Firm Survival: An Empirical Study Concerning Insurance Agencies

This paper analyzes the relevance of firm size, region, sales network, organizational structure, and auditing on the survival likelihood of a specific firm typology, the insurance agency. By applying a configurational comparative method, namely fuzzy-set qualitative analysis (fsQCA), to a unique dataset of 52 insurance agents, representing 52 active Italian insurance agencies, this study demonstrates that, when combined with other variables, organizational structure provides sufficient conditions for insurance agency survival. The different relevance of some specific management areas according to the insurance agency location is highlighted.


Introduction
In Italy, the insurance sector is characterized by the presence of thousands of insurance agencies throughout the country from the north to the south.Each insurance agency is an independent juridical entity and like any commercial enterprise, each of them has to face the problem of increasing competition, developing new distribution channels, and changes in consumer behavior.In a word, even insurance agencies have to face the survival problem.
The literature on firm survival is very extensive and considers different aspects and countries (e.g.Box 2008;Agarwal and Gort 2002;Klepper 2002;Tsoukas 2011;Chung et al. 2013;Bontemps et al. 2012;Paeleman et al. 2015;Guariglia et al. 2016;Huggins et al. 2017).However, the issue of survival in the insurance agency sector is absent from the literature.This paper intends to open a new field of research and fills the gap in the literature.The paper's originality is based on three main aspects: i) the insurance sector and in particular insurance agencies, completely neglected in the literature; ii) the factors considered in the analysis and iii) the unique dataset.Specifically, the factors examined for insurance agencies' survival concern important management areas for commercial companies, such as the sales network, internal auditing and organizational structure.Other factors such as firm size and region are also considered.
The database used in the analysis is a unique dataset.In 2011, 52 insurance agents, representing 52 active Italian insurance agencies, were asked to complete a questionnaire concerning the importance they attribute to specific management areas.In 2018 we checked how many agencies were still active and we examined the factors which had determined their survival.
In order to analyze this dataset and cope with the small sample size problem, we refer to the configurational comparative method devised by Ragin (1987) namely the fuzzy set qualitative comparative analysis (fsQCA), which has gain relevance over the last years in business and management research.fsQCA is a data analysis technique for determining which logical conclusions a dataset supports, combining detailed within-case analysis and formalized cross-case comparison.With respect to other traditional probability-based statistical techniques, fsQCA does not require large samples (limiting the study to a few cases or case studies), while still guaranteeing the general applicability of its conclusions or implications to a larger population.In addition, while traditional statistical techniques are good at drawing out the net effect of single variables, QCA is able to detect different conditions that lead to the same outcome.management areas and company survival, this section contains a brief review of the literature considered relevant to our analysis.A number of studies have empirically evaluated the factors that influence the probability of firms' survival in the market.At firm level, these factors have traditionally been size and age of the firm, both increasing survival probability (Evans 1987;Dunne et al. 1989;Dunne and Hughes 1994;Cefis andMarsili 2006 andFritsch et al. 2006).New firms' survival is also a topic considered in the literature.Many empirical studies at firm level have found that survival probability increases with firm size (Mata and Portugal 1994;Geroski 1995;Audretsch et al. 2000;Agarwal and Audretsch 2001).Kalleberg and Leicht (1991) and Borghesi et al. (2007) analyze the relation between firms' survival and organizational structure.The former consider 441 companies in the computer sales and software, food and drink, and health indutries in South Central Indiana; the second consider a sample of 67,000 firms over the period 1981 to 2003, taking into account firm age, agency cost, survival and firm organizational decisions over time.Shane (1996), Tan and Peng (2003) and Wischnevsky (2004) highlight the fact that organizational structure is necessary to guarantee smooth company management, and rapidity in capturing market changes and coping with moments of crisis.Thus, organizational structure may influence firms' survival.
The relationship between sales networks and firms' survival is also considered in the literature.Winter et al. (2003) highlight the importance of sales networks by considering a sample of Korean furniture manufactures threatened by the financial downturn of 1997.Sepulveda and Gabrielsson (2013) develop a theoretical framework that links resource development and enterpreneurial orientation to network content, structure, centrality and management.In the business-to-business (B2B) context, Jones et al. (2013) find that firms use marketing and sales networks, among other factors, to create value for the firm.Firms' survival may therefore depend on the ability of the sales networks and their involvement and coordination (Reynolds 1987;Lee et al. 2012).Walz (1997) demonstrates that internal auditors contribute to firms' survival through reduction in the cost of the internal function and through the introduction of audit recommendations that increase firm value.Gaeremynck and Willekens (2003) demonstrate the relationship between audit-report type and subsequent business termination for private companies in a non-litigious environment.According to the authors an endogenous relationship exists between bankruptcy and audit-report type.Jain and Martin Jr (2005) examine the relationship between audit quality and post-IPO survival.Karagiorgos et al. (2009) and Karagiorgos et al. (2011) test a hypothesis concerning the role of internal auditing in business success.According to these authors, internal auditing is vital for business survival and success.
Bearing in mind these findings reported in the literature, this study aims to test the following hypotheses: Hypothesis 1 -Insurance agency survival increases in accordance with its size.
Hypothesis 2 -Insurance agency survival is linked to its organizational structure.
Hypothesis 3 -Insurance agency survival is linked to its sales network.
Hypothesis 4 -Insurance agency survival is linked to its internal auditing.

Data
The database used in the analysis is a unique dataset.During 2011, 52 insurance agents, representing 52 active Italian insurance agencies belonging to a primary Italian national insurance agencies' association, were asked to complete a questionnaire regarding three specific management areas (sales network, auditing, and organizational structure).The aim of the questionnaire (see Appendix) was to test the level of awareness of the relevance of these management areas in business continuity.The score attributed to each answer was 1 for positive, 0 for negative and -1 for "I Don't know".39 out of the 52 Italian insurance agencies were still active in 2018, whereas 13 were no longer going concerns.In terms of size, all the firms in the sample were either micro or small (in accordance with the European Commission recommendation, fewer than 10 and 50 employees respectively), because of the characteristics of this business in Italy.

Methodology
Configurational analysis is a multi-faceted approach, based mainly on set theory and Boolean algebra, which addresses configurations as varying case types in order to work out what combination of characteristics may be necessary or sufficient to produce a certain outcome.Comparative qualitative analysis (QCA) makes it possible to bridge the gap between qualitative (case study oriented) and quantitative (variable oriented) approaches, allowing the analysis of a small number of cases.This study uses a specific type of QCA, namely fuzzy set QCA (fsQCA), which allows for any value in the interval [0,1] indicating partial (non)membership in a given set.
QCA follows some typical main steps.The first step is truth table construction, i.e. the construction of a table of all the logically possible combinations of the considered conditions (variables).In the second step, the number of rows in the truth table is reduced, considering only those combinations with a minimum consistency of 0.75 (Ragin, 2006).It is possible to highlight cases that lead to the outcome and drop cases where the outcome is not present.In the third step of analysis, the information contained in the truth table is restated in terms of a parsimonious and encompassing truth-functional proposition set.
The parsimonious solution takes into account all simplifying assumptions, both involving easy or difficult counterfactuals.At this stage of analysis the QCA results are written in a solution formula that logically summarizes the information contained in the truth table.The outcome and conditions are listed in Table 1.The outcome (i.e.survival) is a dichotomous variable distinguishing active firms from those that no longer exist.The conditions are firm size, region and the importance given to management areas (i.e. sales network, organizational structure and auditing).The measurement of firm size is as a fuzzyset condition depending on the number of employees: microenterprises (close to 0) and small firms (close to 1).
Region is a dichotomous condition, which establishes whether the firm is located in the Center or South of Italy (coding a firm as 1, fully in this set) or in the North (coding a firm as 0, outside the set).
Sales, org and audit are fuzzyset conditions specifying the relevance given to each management area -sales network, organizational structure and auditing -(coding a firm as 1, fully in this set and 0, outside the set).

Results and Discussion
This section presents results from the analysis of the conditions that lead firms to the outcome (i.e.survival).The model for analysis is: The first step in fsQCA is the analysis of necessary conditions, which is a separate procedure that looks at which individual factors may be necessary or mostly necessary for the outcome to occur.This entails that the membership score on the outcome be consistently lower than the membership score of the causal factor under consideration.
Table 2 shows the causal conditions relevant to the survival of the 52 agencies from 2011 to 2018, where the outcome of interest is the degree of membership in the set of agencies which survived in this period (survival).Consistency and coverage are the two key parameters for assessing the fit of results (Ragin, 2006).Consistency exceeds 0.75 only for the organization and size variables (see Table 2), meaning that only in these cases is there considerable consistency; coverage, which is a measure of how trivial or relevant a necessary condition is for the outcome, is always higher than 0.7.One can argue that only organization and size are on their own able to assure survival, while the other conditions are not.
Table 3 presents the results of the sufficiency analysis, which indicates a set of sufficient relations between insurance agency survival and a certain subset of conditions.Analysis consistency is 0.850, which indicates a sufficient relation between agency survival and these conditions.Note.Note that symbol ~ represents the "negation" of the characteristic.
Table 3 indicates three solutions resulting from the analysis.The symbol ~ represents the "negation" of the characteristic (e.g."~sales" is equal to "not sales"), while the symbol * represents the logical operator AND.
Configurations connected by * are sufficient conditions to lead to survival.Each of the three configurations reaches the consistency threshold of 0.80 suggested by Ragin (2008).
The first configuration is sales*audit*organization*region, which shows that a combination of each management area considered in the survey (sales network, auditing, and organizational structure) with the insurance agency location in the center or south of Italy is a sufficient condition for its survival.The second configuration is sales*organization*~size*region, which implies that sales network and organizational structure in a microenterprise located in the center or south of Italy represent a sufficient condition for insurance agency survival.Finally, ~sales*~audit*organization*~size*~region implies that micro-size insurance agencies located in the north of Italy have survived when mainly focused on organizational structure rather than on sales network and auditing area.

Conclusions
This study examines firm size, region, and some management areas (sales network, organizational structure, and auditing) which, according to the literature, may influence firm survival in general and therefore that of insurance agencies as well.The analysis employs fsQCA (Ragin, 1987) to identify combinations of conditions (causal configurations) that lead to insurance agency survival in Italy.fsQCA is based on the analysis of set relations, not correlations, and is a way to analyze problems where there are few observations and where the outcome depends on the configurations of some variables.This analysis technique has gain increasing prominence in social science over the last years.
The results obtained basing on a unique dataset of 52 Italian insurance agents highlight the fact that organizational structure in combination with other variables provides sufficient condition for insurance agencies' survival.In particular, a key implication of the analysis is that for microenterprises located in the north of Italy, the entrepreneurship policies of the chief insurance agent should emphasize organizational structure in order to increase the likelihood of survival.For insurance agencies located in the center or south of Italy, other management areas, such as sales network and auditing, play an important role together with organizational structure, in insurance agencies' survival irrespective of the firm size while, in the same region, for micro insurance agencies, organizational structure combined with sales network are relevant for their survival.Thus, organizational structure appears as more relavant than other management areas examined in this research.
This study has some limitations, which can constitute opportunities for future research.For instance, it could be interesting to examine the entrepreneurship policies implemented so as to improve a single management area, bearing in mind the geographical location of insurance agencies.Moreover, future research could investigate whether the distinction between North and Center/South of Italy is a distinctive condition for micro firms operating in other sectors in Italy.

Table 1 .
Outcome and conditions: description and codifications