Essays in Macroeconomics and Firm Dynamics

Abstract

This dissertation consists of three chapters in macroeconomics and firm dynamics. They study macroeconomic questions based on micro-founded models and data. They investigate how businesses react to the technological progress in information technology and artificial intelligence, as well as government interventions in the economy. The first chapter develops a new theory of firm size and firm dynamics. It employs personnel economics approaches within the span of control model in macroeconomics. The model generates increasing concentration and decreasing labor share of income over time without forcing any inefficiency. Personnel management and entrepreneurial risk in the model are the key determinants of the overwhelmingly discussed recent macro trends. A declining supervision cost enables better managers to expand their firms, resulting in a more concentrated economy and a riskier entrepreneurial environment. Risk-taking entrepreneurs are compensated with a growing profit leading to a diminishing labor share. The second chapter studies the impact of state capitalism in the form of state-owned enterprises (SOEs). Focusing on the SOEs in the Middle East and North Africa (MENA), it analyzes big data of more than 50 million firm-level observations covering 74 countries worldwide. It finds that relative to private firms, SOEs are born big but grow more slowly. Moreover, while the SOEs in MENA are less profitable and less productive, they receive cheaper loans. In addition to distorting capital allocation, they weaken economic competition and drop business dynamism. The third chapter studies customer data, which many consider the most valuable resource in today's economy, as an intangible capital, and pioneers a method to measure its economic value. It builds a novel database by merging Compustat with online clickstream data recording clicks of around 200 million users. That provides proxies for data inflow based on visit metrics. This chapter documents that firms' stock of customer data is distributed according to a right-skewed log-normal distribution with a fat tail. Furthermore, it reports a positive correlation between sales and data inflow and between data stock and profit as well as productivity. Relying on these insights, it develops a theory of customer data, where monopolistic firms have three departments: a usual Dixit-Stiglitz production department, a sales department collecting customer data as a byproduct, and a data department investing in software to extract taste predictions from data. It proves that firms with large data stocks have lower labor shares and face unfair antitrust policies. Additionally, this chapter leverages the COVID-19 pandemic lockdown as a natural experiment to identify the model. It finds that, in retail, slightly higher than one-third of firm profit comes from its stock of customer data.