🇹🇷 Turkish project report is available in docs/project_report.md.
This repository presents an econometric analysis of the dynamic relationships between consumer confidence and macroeconomic variables using a Vector Autoregression (VAR) framework.
The project investigates how macroeconomic shocks propagate through the system and how consumer confidence reacts to changes in key economic indicators.
- Project Overview
- Methodology
- Model Pipeline
- Analysis Workflow
- Key Findings
- Repository Structure
- Future Work
Consumer confidence is an important indicator reflecting households’ expectations about economic conditions. Understanding the drivers of consumer confidence provides valuable insights into economic sentiment and potential future consumption behavior.
This project analyzes the dynamic interactions between:
- Consumer Confidence Index (CCI / TGE)
- USD Exchange Rate
- Real Effective Exchange Rate
- Stock Market Index (BIST)
- Industrial Production Indicators
- Inflation
- Economic Confidence Indicators
The analysis focuses on identifying:
- how macroeconomic shocks influence consumer confidence
- how long these shocks persist
- the relative importance of different economic variables
To address these questions, a Vector Autoregression (VAR) model is estimated and evaluated using standard econometric diagnostics.
The analysis follows a standard econometric workflow used in macroeconomic time series modeling.
- Import and structure macroeconomic time series data
- Handle missing values
- Apply log transformations where appropriate
- Augmented Dickey-Fuller (ADF) tests
- Transformation using log-differences to ensure stationarity
- Lag length selection using information criteria
- Estimation of a Vector Autoregression (VAR) model
- Stability diagnostics
- Testing directional relationships between variables
- Identification of predictive relationships in the system
- Evaluating how shocks propagate through the system
- Measuring dynamic responses of consumer confidence to macroeconomic shocks
- Quantifying the contribution of each variable to forecast error variance
- Understanding the relative importance of different macroeconomic drivers
Model adequacy is evaluated using:
- Portmanteau (Whiteness) test for serial correlation
- Normality test of residuals
- Residual ACF inspection
- Q-Q plot analysis
These diagnostics help ensure that the model specification is statistically appropriate.
The econometric analysis follows a structured pipeline from raw macroeconomic data to economic interpretation.
Raw Macroeconomic Data
↓
Data Cleaning & Preparation
↓
Stationarity Testing (ADF)
↓
Log-Difference Transformation
↓
VAR Model Estimation
↓
Lag Selection & Stability Check
↓
Granger Causality Analysis
↓
Impulse Response Functions (IRF)
↓
Forecast Error Variance Decomposition (FEVD)
↓
Residual Diagnostics
↓
Economic Interpretation
The analytical workflow implemented in this repository follows a standard empirical macroeconometric research pipeline.
Data Collection
↓
Exploratory Data Analysis
↓
Stationarity Tests
↓
Model Specification
↓
VAR Estimation
↓
Dynamic Analysis (IRF & FEVD)
↓
Model Diagnostics
↓
Economic Interpretation
This workflow reflects a typical structure used in macroeconomic time series analysis and applied econometrics research.
The empirical results reveal several important insights:
-
Consumer confidence is primarily driven by its own shocks, explaining roughly 80% of the forecast error variance.
-
Exchange rate shocks contribute moderately, accounting for approximately 13% of consumer confidence variability.
-
Macroeconomic shocks dissipate relatively quickly, with impulse responses typically fading within 6–8 periods.
-
Residual diagnostic tests indicate that the VAR model provides a stable and statistically adequate representation of the system.
Overall, the findings suggest that consumer confidence is influenced by macroeconomic conditions but remains largely shaped by internal expectation dynamics.
financial_time_series_modernization/
data/
raw/ — original dataset
interim/ — intermediate processing
processed/ — cleaned or transformed data
notebooks/
01_var_analysis.ipynb — main econometric analysis notebook
docs/
project_report.md — detailed project report
requirements.txt — Python dependencies
.gitignore — excluded files
README.md — project documentation
Several extensions could further enhance the analysis.
Extending the dataset to include more recent observations would allow capturing new economic cycles and structural shifts.
Future work could test for long-run equilibrium relationships between variables using the Johansen cointegration test.
If cointegration is detected, the analysis could be extended using a Vector Error Correction Model (VECM).
Additional econometric approaches could be explored:
- Structural VAR (SVAR)
- Nonlinear time series models
- Regime-switching models
The VAR framework could be extended to perform out-of-sample forecasting of consumer confidence dynamics.
This project was developed as part of a learning and research effort in:
- Financial Time Series Analysis
- Econometrics
- Applied Data Science with Python