Climate change is expected to have significant long-term effects on the physical, financial, and economic environments, posing substantial risks to the financial stability of general insurers. While Dynamic Financial Analysis (DFA) is widely used to assess financial impacts, traditional DFA models often overlook climate risks. To address this gap, this study introduces a climate-dependent DFA framework that integrates climate risk into DFA, providing a comprehensive assessment of its long-term impact on the general insurance industry.
The proposed framework has three key features: (1) it captures the long-term impact of climate change on both assets and liabilities by considering physical and economic dimensions within an interconnected structure; (2) it addresses the uncertainty of climate impacts using stochastic simulations within climate scenario analysis, supporting actuarial applications; and (3) it is tailored to the unique characteristics of the general insurance sector.
This repository contains the results presented in our paper, "Dynamic Financial Analysis (DFA) of General Insurers under Climate Change."
All code used to generate the results can be found in the Rmarkdown file: DFA_model_final.Rmd, which contains the following sections that need to be run in the order specified:
-
Before running the Rmarkdown file, please ensure
knitrandrmarkdownpackages are installed. -
Loading required packages: Load the required packages used in the paper.
-
Files and data path: Define the file paths of all the data sources used in the paper. Those data sources are all stored in the
Datafolder. Unless a new data source is used, users do not need to change anything here. -
Define functions to be used: This section defines all the functions used in the paper.
-
Section 3.1 Data and calibration: This section calibrates the model parameters based on historical data as per Section 3.1 in the paper.
-
Section 3.2 Key simulation results from individual modules: This section generates the results shown in Section 3.2 in the paper.
- The simulation control parameters are defined in Define the control variables. These include the number of simulations, random seed, maximum forecasting horizon, and asset allocations, etc. Users can modify these parameters to suit specific needs. After changing these parameters, only Sections 3.2 and 3.3 need to be rerun, and previous sections do not require rerunning.
- All graphs shown in Section 3.2 in the paper are plotted in Presentations of simulation results.
-
Section 3.3 Risk and return measures: This section presents the results shown in Section 3.3 in the paper.
The miscellaneous data cleansing scripts are stored in the folder Miscellaneous data cleaning codes. These scripts are optional for users to run and can all be executed from within DFA_model_final.Rmd.
Due to the large file size (around 32GB), the data used in this paper cannot be uploaded to GitHub. However, users can download all the data from: https://zenodo.org/records/15098758.
Once the Data folder is downloaded, place it in the same directory as DFA codes.Rproj and DFA_model_final.Rmd. After this setup, users can run the RMarkdown file DFA_model_final.Rmd without further configuration.
The main data folders and the corredsponding data sources used are outlined in the table below.
| Data categories (Folder name) | File names and Descriptions | Sources |
|---|---|---|
| Economic | World_pop.xlsx: World population data by countries |
World Development Indicators (World Bank); downloaded at 23 March 2024 |
AU-GDP-LCU.xlsx: Historical Australian GDP at annual interval between 1960 and 2022 |
World Development Indicators (World Bank); downloaded at 23 March 2024 | |
SSP_scenarios.csv: Country-level GDP projections for each SSP scenario through 2100 |
Sourced from IIASA SSP database; downloaded from https://zenodo.org/records/8116099 at 1 January 2024 | |
AU_population_projections.xlsx: Projections of Australian population for each SSP scenario through 2100 |
Downloaded from IIASA SSP database at 18 March 2024 | |
AU_Consumption_LCU.xlsx: Australian aggregate consumption data at annual interval between 1974 and 2023 |
World Development Indicators (World Bank); downloaded at 23 July 2024 | |
AU_inflation_640101.xlsx: Quarterly Consumer Price Index (CPI) data for Australia from 1948 to 2023 |
Downloaded from Australian Bureau of Statistics | |
Australian_cash_rate_data.xls and Australian_cash_rate_data_2011_2023.xls: Australian cash rates data from 1976 to 2023 |
Downloaded from Reserve Bank of Australia at 1 January 2024 | |
Potential-growth-database.xlsx: Potential GDP growth estimates by countries between 1981 and 2021 |
Downloaded from World Bank Potential Growth Database at 11 June 2024 | |
OECD_Gas_SSP.xlsx and OECD_Oil_SSP.xlsx: Projections of oil and gas production under each SSP scenarios in OECD countries |
Downloaded from IIASA SSP database at 18 March 2024 | |
| Equity return | All-Ordinaries-Yearly-pseudo.xlsx |
Pseudo All-Ordinaries Shares total returns series between 1992 and 2023 |
Woodside_Financials_pseudo.xlsx: Pseudo financial statements from a oil and gas producer (Woodside) between 2014 and 2023 |
Original data is not provided due to licensing restrictions; users can download the original version from FactSet. Alternatively, users can hand-collect the relevant data for calculation based on the published financial statements from Woodside Energy | |
| Hazards loss | ICA-Historical-Normalised-Catastrophe-August-2023.xlsx: Historical records of insured losses from natural disasters in Australia between 1967 and 2023 |
Downloaded from Insurance Council of Australia at 27 August 2023 |
EM_DAT_AU.xlsx: Historical records of both insured losses and total economic damages from natural disasters in Australia between 1985 and 2023 |
EM-DAT database | |
| Market statistics | Quarterly general insurance performance statistics database December 2002 to June 2023.xlsx: Quarterly aggregate financial data on Australian general insurers from 2002 to 2023 |
Downloaded from APRA Quarterly General Insurance Performance Statistics; downloaded at 20 July 2023 |
Annual general insurance institution-level statistics database March 2005 to June 2023.xlsx: Provides annual, institution-level financial information from 2005 to 2023, with key metrics such as insurance losses, premiums and equity bases |
Downloaded from APRA Institution-level Statistics Database; downloaded at 20 July 2023 | |
| Precipitations | era5-x0.25_timeseries_pr,rx1day,rx5day_timeseries_annual_1950-2022_mean_historical_era5_x0.25_mean.xlsx: ERA5 reanalysis data of historical precipitation information in Australia between 1950 and 2022 |
Originated from Copernicus Climate Change Service;cleansed version downloaded from World Bank Climate Change Knowledge Portal at 29 April 2024 |
| FWI | FWI_1940_1949.nc,..., FWI_2016_2022.nc: Raw historical Fire Weather Index (FWI) in Australia constructed based on ERA5 reanalysis data between 1940 and 2022, in NetCDF format |
Downloaded from Copernicus Climate Change Service at 3 May 2024 |
FWI_hist_data.csv: A cleansed CSV version of the historical FWI data (described above) is provided for convenience, particularly for users who may not be able to install spatial data processing packages on their machines |
Derived from FWI data in Copernicus Climate Change Service | |
FWI_ref_data.csv: A cleansed CSV version of the extreme quantiles of FWI data over the historical reference period |
Derived from FWI data in Copernicus Climate Change Service | |
fwixx_hurs_part1,...,fwixx_hurs_part7: Raw projections and historical backcasts of FWI statistics in Australia from an ensemble of CMIP6 model, in NetCDF format |
Downloaded from FWI data under historical and SSP projections in CMIP6 (Quilcaille et al., 2022) | |
mfwixx_ensemble_hist_data.csv and xfwixx_ensemble_hist_data.csv: Processed historical backcasts of FWI statistics in Australia (described above) in cleansed CSV format |
Derived from FWI data under historical and SSP projections in CMIP6 (Quilcaille et al., 2022); detailed source codes can be found in FWI_import_clean.R under the Miscellaneous data cleaning codes folder |
|
ssp126_ensemble_mfwixx_proj.csv,...,ssp585_ensemble_mfwixx_proj.csv: Processed projections of FWI statistics in Australia under different SSP scenarios, in cleansed CSV format |
Derived from FWI data under historical and SSP projections in CMIP6 (Quilcaille et al., 2022); detailed source codes can be found in FWI_import_clean.R under the Miscellaneous data cleaning codes folder |
|
| SST and MSLP | Sea_ERA5_50_59.nc,...,Sea_ERA5_2018_23.nc: ERA5 reanalysis data of historical Sea Surface Temperature (SST) and Mean Sea Level Pressure (MSLP) over the Australian tropical cyclone basin, in NetCDF format |
Downloaded from Copernicus Climate Change Service: ERA5 monthly averaged data on single levels from 1940 to present at 6 May 2024 |
| Near-surface temperature | ERA5.nc: ERA5 reanalysis data of historical near-surface temperature over Australian land, in NetCDF format |
Downloaded from Copernicus Climate Change Service:ERA5-Land monthly averaged data from 1950 to present at 6 May 2024 |
tas_ERA5_hist_dat.csv: Processed ERA5 reanalysis data of historical near-surface temperature over Australian land (described above) in cleansed CSV format |
Derived from the data source listed above | |
| Air temperature | ERA5_air_temp_hpa300.nc: ERA5 reanalysis data of historical air temperature at 300hPa over Australian land, in NetCDF format |
Downloaded from Copernicus Climate Change Service:ERA5 monthly averaged data on pressure levels from 1940 to present |
ta_ERA5_hist_dat: Processed ERA5 reanalysis data of historical air temperature at 300hPa over Australian land (described above) in cleansed CSV format |
Derived from the data source listed above | |
| CMIP6_ensemble_precipitation | cmip6_rx5day_ensemble.xlsx: Projections of Largest five-day cumulative precipitation (rx5day) by CMIP6 ensemble in Australia |
Originated from Copernicus Climate Change Service;cleansed version downloaded from World Bank Climate Change Knowledge Portal at 29 April 2024 |
| CMIP6_ensemble_SST | access_cm2-historical_sea_surface_temperature.zip,..., ukesm1_0_ll-historical_sea_surface_temperature: Raw historical backcasts of sea-surface temperature over the Australian tropical cyclone basin from CMIP6 ensemble, in NetCDF format; the first word before the hyphen (e.g., access_cm2) indicates the name of the component model within the ensemble |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 |
SST_ensemble_hist_data.csv and SST_grad_ensemble_hist_data.csv: Historical backcasts of average sea-surface temperature and sea-surface temperature gradients from CMIP6 ensemble over the Australian tropical cyclone basin derived from the data above, in CSV format |
Derived from the data source listed above; detailed source codes can be found in SST_import_clean.R under the Miscellaneous data cleaning codes folder |
|
access_cm2-ssp1_2_6_sea_surface_temperature.zip,...,noresm2_mm-ssp5_8_5_sea_surface_temperature.zip: Raw projections of sea-surface temperature over the Australian tropical cyclone basin from CMIP6 ensemble under each SSP scenario, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 | |
ssp126_ensemble_SST_proj.csv,...,ssp585_ensemble_SST_proj.csv: Cleansed projections of the average sea-surface temperature over the Australian tropical cyclone basin from CMIP6 ensemble under each SSP scenario derived from the data above, in CSV format |
Derived from the data source listed above | |
ssp126_ensemble_SST_grad_proj.csv,...,ssp585_ensemble_SST_grad_proj.csv: Cleansed projections of the average sea-surface temperature gradients over the Australian tropical cyclone basin from CMIP6 ensemble under each SSP scenario derived from the data above, in CSV format |
Derived from the data source listed above | |
| CMIP6_ensemble_MSLP | access_cm2-historical_sea_level_pressure.zip,...,noresm2_mm-historical_sea_level_pressure.zip: Historical backcasts of sea-level pressure over the Australian tropical cyclone basin from CMIP6 ensemble, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 |
MSLP_ensemble_hist_data.csv: Cleansed historical backcasts of mean sea-level pressure over the Australian tropical cyclone basin from CMIP6 ensemble, in CSV format |
Derived from the data source listed above;detailed source codes can be found in MSLP_import_clean.R under the Miscellaneous data cleaning codes folder |
|
access_cm2-ssp1_2_6_sea_level_pressure.zip,...,noresm2_mm-ssp5_8_5_sea_level_pressure.zip: Raw projections of mean sea-level pressures over the Australian tropical cyclone basin from CMIP6 ensemble under each SSP scenario, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 | |
ssp126_ensemble_MSLP_proj.csv,...,ssp585_ensemble_MSLP_proj.csv: Cleansed projections of the average sea-level pressures over the Australian tropical cyclone basin from CMIP6 ensemble under each SSP scenario derived from the data above, in CSV format |
Derived from the data source listed above | |
| CMIP6_ensemble_near_surface_temperature | access_cm2-historical_near_surface_air_temperature.zip,...,noresm2_mm-historical_near_surface_air_temperature.zip: Historical backcasts of near-surface temperature over the Australian land from CMIP6 ensemble, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 |
historical_ensemble_tas.csv: Cleansed historical backcasts of near-surface temperature over the Australian land from CMIP6 ensemble, in CSV format |
Derived from the data source listed above; detailed source codes can be found in Near_surface_temperature_import_clean.R under the Miscellaneous data cleaning codes folder |
|
access_cm2-ssp1_2_6_near_surface_air_temperature.zip,...,noresm2_mm-ssp5_8_5_near_surface_air_temperature.zip: Raw projections of near-surface temperature over the Australian land from CMIP6 ensemble under each SSP scenario, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 | |
ssp126_ensemble_tas_proj.csv,...,ssp585_ensemble_tas_proj.csv: Cleansed projections of near-surface temperature over the Australian land from CMIP6 ensemble under each SSP scenario, in NetCDF format |
Derived from the data source listed above | |
| CMIP6_ensemble_air_temperature | access_cm2-historical_air_temperature.zip,...,noresm2_mm-historical_air_temperature.zip: Historical backcasts of air temperature (at 300hPa) over the Australia from CMIP6 ensemble, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 |
historical_ensemble_ta.csv: Cleansed historical backcasts of air temperature over Australia from CMIP6 ensemble, in CSV format |
Derived from the data source listed above; detailed source codes can be found in Air_temperature_import_clean.R under the Miscellaneous data cleaning codes folder |
|
access_cm2-ssp1_2_6_air_temperature.zip,...,noresm2_mm-ssp5_8_5_air_temperature.zip: Raw projections of air temperature (at 300hPa) over Australia from CMIP6 ensemble under each SSP scenario, in NetCDF format |
Downloaded from: Copernicus Climate Change Service:CMIP6 climate projections at 29 April 2024 | |
ssp126_ensemble_ta_proj.csv,...,ssp585_ensemble_ta_proj.csv: Cleansed projections of air temperature (at 300hPa) over Australia from CMIP6 ensemble under each SSP scenario, in CSV format |
Derived from the data source listed above |
- Benjamin Avanzi
- Yanfeng (Jim) Li
- Greg Taylor
- Bernard Wong
For any questions or further information, please contact lyf1998130@126.com.