1: % ****** Start of file apssamp.tex ******

2: %

3: %   This file is part of the APS files in the REVTeX 4 distribution.

4: %   Version 4.0 of REVTeX, August 2001

5: %

6: %   Copyright (c) 2001 The American Physical Society.

7: %

8: %   See the REVTeX 4 README file for restrictions and more information.

9: %

10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed

11: % as well as the rest of the prerequisites for REVTeX 4.0

12: %

13: % See the REVTeX 4 README file

14: % It also requires running BibTeX. The commands are as follows:

15: %

16: %  1)  latex apssamp.tex

17: %  2)  bibtex apssamp

18: %  3)  latex apssamp.tex

19: %  4)  latex apssamp.tex

20: %

21: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

22: \documentclass[preprint,showpacs,showkeys,preprintnumbers,amsmath,amssymb]{revtex4}

23: 

24: % Some other (several out of many) possibilities

25: %\documentclass[preprint,aps]{revtex4}

26: %\documentclass[preprint,aps,draft]{revtex4}

27: %\documentclass[prb]{revtex4}% Physical Review B

28: 

29: \usepackage{graphicx}% Include figure files

30: \usepackage{dcolumn}% Align table columns on decimal point

31: \usepackage{bm}% bold math

32: 

33: %\nofiles

34: 

35: \begin{document}

36: 

37: %\preprint{APS/123-QED}

38: 

39: \title{Critical dynamics and global persistence exponent on Taiwan financial market}% Force line breaks with \\

40: 

41: \author{I-Chun Chen}

42: \email{ichun.nancy@msa.hinet.net}

43: \affiliation{Physics Department, National Chung Hsing University, Taichung, Taiwan R.O.C.}

44: \affiliation{Nan Kai Institute of Technology, Nantou, Taiwan R.O.C.}

45: 

46: 

47: \author{Hsen-Che Tseng}

48: \author{Ping-Cheng Li}

49: \author{Hung-Jung Chen}

50: \affiliation{Physics Department, National Chung Hsing University, Taichung, Taiwan R.O.C.}

51: 

52: %\date{\today}

53: 

54: \begin{abstract}

55: We investigated the critical dynamics on the daily Taiwan stock

56: exchange index (TSE) from 1971 to 2005, and the 5-min intraday

57: data from 1996 to 2005. A global persistence exponent $\theta_{p}$ was

58: defined for non-equilibrium critical phenomena \cite{Janssen,Majumdar}, and

59: describing dynamic behavior in an economic index \cite{Zheng}.

60: 

61: In recent numerical analysis studies of literatures,

62: it is illustrated that the persistence probability

63: has a universal scaling form $P(t) \sim t^{-\theta_{p}}$ \cite{Zheng1}. In this work,

64: we analyzed persistence properties of universal scaling

65: behavior on Taiwan financial market, and also calculated

66: the global persistence exponent $\theta_{p}$. We found our analytical

67: results in good agreement with the same universality.

68: 

69: %keywords: Persistence probability; Hurst exponent; Taiwan stock exchange index.

70: \end{abstract}

71: 

72: \pacs{47.27.eb}% PACS, the Physics and Astronomy

73:                              % Classification Scheme.

74: \keywords{Persistence probability; Hurst exponent; Taiwan stock exchange index.}%Use showkeys class option if keyword

75:                               %display desired

76: \maketitle

77: 

78: \section{\label{sec:level1}Introduction}

79: 

80: Problems in economy and finance have attracted the interest

81: of statistical physicists all over the world. Using the

82: tools developed for statistical physics, like phase

83: transitions, critical exponents, mean field approximations,

84: renormalization group \cite{Didier}, persistence

85: probability \cite{Majumdar1,Li,Constantin}.

86: 

87: In recent years the detrended fluctuation analysis (DFA)

88: method \cite{Carbonea,Ausloos,Vandewalle,Matia,Oswiecimka,Schmitt}

89: has become a widely used technique for the

90: determination of (mono-) fractal scaling properties and

91: the detection of long-range correlations in noisy,

92: nonstationary time series \cite{Vandewalle}. In many of non-equilibrium

93: systems, the persistence has been found to decay as a

94: power-law at time series, $P(t) \sim t^{-\theta_{p}}$. Hurst exponent and

95: persistence exponent in these financial time series are

96: investigated in numerical and analytical \cite{Constantin}.

97: 

98: We calculated the experimental data with the daily Taiwan stock exchange

99: index (TSE) from 1971 to 2005, and the 5-min intraday data

100: from 1996 to 2005. In this work, we analyzed persistence

101: properties of universal scaling behavior on Taiwan financial market.

102: 

103: \section{\label{sec:level2}Method}

104: 

105: We consider a set of data recorded the daily Taiwan

106: stock exchange index (TSE) from 1971 to 2005.

107: Let $Y_{i}(t)$ be the stock index at discrete times $i$,

108: $i = 1, 2,...,t_{n}-1$. The final transaction

109: time is denoted by $t_{n}$. Then, the log-return price

110: is defined as

111: \begin{eqnarray}

112: r_{i}(t)=\ln{Y_{i}(t+\Delta t)-\ln{Y_{i}(t)}},

113: \label{1}

114: \end{eqnarray}

115: where $\Delta t$ is time interval. In this paper,

116: we analyzed daily data; $\Delta t=1$ day.

117: 

118: \begin{figure}

119: \includegraphics{ChenFig1}% Here is how to import EPS art

120: \caption{\label{fig:1} Plot of the log-return price

121: $r_i(t)$ vs. $t$ on the daily Taiwan stock exchange

122: index (TSE) from 1971 to 2005.}

123: \end{figure}

124: 

125: Let us denote the value of the TSE at a certain time $t'$ as $y(t')$.

126: $P_{+}(t)$ is the probability that the value of the TSE has never gone down to the

127: value $y(t')$ in time $t$, $y(t'+N\Delta t) > y(t')$ for $N=1,2,...,t$. i.e.,

128: $P_{-}(t)$ is the probability that the value of the TSE has never

129: gone up to the value $y(t')$ in time $t$, $y(t'+N\Delta t) < y(t')$ for $N=1,2,...,t$.

130: The persistence probability $P(t)$ is $[P_{+}(t)+ P_{-}(t)]/2$.

131: 

132: The persistence probability has a power-law behavior

133: \begin{eqnarray}

134: P(t) \sim t^{-\theta_{p}}.

135: \label{2}

136: \end{eqnarray}

137: 

138: \begin{figure}

139: \includegraphics{ChenFig2}% Here is how to import EPS art

140: \caption{\label{fig:2a} Persistence probability with 5 minutes

141: data of the daily Taiwan stock price index (TSE) from 1996 to 2005.}

142: \end{figure}

143: 

144: \begin{figure}

145: \includegraphics{ChenFig3}% Here is how to import EPS art

146: \caption{\label{fig:2b} Persistence probability with 5 minutes

147: data of the daily Taiwan stock price index (TSE) from 1996 to 2005.}

148: \end{figure}

149: 

150: For the cumulative time series of the log-return price variables is defined as

151: \begin{eqnarray}

152: X(i)=\sum_{k=1}^{i} (r_{k}-{\bar r} ),

153: \label{3}

154: \end{eqnarray}

155: where ${\bar r}$ is the average value of log-return price.

156: $X(i)$ is divided into $Ns$ disjoint segments of length $s$.

157: $p_{\nu}(i)=a_i+b_it$ , $a_i$,  $b_i$ is constant. Since the

158: length $N$ of the series is often not a multiple of the considered

159: \begin{eqnarray}

160: X_{s}(i)=X(i)-p_{\nu}(i).

161: \label{4}

162: \end{eqnarray}

163: The generalized $q$~th-order price-price correlation function is defined as

164: \begin{eqnarray}

165: G_q(t)={\langle |Y(t_0+t)-Y(t_0)|^q \rangle}^{1/q},

166: \label{5}

167: \end{eqnarray}

168: where $Y(t)$ is the stock price and the average is over all the initial times $t_0$.

169: $G_q(t)$ has a power-law behavior

170: \begin{eqnarray}

171: G_q(t) \sim t^{H_{q}}.

172: \label{6}

173: \end{eqnarray}

174: where $H_q$ is called the generalized Hurst exponent.

175: 

176: \begin{figure}

177: \includegraphics{ChenFig4}% Here is how to import EPS art

178: \caption{\label{fig:3a} Log-log plot of the generalized

179: price-price correlation function $G_q(t)$ vs. $t$ corresponding

180: to the daily Taiwan stock exchange index (TSE) from

181: 1971 to 2005. }

182: \end{figure}

183: 

184: \begin{figure}

185: \includegraphics{ChenFig5}% Here is how to import EPS art

186: \caption{\label{fig:3b} Log-log plot of the generalized

187: price-price correlation function $G_q(t)$ vs. $t$

188: corresponding to 5 minutes data of the daily Taiwan

189: stock price index (TSE) from 1996 to 2005. }

190: \end{figure}

191: 

192: \begin{figure}

193: \includegraphics{ChenFig6}% Here is how to import EPS art

194: \caption{\label{fig:4a} Plot of the generalized Hurst

195: exponent $H_q$ vs. $1/q$ on differ daily Taiwan stock exchange indices.}

196: \end{figure}

197: 

198: The generalized fluctuation function is defined as

199: \begin{eqnarray}

200: F_s^2(\nu)=[{1\over s}\sum_{j=1}^s X_{s}]^2 .

201: \label{7}

202: \end{eqnarray}

203: The generalized $q$th-order fluctuation function is defined as

204: \begin{eqnarray}

205: F_q(s)=[{1\over 2N}\sum_{\nu=1}^{2N_s} F_s^2(\nu)^{q/2}]^{1/q}.

206: \label{8}

207: \end{eqnarray}

208: 

209: \begin{figure}

210: \includegraphics{ChenFig7}% Here is how to import EPS art

211: \caption{\label{fig:5a} Log-log plot of the normalized

212: fluctuation function $f_q(s)$ vs. $s$ corresponding to the

213: daily Taiwan stock exchange index (TSE) from 1971 to 2005. }

214: \end{figure}

215: 

216: \begin{figure}

217: \includegraphics{ChenFig8}% Here is how to import EPS art

218: \caption{\label{fig:5b} Log-log plot of the normalized

219: fluctuation function $f_q(s)$ vs. $s$ corresponding to 5

220: minutes data of the daily Taiwan stock price index (TSE)

221: from 1996 to 2005.}

222: \end{figure}

223: 

224: \begin{figure}

225: \includegraphics{ChenFig9}% Here is how to import EPS art

226: \caption{\label{fig:6b} Plot of the $P(Z_1)$ vs. $Z_1(t')$ differ

227: time data on the daily Taiwan stock exchange index (TSE).}

228: \end{figure}

229: 

230: 

231: By construction, since we use a linear fit for simplicity,

232: $F_q(s)$ is defined for $s\geq 3$. The scaling form of the correlation

233: function $F_q(s) \sim H_q$ provides the family of generalized

234: Hurst exponents $H_q$. For reasons that will become clearer

235: very shortly we also introduce the dimensionless fluctuation

236: function is defined as

237: \begin{eqnarray}

238: f_q(s)={[F_s^2(\nu)^{q/2}]^{1/2}\over [{1\over N}\sum_k^i (r_{k}-{\bar r})^2]^{1/2}}.

239: \label{9}

240: \end{eqnarray}

241: Let us denote the value of the TSE at a certain

242: time $t'$ as $Y(t')$. We calculate the pdf $P(Z_1)$ of the index changes

243: \begin{eqnarray}

244: Z_1(t')=Y(t'+\Delta t)-Y(t').

245: \label{10}

246: \end{eqnarray}

247: 

248: \section{\label{sec:level3}Discuss and Results}

249: 

250: A "persistence exponent" $\theta_p$ is defined for

251: non-equilibrium critical phenomena. Based on large

252: amounts of data compiled in past years, especially

253: those records in minutes or seconds, it becomes

254: possible to perform relatively accurate analysis

255: and to study the fine structure of the dynamics. In

256: Fig.~\ref{fig:2a} and \ref{fig:2b},

257: we observe the persistence probability that at least up to

258: least up to 250. The slope of persistence probability $P(t)$

259: estimation from the initial time to 200.

260: 

261: The price evolution is multifractal if the exponent

262: hierarchy $H_q$ varies with $q$, otherwise is fractal in

263: the theory of surface dynamical scaling referred to

264: as multiaffine and self-affine, respectively. In particular,

265: for $q = 2$, we recover the fractional Brownian motion

266: case described by the well-known Hurst exponent, $0 < H_2 < 1$.

267: The bridge between these two analyses is provided by

268: the second-order Hurst exponent $H_2$ associated with the

269: correlation function of the stock price, which has been

270: shown to be simply related to the persistence exponent

271: through $H_2 = 1-\theta_p$.

272: 

273: We note that this relation holds for any zero-mean

274: process (not necessarily Gaussian \cite{Hansen,Maslov})that satisfies

275: requirements above.

276: 

277: \begin{table}

278: \caption{\label{tab:table1}Compare $H_2$ , $\theta_p$ value

279: on the daily Taiwan stock exchange index (TSE).  }

280: \begin{ruledtabular}

281: \begin{tabular}{lcr}

282: Time & $H_2$ & $\theta_p$\\

283: \hline

284: 1971 - 1982 (daily) & 0.54 & 0.41\\

285: 1983 - 1990 (daily) & 0.58 & 0.34\\

286: 1991 - 2005 (daily) & 0.50 & 0.42\\

287: 1971 - 2005 (daily) & 0.54 & 0.46\\

288: 1996 - 2005 (5 minutes) & 0.52 & 0.46\\

289: \end{tabular}

290: \end{ruledtabular}

291: \end{table}

292: 

293: \section{\label{sec:level4}Conclusion}

294: 

295: We analyze the daily Taiwan stock exchange index

296: (TSE) from 1971 to 2005 and the 5-min intraday data

297: from 1996 to 2005. The persistence exponent $\theta_p$

298: associated with the power-law decay of the average

299: probability.

300: 

301: Our studies base on the persistence probability

302: analysis of the critical behavior in an economic

303: index, and the numerical estimation of the persistence

304: exponent $\theta_p$  with Hurst exponent $H_2$.

305: 

306: %\bibliography{chen.bib}% Produces the bibliography via BibTeX.

307: 

308: \section{References}

309: \begin{thebibliography}{99}

310: 

311: \bibitem{Janssen}

312: H. K. Janssen, B. Schaub, and B. Schmittmann, Z.~Phys. B 73 (1989), 539.

313: 

314: \bibitem{Majumdar}

315: S. N. Majumdar, Dibyendu Das, Phys.~Rev.~E 71 (2005), 036129.

316: 

317: \bibitem{Zheng}

318: B. Zheng, Mod.~Phys.~Lett.~B 16 (2002), 775.

319: 

320: \bibitem{Zheng1}

321: B. Zheng, Int.~J.~Mod.~Phys.~B 12 (1998), 1419.

322: 

323: \bibitem{Didier}

324: D. Sornette, A. Johansen, Phys.~A 245 (1997), 411.

325: 

326: \bibitem{Majumdar1}

327: S. N. Majumdar, A. J. Bray, S. J. Cornell, and C. Sire, Phys. Rev. Lett. 77 (1996), 3704.

328: 

329: \bibitem{Li}

330: Z. B. Li, L. Schulke and B. Zheng,Phys. Rev. Lett. 74 (1995), 3396.

331: 

332: \bibitem{Constantin}

333: M. Constantin and S. Das Sarma, Phys. Rev. E 72 (2005), 051106.

334: 

335: \bibitem{Carbonea}

336: A. Carbonea, G. Castellia, H.E. Stanley, Phys. A 344 (2004), 267.

337: 

338: \bibitem{Ausloos}

339: M. Ausloos and K. Ivanova, Comput. Phys. Commun. 147 (2002), 582.

340: 

341: \bibitem{Vandewalle}

342: N. Vandewalle and M. Ausloos, Eur. Phys. J. B4 (1998), 257.

343: 

344: \bibitem{Matia}

345: K. Matia, Y. Ashkenazy, and H. E. Stanley, Europhys. Lett. 61 (2003), 422.

346: 

347: \bibitem{Oswiecimka}

348: P. Oswiecimka, J. Kwapien, and S. Drozdz, Physica A 347 (2005), 626.

349: 

350: \bibitem{Schmitt}

351: F. Schmitt, D. Schertzer, and S. Lovejoy, Int. J. Theor. (5)Appl. Finance 3 (2000), 361.

352: 

353: \bibitem{Hansen}

354: A. Hansen, T. Engoy, and K. J. Maloy, Fractals 2 (1994), 527.

355: 

356: \bibitem{Maslov}

357: S. Maslov, M. Paczuski, and P. Bak, Phys. Rev. Lett. 73 (1994), 2162.

358: 

359: \end{thebibliography}

360: 

361: \end{document}

362: