Women’s entrepreneurship policy research: a 30-year review of the evidence

Abstract

This paper focuses on women’s entrepreneurship policy as a core component of the entrepreneurial ecosystem. We use a systematic literature review (SLR) approach to critically explore the policy implications of women’s entrepreneurship research according to gender perspective: feminist empiricism, feminist standpoint theory, and post-structuralist feminist theory. Our research question asks whether there is a link between the nature of policy implications and the different theoretical perspectives adopted, and whether scholars’ policy implications have changed as the field of women’s entrepreneurship research has developed. We concentrate on empirical studies published in the “Big Five” primary entrepreneurship research journals (SBE, ETP, JBV, JSBM, and ERD) over a period of more than 30 years (1983–2015). We find that policy implications from women’s entrepreneurship research are mostly vague, conservative, and center on identifying skills gaps in women entrepreneurs that need to be “fixed,” thus isolating and individualizing any perceived problem. Despite an increase in the number of articles offering policy implications, we find little variance in the types of policy implications being offered by scholars, regardless of the particular theoretical perspective adopted, and no notable change over our 30-year review period. Recommendations to improve the entrepreneurial ecosystem for women from a policy perspective are offered, and avenues for future research are identified.

Appendices

Appendix 1: Stages in the SLR process

Stage	Description
1	A list of the “Big Five” journals in entrepreneurship research was compiled: ERD, ETP, JBV, JSBM, SBEa.
2	Each member of the author team was allocated a discrete 10-year period to search: period 1: 1983–1992; period 2: 1993–2002; period 3: initially 2003–2012 and subsequently updated to include the period up to end December 2015.
3	Within journal searches were conducted by means of a systematic Boolean keyword search using the terms “gender” OR “women” OR “woman” OR female AND “entrepreneur” OR entrepreneurship OR “business” in the title, keywords, and abstract field. We used the academic Scopus database to perform our search. This database covers all the journals in our SLR. As a cross check, in some cases, content pages of each journal issue/volume were examined to ensure no relevant paper was omitted/missed.
4	The resulting articles were then examined, and exclusion criteria were applied. Discussions between the authors throughout the process ensured that any further potential exclusions were discussed and agreed. In this step, we excluded 50 papers in the last (updated) period and 39 papers in the three first periods. This resulted in a total of 165 papers.
5	The common thematic reading guide designed by authors was then applied
6	The author team discussed articles as they reviewed them to ensure consistency of analysis.

1. ^aWe acknowledge that focusing on the “Big Five” journals restricts our study somewhat; however, we wanted to concentrate on those journals deemed to have the most significant scholarly impact in the field and to explore any notable changes over a significant period. Furthermore, in view of the considerable narrative content involved and the in-depth qualitative analysis required, we felt that such focus was warranted

Appendix 2: Reading guide

SLR reading guide
1. Article title
2. Author(s)
3. Year of publication
4 Journal
5.Feminist perspectives	Perspectives	Explicit (x)	Implicit (x)	Comments
	FE	☐	☐
	FST	☐	☐
	PSF	☐	☐
	Other	☐	☐
6. Feminist perspective key findings/notes
7a. Policy implications	Yes (x) ☐ No (x) ☐
7b. If yes in 7a. How are the policy implications being reported in the paper	Explicit ☐ Implicit ☐	Policy implication findings /quotes/notes:
7c: If yes in 7a. Definition of policy implications:	Yes ☐ If yes Explain: No ☐
	How do they address the implications:
8a. Other implications reported?	☐ Future entrepreneurs ☐ Nascent entrepreneurs ☐ Education ☐ Female business owners ☐ Financial capital providers ☐ Researchers ☐ Managers ☐ Others……………………………………………..
8b: Practical implication text:
9a. Ecosystem code	☐P = Policya ☐F = Funding and finance ☐C = Culture ☐M = Mentors ☐U = Universities as catalyst ☐E = Education ☐H = Human capital and workforce ☐L = Local and global markets ☐NO = No implications for ecosystem
9b. Ecosystem explanations (beside policy implications):
10a. Type of implications for future research	☐ Entrepreneurship research ☐ Gender research ☐ Other: …………. ☐ No implications
10b. If yes in 10a. Implication text:
11. Sample size n:
12. Country:

1. ^aNote: Do not code for P here unless you have coded for Yes in 7a

Appendix 3. Logistic regression

We formulated H1: There is a significant relationship between policy implications in articles and feminist perspectives applied. Given the binary nature of the dependent variable, we conducted a binary logistic regression with feminist perspectives as explanatory variables and policy implication as the binary dependent variable. The dependent variable is coded 1 for articles with policy implications and 0 for those with no policy implications. The feminist perspectives are used as the independent variables, all coded as dummy variables. The feminist perspective “other theoretical perspective” containing only three cases was removed from the analyses since there was no variation in this variable. All three of these cases where coded “policy implication.” To ensure we accounted for all the variability in our SLR data, we controlled for other observations collected in our SLR, such as geographical area, time periods and specific type of implication, and type of journal, all coded as dummy variables.

1.1 Results

Categories	Variables	Policy implication
		Coefficienta (std. error)	Wald χ2
Feminist perspectives	Post-structural feminism	–	3.39
	Feminist empiricism	1.023 (0.73)	1.94
	Feminist standpoint theory	0.190 (0.76)	0.06
Journals	SBE	–	6.07
	JSBM	1.134 (0.70)	2.64
	ETP	− 0.240 (0.70)	0.12
	ERD	− 0.286 (0.71)	0.16
	JBV	0.342 (0.70)	0.24
Area	Asia + Africa	–	1.02
	Europe	0.371 (0.68)	0.29
	America	− 0.149 (0.63)	0.05
Time	1982–1993	–	15.57
	1993–2002	1.643 (0.57) *	8.11
	2003–2015	2.472 (0.63) **	15.21
Practical implications	1 = yes, 0 = no	1.454 (0.43) **	11.51
	Constant	− 2.454 (1.2) **	3.75
Model diagnostic
N		162
− 2 log likelihood		158.00
Cox and Snell R2		0.213
Nagelkerke R2		0.303
Model χ2		38.892 **
df		11
Hosmer and Lemeshow χ2		8.002 (p = 0.433)
Overall % correct prediction		76.1%

1. ^aNote: ** and * denote significance at the 1 and 5% levels, respectively
2. The chi-square shows that the overall model is significant at the 0.00 level and it predicts 76.1% of the responses correctly. The Hosmer-Lemeshow goodness of fit (p > 0.05) gives an acceptable outcome (χ² = 8.002, p = 0.433), indicating that the model prediction does not significantly differ from the observed. The − 2 log likelihood was 158.052 and the two pseudo R² values are 0.213 (Cox and Snell) and 0.303 for Nagelkerke, indicating that the fit of the model to the data was moderate. According to the regression results, H1 was rejected: None of the feminist perspectives was significant, indicating that feminist perspectives cannot be used as a determinant whether an article reports policy implications or not. Both the time variable and the practical implication variable were significant at p < 0.05. This indicates that articles in the two last time periods are more likely to report policy implications compared with the first time period. Also, articles reporting practical implications are more likely to also report policy implications than those which do not report practical