Archive:Statistics on enterprise survival and growth prospects between 2008 and 2012

This article has been archive. Up to date data can be founs in the following article: Entrepreneurship - statistical indicators.

Authors: Ilke Van Beveren, Martin Luppes, Linda Rozendaal and Michel Walthouwer (Statistics Netherlands), Pekka Alajääskö and Anton Roodhuijzen (Eurostat, Structural business statistics and global value chains). The article will not be updated.

This article uses data from the 2015 European microdata linking project. This partnership between Eurostat and eight European countries created linked national microdata sets, which enabled the analysis of trends for micro, small and medium — and large — enterprises. The novelty of these microdata sets is that they follow a group of enterprises that existed in 2008 over the time period 2008-2012. It therefore offers a dynamic view on how this period, which included the economic crisis, affected the survival and growth prospects of various types of enterprise. This 'longitudinal analysis' allows us to shed more light on, for example, size-class mobility of small and medium-sized enterprises (SMEs) compared with large ones in various sectors of the business economy. Moreover, by combining this information with international trade in goods statistics, we are able to distinguish between exporting and non-exporting enterprises.  

This article is part of an online publication on Microdata linking in business statistics.

Table 1: Number of enterprises, persons employed and gross value added (GVA) and the share of SMEs, 2012
Source: Eurostat
Table 2: Number of enterprises by 2008-size-class and observed enterprises in 2012
Figure 1: Survival and death rates from 2008 to 2012
Table 3: Size class mobility, transition table for observed enterprises, 2008-2012
Table 4: Export status by size class, 2008-2012
Figure 2: Distribution of enterprises by export status and size class, 2008-2012
Figure 3: Employment dynamics by export status, 2008-2012
Table 5: Distribution of enterprises observed in industry, 2008-2012.png
Table 6: Distribution of enterprises observed in services, 2008-2012
Figure 4: Share of micro enterprises moving to a higher size class in industry and services, 2008-2012.png

Main statistical findings

  • Survival rates are inversely related to enterprise size. On average — and across all countries — SMEs are characterised by a higher number of deaths, both in absolute and relative terms. In the SME category, micro and small enterprises have the lowest chances of survival.
  • Enterprises that survived the financial and economic crisis show limited mobility between size classes from 2008 to 2012: most enterprises remained in the same size class. Changes in size class are more likely to be towards a lower size class. This reflects the size distribution, which is dominated by micro and small enterprises.
  • The majority of enterprises are micro enterprises that do not export goods. Only a small fraction of enterprises that existed in 2008 engaged in exports between 2008 and 2012.
  • Continuing exporters (enterprises that exported in 2008 and 2012) tend to be the largest in size, and they are heavily represented among medium-sized and large enterprises.
  • Non-exporters are generally less likely to change size class than enterprises engaging in exports.
  • Enterprises that stopped exporting between 2008 and 2012 were more likely to drop down in size class; enterprises that started exporting had a higher chance of upward mobility between size classes. For continuing exporters, results are ambiguous: the rates of enterprises increasing and decreasing in size are almost identical.
  • Service enterprises tend to be smaller and have a lower degree of mobility between size classes than enterprises active in industry.

Set up of the project

The size classes used in this article are: micro (0-9), small (10-49), medium (50-249) and large enterprises (250+ persons employed). The number of enterprises, their gross value added (GVA) and employment can be found in Table 1. When reading the article, one should bear in mind that these size classes have different ranges. The 0-9 size class, for example, only has a range of 10 while the 10-49 size class has a range of 40. Moreover there tend to be more enterprises in the lower end of each size class.

The main objective of the microdata linking project is to link— in a harmonised way — data from various microlevel sources for selected European countries. By combining data sources, new insights can be obtained at the country level. As there is no need to collect new data, this avoids increasing the reporting burden on enterprises. Moreover, by harmonising this process across countries, it becomes possible to make cross-country comparisons. It is also important to ensure that the linked microdata sets are extrapolated to the total population of enterprises in order to be able to generalise the findings at the total population level.

This article presents results for eight countries: Austria, Denmark, Finland, Latvia, the Netherlands, Norway, Portugal and Sweden. We combine data on enterprise size (employment) with data on international trade in goods and enterprise growth and survival. The data refer to the period 2008-2012 and to the set of existing enterprises in 2008. We limit our coverage to the non-financial business economy.

Tracking enterprises over time

The data used in this article are unique in the sense that they allow us to follow a particular group of enterprises over time. This differs from earlier research where only consecutive cross-sectional data were available. In each of the eight countries considered, enterprises that were active in 2008 according to the business register were selected. These enterprises were grouped into four size classes on the basis of their employment levels in 2008, and their growth and survival were monitored in the following four years.

Table 2 summarises the data used in the selected countries. Data are not identical but are comparable with the total number of enterprises listed in Table 1. Table 2 shows the size distribution of the original enterprise group as well as the final number of observed enterprises in 2012. There are several reasons why enterprises that were part of the 2008 group are no longer observed in 2012. Firstly, enterprises can exit the market (enterprise death). Secondly, sampling procedures can lead to them not being observed even though they are still active. Thirdly, other reasons can explain why enterprises drop from the sample, for example after undergoing a merger or acquisition or because the enterprise failed to meet its reporting requirements. Unfortunately, we are not able to distinguish between all these various reasons, but we do have information on enterprise survival and deaths.

Figure 1 summarises — for each of the four size classes defined in the introduction — how many enterprises survived the economic crisis. The category ‘unknown’ covers all reasons other than enterprise death that can cause an enterprise to drop from the data. This category varies across countries and is more prevalent among the lower size classes. For most countries, the fall in observations is limited to less than 25 %. Large enterprises are more likely to be observed in both 2008 and 2012 than smaller enterprises. Figure 1 shows a lot of variety in terms of enterprise survival between 2008 and 2012. Portugal and Denmark had disproportionate death rates, particularly among micro and small enterprises. Given the wide differences across countries in the category ‘unknown’, the following sections focus attention solely on those enterprises that remained in the sample between 2008 and 2012.

Size class mobility and export status

This paragraph describes the growth prospects for the 2008 sample of surviving enterprises in each of the countries considered. It looks at these growth prospects from two angles: mobility between size classes; and export activity. Firstly, both aspects are considered in isolation. Table 3 shows enterprise mobility between size classes for those enterprises that were part of the sample in 2008 and 2012. The last four columns of the table, in combination with the row dimension, yield a 'transition table'. It summarises to what extent enterprises have changed size class in the period considered. The bold cells of the table show enterprises that have not changed size class, to the right are upwardly mobile enterprises and to the left are the downwardly mobile.

The majority of observations can be found in the bold cells, suggesting limited changes across employment size classes from 2008 to 2012. In addition, these observations lead to the conclusion that mobility across size class is more prevalent for SMEs than for large enterprises, and SMEs are more likely to end up in a lower size class rather than in a higher one. These findings are consistent with a size distribution dominated by small and micro enterprises (see tables 1 and 2 and related comments). Even though we observe that enterprise movements between employment size classes did not change much between 2008 and 2012, it is not clear what the effect of the economic crisis is on this because we can only observe movements and changes from 2008 to 2012 but not what happens in 2009, 2010 and 2012 individually. This issue can perhaps be addressed in future research using a refined set-up for the same or an extended time period, including an analysis on entrance and exits of enterprises and their movements between size classes during the economic crisis and recovery years.

Table 4 focuses on enterprise export status and combines this information with the size class in 2008. We consider four types of enterprise based on their export status in 2008 and 2012. Continuing exporters are enterprises that exported in 2008 and 2012; non-exporters did not export in 2008 or 2012. First-time exporters did not export in 2008 but are exporters in 2012. Exiting exporters have managed to survive the crisis, but have withdrawn from foreign markets. For methodological reasons, data for the Netherlands are not included.

Figure 2 summarises the distribution of enterprises — for all enterprises that are still part of the sample in 2012 — across size classes for the various types of trading enterprise defined above. Compared with continuing exporters, first-time exporters and exiting exporters between 2008 and 2012, non-exporters in 2008 and 2012 were more likely to be micro enterprises in 2008. At the other end of the size distribution, the majority of large and medium-sized enterprises can be found among the continuing exporters between 2008 and 2012. The difference between first-time and exiting exporters is highest in Finland (8.9 %) and positive in Denmark and Latvia. In the other countries, it is negative, most notably in Portugal (-9.5 %).

Exporters are of special interest for policy makers because of their potential job creation due to demand from markets abroad. This was an important message in the Statistics Explained article comparing enterprises which trade internationally with those who do not. Thus the large positive difference between first-time and exiting exporters is good for employment in Finland, while the large negative difference adversely affects employment in Portugal.

Finally, Figure 3 considers the changes in size class between 2008 and 2012 by export status. Several interesting findings emerge from this. Non-exporters exhibit the lowest (relative) amount of changes in size class between 2008 and 2012. First-time or exiting exporters and continuing exporters are consistently more mobile between size classes than non-exporters. For exiting exporters between 2008 and 2012, downward mobility is clearly more likely than upward movement. The opposite applies to first-time exporters between 2008 and 2012: they are more likely to experience upward mobility. For continuing exporters, results are somewhat ambiguous, perhaps due to the economic crisis. Future research can potentially shed light on this. These findings hold for all countries in the figure.

Enterprise growth by sector

Tables 5 and 6 provide some insights into differences across sectors. Table 5 summarises the size distribution in 2008 and size-class mobility between 2008 and 2012 for industry (mining, manufacturing, utilities and construction, sections B-F of NACE Rev2), while Table 6 looks at the service sectors separately (NACE G-N excl. K + S95). The number of observations in each of the tables adds up to the total of Table 4. We focus only on enterprises that were observed at the beginning and end of the period considered.

In all countries, the number of enterprises active in services is far greater than the number of enterprises active in industry. Depending on the country considered, there are three to five service enterprises for each enterprise in industry. However, service enterprises tend on average to be smaller in size. Service enterprises are also characterised by a lower mobility across size classes than enterprises in industry. In both sectors, downward mobility is more likely than upward mobility, consistent with findings reported in Table 3.

In Figure 4, we compare the upward mobility of micro-sized enterprises (0-9 persons employed) in industry and services for the eight countries covered by the project. With the exception of the Netherlands, upward mobility is in all countries higher in industry than in services. Austrian micro-sized enterprises are most likely to move to a higher size class both in industry and services, and Danish enterprises are second highest in both sectors. The biggest differences between industry and services are found in Austria, Latvia and Portugal. In the latter two, a micro-sized enterprise in industry is more than twice as likely as an enterprise in services to move to a higher size class.

Data sources and availability

New statistics on enterprises have traditionally been produced by carrying out surveys. Microdata linking presents an innovative approach to obtaining new information on the economic performance of enterprises by linking different existing statistical sources at individual enterprise level (microdata level). This approach does not require new surveys to be carried out and thus does not increase the burden placed on enterprises. Due to statistical confidentiality issues it is not possible to publish microdata on the Eurostat website. However in the future Eurostat aims at publishing aggregated tables based on microdata analysis in its database.  


Economic globalisation and the participation of enterprises in international trade in goods are important drivers for growth: as a consequence an increasing number of enterprises are active internationally. This article uses results from the latest European microdata linking project, in which the Structural Business Statistics enterprise population was linked with statistics on international trade in goods. Through this microdata linking it was possible to follow enterprises existing in 2008 for a number of years to 2012. Thus it was possible to see mobility of these enterprises by comparing the initial size class in 2008 to the size class in 2012. This was done for the population as a whole and for the subset of enterprises that export.

See also

Further Eurostat information


Dedicated section

Source data for tables and figures (MS Excel)