Business dynamism and productivity growth in Belgium [ Working Paper 05-21 - 28/05/2021]

  PDF & Download

• Publication (en - 1.63 Mb)

  Authors

Michel Dumont (A)
 
A : Author, C : Contributor

  Publication type

Working Papers

The Working Paper presents a study or analysis conducted by the Federal Planning Bureau on its own initiative.

Besides firm-level productivity growth, the entry and exit of firms and the reallocation of market shares between incumbents can also contribute to the productivity growth of industries. Start-ups may challenge mature firms and rejuvenate industries by introducing new products; services; technologies and organizational forms. Incumbents can react to the competitive pressure from start-ups in several ways but the least productive may be forced to exit.  Reallocation can be productivity-enhancing if resources shift from firms with low productivity towards firms with high productivity. 

Declining business dynamism has recently been put forward as a potential explanation for the long-term slowdown in productivity growth, as witnessed in most OECD countries. This paper considers evidence on business dynamism and the potential link with productivity growth in Belgium.

There are different concepts and data sources for business creation. The broadest concept is VAT registrations, which is obligatory for most market activities in Belgium. A narrower concept is the creation of corporations. An important distinction is between enterprises with employees and enterprises without employees. The latter include self-employed entrepreneurs that do not employ any workers but also corporations that are created for tax purposes or real estate portfolio management. Although self-employed entrepreneurs play an important role in some industries, enterprises without employees are probably more necessity-driven and opportunity-driven than growth-driven and therefore less relevant from the perspective of industry growth than enterprises with employees that comprise enterprises with the ambition to innovate and grow. Data on enterprises with employees, only available for a relatively short period, do not show a decline of entry for Belgium but rather a stabilization at a low level, compared with other EU countries. Data on the creation of corporations, available for a longer period, indicate that the start-up rate in Belgium increased substantially in the 1980s, dropped dramatically in the first half of the 1990s and stagnated at a low level ever since. The exit of enterprises with employees is even lower in Belgium and appears to decline further. Compared with other EU countries, Belgium also has a low share of high-growth firms and, more generally, a low share of growth-driven entrepreneurs.

Periods of low economic growth are often characterized by an increased exit of low-productive firms (‘cleansing’), but also by negative and potentially long-lasting negative effects (‘scarring’) such as a decrease in business creation. This appears to have been the case during the so-called Great Recession that followed the 2007-2008 global financial crisis. Recent data show that the Covid-19 crisis, despite its dramatic economic impact, so far has not resulted in a decrease in business creation or an increase in bankruptcies. The number of corporations created in April and May 2020, at the start of the Covid-19 crisis in Belgium, was well below the level in the same months in 2019. However, from June 2020 onwards, business creation surpassed the level of previous years, to the extent that in the period March-December 2020, more corporations have been created than in the same period for the years 2015-2019. The number of bankruptcies, on a structural downward trend in Belgium, was extremely low from March 2020 onwards, explained by two moratoriums on bankruptcies but also by the fact that the Federal Public Service Finance and the National Social Security Office did not push for the bankruptcy of firms with overdue payment of taxes or social contributions. Moreover, companies and self-employed entrepreneurs received substantial support at the federal, regional, and local level, such as temporary unemployment benefits for affected employees, payment plans for taxes and social security contributions and replacement income for the self-employed (bridging right). So, although the predicted surge in bankruptcies has not materialized yet, it remains to be seen what will happen when support measures are scaled back or wound down.        

Job reallocation – the sum of job creation and job destruction relative to total employment – decreased in most OECD countries, slightly more in market services than in manufacturing industries. Job reallocation also decreased in Belgium but slightly less than in other countries.

A positive correlation between business dynamism and productivity growth does not necessarily prove a causal relationship. Declining business dynamism and productivity growth slowdown may both result from the maturing of industries, a decrease in growth opportunities or demographic changes. Tests of causality of the link between entry, exit and productivity growth in Belgium, reported in this paper, show that the results crucially depend on whether value added, or gross output is considered as a measure of output and provide mixed evidence on the potential causal link between business dynamism and productivity growth.

This paper reports the results of a decomposition of productivity growth, using data on Belgian firms over the period 2003-2017. This decomposition does not aim at detecting causal links but at assessing the relative contribution of four components of industry-level productivity growth: the entry of firms; the exit of firms; the reallocation of market shares between incumbents and firm-level productivity growth. For the decomposition, four alternative estimates of productivity are considered, resulting from two different estimation procedures, and considering two alternative output measures (value added and gross output). 

In the decomposition, incumbents are broken down into three separate age groups: start-ups (1-4 years after entry); young firms (5-9 years old) and mature firms (10 years or more). This breakdown appears to be relevant, as there are substantial differences across age groups, in the extent, and even the sign of, for example, the contribution of reallocation and the contribution of firm-level productivity growth to industry-level productivity growth. A decomposition that considers all incumbents, irrespective of age, would not reveal these differences across age groups, that provide useful information on age-specific patterns.  

The results of the decomposition depend on which productivity estimate is considered, probably explained by well-known biases in the estimation of productivity that cannot be satisfactorily tackled with the data at hand. Especially results with respect to the contribution of the entry of firms appear rather sensitive, with conflicting conclusions even as to the sign of the contribution.

The relative contribution of firm exit is negative, somewhat surprisingly as exiting firms on average tend to have productivity below the industry average. The negative contribution follows from a small group of exiting firms that have above-average productivity and a larger share in industry output than the substantially larger group of exiting firms with below-average productivity but also a smaller share in industry output. Future research could assess whether the high-productive firms that exited were involved in mergers and acquisitions, which would imply that they continued their activities and that the exit is administratively rather than a real exit. 

Somewhat more robust conclusions can be found for reallocation. Start-ups seem to go through a phase of market experimentation and the establishment of a customer base where selection does not occur based on productivity, as many low-productive start-ups witness an increase in their market share in the early years after entry. For mature firms, reallocation is more productivity-enhancing. Mature firms with productivity above (below) the industry average, tend to gain (loose) market shares. This process of market share growth, irrespective of the initial productivity level, and substantial productivity growth of start-ups, appears to be more pronounced in market services than in manufacturing industries.  

The most robust conclusion of the decomposition of productivity growth is the strongly positive contribution of firm-level productivity growth. When the component is broken down by age, productivity growth of start-ups by far contributes most to industry-level productivity growth. This finding confirms the results of previous studies of the crucial role of start-ups in industry-level productivity growth. A more worrying conclusion is the decline in the relative productivity of entrants (in the year of entry) as well as the decreased positive contribution of productivity growth of start-ups (defined as firms 1 up to 4 years after entry).

With the usual caveat of data and methodological limitations, the decomposition analysis in this paper suggests that the low level of business dynamism in Belgium, or its trend over time, does not offer a silver bullet explanation for the structural decline in industry-level productivity growth. Firms appear to witness intrinsic problems to raise, or even sustain, productivity growth. However, the low exit rate in Belgium may indicate the existence of barriers to the exit of low-productive firms that could limit opportunities for start-ups, which are found to be crucial for industry-level productivity growth.  Regulatory barriers and administrative burdens, weak contract enforcement, financing constraints or inefficient bankruptcy procedures are often put forward as factors that raise barriers to entry and exit. The low entry, especially of growth-driven entrepreneurs, justifies policies that support this kind of entrepreneurship during the start-up and scale-up stages.