The Impact of Consumer Animosity on Purchase Unwillingness in a Boycott of Sari Roti

This research was conducted in response to a boycott towards a national bread brand. It was stimulated by a company disclosure in stating they had no relation to any political events in 2016. This research aimed to examine the impact of consumer animosity on product judgment, purchase willingness, purchase unwillingness, and boycott participation. In total, 266 participants took part in this research using convenience sampling. Data were collected using an online survey. The data were analyzed using exploratory factor analysis and Structural Equation Model (SEM). This research finds that there is a significant impact of animosity on product judgment, purchase willingness, purchase unwillingness, and boycott participation. Moreover, there is also a significant impact of product judgment on purchase willingness and boycott participation on purchase unwillingness.


INTRODUCTION
Sari Roti, a national bread brand, has been boycotted by a group of Indonesian Muslims.As the country with the largest Muslim population in the world, a boycott movement by Muslim consumers should be regarded as a serious threat.It also should threaten the company, as it is proven to lower their sales numbers (Luluk, 2016).The boycott was triggered by a disclosure issued by the producer of this bread brand by saying that the company had nothing to do with any political events.Before the event of the boycott movement towards the bread company, there were three events of demonstration conducted by Muslims in Jakarta in 2016.These events were called as '1410' referring to October 14 th , '411' for November 4 th , and '212' for December 2 nd .These actions were done to put pressure on the government to punish the Jakarta governor (at that time) who was accused of blasphemy.After the third protest, the boycott movement was executed.
Some researchers have explored consumer boycott towards foreign products (Barutçu, Saritaş, & Adigüzel, 2016;Klein, Ettenson, & Morris, 1998;Palihawadana, Robson, & Hultman, 2016).Meanwhile, the other researchers focus on consumer boycott towards domestic products (Braunsberger & Buckler, 2011;Friedman, 1996).The existing researches by Ahmed et al. (2013), De Nisco et al. (2013), Fernández-Ferrín et al. (2015), Ben Mrad, Mangleburg, and Mullen (2014), Rose, Rose, and Shoham (2009), and Shah and Ibrahim (2016) showed a significant effect of animosity on attitude, purchase action, purchase willingness, ethnocentrism, and product judgment.In this research, animosity is linked to product judgment, boycott participation, purchase willingness, and purchase unwillingness.Abosag and Farah (2014) used consumer ethnocentrism, religious animosity, boycott participation, brand image, and product judgment to predict consumers' loyalty.They found that boycott participation was positively impacted by ethnocentrism and religious animosity.Meanwhile, Rose et al. (2009) focused on purchase unwillingness from Arab Israeli and Jewish Israeli consumers towards foreign products by Italians and British.They linked animosity on product judgment and purchase unwillingness, product judgment, and ethnocentrism.They distributed their research instruments in northern of Israel where both communities existed.They stated that purchase unwillingness was positively affected by ethnocentrism and animosity.However, it was negatively affected by ethnocentrism and product judgment on both Italian and British products.They also found that product judgment was significantly and negatively influenced by animosity in the case of British products.Then, it was insignificantly influenced by animosity in Italian products.
Similarly, Nijssen et al. (1999) evaluated Dutch consumers' purchase willingness on foreign products.They included several variables such as consumer ethnocentrism, product judgment, interest in foreign travel, and perceived domestic product availability.One of the findings stated that animosity negatively affected purchase willingness.Based on the prior research discussed, the hypotheses used are as follows.
H 1 = Animosity will have a negative impact on product judgment H 2 = Animosity will have a negative impact on purchase willingness H 3 = Animosity will have a positive impact on purchase unwillingness H 4 = Animosity will have a positive impact on boycott participation Moreover, the previous research by Ahmed et al. (2013), Albayati et al. (2012), Ben Mrad et al. (2014), Rose et al. (2009), and Shah and Ibrahim (2016) documented that product judgment affected purchase willingness, purchase unwillingness, boycott participation, and purchase action.In this research, product judgment is linked to purchase willingness and purchase unwillingness.Klein et al. (1998) conducted a research to predict product ownership.They employed consumer ethnocentrism, animosity, product judgment, and purchase willingness.They showed that animosity had an insignificant impact on product judgment.Moreover, there was a negative and significant impact of animosity on purchase willingness.They also mentioned that product judgment had a positive impact on purchase willingness.
Meanwhile, Huang, Phau, and Lin (2010) investigated purchase intention of Taiwan consumers on Chinese and Japanese products.They measured the impact of consumer animosity on quality judgment and purchase intention, and quality judgment on purchase intention.They found a significant impact of animosity on quality judgment and purchase intention, and quality judgment on purchase intention.
Another research in purchase willingness was conducted by Shoham et al. (2006).They focused on Jewish Israeli' reaction towards Arab Israeli' intifada (uprising) action.Animosity was linked to product judgment and purchase willingness.Meanwhile, product judgment was linked to purchase willingness.
As a result, they showed all paths were significant.Therefore, the other hypotheses used are: H 5 = Product judgment will have a positive impact on purchase willingness H 6 = Product judgment will have a negative impact on purchase unwillingness Boycott participation is reported to have a significant impact on purchase willingness, purchase unwillingness, and product judgment (Albrecht et al., 2013;Shah & Ibrahim, 2016).In this research, boycott participation is linked to purchase unwillingness.Albrecht et al. (2013) chose an international soft drink brand that was accused of infringing upon environmental protection policies in India.They found that there was a significant effect of intention in boycott participation on refusal to buy a boycotted brand.Thus, the next hypothesis is: H 7 = Boycott participation will have a positive impact on purchase unwillingness.
Overall, there are seven paths to be examined as illustrated in Figure 1.In this proposed research model, animosity is linked to product judgment, purchase willingness, purchase unwillingness, and boycott participation.In addition, product judgment is linked to purchase willingness and purchase unwillingness.Lastly, boycott participation is linked to purchase unwillingness.
This research raises the case of Sari Roti boycotted by its consumers.The company has never thought that by saying that it has no part in any political activities will be bad for the business.On the other hand, Sari Roti is the first and only bread brand company to be massively marketed nationally.This research aims to measure factors influencing consumers' willingness and unwillingness to purchase Sari Roti, a national-brand bread.To test these dependent variables, the researcher uses animosity, product judgment, and boycott participation as the predicting variables.
Figure 1 The Theoretical Framework

METHODS
Participants of this research are selected conveniently and asked to fill out an online selfadministered instrument.The link of the instrument is distributed personally using WhatsApp, Telegram, and Line applications.
To measure all tested variables, the researcher adapts indicators taken from previous research.Indicators of consumer animosity are adapted from Jin and Furukawa (2006).Then, the indicators of product judgment are adapted from Nakos and Hajidimitriou (2007), and Nijssen et al. (1999).Furthermore, indicators from Altintas et al. (2013) are adapted to measure boycott participation.From Nakos and Hajidimitriou (2007), and Tian (2010), the researcher uses the variable to measure purchase willingness and purchase unwillingness respectively.
The data collected are analyzed in three main stages.The first stage is Exploratory Factor Analysis (EFA) to seek dimensions and valid indicators of each variable.EFA is used to validate the data (Allen & Bennett, 2010).The second stage is reliability test.According to Hair et al. (2006), a construct should be considered reliable if it has a Cronbach's alpha score of 0,7 or more.For these purposes, the SPSS version 22 is used.
The third stage is Confirmatory Factor Analysis (CFA).In this stage, the researcher uses AMOS version 22.It is to reduce and confirm dimensions and indicators.Furthermore, to test the hypotheses, AMOS is still used.Structural Equation Model (SEM) is used to achieve fitness.A model is considered fitted if it has a probability (p) score of >0,05, CMIN/DF score is ≤2,00, CFI score is >0,97, and RMSEA score is <0,05 (Bentler, 1990;Browne & Cudeck, 1992;Schermelleh-Engel, Moosbrugger, & Müller, 2003;Tabachnick & Fidell, 2007).Additionally, a hypothesis is considered accepted if it has a Critical Ratio (CR) score of 1,96 or greater (Hair et al., 2006).

RESULTS AND DISCUSSIONS
This research attracts 306 respondents.However, only 266 respondents complete the online questionnaire with 167 females (62,8%) and 99 males (37,2%).Regarding the age, 191 respondents (71,8%) are between 20-24 years old, and 57 respondents are 19 years old and younger.The rest of the respondents are 30 years old and older.In term of the level of education they have completed, 183 respondents (68,8%) have a high school certificate, 68 respondents (25,6%) have an under-grad certificate, and the remaining respondents have a certificate of post-grad and less than high school.Majority of the respondents (247 respondent or 92,9%) are single and have jobs or still study in a university (137 or 51%).Some of them are studying while working (84 respondent or 31,6%) and the others have jobs (17 respondent or 10,2%).When they are asked about their domicile, predominant respondents indicate that they live in Jakarta (198 or 74,4%).The participants' profiles can be seen in Table 1.In the period between November and December 2016, there were three demonstrations held in Jakarta.The first two were dominated by Muslims, and the last one was varied.Respondents were asked to identify which demonstration they participated in.About 40 respondents (15%) participate in the demonstration held on November 4 th , 63 respondents (23,7%) participates in the demonstration of December 2 nd , and 11 respondents (4,1%) are in the demonstration of October 14 th .Furthermore, 75 respondents (28,2%) do not participate in any of the demonstration.
The sample of this research does not target Muslim consumers in particular.However, the predominant respondents are Muslim (224 respondents or 84,2%).In addition, 261 respondents claim that they have experience of purchasing Sari Roti within the last six months.In detail, 18 respondents (6,8%) purchased the products today, 79 respondents (29,7%) purchased within the last week, 66 respondents (24,8%) purchased within the last month, 102 respondents (38,3%) forgot about the exact time.Then, one respondent (0,4%) has never purchased at all.The religion and experience relating to Sari Roti can be seen in Table 2.About 200 (75,2%) respondents disagree to the boycott against Sari Roti.Furthermore, 158 (59%) respondents consider that the boycott tends to be a political action.Meanwhile, 110 (41,4%) believe that it is a religious action.110 (41%) respondents think that this is a combination of political and religious motives.Next, four items of animosity have the Cronbach's alpha score of 0,910 and factor loadings ranges from 0,865 to 0,919.The result is in Table 3.Two dimensions of boycott participation are developed.The first dimension has a Cronbach's alpha score of 0,949.The factor loadings range from 0,793 to 0,929.The second dimension has a Cronbach's alpha score of 0,471 with factor loading ranging from 0,556 to 0,838.Due to its reliability, the second dimension is dropped for the full model analysis.The result can be seen in Table 4. Product judgment owns five items with a Cronbach's alpha score of 0,0921.The factor loadings are from 0,837 to 0,920.The result is shown in Table 5.
The six items of willingness to purchase are retained and grouped under two dimensions.Firstly, willingness dimension has a Cronbach's alpha score of 0,883 with factor loadings ranging from 0,798 to 0,883.Secondly, purchase unwillingness has a Cronbach's alpha score of 0,701 with factor loadings ranging from 0,775 to 0,912.In further analysis, these two dimensions are treated as two different variables.The result is in Table 6.According to the SEM calculation, it is suggested to link brand trust on product judgment for obtaining fitness.However, in the theoretical framework, it does not exist.This model has a probability score of 0,053, CMIN/DF score of 1,402, CFI score of 0,990, and RMSEA score of 0,039.It can be seen in Figure 2.
Table 7 shows the result of SEM.Five of seven hypotheses tested are accepted.It includes H 1 , H 2 , H 3 , H 4 , H 5 , and H 7 .These paths have greater CR score than 1,96 as required (Hair et al., 2006).However, the other (H 6 ) hypothesis is rejected.
Consumer animosity can be very dangerous for companies as it causes many things.In this research, animosity predicts four other variables like product judgment, purchase willingness, purchase unwillingness, and boycott participation.All predictions are significant.
The first hypothesis predicts the impact of animosity on product judgment.This path has a CR score of -7,057 that indicates significance.Thus, H 1 is accepted.This finding supports the researches by Huang et al. (2010), Rose et al. (2009), Shoham et al. (2006).
The second hypothesis has CR score of -2,703.It has a negative and significant impact of animosity on purchase willingness.Therefore, H 2 is accepted.This finding is significant with research conducted by Nijssen et al. (1999) and Shoham et al. (2006).
The third hypothesis is regarding the impact of animosity on purchase unwillingness.The SEM calculation results in CR score of 2,338.It is considered significant, so H 3 is accepted.This finding supports the research of Rose et al. (2009).
The fourth hypothesis predicts the impact of animosity on boycott participation.The path has CR score of 11,979.This score is considered the highest score among other paths.The more consumers hate the brand, the more they will get involved in boycotting.Thus, H 4 is accepted.The finding is in line with Abosag and Farah (2014).
The fifth hypothesis is about the product judgment on purchase willingness.The path has CR score of 7,719.Thus, H 5 is accepted.The finding supports the researches by Huang et al. (2010) and Klein et al. (1998).
With CR score of -0,685, the sixth hypothesis (H 6 ) is rejected.This hypothesis predicts the impact of product judgment on purchase unwillingness.In this research, product judgment is measured by indicators related to good aspect of raw materials, value for money, trustworthy, durability, and good quality (Nakos & Hajidimitriou, 2007).By these indicators, the expected result is in a negative direction, but the score is too low to be significant.Therefore, the finding is insignificant with previous research by Rose et al. (2009).
Last, the seventh hypothesis is regarding the impact of boycott participation on purchase unwillingness.The path achieves CR score of 4,278.Therefore, H 7 is accepted.This finding supports the research of Albrecht et al. (2013).Consumers who participated in boycotting Sari Roti will easily be predicted to have no intention to purchase the product.

CONCLUSIONS
This research aims to examine factors that influence purchase willingness and unwillingness of a bread brand (Sari Roti) boycotted by Muslim consumers.The boycott was in response to a disclosure issued by the company mentioning that it had no relationship with several political demonstration occurred in Jakarta in 2016.This research includes variables of animosity, product judgment, and boycott participation.
This research finds a significant impact of consumer animosity on product judgment, boycott participation, and purchase willingness and unwillingness.Furthermore, product judgment has a significant impact on purchase willingness.Boycott participation has a significant effect on purchase unwillingness.Then, animosity is indeed very dangerous that it is, directly and indirectly, affecting purchase willingness.
This research has a significant implication for businesses.Religion and politics are two unseparated elements.Playing with these two intentionally or unintentionally will give a boomerang effect to the company.Many things will occur irrationally.Depending on the product judgment, the brand will obtain stability after some time.Looking at the case of Sari Roti, in fact, at the end of the year, the company reports that its sale has increased more than before (Jatmiko, 2017).
Further study may include attitude of consumers towards company disclosure.The disclosure is considered to stimulate the boycott.Besides, it will be important to include those who participate in the demonstration as a sample for this research.Additionally, it will be intriguing to look at the loyalty of Sari Roti customers who purchase the bread in regular basis.

Table 1
Profile of Participants

Table 3 EFA
Result of Animosity

Table 4 EFA
Result of Boycott Participation

Table 5 EFA
Results of Product Judgment

Table 6 EFA
Result of Purchase Willingness

Table 7
Summary Result of SEM