A. Triggers and Vulnerabilities

15. The triggers of systemic crises are similar to idiosyncratic crises—what differs is the scale and potential impact of the triggers. Also, there is no single feature that recurrently stands out as the ultimate trigger of systemic crises (Figure 6 summarizes the triggers of each systemic crisis). The triggers sometimes originated in large and more integrated economies (e.g., the recent global crisis), in line with the more traditional definition of “systemic” noted earlier. At other times, however, the triggers were associated with common shocks to a region or the globe (e.g., the debt crisis, the ERM crisis and the Asian/Russian crisis). In general, the triggers had the potential to hit a large number of countries or they acted as “eye openers” for investors to reassess risks for a whole asset class or region/group of countries. Specifically:

View Full Size

Ultimate Triggers of Systemic Crises¹

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: IMFstaff calculations.¹1/ Time in quarters, t is defined as the quarter when the average systemic crisis indicator goes above one standard deviation. Average systemic crisis Indicator is a simple average of affected countries indicators.

• Download Figure
• Download figure as PowerPoint slide

• Debt crisis: The main trigger was the sovereign default by Mexico in August 1982, following a buildup of balance sheet vulnerabilities caused by the U.S. dollar appreciation and interest rate increases in the U.S. after the oil crisis of the late 1970s. The whole region suffered a sudden stop.

• ERM crisis: A main trigger was the Danish rejection of the Maastricht Treaty criteria in June 1992, which in turn set off deeper and broader scrutiny of the feasibility of the common currency projects as well as a wider belief that the criteria were not consistent with broader fundamentals of the member countries. A speculative attack on those ERM currencies considered to be more vulnerable ensued.

• Asian crisis: The flotation of the Thai baht in July 1997 was the main trigger, which was also a wake-up call on the underlying vulnerabilities in other countries in the region. This, and the ensuing collapse in commodity prices, led to the Russian sovereign default and LTCM crisis, triggering a second wave of that systemic crisis.

• Global crisis: The alarms for the global crisis are well-known, with Lehman’s bankruptcy being the final trigger that globally froze wholesale and securitization markets.

Some common patterns emerge in terms of underlying vulnerabilities/factors, with exogenous factors playing a key role in many affected countries. Underlying factors are grouped into two, i.e., domestic policy factors and externally-driven factors (Table 1). Specifically, key domestic factors that created crisis vulnerabilities are as follows.

Table 1.

Factors Underlying Systemic Crises

Table 1.

Factors Underlying Systemic Crises

Underlying factor	Debt	ERM	Asia	Russia	Global
Domestic factors
Balance sheet mismatches	✔		✔	✔	✔
Public debt sustainability	✔	✔	✔	✔	✔
Unsustainable ER pegs		✔	✔	✔
Asset price bubble	✔	✔			✔
External factors
Monetary policy in major AMs	✔	✔
Commodity prices	✔			✔	✔

View Table

Table 1.

Factors Underlying Systemic Crises

Underlying factor	Debt	ERM	Asia	Russia	Global
Domestic factors
Balance sheet mismatches	✔		✔	✔	✔
Public debt sustainability	✔	✔	✔	✔	✔
Unsustainable ER pegs		✔	✔	✔
Asset price bubble	✔	✔			✔
External factors
Monetary policy in major AMs	✔	✔
Commodity prices	✔			✔	✔

• Financial vulnerabilities were a common theme. For instance, currency and maturity mismatches in balance sheets of crisis countries featured in all but the ERM crisis. Risky lending, contributing to asset price bubbles, created vulnerabilities in the debt, Asian, and global crises. Excessive optimism and complacency in boom times led to erosion of sound risk management practices in financial institutions, compounded by weaknesses in financial regulation and supervision.²¹ Weaknesses in financial oversight exacerbated excessive risk-taking and mispricing of risks, including through proliferation of opaque financial instrument prior to the global crisis.

• Debt sustainability issues were a standard feature across the crises, with external debt averaging as high as 75 percent of GDP among the crisis countries during the debt crisis. In the ERM crisis, average EU public debt was 73 percent of GDP. In the Asian crisis, declining reserve cover at the start of the crisis (to 50-70 percent of short-term external debt) was followed by jumps in external debt due to exchange rate depreciation and output collapse (above 150 percent of GDP in Indonesia in end-1998 and 90 percent in Russia in end-1999). At the onset of the global crisis, average external debt among crisis countries was 122 percent of GDP, with a few outliers displaying very high debt levels.

• Exchange rate policy management was another factor that ran through many crises, often resulting in speculative attacks on the currencies. Many of the crisis countries typically had a pegged exchange rate in conjunction with an inconsistent policy mix, including large fiscal deficits in a number of cases.

  16. Exogenous factors in systemic crises include domestic policy choices of large advanced economies and sharp commodity price changes. Specifically:

  Monetary policy in advanced markets: Policies in advanced economies, while consistent with their domestic policy mandates, may carry systemic implications for the rest of the world. For example, a prolonged period of loose monetary conditions in advanced economies could support a credit boom in their domestic economies and push capital outflows, which, when reversed by a tightening, could lead to sudden deleveraging and stop the capital inflows to other countries, potentially triggering a systemic crisis.²² Such an association is identified with monetary policy in the G-7 countries in three out of the four systemic crises (Figure 7).

  

  

  View Full Size

  

  G-7 Policy Rates Across Systemic Crises

  Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

  Source: IFS and IMF staff estimates.

  • Download Figure
  • Download figure as PowerPoint slide

  • i. The U.S. monetary policy tightening following the oil price shock of the 1970s added to the existing vulnerabilities in Mexico, given the currency and maturity structure of Mexico’s external debt. Higher interest rates resulted in unsustainable debt costs in Mexico and its eventual default, and the ensuing run on other Latin American countries.

  • ii. Likewise, following the reunification of Germany, fiscal spending accelerated, resulting in strong domestic demand and a buildup of inflationary pressures. The Bundesbank raised interest rates in response, resulting in the Deutsche mark (DM) appreciating and a run on other European currencies as many of them were overvalued relative to the DM.

  • iii. In the global crisis, accommodative monetary policy in advanced markets, combined with high saving in many Asian and oil surplus countries, was reflected in low global interest rates (the so-called “great moderation”), feeding into the buildup of systemic vulnerabilities before monetary policy was tightened)²³.

• Movements in commodity prices have also been an important external factor during some of the systemic crises. Sharp drops in commodity prices had severe implications for their exporters during the debt crisis, as well as for Russia in the late 1990s. They were less of a problem during the global crisis given the large reserve buffers many large emerging markets had going into the crisis. Likewise, increases in commodity prices have also resulted in monetary policy tightening in advanced markets to contain inflation, triggering systemic crises as noted above (e.g., debt crisis).

B. Crisis Impact

17. A key feature of systemic crises is the large number of countries with relatively strong fundamentals—or “crisis bystanders”—engulfed by the crisis (Figure 8).

View Full Size

Emerging Markets with Relatively Strong Fundamentals but Affected in Systemic Crises ^1/

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: IMF staff calculations.^1/1/ An emerging market (EM) is considered to have relatively strong fundamentals if it had low or medium pre-crisis external vulnerability (overall vulnerability for the global crisis). See footnote 24 for details on the construction of the vulnerability indicators.

• Download Figure
• Download figure as PowerPoint slide

• Systemic crises affect many countries together, including those with relatively strong fundamentals, as determined by their low or moderate external vulnerabilities. In the absence of a systemic shock, these countries are unlikely to suffer an idiosyncratic economic or financial crisis (Box 3 illustrates the vulnerability indicators for crisis bystanders).²⁴ The focus of the analysis of crisis bystanders is on emerging market countries, but this is purely for data reasons, and does not imply that advanced markets and low-income countries did not have such crisis bystanders.²⁵

  Box 3.

  Vulnerabilities of “crisis bystanders”

  Crisis bystanders are defined as countries with relatively strong fundamentals—i.e., low or medium pre-crisis external vulnerabilities—but nonetheless affected in systemic crises due to their international linkages and exposure to contagion. As noted in |18, the focus here is on emerging markets due to data limitations. Prior to each systemic event, the median crisis bystander had a smaller current account deficit and larger reserve cover (of projected external financing needs in the following year) than the highly vulnerable group, suggesting that the crisis bystanders had a lower likelihood to experience an idiosyncratic crisis. Fiscal indicators— excluded in the definition of crisis bystanders due to data limitations for earlier systemic crises— tell a similar story: crisis bystanders in general had better pre-crisis primary balance positions.

  

  

  View Full Size

  

  Pre-crisis Current Account Balance

  (in percent of GDP)

  Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

  Source: WEO and IMF staff calculations.

  • Download Figure
  • Download figure as PowerPoint slide

  

  

  View Full Size

  

  Pre-crisis Reserve Coverage ^1/

  Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

  ^1/In percent of short-term debt at remaining maturity and current account deficit for the next year.

  • Download Figure
  • Download figure as PowerPoint slide

  

  

  View Full Size

  

  Pre-crisis Primary Balance

  (In percent of GDP)

  Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

  Asian/Russian/LTCM crisis

  • Download Figure
  • Download figure as PowerPoint slide

  From a historical perspective, the strength of fundamentals for the group of crisis bystanders at the onset of the recent global crisis was the result of sustained efforts to improve their policies and policy frameworks following the sequel of financial crises faced by many of them during the 1980s and 1990s. As shown below, their inflation and balance sheets have been steadily improving for more than a decade. Such developments also make these countries more attractive to international investors (and more integrated with the global economy), increasing their exposure to systemic shocks.

  

  

  View Full Size

  

  Source: WEO and IMF staff calculations.1/ In percent of short-term debt at remaining maturity and current account deficit projected for the next year.

  • Download Figure
  • Download figure as PowerPoint slide

• The relatively strong fundamentals of these bystanders make them attractive to international investors who favor countries with fewer capital account restrictions and well-developed liquid markets. But, when global risk aversion rises, the need for investors to deleverage ends up drying up liquidity from these very same markets, without full regard to the relative strength of their fundamentals.

• The global crisis had a particularly large number of such crisis bystanders, hit by the synchronized deleveraging by a number of advanced economy banks.²⁶ The debt crisis, on the other hand, had the most number of highly vulnerable economies, given their deep rooted structural problems (noted in ¶ third bullet).

18. Despite their relatively strong fundamentals, crisis bystanders were not immune to the liquidity crunch and output loss experienced by countries at the crisis epicenter.

• Consistent with the above discussion, crisis bystanders suffered a large foreign exchange liquidity crunch arising from capital outflows (Figure 9). Net capital outflows of the bystanders was about 5% percent of GDP during the global crisis, only slightly lower than the 6% percent of GDP net outflows during the Asian crisis and 7 percent of GDP during the debt crisis.²⁷ The outflows during the global crisis occurred despite stronger initial conditions in many more countries than in past crises.

  

  

  View Full Size

  

  EM Affected Countries: Median Net Inflows

  (peak-to-trough deviation , period T-3 to T+3, T=crisis start, percent of GDP)

  Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

  Source: IMF staff calculations.

  • Download Figure
  • Download figure as PowerPoint slide

• Also, crisis bystanders were not exempt from large output losses, albeit smaller than that for the highly vulnerable countries (Figure 10). The loss in GDP growth (peak-to-trough change) among countries affected by a systemic event was largest during the global crisis. The median output loss was 4% percent for crisis bystanders and 8 percent for highly vulnerable countries, compared with less than 3 percent and 6 percent output loss during the Asian crisis for the two groups, respectively. The difference during the global crisis was the large decline in external demand conditions, resulting in a substantial adverse impact across countries.

View Full Size

EM Affected countries: Median Output Loss by Vulnerability

(peak-to-trough, percent)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

A. Domestic Policy Responses

21. Domestic responses in past systemic events have aimed at restoring market confidence by combining policy adjustment with liquidity injections to absorb external shocks. Policy adjustment—measured by the peak-to-trough correction in the current-account deficit—was achieved through fiscal policy and domestic liquidity provision, the latter measured by the relative changes in the country’s exchange rate, international reserves, and policy interest rate.

22. Policy adjustment was sizable in past systemic crises (Figure 11).

• Greater policy adjustment occurred in countries with larger vulnerabilities affected by the crisis: crisis bystanders’ adjustment averaged about 1.5 percent of GDP, while that for countries with high vulnerabilities averaged above 5 percent of GDP. The bystanders’ relatively stronger fundamentals created room for countercyclical policies (see pre-crisis primary balance in Box 3).

• Despite very weak external demand during the global crisis—import growth from G-7 economies declined by 14 percent in 2009 vis-à-vis an increase by 8V2 percent in 1998 during the Asian/Russian/LTCM crisis—adjustment needs have not been as wrenching as in past crises, thanks to less severe fiscal adjustment, less overshooting of the exchange rate, fewer banking sector problems, and lower inflation.²⁸

23. Domestic liquidity responses to soften adjustment during systemic crises have also varied depending on country-specific circumstances and external conditions.

• Decomposition of exchange market pressure indicators for crisis countries during systemic events highlights the variability of monetary and exchange rate policies across crises (Figure 12). This variability may be linked to domestic conditions, such as the exchange rate regime and the presence of currency and maturity mismatches in balance sheets, constraining countries’ ability to accommodate domestic and foreign exchange liquidity needs to avoid costly currency depreciation.²⁹ The differences in domestic policy responses may also reflect global liquidity conditions (as discussed in 1125).

  

  

  View Full Size

  

  Domestic Liquidity Responses ^1/

  (Median percentage change, 3M-moving average)

  Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

  Source: WEO database and IMF staff calcaulation.^1/A decrease in exchange rate is a depreciation; a decrease in reserves is decumulation.

  • Download Figure
  • Download figure as PowerPoint slide

• In particular, during the Asian crisis—in the absence of liquidity injections by major central banks—policy rates were hiked sharply after unsustainable pegs were abandoned to avoid disorderly currency depreciation, given very large currency mismatches. In the global crisis, policy rates—also aimed at avoiding balance sheet effects—were not hiked as aggressively upfront due to procyclicality risks in an already weak domestic and external environment. Accordingly, rates were generally reduced as soon as initial pressures receded.

• In the Asian crisis countries, foreign exchange intervention to defend the pegs had depleted the stock of reserves, severely constraining monetary policy and injection of domestic and foreign exchange liquidity. In the global crisis, the sizable stock buildup of reserves prior to the crisis enabled upfront use of reserves at the onset of the crisis, helping avoid excessive exchange rate overshooting in most cases and contributing to a more accommodative monetary policy. That said, only about half the emerging markets actually relied on significant depletion of their reserves during the global crisis, suggesting a “fear of losing reserves” rather than a “fear of floating” (Box 4).

Box 4.

Availability and Use of Reserves in Systemic Crises¹

Reserves have been an important factor in mitigating the impact of systemic crises, but with diminishing benefits. Moreover, self-insurance can be costly, and the reserves may not be readily used by countries for “fear of losing reserves”.

• Liquidity buffers have helped smooth consumption during systemic crises, enabling some countries to manage large financing pressures without experiencing a costly crisis. However, as noted in recent Board papers (How Did Emerging Markets Cope in the Crisis and Reserves Accumulation and International Monetary Stability), these benefits tend to diminish with the extent of reserve cover. Large reserve accumulation is also costly at the country level (the average cost of holding reserves for the median emerging market is estimated to be V percent of GDP over 2005-10), and a symptom of imbalances at the global level.

• Also, reserve cover—measured relative to a country’s short-term financing needs—has remarkably increased in the run-up to the recent global crisis, but actual use of reserves has been limited. Some countries demonstrated a reluctance to use reserves in the recent crisis, preferring to draw on foreign currency central bank swap lines, or use liquid assets in their sovereign wealth funds, putting into question the usability of large reserve holdings.

• Specifically, recent empirical work suggests that only about half of emerging markets relied on significant depletion of their international reserves as part of their adjustment mechanism (Aizenman and Sun, 2009). Only nine emerging markets in their sample experienced sizable reserve losses—at least 10 percent of their international reserves—during July 2008 to February 2009, and anyway limited their interventions during the crisis to less than one-quarter of their pre-crisis reserve holdings. In general, countries more exposed to financial than trade shocks refrained from using reserves during the global crisis, exhibiting a “fear of losing reserves” rather than “fear of floating”. This has been associated with the signaling value of reserve holdings and the risk that dwindling reserves may signal greater vulnerability, triggering a run on its remaining reserves, especially in the context of high uncertainty on the duration and extent of the crisis.

View Full Size

Pre-crisis Reserve Coverage ^1/

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

^1/In percent of short-term debt at remaining maturity and current account deficit for the next year.Source: Assessing Reserve Adequacy, IMF (2011)

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Median Cost of Holding Reserves, 2005-10

(in percent of GDP)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: Assessing Reserve Adequacy, IMF (2011)

• Download Figure
• Download figure as PowerPoint slide

¹This box builds on recent analytical work in Assessing Reserve Adequacy, IMF Policy Paper; Feb., 2011.

B. Global Liquidity Response Reserve-currency central banks’ liquidity response

24. The response of large reserve-currency central banks varied, depending on domestic conditions and impact of these crises on their countries. In systemic crises, where a large number of countries are affected and their central banks may be constrained in their ability to inject liquidity due to large depreciation fears and/or balance sheet risks, one option is for major advanced market (e.g., reserve-currency) central banks to help reduce foreign exchange liquidity pressures. However, central banks are driven by their domestic mandates, which may not always align with global needs. Indeed, analysis of global liquidity (average change in GDP-weighted narrow money for the U.S., Euro Area, Japan, and the UK), shows that the responses tended to be uneven (Figure 13):³⁰

View Full Size

Annual Change in Global Liquidity 1/

(In percent of GDP)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Global liquidity is calculated as average quarterly narrow money (converted in US$ and divided by GDP) in the U.S., Euro Area, Japan, and the U.K. Figures for the Euro Area refer to M1.Sources: IFS and staff estimates.

• Download Figure
• Download figure as PowerPoint slide

• During the 1980s debt crisis, global monetary conditions were tightened in response to inflationary pressures in advanced markets following the oil crisis and despite an ongoing, yet somewhat milder, recession in the U.S. and other G7 countries. Thus, monetary responses remained geared to a domestic reaction function rather than global stability considerations.

• In the Asian crisis, liquidity injections from major central banks were rather small, as most advanced economies initially remained strong and even benefited from investors’ “flight to safety.”³¹

25. The scale of the recent global crisis, however, resulted in an unprecedented coordinated response with a crucial role played by the U.S. Federal Reserve (Fed).³² To avoid a global breakdown of dollar funding, the Fed authorized dollar liquidity swap lines with a number of central banks around the world (Box 5). At its peak, following Lehman’s bankruptcy, the amount drawn reached almost $600 billion, accounting for over 25 percent of the Fed’s total assets (Figure 10).³³ This was paralleled by €250 billion swap lines by the European Central Bank (ECB), of which about €200 billion was with the U.S. Fed.³⁴ In October 2008, the Fed also authorized swap lines for $30 billion each to four emerging markets—Brazil, Korea, Mexico, and Singapore—given their large dollar funding shortages and the sizable exposure of U.S. banks to these countries (Aizenman, Yothin, and Park, 2010, and Aizenman and Pasricha, 2010).

26. Market reaction to the Fed swap lines during the global crisis was instantaneous, albeit somewhat short-lived for emerging markets (Figure 14). The swap lines were successful in smoothing disruptions in overseas dollar funding markets, with their announcements and operations associated with improved market conditions across the board (Fleming and Klagge, 2010). The VIX declined sharply after the swap lines’ outstanding amounts reached their peak, although there was no stark decline in the CDS spreads for the advanced economies with capped swap lines as counterparty credit risks lingered (Coffey et al, 2009). CDS spreads for the emerging markets that received the $30 billion swap lines declined only temporarily—through January 2009—in line with the overall decline in risk aversion—but then started to rise again given ongoing market pressures, notably in Europe. Overall, emerging markets also saw only a brief decline in the EMBI-Global spread after the Fed swaps were announced. Exchange rate performance largely mimicked that of bond spreads (Aizenman and Pasricha, 2010). A long-lasting and universal decline in spreads occurred only in April 2009 after the announcement that the U.S. Fed would expand its asset purchases by more than a $1 trillion (March 18, 2009) and the G20 London summit announced a coordinated policy response to the global crisis, including a substantial increase in Fund resources and the overhaul of its lending toolkit (see also ¶29 for additional analysis on spreads).

View Full Size

Federal Reserve Dollar Liquidity Swaps and Market Reaction

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: Bloomberg and Datastream.

• Download Figure
• Download figure as PowerPoint slide

IMF financing

27. Enhanced lending modalities in recent Fund arrangements have helped restore market confidence faster than in past crises.³⁵ Fund financing during the global crisis was larger and more frontloaded, with precautionary arrangements approved at higher access levels (Figure 15). This large and upfront liquidity helped improve market conditions in less than a year, with stabilization of the spreads and net capital inflows to many emerging markets by spring 2009 (Figure 16). By contrast, smaller availability of liquidity at the onset of the Asian crisis contributed to market confidence being restored only after almost two years—after the subsequent Russian and LTCM crisis had also receded.

View Full Size

Frontloading and Access

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: IMF staff calculation.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Market Interest and Spreads

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: Bloomberg, IFS and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

28. Moreover, countries with Fund arrangements weathered the global crisis better than past crises, and in line with the strength of their policies and fundamentals. For countries with Fund arrangements, the current account adjustment during the global crisis was smaller, despite substantially worse cyclical conditions than in past systemic events (Figure 17). The adjustment need was particularly small for crisis bystanders, notably those with FCL arrangements, even though some of them were exposed to significant spillovers from the crisis epicenter. Policy adjustment was well tailored to countries’ macro-critical vulnerabilities by streamlined conditionality and, for the FCL cases, by the signaling value of ex-ante qualification. Indeed, policy responses, such as liquidity support coupled with Fund arrangements, are found to be particularly important determinants of spreads reduction during the recovery phase of the global crisis, based on a pooled regression of affected countries, even after controlling for initial external vulnerabilities (Table 2). As a result, during the recent systemic event, crisis bystanders with Fund arrangements experienced substantially lower output losses than in the past crises, and those with precautionary arrangements outperformed global developments in real GDP growth

Table 2.

Determinants of the change in spreads from peak to trough during the global crisis

(in basis points)

Source: VEE Fall 2008; WEO; Bloomberg; Fund staff calculations. Notes: Robust standard errors in parentheses;

^****** p<0.01,

^**** p<0.05,

^** p<0.1

Table 2.

Determinants of the change in spreads from peak to trough during the global crisis

(in basis points)

Change in cyclically-adjusted primary balance (in percent of GDP)	115.26**	113.27*
Change in interest rate	94.92**	93.34*
Liquidity support: net official financing (in percent of GDP)	29.07*	31.28*
Interaction term: liquidity support*IMF program	106.72	104.74
Use of reserves (in percent of GDP)	-113.99	-116.58
External vulnerability indicator, Fall 2008		103.5
Constant	1017.3***	948.69***
R-squared	0.29	0.30
Obs.	28	27

Source: VEE Fall 2008; WEO; Bloomberg; Fund staff calculations. Notes: Robust standard errors in parentheses;

^****** p<0.01,

^**** p<0.05,

^** p<0.1

View Table

Table 2.

Determinants of the change in spreads from peak to trough during the global crisis

(in basis points)

Change in cyclically-adjusted primary balance (in percent of GDP)	115.26**	113.27*
Change in interest rate	94.92**	93.34*
Liquidity support: net official financing (in percent of GDP)	29.07*	31.28*
Interaction term: liquidity support*IMF program	106.72	104.74
Use of reserves (in percent of GDP)	-113.99	-116.58
External vulnerability indicator, Fall 2008		103.5
Constant	1017.3***	948.69***
R-squared	0.29	0.30
Obs.	28	27

Source: VEE Fall 2008; WEO; Bloomberg; Fund staff calculations. Notes: Robust standard errors in parentheses;

^****** p<0.01,

^**** p<0.05,

^** p<0.1

View Full Size

Peak-to-trough CA Deficit Adjustment by Crisis Episode

(Fund arrangements, average, in percent of GDP)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: WEO and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

(Figure 18). Even among highly vulnerable countries with programs, output losses were more contained after controlling for differences in global conditions.

View Full Size

EMs with Fund Arrangement: Median Output Loss by Vulnerability

(Peak-to-trough deviation from global growth, percent)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: WEO and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

29. Still, there was further scope for the Fund to extend financing support to more crisis bystanders. While countries with Fund arrangements were able to weather the global crisis better, benefiting from the enhanced lending framework, the IMF was able to ringfence only a limited number of strong performing affected countries (Figure 19). Specifically:

View Full Size

Fund Arrangements to Emerging Market “Crisis Bystanders”

(In percent of total crisis bystanders)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: WEO and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

• Only about 40 percent of crisis bystanders requested financing support from the Fund, compared with a 60 percent average in previous systemic events. This lower demand may reflect (i) reserve accumulation by many countries prior to the crisis, or (ii) the monetary policy coordination and liquidity swaps noted earlier. But, reserves may be costly to hold and the availability of swap lines in the future remains uncertain. Also, while the Fund’s lending reforms were helpful, they were lagged relative to when the crisis broke out.

Liquidity support from other official creditors and the private sector

30. Official lending from bilateral and regional creditors has played a critical role complementing Fund financing in past systemic crises. Support from bilateral and other official creditors has allowed effective burden-sharing of global liquidity needs with the Fund. The integral role of regional arrangements, brought to the forefront by the European financing response in the global crisis, highlights the benefits from joint programs between the Fund and RFAs in terms of resources, regional expertise, and program ownership (Figure 20).³⁶ Given the rapidity with which liquidity pressures were faced by many member countries, regardless of fundamentals, further consideration is now being devoted to regional financing on a precautionary basis. These efforts can leverage, on a case-by-case basis, potential synergies with the Fund to ensure effective coordination across the different layers of the GFSN. ³⁷

View Full Size

Burden Sharing (Average financing package in percent of GDP)

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: WEO and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

31. Finally, liquidity commitments by the private sector have secured additional support for crisis resolution during systemic events. Explicit private sector commitments to maintain their exposures can help in some crisis resolution cases support the credibility of a country’s official financing package, mitigate the shock to the real economy, and lower potential moral hazard. While the extent of committed liquidity support by the private sector has varied across crises depending on global and country-specific conditions, it has played an important role in most of them. The recent European Bank Coordination Initiative during the global crisis was effective in securing continued exposures by participating banks in Fund-supported program countries, notably Bosnia, Hungary, Romania, and Serbia (Figure 21).³⁸ Still, the ability to successfully ensure continued private sector liquidity commitments ultimately depends on lending opportunities, loan performance and banks’ overall liquidity situation, including in their parent country.

View Full Size

Rollover of Parent Banks’ Exposures in CESE Countries

Citation: Policy Papers 2011, 041; 10.5089/9781498338813.007.A001

Source: BIS.

• Download Figure
• Download figure as PowerPoint slide

A. Lessons

32. The following key lessons can be drawn regarding systemic crises and the policy responses to restore market confidence:

• Systemic crises are highly contagious and globally costly events. Trade and financial linkages among countries, combined with herding behavior by investors, contribute to rapid crisis propagation. Countries with relatively strong fundamentals were not immune to the costs of systemic crises, although these costs were smaller than for the highly vulnerable countries.

• Timely domestic policy responses were important steps to restore credibility and market confidence. Adjustment in countries with high vulnerabilities was typically more than in countries with low or modest vulnerabilities, the latter having more room for accommodative policies. The experience of the global crisis suggests that the buildup of reserves was useful in preventing exchange rate pressure from affecting consumption, though usability of these reserves is questionable given their diminishing returns, holding costs, and guarded use.

• The multilayered approach to liquidity provision helped mitigate the effects of systemic crises, but only in a limited way.

  • i. The role of global liquidity support by major reserve-currency central banks— effective to restore market liquidity and confidence—has been uneven during systemic events, but justifiably dependent on domestic monetary policy and financial stability considerations, in line with their mandate;

  • ii. During the recent global crisis, the Fed swap lines played a crucial role in addressing elevated pressures on global dollar funding markets. However, the counterpart countries were selective, and their availability remains uncertain going forward. Moreover, the lack of automaticity and restrictions on its use could limit their flexibility and availability during systemic crises;

  • iii. Use of Fund resources in large and frontloaded amounts, and increased recourse to precautionary financing generally helped restore market access more rapidly than in past crises.³⁹ Still, Fund resources were tapped only with a lag and were not fully utilized to ring-fence countries with relatively strong fundamentals.

  • iv. Official sector burden sharing increased during the global crisis, with prominent roles played by bilateral creditors and RFAs. The EU, in particular played a pivotal role in supporting the liquidity needs of the region. However, the focus of such regional financing has largely been on crisis resolution, not prevention.

  • v. Liquidity commitments by the private sector remain a key component of the safety net for crisis resolution cases, further enhancing burden sharing, mitigating crisis costs, and lowering moral hazard risks. As with the official sector, however, private sector involvement was mostly for crisis resolution.

B. Implications for the Global Financial Safety Net

33. The above analysis suggests possible areas of reform complementing recent efforts to strengthen the Fund’s financing role:

• i. Rising risks of systemic instability fuelled by growing trade and financial linkages and unpredictable investor behavior suggest the value of stronger surveillance over systemic risks and greater coordination in deploying crisis response policies; and

• ii. Even-handed and predictable availability of short-term liquidity targeted at countries with relatively strong fundamentals and policies that are hit by a systemic shock would help nip systemic crises in the bud, and may even preempt their occurrence by reducing investor uncertainty about the possible extent of contagion. Demand for existing Fund lending instruments and supply of central bank liquidity tend to lag during systemic crisis events.

• iii. In fact, while major central banks played a crucial role in the recent global crisis, such extraordinary actions cannot be taken for granted given the possibility that constraints from domestic mandates may interfere with decision-making for the global good. Moreover, while central banks may be best placed to alleviate interbank funding problems, they may not be well equipped to deal with funding pressures arising, for example, in sovereign bond markets.

• iv. As for the Fund, recent innovation in the lending toolkit has gone a long way in increasing its ability to respond in a flexible manner to various crisis circumstances. But, to be effective in a systemic crisis, actual short-term liquidity needs of all countries with relatively strong fundamentals and policies need to be met rapidly to cure the systemic shock (even better, to avoid that a large shock turns systemic). Such countries may require greater comfort and predictability on the availability of short-term liquidity to absorb temporary exogenous shocks.

34. These conclusions are corroborated by a growing academic literature. Many academics have argued for automatic provision of emergency liquidity assistance during financial instability, particularly in emerging markets (Annex III has a literature review).

Some argue that such assistance is needed in part because less-developed financial markets, weaker institutions, and lack of easily available information often make these emerging markets especially vulnerable to systemic risk (Mishkin, 2007, Fernández-Arias and Levy-Yeyati, 2010). Others reason that there is presently no mechanism to stabilize market conditions in a global systemic liquidity crisis, and they consider the Fund could play a key role alongside national central banks by providing liquidity (Obstfeld, 2010, Truman, 2010, and Cordella and Levy-Yeyati, 2010). However, the absence of an automatic trigger to access Fund lending, combined with its limited resources, may constrain such a role (Calvo, 2010 and Fernández-Arias and Levy-Yeyati, 2010).

35. There are others, however, who contend that excessive liquidity provision could create moral hazard. Indeed, contrary to the benefits to the liquidity injection in Mexico in 1994-95 described in Annex II, some have argued that such assistance provides incentives to investors to assume excessive risk that are counterproductive to global financial stability (Calomiris, 1998).

• A common moral hazard argument is that international “bailouts” insulate foreign creditors from losses during crises, resulting in excessive risk taking by private creditors. Domestically, the argument is that moral hazard problems could arise both in the official and private sector from the belief that loans from the lender of last resort would be forthcoming during a crisis.

• However, evidence suggests that moral hazard depends on the type of shock (exogenous or idiosyncratic) and whether there are instruments to mitigate the moral hazard behavior (such as conditionality).⁴⁰ Lane and Phillips (2000) and Jeanne and Zettelmeyer (2001) show that IMF resources or implicit subsidies in IMF lending are not large enough to create serious moral hazard.

• Moreover, effective prudential regulation and supervision of the financial sector can help lower moral hazard and reduce excessive private sector risk-taking (Fernández-Arias and Levy-Yeyati, 2010). Indeed, regulatory weaknesses and lack of a macroprudenti al policy framework led to failure to identify the build-up of systemic risk prior to the global crisis. Thus, efforts to strengthen financial regulation should go hand in hand with efforts to strengthen the international financial architecture (see below).

Where do we go from here?

36. First, minimizing potential systemic risks is the first line of defense to preserve global stability. Strengthening financial regulation—including to deal with systemically-important financial institutions via more effective supervision and credible cross-border resolution frameworks—is essential to prevent such crises, especially as more complex financial linkages across borders increase the risk of systemic crises.⁴¹ Likewise, strengthening policies and economic fundamentals through more effective bilateral and multilateral surveillance, in conjunction with sustained policy collaboration that places greater emphasis on policy spillovers and latent systemic risks, will also help lower crisis vulnerability and the ability to accommodate shocks without excessive adjustment and output loss. While these efforts would no doubt contribute to lowering the likelihood of systemic crises, history shows that the possibility of another systemic crisis cannot be ruled out. Thus, preparedness to respond effectively to the next systemic event is of paramount importance.

37. Enhancing the multilayered global financial safety net to create a mechanism to address threats to global stability would be another useful step. Specifically:

• RFAs are gaining prominence as a critical layer of the GFSN, and are already making efforts to enhance their own frameworks. European RFAs have led the way in providing large financing to crisis countries in Europe during the recent global crisis. They are also bolstering their surveillance framework. Concurrently, the ASEAN+3 members are strengthening their institutional frameworks for surveillance and crisis management, proactively developing operational guidelines under the Chiang Mai Initiative Multilateralization. In addition, there is also recognition that RFAs should expand their lending toolkit to incorporate financing for crisis prevention. While RFAs are an integral part of the GFSN, their lending during times of crises are limited to specific groups of countries which—while critical and useful at a regional level—may not be adequate to ensure even-handed and large coverage in a systemic crisis of truly regional or global dimension, given disparities in their resource availability and country coverage.

• Additional efforts are also needed in the multilateral layer of the GFSN, and, on balance, there may be scope for a larger role for the Fund to address the gaps. The Fund can deal with both idiosyncratic shocks and systemic shocks by pooling risks across its near universal membership and mobilizing large amounts of liquidity on a temporary basis. In particular:

  • If faced with a global liquidity crunch, the Fund can make liquidity available by allocating SDRs and extending bilateral financial assistance to allow policy adjustment while cushioning the impacts of the shocks. ⁴² Financial assistance can be used flexibly to meet balance of payments needs arising from currency runs, bank runs, or for budget support if the government faces financing pressures. Moreover, policy signalling from ex ante or ex post conditionality provides an international seal of approval, helping play a catalytic role to mobilize private financing.

  • As the analysis in this paper suggests, during systemic crises, even countries with relatively strong fundamentals could experience liquidity runs. Thus, more predictable provision of short-term liquidity to these countries could help not only mitigate the negative crisis impact, but also strengthen the Fund’s catalytic role by boosting market confidence and perhaps even preventing liquidity runs in the first place. These benefits can be achieved by enhancing flexibility in the current toolkit to provide evenhanded and predictable short-term liquidity support to crisis bystanders during systemic crises.

  • Any enhancement in the lending toolkit would have to be weighed against moral hazard concerns, as well as risks to the Fund. Mitigating such concerns and risks would require tailoring the form of assistance based on country fundamentals and the nature of financing shock. Countries with relatively strong fundamentals suffering an exogenously-driven liquidity shock could be eligible for short-term liquidity support based on qualification criteria (ex ante conditionality), while the financing needs of countries with weak fundamentals and longer-term financing needs would be better-suited for phased financing with ex post conditionality. Increasing the predictability of liquidity provision tailoring ex ante or ex post conditionality to countries circumstances would allow the Fund to make progress in responding effectively to systemic shocks in a way that minimizes the broader risks inherent to any lender of last resort.