In this paper we investigate the development of business ownership (self-employment) rates over time and the effect of business ownership on economic growth, both at the sectoral level. In an earlier exercise, Carree et al. (2002) have presented a two-equation model to analyze the interrelationship between economy-wide business ownership rates and economic development. They apply the model to a data set of 23 OECD-countries for the period 1976-96. The paper showed empirical evidence for a (slow) error-correction process for business ownership rates: countries with business ownership rates more or less than the 'equilibrium' value for the specific stage of economic development of these countries showed, on average, convergence towards the 'equilibrium'.
In addition, it was found that the 'equilibrium' relationship between business ownership rate and stage of economic development (as proxied by GDP per capita) was declining for the larger part of the range of GDP per capita but had the tendency to rise for the highest levels of GDP per capita. The study also provided evidence that countries that had an out-of-equilibrium value of business ownership rate suffered in terms of economic growth foregone.
The analysis performed by Carree et al. (2002) raises an important research question: to what extent do differences in business ownership at the economy-wide level reflect differences in the sectoral structures of economies or differences in business ownership rates at the sectoral (or even industry) level? It is well known that the average business ownership rate in the service sector is much higher than that in the manufacturing sector.
Data in the current paper show that the average rate (including unpaid family workers) for OECD-countries was almost 20% in 1984 for the service sector, while it was less than 7% for the manufacturing sector. This has important consequences for the analysis previously performed. The tendency of business ownership rates to increase may be due just to a shift of economic activity from the manufacturing sector towards the service sector in the course of economic development. As a consequence, the penalty found for deviating from the 'equilibrium' value of (economy-wide) business ownership may really be a penalty for deviating from a certain structural composition of the economy. This study investigates the 'equilibrium' relationship between business ownership rates and economic development, the speed of the error-correction process and the existence (and severity) of the growth penalty when deviating from 'equilibrium' for both the manufacturing sector and the service sector for the OECD- countries in the period 1970-98.
The paper develops an adjusted two-equation model relating business ownership rates and economic growth rates at the sectoral level. Specific attention in the model is paid to whether national statistical agencies have reported to include owner/managers of incorporated business into the data. Four different types of 'equilibrium' relations between business ownership rate and GDP per capita are investigated, two of which have a U-shape (first declining and then rising ownership rates) and two of which have an L-shape (ownership rates continuously declining towards an asymptotic minimum rate).

We have collected data for 21 OECD- countries for the years 1970-1998, as far as the data were available according to uniform definitions. The data show that, on average for OECD- countries, business ownership rates in manufacturing and services are situated in orders of magnitude of 7% and 20%, respectively. This confirms that at least part of the increase in the economy-wide share of business ownership is due to the sectoral shift towards the service sector in developed economies. Results show that the empirical 'fit' of the four different types of 'equilibrium' relationships is not too different, both for the case of manufacturing and that of services. However, results for the speed of error-correction are hardly affected by which type is chosen. The estimated speed of error- correction for the manufacturing sector is  about 8.5% for a four-year period. This estimated speed is twice as high in the service sector: about 16%.
Both in the manufacturing sector and for services there is a positive effect of (lagged) unemployment: countries with high unemployment show higher subsequent business ownership rates both in manufacturing and services. The results imply that there is a significant penalty of the business ownership rate deviating from 'equilibrium' for manufacturing for each of the four types of the 'equilibrium' relationship. For the service sector also a negative effect on growth is found, but it is not always significant and it is far smaller than that for manufacturing.
The analysis confirms the empirical evidence provided by Carree et al. (2002) that dif-ferences in business ownership rates matter and disappear over time slowly. The general idea behind the model is that there can be both too many and too few businesses.
Too many businesses may mean that economies of scale and scope are not benefitted enough from and that there are probably many 'marginal' ventures. Too few business may imply that there is not enough entrepreneurial activity. The estimated equilibrium curves suggest that having too few businesses is the more likely problem in manufacturing, while having too many businesses is the more likely problem in the service sector. It would suggest that, not only, the economic benefits of government promoting new (and small) business may be nation-specific, but also sector-specific. Such promotion seems most beneficial in the manufacturing sector of countries with very low business ownership rates (like the Scandinavian countries). It may be counter-productive to have similar promotion in the service sector of countries with very high business ownership rates (like Italy).

Download a PDF-version of this research report.