Gold Price Shocks and Social Conflict

Causal evidence from the Peruvian Amazon

Causal Inference
Stata
Python
Mining
Spatial Econometrics
Causal analysis using Shift-Share IV to estimate how international gold price fluctuations impact social conflicts in Peru’s mining districts. Presented at LACEA 2024.
Author

Maykol Medrano

Published

October 24, 2025

NoteProject Status

This paper is currently under review. Preliminary version presented at LACEA 2024 Annual Meeting.

Overview

This project investigates the causal relationship between international gold price fluctuations and the incidence of social conflicts in mining districts of Madre de Dios, Peru.

Research Question

Does the boom in extractive industries fuel social unrest? We exploit exogenous variation in international gold prices combined with local mineral endowments to predict conflict probability at the district level.

Methodology

We employ a Shift-Share Instrumental Variable (SSIV) strategy to isolate the component of local economic growth driven purely by external price shocks.

Key equation:

\[ Z_{dt} = \text{Endowment}_d \times \log(\text{Price}_{t}) \]

Where \(Z_{dt}\) is the instrument for district \(d\) at time \(t\).

Data sources: - Ombudsman’s Office (Defensoría del Pueblo): Monthly panel of social conflicts - Sentinel-2 Satellite Imagery: Illegal mining deforestation measurement - Ministry of Economy (SIAF): Local canon transfers analysis

Key Findings

  • A 10% increase in gold prices leads to a 2.5 percentage point increase in the probability of socio-environmental conflict in the subsequent quarter
  • Effect driven primarily by competition for land rights rather than labor disputes
  • Impacts concentrated in districts with weak institutional presence

Implementation

* Define Shift-Share Instrument
gen shift_share_iv = gold_endowment * log(gold_price_int)

* Main Specification (Fixed Effects)
reghdfe conflict_dummy shift_share_iv canon_transfers ///
    population_density, ///
    absorb(district_id year_month) ///
    vce(cluster district_id)

* Export Results
outreg2 using "results/main_table.tex", replace tex
import geopandas as gpd
import matplotlib.pyplot as plt

# Load district data with conflict intensity
peru_map = gpd.read_file("data/peru_districts.geojson")

# Spatial visualization
fig, ax = plt.subplots(1, 1, figsize=(10, 10))
peru_map.plot(column='conflict_intensity',
              ax=ax,
              legend=True,
              cmap='OrRd',
              scheme='quantiles')

plt.title("Spatial Distribution of Social Conflicts (2023)")
plt.show()

Citation

@unpublished{medrano2025gold,
  title={Gold Price Shocks and Social Conflict: Causal Evidence from the Peruvian Amazon},
  author={Medrano, Maykol},
  year={2025},
  note={Under review. Presented at LACEA 2024},
  url={https://maykolmedrano.github.io/projects/gold-shocks.html}
}

APA Format: Medrano, M. (2025). Gold Price Shocks and Social Conflict: Causal Evidence from the Peruvian Amazon [Working Paper]. Presented at LACEA 2024 Annual Meeting. https://maykolmedrano.github.io/projects/gold-shocks.html


Full working paper and replication code available upon request.