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

We contacted by telephone the district manager, division manager, or branch manager representing each branch in the region to illicit support for the current project. Four of the five commercial banks active in the area agreed to participate. The fifth bank exhibited a positive attitude toward the study but was unable to participate in a timely manner due to organizational changes. There are no substantial differences in services offered by the nonparticipating bank as compared to the participating banks. In total, 114 loan officers with relevant experience with small business loans, representing 56 bank branches, were contacted and all agreed to participate in the study. A sample size of 114 exceeds many other conjoint studies, for example, in studies of the decision policies of venture capitalists, with sample sizes of 66 (Shepherd, 1999) and 53 (Zacharakis & Shepherd, 2005).

The sample captures loan officers with a wide range of experience. For example, tenure as a bank employee ranged from 1 to 42 years (mean = 22.9, SD = 9.7), and tenure as a loan officer ranged from 1 month to 36 years (mean = 13.0, SD = 7.4). More than half (50.9%) of the loan officers have some college education. Eighty-six percent of the loan officers are male, with an average age of 45.5 (SD = 8.4). The descriptive statistics and correlation matrix for the sample (level 2 independent human capital variables and control variables) are shown in Table 1.

The mean number of loans approved per week by the loan officers ranged from 0 to 150 (mean = 5.7, SD = 15.9). The mean total capital lent by study participants in the previous year was about SEK 322,000,000 (approximately U.S.$31,000,000).

Data Collection: Experiment and Postexperiment Survey

In the conjoint experiment, loan officers were asked to evaluate a series of hypothetical small businesses that varied in terms of key attributes and decide on the likelihood that they would offer each of them a loan. In order to ensure that all applications were framed similarly, several aspects were explicitly held constant across the hypothetical small businesses. In the instructions to the conjoint experiment, loan officers were instructed that "when making these assessments assume the firm is typical of those from which you would normally receive credit applications and that you are operating under present Swedish economic conditions"; "view the firm as a typical applicant in your bank branch with you as the loan officer"; "the companies applying for credit are growing companies with 20 employees, 20 million SEK in turnover. The company produces and sells five different products ... the company applies for a loan for expanding its business. The expansion equals an amount equal to the firm's equity. This should be regarded as a relatively large investment (i.e., for machinery, production facilities). The estimated ROI (return of investment) for this investment is regarded as normal."

All hypothetical businesses were characterized as selling five products or services, having a market share of 5% in a local market, competition being neither intense nor weak, situated in a city of 120,000 inhabitants, the five biggest customers generate 40% of sales, cash flow matches the industry average, and the company wishes to borrow an amount equal to its equity in order to expand its activities. The instructions specified that all conditions, other than the attributes described in the profile, are to be considered constant across all profiles. This controlled for other possible confounding attributes.

A practice profile was used at the start of the experiment to familiarize respondents with the experiment, but it was not used in the analysis. We pilot tested the instrument with six loan officers. Comments suggested that the instructions and definitions of attribute levels were clear and that the instrument appeared realistic (had face validity). After completing the conjoint experiment, each loan officer completed a survey, the answers to which provided information about their human capital.

Variables and Measures

Probability of Supporting a Loan. Having assessed a hypothetical profile, the loan officers were asked to respond to the following question: "How would you rate the probability that you would support this business's credit request?" It was measured on a 9-point scale anchored by "Not at all Likely" (scored 1) and "Very Likely" (scored 9). This scale captured loan officers' assessments from which their decision policies can be determined.

Decision Attributes. The participants in the experiment were presented with eight attributes describing each small business (scenario) that they were to assess. The first five attributes were generated based on the five Cs and the related frameworks.

1. Business risk: High--the firm tends to invest in high-risk projects that offer the chance of high returns (coded -.5); Low--the firm tends to invest in low-risk projects that typically have small but certain returns (coded .5).

2. Share of investment: High--the firm finances 35% of the capital requirements either through internally generated funds or by additional capital provided by the owner(s) (coded .5); Low--the firm finances 5% of the capital requirements either through internally generated funds or by additional capital provided by the owner(s) (coded -.5).

3. Financial position: High--the firm's liquidity and solvency is well above the industry average (coded .5); Low--the firm's liquidity and solvency is below the industry average (coded-.5).

4. Independence of collateral: High--the firm offers collateral that is independent of the firm's success or failure, e.g., bond, personal guarantee, private property (coded .5); Low--the firm offers collateral that is dependent of the firm's success or failure, e.g., floating charge, receivables (coded -.5).

5. Related business experience: High--the company has previously successfully conducted similar activities and has solid experience, knowledge and skills within the line of business (coded .5); Low--the company has not previously conducted similar activities and has limited experience, knowledge and skills within the line of business (coded -.5).

The remaining three attributes have some overlap with the five Cs, but were added to control for other important factors discussed in the literature on funding.

6. CEO tenure: High--the CEO started his/her position 8 years ago, has a business education, and is regarded as being honest and reliable. (1) Before that the CEO had no experience in the industry or as a CEO (coded .5). Low--the CEO started his/her position 2 years ago, has a business education, and is regarded as being honest and reliable. Before that the CEO had no experience in the industry or as a CEO (coded -.5). An important indicator of the future performance of the business at a particular task is the tenure of the small business owner in the current position. This experience may be beneficial in running an independent business (Wright & Robbie, 1997) and provides benchmarks for judging the relevance of information (Cooper, Folta, & Woo, 1995), which can enhance performance (Davidsson & Honig, 2003). We control for this attribute with the CEO tenure variable.

7. Past performance: High--the company's profitability has hitherto been well above the industry average (coded .5); Low--the company's profitability is hitherto below the industry average (coded -.5). The past financial performance of a business is an important consideration in assessing the ability of a business to repay a loan (Gibson, 1993). Past performance sends important signals about the ability of the management team to formulate and implement growth strategies (Kam, 1990) and is controlled for by this variable.

8. Comprehensiveness of the strategic plan: High--the company has a comprehensive business plan that is documented and followed (coded .5); Low--the company follows a distinct strategic course but its plan for doing so is not documented (coded -.5). In evaluating the creditworthiness of a prospective borrower, a loan officer must determine the ability of the small business to develop and execute effective strategies to adapt to a changing environment (Berger, 1997/1998; Hedelin (1999) found that a realistic business plan was among the most important factors in assessing the ability to repay a loan. We control for this decision criteria because the strategic plan is often used as an indicator of the quality of the management team (Sinkey, 1992).

Other Control Variables. The gender and age of the loan officers were included in the study in order to control for any potential decision variance as a result of these demographic variables.

Independent Variables of the Survey. The questions from the survey were used to measure human capital and included the following variables:

Education. Respondents were asked to state their highest level of completed education using the following categories: junior secondary school, senior high school, some university, bachelor' s degree, and master' s degree. There were few loan officers in the first and last of these categories (9 of 114). We therefore dichotomized the education responses into those that have an undergraduate degree or more advanced degree as their highest level of education (dummy coded 1; n = 25) and those that do not (dummy coded 0; n = 89).

Banking experience. Respondents were asked to state how many years they have worked in the banking industry.

Lending experience. Respondents were asked to state how many years they have worked as a lending officer.