<첫 번째 논문>의 연구요약은 다음과 같다. 거시경제학에서 인플레이션의 dynamics를 연구할 때, New Keynesian Phillip Curve의 이해는 중요한 의미를 가진다. 특히, New Keynesian Phillips curve에서 현재의 인플레이션의 움직임을 설명하는 변수로서 경제주체들의 미 ...

<첫 번째 논문>의 연구요약은 다음과 같다. 거시경제학에서 인플레이션의 dynamics를 연구할 때, New Keynesian Phillip Curve의 이해는 중요한 의미를 가진다. 특히, New Keynesian Phillips curve에서 현재의 인플레이션의 움직임을 설명하는 변수로서 경제주체들의 미래의 인플레이션에 대한 기대치와 더불어 과거의 인플레이션이 중요한가 하는 것은 많은 경제학자들 사이에서 논란의 대상이 되어 왔다. New Keynesian Phillips curve에서 과거의 인플레이션이 중요한 역할을 하는가 하는 것은 중요한 정책적 함의를 가지고 있다. 본 연구에서는 중앙은행의 목표 인플레이션, 혹은 steady-state 인플레이션이 1960~2005년 동안의 표본기간 동안에 일정하지 않은 점을 고려하여 계량분석을 하면, New Keynesian inflation에서 과거 인플레이션의 역할이 통계적으로 유의하지 않다는 것을 보이고 있다. 일반적으로, New Keynesian Phillips curve를 추정할 때, 내생성의 존재를 고려하여 도구변수 추정 혹은 GMM (Generalized Method of Moment)를 사용하는데, 중앙은행의 목표 인플레이션을 대표하는 모형의 파라메타에 구조 변화가 있고 구조 변화의 시점이 알려지지 않은 상황에서의 모형의 추정이 필요하다. 따라서, 본 연구는 저자가 개발한 계량방법론인 ‘Regime switching model with endogenous regressors' 의 의미 있는 응용임과 동시에, 그 응용을 통하여 인플레이션의 dynamics의 이해를 높일 수 있었다.

<두번 째 논문의 연구요약>은 다음과 같다: 본 논문은 국면전환모형(Markov-swtiching models)에 내생성(endogeneity)이 존재하는 경우의 문제해결을 꾀하고 있다. 저자의 2004년도 논문에서도 같은 주제를 다루고 있지만, 본 논문에서는 기존 논문의 가정을 완화하여 보다 현실적인 문제를 다루고 있다. 즉, instrumenting equation 과 우리가 추정하고자 하는 equation에서의 국면전환의 패턴이 다른 보다더 현실적인 경우를 다루고 있다. 본 논문에서는 이러한 상황하에서 joint estimation이 가지는 ‘curse of dimensionality’의 문제점을 해결하는 방안으로 이단계 추정법(two-step estimation procedure)을 유도하였으며, 이 이단계 추정법과 관련하여 발생하는 generated regressor의 문제를 해결하는 방안, 즉 추정치들의 표준오차 교정(standard error correction)이 국면전환 모형에서 어떻게 응용될 수 있는가에 대한 문제 등을 제시하고 있다.

<First paper:> In this paper, we consider a regime-switching hybrid NKPC in order to investigate the nature of the structural break in U.S. inflation dynamics. In doing so, we employ Kim’s (2004) method for estimation of regime-switching models with e ...

<First paper:> In this paper, we consider a regime-switching hybrid NKPC in order to investigate the nature of the structural break in U.S. inflation dynamics. In doing so, we employ Kim’s (2004) method for estimation of regime-switching models with endogenous regressors, in order to deal with the endogeneity problem that results from measurement errors associated with inflation forecasts and output gap measures.
Our empirical results show that, since the estimated breakpoint in 1982, the importance of the forward-looking component has increased, while the backward-looking component has become insignificant. In other words, a hybrid NKPC may well explain the inflation dynamics before the structural break while a pure NKPC would be enough to explain the inflation dynamics after the structural break. The results are robust whether we use ex-post realized inflation or a survey measure of inflation forecasts in place of the unobserved future inflation in the empirical specification of the NKPC.






<Second paper:> In this paper, we propose a two-step MLE procedure to deal with the problem of endogeneity in Markov-switching regression models. While Kim (2004) assumes that a common latent Markov-switching variable governs the regime-switching nature of equations (1) and (2), we allow for the possibility that two potentially correlated or independent latent Markov-switching variables govern the regime-switching nature of equations (1) and (2). The joint estimation procedure provides us with an asymptotically most efficient estimator. However, our Monte Carlo experiments show that the two-step procedure, which ignores the potential correlation between the latent state variables, can be more efficient than the joint estimation procedure in finite samples when there is zero or low correlation between the state variables. Besides, as the number of states for the latent Markov-switching variables increases, the joint estimation procedure is infeasible due to the `curse of dimensionality' in the matrix of transition probabilities. In this case, we have no choice but to employ the two-step procedure introduced in this paper, at the cost of a loss of asymptotic efficiency.

본 연구과제의 결과로 두 편의 논문이 완성되었어며, 이들 논문들의 초록은 다음과 같다.


<논문 1 : "Is the Backward-Looking Component Important in a New Keynesian Phillips Curve?"의 초록:

The specification of the New Keynesian Phillips curves (N ...

본 연구과제의 결과로 두 편의 논문이 완성되었어며, 이들 논문들의 초록은 다음과 같다.


<논문 1 : "Is the Backward-Looking Component Important in a New Keynesian Phillips Curve?"의 초록:

The specification of the New Keynesian Phillips curves (NKPC) has recently been modified to include additional lags of inflation in the specification in order to capture the considerable persistence in postwar U.S. inflation. Given the importance of inflation inertia and the reduction of U.S. inflation persistence since the early 1980’s as reported in the literature, we investigate the nature of structural break in the hybrid NKPC. Empirical results show that since 1982 the importance of the forward-looking component has increased and the backward-looking component has become unimportant. The results are robust whether we use the ex-post realized inflation or the survey measure of inflation forecasts in place of unobserved future inflation. Our empirical results also suggest that the significance of the backward-looking component before the early 1980’s might be due to an omission of the unstable long-run inflation target perceived by economic agents, as implied by Kozicki and Tinsley (2002) and Cogley and Sbordone (2005).

<논문2: "Markov-Switching Models with Endogenous Explanatory Variables II:A Two-Step MLE Procedure"의 초록>

This paper proposes a two-step maximum likelihood estimation (MLE) procedure to deal with the problem of endogeneity in Markov-switching regression models. Joint estimation procedure provides us with an asymptotically most efficient estimator, but it is not always feasible due to the `curse of dimensionality' in the matrix of transition probabilities. A two-step estimation procedure, which ignores potential correlation between the latent state variables, suffers less from the `curse of dimensionality' and it provides a reasonable alternative to the joint estimation procedure. In addition, our Monte Carlo experiments show that the two-step estimation procedure can be more efficient than the joint estimation procedure in finite samples when there is zero or low correlation between the latent state variables.

<논문 1>

Our empirical results show that, since the estimated breakpoint in 1982, the importance of the forward-looking component has increased, while the backward-looking component has become insignificant. In other words, a hybrid NKPC may well e ...

<논문 1>

Our empirical results show that, since the estimated breakpoint in 1982, the importance of the forward-looking component has increased, while the backward-looking component has become insignificant. In other words, a hybrid NKPC may well explain the inflation dynamics before the structural break while a pure NKPC would be enough to explain the inflation dynamics after the structural break.



<논문 2 >

While Kim (2004) assumes that a common latent Markov-switching variable governs the regime-switching nature of equations (1) and (2), we allow for the possibility that two potentially correlated or independent latent Markov-switching variables govern the regime-switching nature of equations (1) and (2) in the paper. The joint estimation procedure provides us with an asymptotically most efficient estimator. However, our Monte Carlo experiments show that the two-step procedure, which ignores the potential correlation between the latent state variables, can be more efficient than the joint estimation procedure in finite samples when there is zero or low correlation between the state variables. Besides, as the number of states for the latent Markov-switching variables increases, the joint estimation procedure is infeasible due to the `curse of dimensionality' in the matrix of transition probabilities. In this case, we have no choice but to employ the two-step procedure introduced in this paper, at the cost of a loss of asymptotic efficiency.