




































 

 

 
26 

† Corresponding author 
© 2016 Conscientia Beam. All Rights Reserved. 

 

THE INDIRECT IMPACT OF OVERCONFIDENCE ON THE PERFORMANCE OF TUNISIAN 
FIRMS THROUGH THEIR FINANCING STRUCTURE 

 

Manel Dahmani1† --- Ghazi Zouari2   
1PHD, Student in Financial and Accounting Methods: LARTIGE, FSEG, University of Sfax, Tunisia 
2Associate Professor in Financial and Accounting Methods: LARTIGE, FSEG, University of Sfax & FARGO, IAE Dijon, France 

 

ABSTRACT 

In the current framework of behavioral corporate finance, this article studies the relationship between the managers’ 

overconfidence and firm performance through the financing structure in the Tunisian context. Our model seeks to identify if the 

financing structure as a mediating variable between the performance and overconfidence. The empirical study is based on a 

sample of 56 firm managers for the year 2014. The results of the conducted regressions confirm the existence of a mediating 

effect of the financing structure on the relationship between overconfidence and the performance of Tunisian firms. 

Keywords: Overconfidence, Financing structure, Firm’s performance. 

 

Received: 20 September 2016/ Revised: 1 October 2016/ Accepted: 5 October 2016/ Published: 7 October 2016 

 

Contribution/ Originality 

This study is one of the few studies which have investigated the indirect impact of overconfidence on the 

performance of Tunisian firms through their financing structure; it is based on a new methodology that uses the 

theory of Baron and Kenny (1986) by utilising mediating variables to assess the relationship between the firm’s 

overconfidence and performance.  

 

1. INTRODUCTION 

The literature review conducted from the perspective of behavioral finance shows that the majority of the work 

examine the direct relationship between leaders’ overconfidence and the performance of the firm by ignoring the 

indirect relationship between these two variables (Bernardo and Welch, 2001; Ben-David et al., 2006; Keiber, 2006). 

For this reason, the research wishes to establish and develop the field of behavioral finance to business psychology 

inspired by adopting a more realistic nature than generally the one used in economics (market finance and 

corporate). So, this is an exploratory study carried out primarily through a literature review. This development is 

similar to a complementary or even, in some respects, a revolution of paradigms. Indeed, some authors (Hermalin 

and Isen, 2000; Heaton, 2002) studying the behavioral approach in the context of rationality (mainly substantive) 

and other (Shiller, 1997) analyze the behavior outside this framework. 

In our study, we will focus on explaining how work performance is not only dependent on the growth and 

skills directly related to the activity of labour but also on the conduct which constitutes an essential element in 

making financial decisions in the firm. To do this, we will undertake a questionnaire with directors of listed 

Tunisian firms. The purpose of this research is to explain the advantage of the behavioral approach apprehended by 

Financial Risk and Management Reviews 
2016 Vol. 2, No. 1, pp. 26-42 
ISSN(e): 2411-6408 
ISSN(p): 2412-3404 
DOI:  10.18488/journal.89/2016.2.1/89.1.26.42 
© 2016 Conscientia Beam. All Rights Reserved. 

 

http://crossmark.crossref.org/dialog/?doi=10.18488/journal.89/2016.2.1/89.1.26.42


Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
27 

© 2016 Conscientia Beam. All Rights Reserved. 

the leaders’ overconfidence as an alternative explanation for the performance of the firm through the funding 

structure. 

We fixed a reality, for a hypothetical-deductive approach, in order to detect the development of a research 

model based on a set of assumptions through the combination of theory on behavioral biases and exploratory 

qualitative study conducted on leaders. Our position is, in essence, positivist (Martinet, 1990; Wacheux, 1996) since 

it is based on a deductive approach to test questionable research hypotheses (Igalens and Roussel, 1998). Thus, we 

are trying in this research to describe the relationships between variables and test them to verify their loyalty when 

put to the test of reality in our sample of leaders. 

Our research shows, then the following two sections: the first presents the theoretical model which posits that 

overconfidence could influence performance. In the midst of this direct relationship, interpose variables are related 

to the financing structure (equity, bank debt, bond debt and external equity). Besides, being influenced by leaders’ 

overconfidence, these variables influence, in turn, performance. The second empirical section is designed to test the 

potential effect of the financing structure as a mediating variable between overconfidence and performance of the 

firm. 

 

2. LITERATURE REVIEW AND HYPOTHESES 

Overconfidence is the illusion of knowledge that makes individuals feel the accuracy of their information and 

capabilities as a selection (stock-picking). It is one of the most documented behavioral means (Daniel and Titman, 

1999). In their synthesis on the foundations of behavioral finance, De Bondt and Thaler (1995) state that 

"overconfidence is one of the most relevant features in the psychology of judgment." These authors define an 

individual with excessive confidence in their own skills, abilities or knowledge. 

This bias, associated with that of optimism, is a central aspect of the current literature on behavioral finance to 

business as highlighted by Baker et al. (2004): "Optimism is an unrealistic overestimation of future events, not 

related to personal skills, while overconfidence reflects an overestimation of the past" (Véronique, 2007). Explicitly, 

in the modeling of these two means, optimism is captured as an average error (overestimation) and overconfidence 

as an underestimation of the variance, but the two terms are often used interchangeably (Fairchild, 2005). In 

addition, these biases are often paralleled (Heaton, 2002; Gervais et al., 2003) and combined, especially in the illusion 

of control when the individual thinks he can control, thanks to his abilities, purely random events. Thus, according 

to Daniel et al. (2001) overconfidence is stronger in activities that involve valuation difficulties, and this is why the 

feedback on the quality of the evaluation is ambiguous. Roll (1986) was the first to introduce the idea of optimistic 

and confident leader in finance. The principle defines the leaders of the acquiring companies as overestimating the 

gains resulting from these transactions and thus on average paying high prices for shareholders of target 

companies, thus leading to the phenomenon of the winner's curse. According to Weinstein (1980) excessive 

optimism is connected to trust; therefore, these two biases are distinct. Considering the definition of Hackbarth 

(2004) optimistic managers overestimate the earnings growth rate, while the on-confident managers underestimate 

the risks of gains. Hackbarth (2004) compares his model in the decisions of a biased and unbiased leader, he finds 

that optimism produces an overestimation of earnings growth rates, while overconfidence reduces the variance from 

expected results. Both effects lead to underestimate the probability of the investment project bankruptcy (Bernardo 

and Welch, 2001) consider that the presence of overconfident leaders play a positive role within the group and 

transmit private information about the area other leaders likely to follow. Indeed, the impact of overconfidence on 

financial decisions and business performance has been studied, both on a theoretical and empirical, since the early 

two miles. These early contributions confirm the effect of the bias, which is not necessarily negative, however. 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
28 

© 2016 Conscientia Beam. All Rights Reserved. 

Yet, few studies have tried to link the three dimensions in one perspective, i.e. trust, financing structure and 

performance. Indeed, this relationship is very complex; it must be specified by including the mediating concept of 

the financing structure. This assumes that the direct relationship between overconfidence and performance is rather 

an indirect relationship through the influence of the financing structure in enterprises. Therefore, the funding 

structure acts as a mediating variable in the relationship between leaders’ overconfidence and business performance. 

Considering the scarcity of work, our research objective is to answer the question: how leader’s overconfidence has 

an indirect effect on the performance of the firm through the firm's financing structure. Several theoretical 

developments and empirical tests have been developed to study the impact of the excess through confidence in the 

major financial decisions, namely the investment decision and the financing (Heaton, 2002; Hackbarth, 2004; 

Malmendier et al., 2005). Several studies show that the direct relationship between the leaders ‘overconfidence and 

the firm performance is unstable. It can be positive, negative or non significant (Bernardo and Welch, 2001; Ben-

David et al., 2006; Keiber, 2006). 

Therefore, it is recommended to introduce a mediating variable, the financing structure (apprehended by the 

various modes of financing, namely: self-financing, bank loans, bond debt and external equity) that allows to 

stabilize the impact of executives’ confidence on business performance. We start with self-financing that plays a 

very important role, it increases the profitability of investments and business performance. 

Indeed, leaders believe that being confident about the decisions they make will help to pursue conflicting 

objectives, namely: maximizing the fundamental value and minimizing capital costs. For this reason, confident 

leaders prefer internal financing (self-financing). However, self-financing is a free, renewable resource for leaders as 

for the company, it allows some financial independence, good control of financial charges, and gives the company 

freedom of action because it is independent of its creditors. Then, it is of lower cost. Besides, confident leaders think 

that self-financing would lead neither to a waste of funds nor to conservatism, but rather it would have a leverage 

accelerator to research new ideas. All these arguments lead us to say that cash is more interesting for leaders and 

for good performance. We then deduce the following hypothesis: 

H1: The use of self –financing positively mediates the relationship between overconfidence managers and 

performance of the firm. 

Hackbarth (2004) develops a model of capital structure based on the trade-off theory to study the impact of bias 

optimism and managerial confidence on the financial policy and the company's value. He believes in optimism by 

overestimating earnings growth rates and trust by an underestimation of the variance results anticipated by 

management. The author examines leaders’ confidence about choosing a high debt ratio and emits mostly debt 

relative to rational leaders, which significantly impacts the capital structure. 

According to Jensen and Meckling (1976) companies ‘debt  is considered as one of the leaders’ control tool that  

reduces discretionary cash flow. 

According to several studies and research, debt plays a very important role in explaining the performance of 

businesses, namely the study of Modigliani and Miller (1958). 

Furthermore, the confident leaders’ use of debt is a way to support business operations, improve productivity 

and stimulate economic growth on the one hand, and increases their performance on the other. Moreover, confident 

leaders prefer this type of funding, that pushes for modernization and innovation, and streamline production 

because debt compresses manufacturing costs. It therefore appears that the debt is required for the performance of 

the company. It is considered as a sacrifice of economic resources today in the hope of obtaining higher revenues in 

the future. From this point of view, should be favored by innovative firms with high quality. 

In addition, debt (bank or bond) opens more prospects other than financing alternatives. Indeed, a company 

issuing a bond sells its signature, hence its "serious" and "reputation" on the market, so that when the company 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
29 

© 2016 Conscientia Beam. All Rights Reserved. 

launches bidding to acquire the equipment they need as part of the project, suppliers know that it has already funds 

This has considerable influence on performance. 

We then deduce the following assumptions: 

H2: The recourse to bank debt positively mediates the relationship between overconfidence managers and 

performance of the firm.. 

H3: The use of bond debt positively mediates the relationship between overconfidence managers and 

performance of the firm. 

The use of external capital is also an important operation for the company; it allows the structures of the company 

to consolidate its business. Indeed, the company can better manage its working capital or even increase it. In 

addition, it will gain in economic credibility, to the extent that it may strengthen its capital base. Finally, this 

operation evokes stability in his favor of shareholders, associates and creditors. The confident leaders feel that this 

transaction as it increases the capital is important in the lives of enterprises to strengthen capital and ensure 

balanced economic development project. 

Moreover, the capital increase is an acquisition mode of the company but also serves to strengthen its balance 

sheet when the company is financially weakened. Thus, it improves the ratio of financial independence and leaves 

room for new debt capacity. This method of financing has a major advantage over the performance of the firm, since 

it is free which automatically leads to an increase in equity and therefore stable resources and turnover of the 

company funds, resulting in a good financial health of the business. However, to our knowledge, no studies have 

been conducted yet to show the intermediary role of the financing structure in the relationship of trust on 

leadership and performance of the firm. We then deduce the following hypothesis: 

H4: The use of external funds positively mediates the relationship between overconfidence managers and 

performance of the firm. 

 

3. EMPIRICAL ANALYSIS 

This section is designed to test the mediating effect of the financing structure in the relations between 

overconfidence managers, on one hand, and the performance of the firm, on the other. First, we will present our 

sample, the dependent and independent variables and the multivariate analysis method (hierarchical). The 

presentation and interpretation of the results of this study will be a second subsection. 

 

3.1. Presentation of Data Variables and Measures 

For our research, a study data from the annual reports of 56 listed companies in the Tunisian stock exchange in 

Tunis for the year 2014 and a questionnaire (conducted for my master thesis) sent to their leaders are conducted. 

These companies belong to the industrial, service, media and travel sectors. 

Our study includes 3 main variables: the dependent variable: the performance of the firm, an independent 

variable: overconfidence and a mediating variable: the funding structure (self-financing, bank debt, bond debt and 

external funds). 

 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
30 

© 2016 Conscientia Beam. All Rights Reserved. 

 
 Figure-1. Conceptual model of the mediating effect of the financing structure in the relationship of overconfidence with the performance of the 
firm 
Source:  Manel (2015). 

 
 

3.1.1. Dependent Variable: The Performance of the Firm 

Performance is the ending result of all efforts of a company the company’s efforts. Indeed, they are doing the 

right thing in the right way, at the right time and cost to produce good results that meet the needs and expectations 

of customers, giving them satisfaction and achieving the goals set by the organization. 

We can say that a successful company must be both effective and efficient1. It is effective when it accomplishes 

the objectives it has set to minimize the resources implemented. It is also effective when it reaches the goals it has 

set to achieve the welfare of its stakeholders ‘parts. Furthermore, performance is measured with qualitative or 

quantitative indicators of outcome. To measure effectiveness, we will be using a criterion that expresses a 

relationship between the result and the goal. To measure efficiency, we will be using a criterion that expresses a 

relationship between the result obtained and the means used. 

To evaluate the performance of a company, it is necessary to set measures at all levels: economic and financial. 

Like previous studies, we define the economic performance of the firm by the Return On Assets "ROA" operating 

income before depreciation and R & D / total assets (Zouari and Zouari-Hadiji, 2014a; 2014b)2 and financial 

performance by the Market to Book "MTB" = market capitalization / book value of equity (Zouari and Zouari-

Hadiji, 2014a; 2014b). 

 

3.1.2. Independent Variable: Overconfidence 

The identification of relevant and operational measure of overconfidence is not an easy exercise to handle in the 

sense that this measure must be earlier and exogenous to the decision in order to determine the direction of 

causality (Véronique, 2007). According Malmendier and Tate (2005b) two different approaches to measure this way: 

the first is an approach called "internal" bias, it is revealed preference by the leader himself through his answers to a 

                                                 
1 According Charreaux (1999). efficiency "refers to the performance of a collective entity appreciated by the welfare provided to its stakeholders, that is  to say by all the 

individuals whose the utility is affected by the decisions of the entity. However, efficiency is a different concept that refers to the means used by the players to achieve 

their goals. a negotiation is deemed effective if the actors have used their best resources to achieve their goals. The aggregation of the effective behavior of individuals 

does not necessarily lead to produce an efficient organizational behavior. In some cases, we also will bead efficiency "informational" to describe the ability of markets 

to reflect the information in prices. " 

2 This measure of performance accounting has the advantage of eliminating the effect of accounting choices related to the treatment of R & D expenses in the financial 

statements largely prone to opportunism of the leaders. 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
31 

© 2016 Conscientia Beam. All Rights Reserved. 

questionnaire or the exercise of such stock options. The second so-called "external" approach is to analyze the 

perception and description of the leader by third parties such as the press. Russo and Schoemaker (1992) propose 

specific questions and ask the leaders to rank the relevance of questions about their work. 

The introduction of different questions depending on the questioned individual aims at improving inter-

individuals comparison, and therefore, detect overconfidence. This requires an individual pre-calibration of the 

relevance and difficulty of the questions. Ben-David et al. (2006) used a simpler methodology, the same question is 

asked to all leaders (expected evolution of the market index, but general issue which can assume relevant to CFOs 

of listed companies that constitute the sample) every quarter for four years. Overconfidence is then measured by the 

difference between volatility perceived ex ante and the actual volatility of the index calculated ex post. 

In our study, we chose the questionnaire as a tool for measuring overconfidence. The questionnaires are 

questions based on the investor profile developed by the research unit on savings (PES) of the company Fern Hill. 

Each item is encoded by a Likert scale of 5 points (from "strongly disagree" = 1 to "strongly agree" = 5). 

 

3.1.3. Mediating Variable: The Financing Structure 

In our study, the mediating variables related to the financing structure are self-financing, bank debt, bond debt 

and external funds. To measure these variables, we use financial statements of listed Tunisian firms for 2014. 

 

3.1.4. Control Variables 

For more results reliability, we have introduced control variables that have a significant effect on performance. 

The multiple linear regression models used in this empirical study retain the size of the company and its industry. 

The firm size is measured by the natural logarithm of total company assets. This measure has been used in several 

studies such as (Nekhili et al., 2012; Zouari and Zouari-Hadiji, 2013;2014a;2014b). The industry is a dummy variable 

taking the value 1, if firms belong to a high-tech sector and 0 if not. This measure has been used by several 

researchers such as Zouari and Zouari-Hadiji (2013;2014a;2014b) and Zouari and Zouari-Hadiji (2010). 

The explanatory and control variables influence the performance of the company and check its multidimensionality. 

They are as distinct from each other and present as shown in Appendix Table 9, a low correlation and / or non-

significant between them. 

 

3.2. The Modeling Assumptions 

The empirical study of this research is based on the use of hierarchical regression models3 to test research 

hypotheses. To verify all assumptions, we need to test the existence of a mediator. The verification of this effect is 

achieved by the construction of three models. According to Baron and Kenny (1986), four conditions to are check a 

complete mediating effect of M in the framework of the X-Y relationship: 

 Condition (1): the variable X must have a significant impact on variable Y. 

 Condition (2): the variable X must have a significant impact on Mr. 

                                                 
3 In this work, the treatment of mediating variables should follow the approach devised by Baron and Kenny (1986). This framework, which aims at testing the 

mediating effect, is implemented via a multiple-hierarchical regression. This analysis consists in assessing the total effect (cumulative) of the explanatory variables on a 

certain criterion. The method can be performed on the basis of several steps. Firstly, it undertakes to test the predictor effect (independent variable) firstly on he criterion 

(dependent variable) and, secondly, on the mediator using partial and simple regressions. Then, the other relationship has to be tested (predictor and mediator on the 

criterion). In this case, a multiple-hierarchical regression has to be applied. It consists in gradually introducing certain independent variables into the regression-

equation: starting with the predictors and control variables (Step 1), then the mediating variable (Step 2). On reaching an increase in the adjusted R² after inserting the 

mediator, one is able to assume the mediator effect on the relationship between the predictor and the criterion Zouari and Zouari-Hadiji (2014a; 2014b).  



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
32 

© 2016 Conscientia Beam. All Rights Reserved. 

 Condition (3): the supposed mediator variable M has significantly influence the variable Y, where the influence 

of variable X on Y is controlled. 

 Condition (4): the significant influence of the variable X on Y should disappear when the effect of M on Y is 

controlled statistically. 

Econometrically, we estimate models 1-3 testing the indirect relationship between the trust and on firm 

performance by the effect of the financing structure (self-financing). These models allow the validation of the 

hypothesis (H1) 

(1) i 

(2)      

(3)     

 

Equations 4-6 test the indirect relationship between the firm’s overconfidence and  its performance by the effect 

of the financing structure (bank debt). These equations are used to validate the hypothesis (H2) 

(4)    

(5)    

(6)  

Equations 7-9 test the indirect relationship between firm’s overconfidence and  its performance by the effect of 

the financing structure (bond debt). These equations are used to validate the hypothesis (H3) 

(7)    

(8)   

(9)   

 

Equations 10-12 test the indirect relationship between firm’s overconfidence and  its performance by the effect of 

the financing structure (external funds). These equations are used to validate the hypothesis (H4) 

(10)     

(11)     

(12)   

With, 

- PERF i: Variable measured by ROA ratios and MTB of firm i, 

- SCONF i: Score overconfidence calculated through a questionnaire of firm i,  

- AUTOF i: This is the result of firm i + amortization, 

- FONDS EXT i: External capital firm i, the capital increase, 

- DETTE BANC i: Bank debt of the company i, 

- DETTE OBLIGA i: Bond debt of the company i, 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
33 

© 2016 Conscientia Beam. All Rights Reserved. 

- TAIL i: The natural logarithm of total assets of firm i, 

- SECT i: A binary variable which takes the value 1 if the firm i belongs to a high-tech industry sector, and 0 

inversely,  

 

- Β0, β1, β2, β3, β4, β5, β6: Parameter estimate, 

-   i: standard error.  

 

3.3. Presentation and Interpretation of Results 

This section aims to present the results of the testing hypotheses that relate to overconfidence to the company's 

performance (ROA, MTB) through the financing structure (AUTOF, DETTE BANC, DEETE OBLIGA, FONDS 

EXT). In order to submit our assumptions, we estimated models of separate regressions for each of the four stages 

of the process of Baron and Kenny (1986). 

Model 1 (reduced model) contains the independent variable and the control variables in predicting the 

performance of the firm. Model 2 on the other side (reduced model), seeks to explain the variation of the mediator 

variable (financing structure) in a third step by the independent variable (SCONF) and control variables. Model 3 

(full model) includes all the variables: the independent variable (SCONF) and mediating variables (AUTOF, 

DETTE BANC, DEETE OBLIGA, FONDS EXT). The control variables (TAIL, SECT) seeking to explain the 

variable dependent, ie the performance of the firm. 

 

3.3.1. Interpreting the Results of the Indirect Relationship between the Overconfidence and Firm 

Performance through Self-Financing 

According to the results in Table 1, the first condition was met, as the Model 1 (which tests the relationship 

between the variable SCONF and ROA) has a low explanatory power (R2 adjusted = 0.041). The overall quality of 

the model is significantly acceptable (F = 0.163 at the 10% threshold), however, when performance is measured by 

MTB, the model in question has a very low explanatory power (R2 adjusted = 0.019), and test Fisher becomes 

insignificant (F = 0.269, p> 10%). The Student’s tests reveal that SCONF variable has a positive and significant 

impact when performance is measured by ROA ( = 0.236, t = 1.704, p <10%), while MTB is not significant ( = -

0093, t = -0661, p> 10%). 

The objective of the second step is to demonstrate the existence of a relationship between SCONF and 

AUTOF. Model 2 shows that overall quality is statistically significant at the 10% level and the SCONF variable is 

positively and significantly associated with AUTOF Tunisian companies ( = 0.214, t = 2.080, p <5%), and hence 

the second condition in the approach Baron and Kenny (1986) holds true. 

 

Table-1. Results of Hierarchical Regression of steps 1 and 2 (model 1 and 2) to Tunisian Company 

 
 
Variables 

Step 1 Model 1 Step 2 Model 2 

Firm’s performance Self-Financing 

ROA MTB 

 
t 

 
t 

 
t 

V. control TAIL -0.201 -1.501 n.s  0.042    0.314 n.s   0.587    5.927 *** 

SECT -0.059 -0.430 n.s  0.227 1.638 * - 0.230 - 2.267 n.s 

V. independent SCONF 0.236 1.704 * -0.093  - 0.661 n.s   0.214  2.080 ** 

R 2 ajusted 0.041  0.019 0.474 

F    0.163 *       0.269 n.s    0.000 * 

         *** Significant at 1%, ** significant at 5%, *: significant at 10%, n.s: not significant 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
34 

© 2016 Conscientia Beam. All Rights Reserved. 

Examining the results of Table 2 reveals a positive and significant relationship between self-financing and the  

two indicators of firm performance (ROA and MTB), from these results we can see that self-financing has a positive 

impact on the economic and market performance of the company (for ROA,  = 0.034, t = 0.045, p <5%; and for 

MTB: = 0.088, t = 0.098, p <10%). 

Model 3 (full model) is used to verify third-mediated condition of self-financing between SCONF and firm 

performance (ROA and MTB). The results of the hierarchical regression analysis indicate that the self-financing 

(mediating variable) remains important in explaining the dependent variable (the two forms of performance) after 

considering the predictor variable. The statistical coefficient of AUTOF variable has a positive and significant value 

to ROA ( = 0.002, t = 0.032, p <10%) and also with respect to MTB ( = 0.065, t = 0.078, p <10%). It appears 

from these results that the third condition is completely verified. 

It only remains to verify the last condition, that is to say the effect of the predictor SCONF on the dependent 

variable (ROA and MTB) which should not be significant once the possible mediator (AUTOF) is considered. 

The results in Table 2 show that the coefficients associated with the variable SCONF are statistically 

significant for the indicator (ROA) of the performance of the firm ( = 0.271, t = 1.872, p <10%) and non-

significant for MTB ( = 0.078, t = - 0.528). It follows that the mediation by self-financing is a partial mediation on 

confidence and performance of the firm. Through these results, we see that the SCONF variable has a positive effect 

when the firm performance is measured by MTB. From these results, the hypothesis 1 is confirmed (partial 

mediation) with Tunisian companies. 

 

Table-2. Results of hierarchical regression steps 3 and 4 (Model 3) to Tunisian companies 

 
 
Variables 

Step 3 Step 3 and 4 Model 3 

Firm’s performance Firm’s performance 

ROA MTB ROA MTB 

  
t 

 
T 

 
t 

 
t 

V. control  TAIL - 0.124 -0.697n.s 0.089 0.506 n.s -0.106 - 0.609 n.s 0.084   0.473 n.s 

 SECT - 0.140 -0.960n.s 0.223 1.556 n.s -0.096 - 0.668 n.s 0.211 1.439 * 

V. Independent  SCONF - - - -  0.271 1.872 * 0.078 -  0.528 n.s 

V. mediator   AUTOF  0.034 0.045 ** 0.088 0.098 *  0.002 0.032 * 0.065 0.078 * 
 

R 2 ajusted  0.053 0.051 0.098 0.095 

F      0.049**   0.095*   0.002*   0.056* 

Adjusted R² variation   0.057 0.076 

*** Significant at 1%, ** significant at 5%, *: significant at 10%, n.s: not significant 

 

From Table 2 (Model 3), we found that the control variables (TAIL, SECT) are not statistically significant for 

both indicators of firm performance (ROA, MTB), except the coefficient of SECT variable that is statistically 

significant for the MTB indicator ( = 0.211, t = 1.439, p <10%). Indeed, the industry has a positive impact on the 

market performance. However, the company size is not an important factor in analyzing the performance of the 

firm. 

According to Table 2, for both measures of performance, Model 3 (full model) has an interesting adjusted 

explanatory power. Thus, this comprehensive model, which takes into account the mediating effect of self-financing, 

also increases the percentage of explained variance from Model 1, in cases where performance is measured by ROA, 

adjusted R2 passes from 0.041 to 0.098. Similarly, when performance is measured by MTB, adjusted R2 passes from 

0.019 to 0.095 and the F statistic becomes significant in Model 3 at the 10% threshold than that in the Model 1 (not 

significant). The adjusted R2 of increase is linked to the consideration of the mediating effect of self-financing, so the 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
35 

© 2016 Conscientia Beam. All Rights Reserved. 

change in adjusted R2 for the two models associated with the addition of the mediating variable is significant (0.057 

and 0.076). 

This shows that this variable is a good predictor of the dependent variable, namely the performance of the firm. 

The mediator variable (AUTOF) allows us to better explain the causal relationship between overconfidence and the 

performance of the firm. 

 

3.3.2. Interpreting the Results of the Indirect Relationship between the Overconfidence and Firm 

Performance through Bank Debt 

Consistent with previous results, the first condition of the relationship between the variable  SCONF and  ROA 

was filled (see Table 3). 

The objective of the second step is to demonstrate the existence of a relationship between SCONF and bank 

debt (BANK DEBT). Model 2 shows that the overall quality is statistically significant at the 10% level and the 

SCONF variable is positively and significantly associated with BANK DEBT Tunisian companies ( = 0.045, t = 

0.098, p <10%), and the second condition of Baron and Kenny (1986) holds true. 

 

Table-3. Results of Hierarchical Regression of Steps 1 and 2 (model 1 and 2) to Tunisian companies 

Variables Step 1 Model 1 Step 2 Model 2 

Firm’s performance Bank debt 

ROA MTB 

 
t 

 
t 

 
t 

V. control TAIL -0.201 -1.501 n.s  0.042    0.314 n.s 0.465  4.670  *** 
SECT -0.059 -0.430 n.s  0.227 1.638 * 0.526  5.158  *** 

V. independent SCONF  0.236 1.704 * -0.093 -  0.661 n.s 0.045    0.098  * 
R 2 ajusted 0.041  0.019                  0.325 
F    0.163 *        0.269 n.s 0.08   

      *** Significant at 1%, ** significant at 5%, *: significant at 10%, n.s: not significant 

 

Examining the results of Table 4 shows a significant positive relationship between bank debt and both 

indicators of firm performance (ROA and MTB). From these results we see that bank debt has a positive impact on 

the economic and market performance of the company (for ROA,  = 0.065, t = 0.096, p <10%; and for MTB: = 

0.015, t = 0.045, p <5%). 

Model 3 (full model) is used to verify third-mediated condition in bank indebtedness between SCONF and firm 

performance (ROA and MTB). The results of the hierarchical regression analysis indicate that bank debt (mediating 

variable) remains important in explaining the dependent variable (the two forms of performance) after considering 

the predictor variable. The statistical coefficient of the variable "DEBT BANK" has a positive and significant value 

compared to the ROA ( = 0.028, t = 0.150, p <10%) as compared to MTB ( = 0.056, t = 0.086, p <10 %). It 

appears from these results that the third condition is completely verified. It only remains to verify the last 

condition, that is to say the effect of the predictor SCONF on the dependent variable (ROA and MTB) which should 

not be significant once the possible mediator (DEBT BANC) is considered. 

The results in Table 4 show that the coefficients associated with the variable SCONF are not statistically 

significant whatever the extent of the performance will be then it was statistically significant at the first step of the 

process of Baron and Kenny (1986) when compared to measuring the performance ROA. The regression coefficients 

of the variable SCONF have insignificant signs relative to ROA ( = 0.012, t = -0.158, p> 10%) and the MTB ( = 

- 0.068, t = -0.490, p> 10%). It follows that the mediation by bank debt is between full confidence on and 

performance of the firm. From these results, the hypothesis is validated with two Tunisian companies. 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
36 

© 2016 Conscientia Beam. All Rights Reserved. 

Table-4. Results of hierarchical regression steps 3 and 4 (Model 3) to Tunisian companies 

 
 
Variables 

Step 3 Step 3 and 4 Model 3 

Firm’s performance Firm’s performance 

ROA MTB ROA MTB 

 
t 

 
T 

 
t 

 
t 

V. control TAIL 0.112 1.166 * 0.180 1.148 * 0.020 1.730 *   0.186 1.174 * 

SECT 0.156 0.060 * 0.412  2.581** 0.068 1.689 *   0.390   2.325 ** 

V. independent SCONF - - - - 0.012 - 0.158 n.s - 0.068 - 0.490 n.s 
V. mediator DETTE  

BANC 
0.065 0.096 * 0.015  0.045 ** 0.028 0.150 *   0.056 0.086 * 

R 2 ajusté 0.056 0.065 0.057 0.039 

F      0.042 **    0.091 *   0.053*    0.098 * 
Adjusted R² variation   0.016 0.002 

*** Significant at 1%, ** significant at 5%, *: significant at 10%, n.s: not significant 

 

From Table 4 (Model 3), we note that the control variables (TAIL, SECT) are statistically significant for both 

indicators of firm performance thresholds of 5% and 10%. These results are consistent with the work of Strahan 

(1999) stipulating that the size of the company and the industry are two indicators necessary for the evaluation of 

the performance of a company. The results of these two variables emphasize their importance among the leaders in 

the financial decisions (financing, investment ...). 

According to Table 4, for both measures of performance, Model 3 (full model) has an interesting adjusted 

explanatory power. Thus, this comprehensive model, which takes into account the effect of the mediator bank debt, 

also increases the percentage of explained variance from Model 1. In cases where performance is measured by ROA, 

adjusted R2 passes from 0.041 to 0.057. Similarly, when performance is measured by MTB, adjusted R2 passes from 

0.019 to 0.039 and the F statistic becomes significant in Model 3 at the 10% threshold than that of the Model 1 (not 

significant). The adjusted R2 of increase is related to the consideration of the mediating effect of bank debt. Thus, 

the variation of adjusted R2 for both models associated with the addition of the variable is an important mediator 

(0.016 and 0.002). This shows that this variable is a good predictor of the dependent variable, namely the 

performance of the firm. 

 

3.3.3. Interpreting the Results of the Indirect Relationship between the Overconfidence and Firm 

Performance through the Bond Debt 

Consistent with the previous results, the first condition of the relationship between the variable SCONF and 

ROA was filled (see Table 5). The objective of the second step is to demonstrate the existence of a relationship 

between SCONF and bond debt (DEBT OBLIGA). Model 2 shows that overall quality is statistically significant at 

the 10% level and the SCONF variable is positively and significantly associated with DEBT OBLIGA Tunisian 

companies ( = 0.065, t = 0.038, p <10%), and the second condition of Baron and Kenny (1986) holds true. 

 

Table-5. Results of Hierarchical Regression of Steps 1 and 2 (model 1 and 2) to Tunisian Companies 

 
 
Variables 

Step 1 Model 1 Step 2 Model 2 

Firm’s performance Bond debt  

ROA MTB 

 
t 

 
t 

   
t 

V. control TAIL -0.201 - 1.501 n.s  0.042   0.314  n.s 0.294  2.340 ** 

SECT -0.059 - 0.430 n.s  0.227       1.638 * 0.332  2.583 ** 

V. independent SCONF 0.236 1.704 * -0.093 - 0.661 n.s 0.065     0.038 * 
R 2 ajusted 0.041  0.019                     0.154 

F    0.163 *       0.269 n.s           0.080  

         *** Significant at1%, ** significant at 5%, *: significant at 10%, n.s: not significant 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
37 

© 2016 Conscientia Beam. All Rights Reserved. 

Examining the results of Table 6 shows a positive and significant relationship between the bond debt and both 

indicators of firm performance (ROA and MTB). From these results, we see that the bond debt has a positive impact 

on the economic and market performance of the company (for ROA,  = 0.027, t = 0.178, p <10%; and for MTB: 

= 0.026, t = 0.046, p <10%). 

Model 3 (full model) is used to verify third-mediated condition of the bond debt between SCONF and firm 

performance (ROA and MTB). The results of the hierarchical regression analysis indicate that the bond debt 

(mediating variable) remains important in explaining the dependent variable (the two forms of performance) after 

considering the predictor variable. The statistical coefficient of the variable DEBT OBLIGA has a positive and 

significant value compared to the ROA (= 0.004, t = 0.030, p <10%) as compared to MTB ( = 0.186, t = 0.076, p 

<10%). It appears from these results that the third condition is completely verified.It only remains to verify the last 

condition, that is to say the effect of the predictor SCONF on the dependent variable (ROA and MTB) which should 

not be significant once the possible mediator (DETTE OBLIGA) is considered. 

The results in Table 6 show that the coefficients associated with the SCONF variable are statistically 

significant for both indicators of the performance of the firm (for ROA: = 0.236, t = 1.678, p <10% and MTB:  

= 0.163, t = 0.012, p <10%). It follows that bond debt is not a mediating variable on overconfidence and 

performance of the firm. From these results, the hypothesis 3 is not validated with Tunisian companies. 

 

Table-6. Results of Hierarchical Regression Steps 3 and 4 (Model 3) to Tunisian Companies 

 
 
Variables 

Step 3 Step  3 and 4 Model 3 

Firm’s performance  Firm’s performance  

ROA MTB ROA MTB 

 
T 

 
t 

 
T 

 
t 

   V. control TAIL - 0.172 - 1.199 n.s 0.086 0.620 n.s -0.202 -1.423 n.s 0.096 0.680 n.s 
SECT - 0.130 - 0.903 n.s 0.311 2.223 ** -0.060 -0.411 n.s 0.288 1.965 * 

 V. independent SCONF - - - -  0.236   1.678 * 0.163 0.012 * 
  V. mediator DETT

E 
OBLIA 

  0.027 0.178 * 0.026   0.046 *  0.004 0.030 * 0.186 0.076 * 

R 2 ajusted 0.022 0.041 0.059  0.028 
F    0.055 *    0.101 *    0.067 *     0.086 * 
Adjusted R² variation   0.018 0.009 

*** Significant at 1%, ** significant at 5%, *: significant at 10%, n.s: not significant 

 

From Table 6 (Model 3), we find that the control variables (TAIL, SECT) are not statistically significant for 

both indicators of firm performance (ROA, MTB), except the variable coefficient SECT which is statistically 

significant for the MTB indicator ( = 0.288, t = 1.965, p <10%). We conclude then that the industry has a positive 

impact on the market performance, as the company size is not an important factor in analyzing the performance of 

the firm. This leads us to also consider that large companies do not have necessarily a high level of confidence on 

leadership and high performance. 

According to Table 6 for both measures of performance, Model 3 (full model) has an interesting adjusted 

explanatory power. Thus, this comprehensive model, which takes account of the bond debt mediate, also increases 

the percentage of explained variance from the Model 1. In cases where performance is measured by ROA, adjusted 

R2 passes from 0.041 to 0.059. Similarly, when performance is measured by MTB, adjusted R2 Passes from 0.019 to 

0.028 and the F statistic is significant at the 10% level in Model 3 than in Model 1 (not significant). 

The adjusted R2 of increase is related to the consideration of the mediating effect of the bond debt. The change 

in adjusted R2 for the two models associated in addition with the mediating variable is significant (0.018 and 0.009). 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
38 

© 2016 Conscientia Beam. All Rights Reserved. 

This shows that this variable is a good predictor of the dependent variable, namely the performance of the firm. But, 

these results do not confirm our hypothesis (H3). 

 

3.3.4. Interpreting the Results of the Indirect Relationship between the Overconfidence and Firm 

Performance through External Funds 

Consistent with previous results, the first condition of the relationship between the variable SCONF and ROA 

was filled (see Table 7). 

The objective of the second step is to demonstrate the existence of a relationship between SCONF and external 

funds. Model 2 shows that the overall quality is statistically significant at the 10% level and the SCONF variable is 

positively and significantly associated with FONDS EXT Tunisian companies ( = 0.257, t = 1.990, p <10%), and 

the second condition of Baron and Kenny (1986) holds true. 

 

Table-7. Results of Hierarchical Regression of Steps 1 and 2 (model 1 and 2) to Tunisian Companies 

 
 
Variables 

Step 1 Model  1 Step 2 Model 2 

Firm’s performance  External Funds 

ROA MTB 

 
t 

 
t 

 
t 

V. control TAIL - 0.201 - 1.501 n.s   0.042   0.314  n.s   0.320   2.565  ** 
SECT - 0.059 - 0.430 n.s   0.227 1.638  * - 0.096 - 0.755  n.s 

V. independent SCONF   0.236    1.704 * - 0.093 - 0.661 n.s   0.257  1.990   * 
R 2 ajusted 0.041             0.019                        0.169 
F    0.163 * 0.269  n.s 0.050  

          *** Significant at 1%, ** significant at 5%, *: significant at 10%, n.s: not significant 

 

Examining the results of Table 8 reveals a positive and significant relationship between external funds and two 

indicators of firm performance (ROA and MTB). From these results, we see that external funds have a positive 

impact on the economic and market performance of the company (for ROA:  = 0.032, t = 0.006, p <10%; and for 

MTB: = 0.091, t = 0.075, p <10%). 

Model 3 (full model) is used to check the condition of third mediation between own external funds SCONF and 

firm performance (ROA and MTB). The results of the hierarchical regression analysis indicate that external funds 

(mediating variable) remain important in explaining the dependent variable (the two forms of performance) after 

considering the predictor variable. The statistical coefficient of the variable FONDS EXT has a positive and 

significant value compared to the ROA ( = 0.015, t = 0.066, p <10%) and MTB ( = 0.022, t = 0.058, p <10%). It 

appears from these results that the third condition is completely verified. The results in Table 8 show that the 

coefficients associated with the variable SCONF are not statistically significant regardless of the extent of the 

performance of the firm retained, while it was statistically significant for the ROA in the first step of the process 

Baron and Kenny (1986), the regression coefficients of the variable SCONF have no significant signs relative to 

ROA ( = 0.082, t = 0.875, p> 10%) and the MTB ( = -0.077, t = -0.527, p > 10%). It follows that the mediation 

by external funds is full on between the overconfidence and performance of the firm. Through these results, we see 

that the SCONF variable has a positive effect on firm performance. The assumption 4 is then validated with 

Tunisian companies. According to Table 8 for the two performance measures, Model 3 (full model) has an 

interesting adjusted explanatory power. Thus, this comprehensive model, which takes into account the mediating 

effect of external capital, also increases the percentage of variance explained from the Model 1. In cases where 

performance is measured by ROA, adjusted R2 passes from 0.041 to 0.089. Similarly, when performance is measured 

by MTB, adjusted R2 Passes from 0.019 to 0.056 and the F statistic is significant at the 10% level in Model 3 that 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
39 

© 2016 Conscientia Beam. All Rights Reserved. 

the Model 1 (not significant). The adjusted R2 of increase is linked to the consideration of the mediating effect of 

external funds. 

The change in adjusted R2 for the two models associated with the addition of the mediating variable is 

significant (0.048 and 0.037). This shows that this variable is a good predictor of the dependent variable, namely the 

performance of the firm. 

 

Table-8. Results of Hierarchical Regression Steps 3 and 4 (Model 3) to Tunisian Companies 

Variables Step 3 Step 3 and 4 Model 3 

Firm’s performance  Firm’s performance  

ROA MTB ROA MTB 

 
T 

 
t 

  
t 

 
t 

  V. control TAIL    0.065 1.723 * 0.057 0.398 n.s  0.188 1.668 * 0.062 0.425 n.s 

SECT - 0.131 - 0.950 n.s 0.238 1.753 * -0.072 - 0.519  n.s 0.221 1.575 n.s 

 V. independent SCONF - - - -  0.082    0.875 n.s -0.077 - 0.52 n.s 

  V. mediator FONDS   
EXT 

  0.032 0.006 * 0.091 0.075 *  0.015 0.066 * 0.022 0.058 * 

R 2 ajusted 0.015 0.027 0.089 0.056 

F    0.009 *    0.056 *    0.075 *     0. 098 * 

  0.048 0.037 

*** Significant at 1% , ** significant at 5%, *: significant at 10% , n.s: not significant 

 

The results in Table 8 (Model 3) show that the control variables (TAIL, SECT) are not statistically significant 

for both indicators of the performance of the firm, unless the statistical coefficient of TAIL variable has a positive 

value and significance compared to ROA ( = 0.188, t = 1.668, p <10%). This result is consistent with the work of 

Strahan (1999) stipulating that the size of the business is a necessary indicator for evaluating the performance of a 

company. We believe that the company size is a decisive and significant criterion for the firm’s performance and 

financing structure. 

From these results, we note that the Tunisian listed firms in the Tunis stock exchange seem to prioritize their 

funding resources by favoring self-financing at the expense of external resources. If they have to resort to external 

funding, they prefer the obligator bank debt. The validation of the funding hierarchical theory is based on the 

existence of asymmetric information that could lead to adverse selection problems from external investors. The role 

of information asymmetry on the choice of Tunisian companies funding is confirmed. The most sensitive 

information asymmetry a reveals that companies rely primarily on debt. However the least affected by adverse 

selection companies prefer the emissions of capital. 

 

4. CONCLUSION 

Under the current behavioral finance companies, this article examined the relationship between the company’s 

overconfidence and performance through the funding structure. The emerging literature on overconfidence 

confirms the marked presence of bias among business leaders. Theoretical approaches, however, do not conclude in 

a systematically negative effect, particularly in the case of a moderate confidence. 

Several studies show the positive effect of managers’ overconfidence on financial decisions (DeBondt and 

Thaler, 1994; Bukszar, 2003; Véronique, 2007). The majority of these studies focus on the direct relationship 

between managers’ overconfidence and performance without taking account of other intermediate factors that may 

be relevant for the understanding of this indirect relationship. 

Empirically, the regression results show that leader’s overconfidence has an impact on the performance of 

Tunisian companies through the mediation of the financing structure. Indeed, these results indicate that Tunisian 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
40 

© 2016 Conscientia Beam. All Rights Reserved. 

companies have an interest to finance their activities and investments using the various modes of financing such as 

self-financing, bank loans, bond debt and equity external funds can be to increase their performance. Moreover, we 

noted that the confidence of the manager plays an important role in improving the performance of Tunisian 

companies. 

The modeling of relations between the three concepts, namely overconfidence / financing structure / firm 

performance, could be summarized as follows. Because the funding structure could act as a mediating variable, 

checking this mediating effect was achieved by developing models based on variables selected in this study. 

In this respect, our results indicate that the SCONF variable is relevant in determining the mediate with the 

methodology of Baron and Kenny (1986). Indeed, considering the mediating variable, the financing structure 

improves in a significant way the explanatory power of the model based on overconfidence / financing structure / 

performance. 

From these results, we note that the mediating variable is a good predictor of the dependent variable, namely 

the performance of the firm. However, hierarchical regressions show that control variables have a significant effect 

on the financing structure and the performance of the firm. This angular part allowed us above all to empirically 

explore the contrast between what the theory says and what is observed in practice. This study limits and still leave 

many questions open about the issue of performance, overconfidence and the funding structure. The model should 

include other variables to represent more fully the economic reality. Finally, the consideration of overconfidence 

opens new perspectives in corporate finance, including a renewal of issues relat related to governance mechanisms 

and value creation associated with the ability to develop investment opportunities. 

 

Funding: This study received no specific financial support. 
 

Competing Interests: The authors declare that they have no competing interests. 
 

Contributors/Acknowledgement: All authors contributed equally to the conception and design of the study.  

 

REFERENCES 

Baker, M., R. Ruback and J. Wurgler, 2004. Behavioral corporate finance: A survey. NBER Working Paper, No. W10863. 

Baron, R.M. and D.A. Kenny, 1986. The moderator-mediator variable distinction in social psychological research: Conceptual, 

strategic and statistical considerations. Journal of Personality and Social Psychology, 51(6): 1173-1182. 

Ben-David, I., J. Graham and C. Harvey, 2006. Managerial overconfidence and corporate policies. AFA 2007, Chicago Meetings 

Paper, NBER Working Paper, N°. 13711. 

Bernardo, A. and I. Welch, 2001. On the evolution of overconfidence and entrepreneurs. Journal of Economics and Management 

Strategy, 10(3): 301-330. 

Bukszar, E., 2003. Does overconfidence lead to poor decisions? A comparison of decision making and judgment under 

uncertainty. Journal of Business and Management, 9(1): 33-43. 

Charreaux, G., 1999. La théorie positive de l’agence: Lecture et relectures, In G. Koenig (Coordonné). De nouvelles théories pour 

gérer l’entreprise du XIXè siècle. Working Papers CREGO from Université de Bourgogne - CREGO EA7317 Centre 

de recherches en gestion des organisations. Economica: 61-141. 

Daniel, K., D. Hirshleifer and S. Teoh, 2001. Investor psychology in capital markets: Evidence and policy implications. Dice 

Center Working Paper No. 2001-10. 

Daniel, K. and S. Titman, 1999. Market efficiency in an irrational world. Financial Analysts Journal, 55(6): 28-40. 

De Bondt, W. and R. Thaler, 1995. Financial decision making in markets and firms: A behavioral perspective, in handbook in 

operations research and management science, Eds., by Jarrow V., Maksimovic R., Ziemba W., Finance. Amsterdam: 

(Elsevier), 9: 385–410. 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
41 

© 2016 Conscientia Beam. All Rights Reserved. 

DeBondt, W. and R. Thaler, 1994. Financial decision-making in markets and firms: A behavioral perspective. NBER Working 

Paper, No. W4777. 

Fairchild, R., 2005. The effect of managerial overconfidence, asymmetric information, and moral hazard on capital structure 

decisions. Available from SSRN 711845. 

Gervais, S., J. Heaton and T. Odean, 2003. Overconfidence, investment policy, and executive stock options, Rodney L. White 

Center for Financial Research Working Paper No. 15-02. 

Hackbarth, D., 2004. Managerial traits and capital structure decisions. EFA 2004 Maastricht Meetings Paper.  

Heaton, J., 2002. Managerial optimism and corporate finance. Financial Management, 31(2): 33-45. 

Hermalin, B.E. and A.M. Isen, 2000. The effect of affect on economic and strategic decision making. USC Center for Law, 

Economics & Organization Research Paper No C01-5, Juillet. 

Igalens and Roussel, 1998. Human resources management research methods. Paris: Economica, Collection Management 

Research. 

Jensen, M.C. and W. Meckling, 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of 

Financial Economics, 3(4): 305-360. 

Keiber, K., 2006. Managerial compensation contracts and overconfidence. EFMA Behavioral Finance Symposium Paper. 

Malmendier and Tate, 2005b. Does overconfidence affect corporate investment? CEO overconfidence measures revisited. 

European Financial Management, 11(5): 649-659. 

Malmendier, Tate and Yan, 2005. Corporate financial policies with overconfident managers. 8th Annual Texas Finance Festival. 

Manel, D., 2015. Bias behavioral, financing structure and performance of Tunisian companies. Master Thesis, FSEG, University 

of Sfax,Tunisia. 

Martinet, 1990. Epistemologies and management science, Éd. Paris: Économica. 

Modigliani and Miller, 1958. The cost of capital, coporation finance and the theory of investment. American Economic Review, 

53(3): 261-275. 

Nekhili, M., B. Sabri and l. Faten, 2012. Ownership structure, voluntary R & D disclosure and market value of firms: The French 

case. International Journal of Business, 17(2): 126-140. 

Roll, R., 1986. The hubris hypothesis of corporate takeovers. Journal of Business, 59(2): 197-216. 

Russo, J.E. and P. Schoemaker, 1992. Managing overconfidence. Sloan Management Review, 33(2): 7-17. 

Shiller, R.J., 1997. Human behavior and the efficiency of the financial system, John B. Taylor and Michael Woodford (Eds). 

Handbook of Macroeconomics. 

Strahan, P.E., 1999. Borrower risk and the price and nonprice terms of bank loans. Federal Reserve Bank of New York in Its 

Series Staff Reports No. 90. 

Véronique, B., 2007. Overconfidence financial leaders and decisions: A synthesis. Finance Control Strategy, 10(1): 39 – 66. 

Wacheux, F., 1996. Qualitative methods and research management. Paris: Economica. pp: 290. 

Weinstein, N., 1980. Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39(5): 806-820. 

Zouari, G. and R. Zouari-Hadiji, 2010. Internal governance systems and R&D investment: An international comparison. 

Corporate Board: Role, Duties & Composition, 6(1): 39-56. 

Zouari, G. and R. Zouari-Hadiji, 2013. Ownership structure, innovation and firm performance: Evidence from Tunisia. 

International Journal of Governance, 3(2): 79-107. 

Zouari, G. and R. Zouari-Hadiji, 2014a. Directors board, R&D investment and the firm’s performance: Evidence from the French 

case. Corporate Board: Role, Duties & Composition, 10(2): 85-101. 

Zouari, G. and R. Zouari-Hadiji, 2014b. The indirect impact of the board of directors composition on the firm's performance: An 

international comparison. Asian Journal of Empirical Research, 4(10): 468-487. 

 



Financial Risk and Management Reviews, 2016, 2(1): 26-42 
 

 
42 

© 2016 Conscientia Beam. All Rights Reserved. 

APPENDIX 

 

Tableau-9. Pearson correlation matrix 

 (1) (2) (3) (4) (5) (6) (7) 

SCONF (1) 1       
AUTOF (2) 0.393 1      
DETTE BANC (3) 0.286 0.363 1     
DETTE OBLIGA (4) -0.219 0.014 0.143 1    
FONDS EXT (5) -0.270 0.043 0.183 0.495 1   
TAIL  (6) 0.253 0.331 0.250 -0.044 -0.051 1  
SECT (7) 0.230 0.153 0.316 0.387 0.313 0.357 1 

 (1) Note: that all the correlations between the explanatory variables are significantly smaller than 0.6 (threshold at which we 

begin to encounter serious problems of multi collinearity). Pearson test and condition index have revealed that these variables 

are also distinct from each other and are non-significant (higher correlation levels 10% and condition index is less than 1000). 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Views and opinions expressed in this article are the views and opinions of the author(s), Financial Risk and Management Reviews shall not be responsible or 
answerable for any loss, damage or liability etc. caused in relation to/arising out of the use of the content. 

 


