-
Econophysics Colloquium 2013
&
Asia Pacific Econophysics Conference 2013
Date: July 29-31, 2013
Venue: POSTECH, Pohang, Korea
Sponsored by
-
1
Invited Speakers
Siew Ann Cheong (Nanyang Technological University, Singapore)
Hiroshi Iyetomi (Niigata University, Japan)
Hawoong Jeong (KAIST, Korea)
Woo-Sung Jung (POSTECH, Korea)
Taisei Kaizoji (International Christian University, Japan)
Beom Jun Kim (Sungkyunkwan University, Korea)
Kyungsik Kim (Pukyong National University, Korea)
Seunghwan Kim (POSTECH, Korea)
Yong-Cheol Kim (University of Wisconsin, USA)
Okyu Kwon (National Institute for Mathematical Sciences, Korea)
Jae Woo Lee (Inha University, Korea)
Sai-Ping Li (Academia Sinica, Taiwan)
Thomas Lux (University of Kiel, Germany)
Rosario Nunzio Mantegna (University of Palermo, Italy)
Tiziana Di Matteo (King's College London, UK)
Gabjin Oh (Chosun University, Korea)
Tobias Preis (University of Warwick, UK)
Aki-Hiro Sato (Kyoto University, Japan)
Seung-Woo Son (Hanyang University, Korea)
Misako Takayasu (Tokyo Institute of Technology, Japan)
Stefan Thurner (Medical University of Vienna, Austria)
Yougui Wang (Beijing Normal University, China)`
Victor Yakovenko (University of Maryland, USA)
Wei-Xing Zhou (East China University of Science and Technology, China)
Scientific Committee
Tomaso Aste (University College London, London)
Damiano Brigo (Imperial College, London)
Carl Chiarella (University of Technology, Sydney)
Guido Caldarelli (Universita' La Sapienza, Rome)
Shu-Heng Chen (National Chengchi University, Taipei)
Siew Ann Cheong (Nanyang Technological University, Singapore)
Carmen Costea (Spiru Haret University, Bucharest)
Michel Dacorogna (SCOR SE Zurich Branch, Zurich)
Tiziana Di Matteo (King's College London, London)
J.Doyne Farmer (Santa Fe Institute, Santa Fe)
Mauro Gallegati (Universita Politecnica delle Marche, Ancona)
Giulia Iori (City University, London)
-
2
Woo-Sung Jung (POSTECH, Pohang)
Taisei Kaizoji (International Christian University, Tokyo)
Janos Kertesz (Technical University of Budapest, Budapest)
Seunghwan Kim (POSTECH, Pohang)
Alan Kirman (G.R.E.Q.A.M, Aix-en-Provence)
Sai-Ping Li (Academia Sinica, Taipei)
Fabrizio Lillo (University of Palermo, Palermo)
Thomas Lux (University of Kiel, Kiel)
Rosario N. Mantegna (Universita' di Palermo, Palermo)
Luciano Pietronero (University of Rome, Roma)
Peter Richmond (Trinity College, Dublin)
Aki-Hiro Sato (Kyoto University, Kyoto)
Enrico Scalas (Universita' del Piemonte Orientale, Alessandria)
Frank Schweitzer (ETH, Zurich)
Eugene H. Stanley (Boston University, Boston)
Hideki Takayasu (Sony Computer Science, Tokyo)
Stefan Thurner (Medical University of Vienna, Vienna)
Constantino Tsallis (CBPF, Rio de Janeiro and Santa Fe Institute, Santa Fe)
Yougui Wang (Beijing Normal University, Beijing)
Yi-Cheng Zhang (University of Fribourg, Fribourg)
Local Organizing Committee
Hawoong Jeong (KAIST)
Jaeseung Jeong (KAIST)
Woo-Sung Jung (POSTECH, Secretary)
Hyungtae Kook (Gachon niversity)
Beom Jun Kim (Sungkyunkwan University)
Kyungsik Kim (Pukyong National University)
Seunghwan Kim (POSTECH, Chair)
Okyu Kwon (National Institute for Mathematical Sciences)
Jae Woo Lee (Inha University)
Gabjin Oh (Chosun University)
Seung-Woo Son (Hanyang University)
Soon-Hyung Yook (Kyung Hee University)
-
3
Guidance for Participants
A. Venue
Invited Talks:
Auditorium, 1st floor, POSCO International Center
Parallel session (Monday afternoon, Tuesday afternoon):
A, E: Auditorium, 1st floor, POSCO International Center
B, F: International Conference Room, 1st floor, POSCO International Center
Parallel session (Tuesday morning):
C: Physics Building (#3) - 109
D: Physics Building (#3) – 111
B. Meals
Breakfast:
Light snack will be provided at 9:00 (Mon) and at 8:30 (Tue & Wed)
Place: The front of the Auditorium, 1st floor, POSCO International Center
Lunch:
Invited speakers: D’medley, 2nd floor, POSCO International Center
Others: Wisdom Restaurant, 2nd floor, Jigok Community Center (Meal ticket will be provided)
Welcoming Reception:
Time: 19:00~21:00, Jul. 29(Mon)
Place: Phoenix, 5th floor, POSCO International Center
Standing Lunch:
Time: 12:30~14:00, Jul. 30(Tue)
Place: APCTP Common Room (501), 5th floor, Hogil Kim Memorial Hall
Banquet:
Time: 19:00~21:00, Jul. 30(Tue)
Place: Grand Ballroom, 2nd floor, POSCO International Center
C. Accommodation
POSCO International Center
(1) Check-in: after 2 p.m. / -Check-out: before 12 p.m.
(2) Wireless Internet: POSCO_IC
Student Dormitory Bldg. 20 & 21 (for women)
(1) Check-in: after 1 p.m. / -Check-out: before 10 a.m.
(2) You should pick up an entry card for dormitory building during registration. It must be
returned before departure.
-
4
(3) Cleaning equipments are NOT provided at dormitory room. Please bring your own
towels, toothbrush, shampoo, etc.
(4) Laundry: Free laundry machines are available on the 1st floor of the building for use
from 7:00 am to 22:00. Laundry room on the 1st floor is for male and the room on the
11th floor is female
(5) Dormitory Regulations strictly prohibit smoking and drinking.
(6) How to use door lock at Dorm Room.
- Pin number is 1234✻. If you need to change the pin number during your stay, please
refer to the following guidance. *Please be sure to reset the Pin Number to 1234✻
before check-out.
-To change the pin number: Press the button ● for 2 sec. → 1234✻→ Press the new pin
number✻→Press the new pin number✻
- To reset: Press the button ● for 10 sec. → press 4560852580✻
(7) Wireless Internet is available.
SSID: postech
ID: visit_34253
PW: a624141
D. Travel Information
Call Taxi
Choose and call one of the following taxi companies. It will be easier to ask a concierge of
your hotel or a Staff of APCTP to call a taxi on behalf of you.
- Haemaji Call Taxi: 054-283-8282
- Yoogil Call Taxi: 054-282-6161
- Pos Call Taxi: 054-252-1111
From Pohang Intercity Bus Terminal to the Incheon / Gimhae Airport
DESTINATION DEPARTURE TIME TRAVEL TIME FARE
Incheon Airport 05:30, 08:20, 11:00
Night 23:30, 01:00, 02:30 5hrs 30mins
KRW44,300 (Night-KRW48,700)
Gimhae Airport
05:00, 05:40, 06:20, 07:20, 08:20, 09:20, 10:20, 11:20, 12:20, 13:20, 14:20, 15:20, 16:20, 17:20, 18:20, 19:20
2hrs KRW11,000
From Pohang Airport to Gimpo(Seoul) Airport
Departure Arrival Flight
10:10 11:25 15:40 17:40
11:00 12:15 16:30 18:30
OZ8332 KE1532 OZ8334 KE1534
*KE: Korean Air http://www.koreanair.com *OZ: Asiana Air http://www.flyasiana.com
http://www.koreanair.com/
-
5
From Pohang to Seoul by Train
1) Pohang Intercity Bus Terminal → Singyeongju Station
Please take a Limousine Bus from the Bus Terminal to Singyeongju station. The Bus starts
at 5:00(Sat.& Sun. excluded), 6:00, 7:00, 7:50, 8:10, 8:50, 9:45, 10:10, 11:50, 12:20, 13:50,
15:00, 16:00, 17:00, 17:30, 18:00, 19:00, 20:00, 20:45, 22:30 and 23:30. The cost is KRW 5,000
and it takes about 40 min. from Pohang to Singyeongju Station.
2) Singyeongju Station → Seoul Station
Please refer to below website for the timetable of KTX (Korea Train Express).
http://info.korail.com/2007/eng/eng_index.jsp
3) Seoul Station → Incheon Airport, please take Airport railroad.
For the airport railroad, please refer to the website, http://english.arex.or.kr/jsp/eng/index.jsp.
E. Excursion:
Departure: 13:30, Jul. 31(Wed), POSCO International Center
http://english.arex.or.kr/jsp/eng/index.jsp
-
6
The available Cafeteria and Restaurant in POSTECH
Jigok Community Center
Location Available time
Freedom
(Student Dining Hall) 2nd floor
Breakfast: 7:30 to 9:30
Lunch: 11:30 to 13:30
Dinner: 17:30 to 19:00
Wisdom
Faculty & Staff
Dining Hall 2nd floor
Lunch: 11:50 to 13:00
(closed on Saturday & Sunday)
Cafeteria 2nd floor
Breakfast: 8:00 to 10:30
Lunch: 11:30 to 15:00
Dinner: 16:00 to 20:30
Yeonji (Korean Restaurant) 1st floor 13:00 to 20:00 (closed on Sunday)
Burger King 1st floor 10:00 to 22:00
Campus Store 1st floor 8:00 to 2:00
POSCO International Center
Location Available time
D’medley
(Buffer Restaurant) 2nd floor 7:00 to 20:30
Phoenix
(Chinese Restaurant) 5th floor
12:00 to 20:30
(closed on Sunday)
Student Union
Location Available time
OASIS 1st floor 8:00 to 19:30
Campus Store 1st floor 8:00 to 21:00
-
7
POSTECH Campus Map
Physics Building
(Science Bldg. #3)
POSCO International Center
Hogil Kim Memorial Hall
(APCTP Common Room)
Jigok Community Center
(Yeonji, Wisdom)
-
8
POSTECH Dormitory Map
-
9
Korean Expression
1. Please call a taxi on behalf of me.
콜택시를 불러 주세요.
2. Please take me to POSCO International center.
포항공대 국제관에 내려주세요.
3. Please take me to Dormitory at POSTECH.
포항공대 기숙사에 내려주세요.
4. Please take me to APCTP.
포항공대 무은재 기념관에 내려주세요.
(포항공대 내 국제관 올라가는 길인 무은재길 정면)
5. Please take me to the Pohang Intercity Bus Terminal.
포항시외버스 터미널에 내려주세요.
6. Please take me to the Pohang Express Bus Terminal.
포항고속버스 터미널에 내려주세요.
7. Please take me to Pohang Airport.
포항공항에 내려주세요.
8. I would like to go to Singyeongju Station.
신경주역에 가려고 합니다.
9. I would like to go to Gimhae Airport.
김해공항에 가려고 합니다.
10. I would like to go to Incheon Intl. airport.
인천공항에 가려고 합니다.
-
10
Program
-
11
Day 1
Invited speaker
Time Speaker Title
Chair : Beom Jun Kim
11:00 ~ 11:30 Rosario Nunzio
Mantegna
Statistically validated networks of market members
trading at the LSE electronic and dealers' market
11:30 ~ 12:00 Tobias Preis Quantifying Economic Behavior Using Big Data
12:00 ~ 12:30 Hawoong Jeong Google knows (almost) everything! - Big-data and
Network Science
12:30 ~ 1:30 Lunch
Chair : Victor Yakovenko
1:30 ~ 2:00 Tiziana Di Matteo Spread of risk across financial markets: better to invest
in the peripheries
2:00 ~ 2:30 Hiroshi Iyetomi Frustrated Correlation Structures Embedded in Well-
Developed Stock Markets
2:30 ~ 3:00 Yougui Wang Incorporating Debt into the Modern Macroeconomics
3:00 ~ 3:30 Break
3:30 ~ 5:00 Parallel session A
Parallel session B
5:00 ~ 5:30 Break
Chair : Rosario Nunzio Mantegna
5:30 ~ 6:00 Misako Takayasu Estimation of flows in Businesses to Business transaction
network
6:00 ~ 6:30 Gabjin Oh Measuring systemic risk through contagion effect of
industry sector
6:30 ~ 7:00 Wei-Xing Zhou Liquidity, trade size and immediate price impact
7:00 ~ 9:00 Welcoming Reception
-
12
Day 1
Parallel session A
Time Speaker Title
Chair : Hawoong Jeong
3:30 ~ 3:45 Jongwook Kim Non-Gaussianity defined by Auto-Correlated Random
Walk
3:45 ~ 4:00 Wen-Jie Xie Extreme value statistics and recurrence intervals of NYMEX
energy futures volatility
4:00 ~ 4:15 Rudi Schaefer Nonstationary correlations: From market states to random
matrix averages
4:15 ~ 4:30 Alejandro Raul
Hernandez Montoya
An statistical analysis of short term price trends symmetry
in daily stock-market index data
4:30 ~ 4:45 Sebastian Poledna Leverage-induced systemic risk under Basle II and other
credit risk policies
4:45 ~ 5:00 Il Gu Yi Fat tailed return distribution and fractal structure in profit
landscapes
Parallel session B
Time Speaker Title
Chair : Misako Takayasu
3:30 ~ 3:45 Hideki Takayasu Dealer model simulation of the intervention event of
foreign exchange markets
3:45 ~ 4:00 aurelien sylvain
christophe cassagnes
Heterogeneous Computation of Rainbow Option Prices
Using Fourier Cosine Series Expansion Under A Mix
4:00 ~ 4:15 Kyubin Yim Bubbles and Crashes in Artificial Double Auction Market
4:15 ~ 4:30 Tat-Shing CHOI Competition as Particle Interactions: From Duopoly to
General Market Structures
4:30 ~ 4:45 Hyejin Youn The Hidden Structure in Urban Economic Diversity
4:45 ~ 5:00 Zhi-Qiang Jiang Trading networks, abnormal motifs and stock manipulation
-
13
Day 2
Invited speaker
Time Speaker Title
Chair : Tiziana Di Matteo
9:00 ~ 9:30 Thomas Lux Hubs and resilience: towards more realistic models of the
interbank markets
9:30 ~ 10:00 Aki-Hiro Sato Parameter estimation methods of a multiplicative
stochastic process for the analysis of financial time series:
An application to inference of tail-risks
10:00 ~ 10:30 Siew Ann Cheong Forecasting Crashes in Financial and Housing Markets
10:30 ~ 11:00 Break
11:00 ~ 12:30 Parallel session C
Parallel session D
12:30 ~ 1:30 Lunch
Chair : Hiroshi Iyetomi
1:30 ~ 2:00 Stefan Thurner DebtRank-transparency: Eliminating systemic risk in
financial networks
2:00 ~ 2:30 Jae Woo Lee Network Topologies of a Financial Market around the
2008 Global Financial Crisis
2:30 ~ 3:00 Sai-Ping Li Volatility Clustering and Stochasticity in Nonlinear Time
Series
3:00 ~ 3:30 Break
3:30 ~ 5:00 Parallel session E
Parallel session F
5:00 ~ 5:30 Break
Chair : Thomas Lux
5:30 ~ 6:00 Victor Yakovenko Statistical Mechanics of Money, Income, Debt, and Energy
Consumption
6:00 ~ 6:30 Okyu Kwon Big Data, Data Science and Econophysics
6:30 ~ 7:00 Seunghwan Kim Understanding complexity of the microstructure of
financial markets
7:00 ~ 9:00 Banquet
-
14
Day 2
Parallel session C
Time Speaker Title
Chair : Sai-Ping Li
11:00 ~ 11:15 Franck Raynaud Statistical properties, networks and information flows in
derivative markets
11:15 ~ 11:30 Leonidas Sandoval Causality relations in a network of financial institutions
11:30 ~ 11:45 Ashadun Nobi Nonlinear dynamic properties and network topology of
global financial indices
11:45 ~ 12:00 Tae Seok Jang Productivity shocks and monetary policy in a two-country
model
12:00 ~ 12:15 Hai-Chuan Xu Short-term Market Reaction after Trading Halts in Chinese
Stock Market
12:15 ~ 12:30
Parallel session D
Time Speaker Title
Chair : Jae Woo Lee
11:00 ~ 11:15 Hongwei Xu Analysis of overlapping community structure in a large-
scale social network
11:15 ~ 11:30 Hang-Hyun Jo Contextual analysis framework for bursty dynamics
11:30 ~ 11:45 Wanting Xiong The Emergence of Fair offers in Ultimatum Game on
Bipartite Networks
11:45 ~ 12:00 Sang Hoon Lee Overlapping Community Detection of Multilayer Networks
12:00 ~ 12:15 Jinzhong Guo Money Circulation and Credit Circulation
12:15 ~ 12:30
-
15
Day 2
Parallel session E
Time Speaker Title
Chair : Siew Ann Cheong
3:30 ~ 3:45 Tian Qiu Correlation dynamics between price return and trading volume
3:45 ~ 4:00 Ming-Xia Li Correlation analysis in Chinese stock trading network
4:00 ~ 4:15 Kihong Chung Generalized Epidemic Process on Modular Networks
4:15 ~ 4:30 Chih-Hao Lin Adaptive Trading for Anti-correlated Pairs of Stocks
4:30 ~ 4:45 Seok-won Ahn Portfolio selection using complex network
4:45 ~ 5:00
Parallel session F
Time Speaker Title
Chair : Wei-Xing Zhou
3:30 ~ 3:45 CHI WUN CHOI A Study on the Rock-Paper-Scissors game in Co-evolving
Networks
3:45 ~ 4:00 Min-Woo Ahn Network structure of national R&D activity in Korea
4:00 ~ 4:15 Eiichi Umehara Relationship between Stock BBS and Stock Market using stock
prices intra-day: As Case of SoftBank
4:15 ~ 4:30 Byounghwa Lee Prediction of Congestion Sites in Pohang
4:30 ~ 4:45 Jianzhong Zhang Scaling Structure in Game-Locked Aggregation
4:45 ~ 5:00 Inho Hong Intra-city Bus Network in Korean Mid-Size Cities
-
16
Day 3
Invited speaker
Time Speaker Title
Chair : Thomas Lux
9:00 ~ 9:30 Yong-Cheol Kim Banking Concentration and Moral Hazard
9:30 ~ 10:00 Taisei Kaizoji Bubbles and crashes: Modeling from the point of
view of statistical physics
10:00 ~ 10:30 Kyungsik Kim Analysis of future prices from the structure of
correlations
10:30 ~ 11:00 Break
Chair : Taisei Kaizoji
11:00 ~ 11:30 Beom Jun Kim Human dynamics of spending: Study of a coalition
loyalty program
11:30 ~ 12:00 Seung-Woo Son Thinking fast and slow in a poker game
12:00 ~ 12:30 Woo-Sung Jung Complex network analysis of social data in Korea
12:30 ~ 1:30 Lunch
1:30 ~ Excursion
-
17
Day 1
Invited Talk (at Auditorium)
11:00 – 11:30 Invited Talk 1 Rosario Nunzio Mantegna
11:30 – 12:00 Invited Talk 2 Tobias Preis
12:00 – 12:30 Invited Talk 3 Hawoong Jeong
1:30 – 2:00 Invited Talk 4 Tiziana Di Matteo
2:00 – 2:30 Invited Talk 5 Hiroshi Iyetomi
2:30 – 3:00 Invited Talk 6 Yougui Wang
Parallel Session A, B (Contributed Talk)
3:30 – 5:00 Parallel session A Auditorium
Parallel session B International Conference Room
Invited Talk (at Auditorium)
5:30 – 6:00 Invited Talk 7 Misako Takayasu
6:00 – 6:30 Invited Talk 8 Gabjin Oh
6:30 – 7:00 Invited Talk 9 Wei-Xing Zhou
-
18
-
19
Invited Talk
STATISTICALLY VALIDATED NETWORKS OF MARKET
MEMBERS TRADING AT THE LSE ELECTRONIC AND
DEALERS' MARKET
Rosario Nunzio Mantegna
University of Palermo, Argentina
We empirically detect and analyze trading networks, which are present among all market members of
the London Stock Exchange (LSE) trading shares of a specific stock in a selected period of time. We
analyze the anonymous electronic book and the networked dealers' market separately, and we
statistically validate a link between two market members if the number of transactions of a selected
stock that occur between the two market members is too large to be explained according to a null
hypothesis of random trading between them. The statistical validation is obtained by generalizing to
directed networks a procedure of statistical validation of networks recently introduced in [1].
Specifically, we separately analyze the trading networks of market members trading five highly liquid
stocks, in the two LSE venues, from daily to yearly time scale, during the calendar year 2005. For the
selected stocks, we find that statistically validated trading networks for the dealers' market are bigger
and more stable over time than those observed for the electronic market. Our results therefore confirm
that anonymity in the electronic order book minimizes the probability of preferential pair trading
interactions and implies that concerns about adverse selection in the dealers' market are somewhat
compensated by other positive aspects, which are specific to the dealers' market, such as the
possibility of exchanging large volumes in a single transaction or obtaining a transaction price within
the spread observed at the anonymous electronic book venue.
This work is done in collaboration with A. Carollo, F. Lillo, M. Tumminello, and G. Vaglica
References:
[1] Tumminello M, Miccichè S, Lillo F, Piilo J, Mantegna RN (2011) Statistically Validated
Networks in Bipartite Complex Systems. PLoS ONE 6(3): e17994. doi:10.1371/journal.pone.0017994
-
20
Invited Talk
QUANTIFYING ECONOMIC BEHAVIOR USING BIG DATA
Tobias Preis
University of Warwick, UK
In this talk, I will outline some recent highlights of our research, addressing two questions.
Firstly, can big data resources provide insights into crises in financial markets which affect
humans worldwide? By analyzing Google query volumes for search terms related to finance
[1,2] and views of Wikipedia articles [3], we find patterns which may be interpreted as early
warning signs of stock market moves. Secondly, can we provide insight into international
differences in economic wellbeing by comparing patterns of interaction with the Internet? To
answer this question, we introduce a future-orientation index to quantify the degree to which
Internet users seek more information about years in the future than years in the past. We
analyze Google logs and find a striking correlation between the country's GDP and the
predisposition of its inhabitants to look forward [4]. Our results illustrate the potential that
combining extensive behavioral data sets offers for a better understanding of large scale
human economic behavior.
-
21
Invited Talk
GOOGLE KNOWS (ALMOST) EVERYTHING! - BIG-DATA AND
NETWORK SCIENCE
Hawoong Jeong
KAIST, Korea
Network science is an interdisciplinary academic field which studies complex networks such
as engineered networks, information networks, biological networks, cognitive and semantic
networks and social networks. This field has received a major boost caused by the availability
of huge network data resources on the Internet. The field draws on theories and methods
including graph theory from mathematics, statistical mechanics from physics, data mining
and information visualization from computer science, inferential modeling from statistics,
and social structure from sociology to understand the complex systems, the problem to be
solved in 21st century. Another research field gaining huge attention nowadays is about big-
data. Big-data is defined as “high-volume, high-velocity, and/or high-variety information
assets that require new forms of processing to enable enhanced decision making, insight
discovery and process optimization.” by Gartner, Inc. This field of research has huge
potential for practical applications but it also promises new discovery in science. However,
these big-data should be combined and analyzed together to be useful, and in this respect,
network science will shed a light on analyzing these big-data in more combined way. In this
presentation, I will briefly review what we can do by combining big-data, especially using
Google and network science together to study various complex systems such as social
network between people, prediction of science and technology trends etc.
-
22
Invited Talk
SPREAD OF RISK ACROSS FINANCIAL MARKETS: BETTER
TO INVEST IN THE PERIPHERIES
Tiziana Di Matteo
King's College London, UK
In this talk I will introduce a methodology and a set of tools to filter complex dependency structures
in financial datasets by using networks [1-2]. The topology of these networks efficiently encodes the
complex dependency structure reducing data complexity while preserving the fundamental
characteristics of the dataset. This methodology has the added advantage of visualizing directly the
complex organization of the dependency structure over the graphic layout of the network.
I will discuss how this approach can be used to build a well-diversified portfolio that effectively
reduces investment risk. Specifically I will show that investments in stocks that occupy peripheral,
poorly connected regions in the financial filtered networks are most successful in diversifying
investments even for small baskets of stocks. On the contrary, investments in subsets of central,
highly connected stocks are characterized by greater risk and worse performance [3].
I will also introduce a general graph-theoretic approach that use these filtered networks to
simultaneously extract clusters and hierarchies in an unsupervised and deterministic manner, without
the use of any prior information and without need to specify any threshold [4-5]. I will show that
applications to financial data-sets can meaningfully identify industrial activities and structural market
changes [6].
References:
[1] T. Aste, T. Di Matteo, S. T. Hyde, Physica A 346 (2005) 20-26.
[2] M. Tumminello, T. Aste, T. Di Matteo, R. N. Mantegna, PNAS 102, n. 30 (2005) 10421.
[3] F. Pozzi, T. Di Matteo and T. Aste, Scientific Reports 3 (2013) 1665.
[4] Won-Min Song, T. Di Matteo, T. Aste, Discrete Applied Mathematics 159 (2011) 2135.
[5] Won-Min Song, T. Di Matteo, T. Aste, PLoS One 7(3) (2012) e31929.
[6] N. Musmeci, T. Di Matteo, T. Aste, in preparation (2013).
-
23
Invited Talk
FRUSTRATED CORRELATION STRUCTURES EMBEDDED IN
WELL-DEVELOPED STOCK MARKETS
Hiroshi Iyetomi
Department of Mathematics, Niigata University, Niigata 950-2181, Japan
We analyze daily stock prices data in S&P 500 and Tokyo Stock Exchange (TSE) to
elucidate correlations among stock price movements in the period of January 2001 through
December 2011. The correlation matrix is purified by random matrix theory and also the
market mode associated with the largest eigenvalue of the matrix is excluded from our study.
Here we take advantage of the concept of community in networks. That is, the purified
correlation matrix is regarded as the adjacency matrix for a stock correlation network. The
network thus constructed has links with weights of either sign depending on whether stocks
are correlated (positive) or anti-correlated (negative). Community is defined as a group of
comoving stocks, which are mutually related with positive correlation coefficients.
The community detection allows us to find that the stocks in S&P 500 are split up into four
communities conflicting to each other with negative correlation coefficients. In TSE, on the
other hand, there exists three communities of stocks forming a conflicting triangle. We also
report temporal change of the frustrated correlation structures embedded in both of the well-
developed markets.
-
24
Invited Talk
INCORPORATING DEBT INTO THE MODERN
MACROECONOMICS
Yougui Wang
School of Systems Science, Beijing Normal University, Beijing 100875, P.R. China
The current financial crisis has been recognized as a crisis of economics. In this talk, I will
argue that this financial turmoil is a crisis of macroeconomics in a strict sense and the culprit
is the bank run of expanded shadow banking system. The resulting attack on the mainstream
economics calls for reconstruction of modern macroeconomics. The deficiencies and faults in
macroeconomics are reviewed and the main challenge facing it is identified as how to
incorporate financial markets into existing canonical models. Some competing and/or
mutually complementary theoretical attempts to meet the challenge are set forth and the
stock-flow consistent model is deemed to be the effective solution. I highlight the role of debt
in the performance of complex and variable contemporary economies, which can retrospect
to even earlier than the birth of money. Once we put debt in a proper position in the new
macroeconomics building, we would better understand not only how the conventional
banking system works but also the way that the bank run of shadow banking jeopardizes the
whole financial system as well as the sustainable development.
-
25
Invited Talk
ESTIMATION OF FLOWS IN BUSINESES TO BUSINESS
TRANSACTION NETWORK
Misako Takayasu
Tokyo Institute of Technology, Japan
We analyze the data of business partnership in Japan and confirm that the network structure
is characterized by scale-free properties. The annual sales and transaction volumes between
pairs of firms are confirmed to be related with the underlying complex network structure. We
approximate the whole business-to-business money flow on the network by introducing a
non-linear interaction model which is formulated by the adjacency matrix of transaction
network.
Based on this model we estimate the flow of transaction for each directed link.
-
26
Invited Talk
MEASURING SYSTEMIC RISK THROUGH CONTAGION
EFFECT OF INDUSTRY SECTOR
Gabjin Oh
Division of Business Administration, Chosun University, Korea
Systemic risk is the risk that a negative feedback of one company is propagated to other
companies through their specific relation channel. To measure systemic risk that is
characterized by interconnected feature between economy units, we employ the generalized
variance decomposition method (GVDM) with the volatility data set of 354 companies listed
on KOSPI index. Based on the contagion behavior of industry sector or conglomerate module
in the financial market, we propose a novel approach to quantify a systemic risk and calculate
quantities of systemic risk for KOSPI market. We find that the systemic risks are closely
related to the financial crisis such as the Asian currency crisis and Subprime mortgage crisis.
In addition, we analyze whether the conglomerate is related to the systemic risk and find that
the conglomerate have influence on both the contagion effect of the real economy sectors
except construction and the systemic risk.
-
27
Invited Talk
LIQUIDITY, TRADE SIZE AND IMMEDIATE PRICE IMPACT
Wei-Xing Zhou
East China University of Science and Technology, China
The trade size has a direct impact on the price formation of the stock traded. Using order
book data from the Chinese market, we show that trades from filled and partially filled limit
orders have very different price impacts. The price impact of trades from partially filled
orders is constant when the volume is not too large, while that of filled orders shows power-
law behavior with an exponent 2/3. When returns and volumes are normalized by stock-
dependent averages, capitalization-independent scaling laws emerge for both types of trades.
However, no scaling relation in terms of stock capitalization can be constructed. We further
propose two regression models to investigate the influence of microscopic factors (trade size,
the bid-ask spread, the price gaps and the outstanding volumes at the bid and ask sides of the
limit order book) on the price impact of buyer-initiated partially filled trades, seller-initiated
partially filled trades, buyer-initiated filled trades and seller-initiated filled trades. We find
that they have quantitatively similar explanatory powers and these factors can account for up
to 44% of the price impacts. Large trade sizes, wide bid-ask spreads, high liquidity at the
same side and low liquidity at the opposite side will cause a large price impact. We also find
that the liquidity at the opposite side has a more influential impact than the liquidity at the
same side. Our results shed new light on the determinants of immediate price impacts.
References:
[1] W.-X. Zhou, Universal price impact functions of individual trades in an order-driven
market, Quantitative Finance 12 (8), 1253-1263 (2012).
[2] W.-X. Zhou, Determinants of immediate price impacts at the trade level in an emerging
order-driven market, New Journal of Physics 14 (2), 023055 (2012).
-
28
Parallel Session A
NON-GAUSSIANITY DEFINED BY AUTO-CORRELATED
RANDOM WALK
Jongwook Kim
APCTP, Korea
We relate the auto-correlation and the non-Gaussian diffusion in the newly proposed discrete
model, which is simply built by the single correlation parameter. The moment generating
function is exactly solved. The rightness of our proposal is tested using high frequency data
of Korea and U.S. stock market.
-
29
Parallel Session A
EXTREME VALUE STATISTICS AND RECURRENCE
INTERVALS OF NYMEX ENERGY FUTURES VOLATILITY
Wen-Jie Xiea), b), c)
, Zhi-Qiang Jianga), b)
, Wei-Xing Zhoua), b), c), d)
a) School of Business, East China University of Science and Technology, Shanghai 200237, China
b) Research Center for Econophysics, East China University of Science and Technology, Shanghai 200237, China
c) Department of Mathematics, East China University of Science and Technology, Shanghai 200237, China
d) Key Laboratory of Coal Gasification and Energy Chemical Engineering (MOE), East China University of Science and Technology, Shanghai 200237, China
Energy markets and the associated energy futures markets play a crucial role in global economies. It is of great
theoretical and practical significance to gain a deeper understanding of extreme value statistics of the volatility
of energy futures traded on the New York Mercantile Exchange (NYMEX). We investigate the statistical
properties of the recurrence intervals of daily volatility time series of four NYMEX energy futures, which are
defined as the waiting times between consecutive volatilities exceeding a given threshold . We find that the
recurrence intervals are distributed as a stretched exponential ( ) , where the exponent decreases
with increasing , and there is no scaling behavior in the distributions for different thresholds after the
recurrence intervals are scaled with the mean recurrence interval ̅. These findings are significant under the
Kolmogorov-Smirnov test and the Cramér-von Mises test. We show that empirical estimations are in nice
agreement with the numerical integration results for the occurrence probability ( ) of a next event above
the threshold within a (short) time interval after an elapsed time from the last event above . We also
investigate the memory effects of the recurrence intervals. It is found that the conditional distributions of large
and small recurrence intervals differ with each other and the conditional mean of the recurrence intervals scales
as a power law of the preceding interval ̅( ) ̅ ( ̅) , indicating that the recurrence intervals have short-
term correlations. Detrended fluctuation analysis and detrending moving average analysis further uncover that
the recurrence intervals possess long-term correlations. We confirm that the “clustering” of the volatility
recurrence intervals is caused by the long-term correlations well known to be present in the volatility. Our
findings shed new lights on the behavior of large volatility and have potential implications in risk management
of energy futures.
-
30
Parallel Session A
NONSTATIONARY CORRELATIONS: FROM MARKET STATES
TO RANDOM MATRIX AVERAGES
Rudi Schaefer
Faculty of Physics, University of Duisburg-Essen, Germany
We propose a definition of state for a financial market and use it to identify points of drastic
change in the correlation structure. These points are mapped to occurrences of financial crises.
In our observation time window we find a wide variety of characteristic correlation patterns,
which can be classified into several typical "market states" using a k-means clustering
analysis. Using this classification we recognize transitions between different market states
and an overall development towards new market states.
In statistical modeling, the nonstationarity of correlations can be taken into account by
averaging the multivariate normal distribution over an ensemble of random covariance
matrices. For this average we consider the Wishart ensemble and show analytical results as
well as a comparison to empirical data.
"
-
31
Parallel Session A
AN STATISTICAL ANALYSIS OF SHORT TERM PRICE TRENDS
SYMMETRY IN DAILY STOCK-MARKET INDEX DATA
Alejandro Raul Hernandez Montoya
University of Veracruz, Mexico
In financial time series there are periods in which the value increases or decreases
monotonically. We call those periods elemental trends and study the symmetry of their
probability distribution of their duration for the indices DJIA, NASDAQ and IPC. We try to
understand if uptrends and downtrends are governed by the same stochastic process.
-
32
Parallel Session A
LEVERAGE-INDUCED SYSTEMIC RISK UNDER BASLE II AND
OTHER CREDIT RISK POLICIES
Sebastian Poledna
Medical University of Vienna, Austria
We use a simple agent based model of value investors in financial markets to test three credit
regulation policies. The first is the unregulated case, which only imposes limits on maximum
leverage. The second is Basle II, which also imposes interest rate spreads on loans and
haircuts on collateral, and the third is a hypothetical alternative in which banks perfectly
hedge all of their leverage-induced risk with options that are paid for by the funds. When
compared to the unregulated case both Basle II and the perfect hedge policy reduce the risk of
default when leverage is low but increase it when leverage is high. This is because both
regulation policies increase the amount of synchronized buying and selling needed to achieve
deleveraging, which can destabilize the market. None of these policies are optimal for
everyone: Risk neutral investors prefer the unregulated case with a maximum leverage of
roughly four, banks prefer the perfect hedge policy, and fund managers prefer the unregulated
case with a high maximum leverage. No one prefers Basle II.
-
33
Parallel Session A
FAT TAILED RETURN DISTRIBUTION AND FRACTAL
STRUCTURE IN PROFIT LANDSCAPES
Il Gu Yi
Sungkyunkwan University, Korea
We study the origin of fractality of profit landscape in stock markets using simple trading
strategy for real financial data and artificial stock prices. The strategy we used is consists of
only two parameters p and q, and if the log return is larger (smaller) than p (-q) then we
decided to sell (buy) some stock. The parameter space lies on unit square (p, q) ∈ [0, 1] ×
[0, 1] and we discretize one into the N × N square grid and calculate the profit Π(p, q) at the
center of each grid. We find local maxima in profit landscapes are distributed in fractal-like
geometry and the number of local maxima M follows the power-law form M ∼ Na, a ≈ 1.6
for real financial data. We test the other artificial time series in order to find the origin of
fractality of profit landscape, and we find that the fat-tailed return distribution is closely
related to the exponent a ≈ 1.6 observed for real stock markets.
-
34
Parallel Session B
DEALER MODEL SIMULATION OF THE INTERVENTION
EVENT OF FOREIGN EXCHANGE MARKETS
Hideki Takayasu
Sony Computer Science Laboratories, and Meiji University, Japan
We pay attention to the historically largest central bank intervention event in the history of
foreign exchange market that occurred in October 2011. The bank of Japan bought USD by
JPY intensively for a few hours intermittently and the market exchage rate showed quite non-
random walk behavior. We apply a generalized dealer model to simulate this extraordinary
market state. We firstly prepare the basic dealer model which reproduces the ordinary market
fluctuations, and then we introduce an intervention dealer into the market. By tuning the
parameters of the model, we can reproduce the whole market bahavior.
-
35
Parallel Session B
HETEROGENEOUS COMPUTATION OF RAINBOW OPTION
PRICES USING FOURIER COSINE SERIES EXPANSION UNDER
A MIX
Aurelien Sylvain Christophe Cassagnes
The University of Tokyo, Japan
In this study we focused on comparing different heterogeneous computational designs for the
calculation of Rainbow options prices using the Fourier-cosine series expansion (FCSE)
method. We also propose a simple enough way to automatically decide ratio of load
balancing at runtime. A GPGPU implementation of the two-dimensional composite Simpson
rule free of conditional statements with some degree of loop unrolling is also introduced. We
will also show how to reduce the integration domain of coefficients appearing in the option
pricing and by doing so, achieve a substantial speed-up versus a straightforward
implementation. Major improvement in scalability when leveraging over all available
resources will serve as our empirical proof for the need of considering mixed computational
architectures.
-
36
Parallel Session B
BUBBLES AND CRASHES IN ARTIFICIAL DOUBLE AUCTION
MARKET
Kyubin Yim
POSTECH, Korea
In order to understand the intrinsic market microstructure properties and origin of market
bubbles and crashes, we introduce an Artificial Double Auction Market(ADAM). We make
ADAM using Agent Based Modelling(ABM) with heterogenous agents. Agents consist of
two types, such as fundamentalist and chartist. Fundamentalist makes strategy using
fundamental value which is independent on market. Chartist makes strategy using trend of
past price which is dependent on market. Specifically, chartist strategy has two options such
as optimistic and pessimistic. Optimisitic(Pessimistic) chartist forcasts that the future price
increases(decreases). And we make our trading system using double auction. Double auction
market is an order-driven market where traders set bids and asks and post market or limit
orders accordingly trader’s specific strategies. We simulate our model during long time
period and analyze the data model generated. As a result, stylized facts in real financial
market, such as market bubble and crash, fat-tails and long memory effect are observed in
market microstructure in ADAM. More specifically, chartist contributes that market goes to
bubble and crash. Whereas, when all agents in market are fundamentalists, there are no
bubble, crash, fat-tails and memory effects. Therefore,our model shows that increase of
chartist makes market extreme states such as bubble and crash.
-
37
Parallel Session B
COMPETITION AS PARTICLE INTERACTIONS: FROM
DUOPOLY TO GENERAL MARKET STRUCTURES
Tat-Shing CHOI
Department of Physics, The Hong Kong University of Science and Technology, Hong Kong
We study competition starting from a duopoly. There are only two players deciding their best
prices so as to optimize their profits in a small market. When they play the game recursively,
they adjust their own price according to information received from the market. By varying
the information parameters in this game, competitive states and cooperative states are
produced. We analyze this game by treating the agents as 1-D interacting particles. The
agents experience local fields due to their current prices. The market information is modeled
as an interaction arising from the price difference with their opponents. Agents move in the
price space following the equation of motion. We then generalize the model to some more
general market structures, and show how their behavior follows that in the duopoly model.
* Supported by the Research Grants Council of Hong Kong (grant numbers 605010 and
604512).
-
38
Parallel Session B
THE HIDDEN STRUCTURE IN URBAN ECONOMIC DIVERSITY
Hyejin Youn
Santa Fe Institute, USA
Cities are now home to a majority of the worlds population, and play a major role in many of
the most pressing challenges facing humanity such as climate change, sustainability, and
economic recovery [1]. Thus, understanding urban dynamics is crucial for effective economic,
social, and environmental policy-making. Recent availability of large, consistent datasets
enables us to study the structure and dynamics of cities in a more quantitative, mathematical
way [2,3]. Here we focus on two salient characteristics of urban areas associated with
innovation and wealth creation – population size and economic diversity [4]. By analysing
the distributions of establishments (work places), the fundamental units of economic analysis,
we show that the diversity of metropolitan areas in the United States manifests a remarkable
universal structure. This suggests that cities are indeed self-similar not only in terms of their
aggregated quantities (such as GDP, patents, crime) [3] but also in their internal abundance
structure, implying a generic mechanism for urban economic development at work. We
generalized Yule-Simon model the economic diversification diminishes its marginals once
the core economy establishes to explain the empirical distribution [5]. Multi-dimensional
allometric scaling reveals the hidden order in this developmental process. We believe our
integrated analysis sheds light on the general development process of urban economy in
association with the systematic economic merits of agglomeration and the division of labor.
[1] UN-HABITAT (United Nations Human Settlements Program), State of the worlds cities
2010/2011.
[2] M. Batty, Science 319, 769-771 (2008)
[3] L. M. A. Bettencourt, G. B. West, Nature 467, 912913 (2010).
[4] J. V. Henderson, Am Econ Rev 64, 640 (1974).
[5] H. A. Simon, Biometrika 42, 425 (1955).
-
39
Parallel Session B
TRADING NETWORKS, ABNORMAL MOTIFS AND STOCK
MANIPULATION
Zhi-Qiang Jiang
East China University of Science and Technology, China
We study trade-based manipulation of stock prices from the perspective of complex trading
networks constructed by using detailed information of trades. A stock trading network
consists of nodes and directed links, where every trader is a node and a link is formed from
one trader to the other if the former sells shares to the latter. Specifically, three abnormal
network motifs are investigated, which are found to be formed by a few traders, implying
potential intention of price manipulation. We further investigate the dynamics of volatility,
trading volume, average trade size and turnover around the transactions associated with the
abnormal motifs for large, medium and small trades. It is found that these variables peak at
the abnormal events and exhibit a power-law accumulation in the pre-event time period and a
power-law relaxation in the post-event period. We also find that the cumulative excess
returns are significantly positive after buyer-initiated suspicious trades and exhibit a mild
price reversal after seller-initiated suspicious trades. These findings can be better understood
in favor of price manipulation. Our work sheds new lights into the detection of price
manipulation resorting to the abnormal motifs of complex trading networks.
-
40
-
41
Day 2
Invited Talk (at Auditorium)
9:00 – 9:30 Invited Talk 10 Thomas Lux
9:30 – 10:00 Invited Talk 11 Aki-Hiro Sato
10:00 – 10:30 Invited Talk 12 Siew Ann Cheong
Parallel Session C, D (Contributed Talk)
11:00 – 12:30 Parallel session C Physics Building - Room 109
Parallel session D Physics Building - Room 111
Invited Talk (at Auditorium)
1:30 – 2:00 Invited Talk 13 Stefan Thurner
2:00 – 2:30 Invited Talk 14 Jae Woo Lee
2:30 – 3:00 Invited Talk 15 Sai-Ping Li
Parallel Session E, F (Contributed Talk)
3:30 – 5:00 Parallel session E Auditorium
Parallel session F International Conference Room
Invited Talk (at Auditorium)
5:30 – 6:00 Invited Talk 16 Victor Yakovenko
6:00 – 6:30 Invited Talk 17 Okyu Kwon
6:30 – 7:00 Invited Talk 18 Seunghwan Kim
-
42
-
43
Invited Talk
HUBS AND RESILIENCE: TOWARDS MORE REALISTIC
MODELS OF THE INTERBANK MARKETS
Thomas Lux
University of Kiel, Germany
This paper uses a toy financial system to study systemic risk in scale-free interbank networks.
Networks are produced according to a fitness algorithm, combined with a representation of
the balance sheets of the banks. Our generating processes for interbank networks are designed
in a way to reproduce the frequently documented features of disassortative behavior, power
laws in the degree distributions and power laws in the distribution of bank sizes. The results
show the presence of a particular shell structure affecting the spread of an endogenous shock.
-
44
Invited Talk
PARAMETER ESTIMATION METHODS OF A
MULTIPLICATIVE STOCHASTIC PROCESS FOR THE
ANALYSIS OF FINANCIAL TIME SERIES: AN APPLICATION
TO INFERENCE OF TAIL-RISKS
Aki-Hiro Sato
Kyoto University, Japan
This study considers Pearson type IV distribution and stochastic differential equations with
both mutually independent multiplicative and additive noises. A one-dimensional Pearson
type IV distribution can be theoretically derived as an equilibrium distribution of the
multiplicative stochastic differential equation. By using the stochastic differential equation,
we propose methods to estimate parameters from observations based on the maximum
likelihood procedure. We employ two kinds of log-likelihood functions: the one associated
with the stationary distribution and the other coming from the conditional distribution that is
calculated by solving the Fokker-Planck equation corresponding to the stochastic differential
equation. Finally we evaluate ruin probabilities under a given risk buffer and numerically
implement the proposed methods in order to assess tail risks for log-return time series of
foreign exchange rates.
-
45
Invited Talk
FORECASTING CRASHES IN FINANCIAL AND HOUSING
MARKETS
Siew Ann Cheong
Nanyang Technological University, Singapore
Drawing insights from statistical and nonlinear physics, we demonstrate that critical
transitions in strongly adaptive systems such as the financial and housing markets can also be
forecasted, after we understand the characters of universal precursors that must be present
prior to crashes. In this talk, we present two forecasting methods, the first based on the
statistical physics of a fusion-fission model of collective interactions [1], the second based on
the universal statistical and nonlinear physics of critical slowing down [2] preceding a crash.
We start by using the results of our time series clustering study of the Singapore Stock
Exchange (SGX) within the year 2008 to justify the use of a statistical model of fusions and
fissions. We then explain precursor signatures that accompany the growth of a giant cluster in
the fusion-fission model, before using these signatures to ‘predict’ the October 2008 crash in
the SGX. Thereafter, we turn our attention to the US housing market, where we use
signatures of critical slowing down and critical fluctuations to ‘predict’ the Asian Financial
Crisis-Technology Bubble Crisis-Subprime Loans Transition-Subprime Crisis sequence of
critical transitions.
References:
[1] Bohorquez, Juan Camilo, Sean Gourley, Alexander R. Dixon, Michael Spagat, and Neil F.
Johnson. "Common ecology quantifies human insurgency." Nature 462, no. 7275 (2009):
911-914.
[2] Scheffer, Marten, Jordi Bascompte, William A. Brock, Victor Brovkin, Stephen R.
Carpenter, Vasilis Dakos, Hermann Held, Egbert H. Van Nes, Max Rietkerk, and George
Sugihara. "Early-warning signals for critical transitions." Nature 461, no. 7260 (2009): 53-59.
-
46
Invited Talk
DEBTRANK-TRANSPARENCY: ELIMINATING SYSTEMIC
RISK IN FINANCIAL NETWORKS
Stefan Thurnera), b), c)
a) Section for Science of Complex Systems; Medical University of Vienna; Spitalgasse 23; A-1090; Austria,
b) Santa Fe Institute; 1399 Hyde Park Road; Santa Fe; NM 87501; USA, c) IIASA, Schlossplatz 1, A-2361 Laxenburg; Austria.
Nodes in a financial network, such as banks, cannot assess the true risks associated with
lending to other nodes in the network, unless they have full information on the riskiness of all
other nodes. These risks can be estimated by using network metrics (as DebtRank) of the
interbank liability network. With a simple agent based model we show that systemic risk in
financial networks can be drastically reduced by increasing transparency, i.e. making the
DebtRank of individual banks visible to others, and by imposing a rule, that reduces
interbank borrowing from systemically risky nodes. This scheme does not reduce the
efficiency of the financial network, but fosters a more homogeneous risk-distribution within
the system in a self-organized critical way. The reduction of systemic risk is due to a massive
reduction of cascading failures in the transparent system. A regulation-policy implementation
of the proposed scheme is discussed.
-
47
Invited Talk
NETWORK TOPOLOGIES OF A FINANCIAL MARKET
AROUND THE 2008 GLOBAL FINANCIAL CRISIS
Jae Woo Lee
Inha University, Korea
We consider effects of the global financial crisis in a local Korean financial market around
the 2008 global financial crisis. We analyze 185 individual stock prices belonging to the
KOSPI (Korea Composite Stock Price Index). We consider three time periods, before, during,
and after the crisis. The complex networks extract from the cross-correlation coefficients
among the stock price time series of the companies. We generate the threshold networks (TN),
the minimal spanning tees (MST), and the hierarchical network (HN) from the fully
connected cross-correlation networks. By assigning a threshold value of the cross-correlation
coefficient, we obtain the threshold networks. The power law of the degree distribution in the
threshold networks is observed in the limited range of the threshold. The threshold networks
during the crisis are fatter than other periods. The clustering coefficient of the threshold
networks follows the power law in the scaling range. We also generate the minimal spanning
trees from the fully connected correlation networks. The MST during the crisis period shrinks
in comparison to the periods before and after the crisis. The cophenetic correlation coefficient
increases during the crisis which indicates that the hierarchical structure increases in this
period. When the crisis hits the market, the companies’ behavior synchronously and their
correlations become stronger than the normal period.
-
48
Invited Talk
VOLATILITY CLUSTERING AND STOCHASTICITY IN
NONLINEAR TIME SERIES
Sai-Ping Li
Academina Sinica, Taiwan
Complex systems display common behavior such as volatility clustering and stochasticity in
their corresponding time series. In this talk, I'll give a brief review of these properties.
Applications to several complex systems, from financial markets to earthquakes and
cardiology will be discussed.
-
49
Invited Talk
STATISTICAL MECHANICS OF MONEY, INCOME, DEBT, AND
ENERGY CONSUMPTION
Victor Yakovenko
University of Maryland, USA
By analogy with the probability distribution of energy in physics, entropy maximization
results in the exponential Boltzmann-Gibbs probability distribution of money among the
agents in a closed economic system. Analysis of empirical data shows that income
distributions in USA, EU, and other countries has a well-defined two-class structure. The
majority of the population (about 97%) belongs to the lower class characterized by the
exponential ("thermal") distribution. The upper class (about 3% of the population) is
characterized by the Pareto power-law ("superthermal") distribution, and its share of the total
income expands and contracts dramatically during bubbles and busts in financial markets.
Globally, inequality in energy consumption per capita around the world has decreased in the
last 30 years and now approaches to the exponential probability distribution, in agreement
with the maximal entropy principle. All papers are available at
http://physics.umd.edu/~yakovenk/econophysics/
-
50
Invited Talk
BIG DATA, DATA SCIENCE AND ECONOPHYSICS
Okyu Kwon
National Institute for Mathematical Science, Korea
Recently, the big data becomes very big issue. We address what big data is and the relation
between data-driven science, which is new science paradigm as a big data era, and
econophysics. As data-driven research, we will introduce two data analysis studies. One is for
stock market data. We observed asymmetric information flow between the stock market
index and their component stocks using a transfer entropy measure. We found that the
amount of information flow from an index to a stock is larger than from a stock to an index.
This finding indicates that the market index is a major driving force in determining individual
stocks. Another one is for transportation data. We investigated the express bus flow in Korea
and its network topology. By using a gravity type model, we found that the bus flow between
cities depends on the square root of the product of the population size of city A and the
population size of city B. On the other hand, the total bus flow of a city depends on only its
population size. These different dependences on population originate from the network
property of the express bus network.
-
51
Invited Talk
UNDERSTANDING COMPLEXITY OF THE MICROSTRUCTURE
OF FINANCIAL MARKETS
Seunghwan Kim
POSTECH, Korea
Economic systems as extremely complex systems have recently become an interesting area of
research for physicists as well as economists. Numerous studies analyzing financial data have
been carried out to understand the nonlinearity and the complexity of economics systems
consisting of heterogeneous interacting agents, which reveals stylized facts different from
random-walk processes based on the efficiency market hypothesis.
The purpose of this talk is to understand the intrinsic characteristics of financial markets, for
example, long-term memory and clustering behavior of volatility data and interactions
between individual stocks using both the nonlinear time series analysis and the agent based
model. The statistical and nonlinear characteristics of financial systems have been analyzed,
which exhibits a strong long-term memory property in the volatility clustering. We also
propose an agent based model (ABM) to understand the intrinsic properties of abnormal
events such as the market crashes and bubbles with a focus on the artificial double auction
market (ADAM) as a trading system. We find that the chartist strategy is mainly responsible
for the fat tails in the return and the bid-ask spread time series, large fluctuations of which are
found to be closely related to the market crashes and bubbles.
-
52
Parallel Session C
STATISTICAL PROPERTIES, NETWORKS AND INFORMATION
FLOWS IN DERIVATIVE MARKETS
Franck Raynaud
Laboratory of Cell Biophysics, EPF Lausanne, France
We present an empirical study of future derivative markets for commodities and financial
assets. One of the main features of the future markets is the Samuelson effect which proposes
that the volatility of future prices increases when the contract reaches its expiration date. In
other words, the fluctuations of future prices decrease along the term structure.
In this study we explore different aspects of the Samuelson effect. First we examine this
pattern of volatility along the term structure as well as the distributions of prices returns. We
observe an ubiquitous behaviour for commodities as well as a segmentation of rare events
along the term structure.
Then we investigate correlation-based networks. The topology of the corresponding
minimum spanning trees reveals a chain-like organization reflecting the presence of a
Samuelson effect. Whereas introducing directionality of price movements between
commodities and financial assets will be a very important goal to reach, we present
preliminary results on shared and transferred information between maturities of the most
important commodity market, the American crude oil market.
-
53
Parallel Session C
CAUSALITY RELATIONS IN A NETWORK OF FINANCIAL
INSTITUTIONS
Leonidas Sandoval
Insper, Instituto de Ensino e Pesquisa, Brazil
After the Suprime Crisis of 2008 and the ongoing Credit Crisis that has been affecting
financial markets worldwide, much attention has been given to the interrelations between
banks and, in general, between financial institutions, with the aim to understand how
volatility may migrate between institutions. Most of the research uses networks based on the
lending and borrowing among banks in order to build models of propagation of crises. Such
networks, although asymmetric, do not display all the complex interrelations among financial
institutions.
The perception the market has of the stock price of a company is often a good measure of the
many complex factors that act on such firm, since it takes into account a myriad of factors
beyond the level of borrowing or lending of that company. So, we hypothesize that the value
of the stock of a firm is a good indicator of its current state and of the expectations the market
has of it. Networks based on the correlation between stocks may be built from the log-returns
of the time series of those stocks, but that is a measure that is both symmetric and not causal,
which may be used to study which stocks behave alike, but gives no results concerning
causality relations between them.
The present work uses the variations of the stocks of the 200 largest financial institutions of
the world in order to build a network based on Transfer Entropy, which is a concept
developed in Information Theory that has been used in a variety of applications, such as the
mapping the relations between regions in the human brain. Transfer Entropy is model-
independent, asymmetric and dynamic, in the sense that it can be used to establish causal
effects among nodes of a network. With some restrictions, it can be reduced to Granger
causality, although it does not have most of the limitations of the former.
-
54
Parallel Session C Such measure has been used by the author in the analysis of the international stock market
indices, using 92 benchmark indices of stock markets across the globe and their lagged
counterparts, with results that corroborate the idea that stock markets tend to influence one
another in a structure that depends much on the opening and closing hours of the stock
exchanges, from East to West and then back to the East, with Europe having the most
influence, contrary to the belief that the American stock market dominates the others.
In the present work, Transfer Entropy is used in order to establish a directed network of
stocks of financial institutions. Such network is then studied using the tools of network theory,
establishing which companies may be seen as most influential according to different
centrality measures. Then, the totality of stocks of financial institutions of some of the
countries that are seen today as representing the greatest risks to the international financial
market, such as Greece, Spain, Portugal, Cyprus and Italy, are added separately, and their
influences on the 200 original stocks is evaluated, pinpointing which institutions are most
affected by the financial institutions of those countries. By doing this, we try draw a map of
risk in financial institutions when affected by sudden changes in some key economies.
-
55
Parallel Session C
NONLINEAR DYNAMIC PROPERTIES AND NETWORK
TOPOLOGY OF GLOBAL FINANCIAL INDICES
Ashadun Nobi
Inha University, Korea
We investigated quantitative nonlinear dynamic and network topology of thirty one global
financial indices from 1998 to 2012. We calculated the Lyapunov exponents and Hurst
exponents by detrended fluctuation analysis. Network analysis has been done by minimum
spanning tree and threshold method. In addition, we constructed hierarchical network by the
average linkage hierarchical clustering algorithm and also calculated cophentic correlation
coefficient. We observed some abrupt changes of nonlinear dynamic exponents and also
network topology due to crisis and also globalization.
-
56
Parallel Session C
PRODUCTIVITY SHOCKS AND MONETARY POLICY IN A
TWO-COUNTRY MODEL
Tae-Seok Janga)
and Eiji Okanob)
a) Duksung Womans University, Dongyang Mirae University, Ewha Womans University, Korea
b) Chiba Keizai University, Japan
This paper examines the effects of foreign productivity shocks on a domestic monetary stance
in a new-Keynesian two-country model. The model shows that a positive productivity shock
derives up output from its natural level with a low natural interest rate. In the IS equation, a
relative rise of the foreign interest rate results in an increase of the output gap, provided that
agents take forward-looking rational expectations. This will provide downward pressures on
the foreign demand side, but put a positive impetus for the domestic output gap. This can
accelerate inflation dynamics in domestic price level. In the end, the monetary authority may
react to this situation by raising the key interest rate. By varying the degree of openness, we
show that impulse and response functions identify shock transmission mechanisms between
symmetric two economies.
-
57
Parallel Session C
SHORT-TERM MARKET REACTION AFTER TRADING HALTS
IN CHINESE STOCK MARKET
Hai-Chuan Xu
Tianjin University, China
In this paper, we study the dynamics of price changes, volume and ask-bid spread after the
trading halts using high-frequency data from Shenzhen Stock Exchange in China. The
method we used is similar with Mu., et al. (2010), Tóth., et al. (2009) and Zawadowski., et al.
(2006), while the definition of filter is not needed because trading halts are extreme events
intrinsically. We deal with all the trading halts of 120 stocks in Shenzhen Stock Exchange
from 2009 to 2010. Both the positive trading halts and the negative trading halts have a
reversal and decay as a power law, which forms a well-established peak. Then we divided the
trading halts into 3 groups according to their halt types, which kind of classification is also
consistent with their halt duration, we find the dynamics of abnormal fluctuation halts (last
for one hour or two hours) are very different from the stockholders meeting halts (last for one
day) and the significant matter halts(lasts for more than one day). These differences
demonstrate that the information asymmetry and traders' behavior play an influential role in
the efficiency of trading halts.
-
58
Parallel Session D
ANALYSIS OF OVERLAPPING COMMUNITY STRUCTURE IN A
LARGE-SCALE SOCIAL NETWORK
Hongwei Xu
The Chinese University of Hong Kong, China
Overlapping community structures are quite common in social networks. Finding out these
communities and the relationships between each other may provide a better understanding of
the structure of social network in a mesoscopic view. Here we propose an efficient method of
finding overlapping communities in huge social networks based on a statistical approach. We
first explore the local community structure around each vertex and find communities which
they belong to, then combine highly overlapped communities which are actually the same
community. We test the method on a friendship network of Sina Weibo with millions of
vertices and find tens of thousands of communities. We construct a web of communities by
exploring their relationships and show some statistical properties of the web.
-
59
Parallel Session D
RELEVANCE OF CONTEXT AND TIME-FRAME IN BURSTY
DYNAMICS
Hang-Hyun Jo
Aalto University, Finland
Inhomogeneous temporal processes in natural and social phenomena have been described by
bursts that are rapidly occurring events within short periods alternating with long periods of
low activity. Such a temporal process can be decomposed into sub-processes, according to
the contexts, i.e. circumstances in which the events occur. Then contextual bursts for each
sub-process are related to context-blind bursts for the original process. This requires to study
contextual bursts in real time-frame as well as in ordinal time-frame, where the real timings
of events are replaced by their orders in the event sequence. By analyzing a model of
uncorrelated inter-event times we find that contextual bursts in real time-frame can be
dominated by either context-blind bursts or contextual bursts in ordinal time-frame, or be
characterized by both. These results on the relevance of context and time-frame give insight
into the origin of bursts.
-
60
Parallel Session D
THE EMERGENCE OF FAIR OFFERS IN ULTIMATUM GAME
ON BIPARTITE NETWORKS
Wanting Xiong
Department of Systems Science, School of Management, Beijing Normal University, China
Simple and intriguing as it is, ultimatum game has been a widely applied analytical tool in
bargaining behaviors between two populations who periodically bargaining pairwise over
their shares of a common pie. This paper examines the dynamics of how fair offers come
about in ultimatum game within the framework of bipartite network where proposers and
responders are divided into two disjoint sets with links representing one game experience
between two agents. links only existing between different sets. Under the postulation that
both fairness motive and adaptive learning play a role in the fair behavior of human players,
we portray proposers as adaptive learners trying to maximize their payoffs and responders as
two types of agents with either pure money concern or fairness motivation. The notion of
“fairness leverage” is introduced as indicators of the influence of “tough” responders who
reject any unfair offers on the behavior group. Through experiments with different network
structures, we find that fair offers would be provided by the proposers as long as a small
proportion of the responders play “tough” against unfair offer.
-
61
Parallel Session D
OVERLAPPING COMMUNITY DETECTION OF MULTILAYER
NETWORKS
Sang Hoon Lee
University of Oxford, UK
We introduce a community detection method for multilayer networks generalized to consider
overlapping communities. A community detection framework incorporating multilayer
structures has been developed [1]. Both time-dependent and multiplex (different types of
interactions) networks can be represented using that framework, and we extend it by allowing
an important notion of node overlapping (nodes assigned to multiple community
memberships) [2,3,4]. The method is applied to various data sets such as temporally varying
weighted fungal networks of different species, networks of brain regions from functional
magnetic resonance imaging (fMRI) signals, and Congressperson networks from similarity in
senators’ roll-call voting patterns and cosponsoring bills.
References:
[1] P. J. Mucha, T. Richardson, K. Macon, M. A. Porter, and J.-P. Onnela, Science 328, 876
(2010).
[2] Y.-Y. Ahn, J. P. Bagrow, and S. Lehmann, Nature 466, 761 (2010).
[3] I. Psorakis, S. Roberts, M. Ebden, and B. Sheldon, Phys. Rev. E 83, 066114 (2011).
[4] J. Yang and J. Leskovec, e-print arXiv:1205.6228.
-
62
Parallel Session D
MONEY CIRCULATION AND CREDIT CIRCULATION
Jinzhong Guo
Department of Systems Science, School of Management, Beijing Normal University, China
In this paper, a multi-agent model has been constructed to illustrate the circulation process of
both money and credit. The system contains individual agents who play the roles of
households and business firms and a virtual bank which creates money as well as credit. As
proverbially elaborated in economics literature, money functions in the working of an
economy through its circulation. Here we argue that credit affects the economy in the same
way. As money is received and disbursed by households, it is circulating among them.
Likewise, with money is borrowed and repaid by firms, credit is also circulating among
debtors. Based on a money exchange model in which agents can be in debt, we portray how
credit circulates from one debtor to another as the bank makes advances and takes them back.
When agents expend money based on both their savings and loans borrowed from the bank,
both the money circulation and credit circulation can be tracked by following each unit of
them and characterized by the holding time of them. The simulation results show that holding
times of money and credit obey the exponential distribution. Thus the velocities of money
and credit can be expressed in terms of corresponding expected value of holding time. The
circulation velocities of money and credit obtained from simulation results are fitted very
well with those theoretical predictions. This agent-based model opens a venue to studying
how individuals’ behavior and decision influence the macroeconomic variables such as the
velocity of money and that of credit.
-
63
Parallel Session E
CORRELATION DYNAMICS BETWEEN PRICE RETURN AND
TRADING VOLUME
Tian Qiu
School of Information Engineering, Nanchang Hangkong University, China
How trading volume dynamically correlates with price return is investigated for the Chinese
and the United States stock markets. A positive correlation is observed for the Chinese stock
markets, while a transition from the positive correlation to the negative correlation is found
for the United States stock markets. For the United States stock markets, negative return-
volume correlations are found for several big financial crashes. Nonlocal dynamics of the
return-volume correlation shows an opposite correlation between the United States and
Chinese stock markets.
-
64
Parallel Session E
CORRELATION ANALYSIS IN CHINESE STOCK TRADING
NETWORK
Ming-Xia Li
East China University of Science and Technology, China
The transactions of a stock over a given time period can be presented as a stock trading
network, where each trader is a node and two nodes are linked if one node sells stock shares
to the other. We construct 11472 networks over time period of five minutes using order flow
data. Strong correlations between network metrics and financial variables are observed in our
results. Furthermore, most network metrics Ganger cause financial variables.
-
65
Parallel Session E
GENERALIZED EPIDEMIC PROCESS ON MODULAR
NETWORKS
Kihong Chung
Department of Physics, KAIST, Daejeon 305-701, Korea
Social reinforcement and existence of communities are two salient features observed in the
emergence of collective behavior through social contacts. To investigate the combined effects
of those two features, we numerically study generalized epidemic process [1] on modular
networks with equal-sized finite communities and adjustable modularity. Our results [2] show
that the system has a continuous phase transition of the bond percolation universality class for
weak social reinforcement, whereas a discontinuous phase transition is observed for
sufficiently strong social reinforcement. We use bimodality coefficient to indicate the
boundary between different types of transition, which is shown to be dependent on the
modularity.
[1] G. Bizhani, M. Paczuski, and P. Grassberger, Phys. Rev. E 86, 011128 (2012).
[2] Kihong Chung, Yongjoo Baek, Daniel Kim, Meesoon Ha, and Hawoong Jeong (in
preparation). This work was supported by the NRF grant (No. 2011-0011550).
-
66
Parallel Session E
ADAPTIVE TRADING FOR ANTI-CORRELATED PAIRS OF
STOCKS
Chih-Hao Lin
Institute of Physics, Academia Sinica, Nankang, Taipei 115, Taiwan
The effect of anti-correlation between stocks in real stock market can be exploited for profit
if one can also properly set the criterion for trading that takes into account the volatility of the
stock pair. This complex problem of resource allocation for portfolio management of stocks
is here simplified to a problem of adaptive trading with an investment criterion that evolves
along with the time series of the stock data. The trend of the stock is modeled with the
standard stochastic dynamics, from which the volatility of the stock provides a criterion for
investment on a two stock portfolio that consists of the anti-correlated pair using mean
variance analysis that optimizes the return. The action of buy and sell of the two-stock
portfolio will be based on the fractional return of the pair : when the fractional return of the
pair is greater than an upper threshold of 1.01, the action “buy” is taken; and when this
fractional return is less than a lower threshold of 0.99, the action “sell” is taken. Since both
the volatility criterion for investment and the fractional return of the two-stock portfolio are
time dependent, the entire trading scheme is adaptive. Comparison of this evolving strategy
of investment with time-average performance of the respective stocks indicates a consistent
superiority.
-
67
Parallel Session E
PORTFOLIO SELECTION USING COMPLEX NETWORK
Seok-won Ahna)
, Gabjin Ohb)
a) Department of English Education, Chosun University, Korea b) Division of Business Administration , Chosun University, Korea
Portfolio management is an essential problem of financial investment literature. Since
Markowitz's portfolio theory introduced, the numerous methods for constructing portfolio set
have been proposed in the traditional technology such as the several clustering algorithm and
the random matrix theory, while there has been relatively little study of network approach.
We used an individual stocks listed on the KOSPI index from 01.03 2000 to 12. 31. 2012. To
make a diverse portfolio sets, we constructed the stock network with the links above given
threshold value. Because the intrinsic network properties such as degree distribution,
clustering coefficient, and etc. will change according to threshold values, we analyze a
network property through the growing pattern of the largest size module created by threshold
value. We find that the increase of the largest size module for stock market which has the
information on interactions between individual firms is much slower than that of random
network. We consider Pearson correlation in order to check the performance of proposed
method and calculated the correlation between the KOSPI and the artificial index. We find
that the correlation value was high enough in overall threshold value to regard it as a
secondary stock index. Besides, the performance of portfolio sets created through the network
approach against the investment rate of return of KOSPI index was much better within certain
parameter regions.
-
68
Parallel Session F
A STUDY ON THE ROCK-PAPER-SCISSORS GAME IN CO-
EVOLVING NETWORKS
CHI WUN CHOI
The Chinese University of Hong Kong, Hong Kong
In complex network, the Rock-Paper-Scissors (RPS) game is one of the interesting research
topics. In the RPS game, three strategies cyclically dominate each other. We propose and
study an Adaptive Rock-Paper-Scissor (ARPS) model, in which each agent adapts by either
rewiring an unfavourable link or switching his strategy in a co-evolving network. We study
the model by computational simulations and establishing a theory. The analytic results
capture the main features in the simulation results, including the emergence of different
phases. We proposed and studied two different rewiring schemes for selecting the rewiring
target randomly and intentionally. Results related to the probability of winning, losing and
drawing in the steady state are also studied.
-
69
Parallel Session F
NETWORK STRUCTURE OF NATIONAL R&D ACTIVITY IN
KOREA
Min-Woo Ahn
POSTECH, Korea
Technology is essential for our life, so R&D activity is crucial for the improvement of our
quality of life. Therefore, many agency, especially government, support R&D activity. We
confirm the structure of the R&D activity supported by government by constructing network
and observing the network structure. We used the data from the NTIS (National Technology
& science Information Service), which includes the information about the research projects
such as the title, category, and keywords, which are supported by government. First, we
construct research project network. Each node is research project, and two nodes are
connected if two projects have common keywords. From this network, we construct 6T
network. Nodes are constructed by combining research projects into one node those are
included in the same category, and two node are connected if there are connection between
two category. We will describe the property of this network and will discuss about the
meaning of the results.
-
70
Parallel Session F
RELATIONSHIP BETWEEN STOCK BBS AND STOCK MARKET
USING STOCK PRICES INTRA-DAY: AS CASE OF SOFTBANK
Eiichi Umehara
Tokyo City University, Japan
Internet stock BBS is a tool that can directly know the voices of other investors. According to
a previous study using daily prices in Japan, the number and the contents of messages posted
on BBS can explain volumes and volatilities of stocks, and also can explain overnight returns
though it is difficult to gain economical profit when considering a commission fee. In this
study, we analyze the relationship between the Internet stock BBS and intra-day stock prices.
Focusing on the intraday stock prices of Soft Bank, we analyze the relationship, using natural
language processing. As a result, we find that the number of messages, the number of bullish
messages, and bullishness is the coincident and lagging indicator of return, the number of
messages is the coincident and lagging indicator of the volatility, and the number of messages,
the number of bull messages, and bullishness are the leading, coincident and lagging
indicator of the volume.
-
71
Parallel Session F
PREDICTION OF CONGESTION SITES IN POHANG
Byounghwa Lee
POSTECH, Korea
I investigated the roads network of the city of Pohang. By using several topological methods,
I obtained the degree distribution of the network which, similarly to that of other cities,
displayed a peak around its average value. Also, basic network measures such as the average
shortest path length and clustering coefficient reflect the basic property of planar networks,
i.e. planarity, which means that the links do not cross each other. I compared roads network
in Pohang with that of other cities, by lo