Funding

Scientific activities are supported by LFIN's participation in several research projects and contracts. These projects are mainly supported by UCLouvain and several international and Belgian public institutions.

FSR Projects

Conic Martingales and Credit Risk Modeling (October 2017-November 2019)

Most often, credit risk is handled by modeling default time as first jump of Lévy processes governed by positive stochastic intensity processes. This specific setup corresponds to the standard Cox framework. However, the Doob-Meyer decomposition of the Azéma supermartingale - which ultimately needs to be modeled in the intensity approach - reveals that Cox setup is just one particular case. This is a sound motivation for investigating alternative classes of default models as conic martingales modelling.

Frédéric Vrins (CORE and LFIN, UCLouvain) is the promoter of this FSR project co-financed by the National Bank of Belgium.

Retail Investors and Multi-asset Protofolio Choices over Time (October 2018-September 2022)

The literature shows that retail investors often make costly investment mistakes, including those related to portfolio diversification. The project aims at characterizing these diversification mistakes and showing how penalizing they can be for the portfolio performance, at understanding whether behavioral biases and/or personal characteristics are responsible for these behaviors and at providing some simple diversification strategies to help retail investors improve their risk-adjusted portfolio returns. Our project will contribute to the literature in addressing this issue in a multi-asset-classes perspective (not only using stocks), in analyzing the way individuals adjust their portfolios over time (for example after receiving/losing some wealth/budget).

Rudy De Winne is the promoter of this project.

Public Institutions Projects

Dynamic Modelling of Recovery Rates with Application to the Risk Management of Financial Products (January 2018-December 2018)

Being extremely simple (straight bonds) or very complex (credit default swaps) the value of derivatives products all depend on a same factor: recovery rate. The later represent the ratio of the invested amount to the face value that will be recovered by the investor in case of default of the underlying reference entity. It is obvious to see that the closer this rate is from 100%, the lower should be the risk premium. Nowadays, the modeling of this factor is performed in an admittedly naïve way: it is essentially considered as a constant, which value needs to be determined. Our goal in this project is to propose a more meaningful approach, accounting for dependency with other risk factors. Several empirical analyses have been performed so far, but none of them had the ambition to understand its impact on derivatives products. This is precisely the scope of this project. It is handled in cooperation with Paris I Panthéon-Sorbonne and the Université d’Evry (Paris).

Frédéric Vrins (CORE and LFIN, UCLouvain) is the promotor of this project cofunded by Wallonie-Bruxelles International and the Ministry of Europe and Foreign Affaires In France.

FNRS:  Aspirant Grants

Dynamic Mode Information-theoretic Approach to Portfolio Optimization

This project deals with portfolio theory based on information theory, and entropy in particular. We wish to address two central issues. (1) We want to properly define the concept of diversification and make connections with the notion of risk. Diversification is commonly recognized as a mean to achieve risk mitigation, but is not properly defined for non-Gaussian data. Moreover, it is not a goal per se for investors, who care about their risk-return profile. However, highly concentrated portfolios are considered as undesirable. This raises the issue of the link between risk mitigation and diversification in a multivariate framework, and how to build properties for "good" diversification measures. To answer those questions, we will take the notion of dependence as a central piece. Mutual information will be taken as a first candidate. A second, non-symmetric, set of measures will be lower tail dependency and other copula-based measures that may be better tailored to our risk context and that received little atatention in the portfolio composition, we detract from the current methods based on statistics constraints and penalization of weight concentration. Instead, we will develop new detraic from the current methods based on statistics constraints and penalization of weight concentration. Instead, we will develop new maximum entropy methods to encapsulate the idea of diversification. We will derive adaptive algorithms to maximize the output entropy. More generally, we intend to design algorithms that provide the portfolio weights generating a return distribution as close as possible from a target function. To our knowledge, this will be the first approach dealing with a functional target instead of scalar targets. We will test our methods by assessing portfolio performances on synthetic and real data, in particular during crisis periods where left tail co-dependence matters most. We will also apply our diversification measures to market concentration determination.

Frédéric Vrins is the promoter of this project. Nathan Lassance is the aspirant.

Real Time Econometrics using Mixed Frequency Data (Ocotober 2018-October 2022)

This project develops machine learning (ML) methods for mixed data sampling by adapting a structured sparsity assumption. From the theory side, we account for both time series dependence and high-dimensionality. We develop theory both for conditional mean and conditional quantile prediction loss functions. Using recently developed theory for inference in high-dimensional models, we provide asymptotic results for our estimators. Lastly, we apply the developed methods to 1. forecast macroeconomic data using a large set of covariates sampled at different frequencies, and 2. model systemic risk of financial institutions.

Rudy de Winne and Eric Ghysels are the promoters of the project. Jonas Striaukas is the aspirant.

FNRS:  PDR Projects

Uncertainty, Macroeconomic Fluctuations and Asset Prices (October 2015-September 2019)

This project studies the relationship between uncertainty, macroeconomic fluctuations, financial markets and asset prices. In particular, we are interested at answering of questions such as: How does uncertainty influences macroeconomic variables such as aggregate level of prices or aggregate economic activity? Does the impact of uncertainty change in bad times? Is the relationship between macroeconomic and financial variables time dependent? What is the impact of uncertainty on asset prices? We answer these questions by building a macro-finance model. In this setting the dynamics of government bond prices, macroeconomic and finance variables are linked by no arbitrage conditions and uncertainty enters in the model in the form of parameter instability. We develop and apply Bayesian econometrics models to explore instability phenomena and to derive new measures of uncertainty. Bayesian econometrics models to explore instability phenomena and to derive new measures of uncertainty. Bayesian econometrics is well known for its flexibility in modelling uncertainty. Even if the model is initially applied to the government market, it can be used to analyze the relationship between uncertainty, macroeconomic dynamics and prices of other contracts such corporate bond, swaps or credit derivatives.
Improving our knowledge in this area is crucial for at least two reasons. From an academic perspective, it is interesting to more deeply understand (1) how the relationship between macro-economic variables, financial variables and asset pricing in time of turbulence, and (2) how uncertainty can influence these variables. This is a new area of research that can improve our understanding of economic models. From an applied perspective, understanding how uncertainty influences macroeconomic conditions or asset prices or the government yield curve is crucial for policy makers, who aim at improving general economic conditions and the efficiency of financial markets.

Leonardo Iania (CORE and LFIN, UCLouvain) is the promoter of this project.

FNRS: CDR Projects

Credit Risk Modelling and Stochastic Recovery Rates ( 2018- 2019)

This project aims at bringing two contribution to credit risk models used in the context of banking supervision. First, the new regulatory framework known as Basel III provides guidelines for better practices in terms of ris management of financial institutions. Among others, it provides banks with so-called "standard formulae" that are used to determine the capital that needs to be set aside to account for unexpected losses. Yet these formulae treat major risk drivers in a simplistic way, under the form of rather arbitrary multiplicative factors. Our ambition here is to analyze the relevance of these formulae in light of recently developed credit risk models. In particular, we would like to evaluate "how bad" do these standard formulae perform in the context of counterparty credit risk, a very topical problem in this context of financial and economic crisis. Second, we would like to enrich these models by accounting for another risk factor, namely the recovery rate. The recovery rate of a firm expresses on a relative basis the outstanding debt that can be recovered from a firm in case of default, via an auction process. Several authors studied the relationship between recovery rates and macroeconomic factors. Surprisingly however, there have been very few attempts of including those in default credit risk models as such. In spite of its acknowledged importance, it is most often disregarded in practice. One of the main reason is, we believe, the lack of specific mathematical tools available to handle this in a sound way. Hence, our second objective in this project is to rely on novel stochastic processes (namely, phi-martingales and lazy clocks) to model the recovery rate risk, incorporate it in existing credit models, and investigate the impact for financial institutions. We shall therefore analyze the dynamics of recovery rate quotes via specific databases, like Moody's and Markit. We shall then calibrate the models and compare the resulting figures to the standard formulae of Basel III.

Frédéric Vrins is the promoter of this project.

Information and (Dis)integration of Financial Markets (January 2016-December 2018)

We analyze two aspects related to information extraction and market functioning, namely the extraction of market expectations in the bond market and the interdepedence between the bond and stock markets. By extracting market expectations from the bond markets and identifying the transmission of shocks from the bond to the equity markets, we aim to better understand the dynamics of financial contagion which is likely to hit the real economy again in the coming years.

Leonardo Iania (CORE and LFIN, UCLouvain) is the promoter of this project.

Uncertainty and Monetary Policy (January 2018-December 2019)

This project will study how uncertainty and market fragmentation can impact the transmission of monetary policy to real activity and financial markets. We will be interested in uncertainty surrounding both economic conditions and the conduct of monetary policy itself.

Leonardo Iania (CORE and LFIN, UCLouvain) is the promoter of this project.

ARC Projects

Negative and Ultra-low Interest Rates: Behavioral and Quantitative Modelling (NeMo) (September 2018-August 2023 )

Interest rates are a cornerstone of economics and finance. They are at the foundation of asset pricing and monetary policy, and more generally of all intertemporal choices made by market participants and institutions every day, with huge consequences for the economic activity and wellbeing of our societies. Until recently, it was assumed (mostly implicitly) that interest rates could only possibly be positive. Notwithstanding, in the wake of the financial crisis initiated in 2008, major central banks of developed countries have been brought to conduct rates policies that turned them negative. The consequences of such a paradigm shift are both potentially huge and not well understood yet. This research project aims at shedding light on these consequences, both from an academic and a policy viewpoint, following three intertwined research lines that bring together a multidisciplinary team of researchers working on behavioral finance, macro finance, and quantitative finance.

Julio Davila (CORE, UCLouvain), Catherine D'Hondt, Leonardo Iania (LFIN and CORE, UCLouvain), Christian Hafner (ISBA and CORE, UCLouvain), Olivier Corneille (IPSY, UCLouvain) andFrédéric Vrins (LFIN and CORE, UCLouvain) and are the promoters of this project.