Macroeconomic Impacts of the US External Imbalances with Two Large Emerging Asian Economies: Japan (1970–1990) versus China (2000–2018)

Abstract

This paper compares the US long-run macroeconomic effects of foreign direct investment from Japan (1970–1990) with foreign portfolio investment from China (2000–2018). Investment from the both countries lowers the US inflation rate, and the effect from Japan is stronger than that from China. While Japan’s investment in the 1970–1990 period accounts for 88% of the unemployment-reducing effect of the overall US financial account surplus, China’s investment in the 2000–2018 period explains 65% of the unemployment-deterioration effect as well as 96% of the inflation-reducing effect of the overall US financial account surplus. The US inflation and unemployment impulse responses exhibit the pertinent supportive evidence. Furthermore, in contrast to Japan’s direct investment, China’s portfolio investment is more sensitive to dollar depreciation, higher inflation rate and lower real interest rate in the USA.

Appendix: Comparative Static Analysis

Totally differentiating Eq. (1) produces the following expression:

$$\left( {\begin{array}{*{20}c} 1 & { - \gamma \beta } & 0 \\ 0 & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ {\theta _{2} } & {\theta _{1} \beta } & { - 1} \\ \end{array} } \right)\left( {\begin{array}{*{20}c} {d\pi } \\ {du} \\ {dr} \\ \end{array} } \right) = \left( {\begin{array}{*{20}c} 0 & { - \gamma \beta } & 0 \\ { - FPI_{S}^{'} } & {\beta + I_{T}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{T}^{'} } & {NCO_{e} ^{'} } \\ 0 & {\theta _{1} \beta } & 0 \\ \end{array} } \right)\left( {\begin{array}{*{20}c} {dS} \\ {du_{n} } \\ {de} \\ \end{array} } \right)$$

(2)

where $NCO_{e}^{'} = FO_{e}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{e}^{'} - FPI_{e}^{'} > < 0,~as~(FO_{e}^{'} - FPI_{e}^{'} ) > < \left( {1 - I_{{FDI}}^{'} } \right)FDI_{e}^{'} .$ In (2), three endogenous variables (inflation rate $\pi$, unemployment rate u, and real interest rate r) are the major macroeconomic indicators for the USA, and they are dependent on three pre-determined variables in the USA (safe-haven status S, total factor productivity T, and the dollar exchange rate e) that influence financial inflow into the USA. Although the Japanese FDI’s crowd-in effect is significant enough, such a crowd-in effect is assumed to be bounded as follows:

$$1 < I_{{FDI}}^{'} < 1 + \frac{1}{{FDI_{u}^{'} }}~\beta \left[ {1 + \left( {\theta _{1} + \theta _{2} \gamma } \right)\left( { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \right)} \right].$$

Thus, the Jacobian determinant for the endogenous variables is positive

$$\left| J \right| = - \beta \left[ {1 + \left( {\theta _{1} + \theta _{2} \gamma } \right)\left( { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \right)} \right] - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{u}^{'} > 0$$

Solving (2) for nine comparative-static derivatives (each of the three endogenous variables with respect to three external exogenous economic shocks) produces the following results:

$$\frac{{\partial \pi }}{{\partial S}} = \frac{1}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 0 & { - \gamma \beta } & 0 \\ { - FPI_{S}^{'} } & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ 0 & {\theta _{1} \beta } & { - 1} \\ \end{array} } \right| = \frac{{\gamma \beta FPI_{S}^{'} }}{{\left| J \right|}} < 0$$

(3)

$$\frac{{\partial u}}{{\partial S}} = \frac{1}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 1 & 0 & 0 \\ 0 & { - FPI_{S}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ {\theta _{2} } & 0 & { - 1} \\ \end{array} } \right| = \frac{{FPI_{S}^{'} }}{{\left| J \right|}} > 0$$

(4)

$$\frac{{\partial r}}{{\partial S}} = \frac{1}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 1 & { - \gamma \beta } & 0 \\ 0 & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & { - FPI_{S}^{'} } \\ {\theta _{2} } & {\theta _{1} \beta } & 0 \\ \end{array} } \right| = \frac{{\left( {\theta _{1} + \gamma \theta _{2} } \right)\beta FPI_{S}^{'} }}{{\left| J \right|}} < 0$$

(5)

$$\begin{gathered} \frac{{\partial \pi }}{{\partial T}} = \frac{{\partial \pi }}{{\partial u_{n} }}\frac{{\partial u_{n} }}{{\partial T}} = \frac{{u_{n}^{'} }}{{\left| J \right|}}\left| {\begin{array}{*{20}c} { - \gamma \beta } & { - \gamma \beta } & 0 \\ {\beta + I_{T}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{T}^{'} } & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ {\theta _{1} \beta } & {\theta _{1} \beta } & { - 1} \\ \end{array} } \right| \hfill \\ = u_{n}^{'} \frac{{\gamma \beta \left[ {\left( {1 - I_{{FDI}}^{'} } \right)\left( {FDI_{u}^{'} + FDI_{T}^{'} } \right) - I_{T}^{'} } \right]}}{{\left| J \right|}} < 0,\;{\text{if}}\;FDI_{T}^{'} > \left| {FDI_{u}^{'} } \right| \hfill \\ \end{gathered}$$

(6)

$$\begin{gathered} \frac{{\partial u}}{{\partial T}} = \frac{{\partial u}}{{\partial u_{n} }}\frac{{\partial u_{n} }}{{\partial T}} = \frac{{u_{n}^{'} }}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 1 & { - \gamma \beta } & 0 \\ 0 & {\beta + I_{T}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{T}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ {\theta _{2} } & {\theta _{1} \beta } & { - 1} \\ \end{array} } \right| \hfill \\ = \frac{{u_{n}^{'} \left\{ {\left( {\theta _{1} + \gamma \theta _{2} } \right)\beta \left[ {I_{r}^{'} + FO_{r}^{'} - FPI_{r}^{'} } \right] - \beta + \left[ {I_{T}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{T}^{'} } \right]} \right\}}}{{\left| J \right|}} < 0,\;,{\text{ if}}\;\beta \left\{ {\left( {\theta _{1} + \gamma \theta _{2} } \right)\left[ {I_{r}^{'} + FO_{r}^{'} - FPI_{r}^{'} } \right] - 1} \right\} > \left[ {I_{T}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{T}^{'} } \right] \hfill \\ \end{gathered}$$

(7)

$$\begin{aligned} & \frac{{\partial r}}{{\partial T}} = \frac{{\partial r}}{{\partial u_{n} }}\frac{{\partial u_{n} }}{{\partial T}} = \frac{{u_{n}^{'} }}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 1 & { - \gamma \beta } & { - \gamma \beta } \\ 0 & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & {\beta + I_{T}^{'} - \left( {1 - I_{{FDI}}^{'} } \right)FDI_{T}^{'} } \\ {\theta _{2} } & {\theta _{1} \beta } & {\theta _{1} \beta } \\ \end{array} } \right| \\ & \quad = \frac{{u_{n}^{'} \left[ {\left( {1 - I_{{FDI}}^{'} } \right)\left( {FDI_{u}^{'} + FDI_{T}^{'} } \right) - I_{T}^{'} } \right]\left( {\theta _{1} + \gamma \theta _{2} } \right)\beta }}{{\left| J \right|}} < 0,\;{\text{if}}\;FDI_{T}^{'} > \left| {FDI_{u}^{'} } \right| \\ \end{aligned}$$

(8)

$$\frac{{\partial \pi }}{{\partial e}} = \frac{1}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 0 & { - \gamma \beta } & 0 \\ {NCO_{e}^{'} } & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ 0 & {\theta _{1} \beta } & { - 1} \\ \end{array} } \right| = \frac{{ - NCO_{e}^{'} \gamma \beta }}{{\left| J \right|}} > < {0,~{\text{as}}~NCO_{e} ^{'} }> < 0$$

(9)

$$\frac{{\partial u}}{{\partial e}} = \frac{1}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 1 & 0 & 0 \\ 0 & {NCO_{e}^{'} } & { - I_{r}^{'} + FPI_{r}^{'} - FO_{r}^{'} } \\ {\theta _{2} } & 0 & { - 1} \\ \end{array} } \right| = \frac{{ - NCO_{e}^{'} }}{{\left| J \right|}} > < 0,~{\text{as}}~NCO_{e} ^{'} < > 0$$

(10)

$$\frac{{\partial r}}{{\partial e}} = \frac{1}{{\left| J \right|}}\left| {\begin{array}{*{20}c} 1 & { - \gamma \beta } & 0 \\ 0 & {\beta + (1 - I_{{FDI}}^{'} )FDI_{u}^{'} } & {NCO_{e}^{'} } \\ {\theta _{2} } & {\theta _{1} \beta } & 0 \\ \end{array} } \right| = \frac{{ - \left( {\theta _{1} + \gamma \theta _{2} } \right)\beta NCO_{e}^{'} }}{{\left| J \right|}} > < {0,~{\text{as}}~NCO_{e} ^{'} } > < 0$$

(11)