HOW GOVERNMENTS GO WRONG: BAD IMPLEMENTATION
Even if a program to encourage entrepreneurship is well conceptualized, things can still go wrong once it is begun. The implementation of these programs requires many decisions. While decision making about programs may seem like an obscure, even arcane topic, it is incredibly important. As we’ll see from many examples in this chapter, program administrators can make seemingly reasonable decisions that turn out to be destructive.
This chapter will consider three of the most common errors in implementation. Ignoring the need for well-directed incentives, not evaluating what is happening with the program, and failing to allow beneficial internationalization are all mistakes that can be extremely costly, as we will see in the pages that follow.
NOT WORRYING ABOUT INCENTIVES
In addition to providing a clear signal of where the market sees the greatest opportunities (a benefit we discussed in the previous chapter), matching funds have another advantage. If a significant share of the matching funds comes from the managers themselves, they are likely to focus on making sure the investments do well. Yet in many cases, overseers of public entrepreneurship initiatives have not demanded such provisions, and the results have often been disastrous. In particular, the people receiving the funds may adopt a “Heads I win, tails you lose” mentality, which leads to unfortunate outcomes.
Experienced investors in entrepreneurial firms pay an enormous amount of attention to the incentives that entrepreneurs have. For instance, entrepreneurs are frequently forced to accept low salaries—less than they could make in corporate positions. They are not allowed to sell their equity until the investors have liquidated their shares. Other positions that generate income are strictly limited.
One might interpret these restrictions as an indication that well-heeled angel investors and venture capitalists are simply exploiting the entrepreneurs. But their purpose is quite different from mean-spiritedness. The primary concern of investors is ensuring that entrepreneurs “do the right thing”: that is, the capital-providers want to ensure that managers take steps that maximize the value of the firm, rather than just benefit themselves. Because it is so hard for even the most diligent investor to oversee all the actions of an entrepreneur, incentives must be correctly aligned. And one of the critical ways to connect the interests of the entrepreneur with that of the firm is to limit the entrepreneur’s ability to cash out before anyone else does.
This concern about incentives is seen when it comes to financing venture capital firms as well. Sophisticated investors in venture funds, such as university endowments, make sure that perverse incentives are avoided. For instance, if the venture capitalist is contributing a substantial share of the capital that the fund is raising, and getting rewarded primarily in the form of a share of the capital gains, seasoned investors will likely be comfortable taking part in the fund. Conversely, if the venture team stands to get rich from their management fees whether the investments succeed or fail, the investors will be much less enthusiastic.
Unfortunately, governments have not always thought as carefully about incentives before establishing entrepreneurship and venture initiatives. Far too often, the programs have been designed so that the private sector participants do well, no matter if the investment generates a good return or not. Alternatively, investments may be linked to the fund’s financial returns, but not to the broader objectives that motivated the launching of the initiative.
Many examples can illustrate the real danger that the fund managers will have the wrong incentives. The Discovery Fund, for instance, was a $76 million fund organized by New York City in 1995, with funding entirely from the public sector and public utilities that focused on doing business in the city.1 The city hired a local venture group, Prospect Street Ventures, to run the fund, which was launched with a great deal of fanfare, including Mayor Rudy Giuliani’s pledge that it would generate 4,000 jobs.
Yet the effort is generally regarded as a failure. While the fund did make some successful investments, such as About.com, many more were failures. Moreover, several decisions seemed puzzling even at the time they were being made—such as leading a $14 million financing round in the web-based broadcaster Pseudo Programs, which was quickly squandered by the firm (which lacked seasoned management) before it went belly-up.2
Beyond the questionable financial returns, questions were raised about the extent to which the fund advanced the city’s social goals. For instance, the fund invested in at least two companies not based in New York City at all. About $3 million was invested in Bondnet Trading Systems, a Connecticut-based operator of an Internet securities trading service. Bondnet was liquidated, and its assets sold off, in 1997. Even some of the New York–based investments seemed to have very limited economic benefits: for instance, the fund’s first deal was putting about $2.5 million into Skyline Multimedia Entertainment, an already publicly traded entity that operated a virtual-reality ride at the Empire State Building and an arcade near Times Square. The fund enabled the firm to begin work on a new virtual-reality ride in Sydney, Australia, a project that soon collapsed. Around the same time, Skyline Multimedia made substantial loans to one of Skyline’s executives. Within a couple of years, the company was trading at only a few percent of the price at which the initial investment was made—though not before the Discovery Fund had invested another million dollars in the firm.
It may have been that at the time of the fund’s formation, New York City–based digital companies—the fund’s putative focus—were underfunded. But the bulk of the investments were made during the bubble period of 1998 through 2000. During these years, local Internet and digital media sector received huge amounts of capital from independent and corporate venture funds: it is hard to believe that any kind of market failure was being addressed. Critics wondered whether the compensation scheme worked out for the fund exacerbated the problems. For instance, the New York Times noted:
Among the major beneficiaries of the fund’s activities have been the people who run it. Executives at Prospect Street Ventures, the New York-based venture capital firm paid by the city to operate the fund, have also been compensated in cash and stock option grants by companies in which the fund has invested.3
It was natural to wonder whether the lack of demand for matching funds and the failure to set a mandate that matched the city’s economic development needs intensified the problems that the fund encountered.
An even more extreme example is the Heartland Seed Capital Fund, an initiative by the State of Iowa to spur local activity.4 The state, seeking to boost venture activity in the region, decided in 1990 to create a $15 million venture fund, which the state’s Public Employees Retirement Fund agreed to fund. But rather than looking for a situation where the incentives would be aligned, the retirement fund apparently selected a group in a classic procurement approach: it issued a lengthy request for proposals and waited for venture groups to respond! McCarthy Weersing, a venture group based on the Atlantic and Pacific coasts—but with no experience or personnel anywhere near Iowa—applied and was selected.
Things soon turned ugly. The venture fund charged a hefty management fee of 3 percent per year (between 1 percent and 2.5 percent is more typical for venture funds), but despite the steady fee stream, seemed unable to find any attractive investments to undertake. (The state later pointed out this might have been easier had the venture fund assigned an investment professional to be in Iowa full time. A similar fund in Indiana, which had an active office there, was making a steady stream of promising investments.)
Once again, the incentives were not well thought through. In particular, the venture group got a hefty management fee whether it made investments or not. And indeed, after three years, the fund managers had collected $1.4 million in fees—yet only had made one investment of $1 million. At this point, the fund requested another half-million dollars from the state: not for investments, but to cover its management fees for its fourth year of operations.
The state now suggested that things were not working out and that perhaps the fund should be amicably dissolved. The venture investors, apparently enraged by the loss of their meal ticket, took a far less congenial approach. First, they sold much of the equity in their sole investment back to the management of the company, and used the proceeds to pay themselves the management fee that the state had refused to pay. Then McCarthy Weersing sued Iowa, demanding not just all the management fees that it would have received had the fund remained in business for a decade, as the original agreement proposed, but also for all the profits that would have accrued, had the venture group been able to find and fund successful transactions!
Unfortunately, these experiences are more the rule than the exception. Even programs that appear to be widely accepted as models often have deep design flaws. For instance, Louisiana in 1983 introduced a “CAPCO” (certified capital company) program, in which insurance companies received huge tax subsidies for setting up venture funds.5 As with the Canadian labor funds discussed in chapter 6, there were few incentives for the insurers to worry about the quality of the managers hired to run the funds, or mechanisms for the public sector to intervene should the investments prove to be flawed. Yet despite these fundamental incentive problems, states from New York to Wisconsin have emulated the structure.
The same sad truth emerges from a systematic study of the performance of almost 7,600 investments in venture capital and buyout funds by institutional investors.6 Among the questions the study examines is the performance of investments when the investor and the fund are located nearby each other. In studies of public markets, a typical finding is that there is “home-field advantage”: investors do better when investing in local stocks. Thus, an investor based in Peoria is likely to get better returns from investments in Caterpillar than from those in Komatsu, presumably because proximity brings insights that distant investors don’t have.
But when we look at venture and other private equity investments, exactly the opposite is true: when putting dollars to work in funds that are nearby, the investor ends up earning returns that are between six and seven percentage points lower each year. What can explain this dramatic disparity from the pattern seen in public markets? Shouldn’t the added insights from proximity be valuable in this context as well? When we look closely, we see that this surprising result is not a general pattern: it is driven almost entirely by public investors (such as pensions and public university endowments) investing in the state or political jurisdiction in which they are based. Presumably, the fund managers are pressured to help out the region by stimulating local entrepreneurial activity. But these seemingly reasonable requests lead to a lowering of standards and inattentiveness to the investors’ incentives, as we saw in the New York City and Iowa examples. All in all, when public authorities channel investment monies to local entrepreneurs and funds, they are following a recipe for mediocre returns.
Clearly, paying careful attention to the incentives offered participants in public programs is essential. Do the entrepreneurs or venture capitalists stand to gain no matter how the investment turns out, as we have seen in some of the examples above? Or do they have a powerful financial stake in the success of the fund? Are the participants driven only to maximize financial returns, or are they steered to address the broader social objectives of the program as well? What if things start going wrong—will they have incentives to stay the course, or instead to behave in a reckless manner? These considerations are critical in the design of investment programs.
THE NEED FOR EVALUATION
Just as venture capital investors carefully analyze the track record of entrepreneurs they are considering funding, government officials should examine the track record of the venture capitalists and entrepreneurs who may receive public funding. Moreover, it is important to look critically at the programs themselves. Far too often, public venture capital programs support underachieving funds and firms. Participants are allowed to linger without a vigorous evaluation.
Evaluations of private sector participants should emphasize two sets of criteria. First, of course, is performance: how well has the fund done to date? But in many cases, evaluating the performance of a venture capitalist or an entrepreneur is not possible until the group has been active for a decade or more.
Thus, it is important to also look at other characteristics that appear to be highly correlated with a participant’s ability to achieve its goals. These include the experience of the team, the presence of a clear product market strategy among funded firms, and a strong desire to seek private financing. By devising new methods to search for such factors, government officials would be better able to distinguish between high-performing and underachieving organizations.
For entrepreneurs and early-stage venture capitalists, a limiting factor is experience. A growing body of research suggests that the surest way to be a successful entrepreneur is to have already run a business (hopefully successfully). It thus comes as no surprise that when experienced venture capitalists sink substantial funds in a company, they often place their own handpicked manager in charge. Similarly, there is a lot of “learning by doing” in venture capital: even those venture capital fund managers who may have accumulated business experience as consultants or as members of large organizations are frequently at a disadvantage when compared to those who have invested in young firms before. The successful operation of an early-stage fund can demand very different management skills than those garnered as a consultant or manager. Because much of the know-how needed for guiding and managing start-up companies can be gained only through experience, the presence of entrepreneurs and fund managers who do not have this background can significantly undermine these initiatives’ ability to succeed.
Another telltale characteristic of underachieving actors is distractions that undermine their ability to focus on their mission. Legal troubles, for instance, can divert substantial amounts of human and financial resources and even cause dramatic changes in the size and structure of the company. And when an organization is ready to raise more capital, the concerns over pending legal battles often impair the company’s ability to attract outside investment dollars.
Research on public programs in a number of nations indicates that another characteristic of underachieving entrepreneurs and venture groups is grants from numerous government sources, with few tangible results to show from previous awards.7 Because a lack of results can easily be attributed to the high-risk nature of technology development, many organizations are able to avoid accountability indefinitely. These grant-oriented organizations are able to drift from one government contract to the next. Such companies appear to treat public venture capital funds in exactly the same manner as other government research grants: venture-oriented funding neither shows notable returns nor meets unique program goals. While government research grants may be a valuable source of financing to small firms, it is important that they be administered separately, using criteria quite different from venture initiatives.
Adding to the problem is that companies with government grant experience appear to have advantages over other firms when applying for future public awards. Past grants, regardless of project outcomes, help a company gain legitimacy in an area of research, as well as acquire the equipment and personnel needed to do future work. There is also a tendency for some government programs to try to piggyback on other government programs, hoping to leverage their grant dollars. In addition, companies gain insight into the grant application process with each proposal they submit. These organizations consequentially often have a greater chance than others of being awarded government grants. The result can be a stream of government funding to organizations that consistently underachieve.
These performance-undermining factors highlight the need for government officials to critically evaluate each company or fund as a vehicle for accomplishing its goals. This evaluation should go far beyond a simple assessment of the feasibility of a business plan or private placement memorandum. In fact, legal troubles, a long history of government grants, or a lack of germane experience will not even be exposed in the written proposal to the government. It is tempting for evaluators, of course, to attribute the failures resulting from such factors to the high-risk nature of the entrepreneurial process. But to a large extent, organizations exhibiting a high potential for underachievement can be weeded out by placing a greater emphasis on telltale signs during the selection process. The technologies in a venture group’s portfolio may be high risk, but the risks of the entrepreneurial team or venture firm itself should be minimized. Regardless of how innovative or enabling a venture capital portfolio may be, if these undermining factors are present, a fund will be hard pressed to succeed.
In sum, research suggests that government officials should closely scrutinize companies and venture funds participating in public programs. Underachieving firms can be weeded out if government officials conduct a comprehensive evaluation of an organization’s past performance and examine the tangible progress attributable to previous public funds.
One illustration of the failure to evaluate companies was the experience of Celltech in the United Kingdom.8 The biotechnology firm was initially funded by the Labour government, which controlled Parliament at the time of the initial investment in 1980. The investment was in part a response to general consternation that the nation was falling behind America in this emerging sector. (These pressures were apparently sufficiently strong that the Conservative Thatcher government, which came to power soon thereafter, acquiesced to the venture.)
Funded initially by the state-run National Enterprise Board, Celltech was also “offered” capital—under government pressure—by a number of leading financial and industrial corporations. The government’s Medical Research Council gave the biotechnology company a right of first refusal to license all genetic engineering and related discoveries coming out of the laboratories it ran. This decision was widely seen as a way for the Council to diffuse criticism about its track record in the 1970s, when its laboratories made some of the foundational discoveries that underlay the biotechnology revolution, but then neglected to patent them. (This agreement was ultimately renegotiated after half a decade, after persistent complaints from other biotechnology firms, which argued that their inability to license public technologies was handicapping their progress.)
Yet despite its generous public backing and preferential access to technology, Celltech proved remarkably unsuccessful for much of its history. To be sure, the firm hired cutting-edge British academics and built modern laboratories, and entered into alliances with leading British firms. But after a decade, the firm’s technologies were little closer to commercialization than on the day Celltech started. Outsider investors blamed a management team too focused on scientific research and a lack of accountability on the part of the government and related shareholders. Had an American venture-backed firm performed so poorly, they argued, it would have been shuttered long ago.
A new management team and a public offering put Celltech on a more stable course. Even here, though, the consequences for the British industry were mixed. One of the keystones of the firm’s strategy was to exploit its access to the capital markets to acquire smaller British firms. While in some cases these deals led Celltech to products that it could develop and market, in many instances (such as Celltech’s acquisition of RibosePharm and merger with Chiroscience) these transactions were counterproductive.9 In some instances, the acquired company was spun out again in weaker shape a few years later; in other cases, progress in the development of the drugs of the acquired unit slowed or stopped.
Ultimately, Celltech was acquired in 2004 for a little over two billion dollars by a little-known Belgian drug company UCB (Union Chimique Belge). Many observers saw this ending as disappointing, For example, Aisling Burnand, chief executive of the UK Bioindustry Association noted, “One thing [the] Celltech [acquisition] definitely shows is that biotech company valuations are way too low, far below the real value in UK biotechnology.” It is, of course, impossible to examine the parallel universe in which the British government decided not to pursue the Celltech initiative. But it is natural to wonder whether the preferential access to, and subsequent acquisitions of, promising entities by this troubled firm did not end up costing the British economy billions of pounds in lost economic opportunities.
A corollary to the necessary evaluation of potential recipients of funding is the need to evaluate public programs themselves on a periodic basis. These should be rigorous and dispassionate analyses of the programs’ success to date. The evaluations should also consider the overall venture capital climate, and whether the economic rationales that justified the program’s creation still apply.
The United States provides a clear example of the consequences of the failure to evaluate venture capital programs systematically. Many observers argue that the U.S. venture capital market is overfunded, and that the industry would have far higher returns if some of the more marginal groups were to exit the industry.10 Certainly, the pool of venture funds today is many thousand times what it was when the SBIC program was established in 1958. Moreover, many of the participants in the SBIC program in the past two decades have been precisely these marginal firms, whose mandates do not differ appreciably from more established groups but are simply unable to raise capital from traditional sources owing to their shaky track records. Given these facts, it is proper to ask whether the program, however valuable in its initial manifestation, has outlived its usefulness. But the program has had no systematic evaluation over the years and remains politically popular.
Instead, lobbyists have repeatedly pushed Congress to hold hearings to consider a major expansion of the program. For instance, in the late 1990s, as the venture capital bubble was expanding, SBIC advocates made a major push to encourage the formation of a new quasi-public entity to be called the Venture Capital Marketing Association, or Vickie Mae. Modeled after the mortgage giant Fannie Mae (now, of course, a basket case), Vickie would have bought investments from SBICs that they could not take public or sell off to corporations. Instead, with an implicit government guarantee behind it, the new entity would sell bonds backed by these hard-to-sell firms (frequently called the “living dead” or “zombies” by venture capitalists). The arguments advanced by William Dunbar, at the time the head of Allied Investment Corporation, an SBIC, were representative of the arguments its advocates offered:
Creating Vickie Mae will allow the SBIC program, one of our country’s most successful programs, to reach its full potential in helping America’s entrepreneurs create the jobs and technologies that are the foundation of America’s greatness.11
There was no effort in this testimony, or in other arguments by program advocates, to review the SBIC program’s strengths and weaknesses, and determine whether the added subsidies were needed.
Fortunately for all of us, Congress—at the time controlled by free-market-leaning Republicans—did not buy the argument. One can only imagine how much worse the bubble years of 1999 and 2000 would have been had some of the most problematic venture firms been making investments secure in the knowledge that they could pass off onto a government-guaranteed corporation any firms that they could not take public. It’s also likely that once the bubble burst, and Vickie Mae had to make good on its obligations, the value of these zombie firms would have shrunk dramatically, and the taxpayers would have been left making up the difference.
Of course, this systematic failure to undertake careful evaluations is not just an American or British phenomenon: it extends across the world. Over the course of the 1990s and first half of the next decade, for instance, various Chinese municipalities and provinces launched dozens of venture funds.12 Many of them represented significant resources for the still-developing nation, such as the 650-million-renminbi fund launched by Jiangsu Province in 1992. But there was virtually no effort to evaluate the success of these early efforts, even if the initiatives faced incredible challenges: limited attractive investments, inexperienced deal teams, and the absence of many of the legal frameworks critical to making venture capital work. It is precisely at these times that assessing the consequences of venture efforts might be the most valuable.
The frequency of failure to evaluate makes the Israeli experience that much more striking. As we’ll discuss later in this chapter, the Yozma program catalyzed the spectacular growth of entrepreneurship and venture capital in Israel. Nonetheless, after five years, the government examined the program and decided to auction off the ownership of Yozma.13 This sale did not represent a shift of fashion, but rather was a planned step to be taken once the market had sufficiently matured. This ability to recognize when programs are no longer needed, and when scarce resources should be allocated elsewhere, is more the exception than the rule.
This brings us to a subtle problem: even if there are evaluations, they may look at the wrong things. In particular, far too often evaluations have relied on the compilation of success stories. Even organizations as august as the National Academy of Sciences have compiled assessments that consist of little more than anecdotes about firms that received government funding and then had commercial success.14
Not only can these misdirected evaluations lead to the wrong decisions about continuing programs, but they can result in programs that are less effective than if no evaluations had been done at all! How can this be? If the people evaluating programs are looking for success stories, the officials running the programs may select firms based on their likely success. In this case, they can claim credit for the happy endings. But this often translates into funding companies that don’t need government funds. In the language of economics, the pressure of evaluation may drive program managers to fund companies for which the marginal contribution of public funds is very low.15
The Advanced Technology Program (ATP), whose failure to be flexible we discussed in the previous chapter, provides a cautionary tale in this regard.16 The initial notion was to target generic, precompetitive technologies that the market had failed to fund. Thus, the idea was to create a diverse array of technologies in a variety of neglected fields.
Over time, however, the mission of the program mutated in an unpromising way. In the late 1990s, bureaucrats decided to target particular industries for focused grants. The industries they initially chose were Internet technologies and genomic sequencing. But these were not two randomly chosen sectors: at the time the program made its decision to proceed, they were flooded with money from investors. The promise of finding the “next Amazon.com” was leading everyone—from dentists to major corporations—to throw money at this sector. (The number of for-profit incubators catering to helping young Internet firms climbed from about 25 in 1997 to 320 in 2000.)17 And the sequencing of the human genome had excited venture capitalists’ imaginations about the possibility of curing long-standing diseases, leading to a smaller but still pronounced funding boom. While there were other sectors that held enormous technological promise and where entrepreneurs were struggling to raise money (alternative energy technologies, for one), these officials were drawn, like moths to a flame, to the sectors that were already overfunded by angel investors and venture funds alike.
It is hard not to attribute this decision to the questions about the ATP the Commerce Department was facing at the time from a skeptical, Republican-dominated Congress. The legislators had hard questions about whether the program was needed, and the program administrators were under tremendous pressure to demonstrate that the effort was a useful economic development tool. What better way to do so than to have a number of success stories to trot out, of companies that received ATP funding and then went public? And where to find companies likely to be successful than in the hottest area of the moment?
This decision may have been rational for the ATP bureaucrats eager to ensure their program’s survival. But it was profoundly at odds with the program’s mission to identify and rectify failures in the market for funding early-stage technologies. While an Internet firm that the program funded was perhaps more likely to go public and generate jobs than an obscure company working on advanced ceramics, the impact of the program’s funds was much less in the online sector: in the late 1990s, everyone was trying to fund the next hot Internet idea. Even if the ATP had not given the online firm any funds, the start-up would still have been able to succeed. It was likely a very different story for the more neglected sectors.
If relying on success stories is not the best route to assess programs, how should these evaluations be done? In undertaking these assessments, one has to ask what would have happened without the subsidies. This may seem pretty daunting: we need to look inside a crystal ball, and figure out what would have happened in the parallel universe in which the program did not exist.
Ask ten economists how to overcome this research problem, and nine of them will give you the same one-word answer: randomization. This approach typically entails selecting some entities for awards that would not otherwise “make the cut,” while not choosing some entities that would otherwise be chosen. The progress of these entities is then compared to their counterparts. The entrepreneurs who received awards that are below the cut-off score, and those who are above the line but did not get awards, are compared to their peers to get a sense of the program’s impact.
The reason for this approach—which may seen as excessively complex and as introducing unnecessary complications—is a fear of unobserved differences. If these are not controlled for, the analysis may be flawed. To see how these considerations can affect conclusions, consider a dean of students, who is trying to persuade the admissions department to let in a hardworking student body. The dean is worried that certain student-athletes are excessively fun-loving: not only do they have poor grades, but their bad example deters their roommates from studying hard. The dean tests this idea by examining whether students who have athletes as roommates also seem to be having too much fun on campus—measured, for instance, by run-ins with the campus cops, or by being on academic probation. Indeed, he finds a relationship. The dean immediately fires off a memo to the college president, demanding that rugby and hockey players no longer be admitted because they are corrupting their fellow classmates.
The college president—trained as an economist—realizes there is a fundamental flaw with the dean’s logic. Just because fun-loving students room together doesn’t mean that the rugby players corrupt their roommates. Rather, it could be that fun-loving students chose to live together, or were placed together by administrators who didn’t want to have to deal with disputes between incompatible roommates. To put it another way, just because the dean found an association between athletes and fun-loving roommates does not mean that the jocks are causing the problem.
The same worry appears when evaluating public programs to encourage entrepreneurship. Just because those entrepreneurs who take part in a government program do better than their peers doesn’t mean the program has made a difference. Rather, the applicants could have been disproportionately the best and the brightest entrepreneurs, who were smart enough to learn about the program and find the time to fill out the application. Moreover, if there was a competition for the rewards, the screening process should have picked out the better groups. Thus, the awardees are not randomly chosen.
The reader might object that there are easier ways to solve this research problem. One idea might be to control for the characteristics of the awardees. This idea is behind a number of the analyses described in this volume, where a researcher matched awardees with, for instance, other firms in the same industry of about the same size. But we still might worry that there are differences in other, unobserved characteristics of the companies that we’re not able to see or control for, and which may affect our conclusions dramatically. By randomly selecting which entrepreneurs receive awards and which don’t, these worries are greatly reduced.
Another objection to randomization is that it’s wrong to knowingly give public money to an inferior entrepreneur. While we have long been comfortable with the use of randomized trials in medical research, where one set of cancer patients gets the experimental drug and the others get the traditional treatment, the introduction of random choices in economic development settings make many leaders profoundly nervous. Whatever the merits of their reluctance, it has blocked attempts to use randomization while assessing public venturing programs.
Fortunately, there is an alternative: the use of an approach called “regression discontinuity” analyses. Essentially, this type of analysis exploits the fact that when program managers do their assessment of potential participants, there are always going to be some applications that fall just above or just below the cut-off line. By comparing these entrepreneurs or venture funds, which are likely to be very similar to each other in everything except for the fact that some were chosen for the program and others not, one can get a good sense of the program’s impact without a randomization procedure. As Adam Jaffe, one of the most vocal advocates of better evaluation approaches, has observed:
I and others have previously harped on randomization as the “gold standard” for program evaluation. I now believe that [regression discontinuity] design represents a better trade-off between statistical benefits and resistance to implementation.18
IMPORTANCE OF A GLOBAL PERSPECTIVE
The final lesson regarding implementation is that governments should emphasize the development of strong interconnections with venture funds elsewhere. Venture capital is an increasingly global business, where strong connections to major markets seem critical to success. Growing a venture capital industry in isolation, however appealing to policymakers, is unlikely to be a winning strategy.
A dramatic example of this globalization is Skype.19 When it received its initial venture financing in 2003, Skype was the very definition of a company in “stealth mode.” In 2000, Skype’s eventual founders, Niklas Zennstrom and Janus Friis, invented a program called Kazaa. It enabled users to readily download music and video content from other users’ computers. Such a peer-to-peer sharing system may sound innocent enough, but rights to the vast majority of the material traded on the Kazaa network weren’t owned by the people doing the trading: the material was movies and music copyrighted by major studios and production companies, which did not take kindly to the loss of revenues they attributed to Kazaa.
By late 2003, with more than 300 million copies of Kazaa downloaded—the most of any program in the world—Zennstrom and Friis had emerged as two of the chief enemies of the music business. It was questionable whether Kazaa actually violated copyright law as it stood at the time: the program essentially served as a platform for traders, rather than directly being involved in trades. Furthermore, the two founders had severed most of their ties with the company. But despite these considerations, the pair were being pursued by music firms, their lawyers, and henchmen. As a result, the company was extremely secretive, not revealing the location of its Europe-based offices and the identities of the Estonian programmers who made up the heart of the firm.
While Skype used the same peer-to-peer technology as Kazaa, it was for a very different application: it offered the ability to call peers essentially for free over the Internet. As long as both callers have microphones and the Skype software, they can readily talk to each other. Once again, the users’ own computers—rather than some network that the firm built—did the hard work of finding the other party, converting the sounds into digital signals, directing them onward. At the time, the firm estimated that it cost Vonage, the leading provider of Internet telephony using a traditional centralized model, $400 to add a customer. A new customer for Skype, by way of contrast, cost one-tenth of a cent.
This value proposition was enough to attract some of the leading venture investors, despite the far-flung and secretive nature of the firm. Bessemer Venture Partners, Draper Fisher Jurvetson, and Index Ventures, among others, led an initial financing round of a few million dollars; larger financings soon followed. Despite the fact that Skype’s far-flung and rapidly changing group resisted close supervision—and that the venture groups had little ability to provide oversight “on the ground” to programmers in Estonia—the investors were willing to bet that the experience of the management team would lead to a successful firm. And indeed their confidence was justified: eBay bought the firm in 2005 for $2.6 billion, giving the “A” round investors more than a hundred-fold return.
The increased globalization of the venture capital industry can be seen along three dimensions:
• The first is in capital commitments by limited partners. Venture capital markets used to be extremely segmented: German limited partners invested in Germany, French investors in France, and so forth. Over time, however, these barriers have broken down, and international capital flows have become far more common. For instance, in Europe, in 1993, only 19 percent of funds raised by venture and buyout groups were from capital sources based outside of Europe. By 2007, the percentage was up to 34 percent. In the most successful markets, such as Great Britain, the domestic shares are even lower: only one-quarter of the capital is from sources within the United Kingdom.20
• The second dimension is the changing location of investments by venture capitalists. In previous years, many venture groups emphasized the importance of investing extremely locally, often within an hour’s drive of the office. Over time, however, long-distance investing has become far more commonplace. The explosive growth of opportunities in India and China, coupled with successful deals elsewhere, has opened individuals’ eyes to the potential of long-distance investing.
• The final, and perhaps most critical, dimension is in the deployment of resources by entrepreneurial firms themselves. In the past few years, it has become commonplace for even the youngest Silicon Valley firm to have an overseas presence. Typically, these groups will employ programmers in India (if a software concern) or design and production experts in China (if a firm selling hardware) almost as soon as they are formed. These extensions allow the entrepreneurs to produce far more output from each dollar invested than they would had they confined their hiring to domestic markets. Moreover, doing work abroad allows the firms to get their products to the market more quickly. As a result, venture capitalists are spending far more time in Asia supervising the farflung operations of their portfolio firms, or even opening offices in these nations.
In some cases, these global connections can arrive without government intervention. Eastern Canada in recent years offers an example. The venture capital industry has expanded, largely because of its close ties to the United States. First, local transactions have appeared increasingly attractive to funds based in Boston and New York: the disparity in valuations between eastern Canada and the United States has meant that stakes in comparable companies have been available at a substantial discount. Second, Ontario- and Quebec-based funds are increasingly attracting limited partners based in the United States as investors in their new funds.
But unless a nation is lucky enough to be proximate to a venture hub, effective government policy is likely to be helpful in catalyzing the globalization process. This point can be illustrated by comparing two case studies: the experiences in Israel and Japan.21
In June 1992, the Israeli government established Yozma Venture Capital Ltd., a $100 million fund wholly owned by the public sector. At the time, there was a single venture fund active in the nation, Athena Venture Partners. While there were certainly well-trained engineers in the nation working on promising technologies, entrepreneurs (and would-be company founders) were suspicious of venture investors. This reluctance was based in part on their interactions with the pioneering venture capitalists in the nation, as well as their general skepticism about selling equity to unaffiliated parties. Instead, they preferred to rely on bank debt for financing. The only problem, of course, was that such financing was rarely available for young, risky ventures.
The key goal of Yozma was to bring foreign venture capitalists’ investment expertise and network of contacts to Israel. The need for this assistance was highlighted by the failure of the nation’s earlier efforts to promote high-technology entrepreneurship. One assessment concluded that fully 60 percent of the entrepreneurs in prior programs had been successful in meeting their technical goals but nonetheless failed because the entrepreneurs were unable to market their products or raise capital for further development.22 Foreign expertise was seen as key to overcoming this problem.
Accordingly, Yozma actively discouraged Israeli financiers from participating in its programs. Rather, the focus was on getting foreign venture investors to commit capital for Israeli entrepreneurs. The government provided matching funds to investors, typically $8 million of a $20 million fund. The venture fund was given the right to buy back the government stake within the first five years for the initial value plus a preset interest rate of roughly 5 to 7 percent. Thus, the incentives of Yozma meant that the government provided an added incentive to the venture fund if the investments proved successful. Moreover, learning from the nation’s misadventures during earlier programs to stimulate the venture industry—when cumbersome application procedures and burdensome reporting requirements discouraged participation—the administration of the program was deliberately made simple.
In addition to the financial incentives, the project adopted a legal structure for the venture funds that foreign investors would be comfortable with. Included were features such as a ten-year fund life, limited partnerships modeled after the Delaware partnerships that are standard practices in the United States and elsewhere, and “flow through” tax status. Had the government not adopted these features—and the Israeli Treasury department resisted them before acquiescing under pressure—it is unlikely that the program would have succeeded in attracting foreign investors.
The Yozma program delivered beyond the wildest dreams of the founders. Ten groups took advantage of this offer, mostly from the United States, Western Europe, and Japan. Many of the original Yozma funds, including Gemini and Walden Ventures, earned spectacular returns and served as precursors to larger, follow-on funds. Moreover, many of the local partners recruited by the overseas venture capitalists were able to spin off and establish their own firms, which global investors were eager to fund because of their impressive track records. (A Yozma “alumni club” allows groups to learn from each others’ experiences while making these transitions.) One decade after the program’s inception, the ten original Yozma groups were managing Israeli funds totaling $2.9 billion, and the Israeli venture market had expanded to include 60 groups managing approximately $10 billion.23 The magnitude of this success is also suggested in figure 6.1, which shows that the ratio of venture investment to GDP is far higher in Israel than elsewhere. In most tabulations, Tel Aviv has surpassed Boston as the urban area with the most venture activity after San Francisco.
Japan is a study in contrast. It has also provided direct financial assistance to entrepreneurial firms, but with a very localized focus.24 Both the Ministry of International Trade and Industry and Japan Development Bank (JDB) developed programs that offer financial assistance to young, entrepreneurial firms. This assistance has taken a variety of forms, from actual operating facilities (“incubators”) to equity investments and loans. For instance, JDB established a fund to provide five-year loans at subsidized rates (typically 3.25 percent and less) to young high-tech firms—loans that the bank secured through these firms’ patents and other intellectual property. And in late 1996, JDB raised capital from more than a hundred corporations and government agencies to make traditional equity investments.
But throughout, there was no effort to encourage the involvement of foreign venture capitalists or others. In fact, the investments were typically structured in ways that were quite alien to outsiders. These efforts were almost uniformly unsuccessful, and venture capital activity in Japan has plummeted in recent years. In large part, this decline reflects the lack of fresh perspectives: given the strength of technological innovation in Japanese firms and the depth of many academic departments in its universities, it is hard to believe that there are not innovations waiting to be commercialized. But the domestic venture industry was never particularly strong at screening, adding value, and monitoring entrepreneurs: instead, it was a very bureaucratic process that was largely in the hands of bank and insurance company affiliates. The incentives introduced by the various government programs were not designed to attract overseas groups with the key skill sets, and these groups have continued to largely ignore the market.
The benefits of global connections are manifold. But within the political process, the imperatives of satisfying domestic audiences can lead to distortions. When Australia legalized the venture capital limited partnership structure in 2002, for instance, legislators worried that foreign funds or firms might exploit the favorable tax treatment these entities enjoyed.25 So they required that each company backed by a venture partnership have at least half its assets in Australia. The pioneering funds found the companies in their portfolio handicapped, as the entrepreneurs could not expand their software development activities in India or their manufacturing operations in China without putting the venture funds’ tax status in danger.
Local venture capital industries can benefit enormously from being well connected to the global market. Such connections are likely to lead to knowledge flows to local venture capitalists, follow-on capital to portfolio firms, and an ability to raise larger follow-on funds. The changing nature of the venture capital process implies that these ties are more readily established than in years past. Meanwhile, to build a venture capital industry in isolation is a recipe for irrelevance and failure.
So far, we have looked at the various elements that government can use to boost entrepreneurial activity—for instance, tax policies, boosts to technology transfer, and subsidies to entrepreneurs—in isolation. Ideally, of course, these elements should be viewed collectively. Are there certain steps that work well together? Are there other combinations that should be avoided?
Here, alas, we must reemphasize the early state of our knowledge. While it is undoubtedly important to understand how the pieces of government policy fit together, we’re still not in a position to say much definitive about their interrelationships.
We can illustrate this point by considering one of the most ambitious research efforts to fit the pieces together, an essay by Christian Keuschnigg.26 He considers a setting where governments can use a variety of weapons:
• Subsidies for investments for venture capitalists
• Subsidies or taxes on sales or profits of large or start-up firms
• Funding of basic research
Practical-minded readers may object that this list leaves out many other strategies that we have discussed. But, as we’ll soon see, an analysis of even this limited number of options proves troublingly difficult.
Keuschnigg points out that policymakers need to be simultaneously active on several fronts. Otherwise, distortions may creep in that leave entrepreneurs worse off than before. For instance, what happens if the government just subsidizes start-ups (for instance, through tax credits)? We might assume that this subsidy will increase the profits of entrepreneurs and venture capitalists. But it may lure more entrepreneurs and venture capitalists into the market, so that, unless the supply of good ideas grows, more firms and financiers are chasing after the same ideas. This competition may depress returns, and ultimately discourage entrepreneurs and venture investors. What seems like a reasonable policy turns out to be self-defeating.
So far, so good. We see that policymakers need to identify the right combination of actions to achieve the optimal effects. But it is at this point that things get complicated. In particular, after running various optimization analyses, Keusch nigg suggests that many of the classic remedies—like supporting research or subsidizing venture capital investments—are not ideal. Rather, to boost entrepreneurship, policymakers should be subsidizing mature firms, perhaps even with some penalties for start-ups! His intuition is that by making the fruits of success really sweet, entrepreneurs will work even harder (and venture capitalists be even more willing to fund them), without the distortions that his model suggests are the by-product of subsidizing start-ups.
The basic idea behind Keuschnigg’s analysis—that piecemeal policies aimed at helping entrepreneurs may have harmful consequences—appears to be reasonable. But it is hard to know how many of his rather puzzling results are a function of the model itself. To arrive at these results, the analysis has to make several assumptions that seem contrary to the behavior of real officials and real entrepreneurs:
• The model requires policymakers to have a lot of information about the production costs, preferences, and innovative potential of the various players in the economy. As we have seen, in many instances, government officials in fact have very little information about the basic actors and their incentives, much less the specific industries they are targeting.
• The government in the model must be able to implement a complex system and subsidies. In actuality, even if government officials want to implement the socially ideal system (and this is a big if!), we know that individuals and firms have boundless creativity when it comes to avoiding taxes and maximizing subsidies.
• While government officials have a great deal of flexibility in developing and implementing complex programs, the model strictly limits the ability of entrepreneurs and venture investors to solve their various problems by entering into complex and creative contracts, as they frequently do in real life.
Thus, Keuschnigg’s model illustrates how far economists need to go in thinking about policy packages, and the challenges researchers face before they will be able to answer the questions that policymakers struggle with. We economists have a way to go before we can present our own “unified field theory” of how government can help entrepreneurs.
FINAL THOUGHTS
The effective implementation of public venturing programs may initially seem a dry and esoteric topic. But as the many examples discussed in this chapter have suggested, effective implementation is vitally important. To successfully promote entrepreneurship and venture capital, public officials must correctly make many choices.
In this chapter, we’ve highlighted three areas where efforts have often gone astray. The design of effective incentive schemes helps ensure that the players receiving the subsidies, whether entrepreneurs or venture capitalists, only benefit at the same time as society as a whole and limits opportunistic self-dealing. Rigorous evaluations can ensure that the right people are attracted to government programs, and that the initiatives themselves are well designed. And a strong international orientation can maximize the chance that best practices are effectively absorbed.
In the next chapter, we’ll turn to a special sort of public venture capital, the sovereign wealth fund. These government investment pools face many of the same issues that the programs we have discussed in this book have grappled with, but their size and visibility introduces additional issues.