چگونه بانک مرکزی برای بحران مالی آماده می شود - تجزیه و تحلیل تجربی اثرات بحران و جهانی شدن در ذخایر بین المللی

Central banks' international reserve holdings have increased significantly in the recent past. While traditional models fail to explain this accumulation of reserves, the more recent literature argues that reserves are used as a lifejacket against financial crises. However, research so far has neglected the question whether and how central banks change their precautionary reserve holdings after the country was affected by a financial crisis. This paper tests the hypothesis that central banks revise their reserve policy in the aftermath of financial crises. Static and dynamic panel data models are estimated for emerging and developing countries covering the period 1970–2010. The evidence suggests that currency crises induce a permanent increase of reserves. This effect is particularly strong for recent currency crises since the Asian financial crisis of 1997–1998. With less robustness, banking crises also induce a positive shift of the reserve level.

One puzzle of the international financial system is the enormous increase in central banks' international reserve holdings since the demise of the Bretton Woods system. In contrast to general wisdom, the transformation to de-jure more flexible exchange rate regimes has not been accompanied by a permanent reduction in the level of reserves. Between 1970 and 2010, worldwide official reserve holdings grew with an average annual rate of 15%.1 This build-up of reserves is mainly due to developing and transition economies. Their share in total worldwide reserves has risen from 22% in 1970 to 65% in 2010. Although this increase is driven by a small number of countries, e.g. China, Japan, Russia and Korea, the phenomenon of reserve accumulation is not restricted to some outliers, but rather observable in the majority of countries (see Fig. 1). The number of systematic accumulators has even increased in the recent past. Whereas from 1982 to 1996 on average 58% of all countries increased their real reserves in a given year, this share amounted to 67% over the period 1997–2010 (Fig. 2). Full-size image (52 K) Fig. 1. Relative increase of international reserves (1995–2010). Figure options Full-size image (35 K) Fig. 2. Share of countries with positive reserve accumulation. Figure options The increase is also observable in commonly used indicators of reserve adequacy, which consider the level of reserves in relation to a scaling variable like imports, GDP or external debt. This shows that the increase cannot be explained by simple rules of thumb: Whereas traditionally a level of reserves covering three to four months of imports was considered to be adequate, in 2010 reserves covered on average almost six months of imports. Even recent models of the optimal amount of reserves fail to explain the actual accumulation. According to Durdu et al. (2009) and Jeanne and Rancière (2011) the optimal level of reserves for a benchmark economy amounts to 10% of GDP. In 2010, however, central banks' reserves averaged 23% of GDP. This unexpected increase in reserves gave rise to a series of papers that investigate two main research questions. The first group of papers analyses the optimality of reserve holdings given that reserves exceed traditional indicators of reserve adequacy (see Bird and Rajan, 2003; IMF, 2011; Jeanne and Rancière, 2011). The second group aims at finding rationales for this unprecedented reserve accumulation (see Aizenman and Lee, 2007; Cheung and Qian, 2009). This paper contributes to the latter strand of the literature. It proposes a new explanation for reserve accumulation: Central banks revise their reserve policy after they have experienced a financial crisis and significantly increase their reserves in the aftermath of a crisis. Questions of the optimality and adequacy of this reserve policy are not touched in this article.2 Existing papers explaining the reserve accumulation can be grouped into two main lines of argumentation: The first argues that the accumulation of reserves is driven by mercantilist motives and the result of an export-led growth strategy (Dooley et al., 2003).3 This view is supported by the literature on the ‘fear of appreciation’ (Levy-Yeyati and Sturzenegger, 2007; Pontines and Rajan, 2011). The second strand highlights the precautionary motive of reserve hoardings. Reserves are seen as a form of self-insurance in the presence of financial crises (see Jeanne and Rancière, 2011; Mendoza, 2004). Aizenman and Lee (2007) contrast both motives and test their empirical relevance. Their results confirm the dominance of a precautionary reserve demand, whereas the mercantilist motive turns out to be economically insignificant. This article contributes to both lines of argumentation: If the experience of a financial crisis alters the reserve policy of a central bank, the additional hoardings might be precautionary. The central bank changes its assessment of the country-specific crisis probability and fears future financial crises. It wants to be better prepared to defend the currency and manage a crisis after a future attack. The reserve accumulation after a crisis might also be driven by mercantilist motives: Since the accumulation of reserves allows to maintain an undervalued exchange rate, it might be regarded as an instrument to ease the negative growth effects of the crisis thanks to a more competitive export sector. Empirical tests of the relationship between financial crises and reserves are rare. Typically, studies confirm that reserves decrease during a currency crisis whereas the long-run effects of a crisis are disregarded. To my knowledge there are only two exceptions: Aizenman and Lee (2007) include two dummies in their reserve demand specification, one for the Mexican Tequila crisis in 1994 and another for the Asian financial crisis of 1997–1998. These are applied to all countries independently whether they were directly affected by these crises or not. They find a significantly positive effect of crises on reserve levels. While these authors focus on the effect of two major financial crises on countries' reserve levels, we adopt a disaggregated, country-centred approach, which examines whether reserves increase in an individual country after being hit by a financial crisis. Bird and Mandilaras (2011) find that recessions do not affect subsequent reserve levels. Having an assistance programme with the IMF, in turn, is associated with higher reserves in the long run. This article extends the existing literature in several ways. First, it tests the hypothesis that reserve holdings are significantly higher in countries that have experienced a financial crisis. We distinguish between restocking of reserves to their pre-crisis level and subsequent reserve accumulation exceeding the pre-crisis level. Second, we identify crises for each country individually. Third, by applying the exchange market pressure index, we investigate the effects of both actual crises and unsuccessful speculative attacks. Fourth, besides currency crises, we also examine the effects of banking and sovereign debt crises. To confirm the robustness of the findings, the results of different estimators and for various subsamples are compared. This paper proceeds as follows. Section 2 examines how reserves and crises are related and postulates the hypothesis of the article. Section 3 describes the data and explains how crises are identified. Section 4 presents and discusses the empirical results. The final section concludes.

The widespread accumulation of international reserves by central banks in recent years is often explained as a precautionary buffer against the risks of international financial integration. Although theoretically plausible, empirical tests of this argument are still scarce. This article analyses empirically whether financial crises affect the level of reserves. It tests whether crisis experience makes a difference. The evidence for a large panel data set of emerging and developing countries shows that currency crises induce a permanent increase of reserves. This effect is particularly strong for recent currency crises since the Asian financial crisis. The more currency crises a country suffered from, the higher the level of reserves is. These findings are robust to different definitions of currency crisis and across different estimation methods. With less robustness, similar effects can be observed in the aftermath of banking crises. Twin crises cause additional reserve increases, although this effect is not significant across all specifications. Central banks revise their reserve policy after the experience of a currency crisis. The enormous accumulation of reserves in the recent past can partly be explained by the incidence of crises. While central banks seem to be better prepared for prospective crises, the financial crisis of 2008–2010 has shown that they are reluctant to use their reserves: Central banks dealt with exchange market pressures rather by exchange rate depreciations than reserve losses (Aizenman and Hutchison, 2010). This “fear of reserve loss”, however, does not question the plausibility of a policy of reserve accumulation in the aftermath of crises. It rather points to the asymmetric effects of reserve changes: Whereas reserve losses may aggravate crises due to the negative signal they transmit, markets perceive reserve increases as improving fundamentals. Transmitting positive signals might be especially important in the aftermath of crises.