Using NCSA resources, Illinois researchers gain better understanding of entrepreneurs

released 08.06.07

By Tracy Culumber, NCSA

Economists have long debated the reasons for entrepreneurial success. Some argue that entrepreneurs are uniquely optimistic and more willing than most people to assume the risks associated with running a business. Others believe that the institutional environment -- for example, bankruptcy laws or the availability of credit -- also affect an entrepreneur's success.

Which has a greater impact on small firm performance: Personal characteristics or the situation within which the entrepreneur operates?

In order to answer this question, researchers Anne Villamil and Stefan Krasa of the University of Illinois Department of Economics and a team of graduate students recently developed a computable model that can distinguish among theories of why people become entrepreneurs. Using the model and computing resources at the National Center for Supercomputing Applications (NCSA), they determined that entrepreneurs need not have personal characteristics that are significantly different from standard values, but the environment in which they operate matters greatly to their success.

Small firms, big calculations

Villamil explains that their dynamic model, which considers individual differences in willingness to bear risk and optimism, is a way to assess how entrepreneurs balance their firm's current return with returns expected in the future. With this model, Villamil and Krasa can evaluate the effect of bankruptcy rules on small firms.

The team applied the model to data from the Survey of Small Business Finance (SSBF). Villamil says that because small businesses often co-mingle personal and business assets, the SSBF data was crucial for the study, making it possible to compute the empirical distribution of firm returns.

"We specifically look at small incorporated firms with median employees of about seven and median assets of $270,000," Krasa says. "We used the SSBF because this data set has detailed information on firms and owners; we are aware of no other data set that contains both types of information."

According to the U.S. Small Business Administration, small firms produce more than 50 percent of U.S. non-farm output, employ 50 percent of workers, and pay 45 percent of total private payroll. Nevertheless, most studies on entrepreneurs use data from large firms because they are required by law to keep more financial statistics.

Using NCSA's Copper cluster computing system, Krasa and Villamil were able to quickly determine if the behavioral predictions of a particular theory could be reconciled with economic data.

"Our problem is especially demanding because we must compute entire distributions for most decision variables, not just point estimates," Villamil says.

"The empirical analysis can be done on a laptop, but running the model with so many policy parameters and integrations gets very long," Krasa adds. "It would have taken years."

Environment trumps personal characteristics

Krasa notes that the results of the computational model were consistent with past studies, which found entrepreneurs to be somewhat optimistic. However, their group was the first to assess the importance of risk-aversion, optimism, and credit constraints in a dynamic model of this size and structure with a risky return distribution. They found that the institutional environment in which firms operate influences entrepreneur success more than personal characteristics. Their findings also show that entrepreneurs are not excessive risk-takers, as some people have asserted.

Villamil says the model explains three puzzling observations they found in the SSBF data. Owners of small firms: (1) Face an unattractive "risk vs. return" tradeoff; (2) have poorly diversified personal financial investments; and (3) tend to "bail out" their companies, rather than declare bankruptcy.

"I was especially surprised to see the degree to which people use personal assets to support their business." Villamil says.

So why do people continue to start small businesses and personally finance them when the chances of losing money are so high?

"They do this because they truly believe the future will be better than the present, and the future is very, very important to these people," Villamil says. "The loss protection afforded by bankruptcy and an owner's ability to preserve the potential for high future gains by injecting personal funds alters the firm's return trade-off."

The result is a low firm bankruptcy rate, and firm capital structures and personal investment patterns that match the SSBF data.

Krasa cites the example of Pat Sullivan, the founder of a company that sells business software for sales people. Sullivan originally developed his software for personal use, but decided to start a company because the bankruptcy laws in Texas would allow him to keep his house, whether or not his business succeeded.

"The point I want to make is that the bankruptcy law reduced the risk for the Sullivans sufficiently so that they were willing to take the risk and start their business," Krasa explains. "This is an example of what it means for the institutional environment to matter greatly."

Krasa adds that this behavior accounts for why entrepreneurs in developed nations -- dynamic settings where bankruptcy institutions may seem lenient but are efficient -- often succeed, while entrepreneurs with the same level of optimism and risk-taking characteristics in developing countries may not succeed.

"Immigrants in the U.S. come here and flourish. Why is that? It is because it is not only the person, but the environment they have entered that will help them to succeed," Villamil says.

Additional avenues for research

The researchers' results -- which were recently presented at a Conference on the Resolution of Financial Distress at the Institute for Advanced Study, the XVI European Workshop on General Equilibrium, and the SAET International Conference on Current Trends in Economics -- indicate that government policies, such as bankruptcy legislation, provide insurance to entrepreneurs against extreme losses, and therefore affect the entrepreneur's willingness to take a risk.

Villamil says that there are many questions that remain to be answered, and that NCSA computing resources will continue to be essential to complete the research. Future computational studies might be aimed at distinguishing between willingness to bear risk and patience, or could examine the impact of alternative bankruptcy rules and other personal characteristics, such as gender.

This research was supported by grants from the National Science Foundation's Social, Behavioral and Economic Sciences Directorate, NCSA, the Ewing Marion Kauffman Foundation, and the University's Academy for Entrepreneurial Leadership.