“Analysis of the intention to use loan facilities from formal financial institutions for large chili farmers through the theory of planned behavior approach”

This study analyzes and examines the factors influencing the intention to use loan facilities from formal financial institutions. This study was carried out quantitatively using the Structural Equation Modeling (SEM) method with the population of large chili farmers in 19 districts in Jember Regency as the region with the best large chili productivity in Indonesia. The results obtained from 116 respondents indicate that the construct of situational temptation does not affect subjective norms. In addition, formal financial institution financing positively affects attitudes toward behavior and subjective norm on intention to use loan facilities, but not on the perceived behavior. Meanwhile, the intention was positively influenced by attitude, subjective norm, and perceived behavior control. The factor with the most significant effect on the intention is perceived behavior control (t-statistics = 4.940). This shows that control of perceived behavior towards intention to use loan facilities from formal financial institutions has the most significant influence among other variables.


INTRODUCTION
Ensuring food security is a major concern for people worldwide (Prosekov & Ivanova, 2018).To address this challenge, it is anticipated that the agricultural industry in Indonesia will play a substantial role in augmenting the availability of food production.The agricultural sector is Indonesia's primary sector, given the agrarian order in its society (Garcia, 2000).Chili plants were included in the top five levels of production of seasonal vegetable commodities in 2021 (Dian & Yotenka, 2022).However, it seems that the price of chili is still fluctuating.Chili is a commodity that contributes to inflation each year (Sativa et al., 2017).This is possibly due to missed harvest schedules; in 2019, the price of cayenne pepper reached Rp85,000/kg, while large curly chilies, usually around Rp13,000/ kg, became Rp58,400/kg (Antaranews, 2019).Even so, prices often drop, which is detrimental to chili farmers, reaching prices down to Rp6,000/kg (Blorakab, 2019).Firdaus et al. (2020) claimed that the fluctuation of chili prices is due to the unequal distribution of chili production throughout the year and the uneven distribution of chili-producing areas in Indonesia.

LITERATURE REVIEW AND HYPOTHESES
This study used the Theory of Planned Behavior (TPB) to evaluate the intention to use loan facilities from formal financial institutions.TPB is an extension of the Theory of Reasoned Action (TRA), which holds that an individual's conduct is determined by his or her intention to perform or refrain from performing a particular activity (Manstead & Parker, 1995).Ajzen (1991) contributed a concept to TRA that did not previously exist: perceived behavioral control.This component is added to comprehend humans' restrictions when executing specific behaviors (Chau & Hu, 2001).In other words, whether an action is carried out or not is governed not just by subjective attitudes and standards but also by the individual's sense of the control he/she can exercise, which stems from his/ her assumption that he/she can wield such control (control beliefs).
Attitude is defined by Fishbein and Ajzen (1977) as the sum of the effects (feelings) a person experiences when accepting or rejecting an object or behavior.The measurement of attitude involves a method that positions the person on a binary evaluative continuum, which contrasts positive and negative or agreement and disagreement (Höög, 2005).According to Ajzen (2002), attitude is an internal state that influences an individual's action toward objects, people, or occurrences.Attitudes are cognitive, affective, and acquired behavioral tendencies to respond favorably or unfavorably to objects, situations, institutions, concepts, or individuals (Eagly & Chaiken, 2007).
The construct of subjective norms refers to an individual's perception of the social pressure or expectation from relevant others, which can shape their inclination to engage in or avoid certain behaviors (Jogiyanto, 2008).Subjective norms refer to the degree to which an individual is inclined to conform to the opinions of others regarding a particular behavior, which is based on normative beliefs.However, if an individual perceives that they have autonomy in determining their actions, regardless of external influences, they may disregard others' views regarding the behavior in question.The term "motivation to comply" was coined by Fishbein and Ajzen (1977) to characterize whether an individual adheres to the views of other influential people in his life.Schifter and Ajzen (1985) added a concept to TRA that had not been included previously, namely perceived behavioral control.This construct was added to the TPB to regulate behavior constrained by a scarcity of resources to execute the behavior.Perceived behavioral control is the new variable.The perceived simplicity or difficulty of executing a behavior characterizes perceived behavioral control (Ajzen, 1991).This condition is referred to by Ajzen (2005) as perceived behavioral control.Numerous variables can disrupt the connection between intention and behavior (Sheeran, 2002).The success of a person's performance and behavior depends on his or her ability to control the factors that can influence behavior, although volitional control influences behavior more than other factors, personal limitations, and obstacles (Ajzen, 1985).External obstacles may also hinder behavior performance (Ajzen, 1991).Ajzen (1991) proposed the notion of perceived behavioral control as a theoretical framework to address the problem of limited volitional control.
Meanwhile, the intention is the desire to conduct a specific behavior (Malle et al., 2001).Intentions are not permanently fixed; they can evolve (Cohen & Levesque, 1990).The longer the interval, the greater the likelihood of intention changes (Jogiyanto, 2008).TPB states that a person can only act following his/her intent if he/she controls his/her behavior (Ajzen, 2002).The intention is a function of two fundamental determinants, the first of which relates to personal factors and attitudes toward behavior, and the second relates to social influence and subjective norms (Ajzen, 2005).TPB explains that an individual can act based on his/ her intentions if he/she controls his/her behavior (Ajzen, 2002).In addition, perceived behavioral control has motivational implications for intentions (Ajzen, 2002).
Several studies have examined the relationship between the previously described TPB construct and finance, particularly regarding business capital or investment.Dewi (2018)  H1: Situational temptation affects subjective norms on the intention to use loan facilities from formal financial institutions.
H2a: Financing of formal financial institutions affects attitude toward behavior on the intention to use loan facility.
H2b: Financing of formal financial institutions affects subjective norms on the intention to use loan facility.
H2c: Financing of formal financial institutions affects perceived behavior on the intention to use loan facility.
H3: Attitude affects intention to use loan facilities from formal financial institutions.
H4: Subjective norm affects intention to use loan facility from formal financial institutions.
H5: Perceived behavior affects intention to use loan facilities from formal financial institutions.

METHOD
The type of this study is quantitative.The method used is a field research survey conducted on several samples from a specific population whose data was collected using a questionnaire (Sekaran, 2007).The unit of analysis in this study is large chili farmers in Jember Regency.In Structural Equation Modeling (SEM), the optimal sample size was determined using the recommendation of Stevens and Edwards (1996), who suggested that the sample size for maximum likelihood estimation should be no less than 15 times the number of observed variables.Based on this criteria, all large chili farmers in Jember Regency were taken as a sample with 116 respondents.
Respondents were asked to choose one of the quantitative answers provided according to their actual conditions.The rating scale represents quantitative information, which can be further analyzed and interpreted in a qualitative way.This scale offers more versatility, as it enables not only the assessment of attitudes but also the evaluation of participants' views on environmental phenomena.For instance, a rating scale can be employed to measure perceptions related to social status, economy, science and technology, agencies and institutions, customer satisfaction, work productivity, among other areas.In this study, the intervening factors analyzed were the attitude towards behavior, subjective norm, and perceived behavior control.The questionnaire filled in by the respondents was quantified first to produce numerical output, which was then analyzed through the SmartPLS ver 3 for Windows.To assess the soundness of the data obtained, the measuring instrument used was subjected to tests of validity and reliability.These tests aim to verify that the instrument accurately measure the intended variables and produce consistent results with the data collected (Kimberlin & Winterstein, 2008).
In this study, an alternative approach to covariance-based structural equation modeling (SEM) called Partial Least Square (PLS) was used for hypothesis testing (Ghozali, 2006).While covariance-based SEM tests causality or theory, PLS is more suited for predictive modeling (Lowry & Gaskin, 2014).This modeling approach allows for testing without a strong theoretical foundation, and assumptions and parameters related to model accuracy can be eliminated (Jogiyanto & Abdillah, 2009;Fornell & Yi, 1992).The PLS method also allows for the use of assumptions, such as non-normal data distribution, various measurement scales, small sample sizes, and non-reflexive indicators.These features make the PLS method a flexible and powerful tool for testing complex models that may not be well-suited to covariance-based SEM (Ghozali, 2006).PLS is an alternative method for a complex relationship among variables, such as in SEM, but the data sample size is small (30 to 100) (Wisz et al., 2008).
The t-test is employed to assess the statistical significance of the independent variables and the inter- Hypotheses equation model cept individually, as well as their impact on the dependent variable (Ghozali, 2006).The critical value for accepting or rejecting the research hypothesis is set at 1.64; if the calculated t-value is below 1.64, the null hypothesis is rejected (Ghozali, 2006).In this study, the alternative hypothesis is accepted at the 5% level of significance.The degree of influence of all the independent variables on the dependent variable is measured by the coefficient of determination, R-square (R 2 ), which ranges from 0 to 1, with a value closer to 1 indicating a greater percentage of the effect (Ghozali, 2006).
The model in this study is a structural, reflective research indicator.The analysis model in the equation is presented in Figure 1.
Based on the equation model, the following equation can be formulated:

RESULTS
This study examined the factors of situational temptation (ST) and financing of formal financial institutions (FFFI) with intervening variables, namely attitude (ATT), subjective norm (SNM), and perceived behavior (PBH) which affect intention (ITN) to use loan facilities from formal financial institutions in Jember Regency.The structural model depicted in Figure 2 is the results from the initial SmartPLS test.Based on the results of the loading factor, several indicators have a value below 0.5 which are declared invalid (see Table 1).Furthermore, the loading factor results with values < 0.5 will be removed from the model, and then the model will be re-estimated (Chin, 1998).This study declared 15 out of 35 or 42.86% of the indicators invalid.This is since the level of understanding of respondents to the questionnaire is very different from that of the respondents from the adopted questionnaire.The cultural background of the north coastal community of East Java (Pandalungan) which tends to be expressive, open, tough, and does not like small talk, is closely related to how they answer the questionnaire.The absence of the respondents' patience, considering that there are relatively many questions in the questionnaire, answering very quickly tends to be rushed, causing many questions to be invalid.The estimation result after eliminating the invalid indicator is shown in Figure 3.
Another approach to assessing discriminant validity involves comparing the square root of the average variance extracted (AVE) of each construct to the correlations between that construct and other constructs in the model (Ghozali, 2006).Discriminant validity is considered adequate if the AVE root value for each construct exceeds the correlation value between that construct and other constructs in the model (Ghozali, 2006), with an expected value of over 0.5.Table 3 displays that the overall reliability of all categories is greater than 0.7, which signifies that all categories in the predicted model meet the requirements for discriminant reliability (Ghozali, 2011).Table 4 presents the R-square value, which is the coefficient of determination for the endogenous construct.As per Ghozali (2011), an R-square value of 0.70 is considered to be strong, 0.50 is regarded as moderate, and 0.25 is considered to be weak.Therefore, the ITN construct is classified in the moderate category, and other constructs are included in the weak category.The procedure of testing hypotheses in SmartPLS relies on the indicators derived from the structural model (Sander & Teh, 2014).In addition, the importance of the path coefficient is determined by both the standardized path coefficient and the t-value (Huang et al., 2013).In order to determine whether the proposed hypothesis is supported or not, the researchers will employ a one-tailed t-test.
If the resulting t-statistic is higher than 1.64 and the standardized path coefficient (p) is greater than 0.05, then the hypothesis will be considered valid (Ghozali, 2006).Table 5 shows the nature of the correlation between constructs and the conclusion of the hypotheses.Based on the hypothesis analysis results, the result shows that the highest value that affects ITN in utilizing loan facilities from formal financial institutions is PBH, with a t-statistics of 4.940.This shows that the effect of PBH in responding to facilities influences the intention to use loan facilities from formal financial institutions higher than SNM or ATT constructs.

DISCUSSION
The hypothesis of situational temptation towards subjective norm on intention to use loan facilities from formal financial institutions is rejected.This is because farmers who plant chilies have already prepared the total cost of farming that will be used in their farming, from seeding and planting to harvesting.In addition, farmers are aware that if they do not prepare farming costs from the beginning, it will interfere with their subsequent farming process, and if this issue arises, their farming productivity will decrease.Chili farmers are businesspeople who always have a logical view.In the agribusiness industry, chili is the highest "caste" in horticulture, including financial readiness or capital.Therefore, they must prepare for the farming process thoroughly.
The normative beliefs and motivation to follow other people's opinions held by large chili farmers cannot directly explain and predict their intention to take loan facilities from formal financial institutions, especially People's Business Credit.Large chili farmers carefully consider alternatives to financing through loans, and some claim that they have never borrowed money from the bank for their farming capital.In addition, their rela-  2015), which asserted that perceived behavior leads young entrepreneurs to get loans from formal financial institutions.Based on the attitude point of view, financing will increase the ability and confidence of large chili in supporting their businesses.Financing from formal financial institutions needs document requirements and good business performance so that when a formal financial institution funds it, the business must be trusted, and with loan funds, the business will undoubtedly grow.This is what encourages a positive attitude from large chili farmers.Large chili farmers have compiled financial reports, calculated turnover, cost of goods, and a business plan.Based on these conditions, a confidence level in formal financial institutions will support the intention to use loan facilities (Mindra & Moya, 2017).
This study indicated that attitude positively affects the intention to use a loan facility.This is supported by research conducted by Pebrina et al. (2021), who used TPB to examine how e-commerce adoption by MSMEs in Banten was affected by attitude.According to the research, attitude toward behavior positively increases intention to use, meaning that the more positive the attitude toward behavior, the greater the intention to use (Pebrina et al., 2021).The result of the current study is also supported by the findings of Astuti and Prijanto (2021), who investigated the factors Influencing Muzaki's intention to pay zakat using TPB.According to the study, attitude significantly affects intention.The results of this study reveal that public interest in using zakat services is proportional to the attitude of zakat service customers.Nonetheless, research conducted by Chrismardani (2016), who also investigated TPB as a predictor of entrepreneurial intention, found that attitude did not positively affect entrepreneurial intention.(Ajzen, 1991).Thus, the results of this study indicate that chili farmers are more likely to register for a loan facility if they strongly believe in the benefits of using a loan facility.Perceived behavioral control is also shaped by the availability of opportunities (accessibility), facilities, and individual endeavors to carry out a behavior (Ajzen, 1991).
In addition, perceived behavioral control is also formed by self-efficacy, which is influenced by preparation, perspective, and the impediments encountered when executing a behavior (Ajzen, 1991).With the influence of perceived behavioral control on intentions, using loan facilities can be an excellent opportunity.This is since perceived behavioral control is the only indicator of intention that can directly influence behavior without passing through intention (Ajzen, 2005).Furthermore, the intention is also a factor in determining the occurrence of a behavior (Ajzen, 1991).

CONCLUSION
The results suggested that the situational temptation factor does not affect the subjective norm on the intention to use loan facilities from formal financial institutions.In addition, financing of formal financial institutions positively affects attitudes and subjective norms on the intention to use loan facilities.However, the financing of formal financial institutions does not affect perceived behavior on the intention to use loan facilities.The intention factor is positively influenced by attitude, subjective norm, and perceived behavior control.Based on the results, the highest value that affects intention is perceived behavior control, with a t-statistic value of 4.940.This indicates that perceived behavior control towards intention to use loan facilities from formal financial institutions has the most significant influence among other variables.
Apart from the results, this study has many limitations.The structural model formed is a development model with the theory of planned behavior, so the model can only be explanatory and cannot confirm the theory of the relationship between variables.In addition, this study used a purposive sampling technique provided by the information taken from a sample selected based on determined criteria, so the research depends on each respondent's answers.In the future, researchers may study ways to enhance people's control over their behavior when using loan services provided by formal financial institutions.This might entail investigating factors that influence such control, like knowledge of finance, availability of information on loan options, and the simplicity of applying for a loan.Additionally, more research could examine the effectiveness of measures that aim to boost perceived behavior control, such as programs that educate people on finance or streamline the loan application procedure.Another possible area for future inquiry could be identifying factors that lead to situations that encourage people to resist using formal financial institutions for loans and how to mitigate these factors so that more people would be inclined to use such services.

Figure 2 Figure 2 .
Figure2shows that eight indicators measure ST construct, while 12 measure FFFI.In addition, the ATT construct is measured by six indicators, SNM

Table 2
AVE values for the constructs, and the data shows that all constructs have √AVE values greater than 0.5.Furthermore, the lowest AVE value is 0.530 for the SNM construct.
presents the Figure 3. Outer loading results after eliminating several indicators