GOVERNMENT POLICY TOWARDS ENTREPRENEURIAL FINANCE: INNOVATION INVESTMENT FUNDS *

GOVERNMENT POLICY TOWARDS ENTREPRENEURIAL FINANCE: INNOVATION INVESTMENT FUNDS * Douglas Cumming Director, Severino Center for Technological Entrepreneurship Lally School of Management and Technology Department of Finance and Accounting Rensselaer Polytechnic Institute (RPI) Troy, New York 12180 Web: http://ssrn.com/author=75390 Email: cummid@rpi.edu This draft: 14 September 2005 * I owe special thanks to the Department of Industry, Tourism and Resources of the Corporate Strategy Branch of the Government of Australia for helpful comments and inspiring and sponsoring this research, and to the Australian Venture Capital Association for their helpful support.

1 GOVERNMENT POLICY TOWARDS ENTREPRENEURIAL FINANCE: INNOVATION INVESTMENT FUNDS Abstract This paper analyses 280 Australian venture capital and private equity funds and their investments in 845 entrepreneurial firms over the period 1982 2005. I focus the analysis on the Innovation Investment Fund (IIF) governmental program, first introduced in 1997. In order to highlight the unique aspects of the IIF, I compare the properties of the Australian IIF program with government venture capital programs in Canada, the UK and the US. The IIF program is unique with a focus on partnering between government-private sector partnerships, as described herein. I analyse the performance of the IIF funds along several dimensions: the propensity to take on risk by investing in early stage and high-tech investments; the propensity to monitor and add value to investees through staging, syndication, and portfolio size per fund manager; and the exit success. For each of these evaluation criteria, I assess the performance of the IIF funds relative to other types of private equity and venture capital funds in Australia. The data analysed show in both a statistically and economically significant way that the IIF program has facilitated investment in start-up, early stage and high tech firms as well as the provision of monitoring and value-added advice to investees. Overall, therefore, the data are strongly consistent with the view that the IIF program is fostering the development of the Australian venture capital industry. However, the vast majority of investments have not yet been exited, and the exit performance of IIFs to date has not been statistically different than that of other private funds. Further evaluation of IIF performance and outcomes is warranted when subsequent years of data become available. Keywords: Private Equity; Venture Capital; Government; Public Policy; Entrepreneurship JEL Classification: G24, G28, G31, G32, G35

2 1. Introduction Governmental bodies around the world are paying increasing attention to the subsidization of research and development through venture capital programs. Venture capitalists (VCs) participate in both the funding and development of early stage enterprises in a variety of knowledge-based high-tech sectors (Sahlman, 1990; Sapienza, 1992; Sapienza et al., 1996; Wright et al, 2000, 2001; Manigart et al., 2002a,b,c; Wright and Lockett, 2003; Shepherd, et al., 2003; Shephard and Zacharakis, 2003; Zacharakis and Shephard, 2001, 2003). In both the United States (US) and Canada, approximately 90% of all venture capital investments are made in technology investments. 1 The perceived importance of developing the knowledge based and high-tech sectors has led to a large number of governmental programs that subsidize these sectors through direct government venture capital funds and tax policy (Poterba, 1989a,b; Lerner, 1999; Gompers and Lerner, 2001a,b; Gilson and Schizer, 2003; Kanniainen and Keuschnigg, 2004; Keuschnigg, 2003, 2004; Keuschnigg and Nielsen, 2001, 2003a,b, 2004a,b,c). 2 This paper analyses one governmental program for venture capital: the Australian Innovation Investment Fund (IIF). The unique institutional features of the IIF program make the analysis of the IIF program worthwhile and instructive for academics and policymakers alike. In order to highlight the unique aspects of the IIF, I first compare the properties of the IIF program with government venture capital programs in Canada, Israel, the UK and the US. With reference to prior research on governmental venture capital programs, I argue the IIF program in Australia has a number of distinct structural features that are superior to alternative governmental programs, particularly those in Canada and the UK. In this paper, I analyse data provided by the Australian Venture Capital Association (AVCAL) and the Thompson Financial Venture Economics Database [hereafter referred to as the AVCAL data ] from 280 Australian venture capital and private equity funds and their investments in 845 entrepreneurial firms. The data are used to provide a graphical and descriptive analysis of the venture capital industry in Australia. The data are also used to provide rigorous statistical and econometric analyses of the Australian venture capital market, and the impact of the IIF program. 1 2 For US statistics, see www.ventureeconomics.com; for Canadian statistics, see www.cvca.ca. See Policy Evaluation in Innovation and Technology: Towards Best Practices Organization for Economic Corporation and Development, http://www.oecd.org/document/23/0,2340,en_2649_34409_1822359_1_1_1_1,00.html

3 The details in the AVCAL data enable an analysis of the performance of the IIF funds along several dimensions: The propensity of IIFs to take on risk by investing in early stage and high-tech investments; The propensity of IIF managers to screen, monitor and add value to investee companies through staging, syndication, and portfolio size per fund manager; IIF exit success and share price returns performance of IIF-backed IPO. The importance of these evaluation criteria are explained in detail in section 4 of this paper. For each of these evaluation criteria, I assess the performance of the IFF funds relative to other types of private equity and venture capital funds in Australia. In brief, the data analysed in this paper indicate the following. Prior to the introduction of the IIF program in 1997, there was scant start-up and early stage venture capital investment in Australia. IIFs are 46% more likely to finance seed stage companies than other types of private funds. IIFs also finance riskier companies in high-tech industries with an industry market/book ratio that is on average 1.2 greater than that of the companies financed by other private funds. IIFs on average are more likely to have one extra staged financing round than other types of private funds, more likely to have one extra syndicated investor than other types of private funds, and finance on average 0.3 fewer firms per manager than other types of private funds. These results are all consistent with the view that the value-added advice to facilitate the professionalization of entrepreneurial investee firms is greater among IIFs than their private counterparts. As such, the data indicate IIFs are fostering the development of the Australian venture capital and private equity industry in a statistically and economically significant way. The data in this paper also indicate a comparative dearth of successful exits in Australia. Exit performance and IPO share price returns have been quite similar among IIFs and non-iifs alike. Policy recommendations in regards to facilitating the exiting of investments are briefly mentioned in the latter part of this paper. This paper is organized as follows. Section 2 discusses the rationales underlying government policy towards venture capital, with comparisons across Canada, Israel, the UK and the US. Section 3 explains in detail the Australian governmental IIF program, and discusses briefly other governmental support programs in Australia. Section 4 explains the criteria upon which the IIF program is evaluated in this paper, and the rationales underlying these criteria. Section 5 introduces the AVCAL data used herein, and provides summary statistics, time-series graphs and univariate comparison tests. Section 6 provides a more rigorous statistical and econometric analysis of the evaluation criteria in a

4 multivariate setting. Section 7 discusses and acknowledges limitations associated with the data and econometric tests considered. Section 8 provides policy recommendations on the basis of the results derived from the analysis of the data. Section 9 concludes. 2. Background and Context: Government Venture Capital Programs in the US, Israel, Canada, and the UK Countries around the world have adopted different forms of direct government investment programs in venture capital and private equity. In order to highlight and provide a context for the unique characteristics of the Australian IIF, in this section I review the main governmental programs (but of course not all programs) for venture capital in the US, Israel, Canada and the UK. Thereafter, in the next sections I focus on Australian governmental policy towards venture capital. I provide this comparative material in this paper so that the appropriate context can be placed on governmental programs for venture capital with a global marketplace perspective, and lessons learned from other countries. Subsection 2.1 reviews the rationales for government programs subsidizing the venture capital market. Subsections 2.2-2.5 review the main areas of governmental programs that have been implemented in the US, Israel, Canada and the UK, respectively. Thereafter, section 3 considers the structure and governance mechanisms of the Australian IIF in the context of other related Australian governmental programs. 2.1. Why Are Governmental Bodies around the World Subsidizing Venture Capital, and What Are the Main Policy Instruments? VCs are often viewed as the primary source of capital for inventive high-tech start-up companies (Gompers and Lerner, 1999, 2001a,b). As small high-tech firms are reported to contribute disproportionately to innovation and economic growth (Cosh and Hughes, 2003), policy makers around the world have become increasingly concerned about the success of their high-tech sectors, and the availability of venture capital. It is further widely believed that the social rate of return to venture capital exceeds the private rate of return as the returns to innovation are not fully internalized by the innovating parties (i.e., there exists broader returns to the development of an innovative society) (Gompers and Lerner, 2001a,b). As such, many policymakers around the world have established government support programs to stimulate venture capital financing of innovative ideas and thereby foster economic growth. Governmental program objectives are consistently identified as including the goal of increasing the aggregate pool of capital for entrepreneurs (i.e., not displacing private investors), and the goal of providing support networks from which entrepreneurs receive value-added assistance for financing and growing their firm (Lerner, 1999, 2002).

5 Broadly classified, public policies towards venture capital come in one of two primary forms: (1) law, and (2) direct government investment schemes. Capital gains taxes are widely recognized as being one of the most important legal instruments for stimulating venture capital markets (Poterba, 1989a,b; Gompers and Lerner, 1998; Jeng and Wells, 2000) (but there are other legal instruments for VC markets 3 ). Poterba (1989a,b) shows U.S venture capital fundraising increased from US$68.2 million in 1977 to US$2.1 billion in 1982 as there was a reduction in the capital gains tax rate from 35% in 1977 to 20% in 1982. VCs invest with a view to exit. As entrepreneurial firms typically do not have cash flows to pay interest on debt and dividends on equity, VCs invariably invest with a view towards an exit and the ensuing capital gains. The most profitable forms of exit for high quality entrepreneurial firms are typically IPO and acquisitions (Gompers and Lerner, 1999; Cumming and MacIntosh, 2003; Cochrane, 2005). Therefore, tax policy in the area of capital gains taxation is particularly important for venture capital finance. 4 A second form of government support is via direct government created venture capital funds. Lerner (1999, 2002b) discusses the ways in which government funds can be successful implemented to work alongside private venture capitalists. One of the most important items identified by Lerner (2002b) is the need for government funds to partner with, and not compete with, private venture capital funds. It is also important for government funds to work towards areas in the market where there exists a clear and identifiable market failure in the financing of companies due to, for example, structural impediments in the market that have given rise to a comparative dearth of capital. Further, Lerner (2002b) suggests it is useful for government funds to be structured in ways that minimize agency costs associated with the financing of small and high-tech firms. For example, it is useful for fund managers to have covenants controlling investment mandates and compensation incentives to add value to all of their investee companies; such covenants and compensation mechanisms have worked extremely well in mitigating agency problems among private limited partnership venture capital funds (Gompers and Lerner, 1996, 1999). 2.2. The US Small Business Innovation Company (SCIC) Program The US Small Business Innovation Company (SBIC) Program, administered by the US Small Business Administration (SBA) is the largest government support program for venture capital in the 3 We do not review these various other instruments in this section, primarily for reasons of space. Other legal instruments that have stimulated VC markets include, for example, changes to legislation in the US in 1979 which increased the scope for pension funds to contribute capital to VC funds (Gompers and Lerner, 1998). As another example, in 2002 Australia introduced legislation allowing venture capital limited partnerships to facilitate VC fundraising (Cumming et al., 2005). 4 For theoretical work on tax policy, venture capital and entrepreneurship, see Kanniainen and Keuschnigg (2004); Keuschnigg (2003, 2004); Keuschnigg and Nielsen (2001, 2003a,b, 2004a,b,c).

6 world. SBICs have invested over US$21 billion in nearly 120,000 financings to US small businesses since the 1960s. Investee companies include such successes as Intel Corporation, Apple Computer, Federal Express and America Online. 5 The SBIC Program invested US$7 billion between 1983 and 1997. The SBIC does not distinguish between types of businesses, although investments in buyouts, real estate, and oil exploration are prohibited. In 1998, the SBIC invested US$3.4 billion in 3,470 ventures, approximately 40% by number and 20% by dollar value of all venture capital financings. SBICs are operated like private venture capital funds and are operated by private investment managers. The difference between a private venture capital fund and an SBIC is that the SBIC is subject to statutory terms and conditions in respect of the types of investments and the manner in which the investments are carried out. 6 For example, there is a minimum period of investment for one year, and a maximum period of seven years for which the SBIC can indirectly or directly control the investee company. Investee companies are required to be small (as defined by the SBA) which generally speaking is smaller than those firms that would be considered for private venture capital financing. SBICs also face restrictions as to the types of investment in which they may invest. Capital is provided by the SBA to an SBIC at a lower required rate of return than typical institutional investors in private venture capital funds. Excess returns to the SBIC flows to the private investors and fund managers, thereby increasing or leveraging their returns. Lerner (1999) shows early stage companies financed by the SBIC have substantially higher growth rates than non-sbic financed companies. Overall, the SBIC program has been quite successful; however, as Lerner (1999) notes, welfare implication of the program in relation to SBIC program expenditures have not been fully studied. 2.3. The Israeli Bilateral Industrial Research and Development Foundation (BIRD) The government of Israel supports venture capital through international cooperation with governmental bodies in other countries. The most successful of these ventures has been the Bilateral Industrial Research and Development Foundation (BIRD). BIRD started in 1977 as an equal partnership with the US government. The BIRD Foundation was seeded with US$110 million to fund joint ventures between Israeli and US firms. BIRD provides 50% of a company s R&D expenses, with equal amounts going to each partner. Its return comes from the royalties it charges on the company s revenue. A similar partnership, started in 1994 between Canada and Israel, is the Canada Israel Industrial Research and Development Foundation (CIIRDF). 7 5 6 7 http://www.sebi.gov.in/commreport/venrep10.html These terms and conditions are summarized at http://www.sba.gov/inv/overview.html http://www.ciirdf.ca/.

7 Any pair of companies, one from each country, may jointly apply for BIRD support, if between them they have the capability and infrastructure to define, develop, manufacture, sell and support an innovative product based on industrial R&D. BIRD and CIIRDF often plays a proactive role in bringing potential strategic partners together. In practice, only 25% of the BIRD funded projects have been successful. 8 This success rate is comparable to private venture capital funds (Gompers and Lerner, 1997; Cumming and MacIntosh, 2003a,b; Cumming and Walz, 2004; Cochrane, 2005). Israel s small high-tech companies and Israeli s high-tech economy has been tremendously successful over the past 20 years. Israel s investment in R&D has been among the highest in the world over the past few years (approximately 3% of GDP), and Israel has more than 3,000 technology-based companies. 9 It is noteworthy that Israel has been particularly successful in creating successful high-tech companies that eventually list on NASDAQ (Rock, 2001). One explanation for the Israeli success story is that they governmental support body has created successful international partnerships and networks (although there exist other explanations related to legal conditions, education, training, culture and the like; see Rock, 2001). 2.4. The Canadian Labour Sponsored Venture Capital Corporation Policy makers in Canada have adopted a unique form of government venture capital funds known as the Labour Sponsored Investment Fund (LSIF) (also known as a Labour Sponsored Venture Capital Corporation, or LSVCC). LSIFs were first introduced in the Province of Quebec in 1983, and subsequently in most other Canadian provinces in the 1980s and early 1990s. The UK has adopted an extremely similar type of fund known as the Venture Capital Trust (VCT) in 1995, as described in subsection 2.5 immediately below. Both the Canadian LSIF and the UK VCT are mutual funds listed on stock exchanges. The LSIF and VCT investors are individuals, and they receive substantial tax incentives to contribute capital to this class of funds (more than 100% in Canada, and up to 40% in the UK, both with limitations on the size of capital contributions that are tax subsidized; see Cumming, 2003, and Cumming and MacIntosh, 2005). In exchange for the tax subsidy, LSIF and VCT managers agree to adhere to a set of statutory covenants that constrain their investment decisions and activities. Cumming and MacIntosh (2005) show that LSIFs have achieved extremely low returns (grossly underperforming even a 30-day risk free t-bill index), and argue that these low returns are a 8 9 http://www.sebi.gov.in/commreport/venrep10.html http://www.ciirdf.ca/benefits.php

8 product of statutory constraints that inefficiently constrain managerial choice and investment choice. In particular (depending on the province of incorporation), LSIFs must reinvest up to 80% of their contributed capital within a 1 to 3-year period of the contribution date, and pay severe penalties for non-compliance. This constraint can adversely affect returns in two ways first, by forcing managers to commit capital to inferior investment projects; and second, by attenuating the due diligence process that lies at the heart of equity investing (Gompers and Lerner, 1999) and adversely impacting the quality of the deal forged with various of a LSIF s portfolio companies. Further statutory constraints exacerbate this problem by limiting the allowable types of investments in entrepreneurial firms, imposing constraints on the structuring of investments and the size of an investment in any given firm, and limiting the geographical situs of investee firms. In addition, in contrast to their private venture fund counterparts, LSIF funds must be structured as corporations, sacrificing the various advantages associated with limited partnership form. Cumming and MacIntosh (2005) further show that, like private venture capital funds (Gompers and Lerner, 1999b), LSIF management remuneration consists of a combination of fixed fees (ostensibly to cover out-of-pocket costs) and carried interest (i.e., a percentage of the appreciation in the value of the fund s investment portfolio, typically paid after certain performance hurdles are met). Unlike either mutual funds or venture capital funds, however, LSIF funds typically split the functions performed by the manager between the investment manager and the advisor. The former either invests the funds assets, or advises the fund s board of directors on the investment of assets, and the latter performs a variety of other administrative and marketing functions. Both LSIF managers and advisors are typically well compensated compared to both mutual funds that invest in publicly-traded securities and private venture capital funds that invest in private equity (with an average MER in excess of 4%, which is approximately twice as high for the typical mutual fund that does not receive tax subsidies). The tax generous tax incentive for individual investors to contribute capital to LSIFs insulates LSIF managers from scrutiny with their fee structures, and incentivizes the managers to maximize the number of funds that they manage. Private venture capital fund managers typically finance only a few investee companies at any one time (Kanniainen and Keuschnigg, 2003, 2004; Keuschnigg, 2004; Cumming, 2004). LSIFs fund managers, by contrast, tend to operate as many funds as private venture capital fund managers manage investee companies. In short, the LSIF structure does not facilitate the provision of advice by fund managers to entrepreneurial investee firms. It is noteworthy that LSIFs often invest or hold much of their contributed capital in the form of debt investments, as opposed to equity investments. Many LSIFs hold a significant percentage of their contributed in cash and short-term liquid assets. Cumming and MacIntosh (2004) show the

9 performance of LSIFs has been worse among funds that typically invest a greater proportion of contributed capital in equity investments; conversely, performance has been better among funds that have invested a greater proportion of the fund s holdings as straight debt investments. Despite extremely low LSIF returns, Cumming and MacIntosh (2004, 2005) show that a significant amount of capital has being contributed to LSIFs. Despite earning rates of return lower than most bond indices, by March 2005 LSIFs had collectively accumulated Can$10 billion of capital. Capital contributions have been sufficiently robust that, in the face of an inability to invest contributed capital within statutory constraints, Canada s two largest LSIF funds have both at one time or another suspended or limited capital contributions. Cumming and MacIntosh (2005a) show that LSIFs have received so much capital that they currently control almost 50% of the total supply of venture capital available for investment in Canada, and have displaced or crowded out other types of private venture capital investors, thereby frustrating one of the primary policy goals of the program (namely, the expansion of the aggregate pool of capital). 2.5. The UK Venture Capital Trust (VCT) Tax subsidized venture capital organizations similar to Canadian LSIF UK VCTs have been introduced in the UK (Cumming, 2003). UK VCTs performance characteristics are strikingly similar to that of Canadian LSIFs. In the autumn of 1995, VCTs were introduced to increase the pool of venture capital in the UK. 10 VCTs are publicly traded companies (listed on the London Stock Exchange) that invest in small private companies, and companies listed on the UK Alternative Investment Market (AIM). The VCT investment vehicle is similar in structure to that of other UK investment trusts. 11 The main difference is that the individuals who invest in VCTs receive special tax breaks (detailed in Cumming, 2003). In exchange for their tax status, VCTs face a number of statutory restrictions on their investment activities (these covenants are explained in detail by Cumming, 2003). Overall, UK VCTs are extremely similar to Canadian LSIFs: VCTs and LSIFs are government created funds that exist because of generous tax incentives offered to investors; investors are individuals; VCTs and LSIFs are mutual funds that invest in private equity; VCTs and LSIFs face statutory covenants governing their behaviour in exchange for their tax subsidies. There are differences in the statutory governing mechanisms between VCTs and LSIFs. Broadly speaking, LSIFs covenants do tend to be more onerous than VCT covenants (for details see Cumming, 2003), 10 VCTs were introduced by legislation in Sections 70-73 and Schedules 14-16 of the Finance Act 1995. 11 See, e.g., http://www.trustnet.com/vct/ <accessed 1 March 2005>. See also Cosh and Hughes (1994, 2003) for further information on small and medium sized enterprises in the UK.

10 but the general effect is similar. The tax incentives to invest are also slightly different: LSIFs have a smaller limit for tax deductible investments, but the tax breaks are larger (as explained in Cumming, 2003). The British Venture Capital Association (BVCA) successfully lobbied UK government regulators in 2002 to further facilitate VCT fundraising efforts through the expansion of tax subsidies and tax-exempt contributions. In the five-year horizon to March 2005, median VCT returns were -40.3%, and median LSIF returns were -5% (Cumming, 2003; Cumming and MacIntosh, 2005). In the one-year horizon to March 2005, median VCT returns were +5.8% and median LSIF returns were -4.1% (Cumming and MacIntosh, 2005). The more recent improved one-year VCT performance appears to be directly attributable to an improvement in portfolio valuations from the years immediately prior to the March 2004 March 2005 period (i.e., portfolio valuations were reduced immediately prior to the most recent year, so the improvement in returns may or may not be persistent in coming years). The similarity of returns performance of VCTs and LSIFs indicates that if policy makers adopt LSIFs and/or VCTs in other countries, the effect is likely to be the same. The tax expenditure is extremely large, and the economic benefits from such expenditures do not appear to match the costs. The weak statutory governance structure is consistent with underperformance. 2.6. Summary of Governmental Venture Capital Programs in Industrialized Economies Outside Australia In sum, international evidence indicates it is quite a challenge to design a successful government venture capital program. Government policy towards venture capital in the US and Israel in the form of private/public partnerships and international partnerships appears to have been quite successful. By contrast, the available evidence indicates substantially fewer social benefits of using tax monies to create governmental venture capital funds of the form of mutual funds that invest in private equity, as done in Canada and the UK. As indicated, there are other governmental policy programs in these other countries which have not been mentioned here (primarily for reasons of conciseness); only the primary programs in these other countries have been reviewed. With this background, the next section considers the Australian policy towards venture capital, with particular focus on the IIF program.

11 3. Government Policy towards Venture Capital in Australia, and the Institutional Details of the Australian IIF Program The focus of the analyses in the remainder of this paper is on the Australian Innovation Investment Fund (IIF) Program. 12 The IIF program is one of eight related programs in Australia; other initiatives include the Renewable Energy Equity Fund (REEF) Program, the Pre-Seed Fund (PSF) Program, the Pooled Development Funds (PDF) Program, the Venture Capital Limited Partnerships (VCLP) Program, the Commercial Ready Program, the Commercializing Emerging Technologies (COMET) Program, and the R&D Tax Concession. 13 These related initiatives are summarized by the Department of Industry, Tourism and Resources (2004). Our analysis of the Australian Venture Capital Industry focuses specifically on the Innovation Investment Funds, and not the other forms of government support in Australia. The other direct subsidization programs in Australia comprise similar attributes and a small share of the formal VC market in Australia, so that those programs are controlled for in our empirical analyses below, but not discussed at length for reasons of conciseness. Our empirical specifications discussed in the subsequent sections of this paper consider as many control variables as possible for companion policy mechanisms alongside the IIF program. As well, note that policy mechanisms other than direct support schemes do not play a material role in our empirical specifications. For instance, the VCLP program was introduced in December 2002, and has only led to fundraising among four funds as at the time of preparation of this report (2005); therefore variables are not included for the VCLP program. Associated changes in tax structures for VCLPs and other entrepreneurial activities (such as early stage R&D expenditures) at that time potentially facilitate VC investment, and this timing effect is picked up in our empirical specifications described in the next section. The IIF was established in 1997 in order to stimulate the financing of small high-tech companies in Australia. The objectives of the IIF fund are stated as follows: By addressing capital and management constraints, to encourage the development of new technology companies which are commercialising research and development; To develop a self-sustaining Australian early stage, technology-based venture capital industry; To establish in the medium term a revolving or self funding program; and To develop fund managers with experience in the early stage venture capital industry. 12 The description of the IIF program in this section is drawn from the Department of Industry, Tourism and Resources (2004) 13 http://www.avcal.com.au/html/public/issues.jsp

12 The IIF Program operates in a manner that is most similar to the US SBIC program, as described above in subsection 2.2. The Australian government held two competitive selection rounds in 1997 and 2000, which led to five IIFs being established in late 1997 (and early 1998) and another four being established in 2001. In total, ten-year licences to nine private sector fund managers were awarded on a competitive basis. The first round of the program was announced in the Government s Small Business Statement in March 1997 and provided $130 million [Australian $ unless otherwise indicated], which has been matched on the basis of a Government to private sector capital ratio of up to 2:1. Five licensed funds were established in round one (A&B, AMWIN, Momentum, GBS (formerly Rothschild) and Coates Myer) and became operational during 1998. The second round of the IIF program enabled funding of $90.7 million, and also matched by private sector capital on the basis of a Government to private ratio of up to 2:1. The Government to private capital ratio was a competitive element in the selection of the round two funds. Four funds were licensed under round two (Foundation, Nanyang, Neo (formerly Newport) and Start-up) and became operational in 2001. In total, the nine licensed funds have total capital of $385.05 million, of which the Australian Government is contributing $220.7 million and the private sector $137.35 million. Annual management fees were fixed at 3% of committed capital for the five round one funds and range from 2.5% to 2.8% among the four round two funds. Management fee levels, like government to private capital ratios, were a competitive element in the selection of the round two funds. As with the US SBICs described in subsection 2.2, the Australian IIFs are administered by licensed private sector fund managers who make all investment decisions, subject to the terms of their licence agreements with the Australian Government and other governing documents. Key elements of the IIF program s operating requirements are: the ratio of Government to privately sourced capital must not exceed 2:1; investments will generally be in the form of equity and must only be in small, newtechnology companies; at least 60% of each fund s committed capital must be invested within 5 years; unless specifically approved by the Industry Research and Development (IR&D) Board, an investee company must not receive funds in excess of $4 million or 10% of the fund s committed capital, whichever is the smaller;

13 distribution arrangements provide for: o o o both the Government and the private investors to receive an amount equivalent to their subscribed capital and interest on that capital; any further amounts to be then shared on a 10:90 basis between the Government and private investors; the private investors component to be shared with the fund manager as a performance incentive; the funds established under the IIF program will have a term of ten years, after which they will be closed in a commercially prudent manner. To be eligible for support under the IIF program, investee companies must: be commercialising the outcomes of R&D activities (as defined by the IR&D Act); be at the seed, start-up, early, or expansion stage of development. have a majority of its employees (by number) and assets (by value) inside Australia at the time a licensed fund first invests in the company; and have an annual average revenue over the previous two years of income that does not exceed $4 million per year and revenue in either of those years that does not exceed $5 million. 4. Evaluation Criteria for the Australian IIF Program This section outlines and describes the criteria upon which the IIFs are evaluated in this paper. The data used and empirical analyses are described and presented in the subsequent sections. 4.1. Criteria #1: Start-up and Early Stage Investments There are different stages of entrepreneurial firm development, and investment at different stages entails different degrees of risk and illiquidity. Simply put, the earlier the stage of entrepreneurial firm development, the greater the probability of entrepreneurial firm failure and bankruptcy, and the greater the illiquidity (time to realization through an exit, such as an initial public offering (IPO) or acquisition) of the investment. Non-governmental venture capital and private equity funds may take on an insufficient number of deserving entrepreneurial projects at the earlier stages of

14 development, possibly due to a lack of experience and expertise, risk intolerance, and/or structural features in the market that make it difficult to commercialize a new technology and exit an investment (Berger and Udell, 1998, 2002; Cressy, 2002; Carpenter and Perterson, 2002; Lerner, 2002). As indicated above, therefore, one objective of the IIF program is to increase the number of start-up and early stage investments. Start-up stage entrepreneurial firms are those carrying out research and developing an initial concept for the product and technology. Early stage firms are those companies with product in testing and/or pilot production, and may or may not be generating revenue, and usually have been in business less than 30 months. 4.2. Criteria #2: High-tech Investments High-tech companies include companies which primarily operate in industries characterized by intangible assets (intellectual property, patents, designs, education and experience of the founding entrepreneurs). It is widely regarded that the risk of investing in high-tech companies are more pronounced than investments in non-technology companies. However, there are positive externalities to investment in innovative companies in that the country in which they operate becomes more innovative, such that the social rate of return to high-tech investment is greater than the private rate of return (Gompers and Lerner, 2001a,b; Lerner, 1999, 2002). Because the social rate of return is greater than the private rate of return, it is appropriate governmental subsidization of high-tech venture capital investment, and particularly early stage high-tech venture capital investment. Note that a widely accepted proxy for the extent to which an industry is high-tech is the industry s market/book ratio (Fama and French, 1993, 1995; Gompers and Lerner, 1999). The empirical analysis uses the market / book measure, as well as specific industry variables, in order to study the extent of high-tech investment in the Australian venture capital industry. 4.3. Criteria #3: Staging The frequency with which capital commitments are staged is an excellent and widely accepted proxy for the degree of advice and monitoring in an entrepreneurial firm (Gompers, 1995; Gompers and Lerner, 1999). Funds that stage their investments more frequently monitor and add more value to their investees, thereby lowering the downside risks and probability of unsuccessful exits, and increase the probability of an IPO by professionalizing the entrepreneurial firm (Gompers, 1995; Gompers and Lerner, 1999). The empirical analysis therefore compares the staging frequency of IIFs versus non- IIFs.

15 4.4. Criteria #4: Syndication As with staging, investment syndication is likewise an excellent and widely accepted proxy for the degree of advice and monitoring in an entrepreneurial firm (Lerner, 1994; Gompers and Lerner, 1999; Lockett and Wright, 1999, 2001; Wright an Lockett; 2003). Syndication facilitates due diligence before taking on an investment, as well as advice and monitoring during the period of investment to bring the company to a successful exit in an IPO or trade sale. The empirical analysis therefore compares the propensity of IIFs to syndicate relative to their non-iif counterparts. It is worth pointing out that investment syndication can potentially create conflicts of interest among the investors (as discussed in Lerner, 1994; Cumming, 2004, 2005). For example, different members in an investor syndicate may free-ride of the efforts of another member of the syndicate. Syndicated investors may also involve different members at deal terms which are not completely fair. However, the vast majority of evidence does not support these conflict of interest concerns. For instance, Cumming and Johan (2005) show that if VC syndicate members provide one hour more every month, the VC manager will also spend about an hour more with the entrepreneur; other empirical studies on syndication are consistent with this finding (Lerner, 1994; Wright and Lockett, 2003). As well, international evidence indicates that investments that are syndicated are more likely to perform better in terms of achieving higher rates of return (Brander et al., 2002; Cumming and Walz, 2004). It is also important to point out that in this paper we do not consider co-investment to be a syndicated investment. A co-investment involves different funds within the same organization and run by the same group of managers investing in the same entrepreneurial firm. A syndicated investment involves funds that are completely independent of one another. Value-added advice is potentially facilitated by syndication, but not co-investment (Lerner, 1994; Gompers and Lerner, 1999; Cumming and Walz, 2004). 4.5. Criteria #5: Portfolio Size Venture capital funds that have larger portfolios per manager (in terms of the number of investee firms per manager) add less value to any given investee firm in the portfolio because time spent with investees is diluted (Kanniainen and Keuschnigg, 2003, 2004; Keuschnigg, 2004; Jääskeläinen et al., 2005; Cumming, 2006). In fact, based on an international sample of venture capital investments in Europe, Cumming and Johan (2005) estimate that VCs with one extra entrepreneurial firm per manager in their portfolio provide on average 2-3 hours per month of less support, 20% less advice, and have 0.2 to 0.3 fewer disagreements with entrepreneurs.

16 4.6. Criteria #6: Exit Vehicle Outcomes Exit performance is a clear and objective way to assess the entrepreneurial firm performance and the success of its venture capital financiers. IPOs and acquisitions are widely regarded as the best exit outcomes (Gompers and Lerner, 1999; Cumming and MacIntosh, 2003; Cumming, Fleming and Schwienbacher, 2005; Cochrane, 2005). In fact, IPOs are typically viewed as more favourable than acquisitions, since the entrepreneur regains control over the firm in an IPO exit, whereas the entrepreneur is typically ousted from the firm in an acquisition exit (and entrepreneurs tend to prefer to regain control regardless of financial gains; Black and Gilson, 1998). Write-offs (liquidations) are of course the worst form of exit. It is important to note that it often takes years before a venture capital investment can be brought to fruition. Investment duration from the time of first investment to exit can take anywhere from 2-8 years, and possibly longer (Gompers and Lerner, 1999, 2001a,b). Performance measures prior to exit are at best questionable, as actual performance may differ substantially from interim reported performance (Cumming and Walz, 2004). As such, it is difficult to fully evaluate the overall performance of a fund that has not yet been fully divested. Hence, IIFs should be reviewed again at the end of their lifecycle. 4.7. Criteria #7: Share Price Returns Performance of IPOs Exits The share price returns performance of IPO exits is also a clear and objective way to assess the entrepreneurial firm performance and the quality of the venture capital support provided to the investee firm prior to the exit (Gompers and Lerner, 1999). Share price returns performance over the long run (over horizons of at least a year) after the IPO will be superior among firms that had higher quality venture capital support prior to the IPO, all else being equal (Gompers and Lerner, 1999; Brav and Gompers, 1997). Finally, it is important to point out that IPO performance on average tends to be worse than the market over the long run; that is, IPOs are overpriced in horizons of 1-3 years (Ritter and Welch, 2002; Ritter, 2003). As such, negative returns following an IPO after 1-3 years is not an unexpected finding, particularly in the aftermath of the Internet bubble. 5. Data and Summary Statistics This paper makes use of data provided by the Australian Venture Capital Association (AVCAL) and the Thompson Financial Venture Economics Database [ the AVCAL data ]. The data

17 comprises 280 Australian venture capital and private equity funds and their investments in 845 entrepreneurial firms. In this section, I provide a graphical and descriptive analysis of the venture capital industry in Australia. The next section (section 6) provides rigorous statistical and econometric analyses of the Australian venture capital market, and the impact of the IIF program. 5.1. Graphical Presentation of the Data A time series of all first-round investments (excluding staged financing rounds) is provided in Figure 1. While the extent of coverage for all venture capital and private equity investments in the AVCAL data is unknown, the AVCAL data provide the most comprehensive look at the history of the Australian venture capital and private equity industry. For example, it is known (Department of Industry, Tourism and Resources, 2004) that the IIFs had financed 66 firms as at 30 June 2004, and the AVCAL data comprise 55 of those 66 firms as at 30 June 2004 (and 57 investments in total including investments up to 2005(Q1)) (the difference is due to incomplete reporting to AVCAL). Moreover, the AVCAL database comprises investments from all of the 9 IIF funds, and the vast majority of private equity and venture capital funds in Australia. The profile of Australian investments over time is quite consistent with patterns observed in the US (Lerner, 2002), Canada (Cumming and MacIntosh, 2005), and Europe (Armour and Cumming, 2005). Venture capital and private equity investments around the world showed a drastic increase in 1999 and 2000 leading up the beginning of the end of the bubble in April 2000. [Figure 1 About Here] Figure 2 provides further details about the stage of investment at the time of first investment for the time series of all venture capital and private equity investments. The definitions of the stages of firm development, among other variables used in the empirical analyses below, are provided in Table 1. Figure 2 highlights as well the start date of the IIF investments in the two rounds. It is very noteworthy from Figure 2 that hardly any start-up and early stage investments existed in Australia prior to the introduction of the IIF program. [Figure 2 and Table 1 About Here] Figure 3 provides a time series profile of the start-up and early stage investments by the identity of the investors. Four categories of investor types are indicated in Figure 3: IIFs, private

18 funds part of venture capital organizations that are associated with IIFs, 14 other governmental program funds (indicated briefly in section 4 above), and non-governmental associated funds. This graphical presentation of the data shows non-governmental funds scantly invested in start-up and early stage companies prior to the IIFs. Non-governmental IIF investments were, however, quite a significant part of the market in 2001. Below, I more rigorously assess which type of fund is more likely to invest in start-up and early stage firms. [Figure 3 About Here] Figure 4 presents a graphical look at the exits data by exit type. A clear limitation of the data is that exits other than IPO exits are not represented prior to 2002. It is noteworthy that the time series of IPO exits is quite dissimilar to that observed in the North American and European venture capital and private equity markets. In other developed countries, venture capital IPO exits were much more common in 1999 and 2000 than the period following the crash of the bubble (Lerner, 2002). By contrast, the AVCAL data indicate that the Australian venture capital market was not sufficiently developed to have as pronounced a boom in IPO exits in the period leading up to the peak of the Internet bubble, and that the drop-off in VC-backed IPOs was only observed in 2001. [Figure 4 About Here] Figure 5 shows the time series of venture capital-backed IPOs by investor type. As in Figure 3, four categories of investor types are indicated in Figure 5: IIFs, private funds associated with IIFs, other governmental program funds, and non-governmental associated funds. The majority of IPO exits appear to have been derived from non-governmental funds. This is expected, as the primary governmental funds (such as the IIFs) were introduced in the recent past, governmental investments are in earlier stage firms (which take longer to bring to fruition in an exit), and many investments have yet to be exited. 15 [Figure 5 About Here] Figure 6 presents the average share price returns of the venture capital-backed IPOs. It is important to point out that these returns are not the returns to the investors from taking the company 14 Seven of the 9 fund managers that operate IIFs also contemporaneously operate at least one other private fund. 15 The Department of Industry, Tourism and Resources (2004) indicates Of the 66 [IIF] investee companies, 52 were still active under the program, with 14 having been merged, divested or written off. Most of the write-offs occurred during 2001 and 2002 as a result of the general fall in the value of technology stocks and there have been none since May 2002. Some of the currently active companies are co-investments, investees of more than one of the licensed funds.

19 public. Rather, these returns are the share price returns from the end of the first day of trading until 30 June 2004. As well, note that 12 of the 55 IPOs were delisted. The returns calculations have been done on the basis that the returns to delisting have been -100% (which may overstate the degree of poor performance, as investors may have been compensated with some value prior to delisting). On average, returns performance is quite negative for all types of funds. This is expected, and similar to IPOs in other countries that went public over contemporaneous periods. 16 [Figure 6 About Here] 5.2. Summary Statistics Table 2 summarizes the AVCAL data by the investee firms. In total there are 845 investee firms that were financed by VCs in Australia. The AVCAL database separates the funds into different types, and these fund types are indicated and defined in Table 1. Likewise, the different entrepreneurial firm characteristics are also identified and defined in Table 1. Table 3 summarizes the data by the investee funds (i.e., averaged values for different fund types). [Tables 2 and 3 About Here] Table 4 provides univariate comparison tests for IIFs versus funds affiliated with IIFs, and for IIFs versus completely unrelated funds. It is useful to investigate comparison tests to examine the differences in proportions, means and medians for each of the evaluation criteria enumerated above in section 4 for IIFs, funds affiliated to IIFs, and for non-iifs for a variety of reasons. These tests indicate the trends in the data and provide preliminary insight into the degree to which the data support value-added provided by IIFs as reflected in the evaluation criteria. The limitation of these tests is that other factors are not controlled for ; that is, the IIF characteristics might be correlated with some other variable, and if so, the attribution of quality or success to an IIF might be incorrectly assigned. This limitation with the comparison tests is overcome with the multivariate tests and econometric and statistical analyses in the next section (section 6). [Table 4 About Here] The univariate comparison tests in Table 4 indicate the following: 16 For example, in Canada over the 1997-2003 period, the median 1-year IPO return was -0.045 for TSX IPOs (on the senior Toronto Stock Exchange), and -0.375 for TSX-V IPOs (on the junior Toronto Venture Exchange).