《 – Reserve Accumulation and Financial CrisesFrom Individual Protection to Systemic Risk.docx》由会员分享，可在线阅读，更多相关《 – Reserve Accumulation and Financial CrisesFrom Individual Protection to Systemic Risk.docx（41页珍藏版）》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、 Reserve Accumulation and Financial Crises: From Individual Protection to Systemic Risk Andreas Steiner Reserve Accumulation and Financial Crises: From Individual Protection to Systemic Risk Andreas Steiner University of Osnabrueck January 2014 Abstract This paper provides a new perspective on the r
2、elationship between countries international reserve holdings and nancial crises: While the “local” view holds that reserves may prevent domestic crises, it overlooks that the accumulation of reserves relaxes the - nancing constraint of the reserve currency country and may cause a nancial crisis in t
3、he center, which is transmitted globally. According to this “global” view reserve accumulation might destabilize the international nancial system. Since the crisis a ects all countries alike, the accumulation of reserves imposes a negative externality on non-accumulating countries. We integrate this
4、 idea in a theoretical model of the optimal amount of reserves and illustrate the gap between local and global optimality: The consideration of systemic risk lowers the demand for reserves. Moreover, if a supranational authority determines the optimal level of reserves, it internalizes the negative
5、externality and accumulates fewer reserves. A macroprudential tax on reserve hoardings might implement the socially optimal solution. Our calibration analysis shows that these considerations are economically signicant: They lower the optimal amount of reserves in the benchmark case by 45%. Keywords:
6、 Financial Crises, Global Imbalances, International Reserves, Macro- Prudential Policies, Systemic Risk. JEL Classication Numbers: F31, F41, D62, E58. Address: University of Osnabrueck, Institute of Empirical Economic Research, Rolandstrasse 8, D-49069 Osnabrueck, Germany, e-mail: asteineruni-osnabr
7、ueck.de. Tel.: +49 541 9692556; fax: +49 541 9692757. 2 1 Introduction Central banks hoardings of international reserves are considered as a form of self- protection against nancial crises. They enable central banks to intervene in the foreign exchange market and help to cushion the economy from ext
8、ernal shocks. This paper turns the tables and shows that the accumulation of reserves might also have a ip-side: Whereas large reserve holdings may indeed protect a country from domestic crises, their accumulation increases the instability of the international nancial system and might cause a global
9、 crisis emanating from the reserve currency country. In the end, central banks attempt to insure against nancial crises via the accumulation of reserves may be counterproductive. Good intensions may result in bad outcomes. The idea is motivated by the global nancial crisis that began as the US subpr
10、ime crisis in 2007 and a ected the rest of the world through trade and nancial linkages. In this re- gard it has been noted that global imbalances have increased the vulnerability of the US (i.e. Aizenman, 2010; Ferguson and Schularick, 2011; IMF, 2009a; Obstfeld and Rogo , 2010; Portes, 2009). Glob
11、al imbalances, in turn, have been partly sustained by central banks accumulation of reserves, especially in Asian emerging markets. Moreover, reserve accumu- lation has lowered US interest rates (see Warnock and Warnock, 2009). Low interest rates, in turn, were a driving force of the US housing bubb
12、le. By implication, the accumulation of reserves has contributed to developments in the US, that eventually turned into a global crisis. The recurrence of nancial crises in the recent past questioned the benets of the increas- ing international nancial integration and challenged countries to nd ways
13、 how to protect the domestic economy from the downside risks of nancial openness. Many countries faced this challenge by the accumulation of foreign reserves. Since the East Asian nancial crisis of 1996-97 the worldwide level of real reserves has more than tripled. Reserves are considered as a preca
14、utionary cushion against the risks of nancial openness, namely sudden stops of capital ows and contagious nancial crises. Their function includes both crisis prevention and crisis management (see Aizenman and Lee, 2007; Obstfeld et al., 2010).1 This unprecedented increase in international reserves a
15、lso reects previous advice given to emerging and developing countries by various sources. In the aftermath of the East Asian nancial crisis the IMF emphasised the importance of reserves as a means of crisis prevention and proposed new measures to evaluate their adequacy (IMF, 2000). Feldstein (1999)
16、 advised emerging markets to rely on large foreign exchange reserves as a form of self- 1There exist other reasons why countries hoard reserves. The mercantillist approach, for example, argues that the accumulation of reserves is the by-product of an export-led growth strategy (see, among others, Do
17、oley et al, 2003). This paper, however, focuses on the relationship between reserves and crises. 3 protection and to count less on assistance by the IMF. Finally, the burdensome conditionality and unpredictability of IMF assistance may explain why many countries prefer to self-insure (see Bird and M
18、andilaras, 2011). The accumulation of reserves, however, contains costs that have been neglected so far: While reserves might e ectively protect the domestic economy from external shocks, their global and continuous accumulation might create systemic risks. Since the accumulation of reserves constit
19、utes a capital inow to the reserve currency country, it increases its external indebtedness. As a matter of fact, holdings of foreign reserves by global central banks are an important source of US external debt: In 2010, 36% of outstanding Treasuries and 15% of total foreign liabilities were held by
20、 foreign o cial institutions. Between 1998 and 2010, despite the Feds policy of quantitative easing, 49% of the increase in outstanding Treasuries were purchased by foreign o cial institutions.2 The rising external indebtedness may create the macroeconomic backdrop for a sovereign debt crisis, a cur
21、rency crisis or a banking crisis. A crisis in the center country could destabilize the international nancial system. Since the reserve currency country is per denition at the centre of the international nancial system, a crisis originating there spreads to other coun- tries and causes a global downt
22、urn. This crisis a ects accumulating and non-accumulating countries alike. Hence, the accumulation of reserves has a negative externality. The accumulation of reserves might induce two types of crises: First, it might lead to overborrowing and overinvestment in the reserve currency country and cause
23、 a nancial crisis when expectations worsen or the reserve accumulation ends. Second, by steadily worsening the net foreign asset position of the reserve currency country, it might result in a currency cri- sis where the reserve currency country deliberately decides to devalue its currency. Whereas t
24、he rst type of crisis follows the lines of the nancial crisis of 2008-10, the second type also has its precedent, namely the breakdown of the Bretton Woods system. A dilemma arises: On the one hand, the recent reserve accumulation is partly due to concerns for nancial stability in a nancially global
25、ised world (see Obstfeld et al., 2010). On the other, policies of reserve accumulation exactly expose the system to additional risks and shocks. Hence, the blessing attributed to the accumulation of reserves might become a curse.3 The relationship between reserve accumulation and systemic risk creat
26、ion has been iden- tied only recently (see Ferguson and Schularick, 2011; Gourinchas et al., 2010; IMF, 2010; 2Data based on Flow of Funds, Federal Reserve, Tables L.106 (lines 11 and 12) and Table L.209 (line 1) and Lane and Milesi-Ferretti, 2007 and update. 3This is an example of the law of uninte
27、nded consequences, which was popularised by Robert K. Merton: Any intervention - in our case a central bank intervention - creates unanticipated and undesired outcomes. In the end, the intended solution may aggravate the problem. 4 Obstfeld and Rogo , 2010; Taylor, 2013). This paper extends the prec
28、eding literature in sev- eral dimensions: First, the line of causality from reserve accumulation to a global nancial crisis is traced both on empirical and theoretical grounds. Second, the paper integrates the idea of negative feedback in a model of the demand for reserves and formalizes the di eren
29、ce between local and global optimality. It solves for the rst best policy chosen by a social plan- ner that internalises the externality associated with local optimality. A calibration analysis quanties these e ects. This paper is organized as follows. The next section elaborates on the link between
30、 reserves and nancial crises, both on theoretical and empirical grounds. The model and its solution are presented in section 3. Section 4 provides a calibration analysis that quanties the di erence between local and global optimum. Section 5 discusses the policy implications of our ndings. Concludin
31、g remarks are o ered in Section 6. 2 Reserves and crises: the links This section illustrates the links between central banks international reserve holdings and the probability of nancial crises. In this context it is important to distinguish between domestic and global crises. A global crisis is den
32、ed here as a crisis that originates in one country - the centre country - and spreads to a number of countries due to their real and nancial linkages with the crisis country. As will be shown later, the reserve currency country could trigger a global nancial crisis. 2.1 Reserves and domestic crises
33、Theoretical models and empirical ndings with respect to the link between the level of international reserves and currency crises show that reserves reduce both the probability and severity of domestic nancial crises. In the rst generation of currency crisis models inconsistent policies lead to a con
34、tinuous loss of reserves and, consequently, to a devaluation crisis when reserves have fallen below some critical value (e.g. Krugman, 1979, and Flood and Garber, 1984). Hence, reserves can avoid a crisis when the policy inconsistency is transitory. Otherwise, reserves can postpone the occurrence of
35、 a crisis and provide a time bu er within which domestic policies can be reconciled with the exchange rate commitment. Models of the second generation emphasize that crisis expectations might become self-fullling. Reserves may signal governments ability and willingness to defend the exchange rate an
36、d prevent a speculative attack. Another strand of the crisis literature stresses the fragility of balance sheets in the presence of currency 5 R mismatches. Reserves reduce these mismatches at the country level and might be used to support the banking and corporate sector during balance sheet crises
37、. Accordingly, the literature traditionally derives central banks demand for reserves from the benets they provide in the face of shocks: Reserves might be used to smooth economic adjustment processes in the face of external shocks (Heller, 1966) and to prevent nancial and sovereign crises (Ben-Bass
38、at and Gottlieb, 1996). Theoretical models show that reserves help to smooth consumption intertemporally in the presence of productivity shocks (Aizenman and Marion, 2004) and mitigate the output e ects of liquidity shock (Aizenman and Lee, 2007). Li and Rajan (2009) show in a theoretical model that
39、 high reserves may o set the negative impact of moderately weak fundamentals and prevent speculative attacks on the currency. According to Jeanne and Ranciere (2011) the optimal amount of reserves for a small open economy increases with the probability and size of a sudden stop in capital ows. Altho
40、ugh these models di er considerably, they coincide in the view that reserves might be hoarded to reduce the incidence and the cost of nancial crises. These theoretical results are conrmed by a series of empirical papers. In a meta-analysis Frankel and Saravelos (2012) review more than 80 papers of t
41、he literature on early warning systems for currency crises. They nd low central banks reserves to be the most reliable warning indicator including the crisis of 2008-10. Concerning the depth of a crisis once it materializes, De Gregorio and Lee (2004) and Aizenman et al. (2012) demonstrate that rese
42、rves reduce the output costs of a crisis. Obstfeld et al. (2009) and Fratzscher (2009) note that low reserves are associated with larger depreciations during the crisis of 2008-10.4 Based on these theoretical and empirical ndings and in line with the literature (see Aizenman and Marion, 2004, and Je
43、anne, 2007) we assume that the probability of a local output-reducing crisis (pL) in country i decreases in the level of reserves (R). Reserves may prevent nancial crises or help to cushion terms-of-trade shocks. Since the e ect of reserves depends on the potential volume of capital ight, reserves a
44、re scaled by the level of external debt (B):5 pL L Bi L i = i + i , i 0, 0 0 and 0 0 implicitly accounts for the major part of the exorbitant privilege, which the reserve currency country enjoys. 13 1 A C i f where the indices g, c and d stand for the good state, crisis state and default, respective
45、ly. In order to focus on a single rst order condition, we assume in line with Aizenman and Marion (2004) that foreign debt reaches the credit ceiling, B = B.20As a consequence, B(1 + r)|B=B = (1 + ). The expected utility is then given by U (.)|B=B = (1 + B R) + 1 + (1 )(1 + ) + (1 + r )R pL (1 q)( +
46、 ) (14) where is the individual discount rate. Government chooses the levels of foreign debt and reserves such that expected utility is maximised given the budget constraints (9) and (13) and the probability of a nancial crisis (1). The rst order condition with respect to R then reads as: dB 1 f dpL
47、 1 dR |B=B = 1 + (1 + r ) (1 q)( + ) dR (15) This equation can be interpreted as follows: The left-hand side represents the costs of a marginal unit of reserves which are given by the di erence between the resources withdrawn from consumption and the additional external debt granted thanks to an inc
48、rease of the credit ceiling. The right-hand side illustrates the benets of a marginal unit of reserves, namely additional interest income and a reduced crisis probability. After calculating the partial derivatives and substituting them into (15), one obtains (1 + rf )R + ( + )2 = ( + )(1 + rf )(1 q)
49、 + (1 + )(1 + ) ( + ) ( rf ) D(16) This equation can be solved for R. However, for following sections this representation turns out to be convenient as a benchmark for comparisons. In subsequent sections, the left-hand side term is referred to as A, the right-hand side term as D. This equation shows that the optim