
[ Talk ]
Resilience to contagion in financial networks
Hamed Amini1, Rama Cont2, Andreea Minca3
( 1SFI @ EPFL, Lausanne, Switzerland, 2Université de Paris VI  VII, France and 3Université de Paris VI, France )
We derive rigorous asymptotic results for the magnitude of contagion in a large financial network and give an analytical expression for the asymptotic fraction of defaults, in terms of network characteristics. These results extend previous studies on contagion in (unweighted) random graphs to inhomogeneous directed graphs with a given degree sequence and arbitrary distribution of weights. We use our result to obtain a criterion for the resilience of a large financial network to the default of a small group of financial institutions. Our results emphasize the role played by “contagious exposures” and show that institutions that are both highly connected and overexposed are those which contribute most to network instability in case of default.
***
[ Talk ]
Statistical Physics of Labor Productivity
Hideaki Aoyama
(Kyoto University, Japan)
While mainstream economics insists that labor productivity is equal among firms and sectors based on the general equilibrium theory, the truth is far from it. Empirical study of the labor productivity in various industrial sectors of Japan and several European countries reveal that labor productivity is distributed with fat tail at the higher end, which is described very well by the power laws. This shows the complete failure of the equilibrium theory and necessitates new theoretical framework in which this empirical fact is explained. Because of this, theories of labor productivity is rapidly being developed during the last few years. In this [ Talk ], we will review these developments and present our latest theoretical framework, developed with H. Iyetomi (Niigata) and H. Yoshikawa (Tokyo), in which firms have limitation on number or workers they can employ. Using the detailbalance condition necessary for the equilibrium, we obtain general form of the productivity distribution of workers.
***
[ Talk ]
How to model asset dynamics on the basis of dependence and anomalous scaling
Fulvio Baldovin
(Dipartimento di Fisica e Sezione INFN, Universita' di Padova, Italy)
Independence and normal scaling play an important role in quantitative Finance. We show that simple generalizations of the Gaussian stability allow to construct sequences of strongly dependent random variables, whose sums have probability density functions obeying anomalous scaling in time. This sets the basis for the construction of an autoregressive scheme producing discrete trajectories in the volatilityreturn space, which correctly reproduce the most relevant stylized facts.
The resulting mathematical model allows for analytical calculations and comparisons with empirical stylized facts, on which calibration protocols can be built.
In particular, we will address the issue of how asymmetry under time reversal arises in the model and interpret the relative parameters in financial terms. Our model reveals also adequate to describe the aftershock statistics occurring after major financial crashes, and accounts for features not covered by the Omori law.
***
[ Short Talk ]
Continuous transition of social efficiencies in the stochastic strategy Minority Game
Soumyajyoti Biswas1, Asim Ghosh1, Arnab Chatterjee2, Tapan Naskar3 and Bikas K. Chakrabarti1
(1TCMP, SINP, Kolkata, India, 2 Centre de Physique Th´eorique, Marseille, France and 3Theory Division, SINP, Kolkata, India)
In a variant of the Minority Game problem [1], the agents can reach a state of maximum social efficiency, where the fluctuation between the two choices is minimum, by following a simple stochastic strategy [2,3]. By imagining a social scenario where the agents can only guess about the number of excess people in the majority, we show that as long as the guess value is sufficiently close to the reality, the system can reach a state of full efficiency or minimum fluctuation. A continuous transition to less efficient condition is observed when the guess value becomes worse. Hence, people can optimize their guess value for excess population to optimize the period of being in the majority state. We also consider the situation where a finite fraction of agents always decide completely randomly (noise trader) as opposed to the rest of the population that follow a certain strategy (chartist). For a single noise trader the system becomes fully efficient with majorityminority crossover occurring every twodays interval on average. For finite fraction of noisy traders we find a ‘rounding off’ of the transition point as generally happens in presence of noise.
References:
[1] D. Challet, M. Marsili, Y. C. Zhang, Minority Games, Oxford Univ. Press, Oxford (2005).
[2] A. Ghosh, A. Chatterjee, M. Mitra, B. K. Chakrabarti, New J. Phys. 12, 075033 (2010).
[3] D. Dhar, V. Sasidevan, B. K. Chakrabarti, Physica A 390, 3477 (2011).
***
[ Short Talk ]
Correlations in financial timeseries
Gayatri Tilak, Tamás Széll and Anirban Chakraborti
(ECP, Paris, France)
We present a perspective on correlation studies in the price timeseries of stocks, using (i) the minimum spanning tree approach and (ii) the multidimensional scaling tools. The first is a method for finding a hierarchical arrangement of stocks by using correlations of asset returns. With an appropriate metric, based on the correlation matrix, a fully connected graph is defined in which the nodes are companies, or stocks, and the “distances” between them are obtained from the corresponding correlation coefficients. The minimum spanning tree is generated from the graph by selecting the most important correlations and it is used to identify clusters of companies. The second is a set of data analysis techniques that display the structure of distancelike data as a geometrical picture, where again the “distances” between them are obtained from the corresponding correlation coefficients. We then use clustering algorithms to identify clusters of companies. Our general aim is to study the dynamical evolution of the correlations in the market.
***
[ Talk ]
Financial Crises and Systemic Risk – Dynamical Systems Approach
Youngna Choi
(Montclair State University, USA)
We build a multiagent model of the economy as a dynamical system on a compact set, and show that the market stability is closely related to the leverage of the agents. The higher the leverage the greater the reaction of market participants to changes in their wealth. This gives rise to a bifurcation mechanism, and eventually a strong dynamical instability in capital markets which is commonly referred as financial crisis. Firstly, we build a wealth dynamical system of the agents and introduce a market instability indicator. This is the spectral radius of the Jacobian matrix of the wealth dynamical system, and its size is proportional to the instability level of the financial market. Thus monitoring the indicator enables us to predict upcoming financial crises. Secondly, we show that when there is an interconnected feedback loop among the agents, the financial crisis can spread to all sectors of the economy, creating systemic risk. We use three financial crises, AsianRussian, the US subprime, and the Eurozone sovereign credit crisis, as case studies of our research to compare financial crises with no contagion, domestic agenttoagent contagion, and international sovereigntosovereign contagion.
***
[ Talk ]
Cash providers networks
JeanEdouard Colliard and Gabrielle Demange
(Paris School of Economics, France)
The role of the OTC markets in the opacity and vulnerability of nancial systems is a growing concern. The mechanism governing the transactions has remained relatively unexplored and some features, such as the fact that institutions are entering into long chains of transactions, are not well understood. We formulate a model for the building up of credit chains, motivated by OTC markets. An institution in need of a volume of cash asks the volume at a given oer rate to a number of institutions, its' neighbors, which it often deals with. These neighbors can provide some amount of cash themselves and possibly serve as intermediaries, asking to their own neighbors part of the needed volume, and so on. This denes a game in which the strategy of a player is the volume and the interest rate oered to his neighbors. We analyze the equilibrium of the game, in particular the incentives to provide liquidity and the determinants of the return oered to neighbors. A basic tradeo is that splitting the volume with one's neighbors saves on lending costs under a convexity assumption (the cost here can be interpreted as a risk measure or as the cost of the required capital associated to the lent amount), but at the risk that the transaction ultimately will not be performed. We rst consider a simple situation, with two layers of transactions at most, to uncover some basic properties and perform comparative statics exercise with respect to the volume asked and the network structure. The general case is studied mainly through simulations. Our model allows us to tackle questions related to the impact of the cost structure (in particular the introduction of a xed cost per transaction) on the equilibrium transactions or on the stability of the system to an exogenous shock (the default propagation in case of an initial default).
***
[ Talk ]
Real trader behaviour in good and bad times
Damien Challet
(University of Fribourg, Switzerland)
Using brokerage data, this [ Talk ] will dig deeper and deeper into trader dynamics, starting from global measures of behaviour down to individual reaction to or anticipation of large price changes. It will also report about the heterogeneity of risk profiles of traders. Finally, it will discuss what the availability of such microscopic data means for agentbased modelling.
***
[ Talk ]
No Title
Rama Cont
(Université de Paris VI  VII, France)
***
[ Short Talk ]
Study of systemic risk involved in mutual funds
Kishore Chandra Dash
(Neelashaila Mahavidyalaya, Rourkela, India)
Systemic risk, may be the risk that contaminated to a whole system, consisting of many interacting agents that fail one after another. These agents, in an economic context, could be firms, banks, funds, or other financial institutions. Systemic risk is a macroscopic property of a system which emerges due to the nonlinear interaction of agents on a microscopic level. A stock market itself is a system in which there are many subsystems, like Dowjones, Nifty, Sensex, Nasdaq, Nikkei and other market indices in global perspective. In Indian market, subsystems may be like sensex, nifty, bse200, bankex, smallcap index, midcap index, S&P CNX 500 and many others. Similarly there are many mutual funds, which have their own portfolio of different stocks, bonds etc. I have in this paper attemped to study the systemic risk involved in a fund as a macroscopic object with regards to its microscopic components as different stocks in its portfolio. It is observed that fund managers do manage to reduce the systemic risk just like we take precautions to control the spread of epidemic.
***
[ Talk ]
Correlations and multifractality in global financial indices
Nivedita Deo
(Dept. Of Physics, University of Delhi, India)
***
[ Talk ]
Dependence of the optimal mixed strategies in the Minority game on the timehorizon of agents
V. Sasidevan and Deepak Dhar
(Department of Theoretical Physics, Tata Institute of Fundamental Research, Homi Bhabha Road, Mumbai400005, India)
We study a variation of the minority game, in which N agents, with N odd, choose repeatedly between two alternatives, and the minority group is the winner. The agents only have the information about number that chose the first option in the past, and cannot communicate with each other. We have shown earlier that if agents use a mixed ’winstayloseshift’ strategy, the steady state reached attains a much more efficient utilization of resources than if they use deterministic strategies [ Physica A 390 (2011), 3477]. However, this strategy can lead to steady states where the payoff are very unevenly spread amonst the agents. When the winning group has exactly (N −1)/2 agents, all agents, using this mixed strategy, and trying to maximize their next day’s gain, would choose the same option on the next day. These states may be called trapping states in which losers remain losers for all future times. We discuss the case where agents try to optimize their individual expected payoffs for the next D days, with D > 1. We show that no trapping states are encountered for D > 1.
***
[ Talk ]
Critical transitions in economics and finance? Some preliminary empirical results
Cees Diks
(University of Amsterdam, The Netherlands)
In our joint project with De Nederlandsche Bank we investigate whether the notion of critical transitions from complex systems theory can help identify financial and economic crises ahead of time. A direct application of a modelfree statistical moving window time series analysis approach, as suggested by Scheffer et al. (2009), turns out to have only limited practical value. In an attempt to understand why this is the case, a heterogeneous agent model is estimated on housing market data. The model estimates, in particular the estimated noise level, provide some insight into why a modelfree approach is difficult, and suggests the use of alternative techniques to forecast critical transitions, such as those currently under development in climatology.
***
[ Talk ]
Chained financial failures at nationwide scale in Japan
Yoshi Fujiwara
(Graduate School of Simulation Studies, University of Hyogo)
I will talk about recent studies based on real data of propagation of financial failures in the past financial crises and the present one due to the earthquake at nationwide scales in Japan. The Bank of Japan and several leading credit research agencies in Tokyo have accumulated a huge amount of data regarding interbank, banksfirms, suppliercustomer and so on for two decades or more. By using such largescale data, we will measure the actually occurred propagation of financial distress on the real data of largescale economic networks comprising of firms, banks, and their relationships at the order of millions and even more. I will also mention about uncovering the structure, temporal change and fragile locations of such networks.
***
[ Talk ]
SEZ and Zipf's Law: An empirical investigation
Kausik Gangopadhyay1 and Banasri Basu2
(1IIM Kozhikode, India and 2ISI Kolkata, India)
Special Economic Zones (SEZ) are common accross countries. SEZs create a country within a country. Do the city size distribution vary in the presence of SEZs? Gangopadhyay and Basu (2010) hypothesize that SEZs may change the city size distribution in case of China. We further examine this phenomenon with data from several countries.
***
[ Talk ]
Systemic risk, macroprudential supervision and regulation
Philipp Hartmann
(European Central Bank)
This presentation deals with economic foundations for macroeconomic supervision and regulation. It first characterises systemic financial risk and surveys the broad analytical approaches for identifying and assessing it. It then discusses the most widely debated macroprudential regulatory instruments to contain systemic risks. Going along reference will be made to areas and approaches that would benefit from combining economics with methodologies from the natural sciences.
***
[ Talk ]
Statistical Mechanics of Labor Markets
Junichi Inoue
(Hokkaido University, Japan)
On the basis of statistical mechanics, we introduce a probabilistic model of labor markets for university graduates in Japan. In order to make a model of the market efficiently, we shall take into account the following assumptions. Namely, each company fixes the number of opening positions for newcomers and the number itself is independent of business year.
The ability of gathering newcomers depends on the result of matching process in past business years. This fact means that the ability of the company is weakened if the company did not make their quota or the company gathered applicants too much over the quota. All laborers, in particular, university graduates who are looking for their jobs can access the public information about the ranking of companies. By assuming the above essential key points, we construct the energy function of each company and describe the probability that an arbitrary company gets students at each business year by a BoltzmannGibbs distribution.
We evaluate the relevant physical quantities such as the employment rate. We find that the system undergoes a phase transition from the `good employment phase' to `poor employment phase' with spontaneous symmetry breaking when one controls the degree of importance for the ranking. With
the assistance of a chaotic map for the inflation rate given by Neugart (2004), we also attempt to draw the socalled Phillips curve.
Relationship between our model and several classes of the urn model willbe discussed.
References:
[1] M. Aoki and H. Yoshikawa, "Reconstructing Macroeconomics: A Perspective from Statistical Physics and Combinatorial Stochastic Processes", Cambridge University Press (2006).
[2] T. Boeri and J. van Ours, "The Economics of Imperfect Labor Markets", Princeton University Press (2008).
[3] R. Gabriele, "Labor Market Dynamics and Institution: An Evolutionary Approach", Working Paper in Laboratory of Economics and Management Sant'Anna School of Advances Studies, Pisa, Italy (2002).
[4] G. Fagiolo, G. Dosi and R. Gabriele, Advances in Complex System, Vol. 7, No.2, pp. 157186 (2004).
[5] M. Neugart, Journal of Economic Behavior and Optimization, Vo. 53, pp. 193213 (2004).
[6] U. Garibaldi and E. Scalas, "Finitary Probabilistic Methods in Econophyscs", Cambridge University Press (2010).
***
[ Talk ]
Firm size distribution and nonGibrat's law observed in the midscale range
Atushi Ishikawa
(Kanazawa Gakuin University, Japan)
In Econophysics, it is well known that firm size distributions follow power law in the largescale range. Here, firm sizes denote sales, profits, assets, the number of employees, and so forth. The power law is known as Pareto's law. In the midscale range, it is also hypothesized that firm sizes follow the lognormal distribution. In many studies, it has been confirmed that these distributions accurately fit empirical data. In EconophysicsKolkata I (2005), Prof. Fujiwara presented an emergence of the power law by using two laws, observed in the largescale range of empirical data, such as Gibrat's law and detailed balance. Detailed balance is invariance of the joint PDF (probability density function) under the exchange of variables at the two successive points in time. Gibrat's law means that the conditional PDF of the growth rate of the variables does not depend on the initial value. After that, by extending Gibrat's law in the largescale range to nonGibrat's law in the midscale range, we have shown an emergence of the lognormal distribution under detailed balance valid not only in the largescale range but also in the midscale range. Here, nonGibrat's law describes the dependence of the initial value on the conditional PDF of the growth rate. In this presentation, we show the validity of above discussions by using firm size data in Japan over 30 years' period (19802009). In the data analyses, we confirm as follows :
[1] Powerlaw exponents in the largescale range and standard deviations of lognormal distributions in the midscale range hardly change over 30 years' period even in the bubble term.
[2] The growthrate distribution of positive profits obeys the following property in the midscale range. The probability of positive growth decreases and the probability of negative growth increases symmetrically as the initial value increases. We call this first nonGibrat's law.
[3] The growthrate distribution of sales and the number of employees obeys the following property in the midscale range. The probability of positive growth decreases as the initial value increases. However, the probability of negative growth hardly changes as the initial value increases. We call this second nonGibrat's law.
[4] In both 2) and 3) cases, nonGibrat's law lead to the lognormal distribution under detailed balance. This is verified analytically and empirically.
[5] On the one hand, the shape of the growthrate distribution of positive profits is linear in loglog scale. On the other hand, the growthrate distributions of scale and the number of employees have wider tails than those of profits in loglog scale. This difference is closely related to the difference between two kinds of nonGibrat's laws.
The results in this presentation should be carefully considered in the understanding and control of systemic risk.
***
[ Talk ]
Hidden Correlation Structure of the Tokyo Stock Exchange Market
Hirosh Iyetomi
(Department of Physics, Niigata University, Japan)
We revisit correlation structure in the Tokyo Stock Exchange market from a network viewpoint. The correlation matrix of stock price changes, purified by the random matrix theory, is regarded as an adjacency matrix for a network. In the stock network thus constructed, each pair of nodes representing the stocks are connected by a link with weight quantified by the correlation coefficient between them; the weighted links can have even negative sign. By minimizing frustration among the nodes, we find that the network is decomposed into four groups. The stock prices comove almost perfectly inside the groups and move oppositely across the groups. Especially three of the groups are strongly anticorrelated to each other and the remaining is rather independent. Such a frustrated triangle group structure of the stocks may give rise to complicated market behavior. Results of detailed analysis of the communities detected here will be also given.
***
[ Talk ]
Conservative selforganized extremal model for wealth distribution
S. S. Manna
(Satyendra Nath Bose National Centre for Basic Sciences, India)
We present a detailed numerical analysis of the modified version of a conservative selforganized extremal model introduced by Pianegonda et. al. for the distribution of wealth of the people in a society. Here the trading process has been modified by the stochastic bipartite trading rule. More specifically in a trade one of the agents is necessarily the one with the globally minimal value of wealth, the other one being selected randomly from the neighbors of the first agent. The pair of agents then randomly reshuffle their entire amount of wealth without saving. This model has most of the characteristics similar to the selforganized critical BakSneppen model of evolutionary dynamics. Numerical estimates of a number of critical exponents indicate this model is likely to belong to a new universality class different from the well known models in the literature. In addition the persistence time, which is the time interval between two successive updates of wealth of an agent has been observed to have a nontrivial power law distribution. An opposite version of the model has also been studied where the agent with maximal wealth is selected instead of the one with minimal wealth, which however, exhibits similar behavior as the minimal model.
***
[ Talk ]
Predatory trading, selective networking and risk minimisation: how to (b)eat the competition
Anita Mehta
(Satyendra Nath Bose National Centre for Basic Sciences, India)
Systemic risk is enhanced in a scenario where there are multiple predators, according to Brunnermeier and Pedersen, 2005. We present a model (Luck and Mehta, 2005; Mehta, Majumdar and Luck, 2005) that examines just such a scenario. In our model, individual traders interact with each other in the presence of a central reserve, which itself can add to, or deplete, their savings. Thus, traders who have less than a threshold amount of wealth at t=0 are doomed to fail relatively quickly even if they are isolated, as 'inflation' eats into their wealth. Those who are wealthier than threshold live on, and survive by preying on each other (while their wealth is still being added to, or depleted by, the reserve); the dynamics of this multiple predator/prey scenario are slower. In fact the dynamics of our model have been shown to be glassy, with the fast and slow phases referred to above being wellseparated.
The main results of our model are as follows: when the model is solved on finitedimensional lattices with nearestneighbour interactions, the wealth is concentrated in a finite number of survivors in the asymptotic limit, who have 'consumed' the competition in their immediate vicinity. By contrast, in the mean field limit, there is at most one survivor asymptotically, and the systemic collapse is complete.
Most realistic interaction environments, however, are somewhere between mean field and finitedimensional; this is the main motivation to solve our model on networks, where our results show the expected trend (Haldane and May, 2011) that increasing complexity decreases systemic stability, as there are fewer and fewer survivors. Our results (Thyagu and Mehta, 2011) interpolate between meanfield and finitedimensional models in this sense. In order to keep the network aspect but enhance systemic stability, we therefore investigate scenarios where selective networking can enhance survival rates of arbitrarily chosen traders. Our conclusions show that networking with 'doomed' traders is the most riskfree scenario, and that if a trader is to network with peers, it is far better to do so with those who have less intrinsic wealth than itself to ensure individual, and perhaps systemic stability.
References:
[1] M K Brunnermeier and L H Pedersen, "Predatory Trading", Journal of Finance, LX, 18251864 (2005)
[2] J.M. Luck and Anita Mehta, "A deterministic model of competitive cluster growth : glassy dynamics, metastability and pattern formation", European Physics Journal B, 44, 7992 (2005)
[3] Anita Mehta, A. S. Majumdar and J. M. Luck, "How the rich get richer", pp. 199204 in 'Econophysics of Wealth Distributions' edited by A. Chatterjee et al, SpringerVerlag Italia (2005)
[4] Andrew G. Haldane & Robert M. May, "Systemic risk in banking ecosystems", Nature 469,351–355 (20 January 2011)
[5] N. Nirmal Thyagu and Anita Mehta, "Competitive cluster growth on networks: complex dynamics and survival strategies", Physica A 390, 14581473 (2011)
***
[ Talk ]
Characterizing price index behavior through fluctuation dynamics
Prasanta K. Panigrahi1, P. Manimaran2 and S. Ghosh3
(1IISER Kolkata, India, 2C R Rao Advanced Institute of Mathematics, Statistics and Computer Science, Hyderabad, India and 3University of KwaZuluNatal, Private Bag X54001, Durban 4000, South Africa)
We study the nature of fluctuations in variety of price indices involving companies listed in BSE and NASDAQ. The fluctuations at multiple scales are extracted through the use of wavelets belonging to Daubechies basis. The fact that these basis sets satisfy vanishing moments conditions makes them ideal to extract local polynomial trends, through the low pass or 'average coefficients'. Subtracting the trends from the original time series yields the fluctuations, at different scales, depending on the level of lowpass coefficients used for finding the 'average behavior'. The fluctuations are then studied using wavelet based multifractal detrended fluctuation analysis to analyze their selfsimilar and nonstatistical properties. Due to the multifractality of such time series, they deviate from Gaussian behavior in different frequency regimes. Their departure from k−3 behavior in such regimes is also analyzed. These deviations and nonstatistical properties of the fluctuations can be instrumental in throwing significant light on the dynamics of financial markets.
***
[ Talk ]
The economic complexity of countries and products
Luciano Pietronero
(ISCCNR and Univ. Sapienza, Roma, Italy)
We discuss a recent new approach to the complexity of countries and products in the spirit of the recent papers by Hidalgo and Hausmann (PNAS 2009). The basic information is represented by the matrix of countries and exported products. The standard economic analysis is essentially based on the GDP but the diversification of this into a series of different products provides an additional element of fitness in the spirit of biodiversification in a fluctuating enviroment. In fact the idea that specialization of countries towards certain specific products is considered as optimal in the standard analysis, but this could only be valid in a static situation. The strongly dynamical situation of the world market suggests that flexibility and adaptability are also important elements. The basic idea is to introduce a metric for the value of the diversification, in particular a Fitness parameter for each country which is able to take into consideration this effect. Such an analysis, selfconsistently also leads to a ranking of the Quality of the products. These concepts are implemented with the use of statistical concepts inspired to the page rank (Google) problem may lead to a novel classification for the fitness of the countries and the quality of products which adds new information with respect to the standard economic analysis which has rarely addresses this point quantitatively. This information can be used in various ways. The direct comparison of the Fitness with the country GDP gives an assessment of the non expressed potential of the country. Also for each country it is possible to define the quality of the products exported and how competitive is this country with respect to the other countries which produce the same product. Finally it is possible to make a planning for the optimal development of a country by considering its potential for adding a new product. The addition of product quality to the product space adds an important new element also in this context.
***
[ Short Talk ]
No Title
H.K. Pradhan
(XLRI Jamshedpur)
***
[ Talk ]
Role of saving in gains and losses in kinetic models of wealth exchange
Parongama Sen
(Department of Physics, University of Calcutta, India)
The distribution of wealth gained or lost after reaching equilibrium for agents in kinetic models of wealth exchange shows that it depends strongly on the saving factor of agents. For uniform saving, the distribution is symmetric while for saving with a distribution, it is not. The sequence of gains and losses of individual agents are mapped to a walk which also shows features which depend on the saving factor. From the analysis of the walk, the probability of gain or loss is obtained as a function of the saving factors of the interacting agents. In general, an agent is likely to gain when he/she interacts with a richer agent. Another intriguing feature which is found is the antipersistent effect in the dynamics which implies the tendency of agents to make a gain immediately after a loss and vice versa. This tendency becomes lesser for agents with higher saving factors.
***
[ Talk ]
Are complex financial systems unstable ? Analyzing cascading failures in banking networks
Sitabhra Sinha
(The Institute of Mathematical Sciences, Chennai, India )
(in collaboration with Maximilian Thess and Sheri Markose)
The recent worldwide financial crisis has focused attention on the need to analyze systemic risk in complex economic networks. We investigate the problem of instability of such systems in the background of the general theory of local stability of arbitrary equilibria in complex networks. We focus on the data of US interbank transactions to empirically investigate the structure of the network and how the topology of connections affects the risk of cascading failures. We observe that the banking network has a prominent coreperiphery organization that can resist largescale collapse when subjected to individual bank defaults (local perturbations) but is vulnerable to systemwide meltdown as a result of the accompanying liquidity crisis (global perturbation).
***
[ Short Talk ]
Cluster analysis and pattern recognition of crosscorrelations in typical Japanese stocks by means of multidimensional scaling
Takero Ibuki, Tatsuya Yamamura, Sei Suzuki and Junichi Inoue
(AoyamaGakuin University, Japan)
We investigate crosscorrelations between typical Japanese stocks collected through Yahoo!Japan website (http://finance.yahoo.co.jp/). By making use of multidimensional scaling (MDS) for the crosscorrelation matrices, we draw twodimensional scattered plots in which each point corresponds to each stock. To analyse these data plots, we utilize the mixture of Gaussians to fit the data set by several Gaussian densities. By minimizing the socalled Akaike Information Criterion (AIC) with respect to parameters in the mixture, we specify the best plausible mixture of Gaussians. It might be naturally assumed that all the twodimensional Gaussians shrink into a single small region when some economic crisis takes place.
This research project just started and is still ongoing, however, we will present the outline of this project in the conference and the justification of above assumption might be numerically examined for the empirical Japanese stock data, for instance, those around 11 March 2011.
***
[ Poster ]
No Title
Paritosh Bhattacharya
(CEM Kolaghat, India)
***
[ Poster ]
Wavelet Variance and correlation analysis of Foreign Exchange Rates of SAARC Countries using MultiResolution Analysis
Abhijit Betal and A. N. Sekar Iyengar
(Plasma Physics Division, Saha Institute of Nuclear Physics, India)
With increasing globalization, the Foreign Exchange (FX) market has been increasing in volume and become more competitive and volatile. In this situation, we have analyzed the foreign exchange rates of South Asian Association of Regional Cooperation (SAARC) countries to understand their basic characteristics and the correlation between them. We have applied the Wavelet Multiresolution Analysis to decompose the wavelet coefficients of foreign exchange rates of SAARC countries and to calculate the variance and to interpret the correlation between them. In this work have we used the Indian Rupee as the base currency and the currencies of other SAARC countries as the quote currency. The analysis of volatilities and trend suggest that after 2009, the rates are almost stable, and at higher wavelet scales the variance values are higher. On the other hand, the wavelet correlation analysis indicates that the NPR (Nepalese rupee) and BTN (Bhutanese Ngultrum) have very little economic relation with the other SAARC countries but PKR (Pakistan rupee) maintains uniform economic relationship with all the SAARC countries. The AFN (Afghani), MVR (Maldivian rufiyaa) and BDT (Bangladesh taka) have very strong economic relationship with them each other.
***
[ Poster ]
No Title
Tanmoy Das
(ISI, India)
***
[ Poster ]
Contagion in Default Protectd Bank Networks
Saumya Deojain
(Lady Shri Ram College for Women, University of Delhi, India)
The interbanking networks that have been studied to characterize systemic risk have largely dealt with interbank loaning. However, over the past few years, interbank loaning has evolved and many new financial instruments have altered the relationships (or links) between banks and hence complicated the study of systemic risk. After the 2008 crisis it is clear that if we want to study systemic risk through network analysis, we must also study some of these instrumentsspecificallycredit default swaps (CDSs). We are looking at the hedging function of CDSs, that is when it is used as a protection that a bank buys from another bank in case of a default of a loan. In our model every bank has a certain number of loans to various firms each of which has a certain probability of defaulting. Banks purchase CDSs from other banks for coverage against these defaults. We discuss a model of a simple network of banks in which links between banks are these CDS transactions. We present results from simulations regarding how the contagion of defaults spreads in the network. We try to characterize the risk of default of a bank through various kinds of exogenous defaults of investments.
***
[ Poster ]
No Title
Manisankar Dhara
(xlri Jamshedpur, India)
***
[ Poster ]
Kinetic models of wealth exchange : A walk scenario
Sanchari Goswami
(Department of Physics , University of Calcutta)
We explored the nature of transactions made in some kinetic models of wealth exchange in depth. Using an equivalent picture of a one dimensional walk in an abstract space for gains and losses, it is found that there is a tendency of individuals to make a gain (loss) immediately after a loss (gain). This antipersistence effect is in fact compatible with human psychology. Moreover we find that if there is no saving factor, this effect is maximum and decreases with saving. This is perhaps in tune with the human feeling of security associated with the saving factor. In the CCM model, where the saving propensity λ is randomly distributed, the antipersistence effect occurs with a simultaneous bias. This simultaneous bias depends on λ. There is also no antipersistence effect in the simulated walk of a single agent of CCM type, This can be shown from the comparative behaviour of the probability of direction reversal of the two models, when the choice of the second agent is made arbitrary. This shows that although we consider only gains and losses for generating the walk in a GLS, the form of the money distribution function is crucial.
In the CCM model, the antipersistence effect makes it distinct from an ordinary biased random walk. Similarly, the antipersistence effect makes the CC model different from an ordinary random walk. The difference in CC and CCM walks turns out to be simply the presence of a bias in the latter. This bias appears as a result of the small positive correlations remaining at large times in the CCM model. The different calculations made in this work shows once again that an agent with λ = λ∗ ≃ 0.469 in the CCM model is identical to a CC walker.
***
[ Poster ]
No Title
Deepak Kumar
(IIT Kanpur, India)
***
[ Poster ]
No Title
Sunil Kumar
(University of Delhi, Delhi110007, India)
Economic systems such as financial markets are among the most intriguing and fascinated complex systems that might be investigated. A widely accepted belief in financial theory is that time series of asset prices are in predictable. This belief is the cornerstone of the description of price dynamics as stochastic processes. Specifically, chaos theory has shown that unpredictable time series can arise from deterministic nonlinear systems. One common theme encountered in these research areas is the time correlation of a financial series. These are some facts which motivated me to work in the field of Econophysics. With the collaboration of my supervisor I have find the multifractal properties in the Indian financial market by using the MultifractalDetrended Fluctuation Analysis (MFDFA). We investigate that both the US (S&P 500) and Indian (BSE & NSE) exhibit the multifractal properties. By comparing the MFDFA results of the original series to those for the shuffled series, we find that the multifractality is due to the contributions of longrange correlations as well as the broad probability density function. The financial markets studied here are compared with the Binomial Multifractal Model (BMFM) and have a smaller multifractal strength than the BMFM.
***
[ Poster ]
Economics, Thermodynamical Probablity and Ecological Augmentation
Subhamoy Singha Roy
(Department of Physics , JIS College of Engineering, WBUT, Kalyani, Nadia741235, India)
I study theoretically in this paper the appropriate modeling of the interface between financial development and environmental evils, summarizes beneath which situation limitless economic growth with limited accepted property is possible, and describes how sustainable development can be achieved. Now neoclassical financial side has only recently measured the crisis of sustainability and how to supply for the wellbeing of potential generations given ecological constraints. It synthesizes the outcome from various ecological growth models. In recent times the ecologists and ecological economists, however, quarrel that the constraints will join. The discuss centres on the importance of thermodynamical probablity to the economics of source use: beside the point in the neoclassical outlook, applicable in the ecological. The corporeal measurement and the value dimension of economic activity have to be treated as theoretically different. In addition of natural variables is surrounded due to biophysical laws (notably, the law of thermodynamical probablity ). The paper thoroughly argues the properties of knowledge and as a result to micro technology and control, deflation a necessary supposition of neoclassical financial side and establishing the relevance of thermodynamical probablity to financial growth theories of resource utilize.
***