Analysis of climate- and financial volatility measures and their relation
Author
Andersson, Mads Bo
Term
4. term
Education
Publication year
2022
Submitted on
2022-06-03
Pages
61
Abstract
Dette speciale undersøger, hvordan målinger af klimavolatilitet og finansiel volatilitet kan konstrueres og hvordan de hænger sammen. Klimadata er daglige temperaturafvigelser, mens finansmarkederne repræsenteres af S&P 500 (G7) og MSCI BRIC (BRIC-landene). Først fjernes deterministisk sæsonvariation med en lineær model. Dernæst udtrækkes stokastisk volatilitet ved hjælp af en tidsvarierende VAR-model med stokastisk volatilitet (TVP-VAR-SV). De afledte klimavolitetsmål sammenlignes med finansielle volatilitetsmål via krydskorrelation, inklusive analyser med og uden COVID-19-perioden. Analysen finder en forsinket afhængighed, hvor stigninger i finansiel volatilitet efterfølges af fald i klimavolatilitet i de efterfølgende år. COVID-19-perioden er forbundet med faldende klimavolatilitet. Forskelle i CO2-udledninger mellem G7 og BRIC synes ikke at kunne spores i klimavolatiliteten. Specialet peger på mulige forbedringer, herunder alternative modeller (fx GARCH og Heston), større geografiske områder, bedre mål for klimaforandringer, mere finansiel data og konstruktion af ukorrelerede finansielle volatilitetsmål.
This thesis investigates how to construct and relate measures of climate volatility and financial volatility. Climate is proxied by daily temperature anomalies, while financial markets are represented by the S&P 500 (G7) and MSCI BRIC (BRIC countries). Deterministic seasonality is first removed using a linear model. Stochastic volatility is then extracted with a Time-Varying Parameter Vector Autoregression with Stochastic Volatility (TVP-VAR-SV). The resulting climate volatility measures are compared with financial volatility using cross-correlation, including analyses with and without the COVID-19 period. The study finds a lagged dependence in which spikes in financial volatility are followed by decreases in climate volatility in subsequent years. The COVID-19 period is associated with a decline in climate volatility. Differences in CO2 emissions between G7 and BRIC do not appear to be reflected in the climate volatility measures. The thesis outlines opportunities for further work, including alternative models (e.g., GARCH and Heston), larger regions, improved climate metrics, more financial data, and the construction of uncorrelated financial volatility measures.
[This summary has been generated with the help of AI directly from the project (PDF)]
Documents