View : 36 Download: 0

일반화 추정 방정식을 이용한 한국 가계부채 패널 자료 분석

Title
일반화 추정 방정식을 이용한 한국 가계부채 패널 자료 분석
Other Titles
Analysis of Korean household debt panel data based on generalized estimating equation approach
Authors
원혜진
Issue Date
2016
Department/Major
대학원 통계학과
Publisher
이화여자대학교 대학원
Degree
Master
Advisors
이동환
Abstract
최근 한국의 몇 년 새 가계의 소득보다 빚이 더 늘어나 가처분소득 대비 부채 비율이 높아진 것으로 나타났다. 전세난으로 집을 사려는 수요가 늘면서 올 2분기 가계의 자금잉여 규모가 5조 원 가까이 줄고 금융부채는 분기 기준으로 사상 최대 폭으로 늘었다. 본 연구는 최근 3년간의 패널자료를 활용하여 가계부채의 현황을 살펴보고자 하며 추세를 파악해 보고자 한다. 먼저 패널자료 간에 독립임을 가정하여 각 연도별로 부채 여부와 부채 증가원인에 영향을 미치는 유의미한 변수를 확인하고 그 차이를 비교한다. 앞선 방법은 패널자료간의 독립성을 가정한 반면 패널자료에 상관관계가 존재 할 것을 가정하여 부채 여부와 부채 증가원인을 반응변수로 하는 일반화 추정 방정식(GEE) 분석을 시행한다. 본 연구의 자료는 통계청, 금융감독원 및 한국은행이 공동으로 주관하는 2012년에서 2014년까지 3년에 걸쳐 패널 조사한『가계금융 · 복지조사』금융부분 설문조사를 자료로 이용하였다. 본 연구를 통해 최근 가계 부채의 부채 여부와 부채 증가원인에 영향을 끼치는 변수가 무엇인지 살펴봄으로써 최근 가계부채의 패턴 변화를 파악해본다. 패널자료를 이용한 GEE 분석을 활용하여 증가세로 이슈화 되고 있는 가계 부채의 추세를 파악하는데 도움을 줄 수 있을 것이다.;This paper investigate the general patterns and trends of Korean household debt by using “Household Financial· Welfare Research”panel data collected over 3 years, from 2012 to 2014, under the jointly supervision of the National Statistical Office, the Financial Supervisory Office, and the Bank of Korea. To identify the relevant variables that are associated with the debt status and cause of dept, we applied the generalized estimating equation (GEE) approach to allow the various correlation structure among the panel data and compare the general linear models based on the independence assumption. According to the analysis by GEE, It seems that as more total assets are possessed, so the possibility to get indebted increases. However, Earnings fractile is not necessarily in such a relation. As a matter of fact, In the sequence of the third group, the fourth group, the fifth group, the second group, the first group of earnings, People were more likely to get themselves into debt. Accordingly, overall, it is more probable that as the income increases, the factors such as real estate arrangement apart from residential housing, stock investment, business funds & debt repayment will be the cause of increased liabilities. At the same time, As more total assets are possessed, so debt repayment will more likely be the cause of increased liabilities. When third to fifth groups in total assets are compared with first group, their burden for debt payment increases at least, approximately more than three times over their burden for marriage, medical, living expense, etc. The similarity of the pattern between earnings fractile and total asset fractile is that both debt payment and another variable(property arrangement excluding residential housing, stock investment, business fund) showed the same pattern in increase & decrease, although they are different in gradient. The cause of increased debt related to residential housing showed rather symmetrical pattern with these two. The aspect of these patterns became obvious, as it goes from 2013 to 2014. Furthermore, because the relation of incomes and assets is mixed complexly, thorough perusal from various angles is needed.
Fulltext
Show the fulltext
Appears in Collections:
일반대학원 > 통계학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE