Challenges and Ways to Develop Insurance Industry in KSA Market

Psychology and ethnic play a central role within Saudi Arabia market. Thus, this editorial argues the cultural and emotion effects on the Saudi consumption of life insurance. Even thought, the present paper intend to understand the reasons of the slowly growth of the purchase of this product and how to increase its rate in the insurance portfolio? Subsequently, this study target to confirm the significant effects of religiosity, optimism, opinion leadership, emotional regulation on Saudi behavior to subscribe in life insurance. Data was collected from two samples. The first sample, count 210 respondents, worn to purify the measurement scales of the variables studied in the exploratory phase. The second sample was administered from belonging 654 policyholders in order to confirm the measures instruments, to verify the hypotheses, validate and re-specify the model. Thus, this study bears a theoretical interest for researchers and it is useful for practitioners in this sector.

he insurance sector in the Gulf region is expected to know an 18,1% growth between 2012 and 2017, reaching a value of USD 37.5 billion, according to reports by Alpen Capital (2013), and the Bahrain Economic Development Board (2012). Moreover, it is expected that the size of the insurance market in KSA will reach about USD 37.5 billion by 2017.
This growth is accounted for by the mandatory areas of the health sector, auto insurance, professional cover, and population progress, as well as by encouraging economic factors in the region.
When compared to similar developed markets, insurance in the Gulf Cooperation Council countries (GCC) in general and KSA in particular is still below potential and expectations. Despite the rise of its insurance spread index from 0,59% in 2007 to 0,86% in 2011, Saudi Arabia is still considered below the desired rate, compared to the global insurance, which amounts to 6,6%. It should be noted that during the year 2013, KSA insurance companies witnessed USD 345 million losses; compared with a profit of USD149 million recorded during the same period of 2012 (the British Foundation of International Business Monitors, specialized in markets studies, 2014 cited by Ethenian, 2014). Nerveless, this industry was conquered by health insurance with 52% of the sector operations.
Therefore, it has become necessary to reconsider the composition of insurance products. Thus, insurers should focus on ways to create a consumption culture of all products, not just the mandatory ones. Consequently, insurance companies need to target profitable branches and segments within the insurance market. In this regard, the study is useful for academics and practitioners interested in this sector. According to the report on the insurance sector in the GCC which was prepared by Alpen Capital in 2011, the density of life insurance in this region grew of 22.2% in the period from 2011 (USD 50.8) to 2015 (USD 113.5).
The Saudis consumption is affected by their psychology and ethnic (Madichie, 2013;Abosag & Farah, 2014;Siamagka & Balabanis, 2015). In this study we diagnose the factors that motivated the development of the life insurance sector in KSA. Specifically, we concentrate on the cultural and psychological constructs affecting the consumption of life insurance products in the Saudi market. Then, we hope to verify the effects of anxiety T International and Multidisciplinary Journal of Social Sciences,5(2) 155 about death and religiosity on behavior underwriting life insurance, as well as the mediator role of opinion leadership and moderator role of regulation emotional.
We have choose these independents constructs to explain the Saudi purchase behavior of life insurance products because, not a few research insist on these variables, within this area (Toukabri et al., 2015;Tienyu et al., 2003;Chen et al., 2015).
To answer to this research questions, we focuses on the study of anxiety about death and religiosity among consumers of life insurance as well as the analyses of the state of optimism in life insurance industry. In the second part, we focus on opinion leadership as a mediator between this explicative variables and life insurance consumer behavior. The last part will be devoted to learn the emotion regulation as a potential moderator of the relationship between emotional states and opinion leadership. This framework will, thereafter, validate through a quantitative study.

Literature Review and Hypotheses
Cognitively, anxiety induced an attention on incidents present (Urien, 2002;Wink & Scott, 2005); whereas behaviorally, this is to use the known methods and routines to cope with life situations, or, take preventive action to keep the status quo rather than taking risks to improve his fate (Urien, 2002). Moreover, Dib & Valette-Florence (2007) admit that the cognitive processes of the threat assessment are responsible for the negative emotion. Urien (2002) emphasized on the cognitive and emotional effects of anxiety about the future. Then, Gentry, Kennedy, Paul & Hill (1995) and Gentry, Baker & Kraft (1995) show that the purchase of certain products (life insurance, house, car) by consumers is a preservation of their identity beyond death.
Studies on the subject of optimism turn to explain self-defense and supposed that it reduce anxiety (Milhabet et al., 2002;Grewal, Mehta & Kardes, 2004). Therefore, reducing anxiety involves two contrasting positions. The first contrast predicts that the optimism is the result of the struggle against anxiety (Taylor & Brown, 1988), and the second confirm that the best defense against anxiety is pessimism (Colvin & Block, 1994).  Religiosity is defined by Johnson, Jang, Larson & Li (2001) as "the extent to which an individual is committed to religion [...] as the attitudes and behaviors of the individual reflect this commitment." The attitudes of the individual and their behavior reflect this commitment. Religious people have value systems differ from the least religious and non-religious. Thus, religiosity is important as it able to influence an individual behavior (Mokhlis, 2008). Thus, it is necessary to integrate this construct in our research model, because the importance of religiosity in Saudi Arabia and its effects on consumer behavior.
Etymologically, opinion leadership means the ability of an individual to conduct or lead other individuals or organizations in order to achieve certain goals. Moreover, the opinion leadership profile characterizes the position of one or more persons to conduct a group. Also, the opinion leadership is known for his skills with other members of the group (Blanchet & Trognon, 1994). Then, under certain conditions, if the degree of emotional contagion from the opinion leadership to the followers is large enough, the effect of verbal communication can be important (Sy et al., 2005). Thus, this ability gives leaders opportunities to influence other members and legitimacy in decision-making (Marc & Picard, 2003).
Emotional regulation refers to attempts made by individuals to influence the sort of emotions they feel (Gross et al., 2006;Mikolajczak et al., 2008). It refers to emotional skills that enable the individual to manage his emotive state and thus to contribute appropriately to the changeable and complex situations of modern life (Gendron, 2006).
After that, we extant the effects of these concepts on underwriting in life insurance and we deduct the hypotheses.

Anxiety about Death
The objective of owning some products is preserving identity beyond death and reducing anxiety about demise (Kasser & Sheldon, 2000). Capelli (2003) notes that anxiety contained in advertising communication positively affects the purchase decision. Anxiety about death affects not only the emotions of the consumer, but also his purchase behavior and his recommendation of products to others. Kopp & Pullen (2002) report the positive effect of anxiety about death on the purchase behavior of products like life insurance. Similarly, Gentry & Goodwin (1995) assert that the design of insurance policies for seniors is a way of symbolically extending the self beyond death. For Urien (2001), the acquisition of products at the end of life is an expression of symbolic immortality. Hence, we can state the following hypotheses: H1.a : Fear of death has a positive influence on the repeat purchase behavior. H1.b : Fear of death has a positive influence on the behavior recommendation. H1.c : Fear of death has a positive influence on the opinion leadership. H2.a : Fear of dying has a positive influence on the repeat purchase behavior. H2.b : Fear of dying has a positive influence on the behavior of recommendation. H2.c : Fear of dying has a positive influence on the opinion leadership. Goleman (1997) states that optimism is the healthiest way to educate people. Optimism motivates consumers internally and not by the perspective of threat or reward. There would be differences in the encoding of information that would play in the adjustment of beliefs maintaining the level of optimism (Sharot, Korn & Dolan, 2011). In the same way, these positive impressions have a positive impact on purchase behavior in life insurance. Besides, they have an effect on both mental and physical health. We can, therefore, propose the following hypotheses: H 3.a : Optimism has a positive influence on the repeat purchase behavior. H 3.b : Optimism has a positive influence on the behavior of recommendation. H 3.c : Optimism has a positive influence on the opinion leadership.

Religiosity
The importance of religiosity in KSA makes it necessary to integrate this construct in our research model. Religiosity is the extent to which a person is involved in religion (Jang & Larson, 2001). It influences an individual cognitively and behaviorally (Mokhlis, 2006). Rodero & Branas (2000) confirm the influence of religious institutions in acquiring life insurance products. Arnould, Price & Zikhan (2004) stress the role of religion in the development of attitudes and the shaping of purchase intentions. Thus, we intend to check the following hypotheses in the present research paper.
H 4.a : Intrapersonal religiosity has a negative influence on the repeat purchase of behavior. H 4.b : intrapersonal religiosity has a negative influence on the behavior of recommendation. H 4.c : intrapersonal religiosity has a negative influence on the opinion leadership. H 5.a : interpersonal religiosity has a negative influence on the repeat purchase of behavior. H 5.b : interpersonal religiosity has a negative influence on the behavior of recommendation. H 5.c : interpersonal religiosity has a negative influence on the opinion leadership.

Opinion Leadership
Opinion leadership is closely related to the leader's aptitude to diffuse innovation (Roehrich, 1993). The literature distinguishes between three perspectives to define opinion leadership, namely the ability to influence others (Goldsmith & De Witt, 2003), the capacity to provide information and to urge one's entourage about a subject (Gilly & Graham, Wolfinbarger & Yale, 1998), and the combination of the two precedent dimensions (Mowen, 1995). To this end, research on opinion leadership confirms that those who occupy a leading position have greater confidence and are more popular, adaptable, attentive and cooperative than others (Vernette, 2006). Hence, the following hypotheses are formulated: H 6.a : Leadership has a positive influence on the repeat purchase behavior. H 6.b : Leadership has a positive influence on the behavior of recommendation.
International and Multidisciplinary Journal of Social Sciences,5(2) 159

The Mediating Role of Opinion Leadership
Leaders are characterized by a coordinating role, a socio-relational role, and a specific interactive role (Barcellini, 2008). Gendron (2008) explains that opinion leaders have the ability to transcend this situation by imposing a vision, and to amend the behavior of their followers so as to make them true agents of change. Vernette (2006) indicates that it is neither the institution nor the charisma that establishes the leader's authority, but the fact that he serves the people around him. The opinion leader must be able to easily get into the skin of his surroundings and provoke an identification reaction or an emotional relationship (Vernette, 2002). One might, therefore, make the following two hypotheses: H 7.a : Leadership mediates the effect between fear of death and repeat purchase behavior. H 7.b : Leadership mediates the effect between fear of death and the behavior of recommendation. H 7.c : Leadership mediates the effect between the fear of death and repeat purchase behavior. H 7 . d : Leadership mediates the effect between the fear of death and the behavior of recommendation. H 8.a : Leadership mediates the effect between optimism and repeat purchase behavior. H 8.b : Leadership mediates the effect between optimism and behavior of recommendation.

The Moderating Effect of Emotional Regulation
Among the possible strategies of emotion regulation, Gross et al. (2006), Gross (1998) and Gross & John (2002 focus on two specific ones, namely cognitive reappraisal (up regulation) and expressive suppression (down regulation). Cognitive reappraisal means the cognitive process by which the evaluation of a situation helps to mitigate or increase the emotional character (Christophe et al., 2009). In most cases, this strategy would reduce negative emotions and increase the positive emotions and the psychological situation arising from welfare. In this case, it approaches the notion of positive revaluation result, found in the literature on coping, and the optimism of selecting the positive aspects of a situation in anticipation of its consequences. As for expressive suppression, it consists in inhibiting one's emotions so that they do not communicate to others information about one's emotional states (Fischer, 2000). Hence the following hypotheses: Thus, the conceptual model of our research is presented in (see figure 1), after elucidating the different hypotheses and concepts.

Sample
The research data was collected through surveys. We chose the area sampling procedure, and used two samples. The first one, whereby the measurement scales of the conceptual constructs were purified, targeted 210 respondents. The second one involved 654 respondents and was applied in the confirmation phase.

Measures
The original scales were in English. So, the measures have undergone the process of translation from English into Arabic, using the back translation method. Respondents were solicited to answer by ticking options on a five-International and Multidisciplinary Journal of Social Sciences, 5(2) 161 point Likert type scale ranging from 1 (strongly disagree) to 5 (strongly agree). For religiosity, the scale of Mokhlis (2006) was used. To measure anxiety related to fear of death, we relied on the scale of Wink & Scott (2005). Optimism was assessed via the scale developed by Sharot, Korn & Dolan (2011). For the concept of opinion leadership we employed the measure of Childers (1986), and for emotional regulation, we adopted the scale of ERQ, developed by Gross & John (2003). Finally, to measure the behavior of subscribing to a life insurance, we opted for the scale of Durvasula, Lysonski, Mehta & Peng (2004).

Validity Check
The principal component analysis confirmed the dimensional structures of our research measures (see table 1). However, few items were removed because of their low correlation with the selected dimensions. A confirmatory factor analysis was carried out for each scale to check the validities (see table 2). The rho convergent validity varies between 0.501 and 0.504 (see table 3). Then, the examination of the results of the φ matrix shows that all correlations among the eight variables are positive and significant. These correlations are established at low levels, which implies that there is no co-linearity between variables and provides evidence of the discriminate validity of each construct. Cronbach's alpha and Jöreskog's rho approved a value more than 0.7, which accepted to judge the good representatively of constructs by these items. Also, the loading of each item on its concept passed 0.70. Moreover, all pairs of concepts verified the conditions of Fornell & Larcker's (1981) test of discriminant validity (see table 4).

The Structural Model
The model was tested using AMOS software, with the maximum likelihood method. The overall model fit was satisfactory (Roussel, Durrieu, Campoy & El Akremi, 2002). These values indicate an acceptable model fit to the data set (see figure 2). Henceforth, it is possible to interpret the results related to structural links. Table 5 shows the results of the verification of relations between model constructs.

Hypotheses Test for Direct Relations
H 1.a about the relationship between fear of death and repeat purchase behavior was confirmed (p = 0.003 < 0.05). Indeed, the value of the Student t expressing this relationship is equal to 2.933 that it is greater than 1.96. In the same line, the hypothesis H 1.b about the relationship between fear of death and behavior of recommendation was confirmed (p = 0.000 < 0.05). The value of the Student t expressing this relationship is equal to 4.133 that it is greater than 1.96. Unlike these two hypotheses, the hypothesis H 1.c about the relationship between fear of death and opinion leadership was infirmed (p = 0.299 > 0.05; Cr = 1.039 < 1.96).
H 2.a linking the fear of dying to the repeat purchase behavior is validated (p = 0.003 < 0.05) as the value of Student's t (Cr = 2.991) is greater than 1.96. In addition, the hypothesis H 2.b , reflecting the link between fear of dying and behavior of recommendation is significant and validated (p = 0.000). The Student t value is equal to 5.224 and therefore greater than |1.96|. Same, the hypothesis H 2.c between fear of dying and the opinion leadership is validated (p = 0.005 < 0.05). The Student t value is equal to 2.824 and therefore greater than |1.96|.
H 3.a linking optimism to the repeat purchase behavior is unverified (p = 0.493 > 0.05) as the Student value (Cr = 0.685) is low than 1.96. Same, the H 3.b hypothesis linking optimism to the behavior of recommendation is unverified (p = 0.210 > 0.05) as the value of Student's t (Cr = 1.253) is low than 1.96. Unlike, the H 3.c hypothesis linking optimism to the opinion leadership was confirmed (p = 0.000 < 0.05) as the value of Student's t (Cr = 3.552) is greater than 1.96.
H 4.a about the relationship between intrapersonal religiosity and repeat purchase behavior was invalidated (p = 0.218 > 0.05). The value of the Student t expressing this relationship is equal to 1.233 that it is low than 1.96. Against, the hypothesis H 4.b about the relationship between intrapersonal religiosity and the behavior of recommendation was confirmed International and Multidisciplinary Journal of Social Sciences, 5(2) 163 (p = 0.042 < 0.05). The value of the Student t expressing this relationship is equal to 2.031 that it is greater than 1.96; under the hypothesis H 4.c about the relationship between intrapersonal religiosity and opinion leadership was invalidated (p = 0.066 < 0.05; Cr = 1.838 < |1.96|).
H 5.a about the relationship between interpersonal religiosity and repeat purchase behavior was rejected (p = 0.134 > 0.05). The value of the Student t expressing this relationship is equal to -1.498 that it is low than 1.96. Against, the hypothesis H 5.b about the relationship between interpersonal religiosity and the behavior of recommendation was confirmed (p = 0.018 < 0.05). The value of the Student t expressing this relationship is equal to -2.356 that it is greater than |1.96|; under the hypothesis H 5.c about the relationship between interpersonal religiosity and opinion leadership was invalidated (p = 0.063 < 0.05; Cr = -1.859 < 1.96).
Finally, H 6.a linking the opinion leadership to the repeat purchase behavior is validated (p = 0.000 < 0.05) as the value of Student's t (Cr = 5.070) is greater than 1.96; in the same case that the H 6.b hypothesis linking the opinion leadership to the behavior of recommendation is validated (p = 0.000 < 0.05) as the value of Student's t (Cr = 7.267) is greater than 1.96 (see table 5).

The Mediator Role of Opinion Leadership
In order to test the effect of this mediation, we need to check the four conditions recommended by Baron and Kenny (1986) (see table 6). Testing the mediating effect of opinion leadership on fear of death and repeated purchase behavior shows that no condition is significant. We can, therefore, state that mediation is absent in hypothesis H7.a. Therefore, the variable 'opinion leadership' is not a mediator in the impact of fear of death on the behavior of repeated purchase. Similarly, testing the mediating effect of opinion leadership on fear of death and the behavior of recommendation shows that mediation is absent in hypothesis H7.b.
Then, in testing the mediating effect of opinion leadership on fear of death on the one hand, and repeated purchase and recommendation of insurance products on the other, only conditions 1 and 2 are significant at 5%. Consequently, mediation is partial in hypotheses H7.c and H7.d.

Toukabri & Ibrahim -Challenges to Develop Insurance Industry
Finally, testing the mediating effect of opinion leadership on optimism, repeated purchase and recommendation behavior shows that all conditions are significant at 5%. Therefore, mediation is total in hypotheses H8.a and H8.b.

The Moderating Role of Emotional Regulation
The test of the moderating effect of the variable 'emotional regulation' on the existing relationship between the two exogenous variables, anxiety towards death and optimism, and the endogenous variable 'opinion leadership' is based on the procedure proposed by Irwin and McClelland (2001).
Results (see table 7) indicate that the moderating effect is noticed in hypothesis H9.a Indeed, the link between fear of death and opinion leadership is not significant (p = 0.250), neither is the link between the moderating variable 'expressive suppression' and the variable 'opinion leadership' (p = 0.007). The interaction between expressive suppression and fear of death exerts, in turn, a significant effect (p = 0.022) on the dependent variable 'opinion leadership'. The research proposal related to this hypothesis is, therefore, confirmed. Contrarily to these hypotheses, results indicate that the moderating effect is absent in hypothesis H9.b because the interaction between expressive suppression and fear of death does not have a significant effect (p = 0.173) on the dependent variable 'opinion leadership'. The research proposal related to this hypothesis is, therefore, infirmed. Whereas, the interaction between expressive suppression and optimism exercise a significant effect (p = 0.020) on the dependent variable 'opinion leadership'. The research proposal related to hypothesis H9.c is, therefore, confirmed.
Although, Results indicate that the moderator in hypothesis H9.d does not validated since the link between optimism and opinion leadership is insignificant (p = 0.474). Thus, the link between the moderating variable 'cognitive reappraisal' (Up-regulation) and the variable' opinion leadership' is excluded and the research proposal related to this hypothesis is, therefore, rejected.
International and Multidisciplinary Journal of Social Sciences,5(2) 165

Re-specification of the Structural Model
The re-specification of the initial model of our study was conducted through changes indices (MI). However, in order to improve our model fit indices, we added a set of links between error terms (including MI> 3.84) to these changes indices. On the other hand, the removal of unconfirmed relationships is justified by the test significance level (p > 0.005), and the Student test (t < 1.96). To this end, we removed the link between intrapersonal religiosity and opinion leadership, the link between fear of death and opinion leadership, the link between interpersonal religiosity and opinion leadership, the link between optimism and repeated purchase behavior, the link between optimism and behavior of recommendation, the link between intrapersonal religiosity and repeated purchase behavior, and finally the link between interpersonal religiosity and repeated purchase behavior.
Indices for the revised model ( figure 3) are broadly acceptable. Indeed, the GFI (0.980), AGFI (0.966) and RMR (0.020) evidence demonstrate that the model accounts for almost all of the variance. RMSEA is 0.005 and the CFI index (1.000) indicates that the tested model is better than more restrictive models. The value of normalized chi-square (1.017) and PNFI (0.636) tell us about the parsimony of the model.

Discussion and Conclusion
The life insurance industry has identified a niche market for the financial protection of homes and businesses. Therefore, it is necessary that the study of behavior in life insurance investigates the emotional aspect of the insured to detect his motives in buying a tailored service. This research is presented as an extension of previous works conducted on consumer behavior in the context of Saudi insurance services.
The empirical results of our research confirm the work of Kasser and Sheldon (2000), Fischler (2001), Urien (2002) and Kopp and Pullen (2002). Our findings assert the influence of anxiety about death on consumer behavior and lifestyle of individuals in a well-defined culture. Anxiety about death encourages consumers to subscribe to insurance policy. The fear of death is a concept that awakens in people a need to counteract this risk that 166 Toukabri & Ibrahim -Challenges to Develop Insurance Industry could occur at any time. The concern of the insured with the prospect of death focuses on the descendants and ascendants that are dependent on him.
Hence, anxiety about death lies at the heart of basic life insurance products. In Saudi Arabia, the need to protect oneself against the risk of death is comparable to one's need to protect oneself against other risks, and pushes policyholders to subscribe to insurance. As suggested by Lifton (1973), symbolic immortality can alleviate the anxiety resulting from fear of death. A contradictory position can increase fear of death, for instance opening the possibility of life after death without providing the certainty of its acquisition or raising the specter of punishment after death (Nelson & Cantrell, 1980). To this end, Capelli (2003) noted that anxiety contained in advertising communication positively affects the purchase decision of the consumer of a service that is the case for Tanner, Hunt and Eppright (1991), Ben-Ari (2000), Gallopel (2002) and Urien (2003) whose have argued that the use of fear in communication affects behavior.
In our research work, we empirically confirmed the theory asserting the critical role of optimism about the behavior of subscribing to insurance. These results are in keeping with the work of Sharot, Korn and Dolan (2011). Hence, optimism proves to be a predicting element in someone's intention to buy insurance services, as well as in his intention to recommend it to his family. Theil (2003) reported that insurance is sold, not bought. This means that it makes more sense to sell insurance through a third party. The insured needs the intervention of an opinion leader to buy insurance products. The information conveyed by the latter can attract more attention to insurance products (Browne & Hoyt, 2000). Therefore, the opinion leader becomes a representative of the insurance service. Our model validates the mediating effect of opinion leadership between emotional variables and behavior involving buying insurance. The results of our test on the moderating effect of emotion regulation on the relationship between emotional variables and opinion leadership go against the work of Gross and John (2003). Indeed, our results indicate that emotional regulation only allows reducing negative emotions including fear of death, but it is not so with optimism and fear of dying.
The present research focuses on the marketing communication led by life insurance companies. It is a subtle combination meant to influence the buying behavior of customers. The focus on marketing communication is crucial to insurance companies in the Saudi market. The communication strategy of these insurance companies remains inexistent. Moreover, the information conveyed is very limited due to lack of available strategies and tools established for customers. Thus, insurers can focus on two areas in their communication campaigns. Firstly, they can try to detect the elements that elicit either anxiety or optimism in the insured. Secondly, they can try to determine the leadership's profiles as their lifestyle, the service delivery system.... Also, Opinion leadership is considered as a lever of satisfaction and loyalty of policyholders. Then; these stimuli will serve as a basis to encourage life insurance purchase in the KSA market.
No research work can escape having limitations. Yet these limitations can often open new tracks for future research. The limitations inherent to our research work are technical and time-related. The technical constraint we encountered consists in the difficulty to integrate other cognitive and affective constructs in order to respect the parsimony of our conceptual model. As for the time constraint, it has prevented us from making a panel to study the behavior of policyholders over a long period of time and after consuming the product. By avoiding these limits we can increase the internal and external validity of our research findings.