Investment Research
- Offer of financial market insights and intelligence
- Explain the complex relationships and interactions between economic variables and financial factors
- Rule-based research of the economy and asset classes
- Explain the cause-effect linkages of financial markets
- Use of cutting-edge technologies, like machine learning
Combine
Adaptive Asset Allocation
With
Systematic Security Selection
To Perform
Beyond Behavioral Bias
Examples of Financial Market Analysis
Recession Forecasting
- Forecast of recessions in the US with multi-factor models and different machine learning techniques for different time lags
- Idea based on the yield curve model of the Fed with statistical probit-model
- Analysis factors with cluster analysis to avoid selecting related factors
- Machine learning models used: K-nearest Neighbors, Support Vector, Gaussian, Decision Tree, Random Forest, Naïve Bayes
Forecasting Trend with L1-Trend Filter
- Analyse stock market trends to overweight or underweight the asset class
- Model based on L1-trend filtering technique
Principal Component Analysis
- Application of the principal component analysis (PCA) on the components of an equity index
- Decompose the dynamics of the returns into orthogonal factors
- Checking the dynamics of the market: top-down vs bottom-up investment phase of the market
- Clustering the stocks with the help of principal component analysis and machine learning techniques in order to improve the analysis of industries and companies
Relative Rotation Graphs
- Relative Rotation Graphs by Julius De Kempenaer
- Analysis of trends in relative strength in a single graph
- Calculation of the JdK ratio and momentum
- Flexibility on the asset universe, the period and the time horizon
Bond Valuation
- Cross-sectional regression model to evaluate the spread of bonds with different models (linear, lasso, elastic net, support vector, extreme gradient boost, light gradient boosting – Microsoft, deep learning with one input, hidden and output layer and with respective 20, 10 and 1 nodes)
- Model used for EUR-denominated and USD-denominated corporate and government bonds
- Estimation of spreads for large universe of issues
- Evaluation and comparison of the different models
Conditional Correlation and Markov Switching Models
- Estimation of the conditional correlation with multivariate Dynamic Conditional Correlation – Generalized AutoRegressive Conditional Heteroskedasticity (DCC – GARCH)
- Computing the probability of an event with low or high correlation between equity and bonds
- Markov Switching Dynamic Regression of the residuals of the DCC – GARCH model to predict the correlation regime switch
- Out-of-sample backtesting of strategies based on Markov Switching models
Portfolio Optimization with Higher Moments
- Mean-variance optimization of different assets in order to perform a global asset allocation
- Optimize the Sharpe ratio in a mean-variance framework
- Portfolio optimization under turnover, return and risk constraints
- Monitoring the weights of the different asset classes
- Controlling the Sharpe ratio, skewness and kurtosis for a universe of assets
