| 摘要: | 政府為了減輕弱勢民眾繳納保費壓力，在2003年訂定「某保險費及滯納金分期繳納辦法?供現階段欠費但無力一次繳納保費的民眾，可以有喘息機會，惟按現行分期繳納辦法所訂資格非常寬鬆，只要累積積欠保費總額超過2千元均可辦理分期繳納，在觀察樣本中，每年未按時繳納違約註銷件數有8千餘件(佔56.43%)。但台灣目前公部門所訂的分期辦法，都只規定欠費金額下限門檻及欠費金額所對應期數，完全不考慮民眾的財務狀況，最後淪為依承辦人主觀判斷給予民眾期數及金額，衍生行政效率不彰或民眾欠費爭議問題。
本研究以客觀基礎評估申請分期，達到有效減少人為之干擾及增加履約率。將可能影響個人信用的基本資料變數歸為羅吉斯迴歸Model Ⅰ 和可能影響分期繳納的個人行歸為羅吉斯迴歸Model Ⅱ，分別運用羅吉斯迴歸(Logistic Regression)分析、KS檢定與ROC曲線分析，估計出顯著變數後，將顯著變數歸納為羅吉斯迴歸Model Ⅲ，再進行一次羅吉斯迴歸分析與模型檢定，得到「年齡」、「性別」、「投保類別」、「分期次數」、「分期金額」、「分期期數」等6項顯著性影響因子。最後建議公部門可藉由FICO評分系統轉換成保險費分期申請客觀評分機制，以系統化方式率先將評分機制導入公部門，可有效提高效率。另外，公部門機關可廣泛建立風險機制在核心業務上，已邁向在事件發生前有所警示，減少事件發生後補救行為或衍生大量之行政成本。
;The Taiwanese government established the "Taiwan Health Insurance Arrears and Late Fees Installment Method" in 2003, so that people who cannot pay off the arrears will have another payment option and the eligibility is very loose, as long as the total arrears exceeds NT$2,000.
Even so, there are still 8,000 cases (approximately 56.43%) of default cancellation due to failure to pay fees on time every year. However, according to the current handling method prescribed by the public authorities in Taiwan, only the minimum threshold and the number of fee periods can only be handled according to the undertaker.
The subjective handling of arrears instalment application cases completely ignores the applicant’s financial status, which leads to problems such as administrative inefficiency and customer litigation disputes.
This research uses an objective basis to evaluate the application phases to effectively reduce human interference and increase the contract performance rate.The basic data variables that may affect personal credit are classified as Logistic regression ModelⅠand the personal behavior that may affect installment payments are classified as Logistic regression ModelⅡ, using Logistic Regression analysis, KS test and ROC respectively Curve analysis, after estimating the significant variables, summarize the significant variables as Logistic regression Model Ⅲ, and perform Logistic regression analysis and model verification again to obtain age, gender, insurance type, number of installments, amount of installment, and number of installments and other 6 significant impact factors.
Finally, it is recommended that the public sector fully introduce the FICO (Fair Isaac Corporation) scoring system, by converting the FICO scoring system into an objective scoring mechanism for insurance premium installment applications, and the first to introduce the scoring mechanism into the public sector in a systematic way, which can effectively improve efficiency.
In addition, public sector agencies can extensively establish risk mechanisms in their core business, and have moved towards warning before the incident, reducing remedial actions after the incident or incurring a large amount of administrative costs. |
