Since the monthly counting of the Geopolitical Risk Index (GPR) started in 1985, the index peaked in ************, immediately after the 9/11 terrorist attack on the World Trade Center and Pentagon in the United States. The attack is perceived to be the deadliest terrorist attack in the 20th and 21st century, and ultimately caused the start of the so-called war on terror, with U.S. invasions of Afghanistan (2001) and Iraq (2003) following in the aftermath. Russia-Ukraine war The GPR was also high in ********** following Russia's invasion of Ukraine at the end of February that year. The attack on an independent state meant that the relations between Russia and the West reached a new low after the collapse of the Soviet Union, and several sanctions were imposed on Russia. 1991: a turbulent year Apart from the 9/11 attacks in 2001, the index reached its highest level in ************. This was a result of the ongoing Gulf War following Iraq's invasion of Kuwait, but also Soviet troops storming the Lithuanian capital in order to stop the country's secession from the Soviet Union. Additionally, a massacre of Tutsi in Rwanda highlighted the growing tensions in the East African country, which ultimately resulted in the genocide in 1994.

OverviewThis archive contains the files to reproduce the results in "Measuring Geopolitical Risk" as well as any additional documentation referred in the paper. Each directory is self-contained. For each directory, download all the files in order to run the necessary scripts. Instructions are given in the README files.Updated data can be found on the geopolitical risk index webpage, which can be found at the following url: https://www.matteoiacoviello.com/gpr.htm For questions or comments, please contact iacoviel@gmail.comData Availability StatementAll the data used in this paper are provided in this repository, with the exception of the Compustat quarterly firm-level data, which can be downloaded from https://wrds-www.wharton.upenn.edu/pages/ with a registered account.Software used The codes here run and have been tested either on Stata/MP 16.0 (for *.do files), on Matlab R2019/A (for *.m files), on R Version 4.04 (for *.R files), and on Anaconda 3 (for *.py, *.ipynb files). Most codes run in seconds/minutes on a personal laptop with 16GB ram, with the exception of the R code to estimate disaster episodes, which takes about 2 days using the standard settings from the Nakamura et al (2013) paper (nIter = 50,000, nRuns = 40). Directory list and list of main input files - if any - in each directory1. Monthly Geopolitical Risk Data Used in the Paper (data_paper)See README.txt file in the directory for detailsdata_gpr_export.dta (Stata format)data_gpr_export.xls (Excel format)2. Replication of Section I: Tables 1-2, Figures 1-8, Appendix Tables A.3-A.6, and Appendix Figures A.1-A.4 and A.10-A.14 (figures_paper) (requires Stata)See README.txt in the directory for detailsinput file: run_figures_tables.do3. Replication of Section III : VAR Evidence - Figures 9-10 and Appendix Figures A.5-A.7 (var_results)(requires Matlab)See README.txt in the directory for detailsinput file: run_all.m4. Replication of Section IV : Country-Specific GPR and Disaster Probability and Quantile Regressions - Tables 3-4 (disaster_regressions)(requires Stata)See README.txt in the directory for detailsinput file: run_replication_country_gpr.do5. Replication of Section V : Firm-Specific Geopolitical Risk - Table 5, Figure 11, Appendix Table A.7, and Appendix Figure A.9 (firm_regressions)(requires Stata)See README.txt file in the directory for details.input file: run_replication_firm_shuffled.do(Note that replication of the results here requires downloading firm-level balance sheet data through Compustat/WRDS. See firm_documentation below for instructions on how to build the firm_level.dta file)6. Auxiliary Material (Section V): Construction of Industry-Specific Exposure to Geopolitical Risk - Appendix Figure A.8 (industry_regressions)(requires Stata)See README.txt file in the directory for details.input file: run_replication_industry.do7. Auxiliary Material: Documentation on how to Build the firm_level.dta file (firm_documentation)See README_BUILD.txt file in the directory for details.8. Auxiliary Material (Section II): Tabulations of Daily Narrative GPR Data from The New York Times (narrative_index)See README.txt file in the directory for details.9. Appendix: Details on the Construction of the Human GPR Index (human_index)See README.txt file in the directory for details.10. Appendix: Audit of Articles Belonging to the GPR Index Described in Appendix Table A.3 (audit_coded)See README.txt file in the directory for details.11. Appendix: Granger Causality Tests --- Appendix Table A.8 (granger_causality)(requires Stata)See README.txt file in the directory for details.input file: run_granger_test.do12. Appendix: Replication of Textual Analysis in Appendix Tables A.1 and A.2 (text_analysis)(requires Matlab, including text analytics toolbox, and Stata for generating the formatted tables in the appendix)See README.txt file in the directory for details.input files: run_find_grams_textanalytics.m and run_app_tables_1_2.do 13. Auxiliary Material: Estimation of the Country Disaster Events from 1900 through 2019 (disaster_estimation)(requires R)See README.txt file in the directory for details.14. Auxiliary Material: Stata File with Firm-Level Geopolitical Risk Data (firm_level_gpr)See README.txt file in the directory for details.15. Auxiliary Material: Search Queries for News-Based GPR Index (news_searches)See README.txt file in the directory

Over the period observed, the Geopolitical Risk Index (GPR) in Israel reached its highest point in ************. This was due to the outbreak of the Israel-Hamas war on October 7th that year. Another notable spike occurred in January 1991, during the Gulf War, at the start of Operation 'Desert Storm' and the Iraqi missile campaign against Israel.

Key information about Australia Geopolitical Risk Index

• Australia Geopolitical Risk Index was reported at 0.165 % in Feb 2025.
• This records an increase from the previous number of 0.082 % for Jan 2025.
• Australia Geopolitical Risk Index data is updated monthly, averaging 0.069 % from Jan 1985 to Feb 2025, with 482 observations.
• The data reached an all-time high of 0.515 % in Mar 2022 and a record low of 0.005 % in Feb 1989.
• Australia Geopolitical Risk Index data remains active status in CEIC and is reported by Dario Caldara and Matteo Iacoviello.
• The data is categorized under World Trend Plus’s Geopolitical Risk Index – Table: Geopolitical Risk Index: By Country.

Listed European real estate stocks - as represented in the GPR *** Europe Index - dropped significantly in value at the beginning of 2020, followed by a period of recovery until early 2022. In the first half of the year, the sector weakened again, with the aggregate market capitalization of real estate in Sweden plummeting between November 2021 and September 2022. The GPR *** is a return index that covers ** real estate companies in Europe. Examples are Germany's Vonovia, Unibail-Rodamco-Westfield from France and Segro from the United Kingdom. These companies own and develop different types of real estate, and saw their stocks decrease significantly as the coronavirus spread to Europe and some governments decided to halt real estate construction.

NARDL estimation results for the effect of domestic (country-specific) GPR and US GPR on stock market returns and volatility.

Graph and download economic data for Global Economic Policy Uncertainty Index: Current Price Adjusted GDP (GEPUCURRENT) from Jan 1997 to Jul 2025 about uncertainty, adjusted, GDP, indexes, and price.

Graph and download economic data for Economic Policy Uncertainty Index for United States (USEPUINDXD) from 1985-01-01 to 2025-09-07 about academic data, uncertainty, indexes, and USA.

This study employs monthly data from January 2000 to October 2022. The estimation sample is dictated by the availability of energy-related uncertainty variables. The core variable of the model is a cryptocurrency environmental attention index (ICEA). This index is calculated by Wang et al. (2022).The index of global geopolitical risks (GPR), developed by Caldara and Iacoviello (2022), is another important variable in the model. Furthermore, the financial stress index (FSI), developed by Office of Financial Research, is utilized. Finally, the energy-related uncertainty index (EU), developed by Dang et al. (2023)

The datasets include standard precipitation index (SPI) at 1-month, 3-month, 6-month, 9-month and 12-month scales over the main terrestrial lands of China from January 1984 to December 2020. The SPI datasets were produced by blending the information from meteorological stations, and precipitation products, as well as topographical and geographical variables based on Gaussian process regression (GPR) models.

The meteorological station data are from the China Meteorological Data Service Centre. Five precipitation products are used: (1) CHIRPS Daily: Climate Hazards Group InfraRed Precipitation With Station Data (Version 2.0 Final); (3) ERA5-Land Monthly Averaged by Hour of Day - ECMWF Climate Reanalysis; (3) FLDAS: Famine Early Warning Systems Network (FEWS NET) Land Data Assimilation System; (4) PERSIANN-CDR: Precipitation Estimation From Remotely Sensed Information Using Artificial Neural Networks-Climate Data Record; (5) TerraClimate: Monthly Climate and Climatic Water Balance for Global Terrestrial Surfaces.

The maps of the difference of the confidence intervals (the upper prediction limit minus the lower prediction limit) at a significance level of 95% are also provided to show the spatial uncertainty of every single SPI map.

The drought events were counted during 1984-2020 at annual and seasonal scales. The variables related to the drought events are presented in “Drought_Event.zip”.

Reference: He, Q., Wang, M., Liu, K., Li, B., & Jiang, Z. (2023). Spatiotemporal analysis of meteorological drought across China based on the high-spatial-resolution multiscale SPI generated by machine learning. Weather and Climate Extremes, 40, 100567.

Background and AimsHepatic and coagulation function are routine laboratory tests prior to curative hepatectomy. The prognostic value of gamma-glutamyl transpeptidase (GGT) to platelet ratio (GPR) and international normalized ratio (INR) in surgically treated patients with intrahepatic cholangiocarcinoma (ICC) remains unclear.MethodsICC patients received curative hepatectomy in two west China centers were included. Time-dependent ROC curves were conducted to compare established indexes with prognostic value for ICC. GPR-INR score was introduced and evaluated using the Time-dependent AUC curve and Kaplan-Meier survival analysis. A novel nomogram based on the GPR-INR score was proposed; Harrell’s C-index, calibration curve and decision curve analysis were used to assess this nomogram.ResultsA total of 653 patients were included. The areas under ROC curves of GPR and INR in OS and RFS were superior to other indexes. Patients with a high GPR-INR score (1,2) presented significantly decreased overall survival (OS) and recurrence-free survival (RFS); GPR-INR sore, along with several clinicopathological indexes were selected into the nomogram, the calibration curve for OS probability showed good coincidence between the nomogram and the actual surveillance. The C-index of the nomogram was 0.708 (derivation set) and 0.746 (validation set), which was more representative than the C-indexes of the GPR-INR score (0.597, 0.678). In decision curve analysis, the net benefits of the nomogram in derivation and validation set were higher than Barcelona Clinic Liver Cancer staging (BCLC) classification and American Joint Committee on Cancer (AJCC) TNM 8th staging system.ConclusionsThe proposed nomogram generated superior discriminative ability to established staging systems; it is profitable to applicate this nomogram in clinical practice.

Comparison of mean absolute percentage error (MAPE) in total gastrointestinal quality of life index (GIQLI) score, physical component summary (PCS) score and mental component summary (MCS) score predicted by multiple linear regression (MLR), support vector machine (SVM), Gaussian process regression (GPR), and artificial neural network (ANN) models in forty new data sets.

Optimized hyperparameters for the Gaussian Process Regression (GPR) model.

MSE = mean square error, MAPE = mean absolute percentage error.

Performance metrics for the GPR model on training and testing datasets, showing high accuracy with low error values and R2 above 99% for both datasets.

We investigate the dynamic volatility connectedness of geopolitical risk, stocks, bonds, bitcoin, gold, and oil from January 2018 to April 2022 in this study. We look at connectivity during the Pre-COVID, COVID, and Russian-Ukraine war subsamples. During the COVID-19 and Russian-Ukraine war periods, we find that conventional, Islamic, and sustainable stock indices are net volatility transmitters, whereas gold, US bonds, GPR, oil, and bitcoin are net volatility receivers. During the Russian-Ukraine war, the commodity index (DJCI) shifted from being a net recipient of volatility to a net transmitter of volatility. Furthermore, we discover that bilateral intercorrelations are strong within stock indices (DJWI, DJIM, and DJSI) but weak across all other financial assets. Our study has important implications for policymakers, regulators, investors, and financial market participants who want to improve their existing strategies for avoiding financial losses.