OVERCONFIDENCE AND ITS INFLUENCE ON DUTCH FEMALE ENTREPRENEURS
Abstract summaryThe aim of this thesis is to find out whether previous findings about overconfidence and entrepreneurs hold in a different setting, namely the Netherlands circa 2009. Using a probit regression model and data from the General Entrepreneurship Monitor, a number of variables are tested to find that overconfidence in one’s knowledge and skills, gender, age, education, fear of failure and personally knowing an entrepreneur, all have an impact on becoming an entrepreneur. The directionality of that impact can be found in table 5, which shows that overconfidence positively influences a person to start a business. Dutch females are found less overconfident than males are.
Student: Nicole Koedooder Student number: 6054137
Supervisor: Laura Rosendahl-‐Huber Program: Economics & Business Track: Finance & Organization
Chapter 1 Introduction
Chapter 2 Theoretical framework
2.1 Overconfidence 2.2 Gender
2.3 Entrepreneurship 2.4 Hypotheses
Chapter 3 Data and method
3.1 Data
3.2 Descriptive statistics
3.3 Dependent and independent variables 3.4 Method Chapter 4 Results 4.1 General analysis 4.2 Regression analysis Chapter 5 Conclusion Acknowledgements Bibliography 3 5 6 6 8 8 9 9 10 11 12 15 16 17
1. Introduction
This thesis finds its motivation in the way entrepreneurial behavior is influenced by an individual’s gender in combination with a cognitive bias called overconfidence. The idea is that when the perception of someone’s own capabilities is biased towards believing they have the required skillset to be an entrepreneur, thereby overestimating the extent of their capabilities, the probability that they will start a business is increased (Camerer & Lovallo, 1999). This thesis explores the way in which overconfidence influences Dutch entrepreneurship levels, and whether the outcome differs between two genders. The research question posed in this thesis is therefore: In what way does overconfidence influence the participation of Dutch females in self-‐employment?
The importance of entrepreneurs compared to people who work for an employer, lies in the entrepreneur’s ability to create jobs for themselves and others, investing in and innovating the existing market, thereby adding value to the economy (Ekelund et al. 2005). Although most entrepreneurs are men (Kelley et al. 2012), women form a subset of entrepreneurs that is worthy of a closer look. This is because female entrepreneurs often reinvest their profits in their families and in society (Brush, 2013). This thesis thus examines the differences in overconfidence between the men and women in the sample. To be able to answer the research question, a regression analysis is performed on data that contains information about real life entrepreneurs.
The Global Entrepreneurship Monitor (GEM) is an organization that collects such data about entrepreneurship each year. Their surveys are widely used by scientists who use these sets of data to write reports and conduct research. According to GEM’s Women’s Report, which was published in 2012, only 7% of the females in the Netherlands are entrepreneurs vs. 14% of the males in the sample. Among those females, 73% pursue an opportunity in the market, while the rest becomes an entrepreneur out of necessity or other reasons. The report also states: “In every economy, women have lower capability perceptions than men as well as a greater level of fear of failure than men”. In the Netherlands, 31% of women vs. 54% of men believe they have the capabilities needed in order to start a business and only 14% of female entrepreneurs in developed countries in Europe predict they will add 6 or more employees in the next five years, which is a lower percentage than their male counterparts (Kelley et al. 2012). For the purposes of this thesis,
a dataset is used that was collected in 2009 by the Dutch national team of the GEM consortium. The 3003 respondents include people from each region and age group, forming a good representation of the Dutch population (Global Entrepreneurship Monitor, 2014). The models used in the analysis are specified using independent variables that – according to the existing literature – are determinants of entrepreneurship. These variables include overconfidence and gender, as well as the following control variables: age, education, knowing an entrepreneur and fear of failure. The analysis consists of a regression on three different models, based on a paper by Koellinger et al. (2007). Each model’s dependent variable is a binary representation of a type of entrepreneur, defined by the stage they are in. The effect of overconfidence on entrepreneurial behavior is estimated by running a probit regression for each type of entrepreneur. Whether the results of this regression are significantly different for males and females is also tested. Previous researches have either focused on various determinants of entrepreneurship and how they might vary across countries, or on gender differences related to overconfidence. This thesis combines the two and focuses on the gender differences related to overconfidence within the context of entrepreneurial behavior. This thesis also focuses on the Netherlands as a subset of the available data and, contrary to past research, does not focus on cross-‐country or cross-‐ cultural differences.
This thesis finds that there is a significant difference in overconfidence between male and female respondents. Men exhibit overconfidence more often than women do. The results of the regression analysis are mostly as expected with regard to previous researches on this topic in different settings. The variable for overconfidence has a small but statistically significant positive influence on entrepreneurship rates for each type of entrepreneur. The coefficient for gender shows that females are less likely to start a business than males are. Age and education positively influence entrepreneurship, though both age squared and post-‐secondary education negatively influence entrepreneurship. Fear of failure has a negative influence in all cases yet knowing an entrepreneur positively influences respondents to starting a business themselves.
Chapter two is a description of the existing literature on entrepreneurship and overconfidence, in which an explanation for each independent variable included in the model is given. In the following chapter, the data and method of research will be described
in detail. The fourth will describe and review the results of the analysis and the final chapter concludes by summarizing and discussing the implications of this thesis.
2. Theoretical framework
This thesis is set out to assess the way in which overconfidence influences entrepreneurial behavior in Dutch females. The next paragraphs discuss the theory on overconfidence and gender in relation to entrepreneurship, as well as the existing theory that suggests there is a relation between entrepreneurship and the factors that influence entrepreneurial behavior.
2.1 Overconfidence
In a cross-‐country research, Koellinge et al. (2007) find evidence that confidence in one’s own skills and knowledge is a significant determinant in deciding to become an entrepreneur and also more present in newly starting entrepreneurs than in established entrepreneurs. This cognitive bias is called overconfidence and causes an individual to overestimate ones skillset and underestimate the risk that is involved with starting a new business (Cooper et al. 1988). Camerer and Lovallo (1999) argue that when we are depending on our own skills, we are subsequently overestimating our performance and are more likely to enter into the market, resulting in excess market entry and lower market shares to be gained. Overconfidence can also be defined as the failure to know the limits of one’s knowledge (Simon et al. 1999). Lowe and Ziedonis (2005) find that entrepreneurs who are overly optimistic tend to continue unsuccessful projects for longer than established businesess would. Herz et al. (2013) define two different forms of overconfidence, namely overoptimism and judgmental overconfidence. The latter is explained as a tendency to overestimate the precision of the information the entrepreneur has on the company and the market. Much unlike overoptimism, which leads to an increase in innovation and entrepreneurial behavior, their findings suggest that judgemental overconfidence influences innovation negatively (Herz et al. 2013).
Overconfidence is found to be higher among men than among women, according to a research where Swedish university students were graded based on their actual skills and knowledge, as well as the amount of confidence they exhibited when asked to answer a
difficult bonus question (Bengtsson et al. 2004). However, other researchers that use a Swedish running competition as their sample, suggest that women within a competitive male-‐dominated business environment are likely to be more optimistic and more confident than men, and that this overconfidence increases their performance as well (Nekby et al. 2007). Similarly, Hardies et al (2013) find that a gender difference in overconfidence is more present among university students than among professionals.
2.2 Gender
According to Blanchflower (2004), the probability of being a self-‐employed individual is higher among men than among women, all across fellow members of the OECD. This is congruent with the findings of other papers by Kelley et al. (2012) and Koellinger et al. (2007). Female entrepreneurs are often older than male entrepreneurs (Llussá, 2010). These factors might be indicative of the reason why women have less confidence in their entrepreneurial skills and are, according to Fairlie and Robb (2009), indicative of low business performance. Asscher (2012) finds that although women still face numerous obstacles when considering becoming an entrepreneur, such as lack of funding and experience and gender discrimination, as well as family responsibilities that restrict the time they can spend on running a business, the number of women who participate in self-‐ employment is growing and barriers are being crossed. These women have greater confidence in their skills and are at least as succesful as their male counterparts (Asscher, 2012), contrary to previous findings. In their study using GEM data on 29 countries, Verheul et al. (2006) find that female and male entrepreneurship rates are significantly influenced by the same factors that also move in the same direction for each gender, such as income lvels and family life. Some factors influence the share of female-‐, and the number of female entrepreneurs differently, so therefore they recommend that governments take this into consideration when implementing entrepreneurship-‐stimulating projects. Their study does find a gender difference in the effects of unemployment and life-‐satisfaction on entrepreneurial activity (Verheul et al. 2006).
2.3 Entrepreneurship
when or how people decide to start a business; these four factors are described in this paragraph.
Since it takes time to earn an income as an entrepreneur, one would not earn an hourly wage or weekly paycheck for example, starting a business requires interest, motivation and an incentive in the form of future return on investment. According to Lévesque and Minniti (2006), people are more inclined to invest the time that goes into starting a business at a young age, rather than later in life. After a threshold point, typically around the age of 35, their interest starts declining, shifting towards spending their time more leisurely. The amount of time we have in our lives is limited, and the amount of time left to earn back the invested capital decreases with age. Another reason could be that waged income might become more appealing as one gets older and more experienced (Lévesque & Minniti, 2006).
Education influences entrepreneurial behavior as well. More so for women than for men, having attained a secondary education or higher increases the probability that an individual will start a business driven by opportunity, while entrepreneurship driven by necessity is decreased by education (Llussá, 2011). Especially nascent entrepreneurs, according to Koellinger et al. (2007) exhibit a positive trend showing that the more educated a person is, the more likely that person is to start their own business. According to Wadhwa et al. (2009), 95 percent of the entrepreneurs they interviewed had obtained a post-‐ secondary degree, of which 45 percent had even more advanced education. Their study finds that most entrepreneurs were at the top of their class in high school (Wadhwa et al. 2009). However, Lazear’s theory suggests that entrepreneurs are not usually schooled experts in one specific skill, but are rather “jacks-‐of-‐all-‐trade” (2005).
Whether or not an individual knows an entrepreneur personally is a determinant of entrepreneurship in itself. Knowing an entrepreneur grants a person access to (some) knowledge about the benefits of the life of an entrepreneur and might therefore positively influence the decision to become self-‐employed (Koellinger et al. 2007). A study among Dutch founders of newly established ventures suggests that entrepreneurs (especially females) have increased satisfaction levels when it comes to their income and leisure time (Carree & Verheul, 2012). That being said, a research among Harvard Business School graduate students begs to differ. Their findings suggest that the probability of becoming an entrepreneur decreases, rather than increases, for students who are brought in direct
contact with peers who have previous entrepreneurial experience. For the post-‐graduates whose business would prove successful however, the effect was more positive (Lerner & Malmendier, 2013).
Finally, fear of failure decreases the likelihood of a person starting a business. Cramer et al. (2002) find that risk aversion is a discouragement for entrepreneurial activity, although appointing causality isn’t as ‘clear cut’ as desired. Accordingly, Ekelund et al. (2005) find a negative relationship between being risk averse and the decision to become self-‐employed as well, factoring in control variables such as parental role models and education. Hardies et al. (2013) investigate gender differences in both overconfidence and risk taking. They find that women are more risk averse than men in all settings (Hardies et al. 2013). In 2012, 39% of the adults surveyed by GEM said that fear of business failure is what actually prevents them from starting a new business (Van Der Zwan et al. 2012).
2.4 Hypotheses
The research question this thesis poses is: In what way does overconfidence influence the participation of Dutch females in self-‐employment? Based on the literature discussed above, the expected findings are as follows:
H1: Dutch females are less overconfident than Dutch males
H2: Overconfidence has a positive effect on the entrepreneurial activity in the Netherlands
3. Data and method
3.1 Data
This thesis uses data collected by the Global Entrepreneurship Monitor’s Dutch national team in 2009. The GEM consortium is a global organization that collects specific data about entrepreneurs, where 70 countries are included in the yearly survey. In the Netherlands, 3003 adults were surveyed over the phone to form a detailed dataset that is representative of the population. However, as the method of collecting is subject to the human error, some observations have missing data and are therefore dropped from the analysis. For instance,
this pertains to people who claimed they were more than 99 years old. This leaves 2970
individual observations that can be used in our regression.
3.2 Descriptive Statistics
Before looking at the method of research more closely, the descriptive statistics of the variables – independent of entrepreneurial status – are given in table 1. The dataset of 2970 respondents includes 75 nascent entrepreneurs, 84 new entrepreneurs and 207 established entrepreneurs. During analysis, the number of observations has dropped to 1787 due to missing values for some of the indicators. The group is divided by gender: 45% is male and 55% is female.
Table 1 Descriptive statistics for all respondents of the 2009 Dutch GEM survey, n = 1787
3.3 Dependent and independent variables
The main interest of this thesis lies in gender differences in overconfidence. Differences between countries and cultures were researched by Koellinger et al. (2007) and Llussá (2010) among others, and are thus left unconsidered in this thesis. The description of each variable below is directly taken from the GEM questionnaire.
Three different phases of entrepreneurship are used as the outcome variables in the regression. Each of those three outcomes is simultaneously tested for a significant difference between males and females. A chi-‐squared test for population differences is used for this purpose. Nascent entrepreneurs (represented in the data by suboanw) are defined as those who are actively involved in starting up a business that is not paying any wages or salaries yet. New entrepreneurs (represented in the data by babybuso) are defined as those
Mean Std. Error Overconfidence .6922 .0313 Female .5098 .0118 Age 52.812 .3751 Education (secondary) .6609 .0112 Education (post-‐secondary) .2641 .0104 Knows an entrepreneur .4219 .0247 Fear of failure .7353 .0462
who manage and/or own a business that is less than 42 months old. Established entrepreneurs (represented in the data by estbbuso) are defined as those who manage and/or own a firm that is more than 42 months old.
The independent variables consist of six factors that are expected to have a significant influence on entrepreneurship rates, as well as some interaction terms. The first variable represents overconfidence. A dummy variable called suskill is added to the model, indicating whether the respondent believes they have the knowledge, skills and experience required to start a new business, and is thus overconfident. Gender is represented by a binary variable where being female equals 1 and being male equals 0. Besides the variables that are crucial to this thesis, several control variables are added to the model.
Age is given in years. Because there appears to be an inverse u-‐shaped relation between entrepreneurial behavior and an individual’s age, according to Lévesque and Minniti (2006), a variable for age squared (agesq) is added to the models. To define the respondent’s highest level of education, four dummy variables are created that represent each possible answer of the variable gemeduc. These dummies are: somesecond for people who have obtained some secondary education, secondary for those who have obtained secondary education, postsecond for those who have attained an education past high school and gradexp for those who went to graduate school. However, since none of the respondents have graduate experience and all of the respondents have attained at least some secondary education, these two variables are dropped from the regression to avoid collinearity. Besides the respondents’ demographics, theorie suggests that the following variables also influence entrepreneurial behaviour and are therefore added to the models. The first is a dummy variable defining wheter the respondent personally knows anyone who has started a business in the past two years, represented in the data as knowent. The second is fearfail, a dummy indicating whether fear of failure would prevent the respondent from starting a business.
Each model also includes a few interaction terms, to account for significant differences in for example the effect of a variable between genders. These interaction terms are created taking the existing theory into account as well as the correlations (see table 4) that exist between them, and consist of the following: gender & overconfidence, age & overconfidence, gender & fear of failure and fear of failure & overconfidence.
3.4 Method
First, a general analysis on the data is performed to see if the first hypothesis holds by means of an independent t-‐test on the differences between men and women. Then, a probit regression analysis is performed on the data, in order to estimate the directionality of the parameters for each variable discussed above. The regression model is specified as follows:
Where 𝑗 = 𝑛𝑎𝑠𝑐𝑒𝑛𝑡−, 𝑛𝑒𝑤−, 𝑒𝑠𝑡𝑎𝑏𝑙𝑖𝑠ℎ𝑒𝑑 𝑒𝑛𝑡𝑟𝑒𝑝𝑟𝑒𝑛𝑒𝑢𝑟𝑠 And 𝑖 = 𝑎𝑔𝑒, 𝑎𝑔𝑒!, 𝑠𝑒𝑐𝑜𝑛𝑑𝑎𝑟𝑦 𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛, 𝑝𝑜𝑠𝑡𝑠𝑒𝑐𝑜𝑛𝑑𝑎𝑟𝑦 𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛, 𝑘𝑛𝑜𝑤𝑠 𝑎𝑛 𝑒𝑛𝑡𝑟𝑒𝑝𝑟𝑒𝑛𝑒𝑢𝑟, 𝑓𝑒𝑎𝑟 𝑜𝑓 𝑓𝑎𝑖𝑙𝑢𝑟𝑒.
After testing this model for each outcome of y, the four interaction terms mentioned in the previous paragraph are added to the model and it is then tested again for each outcome of y. The results are in table 5 where robust standard errors are reported, in case there is a misspecification in the model.
4. Results
In this chapter, a general analysis is performed on the data. First, a one-‐sided independent t-‐ test is performed to see if hypothesis H1 holds. Afterwards, the probit regression analysis is performed and discussed.
4.1 General analysis
To see if the first hypothesis holds, the difference between male and female overconfidence levels needs to be tested for significance. Table 2 shows how the 1787 respondents who answered the question about overconfidence, feel about their own knowledge, skills and experience. At first glance, it appears that men more often than not believe that they have sufficient skills to start a business. The exact opposite seems true for women. In order to see if the first hypothesis holds (recall H1: Dutch females are less overconfident than Dutch
𝑦! = 𝛽!+ 𝛽!𝑜𝑣𝑒𝑟𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 + 𝛽!𝑓𝑒𝑚𝑎𝑙𝑒 + 𝛽!𝑋!+ 𝜀!
males), a two-‐sample t-‐test is performed on the data. This test provides an insight into whether the level of overconfidence is significantly different for men than for women and, more specifically, whether the difference is significantly larger than zero.
Table 2 Frequency table of overconfidence for male and female respondents
Male Female Total No 320 573 893 Yes 525 320 845 Don’t know 31 18 49 Total 876 911 1787
Table 3 shows that the variable suskill, which represents overconfidence, has different means for men and women. The difference between the two means is 0.37, with a standard error of the difference of 0.06. The results of this test indicate that, with a t-‐value of 6.0192 (degrees of freedom of 1785), a 1% significance level and a p-‐value of 0.0000, the difference in means between male and female respondents is in fact significantly different from zero. Since the sample forms a good representation of the Dutch population in 2009, the evidence is in line with the hypothesis.
Table 3 Independent t-‐test on overconfidence
Observations Mean St. Error Male 876 .8824 .0488 Female 911 .5093 .0386 Combined 1787 .6922 .0312 Difference .3731 .0620 4.2 Regression analysis
The independent variables discussed in chapter three are subjected to the pairwise correlation analysis listed in table 4; correlation coefficients that are significant are marked with an asterisk. Since the focus of this thesis lies on the variable overconfidence, its interest is found in the way overconfidence and age, gender and fear of failure move together. The model that is specified in paragraph 3.4 includes these interaction terms.
Table 4 shows that the variables that are used to create the interaction terms also exhibit significant correlation coefficients, except for overconfidence and fear of failure. One could argue that these last two have such differences in the way in which they influence entrepreneurship, that their movements are not related.
Table 4 Correlation tabel independent variables, n = 1787
Age Gender Second Post-‐sec. Knows
an entr. Fear of Failure Overcon-‐fidence
Age 1.0000 Gender 0.0321 1.0000 Education (secondary) -‐0.3035* -‐0.1275* 1.0000 Education (post-‐sec.) 0.3368* 0.1217* -‐0.8423* 1.0000 Knows an entrepreneur -‐0.1010* -‐0.0497* 0.0506* -‐0.0525* 1.0000 Fear of failure 0.1468* 0.1055* -‐0.1007* 0.1033* 0.0035 1.0000 Overconfidence -‐0.0850* -‐0.1410* 0.0273 -‐0.0305 0.0657* 0.0215 1.0000
-‐ * Correlation coefficient is significant at 5%
Table 5 provides an overview of the probit regression results. Each coefficient represents the change in the outcome variable if the value of the variable changes from 0 to 1. Coefficients that are significant are marked with an asterisk. The table shows several interesting findings.
The coefficient for overconfidence is significant and positive for both new and established entrepreneurs, located in the columns for models 2a and 3a. These findings support the second hypothesis (recall H2: Overconfidence has a positive effect on the entrepreneurial activity in the Netherlands). Even though these coefficients are positive, they only have a small contribution to the probability of being an (established or new) entrepreneur. The fact that this effect is smaller than expected can possibly be attributed to the year in which the data was obtained. The survey was held in 2009, the year after the financial market had collapsed and created a worldwide crisis.
In each model, gender negatively influences the probability of being an entrepreneur. Since the value of the dummy variable for gender that is assigned to females is equal to 1, this means that being a female decreases the likelihood of starting a business. This is in line with what the theory explains about why less women are business owners;
they face obstacles such as lack of funding and experience and having more responsibilities in their family life than men would have (Asscher, 2012), they are also more risk averse (Hardies et al. 2013) and less overconfident (Bengtsson et al. 2005) than men.
As expected, the coefficients for the variables age and age squared are significant in every model. Age influences entrepreneurship positively, and age squared influences entrepreneurship negatively. This creates an inverse u-‐shaped effect, which means that the interest in becoming a business owner increases with age, but then, after a certain point, decreases as one gets older (Lévesque & Minniti, 2006).
In most models, having a post-‐secondary education negatively influences the probability of becoming an entrepreneur (although none of the coefficients for education are significant). This might be explained by the fact that a university degree gives a person better employment opportunities and having a stable job eliminates the need to create a job for oneself. As Lévesque and Minniti (2006) explained it, when you reach a certain age, the time you have left in your career to earn back the time and monetary investments that you made in starting your business becomes limited. That awareness can cause one to re-‐ evaluate the benefits of receiving a full-‐time employed wage.
Knowing an entrepreneur proves to positively influence entrepreneurship in all cases. For new and established entrepreneurs, the coefficient for knowing an entrepreneur is significant. The coefficient for fear of failure is significant and negatively influences entrepreneurship in each model. This result is logical because fear of failure, or similarly risk aversion, is what would prevent someone from starting a business (Van Der Zwan et al. 2012).
Only for established entrepreneurs, the coefficients of the interaction terms for gender & overconfidence, fear of failure & overconfidence and fear of failure & gender are significant. The coefficient for gender & overconfidence is positive. This means that for someone who is both female and confident in her own skills, the likelihood of starting a business is increased. The coefficient for fear of failure & overconfidence is negative, which means that a person, who is both overconfident and afraid of failure, is less likely to be self-‐ employed. The coefficient for fear of failure & gender is positive for both nascent and established entrepreneurs and negative for new entrepreneurs. This means that females who are afraid of failure are mostly more likely to start a business or already own a
Table 5 Probit regressions or nascent, new and established entrepreneurs
Nascent entrepreneurs New Entrepreneurs Established Entrepreneurs
Model 1a Model 1b Model 2a Model 2b Model 3a Model 3b Overconfidence .04605 (.0287) -‐.0740 (.1261) .0699 * (.0196) -‐.0306 (.0968) .0573 * (.0248) -‐.0926 (.1351) Female .0037 (.1126) -‐.1195 (.1243) -‐.0341 (.1127) -‐.0322 (.12022) -‐.4090 * (.0860) -‐.6047 * (.0984) Age .1023 * (.0335) .1018 * (.0338) .0719 * (.0259) .0714 * (.0258) .1995 * (.0270) .2017 * (.0277) Age2 -‐.0013 * (.0004) -‐.0013 * (.0004) -‐.0010 * (.0003) -‐.0010 * (.0003) -‐.0020 * (.0003) -‐.0021 * (.0007) Education (secondary) .1951 (.2377) .1955 (.2405) .4276 (.2616) .4387 (.2642) .0080 (.1658) .0224 (.1648) Education (post-‐ secondary) -‐.2484 (.2875) -‐.2410 (.2898) .0165 (.3019) .0248 (.3037) -‐.1519 (.1853) -‐.1253 (.1851) Knows an entrepreneur .0319 (.0353) .0260 (.0380) .0863 * (.0321) .0811 * (.0336) .0849 * (.0322) .0803 * (.0336) Fear of Failure -‐.2792 *
(.1301) -‐.7292 (.4720) -‐.4077 * (.1462) .1497 (.37663) -‐.1364 * (.0558) -‐.6763 * (.2955) Constant -‐3.6492 * (.7774) -‐3.4466 * (.7904) -‐2.9321 * (.6841) -‐2.9413 * (.6878) -‐5.1970 * (.6862) -‐4.9686 * (.7083) Age * overconfidence .0006 (.0020) .0004 (.0017) -‐.0002 (.0020) Female * overconfidence .0815 (.0563) .0688 (.0429) .1447 * (.0550) Fear of failure *
overconfidence -‐.1169 (.0693) -‐.0521 (.0393) -‐.1211 * (.0516) Female*
fear of failure .3069 (.2424) -‐.3544 (.2769) .3198 * (.1521)
Observations 1787 1787 1787 1787 1787 1787
Log likelihood -‐280.1329 -‐278.1701 -‐297.5337 -‐296.3106 -‐562.9462 -‐554.5160
Pseudo R2 0.0999 .1062 .1219 .1255 .1214 .1345
-‐ * Coefficient is significant at 95% -‐ Robust standard errors in parentheses
5. Conclusion
The aim of this thesis is to find an appropriate answer to the question: In what way does overconfidence influence the participation of Dutch females in self-‐employment?
The main results of this thesis are that overconfidence in one’s knowledge, skills and experience influences respondents to start a business in a positive way. This thesis finds a significant difference in overconfidence between male and female respondents. Women exhibit significantly less overconfidence than their male counterparts. Looking at the findings by Bengtsson et al. (2005) and Nekby et al. (2008), this was expected. Women are also less likely to be self-‐employed than men are. However, the women who do consider themselves overconfident are more likely to become entrepreneurs. Overconfidence has a small but statistically significant positive influence on entrepreneurship rates for each type of entrepreneur. Age has a positive effect on entrepreneurship up to a certain point, after
which the negative coefficient for age squared takes over to account for a loss of interest after the age of 40. Education does not have a statistically significant effect on entrepreneurship. Fear of failure has a negative impact in all cases and knowing an entrepreneur positively influences respondents to starting a business themselves.
This pertains to Dutch adults who were surveyed in 2009. For the most part, the results are congruent with the findings from previous researches in different settings (year and country). Any different outcomes, such as the smaller size of the impact overconfidence has on entrepreneurship, might be attributed to the financial crisis of 2008, or any cultural differences that weren’t taken into account in this thesis. It is recommended that further studies use an approach that looks at the properties of overconfidence more profoundly. For example, the use of psychological experiments might prove very insightful.
Acknowledgements
I would like to thank to the Global Entrepreneurship Monitor for the publication of their surveys and datasets, as well as many useful manuals for coping with the data. I would also like to thank my supervisor Laura Rosendahl-‐Huber for helping me write this thesis and giving me very helpful suggestions and feedback.
Bibliography
Asscher, J. (2012). Female Entrepreneurship -‐ An Appropriate Response to Gender Discrimination. Journal of Entrepreneurship, 8(4), 97-‐114.
Bengtsson, C., Persson, M., & Willenhag, P. (2005). Gender and Overconfidence. Economic Letters, 86(3), 199-‐203.
Blanchflower, D. G. (2004). Self-‐Employment: More may not be better. Swedish Economic Policy Review , 11(2), 15-‐74.
Brush, C. (2013). How Women Entrepreneurs Are Transforming Economies and Communities. Retrieved from Forbes: www.forbes.com/sites/babson/2013/02/16/how-‐women-‐ entrepreneurs-‐are-‐transforming-‐economies-‐and-‐communities/
Camerer, C., & Lovallo, D. (1999). Overconfidence and Excess Entry: an Experimental Approach. The American Economic Review, 89(1), 306-‐318.
Carree, M. A., & Verheul, I. (2012). What Makes Entrepreneurs Happy? Determinants of Satisfaction Among Founders. Journal of Happiness Studies, 13(2), 371-‐387.
Cooper, A. C., Woo, C. Y., & Dunkelberg, W. C. (1988). Entrepreneurs' Perceived Chances for Success. Journal of Business Venturing, 3, 97-‐108.
Cramer, J. S., Hartog, J., Jonker, N., & van Praag, C. M. (2002). Low risk aversion encourages the choice for entrepreneurship: an empirical test of a truism. Journal of Economic Behavior & Organization, 48, 29-‐36.
Ekelund, J., Johansson, E., Järvelin, M.-‐R., & Lichtermann, D. (2005). Self-‐employment and Risk Aversion -‐ evidence from psychological test data. Labour Economics, 12, 649-‐ 659.
Fairlie, R. W., & Robb, A. M. (2009). Gender Differences in Business Performance. Small Business Economics, 33(4), 375-‐395.
Global Entrepreneurship Monitor. (2014). Retrieved from GEM: www.gemconsortium.org
Hardies, K., Breesch, D., & Branson, J. (2013). Gender differences in overconfidence and risk taking: Do self-‐selection and socialization matter? Economic Letters, 118(3), 442-‐444.
Herz, H., Schunk, D., & Zehnder, C. (2013). How Do Judgemental Overconfidence and Overoptimism Shape Innovative Activity? Games and Economic Behavior, 1-‐23.
Kelley, D. J., Brush, C. G., Greene, P. G., & Litovsky, Y. (2012). Women's Report. Global Entreprepreneurship Monitor.
Koellinger, P., Minniti, M., & Schade, C. (2007). "I think I can, I think I can": Overconfidence and entrepreneurial behavior. Journal of Economic Psychology, 28, 502-‐527.
Lazear, E. P. (2005). Entrepreneurship. Journal of Labor Economics, 23(4), 649-‐680.
Lerner, J., & Malmendier, U. (2013). With a Little Help from My (Random) Friends: Success and Failure in Post-‐Business School Entrepreneurship. The Review of Financial Studies, 26(10), 2411-‐2452.
Lévesque, M., & Minniti, M. (2006). The effect of aging on entrepreneurial behavior. Journal of Business Venturing, 21, 177-‐194.
Llussá, F. (2011). Determinants of erntrepreneurship: are women different? FEUNL Working Paper Series No. 555.
Lowe, R. A., & Ziedonis, A. A. (2006, February). Overoptimism and the Performance of Entrepreneurial Firms. Management Science, 52(2), 173-‐186.
Nekby, L., Skogman Thoursie, P., & Vahtrik, L. (2008). Gender and self-‐selection into a competitive environment: are women more overconfident than men? Economic Letters , 100(3), 405-‐407.
Simon, M., Houghton, S. M., & Aquino, K. (1999). Cognitive Biases, Risk Perception, and Venture Formation: How Individuals Decide to Start Companies. Journal of Business Venturing, 15, 113-‐134.
Van Der Zwan, P., Hessels, J., Hoogendoorn, B., & De Vries, N. (2012). National Report. Global Entrepreneurship Monitor, The Netherlands.
Verheul, I., Van Stel, A., & Thurik, R. (2006). Explaining Female and Male Entrepreneurship at the Country Level. Entrepreneurship and Regional Development, 18(2), 151-‐183.
Wadhwa, V., Aggerwal, R., Holly, K. Z., & Salkever, A. (2009). The Anatomy of an Entrepreneur: Family Background and Motivation. Kauffman Foundation Small Research Projects Research.
