提升意外保險部門之營運效率研究 - 政大學術集成

全文

(1)國立政治大學商學院國際經營管理英語碩士學位學程 International MBA Program College of Commerce National Chengchi University. 碩士論文. 學. ‧ 國. 立. 政治大 Master’s Thesis. ‧ sit. y. Nat. er. io. 提升意外保險部門之營運效率研究 n. a. v. l C IMPROVING OPERATIONAL EFFICIENCY OF A ni. hengchi U. CASUALTY INSURANCE DEPARTMENT. Student: Stewart Cho Advisor: Professor Jason Cheng-Hsien Tsai. 中華民國一○九年六月 June 2020. DOI:10.6814/NCCU202000750.

(2) 提升意外保險部門之營運效率研究 IMPROVING OPERATIONAL EFFICIENCY OF A CASUALTY INSURANCE DEPARTMENT. 研究生：卓尚緯. Student: Stewart Cho. 指導教授：蔡政憲. Advisor: Professor Jason Cheng-Hsien Tsai. 學. ‧ 國. 立. 政治大國立政治大學. ‧. 商學院國際經營管理英語碩士學位學程碩士論文. er. io. sit. y. Nat. A Thesis. n. a to International MBAi Program Submitted v l C n U h eChengchi National University ngchi. in partial fulfillment of the Requirements for the degree of Master in Business Administration. 中華民國一○九年六月 June 2020. DOI:10.6814/NCCU202000750.

(3) Acknowledgements I am very grateful to many people that have contributed in direct or indirect ways to this dissertation. First, I would like to acknowledge the support of the International MBA program (IMBA) at National Chengchi University (NCCU), and in particular, Professor Jason ChengHsien Tsai, Professor Wei-Kuang Chen, and Professor Tzu-Ting Lin for providing suggestions for me to finish this thesis. I would also like to thank all interview participants in this research who with their support and contribution made the groundwork for this dissertation possible.. 政治大 Many thanks are also owed to Li-Chi Ho and Yih-Chieh Wu for their assistance and patience. 立. ‧ 國. 學. Finally, my deepest gratitude goes to those whose influence may not be so evident but who nevertheless provided the educational and emotional foundation without which this dissertation. ‧. would have never been possible: my parents and my younger brother.. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. i DOI:10.6814/NCCU202000750.

(4) Abstract Improving operational efficiency of a casualty insurance department By Stewart Cho While improving the operational efficiency of a casualty insurance department in the property insurance industry has been extensively investigated, increasing the performance and efficiency. 政治大 studies the improvement of efficiency of the casualty insurance department. The research 立 of an underwriting department in a life insurance company is relatively unexplored. The paper. ‧ 國. 學. focuses on how a non-life insurance company should change in order to improve the efficiency of its casualty insurance department from five different factors, which are artificial intelligence. ‧. (AI) and robots, core systems, internal business processes and delegation of authority policy,. sit. y. Nat. employee training, and distribution channels and marketing strategies. Interview information. al. er. io. from the participants is categorized in order to ascertain the respondent’s opinions are presented. v. n. correctly. The findings suggest that property insurance companies could consider using AI or. Ch. engchi. i Un. robots, introducing a new core system, giving a higher limit of authorities or delegation of authority for its branches to underwrite more insurance products, and providing training that employees needed to handle daily business routines. These findings have implications for nonlife insurance companies to have plans or intend to evaluate what things they should do to increase performance and competitiveness from different perspectives. Keywords: Operational efficiency, Casualty insurance, Artificial Intelligence, Core system, Employee training. ii DOI:10.6814/NCCU202000750.

(5) TABLE OF CONTENTS 1. Introduction .......................................................................................................................... 1 1.1. Research Background Information .................................................................................. 1 1.2. Purpose of Research ........................................................................................................ 2 1.3. Research Questions ......................................................................................................... 3. 立. 政治大. 1.4. Overview of the Paper ..................................................................................................... 3. ‧ 國. 學. 2. Method ................................................................................................................................... 4. ‧. 2.1. Participants ...................................................................................................................... 4. sit. y. Nat. er. io. 2.2. Questionnaire ................................................................................................................... 5. al. n. iv n C 2.3. Data Collection Procedure ............................................................................................... 5 hengchi U 2.4. Analyses of Interview Results ......................................................................................... 6. 3. Background Information of the Research Companies ...................................................... 7 3.1. Overview of the X Company ........................................................................................... 7 3.2. Structure and Responsibility of the Casualty Insurance Department in the X Company 8 3.3. Current Issues about Operations in Casualty Insurance Department of the X Company 8. iii DOI:10.6814/NCCU202000750.

(6) 4. The Five Factors to Improve the Operational Efficiency of the Casualty Insurance Department ............................................................................................................................. 11 4.1. Artificial Intelligence and Robot ................................................................................... 11 4.1.1. Overview of Artificial Intelligence and Robot in Insurance Industry .................. 11 4.1.2. Applications of Artificial Intelligence and Robot................................................. 13. 政治大. 4.1.3. Possible Outcomes of Using AI or Robot in Insurance Industry .......................... 15. 立. 4.1.4. Robotic Process Automation (RPA) ..................................................................... 19. ‧ 國. 學. 4.2. Core Systems ................................................................................................................. 23. ‧. 4.2.1. Possible Business Drivers of Core Systems in Insurance Industry ...................... 24. sit. y. Nat. io. n. al. er. 4.2.2. Possible Benefits of Legacy Systems Modernization........................................... 26. i Un. v. 4.2.3. Core Systems’ Future Applications ...................................................................... 29. Ch. engchi. 4.3. Internal Business Processes and Delegation of Authority Policy .................................. 30 4.3.1. The Purpose of Establishing the Delegation of Authority Policy......................... 30 4.3.2. The Purpose of Underwriting in the Insurance Industry ...................................... 31 4.3.3. Operational Policy, Underwriting Policy, and Underwriting Guidelines in the Insurance Industry ............................................................................................................ 32 4.3.4. Underwriting Performance Evaluation Metrics .................................................... 33. iv DOI:10.6814/NCCU202000750.

(7) 4.4. Employee Training ........................................................................................................ 34 4.4.1. The Purpose of Employee Training ...................................................................... 34 4.4.2. Types of Employee Training................................................................................. 35 4.4.3. Possible Benefits of Providing Employee Training .............................................. 39 4.4.4. Suggested Training for an Underwriter ................................................................ 41. 政治大. 4.5. Distribution Channels and Marketing Strategies ........................................................... 43. 立. 4.5.1. Overview of Distribution Channels in Non-Life Insurance Market ..................... 43. ‧ 國. 學. 4.5.2. Marketing Strategies in the Non-Life Insurance Industry .................................... 45. ‧. 4.5.3. Evaluation of Marketing Performance ................................................................. 46. sit. y. Nat. n. al. er. io. 5. Research Results and Analysis .......................................................................................... 48. Ch. i Un. v. 5.1. Research Design ............................................................................................................ 48. engchi. 5.2. Research Findings ......................................................................................................... 49. 6. Conclusions ......................................................................................................................... 55 6.1. Conclusions ................................................................................................................... 55 6.2. Recommendations ......................................................................................................... 56 6.3. Limitations of the Study ................................................................................................ 58. v DOI:10.6814/NCCU202000750.

(8) 6.4. Suggestions for Future Research ................................................................................... 59. References................................................................................................................................ 60. Appendix ................................................................................................................................. 64. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. vi DOI:10.6814/NCCU202000750.

(9) List of Figures and Tables Figure 1: OECD recommendation on AI .................................................................................. 12 Figure 2: The RPA and AI technology continuum .................................................................... 23 Figure 3: A comparison of adopting legacy core systems and modernizing core systems ....... 28. Table 1: A comparison table of on-the-job training and off-the-job training ........................... 38. 政治大 Table 2: Statistics of written premium income in non-life insurance industry ......................... 44 立 ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. vii DOI:10.6814/NCCU202000750.

(10) 1. Introduction With the advent of new technology in the financial industry, this is a good time for financial institutions to introduce and use advanced technologies to enhance operational efficiency to gain competitive advantages. Therefore, this paper explores how an insurance company could improve its casualty insurance department’s performance and efficiency. The introduction can be separated into four parts: research background information, the purpose of research, research questions, and an overview of the paper.. 政治大 1.1. Research Background Information 立. ‧ 國. 學. According to 金融監督管理委員會保險局【金管會保險局】（2020）, Taiwan has seventeen local property insurance companies and seven foreign property insurance companies at the end. ‧. year of 2019. The market size of non-life insurance in Taiwan in 2019 was approximately. sit. y. Nat. NT$177.13 billion. The non-life insurance market in Taiwan can be divided into fire insurance,. al. er. io. marine insurance, automobile insurance, aviation insurance, engineering insurance, liability. v. n. insurance, credit insurance, accident insurance, and health insurance. The primary insurance. Ch. engchi. i Un. premium income is from fire insurance and automobile insurance in which the automobile insurance accounts for 53.44% while fire insurance shared with 14.61% of the total insurance premiums in 2019. Over recent years, casualty insurance has become a vital role in the non-life insurance market. One of the main reasons is the increasing focus on information security. Aon Public Limited Company [Aon] (2019) found that the number of companies purchasing cyber insurance in the global market grew from 21% to 54% during the period from 2015 to 2019. This trend might have stimulated some companies to buy cyber insurance. Also, as compliance with the 1. DOI:10.6814/NCCU202000750.

(11) establishment of the board of directors of listed companies in Taiwan amended（金融監督管理委員會【金管會】，2019）, more demand for purchasing directors and supervisors liability insurance has arisen. Furthermore, the Executive Yuan has already sent a draft of the Agricultural Insurance Law to Legislative Yuan to review since July 22, 2019（楊芳苓、曾耀民， 2020 ） . Property insurance companies thus have developed a growing number of agricultural insurance products. The industry usually combines liability insurance and other non-life insurance as casualty. 政治大. insurance. However, the combination of casualty insurance might be different from company. 立. to company. For example, some property insurance companies regard liability insurance,. ‧ 國. 學. engineering insurance, and other non-life insurance as casualty insurance. Other corporations. ‧. might add aviation insurance into the category.. Nat. sit. y. 1.2. Purpose of Research. n. al. er. io. The purpose of the study reported here was to explore how to improve operational efficiency. i Un. v. for a casualty insurance department in a property insurance company through analyzing in-. Ch. engchi. depth interview questionnaire respondents answered and reviewing other relevant existing materials. This report aims to propose possible recommendations from different angles to a casualty insurance department in a non-life insurance company for reform and property insurance companies to reconsider their business strategies. To ensure that suggestions for improving a casualty insurance department’s efficiency are to meet property insurance companies’ needs, interviews on experts who work for a non-life insurance company must be carried out to explore feasible solutions. However, the research for improving the efficiency of the casualty insurance department has not been studied yet. 2. DOI:10.6814/NCCU202000750.

(12) The present study intends to examine what non-life insurance companies should do or change to achieve better performance or increase operational efficiency for their casualty insurance department.. 1.3. Research Questions In view of the preceding research purpose, three specific research questions to be addressed in this study are as follows:. 政治大 improve its casualty insurance 立 department’s performance and operational efficiency?. Which areas should an insurance company in the non-life insurance industry change to. . What should a non-life insurer do to change its operating systems, employee training,. ‧ 國. 學. . authority policy, and marketing strategies for improving the efficiency of its casualty. ‧. insurance department?. Nat. y. To simplify underwriting procedures and improve the operational efficiency of a casualty. sit. . al. n. with daily routines?. er. io. insurance department, how should an insurance company introduce AI or robots to deal. Ch. engchi. i Un. v. 1.4. Overview of the Paper For the above objectives to be achieved, the article is structured as follows. The first section deals with the foundation for the development of the research. The research methodology is presented in the next section, along with full details of the participants in the research as well as the instrument and procedures used. Then the results are presented, with an analysis of the description of interviews from respondents. Finally, conclusions are presented, and suggestions are made for non-life insurance companies as references to improve their casualty insurance department’s operational efficiency. 3. DOI:10.6814/NCCU202000750.

(13) 2. Method The purpose of this chapter is to explain the research method in this research. Hence, it can be divided into four parts. The first part is regarding participants. The next section is about the questionnaire. Third, it focuses on the data collection procedure. Furthermore, the final piece is concerning analyses of interview results.. 2.1. Participants. 政治大 companies in Taiwan: X Company, MSIG Mingtai Insurance Company Limited, and Cathay 立 The participants in this research include seven people from three different property insurance. ‧ 國. 學. Century Insurance Company Limited. Those participants were chosen because they all work for non-life insurance companies. They came from different positions like a system. ‧. maintenance worker, managing level of a department with underwriting experience, insurance. sit. y. Nat. sales representative, or underwriter from the casualty insurance department. An IT worker. al. er. io. handles robotic process automation (RPA) 1 , the other is a liability insurance underwriting. v. n. specialist, and they both work for the Cathay Century Insurance Company Limited. Besides,. Ch. engchi. i Un. another two participants work for the MSIG Mingtai Insurance Company Limited, and one of them is a senior liability insurance underwriting specialist, and the other is a junior manager of the casualty insurance department. The last three respondents come from the X Company. One of them is an ex-deputy manager in the casualty insurance department. Another of them is the deputy manager of the industrial insurance marketing department, and the final participant is. Robotic process automation (RPA) in a new technology with artificial intelligence or metaphorical software robots to deal with repeatable tasks and RPA aims at business process automation to transfer jobs from humans to robots. 1. 4. DOI:10.6814/NCCU202000750.

(14) the deputy section chief of the financial channel marketing department.. 2.2. Questionnaire The materials from research questions consisted of six sections. The first section was intended to identify each participant’s current working company has already used AI or robots to deal with underwriting or not, and opinions of using AI or robots to assist in handling daily business routines. The second section asked participants to answer the satisfaction of their company’s operating systems they use to deal with underwriting or any related daily routines, and. 政治大. suggestions to improve the effectiveness of the systems they use. The third section focused on. 立. internal business processes and delegation of authority policy of the different company, and. ‧ 國. 學. participants’ thoughts regarding the feasibility of giving authority to an insurance sales. ‧. representative of providing quotations to customers. The fourth section was designed to assess the opinions of employee training. The fifth part was to collect the answers concerning the. y. Nat. er. io. sit. distribution channels and marketing strategies of each company. In the sixth section, participants were asked to give any recommendations for increasing operational efficiency to a. n. al. Ch. i Un. v. casualty insurance department besides the five factors had asked above.. engchi. 2.3. Data Collection Procedure Data collection for this study occurred over two months. The sources of data included face-toface interviews, written or produced papers, and research reports. The procedure took approximately one and a half hours for each interview. I conducted the interviews individually with each participant. This paper adopted face-to-face interviews instead of other research methods because the research topic focuses on specific areas and does not need samples or use statistical methods to analyze research results and findings. On the other hand, to collect in5. DOI:10.6814/NCCU202000750.

(15) depth, comprehensive understanding of each participant’s views, we think that a face-to-face interview is the appropriate method to receive the information needed.. 2.4. Analyses of Interview Results During the analysis, I read through the interview transcripts, summarize the informant’s views, group theses summaries, and describe those opinions more precisely under headings. The data were then reexamined entirely. Later, the headings were revised and refined to more accurately reflect the interviews. Besides, these categories were renamed to reflect the nature of the informant’s thoughts more precisely.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 6. DOI:10.6814/NCCU202000750.

(16) 3. Background Information of the Research Companies The research companies in this study are three non-life insurance companies in Taiwan, but this introduction will focus on X Company. The first section is about the history and current operating background information of the X Company. The following section turns to the structure and responsibility of the casualty insurance department in the X Company. The final part points out current operating issues in this department in the X Company.. 政治大 The X Company has established for more than 88 years. This company provides its customers 立 3.1. Overview of the X Company. ‧ 國. aviation insurance,. fire insurance,. liability insurance,. marine hull insurance,. engineering insurance, miscellaneous insurance (or. ‧. other non-life insurance),. marine cargo insurance,. 學. with automobile insurance,. accident insurance, and health insurance. The X Company’s. sit. y. Nat. headquarters is in Taipei City, and its branches and corresponding offices are based in most of. al. er. io. the cities in Taiwan. In addition, this organization has a representative office overseas. Now,. v. n. this company belongs to a member of a well-known holding group in Taiwan.. Ch. engchi. i Un. The X Company’s revenue stream comes from issuing non-life insurance policies, inward reinsurance premiums, and investment activities. According to 資誠聯合會計師事務所. 【資. 誠】（2020）, this company generated a total operating income of approximately NT$5,106.87 million in 2018, which had an increase of roughly 5.65% or increased NT$273.155 million premiums compared with the historical record in 2017. Then, the total premium income of 2018 consisted of automobile insurance with 46.90%, fire insurance with 14.57%, marine cargo, marine hull insurance, and aviation insurance with 8.04%. Casualty insurance (including liability insurance, engineering insurance, and miscellaneous insurance) 7. DOI:10.6814/NCCU202000750.

(17) accounted for 24.48% and accident insurance (including health insurance) shared by 4.88% of the total written premiums in 2018. In addition, the inward reinsurance premiums from the overseas market were NT$85.759 million, and they accounted for around 1.13% of the total premium income in 2018. It shows that the major premiums came from automobile insurance and casualty insurance.. 3.2. Structure and Responsibility of the Casualty Insurance Department in the X Company. 政治大. The current structure of the casualty insurance department is divided into four teams. The first. 立. team is the engineering insurance team responsible for underwriting, business development,. ‧ 國. 學. monitoring target budget achieving rate of the direct written premiums per month, preparing. ‧. insurance products documents for submission for review by the government, and any other assigned tasks by the X Company’s executives. Team two deals with all tasks as same as the. y. Nat. er. io. sit. first team mentioned above, but this team primarily focuses on liability insurance, miscellaneous insurance, and nuclear insurance. The third team is in charge of all reinsurance. n. al. Ch. i Un. v. related tasks, especially in dealing with whole reinsurance treaties and facultative reinsurance. engchi. cases’ reports and statements of accounts. The last team handles all cases and tasks associated with claims in the X Company’s casualty insurance department.. 3.3. Current Issues about Operations in Casualty Insurance Department of the X Company The operating issues of the X Company can be separated into five following points. The first issue of this department is lack of efficiency because underwriters need to follow internal underwriting procedures and delegation of authority policies to underwrite insurance policy 8. DOI:10.6814/NCCU202000750.

(18) cases. After the whole process of underwriting procedures, the underwriter can provide. quotations to customers. However, it does not matter that the quotes come from an insurance broker, an insurance agent, or direct customers, customers are eager to receive quotations immediately. This phenomenon has already made a conflict of making profits and pursing the speed of providing quotes to customers. Because some insurance products need to seek for facultative reinsurance support to reduce underlying retention, or some quotes have a relatively high frequency or high severity of loss record. Hence, it is hard for an underwriter to find a. 政治大. balance of taking care of profits of each underwriting case and maintaining a high efficiency to. 立. give quotations at the same time.. ‧ 國. 學. The second issue of the X Company’s department is that operating systems currently use might. ‧. be out of date if compared with competitors’ systems. As a result, customers have already been. Nat. sit. y. used to receive quotations with high efficiency. To achieve the target of providing quotes. n. al. er. io. efficiently, this department might need to consider integrating all currently operating systems. i Un. into a new core system to improve operational efficiency.. Ch. engchi. v. The third issue is that underwriters do not have enough employee training needed when underwriting new insurance products. Some insurance products need highly professional knowledge, or underwriters need to accumulate experience through time to distinguish each insurance product’s potential risks, loss frequency, and loss severity. For example, cyber insurance is a relatively new insurance product introduced in the Taiwan market. Therefore, some underwriters in property insurance companies might not have sufficient knowledge and experience to measure the risks of different cases. Under this circumstance, some companies might rely on insurance brokers or insurance agents to provide supports. However, this strategy 9. DOI:10.6814/NCCU202000750.

(19) might lead to an underwriter to look like a robot or a rubber stamp without learning much and accumulating experiences from underwriting quotations.. The fourth issue is that employees of this department need to spend most of the time dealing with daily business routines, especially in dealing with underwriting. Thus, this situation might lead to employees not having time to focus on developing new insurance products needed by potential customers or the market to miss the timing to grasp opportunities and market shares.. 政治大 employees to follow. For instance, 立 after underwriting a quotation, underwriters have their own The last issue is that this department does not establish a standard operating procedure for. ‧ 國. 學. method to archive files to electronic or paper documents. If an underwriter quit his or her job one day, it might be difficult for a new underwriter to find the historical records he or she needs.. ‧. In addition, this department does not have an existing mentoring program for junior staff to. Nat. n. al. er. io. sit. y. guide them on what to do.. Ch. engchi. i Un. v. 10. DOI:10.6814/NCCU202000750.

(20) 4. The Five Factors to Improve the Operational Efficiency of the Casualty Insurance Department The purpose of this chapter is to investigate which factors might help a casualty insurance department to enhance its competitiveness from five different elements. The first section focuses on artificial intelligence and robots. The next part is to describe the importance of a core system for an insurance company. The third section explores the relationship between the. 治政大 about the non-life insurance employee training. Finally, this section is to provide information 立 delegation of authority policy and underwriting. Section 4 is to analyze the necessity of. company’s distribution channels and marketing strategies in the Taiwan market.. ‧ 國. 學. 4.1. Artificial Intelligence and Robot. ‧. This section will focus on what artificial intelligence is and how this technology can be applied. y. Nat. sit. to the insurance sector. Also, this section will explore the possible outcomes when using. n. al. er. io. artificial intelligence technologies in the insurance industry. The final part will describe the. i Un. v. concept of robotic process automation (RPA) and the reason this robot could help a company to increase operating efficiency.. Ch. engchi. 4.1.1. Overview of Artificial Intelligence and Robot in Insurance Industry The history and development of artificial intelligence research can trace back to the 1950s (Organisation for Economic Co-operation and Development [OECD], 2020). The concept of AI is machines or computers can learn and take actions to maximize its chance to achieve its goal. With the advent of new technology, especially in growth in computing power and memory capacity, which has accelerated artificial intelligence technology, can be applied in more areas, for example, a search engine, virtual assistant, speech or facial recognition system. 11. DOI:10.6814/NCCU202000750.

(21) In addition, AI technology can be used in software or hardware, such as robots, autonomous cars, Internet of things (IoT), and biotechnology areas.. Even though using AI technology can help a company or an organization to deal with many things, there are concerns about the ethical application of AI. As a result, the OECD has a guideline for us to follow, called Recommendation of the Council on Artificial Intelligence (OECD, 2020). The summary of the OECD Recommendation of the Council on Artificial Intelligence is shown as below figure.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. Figure 1: OECD recommendation on AI Source: OECD (2020), The Impact of Big Data and Artificial Intelligence (AI) in the Insurance Sector. 12. DOI:10.6814/NCCU202000750.

(22) Besides, the other guideline for us to follow is Ethics Guidelines for Trustworthy AI, which is written by the European Commission’s Independent High-Level Expert Group on Artificial Intelligence (AI HLEG). The guideline of the Ethics Guidelines for Trustworthy AI has seven main requirements for trustworthy AI, which are (1) human agency and oversight, (2) technical robustness and safety, (3) privacy and data governance, (4) transparency, (5) diversity, nondiscrimination, and fairness, (6) environmental and societal well-being and (7) accountability (High-Level Expert Group on Artificial Intelligence [AI HLEG], 2019).. 政治大. The applications of artificial intelligence are numerous. Image recognition, intelligent robots,. 立. and machine learning are the most widely used areas of artificial intelligence. In the insurance. ‧ 國. 學. sector, artificial intelligence has been applied in many areas. The most common use of AI. distribution sites.. ‧. technology in the insurance industry is underwriting, claims, promoting insurance products, or. sit. y. Nat. n. al. er. io. 4.1.2. Applications of Artificial Intelligence and Robot. i Un. v. The applications of artificial intelligence in the insurance industry can be separated into three. Ch. engchi. types: robo-advisor, regulatory technology (RegTech), and supervisory technology (SupTech). Robo-advisor is a robot that can receive a quote request from a customer. Then, robo-advisor will ask a series of questions after receiving that information from customers, which will provide quotations or investment suggestions to customers. Organisation for Economic Cooperation and Development [OECD] (2017) reported that robo-advisor capabilities could be divided into four main types, which are understanding client needs, proposing a policy, implementing the policy, and monitoring and adjusting the policy. Robo-advisor has the advantage of decreasing operating costs. The robo-advisor does not have an apparent strength 13. DOI:10.6814/NCCU202000750.

(23) to maintain a long-term relationship of trust with a client if compared with a traditional. financial advisor/broker/agent. Also, this new technology might still need time to let clients transfer their old patterns of behavior to make a deal from a human being instead of a robot or artificial intelligence.. On the other hand, robo-advisor can deal with multiple goals from a client simultaneously, for example, it can handle investment, retirement, health care, and protection needs coverage.. 政治大 information to a financial advisor. 立 In this circumstance, robo-advisor can keep clients’ privacy, Besides, some people might think that they do not feel comfortable providing personal. ‧ 國. 學. and this can be appealing to those customers.. ‧. Companies or regulators can use artificial intelligence to help them to monitor regulatory. y. Nat. compliance. According to the Financial Stability Board [FSB] (2017), financial technology. er. io. sit. (FinTech)2 can be separated into two types, which are regulatory technology and supervisory technology. RegTech can help a company or an organization under supervision to use artificial. al. n. iv n C intelligence, blockchain, and other h related technologies e n g c h i U to meet and solve regulatory and. compliance requirements from the authority for monitoring. After the financial crisis of 20072008 or as known as the global financial crisis, financial institutions need to follow stricter regulations to avoid similar financial crisis will happen in the future. Therefore, in order to meet complex rules, RegTech has been developed to help companies to improve compliance response, decrease compliance costs, and simplify compliance procedures. The SupTech is a. Financial technology (FinTech) means a new technology or an innovation that could help to create a new business model, applications, or new products in the financial market or financial institutions. 2. 14. DOI:10.6814/NCCU202000750.

(24) financial technology used by regulators or supervisory authorities to help them to monitor financial institutions more effectively and efficiently. In some cases, the applications of RegTech can encompass SupTech.. 4.1.3. Possible Outcomes of Using AI or Robot in Insurance Industry In the insurance sector, using artificial intelligence technology could result in positive or negative outcomes. 財團法人資訊工業策進會【資策會】（2020）found that there are four possible outcomes when financial institutions adopt artificial intelligence technology. The first. 政治大. possible effect is that changing the structure of employment. The next possible effect is that it. 立. might result in employment discrimination. The third possible outcome is the invasion of. ‧ 國. 學. privacy. The last consequence is that it has changed in working conditions.. ‧. One of the primary purposes of this study is to determine possible outcomes when insurance. Nat. sit. y. companies use AI technology. As a result, several different effects of using AI in the insurance. er. io. sector are discussed as follows.. al. n. iv n C The first advantage of using AI technology industry is that it might help U h e n g cinhthei insurance. . insurance companies to improve operational efficiency, shorten claims processing, and reduce operational costs. In the insurance sector, numerous tasks are manual input, such as underwriting and claims. Artificial intelligence can learn with advanced algorithms, then collect data and help users to make decisions, which can increase automation and save service time and space. Therefore, this scenario mentioned above might lead a company to diminish a certain amount of costs. . The second advantage of using AI in an insurance company is that it could help insurers to improve the internal system’s efficiency and increase the data analysis process’s 15. DOI:10.6814/NCCU202000750.

(25) effectiveness. If an insurance company uses a legacy core system and does not introduce AI into the internal business processes, employees might need to spend much time. dealing with daily routines. Also, the best selling point might be the price. Therefore, the best promotion strategy for an insurance company might be that it could provide the cheapest insurance products on the market. To achieve this, AI can help underwriters to collect data they need more accurately. Then they can give actuaries to adjust the current premium rate of insurance products. . 政治大. The third advantage of using AI in the insurance sector is that it might help insurance. 立. companies to develop parametric insurance (OECD, 2020). The parametric insurance is. ‧ 國. 學. one kind of insurance product with a fixed payment based on the occurrence of an. ‧. objective or a triggering event. The first benefit of parametric insurance is that this insurance can cover risks that traditional insurance cannot include. For example,. sit. y. Nat. io. er. traditional insurance might not pay out to high or low temperatures when no physical damages incurred. The second benefit of parametric insurance is a quicker speed of. n. al. Ch. i Un. v. claims process. Customers can receive a payout when a trigger of a claim has been. engchi. reached. It is difficult for insurers and customers to have reliable tools to decide the trigger event in the past. After a growing development in AI and blockchain, insurance companies can take advantage of predetermined criteria to trigger claim payments because AI can deal with claims with automation, and better data collection and analysis methods have been developed. . The fourth advantage of using AI in the insurance industry is that it might increase a new distribution channel. For instance, a casualty insurance department can analyze its full insurance products first. Choosing some insurance products sells through online websites 16. DOI:10.6814/NCCU202000750.

(26) or applications on smartphones and offers quotations by robo-advisor to clients. Public liability insurance might be one of the ideal products to sell via a robo-advisor because. some public premises in Taiwan are mandatory to buy a public liability insurance policy. The specific aim of this strategy is to target some public premises that have low frequency and low loss severity of risks to increase profits. Therefore, this method can decrease costs on distribution channels and provide quotations at a more competitive premium level. . 政治大. The last advantage of using AI in an insurance company is that it might help insurance. 立. companies to provide better marketing results and customer experience. After issuing. ‧ 國. 學. multiple insurance policies, artificial intelligence can analyze current and the other. ‧. potential customers’ preferences to provide customized insurance products to target customers. Also, insurance companies can use that information analyzed from online. sit. y. Nat. io. er. platforms and forums by AI to set up proper marketing strategies and create innovative insurance products that customers need. Moreover, AI can boost the speed of settling a. n. al. Ch. i Un. v. claim, which might improve customer experience and attract more potential customers to buy insurance policies.. engchi. After focusing on different possible positive outcomes when insurance companies use AI technology, we will discuss various scenarios of potential adverse effects for the insurance industry. . The first probable disadvantage of using AI in the insurance sector is that insurance companies might need to improve data protection of customer’s personal information. Even though AI usage in the insurance industry has been used in several areas, the AI is 17. DOI:10.6814/NCCU202000750.

(27) still the early stage of development. In some circumstances, so there might request human intervention when AI detects unintended consequences or bias. Hence, in order to. protect personal data and avoid the result in an invasion of privacy. Insurance companies might need to upgrade their IT systems to protect the privacy of their customers. . The second probable disadvantage of using AI in the insurance sector is that the insurance company might need to adjust their compliance procedures to meet future compliance requirements. It may be a dilemma for supervisory authorities and regulators. 政治大. to find a perfect balance between encouraging start-ups to develop AI technology and. 立. using strict regulations for supervision. To address this issue, they already have two. ‧ 國. 學. methods for start-ups to determine the appropriate regulatory approaches, which are. ‧. regulatory sandboxes and innovation hubs. However, in order to protect personal privacy, supervisory authorities and regulators might consider making stricter for. sit. y. Nat. io. The three probable disadvantage of using AI in insurance companies is that it might. al. n. . er. companies using AI to comply with in the future.. Ch. i Un. v. increase operating expenses in the short run. Companies need to recruit new employees.. engchi. Those workers must be familiar with AI technology to maintain AI systems. Before introducing AI technology to use, it needs to spend a certain amount of expenses to purchase software and hardware. Then, those systems need to maintain and examine regularly, which will result in maintenance fees and related operating expenses. Besides, those companies still need to recruit new skillful workers to ensure that AI systems can be used smoothly. Moreover, it may be a possible issue is that insurers use AI need to have the ability to integrate existing operating systems with new AI systems. . The fourth probable disadvantage of using AI in the insurance industry is that it might 18. DOI:10.6814/NCCU202000750.

(28) lead to the property insurance industry to be a seller’s market. When AI technology has developed to a new stage, the insurance industry’s current business model might be changed. For example, sharing economy is one of the most well-known examples to use. disruptive innovation to threaten the existing market, such as Uber and Airbnb. In this context, it may lead the market to a winner-takes-all market or an oligopoly, which might have only a few insurance companies for customers to choose their services. Therefore, in this circumstance, the non-life insurance market might become a seller’s market. . 政治大. The last probable disadvantage of using AI in the insurance sector is that it might lead to. 立. some customers not able to buy the insurance policies they need. An insurance company. ‧ 國. 學. using AI can help this firm to distinguish what risks it would like to take and exclude. ‧. those risks that do not want to accept. In the scenario that might increase more exclusions in an insurance policy, it might scarify customers’ rights to buy the insurance. sit. y. Nat. io. n. al 4.1.4. Robotic Process Automation (RPA) Ch. engchi. er. policies they need.. i Un. v. Robotic process automation or it also knows as RPA, is a kind of software that can help humans to deal with repetitive data-driven tasks. RPA can run on the other systems or applications, whose primary purpose is to increase the business process via automation. According to the Institute for Robotic Process Automation and Artificial Intelligence [IRPA AI] (2019), “RPA is the application of technology that allows employees in a company to configure computer software or a “robot” to capture and interpret existing applications for processing a transaction, manipulating data, triggering responses, and communicating with other digital systems” (para. 3). 19. DOI:10.6814/NCCU202000750.

(29) In general, RPA can be used in any industry, such as the financial industry, the insurance sector, or the high tech industry. Lacity and Willcocks (2016) found that RPA is suited for in the “shared services” to achieve business benefits, and it is an ideal method to replace human beings to deal with “swivel chair” processes. According to Lacity and Willcocks (2016), the definition in their study of the shared services is from Accenture Public Limited Company [Accenture] (2005), which is: The consolidation of administrative or support functions (such as human resources,. 政治大 agencies into a single,立 stand-alone organizational entity whose only mission is to finance, information technology and procurement) from several departments or. ‧ 國. 學. provide services as efficiently and effectively as possible. (p. 3). ‧. Besides, Lacity and Willcocks (2016) described “swivel chair” processes as: A human being sitting in a swivel chair at a workstation and taking in work from many. y. Nat. er. io. sit. electronic inputs (like emails and spreadsheets), processing them by accessing more systems, and then updating the completed work to yet other systems, like ERP or CRM. n. al. systems. (p. 14). Ch. engchi. i Un. v. Therefore, if a company or an organization’s major tasks meet three conditions, it could consider introducing RPA to increase operational efficiency, which is repetitive, highly manual, and rule-based.. 20. DOI:10.6814/NCCU202000750.

(30) The benefits of adopting RPA in a company or an organization can be separated into the following points: . The first advantage of using RPA is that RPA is easy to use. It does not need to recruit new employees with coding skills or programming experience. Thus, companies use RPA could consider training their current employees and creating an RPA center of excellence to ensure companies can take advantage of using RPA. The second advantage of using RPA is that RPA can operate in existing systems or. 政治大 applications, which means 立that using RPA does not need to install a new platform to use. . 學. it.. ‧ 國. . The third advantage of using RPA is that RPA could increase operational efficiency and. ‧. improve accuracy. RPA can follow the rules to deal with repetitive tasks that are done by. Nat. sit. y. humans. People might feel bored doing the same things during a period or a long day.. n. al. er. io. With RPA, this software can deal with assigned tasks 24/7 because robots do not take a. i Un. v. vacation or feel tired. Moreover, some professional or strategic tasks need employees to. Ch. engchi. concentrate on handling them. In order to ensure the quality and accuracy of those tasks, RPA could help employees to achieve this goal. . The fourth advantage of using RPA is that it could help to realign the workforce. RPA could help a company to free up its employees. Therefore, employers can assign their employees to deal with strategic activities to increase productivity.. . The last advantage of using RPA is that RPA could improve governance and track audit history. A company uses RPA to help this company to meet regulatory compliance because RPA will follow defined rules to deal with tasks. Therefore, it is easy for IT 21. DOI:10.6814/NCCU202000750.

(31) workers, users, and auditors to check each step of tasks. In addition, auditors can quickly help the IT department to adjust rules for RPA in order to meet regulations and compliance.. However, RPA has its limitations. RPA only can deal with information in a digital form. Besides, RPA needs to follow rule-based or defined rules to process tasks. Therefore, in order to enhance the effectiveness of the traditional RPA, it is suggested that adopting RPA and artificial intelligence at the same time or digital automation tools like optical character. 政治大. recognition (OCR) to overcome the limitation of the traditional RPA. Also, AI can learn. 立. ‧ 國. 學. from RPA when RPA deals with tasks, which can help AI to improve the RPA’s effectiveness. Therefore, RPA and AI could be regarded as complementary to result in better outcomes of. ‧. business process automation, not just collecting data or information to make spreadsheets or. Nat. sit. y. reports. The combination of AI and RPA might be the best strategy to maximize the return on. al. n. suggestions.. er. io. investment, leading to the end-to-end process of offering predictive analytics and making. Ch. engchi. i Un. v. As Osinubi (2018), a report of PwC, pointed out, the RPA and AI technology continuum can be divided into five types, which are (1) macros and scripts, (2) business process automation (BPA), (3) robotic process automation (RPA), (4) intelligent process automation (IPA), and (5) algorithmic business (see Figure 2 below). Hence, based on PwC’s AI technology continuum, it shows that using RPA belongs to the current state or be trending. Intelligent process automation or it means that combining RPA and artificial intelligence technologies might be a future trend.. 22. DOI:10.6814/NCCU202000750.

(32) 立. 政治大. ‧. ‧ 國. 學. Figure 2: The RPA and AI technology continuum. y. Nat. al. iv n C systems inh an insurance company engchi U n. Information technology. er. io. 4.2. Core Systems. sit. Source: PwC (2018), Looking into the Future: Leveraging the Power of AI and Robotics. might be used for policy. administration, underwriting, billing, claims, and reinsurance. However, traditional core systems might not meet new requirements for collecting data or regulatory compliance. Therefore, insurance carriers in order to provide better customer experience and increase operational efficiency may consider core system modernization. The first part focuses on the possible business drivers for systems transformation. The second part gives some possible benefits after modernizing core systems. The final part turns to possible future applications with core systems.. 23. DOI:10.6814/NCCU202000750.

(33) 4.2.1. Possible Business Drivers of Core Systems in Insurance Industry This sub-section will provide two examples that might occur in an insurance company to explain why insurance carriers need to update their existing core systems. The first example is issuing an insurance policy can be very complicated, especially in renewing or issuing multiple insurance policies at one time. The steps of the policy issuance process can be divided into the following points.. Firstly, after receiving a signed proposal from insurance sales representatives, underwriters or. 政治大. a clerk is responsible for issuing insurance policies need to type necessary policy conditions. 立. into an electronic register and get a policy number. After office assistant finishing typing,. ‧ 國. 學. underwriters or clerks need to ensure that all type-in information is correct. In addition, if some. ‧. insurance policies’ conditions are beyond the level of authority, the office assistants need to ask their supervisor’s approval and ensure that insurance contracts match the requirements of. sit. y. Nat. io. er. quotations provided to customers before. According to Taiwan’s Money Laundering Control Act, Taiwan’s Regulations Governing Anti-Money Laundering of. n. al. Ch. engchi. i Un. v. Financial Institutions, and Taiwan’s Counter-Terrorism Financing Act, office assistants must use identification number or government uniform invoice (GUI) number to check whether applicants or insureds or business owner of the applicants are on the list of Financial Action Task Force (FATF) of “Non-Cooperative Countries or Territories” (or known as FATF blacklist) or whether they are on the against sanction lists, law enforcement watch lists, or lists of politically exposed persons (PEPs) from Anti-money laundering database or software. If those individuals or organizations are on those lists mentioned above, insurance companies might not sell their insurance policies owing to regulatory compliance in some circumstances. 24. DOI:10.6814/NCCU202000750.

(34) Furthermore, according to article 44 and article 45 of Taiwan’s Financial Holding Company Act, before insurance companies make a deal with their clients, they need to ensure that applicants or insureds do not have a relationship with them. If those clients have a relationship with insurance companies, those carriers need to make sure they did not engage in transactions with more favorable terms than offered to similarly situated customers. Finally, underwriters or the clerks need to bind printed insurance policies by hand.. The second example is that information in the legacy system is hard to retrieve and analyze.. 政治大. For example, in order to adjust the premium of insurance products to increase the product’s. 立. acceptance of customers, underwriters or the other clerks need to collect historical data like. ‧ 國. 學. policy premium and loss ratio in the five previous years at least of the target insurance products.. ‧. Then, those data will provide to actuaries to review. The actuaries will try to calculate a new insurance premium rate based on those data collected. However, it might take a long time for. sit. y. Nat. io. ask for the IT department for help to get data needed.. n. al. Ch. engchi. er. underwriters or clerks to collect that information, and sometimes this collection might need to. i Un. v. Based on the above two examples, possible business drivers of core systems transformation in. insurance sector might be as follows: . The first possible business driver is product strategy and objectives. For instance, in order to increase productivity or develop some innovative insurance products, legacy systems might not support those demands and deal with issuing multiple insurance policies at one time.. . The second possible business driver is outdated systems and aging programmers. An 25. DOI:10.6814/NCCU202000750.

(35) important issue is that if programmers or IT developers are older, they might not have the ability to maintain the new core systems. In addition, after those programmers retire, it is difficult for companies to recruit new employees with related experience or knowledge to maintain legacy core systems. . The third possible business is that it is a growing trend for a company to focus on corporate social responsibility (CSR). As a result, some insurance companies might try to operate a paperless office with automation. This trend might lead companies to. 政治大 The final possible business 立driver is that carriers tend to increase operational efficiency introduce new technology and upgrade their legacy operating systems.. . ‧ 國. 學. or meet new regulatory compliance. With the advent of new technology and new regulations, legacy IT systems might not be adequate for using innovative InsurTechs or. ‧. meeting new compliance requirements. For example, a new accounting rule of. Nat. sit. y. International Financial Reporting Standards 17 (IFRS 17) will adopt in 2026 in Taiwan. n. al. er. io. （楊絡懸，2020）. Before adopting the IFRS 17, insurance companies might need to. i Un. v. update their core systems, especially in policy administration, claims, reinsurance, and. Ch. engchi. actuarial functions（李靜宜，2020）. To solve this, modernizing existing core systems is one of the possible solutions.. 4.2.2. Possible Benefits of Legacy Systems Modernization According to Krish, Lansing, Münstermann, Olesen, and Vogelgesang (2019), the benefits of modernizing core systems can be divided into three points. A comparison between adopting legacy core systems and modernizing core systems is shown below (see Figure 3). . The first advantage of legacy systems modernization in the insurance industry is that 26. DOI:10.6814/NCCU202000750.

(36) “Increased gross written premiums and reduced churn” (Krish et al., 2019, para. 6). A better core system can enhance faster business processes, provide better customer experience, and reduce procedures. Also, nowadays, customers would like to receive insurance policies or offers as soon as possible. Therefore, how fast an insurance company can serve its clients would be a competitive advantage and attract potential customers to choose using this company’s services. . The second advantage of legacy systems modernization is “Increased operations. 政治大 improved the business process, 立 and repetitive tasks can be dealt with the systems. Hence, productivity” (Krish et al., 2019, para. 7). After upgrading existing core systems, it. ‧ 國. 學. employees can free up to do more strategic activities to increase productivity. Besides, Krish et al. (2019) pointed out that companies with modernized IT systems are more. ‧. productive than their competitors use legacy systems, which shows that companies with. Nat. sit. y. core systems modernization can increase 40% of the total number of insurance policies. al. n. . er. io. per full-time equivalent (FTE) or higher than companies use legacy systems.. i Un. v. The third advantage of legacy systems modernization is “Reduced IT cost” (Krish et al.,. Ch. engchi. 2019, para. 8). As Krish et al. (2019) showed that insurance companies with modernized core systems could save roughly 41% of IT costs lower than those carriers use with their legacy core systems.. 27. DOI:10.6814/NCCU202000750.

(37) 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. Figure 3: A comparison of adopting legacy core systems and modernizing core systems Source: McKinsey & Company (2019), IT modernization in insurance: Three paths to transformation. 28. DOI:10.6814/NCCU202000750.

(38) 4.2.3. Core Systems’ Future Applications After introducing the possible business drivers of modernizing core systems, we focus on the potential applications with modernized core systems. The probable applications primarily discover how big data from core systems can be used in the insurance sector as follows: . The first possible application with big data is that it is easier for underwriters or users to retrieve essential information they need to support the decision-making.. . The second possible application with big data is that it can provide more accurate. 政治大. analysis information for carriers to develop new product designs, collect actuarial data,. 立. and set up marketing strategies and promotion campaigns from new points of view.. ‧ 國. 學. . The third possible application with big data can help insurance companies with. ‧. underwriting and find fraudulent claimants. Big data can help carriers to analyze accurate information to adjust insurance products’ acceptable coverage. Still, it might. sit. y. Nat. io. er. also help insurance companies to distinguish what risks they are unwilling to take. Consequently, it may decrease the premium rate but increase more exclusions of an. n. al. Ch. i Un. v. insurance product for customers. Insurance companies can also analyze current and. engchi. historical data to identify potential fraud rings. . The final possible application is that modernized core systems can support to issue electronic insurance policies and bind printed policies automatically by using proper equipment and software. However, issuing electronic insurance policies have some concerns, such as regulations, policyholders’ acceptance, and data protection.. 29. DOI:10.6814/NCCU202000750.

(39) 4.3. Internal Business Processes and Delegation of Authority Policy This section explores internal business processes and delegation of authority policy in the nonlife insurance industry and how these two factors could influence operational efficiency in a property insurance company. First, this part will explain the delegation of authority policy in a non-life insurance company. Second, it will focus on underwriting in the insurance industry. Third, it will give details regarding underwriting guidelines. Finally, it will provide evaluation metrics to measure the underwriting performance of a department or a company in the insurance sector.. 立. 政治大. 4.3.1. The Purpose of Establishing the Delegation of Authority Policy. ‧ 國. 學. The primary aim of establishing delegation of authority policy is to give authority from general. ‧. manager (GM) or managing director (MD) to each department or every manager to make decisions based on this delegation of authority policy. For example, underwriters from the. sit. y. Nat. io. er. different departments will follow underwriting policies and underwriting guidelines that are for different insurance products like fire insurance, automobile insurance, or liability insurance to. n. al. Ch. i Un. v. judge each insurance policy’s risks. Then, underwriters will decide to provide quotations with. engchi. an appropriate premium based on its risks, or they might reject to offer quotations for customers if underwriters identify risks are not acceptable or the other companies offer unreasonable quotes. Therefore, if a company plans to amend the delegation of authority policy, it usually has to require approval from the general manager or managing director.. 30. DOI:10.6814/NCCU202000750.

(40) 4.3.2. The Purpose of Underwriting in the Insurance Industry Underwriting for an insurance company is to select acceptable cases and provide reasonable quotations to customers to buy insurance policies. The primary purpose of underwriting in an. insurance company can be divided into the following points. . To achieve the maximum profits, one way to realize the maximum profits is to offer the cheapest quotations to customers or introduce an entirely new insurance product with. 政治大 To avoid adverse selection 立or moral hazard, insurance companies need to recognize each potential demand from consumers.. . ‧ 國. 學. quotation request by its risks and decline some unacceptable cases like those that insured already has been diagnosed with a disease.. ‧. To select insurable risk cases from all insurance quotation requests. According to 廖述源. Nat. sit. y. （2007） , the method of selecting an insurable risk case can be separated into three. io. types: preselection, selection during the insurance period, and post-selection.. n. al. er. . i Un. v. Preselection means that underwriters need to evaluate each case’s risks by its risk levels,. Ch. engchi. and underwriters will decide to accept or reject this offer. Selection during the insurance period means that an insurance policy during the insurance period might terminate the contract or change its terms and conditions of the policy because underwriters find that risk levels of this policy increased or underwriting policies of insurers changed. Postselection means that an insurance company or an underwriter might decide to renew, decline, or change its original terms and conditions of an insurance policy after the insurance policy is due.. 31. DOI:10.6814/NCCU202000750.

(41) . To ensure all steps meet internal and external regulations, an underwriter in the insurance sector needs to help the company to make sure all processes are not against laws or compliance because the insurance industry is highly regulated. If an insurance company breaks the rules, this company could be heavily fined. For example, in the worst situation in Taiwan, under article 149 of the Insurance Act in Taiwan, the regulator can order. insurance companies to suspend sales of an insurance product or products or cannot launch a new insurance product. Besides, the regulator can order managers or employees. 政治大. of an insurance company to remove from their positions.. 立. To ensure all insurance policies meet underwriting policies and underwriting guidelines. 學. ‧ 國. . of a property insurance company.. ‧. y. Nat. 4.3.3. Operational Policy, Underwriting Policy, and Underwriting Guidelines in the Insurance Industry. io. sit. According to 廖述源（2007）, insurance operational policies consist of growth policy, stability. n. al. er. policy, and fairness policy. The growth policy means that an insurance company must pursue a. Ch. i Un. v. growing business performance in revenue and profits at the same time. In addition, the stability. engchi. policy implies that an insurance company needs to operate in a sustainable development manner. Moreover, the fairness policy means that an insurance company treats its employees or its customers equally, and this company does not provide special or differential treatment based on an individual or an organization’s status or relationship with others. Therefore, an insurance company needs to find a balance among the three operational policies mentioned above, and this company might need to change its operational policies with the different operational environments accordingly. Still, the company cannot only focus on one policy and ignore the others. 32. DOI:10.6814/NCCU202000750.

(42) Underwriting policy is a policy for an underwriting department of an insurance company to follow, and this policy is based on a company’s operational strategy to formulate. The underwriting policy usually decides on two sides, including dimensions of underwriting policy and limiting factors like capacity, compliance, underwriters, and reinsurance market.. The purpose of underwriting guidelines is for underwriting departments and underwriters in an insurance company to obey and decide whether an insurance offer could be accepted or be rejected by the underwriting guidelines. Also, the primary function of underwriting guidelines. 政治大. is to: (1) ensure all cases to be homogeneous (2) help arrange facultative reinsurance (3). 立. maintain consistency across insurance policies (4) help business development and marketing. ‧ 國. 學. (5) simplify claims settlement (6) establish consistency of underwriting philosophy （廖述源，. Nat. sit. 4.3.4. Underwriting Performance Evaluation Metrics. y. ‧. 2007）.. n. al. er. io. This part will focus on some evaluation metrics to measure the underwriting performance of a. i Un. v. department in an insurance company or an insurance company. According to 廖述源（2007） ,. Ch. engchi. this part provides five evaluation metrics as follows:. 1.. Average underwriting cases = The total number of quotations offered / The total number of underwriters. This metric aims to measure what an underwriter can deal with relative to the total number of quotation requests. Therefore, the higher this metric is and the better the performance will be.. 2.. Average time to underwrite a case = The total spend time for underwriting / The total number of signed cases. The meaning of this metric is an underwriter might create better performance when the value is smaller. 33. DOI:10.6814/NCCU202000750.

(43) 3.. Underwriting error rate = The total number of underwriting error cases / The total number of proposals received. The metric can help to measure an underwriter’s skills and ability. Besides, if the metric value is smaller than the others are, it means that this underwriter has better underwriting performance.. 4.. Average premium per insurance policy = The total number of written premiums / The total number of signed insurance policies. It shows that if this metric value is higher than the others value, a department or an insurance company has better underwriting performance.. 立. Underwriting loss ratio per signed policy = The total number of claims / The total. 學. ‧ 國. 5.. 政治大. number of signed policies. This metric measures underwriting performance by a claim. ‧. that might occur in the number of approved insurance policies. Hence, when this value of an underwriter is lower than the other underwriter’s value, it means that this underwriter. y. Nat. n. 4.4. Employee Training a l. er. io. sit. has a better underwriting performance.. Ch. engchi. i Un. v. This section aims to analyze the importance of employee training in an insurance company. Thus, the first step will explain what employee training is. Next, it will give more details regarding the different types of employee training. Third, it will focus on the possible benefits of implementing employee training for an insurance company. Lastly, it will provide some suggested training for an underwriter who deals with insurance companies’ casualty insurance.. 4.4.1. The Purpose of Employee Training Employee training could provide staff required courses to increase employee’s ability, knowledge, and performance. According to 洪榮昭（1991）, the purpose of employee training 34. DOI:10.6814/NCCU202000750.

(44) in an organization can be separated into ten points:. 1.. To improve knowledge. 2.. To increase judgment about the surroundings. 3.. To enhance practical knowledge of the industry and company. 4.. To increase operational and management capabilities. 5.. To improve leadership and coaching skills. 6.. To improve individual weaknesses (including skills, knowledge, and concepts). 7.. To develop interpersonal and communication skills. 8.. To learn specialized management techniques. 9.. To improve the ability to deal with business. 學. ‧ 國. 立. 政治大. ‧. 10. To develop new skills and abilities. Nat. sit. y. 4.4.2. Types of Employee Training. 1.. al. n. categories as follows:. er. io. According to 廖述源 (2007) , the types of employee training can be divided into five different. Ch. engchi. i Un. v. Pre-service training and on-the-job training: In general, pre-service training is only provided for new employees of a company or an organization. On-the-job training is to provide training for current employees in a company or an organization. In the insurance sector, employees might provide different training if they change their company position. It also has another type of training called off-the-job training, which is the kind of training for an employee at a training place out of the company so that the training providers could be outsourced training institutions or outsourced training providers. A. 35. DOI:10.6814/NCCU202000750.

(45) comparison of on-the-job training and off-the-job training is shown as the below table (see Table 1). 2.. Regular training and irregular training: Regular training is a training that is implemented at a particular time, and it will provide this kind of training repeatedly. For example, under article 14 of Regulations Governing Pre-sale Procedures for Insurance Products in Taiwan, “An insurance product signatory shall attend at least 15 hours per year of insurance-related professional training held by a training institution or recognized by the. 政治大 participate in at least thirty 立hours of on-the-job training every year according to current competent authority” (para. 1). In addition, underwriters and claim adjusters shall. ‧ 國. 學. regulations. Irregular training in an insurance company could be a new product or a new regulation that employees need to know or follow so that insurers will provide unplanned. ‧. training for their workers.. Nat. y. Supervisor training and non-supervisor training: Training can be divided into for. sit. 3.. n. al. er. io. supervisors and employees if it distinguishes from trainees. Hence, the most. i Un. v. significant difference in training for a supervisor or an employee is that supervisors’. Ch. engchi. training might be more regarding leadership skills training and managerial training. However, a company will provide general or other professional training for employees. 4.. Domestic training and overseas training: Employee training can be separated into domestic training and overseas training if we distinguish the difference from the training place. As usual, the insurance company provides training by itself or sends its employees to some domestic insurance training institutions like Taiwan Insurance Institute (TII) or Non-Life Insurance Association of the Republic of China. Overseas training usually will 36. DOI:10.6814/NCCU202000750.

(46) send employees to a seminar held by professional insurance institutions, well-known insurance brokers, or reinsurers abroad. Internal training and external training: In general, insurance companies in order to save money; therefore, they tend to provide the necessary training for their employees internally. However, in some cases, insurers might invite some professionals or offer training for their workers by outsourced training vendors. For instance, an insurance company plans to hold a seminar or training regarding cyber insurance or the future of. 政治大 be a lecturer to give a speech 立 or be a trainer to provide training. Outsourced training. InsurTech in Taiwan, so it might need to invite some professionals in different areas to. 學. providers will hold external training at a place out of the company. The benefit of. ‧ 國. external training is to offer more high-quality training courses for employees.. ‧. Nevertheless, the costs of external training might be more expensive than internal. io. sit. y. Nat. training.. n. al. er. 5.. Ch. engchi. i Un. v. 37. DOI:10.6814/NCCU202000750.

(47) Table 1: A comparison table of on-the-job training and off-the-job training Source:楊純瑜 (2012)、Wikipedia (2020), On-the-job training Category. 立. Off-the-job training. 1. Training is practical. 2. It is easier to teach onthe-job training than offthe-job training. 3. It could motivate employees to learn. 1. It can train most of the employees at the same time. 2. Employees can learn from professionals, and employees in pieces of. because of improvement from training.. training can learn highly specialized knowledge. 4. It could improve the understanding and coordination between supervisors and subordinates. 5. It can help employees to work and train at the same time.. and skills. 3. Trainees could have a competitive mindset to result in better training outcomes. 4. It can develop a sense of group consensus among trainees.. 政治大. n. al. 6. Expenses are lower than the off-the-job training. 7. It can adjust courses depend on an employee’s learning abilities.. er. io. sit. Nat. y. ‧. ‧ 國. 學. Advantages. On-the-job training. Ch. engchi. i Un. v. 38. DOI:10.6814/NCCU202000750.

(48) Disadvantages. 立. 4. Training results cannot be applied in time.. 政治大 cannot be unified.. 5. Specialized and highlevel knowledge and skills cannot be taught.. ‧ 國. y. ‧. n. al. Theoretical. sit. Practical. er. io. Time consumption. many. 4. The content and degree of difficulty of training. Imparted at the actual job At a place other than the location real job location. Nat. Approach. 1. The progress of a task might be delayed because employees are in training. 2. The training expense is more expensive than onthe-job training. 3. It is more challenging to implement for small and medium-sized enterprises.. 學. Location. 1. An excellent superior is not necessarily a good teacher. 2. A trainee might not handle work and training at the same time. 3. It might not prepare training for employees if the participants are too. Less time. iv. More Time. n U e n gbycdoing Learning h i the task. Learning by acquiring knowledge. Effect on production. No, because trainees produce the products during learning. Yes, because training is provided first, which is followed by a performance. Who performs the training. Employees, internal trainers or managers. Experts outside of the company. Cost. Inexpensive. Expensive. Learning method. Ch. 39. DOI:10.6814/NCCU202000750.