The Entrepreneurial Process

by Paul D. Reynolds and Sammis B. White (this chapter with Mary Williams)

Paul D. Reynolds and Sammis B. White, The Entrepreneurial Process: Economic Growth, Men, Women, and Minorities (Westport, Connecticut: Quorum Books, 1997).  Chapter 6, “Fledgling New Firms: Persistence after Birth” written with Mary Williams.

Combinations of factors that increase firm persistence: the process for identifying, and some results

Linear additive regression modeling, using discriminant analysis (as in Norusis, 1992b), provides [fairly poor] predictions of firm persistence (“percentage persisting increases from 63% to 72% ot 76% for the three subgroups and 73% overall; accurate predictions are made for 65%, 59%, and 63% of the cases in these three subgroups and 59% for all cases combined.  On the other hand, a preduction that all new firms persisted would be accurate in 73% of the cases.”)  {Page 146.}  Total cash provided to the new firm by the start-up team before formal financing was acquired and total formal financing are among factors not chosen for inclusion in any of the models, “This would suggest that these [..] factors were either not important, or their presence was highly correlated with the other variables.”

Traditional multiple linear regression models, even when adapted for a dichotomous dependent variable, such as new firm persistence and dicontinuance, assume that the effects of independent variables are additive.  For example, it is assumed that the effect of first-year sales can be added to the impact of a rural location and the effect of a focus on low prices to develop the prediction of the outcome.  The number of potential interaction effects—such as the joint effect of high initial sales with an urban location—is enormous; it is impossible to systematically consider the impact of all potential interactions.  For this reason, a different analysis strategy is utilized to identify significant interactions.

Systematic assessment of possible interactions between factors involves a multistage procedure, CHAID (Magidson, 1992).  It follows a strategy of identifying the single most significant factors associated with new firm persistence and then sorting all cases based on that factor.  For this reasion, each independent variable must be in the form of three to four discrete values.  Each subgroup is then considered in terms of the remaining factors that may affect firm persistence.  This procedure continues until the remaining factors no longer have any impact on new firm persistence, the impact is no longer statistically sifnificant, or the group has fewer than fifty cases.

{Page 147.}

[...]

One major benefit of this procedure is its ability to systematically identify categories of new firms with quite different levels of persistence.  [...]  There is considerable range, from 44% to 100% of the firms persisting.  But when types of firms with similar patterns of persistence are identified, they are often quite different types of firms.

For example, a group of eighty-seven new firms [..] where all (100%) are persisting appears to have strong first-year sales but little growth; no loans; start-up team financing less than $50,000 [actually a step up from the 'none' category]; a young or older start-up team; and operating in construction, manufacturing, or consumer service.  The next group of forty-four [..], where 97% are persisting, have high-growth trajectories; over $200,000 in total financing, with some contributed by the start-up teams; and locations outside metropolitan area.

{Page 152.}

[...]

The two groups with the lowest levels of persistence may reflect quite different reasons for discontinuance.  The lowest persistence (44% [..]) is found among 123 new firms with low initial sales and a low-growth pattern; the start-up team shifted from a nonwork status (students, unemployed, homemaker) to begin the new firm; and most have a post-high school education or a college degree.  They may have not had the skills or the background to operate a new firm, despite the formal education, and were forced to close the firm when resources were exhausted.  The firms with the next lowest persistence (54% [..]) were firms with low growth, but first-year sales were in excess of $75,000, and the start-up team had experience in the world of work.  These firms may have been closed when growth failed to reach expectations.  While both interpretations are speculation, it is reasonable to assume that these firms are deactivated for different reasons.  [...]  Analysis based on the assumption that all firm deactivations reflect the same processes may have limited success [..].

{Pages 152 to 153.}

[..] almost one-third of the new firms are in one group [..], which has a high-growth profile; from none to $200,000 in initial financing; and a start-up team that completed some education beyond high school.  Seven in ten of these new firms were persisting in the follow-up interview.  In a sense, these are the modal type of new firm, and their persistence is about average for the entire group.

{Page 153.}

Return to review of The Entrepreneurial Process by Sammis and White.