Robert L. Axtell

Employment growth through labor flow networks.

It is conventional in labor economics to treat all workers who are seeking new jobs as belonging to a labor pool, and all firms that have job vacancies as an employer pool, and then match workers to jobs. Here we develop a new approach to study labor and firm dynamics. By combining the emerging science of networks with newly available employment micro-data, comprehensive at the level of whole countries, we are able to broadly characterize the process through which workers move between firms. Specifically, for each firm in an economy as a node in a graph, we draw edges between firms if a worker has migrated between them, possibly with a spell of unemployment in between. An economys overall graph of firm-worker interactions is an object we call the labor flow network (LFN). This is the first study that characterizes a LFN for an entire economy. We explore the properties of this network, including its topology, its community structure, and its relationship to economic variables. It is shown that LFNs can be useful in identifying firms with high growth potential. We relate LFNs to other notions of high performance firms. Specifically, it is shown that fewer than 10% of firms account for nearly 90% of all employment growth. We conclude with a model in which empirically-salient LFNs emerge from the interaction of heterogeneous adaptive agents in a decentralized labor market.

Using Agentization for Exploring Firm and Labor Dynamics

Agentization is the process of rendering neoclassical models into computational ones. This methodological tool can be used to analyze and test neoclassical theories under a more flexible computational framework. This paper presents agentization and its methodological framework. We propose that, by classifying the assumptions of a neoclassical model, it is possible to systematically analyze their influence in the predictions of a theory. Furthermore, agentization allows the researcher to explore the potentials and limitations of theories. We present an example by agentizing the model of Gabaix (1999) for the emergence of Zipf laws. We show that the agentized model is able to reproduce the main features of the Gabaix process, without holding neoclassical assumptions such as equilibrium, rationality, agent homogeneity, and centralized anonymous interactions. Additionally, the model generates stylized facts such as tent-shaped firm growth rates distributions, and the employer-size wage premium. These regularities are not considered in the neoclassical model. Thus, allows the researcher to explore the boundaries and potentials of the theory.

An Explanation of Universality in Growth Fluctuations

Phenomena as diverse as breeding bird populations, the size of U.S. firms, money invested in mutual funds, and the scientific output of universities all show unusual but remarkably similar growth fluctuations. The fluctuations display characteristic features, including heavy tails and anomalous power law scaling of the standard deviation as a function of size. Many theories have now been put forward to explain this, all of them based on modifications and extensions of proportional growth of subunits. We analyze data from bird populations, firms, and mutual funds and show that the growth fluctuations match a Levy distribution very well. This was previously suggested by Wyart and Bouchaud and Gabaix, but until now never tested. However, we show that their theory (and indeed all previous theories) are ruled out, at least for these three data sets, because they require size distributions that are too heavy tailed. We introduce a simple additive replication model, in which groups (such as firms) grow by replacing each of their members by a random number of new members. To demonstrate how the individual growth fluctuations can be heavy-tailed even though the sizes are not, we propose a model based on stochastic influence dynamics over a scale-free contact network, and show that it produces the correct behavior. We generalize the model to the case where some groups are preferred over others, and show that this can lead to a breakdown of the anomalous scaling, which appears to be observed for some other data sets.

The Network Picture of Labor Flow

We construct a data-driven model of flows in graphs that captures the essential elements of the movement of workers between jobs in the companies (firms) of entire economic systems such as countries. The model is based on the observation that certain job transitions between firms are often repeated over time, showing persistent behavior, and suggesting the construction of static graphs to act as the scaffolding for job mobility. Individuals in the job market (the workforce) are modelled by a discrete-time random walk on graphs, where each individual at a node can possess two states: employed or unemployed, and the rates of becoming unemployed and of finding a new job are node dependent parameters. We calculate the steady state solution of the model and compare it to extensive micro-datasets for Mexico and Finland, comprised of hundreds of thousands of firms and individuals. We find that our model possesses the correct behavior for the numbers of employed and unemployed individuals in these countries down to the level of individual firms. Our framework opens the door to a new approach to the analysis of labor mobility at high resolution, with the tantalizing potential for the development of full forecasting methods in the future.

Endogenous Dynamics of Multi-Agent Firms

A model is described in which large numbers of simple agents organize into groups that empirically resemble U.S. firms. The agents work in team production environments, regularly adjust their work effort, and periodically seek better jobs or start new teams when it is in their self-interest. Nash equilibria of the team formation game exist but are unstable. Dynamics are studied using agent computing at full-scale with the U.S. private sector (120 million agents). Stationary distributions arise at the aggregate level despite perpetual adaptation at the agent level. Such agent adjustments occur for microeconomic reasons without the need for external shocks.

The Network Composition of Aggregate Unemployment

We develop an alternative framework to the aggregate matching function in which workers search for jobs through a network of firms: the labor flow network. The lack of an edge between two companies indicates the impossibility of labor flows between them due to high frictions. In equilibrium, firms’ hiring behavior correlates through the network, generating highly disaggregated local unemployment. Hence, aggregation depends on the topology of the network in non-trivial ways. This theory provides new micro-foundations for the the Beveridge curve, wage dispersion, and the employersize premium. Using employer-employee matched records, we find that the empirical topology of the network, in conjunction with the supply elasticity, may be a major contributor of aggregate unemployment.

A computational approach to managing coupled human–environmental systems: the POSEIDON model of ocean fisheries

Sustainable management of complex human–environment systems, and the essential services they provide, remains a major challenge, felt from local to global scales. These systems are typically highly dynamic and hard to predict, particularly in the context of rapid environmental change, where novel sets of conditions drive coupled socio-economic-environmental responses. Faced with these challenges, our tools for policy development, while informed by the past experience, must not be unduly constrained; they must allow equally for both the fine-tuning of successful existing approaches and the generation of novel ones in unbiased ways. We study ocean fisheries as an example class of complex human–environmental systems, and present a new model (POSEIDON) and computational approach to policy design. The model includes an adaptive agent-based representation of a fishing fleet, coupled to a simplified ocean ecology model. The agents (fishing boats) do not have programmed responses based on empirical data, but respond adaptively, as a group, to their environment (including policy constraints). This conceptual model captures qualitatively a wide range of empirically observed fleet behaviour, in response to a broad set of policies. Within this framework, we define policy objectives (of arbitrary complexity) and use Bayesian optimization over multiple model runs to find policy parameters that best meet the goals. The trade-offs inherent in this approach are explored explicitly. Taking this further, optimization is used to generate novel hybrid policies. We illustrate this approach using simulated examples, in which policy prescriptions generated by our computational methods are counterintuitive and thus unlikely to be identified by conventional frameworks.