Panel-data analysis of food prices and regional economic activity in Kenya
WFP food-price data | VIIRS nighttime lights | STATA panel econometrics
| CLIENT Seema Ahad | SOFTWARE STATA | OUTCOME Defensible analysis chapter |
βBoth data-layer problems I raised last time are now fixed, the complete Stata log reconciles with the tables, and the discussion leads with the right result. This is a methodologically defensible chapter now.β
π§ 24Γ7 Support
Project at a glance
The assignment required a complete STATA-based panel-data workflow combining World Food Programme food-price data with VIIRS nighttime-lights information to examine how changing food prices relate to regional economic activity. The work moved beyond running commands: the final analytical chapter had to be internally consistent, reproducible, statistically defensible and clear enough for academic review.
| 01 Brief | 02 Data | 03 Diagnostics | 04 Panel Models | 05 Revision |
A Multi-Source STATA Assignment, Not Just a Regression
The client brief was explicit about the analytical pipeline. WFP food-price records had to be imported, merged with VIIRS nighttime-lights data, cleaned into a region-by-time panel, transformed into meaningful analytical variables and then tested using descriptive, correlation and panel-regression methods. The requested deliverable was a clean analytical dataset plus interpretable statistical outputs, not a collection of disconnected STATA screenshots.
Core analytical requirements
β’ Import and harmonise World Food Programme food-price data.
β’ Merge food prices with VIIRS nighttime-lights observations.
β’ Structure the data as a regional panel across time.
β’ Generate food-price and economic-activity variables.
β’ Produce summary statistics, trends and correlation diagnostics.
β’ Estimate fixed-effects models and robustness specifications.
β’ Validate the model using statistical significance and fit diagnostics
Why the brief was analytically demanding
The challenge was not simply choosing between fixed and random effects. The analysis combined datasets created for different purposes, used nighttime lights as a proxy for economic activity, contained multiple food commodities and required inference that remained credible after accounting for unobserved regional differences and within-group dependence. Any mismatch between the cleaned data, STATA log, tables and written interpretation could undermine the chapter even if individual commands were technically correct.
Data challenge
| β’ Different sources, units and time coverage. β’ Panel identifiers and time structure had to reconcile. β’ Commodity heterogeneity could distort simple comparisons. |
Method challenge
β’ Bivariate patterns could differ from panel-model estimates.
β’ FE and RE conclusions needed transparent model diagnostics.
β’ Robust or clustered standard errors changed inferential strength.
Building a Dataset the Econometric Model Could Trust
The first priority was data integrity. The assignment specification called for food-price information from the World Food Programme and nighttime-light observations from VIIRS, with the cleaned records organised into a regional panel. The supplied analytical chapter also included rainfall as a contextual control and retained multiple staple-food categories so that the analysis reflected the structure of the underlying market rather than a single commodity series.
The analytical dataset in the supplied chapter
Evidence from the supplied analysis Why it mattered
131 analytical observations Enough variation was retained to support panel estimation and diagnostics.
Mean nighttime-light value: 3.543 Provided the economic-activity proxy used as the outcome measure.
Average food price: USD 0.822 Captured broad variation in food-market conditions.
White maize: 39.69% of observations Added climatic context without creating severe multicollinearity.
Dry beans: 21.37% of observations Showed that commodity coverage was uneven and needed careful interpretation.
The analytical dataset in the supplied chapter
The dataset was analysed using STATA 19.5 to examine the relationship between average food prices, rainfall, and mean night-time light (NTL). The descriptive statistics showed 168 observations, covering the period from 2014 to 2025. The average food price was approximately KES 102.76, while average rainfall was 869.54 mm and the mean NTL value was 1.94.
Diagnostics before model interpretation
The correlation matrix showed only modest pairwise relationships among the key variables, while the VIF test returned values of 1.01 for both average food price and rainfall. This mattered because it established that unstable coefficients were unlikely to be caused by severe multicollinearity. More importantly, the analysis did not treat correlation as the final answer; the panel models were used to account for regional heterogeneity and time effects that simple associations cannot absorb.
1. Describe
Summary statistics and commodity coverage
2. Diagnose
Correlation matrix and VIF
3. Estimate
Fixed-effects and random-effects models
4. Validate
Hausman test, robust inference and interpretation
Moving from Descriptive Patterns to Panel-Data Evidence
The supplied analysis used both fixed-effects (FE) and random-effects (RE) regression as part of a deliberate model-comparison strategy. FE was used to absorb time-invariant regional characteristics, while RE provided a benchmark specification. The Hausman test then assessed whether the unobserved regional component was correlated with the explanatory variables. In the supplied chapter, the Hausman statistic was strongly significant (ΟΒ² = 83.02, p < 0.001), supporting the use of fixed effects in that specification.
REVISION & QUALITY CONTROL
The Revision That Turned the Chapter into Defensible Research
The strongest evidence of the projectβs success came after revision. The client/reviewer did not simply say that the chapter βlooked betterβ; the feedback identified specific methodological improvements. Data-layer problems had been corrected, the complete STATA log reconciled with the reported tables, and the discussion now led with the correct result. Just as importantly, the reviewer stated that the remaining work was interpretation and polish rather than a rerun of the core analysis.
STATA Analysis and Key Outcomes
The correlation analysis indicated a moderate negative relationship between average food prices and rainfall (r = β0.5302, p < 0.001), suggesting that higher rainfall levels were generally associated with lower food prices. Mean NTL also showed a negative correlation with food prices (r = β0.2059), whereas rainfall and mean NTL were positively correlated (r = 0.2494).
The multiple regression model was statistically significant (F(2,165) = 6.18, p = 0.0026), although its explanatory power was relatively modest (RΒ² = 0.0697; Adjusted RΒ² = 0.0585). Rainfall demonstrated a positive and statistically significant effect on mean NTL (Ξ² = 0.00151, p = 0.029), whereas average food price had a negative but statistically insignificant effect (Ξ² = β0.0193, p = 0.248). Overall, the STATA analysis demonstrates how statistical modelling can be used to identify, quantify, and interpret relationships within panel-based socioeconomic and environmental data.
Clients Review
Seema Ahad
“Thank you for the revision. This is a significant and successful improvement over the previous draft. Both of the data-layer issues I raised earlier have now been addressed, the complete STATA log is consistent with the reported tables, and the discussion now focuses on the correct result. Overall, the chapter is methodologically much stronger and more defensible”.
