GitHunt
LE

lelonxumalo/DSGE-Modeling-of-Social-Insurance-in-Turkey

A Dynamic Stochastic General Equilibrium (DSGE) model analyzing social protection policies in Turkey using a Two-Agent New Keynesian (TANK) framework with dual labor markets (formal and informal sectors).

dsgeinformalityunemploymentvat

DSGE Modeling of Social Insurance in Turkey

A Dynamic Stochastic General Equilibrium (DSGE) model analyzing social protection policies in Turkey using a Two-Agent New Keynesian (TANK) framework with dual labor markets (formal and informal sectors). See writeup on my blog

Overview

This project implements a computational model to study the impact of social insurance expansions on Turkey's labor market, focusing on the trade-offs between:

  • Formal employment (regulated, taxed, with search frictions)
  • Informal employment (unregulated, untaxed, flexible)
  • Unemployment and social protection
  • Fiscal sustainability and policy financing methods

Features

  • Dual Labor Market: Formal sector with search-and-matching frictions vs. informal sector with flexible wages
  • Household Heterogeneity: Savers (capital owners) vs. Spenders (liquidity-constrained)
  • Fiscal Policy: Labor taxes, unemployment benefits, and various financing mechanisms
  • Turkish Calibration: Parameters calibrated to Turkey's economic context (2024-2025)

Scripts

1. 1_turkey_tank.py - Baseline Model

The core TANK-SM model calibrated for Turkey.

Key Features:

  • Formal sector with Nash bargaining and vacancy posting
  • Informal sector as "survival" option
  • 35% labor tax rate and unemployment benefits
  • Policy experiment: +33% increase in unemployment benefits

Output: results/turkey_policy_impact_dual_axis.png

Run:

python 1_turkey_tank.py

2. 2_financing_experiments.py - Fiscal Closure Comparison

Compares different methods of financing social protection expansion.

Scenarios:

  • Baseline: Current policy
  • Debt-financed: Benefits funded via lump-sum taxes on savers
  • Labor tax-financed: Benefits funded by raising formal sector labor taxes

Analysis: Shows who bears the cost and the differential impact on unemployment

Output: results/turkey_tax_experiment.png

Run:

python 2_financing_experiments.py

3. 3_turkey_tank_laffer.py - Laffer Curve Analysis

Traces the relationship between labor tax rates (0-70%) and government revenue.

Key Questions:

  • What is the revenue-maximizing tax rate?
  • At what point does the formal sector collapse?
  • How does informality grow with tax rates?

Output: results/turkey_laffer_curve.png

Run:

python 3_turkey_tank_laffer.py

4. 4_turkey_tank_vat_experiment.py - VAT vs. Labor Tax

Tests whether consumption taxes (VAT) are less distortive than labor taxes for funding social protection.

Hypothesis: VAT is broader and doesn't directly affect the formal sector hiring decision.

Output: results/turkey_vat_experiment.png

Run:

python 4_turkey_tank_vat_experiment.py

Installation

Requirements

  • Python 3.8+
  • GAMS (for GAMSPy backend)

Setup

  1. Clone this repository:
git clone https://github.com/lelonxumalo/DSGE-Modeling-of-Social-Insurance-in-Turkey.git
cd DSGE-Modeling-of-Social-Insurance-in-Turkey
  1. Install Python dependencies:
pip install -r requirements.txt
  1. Install GAMS:
    • Download from GAMS website
    • Follow platform-specific installation instructions
    • Ensure GAMS is in your system PATH

Model Structure

Households

  • Savers (50%): Own capital, work in formal sector, smooth consumption over time
  • Spenders (50%): Hand-to-mouth, work in both formal and informal sectors

Production

  • Formal Sector: Capital-intensive (Cobb-Douglas), subject to labor taxes
  • Informal Sector: Labor-only, lower productivity, untaxed

Labor Market

  • Formal: Search-and-matching with job separation rate of 7%
  • Informal: Walrasian spot market (flexible wages)
  • Unemployment: Benefits provided at 30% of wage (baseline)

Government

  • Revenue: Labor income taxes (35% baseline)
  • Expenditure: Unemployment benefits
  • Closure: Balanced budget via lump-sum adjustments

Key Parameters (Turkey Calibration)

Parameter Value Description
beta 0.985 Discount factor (reflects high real interest rates)
alpha 0.40 Capital share (formal sector)
tau_w 0.35 Labor tax rate (OECD tax wedge)
p_sep_rate 0.07 Formal job separation rate (high turnover)
p_match_eff 0.35 Matching efficiency (structural unemployment)
p_informal_prod 1.2 Informal sector productivity
p_ben_level 0.3 Unemployment benefit replacement rate

Results

All visualizations are saved in the results/ folder:

  • Labor market dynamics (unemployment, informality)
  • Welfare impacts (consumption changes)
  • Fiscal trade-offs (revenue vs. costs)
  • Policy comparison charts

Research Questions Addressed

  1. Does expanding social protection help or hurt the poor?

    • Trade-off: Higher benefits vs. job destruction

  2. Who should pay for social insurance?

    • Rich (debt/lump-sum) vs. workers (labor taxes)

  3. Is there a fiscal limit to labor taxation?

    • Laffer curve analysis

  4. Can VAT reduce labor market distortions?

    • Consumption tax as alternative financing

Citation

If you use this model in your research, please cite:

Nxumalo, Mpumelelo. 2025. Fiscal Limits of Social Protection in Turkey: A General Equilibrium Analysis. https://lelonxumalo.blogspot.com/2026/01/fiscal-limits-of-social-protection-in.html

License

MIT

Acknowledgments

This model builds on the TANK framework and incorporates insights from Turkish labor market research.

Languages

Python100.0%

Contributors

LEMN
MIT License
Created January 3, 2026
Updated February 2, 2026
lelonxumalo/DSGE-Modeling-of-Social-Insurance-in-Turkey | GitHunt