The analysis is carried out over the 2013–2021 period for a strongly balanced dataset of 16463 Italian firms that between 2013 and 2021 have entered a network. To select only those Italian firms that have signed a network agreement during the time span of our analysis, we rely upon the InfoCamere database which has been largely used in several studies in the literature on the Italian network performances. InfoCamere provides information on the number of i) firms entering networks, and ii) networks established since 2010. At the end of 2021 the outstanding number of Italian firms belonging to a network was 42231 for more than 7000 network contracts. Therefore, we first picked up information of those firms that entered a network up to 2021 and then we matched such information with the microeconomic data available on the Bureau van Dijk AIDA database. Due to data availability in the AIDA database, we have been able to retrieve data for a sample of 16463 firms during the period 2013–2021. Firms in the sample are very heterogeneous among them thus allowing us to capture a wide range of variation of business activities and network types across firms.
Figure 2 shows how firms in our sample have gradually joined a network during 2013–2021. While firms were hesitant to join a network in the early years after the Italian law came into effect, the number of firms signing an IFNA grew significantly after 2016. The majority of firms are located in the regions of Lombardia (2433, see Fig. 3), followed by Lazio (2168), Veneto (1577), and Emilia Romagna (1488). On the opposite, Valle D’Aosta (43), Molise (49), and Basilicata (172) are the Italian regions with the lowest number of firms that entered a network in our sample.
We have relied upon the AIDA database to collect firm-specific data used in the existing literature to explore the performance of firms. As well as total revenues and added value which are commonly used as performance indicators, we have also gathered data on variables that are usually correlated to performance, i.e., physical and non-physical capital, number of workers, salaries, and the debt/equity ratio. The number of workers has been used to compute the average size of each firm (micro, small, medium, and large) and the physical and non-physical capital intensity. One of the main novelties of our contribution is to investigate whether the market power of firms could be influenced by an early or delayed entry in a network. Markup is defined by the ratio of marginal costs to output sales prices. However, the lack of data on marginal costs and prices often makes markup calculation a very hard task. We can identify four distinct approaches that attempt to overcome the markup measurement problem (De Loecker et al. 2020). Berry et al. (1995) compute markup through first-order conditions in the context of the optimal pricing problem and estimating price elasticity of demand; this approach cannot be followed in our analysis because it requires knowledge of final prices and output levels. A second way to calculate markup is that introduced by Hall (1988), relying upon the firm’s cost minimization problem (see also De Loecker and Warzynski 2012). This approach does not imply specific return to scale or elasticity of substitution between inputs; however, it requires to work with specific production functions to estimate the output elasticity of at least one variable input. The accounting approach, recently implemented in Karabarbounis and Naiman (2018), considers the profit margin by replacing marginal cost (unobservable) with average cost. This method imposes important restrictions on the cost structure of individual firms, as fixed costs are assumed equal to zero (no economies of scale) and requires constant return to scale. However, given that most of network firms are usually SMEs, we can consider the above assumptions as reasonable. The last approach is the one proposed by Griffith et al. (2010), we compute mark-up as the ratio between added value and total production cost. Such a measure of firm’s profitability can be considered as a proxy of the Lerner index under the assumption of constant returns to scale.3 In this paper, we will compute markup according to Griffith et al. (2010).
Figures 4 and 5 plot the yearly average values of revenues, added value, and markup for those firms that have entered networks or not yet. It emerges that joining a network offers firms benefits in terms of increased revenue and value added compared to not being part of a network, and decrease or stabilize the market power of firms.
Following the criteria provided by ISTAT (Italian National Statistical Institute), we also build regional, size, and sectorial dummies according to, respectively, the geographical location of each firms (North, Centre, South of Italy), the average number of workers during 2013–2021 of each firm (micro-, small-, medium-, and large-sized), and the type of activity carried out by each firms using the ATECO 2007 classification (firms working in the primary, secondary, or tertiary sector of the economy).4 Firms are mostly located in the North of Italy (47%), followed by Centre (30%), and South (23%), and they are mainly engaged in the tertiary (59%) and in the secondary (35%) sectors of the economy; only a share of 6% of the sample works in the primary sector (see Fig. 6). The 86% of the sample consists of micro- and small-sized firms, thus confirming that IFNAs are almost exclusively devoted to small-medium enterprises (SMEs). In addition, to capture whether an early or delayed entry into a network may produce differences in terms of total revenues, added value, and markup, we have generated a dummy variable called “network” which is equal to one for each year a firm belongs to a network of firms and equal to zero otherwise.
Source: our elaboration of InfoCamere data and Aida database.
A detailed description of variables and data sources is provided by Table 2. Descriptive statistics and pairwise correlation coefficients are reported, respectively, in Tables 3 and 4.
Table 2
Data description and sources
Variable | Definition | Source |
Mark-up | Ratio between added value and total cost of production. | Bureau van Dijk Aida database. |
Added value | Added value (thousands of euro). | Bureau van Dijk Aida database. |
Revenue | Total revenue from the sales of goods and services (thousands of euro). | Bureau van Dijk Aida database. |
Ph_capital intensity | Physical capital intensity computed as the ratio between tangible assets and the number of workers. | Bureau van Dijk Aida database. |
Debt_equity | Debt/equity ratio. | Bureau van Dijk Aida database. |
Wages | Total wages (thousands of euro). | Bureau van Dijk Aida database. |
Non-tangible capital intensity | Non-tangible capital intensity computed as the ratio between non-tangible assets and the number of workers. | Bureau van Dijk Aida database. |
North | Dummy variable equal to 1 if the firm is located in the north of Italy, and equal to zero otherwise. | |
Center | Dummy variable equal to 1 if the firm is located in the center of Italy, and equal to zero otherwise. | |
South | Dummy variable equal to 1 if the firm is located in the south of Italy, and equal to zero otherwise. | |
Micro size | Dummy variable equal to 1 if the average number of workers during the 2013–2021 period is < 10, and equal to zero otherwise. | |
Small size | Dummy variable equal to 1 if the average number of workers during the 2013–2021 period is larger than 10 and smaller than 50, and equal to zero otherwise. | |
Medium size | Dummy variable equal to 1 if the average number of workers during the 2013–2021 period is larger than 50 and smaller than 250, and equal to zero otherwise. | |
Large size | Dummy variable equal to 1 if the average number of workers during the 2013–2021 period is larger than 250, and equal to zero otherwise. | |
Ateco 1 | Dummy variable equal to 1 if the firm is employed in the primary sector of the economy, and equal to zero otherwise. | |
Ateco 2 | Dummy variable equal to 1 if the firm is employed in the secondary sector of the economy, and equal to zero otherwise. | |
Ateco 3 | Dummy variable equal to 1 if the firm is employed in the tertiary sector of the economy, and equal to zero otherwise. | |
network | Dummy variable equal to 1 for each year a firm belongs to a network of firms, and equal to zero otherwise. | |
Table 3
Variables | Obs. | Mean | Std. dev. | Min | Max |
Mark-up | 116950 | 9.968 | 1342.058 | -0.809 | 8.769 |
Added value | 110174 | 6.038 | 1.959 | -6.907 | 15.379 |
Revenue | 115947 | 7.106 | 2.024 | -6.907 | 17.168 |
Ph_capital intensity | 101919 | 2.637 | 2.114 | -11.258 | 10.733 |
Debt_equity | 63,976 | -0.201 | 1.676 | -4.605 | 6.901 |
Wages | 105269 | 5.553 | 1.774 | -6.907 | 13.839 |
Non-tangible capital intensity | 88368 | 0.656 | 2.275 | -13.774 | 13.131 |
Table 4
Variables | (1) | (2) | (3) | (4) | (5) | (6) | (7) |
(1) Mark-up | | | | | | | |
(2) Added value | 0.010*** | 1 | | | | | |
(3) Revenue | -0.033*** | 0.824*** | 1 | | | | |
(4) Ph_capital intensity | -0.004 | 0.181*** | 0.224*** | 1 | | | |
(5) Debt_equity | -0.012*** | -0.062*** | -0.025*** | 0.023*** | 1 | | |
(6) Wages | -0.029*** | 0.860*** | 0.842*** | 0.082*** | -0.046*** | 1 | |
(7) Non-tangible capital intensity | 0.008** | -0.000 | 0.025*** | 0.224*** | 0.097*** | -0.064*** | 1 |
*, **, *** denote, respectively, statistical significance at 10%, 5% and 1%.