1 Draft 7/25/04 Assessing Reserve Adequacy in Asia Jung Sik Kim, * Jie Li, ** Ozan Sula,** Ramkishen Rajan, *** and Thomas D. Willett** C I. Introduction and Overview There has been considerable controversy about the huge accumulation of international reserves in recent years by a number of Asian countries. Wide ranges of explanations have been offered for this behavior. Some have argued that it reflects blatant mercantilism that is putting undue adjustment burdens on other regions such as Europe. At the other extreme are interpretations that these accumulations have just reflected prudent reserves management in light of the Asian crises and the absence of the development of strong quasi lender of last resort capabilities by the IMF. In between, * Yonsei University, [email protected]** Claremont Colleges, Claremont Institute for Economic Policy Studies, and Freeman Program in Asian Political Economy *** Adelaide University and Claremont Colleges C Corresponding author, [email protected]
32
Embed
Assessing Reserve Adequacy in Asiafaculty.washington.edu/karyiu/confer/tok04/papers/kim&etal.pdf · Assessing Reserve Adequacy in Asia Jung Sik Kim,* Jie Li,** Ozan Sula,** Ramkishen
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Draft
7/25/04
Assessing Reserve Adequacy in Asia
Jung Sik Kim,* Jie Li,** Ozan Sula,** Ramkishen Rajan,*** and Thomas D. Willett**C
I. Introduction and Overview
There has been considerable controversy about the huge accumulation of
international reserves in recent years by a number of Asian countries. Wide ranges of
explanations have been offered for this behavior. Some have argued that it reflects
blatant mercantilism that is putting undue adjustment burdens on other regions such as
Europe. At the other extreme are interpretations that these accumulations have just
reflected prudent reserves management in light of the Asian crises and the absence of
the development of strong quasi lender of last resort capabilities by the IMF. In between,
* Yonsei University, [email protected] ** Claremont Colleges, Claremont Institute for Economic Policy Studies, and Freeman Program in Asian
Political Economy *** Adelaide University and Claremont Colleges C Corresponding author, [email protected]
2
there are explanations based on smoothing exchange rate movements and providing
temporary stimulus to help overcome domestic recessions. Undoubtedly countries’
behavior is often motivated by more than one objective and the relative importance of
these differential possible motivations may vary across countries and overtime.
Empirical discrimination among the differential explanations is not an easy
matter and we do not attempt to do so here. Rather we take this controversy as
motivation for investigating concepts of reserve adequacy and their implications for
whether current reserve levels of a number of Asian countries are excessive, insufficient,
or about right.
There is of course already an abundance of theoretical literature on this subject
and broad understanding that in a world of substantial capital mobility traditional
measures of reserve adequacy in terms of months worth of imports are of limited
usefulness. There have been several recent empirical studies of the demand for
international reserves, but as Christian Mulder points out at best these can only identify
whether a country is out of step with average behavior. It could still be following the
better policy.
Traditional models of the demand for reserves assumed that the probabilities
and magnitudes of reserve drains were independent of countries’ reserve holdings.
3
However, modern second generation crisis models emphasize the possibility of multiple
equilibria in a world of capital mobility where a countries’ underlying payments
position is neither quite strong nor hopelessly weak, i.e. where it is in a vulnerable zone.
In such circumstances, a country’s reserve level not only influences its ability to finance
speculative runs on its currency, but can also influence their probability of occurring.
Thus we can think of the demand for reserves as being influenced by three
types of considerations (besides costs), the ability to finance underlying payments
imbalances, the ability to provide liquidity in the face of runs on the currency, and the
preventive function of reducing the probability of runs on the currency. All of these
considerations will be influenced in turn by external and internal shocks, the degree of
exchange rate flexibility, the ability and willingness of governments to make domestic
policy adjustments, and the magnitudes of currency pressure that can be quickly brought
to bear. Of course, there is always the possibility of domestic currency holders running
for the exits, but it is widely believed that country’s exposure to currency runs is also
heavily influenced by the extent of foreign capital in the country, especially liquid
capital such as portfolio investments and short-term bank loans. This has led some of
the more sophisticated governments and central banks to develop rules of thumb for
reserve adequacy based on different types of international liabilities. The Bank of Korea
4
provides an example.
In this paper we address the issues both of sensible values for such coverage
ratios and also the likely sufficiency of focusing only on additive coverage ratios as
opposed to considering interactions among the different categories and
interrelationships with other variables considered in the demand of reserves literature.
We draw particularly on two types of empirical literature, the analysis of the
behavior of different types of capital flows during currency crises and empirical models
of currency crises. The latter have implications for the ability of high reserve levels to
protect countries from currency crises.
The former literature raises questions about the primary emphasis in some of
the recent theoretical literature (e.g. Calvo and Mendoza (2000)) and some policy
circles on the particular danger of portfolio investment. In the Asian crises the outflow
from bank loans was much greater. (See Willett et al (forthcoming)).
The second type of literature has clearly demonstrate the importance of high
levels of reserves relative to domestic monetary aggregate and short term foreign debt in
protecting countries from currency crises. While we know there is a big difference
between the effects of very high and very low reserve ratios, we know relatively little
about the relationships in between. Furthermore our knowledge of the interrelations
5
among these ratios and the strength of underlying fundamentals is still rather sketchy.
Sachs, Tornell, and Velasco (1995) assume and find empirical support for the
proposition that high reserves ratios can fully offset weak fundamentals. Bussiere and
Mulder (1999) argue that a significant higher reserve can limit the impact of contagion
even when the country is facing the situation of moderately high current account deficit
and appreciated real exchange rate. They also find empirical support for Greenspan and
Guidotti’s argument to hold reserves in excess of short-term debt by remaining maturity.
Bussiere and Mulder (2003), offer a simple rule of thumb based on the results of
empirical tests: the reserve target should be set at the level of short-term debt, which
should be augmented by 5% for each one per cent of current account deficit and by 1%
for each per cent of overvalued exchange rate. Willett et al (forthcoming b), question the
robustness of the STV conclusion that high reserves can offset weak fundamentals and
point out that it is at odds with our standard crisis models. With fundamentals in the
vulnerable zone high reserves could have a powerful effect in protecting against crises,
but with weak fundamentals first generation crisis models imply that reserve levels
should only influence the timing of crises, not whether they occur. This also suggests
that reserve needs should be related to the state of fundamentals in a non linear manner.
We will also investigate concepts of reserve adequacy where countries may be
6
subject to overly optimistic bubbles of capital flows which subsequently burst. While
extremely difficult to capture in formal optimizing crises models, many observers have
suggested that such behavior has characterized a number of recent currency crises
including Asia in 1997-98. See, for example, Willett et al (forthcoming b).
Clearly a country with strong fundamentals has less need for international
reserves. Reserves cannot be accumulated quickly, however, so in setting current targets
for reserve levels and rates of accumulation countries need to consider not only their
current fundamentals but a guesstimated probability function of the future evolution of
their fundamentals and external shocks.
Another approach would be to adopt the value-at-risk (VAR) methodology that
has become so popular with private sector financial risk managers. Based on historical
behavior this approach calculates the probability of different degrees of financial loss
over specified time periods. For reserve management purposes the analog would be the
probabilities of different size losses of reserves. This is analogous to the traditional use
of the volatility of reserves as an argument in the demand for reserves function. Two
caveats are important, however. First, as Long Term Capital Management found to its
dismay, different types of shocks can give rise to different patterns among returns so
that as the investment-brokers now warn, past performance is not a guarantee of future
7
returns. Typically VAR types statistical exercises use data over much too short time
periods to yield much confidence that they capture most of the range of possible
developments. Thus the designers of VAR methodologies recommend that they be
complemented by stress testing, that is, by imagining various types of shocks and
simulating these effects. This is similar to military planner efforts to calculate what it
would takes to be able to engage effectively in a specific set of actions. We suggest
that the developments during recent crises can offer useful information for these
purposes and illustrate how these experiences can be used for stress testing or scenario
analysis to help countries determine what levels of reserves would be sufficient to
protect themselves from crisis of the order of magnitude of recent ones.
Note that for these purposes vulnerability to capital account crises should not be
judged by standard measures of the variance of different types of capital flows. Such
measures confound the variability of rates of inflow – which will likely be of minor
importance for policy – with the size of capital flows reversals – which are much more
important. Indeed some studies have used very sophisticated econometric methodology
to study the variability of different types of capital flows, but since the data sample was
dominated by periods of capital inflows, the results had little predictive power with
respect to the magnitude of different types of capital outflows during the Asian crises.
8
The greater variability of portfolio investment during the inflow period was not matched
by greater capital flow reversals during the crisis period. The outflows from the banking
sector were for greater, both in absolute terms and as a percentage of previous inflows.
(See Willett et al (forthcoming a)).
We will present analysis of the ratios of reversals for different types of capital
flows during the Mexican, Russian, Brazilian, and Argentine as well as the Asian crises
to investigate whether theirs is a strong case for holding different levels of “reserve
backing” against different types of capital flows and consider alternative measures of
the “size” of recent crises and their implications for levels of reserve adequacy for a
number of Asian countries today. These measures will also be compared with recent
estimates of demand for reserves from work at the International Monetary Fund (Edison
(2003)) and by Aizenman and Marion (2002).
We conclude with a discussion of current policy issues related to reserve
management in Asia and the issue of whether excessive reserve accumulations are
posing serious problem for the global adjustment process.
9
II. New Benchmarks to Reserve Adequacy
For the post war period, the criteria of reserve adequacy was that reserves
should be sufficient to pay for three months of imports. This criteria was appropriate
when the capital flows are limited. However, as the emerging economies had liberalized
the short term capital movement during 1990s, most of the countries are exposed to the
risk of sudden capital outflow. Several years ago, Alen Greenspan, Chairman of Fed and
Pablo Guidotti, deputy Finance Minister of Argentina proposed using short term debt as
a yardstick of reserve adequacy. Bird and Rajan(2002) and Aizenman and Marion(2003)
examined the reserve adequacy using reserve ratio to short term external debt
empirically.
The debt based measures of reserve adequacy monitor only external drains,
and internal drain or capital flight is neglected. Wijinholds and Kapteyn(2001) proposed
a new criteria of reserve adequacy for the emerging market economy by including both
domestic and external drains.
However, these criteria also have some drawbacks in estimating the necessary
reserves for the emerging market economies. First, determining the appropriate level of
reserves for a particular country, one should focus on the most vulnerable items on the
10
balance of payments or main potential drains on reserves. The short term external debt
consists of several components such as bank loan, securities issued from aboard and
trade credit according to the BIS. Among these, it is known that the main drain on
reserve is bank loan because usually it is not renewed during the crisis. In this aspect,
bank loan rather than total short term debt should be concerned in the discussion of
reserve adequacy.
Moreover, all of the bank loan will not be drained during the crisis. No roll over
bank loan will be a main drain. The country will increase its reserve level when it has a
lower roll over ratio. However, it is difficult to get the roll over ratio which could be
applied to the estimation of reserve. Actual capital outflow in bank loan should be
concerned.
Second, some emerging market economies have very high proportions of
foreigner’s stock and bond investment in their domestic financial market. In these
countries, the portfolio outflow should be focused because the portfolio investment
could be another drain during the crisis. In the case of Korea, foreign stock investment
holding is 43% of current total stock value in Korea.
Although people worries about that most of the portfolio investment flowed out
during the crisis, in practice it is difficult that all of the foreign investment to be drained
11
during the crisis, because they will experience big loss by the fall of stock price and
large exchange rate depreciation. In spite of these losses, there could be some portfolio
capital outflow during the crisis. Thus, actual portfolio outflow should be measured for
the estimation of appropriate reserve.
Concerning these factors, in measuring the size of capital account crises and their
implication for the size of prudent precautionary reserves, two simple benchmarks come
to mind. One is the size of actual outflows during the crisis. The second is the change in
the size of net flows from their previous levels. (Because of year to year fluctuations the
average of several previous years should probably be used as the benchmark)
The first measure would be appropriate where the rest of the balance of payments
had been in approximate balance so that previous capital flows had their counterpart
primarily in changes in reserves.
The second measure would be most appropriate where prior to the crises the other
accounts, especially the current account, had adjusted fully to the net capital flows
yielding approximate overall payments balance. In such situation where previous capital
inflows were large, then a sizable fall in inflows could cause a problem. For example if
net inflows fall from 5 to 1 percent of GDP while the current account deficit remains at
5 percent of GDP, there would generate a financing adjustment problem of 4 percent of
12
GDP. This would create a greater need for reserves to smooth as cushion the adjustment
than in the case of a country that went from balance in the capital and current account to
an actual capital outflow of all or two percent of GDP. Thus sudden stops or capital flow
reversals needn’t always require actual net outflows to be a problem.
Of course, the two simple measures just described represent the two extremes of
zero and full adjustment to previous capital flows. Often the actual situation will be one
of partial adjustment. As a rough gauge of the degree of adjustment we could compare
the average change in reserves with the average net capital flows over the preceding few
years. We also need to remember that while the most dramatic of the recent crises have
usually been preceded by large net capital inflows, there isn’t always the case.
Traditionally crises have often been preceded by substantial periods of capital flight and
reserve losses.
III. Reserve Adequacy in Asia
1. New Benchmarks of Reserve Adequacy
(1) Capital Outflow Measure
13
As shown in Figure 1, most of the Asian countries have increased their reserve
holdings since the crisis in 1997. Especially, Korea, China, Indonesia and Japan’s
reserves have been accumulated rapidly since 1997 until recently.
<Figure 1> here
Table 1 shows the reserve adequacy of Asia in 2003 by the previous benchmark
which is the sum of the three months import and short term external debt. In table 1,
most of the Asian countries had excess reserves except Philippine and Hong Kong.
<Table 1> here
However this benchmark did not concern about actual outflow of short term
external debt, domestic drain and portfolio outflow, which could be main drains during
the crisis. In this case, our new benchmarks which are the actual capital outflow and
capital flow reversal could be useful criteria in estimating appropriate level of reserve.
Table 2 indicates the capital outflows in Asian countries during the crisis during
14
1996-1997. 1 During the crisis in 1997, there had been high capital outflows in Korea,
Indonesia and Thailand.
<Table 2 > here
Table 3 illustrates several combinations of different types of capital flows in
Asia, in which Korea had been experienced higher capital outflow based on the sum of
error and other loan.
<Table 3> here
Table 4 shows that the ratios of capital outflow over the GDP during the crisis.
Based on the benchmark of error and other loan, Korea, Thailand and Indonesia had
higher ratios. In table 5, capital outflows were measured by M2 standard during the
crisis and the ratios were higher in Korea, Indonesia, Thailand and Philippine. Table 6
indicates that the ratio of maximum capital outflow over M2.
1 In table 2, error means the domestic drain and portfolio(-) is the portfolio outflow. FDI is the net
outflow of foreign direct investment and other loan is the outflow of short term bank loan. FAC means the
balance of financial account. All data come from IFS data disk of IMF.
15
<Table 4> < Table 5> and <Table 6> here
(2) Capital Flow Reversal Measure
Table 7 and Table 8 shows the case of capital flow reversal which is the new
benchmark. The similar results are found as the benchmark of capital outflow. Korea,
Thailand and Indonesia had experienced higher capital flow reversals during the crisis.
<Table 7> and <Table 8> here
Table 9 and Table 10 indicate the maximum ratios of capital flow reversals
based on M2 standard in Asia during the crisis.
<Table 9> and <Table 10> here
2. Estimation of Appropriate Level of Reserve
In table 11, we estimated the appropriate level of reserve by applying these
16
ratios to the GDP and M2 of 2003. The estimated reserve level means that potential
capital outflow of 2003 in the case of crisis. Based on the results, the maximum capital
outflow might be $93billion and $52 billion in Korea based on the benchmark of M2
and GDP in the case of error and other loan. This outflow exceeded the actual reserve
holding in 2003. It is found that all of five Asian countries had excessive reserve based
on our new benchmarks as shown in Figure 2.
<Table 11> and <Figure 2> here
Even though we concerned about the three months imports, which are shown in
table 1, in addition to the potential capital outflow, most of Asian countries seemed to
hold excessive reserves.2
In Figure 3, the potential capital outflow in China is estimated during the crisis.
Here, we applied the both ratios of Malaysian and Thailand types’ capital outflows.
China’s reserve holding had been excessive based on the Malaysian type capital outflow,
whereas it was insufficient based on the Thailand type.
<Figure 3> here
Conclusively, the results based on the new benchmarks indicate that Asian
countries had excessive reserve holdings in 2003.
2 In the capital outflow estimation, we didn’t concern about the roll-over bank loans which was
negotiated with creditor countries or IMF during the crisis.
17
IV. Conclusion
Asian countries had experienced the crisis in 1997 by opening their financial
market during 1990s. Since then, they have accumulated international reserves.
However excessive reserve accumulation will result serious costs even though it also
has some benefits. Asian countries believe that benefits are greater than the costs, and
continued to increase the reserve.
They believe that reserve accumulation will impose less cost to their economy
which has weak fundamentals than other structural adjustment policies for the
prevention of another crisis. Moreover, export increases by intervening foreign
exchange market and exchange rate volatility might be reduced by smoothing operation.
However, there are serious costs of reserve accumulation. The money supply
and inflation will increase by intervention. Furthermore, the misalignment will result
speculation and finally the country will be exposed to the risk. Therefore, new criteria or
more specific benchmarks should be provided.
The previous benchmarks had three drawbacks. First, the standard is a little bit
ambiguous, not specific. Total short term debt is considered for the appropriate reserve,
whereas no roll over bank loan is one of the main drains. Second the portfolio
18
investment is not included in the estimation of reserve even though it could be flow out
during the crisis. Third, in estimating reserve, the previous benchmarks didn’t count the
actual capital outflows and reversals. It should be considered as the important parts of
appropriate level of reserve because those are related to the crisis.
In this paper, we examined the reserve adequacy in Asia by using new approach
to the reserve holding. New benchmarks of capital outflow and capital flow reversal are
tested and the necessary reserve levels are estimated. The results indicate that most of
Asian countries had excessive reserve holdings in 2003. Even though we concern three
months import payment, it finds that current reserve holdings of most Asian countries
are excessive levels.
19
References Aizenman, Joshua and Nancy Marion, 2002. “The High Demand for International Reserves in the Far East: What’s Going On?” Journal of the Japanese and International Economics, Oct. 2002. Bank of Israel Foreign Currency Department, Annual Report, 2001. Bird, Graham and Ramkishen Rajan, 2003. “Adequacy of Reserves in the Aftermath of Crises”, The World Economy, vol. 26, pp 873-891. Bussiere, Matthieu and Christian Mulder, 1999. “External Vulnerability in Emerging Market Economies: How High Liquidity Can Offset Weak Fundamentals and the Effects of Contagion”, IMF working paper. Bussière, Matthieu and Christian Mulder, 2003. “Which Short-term Debt over Reserve Ratio Works Best? Operationalising the Greenspan Guidotti Rule”.
Calvo, A. Guillermo, and Enrique G. Mendoza, “Rational Contagion and the Globalization of Securities Markets”, Journal of International Economics, No 51, 2000, p79- 113.
Chuhan, Punam, Gabriel Perez-Quiros, and H. Popper. “The Capital Flow Mix: Foreign Direct Investment, Short-term Investment, and Other Ingredients”. Working Paper, Santa Clare University, October 1997. Claassen, Emil-Maria. “The Optimizing Approach to the Demand for International Reserves: A Survey”, Recent Issues in International Monetary Economics, Edited by E. Claassen and P. Salin, 1974. Claessens S., Michael P.Dooley and Warner A. “Portfolio Capital Flows: Hot or Cold?” The World Economic Review, vol.9 1995, p: 153-174 Disyatat, Piti. “Currency Crises and Foreign Reserves: A Simple Model”, IMF working
20
paper, WP/01/18. Edison, Hali, 2003. “Are Foreign Exchange Reserves in Asia too High?” World Economic Outlook, Chapter 2, IMF, Sep. Flood, Robert and Nancy Marion, 2001. “Holding International Reserves in an Era of High Capital Mobility”. Gabriele A., Boratav K. and, Parikh A. “Instability and Volatility of Capital Flows to Developing Countries”. World Economy, v23, 2000: 1031-56 IMF, 1993. “Balance of Payments Manual” Fifth Edition. IMF, 2000. “Debt- and Reserve-Related Indicators of External Vulnerability”. Movchan, Veronika, 2002. “Criteria for International Reserves’ Adequacy: What Level of Reserves does Ukraine Need?” Mulder, Christian and Manuel Rocha, 2001. “A New Approach to Estimating Optimal Reserve Levels”, Working Paper, IMF. Nitithanprapas, Ekniti and Thomas Willett. “A Currency Crises Model That Works: A Payments Disequilibrium Approach,” Working Paper, Claremont Graduate University, 2000. Oh, Junggun, 2003. “Reserve Accumulation and Exchange Rate Policies in East Asia,” paper presented at the Claremont Conference on Exchange Rate Policy, 2004. Patnaik, Ila, 2003. “India’s Policy Stance on Reserves and the Currency”, http://openlib.org/home/ila. Sarno L. and M. P. Taylor, “Hot Money, Accounting Labels and the Permanence of Capital Flows to Developing Countries: An Empirical Investigation”, Journal of Development Economics, vol.59 1997, p:337-364 Wijnholds, J. Onno de Beaufort and Arend Kapteyn, 2001. “Reserve Adequacy in
21
Emerging Market Economics”, Working Paper, IMF, WP/01/143. Willett, T., A. Denzau, A. Budiman, C. Ramos and J. Thomas, “The Falsification of Four Popular Hypothesis about International Financial Behavior during Asian Crises”, The World Economy (forthcoming). Willett, T., E. Nitithanprapas, I. Nitithanprapas, and S. Rongala, “The Asian Crises Reexamined,” Asian Economic Papers (forthcoming).