Abstract

In a policy environment seeking to ensure sustainable value chains and due diligence in agrifood systems, the concept of a living income has emerged as a benchmark for assessing the welfare of smallholder producers. Unlike monetary poverty lines, which are anchored on the definition of a subsistence threshold for attaining a basic basket of consumption goods, the living income concept extends the (Deaton & Zaidi, 2002) approach to estimate the cost of attaining the minimum requirements of a decent life, notably: a nutritious yet affordable diet; durable, healthy housing; full education for all children; adequate healthcare to treat main illnesses; and savings to face shocks and finance future investments. This concept has gained traction among public and private sector actors in global agricultural value chains using it to guide policies and interventions for enhancing economic and social sustainability. However, methodological inconsistencies raise questions about its reliability for welfare analysis and suitability for rural agricultural contexts.

This paper critically examines the validity. consistency and policy relevance of prevailing living income methods and indicators. We present the first systematic assessment of documented approaches for estimating the decent life cost of living threshold, the living income benchmark (LIB), mapping differences in data sources and computational assumptions. We additionally evaluate methods for estimating the living income gap (LIG) indicator – the difference between actual incomes and the LIB – using simulations to assess their performance against established poverty metrics axioms. Finally, we seek to derive conclusions about the utility and operability of this approach for policies targeting small-scale agrifood systems producers. Our empirical application focuses on cocoa producers in Cameroon, where export-oriented cocoa production is central to livelihoods but marked by persistent poverty and inequality.

Primary cost-of-living data were collected from 203 sectoral service providers in the Centre region of Cameroon, interviewing food vendors, schools, healthcare facilities and housing experts. LIBs were calculated according to (Anker & Anker, 2017) and (Van De Ven et al., 2021), the two main documented living income methods, and sensitivity analysis conducted on equivalence scales, food prices, housing data and assumptions underlying non-food, non-housing expenditure.
Household income data was obtained from a representative survey of 4,294 cocoa-producing households across seven regions of Cameroon, enabling estimation of the LIG. We assessed the assumptions underlying existing uncensored LIG indicators against a newly proposed censored indicator. Statistical validation was based on simulations illustrating alignment with the main axioms from the poverty metrics literature (Ravallion, 2016).

We find significant methodological differences in LIB approaches, with 25 percent variation across the two LIBs. This variation is attributed to the use of equivalence scales for estimating food costs; differing assumptions for estimating non-food, non-housing costs; and the role of seasonal fluctuations in food prices. These raise concerns about the comparability of LIB estimates across methods and contexts.

For the LIG, our proposed censored indicator performs more robustly than existing uncensored versions. It satisfies the focus, monotonicity, and transfer axioms, and is distribution sensitive, rendering it suitable for monitoring purposes. The performance of the uncensored indicators is weaker, meeting only two of the four tested axioms, and underestimating the magnitude of income shortfalls.

In a policy environment seeking to ensure sustainable value chains and due diligence of actors in agrifood systems, selecting the appropriate LIG indicators requires careful consideration of the interventions or policies it seeks to inform. Our findings demonstrate the living income approach could effectively reorient the welfare narrative from subsistence to decency in shaping equitable agrifood systems policies. The decomposability of the LIB into food, housing, education, and healthcare components provides an opportunity to assess the affordability of essential goods and services for targeted groups. Statistical validation of the LIG highlights the potential for the censored indicator to reliably monitor deprivation. However, methodological inconsistencies across existing approaches compromise comparability and could create or exacerbate market distortions or lead to misalignment of policies and interventions with desired outcomes. Selecting the appropriate LIG indicator requires careful consideration of the interventions or policies it seeks to inform.

Policy implications are threefold. First, harmonization of LIB estimation methods is critical. We thus offer recommendations for aligning its data requirements with institutionalized data collection and statistical frameworks. Second, selection of LIG indicators must align with policy objectives: censored indicators are better suited for monitoring deprivation among target groups, particularly when using such indicators to estimate and operationalize fair pricing schemes and other sectoral tools aiming to reduce the LIG. Third, policies to bridge or close income gaps must be multifaceted and consider approaches to raise incomes as much as addressing the affordability of goods and services. For this, embedding living income approaches within broader rural development and transformation strategies is critical.