Levy Economics Institute of Bard College
Levy Economics
Institute
of Bard College
Policy Note
2017 / 4
HOW TIME DEFICITS AND HIDDEN
POVERTY UNDERMINE THE SUSTAINABLE
DEVELOPMENT GOALS
ajit zacharias
Introduction
The Sustainable Development Goals (SDGs) include a clear goal to reduce the incidence of pov-
erty by 50 percent by 2030 and a somewhat vague goal to recognize and render support in various
ways to the unpaid provision of domestic services and care of persons undertaken predominantly
by women in their households. Policymakers too often fail to grasp the connections between these
two goals—a failure that stems from an incomplete understanding of poverty and corresponding
blind spot in our official poverty statistics.
The predominant framework for assessing poverty is deeply flawed because it rests on an
implicit assumption that everyone has enough time available to devote to household production
or enough resources to compensate for deficits in household production by purchasing market
substitutes. Official measurements of poverty and poverty reduction are therefore doomed to
paint a biased picture of poverty. Results from our research conducted in seven countries indicate
the extent of bias stemming from this implicit assumption.1
Senior Scholar ajit zacharias is director of the Distribution of Income and Wealth program at the Levy Economics Institute.
The Levy Economics Institute is publishing this research with the conviction that it is a constructive and positive contribution to the
discussion on relevant policy issues. Neither the Institute’s Board of Governors nor its advisers necessarily endorse any proposal made by
the author.
Copyright © 2017 Levy Economics Institute of Bard College ISSN 2166-028X
Our findings suggest that, to obtain a more accurate por- always be a clear inverse relationship between the time spent
trait, assessments of progress in poverty reduction should take on household production and poverty status: some people may
time deficits in household production into account. They also devote the time they gain from reducing drudgery to caring for
highlight the interlocking of deficits in time and income (or their dependent children, so that the total time they spend on
consumption) for a substantial segment of the working popula- household production remains roughly unchanged. In spite of
tion. This indicates that a closer link may exist between the goals such interlinkages between the spheres of poverty reduction
of poverty reduction and support for household production and household production, there does not appear to be a clear
activities than is commonly acknowledged. Failure to recognize articulation of the need for taking household production into
the link in policy design can contribute to failure on both fronts. account in the measurement and monitoring of poverty.
The absence of serious consideration of household produc-
tion is reflected in the dominant thinking about poverty. A clear
A Blind Spot in Measurements of Poverty instance of this can be found in the manner in which conven-
The SDGs envisage a better world in many ways. I will focus tional poverty thresholds are constructed. As originally pointed
on the two goals of ending poverty (Goal 1) and greater gender out by Clair Vickery in a critique of the poverty lines employed
equality (Goal 5). The SDGs’ poverty reduction target (Target in the United States, the thresholds rest on an implicit assump-
1.2) is aimed broadly at all dimensions of poverty recognized by tion about the time spent on household production—specifi-
national governments.2 The qualifier is significant. In practice, cally, to survive with the poverty-level of income, the household
most countries use either income or consumption thresholds to will require a household member to be a full-time homemaker
identify the poor for official statistical purposes. who shops at the cheapest retail outlets, prepares meals from
The goal of greater gender equality has several components. scratch, takes care of children, and so on (Vickery 1977, 30).
Of particular interest to my discussion is the component (Target A research project initiated at the Levy Institute about seven
5.4) pertaining to unpaid care and domestic work (referred to years ago has attempted to unmask the implicit assumption in
hereafter as household production). The target seems to be to a variety of national contexts and modify the official poverty
provide recognition for household production and support to thresholds accordingly. The resulting picture of poverty is dra-
the women who disproportionately bear the responsibility for matically different and sheds new light on certain fundamental
such production.3 The desired extent of recognition and sup- aspects of poverty alleviation strategies. My goal is to provide
port are left to the discretion of national governments. Given an introduction to this body of work, highlight some of the key
that the earlier United Nations declaration on Millennium findings, and outline their implications for the SDGs’ targets.
Development Goals did not contain any reference to household
production, the “recognition-support” statement in the SDGs
may be considered a step in the right direction. The Levy Institute Measure of Time and Income
The two targets are interconnected in several ways. Poor Poverty (LIMTIP)
people tend to lack basic amenities in their households (this Once we accept the idea that surviving with household income
is indeed a characteristic of being poor in most of the devel- (or consumption)4 that is around or below the poverty line
oping world), which contributes to their spending more time requires a certain amount of time to be set aside for household
on household production tasks such as cleaning and cooking production, we have to ask the question: does every household
than their better-off counterparts. If they escape poverty, they have the requisite time? It is unrealistic to expect the answer
may be able to reduce the drudgery associated with household to be in the affirmative. For those who do not have the avail-
production. To illustrate the linkage in the other direction, we able time, the poverty line does not represent the minimum
can consider a policy initiative that makes free childcare avail- amount of monetary resources necessary to avoid material
able to poor parents. This may reduce the time that they spend deprivation—because households with time deficits will have
on household production and, provided that opportunities are to purchase market substitutes to fill gaps in household produc-
available, increase their hours of employment, which in turn tion just to attain the poverty-level living standard. Indeed, to
may allow them to become nonpoor. Of course, there need not consider two households that are identical in all respects except
Policy Note, 2017/4 2
time available for household production—one deficient and the are based on the official national income or consumption pov-
other not—as facing the same poverty line is inequitable toward erty lines in conjunction with our own estimates of national
the household with the time deficit (Zacharias 2017, 263–4). In thresholds for the required time for household production.
addition to being inequitable, this is also an internally incon- Because we are using nationally specific poverty lines, the esti-
sistent procedure for assessing poverty, because while the defi- mates discussed below are not, strictly speaking, comparable
nition of the threshold includes both the minimum monetary across countries.
income and (by assumption) the time required for household
production, the definition of resources includes only mon-
etary income (Zacharias, Antonopoulos, and Masterson 2012, Patterns of Time and Income Poverty
22). The natural way to correct the bias in the poverty line is to We found that the incidence of time deficits among employed
add the replacement cost of the time deficit—that is, the cost individuals was substantial in all the countries that we studied,
of buying goods and services to fill in for the missing house- with the rate falling between 38 percent (Ghana) and 52 percent
hold production—to the poverty line of households with time (Tanzania). This shows that the assumption that households
deficits. In line with the earlier literature (e.g., Harvey and (at least those with employed persons)6 are unlikely to have
Mukhopadhyay 2007), this is the strategy that we followed in time deficits is false. We also found that there was a striking
our studies. Making this correction reveals the poverty that is gender disparity among employed persons even after we con-
hidden by the conventional measures and contributes to better- trol for hours of employment (Figure 1).7 The rates for men
designed poverty alleviation strategies. and women converge at nearly 100 percent when the hours at
The distinctive conceptual feature of the LIMTIP is its the job are very long (over 61 hours per week) in all countries
treatment of time deficits. Previous studies that followed except South Korea, where the equality holds also at the lowest
the approach of Vickery have treated time deficits solely as a hours interval (20 hours or less per week).
household-level phenomenon, because they considered the The higher rate of time poverty among women is due to
household as a monolith in terms of time allocation. But this their bearing a higher share of household production respon-
notion goes against our everyday experience—and the evidence sibilities. This is reflected in the higher average values of the
amassed from time-use surveys from around the world—indi- required hours of household production for women compared
cating that the division of household production tasks is very to men (Figure 2). It is noteworthy how stable men’s average
unequal between members of the household. In general, even hours of household production are across different levels of
when men and women engage in employment for a similar hours of employment and how little variation women’s average
amount of time, women bear a higher share of household tasks. hours display. In short, it does not seem to matter much how
Discarding the monolithic assumption permits us to conceptu- long the hours at the job are for either sex: women dispropor-
alize time deficits as individual- and household-level phenom- tionately bear the responsibilities for household production.
ena. Thus, we can integrate intrahousehold gender disparities in As discussed above, resolving the bias stemming from
the division of household production tasks into the measure- ignoring time deficits in the official poverty line requires us
ment of poverty. This allows for the possibility that everyone in to include the monetized value of time deficits in the poverty
a time-poor household (a household with at least one person line of households with time deficits. Once we carried out the
with a time deficit) need not be time-poor, unlike the case of operation, we expected the adjusted or LIMTIP poverty rate to
income poverty in which everyone in an income-poor house- be higher. However, the size of the difference, or what we prefer
hold is considered as income-poor (Zacharias 2017, 272–6). to call “hidden poverty,” is remarkable (Figure 3). For example,
Combining the information about time and income deficits in Korea is the country with the lowest official poverty rate in this
the manner described above results in the LIMTIP. group, with only 5.4 percent of households. Taking time deficits
We have developed estimates of the LIMTIP for a set of into account almost doubles the measured poverty rate to 10
countries (in a given year): Argentina (2005); Chile (2006); percent, indicating that the extent of hidden poverty is almost
Ghana (2012–13); Korea (2009); Mexico (2008); Tanzania as large as officially recognized poverty. A relatively large num-
(2011–12); and Turkey (2006).5 Our estimates of the LIMTIP ber of Korean households (relative, that is, to the number of
Levy Economics Institute of Bard College 3
Figure 1 Rates of Time Poverty among Employed Persons (18–70 years of age) by Country, Sex, and Weekly Hours of
Employment
Argentina Chile Ghana
100 100 100
80 80 80
60 60 60
40 40 40
20 20 20
0 0 0
20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or
less more less more less more
Weekly Hours of Employment Weekly Hours of Employment Weekly Hours of Employment
Rate of Time Poverty (percent)
Korea Mexico Tanzania
100 100 100
80 80 80
60 60 60
40 40 40
20 20 20
0 0 0
20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or
less more less more less more
Weekly Hours of Employment Weekly Hours of Employment Weekly Hours of Employment
Turkey
100
80
60
40
20
0
20 or 21–35 36–50 51–60 61 or
less more
Weekly Hours of Employment
Female
Male
Source: Author’s calculations based on synthetic datasets (see Note 7).
Policy Note, 2017/4 4
Figure 2 Average Weekly Hours of Required Household Production among Employed Persons (18–70 years of age) by
Country, Sex, and Weekly Hours of Employment
Argentina Chile Ghana
40 40 40
30 30 30
20 20 20
10 10 10
20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or
Hours of Household Production (average weekly values)
less more less more less more
Weekly Hours of Employment Weekly Hours of Employment Weekly Hours of Employment
Korea Mexico Tanzania
40 40 40
30 30 30
20 20 20
10 10 10
20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or 20 or 21–35 36–50 51–60 61 or
less more less more less more
Weekly Hours of Employment Weekly Hours of Employment Weekly Hours of Employment
Turkey
40
30
20
10
20 or 21–35 36–50 51–60 61 or
less more
Weekly Hours of Employment
Female
Male
Source: Author’s calculations based on synthetic datasets (see Note 7).
Levy Economics Institute of Bard College 5
Figure 3 Poverty Rate among Households: Official vs. Figure 4 Monetized Value of Household Time Deficit as
LIMTIP Percentage of Household Consumption Expenditures, Ghana
50.7
Monetized Value of Time Deficit (percent)
50
60
41
40
34.6
30.8 40
Percent
30
23.3 24.5
21.6
20 18.1
16.4 20
11.1 10.9 10
10
6.2 5.4
0 0
Argentina Chile Ghana Korea Mexico Tanzania Turkey 1 2 3 4 5 6 7 8 9 10
Country Decile of Ratio of Consumption Expenditures to
LIMTIP Poverty Threshold
LIMTIP
Official Rural
Source: Author’s calculations based on synthetic datasets (see Note 7). Urban
Source: Author’s calculations based on synthetic datasets (see Note 7).
officially poor households) manage to stay above the official As is well known, the official poverty lines are constructed at
poverty line by putting in long hours at the job at the expense the household level; that is, every person in a poor household
of providing the minimum required care for their dependents is poor irrespective of how unequal the intrahousehold shar-
and homes. ing of income or consumption expenditures might be. In most
Apart from its impoverishing effects, “buying off ” time def- countries, this conception of poverty results in roughly similar
icits can be a relatively expensive proposition for many house- rates of income poverty for men and women, with the excep-
holds that are above the LIMTIP poverty line. Indeed, exercising tion being those countries with a relatively large proportion of
that option may be viable—even for many middle-income fam- single-female-headed households that are disproportionately
ilies—only by cutting back on other expenditures (e.g., clothing poor. However, viewed through the lens of LIMTIP, we can
or healthcare) or going into debt. Consider the case of Ghana, a observe two instances of clear gender disparity: the proportion
country that is classified by the World Bank as a lower middle- of women who bear the double burden of income and time
income country. The average monetized value of the time deficit deficits is much higher than that of men; and the proportion
is a sizeable amount even for households well above the poverty of women with neither income or nor time deficits is notably
line (Figure 4). Expressed as a proportion of total household lower than the proportion for men (Figure 5).
consumption expenditures, the monetized value of the time
deficit falls below 10 percent only in the seventh decile in the
urban distribution and in the eighth decile in the rural distribu-
tion. In the urban seventh decile, only three nonfood budget
shares—education (14 percent), housing (excluding rent), and
transportation (10 percent each)—registered a higher propor-
tion, while in the rural eighth decile none did, though education
(8 percent) came very close.8
Bringing time deficits to the fore also lays bare a gender
disparity that remains hidden in the official poverty statistics.
Policy Note, 2017/4 6
Figure 5 Distribution of Employed Persons (18–70 years of age) by LIMTIP
Argentina Chile Ghana
Male
Female
0% 40% 80% 0% 40% 80% 0% 40% 80%
Korea Mexico Tanzania
Male
Female
0% 40% 80% 0% 40% 80% 0% 40% 80%
Turkey
Male
Female
0% 40% 80%
Four-way Classification of Individuals According to LIMTIP
Income-poor and Time-poor
Income-poor and Time-nonpoor
Income-nonpoor and Time-poor
Income-nonpoor and Time-nonpoor
Source: Author’s calculations based on synthetic datasets (see Note 7).
Levy Economics Institute of Bard College 7
Conclusion aspirations embodied in the SDGs may not be possible in many
In the developing world, the most important pathway out of parts of the world without serious challenges to the status quo.
income poverty for the employed population and their depen-
dents is better employment prospects. A crucial question
highlighted by our findings is whether higher earnings will be Notes
sufficient to offset the impoverishing effects of time deficits. 1. The research findings reported in this policy note were
If not, although officially measured poverty may show prog- produced in collaboration with my colleagues at the Levy
ress toward the SDGs’ target, the reality for the working poor Institute—Rania Antonopoulos, Kijong Kim, Tamar
can be different. We must also recognize that, insofar as the Khitarishvili, Thomas Masterson, and Fernando Rios-
improvement in household earnings is attained by the entry of Avila—as well as collaborators elsewhere: Valeria Esquivel
women into employment, the agenda of gender equality will be (Argentina); Sarah Gammage and María Elena Valenzuela
undermined to the extent that women are likely to encounter (Chile); Monica E. Orozco Corona and Armando Sanchez
the double bind of income and time poverty. Furthermore, the Vargas (Mexico); Tae-hee Kwon (Korea); Emel Memiş
nexus of labor market/household production realities faced by (Turkey); Bernice Ofosu-Badu (Ghana); and Ahmed Makbel
women and men—unintentionally or not—can encourage the (Tanzania). I am also grateful for the financial support of
persistence of the “male breadwinner” model, especially among UNDP-RSCLAC for the Latin American studies; UNDP-
low-income households. Although they desperately need addi- Turkey for the research on Turkey; Korea Employment
tional income, it does not often “pay” for women to be full-time Information Service for the Korean study; and the Hewlett
workers, due to a combination of wage differentials and pre- Foundation for the research on Ghana and Tanzania.
carious work for women, men who are already working very 2. Target 1.2: “By 2030, reduce at least by half the proportion
long hours for slightly better pay, and the lack of social care of men, women and children of all ages living in poverty in
provisioning (Antonopoulos, Masterson, and Zacharias 2012). all its dimensions according to national definitions.”
This mechanism can undermine increases in female labor force 3. Target 5.4: “Recognize and value unpaid care and domestic
participation, with negative consequences for advancing gender work through the provision of public services, infrastruc-
equality in the multiple domains included in the SDGs. ture and social protection policies and the promotion of
These considerations point toward the need for integrating shared responsibility within the household and the family
household production into the measurement and understand- as nationally appropriate.” We should note that unlike the
ing of deprivation. They also highlight the need for policy design goal of poverty reduction, there is no specific quantitative
to be informed by such an integrated perspective. And they make target here—which is probably a reflection of the balance
it abundantly clear that the SDGs’ target of full employment, of forces or compromise among the contending parties for
decent work, and pay equity (Target 8.5)9 is crucial to attain- and against the inclusion of unpaid care work in the SDGs.
ing progress in poverty reduction (Target 1.2) and to easing 4. The Levy Institute measure has been constructed for coun-
the impoverishing effects of time deficits (Target 5.4). Making tries that use income as a measure of resources for assessing
investments in the social care infrastructure and physical infra- poverty and for those that use consumption. To avoid cum-
structure can hasten progress toward all three targets. Enforcing bersome sentences, I will generally refer only to income as
or enacting legislation regarding hours of employment to pre- the relevant resource.
vent overwork, moving toward living wages and away from 5. It should be noted that our samples for Argentina and
starvation wages, extending social insurance and labor protec- Chile are not representative of the entire country: instead,
tions to workers in the informal sector, and providing genuine they represent, respectively, the City of Buenos Aires and
support to small farms can also help us move closer to attaining Greater Santiago. Detailed analysis of the results and infor-
the SDGs. However, the current macroeconomic and political mation regarding sources and methods can be found in
regime that prevails within most countries (and internation- the following list of references: Zacharias, Antonopoulos,
ally) severely limits fiscal expansion and places the interests and Masterson (2012) for Argentina, Chile, and Mexico;
of profits before people. Real progress toward the wishes and Zacharias, Masterson, and Memiş (2014) for Turkey; and
Policy Note, 2017/4 8
Zacharias, Masterson, and Kim (2014) for Korea. The Zacharias, A., T. Masterson, and E. Memiş. 2014. Time Deficits
report discussing Ghana and Tanzania is currently under and Poverty: The Levy Institute Measure of Time and
preparation. Consumption Poverty for Turkey. Research Project Report.
6. Households with at least one employed person constitute Annandale-on-Hudson, N.Y.: Levy Economics Institute of
the vast majority of households in all the countries that we Bard College.
studied. This is likely to be generally true in other countries. Zacharias, A., T. Masterson, and K. Kim. 2014. The
7. The figures were generated using the individual country Measurement of Time and Income Poverty in Korea.
data files compiled for the research project. Each individual Research Project Report. Annandale-on-Hudson, N.Y.:
country data file is a synthetic dataset created by statisti- Levy Economics Institute of Bard College.
cally matching a time-use survey and income/consumption Zacharias, A. 2017. “Measurement of Time and Income
survey. The sources and methods used for Latin America, Poverty.” In I. Hirway, ed., Mainstreaming Unpaid Work:
Turkey, and Korea are discussed, respectively, in Zacharias, Time-use Data in Developing Policies. New Delhi: Oxford
Antonopoulos, and Masterson (2012); Zacharias, Masterson, University Press.
and Memiş (2014); and Zacharias, Masterson, and Kim
(2014). A forthcoming Levy Institute publication will pro-
vide details regarding the estimates for Ghana and Tanzania.
8. The deciles were computed separately for rural and urban
areas. In the rural distribution, the bottom three deciles
consisted entirely of consumption-poor households and
69 percent of the fourth decile was consumption-poor.
Only the bottom decile of the urban distribution was
made up entirely of consumption-poor households, while
in the second decile, about 32 percent of households were
consumption-poor.
9. Target 8.5: “By 2030, achieve full and productive employ-
ment and decent work for all women and men, including
for young people and persons with disabilities, and equal
pay for work of equal value.”
References
Antonopoulos, R., T. Masterson, and A. Zacharias. 2012. It’s
About “Time”: Why Time Deficits Matter for Poverty. Public
Policy Brief No. 126. Annandale-on-Hudson, N.Y.: Levy
Economics Institute of Bard College.
Harvey, A., and A. K. Mukhopadhyay. 2007. “When Twenty-
Four Hours Is not Enough: Time-Poverty of Working
Parents.” Social Indicators Research 82(1): 57–77.
Vickery, C. 1977. “The Time-Poor: A New Look at Poverty.”
The Journal of Human Resources 12(1): 27–48.
Zacharias, A., R. Antonopoulos, and T. Masterson. 2012. Why
Time Deficits Matter: Implications for the Measurement of
Poverty. Research Project Report. Annandale-on-Hudson,
N.Y.: Levy Economics Institute of Bard College.
Levy Economics Institute of Bard College 9
