Investigation of the Factors Influencing the Purchase of Personal Accident Insurance in Tier One Cities in China

Personal Accident Insurance (PAI) has played an important role in the live-lihood security of the residents in China, which is unlike the situation in most developed countries. But in recent years, it has been showing a significant reduction in the growth rate of the annual premium, and it is even −0.16% in 2020 compared to 2019. To study the decisive factors affecting the purchase of PAI, the willingness to purchase PAI and its influencing factors have been investigated based on the data collected from the residents of Shanghai, one of the most developed cities in China. A questionnaire was conducted to collect the data and the chi-square test of independence and histograms have been used to perform analysis. Our results indicate that PAI coverage in Shanghai is about 33%. The unwillingness to purchase PAI has been mainly ascribed to the residents’ weak awareness of the PAI, due to either subjective reasons or objective reasons. Measures have been suggested to change the declining situation.


Introduction
China has a large population in the world, accounting for about 18% of the world population. Along with strong economic development, its health issues have also attracted public attention, and China is marked as a country with an unbalanced economy and healthcare (Wang et al., 2019). Most developed countries have well-developed insurance systems. For example, the United States has three pri-How to cite this paper: Zhou, F. J. (2022). Investigation of the Factors Influencing the Purchase of Personal Accident Insurance in Tier One Cities in China. Modern Economy,13,[519][520][521][522][523][524][525][526][527][528][529][530][531] mary insurances, including health insurance, critical illness insurance, and disability insurance, that covers almost all the loss in the event of a potential accident. PAI is generally a niche market in these countries, and seniors and travelers are the main consumer groups (Anthem, 2018).
In contrast, China has a different situation. In principle, the current insurance system can provide almost all the insurances covering all kinds of potential loss.
Chinese citizens are required to purchase one of the three public insurance programs, including UEBMI (Urban Employee Basic Medical Insurance), URBMI (Urban Resident Basic Medical Insurance), and NRCMS (New Rural Cooperative Medical Care Scheme), based on the type of their registered residence.
But despite the seemingly well-developed insurance system, there exist a lot of problems with low insurance coverage being the most prominent one. First, there are about 300 million migrant workers in China, whose population is approximately equal to that of the whole United States. Second, the take-out industry in China is booming, forming a large part-time occupational group. These occupations are at high risk for accidents because of the imperfect traffic rules and systems, and unsafe driving behavior (Zhang et al., 2006). Third, although the state mandates formal contracts between employers and employees, many companies do not execute them, which causes these wage earners to not even have basic health insurance. Finally, though many higher income groups have health insurance, they make a lot of business travel, domestically and internationally, and the accidental risks can be much higher. While they do have health insurance to cover the loss of potential accidents, the insurance coverage usually has a low cap which can be a fatal blow to the family when faced with severe accidents (Song, 2011;Wan & Xiao, 2018). Thus, PAI is very important in China due to its deficiencies in social development and the healthcare system. However, based on the data from Xinhua News Agency and Beijing Intelligence Research Group (Xinhua News Agency, 2021), the PAI depth of China in 2020 is only about 0.116% (the total PAI premium income/total GDP = ¥117.41 billion/¥101,356.7 billion). Worse still, the willingness for purchasing PAI has been declining for the past three years with a 0.16% reduction in 2020 (Industry Information Network, 2021). Some research has discussed the status quo and deficiencies of PAI in China. Qiao studied the PAI depth in 2018 of 0.076% compared to the life insurance depth of 2.283% in Tianjin-a centrally administered municipality in China (Qiao, 2020). This data implies that, besides the low insurance coverage, there also exists some imbalance in accident insurance among different areas and social groups.
Some studies have been carried out to analyze the reasons for the low willingness to have PAI. For example, Niu and Sheng (2014) ascribed those to the problems with accident insurance products, the lack of professional staff, and the imperfection of relevance. Xiang and Li (2014) (Statista, 2021). Thanks to its huge economic volume and urban mass, this city has been the destination of many immigrant workers, constructors, and tourists. The evaluation of Shanghai citizens' accident insurance can be a good reference to understand the status and needs of PAI for citizens not only in China but also in some other similar developing countries.
This paper is organized as follows. Section 2 describes the methods used for the data collection and analysis and the variable measurement. Section 3 introduces the results of the survey conducted and the chi-square Tests of Independence employed. Section 4 discusses the reasons for the unwillingness to purchase PAI and suggests some measures to change the declining situation. Finally, Section 5 presents our conclusions.

Data Collection Methods
The data collection took place in Shanghai, China. The convenience sampling method was used due to the limit of sources. The target population was all adults living in Shanghai. Responders from various groups were collected, including people of different ages, marital status, education level, registered residence type, etc.

Variable Measure Methods
The survey was conducted through the online platform Wenjuanxing (or "Wen

Data Analysis Methods
Traditionally, the data are collected by picking options on a five (or seven) point Likert ranking scale, which typically uses the following words: strongly agree, agree, neutral, disagree, and strongly disagree. However, the questionnaire has purposely been designed without the neutral option because too many citizens tend to choose the neutral one possibly to avoid any trouble, which may lead to biased results. For graphs in the data section, histograms are generated on Excel's built-in mapping software and are used to present the data of categorical variables. Chi-square Tests of Independence have been used to examine the relationship of categorical variables to the purchasing of PAI.

Results
Finally, the Cochran formula has been used to determine the sample size as follows: where e is the desired level of precision. It is 7.5% when we take a 92.5% confidence; p is the estimated proportion of the population that has the attribute in question. We don't have much information on the subject to begin with, so we're going to assume that 30% of the citizens have PAI based on our preliminary knowledge. This gives us a variability of p = 0.3; q is 1 -p. So we get n = 143. Actually, there have been a total of 175 responses received, in which none has been voided (see Table 1).
As presented in Table 1, the data was collected from a wide range of Shanghai citizens, with all targeted categories, including age, marital status, registered residence types, gender, employment status, education level, and personal annual income. Each category has at least five responders responded. The variables collected are basic information of Shanghai citizens that help determine if the samples are reliable. Besides, they are possibly related to the willingness to purchase PAI (see Figure 1), which will be discussed later.
The age options are based on three age groups: young, middle-aged, and old. The options of personal annual income are based on the statistic average annual income of $10,825 (converted to US dollars according to the Renminbi U.S. dollar exchange rate of 6.5) of Shanghai citizens in 2020. It shows that a small fraction of the responders (32.57%) has PAI. At the same time, the difference factor plays a great role in the purchase of the PAI. Another interesting point is that some factors have response rates not proportional to their population.
To analyze the critical factors affecting the purchase of PAI, the chi-square Tests of Independence were further employed. The data will be examined at the following three levels: personal status factors, socio-cultural influence, and external factors, respectively.

Effect of Personal Status Factors on the Purchase of PAI
where the significance level of 0.05 is used. The chi-square test of independence only applies to categorical variables. Therefore, the age and personal annual income are grouped to form categories like the options in the questionnaire.
The importance of each factor in personal status is evaluated by comparing the p-value with the significance level. The factor is regarded as significant if the p-value is smaller than 0.05. The smaller the p-value is, the more significant the corresponding factor is. It can be seen that age, employment status, and education level show significant relation to the purchase of PAI, while gender, marital status, registered residence type, and personal annual income don't. The responder has a greater chance of purchasing PAI if he or she is older than 50 years old, retired, or with a degree below bachelor's.
It shows that older people are more likely to purchase PAI than younger people. On contrary, the younger group shows less interest despite having more social activities and thus a high chance to get hurt. This is well according to the observation in the developed countries mostly because they are relatively vulnerable to accidents. To our surprise, people with higher education have lower rates of PAI purchases, as is clearly displayed in Figure 1. Only about a quarter of the citizens with a higher degree shows they have PAI while the high school diploma or below is with 60%. This result may be explained by several reasons which will be further discussed in the later section.

Reasons Analysis of Low PAI Coverage
Based on the questionnaire, a great fraction of responders (about 67.43%) claim that they don't have PAI. It is necessary to further analyze the reasons for this lower coverage of PAI. In the questionnaire, a question has been further designed to ask responders to pick all possible reasons from a deliberate list of why not to purchase PAI. Since people by nature are social creatures and they like to socialize with others, their society and culture can have a huge influence on the purchase decision of PAI. These options include citizens' subjective impressions of the accident insurance to somewhat objective conditions, f. e. the pressure to pay the premium. Also, the option "other" was provided to the responders for additional reasons. The results were summarized and shown in Table 3, from the least important to the most important. For a more distinctive comparison, the reasons percentage was also depicted in Figure 2.  According to Table 3, more than half of the responses (64.41%) were especially unsatisfied with the claim settlement which is far more pronounced than the other factors. This is not difficult to understand: the clients purchase PAI to lessen the loss of uncertainty; their confidence gets most hit if there is some problem with the claim settlement. The claim settlement issue has been further ascribed to the complex claim procedure and bureaucracy which intimidate the potential purchasers. It is no surprise since most of the insurance companies in China are state-owned companies, notorious for their low efficiency and bureaucracy. Another with the claim settlement is the false or exaggerated message and spreading resulting from a repudiation of claims.
Another factor, accounting for 29.66% of responses as shown in the second last line, is the bad impression of PAC as a hoax. The responders claimed that the insurance representatives, often under high-performance pressure, are eager to reach a contract and tend to exaggerate the loss coverage or sometimes deliberately mislead it. These false or misleading messages together with representatives' unprofessional performance have led to a poor public's impression of PAI.
About 25.42% of responders believe that the actual dangers and accidents are of little chance to happen to them. This is a cognitive issue, which reflects that Shanghai citizens' insurance consciousness is not strong for a group of people. These responders show overconfidence or blindly confidence in uncertainty. Having accident insurance will bring bad luck also appears to be a relatively small but noticeable reason, accounting for about 12.71% of the response. This reflects that some old or traditional concepts still play a role in their decision.

The Prime Determinant of Low PAI Coverage
Two findings so far have been brought to our attention. The first one is that unsatisfied settlement has been revealed to have a striking effect on purchasing PAI. Insurance companies sometimes do violate the filing clauses and rates requirements due to business pressure or poor-quality control (Zhao & Ren, 2021), especially years ago. But more often today, this happens when there is a big gap between the coverage imagined by the clients and the coverage identified by the insurance company. Most clients are not aware of it until initiating a claim, and then they get very disappointed at being rejected and feel trapped by the insurance company. When they tend to complain about this to the persons around them, it finally becomes the butterfly effect. The situation has been worsening in recent years since more people release their anger online. At the same time, the traditional media and social media, to attract the public attention, tend to make use of and exaggerate the case, leading to a misleading hot topic. These false or exaggerated news and stories have largely affected Chinese people's perception of the personal insurance industry in a negative way and created a bad atmosphere. The second is that, contrary to intuition, citizens with lower education have a relatively higher coverage of PAI. This might be explained for several reasons. For example, people with lower education are usually in occupations with higher risk, which motivates them more likely to have PAI to cover poten-tial accident loss. Contrarily, people of higher education tend to ignore the potential accident loss due to the nature of work and the better health insurance provided by their employers.
Based on the findings above, we conjectured that residents' awareness of PAI could be a critical factor relating to the willingness of purchasing PAI. To elucidate it, we will check in the following section if there is any correlation between PAI awareness and PAI purchase. Table 4 shows statistics of the familiarity with PAI among responders. As presented in Table 4, 79.43% claimed that they are unfamiliar with the current accident insurance policy. In contrast, only 20.57% show somewhat or higher familiarity with PAI, of which approximately 3% are very familiar. This value corresponds well with 32.57% PAI coverage among responders.
In Figure 3, we show the statistic of satisfaction with the PAI system from responders who didn't buy any PAI. About two-thirds (66.49%) were not satisfied with the current PAI system. Based on Table 3 and its analysis, it is a reasonable deduction that most unsatisfaction comes from what they heard from the news, circulating messages, or even first impression since they haven't had a client relationship with the insurance company. It should also be mentioned that this 69.49% dissatisfaction is not far from the unfamiliarity rate of 79.43%. In contrast, the somewhat and above satisfaction rate (30.50%) matches well with the familiarity rate of 20.57% and PAI coverage of 32.57%. In other words, the low familiarity rate of 20.57% generates relatively higher satisfaction of 30.50% and a much higher PAI coverage of 32.57%.
Based on these results above, familiarity seems to have a great influence on the purchasing willingness of Shanghai citizens. To further verify this, the chi-square Test of Independence was used to test the relationship between the purchase of PAI and familiarity with it. A significance level of 0.05 was adopted.
The results of the relationship between the purchase of PAI and the familiarity with it by the chi-square Test of Independence are shown in Table 5. The values in columns 3 & 4 are explained as follows: the observed cell totals, (the expected cell totals) and [the chi-square statistic for each cell]. The chi-square statistic thus got is 11.2756, with the p-value being 0.0103. This p-value of 0.0103 is significantly smaller than the significance level of 0.05, implying that there is a firm correlation between familiarity of PAI and the purchase of PAI. The more familiar with PAI the residents are, the more willing to purchase PAI they are.

Discussion
Based on the findings above, it shows that familiarity (or more precisely awareness) of PAI plays a critical role in deciding the purchase willingness. From a broad perspective, this familiarity or awareness includes the following: 1) the awareness of insurance importance, 2) correct understanding of PAI without prejudice or being misled, and 3) related basic knowledge about PAI.
One of the reasons for this unfamiliarity is that the learning system in Shanghai does not effectively provide its citizen with enough insurance-related know- Republic of China, 2018). In October 2021, the agency issued the "Accident Insurance Business Regulation" to phase out accident insurance products with too low payout rates (CCTV Finance, 2021).
In addition, to ensure that insurance industry personnel handling such procedures have sufficient experience, it's important to strengthen the communica-tion between insurance education departments and insurance regulatory departments, so that insurance regulatory departments can provide a good environment for the development of insurance education in terms of laws and policies (Xing, 2009). Finally, it has to be mentioned that this study has some limitations. First, the convenience sampling method was used to collect data which may lead to bias in results. Second, the limited number of samples due to limited funding and resources may cause bias. Third, the questionnaire has not been equally reached the target groups, like imprisoners, old people et al. Finally, some target groups are unlikely to take the survey due to overwork, privacy concern et al.

Conclusion
In this study, a survey has been conducted on Shanghai residents' overall possessing PAI and the factors affecting their decisions. Based on the results, it's found that the PAI coverage is about 33%, a relatively low value when considering the actual necessity. Further analysis on the reasons that hinder citizens from purchasing PAI has been carried out. It is revealed that the top two reasons are the claim settlement and the bad impression on PAI, accounting for 64.41% and 29.66% respectively. The underlining determinant factor of the lower PAI coverage was studied in detail using the chi-square test of independence. The significant small p-value of 0.0103 provides convincing proof that the low PAI is mainly due to the weak awareness of PAI. In other words, increasing citizens' awareness can effectively improve the chance to have the necessary PAI.
To ameliorate such a situation, a series of measures have been suggested to increase the awareness of insurance importance, correct understanding of PAI, and related basic knowledge about PAI. In addition, authorities should be responsible for supervision in a legal sense for this purpose, and at the same time, promote citizens' protection of their rights.