The challenges and solutions for measuring agricultural labor in household surveys

Today, agriculture remains the backbone of the economies of low-income countries around the world where most of the people live below the international poverty line. The agricultural labor is therefore a key asset, making knowledge about labor productivity a critical component of any poverty reduction strategy.

Measuring work on household farms accurately is critical to the development of effective agricultural policies.  However, experts often disagree on the best-practices for capturing data on household farms’ work in household and agricultural surveys. The menu of available methods is diverse, varying in length of the reference period, the selection of respondents, the sequencing of questions, the seasonal timing of interviews, and even the appropriate granularity of units in which labor is reported. The selection of the appropriate method depends on multiple factors and must be tailored to the survey objectives and the country’s agricultural labor data needs, while bearing in mind other potential users of the data.

The World Bank’s Living Standards Measurement Study (LSMS) team developed a Guidebook on Measuring Agricultural Labor, a reference guide for survey practitioners on measuring labor on farming households in household and agricultural surveys. The methods and recommendations presented in this Guidebook are relevant for integrated and multi-topic household surveys, other smaller-scale household surveys, agricultural sample surveys, and agricultural censuses. It builds on an established set of activities performed by the LSMS team, including previously published guidelines and methodological work conducted in Tanzania, Ghana, and Malawi starting in 2014.

The best-practice survey design choices for agricultural labor modules to measure work on household farms are summarized below.

• The reference period. The reference period for collecting agricultural labor data should be aligned with the reference period for which information on agricultural outputs and non-labor inputs are collected. The reference period should cover at least one entire agricultural season. Where countries and regions experiencing more than one agricultural season in a year, data would ideally be collected separately for each agricultural season, according to the survey objectives.
• The timing and frequency of visits. When feasible, households should be visited twice at different times of an agricultural season, once after planting and once after harvesting. If households can only be visited once, they should be visited after harvesting.
• Household labor on crop production activities. It is recommended that the information on agricultural labor is collected at plot level for each member of the household that worked on any household-listed plot. Any detailed analysis related to agricultural productivity will require information on labor input to be collected separately for each plot. The total number of days worked on the plot and the typical or average number of hours worked per day by each member during the reference period should be gathered without reference to any specific activity. For any given plot, the respondent should be the plot manager.
• Household labor on post-harvest activities. Information on household labor carrying out post-harvest activities should be collected at the crop level for each household member that worked on post-harvest operations for any crop harvested by the household. The respondent for these questions should be the most knowledgeable person on the crop disposition by the household or the household’s farm activities. The total number of days worked on the crop by each member during the reference period should be collected, followed by a question on the typical or average number of hours worked per day for all activities, without reference to any specific activity.
• Hired, free or exchange labor. Hired, free or exchange labor information should be collected at the person-type-plot level, disaggregated by men, women and children. For any given plot, the respondent should be the plot manager. Data should be collected on the total number of people by person-type per plot, the average number of days a typical person-type worked on each plot during the reference period, and the typical or average number of hours worked per day. While wage is not collected for household labor and free or exchange labor, it should be collected for hired labor.

Resources:

• Read the full guidebook here.