Prospects for a Chinese currency area: simulations of Robert Mundell's multi‐currency monetary union

Abstract

Based on Robert Mundell's proposal for an Asian multi‐currency monetary union, this paper provides a framework to analyze whether a prospective Chinese currency union can be sustained after suffering from various shocks. An important assumption is that China and Taiwan can coordinate their economic policies without losing their independent national currencies. A dynamic game approach simulates possible outcomes if a China–Taiwan cooperative mechanism were to be implemented. The results show that the China–Taiwan coalition can be feasible under the shock of currency appreciation.

Keywords:

• multi‐currency monetary union
• policy coordination
• macroeconomic game
• Taiwan
• China

JEL codes:

• C70
• F33
• F42
• F47

Notes

An early version of this paper was presented at the conference of Asian Crisis IV, in Taipei, Taiwan (July 24–25, 2002). Comments and suggestions from the participants, in particular the discussant Chong‐kee Yip, as well as from Marco Lyrio and Joseph Plasmans, are gratefully acknowledged.

1. A famous case is the Chiang Mai initiative. Since 2005, ASEAN, China, Japan, and South Korea have strengthened the network of currency swap and repurchases. The funds are estimated to be about US$70 billion. In addition, the ASEAN +3 Summit has been renamed the East Asian Summit, and is following the European‐style integration model.

2. We define that the two Chinese‐speaking economies, China and Taiwan, as constituting the area.

3. Translated from the Chinese version of Mundell's Lecture (2000c) in Taiwan.

4. The relative prices would be the main factor to decide on the competitiveness among the countries involved in case there was only a single currency in the area (that is, all s_ij equal to zero).

5. The null hypothesis of unit roots cannot be rejected at the 1% level when these series are tested in level, but the statistics in parentheses indicate that these series are stationary after taking first difference. The Johanson co‐integration tests indicate the existence of co‐integration at the 10% or 5% level. The details are available on request.

6. If matrix M_FC shown in equation (11) has more than three positive eigenvalues, multiple equilibria arise, whereas if this matrix has less than three positive eigenvalues, no equilibrium exists (see Engwerda 1998). Here we only report a representative case if there exist multiple equilibria.

7. We simulated different τ (e.g., in FC τ₃=4/5, τ₁=1/5 and τ₂=0) but omitted the results here because they are basically the same.

8. Note that in reality τ may not be continuous, since not every τ can cause sufficient eigenvalues of matrix M_FC in equation (11).

9. The details of the three sensitivity analyses (e.g., values of the loss functions) are omitted here but available upon request.