The role of firm size and firm age in employment growth: evidence for Slovenia, 1996-2013.

• Author: Banerjee, Biswajit
• Position: Report

1. Introduction

   Whether small businesses create more jobs than the larger ones is a subject of much debate. A well-known hypothesis on firm growth, the so-called Gibrat's law, is that growth rates of firms are independent of their size (see Sutton, 1997). However, findings of Birch (1981) and a large number of subsequent studies have led to the conventional wisdom that net employment growth is negatively related to firm size.

   Davis et al. (1996) argue that relating firm-size and net job growth is prone to bias because of regression-to-the mean effects. Transitory shocks to employment or random measurement errors could result in firms being categorized smaller or larger than their "typical" size. If these transitory shocks or random measurement errors were not highly serially correlated, this would lead to upward bias in the estimated growth rate of small firms and downward bias in the estimated growth rate of large firms. To mitigate the effects of regression to the mean, Davis et al. (1996) propose using a classification based on the average size, measured as the simple average of firm size in base year t-1 and current year t. (4) Based on the average size class method, Davis et al. find no systematic relationship between manufacturing plant size and employment growth in the United States.

   Neumark et al. (2011) point out that the finding of Davis et al. (1996) does not hold in all cases. They cite studies on Canada, Greece, Netherlands, Sweden and the United Kingdom based on the average size class method that find greater net job growth in small businesses. In addition, in their own study on the United States for the period 1992-2004, based on the average size class method and nonparametric regression technique, Neumark et al. find that net job growth is greater in small firms and establishments.

   Haltiwanger et al. (2010, 2013) argue that researchers may be confusing firm-size effects with the effects of firm age. New firms are an important source of net job growth and are typically small when they start operations. Thus, in the absence of controls for firm age (which is typical in much of the job flows literature), the finding of a negative relationship between firm size and net growth rates could be attributable to most new firms being classified in the small size classes. When controls are included for firm age and the average size classification is used, Haltiwanger et al. find a positive relationship between net growth and firm size. A similar result is obtained for Austria by Huber et al. (2012), for Canada by Dixon and Rollin (2012), for Hungary by Earle and Telegdy (2011), and for Ireland by Lawless (2014).

   The findings on the relationship between net employment growth and firm age, with controls included for firm size, are less clear cut. Earle and Telegdy (2011) for Hungary, Haltiwanger et al. (2010, 2013) for the United States and Lawless (2014) for Ireland find that young firms grow more rapidly than the more mature firms but that the relationship is relatively flat after the age of 5 years. However, this finding is not supported by evidence from Austria and Canada. Huber et al. (2012) for Austria find that the average marginal impact on net job creation is greater for older firms than for young firms. For Canada, Dixon and Rollin (2012) find that employment growth declines sharply between one-year old and two-year old firms but that after the second year there is a positive relationship between firm age and employment growth.

   A related issue is whether the cyclical dynamics of employment growth are different for firms of different size and age. Moscarini and Postel-Vinay (2012) find that in the United States, Denmark and France, large firms have higher net employment growth than small firms when unemployment is below trend, and lower net employment growth when unemployment is above trend. However, Fort et al. (2013) document cyclical dynamics across firms by both firm size and firm age and obtain a different pattern. They find that young, small firms are cyclically more sensitive than large and mature firms. In particular, the decline in net employment growth in the United States during the 2007-2009 recession experienced by young, small firms was higher than that experienced by large and mature firms.

   In this paper we examine the role of firm size and firm age on job creation and net employment growth in Slovenia across all sectors of the economy during the period 1996-2013 and separately during three sub-periods of distinct cyclical dynamics: 1996-2003, 2004-2008, and 2009-2013. The first period, 1996-2003, includes the middle and late-transition periods prior to European Union (EU) accession. During this period, real GDP growth averaged about 4 percent annually while fluctuating around a mildly declining trend and net employment growth rate in the private sector fluctuated narrowly around a horizontal trend of zero growth. The second period, 2004-2008, spans EU entry in 2004, the two-year stay in Exchange Rate Mechanism II, euro adoption in January 2007, and the eve of the onset of the global financial crisis. A mild expansionary phase began with EU entry in 2004 and both real GDP growth and net employment growth picked up pace and peaked with euro adoption in 2007. A slowdown began in Q4, 2008 as the impact of the global crisis and structural weaknesses in the banking system and fiscal sector began to be felt. The third period, 2009-2013, is one of prolonged recession. Real GDP growth and net employment growth turned sharply negative in 2009 and remained in negative territory in the subsequent years.

   This paper follows the nonparametric regression approach used by Haltiwanger et al. (2010, 2013) and adds to the limited number of studies that have looked at the effects of firm size and firm age on job growth in a multivariate context. Besides, an inquiry in to the features of employment growth in Slovenia is of interest because the country is a member of the euro area. Being a member of a monetary union, Slovenia does not have the option of using exchange rate policy to respond to shocks and the importance of labor market flexibility as an adjustment mechanism becomes greater. The presence of significant institutional rigidities renders the task of macroeconomic stabilization more difficult. The findings on the type of firms where jobs are being created should be of use to policy makers in the formulation of employment promotion strategies. (5) In particular, the analysis will shed light on whether policies targeting firms based on size per se are sufficient or whether attention also should be focused on creating an environment that facilitates start-ups, irrespective of size. This issue is of considerable relevance for all EU member states, since supporting small and medium-sized enterprises (SME) is an important policy objective for them. The European Commission annually publishes fact sheets on trends and national policies affecting SMEs in individual member states (see, for example, European Commission, 2014).

   The rest of the paper is organized as follows. Section 2 describes the data base, measurement of job flows, and methodology. Section 3 presents the empirical findings and Section 4 concludes.

2. Data, job flows definitions, and methodology

   2.1 Data

   The analysis in this paper is based on annual firm-level data from the Slovenian Business Register maintained by the Agency for Public Legal Records and Related Services (AJPES). The data base pertains to all private business entities, their subsidiaries, and other organization segments which perform profitable or non-profitable activities. Enterprises (including banks, insurance companies, investment funds and cooperatives), sole proprietors, legal entities governed by public law, and non-profit organizations have to present their annual reports to AJPES for the purpose of presenting them publicly and for tax and statistical purposes.

   The AJPES data base includes information on firms' financial statements, fulltime equivalent (FTE) employment, industrial affiliation, location, and incorporation year of the firm to the business register. The analysis is confined to firms that have at least one FTE employee.

   The sample data set comprises between 25,000 and 37,000 annual observations on firms over the 1996 to 2013 period. The FTE employment of the firms included in the sample accounted for each year, on average, about 58 percent of the total number of employees in the economy (national accounts measure). The gap is mainly explained by the exclusion of the government sector from the sample and by the fact that the national accounts measure of employees includes part-time workers. The trend in growth of FTE employment for the firms in the sample was broadly similar to that for the growth of the national accounts measure of employees (the simple correlation was 0.893).

   Table 1 shows the distribution of firms and employment by current-year size and age, averaged over the sample period. Small firms, defined as having less than 10 FTE employees, constitute the bulk of the firms in the sample (82 percent) but account for a rather small share of total employment (14 percent)...