The importance of new businesses for regional and economic growth was recognized in Schumpeter’s seminal work, The Theory of Economic Development, in 1934 (Del Sarto et al., 2020). Scholars, practitioners, and policymakers acknowledge that startups significantly contribute to economic expansion, job creation, and societal prosperity (Audretsch et al., 2006; Pradhan et al., 2020).

Nascent ventures face numerous challenges, including operational, competitive, resource, and planning issues, threatening their long-term survival. Startups are particularly vulnerable in their early years, with a 30% failure rate within the first two years (Picken, 2017; Santisteban & Mauricio, 2017). However, this risk decreases over time (Yang & Aldrich, 2017).

The survival of new firms can be analyzed through three main factors: personal attributes, firm-specific characteristics, and the external environment (Brüderl et al., 1992). Challenges to survival often include limited financial resources (Smilor, 1997) and an inexperienced founding team (Gruber et al., 2008).

While the existing literature on startup survival is extensive, it often remains confined to a single theoretical perspective, region, or industry (Andreeva & Garanina, 2016; Baum & Silverman, 2004; Cressy, 1996; Gimmon & Levie, 2010). Most studies have been conducted in advanced economies such as Europe, North America, and Asia, with less emphasis on emerging economies (Azeem & Khanna, 2023). Empirical research employs regression-based models, particularly logistic regression models (LR), to examine the relationship between various antecedents and outcomes, often focusing on internal resources and non-financial performance measures.

Given the limited scope of previous research, future studies should prioritize cross-country analyses. The rise of startups in emerging economies and the expanding research on startup survival offer numerous opportunities for new investigations. These studies can explore the interplay between internal and external factors affecting startup survival in these emerging markets. Therefore, it is essential to include emerging countries in such research endeavors.

This study examines startups in global accelerator programs, including those in emerging economies, using the resource-based view (RBV) and human capital theory. Building on Cooper et al. (1994), it posits that initial financial and human capital allocations predict new venture survival (Van Praag, 2003).

Predicting new venture performance based on initial observable factors is a key interest for entrepreneurship researchers, as it can optimize resource allocation, benefiting both entrepreneurs and society (Dahlqvist et al., 2000).

The survival of startups has been extensively studied from various theoretical perspectives. Azeem and Khanna (2023) highlight the resource-based view (RBV) as the most frequently cited perspective, forming the foundation of our study. RBV posits that a company’s unique, valuable, difficult-to-replicate, and irreplaceable resources and capabilities confer a competitive edge and enhance performance (Barney, 1991; Wernerfelt, 1984). Among these resources, intellectual capital (IC) is an intangible asset intricately linked to a company’s strategy, which is crucial for longevity (Rossi et al., 2016).

Our research is informed by the Resource-Based View (Barney et al., 2001) and in human capital theory (Becker, 1964; Mincer, 1974), underscores the pivotal role of individual expertise and abilities in driving economic productivity. Consistent with Cooper et al. (1994), we argue that a company’s initial financial resources and human capital are fundamental to its survival, especially for nascent enterprises (Van Praag, 2003). This perspective highlights the essential contribution of both tangible and intangible resources possessed by founding teams to the success of a startup (Strotmann, 2007).

Our study examines startups, the engines of innovation and economic growth, focusing on the role of initial intangible knowledge, or intellectual capital (IC), during the critical pre-seed phase. By controlling financial availability, environmental conditions, and founding team characteristics, we aim to underscore the significance of IC in startup longevity.

Nascent enterprises frequently join accelerator programs for financial support and expert guidance (Radojevich-Kelley & Hoffman, 2012). This study investigates the correlation between the intellectual capital (IC) of founding teams - including their skills, knowledge, experience, relationships, and abilities - and the survival rates of these ventures. These startups often emerge in unstable environments to seize new market opportunities (Beckman & Burton, 2008; Davidsson & Honig, 2003). Our objective is to determine how a founder’s initial IC influences the survival chances of these ventures.

Our study focuses on startups in global accelerator programs, utilizing data from the Entrepreneurship Database Program (EDP) at Emory University, part of the Global Accelerator Learning Initiative (GALI, 2020). Launched in 2013, the EDP examines the causal effects of accelerating impact-oriented ventures. It employs a standardized set of questions used by all participating accelerators in their application processes, enhancing responsiveness and allowing observation of the entire pool of serious applicants. Program managers also track which applicants join their programs (Lall et al., 2020). These programs prioritize the IC of founders according to their selection criteria (GALI, 2021).

Most studies on startup survival exhibit a geographical bias, primarily focusing on developed countries (Azeem & Khanna, 2023). Our research aims to address this gap by incorporating geographic and socio-economic diversity. Recognizing the lack of comprehensive studies on startups in emerging economies (Andreeva & Garanina, 2016), we seek to rectify this. Our study expands the scope by examining the impact of intellectual capital (IC) - including both initial financial resources and intangible knowledge - on the global growth of startups, with a focus on the pre-seed stage of accelerator programs.

This study employs a logistic regression model to evaluate the influence of numerous factors on the survival probabilities of startups within our sample. The model also predicts the likelihood of a startup’s success during its initial stages. Our primary objective is to enhance the understanding of how intellectual capital (IC) impacts the success of pre-seed startups, particularly those participating in accelerator programs.

This study offers valuable insights for a range of stakeholders. Policymakers can leverage this knowledge to craft policies that foster startup growth by emphasizing the importance of intellectual capital (IC) in team formation, thereby cultivating a dynamic and prosperous startup ecosystem. Accelerator programs and venture capitalists can utilize these findings to refine their selection and support processes, enhancing the success rates of the startups they endorse. Similarly, practitioners, including startup founders and employees, can apply this knowledge to focus on developing and enhancing their IC, enabling them to better adapt to growing market opportunities. The structure of the remaining sections of the study is as follows: The second section establishes the relevant literature that supports the conceptual framework and the hypotheses under study; the third section discusses materials and methods, followed by the estimation results and their discussion.

Entrepreneurship involves a dynamic process of team formation and adaptation to meet customer needs. Salamzadeh and Kesim (2015) describe business development as a life cycle with stages including idea conception, prototype development, market entry, product sales, and job creation. Wong et al. (2005) argue that preparation for the startup phase is the crucial foundation that shapes emerging ventures. Brüderl and Schussler (1990) suggest that early survival indicates initial success, while long-term endurance highlights the importance of adaptability in later stages.

The resource-based view (RBV) posits that a founder’s attributes and circumstances significantly impact venture performance. Dencker et al. (2009) found that knowledge positively influences startup survival in Germany, with the cumulative experience of founders bolstering knowledge integration. Research has shown a varied positive correlation between talent and performance, dependent on a country’s macroeconomic development (Furlan, 2019; Mayer-Haug et al., 2013). Studies have also demonstrated a differentiated positive impact of talent across different entrepreneurial stages and economic contexts (Kerrin et al., 2017). However, these effects can be nonlinear or varied, contingent on factors such as business development stages, survival duration, administrative maturity, technological orientation, funding sources, performance measures, and the specific conditions of the countries and sectors (Delmar & Shane, 2006).

Human capital theory suggests that an individual’s human capital, comprising their education, work experience, and job training, is vital for achieving organizational goals, securing a competitive edge, and enhancing financial performance (Becker, 1994; Unger et al., 2011). Gimmon and Levie (2010) applied this theory to examine how founder qualities affect the ability to attract external investment and ensure the survival of new high-tech firms. Their research underscores that a founder’s managerial experience and academic credentials are more critical in attracting external investment than their technological expertise.

Cooper et al. (1994) distinguished four types of initial capital in their influential study: general human capital, management expertise, industry-specific knowledge, and financial capital. General human capital, from the entrepreneur’s background, includes knowledge that increases productivity and access to network resources. Management expertise is tacit knowledge from the entrepreneur’s previous general management experience, learning by doing or observing. Industry-specific expertise is also tacit and expensive to obtain without prior industry experience, and it is essential for understanding the context of suppliers, competitors, and customers. Financial capital, the most concrete form of capital, serves as a cushion and allows strategic flexibility.

This study leverages the resource-based view (RBV) and human capital theory, building on Cooper et al.’s (1994) findings, to propose that initial allowances of financial and human capital are reliable predictors of the survival of new ventures (Van Praag, 2003).

Intellectual capital (IC) is widely acknowledged but lacks a precise definition (Bontis, 1998). It gained prominence with the rise of knowledge-based assets, initially quantified as the difference between market and accounting values (Brennan & Connell, 2000). Gary Becker’s modernization of the human capital (HC) concept in 1993 reignited academic interest in HC (Nyberg & Wright, 2015). The resource-based view (RBV) paradigm shifted focus from physical and financial resources to intangible assets (Abeysekera, 2021; Spender, 1996), positioning IC as a strategic resource for emerging technological ventures (Juma & McGee, 2006).

This study defines IC as an organization’s intangible assets, based on Stewart’s (1997) concept of IC as ‘packaged useful knowledge.’ These assets include employee knowledge, adaptability, customer and supplier relationships, brands, intellectual property, product trade names, internal processes, and R&D capabilities. Although not listed in traditional financial statements, these assets create future value and competitive advantage. IC often shows values three to four times higher than their book values (Edvinsson & Malone, 1997).

Bontis (1998) suggests a research framework with three dimensions of intellectual capital (IC): human, social, and structural. This framework, which views IC as valuable interconnected elements, explains IC’s dimensions (Marr & Moustaghfir, 2005). It also enables the empirical examination of IC components’ impact on performance (Felício et al., 2014).

Human capital (HC) attributes, such as education and experience, are associated with small business success (Baptista et al., 2014). Financial intermediaries and venture capital firms value entrepreneurial experience highly when assessing startups, using managerial skills and experience as primary selection criteria (Piva & Rossi-Lamastra, 2018). HC is crucial in knowledge-based companies (Bosma et al., 2004), and the value of specific human capital is evident in new business founders’ entrepreneurial experience, especially among habitual entrepreneurs who have previously founded at least one business (Baptista et al., 2014).

Relational capital (RC), also known as social capital (SC), is an intangible asset that values relationships. It involves cultivating, preserving, and enhancing quality relationships with individuals, organizations, or groups that impact business performance (Welbourne & Pardo-del-Val, 2009). RC includes knowledge from relationships with stakeholders such as customers, suppliers, and industry associations, which influence the organization, add value, and strengthen operations. These relational networks are crucial business resources, enabling entrepreneurs to access otherwise unavailable resources (Bandera & Thomas, 2018; Burt, 2017). A significant aspect of SC is the reputation, experience, and contacts that facilitate entrepreneurs’ access to financing (Baum & Silverman, 2004). New ventures can improve their financing conditions through effective communication with investors and customers (Gardner & Avolio, 1998).

Structural capital (STC) is recognized in the literature as the company’s internalized knowledge. It pertains to the organizational structure and systems that bolster employee productivity (Edvinsson & Malone, 1997). It encompasses all non-human intangible assets of an organization, including culture, philosophy, internal processes, information systems, databases, organizational charts, process manuals, software, planning, strategies, routines, technology, and intellectual property rights such as patents, trademarks, and copyrights. The value of these assets to the company surpasses their material worth (Abdulaali, 2018). Intellectual property is often the only source of competitive advantage for knowledge-based companies (McGee & Dowling, 1994). Structural capital focuses on organizational efficiency, and its value derives from internal infrastructure, processes, and culture on the one hand and from the adaptive and development strategies adopted by the company on the other (Brennan & Connell, 2000).

The resource-based view (RBV) theory highlights human capital as a crucial determinant of firm performance (Barney et al., 2001). HC, characterized by knowledge, is both valuable and challenging to replicate. Studies have demonstrated a positive correlation between knowledge-based intangibles and performance (Davidsson & Honig, 2003; Kellermanns et al., 2016).

Coff (1997) provided moderate evidence supporting HC as a strategic resource. The presence of contradictory findings could be attributed to factors such as path dependence, the inability of cross-sectional studies to capture delayed effects, and the efficiency of the labor market for specific forms of HC (Coff, 1997).

In the context of human capital, both general and specific forms significantly impact a startup’s performance (Cressy, 1996; Gimmon & Levie, 2010). Bosma et al. (2004) investigated the effects of three types of investments - general, industry-specific, and entrepreneurship-specific - on startup survival, profitability, and employment. Their study found that these targeted investments substantially improve these performance metrics.

Research on the impact of initial intellectual capital accumulation on the performance and survival of emerging ventures, especially during the pre-seed phase and within high-impact acceleration programs, is scarce. Most studies focus on startups in developed countries and specific industries (Azeem & Khanna, 2023). This necessitates further validation across a broader spectrum of startups.

Intellectual capital, including knowledge, skills, and experience, is crucial for a startup’s success, particularly in the pre-seed phase. However, current research inadequately addresses this, especially for high-impact startups and those in acceleration programs. Most studies are limited to developed countries, leaving a knowledge gap about startups in diverse geographical contexts. They also tend to focus on specific sectors, missing the broader startup ecosystem (Andreeva & Garanina, 2016). Therefore, expanding research to include startups from various locations, sectors, and development stages is essential to understanding how intellectual capital affects startup performance and longevity.

Our research focuses on startups, which are key drivers of innovation and economic growth. We aim to investigate how initial financial resources and intangible assets, collectively known as intellectual capital (IC), impact a startup’s development during the crucial pre-seed phase. This phase involves setting business objectives, identifying potential hurdles, establishing market positions, and devising strategic plans.

We particularly focus on emerging ventures seeking financial aid and expert advice through accelerator programs. Our goal is to examine the correlation between IC - which includes the founding team’s skills, knowledge, experience, relationships, and capabilities - and the survival rate of nascent for-profit ventures.

We define ‘startups’ and ‘nascent for-profit ventures’ as organizations created in volatile environments to seize new market opportunities. Our primary goal is to determine how a founder’s IC influences the survival prospects of these ventures.

Our key research question is: “How does the initial allowance of intellectual capital of startups influence their survival rate?” Answering this question is crucial as it highlights the role of IC in determining a startup’s future.

Our research focuses on startups in accelerator programs within the EDP sample, emphasizing the intellectual capital (IC) provided by founders. We hypothesize that this initial IC positively influences startups’ survival probabilities during the pre-performance phase, providing a competitive advantage in business navigation, investment acquisition, and survival.

To validate our hypothesis, we use a logistic regression model to quantify the impact of numerous factors on the survival probabilities of startups in our sample and predict their success likelihood during the critical pre-performance phase.

We aim to deepen the understanding of IC’s role in startup success, particularly within accelerator programs across diverse geographical and socio-demographic contexts. The findings could significantly influence founders, investors, and policymakers in designing programs that support and nurture emerging ventures.

Our analysis of the EDP sample, which includes 14,457 for-profit ventures from 164 countries, reveals a strong social orientation and success biases. Ventures that have been operational for at least three years show a survival rate of 31% at the time of application, with over half generating revenue and 78% expanding their workforce beyond the founding members. Notably, 58% of these companies operate on proprietary technology.

About one-third of these ventures have secured external equity investment, while a quarter have taken on debt for startup expenses. Philanthropic contributions support a larger portion. Ventures led by female founders are less likely to secure equity investments but have a higher likelihood of having positive revenues in the preceding year. Over 10% of the ventures in the sample are directed by women.

Ventures led by experienced entrepreneurs or those with previous company-founding experience tend to attract more equity investments and report revenues and employees in the preceding year. Similarly, ventures with founders who hold patents, copyrights, or trademarks also show a higher tendency to attract equity investments and report revenues and employees in the preceding year.

However, as expected, the sample may exhibit a selection bias, as program selectors often favor ventures with more established records (Hallen et al., 2020). Participants in these programs are significantly more likely to report revenues in the preceding year (GALI, 2020, 2021).

The two most frequently used non-financial startup performance measures are ‘survival’ (Adams et al., 2019; Brüderl & Preisendörfer, 1998; Mas-Verdú et al., 2015; Wamba et al., 2017) and ‘growth’ (Haeussler et al., 2019; Vanderstraeten et al., 2016). To provide empirical validation for our research question and working hypotheses, we selected ‘survival’ as our dependent variable (DV), a binomial variable that takes the value of 1 if the startup remains in operation three years after its inception and 0 otherwise (Hyytinen et al., 2015).

Regarding our predictive variables, we focus on the influence of the founding teams’ IC intangibles on the startups’ likelihood of survival, as categorized by three types: human capital, organizational or structural capital, and social capital (Bontis, 1998; Wang & Chang, 2005; Yang & Lin, 2009).

Our logistic regression model integrates the initial intellectual capital endowments, utilizing principal components extracted from an exploratory factor analysis (Aguilera et al., 2006). The resulting variables are ‘F2 SHC experience’, which reflects the founding team’s for-profit venture experience; ‘F3 SC media presence’ (Bandera & Thomas, 2018), indicating the entrepreneur’s network via LinkedIn and Twitter (Song & Vinig, 2012), and two components, ‘F4 STC innovation’ and ‘F5 market rights’ (Alvarez-Salazar & Seclen-Luna, 2023), related to the venture’s structural or organizational capital, covering aspects like intellectual assets, databases, organizational culture, structure, patents, and trademarks (Abdulaali, 2018).

New business survival hinges on three-factor categories: personal, business-specific, and environmental (Brüderl et al., 1992). We consider a range of these factors, including external economic conditions, team dynamics, and venture specifics, to study the effect of a venture’s initial intellectual capital (IC) endowment. Acknowledging their influence on survival probability, we use these factors as control variables in our analysis.

The ‘F1 country context’ factor includes country-specific variables (economic context) derived from World Bank Indicators (WBI). These variables reflect the national economic environment faced by startups. Additionally, ‘startup density,’ as defined by the World Bank (2021), signifies the number of newly registered corporations per 1,000 individuals aged 15-64, offering insight into the entrepreneurial ecosystem.

In terms of firm or venture founding team characteristics in our sample, we include the variable ‘female presence,’ which denotes the degree of female representation or influence in a venture’s founding team. Women’s entrepreneurship plays a significant role in a country’s economic growth, particularly in developing economies (Cardella et al., 2020). In this study, if women constitute two-thirds or more of the founding team, it is considered to exhibit strong female dominance (Barnir, 2014).

Regarding the influence of finances on survival probabilities in the face of financial constraints (Fuertes-Callén et al., 2022; Lee & Zhang, 2011; Wamba et al., 2017), we include two categorical variables ‘has equity’ and ‘has debt’. These variables are coded as follows: 1 signifies that the venture has obtained equity/debt financing either at its inception or in the year following the application; 2 means that the venture has secured financing on both occasions; 0 is used when the venture has not received any such financing. To gain a deeper understanding of the effect of specific financial goals on performance, we incorporate the concept of exhibiting a distinct profit margin, serving as a financial planning proxy (Miner & Raju, 2004) based on the foundational startup survival studies by Cressy (1996). The ‘has target margin’ variable is binary, with a value of 1 assigned if the venture operates with a predetermined target margin and 0 if not.

Business accelerators significantly contribute to the facilitation of new venture creation. Startups seek top-tier accelerators to expedite their developmental journey (Salamzadeh & Markovic, 2018). These accelerators provide early-stage funding and essential mentorship, driven by their confidence in the startup’s potential, personal interest in the concept, or admiration for the entrepreneurial team (Radojevich-Kelley & Hoffman, 2012). The characteristics of accelerator programs are known to influence venture performance. Although we cannot explicitly measure specific accelerator characteristics, we will account for program-specific unobserved heterogeneity in our subsequent analyses, thereby indirectly accommodating program differences.

The EDP examines several variables for accelerator program characteristics. These include ‘application year’ (the submission year of the venture’s application); this variable accounts for temporal variations and the impact of learning over time, and partly alleviates the effect of unobserved heterogeneity; participated - ventures that completed the program (GALI, 2021); ‘accepted initially’ - ventures initially selected into the program; ‘program impact area’ - whether the program has a specific impact area (Lall et al., 2020); and ‘program with curriculum structure’ - whether the program provides a structured curriculum to participants (Kramer et al., 2023). Table 3 exhibits the descriptive statistics for the variables in the LR model.

Table 3

Source	Variable/Label	Type	Mean	Std. Dev.	Min.	Max
EDP Coded	DV Success (survival)	Binomial	0.31	0.46	0	1
EDP Factored	F2 SHC (experience)	Numeric	0.03	0.98	-1.99	1.24
EDP Factored	F3 SC (media presence)	Numeric	0.006	0.99	-1.63	2.03
EDP Factored	F4 STC (innovation)	Numeric	0.017	1.01	-1.31	2.42
EDP Factored	F5 (market rights)	Numeric	0.026	1.01	-1.16	2.18
WBI Factored	F1 Country Context	Numeric	-0.03	0.99	-1.96	2.18
WB Entrepreneurship Data	Startup Density	Numeric	1.29	2.69	0	28.59
EDP Coded	Female Presence	Binomial	0.11	0.31	0	1
EDP Coded	Has Target Margin	Binomial	0.68	0.47	0	1
EDP Coded	Has Equity	Categorical	0.33	0.69	0	2
EDP Coded	Has Debt	Categorical	0.24	0.59	0	2
EDP Survey	Application Year
	2014	Categorical	0.055	0.23	0	1
	2015	Categorical	0.056	0.23	0	1
	2016	Categorical	0.19	0.39	0	1
	2017	Categorical	0.19	0.39	0	1
	2018	Categorical	0.28	0.45	0	1
	2019	Categorical	0.19	0.27	0	1
EDP Survey	Participated# Program Impact Area
	0 1	Binomial	0.41	0.49	0	1
	1 0		0.09	0.29	0	1
	1 1		0.08	0.44	0	1
EDP Survey	Accepted Initially	Binomial	0.19	0.39	0	1
EDP Survey	Program with a Curricular Structure	Binomial	0.75	0.43	0	1

Note. Binomial Variables are assigned a value of 1 when present and 0 when absent. The Application Year represents the survey year. For Equity/Debt Presence, categorical variables are assigned as follows: a value of 1 is given if an investment is present either at inception or at the end of the last calendar year, a value of 2 is given if an investment is present at both inception and the end of the last calendar year, and a value of 0 is given in all other cases.

For the validation of our hypothesis, we use a logarithmic regression model (LR), which is considered suitable when the response variable is dichotomous, and the effect of predictors is linear. Our LR model relies on the reduced-form model: y_i = π_i + ε_i ,i =1,…,n, where π_i is the expected value of Y given (X_1 = x_i1, X_2 = x_i2, … ,X_p = x_ip) (Aguilera et al., 2006). In this case, Y is the probability of survival as a function of a set of available information about the ventures. The analysis of the effects of the dimensions of IC over success as they are operationalized requires a technique that adequately manages the probabilities of attaining a successful performance. Logistic regression (LR) is the appropriate technique when dealing with the relationship between a dichotomous outcome and a set of explanatory variables.

When an LR model estimates a binary response outcome, we assume that its logit transformation has a linear relationship with the predictor variables. Thereby the relationship between the response variable and its covariates is interpreted through the odds ratio from the parameters of the models. Measured in odds ratio (OR), if the parameter in the regression is positive, the OR > 1, and if it is negative OR < 1, indicating the effect of the IV on the chances of survival. The logistic regression model can be written as in Equation 1:

Log( π( X − 1, X − 2,…, X − k )/( 1−π( X − 1, X − 2,…, X − k ) ) )= β − 0+ β − 1 X − 1+ β − 2 X − 2+…〚 +β〛 ) − k X − k (1)

The binary response variable Y is either 0 or 1, and E(Y_1 |Χ_1i, Χ_21, …, Χ_k1). Then π(Χ_1, Χ_2, …, Χ_k) is interpreted as P(Y=1) for a given combination of values of the predictor variables Χ_1, Χ_2, …, Χ_k. We express the model as: 〖Y=π (Χ〗_1, Χ_2, …, Χ_k)+ϵ where ϵ, could only assume two values depending on whether Y is equal to zero or one. The left-hand side of Equation 1 is the log odds ratio, that is, the logarithm of the odds that Y will equal 1, for a given combination of the predictor variables. Our estimation uses the maximum likelihood (ML) method.

To validate our hypothesis, the following logistic regression model (LR) is proposed.

Pr(Survival)b1F2SHC+b2F3SC+b3F4STC+b4F5MarketRights+b6F1CountryContext+b7FemalePresence +b8HasTargetMargin+b9HasEquity+b10HasDebt+b11i.applicationyear +b12participated#programimpactarea+b13acceptedinitially +b14programwithcurriculumstructure+const (2)

The model’s results, expressed in terms of odds ratios (exp^b), are as follows:

The estimated odds ratios exhibit significance at the 1% level, except for the accelerator program-related variable ‘accepted initially’, which demonstrated significance at the 5% level. Furthermore, the variables ‘year 2014’, ‘female dominance in the founding team’, and the interaction between the variables ‘participated#’ program with ‘impact area’ [1 1] and [0 1] were significant at the 10% level.

Pr(Survival)=.95F2SHC+1.08F3SC+1.09F4STC+1−17F5MarketRights+.85F1CountryContext+1.11FemalePresence +1.25HasTargetMargin+1.09HasEquity+1.77HasDebt+1.37.applicationyear(2014) +1.69applicationyear(2015)+1.10applicationyear(2016)+1.70applicationyear(2017) +1.77applicationyear(2018)+2.23applicationyear(2019)+1.29(1.participated1.programimpactarea +1.31acceptedinitially+1.25programwithcurriculumstructure+const (3)

To identify specification errors in the logistic regression (LR) model, we executed a link test, which was not deemed significant at the 5% level. The Hosmer-Lemeshow goodness-of-fit test validated the appropriateness of the LR model for explaining ‘survival,’ with a chi-square statistic of 13.10 (df = 8) and a probability exceeding the chi-square value of 0.11. The McFadden R-squared value for the LR model was determined to be 0.054. The potential issue of multicollinearity in the LR model was investigated and subsequently dismissed; the individual variance inflation factors (VIF) were significantly below the accepted threshold of 10, and the mean VIF was observed to be 2.17.

By employing a logistic regression (LR) model, we were able to measure the influence of various independent variables, including control factors, on the survival probabilities of startups in the EDP sample. This also resulted in a predictive model that can be used to determine the likelihood of success for a startup during its pre-performance phase. The model correctly classified 70.1% of the companies that survived. It had a sensitivity of 16.07%, indicating its ability to correctly identify successful startups, and a specificity of 94.68%, reflecting its ability to correctly identify unsuccessful startups. The probability cutoff point was approximately 0.28, and the area under the receiver operating characteristic (ROC) curve, which incorporates all thresholds, was 66%.

By implementing these practical strategies, startup founders can better navigate the challenges of the initial stages and improve their chances of long-term success. Our study underscores the importance of early recognition and investment in key factors such as intellectual capital (IC), financial planning, and participation in accelerator programs for startup founders. By prioritizing the development of IC assets, including the knowledge and skills of the founding team, startups can build a solid foundation for navigating the early stages. Securing diverse funding sources and establishing clear financial goals from the outset can further mitigate financial constraints and improve survival probabilities.

Engaging in well-structured accelerator programs offers valuable resources, mentorship, and networking opportunities, enhancing the startup’s resilience and performance. Our research contributes to this understanding by demonstrating that these elements significantly influence startup survival, particularly in diverse and globally oriented contexts. By recognizing and leveraging these critical factors early on, founders can better position their ventures for long-term success and growth, adapting to challenges and capitalizing on opportunities as they arise.

For policymakers and accelerator program developers, our findings provide valuable insights for refining selection processes and program development. Leveraging data from the EDP and the predictive strength of our logistic regression (LR) model, we offer preliminary guidance for creating more effective mentoring and support initiatives. The clear distinction between the values of specificity and sensitivity in our LR model indicates that survival and failure probabilities are independent entities, challenging traditional assumptions and uncovering new research opportunities. Further investigation into these factors can enhance the efficacy of support initiatives, benefiting the startup ecosystem by fostering more resilient and successful ventures.