Transition in the Waiting-Time Distribution of Price-Change Events in a Global Socioeconomic System

摘要

The goal of developing a firmer theoretical understanding of inhomogenoustemporal processes -- in particular, the waiting times in some collectivedynamical system -- is attracting significant interest among physicists.Quantifying the deviations in the waiting-time distribution away from onegenerated by a random process, may help unravel the feedback mechanisms thatdrive the underlying dynamics. We analyze the waiting-time distributions ofhigh frequency foreign exchange data for the best executable bid-ask pricesacross all major currencies. We find that the lognormal distribution yields agood overall fit for the waiting-time distribution between currency ratechanges if both short and long waiting times are included. If we restrict ourstudy to long waiting-times, each currency pair's distribution is consistentwith a power law tail with exponent near to 3.5. However for short waitingtimes, the overall distribution resembles one generated by an archetypalcomplex systems model in which boundedly rational agents compete for limitedresources. Our findings suggest a gradual transition arises in trading behaviorbetween a fast regime in which traders act in a boundedly rational way, and aslower one in which traders' decisions are driven by generic feedbackmechanisms across multiple timescales and hence produce similar power-law tailsirrespective of currency type.