Time poverty, work status and gender: the case of Pakistan.
Najam-us-Saqib ; Arif, G.M.
The present study measures time poverty and its incidence across
gender, occupational groups, industries, regions, and income levels
using Time Use Survey (TUS) 2007, the first nationwide time use survey
for Pakistan. In the entire TUS sample, the incidence of time poverty is
14 percent. Women are found to be more time poor than men whether
employed or not. This is because of certain women-specific activities
that they have to perform irrespective of their employment status.
Working women are far more time poor than those not working.. Women
accepting a job have to make a major trade-off between time poverty and
monetary poverty. People working in professions and industries that
generally require extended work hours and offer low wage rates are more
time poor. This entails a situation of double jeopardy for workers who
tend to be money and time poor at the same time. The close association
of time poverty with low income found in this study corroborates this
conclusion. Government can help reduce time poverty by enforcing minimum
wage laws and mandatory ceiling on work hours in industries with high
concentration of time poverty. Eradication of monetary poverty can also
eliminate the need to work long hours at low wages just to survive. A
fair distribution of responsibilities between men and women.is also
needed.
Keywords: Time Poverty, Gender Disparities, Time Use, SNA
Activities, Time Use Survey, Pakistan
1. INTRODUCTION
Time is a valuable economic resource. It may be spent in a variety
of ways, but employed persons spend a significant portion of it in the
labour market for monetary gains. They still have other demands on their
time resource such as self-care, home production of goods and services
and leisure. These demands on time may reach a point where people may be
categorised as time poor. In many developing countries including
Pakistan, working women may be more time poor than men because of their
household responsibilities. Time poverty may also be related to certain
occupations and industries where workers have to work longer hours.
The concept of time poverty is not new to economics literature,
though the revival of interest in this phenomenon and efforts to measure
it empirically are relatively a recent development. Part of the reason
for this renewed interest appears to be the availability of time use
data for a number of countries. The publication of the report on Time
Use Survey 2007 (TUS) has added Pakistan to the list of such countries
[Pakistan (2009)]. Naturally, the availability of this data has
rekindled interest in time use research in Pakistan. The compilation of
this dataset has for the first time opened unlimited vistas for research
on time use in Pakistan.
The present study focuses on analysis of the various aspects of
time poverty among the employed though, for comparison, it has included
the not-working sample as well. The study begins by exploring the
analytical framework used to study time poverty in the next section. In
Section 3, it describes the dataset and discusses its descriptive
statistics. This section also delineates the methodology used in this
study and deals with the question of how to empirically estimate time
poverty. Section 4 presents the results of the present study. The final
section summarises the main findings of this study and in conclusion
presents some policy recommendations.
2. ANALYTICAL FRAMEWORK
Defining time poverty is not a straightforward exercise. It is a
complex matter that involves a number of theoretical and empirical
considerations. Once these issues are clarified, we can move on to the
main focus of our study. The incidence of time poverty among the
employed itself has multiple dimensions that need to be investigated.
Though Vickery's (1977) seminal paper on time poverty is regarded
as a major step towards analytically expounding the concept, the
antecedents of his work can be found in the classical paper by Becker
(1965) who developed a framework that treated time as a household
resource that is used as an input in the production of household goods
and services. (1) However, it may be recalled that time was recognised
by economists as a constrained resource long before Becker's work.
To understand the concept of time poverty, it would be instructive
to begin by looking at the resources that can be used to enhance the
welfare of a household or an individual. As shown in Figure 1, these
resources can be divided into three broad categories, namely, physical
capital, human capital and time.
[FIGURE 1 OMITTED]
The role of physical capital is well known. It generates a stream
of revenue over its lifetime that adds to household income. Becker
(1975) and Mincer (1974) have highlighted parallels between physical and
human capital. According to their theory, investment in human capital
also generates a stream of income over the lifetime of the individual.
Therefore its role in enhancing the welfare of an individual has marked
similarities with that of physical capital and can be easily understood
by drawing parallels between the two types of capital.
As noted earlier, time is also an important household resource that
can be put to a variety of uses. Since Becker's pathbreaking work,
the role of time as an input in household production has been
well-recognised. The literature on household production postulates that
households combine market goods, time, personal and household
characteristics along with other inputs to produce household goods and
services. (2) Oates (1977) and Hamilton (1983) have extended this
approach by showing that community characteristics must also be included
as inputs in the household production function. This implies that if
there is complementarity between time and other inputs, i.e. if time can
be used more efficiently in the presence of the above mentioned inputs
in the household production function, then time poverty will also depend
on these variables.
Time can be used in self-care and leisure as well. Self-care and
leisure may be regarded as utility enhancing consumption activities, but
their role in improving human capital cannot be ignored. Spending time
in rest, leisure and taking care of oneself makes one more productive.
Equally, time spent in productive activities can be used to make leisure
more productive because it generates income that can be spent on goods
that are complementary to leisure, such as books and television.
In addition, time can be used in the market place to directly
generate income. The income thus generated has a direct role with
respect to monetary poverty. More interesting for us is the fact that
employment increases the time used in committed activities which has
strong bearing on time poverty. This raises the spectre of the trade-off
between monetary poverty and time poverty. One more layer of complexity
is added when we recognise the direct substitutability between time and
money. This is evident from the simple fact that time can be bought by
hiring the services of other persons or by purchasing time saving
devices.
The gender dimension of this issue is important as well. In
developing countries, for example, tradition assigns certain activities
such as cooking and childcare solely or primarily to women, so that they
have to perform these activities even if demand on their time increases
as they enter the labour market. If we keep this possibility in mind,
the answer to the question whether getting a job makes women better off
no longer remains a clear cut yes because now the trade-off between time
and monetary poverty as well as personal and social preferences comes
into play.
Economists have long recognised poverty as a multifaceted
phenomenon, though income based measures of poverty are more commonly
known. The United Nations Development Programme and Oxford Poverty
Development Initiative have recently formalised the concept of
multidimensional poverty into a new poverty index called
Multidimensional Poverty Index (MPI). (3) This index takes into account
ten measures of deprivation related to health, education and living
standards but ignores time poverty, which is an important dimension of
overall poverty. However, overlooking time poverty may lead to an
incomplete measurement of overall poverty as it may result in a number
of highly deprived people being classified as non-poor. It may also
hamper a true understanding of the extent of deprivation of those who
are both time poor and income poor.
The above discussion can be summarised into the following points:
* Time poverty is an important aspect of overall poverty because
monetary poverty line provides only a partial measure of poverty.
* It is theoretically possible that some persons could be
monetarily rich but time poor and vice versa.
* There are theoretical grounds to believe that both the household
and community variables are important determinants of time poverty.
* The gender dimension of time poverty is important, especially for
developing countries.
3. DATA AND METHODOLOGY
3.1. The Dataset
This study is based on the Time Use Survey (TUS) 2007 sponsored by
the Strengthening PRS Monitoring Project of the Ministry of Finance and
conducted by the Federal Bureau of Statistics, Government of Pakistan
[Pakistan (2009)]. This is the first nationwide time use survey for
Pakistan. The survey was conducted from January to December 2007 and
covered a cross-section of 19,600 households. It represents both
national and provincial levels with rural/urban breakdown. The
year-round coverage of the survey was designed to capture seasonal
variation in the time use pattern.
While the survey provides useful information about the household
and the community, the prized section of the survey is the diary that
records all the activities of two selected persons from each household
who are ten years of age or older. (4) The activities are recorded over
a period of 24 hours. The entire day is divided into 48 half-an-hour
slots and each person is asked about the activities he/she was engaged
in during each half hour. An elaborate coding scheme is used to classify
the activities reported by the respondents. It is the first time that
such a detailed account of time used in daily activities has been made
available for Pakistan. Some important details of how this data was used
in this study and some of its salient characteristics are described
below.
The individuals aged 10 years and above, who filled the diary to
report their activities during the past 24 hours, form the unit of
analysis for this study. These individuals are grouped into two broad
categories, working or employed and not-working or not employed. The
subsample of 'employed' persons consists of those who have
worked for income or profit at least for one hour during the week
preceding the survey. This definition is consistent with that used by
the Pakistan Bureau of Statistics (PBS). The 'not-working' or
'not employed' subsample is the residual category consisting
of both the unemployed and those who are out of the labour force. This
type of categorisation has recently been used by Kalenkoski, Hamrick and
Andrews (2011) to determine the time poverty thresholds based on pooled
data from 2003-2006 American Time Use Surveys (ATUS).
3.2. Sample Characteristics
Since the major objective of this study is the analysis of time
poverty of the employed sample by gender and other characteristics
related to labour market, it would be instructive to have a brief
description of these characteristics. Table 1 shows the
socio-demographic characteristics of the total sample as well as for the
working and not-working sub-samples separately, while the labour market
specific indicators of the employed sample are reported in Table 2,
where the relevant figures from Pakistan Labour Force Survey [Pakistan
(2010)] have been provided for comparison. Information on monthly income
and sources of income is given in Table 3.
Fifty two percent of the total respondents who filled the diary are
females. The mean age for the total sample is 31 years and the male
sample is on average one year older than the female sample. About 40
percent of the sample is drawn from urban areas and more than half were
married at the time of the survey. There is a gender difference in terms
of the proportion living in urban areas and the marital status, but it
is relatively small (Table 1).
Figure 2 shows the percentage of males and females working at the
time of the survey. Whereas more than two-thirds of the males were found
working at the time of the survey, the corresponding figure for the
females was only 17 percent. Consequently, while three-quarters of the
not-employed sample consists of females, their proportion among the
employed sample is only one-fifth (Table 1).
Another noteworthy gender difference among the not-working persons
is in their marital status. Table 1 shows that approximately 60 percent
of the not-working females are in the 'currently married'
category as compared to only 17 percent for the males. This gap is much
narrower and in opposite direction among the employed persons, the two
figures being 67 percent and 70 percent for women and men respectively.
The overwhelming majority of the employed females (about 75 percent)
live in rural areas, while this figure for the not-working women is
about 60 percent.
[FIGURE 2 OMITTED]
One of the reasons for the higher percentage of working women
living in rural areas appears to be their substantially higher
representation among agricultural workers (48 percent as compared to 29
percent men; Table 2). Within the employed sample, the majority of
females fall in three occupation groups--agriculture workers (48
percent), craft workers (19 percent), and unskilled (elementary) workers
(18 percent). Only 15 percent women are professional or associate
professional workers. Employed males are engaged in four major
occupational categories: agriculture (29 percent) professionals and
associate professionals (24 percent), elementary work (21 percent) and,
craft and machine work (18 percent). In terms of industrial
classification, women are concentrated in agriculture, manufacturing
and, community and social service sectors. In addition to these three
sectors, the employed males have a substantial representation in the
trade sector as well (Table 2).
Table 2 also shows that employment status of 46 percent of the
employed males is reported as 'employee', while the
corresponding figure for women is 39 percent. The most pronounced gender
difference in employment status is found in the 'unpaid family
helper' and 'self-employed' categories. Compared to just
10 percent of the males, around half (47 percent) of the females are
unpaid family workers. On the other hand, 39 percent of males are
self-employed as compared to only 14 percent of females.
Three labour market characteristics of the TUS employed sample are
also compared in Table 2 with the LFS employed sample. While most of the
TUS and LFS figures are fairly close to each other, there are three
noteworthy differences between these datasets. One, LFS shows a higher
representation of females among agriculture workers as compared to TUS
while in case of female craft workers and machine operators' TUS
figures are larger. Two, among industries, LFS reports a higher
percentage of women in agriculture as compared to TUS data, whereas TUS
figures are higher for women working in the manufacturing sector. Three,
the percentage of unpaid family helpers in the TUS is lower than that in
the LFS. An important procedural difference also exists between the two.
The TUS has used female enumerators to report the activities of female
respondents, while this task is performed by male enumerators in the
LFS. Therefore, it may be argued that the reporting of female activities
is more reliable in the TUS.
The gender difference in employment status reflects itself in the
monthly income statistics too (Table 3). More than 43 percent of the
employed women reported no monthly income, (5) whereas 45 percent of
them were earning a monthly income of Rs 4000 or less. This contrasts
sharply with the corresponding figures for the employed males. Among
them, the proportion without any monthly income was only 8 percent while
approximately 60 percent of them were earning more than Rs 4000 per
month. Wages and salaries, and business are the major sources of monthly
income for both the employed men and women.
The differences in the characteristics of working and not-working
women in terms of age and schooling are presented in Appendix Table 1.
It shows that the share of teenagers (10-14 years old) is greater (17.7
percent) among the not-working women sample as compared to the working
sample (7.4 percent). Approximately two-thirds of the working women are
in their prime age, that is, 15--39 years, while the corresponding share
for the not-working sample is 56 percent. The proportion of aged women
among the not-working sample is modestly higher (8.4 percent) than among
the working sample (4.9 percent). In terms of education, it is
interesting to note that the not-working women sample appears to be more
literate than the working women sample. However, the share of degree
holders is relatively greater among the working women.
In short, this description of the characteristics of the employed
and not-employed sample of the 2007 TUS by rural-urban classification
shows a great deal of variation in their demographic profile and
economic activities, which are likely to be closely associated with
their time use patterns and time poverty.
3.3. Methodology
This study proceeds in two steps. The first step consists of an
examination of the time use pattern of the respondents by the type of
activities as classified in the TUS 2007. The focus has been on
differentials in time use pattern by gender, region, work status, and
other labour market indicators. The TUS 2007 organises activities of the
respondents in three broad categories, namely, System of National
Accounts (SNA) activities, extended SNA activities, and non-SNA
activities. The SNA activities consist of employment for establishments,
primary production activities not for establishments, like crop farming,
animal husbandry, fishing, forestry, processing and storage, mining and
quarrying; secondary activities like construction, manufacturing, and
activities like trade, business and services. Extended SNA activities
include household maintenance, care for children, the sick and the
elderly and community services. The activities related to learning,
social and cultural activities, mass media and personal care and
self-maintenance constitute Non-SNA activities.
To proceed to the second stage of the study, which deals with
various aspects of time poverty as discussed in Section 1, it is
inevitable to operationalize the concept of time poverty. What we need
is a working definition of time poverty that makes it possible to
estimate a time poverty threshold using our dataset. The estimated
threshold can then be used to classify people either as time poor or
non-poor. This objective can be achieved by using a methodology that is
similar to that used for estimating monetary poverty.
The first thing that needs to be decided in this regard is whether
to use an absolute or a relative measure of poverty. Both measures are
common in the literature on monetary poverty, though the choice of an
absolute measure is a bit more arbitrary. Often a certain level of per
adult calorie intake equivalent based on "minimal" calorie
requirements is taken as the poverty threshold. Unfortunately, things
get more difficult in case of time poverty as there is no agreed level
of "minimal" time needed by a person to avoid being time poor.
Therefore we have to resort to a relative definition of time poverty
that involves using some measure of the central tendency (such as mean,
median or mode) of time distribution or its multiple as a time poverty
cut-off point. (6)
The issue of the choice of a poverty index comes next. We use the
headcount index, which gives the proportion of people who are time poor.
The results presented using this index are easy to grasp, even by a
non-professional. In addition to being simple and straightforward, it
belongs to the FGT class of poverty indices that possess certain
desirable properties. (7)
Which are the activities that make people time poor if they exceed
a predetermined limit is another question that has to be dealt with.
While it is easy to exclude activities such as, leisure and vacationing
from this list, much more thinking is needed to decide on the activities
that belong to it. The literature on time use describes these activities
in such terms as "necessary or committed activities" and time
spent in these activities as "non-free minutes" [Kalenkoski,
et al. (2007)]. The activities to which an individual has committed as
his economic or social responsibility may be regarded as necessary
activities and time spent in these activities may be counted as non-free
minutes contributing to his/her time poverty [Kalenkoski, et al.
(2011)]. Thus, the figures of non-free minutes (time spent on committed
activities) hence obtained can then be used for defining time poverty
threshold(s) and calculating time poverty.
As noted above, the time use survey data organises activities
performed by the respondents in three broad categories, namely, SNA,
extended SNA, and non-SNA activities. A careful scrutiny of the list of
the activities falling under each of the three broad categories, as
reported at the beginning of this section, reveals that the first two
categories consist of what we may safely call committed activities. For
instance, the major activities included in the SNA, such as employment,
production, trade and business activities, are considered
'committed' because these activities are directly related to
the livelihood and economic wellbeing of working persons and their
households, and they have committed to perform these activities in
exchange for monetary or other economic benefits. Similarly, some social
responsibilities which are essential for the welfare of household
members such as care for children, the sick and the elderly are also
categorised as committed activities, as they must be performed as a
social obligation. These activities are part of the extended SNA
activities. Therefore we add time spent by the respondent in SNA and
extended SNA activities to calculate the total time spent by her/him in
committed activities. Figure 3 shows the link between the concept and
the empirical definition of time poverty as discussed above.
[FIGURE 3 OMITTED]
A poverty line or threshold that is used to calculate the headcount
index is often defined as a multiple of the median of the non-free time
of an individual. In the absence of an agreed cut off point for time
poverty based on sound theoretical grounds, the only option left is to
follow a threshold level commonly used by previous empirical studies of
time poverty. Following Lawson (2007) and Bardasi and Wodon (2006), this
study uses 1.5 times the median time spent in SNA and extended SNA
activities as the time poverty line. Based on this methodology, the time
poverty line is computed as 10.5 hours (630 minutes). The time poor are
those who have spent more than 10.5 hours in a day on the committed
activities (SNA+ex-SNA). (8) In other words, persons who work 63 hours
in a week are deemed to be time poor.
However, it can be argued that the value of 10.5 hours used in this
study as the poverty line is an arbitrary cut-off point. A natural
question then would arise that what difference would it make if a higher
or a lower cut-off point was chosen as poverty line. In the absence of
well-established practices to measure time poverty, alternative poverty
lines have commonly been used in the literature [Bardasi and Woden
(2006)]. Following this practice, two alternative poverty lines have
also been used in the analysis; 9 hours as a lower cut-off point and 12
hours as a higher cut-off point.
4. RESULTS
4.1. Time Use
Table 4 sets out data on the time use patterns for the full sample
as well as working and not-working subsamples separately, controlling
for gender and rural-urban areas. The male sample that filled the diary
spent on average 5 and a half hour a day in SNA activities. In contrast,
the female sample spent 5 hours in ex-SNA and only 1 hour and 15 minutes
in SNA activities. Men spent about half an hour more in non-SNA
activities as compared to women.
Some more details emerge as we look at the time use statistics
separately for the working and the not-working sample. In the
not-working sample, both males and females spent an average of around
half an hour in SNA activities. The real gender difference is observed
in the remaining two categories. On ex-SNA activities, the not-working
male sample spent only half an hour as compared to more than 5 hours
spent by the not-working females. The not-working men spent about 5
hours more than women in non-SNA activities.
The employed males spent 7 and a half hours in SNA activities while
the corresponding time for the female sample was less than 5 hours. On
ex-SNA activities, the employed males spent an average of only 32
minutes in 24 hours whereas the female sample used up, on average, 4
hours and 39 minutes of their day on these activities. The gender gap in
the employed sample in the time spent in non-SNA activities was
substantially smaller as compared to that in the not-working sample.
A comparison of the time use pattern of the working, and
not-working samples reveals that employed males spend almost the same
small amount of time (32 minutes) in ex-SNA activities in both cases. In
contrast, despite having to work around 5 hours a day in the labour
market, the women's lot in terms of shouldering the responsibility
of exSNA activities is not changed substantially after accepting
employment. The time spent by them in ex-SNA activities is reduced, on
average, from 5 hours and 16 minutes to 4 hours and 39 minutes, a gain
of just 37 minutes. This lends credence to the view that some activities
in the developing countries are considered to be women specific which
they have to perform, whatever else they may or may not be doing.
The overall result is that women end up working more hours than men
whether they accept paid work or not. Not-working women spend about 5
more hours in SNA and ex-SNA activities combined as compared to
not-working men. This gender gap persists in the working sample, though
it is reduced to 1 hour and 16 minutes. Men also have more flee time
that they spend in non-SNA activities in both the cases though this
gender gap is much smaller in the working sample.
While the time used in SNA and ex-SNA activities by the males is
almost the same in both rural and urban areas, women living in rural
areas spend more time on both the types of activities as compared to
those living in urban areas. They also have less time available to them
for non-SNA activities as compared to their urban counterparts. This
rural-urban divide in the time spent by women in SNA and exSNA
activities combined on the one hand and non-SNA activities on the other
prevails both among working and not-working sample, though the gap is
much wider among the working women. A working woman living in the rural
area spends, on average, more than double the time in SNA and ex-SNA
activities as compared to a woman living in the urban area.
Tables 5-7 show the time use data by three labour market indicators
of the employed sample, namely the occupational class, industry and
employment status, and gender and rural-urban areas. Service workers and
plant/machine operators, who mostly work in the informal sector, (9)
spent on average 8 and a half hours in SNA activities, approximately 2
hours more than the time spent in SNA activities by professional and
clerical workers. The latter usually work in the formal sector where the
number of working hours is fixed, whereas those employed in the informal
sector usually work longer hours. This difference persists in rural as
well as urban areas. Male workers spent on average more time in SNA
activities than their female counterparts in all occupational
categories. Moreover, male professional and agricultural workers had
relatively more free time than the workers in other occupations.
In terms of industrial classification, workers engaged in trade,
transport and construction sectors spent more time in SNA activities
than those working in other sectors. This pattern of time use is not
influenced much by gender or region.
Overall, the female unpaid family helpers spent 3 hours more in a
day on committed activities than the male unpaid family helpers. The
situation of women working as employees or self-employed was not much
different. Unpaid family helpers spent less time on committed activities
than the other three categories of workers. However, a glance at the
gender distribution of time reveals that female unpaid family helpers
spent a lot more time in ex-SNA activities than their male counterparts
(more than 5 hours vs. only 23 minutes). This resulted in female unpaid
family workers spending more time in committed activities than any of
the remaining three groups of workers.
It is worth focusing on women who spent some time in SNA activities
(Table 8). On average these women spent more than 3 hours with virtually
no difference in rural and urban areas. However, there was significant
difference in this regard between the working and not-working women. In
urban areas, the former spent an average of 5 and a half hours in SNA
activities while the latter used only one hour and 41 minutes. The
working rural women spent on average 5 hours in SNA activities as
compared to 2 hours and 10 minutes used by the not-working sample.
Overall, working women give considerable time to their labour market
activities.
It appears from these simple statistics on the time use pattern
that in Pakistan (rural and urban areas alike) the participation of
women in the labour market does not reduce their time commitment for
ex-SNA activities. Males spend little time in ex-SNA activities, which,
in Pakistani culture, appear exclusively to be for females. Although
women spend much less time than men in SNA activities, their overall
time spent on committed activities (SNA+ ex-SNA) is greater than the
time spent by their male counterparts in these activities.
4.2. Time Poverty
The time use patterns of both the working and not-working samples
are reflected in the time poverty statistics. The last row of the first
panel of Table 9 indicates that, based on a 10.5 hours a day poverty
line, time poverty is 14 percent for the entire TUS sample. As expected,
the employed people (male as well as female) are more time poor than
those in the not-working category, mainly because the latter, in
general, did not spend time in SNA activities (see discussion in the
previous section). This difference is quite large in both urban and
rural areas. Time poverty is substantially higher among not-working as
well as working women as compared to men in the respective categories.
Working women are hugely more time poor as compared to the not-working
women (36.8 percent versus 10.2 percent respectively). This raises the
question whether getting a job is a bane or bliss for women. The answer
depends on the resulting trade off between monetary and time poverty and
its valuation by women. Moreover, if time poverty is computed from the
time used for the SNA activities only, the incidence of poverty among
women is negligible, less than 2 percent.
In urban areas, 12.3 percent people are time poor, while for the
rural areas this figure is 15 percent. Time poverty in rural areas is
higher among females than males. The opposite is true for urban areas.
Within the employed sample, 22.5 percent people are found time poor,
with no major difference between rural and urban areas. However, time
poverty among the employed female sample is double the time poverty
among the corresponding male sample. The difference in rural areas is
around two and a half times. In urban areas, although more females are
time poor than males, the difference is just 5 percentage points. As
noted earlier, it is due to the fact that female participation in the
labour market brings hardly any change in their time allocation for
activities related to household maintenance, care of children and the
elderly.
The second and third panels of Table 9 present results for two
alternative poverty lines; one with a lower cut-off point of 9 hours per
day and the other with a higher cut-off point of 12 hours per day. As
expected the two poverty lines lead, respectively, to higher and lower
estimates of time poverty, though the general pattern of time poverty
across various categories remains generally the same. The change in time
poverty due to change in cut-off point is substantial, for example,
increasing cut-off point to 12 hours per day brings down time poverty
levels to almost negligible in most of the categories, while a decrease
in the cut-off point to 9 hours per day increases considerably the time
poverty of both males and females.
It would be interesting to compare the results reported above with
the time poverty estimates for some other countries. Bardasi and Wodon
(2006) report an overall time poverty rate of 17.6 percent for Guinea,
whereas the corresponding figures for men and women are 9.5 percent and
24.2 percent respectively. The overall time poverty rate estimated by
Lawson (2007) for Lesotho is 7.9 percent, while 8.3 percent men and 6.8
percent women are reported to be time poor. So, time poverty in
Pakistan, based on a 10.5 hours per day cut-off point, is lower than in
Guinea but higher than in Lesotho.
In Pakistan, only a few studies have estimated the money-metric
poverty incidence across the occupational groups. The general conclusion
of these studies is that the level of poverty is higher among unskilled
(elementary workers), skilled and service workers than that among other
occupational categories. (10) The time poverty data presented in Table
10 show higher incidence of time poverty among services workers, machine
operators and workers in elementary occupations than among the clerical,
professional and agriculture workers. This implies that unskilled and
skilled workers along with the service workers are at the receiving end
of both monetary and time poverty.
In the male employed sample, time poverty is less than 10 percent
among the associate professionals, clerical workers and agriculture
workers, whereas one-third of the service workers and plant/machine
operators are time poor. The incidence of time poverty among females is
much higher than that among their male counterparts in all categories of
occupations except professional and service workers. A noteworthy point
is that approximately half of the employed women in elementary, skilled
and semi-skilled occupations are time poor. These differences in time
poverty across occupations persist in rural as well as urban areas. The
case of female agriculture workers is interesting. Table 10 shows that
41 percent of these women are time poor whereas only 9 percent of their
male counterparts fall in this category.
Table 10 also shows the data on time poverty across the type of
industry where the sampled workers were employed. High incidence of time
poverty was observed in trade, transport and manufacturing sectors for
both male and female workers. In the agriculture sector, time poverty
among women was four times higher than that among men. It corroborates
the time poverty data across the occupational categories discussed
above.
One important lesson from the analysis of the time poverty data
across the occupational and industrial classification is that low paid
occupations and sectors get more time of the workers. So these workers
are poor in money-metric terms as well as in terms of time use. They
work for longer hours and get low wages, insufficient to sustain a
decent living standard. Rural women working in the agriculture sector
are particularly in a disadvantageous position in terms of time poverty.
The finding that low paid occupations are associated with high
incidence of time poverty is further reinforced by the monthly income
data. Table 11 shows that, generally, the lower the monthly income the
higher the incidence of time poverty. For the employed sample, the
incidence of time poverty among those who earn a monthly income of Rs
10,000 or more was 16 percent as compared to 30 percent among those who
earn Rs 2000 or less per month, indicating a difference of 14 percentage
points between the highest and the lowest income group. This gap was
wider among women as compared to men, though much smaller between urban
women as compared to rural women. In most of the income groups, women
were found to be more time poor than their male counterparts in rural as
well as urban areas. This indicates a harder trade-off for women between
higher income due to joining labour market and increased time poverty as
compared to their male counterparts. The trade-off between supplying
additional work hours and time poverty is also harder for working women
as compared to working men, but less hard as compared to those women who
have to make a decision about joining the labour market.
The gender dimension of time poverty can be understood more clearly
from the employment status data than from any other labour market
indicators. Figure 4 shows a vast difference between males and females
in the incidence of poverty in all three categories of employment
status: "employees", "self-employed" and
"unpaid family helpers". The time poverty among the female
'unpaid family helpers' is around five-fold the time poverty
among their male counterparts. In the case of employees, the gender
difference in time poverty is around 10 percentage points, favouring the
male. This difference is even greater for the self-employed category.
Finally, education was found to reduce the incidence of time poverty,
particularly among college and university graduates. In addition, the
lowest gender gap in time poverty was found among these graduates
(Appendix Table 2).
[FIGURE 4 OMITTED]
4.3. Determinants of Time Poverty
The analysis carried out in the previous subsection primarily
focused on the incidence of time poverty by gender, the place of
residence and labour market indicators. Studies focussing on the
determinants of time poverty include several other individual, household
and community level variables that can be associated with time poverty.
(11) Due to data limitation, it is not possible to examine the
relationship between time poverty and all these variables. Focusing on
some socio-demographic and labour market characteristics of the sampled
persons who filled the diary, this section has carried out multivariate
analyses to examine the relationship between time poverty and some of
these characteristics. The dependent variable is time poverty which
takes the value 1 if the sampled person is time poor; otherwise it takes
the value 0. Since the dependent variable is binary, logistic regression
rather than OLS is used for the multivariate analysis. Six models have
been estimated. Model 1 is based on the entire sample (working and
not-working persons) while models 2 and 3 are estimated separately for
the male and female samples. Model 4 has included only the employed
sample to analyse the relationship between time poverty and labour
market indicators including occupation, industry, employment status and
income. Models 5 and 6 divide the employed sample between urban and
rural areas respectively to take care of the varying work patterns
between the two types of areas.
Four independent demographic variables, age, sex, marital status
and presence of children younger than 7 years in the household are
included in the regression analyses while the level of educational
attainment is used to study the relationship between time poverty and
human capital. The place of residence represents the influence of
community variables on time poverty. Four labour market indicators,
occupation, industry, income and employment status, are included in
models 4, 5 and 6 to understand their correlation with time poverty.
Three seasonal dummy variables have also been included in models 4 and
6, as working hours in rural areas are considerably affected by changing
seasons. The operational definition of all these variables and results
of the six models are presented in Table 12. (12)
Model 1 includes the entire TUS sample. The results of this model
corroborate the bivariate analysis carried out in the previous section.
All variables included in this model have an independent and significant
effect on the probability of being time poor. The employed persons are
more likely to be time poor than those not employed/not-working. It is
mainly because the not-working sample spends less time on the committed
activities, particularly those falling under the SNA activities
category. Moreover, the economically active women use their time in
household maintenance and child care in addition to SNA activities.
Estimation results of model 1 also show that overall, women are more
likely to be time poor than men. As discussed earlier, the underlying
cause behind this finding is their time use pattern. The quadratic
relationship between age and time poverty also turns out to be
significant. The significant and positive relationship between time
poverty and being married shows that marriage increases the use of time
on committed activities. Same is true for the presence of less than six
years old children in the household. Model 1 shows a positive and
significant relationship between time poverty and having no education or
having education but below the matriculate level. It means that 10 or
more years of education enable individuals to have more free time for
activities like personal care and rest.
The results of models 2 and 3, in which the analysis is carried out
separately for the male and female samples, show no major qualitative
change in the findings except that living in urban areas has a positive
relationship with male time poverty. In the case of the female model,
this relationship turns out to be negative. It shows that males living
in urban areas and females living in rural areas are more time poor than
their counterparts. It is largely because of the involvement of rural
women in farm activities. (13)
In order to learn about the relationship between time poverty and
labour market indicators, model 4 has been modified to include only the
employed sample. In this model, age, sex, marital status, education and
place of residence have signs similar to those in model 1. The Positive
and significant relationship between time poverty and working as
unskilled labourers, service workers and plant/machine operators in the
urban areas (model 5) shows the hard work of these manual workers. It
has been shown earlier that these workers, who are mainly males, spend
little time in ex-SNA activities and work long hours in the labour
market which makes them time poor. The number of such workers is perhaps
too small in rural areas to provide sufficient variation for meaningful
estimation of their effect. Although working women use relatively less
of their time in the labour market, they take all kinds of
responsibilities at home. This dual burden on the sampled women
contributes to their time poverty. They are left with relatively little
free time for personal care and rest. Unpaid family helpers are
generally rural females who, by definition, receive no income for their
work, so that a dummy for this category is likely to be highly collinear
with the income dummy. Dropping the income dummy from the regression for
rural areas (model 6) makes the dummy for unpaid family helpers highly
significant.
The industry, in which a worker is employed, is a strong correlate
of his/her time poverty. Workers engaged in trade, transport and
manufacturing sectors are more time poor than those engaged in other
sectors including agriculture, service and construction. The monthly
income also gives a similar message: the workers in low income groups
are more time poor than the workers in high income groups.
5. CONCLUSIONS AND POLICY IMPLICATIONS
Availability of time use data is relatively a recent phenomenon in
Pakistan. This has allowed us to measure time poverty and look at its
incidence across gender, occupational groups, industries, regions, and
income levels. The study also uses multivariate regression analysis to
examine the relationship between its various determinants. The results
of this study provide some important insights into the phenomenon of
time poverty in Pakistan and lead to some interesting conclusions.
The first important finding of this study is that women spend more
time in committed activities than men whether they are employed or not.
As a result, women are more time poor than men in both the
circumstances. A closer look at time use statistics indicates the reason
behind this occurrence. It appears that there are certain ex-SNA
activities, such as household maintenance, and care for children, the
sick and the elderly, that are women specific probably due to
socio-cultural reasons. Women have to perform these activities
irrespective of their employment status, while Pakistani men are not
usually involved in them. This substantially increases the time spent by
women in committed activities. Since men spend little time in ex-SNA
activities, they have more time available for non-SNA activities
including leisure and personal care as compared to women.
The finding that women generally spend more time in committed
activities and are more time poor as compared to men has two noteworthy
implications that are likely to influence school enrolment decision of
the females. According to the human capital theory, the decision to
enrol in school depends, among other things, on the opportunity cost of
education. The monetary value of the hours worked at home is one of the
components of this opportunity cost. Since women work more hours at home
as compared to men, their opportunity cost of getting enrolled in a
school is likely to be higher, making them less likely to enrol in
school. However, a cancelling factor is simultaneously at play. Women
are also more time poor as compared to men because they work more hours
at home. Hence, assuming that time poverty results in reduced labour
productivity and workers are paid in the labour market according to
their marginal productivity, women would cam less as compared to men for
working the same hours. Consequently, another component of opportunity
cost of education, which consists of the monetary value of the forgone
work in the labour market, would be smaller for women. This would make
them more likely to enroll in school. Thus, the net effect of
women's time spent in committed activities on female school
enrolment could either be positive or negative. However, this issue can
only be sorted out by further empirical research that entails generating
a single dataset that combines information that is available separately
in time use and labour force surveys.
The results of this study also indicate that working women are far
more time poor as compared to not-working women, because time spent by
them in ex-SNA activities does not reduce considerably enough to
compensate for the extra time they devote to their job. Moreover, women
face a harder trade-off between higher labour market earnings and
increased time poverty as compared to men. In other words, while
entering the job market, not only they have to face higher time-poverty
in exchange for reduced monetary poverty, but also the terms of exchange
are more unfavourable for them than for their male counterparts. This
raises the seemingly intriguing issue of whether expanding the job
market for women through economic and noneconomic measures would make
them better off? In the neoclassical framework, the choice of accepting
or rejecting the job offer and number of hours worked will depend on the
decision maker's marginal valuation of leisure as depicted by her
preferences and valuation of time in the labour market as indicated by
the prevailing wage rate. Assuming that women have the same preferences
as men, it can be argued on the basis of the findings mentioned above
that women have to make more difficult choices in their labour supply
decisions as compared to men because women have to spend considerable
time on certain ex-SNA activities even after joining the labour market.
Among the various categories of employment status, the case of
female unpaid family helpers is unique in several respects. Time poverty
among them is around five-folds the time poverty among their male
counterparts. They are more time poor even as compared to fellow women
in other employment status categories. The likely cause of the high
incidence of time poverty among the females in the agriculture sector is
the significant presence of unpaid family helpers.. The apparent reason
for the huge gender gap in time poverty among unpaid family helpers is
that female unpaid family helpers spent a lot more time on ex-SNA
activities than their male counterparts.
People in certain professions such as unskilled, skilled and
services sector are found to be more time poor as compared to people in
other professions. The same is true for some industries like trade,
manufacturing and transport. These professions and industries generally
require extended hours from the workers, while offering low wage rates.
This catches the workers in a situation in which they are both monetary
and time poor at the same time. The close association of time poverty
with low income found in this study corroborates our conclusion.
In the light of these findings, several policy areas emerge where
we need to focus. The first thing that needs to be done is to generate
awareness about a fair distribution of responsibilities between men and
women. If this can be done, a significant portion of the gender gap in
time poverty is likely to be eliminated.
The situation of female unpaid family helpers needs immediate
attention not only due to both the magnitude and the gender gap in time
poverty that they are facing, but also because they are more likely to
be monetarily poor. Generally, these are the women who work along with
other family workers in areas such as agriculture and household help and
maintenance. As the name suggests, they do not receive any payment for
their work. To fully understand their condition, a more thorough study
focusing on this particular group is needed.
Though participation in the labour market, particularly among
women, is not the only reason for time poverty, the findings of the
study show that working people are generally more time poor as compared
to the not working population and time poverty is concentrated in
certain occupations and industries. This opens up an opportunity for the
government to play its part in reducing time poverty. The line of action
is to enforce minimum wage laws to reduce monetary poverty of those who
are more likely to be time poor as well and to put mandatory ceiling on
work hours in the industries which have high concentration of time
poverty. Eradication of monetary poverty in general can also go a long
way in this respect by eliminating the need to work long hours at the
lowest wage rate just to survive. Improving education also has
significant potential in this regard, as high education is found to be
associated with low time poverty.
Appendix Table 1
Socio-demographic Characteristics of Women
Age Working Not-working
10-14 7.4 17.7
15-19 11.1 13.2
20-24 13.7 12.5
25-29 14.8 12.0
30-34 13.9 10.3
35-39 11.2 8.0
40-44 8.4 6.0
45-49 7.3 5.0
50-54 4.7 3.7
55-59 2.6 3.2
60+ 4.9 8.4
All 100 100
Highest Class Passed
No Formal Education 71.8 54.8
< Primary 5.0 9.3
Primary 5.7 14.2
Middle 2.8 7.9
Matriculation 5.2 7.3
Intermediate 3.5 3.9
Degree and Above 6.1 2.6
All 1001 0
Source: Calculated from the micro-data of Time Use Survey, 2007.
Appendix Table 2
Poor among the Employed Sample by Education and Gender
Education Both Sexes Male Female
No Formal Education 26.9 18.7 41.1
Below Primary 20.6 19.8 25.6
Primary 21.6 20.7 30.8
Middle 21.8 21.2 32.2
Matriculation 20.7 19.7 30.2
Intermediate 16.1 15.0 23.9
Degree and above 13.6 13.0 16.4
All 22.5 18.9 36.8
Source: Calculated from the micro-data of Time Use Survey, 2007.
Authors' Note: We are grateful to anonymous referees for their
valuable comments on an earlier draft of this paper. We are also
thankful to the Strengthening PRS Monitoring Project of the Ministry of
Finance, Government of Pakistan for providing access to the micro
dataset used in this study. An earlier version of this paper was
published as PIDE Working Paper 2012:81.
REFERENCES
Aikire, Sabina and Maria Emma Santos (2010) Acute Multidimensional
Poverty: A New Index for Developing Countries. (OPHI Working Paper No.
38).
Bardasi, Elena and Quentin Wodon (2006) Measuring Time Poverty and
Analysing Its Determinants: Concepts and Application to Guinea. In C.
Mark Blackden and Quentin Wodon (eds.) Gender, Time Use, and Poverty in
Sub-Saharan Africa. (World Bank Working Paper No. 73, 75-95.)
Becker, Gary S. (1965) The Allocation of Time. Economic Journal 75,
493-517.
Becker, Gary S. (1975) Human Capital Chicago.
Foster, J., E. J. Greer, and E. Thorbecke (1984) A Class of
Decomposable Poverty Indices. Econometrica 52, 761-766.
Gronau, Reuben (1999) Home Production--A Survey. In Handbook of
Labour Economics. Vol. 1, Chapter 4.
Hamermesh, D. S. and G. A. Pfann (eds.) (2005) The Economics of
Time Use. Contributions to Economic Analysis 271. Amsterdam, San Diego
and Oxford. Elsevier.
Hamilton, B. W. (1983) The Flypaper Effect and Other Anomalies.
Journal of Public Economics 22, 347-361.
Jafri, S. M. Younus (1999) Assessing Poverty in Pakistan. In A
Prifile of Poverty in Pakistan. Mahboob UI Haque Center for Human
Development. Islamabad.
Kalenkoski Charlene M., Karen S. Hamrick, and Margaret Andrews
(2011) Time Poverty Thresholds and Rates for the US Population. Social
Indicators Research 104, 129-155.
Kalenkoski, Charlene M., Karen S. Hamrick, and Margaret Andrews
(2007) Time Poverty Thresholds. Presentation at the 29th Annual
International Association for Time Use Research Conference, Washington,
DC.
Lawson, David (2007) A Gendered Analysis of 'Time
Poverty'--The Importance of Infrastructure. Global Poverty Research
Group. (Working Paper GPRG-WPS-078).
McGinnity, Frances and Helen Russell (2007) Work Rich, Time Poor?
Time-Use of Women and Men in Ireland. The Economic and Social Review,
38: 3, 323-354.
Merz, Joachim and Tim Rathjen (2009) Time and Income Poverty: An
Interdependent Multidimensional Poverty Approach with German Time Use
Diary Data. (IZA Discussion Paper Series No. 4337.)
Mincer, Jacob (1974) Schooling, Experience, and Earnings. Chicago.
Oates, W. E. (1977) On the Use of Local Zoning Ordinances to
Regulate Population Flows and the Quality of Local Public Services. In
O. Ashenfelter and W. E. Oates (eds.) Essays in Labour Market Analysis.
Wiley. New York.
Pakistan, Government of (2010) Pakistan Labour Force Survey
2008-09. Islamabad: Statistics Division, Federal Bureau of Statistics.
Pakistan, Government of (2009) Time Use Survey 2007. Islamabad:
Statistics Division, Federal Bureau Of Statistics.
Qureshi, Sarfraz. K. and G. M. Arif (2001) Profile of Poverty in
Pakistan, 1998-99. (MIMAP Technical Paper Series No. 5).
UNDP (2010) Human Development Report 2010. 20th Anniversary
Edition. United Nations Development Programme. New York. USA.
Vickery, C. (1977) The Time Poor: A New Look at Poverty. The
Journal of Human Resources 12:1, 27-48.
(1) For a more detailed analysis of the economics of time use, see
Hamermesh and Pfann (2005).
(2) For an excellent review of literature on home production, see
Gronau (1999).
(3) For more detail, see UNDP (2010) and Alkire and Santos (2010).
(4) For details of the procedure used to select two individuals
from each household, see, Pakistan (2009).
(5) These women are primarily unpaid family helpers.
(6) This definition is relative with respect to different time
distributions and must not be confused with the measures of relative
poverty that take into account the well being of other people in the
neighbourhood.
(7) For more detail on FGT indices of poverty, see Foster, Greet
and Thorbecke (1984).
(8) Using same methodology, Bardasi and Wodon (2006) have reported
a time poverty line of 70.5 hours per week for Guinean adults (age 15
years and older).
(9) The Labour Force Survey defines the informal sector on the
basis of the type of enterprise and the number of persons working in the
enterprise. The TUS 2007 reveals that the service workers and
plant/machine operators are primarily engaged in the informal sector.
(10) See Jafri (1999) and Qureshi and Arif (2001).
(11) See for example, Bardasi and Wodon (2006), Lawson (2007),
MeGinnity and Russell (2007), and Merz and Rathjen (2009).
(12) The decision to join the labour market (and the number of
hours to be supplied) itself depends on a number of other variables
including wage rate. To the extent that this may introduce endogeneity
in the present context, the coefficients of the regression models that
include employment status as an explanatory variable should be
interpreted with care.
(13) See Tables 5 and 10.
Najam-us.Saqib <saqibphd@yahoo.com> is Senior Research
Economist, Pakistan Institute of Development Economics, Islamabad. G. M.
Arif <gmarif@pide.org.pk> is Joint Director, Pakistan Institute of
Development Economics, Islamabad.
Table 1
Sample Characteristics
Total Sample
Both
Sample Sexes Male Female
Female -- -- 51.6
Mean Age (Years) 30.9 31.4 30.4
Urban 39.4 40.5 38.4
Rural 60.6 59.5 61.6
Marital Status
Currently Married 56.6 53.4 59.7
Unmarried 39.2 44.1 34.5
Others 4.2 2.5 5.7
All 100 100 100
(N) (37832) (18321) (19511)
Not-working/Not Employed
Both
Sample Sexes Male Female
Female -- -- 74.3
Mean Age (Years) 28.4 23.7 30.1
Urban 42.2 45.2 41.1
Rural 57.8 54.8 58.9
Marital Status
Currently Married 41.7 16.8 58.3
Unmarried 47.2 79.8 35.9
Others 5.1 3.4 5.7
All 100 100 100
(N) (21871) (5630) (16241)
Employed
Both
Sample Sexes Male Female
Female -- -- 20.5
Mean Age (Years) 34.3 34.9 32.0
Urban 35.7 38.4 25.0
Rural 64.3 61.6 75.0
Marital Status
Currently Married 68.9 69.6 66.5
Unmarried 28.2 28.3 27.6
Others 2.9 2.1 5.0
All 100 100 100
(N) (15961) (12691) (3270)
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 2
Labour Market Characteristics of the Employed Sample
Working/Employed Sample
Labour Market Characteristics Both Sexes Male
(Percentages) TUS * LFS ** TUS *
Occupation
Professionals 15.4 14.2 18.4
Associate Professional 6.6 5.3 5.5
Clerks 1.5 1.6 1.8
Service and Workshop 5.2 4.9 6.4
Agricultural Worker 33.0 37.4 29.3
Craft Workers 14.2 15.2 13.0
Machine Operators 4.0 4.0 5.1
Elementary (Unskilled) 20.0 17.4 20.7
Occupation (a)
Industry
Agriculture 39.7 44.6 35.3
Manufacturing 12.8 13.0 10.8
Electricity 0.8 1.0
Construction 6.9 6.3 8.7
Trade 14.8 14.6 19.1
Transport 5.1 5.5 6.4
Finance 1.8 2.2
Community and Social Services 17.6 15.7 17.1
Undefined 0.3 2.3 0.4
Employment Status
Employees 44.2 36.0 45.7
Self-employed 34.0 34.2 39.2
Unpaid Family Helpers 17.9 28.9 10.4
Employers 3.9 0.9 4.8
(N) (15961) -- (12691)
Working/Employed
Sample
Labour Market Characteristics Male Female
(Percentages) LFS ** TUS * LFS **
Occupation
Professionals 17.1 3.8 2.6
Associate Professional 5.0 10.9 6.6
Clerks 2.0 0.3 0.2
Service and Workshop 6.0 0.8 0.6
Agricultural Worker 31.3 47.5 60.9
Craft Workers 16.1 19.1 11.8
Machine Operators 5.0 9.1 0.2
Elementary (Unskilled) 17.5 17.5 17.1
Occupation (a)
Industry
Agriculture 36.9 57.0 75.0
Manufacturing 13.3 20.3 11.8
Electricity 0.1
Construction 7.8 0.2 0.4
Trade 17.9 2.2 1.8
Transport 6.8 0.3 0.2
Finance 0.3
Community and Social Services 14.4 19.5 10.6
Undefined 2.9 0.1 0.2
Employment Status
Employees 36.0 38.5 22.2
Self-employed 39.6 13.7 12.8
Unpaid Family Helpers 19.7 47.2 65.0
Employers 1.2 0.6 --
(N) -- (3270) --
Source: * TUS: Calculated from the micro-data of Time Use Survey,
2007.
** LFS: Figures for fiscal year 2007-08 taken from Pakistan Labour
Force Survey 2008-09.
(a): Elementary occupation includes unskilled workers such as street
vendors, cleaners, domestic helpers, and labourers in construction,
agriculture, and mining sector.
Table 3
Distribution of the Employed Sample by Monthly Income and
Sources of Income (Percentages)
Income and
Sources of Income Both Sexes Male Female
Upt to Rs 2000 15.1 9.8 35.6
2001-3000 9.4 10.3 5.7
3001-4000 12.5 14.8 3.6
4001-5000 11.3 13.7 2.2
5001-6000 8.6 10.3 2.0
6001-7000 6.3 7.5 1.7
7001-8000 4.7 5.6 1.2
8001-9000 3.2 3.8 0.9
9001-10,000 2.9 3.4 0.9
10,000 or More 9.8 11.6 2.9
Don't Know/Refusal 1.3 1.4 0.8
No Income (a) 14.8 7.7 42.6
Sources of Income
Wages and Salaries 44.2 45.5 38.8
Business 37.0 43.1 13.1
Transfer Income 3.2 2.7 5.0
Other 0.9 1.0 0.4
No income (a) 14.8 7.7 42.6
All 100 100 100
Source: * TUS: Calculated from the micro-data of Time Use Survey,
2007.
** LFS: Figures for fiscal year 2007-08 taken from Pakistan Labour
Force Survey 2008-09.
(a): These are unpaid family helpers.
Table 4
Mean Time Spent (Hours: Minutes) on Different Activities
by Work Status, Gender and Rural/Urban
Total Sample Employed Only
Non- Non-
Sample SNA Ex.SNA SNA SNA Ex.SNA SNA
Total Sample
All 3:15 2:55 17:50 6:58 1:22 15:40
Male 5:21 0:32 18:07 7:32 0:32 15:56
Female 1:15 5:10 17:35 4:42 4:39 14:39
Rural Areas
All 3:25 3:03 17:32 6:44 1:34 15:42
Male 5:27 0:31 18:02 7:22 0:32 16:06
Female 1:35 5:21 17:04 4:41 4:52 14:27
Urban Areas
All 2:58 2:43 18:19 7:22 1:02 15:36
Male 5:13 0:33 18:14 7:49 0:32 15:39
Female 0:44 4:52 18:24 4:44 3:59 15:17
Not-working
Non
Sample SNA Ex.SNA SNA
Total Sample
All 0:32 3:54 19:34
Male 0:24 0:32 23:04
Female 0:34 5:16 18:10
Rural Areas
All 0:44 4:16 19:00
Male 0:34 0:29 22:57
Female 0:48 5:27 17:45
Urban Areas
All 0:14 3:46 20:00
Male 0:13 0:35 23:12
Female 0:16 4:58 18:46
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 5
Mean Time Spent on Activities by the Employed Sample by
Their Occupation
SNA Ex-SNA
Occupation Rural Total Urban Rural Total Urban
Managers 8:34 8:15 8:04 0:39 0:34 0:32
Professionals 6:23 6:12 6:07 1:06 1:04 1:04
Ass.
Professionals 5:26 5:31 5:34 1:47 1:47 1:47
Clerk 6:49 6:56 7:00 0:39 0:41 0:42
Service
Workers 8:28 8:28 8:28 0:30 0:35 0:39
Agri-workers 6:03 6:02 5:45 1:56 1:56 1:55
Craft Workers 6:31 6:50 7:09 2:09 1:45 1:21
Plant and Mach
Operator 8:27 8:34 8:42 0:28 0:27 0:26
Elementary
Occup. 7:28 7:34 7:44 1:05 1:03 1:00
Non-SNA Male
Ex- Non-
Occupation Rural Total Urban SNA SNA SNA
Managers 14:47 15:11 15:24 8:23 0:26 15:11
Professionals 16:31 16:44 16:49 6:27 0:48 16:45
Ass.
Professionals 16:47 16:42 16:39 6:08 0:43 17:09
Clerk 16:32 16:23 16:18 6:58 0:35 16:27
Service
Workers 15:02 14:57 14:53 8:34 0:29 14:57
Agri-workers 16:01 16:02 16:20 6:45 0:32 16:43
Craft Workers 15:20 15:25 15:30 7:51 0:32 15:37
Plant and Mach
Operator 15:05 14:59 14:52 8:34 0:27 14:59
Elementary
Occup. 15:27 15:23 15:16 7:49 0:33 15:38
Female
Ex- Non
Occupation SNA SNA SNA
Managers 4:38 4:09 15:13
Professionals 4:31 2:53 16:36
Ass.
Professionals 4:38 3:31 15:51
Clerk 4:28 3:02 16:30
Service
Workers 5:11 3:42 15:07
Agri-workers 4:21 5:16 14:23
Craft Workers 4:11 4:57 14:52
Plant and Mach
Operator 8:15 2:14 13:31
Elementary
Occup. 6:23 3:25 14:12
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 6
Time Spent by Industry
Total Rural
Ex- Non- Ex- Non-
Industry SNA SNA SNA SNA SNA SNA
Agriculture 6:15 1:49 15:56 6:15 1:50 15:55
Manfu. 6:49 1:55 15:16 6:24 2:28 15:08
Elect. Gas 6:30 0:42 16:48 6:21 0:41 16:57
Constr. 7:44 0:36 15:40 7:44 0:37 15:39
Trade 8:38 0:30 14:52 8:46 0:33 14:41
Transport 8:24 0:36 15:00 8:15 0:38 15:07
Finance 7:28 0:33 15:59 7:26 0:30 16:04
Com.
Social. Ser 6:30 1:22 16:08 6:30 1:20 16:10
Urban Male
Ex- Non- Ex- Non-
Industry SNA SNA SNA SNA SNA SNA
Agriculture 6:04 1:43 16:13 6:52 0:30 16:38
Manfu. 7:10 1:26 15:24 8:02 0:31 15:27
Elect. Gas 6:34 0:42 16:44 6:26 0:40 16:54
Constr. 7:44 0:34 15:42 7:45 0:35 15:40
Trade 8:32 0:28 15:00 7:04 0:23 16:33
Transport 8:33 0:33 14:54 8:25 0:34 15:01
Finance 7:29 0:33 15:58 7:26 0:32 16:02
Com.
Social. Ser 6:28 1:24 16:08 7:00 0:42 16:18
Female
Ex- Non
Industry SNA SNA SNA
Agriculture 4:45 4:59 14:16
Manfu. 4:16 4:50 14:54
Elect. Gas 8:00 3:10 12:50
Constr. 4:41 3:47 15:32
Trade 5:20 4:04 14:36
Transport 6:52 2:40 14:28
Finance 8:37 1:08 14:15
Com.
Social. Ser 4:47 3:39 15:34
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 7
Time Spent (Hours: Minutes) by Employment Status
Both Sexes
Employment Status SNA Ex. SNA Non-SNA
Employee 7:18 1:11 15:31
Self-employed 7:21 0:52 15:47
Unpaid Family Helper 5:07 2:59 15:54
Employer 8:13 0:31 15:16
Males
Employment Status SNA Ex. SNA Non-SNA
Employee 7:44 0:33 15:43
Self-employed 7:36 0:33 15:51
Unpaid Family Helper 6:09 0:23 17:28
Employer 8:20 0:25 15:15
Female
Employment Status SNA Ex. SNA Non-SNA
Employee 5:20 4:04 14:36
Self-employed 4:25 4:27 15:08
Unpaid Family Helper 4:15 5:12 14:33
Employer 4:30 3:27 16:03
Source: Calculated from the micro-data of
Time Use Survey, 2007.
Table 8
Time Spent (Hours:Minutes) by Women in SNA Activities
Urban Rural Total
Working 5:29 4:56 5:04
Not-working 1:41 2:10 2:03
Total 3:14 3:16 3:15
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 9
Time Poor by Work Status, Gender and Rural-Urban Areas
Working/Employed Not-working/Not-employed
Both Both
Sexes Male Female Sexes Male Female
Poverty line = 10.5 hours per day
Urban 23.2 22.4 27.9 5.6 0.5 7.6
Rural 22.2 16.6 39.8 9.2 0.5 12.1
Total 22.5 18.9 36.8 7.7 0.5 10.2
Poverty line = 9.0 hours per day
Urban 44.7 44.3 46.9 13.2 0.9 19.0
Rural 42.1 36.1 61.2 20.1 1.1 26.2
Total 43.0 39.3 57.6 17.2 1.0 22.8
Poverty line = 12.0 hours per day
Urban 8.7 8.2 11.5 1.7 0.1 2.3
Rural 8.6 5.7 17.8 2.9 0.1 3.8
Total 8.6 6.7 16.2 2.4 0.1 3.2
Total Sample
Both
Sexes Male Female
Poverty line = 10.5 hours per day
Urban 12.3 14.9 9.8
Rural 15.0 12.1 17.7
Total 14.0 13.2 14.7
Poverty line = 9.0 hours per day
Urban 25.2 29.4 21.1
Rural 30.0 26.2 33.4
Total 28.1 27.5 28.7
Poverty line = 12.0 hours per day
Urban 4.4 5.4 3.3
Rural 5.5 4.1 6.7
Total 5.0 4.6 5.4
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 10
Incidence of Time Poverty (% Poor) by Occupation
(Employed only) and Industry
All Areas
Rural Urban
Occupation/Industry Both Male Female Areas Areas
Occupation
Manager 27.9 27.8 39.4 32.2 25.5
Professional 12.5 12.8 11.1 14.4 11.7
Associate Professional 12.8 9.4 19.4 12.2 13.1
Clerks 10.0 9.6 20.4 11.3 9.4
Service Worker 33.6 34.1 19.2 33.0 34.1
Agriculture 18.5 9.3 40.5 18.5 19.9
Craft Worker 24.3 20.1 35.3 26.5 22.1
Machine Operator 32.7 32.6 59.9 31.6 34.0
Elementary 23.6 20.6 43.2 24.3 25.2
All 22.5 18.9 36.8 22.2 23.2
Industry
Agriculture 19.5 10.0 42.3 19.4 21.4
Manufacturing 27.7 22.4 24.9 31.5 23.4
Electricity 13.8 12.6 66.7 18.2 11.6
Construction 17.6 17.5 33.3 17.6 17.5
Trade 32.0 31.9 27.5 34.3 30.6
Transport 32.4 32.3 40.0 32.9 31.8
Finance 16.9 16.7 22.2 12.5 17.7
Services 18.1 16.7 22.9 18.4 17.9
All 22.5 18.9 36.8 22.3 23.2
Source: Calculated from the micro-data of Time Use Survey, 2007.
Table 11
Time Poor by Income Per Month (Rs)
Total Rural
Income Per Both Both
Month Sexes Male Female Sexes Male Female
Upto 2000 29.8 16.5 39.7 29.2 15.3 44.3
2001-3000 23.6 21.7 36.9 21.5 19.2 43.6
3001-4000 22.8 22.1 33.9 21.2 20.4 35.7
4001-5000 23.2 22.6 37.0 20.9 20.2 46.4
5001-6000 21.5 21.7 17.2 19.3 19.1 23.1
6001-7000 29.4 20.2 24.1 19.8 19.8 20.7
7001-8000 17.7 16.9 32.5 13.4 12.3 35.0
8001-9000 16.6 16.2 23.3 13.6 12.5 31.3
9001-10000 17.8 17.9 16.7 13.6 13.6 12.5
10001 or more 15.8 15.3 24.0 12.9 11.0 30.4
Don't Know 22.1 19.8 45.5 14.8 12.7 33.3
Refused 22.7 24.2 11.1 17.3 17.0 20.0
Urban
Income Per Both
Month Sexes Male Female
Upto 2000 24.2 19.7 28.7
2001-3000 28.6 28.5 29.1
3001-4000 26.3 25.9 31.3
4001-5000 27.4 27.1 31.1
5001-6000 24.5 25.3 13.2
6001-7000 21.4 20.9 28.0
7001-8000 22.4 22.0 30.0
8001-9000 20.0 20.3 14.3
9001-10000 21.9 22.3 16.2
10001 or more 17.3 17.0 21.9
Don't Know 38.5 35.1 199
Refused 34.8 42.1 0
Source: Calculated from the micro-data of Time Use Survey, 2007.
Note: 18 percent of the employed sample has no monthly income.
Table 12
Logistic Regression: The Determinants of Time Poverty
Dependent Variable Time Poor = 1
Model 1
(Full Model 2
Sample) (Males)
Constant -4.429 * -6.525 *
Age. (years) 0.084 * 0.056 *
[Age.sup.2] -0.001 * 0.000 *
Gender (male=l) -1.088 * --
Place of Residence (urban=l) 0.094 * 0.462 *
Employment Status (employed=l) 1.772 * 3.557 *
Marital Status (married=l) 0.706 * 0.104
Children < 7 Years in the Household
(Yes= 1) 0.286 * 0.090 *
Education.(belowmatric=l) 0.392 * 0.421 *
Occupation (service workers, machine
operators/unskilled= 1) -- --
Employment Status (unpaid family
helpers=l) -- --
Industry (transport, trade and
manufacturing= 1) -- --
Monthly Income (below the minimum -- --
wage of Rs 7000=1) -- --
-- --
Season (Quarter 1=1) -- --
(Quarter 2=1)
(Quarter 3=1) -- --
N 37815 18308
-2 log Likelihood 25513 12371
Model 3 Model 4
(Females) (Employed)
Constant -6.299 * -2.618 *
Age. (years) 0.130 * 0.050 *
[Age.sup.2] -0.002 * 0.000 *
Gender (male=l) -- -1.064 *
Place of Residence (urban=l) -0.324 * 0.119 *
Employment Status (employed=l) 1.753 * --
Marital Status (married=l) 1.187 * 0.426 *
Children < 7 Years in the Household
(Yes= 1) 0.458 * 0.166 *
Education.(belowmatric=l) 0.236 * 0.375 *
Occupation (service workers, machine
operators/unskilled= 1) -- 0.007
Employment Status (unpaid family
helpers=l) -- 0.097
Industry (transport, trade and
manufacturing= 1) -- 0.763 *
Monthly Income (below the minimum --
wage of Rs 7000=1) -- 0.208 *
-- 0.418 *
Season (Quarter 1=1) -- 0.561 *
(Quarter 2=1)
(Quarter 3=1) -- 0.141 *
N 19507 15959
-2 log Likelihood 12144 15550
Model 5 Model 6
(Employed (Employed
Urban) Rural)
Constant -1.853 * -2.497 *
Age. (years) 0.010 0.070 *
[Age.sup.2] 0.000 -0.001 *
Gender (male=l) -0.439 * -1.375 *
Place of Residence (urban=l) -- --
Employment Status (employed=l) -- --
Marital Status (married=l) 0.433 * 0.386 *
Children < 7 Years in the Household
(Yes= 1) 0.114 * 0.183 *
Education.(belowmatric=l) 0.372 * 0.429 *
Occupation (service workers, machine
operators/unskilled= 1) 0.166 * -0.049
Employment Status (unpaid family
helpers=l) -0.103 -0.148 *
Industry (transport, trade and
manufacturing= 1) 0.567 * 0.857 *
Monthly Income (below the minimum
wage of Rs 7000=1) -- --
-- 0.303 *
Season (Quarter 1=1) -- 0.544 *
(Quarter 2=1)
(Quarter 3=1) -- 0.044
N 5696 10263
-2 log Likelihood 5938 9572
Source: Estimated from the micro-data of Time Use Survey, 2007.
* Significant at 5 percent or less level of significance.
