VoxEU Column Global crisis International Finance

International reserves before and after the Global Crisis: Is there no end to hoarding?

In the aftermath of the global financial crisis new patterns of reserve hoarding have emerged. This column identifies structural changes in international reserve accumulation. Emerging markets with higher savings rates tend to use higher buffers of reserves, partially accounting for the higher levels of reserves in east Asia compared to Latin America. While there is no end in sight for reserve hoarding, some of the newly identified factors may mitigate eventual reserve accumulation.

After the financial crises of the late 1990s many emerging economies recognised the benefits of self-insurance against the volatility associated with financial globalisation, increasing their ratio of international reserves to GDP substantially.1 The evidence in Bussière at al. (2014) supports this view, finding that countries with more reserves relative to short-term debt fared better during the Global Financial Crisis. The takeoff of reserves hoarding by China and other countries in the 2000s added new factors to the list of determinants of hoarding international reserves (IR), including mercantilist motives2 and self-insurance against local residents’ flight from domestic assets in the context of the trilemma.3

In Aizenman et al. (2014) we evaluate whether the Global Financial Crisis (GFC) and recent structural changes in the global economic environment are associated with new patterns of hoarding IR. This possibility is exemplified in the recent experiences of China and South Korea, both of which have undergone large structural changes that have impacted their IR-to-GDP ratios in the past decades. Since the GFC China has experienced a sizable decline in IR to GDP (Figure 1), reflecting a rebalancing of its export-led growth strategy in the face of declining global demand, a liberalisation of its outward foreign direct investment, and the placing of greater emphasis on its sovereign wealth fund (SWF) (Aizenman et al. 2014). In contrast, during the GFC, South Korea found that its sizeable and once-regarded-sufficient stock of IR failed to isolate its economy from massive deleveraging. The ensuing financial panic was ultimately abated only with the help of the Fed’s special swap lines.4 The experiences of China and Korea raise the possibility that the GFC may have induced structural changes in the behavior of IR holding, possibly motivating some countries to supplement their hoarding of reserves with new policies and institutions.5

Figure 1. International reserve holding as a ratio to GDP (a) and to the world total (b)

(a) IR holding as % of GDP

Note: For the country groups, the group’s aggregate IR is divided by the group’s aggregated GDP.

(b) IR holding as % of the world’s IR total

Note: For the country groups, the group’s aggregate IR is divided the world’s total IR.

The impetus of instituting a SWF has been based on the recognition that the primary mandate of the central bank is to conduct monetary policy and ensure financial stability. Hence, the opportunity cost of reserves may be of limited relevance to the central bank’s operations in practice.6 Given these considerations, a higher savings rate could increase the level of IR to GDP,[vii] while the presence of SWFs may lower IR to GDP for a given savings rate.8 Over time, the introduction of a SWF may reduce the exclusivity of IR as the main financial buffer. Effective prudential regulations may reduce external borrowing and the inflows of hot money, thereby reducing the need for IR hoarding for self-insurance purposes. Accessibility to bilateral swap lines may also mitigate the need for IR in times of peril.9

Methodology and results

Against this background we evaluate the stability of factors accounting for IR hoarding and the roles of conditional variables that had not been sufficiently studied before the GFC. We group the explanatory variables into three broad factors.

  • The traditional macroeconomic factors include the propensity to import, the volatility of IR holdings, the opportunity cost of holding IR, and exchange-rate regimes. These variables capture the elements of a vintage international reserve demand equation from the 1970s.
  • The financial factors include domestic financial depth (measured by M2/GDP), external financing, cross-border capital flows, and controls on capital flows.
  • The recently discussed factors includes several factors that have come to the forefront of recent discussions: the existence of national-level SWFs, bilateral currency-swap agreements, the implementation of macro-prudential policies, gross saving, outward direct investment, the composition of trade, the implicit-rivalry incentive (also known as the ‘catching-up-with-the-Joneses’ effect), and the discounted experience of past financial crises.

We confirm that the appropriate level of IR continues to evolve with developments in the global economy. During the pre-GFC period of 1999 to 2006, gross saving is associated with higher international reserves in developing and emerging markets. The outward direct-investment effect is consistent with the efforts of diverting international assets from the IR account into the purchase of more tangible foreign assets, the Joneses effect lends support to the implicit rivalry-hoarding motives, and commodity-price volatility induces IR hoarding against uncertainty while fuel exporters store their proceeds partly in the form of foreign reserves. During the GFC period of 2007 to 2009, many of the variables that are significant contributors to IR hoarding in the previous period become insignificant or display the opposite effect, probably reflecting the frantic market conditions that disrupted their normal economic relationships.10

The results for the post-GFC period from 2010 to 2012 are dominated by the ‘recently discussed factors.’ The negative effect of swap agreements and the positive effect of gross saving on the observed IR-to-GDP ratio are in line with our expectations. In the entire sample period, the existence of macro-prudential policies is found to complement the IR-accumulation policy.

We repeat the exercise using data from developed countries. In line with previous findings in the literature, the developed and developing countries display different IR demand behaviors. We find that the IR hoarding behavior of developed countries is affected by the recently discussed factors, including SWFs, gross saving, the Joneses effect, and trade compositions – even in the pre-GFC period, though their results often differ from those of developing countries. For example, gross saving has a negative impact on IR accumulation of developed countries, possibly because these countries have better accessibility to the global capital market where they can invest their savings. The catching-up-with-the-Joneses effect is robust among both developed and developing Asian countries.  Overall, the demand specification for IR holdings not only evolves over time, but also differs between developed and developing countries. All these suggest that the two groups of countries have different motivations for holding IR because they face different economic realities.

We close with an examination of the adequacy of IR holdings in the post-GFC period as was reflected in exchange market pressures. Our empirical analysis confirms that the “fragile five” countries (Brazil, India, Indonesia, South Africa, and Turkey) held fewer IR than our model predicted in this period. Such a situation reflects the vulnerability of these countries to global economic turbulence in the early 2010s. We test whether and to what degree economies with IR-holding levels below model predictions are susceptible to external shocks, focusing on the induced exchange rate depreciation against the U.S. dollar between 2012 and 2013 – a period dominated by the news coming from the Fed that QE may be tapering soon. We confirm a negative and significant correlation between the exchange rate depreciation against the U.S. dollar and our proxy for over-hoarding of IR given by the prediction error of IR holdings. That is, if a country held an insufficient amount of IR, it tended to experience depreciation in its currency value when adjusting to the tapering news.


We empirically confirm structural changes associated with new patterns of hoarding IR, especially in the aftermath of the global financial crisis. The saving rate has been playing an important role in determining the level of IR hoarding even before the GFC: emerging markets with higher saving rates tend to use higher buffers of IR, partially accounting for the higher levels of IR in East Asia compared to Latin America. Other variables we newly identified include the accessibility to swap lines, implementations of macro-prudential regulations, the existence of a sovereign wealth fund, and the attitude towards outward foreign investment.

While there is no end in sight for reserve hoarding, some of the newly identified factors may eventually mitigate reserve accumulation. The implementation of macroprudential policies may lead to fewer IRs if countries feel less need to hold precautionary IR, or it could lead to more if the policies help prevent a drainage of IR. Interestingly, we found a negative impact of macroprudential policies for developed countries and a positive one for developing countries.  The proliferation of SWFs and possible rebalancing of emerging markets that followed aggressive export-led growth before the GFC may reduce reserve-to-GDP ratios of developing countries, as confirmed in the predictive exercises using the latest data. The robustness of the ‘catching-up-with-the-Joneses’ effect also suggests potential gains from regional and global steps towards deeper use of swap lines and cooperative pooling arrangements.

These predictions should be taken with a grain of salt. Given the dynamic nature of the forces that shape the hoarding of reserves, there is no reason to expect future stability in the patterns of hoarding IR. On top of the relevance of the new determining factors, one should be aware of the possible shift in the hoarding behavior in the future.

Acknowledgements: The support of faculty research funds of the Dockson Chair, University of Southern California, the City University of Hong Kong, and Portland State University is gratefully acknowledged. We also thank Jacinta Bernadette Rico for her excellent research assistance. We are grateful to Yu-chin Chen and the participants at the April 2014 JIMF-USC conference for their insightful comments. All errors are ours. 


Aizenman, J, Y W Cheung, and H Ito (2014), “International Reserves Before and After the Global Crisis: Is There No End to Hoarding?” NBER Working Paper No. 20386.

Aizenman, J, M D Chinn, and H Ito (2010), "The Emerging Global Financial Architecture: Tracing and Evaluating the New Patterns of the Trilemma's Configurations" (with Joshua Aizenman and Menzie Chinn), Journal of International Money and Finance, Vol. 29, No. 4, p. 615-641 (2010).

Aizenman, J, Y Jinjarak, and N Marion (2014), “China's Growth, Stability, and Use of International Reserves. NBER Working Paper #19739. Cambridge: National Bureau of Economic Research, Open Economies Review, 25: 407-428

Aizenman, J and J Lee (2007), “International Reserves: Precautionary Versus Mercantilist Views, Theory and Evidence,” Open Economies Review 18, 191-214.

Aizenman, J, and N Marion (2003), “The High demand for International Reserves in the Far East: What’s Going On?” Journal of the Japanese and International Economies, 17, 370-400.

Aizenman, J, and N Marion (2004), “International Reserve Holdings with Sovereign Risk and Costly Tax Collection,” The Economic Journal, 114, 569–591.

Aizenman J, and G Pasricha (2013), “Why do emerging markets liberalize capital outflow controls? Fiscal versus net capital flows concerns" Journal of International Money and Finance, 2013, 39, pp: 28-64.

Bruno, V, and H S Shin (2014), “Capital Flows and the Risk-taking Channel of Monetary Policy", mimeo, Princeton University.

Bussière, M, G Cheng, M D Chinn, and N Lisack (2014), “For a Few Dollars More: Reserves and Growth in Times of Crises”, National Bureau of Economic Research Working Paper #19791. Cambridge, MA: National Bureau of Economic Research.

Cheung, Y and H Ito (2009), “Cross-sectional analysis on the determinants of international reserves accumulation”, International Economic Journal, Vol. 23, No. 4, p. 447-481. December 2009).

Cheung, Y and H Ito (2008), “Hoarding of International Reserves: A Comparison of the Asian and Latin American Experiences”, in Ramkishen S. Rajan, Shandre Thangavelu, and Rasyad A Parinduri (eds), Exchange Rate, Monetary and Financial Issues and Policies in Asia, World Scientific Press, p. 77-115.

Cheung, Y and X Qian (2009), “Hoarding of International Reserves: Mrs Machlup’s Wardrobe and the Joneses", Review of International Economics, 17(4), 824–843.

Obstfeld, M, J C Shambaugh and A M Taylor (2010), “Financial Stability, the Trilemma, and International Reserves”, American Economic Journal: Macroeconomics, 2: 57-94.

Park, Y C (2010), “Reform of the Global Regulatory System: Perspectives of East Asia’s Emerging Economies", Proceedings, Annual World Bank Conference on Development Economics, pp. 187-212.


1 The growing financial integration of emerging markets during the 1990s and the ensuing crises were identified as key factors in the structural changes in the motives to hold international reserves (IR), among which the weights of financial factors as well as the past crises history increased (Aizenman and Marion 2003, Aizenman and Lee 2007, Cheung and Ito 2008, 2009).

2 See Aizenman and Lee (2008), Cheung and Qian (2009).

3 See Obstfeld et al. (2010), Aizenman et al. (2010).

4 The experience of Korea illustrated the need to supplement reserves hoarding with prudential regulations dealing with balance-sheet exposure of systemic banking. Indeed, unlike the 1997–1998 Korean crisis, the recent crisis did not lead to a further increase in Korea’s reserves-to-GDP ratio but to prudential regulatory changes (Park 2010, Bruno and Shin 2014).

5 The GFC and the resultant quantitative easing (QE) policy by the Fed and other central banks also led to large, hot money inflows to emerging markets in search of yields. Emerging markets reacted to these developments by experimenting with dynamic capital controls aimed at mitigating the resultant appreciation pressure and reducing the exposure of future destabilising outflows. These dynamic policy reactions also included relaxing controls on outward capital flows to defuse greater appreciation pressures from larger inflows, as has been the case in China and other emerging markets (Aizenman and Pasricha (2013).

6 This also explains the failure in the literature to find a stable and economically significant impact of the opportunity cost of reserves on the observed international reserve-to-GDP ratios. 

7 Political economy considerations suggest another channel linking a lower gross savings rate with lower IR to GDP; such a scarcity of saving would make it harder for the central bank to maintain sizeable IR hoardings, as the reserve stock may be an administration’s target of opportunity at times of a fiscal crunch, as has been the experience of Argentina and Venezuela (Aizenman and Marion 2004).          

8 Once the level of IR (as a share of GDP) reaches a level high enough to cover self-insurance needs, countries – usually those with high saving rates – may opt to manage their public saving in their own SWFs. Unlike the central bank authorities, the mandate of SWFs is to secure stable income for future generations; therefore, a SWF generally has a higher risk tolerance than the central bank, and aims for higher-than-expected income and longer-term investments.

9 Although this applies only if the use of swap lines does not entail the stigma effect and if the swap-line arrangements are deemed durable.

10 Nevertheless, the propensity to import and gross saving continue to display strong positive effects (with greater magnitudes).

2,415 Reads