Nesta Working Paper No. 13/01 January 2013

List of Tables

Table 1 EIP Budget....................................................................................................... 10 Table 2 Coaching: UK Approach East Midlands Development Agency, 2005.............. 22 Table 3 Coaching and Mentoring and Consultancy: East Midlands Development Agency 2005 . 23

Executive Summary

Entrepreneurial ability is widely regarded as key factor for success in business and consequently innovation policy has been placing greater focus upon the entrepreneur, his skills and values, a development that has been termed by some the entrepreneurial turn.

This report examines publicly supported policies for entrepreneurship development. Entrepreneurship policies are directed to encouraging socially and economically productive activities by individuals acting independently in business. Their principal objective is to increase a level of entrepreneurial activity which is considered to be below the social optimum. Policies may be implemented directly to address entrepreneurs' needs e.g. business advice programmes or through broadcast methods such as education policy.

Entrepreneurship policy and SME policy both aim at causing two different kinds of effects in the economy. Both seek to improve the performance of economic actors (entrepreneurship policy is focused on the key actors in the business, the entrepreneur, while SME policy seeks to increase the competitiveness of the firm) and both seek to increase the number of economic actors, (entrepreneurship hoping to increase the level of supply of entrepreneurs, while SME policy hopes to increase the number of competitive firms).

In regard then to these two areas for entrepreneurship policy, we have attempted to locate and focus on evaluations that reported on additionality / net effect or that use methods of causal inference to determine the effectiveness and impacts of policy. Such studies tended to be found in the academic literature rather than amongst the reports on the performance of government schemes (that are publicly available).

We have examined the literature and it covers schemes that have been evaluated around a decade ago. Generally we find that recent policy initiatives in this area have not been evaluated. The policies and programmes that have emerged very recently have yet to be examined in detail.

In regard to policies and programmes seeking to effect cultural and behavioural impacts, we believe that the literature can teach a number of lessons. Programmes that seek to increase awareness of entrepreneurship as a career choice can be seen within the area of education policy. However the growing interest in entrepreneurial education has not been matched by sound evaluation evidence. Thus not only have the evaluations undertaken varied greatly, the majority of these studies appear limited in that they fail to include a pre-test post-test method, and few employ control groups or have a longitudinal dimension.

The studies we have uncovered have attempted to address one or more of these concerns and to understand whether or not entrepreneurship education causes individuals to change their intentions to become an entrepreneur. The effect of these schemes appears to be different at different points in of the education system: at school level, there is a negative and significant effect; at the college level the effect is low; at the university level, the effect is positive.

The evaluations noted an important effect on intentions, what we call a sorting effect, by which those who may be unsure about whether they wish to follow an entrepreneurial path will have their minds made up for them, often resulting in a decision not to become an entrepreneur. It almost goes without saying that these schemes which are aimed at cultural and behavioural change are not designed in the short term to impact upon production or efficiency of economic enterprises.

Schemes to provide information and advice of a standardized form are closest in form to those which are addressed at the firm, rather than the entrepreneur. We note within our review two different forms of support. There is a form of support providing advice and information to early stage firms, and a form of support to nascent firms or pre-firms (pre start-up) where the recipient is the entrepreneur or would-be entrepreneur. Schemes of the former type are assessed by reference to the conventional economic impact categories, sales, employment and firm survival. Schemes of the latter type are assessed by reference to the outcomes for the recipient of the support, usually employment status (unemployed, employed (as an entrepreneur)) and income.

Schemes of the former type of scheme are more homogenous in what they provide, while the latter kind is quite diverse and examples are difficult to compare one against another. The former type of scheme, of which the Business Link scheme and the Swedish Innovation Centre are examples, show mixed results. Overall, some schemes find positive impacts in terms of sales, employment and survival while in others impacts were statistically insignificant. The schemes offering support to pre start-up entrepreneurs again varied in outcome. In the short term, all schemes increased employment; however, in the longer term, the US scheme, GATE for example, does not show persistence of employment effects at 5 years.

More specific and situational advice schemes provision is a further sub-category of schemes that seek to address the market failure for advice. The vast majority of programmes of this kind are targeted at those who have elected to run their own business. However, we have found examples of schemes (Active Labour Market Policy Schemes) that target the unemployed - what could be called reluctant entrepreneurs. This scheme is really a combined scheme in that it provides both general training and more specific advice for the "would be" entrepreneur. Such schemes combine both general and specific help to the unemployed. Some minor impacts are noted but the outcomes that have been observed are not a strong endorsement for this kind of scheme.

Coaching schemes are difficult to assess for impact. Those schemes providing assistance to new entrepreneurs running their own business are popular with those who receive the coaching but there is a lack of studies that consider the counterfactual case. It is our belief that as more of these schemes are put into operation, and there is an expectation that they will be, more evaluation will take place. This may reveal the factors that affect success. Incubators also are problematic to assess as the evaluations of them deal with a great variety of schemes few of which are comparable.

While we have found a trend towards entrepreneurial support, we have found no studies that follow up explicitly on the contrast between specific help and more general or operational support, in effect a comparison of two different forms of government assistance. It is a moot point therefore whether these schemes that do have different objectives should be subject to comparison.

1 Introduction

Entrepreneurial ability is widely regarded as key factor for success in business. The entrepreneur is the key mythic actor in the Schumpeterian account of the economic system and entrepreneurship is widely taught in business schools and within companies. Increasingly we see entrepreneurship instruction offered in schools to young children, in general university and college courses and even more widely than that. Some writers have noted that innovation policy is moving its emphasis away from the support of SMEs towards the support of entrepreneurship (Henrekson and Stenkula, 2009), a development that Cox and Rigby have termed "the entrepreneurial turn" (Cox and Rigby, 2012).

The study of initiatives to promote entrepreneurship generally raises interesting issues of public policy that might be regarded as technical in scope, and these we aim to deal with at some length in this paper where they are supported by a body of evidence from programme and policy evaluation. But we feel that we should note briefly a related and growing body of work covering entrepreneurship policy that is achieving public prominence. Entrepreneurship policy, by virtue of being a policy closely associated with a class of social and economic actors, is also quite unmistakeably a political matter. Entrepreneurialism is political creed in a number of countries but no more so than in the United States where a number of writers have extolled the culture of the United States as being more suited to entrepreneurialism than any other. In addition to cultural and social explanations for the origins of and value of entrepreneurialism there is a body of work emerging with the social and psychological sciences that investigates and in a number of cases demonstrates a strong link between entrepreneurial behaviour and genetics (Eckhardt et al., 2006; Nicolaou et al., 2008; Nicolaou and Shane, 2009; Nicolaou et al., 2009; Shane et al., 2010). Nevertheless, there are some who believe that entrepreneurship cannot be taught, that the desire to be entrepreneurial cannot be inculcated because it is "genetic" and that public resources that are allocated to such a task are a waste of money (Shane, 2009). This research we mention but not investigate as policy and programmes have yet to respond to incorporate its insights. We should note however that the UK approach may be making the assumption[^1] - we call the “selectivity of entrepreneurship" - that the firms that will succeed in becoming high growth are a minority, around 6% (NESTA, 2009a; NESTA, 2009c).

This report therefore examines publicly supported policies for entrepreneurship development. The literature often and mistakenly equates entrepreneurship policy and SME policy but as a number of authors (e.g. Storey, 2005) indicate, there is a substantive difference. Whereas SME policies are directed to existing businesses or enterprises, and seek to provide advice and information to help with problems that are firm related (e.g. compliance with regulations, reporting requirements), entrepreneurship policies are directed towards individuals. These are individuals who are either considering, are about to, may wish to consider, or have recently started a new business.

Entrepreneurship policies are therefore directed to encouraging socially and economically productive activities by individuals acting independently in business (Henrekson and Stenkula, 2009). Their principal objective is the increase the level of entrepreneurial activity which is considered to be below the social optimum. Policies may be implemented directly to address entrepreneurs' needs e.g. business advice programmes or through the broader macroeconomic policy frameworks e.g. education policy. Many innovation policies can have a spatial dimension in that they may be provided in a particular location, for example, in a business incubator, in a science park, in a cluster, or in a region. As entrepreneurship policies are focused on the entrepreneur, we have included a brief survey of incubators in our review because at this spatial level, i.e. within incubators, support is generally provided to entrepreneurs and early stages businesses, while at larger scales, the target of support is more likely to be the firm itself rather than the entrepreneur. Our view here is that the incubator is not a distinct form of business support but a mainly locational principle for the delivery of specific activities[^2], including entrepreneurship skills, coaching, and access to other similar firms. Because incubators are highly diverse in the type and balance of the measures they provide, are different also in their location, and have the prospect of many interactions between their constituent activities, the method of meta-analysis / systematic review cannot be realistically applied to a comparative study of the evaluations of incubators. Our review of reports on incubators therefore discusses the most recent literature on incubators and the important general themes emerging. Depending on the level of resource support, entrepreneurial policies can be classified as being hard or soft (Storey, 2005). Hard policies usually provide assistance in the form of finance (loans and grants) while soft measures include counselling activities to entrepreneurs before business start-up, counselling at the start-up phase, facilitating financial assistance, enhancing technology and access to technology and improving access to physical infrastructure, or advice after the start.

Entrepreneurship policy and SME policy both aim at causing two different kinds of effects in the economy. Both seek to improve the performance of economic actors (entrepreneurship policy is focused on the key actors in the business, the entrepreneur, while SME policy seeks to increase the competitiveness of the firm) and both seek to increase the number of economic actors, (entrepreneurship hoping to increase the level of supply of entrepreneurs, while SME policy hopes to increase the number of competitive firms).

2 Conceptual Background

2.1 Entrepreneurialism: Characteristics

The literature sees entrepreneurship as comprising a range of personal characteristics, which we suggest can be considered to form two general categories: a) a value or attitude system of cultural values; and b) a set of skills which we suggest are instrumental in the realization of a value system. The value system is broadly a set of beliefs about ends, namely that the results of entrepreneurship, wealth creation, personal interest, self-realization through business and capital accumulation are the personal priority of an individual. Weber's famous study 1905 (Weber, 1905) is one of the best known but by no means the last to explore inter alia the relationship between belief systems and occupational success. By contrast, the entrepreneurial skill set comprises those aptitudes that give the entrepreneur the means of realizing the ends of the entrepreneur. The relation between ends and means is evidently problematic, given the circularity in both directions in determining the meaning of the terms, viz. entrepreneurs are people who have entrepreneurial skills on the one hand while entrepreneurial skills are those possessed by entrepreneurs on the other.

Our view of a vast literature on the broad issue of what entrepreneurs do and how such a field should be defined is that within the management and entrepreneurship literature on entrepreneurship there is a general agreement that entrepreneurship is concerned with instrumental action and less with values and ends. This view is in our opinion founded on the Austrian view of human action best exemplified by the work of Von Mises where ends and values of human action are assumed subjective and "input" to the calculating faculty of human rationality. How such rationality unfolds within markets and under conditions of scarcity is the Austrian orientation to the study of economics. It is not surprising therefore that this framework for the study of economics places the entrepreneur centre stage, and no more so than in the work of Schumpeter. This instrumental and action oriented approach underpins current work on entrepreneurship. As Shane notes in the Academy of Management Review "entrepreneurship ..[is].. the identification, evaluation and exploitation of opportunities”(Shane, 2012: 12). Furthermore, as Shane and Venkataraman (2000), entrepreneurship is distinct from the study of the firm[^3] including firm formation, a distinction that is important for policy.

While there is a strong assumption that entrepreneurship is therefore focused upon action, within the wider literature on entrepreneurship there is an awareness of and willingness to go beyond the one-dimensional concept of entrepreneurship qua rationality towards richer models that incorporate a variety of characteristics that affect entrepreneurial behaviour and its outcomes. Of note for policy makers and designers of support for entrepreneurship is a truly vast range of work on entrepreneurial characteristics, what might be called dimensions to entrepreneurial behaviour, and the link outcomes of many kinds, and some degree, a focus upon learning by experience and that, to paraphrase Goethe[^4], to achieve success as an entrepreneur, knowing is not enough, one must apply one's knowledge and experience the role by performing it.

We have reviewed some of this work and believe that a number of contributions are helpful in understanding where scope for policy action may exist: on entrepreneurial orientation (EO) (Dess et al., 2011); goal seeking by entrepreneurs (Ardichvili et al., 2003), the social contexts of entrepreneurship, and broader links to labour, taxation and other polices (Henrekson and Stenkula, 2009); on skills development, the OECD review (OECD, 2010) provides a framework that argues for a unique set of (convergent) skills that involve analysing and making use of and adapting new information to combine with existing information: "Entrepreneurship specific skills for creating and running new business ventures and innovative projects in existing firms, such as risk assessment and warranting, strategic thinking, self-confidence, the ability to make the best of personal networks, motivating others to achieve a common goal, co-operation for success, and the ability to deal with other challenges and requirements met by entrepreneurs" (OECD, 2010: 166).

2.1.1 The Entrepreneurial Nexus

While a large body of core of research centres on the entrepreneur as a key economic actor therefore, there has been a tendency for research to work outward from the entrepreneur to consider other factors and policies, what might be termed the supporting cast of policies which assist entrepreneurs achieve their various objectives. These related policies involve such activities as the provision of finance for entrepreneurship, and advice and financial assistance for the firm. They may include to some extent policies that provide these forms of support in a bundle in either time or space (for example, incubators) or both. The attempt to designate areas of action as relevant has been extensively widened and, in some cases, the argument for support to entrepreneurship takes on the form of lobbying for action by government are very broad in scope (Kauffman Foundation, 2012). The next section of this report examines how policy makers justify their policies on entrepreneurship and shows where currently, the limits of action lie.

2.1.2 Systematic Approaches

Entrepreneurship policy while having its own specific rationales can also be considered within the context of broader policies to promote economic growth, development and sustainability. The European Union, for example, has sought to develop its framework for entrepreneurship policy across multi-country context and within a single programme, the Entrepreneurship and Innovation Programme (EIP), consisting of six streams of activity (see Figure 3). This requires systematic policy coordination at all levels and between all levels. The EIP is an umbrella project in effect - to promote entrepreneurship that covers financial assistance to firms, promoting the Enterprise Europe network, and supporting eco-innovation. The programme is important as a form of support to entrepreneurship because of the scale of funding, although of course the amount of money is to be spent across the whole of Europe. Example of activities (parts E and F for instance) directed at promoting entrepreneurship include: a) the Community programme for the reduction of administrative costs; b) IPR Helpdesks (European Helpdesk on IPR and China SMEs Helpdesk); c) E-skills (study; conferences; seminars; best practice exchange); d) European SME week/ European Charter for Small Enterprises: (conferences; workshops dissemination of good practice and information); e) Information campaign on CE marking; and f) SME performance review (studies; workshops; meetings).

A recent evaluation of the programme (Centre for Strategy and Evaluation Services, 2011) examined its relevance, efficiency and effectiveness. However, the real impacts of the programme in terms of the growth of enterprises, generation of employment and contribution to welfare were not able to be assessed as, at the time of evaluation, the programme was still two and a half years from completion. Nevertheless, the review, largely based on a survey and monitoring with high profile interviews about what the participants and beneficiaries believed were useful, came to a number of conclusions on operational performance and early and intermediate inputs, including the following:

• That the programme was particularly effective at the Europe wide level in addressing the needs, problems and issues (points A-F, diagram 2) set out;
• That the overall objectives were coherent and the implementation processes were integrated measures implemented by member states;
• That it was on track to achieve the anticipated impacts expected of it; and,
• That stakeholders believed the budget and the dedicated resources were at the appropriate level.

Table 1. EIP Budget

2.2 Entrepreneurialism: Policy Rationales

2.2.1 Introduction

Why then should government adopt a policy for entrepreneurship or entrepreneurialism? The economic rationale for public intervention in relation to entrepreneurship rests on distortions and market failures in the economy. These result, for the most part, from information asymmetries (Storey, 2005; Audretsch et al., 2007). We consider these policy justifications in turn below, outlining the conditions that give rise to the market failures and what solutions have been proposed. We believe on reviewing the literature that the first two of these arguments provide the rationale for entrepreneurship policy. Access to finance and R&D and innovation related market failures provide the justification for other forms of policies not considered here.

2.2.2 Lack of Awareness

Firstly, there is a lack of or low awareness among (young) individuals of the potential benefits of starting a business (Storey, 2005). This can be met by for example education policy with an entrepreneurial element that can raise awareness and provide some training to the young to the benefit of an entrepreneurial career. In response to this first form of market failure, educational and cultural policies are justified. These have various targets and can include those already in work, the unemployed, and students. The aim of such policies is to promote a change in outlook and value system to the point that the individuals (and society at large) will become more entrepreneurial. Action within this framework could be specific in the sense of targeted at individuals; but it could also be achieved through general diffusion, what might be termed a broadcast model of transmission.

2.2.3 Information Problems

Second is the failure arising from ignorance on the part of business owners of the benefits of obtaining external advice from experts. Such information imperfections can be used to justify public subsidies directed to management training or facilitating specialist advice for small and growing firms. An important distinction to make here is between support to the firm and support to the entrepreneur. Specialist advice provision, which can be provided through business coaching or mentoring, addresses difficult and often strategic questions that face the entrepreneur. This situation specific advice is nearly always given actively, that is to say, the entrepreneur will not know before the advice is given, what kind of problem they or their company will face. More general advice can be given less actively, and may involve subsidized forms of information provision endorsed by government and therefore of assured quality of a very general kind that reduces the observation and search costs faced by the firm.

2.2.4 Access to Finance

A third type of information failure relates to access to finance. The market for credit is particularly vulnerable to information asymmetries. As you firms have little or no track record and or collateral they tend to be credit rationed in financial markets as financial institutions are unable to fully assess their risk. Intervention in the form of credit guarantee schemes rests on this kind of failure. A different study in the series (Access to Finance) has examined this form of support separately.

2.2.5 The Positive Externalities of Investment in R&D and Innovation

Finally, intervention can be justified when it is necessary to align private and social interests. A public subsidy can be utilised to make it privately worthwhile for the firms to undertake projects that are in the interests of society which would not under normal circumstances provide sufficient returns. The classic example is policies to promote innovation. Without subsidies the formation and growth of technology based firms would be sub-optimal. Thus the potential for positive externalities can be used to justify policies to promote Science Parks or the public funding of seed capital programmes focused on technology-based small firms

2.2.6 When there is doubt over market failure: the case of incubators

As we have noted above, significant public support to innovation is given within the contexts of spatial or location based activities, and in the case of entrepreneurship support, this effort is sometimes directed at entrepreneurs within business incubators, sometimes referred to as incubators. However, the general case cannot be made that incubators can only work with government subsidy because the operation of a significant number of privately funded incubators suggests that there is an effective market for services of this kind. On closer examination of the operation of incubators as has been conducted very recently (Dee et al., 2011), it is clear that there is a continuum along which incubators can be placed in terms of the level of government support they receive, at one end of which there are many incubators that operate without public subsidy.

3 Scope

3.1 Variety of Measures

Consideration of the policy support to entrepreneurship leads quickly to the realization that this is an area where entrepreneurship policies in the strict sense of being directly focused on entrepreneurs are strongly connected to other policies considered by other reviews in this series of reviews for NESTA. These other policies have potential but under researched impacts upon entrepreneurship (e.g. the propensity to start a business, the costs of operating a business, the taxation implications for direct or indirect involvement in entrepreneurial activity), but consideration of these activities in conjunction with entrepreneurship schemes is clearly outside the scope of this piece of work. Moreover, policies that are focused on the entrepreneur in a narrow definition are sometimes provided together, making assessment of impact difficult.

3.2 The Selection Criteria of Financial Contribution

A further issue of subject matter clarification is that of the distinction between finance and non-finance schemes. Some of the schemes that might be classified as entrepreneurship are "free at point of delivery" such as educational programmes in the schools. Such schemes are relatively simple to classify. Children do not have money to pay for these; but moving along the spectrum of measures, it is not so easy to designate a measure as either "financial" or "not financial". The example presented by coaching reveals how difficult the distinction between financial and non-financial is in practice. Coaching, which we believe is an important entrepreneurial measure can be subsidized or free. For this reason we have sought to relax the distinction between financial and non-financial as policy measures frequently fit into either category. Rather than consider these separately we feel it is better to consider the measures as single types.

We note also the importance of distinguishing between firm creation and measures focused on existing entrepreneurs. In order to focus our work, it was decided that we would investigate measures that target would be entrepreneurs and those in the earlier stages of an entrepreneurial career. According we did not look at measures that are more widely focused, i.e. on individuals that are already entrepreneurs.

We have covered a range of measures that seek to address a number of market failures. In particular we focused on policies that support culture change and entrepreneurial education and on policies to ameliorate information asymmetries. Some policies we have examined aim to address more than one goal in that they aim to correct more than one market failure. We call these multiple measures. Amongst the multiple measures considered we have noted a number of policies or programmes to provide financial assistance to firms ranging from venture capital to loan guarantees. These polices have been excluded from this report as they covered elsewhere.

3.3 A Framework for the Presentation of Findings

Our review of findings arranges the measures we have examined according to the market failures with which they deal. As access to finance is covered by a separate study, we have therefore three main categories under which we consider the measures. We draw conclusions then at the policy justification level in the first instance.

4 Summary of findings

4.1 Classification of Schemes

As we note above, our presentation of findings is undertaken by grouping the reviews of measures according to the market failure with which they deal. Mostly, this categorization of the measures and policies by reference to the failure to which they are addressed is satisfactory. We do however have one section where measures that are in combination with other initiatives are considered as this type of measure is relatively common.

4.2 Schemes to Promote Cultural and Behavioural Change

Policy makers the world over have recognised the importance of entrepreneurship in the quest for economic development and support instruments like entrepreneurship education to increase levels of entrepreneurial activity. Entrepreneurship education is considered integral to creating a culture for entrepreneurship and there has been a significant increase in the use of entrepreneurship education in schools, colleges and universities in Europe and elsewhere[^5].

Studies of the impact of entrepreneurial education adopt various proxies for entrepreneurship including intentions to become an entrepreneur, the feasibility of entrepreneurial ventures, and competencies associated with entrepreneurship (Lepoutre et al., 2010). However the results of such studies are mixed. Some studies find entrepreneurship education impacts positively on perceived attractiveness and feasibility of starting new business activities while others find evidence that such effects are negative (Oosterbeek et al., 2010). Such inconsistent results, as noted by a number of authors (von Graevenitz et al., 2010; Lange et al., 2011; Lorz, 2011) stem from various methodological short-comings. Few studies adopt a pre-post design or include a control group. Few take account of the effect of the time lag. If the lag between starting a business and graduating is several years, then there may be other motivating factors that explain the decision to start a business. Finally many studies ignore the effect of self-selected participants predisposed to entrepreneurship which is likely to bias results in favour of educational interventions.

Oosterbeek et al. (2010) evaluate the impact of the student mini company scheme (SMC) on students' entrepreneurial competencies and intentions using an instrumental variable approach and a difference in difference framework[^6]. They draw their data from a vocational college in the Netherlands that offered the scheme, part of the Junior Achievement Young Enterprise programme, at one of two of its locations providing similar Bachelor's programme. The latter provided a natural control group but since students may have self-selected into different school locations. Location choice (and thus treatment) is instrumented to account for changes in outcome variables due to unobserved differences between the students of both locations.

Entrepreneurial competencies were measured using the E-Scan Test (Entrepreneur Scan Test), a validated self-assessment test based on 114 items which are converted (loaded) into 10 traits and skills identified as important in the entrepreneurship literature. While traits may be invariant to the programme, skills such as market awareness for example, can be learned and improved through participation and changes are more likely to be observed. The test and a survey that included questions on background and likelihood of becoming an entrepreneur (intention) were administered at the start of the programme and again at the end both in the treatment location and the control. The results show that the SMC participation did not impact on entrepreneurial intention nor stimulate the skills of students. The effect on entrepreneurial intention was negative and significant. In other words, entrepreneurial intention in the control group was higher than for those in the SMC programme. In addition the effects on students' self-assessed entrepreneurial skills and traits were negative and not significantly different from zero. Oosterbeek et al. (2010) suggest that that the SMC programme may have had a discouraging or 'sorting' effect as in participating, students were able to form a more realistic assessment of both themselves as well as what it takes to be an entrepreneur.

Slavtchev et al. (2012) investigate the impact of entrepreneurship education on students' intentions to follow an entrepreneurial career path. They distinguish between short term intention (immediately after graduating) and long term (five years after graduating). Drawing on data collected from two student surveys in 2006 and 2008 respectively at the University of Otago in New Zealand, they identified students completing both surveys and used a difference-in-difference approach to analyse changes in their entrepreneurial intentions in relation to attending entrepreneurship classes. In order to control for potential selection bias they limited the analysis to students who had no entrepreneurship aspirations either prior to enrolling on the course or who had not previously attended such courses. The difference in difference result showed that the intention to become an entrepreneur (or self-employed) immediately after graduation was 9.1% lower for students taking entrepreneurship courses relative to the control, although this result was not statistically significant. An alternative, but significant OLS estimate showed the intention of the treatment group to larger, 12.5% lower after controlling for academic departments.

Entrepreneurial education however seemed to impact on longer term intentions. Fifty percent (50%) of students taking course developed entrepreneurial aspirations compare to only 21.1% of the reference group. The difference in difference estimate was significant and indicated that 28.9% more students taking the course intended to become entrepreneurs or self-employed in the long term relative to those not taking such courses.

Von Graevenitz et al. (2010) develop a learning model in which entrepreneurship education provides signals to students. These signals enable students to evaluate their aptitude for entrepreneurial tasks. The model is tested using data from a compulsory course on Business Planning at the Munich School of Management for all students on the Bachelor of Science curriculum. Students were surveyed at the start of the course and again at the end of the semester. The survey instruments were largely identical and covered issues such as attitudinal measures and assessments, confidence assessments, course impact assessments and ex ante and ex post entrepreneurial intentions. Answers to questions related to the feasibility of starting a business are used to measure signals students received about their entrepreneurial aptitude before and during the course.

The model enables von Graevenitz et al. (2010) to examine whether the variance and strength of student beliefs that they are entrepreneurs change by the end of the course; and the change in intention resulting from entrepreneurship education. Standard deviations of student's beliefs about their entrepreneurial abilities were computed for pre and post course samples of students who responded to at least one survey as well as three restricted subsets of those who participated in both surveys. The results show that the variance increases in all samples but von Graevenitz et al. (2010) are unable show that they are statistically significantly. However they find a significant increase in beliefs for the sample of students who do not update and who received strong signals before the course.

OLS estimations are used to analyse the relationship between the strength of students' intentions to become entrepreneurs (or avoid entrepreneurship) and several variables including pre-course beliefs, sex, religion and whether parents or friends are self-employed. The results show that the strength of students' intentions is positively related to the strength of pre-course beliefs, consistent beliefs (where signals received before and during the course are either both high or both low) and the interaction between the two, all of which are significant at either the 1% or 5% levels. While not employing a control group, von Graevenitz et al. (2010) argue that it was unlikely that students updated their beliefs on information outside the course since the contents of the course were very specific and not duplicated in other courses. However in line with Oosterbeek et al. (2010), these results suggest that entrepreneurial education strengthens (weakens) the intentions to become entrepreneurs and further that the consistency of signals received affects changes in students' intentions to become entrepreneurs.

4.3 Schemes to Provide Information

4.3.1 Schemes to Provide Advice of a Standardized Form

Schemes providing information and advice include general support to entrepreneurship across a whole range of challenges but there are some which are directed at specific types of business, usually SMES, and firms with particular aspirations. Our first example is of a programme in Canada in the Province of Quebec, the OPREX initiative, that aims to improve export performance of firms, particularly SMEs through A) Training and preparedness activities and services; B) Awareness activities and services; and C) Support/guidance activities and services: The review was conducted (Larivière, 2007) for the Quebec Ministry of Development, Innovation and Exports The evaluation notes the strong interest in the use of the scheme, a strong desire on the part of users of the scheme to continue to use it in the future. The evaluation noted that firms targeted were aware that they did not have the knowledge and understanding to undertake export activities successfully and therefore the targeting by the programme appears appropriate. “As regards export training and preparedness, 62% of trained clients said they were not equipped or were poorly equipped to assess the profitability of their export activities, 79% of them said they were not convinced to assign an employee to the "export" function, and more than half claimed that training would not have an impact on their ability to maintain or increase their exports level over the next two years. We also noted that a high percentage of respondents were unaware of certain existing resources that would be able to support them in their endeavours” (Larivière, 2007: v).

This issue is not however touched upon by the evaluation. The approach to evaluation is according to the Storey framework more of a monitoring exercise in that the assessments of impact are descriptive and do not employ any comparison cases, let alone matching or methods to remove selection bias.

The study by Wren and Storey (2002) examines the impact of the UK Enterprise Initiative, a publicly supported advisory assistance programme provided by private consultants, on the performance of SMEs as measured by sales turnover, employment and firm survival. . The study notes, as others have done, that the amount of government money spent on soft support to firms is large, yet the methods employed in the evaluation of these measures are limited in scope and rarely seek to measure additionality – the true net effect. Wren and Storey (2002) cited the rationale for the programme - market failures based on asymmetric information leading SMEs not to use existing forms of advice from the private sector which creates a justification for the government to operate a subsidized and higher quality service. Relative to a control group of firms, the analysis finds that the scheme encouraged firms to use outside consultants and this had positive impacts on firm performance. Using survival rate as a measure of performance the authors observe no effect in terms of the lifespan differences for smaller firms but medium sized firms appear to have a 4% improvement in their survival rates over the longer run. The selectivity corrected regression analyses indicated that assistance had an impact on sales and employment but this varied by firm size. For SMEs, the net impacts were increased sales (from £127K to £151K) and employment (from 3.2 to 3.9 employees). In the case of medium size firms, annualized growth rates rose by about 10% (from £846K to £921K in sales and from 19.8 to 21.8 employees). For larger firms, the estimated impacts on employment were also about 10% (from 146 to 162 employees), but no significant impacts were found for sales.

The policy implication is that assistance to very small firms may be not as effective a way of promoting growth as supporting the larger SMEs which benefited significantly from the advice given under the Enterprise Initiative. In respect to the significant effects of the scheme in terms of sales turnover and employment, the authors suggest that a displacement effect may be occurring. Thus, positive impacts of the scheme in one area may well result in lower turnover in other areas.

Roper and Hart (2005) evaluate the performance effects of assistance to small firms in England provided by the Business Link (BL) programme over the 1996 to 1998 period. The programme was designed as 'one stop shop' approach to supporting SMEs. The overall objective was to co-ordinate existing SME support services by creating a fully integrated, local strategy for promoting business and enterprise. BL therefore sought to increase the use of business support by small firms, rationalise support provision by removing duplication and improve the quality of support services. Using a sample of assisted and unassisted business to control for selection bias, they found little evidence that BL assistance was being targeted effectively at firms with a track record of rapid prior growth and further that BL assistance over the 1996-98 period no significant effect on firms' sales, employment or productivity growth over the 1996-2000 period. By contrast excluding the control for selection bias results in a positive employment growth from BL assistance over the 1996-2000 period.

In a subsequent study of BL, Mole et al. (2008) investigate the kinds of firms using the advisory services programme for SMEs; the types of firms benefitting most from such support; and the impact of programme participation on sales and employment growth. They use a non-experimental approach with a regression model to control for group differences and difference in difference methods to eliminate potential biases from unobserved variables. A probit analysis showed that that BL participation was greater among younger, limited liability firms and among firms receiving BL contact and informational mailings. OLS results indicated that intensive use of BL assistance had a positive and significant impact on employment growth but not on the growth of sales. A difference in difference analysis showed that relative to control firms, BL firms experienced an overall average employment growth of 4.4% (5.4% for firms with less than 20 workers and 7.6% for firms with over 50 employees). Effects varied across firms according to strategic orientation for firms expanding into new markets employment-growth benefits were higher (11.7%) compared to firms serving current markets (4.1%), while it was 6.7% for firms using formal business planning processes compared 3.9% to those that did not.[^8]

Rotger et al. (2012) uses a matching method to analyse the performance of the Denmark's North Jutland Entrepreneurial Network which provides advice and guidance to individuals engaged in the creation of new ventures. The Network offers 3 levels (L) of counselling products: basic counselling (assessing viability of the idea) provided by the local business centre (L1); counselling with private-sector advisors (eg formulating business plan) (L2) and extended counselling during the start up with private-sector start-up consultants (L3). By disaggregating a sample of participants that received at least L1 advice between 2002 and 2005 into three groups: those using only L1, those using L2 after completing L1 and those using L3 after completing both L1 and L2, and applying propensity score matching, Rotger et al. (2012) estimate the marginal effects of different levels by using a group at one level as the control for a group at the next level (e.g. L1 is the control for L2 and L2 for L3). They consider three aspects of performance: survival; employment and output, and conduct separate analyses for 2002-2003 and 2004-2005 because of a change in the programme in 2004[^9]. The analysis found that L2 participation increases two year survival rates by about 8% and four year survival rates by 5.2% while for L3 participants, two year survival rates increased by a further 7.6% and the four year survival rate increased by 6.4%.

In terms of employees the average effects of L2 participation was 0.5 (for the 2002-2003 cohort), equivalent to one job created for every two firms assisted. However this seemed to be a short term effect as the coefficient was not significant in later years. By contrast, the employment impact for L3 firms while not significant initially, became significant in the second and third years cautiously indicating a strengthening of the impact over the longer term. In respect to turnover, although L2 and L3 impacts are positive and significant, they seem to be moving in opposite directions with the L2 impact appearing to be falling over time from 178,000 DKK one year after participation to 127,000 DKK 3 years after participation while the L3 average effect increased from 225,000 DKK after the first year to 357,000 DKK three years after participation.

Norrman and Bager-Sjögren (2011) have carried out a detailed study of a specific measure, the Swedish Innovation Centre using a matched pair analysis. This methodology provides some detailed insight into the net effects on firms of entrepreneurship support. This programme assisted 'innovators in their absolute earliest phases of development with financial capital, advice and networks'. The duration of the programme was the period 1994 and 2003. The authors believe that the SIC was a typical programme in that its main purpose was 'to create a better innovation climate in Sweden ... where people's attitudes to innovators is positive. And where it is easy for an innovator to receive help to develop his or her concept to a commercialized product or service' (SIC, 2002: 24)" (Norrman and Bager-Sjogren, 2011: 604). The programme was therefore supporting entrepreneurial skill development and at the same time, attempting to realize new technologies with support that included financial assistance. The programme was not confined solely to skills development therefore. The authors note also that the programme did not have clear goals and an explanation of how it would work exactly: "Although SIC is a public programme, the official publications of SIC have not revealed any clearly described ‘programme theory' that includes defined targets against which to measure. Turnover (sales) and number of employees were both mentioned as desirable outcomes (SIC, 2004), albeit no explicit evaluation indicators were formulated. This lack of targets blocks estimates of the impact in the case of goal attainment" (Norrman and Bager-Sjogren, 2011: 606). The analysis carried out in the paper focuses upon the impact of the programme as a whole and does not distinguish between financial support to the firm and support in the form of entrepreneurship education.

Approaching the issue of entrepreneurship from a labour market policy[^10] perspective, Caliendo and Kunn (2011) evaluate the effectiveness of two start-up programmes in Germany: the start-up subsidy' (Existenzgründungszuschuss) programme, introduced for the unemployed in 2003 and the 'bridging allowanceʼ (Überbrückungsgeld) programme, implemented in the late 1980s[^11]. Rather than compare the relative success of the two programmes with respect to the success of the businesses outcomes (output growth, employee growth, etc.) they match the labour market outcomes of the formerly unemployed 'entrepreneurs', those participating in the aforementioned programmes with other unemployed individuals. Outcome variables are 'not unemployed[^12] and 'in paid or self-employment' and personal income. Using propensity score matching and difference in difference analysis, they analyse the employment outcomes of participants for up to 56 months after programmes started.

The analysis showed positive and significant effects over time for the outcome variables. By the end of the period of analysis (56 months) participants in the start-up subsidy (bridging allowance) have a 15.6% (10.6%) higher probability of not being registered as unemployed compared to non-participants. In terms of integration into the labour market, that is being either self-employed or regular employed, the employment probability of participants is 22.1% higher for start-up subsidy group relative to the control group and 14.5% in the case of the bridging allowance participants. Positive and significant differences are also registered for different measures of income. Start-up subsidy (bridging allowance) participants have a monthly working income which is €435 (€618) higher than non-participants while for total income (income from self or paid employment plus transfer payments) the differential is €270 (€485).

A separate analysis of the bridging allowance scheme mentioned above (Oberschachtsiek and Scioch, 2011) focused more specifically on support activities (training and coaching)[^13] that are provided in addition to the financial subsidy to help promote entrepreneurship activities among the unemployed. Using a propensity score matching procedure, Oberschachtsiek and Scioch (2011) compare outcomes between a comparison group, individuals who did not receive the extra support but who nevertheless received a bridging allowance to the treatment group. The key outcome variable for their analysis is the duration of self-employment. Thus they compute two measures: the likelihood of quitting self-employment during the first 36 months and the probability of survival between different time periods. In other words, they investigate whether supports schemes enable the development of entrepreneurial activities. If the schemes work, they there should be fewer exits from self-employment either into unemployment or paid employment.

Overall the analysis shows that the treatment effects (difference between the treated and control groups) are relatively low and in most cases statistically insignificant. This suggests that (on average) additional support does not contribute to an increase in the duration of self-employment. In the case of training, statistically significant effects are found for 'exits into unemployment' so that additional support is associated with an increase in exiting self-employment into unemployment. By contrast, coaching significantly reduces exits into dependent employment; business founders who receive coaching support are less likely to enter dependent employment when quitting self-employment.

Fairlie et al. (2012) used data from Project GATE (Growing America through Entrepreneurship) to analyse the impacts of entrepreneurship training. This was a demonstration project organised by the U.S. Department of Labor and the Small Business Administration (SBA) in which 4,198 applicants were randomly assigned to either the programme group or a control group.[^14] Those in the programme group were eligible to receive free best practice training services while no services under the programme were offered to those in the control group. Data in relation to outcome measures were collected in follow-up surveys at 6 and 18, and 60 months after treatment assignment.

Fairlie et al. (2012) estimate the average treatment effects across the entire sample and suggest that GATE had limited impacts on ultimate outcomes. In terms of business ownership, the initial positive effects of business training dissipated over the long run. While the treatment group had a 5.2% higher rate of ownership at the 6 month survey, in the follow-up survey, this declined to 2.2% (statistically insignificant) and after the 60 month period, the levels of ownership between the groups were identical. The results are similar for other measures of business performance. Taking into account the cost of the programme and the absence of any tangible effects, Fairlie et al. (2012) conclude that entrepreneurship training was not a cost effective way of increasing levels of entrepreneurship.

4.3.2 Schemes to Provide More Specific and Situational Advice

4.3.2.1 Coaching

In recent years, governments have begun to support advice to SMEs through subsidy of business coaching. Coaching can be a diverse form of advice giving, including specific information on such issues as marketing (covered by Wren and Storey, 2002) but also more generalized advice to new business owners whose background and experience in their new roles may be limited. While policy initiatives that support advice giving have been justified on the basis of market failures, other justifications that emphasize the subjectivity of the SME and entrepreneur have also been proposed.

These subjective, entrepreneur based assessments of need are based on Austrian (neo-Austrian) conceptions of economic action. They lead to different views of how SMEs and entrepreneurs should be supported. The neoclassical view that views markets as systems that can be affected by information asymmetries leads to attempts to promote standardized approaches, what Lambrecht and Pirnay (2005: 93) term “objective” and are "delivered in an expert and solution oriented way”. By contrast, the neo-Austrian approach focuses on the need to develop help that is specific, and answers the often unique needs of the entrepreneur.

The methods used in the analysis by Lambrecht and Pirnay (2005) did not employ a true counter-factual analysis and only compared firms that had been helped with typical firms (the comparison using match firms and the taking into account of selection bias was not carried out). Nevertheless, the conclusions suggest support amongst the user firms for the measures. The self-reported additionality is 63% in this study, i.e. 63% said that that without the support they would not have used a private consultant. Regarding future use, 26% of the users firms said that they would use an external consultant again, even if the help was to be given without a subsidy. The study reports three main reasons for the hiring of a consultant: the reasons are: quality enhancement of goods and or services' b) diagnosis of the enterprise' and c) organisational improvement. Modelling of changes over time was done using a comparison with a set of firms who did not use the services. This analysis revealed however that there were no "significant impacts on net job creation turnover or financial indicators" form the use of private external consultancies. While this study is focused on a scheme that provided direct and "operational services in a solution oriented way" (Lambrecht and Pirnay, 2005: 105), the assistance given is likely to have been broad enough to constitute entrepreneurship support. The authors' main conclusion is however that support to entrepreneurs needs to be more customized and focused upon "the entrepreneur and his objectives" (Lambrecht and Pirnay, 2005: 107).

Within the support measures provided by the ERDF funded North West ERDF Operational Programme (2007-2013), a coaching scheme was introduced to promote entrepreneurship. The North West Regional Development Authority area introduced a coaching scheme in parallel with the other RDAs across England and Wales. All RDAs were required to implement such schemes (East Midlands Development Agency, 2005).

A box below identifies the UK government's approach and understanding of coaching and its role in the promotion of entrepreneurship.