Economics Letters 111 (2011) 252–255
Contents lists available at ScienceDirect
Economics Letters j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e c o l e t
Foreign exchange market intervention and reserve accumulation in emerging Asia: Is there evidence of fear of appreciation? Victor Pontines a, Ramkishen S. Rajan b,⁎ a b
Flinders Business School, Flinders University, Adelaide, Australia School of Public Policy, George Mason University, Virginia, United States of America and Institute of Southeast Asian Studies, Singapore
a r t i c l e
i n f o
Article history: Received 3 November 2008 Received in revised form 14 December 2010 Accepted 14 January 2011 Available online 22 January 2011
a b s t r a c t Asian central banks react more strongly to currency appreciations than depreciations and more to nominal effective exchange rates (NEERs) than to bilateral US dollar rates. This rationalizes the relative exchange rate stability and the sustained reserve accumulation in the region. © 2011 Elsevier B.V. All rights reserved.
JEL classiﬁcation: F31 F40 F41 Keywords: Asymmetry Emerging Asia Intervention NEER Reserves
1. Introduction Following the Asian crisis of 1997–98 there have been two broad strands of literature on Asian exchange rate regimes. One strand has attempted to examine to what extent the regional currencies have become more ﬂexible, particularly vis-à-vis the US dollar, but also versus other currencies including the trade-weighted exchange rate (for instance, see Cavoli and Rajan, 2009, Chapter 1). The other strand of the literature has attempted to categorize de facto Asian exchange rate regimes using various methodologies (Frankel et al., 2001; Calvo and Reinhart, 2002; Levy-Yeyati and Sturzenegger, 2005; Reinhart and Rogoff, 2004; Shambaugh, 2004). A distinct but related body of work has focused on trying to rationalize the causes and consequences of reserve buildup in emerging Asian economies and elsewhere especially since 1998 (Fig. 1). Apart from the exchange rate valuation changes due to currency composition of reserve stocks, the three main rationale often suggested for reserve accumulation are insurance (i.e. preventing a crisis), mercantilism (i.e. stimulating growth), and reducing exchange rate volatility. The last rationale (i.e. managing exchange rate volatility), while often used by central bankers, is unconvincing as it ⁎ Corresponding author. E-mail addresses: [email protected]
ﬂinders.edu.au (V. Pontines), [email protected]
(R.S. Rajan). 0165-1765/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.econlet.2011.01.022
should imply that, on average, international reserves do not change much. The fact that reserves are being accumulated on a sustained basis suggests that intervention has involved more than just minimizing exchange rate volatility. While the ﬁrst two rationales have very different motivations, they are similar in the sense that they involve the central bank having to intervene in the foreign exchange market by leaning against the wind. Regardless of whether one looks at commonly used benchmarks of reserve adequacy (Bird and Rajan, 2003; Wijnholds Onno de Beaufort and Kapteyn, 2001), or compares reserves holdings against some benchmark model like the Buffer stock model (Aizenman and Marion, 2003), most studies have arrived at the conclusion that Asia holds more than enough reserves as a ﬁnancial safeguard. This in turn implies that the sustained reserve buildup has been because of a desire to keep currencies from appreciating signiﬁcantly. Indeed, many emerging economies have earmarked some portions of their reserves to invest in less liquid, potentially higher return but higher risk assets.1 As Asian central banks move from liquidity management to wealth management, many countries like China have created a sovereign wealth fund (SWF) to attain their goal or are contemplating doing so. Revealed preference therefore suggests that many of the central banks view their reserve holdings as more than
At least prior to the global ﬁnancial crisis of 2008–09.
V. Pontines, R.S. Rajan / Economics Letters 111 (2011) 252–255
Fig. 1. International reserve holdings in selected emerging Asian economies.
2. Central bank intervention reaction function
Billions of US Dollars
Note: Data for 2009 is up to 2009:m7. Source: Based on data fromthe CEIC database.
The paper is organized as follows. Section 2 outlines a simple model of optimal central bank behavior which derives a simple central bank intervention reaction function which is our estimating equation. Section 3 estimates the model for six emerging Asian economies which are known to operate a variety of managed ﬂoats, viz., India, Korea, Philippines, Singapore, Thailand and Indonesia (Cavoli and Rajan, 2009, Chapter 1) over the period. We use monthly data for 2000:m1–2009:m7, a period which corresponds broadly to the start of the rapid accumulation of reserves by regional central banks as well as the global ﬁnancial crisis of 2008–09. Section 4 concludes the paper. To preview the main conclusion, our results conﬁrm the existence of an asymmetry in central bank foreign exchange market intervention responses in selected Asian economies. Asian central banks appear to react more strongly to exchange rate appreciations than depreciations and they react more to nominal effective exchange rate (NEER) changes than to nominal bilateral exchange rate changes (vis-à-vis the US dollar). This ﬁnding that the Asian economies have been managing their currencies asymmetrically against a trade-weighted basket rationalizes the relative exchange rate stability as well as the sustained reserve accumulation in the region.
90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 ASEAN-4
adequate for precautionary purposes and are yet building it up, presumably as a side-effect of exchange rate interventions. Synthesizing what we have learnt from the foregoing body of knowledge, many observers have drawn the conclusion that Asian currencies remain heavily managed and are effectively undervalued in order to sustain export-led growth. This in turn has contributed to a massive reserve accumulation in emerging Asian economies along with the ongoing global macroeconomic imbalances. While this is a reasonable conjecture, some argue it ignores the concerns that small and open economies in Asia and elsewhere have about a currency that is “too weak”. During the crisis period of 1997–98 and its immediate aftermath there was a great deal of discussion on the problems associated with a weak currency, i.e. a rise in unhedged foreign currency liabilities (Rajan and Shen, 2006). This was the reason for the so-called “fear of ﬂoating” both in terms of appreciation (competitiveness) and depreciations (“liability dollarization”). While some corporates and ﬁnancial institutions in Asia remain vulnerable to their home currency depreciations, in aggregate, as these economies have moved from running current account deﬁcits to surpluses and stockpiled reserves in US dollars and Euros, they are arguably more concerned about loss in capital values with a sharp appreciation rather than depreciation of their currencies.2 We are thus left with the tentative conclusion that many emerging economies desire some sort of exchange rate management with a strong bias towards preventing appreciations than depreciations. In other words, whereas Calvo and Reinhart (2002) noted that exchange rate policy in the 1990s in emerging economies is best characterized as “a fear of ﬂoating”, we conjecture that Asian exchange rate regimes in the 2000s can be more precisely described as being a “fear of appreciation” or “fear of ﬂoating in reverse”, a term initially coined by Levy-Yeyati and Sturzenegger (2007). Somewhat surprisingly there has been scant discussion of this possible asymmetry in foreign exchange market intervention in the debate of de facto exchange rate regimes in Asia, a gap that this paper attempts to ﬁll.3 2 There may also be a more persistent problem of currency depreciation passing through into domestic inﬂation, i.e. exchange rate pass-through. While there is some evidence that exchange rate pass-through into emerging economies has been declining over time, it does not seem to have fallen as rapidly as in the case of developed economies (Ghosh and Rajan, 2007). In other words, there is less evidence of pricing-to-market or local currency pricing of imports to emerging economies in Asia and elsewhere. 3 Two notable exceptions are Ramachandran and Srinivasan (2007) and Srinivasan et al. (2008) who ﬁnd evidence in the Indian context that support the existence of asymmetric foreign exchange intervention (Indian Rupee per US dollar).
The central bank is assumed to have full and direct control over a proxy measure of intervention deﬁned as the percent changes in foreign exchange reserves (Rt). The central bank intervenes in the foreign exchange market to minimize the following intertemporal criterion4: ∞
min Et−1 ∑ δ Lt ðRt Þ
where δ is the discount factor and Lt is the period loss function. We follow Surico (2008) and Srinivasan et al. (2008) in specifying the loss function in linear-exponential form: Lt =
1 λ γ 2 2 3 Rt −R + + e˜t −e e˜t −e 2 2 3
where λ N 0 is the relative weight and γ is the asymmetric preference parameter on exchange rate stabilization. e˜t denotes the percent change in the exchange rate (where et is the foreign currency price of one unit of domestic currency and the NEER, respectively), R* is the optimal level of reserves and e* is the Central Bank's target depreciation rate which is assumed to be zero in this case.5 If γ N 0, deviations of the same size but opposite sign yield different losses and, thus, the rate of appreciation is weighted more heavily than the rate of depreciation, h i i.e.,∂Lt = ∂ð e˜t Þ = λ e˜t + ðγ = 2Þð e˜t Þ2 N 0, for e˜t N 0. In other words, a rise in the exchange rate (appreciation) increases the policymaker's loss. It is assumed that interventions can reduce the rate of change (depreciation/appreciation) in the exchange rate.6 Accordingly, e˜t –e = a0 + a1 Rt + εt
4 Data on actual Central Bank intervention are not available for the countries considered. 5 For instance, Singapore loosely targets a certain level of its NEER around a band and crawl (i.e. so-called band-basket-crawl or BBC regime). See Cavoli and Rajan (2009, chapter 1) for a review of the policy pronouncements of the Asian exchange rate regimes. 6 With regard to the issue of why the level of reserves as opposed to deviation matters, of course, we can never be sure as central banks do not make clear their objectives. This said, one might argue that central banks are actually quite sensitive about declines in reserves as they may be viewed by markets as a sign of some concern/weakness (for instance, see Bird and Rajan, 2003). In relation to this, the fact that many regional central banks have rapidly rebuilt their reserves in the last few months once the crisis abated, further adds to the conjecture that central banks arguably are very concerned about the actual reserve levels per se.
V. Pontines, R.S. Rajan / Economics Letters 111 (2011) 252–255
where a1 N 0 and the error term, εt, is independent and identically distributed (i.i.d.) with zero mean and variance σ 2ε . Minimizing Eq. (2) by choosing Rt subject to the constraint (3) leads to the following intervention reaction function: n o γ 2 Rt = R−λa1 Et−1 e˜t + ð e˜t Þ : 2
Replacing the expected values with the actual values, the empirical version of the intervention reaction function can be simpliﬁed as follows:
Table 1 Intervention reaction function and policy preference estimates, 2000:m1–2009:m7. Country India Row (1) Row (2)
Korea Row (1) Row (2)
Rt = c + α e˜t + βð e˜t Þ + vt
where α = −λa1, and β = −λa1γ/2. The reduced form parameters [α, and β] allow us to identify the asymmetric preference on exchange rate stabilization, γ. It can be shown that the asymmetric preference parameter is γ = 2β/α. This parameter is the main concern of our empirical exercise in the next section (Surico, 2008; Srinivasan et al., 2008).
Philippines Row (1) Row (2)
Singapore Row (1)
3. Empirical results
Our estimation is based on monthly data for the sample period between 2000:m1 and 2009:m7 for six emerging Asian economies, viz., India, Korea, Philippines, Singapore, Thailand and Indonesia. This was the period of rapid stockpiling of reserves in the region (i.e. post Asian crisis of 1997–98) and includes the global ﬁnancial crisis of 2008–09 which started to have an impact on emerging Asian balance of payments by early-to-mid 2008. The variables used in the estimation are as follows: the U.S. federal funds rate, Rt = (Δlog Reservest) * 100 and e˜t = (Δlog et) * 100 with et being the nominal exchange rate (US dollar per domestic currency) and the NEER, respectively, such that a rise in each of these two alternative deﬁnitions of the nominal exchange rate denotes a currency appreciation, and vice versa. The data are sourced from the IMF's International Financial Statistics except for the NEER which is sourced from the Bank for International Settlements (BIS). As noted, Eq. (5) is the main equation of interest in the empirical test. The orthogonality conditions implied by the intertemporal optimizationrational expectations paradigm make the generalized method of moments (GMM) the appropriate estimation method. We follow Hansen (1982) and use an optimal weighting estimate of the covariance matrix that accounts for both serial correlation and heteroscedasticity in the error terms. Hence we report robust standard errors. For the most part, a constant, lagged values (1 to 10, 12, and 15 months) of Rt, et as well as current and lagged values (1 to 4, 8, and 15 months) of the U.S. federal funds rate are used as instruments. Table 1 reports the estimates of the intervention reaction function and the asymmetric preference parameter. For each country we present two sets of results — Row (1) using the nominal bilateral exchange rate (US dollar per domestic currency) and Row (2) using the NEER. The J test indicates that the hypothesis of valid overidentifying restrictions is never rejected. The parameters on e˜t , and α, are statistically different from zero in all cases. Of primary interest is the parameter on the squared e˜t of the β coefﬁcient. This is because testing the restriction that H0: β = 0 is akin to testing H0: γ = 0. We ﬁnd β is statistically signiﬁcant in all countries. What are our prior expectations of the asymmetric preference parameter, γ? As noted in Section 2, a rise in the NEER denotes an appreciation, implying γ should be positive. Results are summarized in Table 1. The asymmetric preference parameter is positive and economically and statistically signiﬁcant when either the US dollar rate or the NEER is used (a rise implies appreciation of home currency in all cases). This implies that the central banks in these countries appear to react differently to appreciation and depreciation pressures.
Thailand Row (1) Row (2)
Indonesia Row(1) Row (2)
γ = 2β/α
1.958a (0.160) 1.202a (0.089)
−2.663a (0.231) −0.432a (0.102)
−0.308a (0.050) −0.148a (0.035)
0.232a (0.025) 0.687a (0.123)
0.479a (0.092) 0.568a (0.086)
−0.447a (0.045) −0.131a (0.032)
−0.104a (0.013) −0.019b (0.007)
0.467a (0.074) 0.291c (0.155)
0.459a (0.169) 1.328a (0.138)
−0.872a (0.127) −1.014a (0.093)
−0.284a (0.070) −0.132b (0.054)
0.651a (0.113) 0.259a (0.103)
0.589a (0.123) 0.991a (0.144)
−0.297a (0.90) −0.923a (0.302)
−0.105a (0.037) −0.716a (0.236)
0.707b (0.360) 1.551a (0.529)
0.552a (0.159) 0.506a (0.084)
−0.571a (0.114) −0.437a (0.086)
−0.165a (0.041) −0.997a (0.078)
0.578a (0.196) 4.567a (0.647)
0.681a (0.200) 1.621a (0.151)
−0.894a (0.166) −0.722a (0.104)
0.062a (0.017) −0.041a (0.012)
0.140a (0.020) 0.113a (0.022)
Speciﬁcation: Rt = c + α(et) + β(et)2 + vt. Standard errors using a four-lag Newey–West covariance matrix are reported in parentheses. Row (1) denotes that et is measured using the nominal exchange rate of the US dollar per local currency, while Row (2) denotes that et is measured using the nominal effective exchange rate (NEER). J-test refers to the Hansen's test of overidentifying restrictions, which is distributed as a χ2(m) under the null hypothesis of valid over-identifying restrictions. A constant, lagged values (1 to 10, 12, and 15 months) of Rt, et as well as current and lagged values (1 to 4, 8, and 15 months) of the US Federal Fund Rate. The standard error of γ are obtained using the delta method. Source: Authors. a Denotes rejection of the null hypothesis that the true coefﬁcient is zero at the 1% signiﬁcance level. b Denotes rejection of the null hypothesis that the true coefﬁcient is zero at the 5% signiﬁcance level. c Denotes rejection of the null hypothesis that the true coefﬁcient is zero at the 10% signiﬁcance level.
More to the point, the responses of central banks in these countries to rates of appreciation are much stronger than to rates of depreciations of the same value.7 Speciﬁcally, the parameter ranges from a low of 0.14 for Indonesia to a high of 0.71 for Singapore when the nominal bilateral exchange rate is used, whereas the parameter ranges from a low of 0.11 in the case of Indonesia to a high of 4.57 when the NEER is used. The asymmetric parameter in three cases, viz. India, Thailand and Singapore, is much larger when the NEER is used compared to the bilateral exchange rate as the regressor in the intervention reaction function. This in turn implies that, not only do these three countries intervene asymmetrically, they also tend to pay more attention to managing their trade-weighted exchange rate rather than the US dollar rates. This is consistent with other studies that have estimated the degree of inﬂuence
7 We have also tried the estimations for smaller sub-periods, i.e., pre-global ﬁnancial crisis, and the results remain intact.
V. Pontines, R.S. Rajan / Economics Letters 111 (2011) 252–255
of major currencies on the emerging Asian economies since the Asian crisis and have found evidence of loose pegging to a basket with the Japanese yen and Euro also inﬂuencing movements in the Asian currencies beyond the US dollar (for instance, see Frankel and Wei, 2008; Cavoli and Rajan, 2009 and references cited within). While this ﬁnding for Singapore is consistent with the fact that it ofﬁcially pursues a band, basket and crawl (BBC) regime, with a basket essentially referring to the NEER, the other two countries in the region, Thailand and India are also believed to operate a de facto currency basket arrangement. On the other hand, the results in the cases of Korea, Philippines and Indonesia suggest asymmetric foreign exchange intervention (but primarily against the US dollar as opposed to currency baskets). 4. Conclusion Following Calvo and Reinhart (2002), it has become commonplace to argue that there is a “fear of ﬂoating” among emerging economies in Asia and elsewhere. While this was an important ﬁnding, the sustained reserve buildup in emerging Asian economies since 2000 suggests that emerging economies in Asia are far more sensitive to exchange rate appreciations than to depreciations. Levy-Yeyati and Sturzenegger (2007) conjectured that exchange rate policy has evolved towards an apparent “fear of ﬂoating in reverse” or “fear of appreciation” whereby interventions have been aimed at limiting appreciations rather than depreciations. Our results conﬁrm the existence of an asymmetry in central bank foreign exchange intervention responses to currency appreciations versus depreciations in all six economies. In addition, the Asian central banks of India, Singapore and Thailand appear also to react much more strongly to the NEER rather than the US dollar bilateral rate, implying management of their currencies against some sort of basket of currencies. Overall, this asymmetric exchange rate intervention
explains both the relative exchange rate stability as well as the sustained reserve accumulation in emerging Asia. References Aizenman, J., Marion, N., 2003. The high demand for international reserves in the far east: What's going on? Journal of Japanese and International Economics 17, 370–400. Bird, G., Rajan, R.S., 2003. Adequacy of reserves in the aftermath of crises. The World Economy 26, 873–891. Calvo, G., Reinhart, C.M., 2002. Fear of ﬂoating. Quarterly Journal of Economics 117, 379–408. Cavoli, T., Rajan, R.S., 2009. Exchange Rate Regimes and Macroeconomic Management in Asia. Hong Kong University Press, Hong Kong. Frankel, J., Wei, S.J., 2008. Estimation of de facto exchange rate regimes: Synthesis of the techniques for inferring ﬂexibility and basket weights. International Monetary Fund Staff Papers 55, 384–416. Frankel, J., Fajnzylber, E., Schmukler, S., Servén, L., 2001. Verifying exchange rate regimes. Journal of Development Economics 66, 351–386. Ghosh, A., Rajan, R.S., 2007. A survey of exchange rate pass-through in Asia: What does the literature tell us? Asia Paciﬁc Economic Literature 21, 13–28. Hansen, L., 1982. Large sample properties of generalized method of moments estimators. Econometrica 50, 1029–1286. Levy-Yeyati, E.L., Sturzenegger, F., 2005. Classifying exchange rate regimes: Deeds vs words. European Economic Review 49, 1603–1635. Levy-Yeyati, E.L., Sturzenegger, F., 2007. Fear of Floating in Reverse: Exchange Rate Policy in the 2000s. mimeo. Rajan, R.S., Shen, C.H., 2006. Why are crisis-induced devaluations contractionary?: Exploring alternative hypotheses. Journal of Economic Integration 21, 526–550. Ramachandran, M., Srinivasan, N., 2007. Asymmetric exchange rate intervention and international reserve accumulation in India. Economics Letters 94, 259–265. Reinhart, C.M., Rogoff, K., 2004. The modern history of exchange rate reinterpretation. Quarterly Journal of Economics 199, 1–48. Shambaugh, J., 2004. The effect of ﬁxed exchange rates on monetary policy. Quarterly Journal of Economics 119, 301–352. Srinivasan, N., Mahambare, V., Ramachandran, M., 2008. Preference asymmetry and international reserve accretion in India. Applied Economics Letters 1–14. Surico, P., 2008. Measuring the time inconsistency of US monetary policy. Economica 75, 22–38. Wijnholds Onno de Beaufort, J., Kapteyn, A., 2001. Reserve Adequacy in Emerging Market Economics. Working Paper No.01/143. International Monetary Fund.