VoxEU Column Global governance

Why are reserves so big?

Between 2000 and 2009, developing countries added almost $5 trillion to their foreign-exchange reserves – a number deemed too high by many, prompting accusations of protectionism. But this column argues that developed countries are equally to blame – as well as failures in international coordination. It concludes that remedies therefore require action by both groups.

Foreign-exchange reserves play a crucial role in macroeconomic management. They provide a safety net during times of economic turmoil and, for most developing countries, a means to peg the nominal exchange rate. They also provide a means to manage windfalls from commodity exports or from sudden surges of capital.

A traditional benchmark to assess reserve levels is whether they are enough to cover six months of imports; another is whether they are enough to cover all short-term external debt. Recently, economists have proposed adding 20% of M2 to the benchmarks, since increased financial integration means that a large part of a country’s monetary base can head for the exits during a crisis (de Beaufort Wijnholds and Kapteyn 2001). Recently, the IMF also proposed a measure that combines exports, short-term debt and other portfolio liabilities, and M2 as indicators (IMF 2011)1.

These measures yield hugely different estimates of reserve adequacy, so developing countries now hold between $1 trillion and $4 trillion in excessive reserves. From 2000 to the start of 2011, the nominal stock of foreign exchange reserves in developing countries increased from around $750 billion (11% of GDP) to nearly $6.3 trillion (29% of GDP), a staggering increase compared to a rise from $1.3 trillion (5.1% of GDP) to $3.4 trillion (8.1% of GDP) in OECD countries. The accumulation in developing countries paused only briefly during the Great Recession.

Over 90% of developing country reserves are concentrated in the 20 largest holders, which now have enough reserves to cover over a year of imports or their short-term debt nearly five times over. Even according to the more demanding criteria recently put forward, a majority of these countries have excess reserves2.

Figure 1. 2009 reserve levels in the 20 largest developing country reserve holders ($ billions)

As always, the averages conceal large variation, as shown above. Reserves are below at least one of the two traditional measures in Mexico, Poland, and Turkey, but are particularly high in Asia. China alone accounts for over half the sample’s excess reserves.

How much is too much?

Despite these striking figures, estimates of reserve adequacy must be made with caution. Benchmarks provide a useful guide, but countries differ, including in the probability they attach to crisis and in aversion to risk. Just as some individuals buy minimalist healthcare because they do not fear risk or do not believe they will get sick, and others buy coverage for every conceivable treatment, countries too have different demands for insurance.

Moreover, though holding reserves entails an opportunity cost – they are, by definition, safe and low-yielding assets – it is estimated to be about 0.5% of GDP in the median emerging market (IMF 2011)3, though this opportunity cost is itself subjective since it entails an estimate of the expected yield on higher risk assets. This cost pales in comparison to the deep and prolonged cost and political disruption of financial crises, which can also entail a loss of sovereignty to international creditors in severe cases4.

Reserves cannot completely insure against crises, but countries with large reserves holdings are better able to maintain consumption growth during periods of market pressure. They also have greater fiscal flexibility, allowing them to further mitigate the effects of the crisis6.

The global liquidity glut

Why did developing countries begin to rapidly accumulate reserves ten years ago? Following its financial crisis, savings rates in developing Asia rose steadily from around 31% of GDP in the 1990s to 45% in 2009; investment, which collapsed during the crisis, was slower to return, rising to 41% of GDP, implying a large current-account surplus. Rising oil and commodity prices also played a role, with reserves in oil exporters reaching $1.5 trillion or so in 2010.

However, even a cursory review of the evidence suggests that, while these factors were clearly important, they are not the whole story. Policies in reserve currency countries helped create a “global liquidity glut” that contributed to the reserve build-up in many emerging markets. Though this story differs from the “global savings glut” theory, which hypothesises that increased savings in emerging markets forced lower long-term interest rates on advanced countries, the two explanations are quite complementary.

Beginning around 2000, the US, the UK, and peripheral Europe went on a spending binge that pushed their cumulative current account deficit from -0.5% of world GDP in the 1990s to -2% in 2005-2008. In the US, low interest rates, tax cuts, unfunded war spending, and a housing boom steadily widened the current-account deficit from -1.6% of GDP in the 1990s to -6.1% of GDP in 2006. Meanwhile, the interest rate decline associated with creation of the euro sparked a dramatic rise in demand in peripheral Europe, causing the average current account deficit in Greece, Ireland, Italy, Portugal, and Spain to widen dramatically from -0.9% in the 1990s to below -9% in 2008. The UK also saw a big housing and financial boom.

Meanwhile, prodded by improvements in developing countries and low international interest rates (most evident in deflation-stricken Japan), private capital flows to emerging markets grew from below 4% of emerging market GDP in 2000 to nearly 9% in 2007. This surge of capital (combined with current account surpluses) was, in many instances so big that it resembled a commodity price windfall and could not be absorbed quickly.

Developing countries could have responded in two ways:

  • By allowing the real exchange rate to appreciate, eventually leading to a reduced current account surplus and stopping capital inflows;
  • or accumulating official reserves.

As it happens, real exchange rates in major emerging markets appreciated strongly, by an average of 7.8% (though many countries saw different outcomes; see Figure 2). But the increase in reserves was much more dramatic.

As Figure 2 below shows, over 2000–2007, changes in reserves and real exchange rates in developing countries show no correlation. This puts some doubt on the claim that countries intervened mainly to avoid a loss of competitiveness, but is in line with a large body of literature which suggests that prolonged intervention often fails to influence real exchange rates, though it does impact nominal exchange rates6.

Figure 2. Changes in real exchange rates and reserve levels from 2000 to 2007

The story is incomplete without reference to coordination failures. If -- despite the literature’s findings – the main motivation for reserve accumulation was to prevent real-exchange-rate appreciation, had all emerging markets agreed to allow their exchange rates to appreciate together, losing competitiveness would have been less of a concern. Moreover, advanced countries, particularly those that would have benefited from the increased demand, would have probably let policy interest rates rise faster, thus reducing the capital flows to developing countries. If (as we suspect) reserve accumulation was more motivated by windfalls and precautionary concerns, however, coordination would only have helped had global macroeconomic stability fundamentally improved.

Policy implications

Trying to impose hard and fast limits on reserve accumulation would be both futile and undesirable. Individual countries have different perceptions on risk exposure and risk tolerance, and are willing to pay different amounts for insurance.

Policies should instead focus on the causes of excess global liquidity and volatility. The US, Europe, and Japan – who own the reserve currencies and still account for the large majority of world output and trade – will continue to determine the economic environment within which emerging markets operate. Until they regain their footing, fiscal deficits decline, and international interest rates rise, developing countries will continue struggle with windfalls of foreign money, and to seek insurance against global recessions and sudden stops.

That said, some emerging markets, beginning with China, should take a more serious look at their reserve levels and the associated costs. Excessive reserve accumulation is not only directly costly, but it can also contribute to inflation and overheating credit and asset markets. In addition, efforts to sterilise its effect on the domestic money supply can distort domestic banking systems (Mohanty and Turner 2006), while its effectiveness in preventing real exchange rate appreciation in the long run is at best unproven.

No size fits all, but enhancing international coordination – through the G20’s mutual assessment process, for example – could also help, provided it does not become another mercantilist negotiation or an alibi for inaction.


De Beaufort Wijnholds, J Onno, and Arend Kapteyn (2001), “Reserve Adequacy in Emerging Market Economies”, IMF Working Paper 01/143.
Hutchison, Michael M (2002), “The Role of Sterilized Intervention in Exchange Rate Stabilization Policy”, Mimeo, June.
International Monetary Fund (2011), “Assessing Reserve Adequacy”, International Monetary Fund.
Jurgensen, Philipe (1983), "Report of the Working Group on Exchange Market Intervention [Jurgensen Report]", US Treasury Department.
Mohanty, MS and Phillip Turner (2006). “Foreign exchange reserve accumulation in emerging markets: what are the domestic implications?”, BIS Quarterly Review (September):39-52.
Montiel, Peter J (1998), “The Capital Inflow Problem”, World Bank.
Reinhart, Carmen M and Kenneth Rogoff (2008), This Time is Different: Eight Centuries of Financial Folly, Princeton University Press.
Sarno, Lucio and Mark P Taylor (2001), “Official Intervention in the Foreign Exchange Market: Is It Effective and, If so, How Does It Work?”, Journal of Economic Literature, 39(3):839-868.
Truman, Edwin M (2003), “The Limits of Exchange Market Intervention”, in C Fred Bergsten and John Williamson (eds.), Dollar Overvaluation and the World Economy, Institute for International Economics.

1 Based on historical evidence, the authors derived the following formulas to calculate appropriate the appropriate levels of reserves: in fixed exchange rate regimes, reserves should be equal to 1 to 1.5 times the sum of 30% of short-term debt (STD), 15% of other portfolio liabilities (OPL), 10% of M2, and 10% of exports; in floating exchange rate regimes, reserves should equal 1 to 1.5 times the sum of 30% of STD, 10% of OPL, 5% of M2, and 5% of exports.

2 Twelve of the 16 countries for which there is sufficient data currently hold reserves in excess of short-term debt plus 20% of M2. Fifteen of the top 20 reserve holders are included in the IMF analysis; eight are determined to hold excessive reserves.

3 It should be noted that for countries with low reserves and high debt, the opportunity cost of reserves may be negative (i.e., a net gain) because increased reserves reassure creditors and lower the costs of debt service; for countries with reserves well in excess of benchmarks, such as China and Malaysia, the cost may be as high as 2% of GDP.

4 Reinhart and Rogoff (2008) find that financial crises, on average, reduce per capita GDP by 9%, while returning to the pre-crisis level takes an average of 4 years. In emerging markets, these effects are often even more severe.

5  However, these benefits diminish as reserve levels rise.

6 As Montiel (1998) notes, suppressing nominal appreciation through intervention in foreign exchange markets expands the money supply and increases inflation, implying a real appreciation. Policy makers can sterilize this intervention by selling government bonds, thus removing liquidity and reducing real appreciation pressures. However, the effectiveness of such interventions has been long debated. Jurgensen (1983) and Truman (2003) argue that they are largely ineffective; Hutchison (2002) notes that they are effective in the short term, but not necessarily the long term. See Sarno and Taylor (2001) for a broader survey of the literature.


1,680 Reads