본 연구에서는 국민노후보장패널 4차년도부터 6차년도까지 자료를 바탕으로, 고령가계를 은퇴 여부와 소득 원천 에 따라 2단계로 구분하여 유형별로 살펴보고, 유형별 소비함수를 추정함으로써, 고령가계의 경제적 상태를 정확하 게 이해하고 진단함으로써, 효과적인 지원정책을 모색하고자 하였다. 이를 위해 국민노후보장패널 중 최종적으로 총 1,002가구의 자료를 분석하였다. 잠재프로파일분석을 통해 고령가계를 유형별로 구분하고, 교차분석, 집단별 평균 차이 분석을 통해 유형별 특성을 살펴보았다. 또한, 유형별 소비함수를 분석하기 위해 패널회귀분석을 실시하였고, 확률효과모형을 활용하였다. 본 연구의 결론은 다음과 같다. 첫째, 고령가계는 은퇴 여부와 소득 원천에 따른 2단계 구분을 통해 6개의 유형으로 세분화할 수 있으며, 비은퇴가계는 근로소득형 가계가 많은 반면, 은퇴가계는 사전이 전중심형 가계가 많은 것으로 나타났다. 둘째, 유형별 고령가계의 특성이 서로 다르게 나타나, 유형별 맞춤형 진단 과 대응이 요구된다. 비은퇴가계 중 근로.연금소득형 가계는 근로소득형보다 상대적으로 연령은 높고, 근로소득의 수준은 낮은 것으로 나타나 소일거리를 포함한 경제활동에 지속적으로 참여하고 있으며, 은퇴가계 중 사적이전소득 형 가계는 부채가 소비지출에 양의 영향을 미치는 것으로 나타나 일부 부채를 통해 소비지출을 충당하고 있는 것으 로 나타났다. 이를 통해 개별 유형의 가계가 경험하는 경제적 문제를 해결하기 위해 세분화한 맞춤형 해결방안을 모색할 필요가 있다. 셋째, 고령가계의 유형별 경제 수준의 차이를 살펴본 결과, 은퇴 전 충분한 노후준비가 되어 야 얻을 수 있는 소득원인 재산소득과 연금소득을 중심으로 한 은퇴가계 유형의 경제적 복지 수준이 높게 나타나 은퇴 전 충분한 노후준비의 중요성을 확인하였다. 따라서 아직 은퇴하지 않은 젊은 세대에게 은퇴준비의 중요성을 끊임없이 강조하고, 교육을 제공하는 노력과 함께 가계의 은퇴준비를 지원하기 위한 정책적 대안을 마련할 필요가 있다. 넷째, 고령가계를 대상으로 한 근로소득의 창출은 가계의 경제적 복지에 상당히 중요한 영향을 미친다는 것 을 알 수 있었다. 다섯째, 고령가계의 소비지출은 가계의 유형별로 다른 소득원천에 의해 영향을 받으며, 그 영향 력도 유형별로 다르게 나타났다. 특히 고령가계의 소비지출은 가계의 유형을 결정하는 주된 소득원천보다 근로소 득이나 연금소득과 같은 정기적 성격을 갖는 소득원천에 의존하는 경향이 나타나 정기적인 소득의 중요성을 확인 할 수 있었다. 따라서, 공적연금을 통한 노후소득 보장 수준이 취약한 현실을 고려하여 고령가계를 대상으로 한 일 자리 정책을 적극적으로 지원하는 한편, 근로기간을 연장함으로써, 은퇴 시기를 상대적으로 늦추는 방안을 고려해 볼 필요가 있다.
In this study, based on the data from the 4th to the 6th year of the KReIS, the elderly households were classified two stages, according to retirement status and income sources and consumption functions were estimated. The purpose of this study is as follows. First, it is intended to distinguish the elderly households by disaggregating them by using their retirement status and income sources step by step. Specifically, the types were identified by dividing the elderly households into retirement and non-retirement households according to their retirement status and classifying them again according to income sources. Second, it is intended to identify the key characteristics of each type of elderly households according to their retirement status and income sources. Third, through the panel regression analysis, the consumption function is estimated for each type of elderly household according to their retirement status and income source to identify the impact of each source of income on consumer spending. Through this study, the problems that can ultimately be experienced by the elderly households in economic terms were diagnosed in advance and solutions were sought. This study has the following implications: First, the KReIS data suitable for research purposes were used and, in particular, the three-year data were analyzed using the 4-6th (2011-2015) data to include differences in individual households and differences over time. Besides, this study classified the entire elderly households into more granularity by phasing out the elderly household through two axes: retirement status or income source. Furthermore, the purpose of this study was to assess the condition more accurately and find effective countermeasures by estimating consumption functions differently by looking at the types of sources of income and their influence on consumer spending. We analyzed 1,002 households in the KReIS. Of the total KReIS data, households that meet all three conditions were analyzed and the data were extracted. First, a balanced panel was established for households that responded to all three numbers until 4th and 6th years and used in the analysis. Second, only single-person and married-person households were studied. This is because households that live together with their actual sons and daughters may overstate their income and consumption expenditure. Third, only households whose retirement status remained the same were studied from the 4th to 6th years of the KReIS. In other words, only households that have not been retired or have been retired continuously for 4th and 6th years are included in the analysis data. Households who experienced retirement within the period(4th and 6th years) were excluded from the analysis. This is because if one experienced retirement within the period, there could be a big change in economic conditions, including income and consumer spending. Therefore, it excluded cases where it was difficult to analyze the differences in characteristics of elderly households depending on whether they retired or not. In this study, the latent profile analysis was performed using Mplus to distinguish types of elderly households based on the data extracted. Also, STATA 13 was used to cross-tabulation and analyze the average difference between groups to examine the characteristics of each type based on the type of elderly households. For analysis of consumption function by type, panel regression analysis was performed and the random-effects model was used in detail. This study focused on the elderly households whose life span is increasing after retirement and the importance of that period is becoming increasingly important to them, looking at the economic characteristics of the elderly households and diagnosing problems to find appropriate solutions. The conclusion of this study is as follows. First, elderly households can be subdivided into 6 types. Non-retirement households have more labor-oriented households, while retirement households have more private transfer-centered households. It can be seen that many elderly households now rely on private transfer income, and when studying elderly households in the future, it provides suggestions that analysis is needed considering type-specific characteristics. Second, the characteristics of elderly households by type were different. In the non-retirement households, the labor and pension types are relatively older and the level of earning an income is lower than the labor-oriented type. They can be seen as continuing to engage in economic activities. Besides, private transfer-centered types in retirement households show that debt affects consumption expenditure positively, indicating that some debts cover consumption expenditure. This required a tailored solution to address the economic problems experienced by individual types of elderly households. Third, types focused on property income and pension income that can be accumulated in advance of retirement are higher than other types of economic welfare. Therefore, there is a need to emphasize the importance of retirement preparation for younger generations who have not yet retired, and to provide policy alternatives to support the retirement preparation along with the efforts to provide education. Fourth, it can be seen that the generation of earned income for elderly households has a considerable influence on economic well-being. As the elderly population, which is still motivated and capable of working, is likely to increase with changes in its future demographic structure, efforts need to be made not only in re-employment policy but also in matching legal and social retirement age with actual retirement age. Fifth, the consumption expenditure of elderly households is influenced by other income sources, and its influence is also different by type. Especially, their consumption expenditure tends to depend on the income source having regular characteristics such as earned income or pension income rather than the main income source that determines the type of households. Therefore, considering the reality that the level of guarantee of old-age income through a public pension is low, it is necessary to consider a way to delay the retirement period relatively by proactive supporting the employment policy for elderly households and extending the working period.