Annex A. Additional charts and tables
Scaling-up is a sustainable transformation for 40% to 70% of employment scalers, which remain at their new scale or continue to grow after the first high-growth episode. In addition, 14% to 32% of these employment scalers continue scaling up in turnover in the following 3-year period. The pattern of sustainability repeats for turnover scalers. On average across 4 countries, 59% of turnover scalers consolidate the new scale or continue growing. However, it is slightly less common for turnover scalers to turn into employment scalers than for employment scalers to continue scaling up in turnover. Between 10% and 21% of turnover scalers scale in employment in the 3 years after they scaled up (Figure A A.1). The lower probability of future employment growth for turnover scalers shows that it is more difficult to scale in employment and only a few firms undergo this transition. Employment growth can also be a predecessor of turnover growth. Firms that plan their business expansion might first prepare for the increase in demand by hiring the needed employment and, in the next phase, expand the output.
Note: Turnover scalers grow in employment by at least 10% per year over 3 consecutive years on average, as defined in Box 1.2. The sample includes scalers that end their first 3-year scaling period between 2011 and 2015 in Finland, 2004 to 2015 in Italy, 2013 to 2014 in Portugal and 2006 to 2015 in Spain. The sample is limited to the non-financial business economy. Owing to methodological differences, figures may differ from official statistics.
Source: Calculations based on microdata sources from four countries. See Annex B for more information.
The characteristics of the firm, including size and age, play a similar role in predicting scaling up in turnover scalers as in employment scalers. For example, the probability of being a scaler decreases with a firm’s age. However, one notable difference is that there is a higher proportion of old firms among turnover scalers than among employment scalers (Figure A A.2). One-quarter of turnover scalers are small- and medium-sized enterprises (SMEs) established 21 years ago, while their share among employment scalers is 20%. At the same time, the group of young firms is smaller among turnover scalers. One-quarter of turnover scalers are young firms, compared to 28% of employment scalers.
Note: Turnover scalers grow in employment by at least 10% per year over 3 consecutive years on average, as defined in Box 1.2. The sample includes scalers that end their first 3-year scaling period between 2011 and 2015 in Finland, 2004 to 2015 in Italy, 2013 to 2014 in Portugal and 2006 to 2015 in Spain. The sample is limited to the non-financial business economy.
Source: Calculations based on microdata sources from four countries. See Annex B for more information.
Employment scalers contribute to 47% to 69% of gross job creation and turnover scalers to 51 to 71% of gross turnover growth. Turnover scalers are also major contributors to gross job creation (38-65%). To a lesser extent, employment scalers are also contributors to turnover growth. Between 27% and 47% of gross turnover growth is generated by employment scalers (Figure A A.3). Although the contribution of employment scalers to turnover growth is smaller in size than the equivalent share of contribution to job creation by turnover scalers, it remains substantial when taking into consideration that employment scalers are 50% to 80% less numerous than turnover scalers. An average employment scaler, therefore, contributes a similar share to turnover growth to the average turnover scaler.
Note: Gross turnover creation is calculated as the total turnover added by all non-micro SMEs growing in turnover over the triennium. The contribution by each group of firms is reported as a percentage of the sum of gross turnover growth and gross turnover destruction in absolute value, which implies that, for each country, the positive and negative segments of the sum of the bars is 100 in absolute values. Scalers grow in employment or turnover by at least 10% per year over 3 consecutive years on average, as defined in Box 1.2. The sample includes firms with at least 10 and at most 249 employees. The sample is limited to the non-financial business economy. Owing to methodological differences, figures may differ from official statistics.
Source: Calculations based on microdata sources from five countries. See Annex B for more information.
Structural factors are firm characteristics that change only rarely over time, such as size group, industry, and location of a firm, or they change in an expected, incremental way, such as age. The probability of scaling up for firms within the same groups of structural factors is estimated in two different ways. The first simpler approach consists of calculating the share of scalers within each group, e.g. the share of scalers among firms with 10 to 49 employees. However, this metric may be prone to a composition bias. For example, firms in information technology are more likely to scale but they also tend to be smaller; the higher incidence of scalers among small firms can be only a sector-driven outcome.
The firm-level data allow estimating the relative probability of scaling up controlling simultaneously for structural factors and this way reducing the potential composition bias.1 The outcomes of the estimations generally confirm the results obtained from calculated shares within each group. For example, the propensity to be a scaler falls with the firm age. Firms in older age groups have a 7 to 16 percentage points lower probability to scale up than the youngest firms in Finland (Figure A A.4). This effect is an average difference across firms within the same size class and the same industry class, which means that the age impact is free of the potential influence of other structural factors.
Note: Scalers grow in turnover by at least 10% per year over 3 consecutive years on average, as defined in Box 1.2. The graph indicates the magnitude of the probability to be a scaler compared to the baseline probability of scaling up among young firms. Points represent the estimated values and the vertical lines 90% confidence intervals. If the confidence interval crosses the zero line, the results are statistically insignificant. The vertical axis reports percentage differences in the probability of being a scaler between the given age group and young firms. The negative value implies lower levels as compared to the baseline. The regressions control for a year, size bin, sector and region fixed effects. The sample is limited to firms with at least 10 employees in the first year and surviving for the whole period. The sample includes observations from 2008 and 2018 in Finland.
Source: Calculations based on microdata sources from Finland. See Annex B for more information.
In some countries such as Italy and Spain, the probability of being a scaler among firms in different size classes is between 10-13% and smaller firms have a slightly larger probability to scale up. However, when age and sector variables are included in the evaluation, the marginal probability for firms of being scalers might increase with firm size. For example, in Italy, firms with at least 50 and at most 99 employees have a 1 percentage point higher probability of being a scaler and medium-sized firms with at least 100 and at most 249 employees scale half of a percentage point more than the smallest firms with 10 to 19 employees (Figure A A.5). In the results using simple averages by size class, the composition of age and sector has increased the overall probability of becoming scalers for small firms. Removing these composition effects shows that small size is a poor predictor of scaling up.
Note: Scalers grow in employment by at least 10% per year over 3 consecutive years on average, as defined in Box 1.2. The graph indicates the magnitude of the probability of being a scaler compared to the baseline probability of scaling-up among firms with 10 to 19 employees. Points represent the estimated values and the vertical lines 90% confidence intervals. If the confidence interval crosses the zero line, the results are statistically insignificant. The vertical axis reports percentage differences in the probability of being a scaler between the firms in a given size group and firms with 10 to 19 employees. The positive value shows that the factor is higher than among the baseline, the negative sign implies lower levels as compared to the baseline. The regressions control for a year, age bin, sector and region fixed effects. The sample is limited to firms with at least 10 employees in the first year and surviving for the whole period. The sample includes observations from 2001 to 2018 in Italy.
Source: Calculations based on microdata sources from Italy. See Annex B for more information.
The analysis described in Chapter 4 considers a broad range of firm time-variant characteristics, comparing scalers with similar firms that share the same predetermined factors (size and age class, sector, location) but that do not scale. The tables listed in this section summarise these time-variant characteristics for each country where the analysis was possible and for both types of scalers. The results are organised into tables by topic as follows: innovation in Table A A.1. , human capital in Table A A.2, global markets in Table A A.3, financial indicators in Table A A.4, productivity and profitability in Table A A.5, and workforce characteristics in Table A A.6.
Note
← 1. The probability is estimated with a linear probability model (LPM) at the firm level. The dependent variable is a binary indicator equal to one if the firm is in the final year of a three-year scaling-up period and equal to zero otherwise. The dependent variables are size class, age class and 2-digit sector categorical variables. The model also includes year fixed effects.