Management of International Reserve: a Case Study of China


The remarkable foreign reserves in China during the last decade have been boosted forward the world. Particularly in the aftermath of Asian financial crisis, China enable to succeed in avoiding the financial shock and economic meltdown and recovering in short time among others Asian countries mainly owe to huge protective reserves hoarding. Nevertheless, the soaring foreign reserves have triggered a series of debates on China's reserves adequacy. This paper explores theoretical interpretations for the demand for foreign reserves on the basis of previous literatures and empirically estimates the determinants of reserves holding. The implication of management reserves holding in China is presented as well. Some proposed suggestions based on the management targets can be found at the end of the paper.

Key words: foreign reserves, imports, external debt, Asian financial crisis, management measures

1.0 Introduction

In 2006, China's foreign reserves have reached more than US$850bn which makes China replace Japan as the largest holder in the world. From the last decade, particular in the aftermath of Asian financial crisis, foreign reserves have been growing fast with the well development economy in China.

In 2000, China's foreign reserves accounted for 14.33 percent of its GDP which is consistent with the last year. However, from then on, the ratio has been increased continually and was 61.66 percent of GDP by the end of 2007 (see Table 1.0). Figure 1.0 describes the tendency of the foreign reserves to GDP ratio from 1995 to 2007. As can been seen, after the 97-98 Asian financial crisis, the ratios still maintained in a constant level which is roughly 15 percent. However, since 2000, it had been a rapid soar from 15% up to 50% within only five years. This dramatic growth has attracted the eyeballs from all over the world and aroused a series of debates on the issue whether it is a rational behavior of foreign reserves market. Some believes that it is beneficial to China's economy due to a number of reserves positive functions while others criticise the huge reserves holding bring out potential risk to financial system and international markets.

In fact, increasing reserves holding not only show the Chinese overall economic strength has grown, but have also caused internal and external pressure. Zheng and Yi (2007) analyse the pressure in detail:

Internally, government debts are mounting as a result of official operations of sterilization of net capital inflows and inflationary pressure is increasing. Externally, China faces intensified pressure for a rapid appreciation of currency. The huge foreign reserves can cause imbalances in the country's economy.

Regarding such pressures faced by Chinese government, it is necessary to reconsider whether foreign reserves are excessive and what measures it should take to management large reserves hoarding. To address this question, we have to explore what is the cause behind the huge reserves holding in China.

As we know, the obvious reason for China's fast-expanding reserves holding in recent years is continual surpluses in the current and capital accounts of China's balance of payments. Figure 1.1 illustrates China's imports and exports volume from 1995 to 2007. The comparison in the histogram reflects an imbalance between imports and exports of China. We can analyse the balance sheet of payment further from the structure of reserves composition.

Theoretically, foreign reserves can be divided into creditor's rights-based reserves and debt-based reserves. The former is reserve from current accounts, while the latter is from the capital accounts. Foreign reserves of developing countries are mainly composed of debt-based reserve (Wang, 2006). From table 1.1, it can be seen that in 1995 and 1996, China's foreign reserves were mainly from the capital account which resulted from the increase in foreign direct investment (FDI). Since 1997, the creditor's rights segment of China's foreign reserves is the mainstream. However, during 2003-2004, there is a dramatically rebound in debt-based reserves from US$52bn up to US$111bn. This was consistent with the growth rate of reserves holding. More broadly, it is remarkable that the expectation for renminbi appreciation so high that resulting in a large amount of speculative capital (see Table 2) has inpoured.

Wang, (2006) points out that Central Bank in China has implemented a compulsory foreign exchange settlement policy to maintain the renminbi exchange rate. The principle policy is to enhance foreign reserves in support of exchange rate policy which is an important factor behind expanding foreign reserves. However, this is only one aspect of the huge reserves holding in China. In this paper, we are aiming to explore various aspects hidden in huge foreign reserves holding of China and discuss what measures should be implemented to manage such large foreign reserves and have a better utility of them.

2.0 Literature Review

2.1 Introduction

Since 1960, a series of studies had been published to explore the issue of international reserve adequacy and its management strategy. Among them, Clower and Lipsey, (1968), Grubel have presented explicit reviews of the previous literatures. They have contributed substantially to the stock of knowledge in this field. However, these studies mainly focused on the last 20 years before which may not suitable for the researches today. Therefore, a more comprehensive and renewal review of the literature on the issue of international reserve should be displayed. That is the main purpose of what we are going on here. In the following sections, we will present review both on theoretical and empirical literatures up to date in order to show how its theories and methodologies develop.

2.2 Demand for Reserves: Early Views in 1960s to 1970s

Apart from the motivations of international reserves holding, the optimal demand for foreign reserve is the mainly focus of economists.

Machlup (1966) concludes that There is no sense in which it can be said that the world total of monetary reserves is inadequate. Indeed, since the 1960s, the economists have been attempting to find out the actual demand for the international reserve. And it seems no one can give an affirmative answer on reserve adequacy. In their studies, all econometric estimation of the demand for international reserves are based on the fundamental task that the econometrician is to measure determinants of the demand. Clower and Lipsey, (1968) give a pessimistic view that we can not really hope to judge the overall adequacy of reserves in the present situation in which destabilizing speculation is encouraged. The earlier views of the reserves holding in uncertainty are passive by the reason that in practice all determinants of the foreign reserves are unpredictable, no matter what factors they are faced, such as stability of trade in goods and services, long-term capital flows, or foreign and domestic price and investment. Some economists believe the probability distributions of future values of these variables tend to have smaller variances. It may be true that reserve carry little weight relative to economic environment in the earlier period. Furthermore, in 1971, the study from Flanders suggests that the determination of foreign reserve holdings may be a non-economic phenomenon due to 'policymakers' who just maintain the reserves holding above some minimum critical levels. As can been seen, the mainstream of the earlier period studies is in a passive atmosphere.

After going through the literatures in early 1960s and 1970s, the studies of the issue on international reserve are apparently not enough, particular in the analysis of reserve adequacy. Thus, for the purpose of analysis, the principle task for researchers is to define measurement of reserve adequacy.

All authors recognise that not all forms of international reserves can be convertible with equal speed and capital certainty, thus, it is hard to define a standard measurement to estimate the adequacy in all countries. However, in practice, it has been necessary to define reserve as assets which are acceptable at all times to foreign economic units (Heller, 1966, p296) and measurable. Qualifying under these and similar definitions are national holding of gold, convertible foreign exchange, and unconditional drawing rights with IMF, for which the IMF provides ready and consistent estimates for a large number of countries. These estimates can be seen in the studies of M. Flanders, (1971), Machlup, (1966), Heller, (1966), Kelly, (1970), and P.B Clark, (1971). Moreover, when some authors argue whether these reserve assets should be adjusted for private liabilities towards foreigners, most studies consist with P. B Clark's view that calculations which based on gross and net IMF positions turn out similar results. The theoretical discussions of the nature of reserves holding in the studies are severe. In any case, the studies at that time indeed give a direction to economists to explore in depth.

With respect to the previous literature on the determinants of foreign reserves holding, there were a large number of empirical papers in the 1960s and early 1970s, such as Kenen and Yudin (1965), Kelly, (1970), Flanders, (1971), Herbert, (1971) and Frenkel (1974). These studies on the demand for reserves holding can be divided into three groups according to the methodologies of analysis they employ. They are ratios, regression techniques, and computer simulation. Here, we mainly focus on the ratios indicators and regression methodologies.

It can be realized that all studies of reserve adequacy have to deal with the problem that foreign reserves in different specific countries can not be compared, traced through time, unless they are standardised in some way to reflect difference in countries' size or economic condition. As we know, the scale variable in the most widely used is imports. Besides imports, two more classes of scale variables used in the studies are domestic money supplies and liquid liabilities, net external balance. These variables reflect countries' potential exposure to increase in aggregate demand for goods and services and have a close relation with the reserve demand.

Ratios as a tool to measure reserve adequacy is less efficient because the most difficult problem facing the users of ratios is how to interpret past changes and predict future ones in these ratios. Many analysts using ratios show recognition of the issue involved, but they are more likely to use of ratios as the best empirical tool available for the analysis of reserve adequacy in the past and for the prediction of future demand. According to Brown's analysis (1964), he believes that demand is satisfied when ratios is constant. In contrast to Machlup's study (1966), the efficiency of ratios indicators is suspected in the analysis of the demand for foreign reserves. He shows the difficulties involved in establishing theoretically why any of the ratios he selected should be constant across countries or through time. His statistical analysis demonstrates that all ratios are very different for countries and periods under examination. To sum up, ratios as a tool to measure reserve adequacy is not as efficient as analysts expect.

Therefore, more and more researches trend to employ the regression methodologies rather than the ratios indicators. The mainly regression technique for estimating reserves demand functions is to apply time series and cross-section which also can be seen in Table 2.0. Courchene and Youssef, (1967) used seasonal time series to estimate night countries' reserves holdings as a function of their imports, money supply levels, and long-term interest rates. The main innovation of their study was to move in the promising direction of estimation demand functions for individual countries from time series rather than for a group of countries from a cross-section sample. Another economist Thorn (1967) used the 14 countries sample of specific countries and considers the 1960 reserves to import ratio for each country as it reserve target and test for the years 1954,1957, 1962, and 1964, the equation are as followed:


where R is actual reserves, I denotes as imports, and is the ratio of reserves to imports of countries i, where i=1...14.

The shortcoming of Thorn's model is that the coefficients took on the values zero (a0) and one (a1 and a2) merely reflects the fact that during the period under observation both reserves and imports grew roughly at the same rate. Kenen and Yudin, (1967) has pointed out the weakness and improve by a rewriting form: (2)

The uses of variables and regression technique are largely identical but with little differences. Until Kelly published the study in 1970, there is an important innovation in estimating technique. The study combined annual data for the period for 46 countries in on main regression with a dummy variable for each country. Cross sections were also computed. The empirical results are very good which resulting in high correlation coefficients and most regression coefficients with high statistical significance and the right sign. Another innovation in that period is from Clark, 1970 who uses an autoregressive equation in the tradition of Kenen-Yudin to derive a proxy for individual countries' payment instability. It establishes that the demand for reserves and the speed of adjustment are an increasing function of payment instability. This may be worth pursuing further to find out the more relations in the estimation of demand functions from long run time series. Considering the empirical researches, it suggest that deriving reserve demand functions for individual countries from time series is easier than estimating a more generally valid function from a cross section of countries. That is because larger numbers of structural characteristics of economies which determine the demand for reserves of individual countries can not be measured adequately and incorporated in cross- section studies. Thus, if estimating one single country reserve adequacy, it is reasonable to assume there is only small change through time so that the observed statistical relationships among variables are stable and constant and reflect rational economic behavior.

2.3 The recent empirical literatures in 1980s and 2000s

After 1970s, the literatures focused on using international reserves as part of the management measures of exchange rate regime (Frenkel and Jovanovic, 1981). They develop a model also known as the buffer stock model or inventory model which concentrates on the demand side. It based on the idea that optimal foreign reserves are determined by the balance between macroeconomic adjustment costs and opportunity costs of holding foreign reserves. He believes that ooptimal foreign reserve depends on the variability of international transactions and reserves are a buffer stock to accommodate fluctuations in external transactions. Demand for reserves is a stochastic process driven by payments and receipts. Expected reserve holdings are a positive function of volatility and a negative function of the opportunity costs of maintaining reserves

Frenkel-Jovanovic Model is shown as below:


where Rt is reserve, W (t) is the standard Wiener process1 with mean zero and variance t. dt is the change in reserves in a small time interval, which is normally distributed and the distribution of reserves can be defined as:


where R* is the optimal initial level of reserves, μ is deterministic part of the instantaneous change in reserves and σ is the standard deviation of the Wiener increment in the reserves.

A further econometric formulation can be seen as followed:


where b0 is fixed cost, b1 is adjustment cost, b2 is opportunity cost.

And the empirical evidence is very close to their theoretic prediction which supports the justification of the model. The model predicts that average reserves depend negatively on adjustment costs, the opportunity cost of reserves, and exchange rate flexibility; and positively on GDP and on reserve volatility, driven frequently by the underlying volatility of international trade. Overall these predictions are supported by the literature of the 1980s. But, it failed to explain the recent reserve hoarding in emerging markets.

Since the economic environment has been taken place in last decade, more recent studies are provided by Ben-Bassat and Gottlieb (1992), Bordo and Eichengreen (1998), Aizeman and Marion (2004) to renew the development of reserves management. These papers typically study a much smaller number of countries and a much narrower set of explanatory variables than are considered in this paper. For instance, the Ben-Bassat and Gottlieb, (1992) argue that a central bank determines the optimal reserve level R* by minimizing the following expected cost function:


where π is the probability of a country's default due to economic and financial turmoil. The Equation tells that the total cost of holding reserves is the summation of the expected cost of forgone returns on reserves R and the expected output loss C due to the country's economic or financial crisis. The first condition of the cost function is:


Solving the condition, it can be rewritten as:

R* = (1-π(R*))/ πR (R*) + C/r (8)

It reflects optimal reserve is a function of the cost of failure, net opportunity cost and the absolute and the marginal default probabilities at given reserve level. They also suggest Central Bank should choose the level of reserves to minimize expected costs of holding them, which are the net costs of loss due to economic crisis and opportunity costs of holding international reserves.

Another study of Bordo and Eichengreen (1998) includes only twenty-one countries and considers only country size, trade openness and volatility measures as potential determinants of reserves. It seems a small size of the view and not in a adequately estimation. The updated study from Aizeman and Marion (2004) examine the contribution of reserves and external debt to tax smoothing for the case where future productivity is random, showing the welfare gain associated with hoarding a potentially large stockpile of international reserves. It is noteworthy that sizable precautionary demand for international reserves exists even if agents are risk-neutral: the non-linearity associated with costly taxation and sovereign risk suffice to induce a first order demand for reserves, independent of risk aversion.

They also build up models which belong to the broad optimising model family, and are known as the self-insurance models as against the buffer stock model by Frenkel-Jovanovic Model. Furthermore, in 2005, Aizenmean and Lee go further in the research to extent Ben-bassat and Gottlied model which implies foreign reserves are output stabilizers that can reduce the probability of an output drop induced by a sudden stop of foreign capital and/or the depth of the output collapse when the sudden stop materializes. The study also has empirical findings that the precautionary demand is the main motivation of hoarding reserve against Mercantilism of reserve accumulation (Dooley et al., 2003).

2.4 Literature reviews on emerging market

There is a series of researches arise from the Asians crisis since 1997. The views on the reserves holding in Asian emerging market in the aftermath of crisis are various.

Primarily, it is necessary to reconsider the reasonable reserves levels. It is suggested from the crisis lesson that the currency reserves holding level can be an important indicator and predictor of a financial crisis (Fischer, 2001). A better reserve and debt management system may have reduced the possibility of crisis. The conventional indicators of reserve adequacy based on current account instability may be appropriate for some countries, which may be faced with payment vulnerability due to trade related shocks (Bird and Rajan, 2003). However, for other countries such as the emerging market economies(EMCs), which carry a substantial short-term external debt, the ratio of R/M2 are unable to capture the payment shocks. It is necessary to consider the size of reserves and short-term external debt (Greenspan, 1999). The composition of short-term debt has been found as a key indictor of illiquidity in the financial crises such as that of Mexico in 1994-1995, East Asia in 1997-1998, and Russia and Brazil in 1998-1999 (Dadush et al., 2000; Hawkins and Klau, 2000). Indicators for assessing foreign reserve adequacy for emerging market are not identical for each country. In the next part, we will specify some indicators in empirical estimation to find out how to assess demand for China foreign reserves.

The literature review on the issue of the demand for international reserves leads to a conclusion that the earlier studies in 1960s and 1970s are well developed and several seminal papers gives new direction to the research on international reserve adequacy by showing that econometric estimates of demand functions for reserve on the basis of past experience which can not provide information about optimal reserve supplies in the future. Regarding with the recent research in recent years, the analysis trend of reserve adequacy is more comprehensive and well improved the previous weakness of ratios and regression techniques. The innovation of reserve issues can be seen as a well presentation in the literatures.

3.0 Empirical Specification and Finding

3.1 Introduction

In this part, we are going to perform empirical assessment to rethink the management of international reserves by the Chinese monetary authority. We first employ methodologies to illustrate the mainly factors which are behind the large accumulation of international reserves in the last decade. By taking a retest on foreign reserves and other related macroeconomic variables, we attempt to give a comprehensive cognition of the optimal level of hoarding and how to manage such massive foreign reserves in China.

3.2 Data and Methodologies

In this paper, we design to employ two methodologies which based on the development theories of foreign reserve adequacy suggested by the economists mentioned in the Part 2. In order to estimate whether the current foreign reserves of China is in a suitable level which adapt to balance of payment and exchange rate regime, the ratios as the most obvious tools of analysis will be demonstrated in the empirical evaluation. We are going to look at three indicators for this assessment based on the prevailing theories of reserves from 1960s to 1970s which are reserves to imports; reserves to M2 and reserves to net external balance. Meanwhile, the assessment period will be focus on the last ten years from 1995 to 2007. That is because during the last decade, China has experienced a tough economic environment in the aftermath of Asian financial crisis and entered a rapid development era, hence, this time series data and ratios may be useful and characteristic to well present the economic features.

The original estimated data such as China's foreign reserves, import, export and M2 figures are sourced from the Datastream database and some statistics are gathered from the NBSC (2007). The compare and contrast will be given in the result analysis section. As a significant indicator, although it has the weakness, ratios is still an irreplaceable tool to give a general image of the reserves adequacy of one country.

The second tool of analysis is to use econometric formula to estimate optimal holding according to the models on reserves adequacy. As can be seen in the literature review, there are a number of well developed models that demonstrate different situations of reserves holding. Considering of advantages and shortcomings among these models and the direction of assessment, we intent to employ two mature models suggested by Frenkel and Jovanovic (1981) and Aizenmean and lee (2005) in the empirically evaluation s of reserve management in China.

Frenkel-Jovanovic mode as a buffer stock or inventory model has remarkably successful in explaining foreign reserves holding in the 1980s. It has performed empirically and proceeds to useful outline of optimal reserves. This model hypothesises that the reserves authority will choose an initial level of reserve holding that minimizes its total expected costs which consist of macroeconomic adjustment costs and opportunity costs of holding foreign reserves. In the working paper from Flood and Marion (2002), it points out that the two costs are interrelated due to a higher stock of reserves. The optimal reserve management is to find the cost-minimizing level of reserves to acquire once reserves have reached their lower bound. That is to say, a reasonable level of reserves holding is to optimise the trade-off among the costs produced from the hoarding. Furthermore, Frenkel and Jovanovic (1981) assumed that reserve movements between the occasional re-stockings are generated by an exogenous Wiener process, where the incremental change in reserves in a small time interval is distributed normally.

Based on the previous FJ formal modelling formulations (3) and (4) shown in part 2, a new derivation of the optimal reserves holding developed by Frenkel and Jovanovic (1981) can be presented as followed: 2


Where R* is the optimal reserve; C is a fixed cost per adjustment, related to the capacity to adjust expenditures to income; r is the cost of holding reserves per unit of time, representing the cost of forgone earnings. Finally, σ is standard deviation of the Wiener increment in the reserve time-series process operating between stock adjustments.

In the equation (9), optimal reserve holding increase with the volatility of reserves (), and higher volatility equals that reserves hit the lower bound more frequently. Therefore, the money authority prefers to restock a large amount of reserves and tolerate greater opportunity costs so as to incur the adjustment costs less frequently. It can be also seen that a bigger adjustment cost lead to a larger optimal reserve holdings while a higher opportunity cost shrink them. In a word, it reveals optimal reserve holdings to be a positive function of volatility and a negative function of the opportunity costs of maintaining reserves.

To be more straightforward, the expression (9) can be transformed into the log form which is a successful estimating equation:

12 (10)

For the purpose of empirical test, this estimating equation finally becomes a convenient econometric formulation like below: 3


It is quite straightforward to use the econometric formulation to estimate the variables. Because FJ realise that obtaining a variability measure was free of scale, thus, to make it feasible and simple, they divided the standard deviation by the value of imports and the opportunity cost, r, was approximated by a country's government bond yield. They also define that the component estimating coefficients b0, b1 and b2 are namely fixed cost, adjustment cost and opportunity cost. In particularly, the constant b0 is interpreted to be country specific and regime specific.

In the empirical work from FJ, they estimated the models by using ordinary least squares on various cross-section and panel data sets. They once estimated across 22 developed countries during the 1971 to 1975 period. The result can be seen as below:


R2= .97, n=110, S.E= 0.234

Their empirical evidence is closely consistent with their previous theoretical finding since the estimated coefficients of reserve holdings with respect to and r is very close to the predictions of the theoretical model given by equation (10).

After explaining the methodology, we attempt to replicate the FJ regression equation in China's foreign reserves analysis by using PcGive program. The data sets are in time series which come from Datastream and NBSC as well. It is noticeable that the estimation may not be accurate due to technique issue and scaling issue. But we will try to make the process clear and integral. We now commence our effort to retest this model in China's foreign reserves and discuss the findings in detail in the next section. The output of the estimation can be shown in appendix.

On the other hand, we will go further to estimate the China's foreign reserves via new theoretical model suggested by Aizenman and Marion (2004). Unlike the general framework, there are some limitations of FJ model which lead to under perform in emerging market. The first limitation is using fewer factors as scaling variable for determining reserve adequacy. Crises during the 1990s and beyond have predominantly been crises of the capital account. Reserve adequacy benchmarks accordingly need to be modified to allow more variables influenced on reserves (Bird and Rajan, 2002). The other limitation is that FJ model does not explicitly capture changes in loss aversion. According to the precautionary motive on demand for reserve adequacy, holding reserves may be considered a form of insurance premium, which means that reserves hoarding has the functions to avoid future crises and reduce the probability of an output drop induced by a sudden stop of foreign capital and so on. It can be thought as an approach to stabilise fiscal expenditure in developing countries (Aizenman and Marion, 2004). That is to say, in emerging market, the political factors may have a great impact on reserve holding rather than developed countries. China is the most important developing counties in Asia and we should consider various factors in the management of foreign reserves. Thus, we commence to retest the reserves holding in China on the basis of the extension of theoretical model generalised by Aizeman and Marion (2004).

The estimating equation is as followed:


where: R is actual holding of reserves minus gold; pop is the total population of specific country; gpc is real GDP per capita; exa is the volatility of real export receipts; imy is the share of imports of good and services in GDP; neer is the volatility of the nominal effective exchange rate.

Regarding to the study of Aizenman and Marion, (2003), they point out that the real holdings should increase with the size of international transactions; hence, they expect reserves holding to be positively correlated with the country's population and standard of living. Moreover, reserves holding should increase with the volatility of international receipts and payments which would make it positively correlate with the volatility of export receipts; and it is also expected to increase with the vulnerability to external shocks which means be positive with import. And finally, since the severe fluctuation of exchange rate will reduce the demand for reserves thus, reserves should be negative with the exchange rate. Aizenman and Marion explain the choice of variables in detail in order to construct a structural model.

In the case of China, making the test simply and clear, we intend to apply reduced-form of the model and use a few of significant variables instead of complication. It is remarkable the sample data is during the period from 1998 to 2007. We expect to find out the change of management of reserves holding in the aftermath of Asian financial crisis.

The model we estimate involves variations of the following equation:


where: Rt is the actual reserves holding; ca is the capital account balance; ed is the net external debt; imy is the ratios of import to GDP; RM2 is the ratios of reserve to M2; neer is the volatility of the nominal exchange rate(standard deviation of exchange rate movement).

The estimation output is reported in the appendix. The discussion of the finding will be presented in the next section as well.

3.3 Result Presentation and Analysis

3.3.1 Ratios Indicators

The four key ratios indicators will be introduced in this part to outline the structure of foreign reserves in China. Although ratios indicator can not tell the full condition of the reserves holding, they still play an important role on the analysis of reserves level and reflect the significant problems.

i. Reserves-to-GDP(R/GDP)

The first principle ratio presented here is Reserves to GDP which has mentioned in introduction. Table 1.0 shows the actual data and the R/GDP ratio from 1995 to 2007. In the half end of 1990s, reserves holding in China were in a stable growth level. Consisting with the development of economy, R/GDP has been growing fast, especially after 1998. In 1999, China's foreign reserves accounted for 14.5 percent of its GDP. In the next 2 year, until 2000, the ratio seems maintain in the similar level. However, it has been increased again in 2001 and was reach more than 60 percent of GDP by then end of 2007 (see table 1.0), which made China become the largest foreign reserve holder in the world. Obviously, the reserves holding follow the development of economy.

ii. Reserve-to-M2(R/M2)

Another indicator is R/M2 is considered as a predictor of financial crisis (Kaminsky et al., 1997). A higher level of this ratio will reduce the probability of crisis. Figure 3.0 reflects there is a gradual increase of the ratio after 2000, it shows evidence that china has strengthen the management of reserves holding after Asian financial crisis. The R/M2 can be a useful indicator of the potential impact of capital flight, but, further discussion needs to be supplemented by analysis of other possible sources.

iii. Reserves-to-IMPORT(R/IM) and Reserves-to-EXPERNAL DEBT(R/STED)

The most common ratio in previous foreign reserves literatures, R/M ratio reflects how sufficient of reserves on the foreign import. But Bird and Rajan, (2003) find out it lacks a linear relationship between the occurrences of deficit and value of imports which may fail to reflect the actual condition. However, we still consider it as a main measurement. It can be seen from table 3.0 that the R/IM ratios are very high in China and Korea and Malaysia are less than 3 before the Crisis. It is remarkable that Korea has boosted the ratio from 2.9 up to 8 after the crisis which implies a great adjustment in trade policy.

The last indicator which is considered most relevant for emerging markets (Wijnholds and Kapteyn, 2001) is R/STED. STED is denoted as short-term external debt which debt obligations with an original maturity up to one year. It indicates that countries with low value were most likely to be affected by the financial crisis (Dadush et al., 2000). Table 3.0 summarizes the reserve adequacy indicator: R/IM and R/STED in five main east-Asia countries from 1995 to 2000. They implies that Korea, Malaysia, Hong Kong have all been linked to financial crisis and related to a weak R/STED. China was able to manage a stable exchange rate during the Asian financial crisis due to its comfortable R/STED.

3.3.2. Regression technique

On the basis of FJ model formulation, we retest China's foreign reserve during the period from 1995 to 2007. In this regression, we apply the traditional determinants of the demand for international reserves such as GDP, IMPORT, and EXTERNAL DEBT. It should be noticed that when scale R by these factors, the right-hand side variables have to do the same procedure. In the table 3.1, it reports the results using the three scaling variables mentioned above. As can be seen from the output, we interpret our result is consistent with those FJ model equation (10) since the estimating coefficients is approximately close to the theoretical prediction. Our approach differs from the primarily because we use others scaling factors. In the estimation, reserve volatility still positively influences reserve holdings. It can be seen that the reserve volatility in benchmark regression (a) of table 3.1 has a closest elasticity of 0.58 without scale factor. Meanwhile, the interest rate has a negative and significant coefficient in all regressions as well. Moreover, the value of R2 in first regression also reflect a

The estimating coefficients in other three regressions with scaling variables have the similar values which are around 0.4. It is clear that reserve holding mainly depend on GDP factor since the elasticity of reserve volatility in regression (b) shows the highest value among others. Further, IMPORT has a great impact on reserve holding as well. As the opportunity cost, the coefficients of interest rate have a wide range of value although they are all in negative. In sum, the result of the FJ model in China's foreign reserves is receivable although parts of the parameters are not perfect.

Although FJ model has a good empirical support in reserves holding under certain assumptions, it is noteworthy that it can only predict for developed countries and fail to explain the recent reserves hoarding in emerging market, especially after the Asian financial crisis. Therefore, in the following section, the regression result of the new model (14) introduced in section 1 will be discussed to find out what factors influence on China reserves holding after crisis.

We reconstruct the original model (13) from Aizenman and Marion, (2004) by adding five variables that assume to affect the reserves holding.

In order to assess the responsiveness of international reserves, we run a new form regression which relates change in international reserves to current account, net external debt, import ratios, money supply and exchange rate movement. The output of result is reported in table 3.2.

By regressing the new form model (14) in PcGive, we find out that the criteria of adjusted R square and DW test both show a perfect goodness of fit of the model. When our adjusted R square is very close to 1 and DW is around 2(see table 3.2), it means our regression line is perfectly fit the data offered. From the test result, it can be concluded that the factors of import ratio (LIMY) and M2 money supply ratio (LRM2) have a higher value of coefficients which imply a positive correlation with reserves holding. Although current account balance (LCA) and outstanding external debt (LED) have small value, it still indicates they are positive with reserves holding. Moreover, it is suggested that the volatility of exchange rate has a negative correlation with foreign reserves. In general, the empirical result above is consistent with the suggestion from Aizenman and Marion, (2003). It can be realised from the result that political factors such as fiscal M2 supply, external debt and import volume play an important role to determine reserves hoarding in China. Like Aizenman and Marion suggestion, hoarding of foreign reserves can be viewed as a precautionary adjustment, reflecting the desire for self-insurance against exposure to future sudden stop of foreign capital inflows or economic crisis loss. From the test, we are aware of that political factors may be the dominate determinants to affect the management of foreign reserve in China rather than opportunity cost and adjustment cost.

4.0 Finding and Suggestion

4.1 Finding and extension

From above empirical estimation, we can conclude two important findings in this paper. Firstly, there is a positive effect on the large accumulation of foreign reserves in China according to the ratio indicators. Particular in the Asian financial crisis, China enables to survive in the hazard has to attribute to the huge reserves holding which protected renmingbi from devaluation and stabilise the financial market in Asian.

Secondly, from the result of our regressions on optimal reserves models, we can find out the latter model has a better simulation of the reserves holding in China than the former FJ model. It implies that the new variables (short-term external debt, M2 supply, volatility of exchange rate) added in the latter model is more relative to reserves holding China. It also indicates that adjustment of these significant factors needs to be considered carefully in management reserves holding. The key findings give us an opportunity to rethink the management measures in reserves holding in China. When we are proud of the first place of the reserves accumulation, the more important thing is to rethink the potential risk and a more advanced management.

A Chinese economist Li (2006) give his view on the reserves adequacy. He believes that China's foreign reserves are not excessive and in a suitable level since it needs sufficient reserves to maintain the stability of the yuan and to maintain the confidence of international investors. As can be seen in the Asian financial crisis, the mainly reason that China can survive among other countries should attributes to the massive foreign reserve which prevent yuan from depreciation. He also argues that China's foreign reserves have been rewarded by sufficient returns.

Nevertheless, China's spectacular economic development has presented China with a dilemma. Although foreign reserves could earn credibility and maintain stability of the Chinese currency, the accumulation of huge reserves has created pressure for the yuan to appreciate, and its further appreciation would result in financial losses and instability of the currency. Therefore, the issue whether China foreign reserves are excessive has been explored again. Some analysts suggest it would be reasonable to reduce them until an optimal level is reached. Frankel (2004) has emphasized the opportunity cost of massive foreign reserves and believe that China was presumably paying foreign investors on their inward investment a higher return than it was earning from its investment in foreign reserves. Another argument from Xia (2006) points out that approximately 22 percent of foreign reserves accumulated in 2005 were induced by expectations of the Chinese currency's appreciation, mostly in terms of short-term capital inflows. He maintained that US$700bn in foreign exchange reserves should be sufficient. He believes China foreign reserves are fiercely large in light of the adequate level.

To identify China foreign reserve whether is excessive or not, we need to analyse the essential causes hidden in the rapid growth of holding. In terms of our study, it can be explained by a few points.

i. Exchange rate maintenance

For a long time, China has been blamed for manipulating exchange rates regime to maintain international competitiveness. It intervenes in the foreign exchange market to prevent real appreciation of the yuan. However, from July of 2005, China reformed the exchange rate regime by moving into a managed floating exchange rate regime based on market supply and demand with reference to a basket of currencies. RMB has no longer been pegged to the US dollar. After the reform, the RMB exchange rate has become more flexible.

It is noteworthy that since 2005, the expectation for renminbi appreciation has been growing continually and a large amount of speculative capital has swarmed in. Therefore, the Central Bank has adopted a compulsory foreign exchange settlement policy to maintain the relative stability of the renminbi exchange rate, which is an important factor behind expanding foreign exchange reserves.

ii. Financial Stability

As we know, China did not face financial shock during the 1997-1998 Asian financial crisis due to the strong capital controls. Witnessing the damage inflicted on the financial and corporate sector of its neighbours it took measure to tighten the strict foreign debt management as well. At the end of 1997, its foreign debt was 373 US$ billion. Its capacity to service foreign debt, as measured by the growth of foreign reserves, has kept pace with the growth of foreign liabilities. Consequently, the key debt ratios have been kept well with the generally accepted safety limits. In particular, the foreign reserves at that time are four times larger than short-term foreign debt which protects China from economic meltdown. That is why china favors having sizeable reserves holding because it insists that large foreign reserves help to maintain the foreign confidence deeded for attracting foreign direct investment and securing foreign loans at good terms (Ford and Huang, 1994).

It is worthy to rethink the reasons behind such huge foreign reserves in China. For the precaution functions of reserves holding, it is difficult to tell what level should be optimal or adequacy to reach a safety protection. On the other hand, if consider the economic effect, we need to have a standard for foreign reserves to minimise the cost induced by holding. In the special case of China, it is more complicate to confirm a optimal level of reserves because its has a complex economic and political environment. In this paper, we suggest to focus on the management of its reserves holding.

4.2 Suggestions on Management China Foreign Reserves

Rregarding huge China's foreign reserves, there are theoretical misconceptions that magnifying the function of reserves holding and the more reserves the safety economics are. Under this the view of passive precautionary motive, it is suggested that reserves should not be used until the critical moment. Consequently, foreign reserves have in reality not been effectively utilized although there has been a large pool of them. Reflecting such a view, in the management of foreign reserves, factors such as the economic benefit, social benefit, risk control, efficiency of operation have been ignored. This has led to the management of reserves rigid and inflexible.

As mentioned before, foreign reserves are the important part of a country's monetary policy. Change in reserves is a tool for balancing international payments, and it is usually a result of intervention in the foreign exchange market by the central banks.

However, management of the foreign reserve is not only a policy objective, but also a protection of the export competitiveness of domestic enterprises. This is related to measures to boost economic growth and to ensure full employment, both of which are key internal factors in maintaining the balance of development. Therefore, the rapid expansion of reserve should be considered not only for political motives but also the economic motives. It should be worthy to rethink the reserves management measures comprehensively and feasibly.

The present condition of foreign reserves in China has been concerned broadly. There is a worry about the potential danger in massive reserves holding which may result in excess load for China. Therefore, it is necessary to adjust measure so as to reduce the risk and even gain profit from the large reserves. For purpose of that, controls the scale of foreign reserves and strengthen monitoring are the key principle of the management.

In the present situation, it is more important for China to implement feasible management and make good use of them.

China has long implemented the traditional trade strategy which led to sustained trade surplus and rapid expansion of foreign exchange reserves. It would not only exacerbate the pressure for renminbi appreciation, but also carries great risk. Moreover, reforming and maintaining stable renminbi exchange rate regime will be good to the foreign exchange markets which can slowdown the rapid growth in foreign exchange reserves.

Apart from reducing its adverse impacts on the macroeconomic, another task is to boost economic growth and establish a new mechanism to transform and use reserves. Further, risk management of foreign reserves should be strengthened and a risk management framework should be established and improved as well. More broadly, the efficiency of reserve operation should be lift. Finally, the capital structures of foreign reserves should be considered in a rational way.

According to these directions, we propose some available measures: a) spending and investing foreign reserves; b) gradual liberalization of the capital account; c) reserves hoarding diversification.

Since foreign reserves can be regarded as indirect debts of the government against companies and individuals, large foreign reserves could be used to import petroleum and other essential raw materials, high technology, even to invest in key sectors abroad. However, expending and investing reserves would increase liquidity risks problems. Thus, policy decisions on expenditure and investment should be carefully considered based on asset risks.

With regarding to capital liberalization, China government always keeps eyes on hot money. It allows efficient capital inflows and reasonable capital outflows through a strictly monitoring system. However, it is difficult to distinguish hot money inflows from normal short-term inflows. Moreover, RMB appreciation and net capital inflows could easily reverse if the government lifts control on capital outflows immediately. In case it happens, foreign reserves holding would still not be sufficient to satisfy the extremely high demand for foreign exchange and this could lead to a financial crisis. Therefore, capital account needs to be monitored before such thing happen.

Among the options, diversification would allow China to diversify reserves out of US dollar securities and switch into others countries bonds, given concerns over a weakening US currency. However, US dollar is still the main currency to be used in international payment. No other currency has the similar high level of liquidity as US dollar. Therefore, diversification of reserves asset may add volatility of foreign exchange markets and can catalyze abrupt exchange rate movement. As the result, we suggest the policymakers in China adopt an international reserve diversification standard including disclosure fully of the composition and having gradually adjustment from small scale reserves.

5.0 Conclusion

Accumulation of foreign reserves in China was beneficial over the past decade, especially in and in the aftermath of Asian financial crisis. They help to maintain the stability of financial market, expand international trade, attract foreign investment, lower the cost of operation, and enhance financing capacity. Due to a number of positive functions, China has accumulated more and more foreign reserves and surpassed Japan to become the first place in the world from 2006. However, the recent rapid soar has brought about potential risks and worrisome problems for the Chinese economy such as loss of interest and the shrinking of the real value of foreign exchange reserves. In particular, large speculative capital inflows such as hot money have increased risks to the financial system and foreign exchange rate market.

It is suggested that large scale foreign reserves in China are more relate to the political factors such as the monetary policy and trade policy. Therefore, it is necessary for China government to make good use of foreign reserves and implement feasible management measures. The suggestions such as a strict control on capital flow, encouraging expenditure and investment with reserve, diversification on reserves holding asset etc are presented from the paper in order to provide available means on reserves management which may reduce risk and address the problem at present. However, these measures' feasibility analysis should be put forward in depth so as to find out which is more suitable on the basis of Chinese political situation and financial environment.

Foreign reserves continue to increase with a rapid rate in the beginning of 2008. It is sensible to realise that reserves holding should be controlled in a range of reasonable level to maintain the economy in a stably condition. Consequently, Chinese government must beware of this and spend more time on reforming the reserves holding management. It is expected that creative and available options can be implemented in the near future.


1 Kenen and Yudin (1965) treated reserves as a discrete-time random walk process which is probably similar to the continuous time Wiener process. Moreover, they argue that central bank reserve holding are sensitive to the volatility in the balance of payments rather than the absolute size of the gap between international payments and receipts. Claassen (1975) also used an inventory model of international reserve choice to explain the restocking costs. Under the FJ null hypothesis, foreign reserves follow a Wiener process until hit the lower zero boundary. After, in a one-step adjustment, reserves jump back up to their optimal restocking level adhere to Wiener process again.

2 Rajan and Bird, (2003) points out this equation is a second-order Taylor-series approximation of optimal reserves holding which isunder the assumption that the special case of no reserve drift between stock adjustments.

3 A key step in taking equation (11) to data is the additional assumption that observed reserves, Rt , are proportional to optimal reserves up to an error term that is uncorrelated with and r. Hence,.


  • Aizenman, J. and Marion, N., 2003, The high demand for international reserves in the Far East: what's going on? Journal of the Japanese and International Economies, Vol.17, pp. 370-400.
  • Aizenman, J. and N. Marion. 2004. International Reserve Holdings with Sovereign Risk and Costly Tax Collection. Economic Journal 114(497): 569-91.
  • Aizenman, J. and Lee, J. 2005, International Reserves: Precautionary vs.
  • Mercantilist Views, Theory and Evidence, IMF working paper, WP/05/198.
  • International Monetary Fund.
  • Bird, G and Rajan, R., 2002, The Evolving Asian Financial Architecture, Princeton Essays in International Economics, No.226, International Economics Section, Princeton University
  • Bird, G and Rajan, R., 2003, Too much of a good thing? The adequacy of international reserves in the aftermath of crises, The World Economy, Vol. 26 No.6, pp.873-91
  • Bordo, Michael and Barry Eichengreen, 1998, The Rise and Fall of a Barbarous Relic:The Role of Gold, Mimeo, UC-Berkeley.
  • Claassen, Emil, 1975, Demand for International Reserves and the Optimum Mix and Speed of Adjustment Policies, American Economic Review, Vol. 65, pp.446-453
  • Clower, R. and Lipsey, R., 1968, The present state of international liquidity theory, American Economic Review, May, Vol.58, issue.2, pp. 586-95
  • Courchene, T. J., and Youssef, G. M., 1967, The demand for international reserves, Journal of Political Economics, August, Vol. 75(4), pp. 403-413
  • Dadush, U., Dasgupta, D., Ratha, D.,2000, The role of short-term external debt in recent crisis, Finance & Development, Vol. 37 pp.54-7.
  • Dooley, Michael P., Folkerts-Landau, David and Garber, Peter M., 2003. An Essay on the Revived Bretton Woods System, National Bureau of Economic Research (Cambridge, MA), Working Paper No. 9971,
  • Fischer, S., 2001, Opening Remarks. IMF/World Bank International Reserves: Policy Issues Forum, International Monetary Fund, Washington, DC
  • Flanders, M. J., 1971, The demand for international reserves, Princeton studies in international finance, No. 27, Princeton: Princeton University Press
  • Flood, R. and Marion, N., 2002, Holding International Reserves in an Era of High Capital Mobility, IMF Working paper, No: WP/02/62
  • Ford, J.L and Guobo, Huang, 1994, The Demand for International Reserves in China: An ECM Model with Domestic Monetary Disequilibrium, Economics, Vol.67, August, pp.379-397
  • Frankel, Jeffrey., 2004, On the Renminbi: The choice between adjustment under a fixed exchange rate and adjustment under a flexible rate, paper presented in a seminar on China's Exchange Rate System, Dalian, China, 26-27 May.
  • Frenkel, J. A., Jovanovic, B., 1981, Optimal international reserves, The Economic Journal, Vol. 91, No. 362, pp.507-14.
  • Frenkel, J., 1983, International Liquidity and monetary control. In International
  • Money and Credit: The Policy rules, Ed. G. von Furstenberg. Washington, DC:
  • International Monetary Fund.
  • Greenspan, A., 1999, Currency reserves and debt, paper presented at Remarks before the World Bank Conference on Recent Trend in Reserve Management, Washington, DC
  • Heller, H.R., 1966, Optimal International Reserves, Economic Journal, Vol.76(302), pp. 296-311
  • Hawkins, J., Klau, M., 2000, Measuring potential vulnerabilities in emerging market economies, Bank for International Settlements, Basel, BIS Working Paper, No. 91
  • Heller, R.H. and M.S. Khan, 1978, The Demand for International Reserves Under Fixed and Floating Exchange Rates, International Monetary Fund Staff Papers, Vol. 25(4), pp. 623-49.
  • Herbert G. Grubel, 1971, The Demand for International Reserves: A Critical Review of the Literature, Journal of Economic Literature, Vol. 9, No. 4, pp. 1148-1166
  • IMF, 2005, IFS Yearbook 2005, Washington DC: IMF publication.
  • Kaminsky, G., Lizondo, S., Reinhart, C., 1997, Leading indicators of currency crises, IMF Staff Papers, No. WP/97/79.
  • Kenen, P. and Yudin, E., 1965, The demand for international reserves, Review Economic Statistic, August, Vol. 47, pp. 242-250
  • Kelly, M. G., 1970, The demand for international reserves, American Economic Reviews, September, Vol. 55(4), pp. 655-667
  • Li, Yang, 2006, China's foreign exchange reserves with sufficient return, Shanghai Securities News, 17 March (In Chinese)
  • Machlup, Fritz, 1966, The Need for Monetary Reserves, Banca Nazionale del Lavoro Quarterly Review, Vol.78, September, pp. 175-222
  • National Bureau of Statistics of China, 2007, Statistic and Data, [online] available from:
  • Thorn, R., 1967, The demand for international reserves A note in behalf of the rejected hypothesis, Review economic statistic, Nov, Vol. 49, pp. 623-627
  • Wijnholds, J.O., Kapteyn, A., 2001, Reserve adequacy in emerging market economies", International Monetary Fund, Washington, DC, IMF Working Paper, No. WP/01/143.
  • Xia, Bin., 2006, $700bn foreign exchange reserve is enough, Shanghai Securities News, April.