pay sectors than men. Table 3 provdes smple descrptve statstcs on wage dfferentals and gender segregaton by ndustry of employment. Results n Table 3 ndcate that whle there s some gender segregaton n Macedona wth men (women) beng more segregated n the manufacturng and constructon (educaton/heath/and socal servces) overall segregaton s not hgh. Indeed, the ndex of dssmlarty (one of the most used statstc for segmentaton) by 2 Smlar results usng a standard Mncer equaton on the natural logarthm hourly wages controllng for gender, age, age squared, educaton, regon, and ndustry of employment and other nteracton terms between a gender dummy and educaton categores ndcate that the gender gender wage gap for employees s about 27.3%. These results are avalable upon request. Also refer to Angel-Urdnola and Macas (2008) for smlar estmates. 4

7 ndustry s only at Results ndcate that the gender wage gap s qute hgh n segregated ndustres such as manufacturng and constructon (at 32% and 58% respectvely), as well as n non-segregated ndustres such as wholesale, hotels/ restaurants, and agrculture (oscllatng between 15 and 56%). Fnally, the gender wage gap s rather small (5 to 7%) among employees workng n utltes (electrcty, water, and gas) and transport and communcatons, both whch are generally men-segregated ndustres. The aforementoned results ndcate a rather weak correlaton between gender segregaton and the gender wage gap n Macedona. 3. A Framework to Decompose the Gender Wage Gap Ths secton presents a new albet smple framework that decomposes the wage gap nto three factors: () a wage factor, that measures the extent to whch the gender gap exsts because wages among low-sklled women are below those of lowsklled men; () a segmentaton factor, that quantfes the extent to whch the wage gap exsts because the share of hgh-sklled workers dffer by gender; and () a returns to educaton factor, that measures the extent to whch the gender gap exsts due to dfferences by gender n returns to educaton. The workforce s made up low-sklled (L) and hgh-sklled () workers that can be male (M) or female (F). Let = {,L} and =(M,F}. Let S denote the share of workers by gender where = 1 and let S denotes the share of workers by skll level where S = 1. The share of workers accordng to ther gender and skll level s denoted by s, wth s = 1, s = S, and S s = S. Let denote the average wage rate for ndvduals wth sklls and gender ; w L where w w. Returns to hgh-sklled workers by gender a prox for returns to educaton, denoted by r, s defned as: L w w r = (1) L w where Let s denote the average wage n the labor market as: W = S w, (2) L s L s w = + w w. (3) S S 3 The ndex s a measure from 0 to 1, where the hgher the number, the more segregated the two groups are. The formula for computng the Index of Dssmlarty by ndustry s D = 0.5 ( M M ) ( F F) where M (F) s the male (female) populaton of employees and M (F ) s the male (female) populaton of employees n ndustry. 5

8 L w L s By addng and subtractng the term w and by multplyng and dvdng by S both terms n the rght sde of (3) and smplfyng, we can obtan a formula to calculate the average wage rate by gender w as follows: L L L s w w L s s w = w + w ( + ) = L S w S S s L w + 1 r (4) S Equaton (4) ndcates that the average wage by gender s equal to the average wage of ther low-sklled workers tmes a hgh-sklls premum factor (normally greater than one) that depends on the returns to hgh skll labor, r, and on the share of hghsklled workers n the labor market, denoted by s S. The share of hgh-sklled workers s our prox for labor market segmentaton accordng to the worker s sklls level. Let s also defne the gender wage gap as: wm w f GAP = (5) W Replacng (4) nto (5), we get a smple decomposton of the gender wage GAP: L L w = M s + M w F s + F GAP 1 r M 1 rf (6) W S M W S F Equaton (6) decomposes the gender wage gap nto three man factors: 1. Wage level of low-sklled workers: proxed by w L W 2. Segmentaton: share of hgh-sklled workers by gender, proxed by s S 3. Returns to educaton: proxed by r : The decomposton presented by equaton (6) s useful to smulate changes n the wage gap due to polces that affect the wage level of low-sklled workers relatve to the populaton and labor market segmentaton (through nvestments on educaton, for nstance). Wthn the proposed framework, returns to educaton by gender as gven by the labor market. Note that equaton (6) could have been wrtten usng the wages of hgh-sklled workers as bass for the analyss. The choce of wrtng equaton (6) n terms of the wages of low-sklled workers s rather opportunstc as t provdes an advantage for analyzng changes n the wage level the low-sklled workers after the ntroducton of a mnmum wage. To smulate the effect of the ntroducton of the mnmum wage, we assume full complance so that all workers earnng below or at the mnmum wage would earn at the mnmum wage after ts ntroducton, so that: 6

9 w c = MW f w MW, (7) where c denotes the counterfactual wage of worker of level of sklls and gender after the ntroducton of a mnmum wage. Smulatons re-calculate the counterfactual wage gap usng the new level of ndvdual wages as specfed by equaton (7). Whle full complance s a rather hard assumpton, n practce s rarely acheved. As a consequence, the results of the smulatons presented here are lkely to overestmate the mpact of mnmum wages on the wage gap. 4. Results Ths secton analyzes the gender wage gap n Macedona based on the framework presented n secton 3 and then smulates the effect of ntroducton of a mnmum wage n the gender wage gap. Table 4 provdes the results of the decomposton. The frst row n Table 4 presents the average wage for low-sklled employees (by gender) as a proporton of the average wage. A value 1.03 for low-sklled males ndcates that ths group earns wage rates that are 3% above average. The same value s at 0.71 for females, suggestng that ths group earns 20% lower wage rates than average. Columns 2 and 3 n Table 4 ndcate that ths same feature occurs among employees n the prvate and non-prvate sectors. 4 The second row n Table 4 presents the share of hgh-sklled workers by gender. As expected, the share of hgh-skll workers n the non-prvate sector s larger than n the prvate sector (ths s probably drven by the fact that workers n the publc sector tend to be more educated). Interestngly, the share of hgh-sklled employees s hgher among women (19% among women vs. 28% among men), especally n the non-prvate sector (28% among women vs. 44% among men). The thrd row n Table 4 quantfes the returns to educaton of hgh vs. low-sklled workers by gender. Results ndcate that female employees dsplay hgher returns to educaton than male employees, especally n the non-prvate sector: wages of hghsklled males (females) are 38% (71%) hgher than those of low-sklled males (females). The fourth row n Table 4 dsplay a hgh-sklls premum factor, based on both educaton endowments and returns. Results ndcate that ths factor s generally hgher for women (1.20 vs for men), manly n the non-prvate sector (1.26 vs for men). 5 By multplyng the values n column (1) tmes the values n column (4) we get a number call t a gender factor. The gender wage gap can be calculated as the male factor mnus the female factor. For male employees, the factor equals 1.10 (the hgher the number the better). The factor for women equals The value of the factor for men (1.10) mnus the value of the factor for women (0.85) equals to the wage gap (0.25 or 25%). Results of the decomposton ndcate mportant dfferences between prvate and non-prvate employees. In partcular, among prvate employees the advantages n wage levels enoyed by low-sklled men over low-sklled women are offset by advantages 4 The choce of dong ndependent analyss for the prvate vs. non-prvate sector s not nnocuous. Angel- Urdnola and Macas (2008) fnd mportant dfferences n earnng by gender and n returns to educaton n the prvate vs. non-publc sectors. 5 Women n Macedona wth hgher levels of educaton are more lkely to partcpate n the labor market (Angel-Urdnola and Macas, 2008). 7

10 enoyed by women, who n turn dsplay hgher returns to educaton and a hgher endowment of hgh-sklled workers. As a result, the wage gap n the non-prvate sector s only at 6.3%. A nce feature of the decomposton s that t helps provde some further nsght as to what are the man drvers of the wage gap. In ths case, most of the gender wage gap s explaned by large dsadvantages n wage levels for low-sklled workng women. Nevertheless, the decomposton s slent about the reason for such dfferences. As such, t should be used only as a descrptve tool. To recaptulate, results n Table 4 ndcate that the gender wage gap (at 25%) s manly explaned by very low wage rates among low-sklled women as compared to the average wage rates a dsadvantage that s not evdent among low-sklled men. Ths phenomenon domnates other advantages dsplayed by female employees, such as hgher returns to educaton and a larger endowment of hgh-sklled labor. Results are manly drven by what occurs n the prvate sector. In the non-prvate sector, the gender wage gap s low (at 6.3%) despte the fact that low-sklled men stll earn hgher wages than low-sklled women. Ths occurs because men s advantages are offset because women s hgher returns to educaton and endowments of hgh-sklled labor. Fnally, Table 5 presents a seres of smulatons of how the gender wage gap would change wth the ntroducton of a mnmum wage. For smulatons, a level of the mnmum wage between 0.6 and 1 medan wage (ths s, between 30 and 50 Dnars per hour) s used. Ths s a range of the mnmum wage level that s common n developng economes (Maloney and Nuñez, 2006). As llustrated n Table 5, a low level of a mnmum wage (equvalent to 30 Dnars per hour) would not have much of an effect n the gender wage gap. owever, results ndcate that the ntroducton of a mnmum wage between 40 and 50 Dnars per hour could contrbute to a decrease n the gender wage gap of 15 to 23%. The ntuton of ths result s very smple: as dscussed before, low-sklled women earn much less than lowsklled men, despte women dsplayng hgher human captal endowments and returns to educaton. As such, a mnmum wage s lkely to become more bndng for low-sklled women than for low-sklled men and thereby lkely to mprove the wage level of lowsklled women much more than t would for low-sklled men. owever, n order to get a sgnfcant mprovement n the wage level of low-sklled women, a rather hgh level of a mnmum wage (close to the medan wage) s needed. 5. Concluson Mnmum wage polcy poses a tradtonal trade-off. Rasng the mnmum wage allows for the possblty of ncreasng the earnngs of workers at the lower tal of the wage dstrbuton by more than average and thus promotng postve effects n labor supply, especally among low-sklled workers. owever, a mnmum wage set above the market-clearng prce of labor wll lead employers to move back along ther demand curves, causng a reducton n employment. Evdence of mperfect competton n Macedona (due to hgh levels of dscrmnaton and low moblty) leads to the hypothess that frms may be payng workers, and especally low-sklled women, below ther margnal product of labor, whch causes ther supply of labor to be below that n a compettve settng. If so, the ntroducton of a mnmum wage at or below the 8

11 compettve market wage would lkely ncrease the overall level of employment. Results n ths paper ndcate that a large wage gap between male and female employees exsts, manly n the prvate sector, due to the fact that low-sklled males earn wages that are hgher that those of low-sklled women; despte lttle labor occupatonal segregaton and despte the fact that women dsplay hgher human captal ndcators. The ntroducton of a mnmum wage between 40 and 50 Denars per hour would contrbute to a decrease n the gender wage gap of between 15 and 23%. Ths occurs because ntroducng a mnmum wage s lkely to be more bndng for low-sklled women than for low-sklled men and thereby lkely to mprove the wage level of low-sklled women by more than t would among low-sklled men. owever, n order to get a sgnfcant mprovement n the wage level of low-sklled women, a rather hgh mnmum wage (close to the medan wage) would need to be ntroduced. As such, whle the ntroducton of a mnmum wage may lkely contrbute to ncrease the wages of low-educated women and ther levels of partcpaton n the labor market; t may also contrbute adversely to employment. The employment effects of mnmum wage polcy should be treated and assessed wth partcular care n Macedona gven ts very low employment and hgh unemployment rates. Besdes mnmum wage polcy, other polces amed at strengthenng market competton and at mprovng women s ob condtons and wages should be explored. Promotng hgher wages for low-sklled women workers s lkely to boost ther partcpaton, reduce the gender wage gap, and eventually contrbute to poverty reducton (Angel-Urdnola and Wodon, 2006). Fnally, the queston of whether or not Macedona has regonal markets wth monopsony power needs further testng, and more research amed to understand the demand-sde of the labor market n the country should be conducted. References Aaronson, D.; French, E.; MacDonald, J. (2008). The Mnmum Wage, Restaurant Prces, and Labor Market Structure. Journal of uman Resources, vol. 43(3) Angel-Urdnola, D. (2008). Can a Mnmum Wage Increase have an Adverse Inpact on Inequalty? Evdence from two Latn Amercan Economes. Journal of Economcs Inequalty, vol. 6 (1): Angel-Urdnola, D. and Macas-Essedn, V (2008). Macedona Employment Profle, Mmeo. The World Bank, 2008 Angel-Urdnola, D; and Wodon, Q. (2006). The Gender Wage Gap and Poverty n ColombaPrevew. Labour, vol. 20 (4): Becker, G The Economcs of Dscrmnaton. Chcago Unversty Press. Chcago, IL. 9

12 Brown, C., Glroy, C., and Kohen, A. (1982). The Effect of the Mnmum Wage on Employment and Unemployment. Journal of Economc Theory, vol. 20: Card, C., and Ashenfelter, O. (1999). Mnmum Wages, Employment and the Dstrbuton of Income. andbook of Labor Economcs, vol. 3b, chapter 32. Cotter, D; ermsen, J.; and Vanneman R. (2004) "Gender Inequalty at Work," The Amercan People: Census 2000, ed. Reynolds Farley and John aaga, New York: Russell Sage Foundaton. CRPM (2008). Constrants to Labor Force Partcpaton n Macedona: A qualtatve approach. Mmeo. World Bank, Washngton DC. orrgan, W., and Mncy, R. (1993). Uneven Tes. Edted by Sheldon Danzger and Peter Gottschalk. Russell Sage Foundaton. New York. Johnson, G., and Stafford, F Alternatve Approaches to Occupatonal Excluson. Women's Work and Wages. Research n Gender and Socety, vol. 2: Josh,. and Pac, P. (1998) Unequal pay for women and men: Evdence from the Brtsh brth cohort studes. Edted by Gerald Makepeace and Jane Waldfogel. Cambrdge and London: MIT Press Kaas, L.; Madden, P.(2008). oldup n Olgopsonstc Labour Markets--A New Role for the Mnmum WagePrevew; Labour Economcs, vol. 15(3): Machn, S., and Mannng, A. (1994). The Effects of Mnmum Wages on Wage Dsperson and Employment: Evdence from the U.K. Wage Councls. Industral and Labor Relatons Revew, vol. 47, no.2: Maloney W., and Nuñez, J.. (2001), Measurng the mpact of mnmum wages. Evdence from Latn Amerca Polcy Research Workng Paper No World Bank. Washngton. DC. Oaxaca, R. (1973). Male Female Dfferentals n Urban Markets. Internatonal Economc Revew, Vol. 14:

13 Table 1: Descrptve statstcs on the sample [employed workng age populaton] Male Female All Employment Rates Employed Indvduals Sample sze 6,536 4,445 10,981 Weghted sample 240, , ,564 Age Group % % % % Educaton % Low Skll % gh Skll* Regon % Skope % Btola % Veles % Kumanovo % Ohrd % Strumka % Tetovo % Shtp Industry % Agrculture/mnng/fshng % Manufacturng % Elec/Gas/Water % Constructon % Wholesale/retal % otels/restaurants % Transport/communcaton % Fn & Real State Svs % Publc Admn % Educaton/health/Socal Work % Other servces Ownershp % Prvate % Non prvate** Total Source: Author s estmates usng 2006 Macedona LFS data. * gh-skll workers are defned as those wth at least complete hgher educaton (unversty and above). ** Includes publc and sem-prvate companes such as utltes (gas, water, and electrcty). 11

14 Table 2: Average wage rates (n Dnars per hour) and populaton shares by gender and level of educaton. All Employees Prvate, Non agrculture Non-prvate Populaton shares Average Wages Rate Populaton shares Average Wages Rate Populaton shares Average Wages Rate Gender Male 59% % % 70.5 Females 41% % % 66.2 Educaton gh-sklled 23% % % 88.1 Low-sklled 77% % % 58.9 Gender and Educaton Low-sklled females 30% % % 52.5 Low-sklled males 47% % % 62.0 gh-sklled females 11% % % 83.7 gh-sklled males 11% % % 92.5 Total 100% % % 68.9 Source: Author s estmates usng 2006 Macedona LFS data. * gh-skll workers are defned as those wth at least complete hgher educaton (unversty and above). Table 3: Uncondtonal gender wage gap and occupatonal segregaton [low-sklled workers only] Share of employment Average wage rate Male (1) Female (2) Male (3) Female (4) M/M F/F (5) Gender Gap (3) / (4) Agrculture/mnng/fshng % Manufacturng % Elec/Gas/Water % Constructon % Wholesale/retal % otels/restaurants % Transport/communcaton % Fn & Real State Svs % Publc Admn % Educ./health/Socal Work % Other servces % Index of segregaton 0.33 Source: Author s estmates usng 2006 Macedona LFS data. The formula for computng the Index of Dssmlarty by ndustry s: D = 0.5 ( M M ) ( F F) where M (F) s the male (female) populaton of employees and M (F ) s the male (female) populaton employed n ndustry. 12

15 Table 4: Results of the decomposton of the Gender Wage Gap All Employees (1) Prvate, Non agrculture (2) Non-Prvate* (3) Males Females Males Females Males Females (1) w L W (2) s S (3) r (4) s r + S (1) x (4) Wage Gap 24.8% 42.5% 6.3% Sample Sze N 6,536 4,445 3,086 2,348 3,264 2,037 Weghted N 240, , ,770 86, ,027 74,406 Source: Author s estmates usng 2006 Macedona LFS data. *Includes publc and sem-prvate companes such as utltes (gas, water, and electrcty). Table 5: Smulatons of changes n the Gender Wage Gap after the ntroducton of a mnmum wage [all employees] MW=30 Dnars/hours (0.6 the medan wage) MW=40 Dnars/hours (0.8 the medan wage) MW=50 Dnars/hours (medan wage) Males Females Males Females Males Females (1) w L W (2) s S (3) r (4) s 1 + r S (1) x (4) Wage Gap % 19.0% % reducton n the Wage Gap 2% 15% 23% Source: Author s estmates usng 2006 Macedona LFS data. *Includes publc and sem-prvate companes such as utltes (gas, water, and electrcty). 13

16 Fgure 1: Oaxaca Decomposton [Low-sklled Employees only] 1.8 Cumulatve Probablty lhwage c.d.f. of female c.d.f. of male Source: World Author s estmates usng 2006 Macedona LFS data. 14