Projects for Master theses

• Project #1: Stochastic declustering diagnostic of endogenous versus exogenous origins of successes and shocks in social systems
  
  
• Project #2: Drawdowns and financial risks. A major concern for financial risk management is the existence of "outliers" (also called "kings'' or "black swans'') in the distribution of financial risks measured at variable time scales such as with drawdowns. These outliers are identified only with metrics adapted to take into account transient increases of the time dependence in the time series of returns of individual financial assets. They seem to belong to a statistical population which is different from the bulk of the distribution and thus require some specific treatment. The Chair of Entrepreneurial Risk proposes a Master thesis whose main purpose is the study of the challenging question of identifying and quantifying the rallies and drawdowns observed in the times series of financial return. The work will also include the study of the statistical distribution of the rallies and drawdowns and the development of asset allocation methods adapted to account for these phenomena. The applicants must have good computational skills (knowledge of Matlab® is mandatory) as well as a good level in finance, mathematics and statistics.
  
  
• Project #3: Study of a nonlinear model of the price of an asset: Kalman filter calibration to real data to develop a trading program of financial stocks
  
  
• Project #4: Porfolio optimization and genetic algorithm: tests with different objective functions, survival bias, non-stationarity
  
  
• Project #5: Portfolio tests of stock market predictions and financial crashes
  
  
• Project #6: Endogenous versus exogenous dynamics and scaling laws in YouTube dynamics of the fame of video
  - Finding correlations between the most viewed videos each day and relevant news stories in order to gauge whether or not YouTube can act as a sort of social barometer.
  - Looking at the microdynamics of the ranks of videos. Analyzing on the basis of Rank Clocks (Nature: http://www.nature.com/nature/journal/v444/n7119/full/nature05302.html )
  - Developing a measure of the relative importance or impact of the different "featuring" mechanisms on YouTube.....in other words, what is the effect of being featured on the "Most Viewed" part of the website as compared to the "Most Discussed" section of the site?
  
  
• Project #7: Factors of Success and Failure in Open Source Projects
  (http://www.er.ethz.ch/research/computer_risks)
  
  
• Projet #8: Rating the social networks of citations of scientists using the endo-exo dynamical approach
  
  
• Project #9: FTS-GARCH model of financial bubble
  Here FTS means "finite-time-singularity" and refers to the dynamics of financial prices during bubbles, which results from self-reinforcing positive feedbacks.
  The goal is to couple the theory of super-exponential bubbles developed in our group with the standard GARCH model of financial volatility clustering (Nobel prize winning work Engle 2003).
  
  
  [image reference is broken]

  This is a simple GARCH model just with a different mean dynamics (the bubble dynamics) than is usually used. The goal is to estimate this FTS-GARCH model directly estimated with Maximum likelihood and develop formal tests to determine the exponent n of the price that represents the strength of the positive feedback.
  The gain would be to have a model which combines the standard of the financial industry to assess volatility risk together with the exceptional risk component of a financial bubble.
  
  

• Project #10: Agent-based models of financial volatility and stochastic resonance
  
  
• Project #11: New explanation of the conjunction fallacy
  The conjunction fallacy is a logical fallacy that occurs when it is assumed that specific conditions are more probable than a single general one. The most oft-cited example is this fallacy originated with amos Tversky and Daniel Kahneman:
  Linda is 31 years old, single. outspoken and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations. Which is more probable
  - Linda is a bank teller,
  - Linda is a bank teller and is active in the feminist movement?
  

  Method: probability theory using "typical" as mode of multivariate probabilities of the properties considered and some hidden implicit attribute (IQ for instance). This explanation for the conjunction effect would be similar to those invoking the "reason-based" explanation for the disjunction fallacy but here there is a clear mathematical reason why the conjunction fallacy could/should appear.
  
  

• Project #12: Understanding the fundamentals of the US and European financial systems
  - understand liquidity, monetary base, funding and leverage
  - look for good and alternative measures
  - discuss the central bank mechanisms (including ad hoc like quantitative easing)
  - look for feedback mechanisms
  - look for signatures of criticality
  Contact: Cauwels Peter pcauwels@ethz.ch - http://www.er.ethz.ch/people/cauwelsp
  
  

• Project #13: Statistical analysis of civil conflicts and wars
  
  

• Project #14: Agent-based models of human cooperation in public good situations
  
  
• Project #15: Linking three stylized facts of financial stock markets: leverage effect, gain-loss asymmetry and cross-sectional dependence of stock returns. The goal is to develop the theory and empirical tests of the gain-loss asymmetry, as a result of the two other facts (leverage effect and cross-correlations). The gain-loss asymmetry refers to the remarkable observation that, for the S&P500, it takes typically 10 days to lose 5% but it takes 20 days to gain 5%.
  
  
• Project #16: PRECURSORS AND DECAY OF HAPPINESS: The goal is to test the endo-exo theory that we have developed elsewhere for commercial successes and to understand the power law distribution of waiting times between cause and actions to the dynamics of happiness!
  see http://www.uvm.edu/~cdanfort/research/dodds-danforth-johs-2009.pdf
  How Happy Is the Internet?  http://sciencenow.sciencemag.org/cgi/content/full/2009/803/3
  
  
• Project #17: Empirical tests of punctuated evolution due to delayed carrying capacity in economics and biology. For instance, use data on the five known developments and mass extinctions of biological species, data on successions of S-curves for innovations, and so on.
  This project is based on the paper: V.I. Yukalov, E.P. Yukalova and D. Sornette, Punctuated Evolution due to delayed Carrying Capacity, Physica D 238, 1752-1767 (2009) (http://arXiv.org./abs/0901.4714)
  
  
• Project #18: Assessment and prediction of oil peak, gold peak, copper peak and the fate of civilization. Excluding soil and water, which allows us to grow our food, oil and gold must be the two most important minerals of our present civilization and their production have spanned only a few centuries, much shorter than western civilization. It is amazing to think that we are presently at a key time (peak or plateau) of our civilization in term of supply and most do not realize it. Paul Valery wrote in 1931 "the time of a limited world begins". The master thesis will consist in reviewing data provided by the ER chair and applying different models to assess the different scenarios.
  
  
• Project #19: Analysis of entertainment shoping on the internet. The project aims at analyzing the bidding behavior in online "entertainment" auctions. The dynamics of individual subjects up to groups of bidders should be revealed and possible behavioral patterns be identified. Furthermore the attraction- and bidding dynamics of specific auctioneered producsts should be analyzed. Fluent English and good knowledge in statistics, the Matlab programming language as well as SQL is required. The project is based on data obtained from www.swoopo.com
  
  
  
• Project #20: Real-time "paper-money" implementation of trading strategies using reverse engineering of financial markets with Agent-Based Models (ABM)
  Goals:
  Understanding, Testing and Ideally Extending / Evolving a Trading Strategy based on Reverse Engineering Financial Markets with Agent-Based Models (ABM)
  Tasks:
  - Getting to know the concept of ABM and understanding their potential when used for reverse engineering financial markets.
  - Learning the fundamentals of genetic algorithms (and similar evolutionary strategies).
  - Getting familiar with the online trading platform "Interactive Brokers" for performing (automated) trading according to the stratgegy (paper trading and optionally real trading).
  - Developing and understanding of strategies performance analysis and apply them for the developed strategy.
  Requirements:
  - Understanding C++ and A plus knowing the boost library
  - A plus: having experience in using Interactive Brokers or similar trading platforms.
  
  
• Project #21: Multiple Criteria Decision Analysis applied to the financial portfolio investments
  The thesis will be based on the ELECTRE method (http://en.wikipedia.org/wiki/ELECTRE) and will apply thius method in the context of financial markets.
  
  Goal:
  To develop a robust and and efficient asset allocation methodology based on multiple criteria. Another domain of application will be the evaluation of real options for new projects.
  
  ELECTRE is a family of multi-criteria decision analysis methods that originated in Europe in the mid-1960s. The acronym ELECTRE stands for: ELimination Et Choix Traduisant la REalité (ELimination and Choice Expressing REality). The method was first proposed by Bernard Roy and his colleagues at SEMA consultancy company. A team at SEMA was working on the concrete, multiple criteria, real-world problem of how firms could decide on new activities and had encountered problems using a weighted sum technique. Bernard Roy was called in as a consultant and the group devised the ELECTRE method. As it was first applied in 1965, the ELECTRE method was to choose the best action(s) from a given set of actions, but it was soon applied to three main problems: choosing, ranking and sorting. The method became more widely known when a paper by B. Roy appeared in a French operations research journal.[1] It evolved into ELECTRE I (electre one) and the evolutions have continued with ELECTRE II, ELECTRE III, ELECTRE IV, ELECTRE IS and ELECTRE TRI (electre tree), to mention a few.[2] Bernard Roy is widely recognized as the father of the ELECTRE method, which was one of the earliest approaches in what is sometimes known as the French School of decision making. It is usually classified as an "outranking method" of decision making.There are two main parts to an ELECTRE application: first, the construction of one or several outranking relations, which aims at comparing in a comprehensive way each pair of actions; second, an exploitation procedure that elaborates on the recommendations obtained in the first phase. The nature of the recommendation depends on the problem being addressed: choosing, ranking or sorting. Criteria in ELECTRE methods have two distinct sets of parameters: the importance coefficients and the veto thresholds.
  
  References:
  [1] Roy, Bernard (1968). "Classement et choix en présence de points de vue multiples (la méthode ELECTRE)". La Revue d'Informatique et de Recherche Opérationelle (RIRO) (8): 57-75.
  
  [2] Figueira, José; Salvatore Greco, Matthias Ehrgott (2005). Multiple Criteria Decision Analysis: State of the Art Surveys. New York: Springer Science + Business Media, Inc. ISBN 0-387-23081-5.
  
  
• Project #22: Study of economic growth and of human footprint on the planet by night light from NASA for China, India and the US. The goal is to explore how satellite images of global night lights from the years 2001 to present can be used to estimate economic activity and human footprint at the sub-regional level in the US, India and China. The night lights based estimates of economic activity will then be spatially analyzed and compared with sub-regional economic indicators where available for selected years. A major research goal will be to quantify the rate of growth at the global and regional level, and relate with other economic and demographic variables. In this way, we would like to understand the spatio-temporal dynamica of human growth on the planet.
  
  
• Project #23: In macro-economics, modellers use essentially conditions of monotonous preferences and there are essentially four fundamental variables that macro-economists consider.
  1) more wealth is better
  2) less inequality is better
  3) less inflation is better
  4) less unemployment is better.
  
  RESEARCH QUESTION: at the micro-economics level, we have shown that people are "disadvantageous inequity adverse" [1,2]. How does it renormalize into the adverse effect of INEQUALITY at the macro level? The research will consist in developing agent-based models with economic agents endowed with "disadvantageous inequity aversion" who compete for scarce resourdes in an economic set-up of production and consumption. As the agents see their wealths evolve and form a broad distribution of emergent inequalities, the question is to investigate what controls the growth of such inequalities and the macro-welfare associated with the level of inequality. Measures of macro-success will be used and novel ones will be developed if necessary.
  
  [1] M. Hetzer and D. Sornette, The co-evolution of fairness preferences and altruistic punishment evolution & human behavior, submitted 9 September 2011 (http://www.ssrn.com/abstract=1468517)
  
  [2] M. Hetzer and D. Sornette, A theory or evolution, fairness, and altruistic punishment (2011), submitted to Games and Economic Behavior, 13 September 2011
  
  
• Project #24: Comparison with theory and insights from the empirical size distribution of banks in the USA: small versus large banks and the economic meaning of deviations from Zipf law.
  
  References:
  
  [1] A. Saichev, Y. Malevergne and D. Sornette, Theory of Zipf's Law and beyond, Lecture Notes in Economics and Mathematical Systems, Volume 632, Springer (November 2009), ISBN: 978-3-642-02945-5 (http://www.springer.com/series/300) (http://arxiv.org/abs/0808.1828)
  
  [2] Y. Malevergne. A. Saichev and D. Sornette, Zipf's law and optimum long-term growth, Journal of European Economic Association (JEEA) (submitted 9 August 2011) (http://ssrn.com/abstract=1083962 and http://arxiv.org/abs/1012.0199)
  
  [3] "On the Size Distribution of Financial Institutions"
  JOHN GODDARD, University of Wales System - Bangor University
  Email: abs006@bagor.ac.uk
  HONG LIU, University of Glasgow - Glasgow Business School
  Email: hong.liu@glasgow.ac.uk
  DONAL G. MCKILLOP, Queen's University Management School
  Email: dg.mckillop@qub.ac.uk
  JOHN O.S. WILSON, University of St. Andrews
  Email: jsw7@st-andrews.ac.uk
  This study examines the firm size distribution of US financial institutions. A truncated lognormal distribution describes the size distribution, measured using assets data, of a large population of small, community-based commercial banks. The size distribution of a smaller but increasingly dominant cohort of large banks, which operate a high-volume low-cost retail banking, exhibits power-law behavior. There is a progressive increase in skewness over time, and Zipf's Law is rejected as a descriptor of the size distribution in the upper tail. By contrast, the asset size distribution of the population of credit unions conforms closely to the lognormal distribution.
  
  [4] T. Maillart, D. Sornette, S. Spaeth and G. von Krogh, Empirical Tests of Zipf's Law Mechanism In Open Source Linux Distribution, Physical Review Letters 101, 218701 (2008) (http://arxiv.org/abs/0807.0014)
  
  [5] Qunzhi Zhang and Didier Sornette, Empirical Test of the Origin of Zipf's Law in Growing Social Networks, Physica A 390, 4124-4130 (2011)
  
  [6] Ryohei Hisano and Didier Sornette, Predicted and verified evolution of power-law exponent in product market, Phys. Rev. E (April 2011; revised June 2011) (http://arxiv.org/abs/1101.5888)
  
  

• Project #25: Are crises good for entrepreneurship?
  Steve Jobs and Steve Wozniak started Apple in 1976. America was then at a low point, following the 1973-74 recession, the Arab oil embargo, the Watergate scandal, and the fall of Saigon. Other US companies that were started in bad times include General Electric, IBM, Hewlett-Packard, and Microsoft. Question: Do tough times beget a disproportionate number of great companies? If so, why?
  

  "The crisis is a catalyst for change in the technological environment. Things that we only gave half a thought to in the past are suddenly being addressed very quickly." Many innovations are geared to optimizing processes and reducing non-personnel costs. Fundamental issues are also being broached: "Without innovating", said one  manager "it won't be possible to prosper over the next few years". One characteristic shared by all the companies surveyed is that none reduced spending on research and development. On the contrary, some even increased it sharply because, as another executive emphasized, "The capital market is looking longer term – at least for now. The winners are making the most of the opportunities arising from modified investor perspectives." [John Mauldin's report, 1 Dec 2011].
  

  Study firm databases on firm creation rates and compare with financial market conditions, economic conditions to test the hypothesis that crisis and recession times are indeed disruptive and create opportunities to create the best and enduring companies.

  In other words, is this true that there is an abnormal number of great companies that are created in stressed times?

• Project #26: Liquidity risk
  Market liquidity is a hard-to-quantify and often neglected financial risk. The consequences, however, of a liquidity dry-up can have a dramatic impact on the profit and loss account of a trading book. During the 2008 financial crisis, the combination of mark-to-market accounting and a liquidity squeeze of structured credit products (like Asset Backed Securities) created black holes in investment banks' balance sheets. This mechanism was responsible for a rapid acceleration and a viral spreading of the financial crisis. Regulators and politicians had to handle this crisis with only very limited tools at hand. They fought the symptoms, changing the rules Ad Hoc, by temporarily allowing a change in accounting from mark-to-market to mark-to-model.
  
  It is clear that a much better understanding of liquidity risk is needed. We will start this challenging research subject by analysing options bid-offer spreads. We have seen that when options get more and more out of the money the bid-offer spread actually explodes. The preliminary results suggest that in an option trading book liquidity losses present a risk comparable to the classical market risk (an adverse movement of the underlying). However, where market risk can be hedged, liquidity risk cannot (at least not in our current understanding).
  
  Liquidity risk research is a fairly new domain in finance. The results, however, are very relevant and highly anticipated by the financial sector, investment  banks, regulators, central banks. The purpose of this research subject is twofold:
  1. Study the current state of affairs of liquidity risk research;
  2. Create new models of liquidity risk based on an analysis of out-of-the money options.
  
  Contact: Cauwels Peter (pcauwels@ethz.ch)
  http://www.er.ethz.ch/people/cauwelsp

• Project #27: Transient memory of emerging markets
  Financial time series are well known to exhibit a number of empirical properties, one of the most important being the absence of linear autocorrelations of returns. Uncorrelated returns ensure the absence of arbitrage opportunities and therefore the efficiency of financial markets. This property has been extensively tested over the last half century. But while for US or Europe markets, the almost absence of return correlations was consistently shown to be true, the results of tests on emerging markets are controversial. The proposed project consists in empirical testing of the presence of transient memory effects in emergent markets. As the project heavily relies on data analysis, good computational skills (knowledge of Matlab, Python ot R) is mandatory for the candidate, as well as a good level of statistics.
  
  _____________________________
  
  
  

If interested in a project, please contact

Prof. Dr. Didier Sornette
Entrepreneurial Risks, ETH Zurich
KPL F 38.2
Kreuzplatz 5, 8032 Zürich

Phone: +41 44 632 89 17
E-Mail: dsornette@ethz.ch