Why start-ups?

• Author: Bankman, Joseph
• Position: Economic rationale for start-up companies

The prototypical start-up involves an employee leaving her job with an idea and selling a portion of that idea to a venture capitalist. In many respects, however, the idea should be worth more to the former employer. The former employer can be expected to have better information concerning the employee-entrepreneur and the technology, have opportunities to capture economies of scale and scope not available to a venture capital-backed start-up, and will receive more favorable tax treatment than the start-up should the innovation fail. In connection with an auction of the idea, the former employer should have both a more accurate estimate of its value and receive an element of private value not available to the venture capitalist. In turn, this should give rise to a powerful winner's curse: each time a venture capitalist wins the auction, it will have paid more than a party that has better information and receives an element of private value. The puzzle, then, is why do we ever observe start-ups? Professors Joseph Bankman and Ronald J. Gilson suggest three interrelated explanations. First, the venture capitalist may have superior information with respect to some subset of employee innovations. Second, employer bids on employee innovation can create an incentive for employees to establish internal property rights in their research efforts that may reduce the future output of the employer's research and development efforts. Finally, employees are not homogenous. The attractiveness of venture capital financing depends critically on employee personal characteristics, such as risk aversion. The employer sets the internal payoff to discovery--its bid--to equalize the marginal benefit of retaining employees who might otherwise leave to the marginal cost of establishing unfavorable incentives for future research and development for those employees who do not find venture capital financing a close substitute for continued employment. In some cases, this calculus might lead to a "no bid" policy. In virtually all cases, the pay-off will be set too low to retain all employees, and start-ups ensue.

Communities are defined by their mythology. In Silicon Valley, the defining myth takes as its stage David Packard's or Steve Jobs' garage. Palo Alto's Roland is the engineer who, with nothing but an idea and strength of character, leaves his job with an established company, starts a firm that becomes an industry leader, and in the process becomes fabulously wealthy. In this community, the myth is taken seriously. Over and over again, people set out on the path of heroes: They leave their comfortable, secure jobs, and start from scratch.(1) In this article, we seek to better understand the economic foundation for the high-tech start-up phenomenon.

The task is interesting because, as a matter of culture, high-tech entrepreneurs are the cowboys of our age. In the United States, as Willie Nelson has told us, our heroes have always been cowboys.(2) More to the point of our inquiry here, the presence of start-ups poses an intriguing puzzle as a matter of industrial organization. Each time an engineer leaves her established employer and organizes a venture capital-backed start-up, an alternative was also available: The employer could have paid the engineer to stay with the company and develop the new idea there. The puzzle arises because, on a first run-through, it is easy to show that the employer's proposal should be successful every time. Characterizing the initial transaction as an auction by the engineer of a share in her intellectual property, the idea should be worth more to the employer than to a venture capitalist who would fund the startup. Accordingly, the employer should always win the auction. The puzzle, then, is how do venture capitalists ever win the auction? Why do we ever observe start-ups? Our analysis suggests that venture capitalists win, and we observe start-ups at all, at least in part because employers limit the circumstances in which they bid.

In Part I, we motivate the analysis by demonstrating the employer's bidding advantage. In addition to information and scope economies available to the employer but not a start-up, we stress that the asymmetrical character of the corporate income tax favors an established profitable company (the employer) over a start-up as a vehicle for developing the employee's innovation. In Part II, we canvas a range of explanations for why the employer might decline to bid for ideas that are worth more to it than to the next highest bidder. We note that, for a certain subgroup of innovations, the employer's bidding advantage is offset by certain skills possessed not by it but by venture capitalists. In addition, a practice of establishing large rewards to those employees who develop an idea that could be sold to a venture capitalist may have dramatic effects on the employer's overall research and development effort. In order for the employee to have a property right that can be sold externally, and thereby trigger a bid by the employer, the employee must have established an internal property right. The behavior resulting from the employees' incentive to create internal property rights gives rise to costs for an employer that participates in the venture capital auction. In Part III, we note that the attractiveness of venture capital funding to an employee depends critically on certain employee characteristics, such as risk aversion, or the desire to be one's own boss. For employees with these characteristics, venture capital financing and employment with an established firm are not close substitutes. Employers have only a rough idea of the dispersion of these characteristics among employees. Employers will adopt a bidding strategy that balances overpaying to retain employees who do not find venture financing a close substitute for continued employment, and establishing unfavorable research and development incentives for this class of employees, against the benefits of retaining employees who find venture capital funding more attractive.

1. THE EMPLOYER'S ADVANTAGE: WHY WE SHOULD NOT OBSERVE START-UPS

   To set up the puzzle, assume that we have a risk-neutral, established employer and a group of risk-averse engineers. Each of the engineers is competent at pursuing the employer's ongoing business--normal science--and each has a positive probability of developing an innovation that would give rise to a new business. Neither the engineers nor the employer can identify ex ante which of the engineers actually will develop the innovation.(3) The obvious risk-sharing arrangement is for the employer to hire all the engineers, pay each the mean expected value of the group, and receive in return the engineers' contribution to the employer's ongoing business and all discoveries made by any of them.

   Of course, once the uncertainty is eliminated--when the passage of time reveals which engineers have won the innovation lottery--the engineers' attraction to ex ante risk sharing gives way to an incentive for ex post opportunism. The winning engineer wants to keep the discovery for herself. Thus, the risk-sharing arrangement must have some characteristics that protect it from the corrosive effect of learning which engineer wins. And it should be apparent that, whatever the arrangement, it is not perfect. There must be a problem in prospectively transferring ownership of the engineers' discoveries to the employer. If the property rights could be fully specified and transferred ex ante, we would not observe start-ups. Thus, the puzzle is premised on a contracting failure.(4)

   However, contracting failure alone is insufficient to account for the existence of start-ups. Even if the employee retains some property right in her discovery, commercializing the idea requires starting a business. In return for some portion of the future earnings from the innovation, other parties still must put up the factors of production the employee lacks: (1) capital; and (2) managerial, manufacturing, and marketing expertise. A conventional solution to this problem is venture capital financing. A venture capital fund contributes capital to the start-up as well as noncapital contributions such as management consulting, monitoring, and reputation, in return for an ownership stake in the start-up.(5) But suppose we recharacterize the venture capital process as an auction. The engineer leaving her position with the employer is selling off a portion of her innovation to the highest bidder. When all else is equal, the employer has advantages--tax, information, and scope--that should result in it consistently winning the auction.

   1. The Employer's Tax Advantage

      Under current law, the tax treatment of corporate investment depends on the tax history of the organization making that investment. An investment made by a company with sources of past or present income is subject to a symmetrical tax regime: The expenses of the investment are deductible and produce tax savings, while the gains from investment are subject to tax. An investment carried out by a start-up, in contrast, is subject to an asymmetrical tax regime: Gains are fully taxed, but losses may be deducted only against future income. As discussed in more detail below, this difference in tax regime operates in the circumstance of interest here to provide a substantial subsidy for an employer's efforts to develop an employee's innovation internally.(6)

      Under Internal Revenue Code (I.R.C.) [sections] 172, a start-up may deduct expenses only against income--expenses in excess of current income (a net operating loss) may generally be carried forward for fifteen years and deducted against future income.(7) In contrast, a company with sources of past or present income from other activities--we will call it an "established company"--may deduct expenses of the start-up activity that exceed start-up activity income as they are incurred: The expenses of the start-up activity...