본 연구는 국민연금관리공단과 국민연금연구원의 국민노후보장패널(KReIS)의 1차년~3차년도 자료를 이용하여필요노후생활비와 개인연금불입액간의 동태적인 변화를 살펴보기 위해 종단분석 방법 중 하나인 자기회귀교차지연모델을 적용하였다. 이와 더불어 필요노후생활비와 개인연금 불입액의 관계가 부동산과 금융자산 소유여부에 따라달라지는가를 살펴보기 위해 부동산과 금융자산 소유여부에 따른 다집단 분석을 실시하였다. 그 결과 필요노후생활비의 자기회귀 계수는 모두 통계적으로 유의미하였으나 개인연금불입액의 자기회귀계수는 부분적으로 통계적 유의성이 확인되었다. 필요노후생활비의 개인연금불입액에 대한 교차지연계수는 검증되지 않았지만 개인연금불입액의필요노후생활비에 대한 교차지연계수는 통계적 유의성이 검증되었다. 그러나 부동산과 금융자산소유여부에 따른 차이는 검증되지 않았다. 본 연구의 결론은 구체적으로 다음과 같다. 첫째, 필요노후생활비와 개인연금불입액간의 자기회귀계수가 검증됨으로 필요노후생활비와 개인연금불입액은 시간의 흐름에 따라 점차 증가하는 것으로 나타났으며 이전 시점의 필요노후생활비와 개인연금불입액은 이후 시점의 필요노후생활비와 개인연금불입액에 영향을 주는것으로 나타났다. 결국 개인연금의 경우 한번 가입하게 되면 지속적으로 유지하게 되는 경향이 있으며 불입액이 점차 증가한다고 할 수 있다. 또한 필요노후생활비의 경우 한번 산정하게 되면 시간이 흐를수록 즉, 노후와 가까워질수록 좀 더 현실적으로 평가하여 증가하는 경향이 있다. 둘째, 본 연구에서는 필요노후생활비의 개인연금불입액에대한 교차지연계수는 통계적으로 검증되지 않았지만 개인연금불입액의 필요노후생활비에 대한 교차지연계수는 검증되었다. 그러나 필요노후생활비의 1차년도 값과 개인연금불입액의 1차년도 값의 상관계수는 통계적으로 유의미하지않는 것으로 나타나, 필요노후생활비와 개인연금불입액의 관계는 동일시점에서는 검증되지 않는다는 사실을 보여주고 있다. 셋째, 본 연구에서는 부동산과 금융자산 소유여부에 따라 필요노후생활비의 개인연금불입액에 대한 영향력은 달라지지 않는 것으로 나타났다. 본 연구에서는 필요노후생활비의 개인연금불입액에 대한 영향력은 확인되지 않았지만 개인연금불입액의 필요노후생활비에 대한 영향력은 검증됨으로써 개인연금불입액과 필요노후생활비의 관계를 어느 정도 설명했다는 점에서 의의가 있다. 또한 필요노후생활비와 개인연금불입액의 자기회귀효과와, 개인연금불입액의 필요노후생활비에 대한 교차지연효과 등이 검증됨으로써 개인연금과 노후자금에 대한 동태적인 변화를 위해서 종단연구가 필요하다는 사실을 보여주었다. 특히 1차년도 필요노후생활비와 1차년도 개인연금불입액의 상관관계는 통계적으로 유의미하지 않았지만 개인연금불입액의 필요노후생활비에 대한 교차지연계수인 1차년도의 2차년도에 대한 영향력 그리고 2차년도의 3차년도에 대한 영향력이 통계적으로 유의미한 것으로 나타나 개인연금 불입액과필요노후생활비간의 종단연구의 효용성을 확인할 수 있었다.

The low interest rates and aging environment are considered to be the most representativeelements to threaten the financial aspects of individuals in Korea. In particular, preparationfor living expenses after retirement has become one of the major concerns with rapid progress ofaging society, which indicates increase in the period of survival after retirement and longevityrisk, a possibility that retirement assets are depleted ahead of the time before death. However,Korean people's preparation for living costs for aging is still lacking. The longevity risk refersto the expense related to lifespan, and it also means risk of living longer than the preparedresources. The lack of the adequate preparation for the rapid progress of aging leads to increasein the longevity risk, and to decrease this risk, so the need for pension products is emphasizedin terms of guaranty of predetermined income for a certain period of time. In this sense, lifeannuity is expected to provide ‘longevity insurance’ to hedge the longevity risks in that it canguarantee a fixed income until the time of death. However, despite these advantages, personalpension has not widely been used as a part of retirement savings plan. In particular, the lifeannuity has been confronted with the “annuity puzzle” phenomenon due to inactivation offinancial market despite its characteristic of the most obvious countermeasure for the longevityrisk. This study attempted to identify that one of the causes of the “annuity puzzle” is the lackof awareness of the pension and retirement funds through checking if the amount paid forpersonal pension increases as evaluation finds that the need for retirement funds increases. However, although the need for living costs after retirement is well perceived, the necessity of pension can be underestimated because other assets such as real estate and financial assetsthat can be easily turned into cash upon demand exist. Accordingly, this study also tried toinvestigate the effect of the presence of real estate and financial assets on the relationshipbetween the need for retirement funds and the amount paid for personal pension. In addition,it is necessary to take the passage of time into consideration since the recognition of the needfor retirement funds tends to change over time. In other words, it is possible that the need forthe retirement funds is likely to change specifically with age. Accordingly, autoregressivecross-lagged modeling (ARCL), one of the longitudinal analysis methods was applied in thisstudy through using 1st to 3rd year-data of Korea retirement and income study (KReIS) fromNational Pension Service and National Institute of Pension. From its results, both autoregressivecoefficient of personal pension payments and appropriate living expenses needed in the old agewere found to be statistically significant, and the cross-lagged coefficient on appropriate livingexpenses needed in the old age of the personal pension payments was verified. On the otherhand, differences depending on the presence of real estate and financial assets were not verified. The implications based on these results are as follows. First, the verification of autoregressive coefficient between appropriate living expenses neededin the old age and personal pension payments showed that appropriate funds needed in the oldage and personal pension payments increase as time progresses, and appropriate living expensesneeded in the old age and personal pension payments of an earlier time turned out to have aneffect on those of a later time. In conclusion, it can be safely said that personal pension tendsto be maintained consistently and its payments gradually increase once it is signed up. Second, the cross-lagged coefficient on appropriate expenses needed in the old age of thepersonal pension payments was verified statistically in this study. However, the correlationcoefficient between 1st year values of appropriate expenses needed in the old age and those ofpersonal pension payments turned out to be statistically insignificant, which indicates the factthat the effect of awareness of the retirement funds on the increase in pension payments is notverified in the same point of time. Third, this study results also found that the presence of real estate and financial assets havelittle effect on the personal pension payments of appropriate expenses needed in the old age. Asa method to prepare living costs needed in the old age, encashment of real estate and financialassets can be considered in addition to pension. However, the effect of real estate and financialassets was not verified in this study. It is expected that this study will provide basic materials to seek conditions required toactivate personal pension as a part of retirement savings plan. In addition, it can provide more in-depth implications in customer's awareness of the retirement funds and selection of personalpension through longitudinal analysis of dynamic causation on the personal pension paymentsof retirement funds away from the framework of cross-sectional analysis on the existingpersonal pension and retirement funds.

private pension, appropriate expenses needed in the old age, personal pension payments, autoregressive cross-lagged modeling, longitudinal analysis