Overview
Consumption expenditures remain the prevailing measure of welfare in many low- and middle-income countries (LMICs). The insight they provide is used for official estimates of poverty and wellbeing, as well as to inform a wide range of policies and programs, therefore it is critical to ensure consumption expenditures are measured accurately.
However, the information needed to calculate household consumption expenditures is difficult to collect, and it is one of the reasons why we have focused on developing, testing and validating methods to accurately capture data on the expenditure of households on goods and services.
Work Areas
1. Food Consumption at Home
This area has been of particular interest for our team, methodological research is crucial due to the fact that food represents the majority of consumption expenditures in most LMICs. Hence, several designs decisions have been tested across multiple survey experiments.
Mode of Data Capture: Diary vs. Recall
One of the enduring debates when collecting data about food consumption at home is whether it is best to capture it through the administration of a diary – in which the respondent is responsible for recording their consumption - or through recall - where an interviewer asks the respondent about their food consumption within a particular period.
Our team has conducted several survey experiments comparing these two methods, the most prominent are two studies conducted in Tanzania almost a decade apart, in 2012 and 2022, as well as a similar study performed in Niger in 2017. These studies provided the basis for the official recommendations included in the United Nations-endorsed guidelines, stating that in low-income settings, the recall approach is generally preferable over diary methods.
Level of Respondent: Household vs. Individual
Another important design decision when collecting information about food consumption at home is the level at which it should be gathered. The traditional approach has been to ask one respondent to report – collectively - the consumption of all household members.
However, there are distinct analytical advantages to collecting this information at the individual level, as it provides a more holistic understanding of intra-household distribution of food and individual nutrition. The LSMS team tested the implication and feasibility of collecting individual consumption information in the 2012 experiment in Tanzania, as well as in a recent methodological study conducted in Bangladesh.
Length of Reference Period
The period for which information on food consumption at home is asked and recorded also has substantial implications on the accuracy of the data collected. In the context of the recall approach, longer reference periods can lead to bias through “telescoping” of reported consumption outside of the reference period or omitting some consumption within. For diaries, a longer reference period (when the diary is recorded) often results in decreasing consumption over time because respondents become less diligent in filling the diary as time goes on.
The impacts of the reference period used to collect food consumption at home have been tested by our team through the experiments in Tanzania and Niger referred to above. The results of these studies have also informed the official guidelines endorsed by UNStats that promote a reference period of 7-days for recall and no more than 14-days for diary.
Length of Item List
Capturing information on food consumption at home, particularly through recall, involves iteration through a long list of food items. The length of this list has important tradeoffs: more items allow the collection of more detailed, holistic and disaggregated information, but this could also result in respondent fatigue and, as a consequence, in food consumption underreporting.
A careful balance is required to establish a list that is comprehensive enough to produce meaningful and reflective estimates of food consumption without overburdening respondents. The experiments conducted by our team in Tanzania and Niger explored variations in the list of food items.
2. Food Away from Home (FAFH)
Household expenditures on food consumed away from home, such as restaurants, bars and cafeterias, is a growing component of expenditures in many LMICs, but collecting this information presents several methodological challenges.
The level of data collection for FAFH consumption is important given that it is usually an individual activity, not a collective one, so asking a single member to report FAFH consumption for all the household members, may not be reliable. The disaggregation of FAFH items is also problematic, since it is difficult for respondents to report specific food items for meals they did not prepare themselves, which is not the case when the cooking is done at home.
Due to the importance of these aspects in the design of a survey questionnaire, our team conducted studies in Peru and Vietnam. These findings also informed the recommendations included in the United Nations-endorsed guidelines for consumption measurement.
3. Nonstandard Units
When collecting information on food consumption it is important to capture the quantity of food items that are consumed. This allows the assessment of prices or unit values, as well as the intake of calories and key nutrients. However, collecting information on food quantities in many LMICs is challenging due to the prevalence and use of nonstandard units - such as cups or pieces – that are difficult to collect and analyze.
Collection is challenging because these units are not always the same and can vary substantially across regions, and even from person to person. At the analysis stage, quantities in nonstandard units must be converted to standard units to aggregate or apply other metrics, like calories, therefore, conversion factors are needed.
Our team has developed comprehensive guidelines for survey practitioners to effectively collect quantities in nonstandard units and also to gather conversion factors to analyze the units.
4. Nonfood Expenditure: Education
Education expenditure is an important non-food component of the consumption aggregates – the total amount spent by households’ members on goods and services. In multi-topic household surveys that usually have a dedicated module for collecting education information at the individual level, decision has to be made whether to collect education expenditure data as part of the dedicated individual-level education module or collect them at the household-level under the non-food expenditure module.
The LSMS team and the UNESCO Institute for Statistics have prepared an official guideline for collecting education expenditure data in household surveys.