A partnership with academia

Berlin Working Papers on Money, Finance, Trade and Development

In cooperation with

Working Paper No. 01/2010

Risk factors in international financial crises: early
lessons from the 2008-2009 turmoil

Sebastian Dullien

July 2010

1

Risk factors in international financial crises: early lessons
from the 2008-2009 turmoil

By Sebastian Dullien
1

Abstract

This paper analyses the global transmission of the recent economic and financial
crisis as a function of macroeconomic factors such as per capita gross domestic
product, current-account positions prior to the crisis, exchange-rate regimes,
inflation prior to the crisis and financial openness. It finds that large current-account
imbalances (both surpluses and deficits) were a risk factor in the current global
economic turmoil. It also finds that countries that use currency boards have suffered
much more from the crisis than countries with other exchange-rate regimes.
Financial openness appears to have increased the risk of experiencing a deep
recession, while higher inflation prior to the crisis seems to have mitigated its
impact.

Keywords: crisis of 2008/9, exchange rate regimes, emerging markets, developing
countries

JEL classifications: E63, F31, F32

Introduction

There is a growing body of literature on the various impacts of the economic and financial

crisis on countries around the world. Much has been written on its impacts on world trade, on

commodity producing countries, on countries which have close trade linkages with the United

States, and on countries which rely heavily on remittance flows from developed countries.2

This paper aims to shed light on the spreading financial turmoil from a different angle: it

attempts to examine the international transmission of the subprime crisis in the United States

to determine which macroeconomic characteristics, beyond sectoral specialization and trade

specialization, make countries more vulnerable to the contagion effects of a global financial

and economic crisis. It looks at economic aspects which can be influenced by policymakers,

such as the exchange-rate regime, inflation, the current-account balance and capital-account

openness. In so doing, it adds to the debate on the choice of exchange-rate regimes, on

macroeconomic management, including under- or overvaluation of a currency, and on capital

account convertibility.

1 HTW Berlin – University of Applied Sciences, E-Mail: sebastian.dullien@htw-berlin.de
2 For a recent overview of a number of these issues, see Ocampo et al., 2010.

2

The paper is structured as follows. After a brief discussion on measuring the impact of the

crisis on individual countries, it provides a quantitative description of the most important

stylized facts of the global spread of the crisis, building on economic data for 181 countries

covered by the World Economic Outlook of the International Monetary Fund (IMF).3 It then

uses econometric techniques to determine which macroeconomic features helped some

countries to be more resilient to the financial and economic crisis than others. This section

also looks at the factors that might have played a role in determining whether a country

should turn to the IMF to cover financing needs in the recent crisis. The final section seeks to

offer tentative explanations for the empirical observations. Other contributions in this book

dwell on the wider implications of the findings, though these will also require further research

as more data become available.

1. Empirical analysis of the crisis

For determining the negative impact of the crisis, the following three criteria have been used

throughout the paper:

1. The change of trend in the GDP growth rate from the average of the years prior to the

crisis (2003–2007) to the average of the crisis years 2008–2009. This measure has

been chosen because the crisis hit different countries at different points in time. World

trade was already severely affected in the last quarter of 2008, and some countries

already had trouble financing their foreign deficit that year. However, due to the base

effect, this drop is partly reflected in the annual GDP growth rate in 2008 and partly in

2009. Looking only at the growth rate of one of these two years would have distorted

the picture.

2. The simple average growth rate of GDP for the years 2008 and 2009. Again, looking

at both years together gives a better picture than looking only at 2009 when most of

the decline occurred.

3. The fact that a country had to turn to the IMF for borrowing. Especially after the huge

wave of criticism of the IMF’s policies during the East Asian crisis of 1997–1998,

borrowing from the IMF has come to be seen not only as a national humiliation, but

also, increasingly, as an economic evil best avoided. Thus, being forced to accept IMF

lending can be viewed as a sign that a country has been severely affected by a crisis.

3 For this study, Zimbabwe has been excluded from the data set as it is an outlier for a number of the data points
considered, and the country’s recession is by most accounts largely independent of the global crisis.

3

Of course, there are other important negative economic and social consequences of the crisis,

such as rising unemployment and poverty, and increasing government debt. However, limited

availability of up-to-date data on these aspects constrains the analysis here. Unemployment

data are often not comparable between countries, and recording of unemployment figures,

especially for developing countries and emerging-market economies, are often inexact, as

employment in the informal sector is not always well covered. Moreover, the impact of the

crisis on the labour market may exhibit different time lags in different countries. In some

countries, retrenchment of workers is an easy and quick process, while in others it takes much

longer due to the legal regime or conventions. In addition, some countries have passed

measures temporarily stabilizing labour markets. Thus data currently available on labour

market performance are not an adequate indicator for measuring the impact of the crisis at this

particular point in time; its full impact can only be evaluated later.

Reporting of government debt and government budget deficits outside the OECD countries is

also not very exact and up-to-date, and the IMF’s World Economic Outlook therefore provides

such data for only a limited number of countries. Similarly, due to the lack of reliable, up-to-

date statistics for the incidence of poverty across countries, it is difficult to assess to what

extent poverty has increased as a result of the crisis. While there have been a number of

estimates (i.e. Chen and Ravallion, 2009), these are necessarily only very rough. These

indicators have therefore been omitted from this paper; instead, the paper focuses on the drop

in GDP and the extent of IMF involvement.

The analysis in this paper is based on data assembled from various sources. Data on GDP,

inflation and current accounts have been taken from the IMF’s World Economic Outlook

database (January 2010). Data on capital-account openness have been derived from Chinn and

Ito (2008). And data on exchange-rate regimes have been taken from the IMF’s classification

of exchange-rate regimes (IMF, 2009) and modified to include an additional group of

countries in the European Monetary Union (EMU).4 Altogether, the sample comprises 179

countries.

1.1 Descriptive statistics

Before we turn to a rigorous econometric analysis, it is useful to take a brief look at the data.

At the beginning of the crisis, it was often argued by the IMF and financial sector analysts

4 The IMF classifies EMU countries as “independently floating”. While this might be an appropriate description
of EMU as a whole, it is misleading when looking at the performance of individual member countries such as
Greece, as that country has a fixed exchange rate with its main trading partners.

4

that the emerging-market economies and developing countries might be decoupled from

developed economies, particularly the United States, and may therefore be able to cope with

the turmoil more effectively. While this hope proved to be illusory, at least some emerging-

market economies have performed much better than other parts of the world. Asian countries,

in particular, have managed to recover very quickly and briskly from the crisis, with parts of

Latin America following. In contrast, economic data for most of the members of the

Commonwealth of Independent States (CIS) and the new member States of the European

Union (EU) have shown few real improvements. Also the United States and the Western

European industrialized economies have proved to be laggards, with vulnerable economic

recovery (IMF, 2010).

Beyond these regional features, however, the impact of the crisis has clearly varied with the

state of development of the economies in question.5 On examining the different categories of

countries, namely low-income countries (GDP per capita below $975), lower middle-income

countries (GDP per capita between $976 and $3,855), upper middle-income countries (GDP

between $3,858 and $11,905) and high-income countries, we found fairly large variations in

the fall in the growth between the years 2003–2007 and 2008–2009: high-income countries

experienced a drop in the growth rate of 5.2 percentage points,6 upper middle- income

countries saw an almost equally large drop of 4.9 percentage points, while lower middle-

income countries saw growth decline by 2.7 percentage points and lower income countries by

only 1.2 percentage points. The group of high-income countries was the only category which

recorded an average annual negative growth rate for the years 2008 and 2009 of minus 0.7

per cent. This group therefore was solely responsible for the contraction of world GDP in

2009.

The crisis has also seen a resurgence of borrowing from the IMF. After years of not being

able to find borrowers, the IMF has started to lend again, supported by a pledge by its

shareholders to provide more funding as part of internationally coordinated crisis-fighting

efforts. Net disbursements by the Fund have been higher than at any time since the mid-

1980s, with net payouts totalling more than 20 billion in Special Drawing Rights (SDRs)

(about US$ 30 billion) in 2009 (figure 1). Also, the number of countries borrowing from the

IMF has risen sharply: out of 179 countries in our sample, 53 received IMF funding in 2009 –

a share of almost 30 per cent.

5 For descriptive statistics on the impact of the crisis on different economies, see the table in the annex.
6 All data for each country group refer to simple, unweighted averages for the country group in question.

5

Figure 1: Net IMF loan disbursements

-30 000

-20 000

-10 000

0

10 000

20 000

30 000

1
9

8
4

1
9

8
5

1
9

8
6

1
9

8
7

1
9

8
8

1
9

8
9

1
9

9
0

1
9

9
1

1
9

9
2

1
9

9
3

1
9

9
4

1
9

9
5

1
9

9
6

1
9

9
7

1
9

9
8

1
9

9
9

2
0

0
0

2
0

0
1

2
0

0
2

2
0

0
3

2
0

0
4

2
0

0
5

2
0

0
6

2
0

0
7

2
0

0
8

2
0

0
9

S
D

R

m
il

li
o

n

Source: author's calculations, based on data from IMF website

The impact of the crisis has clearly varied with the size of the external imbalances of

individual countries. Dividing the sample into four country groups according to their current-

account positions prior to the crisis (those with a high current-account surplus of more than 5

per cent of GDP, those with a current-account surpluses of less than 5 per cent of GDP, those

with a current-account deficit of more than 5 per cent of GDP and those with a current-

account deficit of less than 5 per cent of GDP), it can be observed that countries with large-

current account imbalances – surpluses or deficits – have been hit harder than those with

moderate imbalances. The group with very high surpluses experienced a drop in the growth

trend by 4.2 percentage points, followed by an only slightly smaller drop in the growth trend

of 3.9 percentage points for the group with very high deficits. In contrast, countries with

6

moderate deficits and those with moderate surpluses experienced a decline of only 2.2

percentage points and 3.1 percentage points respectively (figure 2).

Figure 2: Change in GDP growth between 2003–2007 and 2008–2009 by current-

account position of countries (Percentage points)

The exchange-rate regime also seems to have an impact on the vulnerability of a country to

the contagion effects of a crisis. After the Asian crisis in the 1990s, the notion of the stable

corner solutions (“corner solution paradigm”) came into vogue. According to this proposition,

in the long run only two currency regimes would be stable: the completely fixed or the

completely flexible exchange rate.7 Proponents of this hypothesis understood by “completely

fixed” any regime which was then seen as providing an irrevocably fixed exchange rate,

thereby providing no room for speculation. In addition to dollarization, currency boards and

monetary union were also seen as belonging to this category of exchange-rate regimes,

7 Early proponents include Eichengreen, 1994, and Obstfeld and Rogoff, 1995.

7

because, in principle, under these regimes the authorities have the necessary means in the

form of reserves to prevent any crack in the exchange-rate peg.8

In order to get an idea of the initial impact of the exchange-rate regime on the vulnerability of

countries, the sample was divided into nine groups, using the IMF’s classification of

exchange-rate regimes plus a separate group for countries in the EMU.9 Again, the results are

quite revealing. The (small) group of dollarized economies, including countries such as

Ecuador, Montenegro and Panama,10 managed the crisis relatively well: their GDP growth fell

by only 0.6 percentage points, and growth continued at an average rate of 3.7 per cent in

2008–2009 – above average in the overall sample. None of these countries had to seek IMF

support. However, before taking this result as a strong endorsement of dollarization, it must

be borne in mind that the countries which lacked a legal tender of their own had been growing

less rapidly in the years prior to the crisis than other countries of similar income levels (see

annex table). In addition, abandoning the national currency deprives policy makers of the

possibility of domestic financing of investment, as noted in Dullien (2009). Countries having

the other types of exchange-rate regimes originally considered as “completely fixed” have

performed comparatively badly during the crisis. The group of currency board countries,

including Bulgaria and Estonia, but also some smaller Caribbean countries, have been the

worst affected. GDP growth there declined, on average, by a whopping 6 percentage points.

In addition, these countries experienced a contraction in average annual GDP of 1 per cent in

2008 and 2009.

Interestingly, the exchange-rate regimes that, on average, produced the best outcome during

the crisis are those in the “middle ground” which were once seen as not sustainable. Countries

which had exchange-rate regimes classified as “conventional fixed peg” (except currency

boards, monetary union and dollarization), “pegged exchange rate within horizontal bands”,

“crawling pegs” or “crawling bands” saw their GDP growth rates decline by an average of

only 3 percentage points, and they achieved an average annual GDP growth rate of 3 per cent

in 2008 and 2009, while those with exchange-rate regimes closer to the “corners” saw their

8 Of course, the Argentine crisis of 2001-2002, which resulted in its exit from a currency board, showed that
such a regime is certainly not an “irrevocably fixed” exchange-rate regime.
9 The IMF classifies EMU countries as “independently floating”. While this might be an appropriate description
of EMU as a whole, it is certainly misleading when looking at the performance of a single member country such
as Greece as that country has a fixed exchange rate with its main trading partners.
10 Countries are counted as “dollarized” if they have adopted a foreign currency. Thus, Montenegro is considered
as having a “dollarized” economy even though it uses the euro.

8

GDP growth rate decline by 3.8 percentage points and recorded an average annual GDP

growth rate of only 1.2 per cent.

Table 1: Impact of the crisis by different exchange-rate regimes

1.2 Econometric estimates

Descriptive statistics like those above can be misleading. For example, currency board

countries as a group also usually have high current-account deficits. The question is therefore

whether the factors analysed above have a direct influence on their own, or only an indirect

influence. This can only be answered by means of rigorous econometric testing. Thus, as a

first step, a regression was run with the change in GDP growth between 2003–2007 and

2008–2009 as the dependent variable, and the current-account balance prior to the crisis

(2007), the inflation rate prior to the crisis (2007), GDP per capita, the variable for capital-

account openness, a dummy for an IMF programme in 2009 and dummies for the different

types of exchange-rate regimes as independent variables. In a general-to-specific-approach,

Change in GDP

growth, 2008–

2009

compared

with 2003–

2007

(Percentage

points)

Average

annual GDP

growth, 2008–

2009

(Per cent)

Average annual

GDP growth

2003–2007

(Per cent)

Dollarized economies -0.6 3.7 4.3

Currency board arrangements -6.0 -1.0 5.1

Free float -4.2 -0.1 4.1

Managed float -3.2 3.0 6.2

European Monetary Union -4.3 -1.3 3.0

Others ("middle ground")* -3.0 3.0 5.9

* Other conventional fixed peg arrangements, pegged exchange rate within horizontal

bands, crawling peg, crawling band.
Source: Author’s calculations, based on the IMF’s World Economic Outlook Database
(accessed in January 2010) and IMF (2009)

9

variables that were not significant, at least at a 10 per cent level, were eliminated. In addition,

both the current-account balance and the absolute value of the current-account balance were

alternatively included in order to allow for the possibility that large surpluses also make a

country vulnerable. The final equation for the change in the growth trend during the crisis

reads as follows:

,

where is the percentage point change in the average annual growth rate between

2003–2007 and 2008–2009, is GDP per capita in current US$ 1,000, and

is the absolute value of the current account in 2007 as a per cent of

GDP.

From this it can be observed that only per capita GDP levels and current-account imbalances

had a clearly negative influence on the way a country was affected by the crisis (both

coefficients are significant at the 5 per cent level), where the impact was measured as a

change in the trend growth rate. Countries with higher per capita incomes have been hit

significantly harder by the crisis than those with lower incomes. Interestingly, the current-

account balance as a per cent of GDP was insignificant in explaining the change in GDP

growth, while the absolute value of the current-account balance as a per cent of GDP turned

out to be highly significant. Hence, not only current-account deficits appear to have

contributed to the propagation of the crisis, but also current-account surpluses.

In a second step, a regression analysis was undertaken of the current-account balance prior to

the crisis (2007), the inflation rate prior to the crisis (2007), GDP growth rate prior to the

crisis (2003 to 2007), GDP per capita, the variable for capital-account openness, a dummy for

an IMF programme in 2009 and dummies for the different types of exchange-rate regimes as

possible factors influencing the average annual rate of GDP growth in 2008-2009. As before,

variables which turned out to be statistically insignificant were eliminated, and both the

current-account balance and the absolute value of the current-account balance were tested.

The resulting equation reads:

10

Where is the average annual growth rate of GDP in 2008 and 2009,

is the average annual growth rate of GDP during the period 2003–2007,

is the current account position as a per cent of GDP in the year 2007,

is the rate of inflation in 2007 and is a dummy for the country using a currency

board.

All variables were significant at the 5 per cent level, except inflation and the GDP growth rate

for the period 2003–2007 which were significant at 10 per cent.

A few of the results are notable. First, again GDP per capita turned out to be a very strong

predictor of lower growth in the crisis years, even when controlling for growth prior to the

crisis. One reason might be that the crisis originated in some of the most developed countries.

Second, the current-account deficit, not the absolute value, seems to be a significant variable.

A larger deficit prior to the crisis led to lower growth during the crisis years. Third, countries

with a currency board in place had a significantly lower growth rate in 2008–2009 (by an

annual two percentage points on average), even after controlling for the effects of the huge

current-account deficits some of the currency board countries such as Lithuania and Estonia

were running prior to the crisis. Third, inflation prior to the crisis seems to have influenced

the impact of the crisis, but not in the way that would be predicted by standard theory. In

actual fact, a higher rate of inflation prior to the crisis was correlated with a higher growth

rate during the crisis (even when controlling for GDP growth prior to the crisis).

Another interesting feature seems to be the lack of any correlation between the depth of the

crisis in a country and its request for IMF support. This result would mean first that countries

seem to have sought IMF support regardless of the scale of their economic downturn, and

second, that the IMF programmes do not appear to have significantly influenced the growth

outcomes of those countries compared with other countries having similar characteristics.

11

In a third step, a probit approach was used to test which characteristics increased the

probability of a country seeking IMF support. Again, all variables were initially included and

subsequently eliminated. In the end, the probit model for the probability of an IMF

programme was estimated (table 2).

Table 2: Probit model: Probability of IMF intervention

Variable Coefficient Standard error

Constant -0.575 0.151 (***)

-0.056 0.011(***)

-0.392 0.118 (***)

*** significant, at 1 per cent
level

Only two variables are significant for explaining the need for an IMF programme: the current-

account balance and the GDP per capita. The larger the current-account deficit prior to the

crisis, the larger was the probability of a country seeking IMF assistance in response to the

crisis. In fact, looking at the descriptive statistics, it can be seen that only 2 out of the 53

countries which borrowed from the IMF in 2009 had a current-account surplus prior to the

crisis. In addition, the richer a country in per capita terms, the less likely it was to seek IMF

intervention. This is an interesting result, as IMF intervention was considered most likely for

emerging-market economies. During the crisis, however, the Fund has lent strongly also to

lower income countries. None of the exchange-rate regime dummies proved to be

significant.11

11 However, some of the exchange-rate regime dummies showed a 100 per cent correlation with no IMF
programmes. For example, no dollarized country turned to the IMF in the latest crisis. However, interpreting this
fact in economic terms is not straightforward. While proponents of dollarization might claim that this shows the
greater stability of dollarized economies, it is just as plausible that dollarized economies lack the channels for
intervention through an IMF loan, or that the number of dollarized economies was too small (5 out of 179) to
enable a reliable conclusion to be drawn.

12

Finally, the group of worst performers during the crisis was selected and another probit

estimation run on the characteristics of this group. To this end, a threshold of an annual

contraction by more than 3 per cent for 2008–2009 was chosen (a total contraction of more

than 6 per cent), which produced 12 countries: Armenia, Botswana, Estonia, Hungary,

Iceland, Ireland, Italy, Japan, Latvia, Lithuania, Seychelles and Ukraine. The probit

estimation for these countries yielded the results presented in table 3, with KOpen referring to

capital-account openness as measured by the Chinn/Ito index.

Table 3: Probit model: Probability of a deep recession

Variable Coefficient Standard error

Constant -2.671 -2.671 (***)

-0.028 0.013(**)

0.154 0.857 (*)

0.092 0.055 (*)

KOpen 0.262 0.122 (**)

*** significant at 1 per cent level
** significant at 5 per cent level
* significant at 10 per cent level

Thus again, having a higher GDP per capita generally increases the risk of experiencing a

severe recession. A large current-account deficit prior to the crisis is also an important risk

factor. Having a relatively open capital account seems to be another risk factor for suffering

severe consequences of a global financial and economic crisis. Our regression analysis

revealed yet another factor: experiencing very strong growth in the years 2003–2007 (i.e. just

prior to the crisis) also seemed to have increased the risk of the crisis plunging a country into

a deep recession. This finding hints that a boom prior to the crisis might have led to

imbalances, which made the economy in question more vulnerable (as it might have been part

13

of a boom-and-bust cycle). Finally, having a very open capital account, as measured by the

Chinn/Ito index, significantly increased the risk of experiencing a very deep recession as a

consequence of the United States subprime crisis.

1.3 Summing up the empirical evidence

Thus, the findings may be summarized as follows:

1. In terms of impact on GDP and GDP growth, the crisis appears to have affected high-

and upper middle-income countries more than poorer countries, even though there

may have been greater suffering in lower income countries, as a drop in GDP growth

might be more severe in an environment without social safety nets and widespread

poverty as a result of the crisis.

2. Large current-account imbalances – not only deficits – seem to be an important risk

factor for vulnerability to crisis transmission.

3. Currency boards seem to be an additional risk factor, in addition to the impact a

currency board might have on the external balance by increasing the current-account

deficit.

4. An open capital account appears to exacerbate vulnerability.

5. Inflation, long seen as a prime concern for macroeconomic stability and an important

factor in increasing countries’ vulnerability to financial and currency crises, does not

seem to be as significant a factor as was previously thought.

6. Higher per capita incomes make IMF intervention less likely.

7. IMF programmes cannot be shown to have significant positive or negative effects on

the depth of a crisis

2. Tentative explanations and conclusions

From a theoretical point of view, and against the background of the Washington Consensus,

these results provide the basis for considerable rethinking. First, the benefits of free global

capital flows are very difficult to detect in this data set. Economic textbook theory tells us that

open capital accounts can do two things. First, they can help countries which lack capital to

import capital to grow faster. They can borrow from abroad, invest and hence boost growth.

14

As marginal productivity of capital is higher than in countries which are capital-abundant,

they can easily use the proceeds from their investments to service their debt. Second, open

capital accounts can help countries weather asymmetric shocks. If an unexpected shock

lowers national income, borrowing from abroad can be used to smooth national consumption,

thus increasing welfare.12 As long as domestic consumption has an influence on domestic

output, this should also help reduce the volatility of overall output. Countries which are more

financially open can more easily borrow from abroad, and therefore should be able to

withstand a crisis – such as the recent one – better.

However, the data presented in this paper do not confirm this story. Whether importing capital

is a sensible strategy for sustainably accelerating economic growth has been disputed for a

number of years (see, for example, Prasad, Rajan and Subramanian, 2007). The data set used

in this paper raises doubts about the ability of capital inflows to smooth the economic cycle.

While an open capital account per se does not seem to have a significant influence on the

depth of a crisis for the whole sample, it seems to increase the probability that a global

economic and financial crisis can push a country with such an account into a deep recession.

Moreover, using the possibility of global capital flows, either as an exporter or an importer of

large amounts of capital (as reflected in a large current-account imbalance) clearly and

strongly adds to a country’s vulnerability to a crisis. One plausible explanation would be that

in a financial crisis, such as the current one, access to foreign finance might not be possible

due to a sudden increase in risk aversion among investors, thereby hurting countries that have

relied on external capital inflows. The significant impact on countries with large surpluses

might be explained by the fact that the large surpluses possibly hint at macroeconomic

imbalances in these countries prior to the crisis in the form of permanently insufficient

domestic demand. With borrowers being cut off from the global financial markets during the

current crisis, countries that relied on other countries’ demand growth for their own economic

growth were hit disproportionally, due to the lack of internal demand growth momentum to

make up for the loss of external demand.

The probability of entering a very deep recession might increase in proportion to the openness

of the capital account. This is because capital controls are usually geared more towards short-

term capital flows, and hence a more open capital account means a larger share of volatile

short-term inflows in the overall capital inflows of a country. Given that the benefits of free

capital flows do not seem to materialize as promised to the countries which – at least in the

12 For a typical detailed explanation, see Feenstra andTaylor, 2008, chap. 17.

15

textbook model – should profit most from them (because they have made most use of

international capital flows), there might be a case for introducing controls and limits on global

capital flows.

Of course it may seem somewhat inappropriate to use the recent crisis as evidence against the

textbook argument of the cushioning effects of global capital flows. After all, the textbook

argument is in general about supply-side shocks to national output, while the origin of the

latest crisis has clearly been a financial one. However, given the magnitude of the crisis and

the fact that most of the economic crises of the past few decades arguably had financial

origins, one has to question the relevance of the argument in favour of insuring against

national supply shocks compared to potential shocks created by international capital flows for

an individual economy.

If one agrees with the necessity of proactive macroeconomic management to limit current-

account imbalances, and the need for bold policy action to counteract potential crises, the

other results are rather easy to explain: moderate rates of inflation (instead of low rates) are

not necessarily a problem, but might provide more space for monetary policy to implement

rate cuts before the zero bound limits further actions. Such a stance could be considered as

supporting the conclusions drawn by a recent IMF paper on the optimum rate of inflation

(Blanchard, Dell’Ariccia and Mauro, 2010). Currency boards are a danger as they create a

false sense of security and make proper macroeconomic management aimed at limiting

current-account imbalances virtually impossible.

More puzzling is the fact that IMF involvement does not seem to have any explanatory power

for the depth of a recession or a slowdown in growth. This result might be uncomfortable both

for the IMF itself as well as its critics. If it turns out to be robust, it would mean that IMF

involvement does not necessarily stabilize economic growth (as measured in GDP terms), nor

does the conditionality attached to IMF programmes exacerbate the short-term impact of a

crisis, as was repeatedly claimed for IMF programmes during the Asian crisis (Stiglitz, 2002).

It might also indicate that there has been a change in the way the IMF designs its adjustment

programmes so as to reduce their negative short-term impact on GDP growth compared with

the IMF programmes of previous decades, as some observers claim (Schieritz, 2010). Turning

this evidence against the IMF would imply that its programmes, while not exacerbating the

economic situation, have not contributed much towards economic stabilization in the latest

crisis.

16

From an economic policy perspective, this means that emerging-market economies and

developing countries should think twice about opening up their capital accounts. Should they

decide to open their capital accounts, countries should undertake active macroeconomic

management to prevent the emergence of large current-account imbalances, even if this comes

at the price of higher inflation. Finally, the results are a clear warning against creating a

currency-board framework. Far from providing a stable macroeconomic environment, as

some proponents have long argued, empirically such a framework seems to amplify shocks.

References

Blanchard O, Dell’Ariccia G and Mauro P (2010). Rethinking macroeconomic policy. IMF

Staff Position Note 10/03, Washington, DC.

Chen S and Ravallion M (2009). The impact of the global financial crisis on the world’s

poorest. VoxEU; available at: http://www.voxeu.org/index.php?q=node/3520.

Chinn MD and Ito H (2008). A new measure of financial openness. Journal of Comparative

Policy Analysis, 10(3): 307–320.

Dullien S (2009) Central banking, financial institutions and credit creation in developing

countries. UNCTAD Discussion Papers 193, United Nations Conference on Trade and

Development, Geneva.

Eichengreen B (1994). International monetary arrangements for the 21st century. Washington,

DC, Brookings Institution.

Feenstra RC and Taylor AM (2008). International Economics. New York: Worth Publishers.

IMF (2009). De facto classification of exchange rate regimes and monetary policy

frameworks (as of April 31, 2008); available at:

http://www.imf.org/external/np/mfd/er/2008/eng/0408.htm.

IMF (2010), World Economic Outlook, April, Washington, D.C.

Obstfeld M and Rogoff K (1995). The mirage of fixed exchange rates. Journal of Economic

Perspectives, 9 (4): 73–96.

Ocampo JA et al. (2010). The great recession and the developing world. Initative for Policy

Dialogue Working Paper; available at:

17

http://www0.gsb.columbia.edu/ipd/pub/Crisis_Complutense%5B1%5D_Great_Recessio

n.pdf.

Prasad ES, Rajan R and Subramanian A (2007). Foreign capital and economic growth.

Brookings Papers on Economic Activity, 38: 153–230.

Schieritz M (2010). Waschlappen aus Washington. DIE ZEIT Nr. 14/2010: 16.

Stiglitz J (2002). Globalization and its Discontents. New York: W.W. Norton & Co. Inc.

18

Annex

Table A.1: Impact of the crisis, by country categories

Number of
countries

Number of

countries

with IMF

lending in

2009

Change in

GDP

growth,

2008–2009

vs. 2003–

2007

(percentage

points)

Average

annual GDP

growth,

2008–2009

Average

current-

account

position,

2007

Absolute

average

current-

account

position in

2007

Average

inflation,

2007

Average

annual

growth,

2003–2007

Average

Chinn/Ito

index for

capital-

account

openness

Total sample 179 53 -3.5 2.0 -3.1 10.7 6.1 5.4 0.5

Per capita GDP less

than $976

44 26 -1.2 4.3 -7.0 8.8 8.4 5.4 -0.5

Per capita GDP

between $975 and $3

855

45 14 -2.7 3.3 -3.5 9.4 6.8 6.1 0.3

Per capita GDP

between $3 856 and

$11 905

42 10 -4.9 1.0 -5.0 13.0 6.3 5.9 0.2

Per capita GDP more

than $11 905

48 3 -5.2 -0.7 2.4 11.7 3.3 4.5 1.8

Current-account

surplus of less than

5% of GDP

21 0 -3.1 1.8 2.2 2.2 3.8 5.0 1.2

19

Current-account

surplus of more than

5% of GDP

40 2 -4.2 2.6 15.8 15.8 7.5 6.8 0.4

Current-account

deficit of less than 5%

of GDP

34 14 -2.2 2.6 -2.4 2.4 6.6 4.8 0.1

Current-account

deficit of more than

5% of GDP

84 37 -3.9 1.3 -13.8 13.8 5.9 5.2 0.5

Severe recession 12 6 -11.5 -5.4 -8.6 11.7 5.8 6.1 1.8

IMF lending in 2009 53 53 -2.9 2.2 -12.3 12.9 6.8 5.1 0.1

Dollarized economies 5 0 -0.6 3.7 -7.9 9.3 3.7 4.3 1.4

Economies with

currency boards

13 7 -6.0 -1.0 -18.2 27.9 3.9 5.1 0.8

Economies with

managed floating

currencies

43 17 -3.2 3.0 -5.0 9.8 8.3 6.2 0.6

Economies with free

floating currencies

23 4 -4.2 -0.1 -1.4 6,1 3.5 4.1 1.3

European Monetary

Union

15 0 -4.3 -1.3 -2.1 6.6 2.2 3.0 2.1

Economies with

"middle ground"

exchange rate

regimes

78 24 -3.0 3.0 -0.1 10.8 6.9 5.9 -0.2

Source: Author’s calculations, based on IMF (2009) and the IMF World Economic Outlook database (accessed January 2010).

Publisher: Competence Center “Money, Finance, Trade and Development “ 
HTW-Berlin – Treskowallee 8, 10318 Berlin
Prof. Dr. Sebastian Dullien, Prof. Dr. Jan Priewe

http://finance-and-trade.htw-berlin.de

ISSN:

© All rights reserved. Reproduction for educational and non- commercial purposes is
permitted provided that the source is acknowledged.