Abstract
This article utilizes a rare Time Use Survey (TUS), focusing on Pakistan, to uncover productive labour market activities that often go unnoticed in mainstream labour force surveys (LFS). Leveraging rich time use data along with labour force classification question, we identify and analyse the invisible workforce. Moreover, employing the multinomial logit model, I examine the determinants—such as human capital accumulation (or lack thereof), mobility constraints and financial well-being—of the invisible labour force among women and men aged 10–74 years in Pakistan. The findings reveal significant gender disparities within the invisible workforce, with women constituting a staggering 88% of its members. These women predominantly engage in part-time work concurrently with other activities within their own dwellings, particularly in sectors such as textiles, crafts and animal husbandry. Furthermore, I uncover that the lack of human capital and mobility constraints significantly increase the probability of participation in the invisible workforce. This article tackles the challenge of accurately measuring women’s engagement in productive work by identifying and examining the ‘invisible workforce’ through a unique survey method. Notably, this TUS stands out as the only one available in the South Asian context that integrates LFS questions to identify and study the invisible workforce. The implications of these findings extend to the development of more inclusive measurement frameworks and the promotion of gender equality in labour force participation.
Introduction
Accurately measuring the composition of the labour force, especially in relation to informal employment, is a significant challenge faced by labour statisticians. In addition, one critical issue that arises from inaccurate measurement is the exclusion of women from the labour force, particularly in the context of informal work. The heterogeneity of informal work, characterized by variations in time, location and regularity, poses difficulties in its documentation and understanding the decision-making processes of individuals involved (Hirway & Charmes, 2006; Radchenko, 2014, 2017; Sahoo & Neog, 2017). Unfortunately, existing measurement approaches often fail to capture the full extent of women’s participation in the labour force, leading to their exclusion from official statistics.
Previous studies have extensively explored measurement issues within the labour force and household surveys, recognising the limitations in accurately accounting for women’s engagement in informal work (Charmes, 2004; Hirway, 2002; Jain, 2008). To address this limitation, researchers have also explored the potential of Time Use Surveys (TUS) in providing a more comprehensive understanding of women’s work (Floro & Komatsu, 2011; Hirway, 2003; Hirway & Charmes, 2006; Hirway & Jose, 2011; Ironmonger, 2008; Saha, 2003). By acknowledging the problem of excluding women from labour force statistics due to inaccurate measurement, the literature highlights the need for understanding what this labour force is doing and why they remain excluded. These insights underscore the importance of developing inclusive measurement frameworks to ensure that women’s labour force participation is accurately accounted for and valued.
This article aims to address the exclusion of women from labour force statistics by identifying the ‘invisible workforce’; and studying their characteristics and work. Using the Pakistan TUS 2007, this article analyses the factors influencing participation in the invisible workforce. Following the approach of Floro and Komatsu (2011), I distinguish between the ‘invisible workforce’ who perform productive activities, as per the time use data, but do not report labour market work (LMW) 1 in surveys and the ‘visible workforce’ who accurately report their participation. While the dataset is old, we employ it as it has the unique ability to identify the workforce that remains invisible otherwise.
This article is the first, to the best of my knowledge, to empirically estimate the choice of participation in the invisible and visible workforce specifically for a South Asian country. Contrary to Floro and Komatsu’s (2011) findings on African labour, where married men with less than 8 years of education had a higher probability of participating in LMW despite reporting themselves as non-participants, the characteristics of the South Asian labour market differ significantly. Moreover, this study investigates the reasons behind the invisibility of such work by examining the type of work performed, its location and simultaneous activities.
Even today, the female labour force participation in Pakistan is one of the lowest in the South Asian region (Figure 1). Female labour force participation is stigmatized in patriarchal societies like Pakistan, which contributes to a reportedly low female labour force participation, as shown in Figure 1.
Female Labour Force Participation Rate (%) of South Asian Countries.
This article contributes to two strands of the inter-related literature. First, regarding the measurement of labour force participation and second, concerning the determinants of low female labour force participation. Previous studies highlight the plight of home-based workers, simply assuming that a large proportion of their productive activities remain undocumented. There is extensive qualitative and descriptive literature examining the characteristics and profile of these home-based female workers (Akhtar, 2011; Bajaj, 1999; Hassan & Azman, 2014; Hiralal, 2010). While they assume that the productive activities of all home-based workers remain invisible, they do not empirically test if these home-based workers actually remain invisible in the employment statistics. This study uses the Pakistan TUS, which in addition to daily time diary, has labour force survey (LFS) questions included in it, to identify both invisible and visible workforce and examine their determinants.
Literature identifies various factors explaining the underestimation of LMW. Unconscious factors are related to the respondent’s lack of understanding of the term ‘work’. Elements contributing to this include the precarious and scanty nature of the work, atypical work location and frequent overlapping of LMW with other activities such as unpaid reproductive work 2 (URW) as the primary factors contributing to underestimation of labour force statistics. I analyse each of these factors descriptively, taking advantage of the detailed time diary that incorporates the information of simultaneous activities and their location.
The conscious factors point towards the reasoning behind individuals who are pushed to intentionally hide their LMW. This is particularly observed in patriarchal societies, where gender roles defined by the society assign financial responsibilities to men and household and care work to women. Any deviation from the assigned roles imposes social costs (Akerlof, 2017; Akerlof & Kranton, 2000) would result in penalty by the society possibly instigating these women to intentionally hide their work. However, this aspect is beyond the scope of this article and is left for future studies.
The findings highlight the exclusionary dynamics within labour market statistics, particularly in relation to women’s work. Females constitute a significant majority of the invisible workforce, accounting for 88%. By employing the TUS method instead of the LFS method, the inclusion of invisible workers significantly raises the proportion of employed women (30% compared to 17.5%). The invisible workforce comprises a large number of women engaged in textiles (62%) and animal husbandry (32%), predominantly operating from their own dwellings. In contrast, men in the invisible workforce are involved in animal husbandry (60%) and farming (30%) either at their own dwellings or on farms. Moreover, the study indicates that both men and women may unknowingly misreport their LMW due to various factors. These include limited education, which restricts their comprehension of LFS questions, atypical workplace arrangements (especially for women working from home) and the absence of remuneration associated with their work. This particularly affects unpaid family workers who tend to assume that the labour force only includes individuals who earn a positive income. These findings shed light on the systematic exclusion of certain forms of work, particularly those performed by women, from official labour market statistics.
Reasons for Underestimation of Labour Force
Here, I examine the factors why LMW may be misreported unintentionally. At the data collection end, at times the labour survey questions fail to capture the precarious and intermittent jobs, like day labourers or seasonal farm workers because of the nature and timing of their work (Hirway, 2003).The International Labour Organization (ILO) improved the definition of ‘employed’ to include the marginal activities in the labour force estimates. The 1-hour criterion was introduced in 13th ICLS Resolution to interpret ‘some work’ in ILO’s definition of employed. This implies that as little as 1 hour of economic activity over a short reference period of one week or a day need to be accounted for in LFSs. The objective was to ensure that all type of jobs (casual and temporary, standby work or work in informal sector and other informal employments), regardless of the time spent (part-time and short-time), shall be included in the employment statistics (Hussmanns, 2007).
The LFS of Pakistan included probing questions in 1990 and added a section to reckon its size and composition in 1995. Following the ILO’s 1-hour criterion, Pakistan’s LFS asks; ‘Did you do any work for pay, profit or family gain during last week, at least for an hour on any day?’ Further, to assure no unpaid family worker is left out, further probing questions were included; ‘Did you help to work for family gain in a family business or family farm during last week?’ (Labour Force Survey, 2015). These changes were an effort to assure that the productive work does not remain unaccounted for.
At the respondent’s end, there may be several issues due to which they would unintentionally misreport economic activities. Factors such as casual and scanty work, non-standard workplace and simultaneous performance of LMW and other activities hamper respondents’ understanding of the term ‘work’. A rare or seasonal economic activity, may not be recalled easily. Similarly, some activities like waste picking and street vending are, at times, not acknowledged as ‘work’. In addition, an atypical workplace such as street, back garden or own dwelling combined with overlapping simultaneous LMW and URW, makes it arduous to recognize and identify economic activities. For instance, a woman looking after children and teaching them while simultaneously stitching or doing embroidery may just view the LMW as an extension of the URW. These scenarios are extremely relevant for the unpaid family workers, self-employed and home-based workers who are largely women (International Labour Organisation, 2015; World Bank, 2011). Women have a higher tendency to work from home, frequently overlapping the LMW and URW because of two restrictive codes prevailing under patriarchy.
Hence, the empirical section will be categorically examining the types of activities these individuals are performing, their location and how and when are they performed. In addition, empirically, socio-demographic factors associated with probability to participate in invisible workforce are also examined.
Data
Pakistan’s TUS 2007 is a nationally representative dataset. The survey, conducted among 19,247 households between January and December 2007, includes data on the demographic and socioeconomic indicators of a household. From each household, two respondents, age 10 years or above, are selected systematically through a selection table based on age rank.
The time diary section comprises of 144 30-minute episodes, with each accounting for up to three activities performed simultaneously or sequentially. All the activities are classified by the Pakistan Bureau of Statistics which follows the United Nations Statistical Division System of National Accounts (SNA) (Government of Pakistan, 2007). There is a location variable identifying where the activity took place; own dwelling, someone’s dwelling, field/farm, other workplaces or educational establishments. Face-to-face interviews were conducted throughout the year including weekdays and weekends. Labour market questions in this survey are identical to the LFS questions that helped compare the time use patterns by employment status, using population weights. In my analysis of labour force participation choice, invisible or visible, the sample is restricted to individuals aged 10–74 years. The Labour force statistics in Pakistan are based on the age bracket of 10+ years with no upper limit. However, respondents beyond the age of 74 working in the labour market serve as outliers and have been omitted from the analysis. This results in a sub-sample constituting of 17,784 female and 16,384 male respondents.
As a quality check, I make sure the time diary data is for 24 hours (1440 minutes) and no observation was found to be violating this condition in this data. In addition, observations which reported the previous day was not a normal day for any reason were filtered out. For instance, if it was a religious holiday, or respondents were occupied with other social/cultural activities, then it was not reported as a normal day.
Identification of Invisible Workforce
Following the methodology of Floro and Komatsu (2011), TUS is used to accurately determine the employment status of the respondents. Figure 2 exhibits the comparability of the employment status classification using labour market questions from two surveys, TUS and LFS. The employment status is classified using LFS questions and time diary data. As per the time diary method, the total time spent on the activities considered as ‘work’ is calculated for the day. Based on the definition of ILO, anyone working for at least more than 1 hour (60 min) in short reference period, that is, any given day in case of Pakistan is classified as employed. Therefore, anyone who, on average, works for more than 1 (60) hour (minutes) is considered employed. Part-time workers are classified as those who work for more than 1 (60) hour (minutes) but less than 5 (300) hours (minutes) on a given day and hence, anyone who works more than 5 (300) hours (minutes) is classified as full-time worker. A series of subsequent questions help in differentiating between the people who are unemployed and not in the labour force.
The Classification of Employment Status by Two Comparable Surveys Using Identical Labour Market Questions.
Table 1 compares the employment classification using the two methods explained above. The individuals who are classified as employed (full-time or part-time workers) in both the methods are defined as Visible Workforce. On the other hand, individuals classified as unemployed or not in the labour force as per the LFS method yet are identified as employed by the time diary data are defined as invisible workforce.
Identification of Visible and Invisible Workforce by Gender.
The employment classification as per the two methods is given in Tables 1 and 2. A significant gender bias in the invisible workforce can be observed. Table 1, Panel A shows that of the total women employed, 46% are participating in the invisible workforce with remaining in visible workforce. A large majority of these women work part-time whereas when asked if they did any productive work, they negate it and hence, are classified as not in the labour force. Panel B indicates that while some men are also engaging in invisible work, they are a very small percentage, only 2.8% of the total employed men. In Table 2, I am comparing the labour force statistics using both classification methods, LFS and TUS. The employment rate of women increases from 18 to 30% when we account for the invisible workforce whereas for men the change is trivial, 73 to 75%. It is important to note that for women, the change is primarily in the part-time workers whose percentage increased from 10 to 22 across the two classification methods.
Classification of Employment Status Using Labour Force Survey and Time Use Survey by Gender.
(a) Employment status classified using the labour market questions, in the time use survey, which are identical to labour force survey.
(b) Employment status classified using time diary responses.
Characteristics of Invisible and Visible Workforce
This study examines three employment categories: not in the labour force (NLF), invisible workforce and visible workforce. Descriptive statistics provided in Tables 3 and 4 detail individual and household characteristics. The majority of households, averaging 7 members including 2 children and predominantly headed by males, indicate a young demographic with over 70% having children under the age of seven, suggesting a high demand for care work.
Individual Characteristics of the Sample, 10–74 Years, by Employment Category and Gender.
Household Characteristics of the Sample, 10–74 Years, by Employment Status and Gender.
The invisible workforce primarily comprises married, working-age women with no formal education and negligible personal income. These women share similar demographics with those in the visible workforce, except that the latter have their own sources of income. Despite most of the invisible workforce reporting no personal income, about 50% of women in the visible workforce also fall into this category. Of those who are employed, 79% earn a monthly income in the lowest quartile.
Conversely, the profiles of men in the invisible and visible workforces are strikingly different. The invisible workforce primarily consists of young, single teenage boys with primary or secondary education from female-headed households. In contrast, the visible workforce includes married, working-age men with varied educational backgrounds. For women, however, there are no significant differences in individual characteristics between those in the visible and invisible workforces, unlike for men.
Table 2 illustrates that women in the invisible workforce tend to reside in multi-generational households with higher average household income and greater wealth indicators compared to those in the visible workforce. Conversely, men in the invisible workforce typically live in households with a larger average number of members, including infants and elderly dependents, suggesting increased financial obligations and potentially necessitating young boys to engage in marginal work to contribute to family support. This underscores the socio-economic complexity within different workforce segments and their familial contexts.
Understanding the Activities Performed by Invisible Workforce
This section delves into an insightful analysis of the unconscious factors influencing the perception of ‘work’ in both visible and invisible labour sectors. As evidenced in Table 1, a significant proportion of the invisible workforce is categorized as part-time workers, suggesting that their work is often sporadic and of limited duration. To further understand the nature of activities undertaken by both invisible and visible workforce, Table 5 provides data on participation rates and average time spent in various types of activities.
Types of Labour Market Activity Performed by Invisible and Visible Workforce by Gender (Participation Rate and Conditional Mean Time Spent).
(a) Wage and salary employment.
(b) Outworkers, contract worker, home-based workers and unpaid workers.
(c) Employer/self-employed.
(d) Crop farming, market gardening and kitchen gardening.
(e) Tending animals, fish farming, forestry and gathering of wild products.
(f) Food processing and beverage preparation
(g) Crafts and textiles
(h) Petty trade and other activities in mobile locations.
(i) Building of dwelling, fitting, installing and repairing tools and machinery and other services.
Among women in the invisible workforce, the primary participation is observed in textiles (62%) and animal husbandry (32%), while men predominantly engage in animal husbandry (60%) and farming (30%). On an average day, women dedicate approximately 137 minutes to textiles and 114 minutes to animal husbandry, translating to roughly 1–2 hours spent on these activities daily. Notably, the time allocation of women in the invisible workforce does not significantly differ from that of their counterparts in the visible workforce. This underscores the parity in time commitment across different sectors, despite the invisibility of certain types of labour.
The heatmaps depicted in Figure 3 reveal distinct patterns of workplace participation between women and men in both the invisible and visible workforce. Women in the invisible workforce predominantly engage in their labour activities from their own dwellings, while men typically work in fields. Conversely, in the visible workforce, women are observed to work either in fields or from their own dwellings. In contrast, men in the visible workforce exhibit a broader spectrum of workplace engagement, with their activities mainly conducted in fields or other external workplaces.
Work-Location Intensity Heatmaps of Invisible and Visible Workforce by Gender.
Given these observations, it is reasonable to argue that non-standard workplaces, such as working from home or in fields, could significantly influence the accurate reporting of economic activities. The disparity in workplace settings between genders within the invisible and visible workforce underscores the potential for misreporting or overlooking certain types of labour, particularly those performed in less conventional or informal settings.
I devised a variable to capture LMW conducted either concurrently or consecutively with other activities, including URW, self-maintenance, social/cultural activities, engagement with mass media or learning. Simultaneous LMW is identified if LMW overlaps with another activity within the same 30-minute time-diary slot, while, sequential LMW occurs when LMW and another activity are carried out in two consecutive time slots at the same location. I further disaggregated simultaneous and sequential LMW into two distinct variables: the first encompasses LMW performed alongside URW, while the second encompasses LMW performed alongside any other activity, such as community work, social/cultural activities, or leisure pursuits. The reason for this disaggregation is that most URW is typically carried out within the household by women. When there is more overlap between LMW and URW compared to other activities, it suggests that LMW is primarily associated with women’s responsibilities within the household. Conversely, for all other activities (which could encompass a wide range of tasks), both men and women are expected to participate to a similar extent.
Table 6 presents data on participation in simultaneous and sequential LMW. In Panel A, participation rates and the percentage of LMW performed simultaneously or sequentially with URW are outlined. Notably, there’s no discernible difference in participation rates for simultaneous LMW with URW between men and women, but significant disparities emerge in sequential work, observed in both genders within the invisible workforce. Particularly, participation rates in sequential work are notably higher for the invisible workforce, suggesting that individuals in this category often undertake LMW and URW sequentially at the same location. Men in the invisible workforce exhibit three times higher participation rates in both sequential and simultaneous work compared to their counterparts in the visible workforce. For the invisible workforce, approximately 31% of LMW performed by women and 20% by men is carried out sequentially or simultaneously with URW.
Table 6, Panel B reports proportion of LMW performed simultaneously or sequentially with other social and leisure activities excluding URW, higher participation rates are observed for both men and women. Women in invisible work engage in LMW concurrently with other activities as high as 42%, while men engage at a rate of 28%. When aggregating all simultaneous and sequential LMW performed with any activity (URW and others), these numbers increase to 57% for women and 29% for men in the invisible labour force, significantly surpassing figures for the visible workforce.
Labour Market Work Performed Simultaneously and Sequentially by Employment Status and Gender.
(a) Community work, educational activities, social/cultural activities and leisure.
*p < .1, **p < .05, ***p < .01 ; t-test of mean differences of same gender across employment classification, that is, mean difference of women in invisible and visible workforce.
In summary, the statistics presented in this section substantiate the theoretical argument that economic activities may go unnoticed due to the high incidence of overlapping LMW with URW and other activities. This overlapping nature may lead respondents to perceive LMW as merely an extension of URW or other non-economic activities, thereby potentially distorting their comprehension of the term ‘work’. Consequently, when directly queried, respondents may fail to acknowledge these activities as constituting part of their work, contributing to their invisibility in labour market discourse.
Econometric Methods
In this section, I delve into the factors influencing labour supply within both the invisible and visible workforce, specifically focusing on men and women. Beyond inadvertent misreporting, we anticipate that life cycle stage and human capital accumulation will emerge as significant predictors of labour-supply decisions for individuals across genders.
The life cycle stage encompasses various life transitions, such as entering the workforce, starting a family, or nearing retirement, which often exert considerable influence on individuals’ labour market participation. Similarly, human capital accumulation, including education level and vocational training, serves as a pivotal determinant of labour supply, shaping individuals’ skills, qualifications and job prospects. By examining these factors, we aim to unravel the underlying determinants of labour supply dynamics within both visible and invisible sectors, shedding light on the nuanced interplay between socio-demographic characteristics and economic decision-making.
Multinomial Logit Model
Since the dependent variable is categorical, I will use the multinomial logit model (Kwak & Clayton-Matthews, 2002), which is also established to be more accurate than the multinomial probit model (Kropko, 2007). The standard errors are clustered at the community level. Thus, the probability of individual i of household h to participate in choice (j)—not in labour force (1), invisible workforce (2) and visible workforce (3) is given by the following:
where
Multinomial logit assumes the error to be uncorrelated across equations which implies that choice probabilities must satisfy an independence of irrelevant alternatives (IIA) property.
Determinants
In this study, I focus on two key variables: human capital accumulation and financial well-being. Human capital accumulation is gauged by two factors: education level and vocational training. Education level is categorized into four tiers: no formal education, primary education (five years of schooling), secondary education (five to eight years of schooling) and tertiary education (college education or higher). I use no formal education as the reference category for the education level variable. Additionally, I employ a binary variable for vocational training, which takes the value of 1 if the respondent has received any formal training.
Research consistently demonstrates that individuals with lower levels of education are more likely to engage in marginal sectors and other low-skilled work opportunities (Amaral & Quintin, 2006). Therefore, education level serves as a crucial indicator of human capital accumulation, reflecting individuals’ skill sets and qualifications. Furthermore, vocational training provides insights into whether individuals have acquired specialized skills beyond their formal education, which can enhance their employability and earning potential. By considering both education level and vocational training, we aim to capture a comprehensive understanding of human capital accumulation and its implications for labour market participation and outcomes.
Given the patriarchal norms that often discourage women from participating in LMW, financial necessity may emerge as a powerful motivator pushing women into the labour market. To gauge the financial position of households and its impact on labour market participation, several household variables are examined. First, the receipt of remittances is considered, with a binary variable coded as 1 if the household has received remittances. Remittances can serve as a significant source of income for households, potentially alleviating financial constraints and influencing labour market decisions. Second, the durable assets quartile which reflects the accumulation of durable assets within the household, such as TVs, radios, bicycles, cars and other possessions. This variable is indicative of the household’s overall economic well-being and can provide insights into their financial stability. Lastly, the household construction material is examined, categorized based on the type of material used in construction (e.g., mud, a mixture of mud and bricks or bricks only). This variable offers insights into the socioeconomic status of the household, with households residing in more durable and modern structures likely to have greater financial stability and resources.
In addition to the aforementioned variables, I consider a range of individual and household-level characteristics that may influence labour supply decisions. These include age, marital status, student status, household composition, seasonal work months, local development indicators and regional variables. Age is categorized into distinct life-cycle stages: teenagers (10–17 years), adolescents (18–24 years), adults (25–64 years) and older individuals (65–74 years). It is anticipated that working-age adults (25–64 years) will exhibit higher participation in the visible workforce, as they are typically in their prime working years. In contrast, younger (10–17 years) and older (65–74 years) individuals may engage in irregular work, as financial responsibilities may not be their primary concern.
Marital status serves as an indicator of gender roles within the household. Married individuals, particularly men, may be more likely to participate in the labour market, as they have financial obligations to fulfil for their families. Additionally, married women may face increased pressure to engage in URW compared to their single counterparts. Student status is considered, with individuals identified as students expected to have a lower probability of participating in the labour market, whether visible or invisible, as they are primarily focused on their education.
In examining the determinants of labour supply, household composition emerges as a pivotal factor. The presence of young children typically signifies a higher demand for URW within the household, as caregiving responsibilities often fall disproportionately on women. Conversely, the presence of young girls may indicate a more equitable sharing of URW duties within the household. Furthermore, in rural areas, the labour market experiences seasonal fluctuations, particularly during specific crop sowing or harvesting months. To account for this, I include dummy variables for the seasonal months of April–May, July–August and October–November (Ilahi & Grimard, 2000). These variables capture the surge in labour market opportunities within the agricultural sector during these months.
Additionally, I control for local development indicators, such as access to markets, primary and secondary schools, clinics and train stations. These indicators reflect the infrastructure and services available within the locality, which can influence labour market participation and opportunities. Regional variables, including urban/rural designation and province, are also considered to capture broader geographical influences on labour supply dynamics. Urban areas may offer different employment opportunities and resources compared to rural areas, while provincial variations may reflect differences in economic structure, policy environments and social norms.
By controlling for these diverse factors, I aim to construct a comprehensive model that elucidates the multifaceted determinants of labour supply within both the visible and invisible workforce, considering individual characteristics, household dynamics and contextual factors.
Results and Discussion
To ascertain the factors influencing the likelihood of participation in either the visible or invisible workforce, a multinomial logit model is estimated for the full sample of men and women aged 10–74 years. The baseline category for the regression is individuals not in the labour force. Table 7 (Columns 1–4) presents the marginal effects of the multinomial logit estimates, with standard errors clustered at the community level. First, the results for women are examined (Columns 1 and 2). Human capital accumulation emerges as a significant determinant of transitioning from the invisible to the visible workforce for women. Specifically, any level of formal education (primary, secondary and tertiary), compared to the reference group (no formal education), is associated with a lower probability of participation in the invisible workforce. Conversely, for the visible workforce, only tertiary education (college graduates or above) exhibits a positive probability compared to no formal education. This suggests that a significant portion of women participating in the labour market are uneducated.
Marginal Effect of Multinomial Logit Estimates for Full Sample (Men and Women) and Married Couples Age 10–74 Years.
(a) Reference group: No formal education.
(b) Self-excluding community average of women’s travel time for the purpose of labour market work or education.
(c) Self-excluding community average of women’s travel time for the purpose of community work, social and cultural activities and leisure.
(d) Reference group: Mud house.
(e) Reference group: Lowest quartile.
(f) Reference group: Nuclear household.
(g) Other control variables include age, student, seasonal work months, local development indicators (access to market, clinic, primary and secondary school) and regional variables (urban, provinces).
Community-level travel norms among females exert a significant and positive influence on participation in the invisible workforce. However, for the visible workforce, only the average travel time for LMW and education displays a positive association, while the travel average for social/cultural and leisure activities shows no statistically significant effect. These findings align with existing literature, which suggests that safe travel is a major obstacle to females’ education (Borker, 2021) and their participation in the labour force (Martinez et al., 2018). Interestingly, the mode of transport shows no significant association with participation in LMW, whether in the invisible or visible workforce.
The significance of financial well-being indicators underscores the role of economic constraints as drivers of female labour force participation, whether in the invisible or visible workforce. Three key indicators are considered: household construction material, durable assets index and remittances received by the household. Interestingly, the highest predicted participation of women in both the invisible and visible workforce is associated with households residing in mud houses, belonging to the lowest quartile of durable assets and not receiving any remittances. This suggests that economic necessity plays a pivotal role in motivating these women to enter the labour market.
Household composition emerges as a significant predictor of labour market participation. The presence of children, indicating a greater need for caregiving responsibilities, reduces the probability of participating in LMW, whether in the invisible or visible workforce. Conversely, the presence of young girls (8–17 years) or adult women (18–64 years), who typically share the burden of unpaid regular work (URW), increases the probability of LMW participation. However, the presence of adult men (18–64 years), who traditionally bear the financial responsibility within households, reduces the probability of women’s participation in the visible workforce. This finding strengthens the argument for added-worker effects, suggesting that women may only enter the labour force to supplement household income (Bredtmann et al., 2018).
The marginal effects of the multinomial logit model for men are reported in Columns (3–4) of Table 7. Similar to women, human capital accumulation plays a crucial role in transitioning men from the invisible to the visible workforce. Individuals with primary education exhibit a higher probability of participating in LMW, whether in the invisible or visible workforce, whereas secondary education emerges as a significant predictor specifically for the visible workforce. This suggests that, akin to women, men are driven into the workforce due to economic necessity. Moreover, the highest probability of working is observed among men residing in mud houses, owning the lowest quartile of durable assets and receiving no remittances. This further underscores the influence of economic need on men’s labour force participation.
Interestingly, residing in a multi-generational household lowers the probability of participation in the visible workforce compared to a nuclear household. This could be attributed to the co-residence with other adult males who share the financial responsibility, a scenario less likely in a nuclear household. Furthermore, the presence of children aged 0–3 years reduces the probability of participating in invisible work but increases it in the visible workforce. This implies that young children signify increased financial needs, prompting men to transition into visible LMW. Additionally, a greater number of adult men (18–64 years) in the household implies that one is not the sole earner, leading to more participation in invisible work. However, with younger boys, there is greater responsibility, resulting in a higher probability of participating in the visible workforce.
In conclusion, the analysis highlights several key factors influencing labour force participation among both women and men in the visible and invisible workforce. Human capital accumulation, as indicated by education level, emerges as a significant determinant, with higher levels of education associated with a transition from invisible to visible work. Community-level travel norms, financial well-being indicators and household composition also play crucial roles. Notably, economic necessity appears to be a driving force for labour force participation, especially for women, with households facing greater financial constraints more likely to engage in both visible and invisible work. Furthermore, household dynamics, such as the presence of children and adult men, significantly impact labour force decisions, underscoring the complex interplay between economic, social and familial factors. Overall, these findings provide valuable insights into the multifaceted nature of labour force participation and highlight the importance of addressing barriers to entry for both women and men across different sectors of the workforce
Conclusion
Accurately estimating labour market participation is crucial for informed policy decisions, yet there remains a lack of rigorous empirical evidence on the determinants of women’s absorption into low-quality, home-based jobs. In Pakistan, where female labour force participation is already low, this paper sheds light on a concerning trend: 45% of employed women remain undocumented. This not only leads to underestimated gross national product and female labour force statistics but also hinders effective policymaking. Drawing upon the Pakistan Time Use Data from 2007, this study identifies the ‘invisible’ workforce and delves into the reasons behind the concealment of these economic activities. Building upon existing literature from outside South Asia, the analysis explores unconscious factors such as the nature of the work, simultaneous and sequential labour market engagements, and the work location. Additionally, it investigates the determinants of labour supply choices in both the invisible and visible labour market sectors for both men and women. By uncovering these hidden dynamics, the study aims to provide valuable insights for policymakers seeking to promote inclusive economic growth and enhance labour market opportunities, particularly for women, in Pakistan and beyond.
The analysis unveils a stark gender bias in the participation of the invisible workforce, with women comprising a staggering 88% of its total. Accounting for the invisible workforce yields a significant increase of 12 percentage points in the employment rate for women, particularly among part-time workers. Notably, women predominantly engage in activities such as animal husbandry and textile work within their own dwellings, often concurrently with URW or other social and cultural activities (57%). As a result, the unconscious factors tied to the nature and location of their work make it challenging for respondents to recognize these activities as ‘work’, thereby posing substantial hurdles for their accurate inclusion in LFSs.
Additionally, the study reveals that women’s inclusion in the invisible workforce, as opposed to the visible workforce, is largely influenced by factors such as lack of human capital and mobility constraints. Economic necessity emerges as a significant driving force behind women’s labour force participation. In contrast, men demonstrate insignificant mobility constraints and may transition from the invisible to the visible workforce through greater accumulation of human capital and financial need.
These findings underscore the importance of accurately recognizing and measuring women’s work in the labour force, taking into account the unconscious factors that shape their participation in the invisible workforce. Addressing these issues is paramount for the development of policies that promote gender equality and facilitate women’s full and meaningful participation in the labour market.
Footnotes
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author received no financial support for the research, authorship and/or publication of this article.
