Essays on wavelet-based approaches for analyzing stock price dynamics
submitted by Christian Vial (from Lucerne)
Financial markets - and stock price movements in particular - are often described using simplified assumptions. However, financial markets are complex systems involving various interacting components. Agents in these markets have heterogeneous traits and differ in many respects. Among other characteristics, they have unique preferences, interpret information differently, pursue disparate investment goals, and focus on different investment horizons. These heterogeneities impact agents' buying and selling decisions and ultimately stock prices. As a result, those heterogeneities directly influence interdependencies between stocks and their price dynamics. Existing methods have not been sufficiently able to capture and explain these complexities. For this reason, the present thesis examines stock market mechanisms and interaction patterns using alternative mathematical filtration methods. The focus lies on investigating the price fluctuations of and the interdependencies between stocks across different timescales (time horizons). This analysis is directly linked to the assumption that market agents operate on different investment horizons. Chapter 1 studies changes in US stock correlations for different time horizons using wavelet decomposition. Wavelet decomposition is a method that allows filtering the dynamics of a time series within certain frequency ranges (time horizons). The empirical observations in this study indicate that stock market correlations do not remain constant across different time horizons. A major deficiency of the analysis in Chapter 1 is the significant degree of randomness hidden in correlation matrices. Chapter 2 therefore examines correlation structures using Random Matrix Theory (RMT). RMT analysis reveals that stock markets are governed by collective market behavior and sectoral factors across different timescales. Based on these insights, Chapter 3 studies portfolio strategies for minimizing risk at specific time horizons (scale-based portfolio strategies). The study demonstrates that (portfolio) variances can be minimized within a targeted frequency range using these scale-based portfolio strategies. Based on these findings, an optimization-method for simultaneous variance-minimization across different frequency bands is proposed.
Year of publication: |
2020
|
---|---|
Authors: | Vial, Christian |
Publisher: |
2020: St. Gallen |
Subject: | Theorie | Theory | Börsenkurs | Share price |
Saved in:
freely available
Extent: | 1 Online-Ressource (ix, 206 Seiten) Illustrationen |
---|---|
Type of publication: | Book / Working Paper |
Type of publication (narrower categories): | Hochschulschrift ; Aufsatzsammlung ; Graue Literatur ; Non-commercial literature |
Language: | English |
Thesis: | Dissertation, University of St.Gallen, 2020 |
Notes: | Enthält 3 Beiträge Zusammenfassung in deutscher und englischer Sprache |
Source: | ECONIS - Online Catalogue of the ZBW |
Persistent link: https://www.econbiz.de/10012316181
Saved in favorites
Similar items by subject
-
Do diversity reputation signals increase share value?
Cook, Alison, (2014)
-
Identification and evaluation of factors of dividend policy
Omerhodžić, Sead, (2014)
-
Non-financial human capital disclosure and share price
Mohamat Sabri Hassan, (2019)
- More ...
Similar items by person