이 연구는 국민노후보장패널조사(KReIS)의 1-3차년 자료를 이용하여, 중․고령자 단독가계의 의료비지출규모와의료비지출비중 추이에 가구주의 소득, 건강상태 및 기타 인구사회학적 특성의 변화가 미치는 영향을 분석하였다.
1차년도 (2005년) 조사 당시 50세 이상 단독가계의 가구주로서 2차년도 (2007년), 3차년도(2009년) 조사에도결측값 없이 응답한 중․고령자 단독가구주 683명 1,794건의 패널자료가 분석에 사용되었다. 의료비지출의 선형회귀모형을 OLS, 개인고정효과 및 개인랜덤효과 모형으로 분석한 결과는 다음과 같다.
첫째, 의료비 소비는 정상재이며 필수재의 성격을 보인다. 둘째, 중․고령자 단독가계의 의료비지출규모는 질환이나 장애가 있을 경우, 연령이 높을수록 증가하였으며, 이것은 건강상태가 의료비지출에 주요한 영향요인임을 알 수있다. 셋째, 조사시점이 최근으로 올수록 의료비지출규모가 감소하였는데, 이는 2008년 7월에 시행된 노인장기요양보험과 2008년 경제위기로 인한 소비지출의 위축이 원인이 되었을 것으로 보인다. 넷째, 교육수준이 높을 경우의료비지출비중이 낮은 것으로 보아 교육을 통한 인적자본이 의료비지출의 중요한 영향요인임을 알 수 있다. 다섯째, 여성은 남성에 비해 전반적으로 의료비지출규모가 크며, 남성은 의료비지출의 소득탄력성이 여성에 비해 훨씬높을 뿐 아니라 질환이나 연령에 따른 효과도 여성에 비해 더 뚜렷했다. 여섯째, 자산, 부채, 취업 등의 경제적 지위도 의료비지출과 유의한 관련을 보이나 이것은 인과관계라기보다는 내생성으로 인한 상관관계인 것으로 보인다.
이러한 연구결과는 저소득층 중․고령자 단독가계 가구주를 위한 질병 예방, 치료 및 의료비 지출절감정책, 질환이 있는 중․고령자 단독가계 가구주를 위한 치료 및 의료비 지출절감정책, 예비노인기부터 건강관리 및 질병예방지원정책, 저교육 수준의 중․고령자 단독가계 가구주를 위한 청․장년기부터의 건강관리와 질병예방교육 프로그램,여성가구주 전체와 특히 저소득층 남성가구주를 위한 의료비지출절감정책이 필요함을 시사한다.
Korea is approaching an aged society faster than most OECD countries. By 2014, the 65and older population will account for 14 percent of the population. In 2005, every fifth Korean household was a single-person household mainly due to the elderly living alone. The elderly are facing many problems today, but healthcare and economic hardship are their utmost concern.
This study uses the 2005, 2007 and 2009 surveys of the Korean Retirement and Income Study (KReIS) to explore how income, health status, and socio-demographic characteristics determine health care expenditures among older Koreans in single-person households. This study is different from previous studies in several aspects: first, it takes advantage of panel data analysis; second, the sample includes the near elderly, which age group is identified as significant consumers of health care; and third, it adds to the literature by contrasting men against women as health care consumers.
The sample consisted of 1,794 observations from 683 single householders aged 50 years or older in 2005. Both the size and share of health care expenditures, the latter of which was defined as a fraction of total consumption expenditures, were regressed on income, health status, and other socio-demographic characteristics of the household. The socio-demographic variables were selected based on the literature and included age, gender, cohort, education,area of residence, assets, debts, home ownership, employment status, health insurance coverage,and receipt of other social insurance and welfare benefits. Regression coefficients were obtained through Ordinary Least Squares (OLS), individual fixed effects, and random effects models.
The main findings can be summarized as the following:First, in all models income is positively associated with health care expenditures, with the fixed-effect income elasticity at 0.14, suggesting that health care consumption is a normal good and necessary.
Second, having disease or disability is positively associated with both the size and the share of health care expenditures. In addition, after controlling for the survey year and birth cohort, older age is positively associated with the size of health care expenditure. This suggests that health status is a significant predictor of health care expenditures.
Third, the size of health care expenditures has decreased over the 2005-2009 period. It can be inferred that macroeconomic and institutional factors like nationwide expansion of long-term care insurance in 2008, and economic recession during the late 2000’s played a role in reducing health care expenditures by the elderly.
Fourth, education is negatively associated with the share of health care expenditures as a fraction of total consumption expenditures of the household. This appears to suggest that human capital plays a role in reducing the need for health care expenditure at an older age.
Fifth, when income, health status, and other socio-demographic characteristics were controlled,women spend significantly more on health care than men. This gender difference is robust in both dependent variables - the size as well as the share of health care expenditures. Income elasticity of health care expenditures is also much greater for men than for women. Furthermore, health care expenditures by men vary more markedly by age and health status compared to those by women.
Sixth, economic status such as assets, debt, and employment status are also significant,most likely due to endogeneity arising from inherent healthiness.
Findings from this study offer several policy implications. First, it is suggests that the health care policy for the older single-person households target low-income households in all phases of prevention, treatment, and cost reduction. Second, such policy should particularly focus on the older single-person households who have disease or disability. Third, it is important to provide support for disease prevention before the individuals reach old age and develop educational programs on health management and disease prevention for the middle-age population.
Fourth, this study suggests that policies that aim to reduce health care costs should be genderspecific.
Such policies should be directed at older women in general and older men especially in low-income households.