Does venture opportunity variation matter? Investigating systematic process differences between innovative and imitative new ventures

Abstract

The central thesis in the article is that the venture creation process is different for innovative versus imitative ventures. This holds up; the pace of the process differs by type of venture as do, in line with theory-based hypotheses, the effects of certain human capital (HC) and social capital (SC) predictors. Importantly, and somewhat unexpectedly, the theoretically derived models using HC, SC, and certain controls are relatively successful explaining progress in the creation process for the minority of innovative ventures, but achieve very limited success for the imitative majority. This may be due to a rationalistic bias in conventional theorizing and suggests that there is need for considerable theoretical development regarding the important phenomenon of new venture creation processes. Another important result is that the building up of instrumental social capital, which we assess comprehensively and as a time variant construct, is important for making progress with both types of ventures, and increasingly, so as the process progresses. This result corroborates with stronger operationalization and more appropriate analysis method what previously published research has only been able to hint at.

Appendices

Appendix

Table A1 Mean, standard deviations and correlations: innovative ventures (n = 40)

Table A2 Mean, standard deviations and correlations: imitative ventures (n = 219)

Table A3 Longitudinal growth model results initial status mean variation

Table A4 Residual results

Technical appendix

The statistical development to be drawn upon in longitudinal growth modeling is termed random coefficient models. This technique goes beyond traditional equation modeling of longitudinal data and its focus on auto-regressive models (cf. Maruyama 1998). It goes further in terms of including both a mean and a covariance structure. The growth model is multilevel in that an individual’s observation over time is correlated. A second part of the model describes individual variation in growth parameters in terms of person-specific, time invariant covariates. Compared to a single regression that gives an intercept and a slope estimate of all individual units, this approach gives an intercept and slope estimate for each individual unit (Muthén and Khoo 1998). In addition, this approach accounts for similarities among individuals by stipulating that all individuals’ random effects come from a single, common population. In statistical terms, the growth model is specified as:

$$ \gamma it=\alpha i+\beta i+\zeta it $$

(1)

Here αi and βi are individual specific parameters describing initial level (intercept) and rate of growth (slope) of the entrepreneurial opportunity recognition process and $\zeta it$ represents time varying residuals. The regression intercept and slopes are random parameters that vary over individuals. No specification of linear growth is necessary; linearity can be estimated whenever there are data on a sufficient number of time points (see Muthén and Khoo 1998 for a technical description).

It is convenient to view initial status and growth rate as latent variables. In order to understand this it is useful to consider the specification of this model into two parts. The first part includes terms contributing to the means of the observed variables and terms contributing to variances and covariances among these variables. First the mean structure, intercepts in the regression of the y’s (gestation behavior) on the two latent variables (Igestb and Ggestb) are parameters which should be held equal across time to reflect that these are measured at all time points (i.e. the same metric). The latent variable (α = initial status) is fixed at zero. The growth in the observed variable means over time is captured by the latent variable (β) (Muthén and Khoo 1998). Extending the model like this with time-invariant covariates (wi = resources and strategy), the individual variation in these parameters is specified as:

$$ \alpha i=\alpha+\gamma\alpha wi+\delta\alpha i $$

(2)

$$ \beta i=\beta+\gamma\beta wi+\delta\beta i $$

(3)

When further expanding the model by adding a time-varying covariate (vit) to the growth curve of Eq. 1, introducing time-specific deviations from the growth curve the following equation is used:

$$ Yit=\alpha i+\beta it+ ytvit +\zeta it $$

(4)

Assuming for simplicity that there is no time-varying covariate (v), the model can be seen to imply growth in means and variances as a function of time (t).

$$ \hbox{E }(yit \vert wi)=\alpha+\gamma\alpha wi+ \beta (\gamma\beta wi) t $$

(5)

$$ V (yit \vert wi)=\sigma^{2}_{\alpha}+2t\sigma_{\alpha \beta}+ t^{2}\sigma^{2}_{\beta}+\sigma^{2}_{\zeta} $$

(6)

The growth model can be viewed as a structural equation model with latent variables (cf. Maryama 1998). The αs and βs can be viewed as latent variables instead of random parameters. They are both unobserved variables varying across individuals. Another application is when (t) is not varying across individuals. In this situation, (t) in Eq. (4) can be considered as a fixed regression parameter (factor loading) for the variable βi. According to this, beta can be estimated when fixing the first two (t) values, thereby capturing also non-linear growth (Muthén and Khoo 1998).