Gender And Ownership In Uk Small Firms

1042-2587-94-183$1.50 Copyright 1994 by Baylor University

Gender and Ownership in UK Small Firms Peter Rosa Daphne Hamilton The emphasis on the individual "female entrepreneur" in much of the small business literature in the last decade disguises the fact that many women in business ownership are in partnership with others, usually with men. How "gender" impinges on the process of small business ownership has been littie studied. The paper examines gender and ownership using evidence from a three-year study on the impact of gender on smail business management, involving interviews with 602 male and female UK business owners, drawn from three industrial sectors. Difficulties were encountered in interpreting sex differences as "gender" trends, owing to significant sectorai variation. Nevertheless, some marked gender differences were identified. These referred to differential patterns of kinship with the respondent; the allocation and perception of specialist roles within the business; and the fact that female owners are less likely to be associated with more than two businesses. Overall sole traders were in the minority in both sexes, implying that most owners shared responsibility and management in some way with other owners. The paper concludes with methodologlcai implications of co-ownership for the sampling and analysis of small business owner/managers from a gender perspective.

i n July 1986 an article appeared in the business magazine Venture, entitled "Women Entrepreneurs, the new business owners." It reported True, women entrepreneurs are rare in a wide field of men. But startups have swelled their number so that now the woman who heads her own company is no longer the exception. Indeed for an ambitious and capable woman, entrepreneurship is a sensible choice, devoid of conflict between her professional and working life (p. 33). Implicit in this quotation is the view of the female entrepreneur as an individual owner, pioneering new opportunities in a previously male-dominated area of economic activity, and in control of her own enterprise. This theme is implicit in much of the academic small business literature. Solomon and Femald (1988, p. 24) estimate that there are some three million female-owned businesses in the United States. Their article focuses on the characteristics of the "new phenomenon" of the female entrepreneur (p. 24). There are many other similar studies emphasizing female individualism through identifying the characteristics ofthe female entrepreneur (for example Carter & Cannon, 1991; Brush, 1988; Birley, Moss, & Saunders, 1987; Cromie & Hayes, 1988). This preoccupation with the female entrepreneur is rooted in the debate of whether proprietorship represents an opportunity of liberation for women, or whether it is yet another form of subordination of women in an exploitative capitalist system in which proprietorship "promises autonomy, but in fact offers serfdom" (Bechofer, quoted by Goffee & Scase, 1983. p. 641). Spring, 1994

11

The positive role of proprietorship for women relies on the idea that in the capitalist system, members of economically and socially deprived groups can " e s c a p e " deprivation through business ownership, which provides opportunities for self-determination through owning and controlling resources, and through an increased ability to flexibly interface work and domestic life (Goffee & Scase, 1983, p. 626). By exploiting social and economic conditions in the 1970s and 1980s it has been easier for women to aspire to self-determination through proprietorship. Prominent, successful female entrepreneurs have not only bettered their own lives, but have made a dramatic impact on the perception of women in society as a whole. As Goffee and Scase put it, "female entrepreneurs have a symbolic importance which implicitly questions popular conceptions of the position of women in society" (p. 627). On the negative side, proprietorship could be seen as supporting capitalist institutions which "sustain the dominance of men over women"; and as reinforcing economic deprivation of women in employment, from whose ranks the entrepreneurs often come (Goffee & Scase, 1983, p. 627). Moreover, as these authors show in their analysis of case studies, subordination does not necessarily cease with proprietorship. The influence of husbands and male relatives can undermine the quest for self-determination. Many female proprietors may feel that they "have failed to obtain full legitimacy and credibility as proprietors" as a result of experiences in dealing with bank and businesssupport officials, customers, suppliers, and clients. A researcher's views on this debate may be fundamental to the way in which "ownership" is treated in a study. A belief in the benefits of proprietorship for women tends to stress "individuality." The struggle for self-determination through proprietorship is a personal one, to be undertaken alone. It is an exploration of "self" rather than a competitive " g a m e . " Likewise business success is inseparable from the achievement of the female entrepreneur who owns and manages the business. The rewards are the consequences of her struggle and achievement. This line of argument tends to be undermined by co-ownership, especially if the co-owner is male. Consequently there tends to be little consideration of this comphcation in the literature where the female entrepreneur is the focus of interest. If a researcher takes a less unidimensional view of proprietorship, then the concept ofthe "lone female entrepreneur" is perhaps too limited. A woman in business is not an island, even if a sole legal owner, and cannot "escape" from the wider society. The social conditions that reinforce gender inequalities may impinge on her business and personal life at several levels (Hamilton, Rosa, & Carter, 1992). These may wel! be manifest in the types of relationships she has with others in both the business and domestic domains. In this scenario, the analysis of co-ownership is vital in our attempts to study gender from a more holistic perspective (Hamilton, 1990). The preoccupation with "individualism" in entrepreneurial and small business studies has disguised the potential complexity of small business ownership, a complexity compounded by the influence of gender. Large numbers of small businesses are partnerships. Partnerships may comprise: Married couples; Non-married couples; With relatives; With nonrelatives; With relatives and non-relatives; Single sex; Mixed sex. Where partnerships comprise both men and women, gender roles and relationships can be potentially extremely intricate. Sole traders are not necessarily on their own either. Many have domestic partners that contribute significantly to tbe business even though they may not be legal co-owners (Goffee & Scase, 1983; Meijer, Braaksma, & Van Uxem, 1986; Meittinen, 1986). Even single people with no domestic partner contributing to the business may have father, mother, sisters or brothers contributing. The question of "ownership" is thus fundamental in any study of gender and small business management. It is an area we know little about, because previous studies, by 12

ENTREPRENEURSHIP THEORY and PRACTICE

focusing on "female entrepreneurs" alone, have tended to avoid the issues raised by co-ownership. By focusing on "females" or "women," moreover, males have been relegated to a functional research role as "controls," or represented as examples of gender antagonists (e.g., the interfering or hostile husband). Consequently even less is known about male business ownership, and the way women and men interact as business co-owners. Because so little is known, we regard it as important to find out what small business ownership empirically looks like. Until this is done, it is perhaps premature to address theoretical debates in detail.

MATERIALS AND METHODS Ownership is one element in a study of gender in small business management drawing to its close at the University of Stirling. Over six hundred owner/managers have been interviewed, using a structured questionnaire comprising over 100 questions (some 650 variables). Each interview took approximately an hour. The sample is a random statistical "quota" sample, half male and half female. It was also stratified by sector. One third are from the textile and clothing sectors; one third from business services; and one third from hotel and catering. The sectors were limited to enable us to control for sectoral effects in evaluating gender relationships. For contacted businesses that had more than one owner., procedures were devised and implemented to randomly determine whether a male or female owner should be interviewed. In co-owned businesses, only one owner/manager was interviewed. Included in the questions on ownership were details on other owners in the business (sex of other owner; when met; whether they are active; their managerial roles). Similar questions were also asked about owners that had left the business. Identifying gender relationships using statistical analysis is a deceptively simple process. A "gender" relationship is usually deemed to exist if a statistically significant difference between male and female respondents is found for a given variable. All this really tells us, however, is whether a statistically significant "sex difference" exists. A sex difference may indeed correspond to a genuine "gender" difference, that is, one which is attributable to the action of socially determined forces that differentiate the sexes. Alternatively, a significant sex difference may merely reflect the influence of specific economic forces. The most important of these are sectoral factors (see Hamilton, Rosa, & Carter, 1992). It is possible that most apparent gender differences at an aggregate level may be explicable in terms of underlying trends of sectoral variation. To give a hypothetical example, at an aggregate level women could appear to be significantly more likely to own a business in partnership with a spouse than men are. A universal gender explanation is implied. If, however, female respondents are clustered in sectors where married couples are a dominant form of ownership (for example hotel and catering), and men are not, then the difference at the level of the total sample merely reflects the sampling imbalance, rather than a genuine gender-related trend. Our decision to obtain an equal quota sample for three industrial sectors has reduced the sectoral problem. Unfortunately, however, there is still considerable variation in markets within each of our chosen sectors (for example a respondent owning a business in the business services sector could be a solicitor, accountant, employment agent, a provider of typing services and so on). The analysis that follows, therefore, will consistently examine sex differences in relation to sectoral trends. In spite of procedures to maximize randomness in selecting quotas, employing a quota sample still means that the sample is not randomly drawn in a strict sense. Results and statistical tests of significance cannot be interpreted in the conventional way, as Spring, 1994

13

population estimates (Moser & Kalton, 1971, pp. !27fO. In this study, however, we are not trying to estimate population parameters. We feel that any such attempt is (arguably) meaningless in the context of small business proprietorship, due to the considerable sectoral variation just outlined. The probability of any two geographical areas having the same sectoral blend of businesses, and of key personal factors such as social class origins, ethnicity, and capital ownership negated any attempt at matched sampling too. Rather, the onus is on deciphering the nature of relationships with gender in tbe business context. Tests of significance are used instead as surrogate guides for assessing strengths of relationship. Nevertheless our confidence in the approximate representativeness ofthe sample obtained was considerably enhanced on observing consistent similarities when statistical results were compared between those based on the first 340 respondents and those based on the remaining 262. Statistical significance is tested through chi square. Tbis is usually based on the overall table. Where a significant "separate effect" is recorded in a table, this means that the chi square was performed for that row versus all other categories pooled. RESULTS

,,

The results summarized in Table 1 indicate that a majority of businesses were co-owned (61%), and that a significant minority of businesses (41%) were of mixed-sex ownership. Both male and female single-sex partnerships were relatively uncommon. The most interesting finding is the rarity of all-female co-owned businesses (4.7% compared to 14% for all-male co-owned businesses). If we consider businesses with three or more owners, none was all female, compared to 5% all-male owned. The absence of all-female multiple-owned businesses may reflect the low concentrations of more senior women managers in corporations in the UK (Davidson & Cooper, 1992, p. 13). Many businesses with three or more male owners are started by corporate managers who leave together to start a business. Some may occasionally include a female manager in the partnership. The chance of getting three or more female managers all together within one firm, department or branch, and if so, willing to leave to start a business, is rare indeed. Another explanation may be the absence of any tradition of bringing together female family members to run a business. We have yet to encounter a "Mary Smith and Daughters, accountants" for example. Even male parterships, however, are much less frequent than mixed-sex partnerships, which, as Table 1 shows, form as much as 4 1 % of the sample. Table 2 shows the distribution of ownership by legal type. The results again confirm that single ownership is in the minority in each sex. Distributions are surprisingly similar by sex in the total sample and in two ofthe sectors examined. Only in one sector, textile and clothing, are sex differences significant, and displaying the more usual picture of sole trading being more common in women proprietors, and limited companies more common in men. This is specified by Brush (1992, p. 13) who states that "comparable to male-owned businesses, women most often choose sole proprietorship as the preferred form of business structure." The fact that the proportion of sole ownership can vary appreciably from one study to another is interesting. One explanation may be that industrial sectors vary in the proportion of businesses that are sole trading. The proportion of sole traders in a study will thus reflect the sectoral mix of that study. Another reason may be that sole trading may be artificially high due to the fact that a sample chosen is non-random. A recent study by Challis (1992), for example, also from the UK, reports a much higher percentage of small traders than in the current study (close to 60% for each sex). Challis's 14

ENTREPRENEURSHIP THEORY ond PRACTICE

Table 1 Distribution of Ownership Categories by Gender ALL SECTORS % ONE OWNER: Female Male TWO OWNERS: Both female Both male Mixed sex THREE OR MORE OWNERS: All female All male Mixed sex

19.6 19.6

4.7 9S 29.4 0.0 12.2 100.0

TOTAL

sample, however, was drawn predominantly from very small businesses with selfemployed owners that had gone to local support agencies for help. He also covered all sectors. Sectoral and sample variation in the incidence of sole trading may thus indicate that it is not so much that women prefer sole trading, hut that in sectors where women are commonly self employed, sole trading is the dominant form of business ownership. It is equally possible that men in those types of businesses will also prefer sole trading.

Table 2 Legal Status of Business by Sex of Respondent Textile/Clothing Business Services Hotel and Catering ALL SECTORS

Sole trader Partnership Ljnijied company TOTAL CHI SQUARE (p)

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

49.0 31.6 19.4 100.0 98

37.6 25.7 36.6* 100.0 101

36.0 36.0 28.0 100.0 100

38.6 32.7 28,7 100-0 101

34.3 49.5 16.1 100,0 99

33.0 50.0 17.0 100.0 100

39.7 39.1 21.2 100,0 297

36.4 36.1 27.4 100,0 302

0.03

0.73

0.99

0,42

* Separate Effect Significant: «.OO1. Figures do not itemize 3 "others." franchises, cooperatives.

Spring 1994

15

Becoming an Owner Only 61% of the sample overall started from scratch (Table 3). A sizeable minority (about a quarter) bought or bought into the business. There is some marked sectoral variation, with owners from the hotel and catering sector being much more likely to have bought a business than those from the other sectors (true for both sexes). This again is an example of the peculiar conditions of specific sectors, which erode any gender effects, and which complicate our ability to interpret the impact of gender. Significant differences by sex were only found in the textile/clothing sector. Though the overall differences were non-significant, separate effects showed that men were significantly more likely to have inherited the business or bought the business, while women were significantly more likely to have started one. Other data from the survey show that in this sector in particular, female businesses tend to be younger, and smaller, which may account for these differences.

Other Owners Those respondents whose businesses were co-owned were asked to provide details of up to six other owners. The number of other owners whose details were obtained totalled 540. Table 4 demonstrates that about half the sample of co-owners were associated with one other only, but that the proportion associated with three or more is still considerable, at about a quarter. Table 4 shows highly significant differences by sex in the total sample and in two ofthe three sectors. The genera! trend is for female owners to be more commonly associated with only one other owner, and male respondents with four or more other owners. These trends are most marked in business services, where traditional financial partnerships are more common amongst male respondents.

Table 3 Mechanism of Business Entry by Sex of Respondent Textile/Clothing

Inherited firm Bought firm Bought into firm Married into firm Started up nrm Other TOTAL CHI SQUARE (pl

Business Services Hotel and Catering

ALL SECTORS

Women

Men

Women

Men

Women

Men

Women

Men

1.0 3.0 4,0 4.0 87.9 0.0 100.0 100

12,0*' 12.0* 3,0 3.0 68.0" 2.0 100.0 101

3.0 8.0 7.0 5.0 68.0 9.0 100.0 100

2.0 4.0 9.9 0.0 74.3 9.8 100.0 101

4.1 52.0 8.2 1.0 31.6 3.1 100,0 98

6.1 43.4 8,1 1,0 37.4 4.0 100.0 99

2.7 20.9 6.4 3.4 62.6 4.0 100.0 297

6.7 19.7 7,0 1.3 60.0 5,3 100.0 300

0,18

0, 19

0.88

O.I

Separate Effect Significant: ^.01 >.OO1. * Separate Effect Significant: =£.001.

16

ENTREPRENEURSHIP THEORY a n d PRACTICE

Table 4 Number of the Other Owners by Sex of Respondent

one other iwo others 3-4 others more than 4 oihers TOTAL CHI SQUARE (p)

Textile/Clothing

Business Services

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

56.0 !8.7 25.3 0,0 100.0 75

44.9 24.7 24.7 5.6100.0 89

54.1 16,5 17.6 11.8 100.0 85

26.1-" 26.1 18.3 29.6-" 100.0 115

56.5 13.0 25.0 5.4 100.0 92

58.3 26.2' 15.5 0.0* 100.0 S4

55.6 15.9 22.6 6.0 100.0 252

41.3*** 25.7'* 19.4 13.5" 100.0 288

0.12

Hotel and Catering ALL SECTORS

0.00

0.01

0,00

* Separate Effect Significanl: =s.O5 > . O I . •* Separate Effect Significant: ^ . 0 1 >.OO1. *•* Separate Effect Significant: =£.001.

Distribution of Sex of Other Owners Table 5 shows that the sex of other owners is not distributed randomly when associated with the sex of the respondent. For both sexes other owners are predominantly male, though male respondents are much more likely than females to be associated with another female owner. This overall trend is replicated in the hotel and catering sector and the textile and clothing sector. In the case of business services, however, there is no significant difference at all between the distribution of other owners in regard to sex. It is unclear why business services should stand out in this way.

Table 5 Sex of Other Owners by Sex of Respondent Textile/Clothing

Women Men TOTAL CHI SQUARE (p)

Spring, 1994

Business Services Hotel and Catering

ALL SECTORS

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

24.0 76,0 100.0 75

39.5 60.5 100,0 86

22.4 77.6 100,0 85

21.7 78.3 100.0 115

24.1 75.9 100,0 87

48,2 51.8 lOO.O 83

23.5 76.5 100.0 247

34,9 65.1 100.0 284

0.04

0.92

0,00

0.00

17

Original Founder by Sex of Other Owner Two-thirds of the sample of other owners were original founders (Table 6), a trend which surprisingly differs little between the sexes overall. This can be related back to Table 3, where the proportion marrying into the business is also surprisingly low for women as well as men. This may mean that more women and men start businesses together than we might expect from conventional wisdom, where men are routinely thought to be the founders, and women later entrants either through marriage or to support their husbands. Significant differences are lacking in two of the three sectors. In the textile and clothing sector women are more likely than men to have been original founders. This may again be related to the greater age of businesses owned by male respondents in this sector.

The Relationship of Other Owners to the Respondent Women respondents are much more likely to be related to one or more of the other owners in their business than men are (Table 7). This variable, however, shows enormous sectoral variability. Over three-quarters of the respondents in the textile and clothing and hotel and catering sectors are related in some way to other owners, in contrast to business services, where the proportion of respondents related to other owners is appreciably lower, especially in men. In contrast to the overall result, only one of the sub-sectors, business services, shows a significant difference in the degree of relatedness between the sexes. In business services, females are more likely than males to be related to another owner. Table 8 explores the nature of the relationship of related other owners to the sex of the respondent. In all sectors it is the domestic partner which tends to be the dominant category of relationship. In all sectors, too, it is the female respondent who is much more likely to be related to a domestic partner than male respondents. This relationship is significant overall and also in the textile and clothing sector. This means, therefore, that women are much less likely to be associated with other cateogories of relatives than men are. This is particularly evident in certain categories of kinship. In the textile and clothing sector, for example, nearly a fifth of male respondents have other owners who are their children, usually or almost invariably sons rather than daughters, whereas women respondents have hardly any examples of children being co-owners in their firm.

Table 6 Whether Original Founder by Sex of Other Owner Textile/Clothing

Yes No TOTAL CHI SQUARE {p)

18

Business Services Hotel and Catering

ALL SECTORS

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

78.8 21.2 100.0 52

62,0 38,0 100.0 108

54.5 45.5 100,0 44

56.6 43.4 100.0 152

67.8 32.2 100.0 59

78.3 21.7 100.0 106

bl.l 32.3 100.0 155

64.5 35.5 100,0 366

0.03

D.KI

0.14

(1.47

ENTREPRENEURSHIP THEORY a n d PRACTICE

Table 7 Whether Related to Respondent by Sex of Respondent Textile/Clothing Business Services Hotel and Catering ALL SECTORS

Yes No TOTAL CHI SQUARE (p)

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

77.3 22.7 100.0 75

82.1 17.9 100.0 84

59,3 41.7 100.0 84

28.1 71.9 100.0 114

82,6 17.4 100.0 86

74.0 26.0 100.0 77

72.7 27.3 100.0 245

57.5 42.5 100.0 275

0,45

0,00

0,19

0.00

In all three sectors as well as in the overall sample siblings are relatively common as other owners where the respondent is a male, and extremely uncommon where the respondent is female. Breaking down siblings into brothers and sisters showed that nearly all the relationships mentioned happen to be brothers rather than sisters. Of note, too, referring to the textile and clothing sector, is the fact that female respondents are more likely than male to have other owners who are related outside the nuclear family. This probably reflects the influence of in-laws. This is not surprising given that a male domestic partner is much more likely in this sector to have relatives drawn from his nuclear family. On marrying the person, the woman obviously acquires

Table 8 How Related to Other Owner by Sex of Respondent Textile/Clothing Business Services Hotel and iCatering ALL SECTORS

Parents Children Siblings Other relative Domestic partner TOTAL CHI SQUARE (p)

Women

Men

Women

Men

8,6 1.7 8.6 17,2 63.8 100.0 58

17,4 I4.5*'' 21.713.0 33.3*** 100.0 69

22.4 0.0 6.1 2,0 59.4 100.0 49

21.9 0.0 12.5 6,2 65.4 100,0 32

)

0.71

Women

16.9 n.3 1.4 5,6 64,8 100,0 71 0,0f

Men

Women

Men

17,5 7,0 15.8**' 5.3 54.4 100.0 57

15.7 5.1 5.1 8.4 65.7

18.4 8.9 17.7*** 8.9 46.2*** 100.0 158

100.0 178 O.(K)

* Separate Effect Significant: 'S.OS >.O1. * * Separate Effect Significant: ^ . 0 1 >.OO1. * * * Separate Effect Significant: 'S.OOl.

Spring, 1994

19

these relationships with marriage. Tn the business services sector, tbe fact that relatives are much less common than other sectors reflects the professional nature of many of these businesses. It is interesting that proportionally far fewer female respondents have relatives in the business services sector other than domestic partners. This may reflect the structure of ownership of many of the business services' professional practices, where sons and fathers are often co-owners of the business and daughters seldom are. For example, a firm called Hogg, Hogg and Hogg, Solicitors, would usually refer to three male Hoggs. Given the rarity of involving daughters in such practices, it is almost inevitable that the only way a woman can enter them through a kinship route is through marriage. This is clearly reflected in Table 7.

Active Management/Business Roles Table 9 shows that male other owners are more likely to be active managers in all sectors (significant in business services). More surprising is the fact that three-quarters of female other owners were active managers too, a much higher proportion than we expected. Differences by sex of owner were also small overall when the respondents were asked whether each owner had a specialist role in the business (Table 10). Sex differences within sectors were also non-significant, though men in each sector were consistently more likely to have a specialized role. If a specialist role was indicated for an owner, the respondent was asked to itemize what the role actually was in an openended question. An impressive variety of answers was obtained, which have rendered meaningful statistical analysis by gender difficult due to low cell numbers. An example of one gender trend, however, is discernible in the figures. A specialist role in the hotel/catering sector is cooking food. There are two labels used to describe this role in English. A person doing the cooking is either the "cook" or the "chef" The word "chef" conveys considerably more status and prestige than "cook." Table 11 demonstrates a highly significant gender difference. Of 127 respondents with a specialist role in the hotel and catering sector, 27% are "cooks" if women, but only 6% if men. Conversely, 19% of male owners are "chefs" but only 4% are women. This distinction is clearly irrelevant in other sectors where no cooking takes place as a business role. Specific business roles are thus heavily influenced by sectoral factors, but within relevant sectors can be associated with significant gender components too.

Table 9 Whether an Active Manager by Sex of Other Owner Textile/Clothing

Yes No TOTAL CHI SQUARE (p)

20

Business Services

Hotel and Catering

ALL SECTORS

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

74.5 25.5 100,0 51

86,1 13,9 100.0 108

72.7 27..1 100.0 44

87.2 12.8 100.0 156

75.0 25.0 100.0 60

68.2 31.8 100.0 107

74.2 25,8 100.0 155

81.4 18.6 100.0 371

0,07

0.02

0.36

0,06

ENTREPRENEURSHIP THEORY a n d PRACTICE

Table 10 Whether They Had a Specific Role by Sex of Other Owner Textile/Clothing Business Services Hotel and Catering ALL SECTORS

Yes No TOTAL CHI SQUARE (p)

Women

Men

Women

Men

Women

Men

Women

Men

64.6 35.4 1000 48

73,6 26.4 100,0 106

59.5 40.5 100.0 42

67.6 32.4 100.0 142

61,4 38.6

69.1 30.9 100.0

61.9 38.1

69.9 30.1 100.0 345

0,26

100.0 57

97

0.33

100.0 147 0,0B

Owners Who Have Left the Business

^

A minority indicated that at least one other owner had left the business during the time they had been involved with the ownership and management of the business (Table 12). Departed owners were significantly more common in the case of male respondents overall, and in two of the three sectors (hotel and catering was again the exception). Considerable information was obtained in the questionnaire on the characteristics of other owners who had left the firm. Most of these results will not be reported in this paper. However, it is interesting to note that whereas in Table 7 two-thirds of the other owners were related in some way to the respondent, when the other owners who had left

Table 11 "Cook" Versus " C h e f as Specialist Role in the Hotel and Catering Sector: by Sex of Other Owner ALL SECTORS Specialist role Qwf

m* TOTAL CHI SQUARE (p)

Women

Men %

% 4.1

19.2 6.4 74.4

100.0 49

100.0

78

0.001

Separate Effect Significant: ^.05 <.O1.

Spring 1994

21

Table 12 Whether Owners Have Left the Business by Sex of Respondent Textile/Clothing

Business Services Hotel and Catering ALL SECTORS

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

8.2 91.8 100.0 98

19.8 80 2 100.0 101

14,0 86.0 100.0 100

28.7 71.3 100.0 94

13.2 86.8 100.0 91

12.2 87.4 100.0 87

11.8 88.2 100.0 289

20.6 79.4 100.0 282

Yes No TOTAL CHI SQUARE (p)

0.02

0.01

0.00

0.91

were examined, only 22% were related to the respondent. This indicates that it is usually non-relatives who leave the firm. The sexes do not differ very much in the proportion of ex-owners who are related to the respondent. There is. however, a notable difference when the distribution of relatives who have left is considered. In the case of female respondents, ex-owners include father, mother, sister and former domestic partners. There is no instance of a brother or of another relative having left a firm where a female respondent was working. In contrast, only two former male owners fall into these categories. The majority of male former owners were either brothers or other relatives, categories which do not appear at all in relation to female respondents.

Multiple Business Ownership Just under a fifth of respondents overall own more than one business. Men are far more likely to own more than one business than women are (Table 13). This trend is highly significant in the total sample, and is either significant or close to being significant within each sector. Unlike most of the trends illustrated in this paper where at least

Table 13

'

'

Whether Currently Owning Another Business by Sex of Respondent Textile/Clothing

Yes No TOTAL CHI SQUARE (p)

22

Business Services Hotel and Catering

ALL SECTORS

Women %

Men %

Women %

Men %

Women %

Men %

Women %

Men %

7.5 92.5 100.0 93

17.5 82.5 100.0 97

9.3 90.7 100.0 97

22.9 77.1 100.0 96

8.9 91.1 100.0 90

184 81.6 100.0 87

8,6 91.4 100.0 280

19.6 80.4 100.0 280

0.04

0,01

0.06

0,00

ENTREPRENEURSHIP THEORY a n d PRACTICE

one sector does not follow the general trend of the total sample, all three sectors show the same type of trend in this particular case. Of those who own more than one business, there is a highly significant gender difference in the distribution of numbers of additional businesses owned (Table 14). The table shows quite clearly that men are much more likely to own multiple businesses than are women. CONCLUSIONS . If we consider just the total sample, and statistically compare the sexes, trends appear to be relatively clear. Surprisingly there were no significant differences by sex for legal ownership, on mechanisms for business entry, and for being an active manager. Femaie respondents tend to be significantly associated with just one other owner, while male respondents are much more likely to be associated with four or more other owners. Other owners are much more likely to be male for both sexes but male respondents are much more likely to be associated with women (not surprisingly, as the other owner is often a domestic partner). Women are significantly more likely to be related to other owners. This is more likely to be a domestic partner than for men, and less likely to be siblings. Although most owners were active managers, men were more likely to be active. More men tended to have a specific role, and to have left an ownership position in the business. Multiple business ownership was significantly more likely amongst men. These trends, however, are much more complex when sector is examined. The variation by sector tends to be much greater and more significant than gender, which means that gender trends are often eroded or compromised by the interaction of sectoral forces. For example, in Table 2, both sexes are much more likely to have bought a

Table 14 Number of Other Current Businesses by Sex of Respondent (of Those Owning Others) ALL SECTORS

One other Two others

Women

Men

83.3 12.5

It.9

More than two

TOTAL CHI SQUARE (p)

28.6* ioo.o 56

24 0.04

* Separate Effect Significant: ^ . 0 5 < . 0 I . Spring, 1994

23

business if they come from the hotel and catering sector. This is determined ultimately by the physical properties of hotels which form permanent stages for entrepreneurial actors to play and exit, and are rarely huilt from scratch when an entrepreneur starts the venture. If we take all the tables in this paper with a full sectoral breakdown, and allow for "separate effects" too, there is scarcely a table without some significant gender trend (see Table 2 for one sector only being significant; and Table 3 for the impact of separate effects in the light of overall non-significance). We hypothesize that a sex difference can be interpreted in terms of social structural gender differentiation if the difference by sex is significant and in the same direction in all three sectors. Unfortunately none of the variables examined in this paper fell into this category. A gender difference is also strongly implied if two sectors are significant, and a third, while not significant, shows trends consistent with the other two sectors. Table 7 demonstrates this kind of pattem, as does Table 8 in the case of "siblings" (where men are much more likely to be related to other owners as siblings than women are), and Table 13 for higher incidences of male multiple business ownership. Most of the tables do not follow even this reduced indicator of gender. Usually the trend strong in two sectors is absent in the third (e.g., see Table 12 in the case of hotel and catering). In some cases the third sector contradicts the other two (see Table 9 where the dominant trend is reversed in the hotel and catering sector). Lack of consistency in trends between sectors is an example of "discordant variation." Discordant variation by sex should not lead us to necessarily conclude that no gender trends exist. This is possible, but we showed in Table 11, when referring to the "cook/ chef" gender difference, that instances may arise where gender relevantly affects only one sector. (Cooks and chefs have no relevance to management practice in the textile or business services sectors.) Some interesting results relate to kinship and ownership, and reflect institutionalized social trends. We noted that levels of inheritance were surprisingly low, and in the total sample were only marginally higher for men. In the traditional textile/clothing sector, however, men were much more likely to inherit the business. This points to a persistence in the traditional reluctance to pass on resources to daughters. We highlight too the rarity of any co-ownership with sisters. It would appear that women are mostly confined to partnerships with domestic partners, while men have a much wider choice of relatives to share the business with. The power of the family business in general is shown by the fact that owners that left tended to be non-relatives, a trend not significant by sex of respondent. The role of kinship in small business ownership appears to differ significantly by gender, and invites farther research. The study also demonstrates that the truly individual owner-manager, particularly the individual woman in business, is less common than the literature implies. Two-thirds of businesses involve partnership of some kind. Even sole traders tend to have domestic partners and other relatives who are contributing to the business in some way (we have a great deal of evidence for this which could not be reported in this paper). The diversity of sole trading from one sector to another and one study to another questions, as we discussed earlier, the view that women prefer sole trading as a legal state (Brush, 1992). Indeed women are not always proportionally the majority of sole traders, as we show in Table 2, where male sole traders are more common in business services. If sole trading is the dominant form of legal ownership in sectors associated with new smaller businesses, often low performance businesses, then this is likely to be the form of legal ownership adopted by any business owner for that business type. If women tend to be more commonly sole traders, then this may not so much be a symptom of preference, but of structural disadvantage, that women tend to be clustered in small low-performance sectors. 24

ENTREPRENEURSHIP THEORY and PRACTICE

The most spectacular and consistent gender difference involves multiple business ownership. Why men tend to own more businesses than women begs many genderrelated explanations. (For example, men are reputed to have more resources; are less likely to value financial security; have more opportunities and choice of establishing new ventures; and so on.) A major possibility too is that in practice many women may not have the total control of their own resources necessary for lateral business expansion. Most women either share their business resources with legal partners, and/or share their domestic resources with their domestic partners. This is also true for male business owners to a large extent, but social custom arguably allows men more freedom to separate and "ring fence" resources for their own ends. These conclusions, however, need to be modified to take into account the inherent heterogeneity of female business owners in their relationships with spouses and partners. For example, the detrimental impact of the husband to a woman's business, indicated as common by Goffee and Scase (1983), is often contradicted in our study. The results are also open to more positive interpretations. Although there may be more male multiple business owners proportionally, there are still a significant number of female multiple business owners too. If we speculate that women in business have started from a much lower tradition of achievement, and have a smaller resource base on average, then these figures are remarkable, and may be an indicator that the gender gap may be narrowing (Birley, 1989). This positive note, however, is a cautious one. One of the most clearcut differences was the total absence of all-female owned businesses with three or more owners, in contrast to a significant representation of all-male owned businesses with three or more owners. This may reflect the rarity of opportunities within the labor market or within families for groups of women to get together to start businesses.

Methodological Implications The fmdings also have implications for the methodology of small business gender studies. As most variables vary so widely between sectors, precise "gender" findings of a study may be a consequence of the sectoral mix of the sample as much as genuine social trends impinging on the small business community. We concur with Thomas, Uribe-Echevarria, and Romijn (1991, p. 5) when they state; The so-called small scale sector is far from homogenous and does not constitute a single sector of the economy in any real sense. Reality shows a sectoral labour market, regional heterogeneity that cannot be ignored. A most disaggregated view of small scale economic activity should therefore reduce the emphasis on broad and highly aggregated analysis and policy. If sectoral differences are large, then it becomes almost impossible to talk about a "random" typical sample of small businesses from which to draw general conclusions. Detailed descriptive results become relatively meaningless outside the context of that study. To say that X% of men compared to Y% of women are active managers, for example, would be a sterile statistic, as another sample with another sectoral mix could show quite different figures. In this scenario, it is the relationships of variables that are important to enable us to generalize from gender trends, rather than inferences from detailed figures. Secondly, the practice of comparing samples of female and male "entrepreneurs" without taking into account wider ownership issues in the sampling and analysis phases can be questioned. Fischer (1992, p. 6), for example, extracted 2724 firms from the Dun Spring, 1994

,

25

and Bradstreet database in Canada and identifled 60 female-owned firms out of 508 usable replies. She does not explain what is meant by "female owned," or whether the male-owned sample is totally male-owned, or a mixture of male and female co-owners. What does Bmsh (1988, p. 612) mean by "women owned enterprises now account for more than 25% of all small businesses?" Even Kalleberg and Leicht's otherwise commendable and thorough longitudinal analysis of Indiana firms does not elaborate on what is meant by "businesses headed by women" (1991. pp. 136, 139), or what differential success and performance between male- and female-headed businesses really means in the light of the authors' failure to control fully for co-ownership. This paper shows that a majority of owner-managers, whether female or male, share decision making in some way. Indeed a majority of co-owned businesses were of mixed sex. What are we comparing when we compare "random" sam'les of female and male entrepreneurs or business owners? Are we comparing management of female and male ways? Or are we comparing perceptions of a common process of decision making, or merely the compromise between the negotiated decisions between male and female owner-managers within the flrm? These questions are obviously much more complex to answer than is implied by a simple male versus female sampling strategy, where there is a strong implication that each owner-manager in the sample is entirely responsible for the decisions of the firm he or she manages and that he or she does not share the burden with others. Future studies should seriously consider the possibilities of adopting a sampling framework based on the categories outlined in Table 1, rather than a simple male versus female approach. Finally, the evidence from this paper suggests that, for many women, proprietorship tends to provide only a limited opportunity for "liberation through self-determination" (Goffee & Scase, 1983, p. 627). Most women have to accommodate their own ambitions in business ownership with those of others. The impact of "gender" on small business ownership is thus not interpretable in simple unidimensional terms, but can only be fully understood by taking a more holistic approach to research (Hamilton, Rosa, & Carter, 1992).

REFERENCES Birley. S. (1989). Female entrepreneurs: Are they really any different? Journal of Small Bu.siness Management, January. 32-36. Birley, S., Moss, C , & Saunders. P. (1987). Do women entrepreneurs require different training? y^mer/can Journal of Small Business, /2(1), 27-35. Brush. C. (1988). Women entrepreneurs: Strategic origins impact on growth. In B. A. Kirchoff. W. A. Long, W. E. McMullan. K. H. Vespur & W. E. WetzeL Jr. (Eds.). Frontiers of entrepreneurship research, pp. 612-625. Wellesley, MA: Babson College. Brush, C. (1992). Research on women business owners; Past trends, a new perspective and future directions. Entrepreneurship Theory and Practice. 16(4), 5-30. Carter, S., & Cannon. T. (1991). Women as entrepreneurs. New York: Academic Press. Challis. B. (1993). Self reported motivation towards entrepreneurship: A comparison between male and female proprietors of recently formed businesses. MSC in Entrepreneurial Studies Dissertation. University of Stirling. Cromie. S., & Hayes, J. (1988). Towards a typology of female entrepreneurs. The Sociological Review, 36(\), 87-113. 26

ENTREPRENEURSHIP THEORY a n d PRACTICE

Davidson, M. J., & Cooper, G. L. (1992). Shattering the glass ceiling: The woman manager. London: Paul Chapman. Fischer, E. (1992). Sex differences and small business performance among Canadian retailers and service providers. Journal of Small Business and Entrepreneurship. 9(4), 2-13. Goffee. R.. & Scase, R. (1983). Business ownership and women's subordination: A preliminary study of female proprietors. The Sociological Review. 3{4), 625-648. Hamilton. D. (1990). An ecological basis forthe analysis of gender differences in the predisposition to self employment. University of Stirling, SEF monograph 73/90. Hamilton. D.. Rosa, P., & Carter, S. (1992). The impact of gender on the management of small business: Some fundamental problems. In R. Welford (Ed.), Small businesses and small business development—A practical approach, pp. 33-40. Bradford: European Research Press. Kalleberg, A., & Leicht, K. (1991). Gender and organisational perfonnance: Determinants of small business survival and success. Academy of Management Journal. 34(1), 136-161. Meijer, J., Braaksma, R., & Van Uxem, F. (1986). Contributing wife partner in business, ln R. Donckels & i. Meijer (Eds.), Women in .tmall business, pp. 65-77. Maastricht: Van Gorcum. Meittinen, A. (1986). Contributing spouses and the dynamics of entrepreneurial families. In R. Donckels & J. Meijer (Eds.), Women in .small busine.ss, pp. 78-86. Maastricht: Van Gorcum. Moser. C. A., & Kalton. G. (1971). Surx'ey methods in social investigation (2nded.). London: Heinneman. Solomon, T., & Femald, L. (1988). Value profiles of male and female entrepreneurs. International Small Business Journal, 6(3), 24-33. Thomas, H., Uribe-Echevama, F., & Romijn, H. (1991). Smalt-scale production. London: Intermediate Technology Publications.

Peter Rosa is a Senior Research Fellow at the Scottish Enterprise Foundation, University of Stirling, Stirling. Scotland FK9 4LA. Daphne Hamilton was Research Fellow at the Scottish Enterprise Foundation. University of Stirling and is now Lecturing at the Department of Political Science and Social Policy at the University of Dundee, Scotland. The research was funded by the UK Economic and Social Research Council. The research team also consisted of Sara Carter and Helen Bums. Their contribution during the iife of the project made this paper possible, but all errors and unsubstantiated opinions remain our own.

Spring, 1994

27