Searcharxiv⌕ Search

arXiv subjects

Vasiliki Plerou

Publications and source records attributed to Vasiliki Plerou.

11 recordsLinked to original sources
arXiv2004

On the Origin of Power-Law Fluctuations in Stock Prices

We respond to the issues discussed by Farmer and Lillo (FL) related to our proposed approach to understanding the origin of power-law distributions in stock price fluctuations. First, we extend our previous analysis to 1000 US stocks and perform a new estimation of market impact that accounts for splitting of large orders and potential autocorrelations in the trade flow. Our new analysis shows clearly that price impact and volume are related by a square-root functional form of market impact for large volumes, in contrast to the claim of FL that this relationship increases as a power law with a smaller exponent. Since large orders are usually executed by splitting into smaller size trades, procedures used by FL give a downward bias for this power law exponent. Second, FL analyze 3 stocks traded on the London Stock Exchange, and solely on this basis they claim that the distribution of transaction volumes do not have a power-law tail for the London Stock Exchange. We perform new empirical analysis on transaction data for the 262 largest stocks listed in the London Stock Exchange, and find that the distribution of volume decays as a power-law with an exponent $\approx 3/2$ -- in sharp contrast to FL's claim that the distribution of transaction volume does not have a power-law tail. Our exponent estimate of $\approx 3/2$ is consistent with our previous results from the New York and Paris Stock Exchanges. We conclude that the available empirical evidence is consistent with our hypothesis on the origin of power-law fluctuations in stock prices.

cond-mat.dis-nn↗
arXiv2001

Symmetry Breaking in Stock Demand

Scale-free distributions and correlation functions found in financial data are reminiscent of the scale invariance of physical observables in the vicinity of a critical point. Here, we present empirical evidence for a transition phenomenon, accompanied by a symmetry breaking, in the investors' demand for stocks. We study the volume imbalance $Ω$ -- difference between the number of shares traded in buyer-initiated and seller-initiated trades in a time interval $Δt$ -- conditioned on $Σ$ which is defined as the local first moment of $Ω$ in $Δt$. We find that the conditional distribution $P(Ω| Σ)$ undergoes a qualitative change in behavior as $Σ$ increases beyond a critical threshold $Σ_c$. For $Σ<Σ_c$, $P(Ω|Σ)$ displays a maximum at $Ω=0$, i.e., trades in $Δt$ are equally likely to be buyer initiated or seller initiated. For $Σ> Σ_c$, $Ω=0$ becomes a local minimum and two new maxima $Ω_{+}$ and $Ω_{-}$ appear at non-zero values of $Ω$, i.e., trades in $Δt$ are either predominantly buyer initiated or predominantly seller initiated. We interpret these results using a Langevin equation with multiplicative noise.

cond-mat.stat-mech↗
arXiv2001

Quantifying Stock Price Response to Demand Fluctuations

We address the question of how stock prices respond to changes in demand. We quantify the relations between price change $G$ over a time interval $Δt$ and two different measures of demand fluctuations: (a) $Φ$, defined as the difference between the number of buyer-initiated and seller-initiated trades, and (b) $Ω$, defined as the difference in number of shares traded in buyer and seller initiated trades. We find that the conditional expectations $ _Ω$ and $ _Φ$ of price change for a given $Ω$ or $Φ$ are both concave. We find that large price fluctuations occur when demand is very small --- a fact which is reminiscent of large fluctuations that occur at critical points in spin systems, where the divergent nature of the response function leads to large fluctuations.

cond-mat.stat-mech↗
arXiv2000

Identifying Business Sectors from Stock Price Fluctuations

Firms having similar business activities are correlated. We analyze two different cross-correlation matrices C constructed from (i) 30-min price fluctuations of 1000 US stocks for the 2-year period 1994-95 and (ii) 1-day price fluctuations of 422 US stocks for the 35-year period 1962-96. We find that the eigenvectors of C corresponding to the largest eigenvalues allow us to partition the set of all stocks into distinct subsets. These subsets are similar to conventionally-identified business sectors, and are stable for extended periods of time. Using a set of coupled stochastic differential equations, we argue how correlations between stocks might arise. Finally, we demonstrate that the sectors we identify are useful for the practical goal of finding an investment which earns a given return without exposure to unnecessary risk.

cond-mat.stat-mech↗
arXiv2000

Statistical Properties of Share Volume Traded in Financial Markets

We quantitatively investigate the ideas behind the often-expressed adage `it takes volume to move stock prices', and study the statistical properties of the number of shares traded $Q_{Δt}$ for a given stock in a fixed time interval $Δt$. We analyze transaction data for the largest 1000 stocks for the two-year period 1994-95, using a database that records every transaction for all securities in three major US stock markets. We find that the distribution $P(Q_{Δt})$ displays a power-law decay, and that the time correlations in $Q_{Δt}$ display long-range persistence. Further, we investigate the relation between $Q_{Δt}$ and the number of transactions $N_{Δt}$ in a time interval $Δt$, and find that the long-range correlations in $Q_{Δt}$ are largely due to those of $N_{Δt}$. Our results are consistent with the interpretation that the large equal-time correlation previously found between $Q_{Δt}$ and the absolute value of price change $| G_{Δt} |$ (related to volatility) are largely due to $N_{Δt}$.

cond-mat.stat-mech↗
arXiv1999

Economic Fluctuations and Diffusion

Stock price changes occur through transactions, just as diffusion in physical systems occurs through molecular collisions. We systematically explore this analogy and quantify the relation between trading activity - measured by the number of transactions $N_{Δt}$ - and the price change $G_{Δt}$, for a given stock, over a time interval $[t, t+Δt]$. To this end, we analyze a database documenting every transaction for 1000 US stocks over the two-year period 1994-1995. We find that price movements are equivalent to a complex variant of diffusion, where the diffusion coefficient fluctuates drastically in time. We relate the analog of the diffusion coefficient to two microscopic quantities: (i) the number of transactions $N_{Δt}$ in $Δt$, which is the analog of the number of collisions and (ii) the local variance $w^2_{Δt}$ of the price changes for all transactions in $Δt$, which is the analog of the local mean square displacement between collisions. We study the distributions of both $N_{Δt}$ and $w_{Δt}$, and find that they display power-law tails. Further, we find that $N_{Δt}$ displays long-range power-law correlations in time, whereas $w_{Δt}$ does not. Our results are consistent with the interpretation that the pronounced tails of the distribution of $G_{Δt} are due to $w_{Δt}$, and that the long-range correlations previously found for $| G_{Δt} |$ are due to $N_{Δt}$.

cond-mat.stat-mech↗
arXiv1999

Ivory Tower Universities and Competitive Business Firms

There is nowadays considerable interest on ways to quantify the dynamics of research activities, in part due to recent changes in research and development (R&D) funding. Here, we seek to quantify and analyze university research activities, and compare their growth dynamics with those of business firms. Specifically, we analyze five distinct databases, the largest of which is a National Science Foundation database of the R&D expenditures for science and engineering of 719 United States (US) universities for the 17-year period 1979--1995. We find that the distribution of growth rates displays a ``universal'' form that does not depend on the size of the university or on the measure of size used, and that the width of this distribution decays with size as a power law. Our findings are quantitatively similar to those independently uncovered for business firms, and consistent with the hypothesis that the growth dynamics of complex organizations may be governed by universal mechanisms.

cond-mat.stat-mech↗
arXiv1999

Scaling of the distribution of fluctuations of financial market indices

We study the distribution of fluctuations over a time scale $Δt$ (i.e., the returns) of the S&P 500 index by analyzing three distinct databases. Database (i) contains approximately 1 million records sampled at 1 min intervals for the 13-year period 1984-1996, database (ii) contains 8686 daily records for the 35-year period 1962-1996, and database (iii) contains 852 monthly records for the 71-year period 1926-1996. We compute the probability distributions of returns over a time scale $Δt$, where $Δt$ varies approximately over a factor of 10^4 - from 1 min up to more than 1 month. We find that the distributions for $Δt \leq$ 4 days (1560 mins) are consistent with a power-law asymptotic behavior, characterized by an exponent $α\approx 3$, well outside the stable Lévy regime $0 < α< 2$. To test the robustness of the S&P result, we perform a parallel analysis on two other financial market indices. Database (iv) contains 3560 daily records of the NIKKEI index for the 14-year period 1984-97, and database (v) contains 4649 daily records of the Hang-Seng index for the 18-year period 1980-97. We find estimates of $α$ consistent with those describing the distribution of S&P 500 daily-returns. One possible reason for the scaling of these distributions is the long persistence of the autocorrelation function of the volatility. For time scales longer than $(Δt)_{\times} \approx 4$ days, our results are consistent with slow convergence to Gaussian behavior.

cond-mat.stat-mech↗
arXiv1999

Universal and non-universal properties of cross-correlations in financial time series

We use methods of random matrix theory to analyze the cross-correlation matrix C of price changes of the largest 1000 US stocks for the 2-year period 1994-95. We find that the statistics of most of the eigenvalues in the spectrum of C agree with the predictions of random matrix theory, but there are deviations for a few of the largest eigenvalues. We find that C has the universal properties of the Gaussian orthogonal ensemble of random matrices. Furthermore, we analyze the eigenvectors of C through their inverse participation ratio and find eigenvectors with large inverse participation ratios at both edges of the eigenvalue spectrum--a situation reminiscent of results in localization theory.

cond-mat.stat-mech↗
arXiv1998

Conductances, Conductance Fluctuations, and Level Statistics on the Surface of Multilayer Quantum Hall States

The transport properties on the two-dimensional surface of coupled multilayer heterostructures are studied in the integer quantum Hall states. We emphasize the criticality of the surface state and the phase coherent transport properties in the thermodynamic limit. A new, stable numerical algorithm for large scale conductance calculations in the transfer matrix approach is discussed in detail. It is then applied to a directed network model describing the quantum mechanical tunneling and impurity scattering of the multilayer edge states. We calculate the two-probe conductance in the direction parallel to the external magnetic field, its fluctuations and statistical distributions as a function of the interlayer tunneling strength. Using finite size scaling, the asymptotic scaling functions of the ensemble averaged conductance and the conductance fluctuations are calculated for a fixed aspect ratio and found to be in remarkable agreement with the analytical results obtained using the supersymmetric nonlinear $σ$-model approach. The conductance distribution is determined in the quasi-one-dimensional metallic, insulating, as well as the crossover regime where comparisons are made to that at the single-layer quantum Hall transition. We present, for the first time, a detailed study of the level statistics in the eigenvalue spectrum of the transfer matrix. Coexistence of metallic and insulating statistics is observed in the crossover regime, which is attributed to the emergence of a finite range level repulsion in the crossover regime, separating the metallic (Wigner-surmise) behavior at small level spacings from the insulating (uncorrelated or Poisson) behavior at large level spacings.

cond-mat.mes-hall↗