Political Empowerment, Rule of Law, and Women's Entry into Entrepreneurship

Abstract

Women's political leadership may contribute to women's entrepreneurship by removing existing constraints on the economic behavior of women, assuming these changes are then enforced. We examine the association of women's political power and a country's rule of law with women's entrepreneurial entry, using the lobal ntrepreneurship onitor dataset combined with other indicators. Both variables are positively associated with women's entry into entrepreneurship, and the association between political empowerment. Entry into entrepreneurship is moderated by rule of law, with higher levels of women's political power having greater effects in countries with higher levels of rule of law. Implications are discussed.

Notes

1. Publicly available GEM surveys from 2001 to 2008 include over 60 countries. However, usable data for the purpose of our study were available for only 53 countries. Since its inception in 1998, GERA has been collecting individual‐level data alongside country‐level data across countries that reflect the incidence as well as determinants and outcomes of entrepreneurial activities in participating nations. Although an unbalanced panel, GEM lists over 60 countries spanning eight years (2001–2008) by surveying adult populations. Consequently, GEM is the most comprehensive and harmonized data set listing internationally comparative data on individual‐level entrepreneurial behaviors. Since describing GEM's operationalization of its survey and listing the extent of its usage in various research is beyond the scope of this study, we direct the attention of our readers to Levie and Autio (2011), Minniti, Arenius, and Langowitz (2005), and Reynolds, Bosma, and Autio (2005).

2. The World Bank Governance Indicators lists a total of six distinct country‐level aggregate indicators—voice and accountability, political stability and absence of violence and terrorism, government effectiveness, rule of law, control of corruption, and regulatory quality. Each of six aggregate WGI measures is constructed by averaging together data from the underlying sources that correspond to the concept of governance being measured. We used the last indicator for the purpose of our study.

3. WGI uses nine different underlying representative sources to measure key concepts that define rule of law, including the extent of violent and organized crime, fairness and speediness of judicial process, enforceability of contracts, intellectual and private property rights protection, reliability of police services, respect for law, government respect for contracts, settlement of economic disputes, etc.

4. The minimum and maximum values reported in the text and tables for both rule of law and political empowerment were obtained from the average values over 2002–2008. However, the minimum and maximum in terms of the raw values for each of these two predictors for all countries and for all years (year_country matrix) are slightly different—minimum of −1.18 for Ecuador in 2008 and maximum of 2.01 for Denmark in 2007 for rule of law and minimum of 1.80 for Egypt in 2008 and maximum of 47.30 for Sweden in 2006 for political empowerment.

5. z‐Scores were used to create the interaction term to minimize issues related to multicollinearity.

6. Countries with GDP per capita less than $10,000 (USD) were categorized as developing, whereas those with higher GDPs were categorized as developed nations.

7. Results for all robustness checks are unreported here for brevity but are available from the authors upon request.

8. The sample command in Stata generates randomized subsamples out of the entire dataset by retaining a given percentage of data as defined by the user.

9. The mentioned effects were observed to be statistically significant on subsamples with 20 percent or lower observations when OLS regressions were employed on them. This was an expected finding—significant variance in dependent variables across groups (our data were grouped by country and year) necessitates multilevel analyses. Smaller subsamples do not show significant variance, suggesting that group effects are not dominant; hence, use of OLS—which considers data as pooled without any distinctions to groups—is adequate.

Additional information

Notes on contributors

Sonia Goltz

Sonia Goltz is professor of Organizational Behavior in the School of Business and Economics at the Michigan Technological University.

Mari W. Buche

Mari W. Buche is associate professor of Management Information Systems in the School of Business and Economics at the Michigan Technological University.

Saurav Pathak

Saurav Pathak is the Rick and Jo Berquist assistant professor of Entrepreneurship and Innovation in the School of Business and Economics at the Michigan Technological University.