Barriers to ethnic minority and women’s enterprise: Existing evidence, policy tensions and unsettled questions

This article presents an overarching review of the evidence regarding enterprise diversity. It discusses the context of ethnic minorities and women in enterprise and summarises research evidence relating to their relative access to finance, market selection and management skills. Policy within the field of diversity and enterprise is characterised by a number of tensions and unresolved questions including the presence of perceived or actual discrimination, the quantity and quality of ethnic minority and women-led businesses, potential market failure in the support provided to diverse enterprises and the substantive uniqueness of ethnic minority and women-led enterprises. Particular implications for policy and practice as well as directions for future research are discussed.


Introduction
A notable feature of the enterprise policy discourse in the United Kingdom is the longstanding concern that entrepreneurial ambitions, participation and fortunes are unevenly distributed across social groups (see, for example, Parker, 2004;Ram and Jones, 2008;Southern, 2011). Two groups have captured the particular attention of policy-makers: ethnic minorities and women, and both have been the focus of concerted efforts by successive governments aiming to increase enterprise levels (Small Business Service, 2003). The rationale for supporting both groups is subtly different.
Policy interest in women's enterprise has concentrated on the potential economic gains that could accrue from increased rates of participation and productivity. Ethnic minority businesses (EMBs), many of which operate in highly visible sectors such as retailing and catering, have been particularly valued for their role in promoting social cohesion and multiculturalism. While policy-makers typically regard these two groups as distinctive -and each has developed its own separate set of activists and advocacy groups -policy initiatives have often occurred in tandem. This was most clearly seen in the simultaneous establishment, management and reporting of the Ethnic Minority Business Task Force and the Women's Enterprise Task Force (2007)(2008)(2009). For both groups, the prevailing policy discourse emphasises boosterism (Blackburn and Ram, 2006), an approach underpinned by the view that 'all enterprise is good, more enterprise is better'. Despite longstanding policy interest, research suggests that attempts to boost ethnic minority and women's (EMW) enterprise have yielded only modest changes in the numbers of diverse business owners (Alexander et al., 2009;Kelley et al., 2011;Marlow et al., 2008;Ram and Jones, 2008).
Entrepreneurship research also regards these two groups as separate and distinctive, a bifurcation which evolved from and has widened with the growing specialisation within entrepreneurship scholarship. As business owning groups, ethnic minorities and women have both been researched extensively, but largely in isolation from each other. Few studies have focused simultaneously on the experiences of both EMB owners and women business owners, to the extent that there appears to be little overlap between the two specialist research areas. Nevertheless, there are considerable practical similarities, not least because the lower enterprise participation and performance of women and some ethnic minorities are due, in part, to lower levels of resources and other factors necessary for business entry and growth. These chiefly pertain to money, markets and management skills, dubbed the '3Ms' by Bates et al. (2007) who identified them as vital building blocks for business that ethnic minority enterprises are frequently unable to adequately access. Brush et al. (2009) in turn espouse and elaborate the 3Ms framework as a springboard for researching women's entrepreneurship, thereby, inadvertently perhaps, integrating the conceptual underpinnings of both EMW's enterprise.
The Enterprise Research Centre's focus on diversity and small and medium-sized enterprises (SMEs), one of its six thematic research areas, allows us to bring together research evidence relating to ethnic minority and female business ownership to review the connections between these two separate areas of inquiry. This article presents the results of an examination of the extant research evidence on barriers to EMW's enterprise, its purpose being to identify key gaps that require prompt research attention, highlight policy tensions common to both groups and outline a research agenda accordingly. Following Bates et al. (2007), the review of prior research on EMW's enterprise is structured around three themes central to both research areas, pertaining to access to external finance (money), markets and market selection, and management skills and outcomes. Distilling two extensive bodies of the literature into three research themes inevitably entails losing much of the broad landscape of each research area, but this is compensated by the ability to focus in detail on these three issues that are central to both fields. In doing so, we are able to draw broad comparisons in the experience of entrepreneurship as perceived by each group and also draw attention to particular research questions, methodologies and insights that have proven fruitful in one area that may profitably be used in the other.
As one of the first reviews to bring together the research evidence pertaining to EMW's business ownership, this article affords an opportunity not only to explore how business ownership is experienced by diverse social groups but also to consider the consequences of research practices that separate social groups into distinctive areas of study, in which minority subjects may be represented as being 'in deficit' to a mythologised norm (Ahl and Marlow, 2012;Ogbor, 2000). Bringing together these two separate strands of scholarship raises two immediate questions: first, a consideration of why these research strands have not been united previously and, second, to reflect on the likelihood that future studies will move beyond a focus on the experiences of a single group to encompass multiple dimensions and modalities of social relations (McCall, 2005). In answer to the first question, it is clear that while intersectionality -the interaction of multiple identities and experiences of exclusion -has emerged as a major paradigm in social research, particularly women's studies (Walby et al., 2012), it has made minimal impact either on small business research or on public policy relating to SMEs. So far, the two research strands have developed to their current state because of the advantages conferred by specialised scholarship, but there now may be benefits in exploring whether the intersection of multiple inequalities provides opportunities for advancing entrepreneurship theory and informing public policy relating to SMEs. There are similarities here with the concept of 'mixed embeddedness' (Kloosterman, 2010) in ethnic minority and immigrant business research, which emphasises the variety of political, spatial, economic and regulatory contexts in which minority firms operate. Although the approach has gained popularity among researchers (Ram et al., 2013), its traction with policy-makers is negligible (Ram and Jones, 2008). These issues are considered in the conclusions of this article, which also reflects on some policy tensions and highlights directions for future research.

Ethnic minority-owned businesses
Within the broader social context of the 'superdiversity' in modern Britain (Vertovec, 2007), EMBs are a complex and rapidly changing group of enterprises that include longstanding immigrant communities, notably South Asians and African-Caribbeans, and comparatively new arrivals from eastern Europe and Africa (Clark and Drinkwater, 2010;Ram et al., 2013;Sepulveda et al., 2011). While the term 'EMB' is a convenient way of describing enterprises owned and managed by ethnic minorities, and the term adopted in this article, some caution is necessary in its use. Since many EMBs themselves eschew the ethnic label, ethnicity should not be taken as the defining characteristic of EMBs. We recognise that the processes that attend the labelling of such groups are rarely remarked upon; yet they can have profound effects on the way in which immigrant and other groups are constructed and perceived by the wider population. Often unarticulated philosophical assumptions lurk behind the taken-for-granted terminology that permeates policy discourse in this area. This neglect leaves researchers and policy-makers in a 'treacherous bind' which involves managing the tension of 'how [they] can work with inadequate racial and ethnic categories that are to hand, whilst also finding ways of identifying and disrupting the ways in which the same categories can "essentialise"' (Gunaratnam, 2003: 29). Our review is cognisant of this double-bind.
Recent government estimates suggest that there are almost 300,000 EMBs comprising around 6% of the small firm population in the United Kingdom (BIS, 2013b). Other estimates suggest a share of 8%, but indicate further that EMBs are unevenly distributed across the United Kingdom with higher concentrations in the main urban areas, notably London, Birmingham, Manchester and Leeds (IFF Research, 2011;cf. Regeneris Consulting, 2010). UK government estimates suggest that EMBs contribute about £3 billion worth of gross value added (GVA) to the UK economy (BIS, 2013b), around 6%-7% of total GVA attributable to all SMEs in the United Kingdom (Regeneris Consulting, 2010). Other economic contributions include the revival of declining sectors and places (McEwan et al., 2005), as well as the enhancement or development of new conduits for transnational trading links (Mascarenhas-Keyes, 2008). EMBs also play an important role in the social adaptation and integration of new migrants in their local economies and communities (Jones et al., 2012;Zhou, 2004). Yet, EMBs continue to face barriers in relation to access to finance, the often narrow markets and sectors that they serve, and management competencies and practices. Below, we review the extant evidence pertaining to these fundamental issues.

Finance
Historically, access to finance has been cited as one of the most significant barriers facing EMBs (Bates, 2011;Pierce, 1947). Despite this, there has been remarkably little research specifically examining this major concern. UK evidence suggests, however, that ethnic minority groups tend to have widely divergent experiences of external finance (Fraser, 2009;Ram et al., 2002; on EMBs in the United States, see Blanchflower et al., 2003). Research has focused on bank lending as the major providers of external business finance. Drawing on large and representative datasets examining bank lending to EMBs, Fraser's (2009) study proposed three key observations.
First, the experience of unfavourable credit outcomes varies among entrepreneurs from different minority ethnic groups. Black African firms are more than four times as likely as White firms to be denied a loan outright, Black Caribbean firms 3.5 times as likely, Bangladeshi firms 2.5 times as likely and Pakistani firms 1.5 times as likely. Indian firms had a slightly lower loan denial rate than White firms. Furthermore, discouragement, a situation where a firm would like to apply for finance but does not for fear of rejection (Kon and Storey, 2003), was found to be highest among EMBs. Here, 44% of Black African, 39% of Black Caribbean, 31% of Bangladeshi, 21% of Pakistani and 9% of Indian firms compared to 4% of White firms reported that the fear of rejection had stopped them from applying for loans that their businesses actually required.
Second, standard risk factors (e.g. age of business, financial track records) rather than direct discrimination largely accounted for discrepancies between different ethnic minority groups. As Fraser (2009) noted, 'In particular Black African firms are significantly more likely to miss loan repayments and/or exceed their agreed overdraft limit and this behaviour seems to largely account for their much higher loan denial rates' (p. 601).
Finally, even after controlling for the effect of other explanatory factors such as poorer credit worthiness, Fraser (2009) found that ethnicity remained a significant factor in discouragement, particularly for Black Caribbean firms and, to a lesser extent, Indian firms (see also Blanchflower et al., 2003). Given that modern lending relies to a large extent on more objective statistical estimations of customer riskiness than relationship lending where the subjective judgement of the lending bank employee may be pivotal (Allen et al., 2004;Berger and Udell, 2002;Mester, 1997), and the regulatory context is such that discrimination by race, ethnicity or gender is prohibited, Fraser (2009) argues that EMBs may be discouraged by their own misperceptions of the possibility of discrimination.
While the concept of discouragement has itself received little research attention both in terms of conceptual elaboration and wider empirical application, there is clearly a significant ethnicity element requiring further investigation. Indeed, intra-minority differences in the perceptions of bank borrowing have scarcely been addressed in the extant literature. There is also scope for experimental research methodologies in assessing the bank lending experiences of minority entrepreneurs, following the experimental techniques used to assess the impact of gender on bank lending decisions Wilson et al., 2007).
For policy and practice, the importance of fostering mutual understanding between banks and EMBs by improving engagement and information flows cannot be overemphasised. Programmes of action learning and engaged scholarship may be expected to enable EMBs of various ethnic backgrounds to appreciate the changing nature of modern bank lending with more objective systems and less scope for prejudice. The knowledge or assurance that applications are mostly considered using statistical methods and objective risk variables may abate EMB discrimination fears and misperceptions, thereby encouraging application. Banks, however, may learn of the particular sources of concern held by EMBs and therefore, re-consider their lending procedures.
Indeed, research suggests that while the substitution of statistical techniques as a replacement for subjective judgement by local bank managers in predicting risk has led to increases in bank lending to small businesses (Frame et al., 2001), there have also been misgivings about this practice. Avery et al. (2004), for example, contend that misspecification errors in statistical models may lead to unpredicted credit losses on the part of the bank, as well as high borrowing costs and unnecessary customer rejections. While ethnicity is not itself directly factored into these models, certain variables associated with poor credit scores, such as problematic postcodes (Fraser, 2009) and sectors (Ram and Jones, 2008), may be coterminous with ethnicity and hence, adversely affect EMB applications. Other situational household and social factors central to enterprise that banks may consider have been found to vary non-randomly across minority ethnic groups (Borooah and Hart, 1999). Accordingly, structural financial disadvantage may neither be independent of, nor incidental to, ethnicity. Such effects may be culturally disaggregated in a way that may be occluded by ethnic labels. Basu and Altinay (2002), for example, report significant differences between Asian entrepreneurs of various origins. However, besides fragmented findings and anecdotes, a granular audit of the extent to which unfavourable structural or 'strictly business' factors are variously ethnicised, if at all, is yet to be documented.

Markets
Obvious but easily overlooked is the simple truth that, without a viable volume of custom, even the most richly capitalised and expertly managed firm is unable to succeed. Certainly, the initial wave of policy-based EMB surveys tended to pay more attention to lack of financial and human capital than to lack of markets, with the effect of potentially over-emphasising the former at the expense of the latter. Beyond an assessment of their independent effects, a deeper investigation of path dependence, interactions and/or nestedness of these and their effects on various aspects of minority enterprise may illuminate the situation further. For example, small-scale retailing and catering make relatively modest demands on capital and expertise and hence, are highly popular with underresourced immigrant entrepreneurs who may not have access to large supplies of capital and may also be discouraged from pursuing bank finance or have had rejected applications. With low cost entry, however, the supply of minority retail and catering businesses has tended to out-run market potential, creating excessive competition and market saturation, thereby stifling the earning capacity of competing EMBs (Jones et al., 2000). Heightening this problem is exposure to fierce competition from corporations (supermarkets and restaurant chains). For the most part, however, the most intense competition faced by South Asians and Chinese is with one another .
Although the latter is also a problem for African-Caribbean entrepreneurs, their exposure is less acute because of their lesser concentration in catering and small retail (Ram and Jones, 2008). African-Caribbean self-employment is not only much less restricted to the disadvantaged retail and catering sectors, it is also much smaller than South Asian self-employment in general (Borooah and Hart, 1999;Ram and Jones, 2008). Historical studies have suggested that the need to drum up custom proved a challenge at the outset for immigrant newcomers searching for market space amid deeply rooted incumbent native firms. The relatively higher prevalence of South Asian retailers since the 1970s was in part driven by their own co-ethnic communities with their demand for Asian food, clothing and other ethno-cultural artefacts supplemented by demand for non-specific items purchased by fellow Asians -a demand which emanated from a combination of ethnic loyalty and neighbourhood proximity (Aldrich et al., 1981).
In contrast, it has been argued that the comparatively slight volume of Black Caribbean retailing stems from a relative lack of this ethnic customer particularism (Ward, 1991). While Asian entrepreneurs contended with narrow ethnic markets, cultural factors would have rendered the effective customer base for potential Black entrepreneurs too narrow to sustain even a small ethnic retail operation. Nevertheless, from the 1980s, Asian and other ethnic minority retailers have increasingly spread into more expansive White residential markets (Ward, 1985). Increasingly, this caused a shift towards non-ethnic general purpose necessities, with food retailing and newsagents in the vanguard. More spectacularly, perhaps, over the years, South Asian and Chinese entrepreneurs have managed to cater for a vast demand for exotic cuisine (Barrett et al., 1996). Little research has been carried out, however, to investigate why Africans and West Indians have not been as successful in commodifying their cuisines and/or targeting wider markets. In particular, whether it is the size of the immediate ethnic market that affords significant springboard effects or whether there are other cultural entrepreneurial drivers that play a larger role has not been established.
Indeed, besides food and retail, globalisation dynamics in the apparel industry afforded opportunities for sectoral diversification that were seemingly seized by South Asian and (in London) Turkish entrepreneurs. Setting up operations in defunct industrial premises, these businesses took advantage of locational proximity to outcompete foreign suppliers. Notwithstanding such advances, structural disadvantage springs from the fact that UK producer firms are in direct competition with low-cost labour in the developing world, with only geographical proximity operating in their favour. With such overseas competition, especially the ascendant role of Chinese producers, survival by EMBs in the apparel sector has been increasingly precarious and dependent on extreme cost-cutting measures, sometimes even non-compliance with minimum wage requirements (Jones et al., 2006;. Coupled with this, their customers are mostly large-scale high street retailers, whose terms and conditions tend to be disadvantageous (Ram and Jones, 2008).
In the low value-added, low-profit areas in which EMBs operate, commercial survival is often a painful struggle. Indeed, while there is evidence of some diversification of markets and sector, EMBs remain in disadvantaged spaces characterised by parochial ethnic silos, narrow localitybased markets. Even for firms operating in White residential areas, cannibalistic competition between EMBs and their low-pay 'open-all-hours' operations . Exposure to unequal competition from corporations constrains scalability and returns. (Jones et al., 2000), global competition from cheaper imports and exploitative monopsonistic buyers (Ram and Jones, 2008), as well as unequal treatment in industrial markets (Ram and Smallbone, 2003;Worthington, 2009) also pose significant threats to small EMBs.
US research also finds that low performing EMBs, which again is uneven between different ethnic minority groups (Fairlie and Robb, 2008), have limited labour markets in that they frequently recruit fellow ethnics, while more successful EMBs have a more diverse workforce with significant shares of White employees (Bates, 2011). These labour market dynamics have rarely been studied in the United Kingdom. In general, however, future prospects of EMBs within the various ethnic groups critically depend upon the creation of new market opportunities by targeting broader markets, as well as diversifying into markets and sectors higher up the value-added chain. While it hardly needs stating that such a break-out requires financial and human capital on a scale as yet unavailable to most EMBs (Ram and Jones, 2008), a thorough understanding of the whole gamut of barriers faced by the various minority businesses in broadening their markets and the role policy could play is yet to be established.

Management
While skills and other entrepreneurial capabilities have been identified as an important building block for the creation of viable small businesses, access to educational and business-related experiences is a significant barrier facing EMBs (Bates et al., 2007). Evidence from the United States suggests that the possession of higher level qualifications is associated with the higher presence of recent African American graduates in 'emerging' business lines, including new business sectors like media, information technology (IT) and engineering (Bates, 2011). Besides education, however, applied managerial skills are also crucial for success in business. Reiterating the importance of context and the interrelatedness between the key factors, a recent study in the USA observed differences between the fortunes of Latino construction entrepreneurs in Philadelphia and those in North Carolina (Iskander and Lowe, 2010). In Philadelphia, Latino entrepreneurs were confined to informal and unskilled small-scale residential works as powerful labour unions that controll industry training and credentialising processes are largely closed to immigrant workers. Here, EMBs evidently face obstacles related to both the labour market and consumer market. In contrast, the relative absence of such limitations in North Carolina enabled Latino workers to gain experience; first, in supervisory and managerial roles in mainstream construction companies and later, as independent entrepreneurs embedded in the networks of prime contractors as key subcontractors (Iskander and Lowe, 2010).
In recent British studies, three trends are evident. First, although the uptake of information and communications technology (ICT) among EMBs in traditional sectors has remained relatively low (Beckinsale et al., 2011), EMBs are increasingly to be found in non-traditional sectors (notably, pharmaceuticals, IT and the media) and run by highly credentialised owners (Ram and Jones, 2008). The possession of higher level skills creates opportunities to engage in knowledge-based entrepreneurial activities where knowledge is a primary asset and a source of competitive advantage (Thompson et al., 2010).
Second, there is a lengthy and persistent history of mismatch between qualifications and occupation in the field of self-employment with highly qualified individuals in activities completely unrelated to their specialised expertise (Aldrich et al., 1981;Jones et al., 1994;Ram et al., 2002). Studies of new migrant business repeat this finding Sepulveda et al., 2011), and the prevalence of 'necessity entrepreneurship' among EMB owners in the United Kingdom is illustrated in a series of recent studies (for a review, see Ram and Jones, 2008).
Although the pertinence of the necessity/opportunity construct in deprived contexts has been questioned, not least since motivations may change post-entry (Rosa et al., 2009;Williams and Williams, 2011), the propensity towards necessity entrepreneurship among highly educated ethnic minorities and the labour market dynamics in play here remains relevant, albeit relatively underresearched. For example, Fraser (2009) finds high levels of financial delinquency among Black African entrepreneurs, despite the fact that they had relatively higher rates of professional and postgraduate qualifications. It may be the case, here, that applied professional experience may be an integral complement to formal qualifications, and lack of it may actually render the qualifications economically futile even where the agent pursues self-employment.
Finally, the market 'break-out' that is often called for in policy discourse is heavily dependent on the more strategic deployment of the entrepreneur's labour. This is likely to require not just experience and skills, but a wholesale re-direction of the proprietor's efforts, if not the nature of the business itself. This can be challenging for the many EMBs that operate on a tight budget with little paid assistance (Jones et al., 1994) and with high propensity for ethnic homophily (Ruef et al., 2003). Yet, a broader personnel base is often a feature of EMBs that have diversified, something which is evident in ethnic minority suppliers to large organisations . Intriguingly, Wang and Altinay's (2012) study of entrepreneurial orientation (EO) in EMBs suggested that while co-ethnic advice and labour had no effect on the development of a firm's EO, access to co-ethnic suppliers of facilities and utilities had a significant effect and promoted subsequent employment growth.
Self-employed minorities engaging in a variety of low value-added activities may therefore, not gain the skills and capabilities necessary to strategically steer the business to success. Rather, it is the employment and deployment of workers in more specialised and professional roles, especially workers recruited from diverse ethnic groups, that are associated with better EMB performance (Bates, 2011). Whether the division of labour and diversification of workers are causally related to firm performance and growth has yet to be established. Furthermore, the extent to which EMBs in the United Kingdom are shunning or are prepared to shun parochial and patriarchal tendencies and move towards a more diverse workforce, more professionalised delegation and more broadminded strategic growth, and the complexities such moves entail are relatively unexplored by enterprise research (Jones and Ram, 2010).

Women-owned businesses
Given the relatively high proportion of joint male-female business ownership, particularly among marital couples and sibling relations, women-owned businesses can be difficult to define and precisely enumerate. However, recent estimates by the Office for National Statistics (ONS) indicate that women comprise about 29% of the United Kingdom's self-employed population, and 22% of incorporated businesses are women-led (BIS, 2013b;Causer and Park, 2009). Women-owned businesses contribute about £75 billion to GVA, about 16% of the approximate GVA that all UK SMEs generate (BIS, 2013b). At the regional scale, self-employment rates for men and women are closely correlated, and the highest rates of female self-employment in the United Kingdom are in those regions where male self-employment rates are also highest (South East, South West and London) (BIS, 2013a).
Despite many initiatives to increase women's participation in enterprise (Alexander et al., 2009;Causer and Park, 2009), men are still almost twice as likely to start businesses as women (Kelley et al., 2011;Marlow et al., 2012). The 'global gender gap' debate underpins much of the policy interest in women's enterprise as it identifies a clear economic rationale for the encouragement of women to become independent business owners. Relative to other high-income countries, particularly the United States, UK rates of female business ownership have been persistently low (Hausmann, 2013;Kelley et al., 2011;Xavier et al., 2013). While there have been strong critiques of the United States as an appropriate comparator for UK entrepreneurship policy (Marlow et al., 2008), and pointed resistance to the view of women entrepreneurs as 'lacking and incomplete men' (Ahl and Marlow, 2012: 543), the scale of the UK enterprise 'gap' is commonly illustrated by reference to both men in the United Kingdom and women in the United States. These estimates suggest that an additional 150,000 businesses would be created if rates of business ownership among women were the same as men, and an additional 900,000 businesses would be created annually if the United Kingdom had the same rates of women's business ownership as in the United States (Alexander et al., 2009).
While the number of women engaged in self-employment and business ownership has risen in recent years, much of the growth has been in women working part-time and among those wanting more flexible working hours to complement domestic commitments (Causer and Park, 2009). Studies suggest that a substantial proportion of self-employed women 'may be working very few hours -as little as an hour per week', fitting in flexible self-employment around family commitments (Causer and Park, 2009: 47). Indeed, more women than men use their home as a business base (Mason et al., 2011), and some women do not even consider their self-employment as 'proper work' (Marlow, 2006). Research has found that although the career transition experiences of women are complex (Patterson and Mavin, 2009), many women use self-employment as a temporary solution within a broader career path and are therefore, likely to exit business for personal reasons other than business failure . While some studies report under-participation of women in self-employment, other data suggest that the participation of women may be widespread, but masked by dual ownership. In 2012, on top of the 19% of SME employers that were women-led, a further 23% were equally led by men and women, suggesting that about 42% of SME employers in the United Kingdom are at least equally, if not wholly, led by women (Rhodes, 2013). When entrepreneurship studies focus only on the individual, besides overlooking the important contributions of the family and the household to the business (Alsos et al., 2014), the much wider participation of women in enterprise is disguised (see also, Jennings et al., 2013). Thus, existing levels of self-employment among women are not accurate indicators of entrepreneurial activity and economic contributions. With much of the extant research evidence suggesting a bimodal profile of male-owned and femaleowned businesses with regard to size, age, income and other performance measures, efforts towards generating a richer and more robust understanding of women's entrepreneurship have been called for (Hughes et al., 2012).
That women are under-represented in enterprise can be seen to be a complex research issue obscured by the use of gender labels. Arguing that the use of biological sex to dichotomise businesses is overly simplistic, Bird and Brush (2002) propose a more nuanced approach which considers gender as a mental (cognitive and affective) perspective that influences the process of organisational creation and operations but is not necessarily isomorphic with biological sex (see also Marlow and Swail, 2014). Alongside these issues on the conceptualisation of gender in enterprise research, the extent and causes of female 'under-performance' have been subject to extensive debate and empirical research (Chell and Baines, 1998;Johnsen and McMahon, 2005;Marlow and McAdam, 2013;Watson, 2002). Towards advancing this research, Brush et al. (2009) argue that while Bates et al.'s (2007) framework is useful, extending and mediating the 3Ms with issues pertaining to family and household (motherhood) as well as the meso-context and social and cultural norms at the macro-level could further a holistic understanding of women's enterprise.

Money
Securing external finance has long been regarded as the major obstacle preventing women from starting and growing a successful enterprise. Although regulatory developments have sought to make access to finance gender neutral by annulling formal gender identification, women still perceive higher financial barriers (Roper and Scott, 2009) and are therefore, more likely to be discouraged borrowers (Freel et al., 2012;Treichel and Scott, 2006). Indeed, while sources of finance for male-and female-led businesses are largely similar (Irwin and Scott, 2010), studies show that women-owned businesses start with substantially lower levels of overall capitalisation from personal and external sources, use lower ratios of debt finance and are much less likely to use private equity or venture capital (Carter and Rosa, 1998;Carter and Shaw, 2006).
Further evidence suggests that initial under-capitalisation has a long-term effect constraining future business growth prospects (Alsos et al., 2006;Carter et al., 2001;Rosa et al., 1996;Watson, 2002), not least because under-capitalisation may mire the firms in sub-optimal scales (Quadrini, 2009), which leave them more susceptible to failure (Headd, 2003). Understanding gender-based differences in finance usage is therefore, crucial and has been a major research focus. Three main explanations often cited in the literature include structural dissimilarities, supply-side discrimination and demand-side risk aversion.
Structural dissimilarities. Differences in finance usage between male-and female-owned businesses are most often explained as a product of differences in business size, age and sector , with women more likely to found firms in sectors and sizes with low capital requirements (Carter and Shaw, 2006;Coleman, 2000). While this explains much of the difference, it is not a complete explanation; studies using matched samples of men and women have reported residual gender-based finance differences even after structural factors have been controlled (for a review, see Carter et al., 2007). Importantly, one recent study also suggests the presence of second-order gender effects in US small business borrowing costs, arguing that 'the "gendering" of structure is itself a gender effect' (Wu and Chua, 2012). Indeed, the extent to which structure and gender are coterminous remains an important question for research.
Supply-side gender discrimination. Although there have been high-profile accusations of gender discrimination by lenders (Hertz, 2011), there is virtually no evidence to unequivocally support this claim. Extant evidence from Italy (Alesina et al., 2013), France (Orhan, 2001) and the United States (Coleman, 2000) indicates that certain supply-side practices, in particular collateral requirements, may disadvantage women business owners more than their structurally identical male peers. Historically, results such as these were used to explain higher levels of female dissatisfaction with banks, and accompanied by claims of sexual stereotyping (Buttner and Rosen, 1988;Hisrich and Brush, 1990) and conjectures of unconscious discrimination (McKechnie et al., 1998). Recent studies deploying more sophisticated and experimental methodologies found that differences in finance usage were not the result of systemic supply-side gender bias, but rather the outcome of a co-produced lending decision about businesses which could often be started -although not necessarily successfully sustained -with minimal capital outlay Wilson et al., 2007).

Demand-side risk aversion.
Other studies suggest that the lower uptake of debt finance among women entrepreneurs may simply emanate from a reluctance to actually request it (Marlow and Carter, 2004). While some studies have drawn on the psychological literature to claim the existence of higher levels of risk aversion among women (Powell and Ansic, 1997), more credible analyses of women's entrepreneurial endeavours have argued that women's disinclination to engage in fastpaced business growth (Bird and Brush, 2002;Cliff, 1998) and reluctance to take on high levels of business debt (Watson and Robinson, 2003) are symptomatic of wider socio-economic gender differences. Feminist analyses have shown how entrepreneurship theory has been informed and shaped by prevailing hetero-normative assumptions that serve to valorise the successful male entrepreneur while positioning women as 'lacking' (Ahl and Marlow, 2012: 543). Assertions of female 'risk aversion' illustrate how the subordination of women may be reproduced within entrepreneurship theory by simplistic descriptions that offer little explanation and even less critical reflection on 'the idea of who can be and what might be an entrepreneur' (Ahl and Marlow, 2012: 543).

Markets
That many women are relatively loath to engage in or pursue business growth (Bird and Brush, 2002;Cliff, 1998;Rosa et al., 1996;Watson and Robinson, 2003) suggests that women's enterprise will by definition, and by choice, have narrow markets and smaller operations. Studies have consistently demonstrated that women-owned firms are typically smaller, over-represented within service sectors and more likely to be part-time and to operate from a home-base (Hughes et al., 2012). Women-led businesses may hence routinely contend with efficiency issues, not least scale economies and cost optimisation, that ordinarily determine not just how well a business performs but also whether it manages to survive at all.
In contrast, Cliff (1998) argues that women-led enterprises espousing guardedly managed or deliberately suppressed growth should out-survive firms pursuing high growth but risky strategies. Marlow and McAdam (2013), however, observe that the fact that the overall share of women's enterprise has changed very little over the last two decades despite high entry into self-employment is indicative of 'high levels of churn'. While non-business reasons, such as return to employment in line with parenting demands, do account for some exits , it is undoubted that classical micro-economic fundamentals regarding competitive efficiencies are responsible for a substantial share of exits. The extent of the trade-offs between scale inefficiencies vis-à-vis control efficacies in determining the longevity of women-led businesses is therefore, an open question.
Although business growth is a paramount question for policy, some have suggested that growth aversion among women is so intrinsic that even 'providing courses specifically for women that are designed to foster growth may be largely a wasted effort' (Watson and Robinson, 2003: 774). In fact, Hundley (2000) finds that while male entrepreneurs pursued higher earnings, women pursued self-employment to be able to flexibly work around household (caring) demands. Thus, business growth may not be a priority or indeed, an objective for women entrepreneurs.
While this may also be said of many male business owners (Marlow and McAdam, 2013), Hundley's (2000) finding that more ambitious women tend to favour organisational employment over self-employment while more ambitious men elect entrepreneurship over salaried work suggests fundamental differences with significant effects on enterprise outcomes. Other research, however, finds that women entrepreneurs are as motivated and growth-driven as their male counterparts (Fischer et al., 1993;Gundry and Welsch, 2001), including in high-technology sectors (Marlow and McAdam, 2013), which suggests that barriers to expansion may be external. Whether there are significant gender differences in entry motives and growth aspirations, and the nature (internal or external) and causes of the apparently suppressed business expansion ambitions among women remain unsettled.
Be that as it may, women's participation in the economy is traditionally limited to highly gendered and under-valued sectors and occupations -a situation that has remained a persistent challenge to gender equity within employment and in society more generally (Grimshaw and Rubery, 2007;Perrons, 2009). According to Grimshaw and Rubery (2007), under-valuation of women's economic activities may be summarised under five Vs: visibility -limited sectoral diversity; valuationendemic cultural deprecation of women's work; vocation -the association of women's work with 'natural talents' as well as socialised or affective dimensions as mothers and carers rather than professional skills; value-added -women are more likely to work in low value-added and high labour intensive areas (e.g. care for children and the elderly) with little scope for the use of technology to enhance productivity and variance -the perception of women's work and work patterns (due to household demands) as fundamentally different from men's.
A key explanation of female under-participation and under-performance in enterprise is that women's employment trends are reproduced in self-employment (Marlow, 1997). In a debate described as 'the female under-performance hypothesis' (Du Rietz and Henrekson, 2000), studies have sought to explain the profiles of women-owned enterprises arguing that the performance potential of women's businesses is constrained by specific socio-economic influences which position their firms in particular gendered spaces (Marlow and McAdam, 2013). This limits the markets for women's enterprise in terms of both sectoral diversity and the returns accruable to women's enterprises. Indeed, the Organisation for Economic Co-operation and Development (OECD) (2002) estimates that 75% of women are concentrated in just 19 highly feminised sectors (more than 70% female employment) out of 114 sectors, compared to 30 sectors for men. UK data from the ONS similarly illustrate the prevalence of highly feminised sectors as the location for much female self-employment; 43% of selfemployed women were found in just 13 sectors, the largest of which were child-minding, teaching, cleaning, retail and hairdressing. In contrast, the construction sector accounts for 2% of female selfemployment compared to 27% of male self-employment (Causer and Park, 2009).
Remarkably, information and communication and financial and insurance services account for about 23% of both male and female self-employment, suggesting that more diversified increases in female self-employment may be underway. The growing number of women entering the liberal professions such as accounting, law and medicine similarly has the potential to lead more women into self-employed private practice in higher value-added sectors (Marlow and Carter, 2004). The extent to which these successes in diversifying women's careers are translating to the enterprise scene makes for an interesting area for research. Indeed, while discrimination in the work-place may lead women into entrepreneurship, the broader market, not least for finances, may not be gender neutral as the unsettled financial discrimination question suggests.

Management
Within recent studies examining the performance of women-led enterprises, much of the variability in performance has been attributed to resource endowments and business structure (Birley et al., 1987;Du Rietz and Henrekson, 2000;Johnsen and McMahon, 2005;Robb and Watson, 2012;Watson, 2002;Watson and Robinson, 2003;Zolin et al., 2013). Women entrepreneurs have been found to have less managerial experience than their male counterparts (Boden and Nucci, 2000;Brush, 1992;Zolin et al., 2013), to use 'less than optimal or perhaps "feminine" management practices or strategies' (Ahl, 2006: 603) such as bounded growth (Cliff, 1998) and to engage in sectors where advanced managerial skills and competencies are not required (Marlow and McAdam, 2013). Here, as is the case with finance (Wu and Chua, 2012), the issue that strongly emerges is the impact of gendered structure. Rather than under-performance as such, Marlow and McAdam (2013) argue that women's enterprise is constrained by strong socio-economic influences that funnel women into unpropitious sectors, and since this is not a question of managerial competency, a focus on education and similar human capital enhancements fails to address the fundamental issue. Indeed, while Hundley (2001) finds that women without children earn more than men suggesting that parenting is a pivotal determinant of economic performance, Gorman and Kmec (2007) argue that women face stricter performance standards and have to, and do indeed, try harder.
Beyond this complex debate, Bird and Brush (2002) have developed a gendered perspective on managerial practice. Since this concept of gender is a set of abstract constructs and continua, it may not correspond with dichotomous biological sex, hence, perhaps, why no latent gender effects have been observed after accounting for structure (see also, Watson, 2012). There may be scope, yet, for empirical research to explore these perspectives, not least qualitatively. In addition, while the debate on women's participation and contributions in corporate boardrooms has been vibrant (Terjesen et al., 2009), research into managerial dynamics within jointly owned (male and female) enterprises and how such firms perform relative to male-or female-owned businesses has not been forthcoming.

Conclusion: policy implications and questions for future research
That the recent efforts towards boosting EMW's enterprise have only returned modest changes suggests that policy in this area remains under-developed and problematic (Arshed et al., 2014). In reviewing the existing literature, this article highlights some of the loose ends that have contributed to the obscurity and tensions surrounding EMW's enterprise research and policy. Among these, we identify four key issues and suggest implications for entrepreneurship policy and practice as well as directions for further research. These are, first, whether discrimination is real or merely perceived; second, quality versus quantity within EMW enterprise; third, market failure and government failure in upholding EMWs and their enterprises and finally, whether EMWs are different 'in kind' or 'in degree' and therefore whether they require specialist attention.
The question of discrimination remains relevant. Research on access to finance has contributed to the emerging consensus that the divergent experiences of EMBs are attributable to business reasons rather than direct discrimination (Fraser, 2009). Nevertheless, the perception of unequal treatment continues to linger, both in finance and in other industrial markets more widely. Studies elsewhere also find residual gender and ethnicity differences and attribute these to discrimination (Alesina et al., 2013;Blanchflower et al., 2003). Indeed, discrimination remains an important concern for finance institutions and policy in the United Kingdom, and dissecting social inequity is a priority area for ongoing research in a variety of disciplines. The lack of accurate data and precise methodologies for identifying discrimination preclude any definitive statement. More focused studies of the demand and supply-side experiences would be valuable. The former could include more focused investigation of minority firms (and a control group) at the point where loan applications are made. Similarly focused examinations of the supply-side would also be appropriate. This should take the form of case studies of the banking process at the point where applications are made by EMBs (Ram et al., 2002).
Besides perceived discrimination, EMW enterprise is characterised by systematic differences within structural factors that are pivotal to enterprise fortunes, including size (Watson and Robinson, 2003), sector (Marlow and McAdam, 2013;Ram and Jones, 2008) and geographical location (Fraser, 2009). Recent research suggests further that significant second-order gender effects are also detectable (Wu and Chua, 2012). A closer empirical examination of the relationship between ethnicity, gender and business structure, in particular the extent to which gender and ethnicity may be coterminous with business structural dimensions, would not only help alleviate persistent concerns regarding access to and cost of borrowing but would also contribute to the development of more discerning policy interventions with higher likelihood of success. For example, younger and smaller EMW enterprises operating in problematic sectors and postcodes may be advised, including by declining banks, to seek finance from community development finance institutions that may be able to offer both friendly finance and business counselling.
The second key issue pertains to the quality and quantity of EMW's businesses. EMW enterprise policy has been marked by a contradiction between the quantitative expansion of businesses that may be undermined by the low quality of a substantial share of EMW enterprises in respect of profits, turnover and other performance indicators. Within the 'under-representation' debate, simplistic policy invocations to become self-employed should be questioned. Certainly, if necessity entrepreneurship, emanating from labour market discrimination for example, drives a large number of EMWs into self-employment, then high levels of the resultant low-quality entrepreneurship among EMWs should not be viewed as an unqualified indicator of upward social mobility. In fact, encouraging further entry into the low quality sectors that most EMW enterprises are already crowded in may actually be counterproductive (Marlow and McAdam, 2013;Ward, 1985). The US experience of focusing on encouraging well-credentialised minority entrepreneurs in high valueadded sectors (Bates, 2011) highlights a quality versus quantity entrepreneurship angle that may be instructive to UK enterprise policy.
Importantly, within the under-representation debate, the quantity of EMW enterprises may be hard to properly establish. Indeed, a focus on women-owned and male-owned enterprises, a label not often used by entrepreneurs themselves, masks the substantial role of different household members in supporting both male-owned and female-owned enterprises. In particular, female spouses are understood to make significant but silent contributions to family firms which would normally count as male-led. Although businesses and households have been traditionally regarded as separate spheres, there is growing realisation that the two institutions are inextricably linked and compelling arguments to embed entrepreneurship research within the context of the family (Carter, 2011;Jennings et al., 2013). Cases of joint male and female partners are increasing, and developments elsewhere in entrepreneurship research contend that identifying firms with single individuals and crediting them with enterprise outputs epitomise attribution errors (Dimov, 2007). There is much scope, therefore, for further theoretical elaboration and empirical research in this area towards a more sophisticated account of the participation and contribution of households in enterprise.
Indeed, it is appreciable that entrepreneurship requires distinctive skills, aptitudes and indeed preferences, such that agents otherwise productive as employees may not be productive as the designated entrepreneurs. By the same token, whether policy interventions may succeed in changing individuals and cultures and therefore, boosting enterprise outcomes for women has been questioned (Marlow and McAdam, 2013;Watson and Robinson, 2003). These constitute the third key issue that pertains to market failure and government failure. Market failure may emanate from issues related to discrimination and other biases where market dynamics are unable to eliminate EMW disadvantage and may even reinforce such disadvantage further. On the other hand, policy attempts to alleviate the disadvantage may prove unsuccessful and hence, constitute 'government failure' as would the non-attempt to correct market failure that then allows the disadvantage to persist.
There clearly is an imbalance between supply and market demand, with much of EMW enterprise overcrowded in easy-to-enter, low value-added market sectors. Within gender, discussions about how best to support the growth of women's participation in enterprise and the existence of gender-based market failure have been largely conflated with much wider debates about the role of women in society and in the economy. These debates encompass relatively modern concerns about the need to introduce gender-based quotas on the boards of publicly listed companies, the 'leaky pipeline' of women in Science, technology, engineering and mathematics (STEM) industries, as well as the persistent discussions about the relative social importance of economically inactive, stay-at-home mothers. The conflation of complex and overlapping issues has been unhelpful in delineating a clear view of the existence of market failure in women's entrepreneurship. The paucity of research is similarly unhelpful in this regard.
While some studies suggest that starting one's own business provides individuals with the means to avoid work-place discrimination and glass ceiling thresholds, more critical studies argue that the context of female self-employment simply reproduces gendered outcomes coupled with relative impoverishment. Furthermore, the tendency of EMW enterprises to employ co-ethnics and women may do little to eliminate wider inequality and may also engender 'divorced diversity' as opposed to 'diffuse diversity'. The lack of a clear evidence base has prevented systematic analysis of the extent of such market failures and how best they may be addressed.
Policy interventions may unwittingly exacerbate market failures. For example, encouraging ethnic minority start-ups may only increase enterprise within ethnic enclaves and therefore, not contribute to a social integration agenda. Ethnic minorities may also not take up business support to the expected extent for a variety of reasons: for instance, they may fear contact with official agencies, particularly if they are concerned about their immigration status as in the case of refugees and asylum seekers, or such provision does not exist in their country of origin and they do not see the relevance of the provision to them (Mascarenhas-Keyes, 2008). Another source of government failure is where the various barriers to minority enterprise are significantly different across ethnic groupings as the present review has revealed. While future research should pursue more granularity to unpack differences, policy may encounter practical and political difficulties in not only customising support to the various groups but also identifying and targeting some minority groups over others for support. Ram et al.'s (2013) study of a state support agency attempts to get to grips with the phenomenon of 'superdiversity' hints at the problems of providing a coherent policy towards and an ever more differentiated set of 'minority' entrepreneurs.
It is important to ascertain that there is a market failure that justifies the provision of specialist business support for ethnic minorities and for women. This leads to the fourth issue -whether EMW enterprise requires specialist attention. There has been a dilemma between the use of mainstream enterprise policy and specialist interventions particular to EMW entrepreneurs. For minorities, ethnic managerialism may be fraught with difficulties and likely to have unintended consequences leading to new forms of exclusion (Law, 1997). On the other hand, the 'ethnic blind' mainstream approach ignores the unique challenges minority enterprises face which may leave them unable to benefit from interventions. Changes introduced by the present coalition UK Government with regard to publicly funded business support institutions may provide a unique opportunity to study the differences in the various interventions and their effects on EMW enterprise (see, for example, Ram et al., 2012b).
Discussions about mainstream or specialist provision for female start-ups and women-led businesses centre on two issues. First, there is well-founded concern about the extent to which women wish to engage with specialist institutions. Many women view the prospect of women-only business support mechanisms with scepticism, although evidence suggests that single sex business coaching may be highly effective (Fielden and Hunt, 2011). Second, there is an equally wellfounded concern that mainstream support, particularly support focused on businesses identified as high-growth firms, in practice excludes women-led businesses as relatively few meet inclusion thresholds of selective programmes and initiatives. In this regard, gender-blindness may be disadvantageous to women-owned firms which are typically smaller in scale.
Enterprise research itself also faces its own dilemmas with regard to how it studies ethnic minorities and women's enterprise. While the emergence of intersectionality has so far gained little traction in entrepreneurship research (or policy), it has become a major paradigm particularly within women's studies (McCall, 2005;Walby et al., 2012). The benefits of intersectional approaches are clear given that the separation of research effort focusing either on ethnic minority or on women's enterprise has given little voice to the experiences of entrepreneurial groups at the intersection of two or more identity categories. While the methodologies of intersectionality are still emerging (Walby et al., 2012), the approach is likely to prove fruitful to future EMW studies. A similar observation can be made to the conceptual approach of 'mixed embeddedness' which is rather more established in the field of ethnic minority and immigrant business research (Jones et al., 2014;Kloosterman, 2010;Ram et al., 2013). The concept has been proved fruitful in sensitising researchers to the importance of the political, economic and regulatory contexts in which firms are embedded, as well the more usual concern with the social ties and resources of minority entrepreneurs. But mixed embeddedness too needs be more attentive to importance of racism, the agency and of minority entrepreneurs and workers, and gender (Jones et al., 2014).
In their proposed framework for women's entrepreneurship research, Brush et al. (2009) explore whether a new theory is required to understand women's enterprise or whether mere expansions of extant theory should suffice. Indeed, SMEs of various backgrounds have been found to be broadly similar (Irwin and Scott, 2010;Marlow and McAdam, 2013). Differences within the groups may be more significant than differences between the groups, thereby rendering the differentiation of enterprises by gender or ethnicity unnecessary or erroneous (Ahl, 2006). However, that many studies still find residual effects that may be solely attributable to gender and ethnicity (Alesina et al., 2013;Blanchflower et al., 2003;Borooah and Hart, 1999), as well as significant second-order effects (Wu and Chua, 2012), and that gender (Bird and Brush, 2002) and ethnicity (Basu and Altinay, 2002) are more complex than presently understood or typically operationalised only underscore the fact that there is much scope for theory development and empirical research into EMW's enterprise within the larger domain of entrepreneurship. Perhaps one way of addressing this question is to pay attention to 'metatheory', which relates to the philosophical assumptions that underpin the practice of research. Useful lessons can be learnt from the cognate field of organisation and management studies, which has extended its reach by embracing a proliferating array of topics, modes of theorising and methods. This pluralisation raises questions about what is considered 'good' or 'acceptable' research (Cunliffe, 2011). Explicating the philosophical and political context of organisational (and minority entrepreneurship?) research is an important means of accommodating diversity and producing credible research (Amis and Silk, 2008;Easterby-Smith et al., 2008;Leitch et al., 2010). In short, metatheoretical positioning will have practical consequences for the conduct of research (Cunliffe, 2011). Isolated studies of minority entrepreneurs that are sensitive to such philosophical matters are beginning to emerge (Jones et al., 2014;Scholin et al., forthcoming); they suggest that greater attention to metatheory might be helpful in grasping the 'diversification of diversity' (Vertovec, 2007) that increasingly characterises research on ethnic and women-owned businesses.