Unbundling The Influence Of Human Capital On The New .

DLSU Business & Economics Review 28(3) 2019, p. 47-51RESEARCH ARTICLEUnbundling the Influence of Human Capitalon the New Venture’s PerformanceFerran GionesUniversity of Southern Denmark, DenmarkLa Salle–Ramon Llull University, Spainfgiones@mci.sdu.dkBrian GozunDe La Salle University, PhilippinesFrancesc MirallesLa Salle–Ramon Llull University, SpainAbstract: We use a longitudinal dataset on new ventures to assess the effects of human capital factors (education, workexperience, and entrepreneurial experience) on new ventures’ performance. Our results show how the influence of humancapital factors are dependent on the context (high vs. non-high technology industries) and illustrate the different effects ofgeneral and specific human capital factors. The findings help to clarify the existing debate on the influence of human capitalwhere we introduce a longitudinal perspective that contributes to uncover the influence of factors such as prior experience,in particular if in the same industry, as a positive influence on new ventures’ future performance.Keywords: Human Capital, Entrepreneurship, Venture Performance, Panel DataJEL Classifications: L26, M13, O30Human capital is part of the initial set of resourcesof all new ventures where a combination of theentrepreneur’s education and experience is expectedto have been transformed into knowledge and skillsthat define the specific characteristics of the newventure’s human capital. There is a shared assumptionthat human capital, as a key and single resource, has apositive influence on the new venture’s performance(Unger, Rauch, Frese, & Rosenbusch, 2011).There has been a number of researches exploringthe linkages between human capital and ventureperformance and the results have provided a varietyof insights. For example, research on the influenceof prior experience has found its effects on the newCopyright 2019 by De La Salle University

48firm’s capacity to identify and exploit new businessopportunities (Eesley & Roberts, 2012); or that whilegeneric human capital (such as education) would havea negligible influence on new venture performance,specific human capital (such as entrepreneurialexperience) would have positive effects (Rauch &Rijsdijk, 2013). On the other hand, other researcherscould not find a significant impact either from educationor work or entrepreneurial experience on new venturegrowth (Stuetzer, Goethner, & Cantner, 2012).In an attempt to gain clarity through furthercontextualization of the findings, researchers have alsostudied specific knowledge intensive contexts suchas high-tech industries where greater human capitalis expected to provide differential capabilities to newentrants (Strehle, Katzy, & Davila, 2010). Researchfindings have supported the expectation that specificsources of human capital, such as entrepreneurialspecific experience or industry-specific experience,could have a significant positive impact on performancein such context (Colombo & Grilli, 2005).As a result, despite the ongoing research on therelationship between human capital and new ventureperformance, we still have a limited understanding onthe magnitude and contingencies of the impact of thedifferent components of human capital on new ventureperformance (Unger et al., 2011). We know littleon how the different components of human capital,general or specific, impact on new venture performance(measured either as revenue or employment creation),and whether these relationships are contingent tohigh-technology contexts, or also hold for entrants innon-high-tech industries.Building on human capital theory (Becker, 1975),the purpose of this work is to systematically identifythe relationships and effects of different humancapital components in new ventures by introducingthe moderating role of the context (high or non-hightechnology industries), and exploring the time-effectson new venture performance.Material and MethodsThe data comes from the Kauffman Firm Survey(KFS), a longitudinal panel data set that tracks a sampleof new firms created in 2004 (baseline) in the US andwe follow the ventures in their first three years ofoperations (2004-2007). For more information on theF. Giones, et al.KFS survey and data design see Robb and Reynolds(2009).For our dependent variable new ventureperformance, we used revenue generation (log offirm’s revenue (M 10.28, SD 2.13)) as Model 1and employment creation (log of firm’s employees(M .75, SD .84) as Model 2 in line with prior researchsuggestions (Delmar, Davidsson, & Gartner, 2003).The human capital independent variables are measuredusing generic human capital components: education(1 to 10, from less than high school to professionalschool or doctorate), work experience (in numberof years); and specific human capital components:entrepreneurial experience (number of businessesstarted), and entrepreneurial experience in the sameindustry (whether previous businesses started in thesame industry).The panel data design and measures used offer thepossibility of overcoming two limitations: endogeneity(Colombo & Grilli, 2005) and the time lag of humancapital influence (Rauch & Rijsdijk, 2013). Giventhat an entrepreneur’s human capital factors are timeinvariant, we cannot rely on the Hausman test to choosebetween fixed or random effects (Bell & Jones, 2015)Random Effects modelling provides everything thatFixed Effects modelling promises, and much more.Crucially it allows the modelling of time-invariantvariables, and does so in a more parsimonious andexplicit way than an alternative, Plümper and TroegersFixed Effects Vector Decomposition (2007. We,therefore, introduced instrumental variables that couldcapture otherwise unobserved sources of variationand these are: firm size, market approach (product orservice), R&D intensity (as % of employees in R&Dfunction), and number of patents (Garcia-Castro, Ariño,& Canela, 2010). We were able to capture the time lagin the effects of an entrepreneur’s human capital (as atime-invariant input variable) by observing the changesin the first wave of data (wave 1) and three years after(capturing the changes across wave 1 to wave 3).The sample classification between high and nonhigh technology contexts is done by selecting theindustries that are classified as technology employers(Chapple, Markusen, Schrock, Yamamoto, & Yu, 2004)or technology generators (Paytas & Berglund, 2004)as suggested by Coleman & Robb (2012). Controlvariables for the age and gender (where 1 is male, and0 is female) of the entrepreneur, and time (data waveyears) are also introduced.

Unbundling the Influence of Human Capital on the New Venture’s PerformanceEmpirical Analysis and ResultsWe run a regression analysis using random effects(RE) for the two measurements of the dependentvariables (described as Model 1 and Model 2 in Table1), the correlation between revenues (log) and numberof employees (log) is 0.47. For each of the models,we run a separate analysis for firms classified as hightechnology new ventures and for those classified asnon-high technology new ventures. We also run aseparate test for the first data wave (one year after) andthe third data wave (three years after).The Chi2 test for all the regression analyses providedsupport to assume that none of the coefficients wouldbe 0 (Prob Chi2 0.00). The overall R2 of the differentmodels provides a measure of the influence of humancapital factors and the time effects on the performanceof the new venture (0.18 overall R2 0.35). Thebetween R2 shows that the between firm’s differencesexplain better the variability of the revenue’s growth(Model 1b: 0.26 between R2 0.33) and employees’growth (Model 2b: 0.32 between R2 0.34).The descriptive statistics and regression results canbe seen in Table 1. Model 1 (revenues as performancemeasure) shows the positive influence of educationon high-tech firms’ revenues. For the non-high-techfirms, it shows a weak, but significant, influence ofentrepreneurial experience, but a stronger positiveinfluence from entrepreneurial experience in thesame industry. Model 2 (number of employees asperformance measure) shows a weak but positiveinfluence of education (both in Model 2a and 2b), andweak influence of entrepreneurial experience in thethree-year data (Model 2b) on high-tech firm’s numberof employees. For non-high-tech firms, the results showthe influence of specific human capital factors with aweak but positive effect of entrepreneurial experienceand a stronger positive effect of entrepreneurialexperience in the same Industry.Overall, we find support on the influence of humancapital factors on new venture performance, but thissupport is not consistent across all the different factors(general or specific) that were studied as well as thedifferent models that were explored. This suggeststhat part of the current debate on the influence ofhuman capital factors can now be now clarified.While general factors, such as education, are seen tohave some influence on high tech context startups,it is the specific factor entrepreneurial experience49that makes a difference in explaining performancedifference between firms in non-high-tech contexts.Consistent across the models and contexts studied,work experience does not have a statistically significantinfluence.The instrumental variables provide additionalinformation on the influence of the firm’s size impactson new venture performance, but the coefficients areeither of similar magnitude or weaker than the directimpact of human capital factors in non-high-techfirms (Model 1a/1b). The type of firm, by R&D focusor the market approach (product/service), providesadditional information to other sources of influenceon new venture performance.Discussion and ConclusionThis article contributes to entrepreneurship researchon new venture performance by providing furtherinsights into the influence of human capital factors.First, using general (education and work experience)and specific human capital factors (entrepreneurialexperience and entrepreneurial experience in the sameindustry), we can assess and compare their effectson different contexts (high vs. non-high technologyindustries). Second, the longitudinal research designand the use of two measures for the performancedependent variable overcomes a limitation from priorresearch in the area (Rauch & Rijsdijk, 2013), andprovides additional confidence in the observed resultsin relation to prior studies. As a result, this researchfindings on the different effects of general and specifichuman factors on high-tech and non-high-tech newventures contribute to clarify the different findingsof previous studies that do not differentiate acrosscontexts (Unger et al., 2011) or that had their findingslimited to high-tech firms (Colombo & Grilli, 2005).Furthermore, we find that the results haveimplications to advance the discussion of humancapital for new venture performance. We were able toclarify the effects of general and specific human capitaldepending on the context of the firm. Human capitalfactors such as education or experience are seen to havean influence (or not) on new venture performance. Inaddition, we were able to decipher the time lag effectsof initial human capital by observing that effectsare sustained across time and contribute to explaindifferences in performance beyond the first year.