Empirical Model

We used the following empirical model to analyze the effects of changes in US monetary policy on East Asian economies. Most East Asian countries can be treated as small, open economies. In order to reflect this structure in the model, we assume a block-exogenous VAR model in which the US variables are treated as exogenous to East Asian variables.¹ In addition, by constructing a block-exogenous VAR model, we can also save the degree of freedom (compared to considering all the interactions between two countries).

US monetary policy and East Asian economies endogenously respond to the US or world structural shocks. As a result, simple analysis on the relation between US monetary policy and East Asian economies can be misleading since the simple timing relation between US monetary policy and East Asian variables can originate from non-monetary structural shocks. Therefore, exogenous US monetary shocks are identified in the following model, by using the Christiano, Eichenbaum, and Evans (1999) method.

The empirical model assumes that the economy is described by the following structural equation system

where G(L) is a matrix polynomial in lag operator L, and y(t) is a m x 1 data vector with m the number of variables in the model. In addition, var(e(t)) = Λ where Λ is a diagonal matrix and the diagonal elements are variances of structural shocks.

We assumed that the equation system (1) can be expressed as

where y₁(t) and e₁(t) are m₁ x 1 vectors, y₂(t) and e₂(t) are an m₂ x 1 vectors, G₁₁(L) is an m₁x m₁ matrix, G₂₁(L) is an m₂ x m₁ matrix, and G₂₂(L) is an m₂ x m₂ matrix.

We assumed that G₁₂(L)=0. This assumption is the restriction of block-exogeneity, which implies that y₁(t) is not affected by not only current but also lagged y₂(t). In the empirical model, y₁(t) is the US variables and y₂(t) is variables of an East Asian country. Therefore, the US variables are not affected by the variables of a small, open East Asian country.

In the US block, the Christiano, Eichenbaum, and Evans method was applied to identify US monetary policy shocks. y₁(t) is [IP_US, CPI_US, CMP, FFR, NBR, M]´ where IP_US is Industrial Production, CPI_US is Consumer Price Index, CMP is Commodity Price, FFR is the Federal Funds Rate, NBR is Non-Borrowed Reserves, and M is monetary aggregate. Following Christiano, Eichenbaum, and Evans, we identified FFR shocks as monetary policy shocks by imposing restrictions on contemporaneous structural parameter G11(0) that IP, CPI, and CMP don't respond to FFR contemporaneously and FFR does not respond to NBR and M contemporaneously (for contemporaneous structural parameter).²

For the variables of a East Asian country (y₂(t)), CPI, industrial production (IP), call rate (or interbank rate) (CR), and exchange rate against the US dollar (ERA) are considered. CPI and IP are important economic variables indicating the price level and production level of the country. CR and ERA are included since those variables show monetary policy and foreign exchange policy for each country.³ The basic model includes these four variables only since the estimation period is relatively short. Then, the basic model is extended to include M1 and foreign exchange reserves (RES) additionally, one by one, in order to infer more detailed effects on foreign exchange policy responses of a East Asian country.

Since the variables on the right hand side are different in the reduced form Block-Exogenous VAR model, Ordinary Least Square provides inconsistent estimates. We estimated the reduced form Block-Exogenous VAR model with seemingly unrelated regression and then transformed to structural VAR model.

Since the estimation period is short, we used monthly data. The estimation period is January 1999– June 2007, derived from the period after the Asian financial crisis. We considered the period before the recent global financial crisis since the US takes unusual monetary policy (i.e., quantitative easing) during such times. A constant and two lags are assumed in the VAR model.

Download this Paper [ PDF 424.7KB| 25 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.