This project presents a global panel data analysis examining the relationship between real interest rates and unemployment across countries over multiple decades, with a particular focus on major economic crises.
Using fixed‑effects econometric techniques, the analysis finds that real interest rates—both contemporaneous and lagged—do not exhibit a statistically significant direct effect on unemployment once country and year heterogeneity are accounted for. Instead, unemployment dynamics are shown to be largely structural, driven by persistent country‑specific characteristics and global shocks rather than by monetary conditions alone.
Crisis‑focused models reveal that the COVID‑19 pandemic (2020) produced a clear and significant increase in unemployment worldwide, while the effects of the 2008 Global Financial Crisis were more complex and heterogeneous across countries. Interaction analysis further indicates that monetary policy effectiveness did not change materially during the Global Financial Crisis, suggesting that labor‑market disruptions during crises are driven primarily by non‑monetary factors.
Overall, the findings highlight the importance of institutional structure, labor‑market resilience, and crisis management policies in shaping employment outcomes, offering insights relevant for policymakers, economists, and applied data analysts.
``
This project investigates the relationship between real interest rates and unemployment using a global panel dataset covering multiple countries over several decades. The analysis focuses on three central research questions:
- How are real interest rates associated with unemployment across countries?
- Do monetary policy effects on unemployment occur with a delay?
- Do major global crises—specifically the 2008 Global Financial Crisis and the 2020 COVID‑19 pandemic—alter unemployment dynamics or monetary transmission?
To address these questions, the project combines descriptive analysis, fixed‑effects panel regressions, and crisis‑based robustness checks. The objective is not only to estimate statistical relationships, but also to develop an interpretable and policy‑relevant narrative around global labor markets.
macroeconomic-panel-analysis/
│
├── notebooks/
│ ├── 01_data_cleaning_and_merging.ipynb
│ └── 02_analysis_and_regression.ipynb
│
├── data/
│ └── processed/
│ └── final_panel_dataset.csv
│
├── figures/
│ └── (optional: saved plots)
│
├── README.md
└── requirements.txt
The analysis is organized into two Jupyter notebooks, each with a clear and distinct purpose.
1️⃣ Data Cleaning and Merging 📓 Notebook: 01_data_cleaning_and_merging.ipynb
This notebook covers:
Loading data from multiple macroeconomic sources Cleaning and standardizing unemployment, interest rate, GDP, and exchange rate data Reshaping datasets into a country‑year panel format Merging all datasets into a unified panel Exporting the final cleaned dataset used for all subsequent analysis
📦 Final cleaned dataset: data/processed/final_panel_dataset.csv
2️⃣ Analysis and Regression Results
📓 Notebook: 02_analysis_and_regression.ipynb
This notebook covers:
Exploratory data analysis and professional visualizations Baseline pooled OLS regression Fixed‑effects regression (country & year) Lagged interest rate robustness tests Crisis dummy models (2008 & 2020) Crisis interaction model Interpretation of results and final conclusions
data/processed/final_panel_dataset.csv
🧹 Data Preparation The project integrates multiple macroeconomic datasets, including:
- Unemployment rates
- Real interest rates
- GDP
- Country‑level indicators
Because these datasets originate from different sources and formats, extensive preprocessing was required:
- Unemployment data were transformed from wide to long (country‑year) format
- Real interest rate data were standardized and cleaned
- GDP data were reshaped from wide to long format
- Cross‑sectional country datasets were merged appropriately
The final result is a clean, harmonized country‑year panel dataset suitable for econometric analysis.
Exploratory analysis reveals strong heterogeneity in unemployment outcomes across countries and over time. Key findings include:
- Persistent structural differences in unemployment levels across countries
- Smooth global averages masking substantial country‑specific variation
- Clear labor market disruptions during major global shocks
Five professional visualizations support these insights:
- Global unemployment and real interest rate trends
- Interest rate–unemployment scatter with regression line
- Country‑level unemployment trajectories
- GDP versus unemployment relationship
- Unemployment distributions across crisis periods
These plots motivate the use of fixed‑effects and crisis‑focused econometric models.
Below are selected figures highlighting the main empirical patterns discussed in the analysis.
All figures are generated in 02_analysis_and_regression.ipynb.
This figure shows the evolution of average unemployment rates and real interest rates over time, with major global shocks highlighted.
A scatter plot with a fitted trend line illustrating the weak and heterogeneous relationship between real interest rates and unemployment across countries.
Distribution of unemployment rates before 2008, during the post‑2008 period, and after the COVID‑19 shock, highlighting structural breaks.
Correlation structure among the main variables used in regression analysis.
``
The empirical strategy follows a layered approach:
-
Baseline Models
-
Pooled OLS as a benchmark specification Two‑way fixed effects regression controlling for country and year heterogeneity
-
Robustness and Crisis Models
Lagged interest rate model to test delayed monetary transmission Crisis dummy models isolating the 2008 and 2020 shocks Crisis interaction model examining whether monetary transmission changes during crises
All regressions use cluster‑robust standard errors at the country level, consistent with best practices in applied panel data econometrics.
- Real interest rates do not exhibit a statistically significant direct or lagged effect on unemployment once country and year fixed effects are included.
- GDP is negatively associated with unemployment, though the effect is modest in magnitude.
- The COVID‑19 pandemic (2020) produced a strong and statistically significant increase in unemployment across countries.
- The 2008 Global Financial Crisis shows weaker and more complex labor‑market effects.
- Crisis interaction models indicate that monetary policy effectiveness did not change significantly during the Global Financial Crisis.
- Persistent country‑specific structural factors dominate unemployment dynamics.
- Overall, the findings indicate that global unemployment dynamics are driven primarily by structural country characteristics and major economic shocks, rather than by interest rate movements alone. While monetary policy remains an essential stabilization tool, its direct influence on unemployment appears indirect, heterogeneous, and context‑dependent in a global panel setting.
- The COVID‑19 pandemic stands out as a uniquely severe labor market shock, while the effects of the 2008 crisis appear to have propagated through more complex institutional and structural channels. These results highlight the importance of labor market resilience, institutional quality, and complementary fiscal and structural policies alongside monetary intervention.
- Python
- pandas
- numpy
- matplotlib
- seaborn
- statsmodels
- AWAIS PASHA
- Macroeconomic & Data Analysis Project
- Islamabad, Pakistan
This repository is intended for:
- Academic coursework
- Applied econometrics practice
- Portfolio demonstration
- Interview and review discussions