주택담보대출 차입자의 금리선택에 관한 연구 (A Study on the Determinants of Choice between ARMs and FRMs)

서지정보

· 발행기관 : 한국금융학회
· 수록지 정보 : 금융연구 / 23권 / 2호 / 133 ~ 151페이지
· 저자명 : 박성욱, 박갑제, 김태혁

초록

이 논문은 은행에서 실제로 집행된 주택담보대출자료를 이용하여 주택담보대출 차입자의 위험관
련 특성변수와 금리차가 차입자의 금리선택에 미치는 효과를 실증적으로 추정하고 있다. 추정결과
를 요약하면 다음과 같다. 첫째, 고정금리형 주택담보대출금리와 변동금리형 주택담보대출금리의
차이가 차입자의 금리선택에 유의적인 영향을 미치는 것으로 나타났다. 둘째, 위험회피도와 관련
되는 변수들 중 나이, 성별차이, 직업유형 및 주택구입가격 등이 금리선택에 유의적으로 영향을 미
치는 것으로 나타났다. 즉, 나이가 많을수록, 남성일수록, 자영업 및 개인기업일수록, 그리고 부
(wealth)의 수준이 높을수록 변동금리형을 선택하는 것으로 나타났다. 반면, 결혼여부, 맞벌이 여부
등은 금리선택에 유의적인 영향을 미치지 않는 것으로 조사되었다. 셋째, 대출금액과 대출만기가 증
가할수록 즉, 위험수준이 증가할수록 상대적으로 안전한 고정금리형을 선택하는 것으로 나타났다.

영어초록

This paper examines the determinants of borrower demand for ARMsadjustable
rate mortgages). To do this, this paper estimates a probability
model of the choice between adjustable-and fixed-rate mortgage using the
home mortgage data compiled by a bank that operates in the Korea bank
markets. The size of sample in this paper is 30,364. The explanatory variable
in the probit and logit probability model includes borrower characteristics
and the FRM-ARM rate differential. To obtain the FRM-ARM rate differential,
we must estimate the unobservable interest rate not chosen by the
borrowers. In order to solve this problem, we adopt the Heckit and OLS
method. And then the FRM-ARM rate differential are calculated by difference
between the predicted values of interest rates. But the Heckit method shows
the poorer performance in predicting the observed interest rate than the OLS
method. Thus, we focus on the estimated results of regression model based
on the FRM-ARM rate differential that is calculated by the OLS method.
An important estimation results are as follows. First, our estimation
results shows that the FRM-ARM rate differential significantly affects the
choice between ARM and FRM. As the FRM-ARM rate differential is escalated,
the borrowers are likely to finance through the adjustable-rate mortgage
rather than the fixed-rate mortgage. This result is the same one as
expected by the theory. Second, our results reveals that the variable related
to the risk aversion such as age, sex and wealth are key variables in explaining
the choice. According to the theoretical studies of Alm and Follain(1987)
and Bruckner(1986), it is suggested that a borrower’s risk aversion, his discount
rate for future consumption may play roles in the FRM-ARM choice.
This imply that an individual with low risk aversion and a high discount
rate will find it easy to live with uncertain future mortgage payments. The
estimated positive coefficient of variables such as age and wealth indicates
that a higher age and wealth raises the ARM choice probability. Third, as
the risk level arises, the borrowers are likely to be serviced in FRM. The
reason that the interpretation above can be possible is because the estimated
coefficient of variable such as maturity and Loan serviced are positive.
This paper is the first in that it empirically explores the determinants
of choice between the ARM and FRM in the Korean bank markets. But this
paper has also the limitation in that it uses the home mortgage data of only
a bank rather than nation wide data.

참고자료

· 없음