Technical Report: Iraq MPCA Vulnerability Model Review 2021 (December 2021)


Executive Summary

A new socio-economic vulnerability assessment tool (SEVAT) used to target IDP (living out-ofcamp), returnee, and host community households eligible for multi-purpose cash assistance (MPCA) in Iraq was developed in 2021. Actors delivering MPCA in Iraq have used a harmonised tool to identify households since 2016 and the previous revision took place in 2019. The vulnerability model review process was led by the Cash and Livelihoods Consortium for Iraq (CLCI)1 , REACH Initiative and the Iraq Cash Working Group (CWG) with the support of a vulnerability model task force composed of programmatic and technical reviewers.

The 2021 revision of the SEVAT was prompted by the contextual shifts that have taken place in Iraq since the creation of the 2019 models and the availability of a new and fit-for-purpose dataset. The notable contextual shifts include large-scale returns of displaced households since 2019, the onset of the COVID-19 pandemic and related containment measures in early 2020 and the devaluation of the Iraqi Dinar in late 2020 with resulting macro-economic and inflationary consequences. The REACH Initiative, in coordination with OCHA and the ICCG and with the support of 20 INGO and NNGO partners, conducted a Multi-Cluster Needs Assessment (MCNA) in August 2021 to establish a comprehensive evidence base on the prevalence and type of vulnerabilities to inform Iraq’s Humanitarian Needs Overview (HNO), which contained sufficient coverage and relevant variables to carry out a model review.2

This technical report describes all steps and considerations taken during the development ofthree regional vulnerability models, including data collection, variable screening, variable selection process, thresholds, and inclusion and exclusion errors. Consistent with the previous model, the 2021 vulnerability model uses a proxy-means test (PMT) method and consumption as the indicator for vulnerability.

Proxy-means testing is a method of predicting or estimating a household’s consumption based on observable or accessible household characteristics and behaviours. The PMT method is employed when reliable and accessible data on socio-economic status (such as tax returns, pay slips or bank account balances) are unavailable or difficult to obtain, such is the case among conflict-affected populations in Iraq. Using detailed survey data, proxy means testing relies on multivariate regression analysis to generate a formula for estimating household consumption using proxy indicators, such as housing type or negative coping strategies. Households whose predicted consumption falls below a certain defined threshold are determined likely to be vulnerable and are therefore eligible to receive MPCA.

The 2021 SEVAT revision resulted in the creation of three regional models containing: a constant: the region’s dependent variable of the consumption baseline in the form of a log10 value, variables and their coefficients; a selection of significantly correlated independent variables and corresponding coefficients, the explanatory power: R-squared and adjusted R-squared values to indicate each model’s ability to account for variance in consumption on the basis of its independent variables, and error rates: inclusion and exclusion errors calculated on the basis of