The motives behind foreign exchange hoarding: a test on neo-mercantilist and precautionary arguments

Size: px

Start display at page:

Download "The motives behind foreign exchange hoarding: a test on neo-mercantilist and precautionary arguments"

Clarence Simmons
5 years ago
Views:

1 MIEF Analytical Paper The motives behind foreign exchange hoarding: a test on neo-mercantilist and precautionary arguments Franco Gomes Ortiz Sebastián Herrador ABSTRACT This paper studies the extent to which international reserve (IR) hoarding countries relied on previously accumulated reserves to fight currency depreciation in the midst of the 2008 financial crisis. Contrary to previous studies, which deemed the growing tendencies for IR accumulation as an exclusive sign of precautionary behaviour, evidence in this paper suggests that some monetary authorities may have been influenced by neo-mercantilist motives to the extent that they were willing to allow sharp exchange rate depreciations following the financial crisis. INTRODUCTION With the fall in the early 1970s of the fixed exchange rates supported by the Bretton Woods system, international finance moved into an era of increasing uncertainty as more and more economies accepted the idea of floating exchange rates and de-pegged their currencies from the dollar. Nonetheless, the instability inherent to a system of floating exchange rates brought a new set of problems and policy tools to deal with them. In particular, for emerging markets economies (EMEs) a long history of crises, especially leading into the late 1990s and early 2000s, provided strong incentives for monetary authorities to adopt structural changes so as to increase the rate at which international reserves were accumulated. Almost on a parallel level we saw the results of the commonly associated natural effects of the large current account surpluses emerging markets ran in the 1990 s and 2000s leading to large increases in the foreign exchange reserves being held by central banks. Some economies kept lower levels of reserves even at fast 1

2 growing rates, for example Japan and Korea, and only started hoarding reserves at point in which that growth had substantially slowed down. As such, because we know that similar to private banks, central banks also - usually - enjoy autonomy to decide how to allocate their financial portfolio, why do we see this particular portfolio allocation choice across many countries. Such question has raised a debate between two theoretical responses. On the one hand, some authors argue that central banks have accumulated reserves to preserve their economies from asymmetric shocks, for example, the Asian crisis in the late 1990s. In their favor, one can argue that much of the hoarding has happened in the aftermath of such financial crises. On the other hand, a part of the literature points out the competitive neo-mercantilist behavior displayed by some economies, which maintain their exchange rates artificially low so that domestic product have a competitive advantage on international trade. However, despite this theoretical discussion, it is considerably hard to identify the motives for reserve accumulation in the real world, and the attempts at this issue so far have been only partially satisfactory. As such, we propose a novel approach to test whether countries were accumulating reserves to protect themselves from asymmetric shocks, or if the behavior is more consistent with the neo-mercantilist approach of maintaining exchange rates artificially low. To do so, we look at the behavior of exchange rates, reserves and interest rate differentials for six economies before, during and after the fall of Lehman Brothers in the US, an event that raised uncertainty in international financial markets and caused abrupt capital flows across various countries. The next section discuss the existing literature on foreign reserves accumulation, our empirical research and our interpretation of the data. LITERATURE DISCUSSION International capital flows are inherent to open market economies. On the positive note, the international flow of capital creates opportunity: it allows better geographical allocation of resources, and makes possible investment ventures that may have been considered too risky for the national levels of savings. On the other hand, one cannot oversee the possible pervasive connection between the money markets and the real economy. As the market s perception of sovereign risk changes, speculative capital or hot money may move 2

3 away as fast as it moved toward national economies, posing a liquidity threat to domestic credit, and constraining business. Furthermore in a world of wide spread use of floating exchange rate regimes, sudden stops of capital flows or speculative attacks have an added layer of complication. As the economy becomes more integrated with foreign markets, the values of assets and liabilities held by nationals shifts from being denoted exclusively in national currency to a mix of national and foreign currencies. As such, the natural adjustment in exchange rates following such an event leads to a real economy disruption. National governments have a implemented myriad of tools to control the volatility of their exchange rates, some more intrusive than others, including: managing interest rates, changing reserve requirements, using foreign exchange reserves to increase or decrease liquidity in the domestic market, and taxation on currency conversion (Tobin tax). As discussed in the previous section, our policy instrument of interest in this paper are the foreign exchange reserves. In Dooley, Folkerts-Landau, and Garber (2005) the authors summarize the debate about the motives behind the unprecedented accumulation of reserves by emerging economies world wide. On the one hand, under the previous discussion about the uncertainty related to capital flows and their potentially pervasive real economy reverberations, the accumulation of reserves would be a precautionary measure undertaken by governments to fight off such speculative attacks. Furthemore, on top of the insurance facet, holding foreign exchange reserves would also increase the credibility of the national market to investors, reducing the risk of investing and thus attracting more capital. On the other hand, another hypothesis for the underlying reasons for the hoarding of foreign exchange reserves is export competition. In this narrative, countries would accumulate reserves to keep their currencies artificially low, and thus have an edge over its industry competitors in international trade. This is commonly referred to in the literature as the neo-mercantilist motive for accumulating reserves. Aizenman and Lee (2006) use a novel framework to show that the two motives for accumulating reserves have two very different impacts. While the precautionary motive adds to the predictability and stability of the economy and reduce the risk of an output fall following a sudden stop of capital, the neo-mercantilist motive would create a less than optimal environment that dissipates any real gains in productivity. 3

4 However, despite these two motives having clear-cut differences in theory, it is not so easy to distinguish between them in practice. For example, the hoarding of reserves was absent in Japan and Korea during periods of fast growth, however, as growth subdued in these economies the hoarding of reserves initiated. Such behavior is consistent with both the precautionary and neo-mercantilist hypotheses and, as such, a different test must be applied. In the following section we present a novel methodology to test the hypothesis presented above, and interpret our results. (AUTHORS NOTE: THIS ANALYSIS MAY BE EXPANDED IN THE FINAL DRAFT BY USING MORE COUNTRIES AND DIFFERENT SPECULATIVE/SUDDEN STOP EVENTS.) DATA (1) Increase in reserves note: despite not being part of our sample, China is introduced in the graph for comparison purposes. We select six countries (Brazil, India, Indonesia, Korea, Mexico and South Africa) that have displayed sizeable increase in their reserves to GDP ratio in the past 3 decades. Under our categorization, such countries can be considered reserve hoarders as their levels of reserves have increase at least by a fact or 2 since the 1980s. (2) Pressure to Depreciate As we strive to assess the policy response in our selection of countries with regards to asymmetric shocks in the shape of capital 4

flight and currency depreciation, it is important to understand how their corresponding exchange rates faced an increasing pressure to depreciate just as uncertainty began to build up the precedent

5 flight and currency depreciation, it is important to understand how their corresponding exchange rates faced an increasing pressure to depreciate just as uncertainty began to build up the precedent months before the financial crisis finally detonated. To this end we proceeded to analyze the evolution of overnight interest rate differentials between our countries of interest and the US. In particular, we pay close attention to the period between January 2007 and December We begin by stating the uncovered interest rate parity (UIRP) condition: Δ E e = [(E /E) 1] = R R * where R is the nominal i th country interest rate ( i.e. the immediate rates less than 24hrs interbank rate for the i th country), R * is the U.S. effective federal funds rate, E is the spot nominal exchange rate in LCU / USD, and E e denotes the expected exchange rate next period. Note that,we to assume that the effect of a risk premia ρ on exchange rate determination is negligible. This assertion is supported by the fact that we are looking at overnight rates so ρ is relatively too small to have a significant effect over Δ E. What we see is that, leading up to the core of the financial crisis, US bonds were undoubtedly perceived to be the relatively safest assets. The subsequent drop in R * was inevitable given that capital fled the relatively riskier economies and was being used to purchase treasuries. As the price of these treasuries kept going up, the downward pressure on R * US continued to build up and this lead to an imminent decline of almost 300 bpts over the course of a year. Similarly, the sudden sell-off of EME bonds also lead to a rise R EME. (See Appendix 1) This two tendencies together lead to a widening in the gap between R and R * which would contribute to the pressure on the exchange rate to depreciate as Δ E. The following 5

6 table shows the degree to which each of the currencies in question depreciated in the last quarter of 2008: CURRENCY DEPRECIATIONS % Change in Exchange Rates [LCU/USD] from August to December 2008 India (Rupee) Indonesia (Rupiah) Mexico (Peso) Brazil (Real) Korea (Won) SouthAfrica (Rand) 13% 23% 33% 49% 31% 30% These depreciations are the direct result of the enormous dimension of capital flight during this period as manifested by the abovementioned pressure on the interest rate gap. To illustrate this relationship the following charts depict the monthly percentage change in the gap between R and R * for each of the countries in our sample and how it is related to the eventual depreciation of the currencies in question: 6

7 These figures show that the pressure on these currencies to depreciate began to build-up considerably from January 2008 onwards. In light of these differentials, one would expect a dramatic fall in the international reserves of those countries that had amassed reserves for precautionary reasons. However, the degree to which the foreign reserves of our sample countries decreased did not reach the necessary level to prevent their currency to depreciate: DECLINE IN RESERVES % Change in Foreign Reserves [ex Gold] from August to December 2008 India Indonesia Mexico Brazil Korea South Africa -12% -10% 0% -4% -16% 1% The table above shows that Mexico, Brazil and South Africa did little to nothing to prevent their currency from depreciating. Although Korea, India, and Indonesia did use their international reserves to smoothen and postpone the exchange rate depreciation, this was clearly not enough to prevent it. (see Appendix 2 for charts on Exchange Rate vs Foreign Reserves) (3) CAPITAL CONTROLS High (2/42) Upper Middle (3/26) Korea (gate) Brazil (gate) India Mexico (gate) Lower Middle & Low (2/32) Indonesia 7

8 South Africa (gate) Note: Following Klein (2012), Open ( Walls ) countries have, on average, capital controls on less than 10 percent (more than 70 percent) of their transactions subcategories over the sample period and do not have any years in which controls are on more than 20 percent (less than 60 percent) of their transaction subcategories. Gate countries are neither Walls nor Open. We use Fernandez, Klein, Rebucci, Schindler and Uribe (2005) categorization of capital controls to test the randomization of data sample of countries. Our goal is to avoid selecting a sample of countries that had only very high or very low capital controls. (AUTHORS NOTE: OTHER COUNTRIES MAY BE ADDED TO THIS SAMPLE OR TAKEN DOWN AFTER PROF. ARROYO s REACTION TO THIS DRAFT) INTERPRETATION (1) The reserve hoarding countries in our sample have continued to amass foreign reserves even immediately after experiencing high levels of capital flight in the midst of the crisis in (2) Although there have been some efforts to smooth the rate of depreciation in some countries, we see that all countries allowed their currency to depreciate between 10-40% in a period of less than 5 months. (3) In particular, in countries like South Africa, Mexico and Brazil we reject the hypothesis that they were accumulating reserves for precautionary motives. (4) Hypothesis for the observed behavior: a. Defending currency might mean that there is a high ratio of liabilities in FX, and a sharp depreciation would overburden the financial system) b. Large share of exports in the economy (this sector would benefit from a weaker currency) c. More? TENTATIVE CONCLUSIONS To summarize, all countries in our sample suffered sharp devaluations of their currencies in the aftermath of the 2008 crisis, nonetheless, the pattern of use of foreign exchange reserves within the sample shows contrasting policies. On one end, Korea used about ⅙of its reserves to smoothen or postpone the devaluation of the won. On the other extreme, South Africa s central bank did not interfere, and accumulated more reserves over the period. The overall effect of the outflow of capital in both exchange rates is quite similar, at about 30%. 8

9 Such results lead to some tentative conclusions and other unanswered questions. First, there is strong evidence - even if not definitive evidence - to suggest that South Africa s and Mexico s accumulation of reserves is not for precautionary or self-insurance motives. Should that have been the case, we would expect to see some sort of reaction from the central banks to the sharp devaluation, which did not happen. Furthemore, the accumulation of reserves by South Africa in this critical period suggests that maintaining a stable exchange rate was not a priority for the South African government, an analysis that can also be extended to Mexico. The other countries display different patterns of foreign exchange reserves use. Korea, India, Indonesia, and Brazil use their reserves to smoothen the 2008 devaluation, but the size of the intervention is differs considerably. Brazil uses less than 1/20 of its international foreign exchange reserves in the four months after the turning point of the crisis, whereas Korea uses 4 times as much. Considering solely these two countries, the size of the impact on the exchanger rates are are large, Brazil suffered the sharpest devaluation (49%), whereas Korea s won lost 30% of its value. We take these interventions as evidence that these countries were trying to maintain their exchange rates protected from speculative attacks - a stronger evidence in the Korean case, and a weaker evidence in the Brazilian case. Nonetheless, a strong argument against our conclusions is that the countries gave-up fighting their exchange rates while still having considerable amounts of reserves under their belts and thus being able to keep the fight. While that might be so, the idiosyncrasies of the 2008 crisis may play an important role in the pattern of foreign exchange reserves use, and the expectations of the agents that were formulating the policies at the time. We know that the devaluation of currencies was a world wide event that happened in a very short time-frame. Information about the severity of the crisis at the time was not reliable, and it may be the central bank s who did initially intervene but at a later point allowed their currencies to devalue sharply did so because of they were uncertain that they could prevent the capital outflow, and did not want to exhaust their reserves fighting a lost battle. This view is consistent with the fluctuation in exchange rates after december 2008, when 3 out of 6 (Brazil, Indonesia and South Africa) exchange rates returned to their original pre-crisis levels. Even though analyzing the specific case of the 2008 crisis shed much light on the behavior of central bank s in face of a sudden stop of 9

10 capital flows, the episode itself does not provide information about consistent behaviors by central banks. As such, more research is needed to ascertain how central bank s reacted to smaller and possibly more manageable sudden stops. In that connection, a possible expansion of this study is to investigate the use of reserves in other - less apocalyptical - financial shocks. REFERENCES To be included in the final draft. APPENDIX (1) Evolution of the Exchange Rate pressure to depreciate as measured by % changes in interest rate differentials (2) Exchange rates vs Foreign Reserves 10

11 11

Ten Lessons Learned from the Korean Crisis Center for International Development, 11/19/99. Jeffrey A. Frankel, Harpel Professor, Harvard University

Ten Lessons Learned from the Korean Crisis Center for International Development, 11/19/99 Jeffrey A. Frankel, Harpel Professor, Harvard University The crisis has now passed in Korea. The excessive optimism