Swiss Finance Institute @ EPFL

The Swiss Finance Institute @ EPFL has been created to foster research in finance and to develop a strong offering of programs in finance and financial engineering at the Ecole Polytechnique Fédérale de Lausanne. The focus is on the areas within finance that have a natural interaction with mathematics, statistics, engineering, and science, namely, mathematical finance, financial econometrics, and entrepreneurial finance.

The Swiss Finance Institute @ EPFL participates in two teaching programs, The Master in Financial Engineering at EPFL, which is a highly selective 2-year master program, and The PhD in Finance, which is organized jointly with the Swiss Finance Institute and the Universities of Geneva and Lausanne.

The Swiss Finance Institute @ EPFL benefits from the institutional support of the Swiss Finance Institute, a private foundation created in 2006 by Switzerland’s banking and finance community in cooperation with leading Swiss universities, and from Swissquote, who endowed the Swissquote Chair in Quantitative Finance.

Export event



CM 1 113

Stochastic Processes in Domains with Boundaries and Some of Their Financial Applications

In this talk we consider two connected problems:
First, we study the classical problem of the first passage hitting density of an Ornstein-Uhlenbeck process. We give two complementary (forward and backward) formulations of this problem and provide semi-analytical solutions for both. The corresponding problems are comparable in complexity. By using the method of heat potentials, we show how to reduce these problems to linear Volterra integral equations of the second kind. For small values of t we solve these equations analytically by using Abel equation approximation; for larger t we solve them numerically. We also provide a comparison with other known methods for finding the hitting density of interest, and argue that our method has considerable advantages and provides additional valuable insights.
Second, we study the non-linear diffusion equation associated with a particle system where the common drift depends on the rate of absorption of particles at a boundary. We provide an interpretation as a structural credit risk model with default contagion in a large interconnected banking system. Using the method of heat potentials, we derive a coupled system of Volterra integral equations for the transition density and for the loss through absorption. An approximation by expansion is given for a small interaction parameter. We also present a numerical solution algorithm and conduct computational tests.

By: Alexander LIPTON, SilaMoney, MIT & EPFL

Machine Learning, Asset Pricing and FinTech

Combination of 2 papers:
Paper 1. Predicting Stock Returns with Machine Learning
We employ a semi-parametric method known as Boosted Regression Trees (BRT) to forecast stock returns and volatility at the monthly frequency. BRT is a statistical method that generates forecasts on the basis of large sets of conditioning information without imposing strong parametric assumptions such as linearity or monotonicity. It applies soft weighting functions to the predictor variables and performs a type of model averaging that increases the stability of the forecasts and therefore protects it against overfitting. Our results indicate that expanding the conditioning information set results in greater out-of-sample predictive accuracy compared to the standard models proposed in the literature and that the forecasts generate profitable portfolio allocations even when market frictions are considered. By working directly with the mean-variance investor’s conditional Euler equation we also characterize semi-parametrically the relation between the various covariates constituting the conditioning information set and the investor’s optimal portfolio weights. Our results suggest that the relation between predictor variables and the optimal portfolio allocation to risky assets is highly non-linear. 
Paper 2. Who Benefits from Robo-advising? Evidence from Machine Learning
We study the effects of a large US robo-adviser on investor performance. Across all clients, the robo-adviser reduces investors holdings in money market mutual funds and increases bond holdings. It reduces the holdings of individual stocks and US active mutual funds, and moves investors towards low-cost indexed mutual funds. Finally, it increases investors’ international diversification and investors’ overall risk-adjusted performance. From sign-up, it takes approximately six months for the robo-adviser to adjust investors’ portfolios to the new allocations. We use a machine learning algorithm, known as Boosted Regression Trees (BRT), to explain the cross- sectional variation in the effects of PAS on investors’ portfolio allocation and performance. The investors that benefit the most from robo-advising are the clients with little investment experience, as well as the ones that have high cash-holdings and high trading volume pre-adoption. Clients with little mutual fund holdings and clients invested in high-fee active mutual funds also display significant performance gains. 

By: Alberto ROSSI, University of Maryland


Swissquote Conference 2018 on Machine Learning in Finance

The ninth annual Swissquote Conference on Machine Learning in Finance took place at EPFL on 9 November 2018. The conference featured current research and insights on machine learning in finance provided by leading experts from academia a