3 Measuring entrepreneurship.

Given the disagreement among scholars on the definition of entrepreneurship, is it then possible to measure the level of entrepreneurial activity? Clearly, the answer to this question is not straightforward and we need to address several issues before we may provide a satisfactory response.

As argued above, entrepreneurship is probably best considered a multifaceted concept which implies that it is not possible to construct an aggregate measure reflecting the "total" amount of entrepreneurial activity in a given area. For instance, even if we could count the number of innovations made by the entrepreneurs and measure the amount of uncertainty-bearing by entrepreneurs, no aggregate measure of entrepreneurship can be combined from these two indicators. We would be comparing apples and oranges.

However, the use of specialized measures may enable us to identify some dimensions of entrepreneurship and then compare the level of entrepreneurial activity across areas in one dimension at a time. In fact, this describes the current state of research in which a multitude of measures of entrepreneurship are used.

Using a specialized measure complicates the empirical analysis for (at least) two reasons. First, since many dimensions of entrepreneurship described in the theoretical literature are not quantifiable, or are imperfectly measured, researchers need to be precise about what dimension of entrepreneurship that is actually analyzed in a given context. Second, any empirical study trying to measure the "total" amount of entrepreneurship needs to consider a wide range of indicators in order to measure as many dimensions of entrepreneurship as possible.

The aim of this chapter is to review the measures of entrepreneurship available in the literature and relate these indicators to the economic theories of entrepreneurship in order to discuss what is actually being measured. Also, we present some new indicators of entrepreneurship which may better capture Schumpeterian and Kirzernian entrepreneurship using the European Community Household Panel (ECHP). The empirical exercise in Chapter 4 shows that measurement matters a great deal when different measures of entrepreneurship are compared across a group of countries.

3.1 The Self-Employment Rate

The starting point is the overall rate of self-employment, defined as the number of self-employed relative to the labor force. The self-employment rate has been used to compare entrepreneurship across countries by, e.g., Acs et al. (1994), Blanchflower (2000), Blanchflower (2004), Le (1999), OECD (1998), OECD (2000), Parker (2004) and Parker and Robson (2004). (1) Note that there is some disagreement in the literature on what to use in the denominator when computing the self-employment rate. For instance, Blanchflower (2004), Le (1999), OECD (1998), and OECD (2000) use total employment, while Parker (2004) and Parker and Robson (2004) use the labor force.

As seen in the first column of Table 3.1, the Mediterranean countries, Mexico and Korea have the highest self-employment rates among the OECD countries, whereas the countries of Central Europe and the Nordic countries are all found below or at the median of the sample. No distinct picture seems to emerge with respect to the Eastern European and the Anglo-Saxon countries. One notable feature of Table 3.1 is, however, that the US economy does not appear to be especially entrepreneurial in terms of the self-employment rate in contrast to its reputation of being the world's leader in entrepreneurship.

The extensive use of the self-employment rate is mainly due to the availability of this measure in most countries. But what is actually included in the definition of the self-employed? The OECD Labor Force Statistics define individuals as self-employed if they work for profit or family gain given in cash or in kind, whereas employees work for a wage or a salary, cf. OECD (2005a).

According to OECD (2000), the OECD Labor Force Statistics collect their information from household interview surveys where the interviewee's main occupation is determined by selecting one of the following categories: (1) employee; (2) employer; (3) own-account worker; (4) member of producers' co-operatives; (5) contributing (or unpaid) family worker; and (6) worker not classifiable by status. (2)

The OECD Labor Force definition indicates that employers, own-account workers, and unpaid family workers should be characterized as self-employed since they work for profit or family gain. However, OECD (1998) and OECD (2000) exclude unpaid family workers from the self-employment rate, arguing that these individuals are assistants of entrepreneurs and not entrepreneurs themselves. Blanchflower (2000) argues, on the contrary, that it may not be appropriate to leave out unpaid family workers just because they are remunerated in some other way, e.g., through a future ownership share in the family business. The issue of unpaid family workers seems to be important as an average of 11.6% of the self-employed in 1996 are unpaid family workers and Blanchflower (2000) reports considerable differences in this number across countries. (3)

Another potential limitation of the OECD self-employment rate is that it may not include all owners of incorporated businesses. These individuals are, in legal terms, employees of the corporation and may therefore identify themselves as employees in the household survey interviews. To address this problem, OECD (2000) defines owner-managers as individuals with a job in an incorporated business who: (a) have controlling ownership of the enterprise; or (b) have the authority to act on its regard with respect to contract negotiation with other firms and employees. The problem is that in some countries, owner-managers are classified as self-employed and in other countries as employees. For instance, the labor force surveys in Australia, Japan, Norway and the United States use a narrow definition of the self-employed which excludes the owner-managers, whereas most European countries include these individuals as self-employed, cf. OECD (2000). Unfortunately, OECD (2000) does not report separate numbers for the owner-managers, employers and own-account workers for each country. This is potentially important as owner-managers constitute, e.g., 31.4% of the self-employed in the United States (OECD, 2000).

Furthermore, because the OECD Labor Force Statistics categorize individuals according to their main job, the wage workers with self-employment as their secondary occupation are not included in the numerator when estimating the self-employment rates. According to Meyer (1990), the employed with a side business amount to 24.6% of the individuals whose main occupation is self-employment in the United States.

The before-mentioned self-employment rate from Table 3.1 includes all sectors. However, most empirical cross-country studies on entrepreneurship exclude the agricultural sector since the farming business is influenced by heavy subsidies, historic traditions of family ownership, and a relative high proportion of unpaid family workers compared to the rest of the economy (Blanchflower, 2000; Parker and Robson, 2004). However, while the average difference between the aggregate self-employment rate and the non-agricultural self-employment rate is small in Table 3.1 (3 percentage points), the discrepancy between these two measures tend to be large when the self-employment rate is large. For instance, the Turkish self-employment rate is approximately cut by a third when the agricultural sector is removed from the analysis. (4)

The self-employment rate is a widely used measure of entrepreneurship, but what dimension of entrepreneurship is it actually measuring? The crucial conceptual difference between self-employment and ordinary employment lies in the type of remuneration received, where the self-employed remuneration is uncertain. Hence, the OECD definition of the self-employed is closely related to the Knightian entrepreneur who, in his role as residual claimant, assumes all the uncertainty connected with the firm, thereby leaving all other stakeholders insured. In principle, all individuals who own (shares in) a business are entrepreneurs in the Knightian sense. For practical purposes, we must, however, require that an owner faces substantial uncertainty in order to be characterized as a Knightian entrepreneur.

In situations with no or little uncertainty, the self-employed is closer to Say's entrepreneur, the manager, than the Knightian uncertainty-bearer. However, a business-owner does not need to be a manager since mangers can be hired without reducing the true uncertainty.

Clearly, the self-employed are only a subset of the Knightian entrepreneurs in the economy and a measure that includes business owners may better proxy the number of Knightian entrepreneurs in a given area.

Taken at face value, the definition of the self-employment rate allows no inference to be drawn about Schumpeter's innovative entrepreneurship. Nothing guarantees that individuals working for profit or family gain are engaged in any of the five tasks characterizing Schumpeterian entrepreneurship. According to the OECD definition, self-employed individuals are motivated by profit and business opportunities in accordance with Kirzner's definition of entrepreneurship. Under this interpretation, the self-employment rate may also be a measure of the degree of Kirznerian entrepreneurship in the economy.