A Study of the Default Factors on the Student Loan-Using Logistic Regression Model
碩士 === 國立高雄應用科技大學 === 金融資訊研究所 === 99 === Banks undertook the service of student loan in order to fulfill the responsibility of the social welfare, to support the government’s policy and to assist the students to complete their education. Different from all other loan services, the student loan servi...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2011
|
Online Access: | http://ndltd.ncl.edu.tw/handle/93655158679768897930 |
id |
ndltd-TW-099KUAS8213042 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-099KUAS82130422015-10-16T04:02:39Z http://ndltd.ncl.edu.tw/handle/93655158679768897930 A Study of the Default Factors on the Student Loan-Using Logistic Regression Model 就學貸款違約風險因子之研究-應用羅吉斯迴歸模型 Jun Yen 鄢駿 碩士 國立高雄應用科技大學 金融資訊研究所 99 Banks undertook the service of student loan in order to fulfill the responsibility of the social welfare, to support the government’s policy and to assist the students to complete their education. Different from all other loan services, the student loan service is provided to the qualified applicants without a guarantee. Due to the expansion of the educational institutions and the impact of low birth rate; it is no longer a dream to go to university. However, the recent financial and economical crisis caused the raising of the unemployment rate and the increasing of the tuition fees; the average income families are not able to afford the high education cost.Therefore, the student loan service is becoming more critical to the financially disadvantaged families. According to the statistics of the Department of Education in School Year 98, the number of the students who applied the student loans was 817,000 and the total amount of the loan was about 32 billion. Both of them have reached the recorded high. Based on the statistics of the student loan in the Bank of Taiwan at the end of December, 2011- the NPL ratio was 0.62%, the amount of the overdue loans was 1.593 billion, and its NPL ratio was 1.1%. This showed the problem of the overdue loans was very severe. Therefore, how to reduce the NPL ratio, improve the quality of educational loan and achieve the performance of the bank are becoming the important issues to the banks. This study is based on the domestic major banks for the student loan and targets on the students of high school and above who applied the student loans in Pingtung area. Also it focuses on the students who have repaid the loans since August 1st, 2009 as the study candidates. This study randomly selected 5,000 samples which consisted of 3,679 of the normal cases and 1,321 of the default cases. This study established 21 possible causes of the student loan default risks. It analyzed the degree of impact to the occurrences of late payment of the loan and predicted the probability of the possible default through the logistic regression analysis. This study shows: the significance of this study is 0.05, birthplace、amount of the loan、interest rate、the type of guarantors、the attributes of schools、the type of schools、tell to catch up in accordance with law、pre-authorized withdrawal、the reduction of tuition and schooling、low income earners、the application for extension payment and early repayment are the 12 possible significant factors that will affect the delay payment of the loan. These factors provide a very significant explanation of the default of the student loan.Of the fitted logistic regression model in this study, to the default cases is 92.43%, to the normal cases is 98.18% and the overall accuracy is 96.66%. This study shall benefit to the banks and will greatly assist the lending service in the future. Yen-shin Cheng 程言信 2011 學位論文 ; thesis 126 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立高雄應用科技大學 === 金融資訊研究所 === 99 === Banks undertook the service of student loan in order to fulfill the responsibility of the social welfare, to support the government’s policy and to assist the students to complete their education. Different from all other loan services, the student loan service is provided to the qualified applicants without a guarantee. Due to the expansion of the educational institutions and the impact of low birth rate; it is no longer a dream to go to university. However, the recent financial and economical crisis caused the raising of the unemployment rate and the increasing of the tuition fees; the average income families are not able to afford the high education cost.Therefore, the student loan service is becoming more critical to the financially disadvantaged families.
According to the statistics of the Department of Education in School Year 98, the number of the students who applied the student loans was 817,000 and the total amount of the loan was about 32 billion. Both of them have reached the recorded high. Based on the statistics of the student loan in the Bank of Taiwan at the end of December, 2011- the NPL ratio was 0.62%, the amount of the overdue loans was 1.593 billion, and its NPL ratio was 1.1%. This showed the problem of the overdue loans was very severe. Therefore, how to reduce the NPL ratio, improve the quality of educational loan and achieve the performance of the bank are becoming the important issues to the banks.
This study is based on the domestic major banks for the student loan and targets on the students of high school and above who applied the student loans in Pingtung area. Also it focuses on the students who have repaid the loans since August 1st, 2009 as the study candidates. This study randomly selected 5,000 samples which consisted of 3,679 of the normal cases and 1,321 of the default cases. This study established 21 possible causes of the student loan default risks. It analyzed the degree of impact to the occurrences of late payment of the loan and predicted the probability of the possible default through the logistic regression analysis.
This study shows: the significance of this study is 0.05, birthplace、amount of the loan、interest rate、the type of guarantors、the attributes of schools、the type of schools、tell to catch up in accordance with law、pre-authorized withdrawal、the reduction of tuition and schooling、low income earners、the application for extension payment and early repayment are the 12 possible significant factors that will affect the delay payment of the loan. These factors provide a very significant explanation of the default of the student loan.Of the fitted logistic regression model in this study, to the default cases is 92.43%, to the normal cases is 98.18% and the overall accuracy is 96.66%. This study shall benefit to the banks and will greatly assist the lending service in the future.
|
author2 |
Yen-shin Cheng |
author_facet |
Yen-shin Cheng Jun Yen 鄢駿 |
author |
Jun Yen 鄢駿 |
spellingShingle |
Jun Yen 鄢駿 A Study of the Default Factors on the Student Loan-Using Logistic Regression Model |
author_sort |
Jun Yen |
title |
A Study of the Default Factors on the Student Loan-Using Logistic Regression Model |
title_short |
A Study of the Default Factors on the Student Loan-Using Logistic Regression Model |
title_full |
A Study of the Default Factors on the Student Loan-Using Logistic Regression Model |
title_fullStr |
A Study of the Default Factors on the Student Loan-Using Logistic Regression Model |
title_full_unstemmed |
A Study of the Default Factors on the Student Loan-Using Logistic Regression Model |
title_sort |
study of the default factors on the student loan-using logistic regression model |
publishDate |
2011 |
url |
http://ndltd.ncl.edu.tw/handle/93655158679768897930 |
work_keys_str_mv |
AT junyen astudyofthedefaultfactorsonthestudentloanusinglogisticregressionmodel AT yānjùn astudyofthedefaultfactorsonthestudentloanusinglogisticregressionmodel AT junyen jiùxuédàikuǎnwéiyuēfēngxiǎnyīnzizhīyánjiūyīngyòngluójísīhuíguīmóxíng AT yānjùn jiùxuédàikuǎnwéiyuēfēngxiǎnyīnzizhīyánjiūyīngyòngluójísīhuíguīmóxíng AT junyen studyofthedefaultfactorsonthestudentloanusinglogisticregressionmodel AT yānjùn studyofthedefaultfactorsonthestudentloanusinglogisticregressionmodel |
_version_ |
1718090570628333568 |