Breadth and New Venture Growth
A fundamental aspect of new venture strategy is the level of aggressiveness with which the new venture will enter the market, with broader breadth strategies being more aggressive than narrow breadth strategies. Analyzing a small sample of corporate-sponsored new ventures, Biggadike (1979) used a combination of posture and marketing mix to determine that entry on a large scale was a key to success, at least in fast growing markets. This study was followed by Sandberg's (1986) study that found a positive relationship between new venture performance and firm's that pursued wider markets with a broad array of products. This broad breadth approach to new firm success has been supported in other literature tracking new internal corporate ventures (Miller & Camp, 1985; Kekre & Srinivasan, 1990) as well as young, independent firms (McDougall, et.al., 1994). Given this evidence, we examine this aspect of new venture strategy breadth as a combination of factors related to product breadth and the number of products offered to the market.
Breadth. A broader product line is an indication of the market that the new firm is attempting to capture. In studies of new corporate-sponsored ventures, this relationship has been associated with improved performance (Biggadike, 1979; Kekre & Srinivasan, 1990; Miller & Camp, 1985). Buzzell & Gale (1987) found that a broader product line was related to increased market share and was a key element of successful growth. However, as suggested earlier, these ventures were supported by existing corporations and were therefore endowed with unique resources not traditionally available to an independent new venture. Little direct evidence is available on new ventures at the point of formation with respect to the number of products they offer and the breadth or focus of the product offering to the market. Banks have a wide variety of both deposit and loan products available for marketing. However, banks will only have sales for those products that they wish to support and (as other types of retail outlets) will either focus their efforts on a few products or attempt to satisfy a wider market by supporting a wider variety of product offerings. Given the empirical evidence, we would expect the following:
H1: New venture breadth will be positively associated with new
Initial Capital. While it can be effectively argued that the ability of the founding team to attract capital is a significant indicator of both the ability of the firm's management as well as an indicator of future intentions, it is also relatively apparent that a firm with high initial capitalization would tend to have the opportunity to grow faster and perform better than other firms regardless of the strategic posture chosen (Arshadi & Lawrence, 1987; Duchesneau & Gartner, 1990; Eisenhardt & Schoonhoven, 1990). A study of extant entrepreneurship research found that increased initial capital was associated with better performance in more than three-fourths of the studies examined (Cooper & Gimeno-Gascon, 1992). "Startup capital serves three
purposes: 1) to purchase the assets needed to operate a business; 2) to sustain a business during its early period when cash flows are likely to be negative; and 3) to buffer against management mistakes, environmental uncertainties, and other unforeseen difficulties" (Castrogiovanni, 1996). Therefore, we hypothesize the following:
H2: Initial capital will be positively associated with new venture growth.
Breadth and Initial Capital. The theoretical underpinnings of new venture entry suggests that firms focus on a limited set of products and services. This is suggested to limit potential retaliation, more properly serve a niche market, and husband resources while administrative systems are developed (Abell, 1980; Porter, 1980, 1985; Scherer & Ross, 1990; Vesper, 1990). The foundation of this approach is the assumption that new firms will have only limited financial resources and will be unable to withstand the shocks that larger, better endowed, established firms can handle (Hambrick & D'Aveni, 1988; Venkataraman, Van de Ven, Buckeye & Hudson, 1990). However, much of the new venture research on breadth has utilized corporate-sponsored ventures (Biggadike, 1979; Miller & Camp, 1985; Kekre & Srinivasan, 1990). These ventures are unlike independent ventures as they are supported both financially and often administratively by a parent company. Therefore, it is not surprising that aggressive strategies have been so universally and directly attributed to the success of these new ventures. These firms can respond to retaliation, obtain well developed administrative systems, and can fund shocks that might eliminate independent firms.
Research on independent ventures is more mixed. This research stream has had a more difficult time tapping into the initial financial resources of the new venture, has typically measured various aspects of the firm several years after firm birth (usually firms that are less than eight years old), and has traditionally examined the environment as a critical factor without accounting for the initial capital of the firm (Chandler & Hanks, 1994; McDougall, Covin, Robinson & Herron, 1994; Romanelli, 1989; Venkataraman, Van de Ven, Buckeye & Hudson, 1990). The exception to this is a wide-ranging study by Eisenhardt and Schoonhoven (1990) that did include initial capital as a control variable, but primarily examined the interaction of strategy and the environment.
Given this, we suggest that the dissonance between the theory of new venture entry and the empirical results may be due to the interaction of the breadth of the market entry and the capital the firm has acquired at the point of inception. Examining the available research, we would expect that the broader the breadth of the initial strategy of a new venture, the more initial capital that would be required. This might provide an effective explanation for the similarity of findings in corporate-sponsored venture research as well as the inconsistent findings in independent-firm research. Therefore, we hypothesize the following:
H3: The level of financial resources available to the venture
at inception will moderate the relationship between new venture breadth
We propose to examine these questions utilizing a multiple linear regression
methodology on a newly developed data base of banks formed in the United
States between 1985 and 1988. The database allows us to examine performance
over a 5-year period for each of the over 600
independent banks that were chartered during those years. The data has been gathered over the past two years with the direct assistance of the U.S. Office of the Comptroller of the Currency (OCC), in addition to data from: 1) the Board of Governors of the Federal Reserve System ; 2) the Federal Deposit Insurance Corporation (FDIC); 3) the U.S. Census Bureau; and 4) Polk's Bank Directory. Bank formations provide researchers a unique opportunity (both for the availability of secondary data at the point of firm formation as well as the number of new firms formed each year in a single industry) to examine the new venture at the point of inception. While it is often difficult to establish the specific date in which a new venture begins operations (Vesper, 1990), the date of inception for banks is clear as regulations require that all banks be chartered prior to beginning operations.
Using data provided by the U.S. Office of the Comptroller of the Currency, we identified 814 bank formations during the period 1985-1988. Of these, 120 were identified as subsidiaries of other banks or bank holding companies (and therefore were excluded from the sample). Also excluded were trust banks and credit card banks as these are established with unique missions and are not generally comparable to commercial banks. Sixty-six banks had merged, been closed by regulators or had submitted to voluntary liquidation within the five-year time frame of this study. Due missing information, the final dataset consisted of 491 independent new bank ventures, although complete performance and product data was not available on all of these. Variables included: 1) Breadth; 2) Initial Capital; and 3) A group of control variables designed to control for the environmental conditions at founding.
Breadth. Breadth is a combination of both the breadth and number of products offered to the market (Romanelli, 1989). A factor analysis was performed on the three variables strongly suggested to affect this construct as outlined in the literature and available to the researchers (Romanelli, 1989; Sandberg, 1986). These three variables (number of actual products offered, breadth of loan product mix, and breadth of deposit product mix) loaded on one factor with a Cronbach alpha of .70. The factor score from the varimax-rotated solution was used as the measure.
Breadth of Loan & Deposit Product Mix. A concentration measure was developed as an indicator of the breadth of activity within the firm (Arshadi & Lawrence, 1987; Lindley, Verbrugge, McNulty, & Gup, 1992; Reger, Duhaime & Stimpert, 1992). This was measured as a Herfindahl-type index separately calculated for loans and then for deposits. Actual dollar sales during the first year of operation was recorded for each product offered and a Herfindahl measure was calculated as the sum of the squared percentage of either total loans or total deposits This index ranges from 0 to 1 with a higher score equating to more concentration (Boyd, 1990). This data was gathered from the Board of Governors of the Federal Reserve System.
Number of Products. Banks have a wide variety of potential products that may be marketed. All potential products were examined and each product was coded as a 0 for no sales or a 1 for those products that had sales during the first year of operation. These codes were summed for a total number of products being actively offered to the market (Romanelli, 1989). This data was gathered from the Board of Governors of the Federal Reserve System.
Initial Capital. Initial capital was operationalized as the actual dollar
amount raised at the point of founding (Eisenhardt & Schoonhoven, 1990;
Cooper, Gimeno-Gascon & Woo, 1994). This data was reported on the original
charter application to the OCC.
Growth. Growth measures are difficult but critical measures of new venture performance. While profitability measures may be of interest to general organizational researchers, it is problematic in new venture research (Brush & Vanderwerf, 1992). 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 are initially zero. Using Eisenhardt and Schoonhoven's (1990) measure, we gathered this data from the Board of Governors of the Federal Reserve System and operationalized firm growth as the difference in sales (deposits) at year 5 as compared to sales at founding. Since the sales at founding were 0 for all ventures, the measure of growth in any year reduced simply to the sales 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.
Control Measures. As this research focuses upon the strategic decisions made at the inception of a new venture, we felt that it was incumbent upon us to eliminate alternative explanatory variables. As such, we control for four potential environmental areas: 1) Munificence; 2) Dynamism; 3) Competitive Intensity; and 4) Date of Inception.
Munificence, dynamism, and competitive intensity were calculated for the county in which the new bank was based. The authors knowledge of the industry as well as previous new bank research (Arshadi & Lawrence, 1987; Lindley, et.al., 1992) suggests that 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 FDIC on a county by county basis. Despite recent press reports and the efforts by the largest banks in the country, most banks (and especially those that are new) do not operate on an interstate level. No bank in the sample began operations in multiple counties and indeed, an examination of these banks three years after formation found that 97.3% were still operating solely within one county.
Our munificence measure was adopted from the literature (Boyd, 1990; Dess & Beard, 1984) and was measured as the annual percentage increase in countywide bank deposits (value of shipments) for the five-year period prior to the bank opening. This was calculated as 100 multiplied by the regression coefficient of the trend in the natural logarithm of the county bank deposits over the five year period and was gathered from the annual Data Book: Operating Banks and Branches published by the Federal Deposit Insurance Corporation which tracks all market data by county and state within the United States. Dynamism was measured as the standard error of the estimate for the regression of the natural logarithm of the county bank deposits for the five-year period prior to the bank opening, divided by the average (for all years) natural logarithm of county deposits for the same period (Dean, 1995). Dividing by the county average adjusts this measure for the size of the countywide banking industry (Dess and Beard, 1984). Competitive intensity aims to measure the density of banking in each county (Swaminathan, 1996). The measure was calculated as the population in the county divided by the number of bank branches in the county in the year of formation (Arshadi & Lawrence, 1987). This information was gathered from the U.S. Census and the previously mentioned Data Book: Operating Banks and Branches published by the Federal Deposit Insurance Corporation.
Previous research has suggested that imprinting at the time of founding and the general conditions of the country can have an impact of the success/failure of new firms (Carroll & Delacroix, 1982; Carroll & Hannan, 1989; Stinchcombe, 1965). As this research is longitudinal, encompassing four years worth of foundings, this variable will be coded as the year of initial operations for each new venture.
An analysis of the individual variables resulted in the need to transform
two of them as they violated the tenets of normal distributions necessary
for effective linear regression analysis (Neter, Wasserman & Kutner,
1990). The dependent variable (growth) and initial capital were both skewed
and somewhat kurtotic. Examination of scatter plots suggested a natural
log transformation would be effective. After transformation each variable
was improved without affecting the distributions. All variables were
tested within a linear regression model as outlined below.