Official Japanese intervention in foreign exchange markets

Official Japanese intervention in foreign exchange markets

Economics Letters North-Holland 115 15 (1984) 115-120 OFFICIAL JAPANESE INTERVENTION FOREIGN EXCHANGE MARKETS Leaning Against the Wind? Michael M...

315KB Sizes 0 Downloads 44 Views

Economics Letters North-Holland

115

15 (1984) 115-120

OFFICIAL JAPANESE INTERVENTION FOREIGN EXCHANGE MARKETS Leaning Against the Wind? Michael

M. HUTCHISON

IN

*

Uniuersity of Oregon, Eugene, OR 97403, USA Received

30 August

1983

Several recent studies have suggested that official Japanese intervention in foreign exchange markets is limited to simple ‘leaning against the wind’. This note provides evidence, however, that Bank of Japan intervention is systematically biased against yen appreciation.

1. Introduction Since the introduction of managed floating of the Japanese yen in 1973, the Bank of Japan has generally followed an intervention policy in foreign exchange markets of ‘leaning against the wind - purchasing foreign exchange (selling yen) when the yen is appreciating and purchasing yen when the yen is depreciating. This intervention policy for Japan is well documented by Quirk (1977) and Dornbusch (1980), and suggests that the Bank of Japan is following the three principles adopted by the International Monetary Fund for the guidance of members’ exchange rate policies. These principles are: (A) A member

shall avoid manipulating exchange rates or the international monetary system in order to prevent effective balance of payments adjustment or to gain an unfair competitive advantage over other members.

* I would like to thank Randall Stone for helpful suggestions. 0165-1765/84/$3.00

Eberts,

Stephen

Haynes,

0 1984, Elsevier Science Publishers

Raymond

F. Mikesell,

B.V. (North-Holland)

and Joe

116

M.M. Hutch&n

/ Japanese intervention in foreign exchange markets

(B) A member should intervene in the exchange market if necessary to counter disorderly conditions which may be characterized inter alia by disruptive short-term movements in the exchange value of its currency. (C) Members should take into account in their intervention policies the interests of other members, including those of the countries in whose currencies they intervene. ’ The evidence presented below, however, suggests that exchange rate policy in Japan extends beyond simple ‘leaning’ in the sense of a policy designed purely to moderate the speed of exchange rate movements. Specifically, evidence is presented which suggests that the Bank of Japan intervenes much more aggressively when the yen is appreciating than when it is depreciating. This is the first empirical support for the hypothesis that systematic official Japanese intervention is biased against yen appreciation.

2. The model The model employed here postulates simple leaning behavior within a partial adjustment framework, I;’ = a - p,,

optimal

I;‘=yI~+(l-y)I:_,+U,,

lagged response

intervention

against

policy,

function,

the wind

(1) (2)

where I;’ = Bank of Japan purchases of foreign financial assets (a minus sign denotes sales of foreign financial assets) at time t, . = percent change in the yen/dollar rate (approximated by the first e, difference in the log of the yen/dollar rate), u, = an error term. Combining

(1) and (2) gives

I:=cry-py~,+(l-y)I:_,+u,. ’ IMF Executive Board Decision no. 5392-(77/63), pp. 107-109).

(3) adopted

April 1977. See IMF (1977,

M.M. Hutchison / Japanese intervention in foreign exchange markets

117

Three hypotheses are considered within this framework. First, is the Bank of Japan intervening in foreign exchange markets in the manner hypothesized above, i.e., purchasing foreign financial assets when the yen is appreciating and selling foreign assets when the yen is depreciating? If leaning against the wind policy is being followed, the expected sign on er is negative. Second, is the Bank of Japan reacting differently in response to yen appreciation than it is to yen depreciation? This hypothesis, which may be labeled ‘asymmetric leaning’ behavior, is investigated by introducing an interaction variable (0, * 6,). This allows the ti, slope coefficient to take on two separate values, one for intervention in response to a percent change in the yen rate and an additional amount of response to yen appreciation, i.e., D, * k, = k,

if

2< 0

= 0,

otherwise,

where D,=l,

if

e,
= 0,

otherwise.

If the interaction variable is significant, constraining one coefficient (j3y) to estimate both Japanese official intervention response to yen appreciation and that to yen depreciation is incorrect. A significant negative interaction variable coefficient would support the asymmetric leaning hypothesis. The third hypothesis to be investigated concerns Bank of Japan intervention behavior in relation to IMF principle (C) above, that states: ‘Members should take into account in their intervention policies the interests of other members, including those of the countries in whose currencies they intervene.’ In order to test whether Japanese official intervention is coordinated with U.S. official intervention or, on the other hand, tends to offset U.S. intervention operations, U.S. intervention is entered directly into the regression equation. A significant negative sign on the U.S. intervention coefficient, ’ ItuS*, would suggest that Bank of Japan intervention behavior is complementary and not offsetting, i.e., a negative coefficient signifies that the Bank of Japan is purchasing yen in yen support operations during the same period that the U.S. treasury is purchasing foreign currencies (including the yen). * A positive vale for U.S. intervention (Z,USA) denotes U.S. purchases of foreign exchange and a negative value denotes U.S. sales of foreign exchange.

118

M.M. Hutchison / Japanese intervention in foreign exchange markets

3. Empirical

results

The empirical results are presented in table 1. The data covers the period from March 1973 to November 1981 and consists of monthly observations. 3 The models are estimated using OLS regression analysis. 4 The basic model is estimated in regression (1). These coefficient estimates are consistent with the findings of Dornbusch (1980) and Quirk (1977). The results lend support to the partial adjustment model and to the postulated leaning against the wind foreign exchange market intervention behavior by the Bank of Japan. The coefficient estimate on the percent change in the yen/dollar rate suggests that in response to a 1% appreciation (depreciation) in the yen, the Bank of Japan sells (purchases) 41,861 million yen (approximately U.S.$ 167 million) in the current month. 5 Models (2) and (3) in table 1 introduce the interaction term which allows the slope coefficient to take on a separate value for yen rate appreciation (2, < 0). The interaction term is significant at the 10% level of confidence in both models and supports the asymmetric intervention hypothesis, e.g., the estimate in regression (2) suggests that the Bank of Japan purchases 23,697 million yen in response to a 1% yen depreciation during a given month, but sells an average of 59,256 million yen in response to a 1% appreciation. This result indicates that Japanese official

3 The proxy used in the empirical analysis for Japanese spot intervention is changes in the Foreign Exchange Fund accounts maintained by the Bank of Japan. This series is published monthly by the Bank of Japan in Economic Statistics Monthly. The advantage of the Foreign Exchange Fund series over changes in gross international reserves as a proxy for official intervention is that, unlike the latter, the Foreign Exchange Fund accounts exclude extra-market transactions (primarily comprised of interest accruals on official holdings and receipts arising from certain U.S. military purchases of Japanese goods and services) and include the so-called hidden reserves, i.e., changes in official deposits of foreign exchange with Japanese commercial banks [Quirk (1977, p. 643)]. U.S. intervention is proxied by changes in U.S. gross international reserves adjusted for interest accruals, where interest accruals on official U.S. holdings of foreign assets is estimated by the rate of return on 3-month U.S. Treasury Bills. The unadjusted U.S. gross international reserves series and the yen/dollar spot exchange series are from various issues of International Financial Statistics. 4 An instrumental variables technique was also utilized in order to capture the feedback causality from I,? to e,. Little evidence of simultaneity bias was found, however, and the basic results presented above were not modified. s Using monthly data covering the March 1973-October 1976 period, Quirk (1977) estimates a response in the current month of U.S.% 160 million.

M. M. Hutchison / Japanese intervention

in foreign exchange markets

119

intervention in the foreign exchange market is much more aggressive in response to yen appreciation than it is to yen depreciation. Turning to the estimate of Bank of Japan response to U.S. intervention policy, the results reported in regression (3) suggest that the Japanese authorities tend to support and complement, rather than offset, official U.S. intervention. Specifically, the significant negative coefficient signifies that the Bank of Japan is supporting the yen during the same periods that the U.S. Treasury is involved in yen support operations.

Table 1 Regression estimates of Japanese 1981 (t-statistics in parentheses). Independent variables

Intercept

I:-

1

official

intervention

functions,

(2) - 103 (- 0.49)

0.4‘5b (5.87)

(3) -468 a (1.63)

a b ’ d e

0.66 a

(- 1.79) - 41,674 b ( - 6.12)

D * ti,

DF’ DWs

0.44 b (6.09) -

,

RZd P.e

- 333 (-1.13)

0.45 b (6.08)

IUS.

e,

1973-November

Dependent variable: I:

Regressions (1)

March

0.41 - 0.09 101 2.17

Significant at the 10% level of confidence. Significant at the 1% level of confidence. Significant at the 5% level of confidence. The coefficient of determination. The estimate of first-order autocorrelation. ’ The degrees of freedom for each regression. s The Durbin-Watson statistic.

- 23,697 = (- 2.01)

- 25,788 ’ ( - 2.20)

- 35,559 a (- 1.85)

- 32,409 a (-1.70)

0.43 - 0.09 100 2.17

0.45 - 0.07 99 2.14

120

M.M. Hutch&n

/ Japanese interuentron UIforeign exchange markets

4. Conclusions Several recent studies have suggested that Bank of Japan intervention behavior in foreign exchange markets is based solely on an attempt to moderate the speed of exchange rate movements [e.g., Quirk (1977) and Dornbusch (1980)]. The results presented in this paper suggest that official Japanese intervention policy extends significantly beyond simple leaning against the wind. Specifically, the empirical evidence suggests that official Japanese intervention is more aggressive in attempting to siow yen aF?reciation than it is in attempting to slow yen depreciation. For a given percent appreciation of the yen, the Bank of Japan sells roughly twice as many yen as it will purchase in response to a comparable percent depreciation. This phenomenon may be termed ‘asymmetric’ leaning against the wind intervention behavior. The policy implications of these results are clear. To those analysts and policymakers who believe the yen to be undervalued, 6 this is the first empirical evidence which suggests that one possible cause is systematic official Japanese intervention biased against yen appreciation.

References Bank of Japan, Economic Statistics Monthly, various issues. Bergsten, C. Fred, 1982, What to do about Japan, Remarks presented to the Japan Society, New York, May 5. Dombusch, Rudiger, 1980, Exchange rate economics: Where do we stand?, Brookings Papers on Economic Activity 1, 143-185. International Monetary Fund, 1977, Annual report of the executive directors for the fiscal year ended April 30, 1977 (Washington). International Monetary Fund, International Financial Statistics, various issues. Quirk, Peter J., 1977, Exchange rate policy in Japan: Leaning against the wind, IMF staff papers, Nov., 642-664.

6 See, for example, Bergsten (1982). He attributes a major part of Japanese-U.S. trade frictions to a yen undervaluation, but suggests that the major cause ‘lies in the monetary realm’.