https://orcid.org/0000-0003-4015-4939
![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
Abstract
This paper analyzes the drivers of firms' technical efficiency among exporting Italian small-to-medium enterprises (SMEs). It fills a gap in the literature on international SMEs' performance by relating their profit and cost efficiency to a set of core determinants. We find that profit efficiency decreases as export intensity grows unless a firm achieves a medium scale. The evidence highlights another interesting trend: regardless of firm size, workforce experience shows a non-monotonic relationship with both a firm's profit and the cost-efficiency score. Further, a firm's debt sustainability affects its efficiency performance negatively in terms of profit but positively in terms of cost. Conversely, a higher financial burden leads to better profit efficiency for medium-sized enterprises, but this effect is reversed for smaller firms. These results have important policy implications, not only for policymakers but also for firms' management, pursuing a competitive advantage to navigate the stormy international market.
Acknowledgements
We thank the Editors and two anonymous referees for their critical comments and suggestions. We also thank participants at the 60th SIE annual conference. All remaining errors are ours. Both authors contributed equally to each section of the article.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Hereafter, we use ‘technical efficiency’ and ‘efficiency’ interchangeably.
2 Literature in this field adopting the technical efficiency method relies mainly on binary variables to record the effect of being an exporter on technical efficiency (Alvarez and Crespi Citation2003; Charoenrat and Harvie Citation2014; Pérez-Gómez, Arbelo-Pérez, and Arbelo Citation2018). For medium and large enterprises, meanwhile, see, e.g. Kutlu, Tran, and Tsionas (Citation2020).
3 One strand of the literature provides evidence in support of the learning-by-exporting (LBE) hypothesis (Lu and Beamish Citation2001; Castellani Citation2002; Girma, Greenaway, and Kneller Citation2004; Crespi, Criscuolo, and Haskel Citation2008; De Loecker Citation2013; Lööf et al. Citation2015; Tse, Yu, and Zhu Citation2017) identifying an enhanced role of export intensity on productivity. Conversely, other empirical studies provide evidence against the LBE hypothesis and in favor of the self-selection assumption, namely that only efficient companies can become exporters (Delgado, Farinas, and Ruano Citation2002; Wagner Citation2007; Movahedi, Shahbazi, and Gaussens Citation2017).
4 Indeed, Ilmakunnas and Miyakoshi (Citation2013) reported that experience offsets the negative effects of seniority for higher-skilled workers.
5 Interestingly, exporting SMEs may have better financial health (Greenaway, Guariglia, and Kneller Citation2007) and ceteris paribus suffer less from the financial constraints typically present in SMEs (Berman and Héricourt Citation2010). To a great extent, Wagner (Citation2014), in his broad literature review, pointed out that exporting firms are less financially constrained than non-exporting firms.
6 Corporate finance literature provides mixed evidence on the relevance of capital structure to efficiency performance. Some studies have shown that companies with greater leverage are closer to the efficiency frontier (Margaritis and Psillaki Citation2010; Hanousek, Kočenda, and Shamshur Citation2015; Fernandes, Vaz, and Monte Citation2018). This strand of the literature supports both the agency-cost and efficiency-risk hypotheses. According to other studies, the franchisee value hypothesis may supersede the above two hypotheses: efficient companies prefer to use internal financial resources and retain a few resources of debt to preserve the economic rent arising from such efficiency (Seelanatha Citation2010). Agostino, Ruberto, and Trivieri (Citation2018) showed that the impact of lasting lending improves technical efficiency for low levels of indebtedness and this effect reverts when the leverage reaches significant levels.
7 Psillaki and Daskalakis (Citation2009) documented that Italian SMEs are the most leveraged companies among those based in France, Greece, Italy, and Portugal.
8 Hanousek, Kočenda, and Shamshur (Citation2015) found that very large firms are less technically efficient than smaller firms. However, the authors adopted a different measure of firm size (i.e. the log of total assets), which is strictly related to the capital variable included in the production function. Their sample includes large players with more than 40,000 employees.
9 Schiersch (Citation2013) interpreted this as evidence that small firms operating efficiently over time tend to grow (noisy selection effect). However, the transition process to large scale induces inefficiency in the organizational structure of medium-sized enterprises (the inefficient hierarchical structures effect), which is compensated for by economies of scale for firms that become larger.
10 In a preliminary analysis, the specification also considered the non-linear term of Export, but neither of the associated coefficients (the square and the cube) were significant.
11 In general, dichotomization can induce bias in the estimates, information losses and inefficiency (see for instance, MacCallum et al. Citation2002). However, it may be difficult to define a cut-off point for classifying exporters as small or large. Likewise, small exporters would then grow in size in the future.
12 Expertise is a highly skewed variable mitigated by the log transformation. Therefore, the associated coefficient approximately represents the change in efficiency score when Expertise increases by 1%.
13 The specification including the quadratic term must be carefully interpreted. Consider that with the linear-log specification along with a quadratic term in a logarithm, changes in the dependent variable would be as follows:
14 In 2007, a reform of the Italian pension system allowed workers to allocate future accrued severance pay to a complementary pension fund (For more detail, see http://www.eurofound.europa.eu/observatories/emcc/erm/legislation/italy- //www.eurofound.europa.eu/observatories/emcc/erm/legislation/italy-severance- pay redundancy-compensation). The 2007 reform also required firms with more than 50 employees to accrue severance pay in a public fund owned by the national Social Security Agency. These new rules affect our inference because Expertise includes only the severance pay accrued in the firm's balance sheet, and thereby does not account for the amounts moved to pension funds. However, we believe that the bias is likely negligible for two main reasons. First, according to the Italian Pension Funds Supervisory Commission (Covip), even 10 years after the reform, Italian firms still held more than 55% of the value of the new TFR. Second, Expertise is a stock variable and accounts for the exact value of the severance pay accrued by a firm until 2006. In our dataset, the average firm age is 37.88 years and only 0.7% of workers have seniority of less than 10 years.
15 The analysis of Diaz and Sánchez (Citation2008) also includes large firms (more than 500 employees).
16 The interaction term is calculated by multiplying the demeaned continuous variables of interest (Expertise and Debt Sustainability) with the discrete variable Size. For the interaction of Expertise with Size, the associated coefficient was never significant.
17 With regard to Export, neither its squared term nor the interaction term were ever significant.
18 According to Aidaf (2018), there are approximately 784,000 family firms in Italy, which account for more than 85% of all businesses in the country. Data are available online at http://www.aidaf.it/aidaf/le-aziende-familiari-in-italia.
19 In a preliminary analysis, the specifications also included a firm independence indicator (i.e. the BvD independence indicator) reflecting the degree of independence from shareholders. However, the variable was never significant.
20 Ibid, footnote 19.
21 In fact, our dataset does not include the number of foreign markets to which each firm exports. As a consequence, with the same level of export intensity, technical inefficiency may diverge depending of the number of foreign markets on which a firm sells its products.
22 In both specifications 3 and 4 of Table , the coefficient of Export represents the effect of the intensity of foreign trade when both Size1 and Size2 are equal to 0 and Size3 is equal to 1. Therefore, in specification 3, the beta associated with Export (−0.957) signifies that as long as Export increases, the inefficiency of medium-sized enterprise decreases. In other words, medium sized exporters improve efficiency if the share of exports increases. Conversely Size1 ×Export and Size2 ×Export are positive, implying that for smaller firms a large export intensity reduces the firm's technical efficiency.
23 In this regard, the coefficient of Expertise and Expertise are both negative but the quadratic term is lower than the linear term.
24 In a further specification, the interaction effect between Size and Expertise is never significant, implying that even for medium-sized enterprises the transfer of knowledge does not hold.
25 In detail, both the linear interaction terms (Debt Sustainability × Size1 and Size2) are negative and statistically significant. Conversely, the squared interaction terms (Debt Sustainability2 ×Size1 and Debt Sustainability2 ×Size2) are significant and positive.