LETTER • OPEN ACCESS You may also like
On the use of household expenditure surveys to - Making of KCl liquid fertilizer from liquidwaste manufacture of seaweed and
galvanized industry
monitor mismanaged plastic waste from food R Pasae, Maming and E Soekendarsi
packaging in low- and middle-income countries - Institutional capacity building of Bina HangTuah youth organization and Tambaan
Indah waste bank in the context of plastic
waste management in Tambaan Urban
To cite this article: Jim Allan Wright et al 2022 Environ. Res. Lett. 17 124029 Village, Pasuruan City, East Java
Province, Indonesia
Rudianto, E Yudaningtyas and R D
 Kasitowati
- Material flow analysis of China’s five
View the article online for updates and enhancements. commodity plastics urges radical waste
infrastructure improvement
Xiaomei Jian, Peng Wang, Ningning Sun
et al.
This content was downloaded from IP address 197.255.68.123 on 19/01/2023 at 09:55
Environ. Res. Lett. 17 (2022) 124029 https://doi.org/10.1088/1748-9326/aca623
LETTER
On the use of household expenditure surveys to monitor
OPEN ACCESS mismanaged plastic waste from food packaging in low- and
RECEIVED
22 August 2022 middle-income countries
REVISED
23 November 2022 Jim Allan Wright1,∗, Simon Damkjaer2, Heini Vaisanen3, Quaranchie Adama-Tettey4,
ACCEPTED FOR PUBLICATION Mawuli Dzodzomenyo5, Allan G Hill6, Lorna Grace Okotto7, Joseph Okotto-Okotto8
25 November 2022
and Peter Shaw9
PUBLISHED 1
6 December 2022 GIS and International Development, School of Geography and Environmental Science, University of Southampton, Building 44,
Highfield, Southampton SO17 1BJ, United Kingdom
2 Research Fellow, School of Geography and Environmental Science, University of Southampton, Building 44, Highfield, Southampton
Original content from
SO17 1BJ, United Kingdom
this work may be used 3
under the terms of the Social Statistics and Demography, University of Southampton, Building 58, Highfield, Southampton SO17 1BJ, United Kingdom
Creative Commons 4 Behaviour Change Communication Practitioner, Ministry of Sanitation and Water Resources, Accra, Ghana
Attribution 4.0 licence. 5 Ghana School of Public Health, University of Ghana, Legon, Accra, Ghana
Any further distribution 6 Population and International Health, Social Statistics and Demography, University of Southampton, Building 44, Highfield,
of this work must Southampton SO17 1BJ, United Kingdom
maintain attribution to 7 School of Spatial Planning and Natural Resource Management, Jaramogi Oginga Odinga University of Science and Technology, Bondo
the author(s) and the title
of the work, journal (Main) Campus, P O Box 210-40601 Bondo, Kenya
citation and DOI. 8 Victoria Institute for Research on Environment and Development (VIRED) International, P O Box 6423-40103, off Nairobi Road,
Rabuor, Kisumu, Kenya
9 School of Geography and Environmental Science, University of Southampton, Building 44, Highfield, Southampton SO17 1BJ,
United Kingdom
∗ Author to whom any correspondence should be addressed.
E-mail: j.a.wright@soton.ac.uk
Keywords: domestic waste, food consumption, household expenditure survey, low and middle income countries,
mismanaged plastic waste
Supplementary material for this article is available online
Abstract
Background: substantial increases in plastic production have resulted in plastics proliferating of in
the environment, with subsequent seabed plastic deposition and ingestion by marine fauna. There
is an urgent need to monitor mismanaged plastic waste from household consumption. Household
expenditure survey analysis has quantified mismanaged plastic waste generated from household
packaged (bottled or bagged) water consumption, but not from consumption of other products.
Methods: to evaluate whether household expenditure surveys can quantify mismanaged waste from
other widely consumed commodities, we quantify mismanaged plastic waste from the domestic
consumption of cooking oil alongside packaged water in urban Greater Accra, Ghana, and all cities
nationally in Kenya using two household expenditure surveys. Results: household survey-derived
estimates indicate packaged water consumption generates considerably more plastic waste than oil
packaging in Greater Accra, whereas oil packaging generates more plastic waste than packaged
water in urban Kenya. Conclusion: by successfully transferring a survey analysis protocol from
packaged water to cooking oil, we conclude that there is ample potential for expenditure surveys to
be used internationally to quantify mismanaged plastic waste from households. However,
uncertainties affecting mismanaged waste estimates need to be accounted for.
1. Introduction production, with over 360 million tons produced
globally in 2020 [3]. The resultant proliferation
Growing concerns regarding plastic in the of plastics has led to widespread impacts [4] and
environment [1] are related to mismanaged waste contamination [5]. Impacts include seabed depos-
[2]. This problem stems from the high level of plastic ition of plastics, micro-plastic ingestion by and
© 2022 The Author(s). Published by IOP Publishing Ltd
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
entanglement of marine fauna [6], with subsequent with those from packaged water consumption among
micro-plastic transfer tomarine invertebrates [7] and such households.
ultimately species at higher trophic levels including This study focused on packaged water and cook-
humans [8]. ing oil in Greater Accra Region, Ghana and urban
The uncontrolled release of plastics into the envir- Kenya, thereby exploring household expenditure sur-
onment relates in part with available waste manage- vey use beyond West Africa. As case study products,
ment systems, which differ markedly between eco- we chose cooking oil given its ubiquitous use world-
nomically developed and developing countries [9]. wide, while packaged water consumption is rapidly
Open burning of solid waste in low- and middle- increasing in many countries (e.g. Laos, Turkey, and
income countries (LMICs), for example, releases Indonesia), forming the main drinking water source
atmospheric particulates that adversely affect citizens’ for most urban Ghanaian households (58%) in 2019
health [10]. In the absence of waste services, predicted [16]. Water is sold in bags in Ghana [17], typically in
population growth and changing urban consumption 500 ml plastic sachets [15]. In Kenya, packaged water
patterns in LMICsmay exacerbatemismanaged waste is mostly bottled [18]. In Ghana and Nigeria, palm
disposal. oil is generally packaged by vendors or smallholder
Extensive reliance on small-scale, ad hoc stud- farmers [19, 20]. Thus, these case studies present
ies or country-specific datasets means that ongo- contrasting situations for evaluating the potential of
ing, international monitoring of domestic waste and household surveys in estimating plastic packaging
its management lacks consistency in study designs waste.
[11]. Since primary fieldwork to monitor domestic
waste generation and disposal is costly, household
expenditure surveys could form a more cost-effective 2. Method
and internationally standardised means of monitor-
ing mismanaged waste arising from domestic con- 2.1. Secondary data sources
sumption of specific products. Data pertaining to household consumption patterns
Household expenditure surveys ask a nation- and waste disposal practices are available for both
ally representative sample of households to estim- Ghana [21] and Kenya [22]. The Ghana Living Stand-
ate spending on consumption of goods and services. ards Survey 7 (GLSS7) is the latest in a series of
They include Household Budget Surveys (HBS), nationally representative surveys quantifying house-
Income and Expenditure Surveys (IES), and the hold consumption. In 2016–17, the GLSS7 sampled
World Bank’s Living Standards Measurement Sur- 1000 enumeration areas (EAs) across Ghana’s ten
veys (LSMS) [12]. Consumption modules have also regions based on probability-proportional-to-size
been implemented within some Multiple Indicator sampling. In each EA, 15 households were randomly
Cluster Surveys [13]. There are now 137 openly selected, resulting in 14 009 sampled households [23].
available LSMS surveys for 37 countries (figure 1); Respondents reported food and beverage purchases
HBS and IES have similar availability. By 2013, 43 in the preceding week over six consecutive weeks
of 49 sub-Saharan African countries had conduc- and their main means of solid waste disposal. Only
ted at least one consumption and expenditure survey the urban Greater Accra region was evaluated in
[12], although surveys take place only every three the present study, the Ghanaian region where sachet
to ten years, so can be outdated. Since many HBS, water consumption is concentrated [15]. In Kenya,
IES, and LSMS also record household waste services given the more even regional distribution of pack-
or disposal [14], they have potential for character- aged water consumption, we analysed data for all
ising waste generated from consumption of specific cities.
products, particularly in households lacking waste The Kenya Integrated Household Budget Survey
services. We previously used household expenditure (KIHBS) 2015–16 is also a nationally representat-
surveys in Ghana, Liberia, and Nigeria to character- ive cross-sectional household survey. Its clustered
ise plastic waste generated through packaged (bagged sampling design included 24 000 households in
or bottled) water consumption among households Kenya’s 47 counties [24] chosen from2400 randomly-
lacking waste services [15]. However, it is unclear sampled clusters. In each cluster, 16 households
if household expenditure surveys could be used to were randomly selected, of which ten were sub-
quantify plastic waste generation from household sampled to participate in the main KIHBS. The
consumption of other products and in countries out- survey response rate was 91%. The questionnaires
side West Africa. addressed sociodemographic characteristics, wealth,
This study therefore aims to assess (a) whether economic standing, and consumption expenditure.
household expenditure surveys can be analysed to Adult household respondents reported spending on
quantify plastic waste generation from consumption itemised services and goods, including cooking oil
of a different product (cooking oil) among house- and packaged water, in the seven days before inter-
holds lacking waste services; (b) how quantities of view. Householders also completed diaries recording
plastic waste generated from cooking oil compare their purchases and expenditures.
2
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
Figure 1. The global distribution of living standards measurement surveys with available micro-data per country, September 2021.
2.2. Field data collection High, medium, and low plastic use scenarios
For water sachets inGreater Accra, each 500ml plastic were generated to convert quantities purchased to
sleeve was estimated to weigh 1.7 g of high- or low- amounts of packaging), reflecting the uncertainty
density polyethylene (HDPE/LDPE) [15]. For other associated with the survey measurement units. Since
packaging units, 97 and 16 samples of water and oil most water purchases in Greater Accra were recor-
packaging were collected from markets and kiosks in ded in standard units (sachets or bags of 30 sachets),
Kisumu (Kenya) and Greater Accra (Ghana) respect- the associated low, medium, and high scenarios were
ively, from July to September 2021 (Online Resource almost identical. In Kenya, where measurement units
1). The empty containers were cleaned and weighed, for water purchases were mostly in litres, we gen-
and the lids and labels were removed. erated high (0.5 l bottles), medium (1.0 l bottles),
and low (1.5 l bottles) waste scenario estimates.
2.3. Data analysis In Greater Accra, oil purchases mostly comprised
For both cooking oil and packaged water, purchases vendor-measured units, so we assumed the wide-
were used to infer packaging by classifying the meas- spread use of sachets or hand-tied bags for low and
urement units into four groups: medium scenarios and plastic bottle use for a high
waste scenario. In Kisumu, our fieldwork showed that
(a) Specific containers: in Ghana, these include oil was mostly sold in plastic jerrycans. Therefore, in
manufacturers’ sachets for packaging oil and Kenya, our three scenarios assumed: all oil sold in
water, reused branded water bottles, and plastic small 1 l jerrycans (high waste); all oil sold in the
bags hand-tied by vendors. largest possible jerrycan (medium waste), and half of
(b) Measurement and packaging by vendors (e.g. the oil sold in the largest possible jerrycan and half
tins, heaps, beer bottles, branded bottles). sold in the largest possible plastic bottle (low waste).
(c) Standard units (e.g. litres or gallons) In Ghana, where each household was visited six
(d) Unspecified units (e.g. ‘single’; ‘piece’). times, we also used logistic regression to test for sig-
nificant differences in the reporting of water or oil
For each product, the mean packaging ineffi- purchases per survey visit, using regression coeffi-
ciency (weight of plastic packaging per unit volume cients to adjust estimates where required. In Kenya,
of product) was calculated to convert the consump- where the KIHBS recorded the retail outlet for each
tion data to plastic waste quantities. Published data on purchase, we also cross-tabulated purchases against
cooking oil and water packaging volumes andweights retail outlet. Finally, we estimated the annual weight
were also obtained for comparison. Search terms of plastics generated through oil or packaged water
relating to the two products (‘oil’ or ‘water’), contain- consumption according to the main household waste
ers (e.g. ‘bottle’, ‘sachet’, ‘container’) and weight were disposal method, differentiating waste collected via
combined and searches made using Web of Science, service providers from other forms of disposal. Ana-
Google Scholar and Google. Promotional statistics lyses of plastics by waste disposal method accoun-
and data presented without a stated methodology ted for the GLSS7’s multi-stage cluster design via the
were excluded. svy commands in Stata [25] to generate nationally
3
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
representative estimates with 95% confidence inter- broadly comparable with published estimates for
vals reflecting sampling error. Where multiple pack- other LMICs (figure 4; Online Resource 1). Only
aging assumption scenarios were used, the estim- one set of measurements for oil containers (from
ate from the medium waste scenario was presented Brazil) was identified from literature, precluding sim-
alongside the widest confidence intervals from any of ilar comparisons for oil containers. Smaller-capacity
the three scenarios used. Finally, to explore the eco- containers were generally less efficient than larger
nomic drivers of lightweight packaging, we examined containers. Themost efficient containers were sachets
the price of 500 ml water sachets versus bottles in and hand-tied bags, which are used extensively in
Greater Accra Region via regression analysis. Ghana for both oils and water. In Kisumu, the jerry-
cans widely used for oil were less efficient than bottles
3. Results (figure 4(b)).
3.1. Quantification of plastic waste generation 3.2. Plastic waste generation from cooking oil and
using household expenditure survey data packaged water consumption
concerning cooking oil and packaged water Most plastic waste from sachets in Greater Accra is
consumption generated in households that have a formal waste col-
Overall, 5% of urban households reported buying lection system in place, so relatively small proportions
packaged water in Kenya during the seven days before are dumped indiscriminately or burned (figure 5(a)).
interview and 57% reported buying vegetable oil. Unlike the GLSS7, the KIHBS records the waste col-
In Greater Accra, households reported purchasing lector type, thereby highlighting the contribution
water sachets in 71% of the weekly survey visits. of private companies and informal waste collectors
Logistic regression indicated no significant differ- (figure 5(b)).
ences in water purchase reporting by GLSS7 sur- Consumption of palm or vegetable oils in urban
vey round, but respondent fatigue was evident in Greater Accra mostly takes place within house-
reported oil purchases following the first visit (odds holds with waste collection services (figure 6(a)).
ratio = 1.14; p = 0.03), so reported purchases from The confidence intervals for plastic waste generated
subsequent visits were adjusted upwards by 1.09. Oil are wider than those for packaged water, reflecting
purchases were more prevalent in urban Kenya, and the uncertainty in interpreting measurement units
packaged water purchases were more prevalent in (figure 5(a)). In Greater Accra, the total amount of
Greater Accra. plastics generated from oil consumption is much
In Greater Accra, most water purchases were lower than that of packaged water, as is the num-
recorded as individual 500 ml sachets or sachet bags ber of reported oil purchases (figure 3). In Kenya,
(30 × 500 ml sachets) (figure 2(a)). In Kenya, water private collectors handle more waste from vegetable
purchases were typically expressed in metric units oil than other collectors do (figure 6(b)). Among
(figure 2(b)). After controlling for the discounting Kenyan households that lackwaste collection services,
of bulk purchases, regression analysis of packaged plastic waste from oil consumption exceeds that from
water prices indicated that bottled water sold for sig- packaged water consumption.
nificantly more than sachet water in Greater Accra
Region (GHS1.96 or $0.46 versus GHS0.36 or $0.08, 4. Discussion
p < 0.001).
A wide variety of cooking oil measurement units 4.1. Plastic waste insights from household
is apparent in Greater Accra, particularly for vendor- expenditure surveys in Ghana and Kenya
measured units of palm oil (figure 3(a)). Oil is also Our household survey analysis provides insights into
packaged by manufacturers in sachets rather than in domestic generation of mismanaged plastic waste.
bottles. Some units signify container reuse, such as The analysis shows that plastic waste from sachet
the use of beer, Fanta (soft drink), or Voltic (water) water consumption (figure 5(a)) far exceeds that from
bottles. Similarly, olonkas (American tins) indicate oil (figure 6(a)) in Greater Accra, whereas in urban
that vendors reuse containers for measuring. In con- Kenya, the opposite occurs (figures 5(b) and 6(b)).
trast, in Kenya, measurement units for oils were Although plastic sachets weigh little, sachet water is
largelymetric, although some vendor-measured units among the most frequently purchased items recorded
such as gorogoros (tins for measuring out quantit- in the GLSS7. As noted in the USA [26], the ability to
ies, cf. Ghanaian olonkas) were evident. In the KIHBS assess the contribution of different food products
(figure 3(b)), vegetable oil purchased from supermar- to mismanaged plastic waste could provide evidence
kets was measured in litres in 83% of reported cases, to target regulatory policy concerning food packaging
while vendor-measured units were more common in and suitable waste management facilities at specific
open markets and kiosks. products of concern. The household survey analysis
In terms of container inefficiency, polyethylene also quantifies the importance of Kenya’s private pro-
terephthalate (PET) bottles (without lids) used for viders, either working informally or contracted via
water packaging in Greater Accra and Kenya were public–private partnerships [27]. In Kenya, where
4
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
 
(a) 
 
(b)  
Figure 2.Measurement units for water transactions per week reported by (a) 1,253 households in urban Greater Accra Region,
Ghana, totalled over six weekly visits in 2016–2017; (b) 956 households in Kenya totalled over one week in 2015–16 (x-axis codes
are 1: specific container; 2: units for vendor-measured purchases; 3. Imperial/standard unit; 4: unspecified units).
the KIHBS differentiated service providers by sector, Non-standard measurement units highlight the
private providers collected more oil and water pack- widespread practice of vendors packaging goods via
aging waste than local governments or community reused containers. In both urban Kenya and Greater
associations (figures 5 and 6). Accra, containers were widely reused in markets
5
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
Figure 3.Measurement units for vegetable or palm oil transactions reported by (a) 1,253 households in urban Greater Accra
Region, Ghana, totalled over six weekly visits in 2016–17; (b) 7551 households in Kenya, totalled over one week in 2015–2016
(x-axis codes are 1: specific container; 2: units for vendor-measured purchases; 3: imperial/metric units; 4. Unspecified).
to measure product quantities and, in some cases, out commodities for sale, reflecting reported con-
reused as packaging. Vendors’ resale of smaller, more tainer reuse by vendors elsewhere inWest Africa [29].
affordable quantities of food has been described as Similarly, soda, beer, and water bottles were some-
the ‘Kadogo economy’ in Kenya, meaning ‘tiny’ in times reused to sell vegetables or palm oil (figure 3).
Kiswahili [28]. Vendors used margarine tubs or tins However, systematic recording of container reuse
(Gorogoros in Kenya orOlonkas in Ghana) tomeasure would likely require ancillary fieldwork. Some widely
6
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
Figure 4. Container inefficiency (grams of plastic per litre of capacity) versus container capacity based on literature review and
fieldwork in Greater Accra, Ghana and Kisumu, Kenya, for (a) packaged water and (b) cooking oils.
used non-standard packaging units (e.g. hand-tied as oils, has not. In the USA, premium bottled
bags) were not pre-coded into theGLSS7 survey ques- water brands have higher packaging inefficiency than
tionnaire, but could be systematically recorded in budget brands [26, 31]. Packaging inefficiency for
future surveys. bottled water from Mexico and India is similar to
Widespread use of lightweight plastic sachet pack- that of mid-range US brands [31]. The present study
aging for both oils and vended water is evident in suggests that higher-volume purchases in LMICs
Greater Accra, whereby savings on packaging mater- entail less packaging per litre than smaller purchases
ials reduce costs to consumers. Although Ghana’s (figure 4). The GLSS7 data suggest that food or
sachet water industry has previously been well beverage products manufactured for LMIC markets
documented [30], the use of sachets as a lightweight sometimes use less packaging than their high-income
alternative to PET bottles for other products, such country equivalents. However, the substitution of
7
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
Figure 5. Estimated total annual weight of plastics per year from packaged water consumption by urban households in (a) Greater
Accra region, Ghana, in 2016–17 by main form of refuse disposal (vertical lines indicate widest 95% confidence intervals from
low, medium, and high plastic use scenarios) and (b) Kenya (1: waste disposal potentially managed appropriately; 2: mismanaged
waste disposal).
PET bottles with sachets inhibits their reuse poten- the reuse of PET bottles permits greater value to be
tial. Sachet sleeves, thin films of HDPE or LDPE, are gained from the materials and energy used for their
recycled to some extent in Greater Accra and collec- production [33] in alignment with waste hierarchy
ted by waste pickers [32], but not reused. In contrast, aims and principles [34].
8
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
Figure 6. Estimated total annual weight of plastics per year from palm or vegetable oil consumption by urban households in
(a) Greater Accra region, Ghana, in 2017 by main form of refuse disposal and (b) Kenya (vertical lines indicate widest of 95%
confidence intervals for low and high plastics scenarios; 1: waste disposal potentially managed appropriately; 2: mismanaged
waste disposal).
For urban water delivery, a three-axis frame- plastic waste from packaged water as a fourth axis.
work has been proposed that characterises its quant- Successive household surveys also provide oppor-
ity, availability, and delivery [35]. Given the wide- tunities to monitor changes in goods consumed by
spread use of packagedwater in someLMIC cities, our households lacking waste services in different years
approach could enable the addition of mismanaged within a given country.
9
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
4.2. Implications for using household expenditure consumption [14]. However, products with similar
surveys to quantify mismanaged waste from purposes may be packaged differently. For example,
domestic consumption water sold in bottles uses more plastic per unit
In applying an analysis protocol developed for pack- product than sachet water and generates PET rather
aged water to cooking oil, we also identify some thanHDPE plastics [15], bothmay be recorded under
generic issues affecting plastic waste quantification a single packaged water commodity code. While the
via household expenditure surveys. A typical survey consumption of food away from home (FAFH) is
may use standard (e.g. litres) or non-standard units growing in LMICs [40] and many surveys include
(e.g. bunches) to record food quantities consumed commodity codes for FAFH foods, few explicitly cap-
[36]. The quantification of domestic waste associ- ture meals or snacks consumed outside the home
ated with packaging requires a conversion table to [43]. Waste disposal facilities for FAFH will also not
translate non-standard quantities of specific products be captured by LSMS, preventing mismanaged waste
into quantities and resin types of plastic waste. Such quantification for FAFH via surveys.
tables currently do not exist; thus, primary fieldwork There are also challenges in measuring domestic
is required to quantify the weights of different pack- waste disposal through household surveys. While
aging types associated with commonly consumed LMIC households in cities such as Lusaka, Bulawayo,
foods. However, a multi-country study [37] found and Mombasa reportedly separate waste for burning,
that food metrics could be constructed for only 53% recycling, or reuse [44], recycling rates are generally
of recent household surveys because of the difficulties low; for example, only an estimated 6% of domestic
in constructing conversion tables for non-standard waste is recycled in Lusaka [44].Markets for the recyc-
units. Our case study suggests that packaged water ling or reuse of plastic waste vary by product and
is sold in standardised units (figure 2), whereas the plastic type. For example, there is extensive informal
units for oils are highly variable (figure 3). Where reuse of water PET bottles in urban Senegal but no
present, such heterogeneous measurement units pose reuse of plastics from sachet water [29]. However, in
challenges for the analysis of food and beverage most LSMS, only the main method of waste disposal
packaging. is typically recorded. Without specific questions on
Some generic LSMS data quality issues impact waste separation and secondary waste disposal meth-
their use in domestic waste quantification. For ods, household surveys may inaccurately estimate
example, ‘prestige errors’ may lead to over-reporting mismanaged waste. Furthermore, while surveys such
of socially desirable commodities [38]. Many LSMS as theKIHBS andGLSS7 identify collectedwaste, they
use aweekly recall period for food consumption ques- do not record whether it is safely contained following
tions administered overmultiple, repeated household collection, so overall waste mismanagement may be
visits. Respondent fatigue has been noted in some underestimated.
household surveys with repeated visits [39] and was Household consumption surveys also lack geor-
apparent in household reporting of oil purchases in eferencing. Although Demographic and Health Sur-
the GLSS7. Reported estimates may also be subject veys are released with GPS coordinates for household
to both ‘telescoping’, whereby events before the recall clusters [45], they lack consumption modules [12].
period are reported within it, and ‘memory decay’, Conversely, LSMS surveysmeasure consumption, but
whereby consumption details are progressively for- their lack of detailed spatial referencing restricts spa-
gotten as the recall period lengthens [40]. tial analysis to the provincial level. Studies modelling
The most widely used recall period in the LSMS the onwards environmental fate of plastics typically
is a week [37], which makes it challenging to use river catchments as spatial units [1], and the lack
differentiate between products purchased at differ- of detailed LSMS georeferencing inhibits the poten-
ent frequencies. Less frequent bulk purchases often tial integration of derived mismanaged waste estim-
entail additional secondary packaging [41], whereas ates with environmental transport models.
packaging efficiency is generally greater for bulk items
(figure 4). However, estimating waste from food con- 4.3. Future approaches to quantifying mismanaged
sumption as opposed to waste disposal events may domestic waste via household surveys
be preferable. While many foods are frequently pur- Several potential approaches could be used to char-
chased, some waste disposal events may occur less acterise mismanaged domestic waste via household
frequently as households accumulate waste before surveys. First, future analyses could focus on a lim-
disposal [42]. ited number of case study products sold in stand-
The commodity codes used in LSMS also likely ardised packaging of particular concern for national
impact the quantification of packaging associated waste management. In the USA, such analyses have
with domestic waste streams. There is international identified specific products for which packaging can
variation in the number and nature of codes used be reduced [26]. In Liberia, Nigeria, and Ghana, for
[37], which reflects the local context and the sur- example, water sachets are typically sold in stand-
vey’s primary purpose [40]. Commodity code sys- ardised quantities (individual 500 ml plastic sleeves
tems are often harmonised to reflect the purpose of or 15 l bags of 30 sachets), which facilitates the
10
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
quantification of plastic waste generated and cross- [48] are of merit, although these are scarce in
tabulation with waste service type(s) [15]. A second LMICs.
approach would be to classify commodities into
groups based on their packaging, perhaps in conjunc- 5. Conclusions
tion with data on the retail channels through which
they are sold (where available). Analogous to the There is considerable potential to exploit existing
World Food Programme’s Food Consumption Score international household consumption surveys for the
[12], commodities can be grouped by likely packaging quantification of domestic waste. This case study of
type (e.g. composites, LDPE, HDPE, PET, card/pa- cooking oils and packaged water in Greater Accra,
per, and glass). However, this would require an expert Ghana, and Kenya demonstrates quantification of
understanding of the commodity packaging. domestic mismanaged plastic waste, container reuse
A third approach records both packaging and and the use of sachets as lightweight, inexpensive
waste management facilities through a community or packaging. It also highlights the contributions of dif-
market surveymodule, similar to those used in house- ferent service providers in Kenya to domestic waste
hold surveys for recording prices or neighbourhood collection. There are nonetheless challenges arising
facilities. Market survey observations are sometimes from variations in survey implementation and con-
used to convert non-standard units in food consump- tent. Some of these challenges could be addressed
tion modules into metric units [46]. However, this through the adaptation of household survey proto-
approach would increase implementation costs and cols, for example, by aligning measurement units on
present challenges in identifying retail outlets used by questionnaires with widely used forms of packaging.
the sampled households. Although community and There is also a need for complementary fieldwork,
price modules are typically implemented within the particularly for market surveillance of packaging,
EAs selected for an LSMS, these small areas do not to enable more robust quantification of misman-
generally encompass the services or larger retail out- aged plastic waste. Data enabling the conversion of
lets used by these households [47]. nonstandard product measurements into standard
Minor adaptations to existing household survey units via conversion factors are required for a wide
designs could enhance their use in quantifying mis- range of commonly purchased products and their
managed plastic waste and thereby inform initiatives packaging.
to better manage the resources and materials con-
sumed in plastic packaging production. Becausemost Data availability statement
surveys only record the main mode of waste dis-
posal, an additional question pertaining to product- All data that support the findings of this study are
specific waste separation would be valuable. Simil- included within the article (and any supplementary
arly, more response categories for waste service oper- files).
ators would enable a fuller quantification of their
benefits (such as the separate categories for local gov- Acknowledgments
ernment, community association, and private waste
operators in the KIHBS). Given that the measure- This research was undertaken through the ‘Expand-
ment units and packaging varied for oil purchases ing safe water and waste management service access
by retail outlet in Kenya, recording the market chan- to off-grid populations in Africa’ project funded
nels for purchases has merit. A review of survey through a UKRI Collective Fund award via the Global
codes for non-standard measurement units could aid Challenges Research Fund (ref: ES/T008121/1). The
in the understanding of waste. In the GLSS7, for support of the UK Economic and Social Research
example, there was no predefined code for hand-tied Council (ESRC) is gratefully acknowledged.
plastic bags, despite their widespread use for selling Author contributions were as follows:
palm oil. Proposed commodity codes could be also Conceptualisation—JW, AH, HV; Investigation—
reviewed against heterogeneity of packaging. In the MD, QAT, JOO, LO; Writing—Original draft: JW &
GLSS7, for example, there are separate commodity PS; writing—review & editing: HV, SD, AH, MD,
codes for bottled and sachet water purchases that JOO, LO, QAT; Formal analysis, visualization—JW,
facilitate plastic waste quantification. In other coun- PS, & HV; Project administration—JW, MD, JOO,
tries, even when bagged water is sold, there is often LO; Funding acquisition—JW, AH, HV, JOO, LO,
only a single commodity code for recording packaged MD, PS.
water.
Finally, estimates of collected and mismanaged Conflict of Interest declaration:
waste from household surveys need to be evalu-
ated through triangulation with other data sources; The authors declare that they have NO affiliations
retail sector records on product sales and regulator with or involvement in any organization or entity
records on waste composition [26], combined with with any financial interest in the subject matter or
bespoke household surveys to assess waste generation materials discussed in this manuscript.
11
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
Ethical statement burning of municipal solid waste (MSW) in Indian cities
Environ. Sci. Technol. 49 12904–12
This secondary data analysis of human subjects data [11] Kaza S, Yao L, Bhada-Tata P and Van Woerden F 2018What
a Waste 2.0: A Global Snapshot of Solid Waste Management to
was approved by the ethics committee of the Fac-
2050 (Washington, DC: World Bank)
ulty of Environmental and Life Sciences, University [12] Carletto C, Zezza A and Banerjee R 2013 Towards better
of Southampton, UK (Ref: 56 807; approval date: 10 measurement of household food security: harmonizing
May 2020). All procedures performed in studies ana- indicators and the role of household surveys Glob. Food
Secur. 2 30–40
lysing data on human participants were in accordance
[13] Natali L and de Neubourg C 2014 Household welfare
with the ethical standards of the institutional and/or measurement in Bangladesh: a tale of two short consumption
national research committee and with the 1964 Hel- modules 2014-17 (Unicef Office of Research - Innocenti)
sinki Declaration and its later amendments. [14] United Nations 2018 Classification of individual
consumption according to purpose (COICOP) 2018
Division S; Report No. M99 (New York: Department of
ORCID iDs Economic and Social Affairs)
[15] Wardrop N A, Dzodzomenyo M, Aryeetey G, Hill A G,
Jim Allan Wright https://orcid.org/0000-0002- Bain R E S and Wright J 2017 Estimation of packaged
water consumption and associated plastic waste
8842-2181 production from household budget surveys Environ.
Simon Damkjaer https://orcid.org/0000-0002- Res. Lett. 12 074029
0758-6038 [16] Ghana Statistical Services, ICF 2020 Ghana Malaria Indicator
Heini Vaisanen https://orcid.org/0000-0002-5494- Survey 2019 (Accra: Ghana Statistical Services and ICF.)
[17] Stoler J 2012 Improved but unsustainable: accounting for
0415 sachet water in post-2015 goals for global safe water Trop.
Mawuli Dzodzomenyo https://orcid.org/0000- Med. Int. Health 17 1506–8
0002-4709-3182 [18] Robinson J and Howland O 2022 Visible and invisible risks:
Allan G Hill https://orcid.org/0000-0002-4418- exploring washing and hygiene practices with women living
on low income in Kenya Glob. Public Health 17 1002–15
0379 [19] Akinbami C A O and Momodu A S 2013 Health and
Lorna Grace Okotto https://orcid.org/0000-0003- environmental implications of rural female entrepreneurship
0109-3942 practices in Osun state Nigeria AMBIO 42 644–57
Joseph Okotto-Okotto https://orcid.org/0000- [20] MacArthur R, Teye E and Darkwa S 2021 Quality and safety
evaluation of important parameters in palm oil from major
0002-4656-8369 cities in Ghana Sci. Afr. 13 e00860
Peter Shaw https://orcid.org/0000-0003-0925- [21] Ghana Statistical Service 2019 Ghana Living Standard Survey
5010 (GLSS 7) 2017 (available at: https://www2.statsghana.gov.gh/
nada/index.php/catalog/97) (Accessed 1 December 2022)
[22] Kenya National Bureau of Statistics 2022 Kenya Integrated
References Household Budget Survey 2015-2016 (available at:
https://statistics.knbs.or.ke/nada/index.php/catalog/13)
[1] Lebreton L and Andrady A 2019 Future scenarios of global (Accessed 1 January 2022)
plastic waste generation and disposal Palgrave Commun. 5 6 [23] Ghana Statistical Services 2018 Ghana living standards survey
[2] Jambeck J R, Geyer R, Wilcox C, Siegler T R, Perryman M, round 7: poverty trends in Ghana, 2005-2017 (Ghana
Andrady A, Narayan R and Law K L 2015 Plastic waste Statistical Services)
inputs from land into the ocean Science 347 768–71 [24] Kenya National Bureau of Statistics. Basic report 2018
[3] Tiseo I 2021 Global plastic production 1950–2020 2015/16 Kenya Integrated Household Budget Survey (KIHBS)
(Hamburg: Statistica) (available at: www.statista.com/ (Nairobi)
statistics/282732/global-production-of-plastics-since-1950/) [25] StataCorp 2019 Stata statistical software 16 College Station
[4] Li W C, Tse H F and Fok L 2016 Plastic waste in the marine (Texas: Statacorp LLC)
environment: a review of sources, occurrence and effects Sci. [26] Becerril-Arreola R and Bucklin R E 2021 Beverage bottle
Total Environ. 566 333–49 capacity, packaging efficiency, and the potential for plastic
[5] Hurley R, Horton A, Lusher A and Nizzetto L 2020 Plastic waste reduction Sci. Rep. 11 3542
waste in the terrestrial environment Plastic Waste and [27] Sibanda L K, Obange N and Awuor F O 2017 Challenges of
Recycling ed T M Letcher (New York: Academic) ch 7, solid waste management in Kisumu, Kenya Urban Forum
pp 163–93 28 387–402
[6] Eriksen M, Lebreton L C M, Carson H S, Thiel M, [28] Muthuuri P, Kassim J, Richard P and Mwarania F 2021 Last
Moore C J, Borerro J C, Galgani F, Ryan P G and Reisser J mile access to enriched children’s complementary food:
2014 Plastic pollution in the world’s oceans: more than 5 mitigating malnutrition in Kenya Front. Public Health
trillion plastic pieces weighing over 250,000 tons afloat at sea 9 604864
PLoS One 9 e111913 [29] Valentin M 2010 Bouteilles et sachets en plastique: pratiques
[7] Wright S L, Thompson R C and Galloway T S 2013 The et impacts des modes de consommation d’eau a boire au
physical impacts of microplastics on marine organisms: a Senegal Autrepart 55 57–70
review Environ. Pollut. 178 483–92 [30] Stoler J, Weeks J R and Fink G 2012 Sachet drinking water in
[8] Schwabl P, Köppel S, Königshofer P, Bucsics T, Trauner M, Ghana’s Accra-Tema metropolitan area: past, present, and
Reiberger T and Liebmann B 2019 Detection of various future J. Water Sanit. Hyg. Dev. 2 223–40
microplastics in human stool Ann. Intern. Med. 171 453–7 [31] Amazonas M, Fisher G and Zhang W 2018 Research Shows
[9] Mukhtar E, Williams I and Shaw P 2018 Visibility of Which PET Water Bottle Design Attributes Impact Recycling
fundamental solid waste management factors in developing (Holland, OH: Plastic Technologies Inc.)
countries Detritus 1 162–73 [32] Wright J, Dzodzomenyo M, Fink G, Wardrop N A,
[10] Nagpure A S, Ramaswami A and Russell A 2015 Aryeetey G C, Adanu R M and Hill A G 2016 Subsidized
Characterizing the spatial and temporal patterns of open sachet water to reduce diarrheal disease in young children: a
12
Environ. Res. Lett. 17 (2022) 124029 J A Wright et al
feasibility study in Accra, Ghana Am. J. Trop. Med. Hyg. [41] Sohrabpour V, Hellstrom D and Jahre M 2012 Packaging in
95 239–46 developing countries: identifying supply chain
[33] Syrigos A and Adamides E 2022 Environmental and needs J. Humanit. Logist. Supply Chain Manage.
economic assessment of the PET bottles manufacturing 2 183–204
process: a case study Sustainable Design and Manufacturing [42] Shaw P J, Lyas J K and Hudson M D 2006 Quantitative
ed S Scholz, R Howlett and R Setchi (Singapore) pp 202–13 analysis of recyclable materials composition: tools to support
[34] European Council 2008 Directive 2008/98/EC of the decision making in kerbside recycling Resour. Conserv.
European parliament and of the council on waste and Recycl. 48 263–79
repealing certain directives, 2008/98/EC [43] Farfán G, Genoni M E and Vakis R 2017 You are what (and
[35] McDonald R I, Douglas I, Revenga C, Hale R, Grimm N where) you eat: capturing food away from home in welfare
Grönwall J and Fekete B 2011 Global urban growth and the measures Food Policy 72 146–56
geography of water availability, quality, and delivery Ambio [44] Mbiba B 2014 Urban solid waste characteristics and
40 437–46 household appetite for separation at source in Eastern and
[36] Smith L C and Subandoro A 2007 Measuring food security Southern Africa Habitat Int. 43 152–62
using household expenditure surveys (Washington, DC: [45] Perez-Heydrich C, Warren J L, Burgert C R and Emch M E
International Food Policy Research Institute) p 147 2013 Guidelines on the Use of DHS GPS Data Contract No.
[37] Smith L C, Dupriez O and Troubat N 2014 Assessment of the SAR8 (Calverton, MD: ICF International)
Reliability and Relevance of the Food Data Collected in [46] Oseni G, Durazo J and McGee K 2017 The Use of
National Household Consumption and Expenditure Surveys Non-Standard Units for the Colleciton of Food Quantity: A
(Washington, DC: International Household Survey Guidebook for the Measurement of Food Consumption and
Network) Agricultural Production in Living Standards Surveys
[38] Deaton A and Grosh M 2000 Consumption Designing (Washington, DC: World Bank)
Household Survey Questionnaires for Developing Countries: [47] Frankenberg E 2000 Community and price data Designing
Lessons from 15 Years of the Living Standards Measurement Household Survey Questionnaires for Developing Countries:
Survey ed M Grosh and P Glewwe (Washington, DC: World Lessons from 15 Years of the Living Standards Measurement
Bank) pp 91–134 Survey ed M Grosh and P Glewwe (Washington, DC: World
[39] Schundeln M 2018 Multiple visits and data quality in Bank) pp 315–38
household surveys Oxf. Bull. Econ. Stat. 80 380–405 [48] Miezah K, Obiri-Danso K, Kádár Z, Fei-Baffoe B and
[40] FAO, World Bank 2018 Food Data Collection in Household Mensah M Y 2015 Municipal solid waste characterization
Consumption and Expenditure Surveys: Guidance for Low and and quantification as a measure towards effective
Middle Income Countries (Rome: Food and Agriculture waste management in GhanaWaste Manage.
Organisation) 46 15–27
13