최근 주식투자 인구가 확대되고 있어 투자자에 대한 이해가 중요해지고 있다. 본 연구는 주식투자자를 직접과 간접 투자자로 구분하고 이들 수요에 세대별 인구구조가 미치는 영향을 분석하였다. 주요 분석 내용은 다음과 같다. 첫째, 세대별 인구구조가 직·간접 주식투자 수요에 미치는 영향이 상이하게 나타났다. 특히 노년 세대 규모는 직접 주식투자 수요에 양(+)의 영향을 미쳤으며, 이는 생애주기 가설과 달리 직접 주식투자에 대한 노년 세대의 수요가 있음을 의미한다. 둘째, 노년 세대의 시작연령을 60세에서 75세까지 1세씩 상향하여 분석한 결과, '73세 이상'의 노년 세대까지 직접 주식투자에 양(+)의 영향을 미치는 것으로 나타났다. 이는 60~73세의 인구의 직접 주식투자에 대한 수요가 지속됨을 의미한다. 반면, 간접 주식투자 수요에는 세대별 인구비중이 유의한 영향을 미치지 않았다. 셋째, 노년 세대와 기대수명 증가율의 '상호변수(interaction variable)'는 직접 주식투자 수요에 유의하게 음(-)의 영향을 주는 것으로 나타났다. 이는 기대수명 증가에 따른 장수위험(longevity risk) 등으로 노년 세대의 직접 주식투자에 대한 수요가 감소하는 것으로 보인다. 본 연구는 노년 세대의 직접 및 간접 투자에 대한 이해도를 제고하였다는 점에서 의의가 있으며, 타겟데이트펀드(Target Date Fund) 등 금융상품의 설계에서도 활용될 수 있다.

It has become more important to understand investor behaviors as the size of stock investor increases in Korea. Individuals invest in stocks in two ways: direct and indirect. For example, individuals can invest in stocks through publicly placed equity funds which invest principally in stocks.
According to the life cycle, older investors will have a lower demand for equity compared to middle-aged investors due to risk aversion increasing with age. The life cycle of risk aversion is widely accepted in academia, as well as in the marketplace for practical investment advice.
On the other hand, 2011-2018 Household Finance Welfare Survey by Statistics Korea says that the median of household owner's asset by three generation, i.e young, middle and old, is 226.72 million won and 492.28 million won and 161.85 million won respectively. In contrast, according to the 2017 Korean Labor and Income Panel Study by Korean Labor Institute, the financial income per month of three generation is about 0.1 million won, 0.3 million won and 0.4 million won, respectively. It implies that the old generation has the most financial income.
This study intends to fill the gap between the life cycle risk-aversion hypothesis and the market observations. This study empirically investigates how demographic structure affects direct and indirect equity demands. For the investigation, first of all, we divide investors into two groups such as direct investors and indirect investors. Second, we increase old generation starting ages from 60 to 75 years old.
Main findings are as follows. First, demographic structure gives different impact on direct and indirect equity demands. Particularly old generation size gives significant and positive impact on direct equity demands. This result is against the 'life cycle risk-aversion hypothesis' which argues weaker demands for equity from the old generation.
Second, through increasing the starting age of old generation by one year, we find that from 60 to 73 age cohort gives positive impact on direct equity demands. Meanwhile we do not find significant impact of demographic structure on indirect equity demands.
Lastly, the interaction between life expectancy and old generation size provides negative impact on direct equity demands. It means that longevity from longer life expectancy lessens the demands for equity. Our work improves understanding on old generation's impact on direct and indirect equity demands. Also the empirical results can be used for designing Target Date Funds.
This study shows that direct and indirect investors are different in terms of stock investment behavior. Therefore, we need more investigations in order to enhance the understanding about stock investors particularly for old generation.

본 연구에서는 1976년부터 2020년까지의 분기별 자료를 이용해 과연 경제성장률이 명목 또는 실질 단기금리와 어떤 관계를 갖고 있는지를 살펴보았다. 레짐스위칭 모형 추정결과 두 변수 간의 시변적인 동조화 확률은 전반적으로 높은 양(+)의 값을 가지나 외환위기 전후로 크게 하락한다. 인과관계분석에 따르면 GDP 성장률이 떨어지는 경우 외환위기기간을 제외하고는 금리가 하락하는데 금리 하락폭은 외환위기 이전 기간에 비해 외환위기 이후 기간에 훨씬 크며 통계적으로 유의적이다. 또한 동일한 크기의 충격에 대한 금리 하락폭은 외환위기 이후부터 코로나 사태 이전 기간 동안에 점차 커지는 경향을 보인다.

This paper analyzes the relationship between the real GDP growth rate and nominal or real short-term interest rates using quarterly data from 1976 to 2020. First of all, depending on how the real interest rate is measured, real interest rates show a big difference in the period before the currency crisis, while relatively similar movements in the period after the currency crisis. Meanwhile, the regime switching models show that high growth and high interest rates in the period before the currency crisis rapidly change to low growth and low interest rates in the period after the currency crisis. According to the time-varying regime switching model, the probability of time-varying synchronization between the two variables not only has a positive (+) value even before the currency crisis, but also has been on the rise overall until recently, although it has fallen significantly just before and after the currency crisis.
This study also examines whether the economic growth rate can affect interest rates through the causal relationship analysis as well as correlation and synchronization analyses. The estimation results of the linear VAR model with crisis dummy variables and trend terms indicate that interest rates fall when economic growth falls, as claimed by the new classical school. In the case of the time-varying VAR model as well as the linear VAR model, interest rates fall significantly in both nominal and real interest rates due to the impact of falling economic growth during the post-crisis period compared to the pre-crisis period. This seems to be related to a major change in monetary policy, which changed price stability from the previous monetary target to a single final goal under the sixth revision of the Bank of Korea Act on December 31, 1997, and disclosed the call rate as an operating goal from May 1999. However, even though the economic development clause(Article 3, Paragraph 2), which had been stipulated together with the stability of the currency value(Article 3, Paragraph 1) before the 6th revision of the Bank of Korea Act, was omitted in the 6th revision, the economic growth rate and interest rate seem to have a closer positive relationship after the revision. This fact means that the central bank is always concerned with growth rates and use its operating target, as the Taylor rule shows, regardless of the text of the law.
In addition, the rate of decline in interest rates in response to shocks of the same size tends to gradually increase during the period from the post-currency crisis to before the outbreak of COVID-19. These empirical analysis results have important implications in relation to monetary policy at a time when the potential growth rate continues to decline. If interest rates fall more as the growth rate decreases, they face a zero lower bound problem as interest rates get closer to zero. In this case, as the aggregate demand curve goes upward rather than downward, the automatic adjustment function of the economy is lost, and if left unattended, the economic recession will worsen. In this situation, advanced countries such as the United States have overcome this difficulty through non-traditional monetary policies such as quantitative easing, credit easing, and forward guidance. But it is unclear whether the Korean economy, which is a small open economy and the won is not the key currency, can implement these policies without major side effects.

본 연구는 단기이자율 확률과정의 특성을 결정하는 핵심 파라미터인 수준효과를 포함하는 단기이자율 확률변동성모형에 대한 효율적인 추론전략과 MCMC(Markov chain Monte Carlo)를 이용한 베이지언 사후표본추출 알고리듬을 제시한다. 수준효과를 설명하는 CEV형태의 확산함수에 포함된 파라미터들의 약식별성과 단기이자율의 단위근에 가까운 지속성으로 인한 파라미터들 간의 강한 상관관계는 단기이자율의 안정성을 특징짓는 수준효과 파라미터들에 대한 식별과 효율적인 추정을 어렵게 한다. 본 연구는 이에 대한 해결방법으로 모수재설정 단기이자율 확률변동성모형과 베이지언 MCMC알고리듬을 제시하고 모의실험을 통한 성과평가를 함께 제시한다. 3개월 만기 미국재무부채권 수익률 주간자료에 대한 실증분석을 통해 본 연구에서 제시하는 모수재설정 모형이 기존 연구에서 제시된 모형설정들보다 수준효과 파라미터 식별과 추정에 효율적임을 확인할 수 있었다. 또한 모수재설정 모형을 이용하여 얻은 수준효과 파라미터의 크기에 대한 베이지언 가설검정 결과는 수준효과의 크기가 0.5라는 귀무가설을 강하게 지지하는 것으로 나타났다.

We propose an efficient Bayesian inference strategy and a new MCMC algorithm for stochastic volatility models of short term interest rates with level effect. The strong correlations among parameters caused by weak identifiability of level effect parameter and near-unit root persistency of short rates makes difficult the identification and efficient inference of the key parameter, the level effect parameter of CEV diffusion function, characterizing stationarity of short term interest rates.
Our strategy for efficient Bayesian inference for stochastic volatility models with level effect is a kind of centered parameterization of moving the level effect component which is difficult to identify, included in the diffusion function of CEV observation equation to the log-volatility state equation. Our centered parameterization is contrast to the non-centered parameterzation suggested in Strickland et al. (2008), Kastner and Frühwirth-Schnatter (2014), etc. as an efficient inference strategy for stochastic volatility models. Considering that basic stochastic volatility model with level effect is a model with the characteristics of partial NCP in itself, and that the information of the observed data may not be sufficient to identify the level effect parameter, the non-centered parameterization of Strickland et al. (2008) which moves the parameters of state equation to the observation equation, can make it even more difficult to identify the level effect parameter. The reparameterized model proposed in this study can solve the problem of weak identification problem of level effect parameter and the problem of inefficient inference due to the strong correlation between level effect parameter and mean parameter of log-volatility state equation.
We validate the performance of our reparameterization strategy and Bayesian inference algorithm through both simulation experiments and empirical analysis comparing the efficiency of posterior sampling for the level effect parameter with that of the suggested parameterizations in the existing literatures. Empirical analysis results on weekly data of 3 month U.S. T-bill yields reveal that our reparameterized model for Bayesian inference performs better in the identification and efficiency of posterior sampling of the level effect parameter than existing reparameterized models. The better performance in posterior sampling can facilitate efficient Bayesian hypothesis tests for parameters of interest. A Bayesian hypothesis test result for the size of level effect parameter supports the null hypothesis of 0.5, which implies the drift-induced stationarity of 3 month U.S. T-bill yields.