METHODOLOGY

The research sample consists of independent banks formed in the United States in 1988. We chose the year 1988 as this provided the opportunity to track performance over a 6-year period subsequent to formation. Bank formations provide a unique opportunity to examine the new venture at the point of inception for several reasons. First, during a typical year, in excess of 200 new banks are formed in the United States (Associated Press, 1995). Second, while it is often difficult to establish a specific date in which a new venture begins operations (Vesper, 1990), for banks, the date of inception is clear as regulations require that all banks be chartered prior to beginning operations. Third, due to the regulatory nature of the banking industry, reliable secondary data regarding their performance and founding conditions is available.

Using the monthly listing of new bank formations published by the Board of Governors of the Federal Reserve System, we identified a population of 212 new bank formations for the year 1988. After excluding the 39 banks that were subsidiaries of existing banks or bank holding companies, 173 independent banks remained in the database. We excluded seven trust banks and credit card banks as these are established with unique missions and are not generally comparable to traditional banks. Sixteen banks had to be excluded because of missing data. Of the remaining 150 banks, 35 had been closed, merged or acquired by 1994. To have eliminated these 35 ventures from our sample and examined only the survivors would have created a restriction of range issue and would have unduly biased the study. In conducting research on established firms, researchers often regard closings, mergers and acquisitions as failures and treat them accordingly. In new venture research, a merger or acquisition may not represent a failure, but may be the result of superior performance. The 12 banks that were closed by banking regulators were clearly failures and could be treated as such in our analysis. The performance data for the 23 mergers and acquisitions were examined independently by two researchers who then judged each as a success or a failure. The raters agreed initially on 22 of the 23 cases, for an inter-rater reliability score of .91. After some discussion, agreement was reached on all 23 cases. Thirteen were identified as failures. Ten banks were found to have been performing well prior to acquisition or merger and were identified as successes. Since no appropriate performance measures would exist for these success ventures for the years following their merger or acquisition, these ten successes had to be eliminated from the sample. This yielded a final data set of 140 ventures.

 

Measures

With few exceptions, all of the data was obtained from the annual editions of Polk's Bank Directory (North America) - Fall Edition for the period 1989-1994. However, in the past Polk's Bank Directory did not always acquire full performance data (especially net income). In instances of missing data, performance data from Rand McNally Bankers Directory was used.

External Control Perspective. The environment in which banks function is defined by geographic boundaries. Competition is based upon branch location and is tracked by the Federal Reserve and the U.S. Office of the Comptroller of the Currency (OCC) on a county by county basis. The external control perspective was developed using the measurers of munificence and dynamism. Our munificence measure was adopted from Dess and Beard (1984) and Boyd (1990). Munificence was measured as the growth in the natural log of sales growth (deposits) over the 5-year period prior to firm inception utilizing the regression slope coefficient. Dynamism was measured as the standard error of the estimate for the regression of the logarithm of the value of deposits from 1984 to 1988, divided by the average logarithm of deposits for the same period (Dean, 1995).

Strategic Management Perspective. At the time of formation, bank founders are faced with two very important strategic decisions. These two decisions are the class of bank charter and the initial number of branches. The class of bank charter chosen by the founders is a strong indicator of the breadth of operation anticipated, as well as an evaluation of the needs for competition. Each of the three classes of charter affords a differing level of perceived protection to the potential customer (i.e., FDIC-insured deposits). A national bank charter has the most stringent requirements (e.g., supporting documentation, percentage amount of deposits to be held on reserve), but affords a higher level of legitimacy. The state-chartered, member has less stringent requirements than a national charter, but still affords the potential customer with FDIC protection. The state-chartered, non-member requires a substantially less rigorous review prior to startup, but does not offer FDIC insured deposits protection to potential customers (although, in some cases state agencies provided a smaller level of protection). The class variable was coded as 0 for State-charted, non-member, 1 for State-chartered, member, and 2 for Nationally chartered banks. The second proxy used for a strategic management measure was number of branches the new bank opened as operations began. Within the area that a bank operates, it can choose the level of market aggressiveness. While many of the new banks in our sample chose to open cautiously with only one branch (79%), others chose to enter the market with as many as three new branches.

Resource Perspective. Three variables were utilized to indicate the resource perspective: (1) initial financial resources, (2) the number of top management positions filled, and (3) the number of external members of the board of directors. Initial financial resources was operationalized as the initial capitalization of the new venture (Cooper, Gimeno-Gascon and Woo, 1994). The number of top management positions filled at the inception of the new venture was recorded as the number of unique individuals reported in titled top management positions. Thus, if one person was filling the role of CEO, Chief Loan Officer and Operations Officer, then we coded that as one position filled out of three possible positions. It is assumed that one person is filling multiple roles generally because a specific individual had not yet been recruited to the organization. Finally, the external members of the board of directors was determined as those members of the board who were not employed by the organization and therefore provided an external resource for the new venture.

Performance. In keeping with current research efforts to utilize multiple measures of performance, we obtained two measures of performance to analyze the separate effects of new venture profitability and firm growth. We utilized ROA as the measure of profitability and growth in deposits as a measure of sales growth as these are the predominant measures of performance in banking industry and are reported in a standard, well-regulated manner (Garcia, 1985; Roussakis, 1984).

ROA was computed as net income divided by assets. However, firm growth was somewhat more problematic. As noted by Eisenhardt and Schoonhoven (1990) regarding their research on founding conditions of semiconductor firms, percentage growth measures cannot be computed from founding, since all sales growth are initially zero. Using Eisenhardt and Schoonhoven's (1990) measure, we operationalized firm growth as the difference in sales growth (deposits) in each year of life relative to sales growth at founding. Since the sales growth at founding were 0 for all ventures, the measure of growth in any year reduced simply to the sales growth in that year. As noted by Eisenhardt and Schoonhoven, this measures the absolute change in size of each firm from a common starting point, the founding of the firm, and fits with our interest in understanding why some ventures grow more than others.

For failed ventures, performance subsequent to failing was coded in one of two ways. If the venture had a negative value for ROA in its last year, that negative figure was recorded for each year subsequent to failing. If the performance was a positive value, then a 0 was recorded for each year subsequent to failing.

 

RESULTS

Table 1 presents the results of the hypotheses testing. The top portion of the table presents the results for sales growth, and the bottom portion presents the results for ROA. Columns two and three are the R2 and F-statistic for the full regression model (external control, strategic management, and resource perspectives). Columns four and five are the change in R2 and F-statistic for the addition of the external control variables to a restricted model of strategic management and resource variables. Columns six and seven reflect the addition of the strategic management variables to a restricted model of external control and resource variables. Columns eight and nine reflect the addition of the resource variables to a restricted model of external control and strategic management variables.

The hypotheses were analyzed using hierarchical regression. The changes in R2 were tested using a procedure recommended by Cohen (1968) which accounts for the degrees of freedom. Hypothesis 1a suggested that a set of variables representing initial founding conditions will explain a significant amount of variance in new venture performance. This was supported for sales growth in all six years as the F-values ranged from 17.65 (p<.001) in year 1 to 9.43 (p<.001) in year 6, however, the hypothesis was supported for ROA only in the first two years after founding. Total R2s for both sales growth and ROA tended to diminish over time, offering support for Hypothesis 1b. The results of hypotheses 2-4 differed for sales growth and ROA. Support for the external control perspective (hypothesis 2) was supported in years one, two and three for sales growth. In year one, the addition of external control variables added almost 6 percent of the variance in sales performance. There was no support of hypothesis 2 for ROA. For both sales growth and ROA, the overall effect of environmental variables appeared to diminish over time. Support for the strategic management perspective (hypothesis 3) was supported in all six years for sales growth. There was also support for hypothesis 3 in years one and two for ROA. After peaking in year two, the effects of strategy on sales growth diminished. A diminishing effect was also noted in ROA. Strong support for the resource perspective (hypothesis 3) was shown for all six years for sales growth. Resource variables explained from 21% of the sales growth performance variance in year six to 36 % in year two. In contrast, hypothesis 3 was not supported for ROA.

 

DISCUSSION & CONCLUSIONS

We began this study by suggesting that three distinct theoretical perspectives might be used to explain the impact of initial founding conditions upon new venture performance. Furthermore, we suggested that while each perspective would explain a significant amount of unique variance, that differences in their relative explanatory power over time would assist researchers in focusing their efforts. Using data from new bank foundings, we investigated the effect of variables representing each perspective over a six-year period. This research provides empirically-based insights into several areas of entrepreneurship research. First and foremost, founding conditions and initial decisions appear to have a significant and lasting effect upon the performance of new ventures. In agreeing with the findings of Cooper, Gimeno-Gascon and Woo (1994), we were also able to address some of their concerns and extend the literature. This study incorporated ventures from their actual inception, tracked failed firms and examined performance over a six-year time period. The impact of initial resources was found to be especially strong, while at the same time we were able to detect significant and lasting importance to the strategic choices made by management at the inception of the firm. This should tend to empirically validate and encourage continued work in this area of entrepreneurship (Chandler & Hanks, 1994; Covin & Slevin, 1989; McDougall, et. al., 1994).

A second finding of this research was the diminishing effect of initial founding conditions. While intuitively we would expect this to be the case as conditions and strategies change, some research has suggested that the effect of initial conditions will amplify with time (Eisenhardt & Schoonhoven, 1990). It appears to us that additional research will need to be performed in this area. For instance, splitting our sample into high performers and low performers may indeed provide an indication of the positive-feedback response suggested by Eisenhardt & Schoonhoven. That is, the positive-feedback mechanism may only be in effect at higher threshold levels of initial resources/conditions/strategies.

A third important area this research addressed was that of performance. As suggested by researchers, and especially entrepreneurship researchers, performance is a multi-dimensional construct (Cameron & Whetton, 1983; Cooper & Gimeno-Gascon, 1992). Indeed, we found substantial explanatory differences in the performance variables of profitability and new venture growth. While the initial resources that a firm had obtained at inception was strongly predictive of growth over the next six-year period, it provided no explanatory power for venture profitability. On the other hand, the strategic choices made by the firm were predictive of new venture growth even six-years after inception, while being predictive of the firm’s profitability for the first two years after founding. We note, as others have, that ROA for the sample was very low throughout the six-year time period that was studied. It has been suggested that new ventures focus more on growth than profitability and this indeed appears to be the case with our study (Timmons, 1994).

Finally, we believe that bank formations provide researchers a unique opportunity (both for the availability of secondary data as well as the number of new firms formed each year in a single industry) to examine the new venture from the point of inception. This study has allowed us to examine independent, new ventures through a series of well-developed theoretical lenses. Examination of new venture formation and performance is extended and perhaps reoriented when we are given the opportunity to test theory with empirical examination.

Previous Page | Main Menu | Next Page

 

1997 Babson College All Rights Reserved
Last Updated 1/15/97 by Geoff Goldman & Dennis Valencia

To sign-up for the Center for Entrepreneurial Studies' publication lists,
please register with the
Entrepreneurship WebTeam.