Theory of Zipf's law and beyond (Q1039386)

Zipf's law usually refers to the formula \(\Pr\{S>s\} \sim s^{-1}\), where \(S\) is some stochastic variable. The distribution most studied from the perspective of Zipf's law is that of firm size, where size is proxied by sales, income, number of employees, or total assets. One of the most promising explanations proposed to understand the origin of Zipf's law is based on Gibrat's law of proportionate growth, which states that size of a firm and its growth are independent. A goal of the present book is to demonstrate that birth as well as death processes are especially important to understand the economic foundation of Zipf's law and its robustness. The authors consider two different mechanisms for the exit of a firm. The first is when the firm total asset value becomes smaller than a given minimum threshold, and the second, when an exogenous shock occurs independently of the size of the firm. The book is built on the realistic description of the behavior of the asset value of firms, which assumes that the births of firms occur according to a random point process characterized by some intensity. The approach in terms of the dynamics of birth-death with stochastic growth, developed by the authors, leads to specific predictions of the conditions under which deviations from Zipf's law occur, which help rationalize the empirical evidence documented in the literature. The book is organized as follows. Chapter 2, appearing directly after introduction, presents the continuous version of Gibrat's law and some peculiarities of the stochastic behavior of the geometric Brownian motion of firm's asset values, resulting from this law. Chapter 3 describes the proposed model for the current density of firm's asset values, taking into account the random flow of the birth of firms. In chapter 4, the authors gather useful properties of the geometric Brownian motion, which play a significant role for the understanding of the roots and conditions of the validity of Zipf's law. Chapter 5 discusses in detail how the occurrence of the death of firms, when their value falls below some low level, influences the validity of Zipf's law. In Chapter 6, the authors derive an equation for the steady-state density of firm asset values, which enables them to explore in detail the consequences of deviations from Gibrat's law at moderate values on the validity of Zipf's law at higher asset values. Chapter 7 is devoted to discussing the condition for the existence of Zipf's law and the circumstances under which deviations from it occur, when taking into account the possibility of sudden death of firms occurring even for large sizes. Chapter 8 provides the most general treatment taking into account time dependence of birth rates, sizes at birth, and minimum firm sizes. In chapter 9, the authors analyze the extent to which the results presented in previous chapters can be used for description of a single realization of a finite economy. Chapter 10 concludes and outlines future research directions.