The Rich Get Gas, the Poor Go Back to Firewood: Companion Data and Workbook (Nigeria GHS-Panel, 2018-2019 to 2023-2024)

Description

This dataset is the companion computation workbook for the paper "The Rich Get Gas, the Poor Go Back to Firewood: Transition Incidence and the Liquidity of the Clean Energy Escape in Nigeria's 2023 Fuel Subsidy Reform." It reproduces every table and figure in the manuscript from the underlying weighted household counts of the Nigeria General Household Survey Panel (GHS-Panel, World Bank LSMS-ISA), Waves 4 (2018-2019) and 5 (2023-2024), with the pre-reform placebo drawing added on Wave 3 (2015-2016). The balanced analytic sample is 4,682 households, weighted to approximately 25.2 million. Each sheet corresponds to a numbered exhibit in the paper including the household energy portfolio shares (Table 2); the cooking fuel and main electricity transition matrices (Tables 3-4) with a cell by cell Reconciliation tab that ties the single wave margins to the matched pairs row totals; the two-speed income gradient and its cluster-bootstrap confidence intervals (Tables 5-6); pre-reform budget shares by decile (Table 7); the difference-in-differences pre-reform placebo (Table 8); the state-level dose-response (Table 9); the assembled state-mean petrol (PMS) price series and the Nigeria Longitudinal Phone Survey timing check (Figures 1 & 9); and the Leontief cost-push pass-through with the informal-transport stress test (Tables A1-A3). Every share is implemented as a live formula (weighted count ÷ weighted total), not manually entered, so any headline number can be traced back to the household flows that produced it; the central pump price shock (δ = 1.77) and every denominator are documented on the tabs. The workbook additionally embeds the 14 manuscript figures and editable Excel versions of the headline charts. This workbook documents the point estimates and their reconciliation. It does not contain the GHS-Panel microdata, which are publicly available from the World Bank Microdata Library (LSMS-ISA, NBS NADA), nor the inferential estimation code (EA and state-cluster bootstraps, inverse probability weighting, Lee bounds, and the Table 8-9 regressions), which is distributed as a separate replication script.

Steps to reproduce

1. Download the Nigeria GHS-Panel Waves 3 to 5 microdata from the World Bank Microdata Library (LSMS-ISA, NBS NADA). 2. Construct the balanced Wave 4 to Wave 5 panel from "sect11 plantingw4" (cooking & electricity, W4), "sect9 harvestw5" (cooking & electricity, W5), "sect8 plantingw5" (subsidy opinion, W5) and "totcons final" (consumption & weights, W4), matched on the household identifier; the three-wave placebo adds "sect11 plantingw3" and the Wave 3 consumption aggregate. 3. Classify primary cooking fuel (firewood, charcoal and other solid fuel = biomass; kerosene; LPG and other modern fuels = clean) and main electricity source (W4 coded variable; W5 free text field under the documented keyword rule). 4. Weight by "wt wave4"; compute weighted category counts and divide by the weighted total to obtain each share with the live formulas on each tab. 5. Verify against the workbook with each tab lists weighted counts, denominators and the resulting shares; the Reconciliation tab ties the Table 2 single wave margins to the Tables 3 to 4 matched pairs row totals cell by cell.

Institutions

• University of Bradford

  England, Bradford

Categories

Licence