Prof. Poh-Kam Wong, National University of Singapore Entrepreneurial Firm Formation and Income Equality: A Cross-Country Analysis 1. Introduction Wealth distribution is a fundamental concern in economics. Surprisingly, one feature of economic systems often overlooked when discussing wealth distribution is the role of entrepreneurship. In the conceptual (Meh, 2005; Cagetti and De Nardi, 2006) and empirical (Quadrini, 1999) literature, there is support for the conventional wisdom that entrepreneurship is associated with higher income inequality. However, the empirical evidence on economies outside the US appears to be ambiguous. This paper attempts to close this gap by exploring the empirical relationship between entrepreneurship and income distribution, utilizing cross-national data. The firm formation attribute of entrepreneurship is emphasized and comparisons are made between different measures of entrepreneurial activity. 2. Entrepreneurship and Income Distribution in the Literature A number of studies in the income distribution literature have sought to examine the association between entrepreneurship and income inequality. However, the empirical evidence that emerged from these studies appear to be ambiguous. By introducing entrepreneurship as an additional dimension in the Kuznets (1955) curve (which hypothesizes a U-shaped relationship between macroeconomic development and equality), Deutsch and Silber (2004) implied that rising levels of entrepreneurship are associated with widening income inequality. This supposition that entrepreneurship causes inequality appears to be supported by empirical work using Italian and US data (Quadrini, 1999, 2000; Cagetti and De Nardi, 2003, 2006; Budria et al., 2002; Quintano et al., 2005). However, other studies have found that the direction of the relationship between inequality and entrepreneurship to be dependent on moderating factors, such as the tax regime (Meh, 2005; Kanbur, 1982). A number of economy-specific studies using data from non-OECD economies have postulated the entrepreneurship should redistribute income more equitably. Berkowitz and Jackson’s (2006) empirical work on Russian and Polish income distribution indicate that new firm creation is associated with both larger income and a larger portion of income distributed to the lower quintiles. Kimhi (2009) found that a uniform increase in entrepreneurial income reduced household income inequality in southern Ethiopia. However, increasing the number of entrepreneurs per se did not affect income inequality. A recent contribution to this discussion is found in a conceptual paper by Spencer et al. (2006), attempting to link different views of entrepreneurship to their potential wealth distribution effects. Drawing on illustrative case studies of disruptive technologies that have spawned new industries, they argue that it is new firms that embody Schumpeter’s (1934) original emphasis on the wealth distribution function of independent entrepreneurs, a function overlooked by subsequent theories of entrepreneurship that have moved in divergent directions. 3. Model, Measures and Data for Empirical Analysis The contrasting findings from the literature suggest that the relationship between entrepreneurship and income distribution may not be linear. To verify this, we perform a cross- country analysis, hypothesizing that there is a non-linear causality effect of entrepreneurship rate on the degree of income inequality. The basic analysis model takes the general form: Inequality = + 1i (Control Variables) + 2 Entrepreneurship + 3 Entrepreneurship 2 We have taken advantage of comprehensive cross-economy databases released by international organizations to construct our analysis dataset. Data on Income Inequality are taken from the UNU / WIDER World Income Inequality Database (WIID) version 2.0c. Values for 2005 are used where available; where 2005 data are not given, the latest available year is used provided that the year is no earlier than 2001. The control variables included are: GDP per capita & GDP per capita squared – Control for the Kuznets (1955) curve effect. Growth in Real GDP – The literature establishes a relationship between growth and inequality although there are contrasting findings on the direction of correlation (eg. Li and Zhou, 1998; Perotti, 1996; Barro, 2000) Change in Wages – Controls for effect of wages as the price of labour. Ratio of FDI to GDP and Ratio of FDI:GDP squared – This measure proxies the openness of an economy. As reviewed by Anderson (2005), time series studies have established a link between openness and inequality, but the evidence is less conclusive in cross-country studies. A squared term is included to control for potential non-linearity as found by Calderon and Chong (2001). Several different measures of entrepreneurship are used, drawing from the World Bank Group Entrepreneurship Survey (Klapper et al, 2008) and the Global Entrepreneurship Monitor (GEM). Our primary interest is focused on entrepreneurship as defined by firm formation. From the World Bank Survey, we extracted the variable New Business Density, which measures the ratio of total new business registrations to national population, Three GEM measures, all of which are similarly propensity-based, are examined. Total Early Stage Entrepreneurial Activity (TEA) is the proportion of adult population involved in early stage entrepreneurship. TEA is divided into two sub-components: Nascent Entrepreneurship – those involved in attempting to start a business - and Start-Up Entrepreneurship – those who have started a new business in the last 3.5 years. 4. Preliminary Results Preliminary regression results are shown in Table 1. Broadly, there is support for the Kuznets curve hypothesis. However, the other control variables included were non-significant. Comparing the World Bank and GEM measures, the squared terms for the GEM propensity measures were found to be non-significant and have been consequently dropped to improve model fit. In contrast, the squared terms for the World Bank measure was significant, showing an inverse U-shaped relationship between new business propensity and inequality. The results for the GEM measures highlight that it is Start-Up Entrepreneurship that is linked to inequality, with a positive and significant relationship. The pre-start-up processes of nascent entrepreneurship do not seem to influence income distribution. The final version of the paper will include more in-depth analysis of interaction terms between control variables and entrepreneurship measures to more clearly establish the effects of entrepreneurship on income distribution in economies of different development levels. Policy implications and suggestions for further research will be discussed.