# Abstract

This study analyses the woman's status labour market of Khyber Pakhtunkhwa. Four working states: self-employed, paid employees, and unpaid family helpers were investigated. Data were collected about individuals and household characteristics of women aged between (15-60) years from the Pakistan Social and Living Standard Measurement Survey (PSLM, 2014-15). The estimated results based on Multinomial Logit (MNL) suggest a positive and significant impact of women’s age on all working categories in the labour market. The woman who owns a house, or the married woman, with multiple children or having a combined family system, or the residents of the countryside have less likelihood to take part in paid works. Participation in paid works decreases with the increase in the number of children, whereas participation in self-employment increases with the increase in the number of children. The probability of female participation in all four working states increases with the increase in the number of working individuals in the family. Whereas, probability of women's participation in the labor market decrease with the Joint family system, house owning, marriage, or higher household income.

# Key Words:

Labour Market, Women Status, Labor Force Participation, Multinomial Model

# Introduction

The labour market has a key role in the macroeconomy, and that is because of its forward and backward interconnections. The forward linkage with the labour market expansion develops those sectors that utilize labour and generate demand for labour. The growth of the labour market promoting all those sectors of labour supply, i.e. education, skill and training, and creating opportunities for employing both the skilled and unskilled individuals is associated with backward linkage. Similarly, the output level, the standard of living, production and average income increase, which is is essential for the economic development of a country. For the last more than thirty years, women labour force participation has been witnessed to significantly increase all over the world. In advanced countries, an increase of 4% to 70% in the female’s active labour force has been comprehended (Hotchkiss, 2006). Following the SPDC Annual Report (2008), Pakistan in the years 2004-2007 witnessed high growth rates of women labor force participation; still at lowest level. The participation rate estimated was 49 percent in 1971-72 and was 52.5 percent in 2007-08.

The participation rate of females increased from 9 to 22 percent in the time period from 1971-72 to 2007-08. Almost 78 percent of the total females are economically inactive in the labour force, in contrast to seventeen percent of males that are inactive. The data of key indicators for labour market studied from 1980 to 2007 for sixteen selected countries has shown the net growth of female participation rate in the work force. The rate of participation increased from 5.5% to 16% in the time period of 1980 to 2000 and up to 20.8% in the year 2007. The women input rate in Thailand was found to be 69.3%, in the Philippines 70.6 %, in China 49.8%, and in Korea 49.3% in the year 2007, in the workforce (ILO, 2014).

Materials and Methods

Data

The micro-level data of “the Pakistan Social and Living Measurement (PSLM)” 2014-15 supervised by the “Pakistan Bureau of Statistics (PBS)” is used for this study. The important information about 25,999 households from 25 districts of Khyber Pakhtunkhua-Pakistan.

Model Specification

In order to test hypothesis multinomial logit model was used. The variables description is given as: Women Employment Status = Dependent variable is female’s status of employment in labour market comprising working and not working females. Employment status= 1 if the working female and if 0 for non-working female. The working categories are coded as Y: 0= not working and 1= if the self employed   female in non agri, 2 = if the female is paid employed, 3 = if the self-employed female in agriculture sector, 4= if female is unpaid family helper. The family size, household monthly earning, age, marital status, area, and the education of family head are the control variables.

Greene (1992): the probability of women involvement in the working states is studied as a function of the observed characteristics by using multinomial logit model:

$y-y_0=m(x-x_0)$

The final equation is as follows:

$y-y_0=m(x-x_0)$

Table 1 explains the explanatory variables. The minimum age in the PSLM for all males and females is zero, and the maximum age is 99 years. However for the current study age range from the minimum to maximum is 15-60 years. The level of education considered for this study is primary, secondary, and higher. The minimum years of schooling completed are 0 or less than one years (Montessori) and (PhD), i.e. 19-years of education are considered the highest level of education. Dummy variable married is considered for marital status, which is 1 if married and 0 if un-married which includes (divorced, widowed and singles). The dummy variable for women head shows value one and zero in the summary figures. The number of dependents in data i-e kids aged below 6 years and elders who are above sixty years of age per household, is 23. Sixteen is the absolute number of working individuals in the family who share one kitchen. The maximum total number of kids aged 6-10 years is 12. The value for respondents who are residents of a rented house is zero, and the woman who owns a home is one, one percent of the respondents possess accommodation. Zero is for the nuclear family systems whereas, for individuals having combined-family systems (i.e., two families, more than four couples in a house, or more who shares a single kitchen) is equivalent to one. The maximum revenue of the participants in a household is 1350000 Rupees, the minimum wage is zero, and 10 is the logarithm of family circle. The zero for least amount and one for highest are shown by the location dummy for rural and non-rural areas.

Table 1. Description of Explanatory Variables

 Variables           Description Age              (15-60) years Age2            age quadratic Edu                the completed years of schooling. Married1 =married and 0= unmarried (widow, divorced or single). Women head1 = head of household,0 otherwise. Rural/Urban 1 =Urban area and 0= Rural. HHedu1= head of the household  is educated 0= if not Ln(Income)       log of  family monthly income Ln(Income)2          log of family monthly income squared Working people      a household number of working individuals Co-residence    1= joint family, 0 = nuclear Childrenkids in number aged (6- 10 years) Dependents      household dependents in numbers(age<5&>60and not working) Own House      1 =living in own house 0 = otherwise(rented house, on subsidized rent)

The total number of not-working and working-women aged (15-60) according to the sample is 25999. Approximately 25% of the total sample is the ratio for working women which is equal to 6349 women out of the total. While the not working women are 19650, about 75% of the total sample women.

Table 2. Women Category-wise Observations

 Category Observations Not working 19,650 Self employed Non-Agriculture Sector 1,107 Paid Employed 2,417 Self Employed in Agriculture Sector 1,311 Unpaid Family Worker Total 1,514 25999

Source: Author’s calculation based on PSLM 2014-15

Results and Interpretation

Female Participation by Multinomial Logit Results

The determinants of female labour

participation in the Khyber Pakhtunkhwa is examined by the logit regression result is presented in the table below;

Table 3. Summary Statistics

 Variables N Mean SD. Min Max Age 25,999 1.85 2.59 5 60 Age2 25,999 1173 00.6 25 600 Married 25,999 0.718 0.450 0 1 Location 25,999 0.0909 0.287 0 1 Education 25,999 6.088 0.027 1 6 No.of Working 25,999 1.990 1.823 0 16 No.of children 25,999 4.920 1.485 0 12 No.ofDependents 25,999 8.773 2.179 0 23 Co-residence 25,999 3.586 0.464 0 1 Ln(income) 25,999 0.315 0.772 4.419 10.40 Ln(income)2 25,999 87.85 13.70 19.53 108.1 Women head 25,999 0.0760 0.265 0 1 Own house 25,999 0.895 0.306 0 1

Source: Author’s calculation based on PSLM 2014-15

Table 2 sums up the statistics of all the explanatory variables utilized in the estimation procedures. The stylized facts of Pakistan data are clearly reflected. 25999 is the total count for all observations of women.

Table 4 describes the estimated outcome of the women employment status in Khyber Pakhtunkhwa.

Table 4. Multinomial Logistic Regression Results

 Explanatory Variables Self Employed Non-Agriculture Paid Employee Self-employed Agriculture Unpaid family helper Age 0.00038 0.0311 0.00111 0.0015*** age2 4.41e-05* 0.000301* 0.000016* 0.0002** Education 0.0081* 0.02* 0.0035* 0.00182* Woman head 0.0039* 0.158* 0.0035* -0.0012* Married -0.011 -0.061* 0.0123* 0.0081* No.of dependents -0.023* -0.0377* -0.0115* -0.007* No.of working people 0.0015* 0.017* 0.007* 0.016* Ln(income) 0.0904* 0.3933* 0.0207* 0.052*** Ln(income2) -0.0036* -0.247* -0.002* -0.0034** No.of children -0.035** -0.0006* 0.007 -0.0035* Coresidence -0.055* -0.058* -0.016* -0.021* Location 0.0201* -0.008* -0.0287* -0.0085** Own house -0.016* -0.019* 0.011* 0.005 Constant -16.50** -27.99*** -6.135 -14.40** MultinomialLogistic regression       Number of obs   =       8562 LR chi2(48)     =    2909.03Prob > chi2     =     0.0000 Log likelihood = -6395.4722                                             Pseudo R2       =     0.1853

***, **, * Represent significance level at 10%,5% and 1% respectively.

This study takes the four working states for women; not-working as the base category and as the dependent variable. It is an established concept among researchers that the grouping of family characteristics, the individual’s personal characteristics, and a number of demand-side and supply-side factors form a basis for the decision-making process to participate in any economic activity. The interpretation for the results of involvement in various forms of employment in the workplace market was computed by the marginal effects. (Table 2) shows that 31.8 is the mean value of age for women. As per the results, the probability of participation in all forms of service work increases with the addition of years of age from its mean. Although the likelihood of being not-working in contrast to being an unpaid family worker for women is very low (i.e. being an unpaid family worker for women has a high probability equal to 0.15 percentage points (pp)), their magnitudes are different. Age square is included in determining a non-linear association between the likelihood to participate in any state of employment and age. Results indicate that any increase in the women age-square,  the probability of being in the self-employed in non-agriculture was very low (4.41e-05pp) as compared to those who are not part of the workforce. However, the immensity remained very low, and the age-squared had a significantly positive impact on determining the probability of women for labour supply. These findings are in accordance with the literature ((Naqvi et al. (2002), Ejaz (2011)).

As per the study of (Hafeez and Ahmad (2002) Safana et al. (2011) the probability of participation in the labor marketplace has a direct relationship with education, in our analysis, 50% of women are uneducated in the total sample, which refers to the 6 years of education as the lower mean value for them, the results further indicate that any addition to the years of education from the mean values leads to an increase in the likelihood for women being in a self-employment in the agriculture sector (0.35pp) and in paid employment (2.41pp). The married women in our sample constitute 71%. The likelihood of salaried work for the wedded women was lower than the unmarried in the labour market, i.e. (0.61pp). It means that majority of married females are more likely to work as unpaid family helpers (0.81pp) or of self-employment in the agriculture sector (0.123pp). The findings of the current study are congruent with the previous studies (Naqvi et al. (2002), Ejaz (2007)). Children between 6 to 10 years of age are considered a strong determinant of women’s labour supply. Almost 45% of women had 2 or more children in our analysis. (Bradbury and Katz (2005) suggest that majority of women work in the agriculture sector or are in unpaid jobs as compared to non-working women.

The results further reveal that an increase in standard deviation for the kid's number in the family lower the chance for women to be in a salaried service by (0.089pp)(1), and the voluntary household assistant (with school-going children) has less likelihood to participate as compared to the ones who are not the part of labor state. This demonstrates a consistent result with the previous literature (Naqvi et al. (2002), Bradbury and Katz (2005)), which states that the number of children has negative relation with the female labour supply. In this research, the dependents include all infants of age (less than 5 years and elders greater than 60 years) in the household. Another study explains the negative relation of a number of dependents and women participation in labour market (Faridi and Basit (2011), Faridi et al. (2009)). Study findings suggest that the likelihood be (paid employment, unpaid and self-employed) lowers with an increase in the standard deviation in the number of dependents in a family. “Co-residence” means “To live in a joint-family system with parents or with in-laws, consisting two or more families who share one kitchen”. In our analysis, 31% of the women are living in a combined family system. Results show that women living with joint families have a ghost of a chance to be in any of the state of employment comparative to non-working women. The monthly income of the household is another strong determinant that affects the women supply to the labour market. Results show that the logarithm of the monthly household income has a significantly positive relationship with the women’s decision to work in the labour market. This suggests that the probability of women employment increases with an increase in family monthly earnings (log-level). Conversely, the lower probability for women participation in any employment category can be the outcome of the increase in its quadratic term.

The marginal effects of household (logarithm income) show that the probability of females being in any labour market state increases with a one percent increase in monthly family income, with a falling rate comparative to not-work. Although probability of being in paid employment increases with a high magnitude of 3.43(pp), it reveals that household monthly earnings have a significant positive relation with women decided to work; this is consistent with the findings of Ejaz (2011),Esfahani (2012)).

The probability of being a volunteer household-helper, who does not get paid, remains high by 0.5 pp and self-employment in the agriculture sector by 1.1pp. Whereas the female’s possession of a home employs 1.9 pp less likely to be in a paid-work relative to not being part of a labor market.

The results suggest that number of working people also have an influence on the probability of women’s participation in labor market. In a household, the rise in the number of working people by one standard deviation, Females are more likely to take part in self-employment non-agriculture sector, paid employee,  free household work and self-employment in agriculture-sector by the magnitude of 0.15,1.7,1.6, 0.7 (pp) respectively, relative to not work. The results of Naqvi et al. (2002) and Ejaz (2007)’s work support these findings.

Females as the head of the home are more likely to participate in all working states except unpaid family work. This shows that being a supervising figure in the house, women do not like to work with no any financial-reward or voluntarily as un-paid household workers/helpers. However, in order to meet large family expenses or to offer a superior education to children, some married women as leaders of the home may decide to work in labour market. These findings adhere to the previous studies of (Ejaz (2007), Azid et al. (2010)); which established a positive relationship between women heads of the household and the probabilities of being in employment. To check the role of urban/rural impact on women participation, their dummy was utilized as a regional control. Women belonging to city areas had 0.8 (pp) fewer chances to be in paid employment. Conclusively, these findings suggest that women who are married, is head of the house, have 2 or more children, lives in a combined family system or belongs to non-rural areas has a lower probability of being in remunerative employment.

Conclusions

This study identified and analyzed the major determinants of women’s participation in the labor marketplace of Khyber Pakhtunkhwa. Data analyzed for women (aged 15-60) was taken from PSLM Survey (2014-15), using the multinomial Logistic regression model. The major findings suggest that for women, attainment of higher education leads to an increased participation rate in the labour market. The probability of participating is higher if a woman is unmarried, has a high household income, owns a house, has a joint family and belongs to a rural locality. The empirical study suggests that family labour supply has a positive relationship with the number of working people in a household. The more the number of kids and dependents, the less will be the likelihood of involvement (in any state of service) of women in the labor marketplace. The research, therefore, discloses that education has a vital role in the women labor supply and the data collected shows that 50% of females are uneducated, so the government should work on women education to enable them to participate in labour force and in the monetary development of the country.

Ahmad, E., % Hafeez, A. (2007).Labour supply and earning functions of educated marriedwomen: A case study of Northern Punjab. The Pakistan Development Review 46(1), 45-62.

Ahmad, M. (2001). Estimation of distribution of income among various occupations/professionsin Pakistan. Pakistan Economic and Social Review 39(2), 119-134.

Azid, T., Khan, R. E. A., % Alamasi, A. M. (2010). Labor force participation of married women in Punjab (Pakistan). International Journal of Social Economics 37(8), 592-612.

Becker, G. S. (1965). A theory of the allocation of time. Economic journal 75(299), 493-517.

Chiappori, P. A. (1992). Collective labor supply and welfare. Journal of political Economy 100(3), 437.

Bradbury, K., % Katz, J. (2005). Women's rise: A work in progress. The Federal Reserve Bank of Boston Regional Review 14(3), 58-67.

Ejaz, M. (2007). Determinants of female labor force participation in Pakistan an empirical analysisof PSLM (2004- 05) micro data. Lahore Journal of Economics Special Edition, 203-233

Ejaz, M. (2011). The determinants of female labor force participation in Pakistan: An instrumentalvariable approach. Centre for Research in Economics and Business Working Papers No. 01-11.

Esfahani, H. S., % Shajari, P. (2012).Gender, education, family structure, and the allocation oflabor in Iran. Middle East Development Journal 4(02), 1-40.

Faridi, M. Z., Malik, S. % Basit, A. (2009). Impact of education on female labour force participationin Pakistan: Empirical evidence from primary data analysis. Pakistan Journal of SocialSciences 29(1), 127-140.

Faridi, M. Z., % Basit, A. (2011). Factors determining rural labour supply: A micro analysis. Pakistan Economic and Social Review 49(1), 91-108.

Finegan, T. A. (1962). Hours of work in the united states: A cross-sectional analysis. The Journalof Political Economy 70(5), 452-470.

Gronau, R. (1973). The effect of children on the housewife's value of time. Journal of Political Economy 81(2), S168-99.

Hafeez, A., % Ahmad, E. (2002). Factors determining the labour force participation decision ofeducated married women in a district of Punjab. Pakistan Economic and Social Review 41(1),75-88.

Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica: Journal of the econometric society 47(1), 153-161.

Hotchkiss, J. L. (2006). Changes in behavioral and characteristic determination of female laborforce participation, 1975-2005. Economic Review-Federal Reserve Bank of Atlanta 91(2), 1-20.

ILO. (2014).KEY INDICATORS OF THE LABOUR MARKET. International Labour Office,Geneva.

Khadim, Z., % Akram, W. (2013)Female Labor Force Participation in Formal Sector:An Empirical Evidence from PSLM (2007-08).Middle-East Journal of Scientific Research 14(11), 1480-1488, 2013

McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior.In Z. P. (Ed.), Frontiers in Econometrics,Academic press.pp. 105-142

Mincer, J. (1962). Labor force participation of married women: A study of labor supply, Aspects of labor economics. Princeton University Press.pp.63-106.

Naqvi, Z. F., Shahnaz, L., % Arif, G. (2002). How do women decide to work in Pakistan? The Pakistan Development Review 41(4), 495-513.

Safana, S., Sial, M. H., % Awan, M. S. (2011). Female labor force participation in Pakistan: A case of Punjab. Journal of Social and Development Sciences 2(3), 104-110.

SPDC. (2007-08) Social development in Pakistan, Annual Review, Social Policy and Development Centre Karachi.

Sutradhar, R., Dali, R. K., Sarker, M. E., % Hossain. (2017). Socio-economic and Demographic Factors associated Women's Labor Force Participation in Rural Bangladesh.Antrocom Journal of Anthropology 2(2017), 129-137.

Yakubu, Y. A. (2010). Factors influencing female labor force participation in South Africa in 2008. African Statistical Journal 11, 85-104.