Can stock returns be predicted? This master's thesis investigates the question empirically for the German stock market. Using monthly price data for all DAX constituents from 1983 to 2013, it develops and tests its own forecasting models: a multivariate regression based on micro- and macroeconomic variables (P/E ratio, EPS, credit spread, oil price, GDP and others), an ARMA time-series model and a naive benchmark. The complete methodology is demonstrated using the BASF share, with forecasting quality assessed through in-sample and out-of-sample tests via RMSE, MAE and MAPE.
The study first provides a solid grounding in the theoretical foundations — capital market efficiency, portfolio management, technical and fundamental analysis — together with a comprehensive literature review. The findings are as honest as they are insightful: over the long run, returns can be approximately explained by a few economic factors, while in the short run the market remains largely unpredictable. Essential reading for economics and finance students, analysts and anyone interested in quantitative equity valuation.
Table of Contents
A Introduction
B Theoretical Foundations
I. Capital Market Efficiency
II. Portfolio Management
III. Financial Analysts
C Equity Valuation
I. Technical Analysis
1. Dow Theory
2. Elliott Wave Theory
3. Chart Patterns
4. Indicators
II. Fundamental Analysis
1. Global Analysis
2. Industry Analysis
3. Company Analysis
D Empirical Study
I. Own Empirical Study
1. Introduction and Literature Review
2. Data
3. Test for Stationarity
4. Statistical Properties of the Data
II. Methodological Foundations of Time Series Analysis
1. ARMA Model
2. Estimation of ARMA Models
III. Methodological Foundations of Multivariate Regression Analysis
1. Model Formulation
2. Estimation and Testing of the Regression Function
3. Testing of the Regression Coefficients
IV. Empirical Results
E. Summary and Conclusion
Research Objectives and Core Topics
This thesis investigates the predictability of stock returns on the German stock market, aiming to determine whether empirical models can outperform historical averages or simple random walk expectations. The research addresses the tension between the efficient-market hypothesis and various analytical approaches used by market participants to forecast future developments.
- Theoretical foundations of capital market efficiency and portfolio management.
- Comprehensive analysis of technical and fundamental equity valuation methods.
- Methodological development of time series (ARMA) and multivariate regression models.
- Empirical evaluation of forecasting quality using in-sample and out-of-sample data for DAX constituents.
Excerpt from the Book
1. Introduction and Literature Review
Empirical investigations into the forecasting of stock returns have been a fascinating challenge for all kinds of capital market participants for many decades. In the literature there is an almost endless list of the most varied studies on the prediction of stock returns. Most studies relate largely to the US capital market and the associated micro- and macroeconomic data sets.79
For practitioners in asset management as well as for academics, it is of the greatest interest to study forecasts of stock returns, whether to optimise the composition of a portfolio or to test the efficiency of the capital market. Over the past decades, opinion has frequently shifted as to whether stock returns can be explained.
The random walk theory, which has been assumed by many economists in their studies, states that future stock returns cannot be forecast, at least not with the information currently available. Since future information cannot be foreseen, share prices and returns cannot be predicted.
The random walk theory is also supported by the studies of Working (1934) and Cowles/Jones (1937), which show that US share prices and other economic data follow a random course. In the 1930s, Graham and Dodd (1934) were the first to examine, in their book "Security Analysis", the relationship between high valuation ratios (e.g. the P/E ratio) and future stock returns. They argue that only a company's distribution policy and the earnings it generates are decisive for how a share price will behave in the future.808182
Summary of Chapters
A Introduction: Outlines the core challenge of predicting stock returns and introduces the fundamental theories and hypotheses governing capital market research.
B Theoretical Foundations: Explores the concepts of capital market efficiency, the structure of portfolio management, and the role of financial analysts.
C Equity Valuation: Details the methodologies behind technical and fundamental analysis used to evaluate stocks and derive potential price forecasts.
D Empirical Study: Presents the author's own research, covering the literature review, data processing, model methodologies, and the resulting performance analysis of the forecasting models.
E. Summary and Conclusion: Summarizes the thesis findings, stating that despite the extensive analytical effort, no model consistently outperformed simple benchmarks in predicting monthly returns.
Keywords
Stock returns, predictability, German stock market, DAX, capital market efficiency, random walk hypothesis, technical analysis, fundamental analysis, ARMA models, multivariate regression, forecasting quality, in-sample, out-of-sample, finance.
Frequently Asked Questions
What is the core focus of this Master's thesis?
The thesis investigates whether stock returns on the German stock market are predictable using historical economic data and various quantitative modeling techniques.
What are the primary analytical fields covered?
The study covers capital market theory, portfolio management, financial analysis, technical analysis (charts, indicators), and fundamental analysis (global, industry, and company analysis).
What is the central research goal?
The goal is to develop and test forecasting models, specifically ARMA models and multivariate regressions, to see if they can effectively predict monthly returns of DAX stocks.
Which methodologies are employed for the study?
The author uses time series analysis (ARMA models) to account for dependencies in returns and multivariate regression analysis to assess the influence of micro- and macroeconomic variables.
What does the main part of the thesis discuss?
The main part explains the theoretical frameworks, the setup of the empirical study including data transformation and stationarity testing, and the performance testing of different forecasting models.
Which keywords best describe this research?
The research is best characterized by terms such as stock return predictability, capital market efficiency, DAX constituents, time series analysis, and quantitative forecasting models.
Why did the author specifically choose BASF for the detailed analysis?
The author selected BASF because complete data sets were available, and the stock is considered representative of the broader market performance across DAX constituents.
What was the general conclusion regarding the predictability of returns?
The author concludes that none of the models achieved adequate forecasting quality, suggesting that short-term stock returns are largely influenced by incalculable factors, making consistent outperformance difficult.
- Quote paper
- Marc Seibert (Author), 2015, Are stock returns predictable?, Munich, GRIN Verlag, https://www.grin.com/document/1737281
