The role of human capital in loan officers' decision policies.

The similarity-attraction paradigm asserts that individuals tend to be attracted to people who are similar to themselves (Byrne, 1971). Individuals are attracted to others who have similar demographic or psychographic characteristics like gender, race, age, education, or personality because they assume that these visible characteristics are associated with characteristics that are more difficult to discern, such as values, priorities, and abilities that are similar to their own (Harrison, Price, Gavin, & Florey, 2002; Klein, Beng-Chong, Saltz, & Mayer, 2004). Similarity is often perceived, accurately or inaccurately, based on shared observable characteristics (Jackson et al., 1991; McNeilly & Russ, 2000). Similarity provides a source of predictability, familiarity, comfort, and validation (Klein et al., 2004; Williams & O'Reilly, 1998) which can significantly influence the decisions made by individuals. For example, education similarities between sales representatives and managers are shown to affect the managers' ratings of the representatives' performance (McNeilly & Russ, 2000). In another recent study, participants selected coworkers with similar skills and values for participation in a mutually beneficial training exercise, while they selected dissimilar coworkers for participation in a competitive activity (Glaman, Jones, & Rozelle, 2002) Orpen (1984) showed that the perceived similarity of an interviewee to the interviewer influenced the interviewer's decision of whom to extend a job offer.

Because of the short history of entrepreneurial firms, it is often difficult for loan officers' to gather reliable, detailed performance data. However, there are many attributes that a loan officer may easily assess when considering an application of a prospective borrower, such as gender, age, education, and previous work experience. The similarity attraction paradigm suggests that loan officers will value attributes of the small business owner that are similar to their own. Loan officers with a high level of banking, lending, or small-business loan experience are more likely to be attracted to entrepreneurs with a high level of experience, knowledge, and skills in the line of business for which they are seeking a loan. Because of the desire to judge oneself and one's performance positively (Pfeffer, 1998), the more experienced loan officers will be inclined to make positive attributions regarding the benefits of their own work experience (e.g., shrewd judgment, problem solving skills, tacit knowledge) on their own job performance. As a result, when they see small-business owners with a similar level of experience, they will be more prone to assign comparable positive attributions regarding the work experience of the small-business owners (Kelley, 1971). A loan officer with less experience will comparatively devalue the need for such experience because it is contradictory with his or her own human capital attributes. In this way, similarity attraction affects the emphasis placed on the small business owners' human capital criteria in the decision-making process. Thus,

Hypothesis 3: In determining whether to offer a loan to a small business, loan officers with more specific human capital in terms of (a) banking experience, (b) lending experience, and (c) recent small business loan experience will place greater emphasis on the level of the small business manager's previous related business experience, than those loan officers with less specific human capital.

Using the similarity-attraction paradigm, we argued that loan officers with a high level of experience would consider business owners with a high level of experience to be similar to themselves. Hypothesis 3 considers only the specific human capital of both the loan officers and the business owners. The particular type of human capital (i.e., general [GHC] or specific [SHC]) is critical to the level of attraction of the applicant to the loan officer. We would expect loan officers to consider an applicant similar to themselves if they both have a comparable level of work experience (SHC) but not necessarily if one has a high level of work experience (SHC) while the other has a high level of education (GHC). What is more, we would expect that when controlling for age, the loan officers' level of education (GHC) will be inversely related to their number of years of experience (SHC) because of "lost" years of work while attending school. Therefore, a loan officer with a high level of education (GHC) will be more likely to have a lower level of specific human capital and will consequently perceive dissimilarity with a highly experienced loan applicant's level of specific human capital. This dissimilarity in specific human capital results in less attraction towards and ultimately less emphasis placed on that applicant's level of business experience. Therefore,

Hypothesis 4: In determining whether to offer a loan to a small business, loan officers with more general human capital in terms of education will place less emphasis on the level of the small business manager's specific human capital, than those loan officers with less general human capital.

Method

Research Context and Design

This study collected data from loan officers at several Swedish banks. In 2000, loans extended to the Swedish commercial and industrial sector by Swedish banks totaled 542 billion Swedish krona (SEK) (approximately US$52 billion) (Svenska Bankforeningen, 2003) and increased to 838 billion SEK in 2006 (Svenska Bankforeningen, 2007). In comparison, risk capital and venture capital investments amounted to a relatively nominal 19.4 billion SEK in 2000 and 39 billion SEK in 2006 (EVCA Yearbook, 2007). The total loaned to the business sector makes up approximately 41% of all bank loans (Svenska Bankforeningen, 2007). The commercial loan numbers include both small and large businesses and are difficult to categorize by business size, as the Swedish banks typically track business loans by the size of the loan rather than the size of the business. However, the importance of the SME sector is evidenced in the annual report of one particular Swedish bank--the bank states more than 98% of the bank's business clients are SMEs (ForeningsSparbanken, 2006).

As in most countries, the Swedish banking system is regulated with the purpose of protecting deposits made by the public. Swedish banks, and others within the European Union, are permitted to lend up to 12.5 times their equity, ensuring that a certain amount of credit is covered by the equity of the banks. Before credit is granted, the bank is required by law to be informed of the borrower's financial situation. The law does not regulate the content and depth of a credit assessment, which is the sole responsibility of banks and their loan officers. Therefore, the banks inevitably specify limitations and parameters through bank-specific loan policies.

To examine the assessment policies of the loan officers, we relied on a metric conjoint experiment. Conjoint experiments require respondents to make a series of judgments based on a set of attributes. In our case, the judgment consisted of the probability that participating loan officers would support credit requests by a series of hypothetical small firms. On the basis of these judgments the underlying structure of their decision policies can be decomposed by means of hierarchical linear modeling. Conjoint analysis has been used in hundreds of studies of judgment and decision making (Green & Srinivasan, 1990), including venture capitalists' assessment of portfolio companies (e.g., Shepherd, 1999). Conjoint analysis is particularly appropriate for this study, because as a real time method it overcomes many of the potential problems associated with post hoc methods, such as self-reporting and recall biases. It also allows us to examine how respondents process contingency relationships (Hitt & Barr, 1989; Zacharakis & Shepherd, 2005), which is of particular interest to this article, without relying on the respondents' (generally inaccurate) introspection (Fischhoff, 1982).

Sample

The sample consisted of loan officers' actively involved in assessing loan applications by businesses, including those from small- and medium-sized businesses. We focused on loan officers that were employed by the major commercial banks in Sweden and worked in branches within a 100-mile radius of the first author's institution. A geographical criterion for selecting the sample was necessary because we wanted the experiment to be conducted in the presence of one of the authors to ensure that any of the respondents' questions could be immediately addressed, to conduct a postexperiment interview, and to enhance the response rate. One hundred miles represented a maximum commute of 2 hours. We also wish to point out that this region is well known for its prevalence of small businesses.