This chapter takes our empirical definition of scale-up to the data: Swedish register data on over 700,000 firms for the period 1997–2021. 1.28% of firms meet the HGF criteria. Among these HGFs, it is rare for a firm to satisfy all 7 scale-up conditions (in line with the idea of ‘too many exceptions’). 25.89% of HGFs satisfy 5 or more of the 7 conditions for scale-ups, while 60.75% of HGFs satisfy 4 out of 7 conditions. Our analysis highlighted how missing values can cause problems when investigating which HGFs are scale-ups.

7.1 Data Description

We use a Swedish register database that includes all limited liability firms active at some point for the years from 1997 to 2021 (n = 8,294,726 firm-year observations).Footnote 1 This reduces our dataset to 6,888,528 firm-year observations, covering 739,094 firms. The database contains audited annual report information, such as yearly revenues, as well as costs for machinery, buildings, employees and salaries, and intellectual property, among other variables. It also includes information on firm age, financial strength, location, industry classification, and much more.

This database has been widely used to study, for example, the effects of employment protection (Bornhäll et al., 2017), barriers to firm growth (Bornhäll et al., 2016), firm growth paths (Coad et al., 2018, 2022) and high-growth firms (Daunfeldt et al., 2014; Daunfeldt and Halvarsson, 2015).

HGFs are defined as firms with an average annualized employment growth greater than 20% over three years and with ten or more employees at the beginning of the observation period. We divide our data into seven three-year periods: 2000–2002, 2003–2005, 2006–2008, 2009–2011, 2012–2014, 2015–2017, and 2018–2020. During these seven three-year periods, we identify 9,448 firms (1.28 %) that meet the HGF criteria, of which 12% (1,157 firms) meet the criteria during more than one period.

7.2 Results

7.2.1 Frequency of HGFs

Table 7.1 shows that HGF events are rare and that only 1.28% of all Swedish firms ever experience a high-growth period. Also, previous work on Swedish data has shown that HGFs are ‘one-hit wonders’ (Daunfeldt and Halvarsson, 2015) and unlikely to repeat their HGF episode. These findings are supported by our results, showing that only 0.16% (1,157) of the firms manage to meet the HGF criteria for more than one period.Footnote 2 If HGFs are rare and lack persistence, we can expect this from scale-ups too. Scaling is likely to be a short-lived episode that is more akin to a ‘stage of growth’ (Blank, 2013) than a durable firm-specific trait.

Table 7.1 Frequency of high-growth firms
Full size table

7.2.2 How Many HGFs Satisfy the Conditions for Being Scale-ups?

Table 7.2 shows many interesting findings. Regarding Condition 1 (non-negative change in marketing): we had severe problems of missing values for this case. We could only calculate the growth in “cost of sales” for 598 of the high-growth events. 543 (90.80%) of those had a non-negative growth in cost of sales during their high-growth period. This also highlights the difficulties in getting data to test this condition properly.

Table 7.2 Decomposing the sample of HGFs to see how many HGFs satisfy the conditions for scale-ups
Full size table

Regarding the age distribution of HGFs, some key percentiles are as follows: 25th percentile: 6 years; 50th percentile: 11 years; 75th percentile: 18 years. Therefore, it might be appropriate to take age 10 as a cut-off point (following previous work e.g. Coad et al., 2016; Vandresse et al., 2023). Firms older than 10 years would probably not correspond to a common understanding of what is a scale-up. However, we must stress that our methodology allows for exceptions too. Restricting in this way means we still have about half of high growth events (5,296 out of 10,732) remaining in the category of potential scale-ups.

Condition 3 focuses on a non-negative change in intangible assets. Intangible assets are our best-available proxy for software, which is crucial for scale-ups (De Ridder, 2023). Data for intangible assets show that 7,987 (85.25%) firms had a non-negative growth of intangible assets during the HGF period compared to 1,382 (12.88%) firms that had a decline in intangible assets.

Condition 4 relates to non-negative growth of inventory. 8,292 (88.52%) out of 9,367 high-growth events were associated with a non-negative growth of inventory. 1,075 (11.48%) of the high-growth periods had a decline in inventory.

Condition 5 focuses on the industry classification of the HGFs, in particular whether they are in the digital/ICT sector. This is operationalized using the NACE codes in the classification scheme in OECD (2011). 492 (5.67%) out of 8,680 HGFs are active in the ICT sector. 8,188 (94.33%) are active in some other sector. This shows that scale-ups appear in a variety of industry sectors as opposed to solely in Information Technology. In that way, they are similar to HGFs that appear across all types of industries (Daunfeldt et al., 2015).

What is notable about this finding, however, is that much of the scaling is in great part thanks to general purpose technology that these scaling firms leverage to increase their efficiencies over time. Further examining the details of industry-based differences and how scaling is possible may offer greater conceptual insights. Studying the limited number of firms that can scale, despite what seems to be a challenging environment to scale, may also open up new avenues for understanding what entrepreneurs can accomplish within what may appear to be externally objective challenges, but where subjective perceptions of opportunity may vary greatly (McKelvie et al., 2018). These deeper examinations may also promulgate genius ways in which entrepreneurs leverage technology or investments to scale.

Condition 6 stipulates that growth of sales is faster than growth of employees. In our data, 4,193 (44.87%) of HGFs experienced higher growth in sales than in employment during their high-growth phase. 5,151 (55.13%) experienced the opposite. No HGFs had equal growth in sales and employment.

Condition 7 requires that HGFs have gross margins that are high, above 40% (or alternatively above 30%). We consider two indicators of financial performance: gross margins (which suffers from problems of missing values) and operating margins (where there are far fewer missing values). Our preferred indicator (that will be used in Table 7.3) is the number of HGFs with an operating margin above 30%.

Table 7.3 Number and percentage of the sample of HGFs with non-missing values for the 7 conditions
Full size table

Table 7.3 shows the number and percentage of the sample of HGFs with non-missing values for the 7 conditions. There are 502 HGFs with non-missing values for all of the 7 conditions. From the set of 502 HGFs, there is only one that satisfies zero conditions. Clearly, this would not be a scale-up. In contrast, only 2 HGFs satisfy all 7 conditions. The most common case is for HGFs to satisfy 4 out of 7 conditions (34.86% of HGFs). There are 25.89% of HGFs (= 21.71 + 3.78 + 0.40) that satisfy 5 or more of the 7 conditions for scale-ups, while 60.75% of HGFs satisfy at least 4 out of 7 conditions.