Author:

Abstract:

Gabaix (2011) coined the term 'granularity' to describe how idiosyncratic shocks to large firms impact the aggregate volatility. Central to the concept of granularity is the fact that at the micro-level the distribution of firm sizes is very skewed. These large firms are so disproportionally large that they do not average out in the aggregate. Therefore, the common notion of the 'representative firm' no longer holds. Shocks to granular firms can impact aggregate volatility (such as GDP), either directly or because shocks to these firms propagate through the production network. In the academic literature, notably Acemoglu et al. (2012, 2017, 2020), Baqaee et al. (2018, 2019), Carvalho et al. (2019a, 2019b, 2021) and di Giovanni et al. (2014, 2017) made key contributions in this field in the last decade. Granular firms do not only impact aggregate volatility, they can also have an impact on the specialization patterns of countries (Gaubert and Itskhoki, 2021) Large firms do not only shape export patterns but also the domestic specialization patterns. Large firms need an ecosystem to flourish and each ecosystem needs large firms. National (innovation) policies aim to boost those sectors where the country is specialized in. Since the seminal work of Porter (1998), the importance of clusters for an economy has been widely researched. Using recent data from the EU Cluster Observatory, Ketels et al. (2021) assess the impact of clusters on regional prosperity across Europe. In addition to geographical clusters, more recent literature also looks at cluster policy initiatives (Wilson et al., 2022). These organizational clusters bring together companies and institutions that are not necessarily geographically connected, with the aim to boost innovation and competitiveness within the cluster and beyond. The Flemish Spearhead Cluster policy is an example of such a cluster initiative. The initiative is focused around 7 cluster areas, in each of which Flanders is specialized. This PhD thesis taps into the above mentioned research on granularity and cluster performance and contributes to it in several ways. Chapter 1, "The granular nature of emerging market economies: The case of Kazakhstan" builds on the methodology developed by Gabaix (2011) and Blanco-Arroyo et al. (2018) in order to assess the granular hypothesis and to quantify the number of granular firms. My co-authors and I investigate to what extent idiosyncratic shocks to Total Factor Productivity (TFP) explain aggregate fluctuations in TFP. This study contributes to the existing literature as it is one of the first to analyze granularity in an emerging economy, namely Kazakhstan. Making use of a confidential quarterly firm-level database between 2002 and 2018, including information on the ownership structure, we find that the Kazakh economy is highly granular, with the largest 30 firms explaining nearly 80 percent of the growth in aggregate TFP. Chapter 2, entitled "Granularity at work: The impact of large firms on aggregate employment and business dynamism" looks at granularity from a different angle. Rather than identifying large firms based on their sales, I base it on the employment size (FTE). I also move from an economy in transition (Kazakhstan), to a small open economy (Belgium). Using confidential quarterly data from the National Social Security Office (NSSO), I assess whether the Belgian economy is granular and characterize the type of firms that are considered granular. In addition, I also investigate the role these granular firms have in observed aggregate phenomena, such as the decline in business dynamism (Bijnens et al., 2020). This paper links the literature on granularity with the literature on labor economics and provides new insights to understand the evolution of business dynamism. I find that the employment agencies sector is granular and responsible for 55 percent of the observed decline in business dynamism over the period 2006-2019. In Chapter 3, "Do cluster initiatives boost productivity? The impact of cluster membership on firm productivity", my co-authors and I apply a Difference-in-Differences method in order to assess whether cluster membership has an impact on firm level productivity. For this research we use firm-level accounting data for the period 2013-2020. This paper is the first to analyze the firm-level impact of the Flemish cluster policy on productivity. We also apply an innovative method to define cluster membership. We find that becoming a member of a cluster has an average positive impact on firm level TFP of 1 to 3 percent, depending on the econometric specification.