INTERNATIONAL BANK LENDING INTERNATIONAL BANK LENDING AND THE EAST ASIAN CRISIS 1 Jacques Cailloux...

March 2000

INTERNATIONAL BANK LENDING

AND THE EAST ASIAN CRISIS 1

Jacques Cailloux and Stephany Griffith-Jones

Institute of Development StudiesUniversity of Sussex

FalmerBRIGHTON

BN1 9REtel: 44 1273 606261fax: 44 1273 621202

e-mail: [email protected]@ids.ac.uk

1 We would like to thank Stephen Spratt for helping us in gathering the data.

2

"There is a structural fault in the nature of capital flows. Short-term debt flows, especiallybank finance, are highly volatile. Unless the problem is tackled the potential for future criseswill remain". Mervyn King (1999)

"There is growing agreement that an excessive build-up of short-term debt was a proximatecause of the recent crises, particularly in East Asia" Dani Rodrik and Andres Velasco (1999)

1 Introduction

As discussed in chapter XX, the proximate cause of the crises in East Asia was a sharp

reversal in capital flows, and the dominant category in the reversal of these flows were bank

credits, which represented 92 % of the outflows in the four most affected countries during

1997 and 1998 and 72% of the reversals between 1996 and 1998 (see table 2 in chapter XX)

As we also highlighted, a particular feature of bank lending to East Asia was that a very high

proportion of those loans was short-term. This implied a high accumulated stock of short-

term debt, (particularly bank debt), which made the East Asian countries particularly prone to

large reversals. We have discussed that the higher the stocks of liabilities the higher the scale

of potential outflows; this is particularly true for short-term bank debt, where the potential for

reversal is extremely high as creditors have the simple option of not renewing loans. As the

empirical analysis of Radelet and Sachs (1998) and particularly Rodrik and Velasco, op. cit.,

confirm, short-term debt (and especially short-term debt to banks) as a proportion of reserves

is a strong predictor both of likelihood and severity of crises.

There has been surprisingly little detailed empirical analysis of the nature and causes of

reversal of bank loans to the East Asian countries, for example by nationality of banks and by

individual banks (an exception is Kaminsky and Reinhart, 1999). Given the importance of

these flows and their reversal in East Asia, we develop such a detailed empirical analysis

here, which gives us a basis for more in-depth examination of the causes of these large

reversals.

In section 2, we present the broad picture of bank lending to East Asia, focusing on the main

players. Section 3 then, analyses the different reasons behind the rapid build-up of short term

external debt. Section 4, in turn, presents some explanations of the withdrawal of bank loans.

Section 5 draws policy implications.

3

2 Recent trends in bank lending to East Asia

In this section, we analyse the maturity structure and sectoral distribution of bank loans, and

then examine the behaviour by creditor banks' nationality. Our analysis focuses on the period

between end June 1995 and end June 1999 and draws on data published by the BIS.

2.1 Global trends to East Asia

As can be seen from table 1, international lending to East Asia (Indonesia, South Korea,

Malaysia and Thailand) rose sharply during the years just preceding the crisis. However, it is

worth underlining that the stock of external debt at the beginning of the 1990’s was fairly

small in all four countries (with the exception of Korea) but that it had been increasing very

rapidly throughout the period. For example, total lending to the region between December

1993 (first period for which BIS data is available on the four countries) and June 1997 more

than doubled (an increase of 133%)2. By the end of 1996, to give an order of magnitude, total

external debt amounted to roughly 25 percent of the region’s GDP, compared to a level of

13% of GDP in Latin America3.

From June 1997 onwards and as the crisis unfolded, total lending fell dramatically, reaching

at the end of June 1999 a level below June 1995. The largest fall was experienced during the

first semester of 1998 with a decline of almost 20 percent of total claims compared to a

decline of 8 percent during the previous semester. The reduction in exposure continued

throughout 1998 and the first half of 1999 with successive falls of 11 and 6 percent. Between

June 1997 and June 1999, the exposure of international banks to East Asia had fallen by 38

percent!

2 BIS (1993) and table 1.3 Only three Latin American countries, Mexico, Argentina and Brazil have been selected here. The percentagesare calculated as the proportion of total claims from BIS reporting banks to the countries’ GDP.

4

Table 1 : International claims by nationality of reporting banks on Asia-4(in millions of US dollars, and percentages of grand total)

GrandTotal Japan EU US

Localsubsidiaries

USD % USD % USD % %end June 99 160563 53395 33 66542 41 12450 8 18end Dec 98 171695 62387 36 70393 41 12044 7 16end June 98 192537 71989 37 76362 40 13541 7 16end Dec 97 238977 84027 35 88128 37 18750 8 20end June 97 261056 95123 36 92000 35 20929 8 20end Dec 96 247857 92094 37 83183 34 22020 9 20end June 96 226842 89817 40 70899 31 19462 9 21end Dec 95 201655 86594 43 59758 30 15988 8 19end June 95 180167 80105 44 57841 32 13595 8 16Note : The grand total in the first column of the table also includes the international claims of domesticallyowned banks in Switzerland, which are not shown separately in this table, as well as those of local subsidiariesand branches of banks which have their head- office outside the BIS reporting area. The last column is thus theshare of bank lending mainly from local subsidiaries and branches of banks which have their head- officeoutside the BIS reporting area.Source : BIS, The maturity and sectoral and nationality distribution of international bank lending , variousissues.

As far as source countries are concerned, although Japan is the main source of bank lending

to the region at the beginning of the period, its share actually declines fairly rapidly over time

in favour of European lenders (see Table 1). Indeed, while between June 1995 and June 1997

Japanese debt only increased by 18.6 percent, European debt rose by 61.4 percent. As a result

of the very rapid increase in European lending, at the onset of the crisis, Europe had caught

up with Japan in terms of exposure to the region, with $92 billion (compared to $95 billion

for Japan) by June 1997. During and after the crisis, Japanese claims contracted quicker than

those of European banks, so that by the end of June 1999, the share of Japanese claims on the

Asia-4 was significantly smaller than European claims (see again Table 1). Interestingly, the

share of US banks was both stable and relatively small all along the period.

We now study in turn each of the four most affected countries.

2.2 South Korea

As of December 1996, Korea had accumulated about $100 billion of debt with international

lenders, that is 40% of the four East Asian countries’ total claims. Korea was thus the main

recipient of international bank lending to the region although its claims only accounted for

20% of its GDP, the smallest share of the four countries.

5

Table 2 provides data on bank lending by nationality to Korea. In June 1997, Japan

represented 23% of total bank lending, the EU 36% (of which France and Germany

represented 10% each, see annex 1, table 1) and the US 10%.

As regards total declines since the crisis broke out, all the big creditor countries' banks see

very large declines of similar magnitudes. Thus, from their peak to June 1999, Japanese and

US banks' exposure fall by 37% and 36% respectively. German, French and UK banks fall

also significantly, though somewhat less, by 30, 19 and 23% respectively. This seems to

provide some evidence of herding between big players. It is interesting that some of the

smaller creditor countries banks claims fall much more with Belgium declining by 85% from

peak and Finland with an 88% fall from peak, whilst banks of two creditor nationalities

(Netherlands and Canada) behave in a somewhat counter-cyclical manner actually increasing

their lending during the crisis period or its immediate aftermath.

Finally, an important feature to highlight is that bank exposure to South Korea not only fell

sharply during the crisis period, but continued to do so till June 1999. This trend is true also

for the major creditor countries, except for the US and French banks, where lending

recovered somewhat from the end of December 1998.

Table 2 : International claims on South Korea by nationality of banks(In millions of US dollars and as a share of grand total)


Localsubsidiaries

USD % USD % USD % %end June 99 63482 15018 24 24925 39 6420 10 27end Dec 98 65293 16925 26 26194 40 6291 10 24end June 98 72444 18934 26 28118 39 7409 10 25end Dec 97 94206 20278 22 33676 36 9533 10 33end June 97 104148 23732 23 37043 36 9961 10 32end Dec 96 99953 24324 24 33819 34 9355 9 32end June 96 88027 22512 26 26859 31 9582 11 33end Dec 95 77528 21454 28 23611 30 7590 10 32end June 95 71430 20874 29 24505 34 7122 10 27Note : See table 1Source : See table 1

An important feature of the build-up and reversal of bank loans to Korea worth highlighting

is that approximately 60% of the increase of bank lending between June 1995 and the peak in

June 1997 and around 90% of the decline in bank lending between June 1997 and June 1999

is explained by lending of up to one year.

6

Similarly, around 60% of the increase in bank lending (between June 1995 and the peak of

June 1997) and around 80% of the outflow was accounted for by inter-bank lending (see

annex 1, table 2). Therefore we can clearly see that in South Korea, a very significant

proportion of the reversal of capital flows was explained by a very sharp reversal of short-

term inter-bank loans, which had previously increased very sharply.

Bank loans to the non-bank private sector increased and fell quite significantly in the same

period, but accounted for a clearly smaller proportion of the outflows. Curiously, bank

lending to the public sector was counter-cyclical, falling between December 1995 and

December 1997, and increasing till June 1999.

2.3 Thailand

Thailand is the second largest borrower of the region following Korea, with a share in total

claims of 28% (compared to 40% for Korea). At the end of December 1996, total claims on

Thailand amounted to about $70 billion or 37% of its GDP (which is a very high level).

Thailand is the only country among the Asia-4 to experience a reversal in total lending,

though of a small scale, as early as the first semester of 1997. This is mainly due to US and

UK banks which lowered their total claims by about 20 and 10% respectively during that

period (see annex 1, table 7).

As can be seen from table 3, in June 1997 European claims accounted for 29% of total claims

while the US and Japan's shares amounted to 6 and 54% respectively. Total claims on

Thailand fell by 50% between June 1997 and June 1999. The sharpest fall was experienced

by claims on US banks with a decline of 70% followed by Japan with 51%. European banks

also sharply reduced their exposure to Thailand and to a greater extent than in other countries.

Indeed, UK claims fell by 48% while French and German claims fell by 43 and 39%

respectively (see annex 1, table 7)

7

Table 3 : International claims on Thailand by nationality of banks(In millions of US dollars and as a share of grand total)

GrandTotal

Japan EU US Localsubsidiaries


As of June 1997, 66% of total lending was of maturity shorter than one year. About 76% of

the decline in total claims are explained by the reduction in short-term debt (annex 1, table 8).

As opposed to Korea, but in a similar way to Indonesia, foreign bank lending was mainly to

the corporate sector. It accounted, in June 1997, for 60% of the external debt. However, it is

the decline in interbank lending which explains most of the fall in total lending (about 55%)

(see again annex 1, table 8) in a similar way to Korea.

It is worth underlining that all major lenders were still reducing their exposure during the first

semester of 1999.

2.4 Malaysia

Malaysia experienced a peak in external indebtedness in June 1997. Total debt almost

doubled between June 1995 and June 1997 to reach about $29 billion. By the end of

December 1996, Malaysia’s claims only represented 9% of the region total external claims

and accounted for about 22% of its GDP, a share just above that of Korea. Total claims then

declined by 35% between June 1997 and June 1999.

8

Table 4 : International claims on Malaysia by nationality of banks

(In millions of US dollars and as a share of grand total) Grand

Total Japan EU USLocalsubsidiaries


As in the other countries, short-term debt played an important role in both the build-up and

reversal of foreign loans. Indeed, the increase in short term debt explains more than 60% of

the total debt growth and more than 80% of the fall. Both loans to local banks and the

corporate sector sharply increased during the boom period but it is largely inter bank lending

which fell dramatically (more than 60%).

As far as source countries are concerned, in the same fashion as with the other countries,

Europe accounted, in June 1997, for the bulk of the Malaysian external debt with a share of

44% while the US and Japan accounted for 8 and 36% (table 4).

2.5 Indonesia

In a similar way to the other three countries, Indonesia enjoyed a sharp rise in total claims

during the period preceding the crisis. Between June 1995 and June 1997, they increased by

45%. Most of the increase is explained by additional EU lending. As of December 1996, total

claims accounted for 25% of Indonesia’s GDP and 22% of the region’s liabilities.

At the onset of the crisis, the share of Japanese and European claims amounted to almost the

same amount (about 40% each). However, there are clear differences in the way they pulled

out. Indeed, the data in table 5 show that European bank lending exhibited a much more sable

trend with a fall of only 8% between June 1997 and June 1999 compared to a fall of 39% for

Japanese lending. Table 3 in annex 1 provides further information on the behaviour of

9

European banks. The large creditor countries (Germany, France and the UK) reduced their

exposure by about 20%, slightly less than their US counterparts (-24%).

Table 5 : International claims on Indonesia by nationality of banks(In millions of US dollars and as a share of grand total)


Localsubsidiaries


About half of the increase in total claims was due to short-term loans while 86% of the fall

was caused by a reduction in short term lending. As far as sectors are concerned, the vast

majority of debt was held by the private sector (about 70 percent). However, the reversal was

experienced both in the banking sector and the non-bank private sector which explained

respectively 53% and 45% of the total fall. Interestingly and as in the three other countries,

lending to the public sector was counter-cyclical.

As a conclusion, the analysis on individual countries that we have carried out above shows

that all four countries experienced quite similar trends both before and during the crisis.

Japan was the largest lender to the region in the mid-1990s but at a similar level to Europe

taken a whole. Since the mid 1990s, the share of Japanese exposure fell, both before, during

and after the crisis, while that of European banks grew. Within Europe, the largest lender

countries are France and Germany with similar levels of exposure and then the UK. The US

only represents around 10% or less of total claims.

Short term bank lending amounted for the majority of total claims in all countries. It then

largely explained the fluctuations in total lending. However, it played a more important role

in the reduction of exposure by foreign lenders than in the build up. Indeed, it accounted on

average for about 60% of the increase and 80% of the decline in bank lending to the region.

10

Similarities were also found in terms of the allocation of foreign debt at the sectoral level.

With the exception of South Korea where inter-bank lending was dominant, in all the three

other countries it is the non-bank private sector which was the largest borrower.

In the next section, we analyse the explanatory factors behind the very rapid build-up of

short-term debt.

3 Why did banks lend so much and so short term?

As we have seen in the previous section, all four Asian countries experienced a very rapid

increase in their external commercial debt, a large part of which was of a short-term nature.

There are a number of factors which have been widely discussed as being clearly important in

the build-up of external debt. As Sachs and Radelet (1998) rightly stress, the excellent growth

record and wide ranging financial deregulations all clearly contributed to strengthen banks'

willingness to lend. It is clear that these countries were seen and saw themselves as very

creditworthy and thus did not worry about borrowing short-term as they thought they could

roll-over the loans very easily (in the same way as Mexico before the 1994 crisis). In this

section, we focus our analysis on other, complementary, explanations. We analyse in turn the

role of moral hazard in its international dimension, the impact of the 1988 capital asset ratio

implemented by the Bank for International Settlements (BIS) and finally the role of Japanese

banks.

3.1 The moral hazard hypothesis

In some of the literature which emerged in the aftermath of the Asian crisis, the moral hazard

argument played a relatively important role in explaining the size and to some extent the

maturity of international lending to the most affected economies. For example, Corsetti et al

(1998) and Krugman (1998) argue that moral hazard, under the form of implicit guarantees

by governments, contributed to excess borrowing from abroad. This was seen as particularly

true for Thailand and Korea where most of financial institutions had guaranteed liabilities at

the beginning of the crisis (Krugman, 1998). According to this analysis, it is thus the

international dimension of moral hazard which had an impact on the behaviour of

international banks. Indeed, these institutions based their lending decision on the assumption

that local governments would guarantee those loans or if not, that IMF - like institutions

would bail out borrowers. Following this line of analysis, moral hazard then lead to

11

excessive risk taking with less monitoring of the projects quality, which itself lead to

financial collapse.

The argument of implicit guarantee has also been stressed by McKinnon and Pill (1996 and

1998) who argue that, even with a clear engagement from authorities – national or

international -not to bail out, the impact of financial crises can be so costly that these

institutions will ultimately have to bail out the financial system anyway.

The authors then argue that, in the context of a badly regulated financial system, i.e. where

the financial authorities are unable to monitor the moral hazard problem then, implicit

guarantees lead to overborrowing and increased instability. Interestingly, the authors argue

that good supervision and regulation of the financial system might not be sufficient as

exemplified by the financial crises that occurred in developed countries. The authors propose

that monetary authorities should also monitor the composition of bank credit expansion as

well as the aggregate lending levels to property, construction and personal consumption.

These levels should be limited in order to prevent speculative borrowing that have a

destabilising impact on the economy. As for the international dimension of moral hazard,

McKinnon and Pill (op. cit.) propose that "direct and indirect measures to restrain inflows of

foreign financial capital may be necessary" (McKinnon and Pill, 1998, p 352).

However, it is worth stressing that the moral hazard hypothesis has been challenged more

recently both from an empirical and theoretical point of view, including by Krugman himself.

At a theoretical level, Krugman (1999) argues that, if implicit guarantees were the main

causal factor of the East Asian crisis, one should have observed a decline in investments

which had no guarantee at all (such as equity investment) in favour of other forms of loans

which were assumed to have some form of guarantee either explicit or implicit. This pattern

clearly was not witnessed in the case of East Asia. Krugman (op. cit.) adds that if the moral

hazard argument had played an important role, it would have lead to substantial bad loans

earlier on in the most affected economies. However, it has been recognised that most of bad

loans were in fact a result of the sharp recessions and depreciations and not the other way

round.

On the empirical side, two recent studies find conflicting evidence. On the one hand, the IIF

(1999) analyses the impact of the Mexican bail out in 1995 on the large flows of bank lending

12

towards emerging markets which took place between 1996 and mid 1997 at low spreads. The

findings do not support the moral hazard hypothesis; on the contrary, declines in emerging

market spreads are mainly explained by the sharp increase in international liquidity

conditions.

Sarno and Taylor (1999), on the other hand do find evidence in favour of the moral hazard

hypothesis. The study finds evidence of asset price bubbles in all affected countries and a

high degree of reversibility in capital flows to East Asia. According to the authors, these two

findings are consistent and are necessary conditions for the moral hazard interpretation to

hold. However, this analysis is problematic. First, although moral hazard might lead to a

sharp increase in the demand for speculative and risky assets itself, leading to an asset price

bubble, the latter can have other causal factors. This is all the more true that the bubbles

which are tested are stock market bubbles. Indeed, as underlined above, the stock market

does not typically have any insurance guarantee. Second, Sarno and Taylor (1999) do not

provide a clear theoretical explanation of the relationship between a high degree of

reversibility and moral hazard. Furthermore, they find a high reversible component in

portfolio flows (consistent, according to the authors, with the moral hazard hypothesis) while

bank credits have a relatively higher permanent component. This is a rather disturbing result

as one would expect that bank credits and not portfolio flows should confirm the moral

hazard hypothesis.

3.2 A regulatory bias in the BIS capital accord ?

There is still no definite explanation behind the large share of short-term debt to total debt

accumulated by the Asia-4. One argument has focused on a potential distortion in the Basle

Capital Accord which would make short-term lending more profitable for creditor banks.

Indeed, inter-bank lending to non-OECD countries of a residual maturity of up to one year

has a weighting of only 20% for capital adequacy purposes, whilst loans over one year

require a 100% weighting4. Short-term lending to non-OECD countries thus only requires a

1.6% capital cover against 8% for medium to long-term lending. The underlying rationale of

those weightings is that the shorter the loan maturity, the higher the probability to recover it.

4 OECD interbank lending receives a 20% weighting irrespective to its maturity.

13

However, from a debtor country standpoint, these weightings could shift the overall maturity

structure of the country's external balance sheet towards short-term

It is worth underlining that the weighting structure could have an impact not only on the

maturity but also on the relative size of inter-bank lending compared to the corporate sector.

This issue was addressed by a working group established by the Basle Committee on banking

supervision, as part of a broader review of the current banking supervisory framework and

the lessons that could be drawn from the Asian crisis (See Basle Committee on Banking

Supervision, 1999). Four studies were conducted to try to assess these potential distortions

both in terms of volumes and maturity.

The first one analyses the changes in the overall lending maturity of countries following their

membership to the OECD. These countries are Mexico, South Korea, the Czech Republic

and Hungary (although Poland joined the OECD in 1996 it was not included in the sample as

it was still affected by a past rescheduling). In the case of Mexico and South Korea, there

was some evidence that their change in status from non-OECD to OECD members reduced

the proportion of short-term debt to total debt. However, the lack of observations and the fact

that Mexico and South Korea suffered a financial crisis just after joining the OECD

somewhat weakened the robustness of the results.

The second study assesses the impact of joining the OECD on the size of inter-bank lending.

A significant rise in the proportion of inter-bank lending to total lending was observed in

three countries out of four following their accession to the OECD. But this increase as

underlined by the study, could partly have also been a result of the liberalisation measures

undertaken under the OECD codes.

Thirdly, the maturity structure of OECD countries was compared to non-OECD countries that

have the same ratings. It was found that for relatively highly rated countries, the share of

short-term debt was higher in non-OECD countries than in OECD ones, confirming the

hypothesis of a short-term bias for non OECD countries. This relationship was not as clear in

the case of lower ratings.

Finally, the report assessed whether the lower weights applied to inter-bank lending lead to

more lending being channelled to banks rather than to the corporate sector (see table 8 on

14

how weights are applied to banks under the 1988 system). The study was conducted both

within the OECD where the bias was expected to be the strongest (because inter-bank lending

requires a small risk weight for short-term as well as for long-term lending) and to non-

OECD countries. No evidence was found that the capital accord significantly impacted upon

the lending composition.

Further work could compare the trends in total short-term lending to non-OECD countries

with the trends in lending to the banking sector and the corporate sector. This analysis could

reduce the data limitation due to the non availability of data on the maturity of inter-bank

lending. Indeed, a higher correlation between short-term lending and lending to the corporate

sector would confirm the hypothesis of a short-term bias.

Reisen (1999) adds that new OECD members benefited from cheaper borrowing through

lower interest rates due to the lowering of risk weights.

There is thus some evidence that the Basle Accord has encouraged more short-term lending

to non-OECD countries, though the evidence is somewhat constrained by data limitations

(see also, Cornford, 2000). It is worth underlining that the new Basle proposal has suggested

amending the current system, especially as a response to the criticisms that arose after the

Asian crisis. A discussion of the 1999 proposal is provided in section 5.

3.3 The role of Japanese Banks

As pointed out in the first section, Japan is the first lender to each Asia-4 country5. In June

1997, Japanese banks had a slightly higher exposure to the Asia-4 than total Europe did. Part

of the explanation behind the rapid growth of international credit to these four countries must

thus lie in the behaviour of Japanese banks. Interestingly, in 1997, the position of Japan in the

region was quite unique compared to the other countries. Indeed, while all other countries had

their largest exposures in South Korea then Thailand, Indonesia and Malaysia (with the

exception of the US where Indonesia was also second exposure) Japan’s largest exposure was

in Thailand ($ 16.1 bn), followed by Indonesia ($ 10.9 bn), South Korea ($ 9.5 bn) and

Malaysia ($ 4.6 bn).

5 This is not true if Europe is taken as a whole.

15

The international expansion of Japanese banks needs to be put within the broader context of

the so called "Main Bank" and its role within the industrial group with which it has cross

holdings. Indeed, as underlined by Peek and Rosengren (1998), the wave of Japanese FDI to

the US during the 1980s was associated with a sharp increase of Japanese lending, most of

which was initially made to Japanese affiliates. The main bank which provides most of the

financing to group members thus tends to behave similarly domestically and internationally.

However, Japanese banks later diversified their portfolio and expanded their loans to a wider

US customer base which were non affiliates. During the first half of the 1990s, Japanese FDI

shifted to Asia. Ito (1999) underlines that, between 1990 and 1996, the Asia-4 experienced

increases of Japanese FDI of a scale ranging from two to four fold. As mentioned earlier,

Japanese lending to the region also increased during that period. This is confirmed by Graph

1 which compares the cumulative inflows of FDI and bank loans to the Asian-4 between

1989 and 19966.

Graph 1 : Global cumulative inflows of FDI and bank loans to the Asian-4 countries,1989-1996 (in billion of 1990 US dollars)

0

50

100

150

200

1989 1990 1991 1992 1993 1994 1995 1996

FDI Loans

Source : Table 2, chapter XX.

Interestingly and paradoxically, the financial turmoil that started in Japan in late 1990 -

beginning 1991 barely affected that trend. This is in sharp contrast with the decline of

Japanese lending to the US, observed during the first half of the 1990s. At that time, the fall

was mainly explained by the need for Japanese banks to rebalance their loans in favour of

domestic companies so as to avoid a major credit crunch in Japan (see Peek and Rosengren,

1997). The sharp increase in Japanese lending to East Asia was thus mainly lead by the wave

of FDI to the region and the banks’ strategies to strengthen their position in the region (to

16

facilitate both production and distribution). These new loans were typically issued with

guarantees from the parent company. This is particularly true for Malaysia, Thailand and

Indonesia but less true for South Korea due to its relatively high barriers to FDI.

Peek and Rosengren (1998) add two other important driving forces behind the rapid increase

in Japanese lending which must have been particularly important in the years just preceding

the crisis. First, East Asia enjoyed much higher yields than other regions and than financial

centres. The authors show that Japanese banks abandoned most of the low margin wholesale

markets, especially in Hong Kong and Singapore, in favour of more profitable markets.

Second, Japanese banks developed customer relationships with non-affiliates as these retail

operations also proved more profitable. In sum, Japanese banks decreased their international

exposure due to the fall in asset prices in Japan but kept a strong position in Asia as that

region enjoyed relatively higher yields than other regions in the world, especially higher than

Japan where interest rates were so low. The fixed or quasi fixed exchange rate regimes

adopted by most of the countries in the region also increased significantly the confidence of

Japanese lenders.

The fact that Japanese banks may behave similarly abroad and domestically could also

explain why so much of lending was short-term. Indeed, short-term lending to group

affiliates is common practice as loans are routinely rolled over insuring thereby long-term

finance (see for example Aoki and Patrick, 1994). Furthermore, under the Japanese

legislation commercial banks which provide the vast majority of lending both domestically

and internationally, can only lend short term (less than a year). Long term lending is being

provided by the long term credit banks.

Another factor that might have also played a role (see section on regulatory bias), is the

incentive for Japanese banks facing low capital-asset ratios to lend short-term. This was

particularly true at the beginning of the 1990's because their capital adequacy ratio was so

close to 8% (see Reisen, 1999).

McCauley and Yeaple (1994) underline that the Basle Accord did have an impact on

Japanese banks lending at least in the inter-bank market, but not in a way one would expect.

6 Due to the lack of data we use global flows from the IFS database rather than Japanese flows.

17

Although the weighting for inter-bank lending is only 20%, Japanese banks still reduced their

exposure to that market because the yields were not high enough. Indeed, the low weights

were not sufficient enough to promote inter-bank lending as this market's profitability was

too low.

As a conclusion, we have shown that the rapid build up of short-term foreign debt in the East

Asian region can be explained by a mix of factors. Apart from the high growth and very good

rating records of the four countries, which was clearly a major factor, we have analysed the

relative importance of moral hazard, the bias in the BIS capital accord and the behaviour of

Japanese banks. Interestingly, the latter factor which has received least attention in the

literature seems to have counted for a great deal. However, more research needs to be

conducted in that area.

4. Why did East Asia experience such large reversals?

In this section, we try to identify the reasons behind the large reversals in bank lending to the

region starting with recent work which has been done in an attempt to shed new light on the

rationale behind the large reversals. Following a brief review of the common lender channel

literature (section 4.1), we provide new evidence, based on data from individual banks, on the

driving forces underlying the reversals in bank loans.

We do not analyse in depth the more widely discussed causes which have been highlighted in

the literature and are now accepted as having played an important role 7. These factors are

clearly important. For example, the factor that has reached the largest support in the literature

is certainly the financial panic argument. This was put most forcefully by Radelet and Sachs

(1998) and Krugman (1999) and formalised notably by Chang and Velasco (1998).

According to the authors, the scale of the reversal experienced by the East Asian countries

cannot be attributed only to changes in economic fundamentals. Indeed, in a bank run style

panic taking place at the international level, short-term creditors can precipitate a solvent

borrower into a financial crisis by withdrawing their loans. The economy then moves from a

good to a bad equilibrium because of a very sharp loss in investor’s confidence. It is now

7 For our analyses on these factors, see Griffith-Jones et al (1998).

18

widely accepted that the most affected East Asian countries suffered from a run by

international banks and investors.

4.1 The common lender hypothesis

Recent studies argue that Japanese banks, in a similar fashion to US banks in the 1980s in the

case of Latin America, due to their dominant position in East Asia, played an important role

in aggravating the crisis (see for example Kaminsky and Reinhart, 1999). Kaminsky and

Reinhart’s hypothesis is that if a country enters a crisis, the lender will need to pullout from

other borrowing countries in order to rebalance its portfolio and meet capital adequacy

requirements. Then, all countries that have a common lender are exposed to a crisis should it

occur in any one country. The literature on the common lender channel argues furthermore

that the explanation of a fall across countries is a response to an almost "mechanical"

readjustment (see Van Rijckeghern and Weder, 1999).

The following table clearly shows the dominant position of Japan in the region, although as

shown the section 2, it has been decreasing over the period.

Table 7 : Share of Japanese liabilities in total liabilitiesas of June 1994 as of June 1997 as of June 1999

Indonesia 54% 39% 32%

Korea 29% 23% 24%

Malaysia 40% 36% 33%

Thailand 57% 54% 53%

Source: Kaminsky and Reinhart, 1999 and Table 1

Van Rijckeghern and Weder (1999) also test for the common lender hypothesis and find

similar results to Kaminsky and Reinhart (op. cit.) : the hypothesis is confirmed both in the

cases of Mexico and Asia 8.

The results found are quite robust and convincing. Evidently, and as shown in our first

section, aggregate data on countries' exposure to East Asia, show that Japanese loans reversed

in all four countries at the time of the crisis. However, if the common lender channel was to

8 In order to test the importance of the common lender channel, the authors measure the impact of a crisisoccurring in one country on the country’s lenders exposure in other countries.

19

hold, this trend should also be reflected at the level of individual banks : banks should

withdraw on the same scale in countries to which they are exposed following difficulties

experienced by one of their main borrower. We analyse this hypothesis in the next section.

4.2 New evidence from individual banks

We have gathered data on selected individual banks' exposure to East Asia during 1997 and

1998. The data was collected from the annual reports of the banks which have devoted a

special section to their exposure to East Asia. Although we do not have all banks involved in

Asia we believe the sample provides a satisfying overall picture.

Table 6 reports international banks' exposure to Asia at the end of 1997 and 1998. Due to the

lack of data, we focus on France, the UK, Japan and the US. The US is the least

representative as we did not find data on many banks. A number of caveats in the data need

to be underlined, especially in the case of Japan. First, in some cases, the data include on and

off balance sheet exposure. Second, the data on Japanese banks report their positions at the

end of the 1997 and 1998 fiscal years, that is end March 1998 and end March 1999, and thus

do not exactly cover the crisis. Third, the data was available either directly in US dollars or

in Yen. But Japanese banks’ exposures include both loans in foreign currencies (mainly in

dollars) and in local currencies. There is thus potentially a bias due to exchange rate problems

(This could also be true for other countries). Fourth, the definition of the exposure varies

from one bank to another and might thus limit comparisons between banks and countries.

Bearing in mind these words of caution, we believe that the data set still represents a valuable

contribution and allows for a better understanding of the micro-economic behaviour of banks

during the crisis.

In order to assess how representative the sample is, we compare total exposure by country to

other sources of data. As can be seen from table 6, we find an outflow from the four

countries of $32 billion during 1998. The IIF (1999) finds an outflow of $36.1 billion

including the Philippines. However, the BIS reports an outflow, for the 4 countries, of $67.3

billion. At the country level, we find similar exposures to those reported by the BIS for

France ($25.3 bn compared to $23.3 bn), and the UK ($14.1 bn compared to $15.7 bn). As far

as Japan is concerned, the difference is bigger as we find a total exposure of $41.1 bn against

$ 88.7 bn reported by the BIS. We can thus conclude that although the coverage is not fully

20

comprehensive, the data gathered is still representative enough to draw some analysis on the

behaviour of individual banks by country.

As can be seen from table 6, French, British and American banks behave as expected. All

banks diminish their exposure significantly to all countries. However, the scale of the

declines varies between banks and borrowing countries. For example, while Barclays

reduced its exposure to the 4 Asian countries by 40%, HSBC’s exposure declined by far less

(-16%). The declines do not seem to be related with the stock of claims as, although HSBC

had a position six times larger than Barclays, other banks with similar exposure to HSBC

(Chase Manhattan, Société Générale) also pulled out very significantly.

The specific position of HSBC as being largely based in the region and characterised by long

standing ties with local customers could be a plausible explanation for the smaller decline.

Furthermore, as pointed out by an economist from a London bank, the fact that HSBC has

balance sheets in East Asia leads the bank to assess its risk exposure differently. Indeed, it is

more likely that a bank such as HSBC with a local physical presence will adopt a more long

term strategy and reduce its exposure to a lesser extent than other banks during a crisis if it is

viewed as a cyclical phenomenom9.

9 Interview material

21

Table 6 : Exposure of individual international banks to the 4 most affected countries, 1997-1998

South Korea Indonesia Malaysia Thailand TotalCountry Bank End

1997End

1998% change End

1997End

1998% change End

1997End

1998% change End

1997End

1998% change End

1997End

1998% change

France BNP 1.7 1.6 -6% 1.0 0.8 -20% 0.6 0.5 -17% 1.0 0.7 -30% 4.3 3.6 -16%Crédit Agricole 1.9 1 -47% 0.7 0.5 -29% 0.2 0.1 -50% 1.1 0.6 -45% 3.9 2.2 -44%Crédit Lyonnais 3.7 2.4 -35% 1.3 1.0 -23% 0.3 0.2 -33% 0.6 0.5 -17% 5.9 4.1 -31%Paribas 1.5 1.0 -33% 0.4 0.4 0% 0.3 0.3 0% 0.4 0.3 -25% 2.6 2.0 -23%Société Générale 4.5 2.7 -40% 0.8 0.6 -25% 0.9 0.6 -33% 2.4 1.8 -25% 8.6 5.7 -34%Total 13.3 8.7 -35% 4.2 3.3 -21% 2.3 1.7 -26% 5.5 3.9 -29% 25.3 17.6 -30%

UK Barclays 1.0 0.8 -20% 0.5 0.2 -60% Na na 0.5 0.2 -60% 2.0 1.2 -40%Natwest 1.8 1.3 -28% 0.4 0.3 -25% Na na 0.4 0.2 -50% 2.6 1.8 -31%HSBC 4.1 3.6 -12% 1.8 1.4 -22% Na na 2.8 2.3 -18% 8.7 7.3 -16%Lloyds TSB 0.5 0.8Abbey National 0.8Total 6.9 6.2 -10% 2.7 1.9 -30% 3.7 2.7 -27% 14.1 11.1 -21%

Japan Bank of TokyoMitsubishi

3.4 2.7 -21% 2.6 1.2 -54% 1.3 1.1 -15% 4.5 2.7 -40% 11.8 7.7 -35%

Dai Ichi Kangyo 1.6 1.7 6% 1.4 1.0 -29% 0.4 0.4 0% 1.8 1.6 -11% 5.2 4.7 -10%Sanwa 0.9 2.0 122% 2.3 0.9 -61% 0.8 0.7 -13% 2.2 1.7 -23% 6.2 5.3 -15%Fuji 0.9 1.2 33% 1.4 0.8 -43% 0.6 0.5 -17% 1.8 1.1 -39% 4.7 3.6 -23%Sumitomo 1.6 1.2 -25% 1.2 0.9 -25% 0.6 0.6 -0% 2.0 1.4 -30% 5.4 4.1 -24%Tokai 0.4 0.3 -25% 0.6 0.5 -17% 0.6 0.5 -17% 1.4 0.8 -43% 3.0 2.1 -30%Sakura 0.7 1.0 -43% 1.4 1.4 0% 0.3 0.3 0% 2.4 1.5 -38% 4.8 4.2 -13%Total 9.5 10.1 6% 10.9 6.7 -39% 4.6 4.1 -11% 16.1 10.8 -33% 41.1 31.7 -23%

USA Chase Manhattan 5.3 2.4 -55% 2.2 1.2 -45% 0.9 0.6 -33% 1.4 0.9 -36% 9.8 5.1 -48%Citicorp 2.6 <0.8 na 0.6 na na 0.7 na na 0.3 na na 3.5 na naJ.P. Morgan 3.4 1.7 -50% 0.8 0.2 -75% 0.4 na na 1.2 0.2 -83% 5.8 2.1 -64%Total (1) 11.3 4.1 -64% 3.6 1.4 -61% 2 0.6 -70% 2.9 1.1 -62% 19.1 7.2 -62%Grand total 41.0 29.1 -29% 21.4 13.3 -38% 8.9 6.4 -28% 28.2 18.5 -34% 99.6 67.6 -32%

Note : Exchange rate : 31/12/97, 1 ECU = 1.10 USD, 1FRF = 0.17 USD; 1GDP = 1.65 USD ; 31/12/98, 1 EURO = 1.17 USD; 1GDP = 1.66 USD ; 1 FRF = 0.18 USD31/03/98, 1 JPY = 0.007524 USD; 30/09/98, 1 JPY = 0.007369 USD; 31/03/99, 1 JPY = 0.008396 USD.Some banks include their on and off balance sheets credit exposure while others do not. Data exclude for most banks sovereign risk, trade credit risk and export credit agency guarantees(1) The total for the USA needs to be interpreted with care due to lack of data.

22

Apart from HSBC and BNP, all European and US banks reduced their exposure on a large

scale across the four countries (between 26 and 62% of their claims).

The case of Japan is much more striking. Contrary to European and US banks, Japanese

banks do not behave similarly. Three out of the seven banks studied reduced their exposure by

15% or less while the remaining four cut their exposure by more than 20% (up to 30% in the

case of Tokai Bank). Bank of Tokyo-Mitsubishi is the largest lender to the region both

compared to other Japanese and international banks. It is also the Japanese bank that reduced

the most its exposure. This is due to the different historical background of Bank of Tokyo-

Mitsubishi. Indeed, it is the most long established Japanese bank at an international level

because of its role in the promotion of exports of national industries10. More differences

appear when looking at each of the four borrowers. Five out of the seven banks actually

increased or maintained their exposure in one or more borrowing countries11.

Surprisingly, we find an increase in total Japanese lending to Korea, increase which is not

reflected in the BIS data. It results from the significant rise in exposure by three Japanese

banks, DKB, Sanwa Bank and Fuji bank. According to an economist from a Japanese bank

based in Tokyo, this can be explained by the negotiations held in January 1998 between the

Korean government and private creditors which resulted in the exchange of $24 billion of

short-term debt for government guaranteed loans. Under this agreement, Korean banks

exchanged their short-term non trade credit for new loans with longer maturities. Sanwa bank

was one of the 13 international banks which negotiated the deal12.

The differences in behaviour between Japanese banks can be explained by several factors.

First, it is quite likely that lending decisions in times of crisis might differ whether the

borrower is a group affiliate or not. Indeed, as stated in the annual report of the Bank of

Tokyo-Mitsubishi, much lower provisions were made for bad loans which had been given to

group affiliates : “From a credit perspective, we regard affiliates of Japanese corporations

differently from other types of borrowers since, in many cases, we obtain some kind of

10 Interview material11 It is also worth noting that the scale of the declines in exposure to different countries for a single bank varieswidely. For example, the Bank of Tokyo reduced its exposure to Indonesia by 54%, that is more than twice thanto South Korea (-21%) and to Malaysia (-15%).12 Interview material

23

guaranty or other form of assurance of support from the affiliated Japanese corporations”

(Bank of Tokyo-Mitsubishi, 1999). Table 7 reports exposures of Dai Ichi Kangyo Bank

(DKB) to the Asian-4 between end March 1998 and end March 1999.

Table 7 : Loans outstanding in Asia by location of borrowers (Millions of US dollars)Japanese affiliates Non Japanese Public Total1997 1998 %

change97/98

1997 1998 %change

97/98

1997 1998 %change

97/98

1997 1998 %change

97/98Thailand 829.4 788.4 -4% 948.9 734.7 -14% 56.0 41.1 -2% 1834.3 1565.0 -15%Indonesia 430.0 339.2 -10% 914.0 597.0 -30% 77.8 107.5 9% 1421.8 1045.3 -26%Malaysia 95.0 152.8 21% 271.5 48.7 -58% 18.6 179.7 105% 385.1 382.9 -1%Korea 54.4 47.0 -1% 1290.3 1660.7 21% 263.8 26.9 -504% 1608.5 1736.3 8%Source : 1999 and 1998 Annual Reports from Dai Ichi Kangyo Bank, www.dkb.co.jp

The data show that DKB reduced its exposure to Japanese affiliates to a lesser extent than to

non-affiliates with the exception of Korea. Malaysia is a striking example where lending to

affiliates actually increased (+21%) while it was cut by more than half to non affiliates (-

58%). DKB’s 1998 Annual Report, in a similar way to Bank of Tokyo-Mitsubishi,

differentiates between affiliates and non-affiliates : “Many Japanese corporations have

operations in the region and will continue to require financial services. For these reasons, the

Bank continues to regard Asia as an important area in its international activities and will

continue to promote its operations there” (DKB, 1998, p25). As far as Korea is concerned,

and as mentioned above, the increase to non-Japanese affiliates (mostly to banks) reflects the

deal reached between the Korean government and private creditors in January 1998.

Another factor that might explain these differences is that the banking crisis in Japan has not

affected Japanese banks equally. Healthier banks might thus behave differently from more

fragile ones as they do not face the same capital asset constraints.

Another question worth investigating is why would loans be renewed during the crisis?

Although the answer to that question is not completely clear-cut, there are at least two reasons

that can be identified apart from the rescheduling agreements. First, as underlined by Hoshi et

al. (1990, 1991), one of the specificity of the main bank system is that in times of crisis it is

likely to keep on providing finance to its customers. This is due to the more long term

perspective they adopt. Second, Peek and Rosengren (1998) point out that some banks might

be rolling over loans to help borrowers in difficulty to meet their repayments so that the

24

lending bank does not have to report an increase in its bad loans. It is worth saying that this

particular behaviour is not unique to the Japanese case.

Why then did they pull out? The banks that pulled out did it mainly because they needed to

improve their balance sheet and profitability. The deterioration in their balance sheet was

caused by at least three factors. The two first ones are more specific to the Japanese economy.

First, it is important to underline that Japan hit the worst of its crisis in the first half of 1997.

This was highlighted by the failure of major commercial banks and later, in November 1997,

by the collapse of four major stockbrokers. The deepening of the Japanese banking crisis lead

the Financial Authorities to call for the restructuring of banks balance sheets13. This led to a

massive reorganisation of international portfolios through the reduction of overseas operations

in Asia (see the fall in the second half of 1997 table 5) and in Europe and the US. In 1998, a

new legislation from the financial authorities formally called for the restructuring of balance

sheets which lead to further withdrawals of loans. According to an economist from a Japanese

bank based in London, during the second half of 1997 and first half of 1998, non-affiliates

were the primary targets of the reductions in exposure. According to this economist, the end

of the fiscal year (March 1998) and of the interim report (September 1997 and September

1998) were used as benchmarks for the reallocation of portfolios due to publication

requirements and thus played an important role in the timing of the reversals.

Second, Kwan (1999) points out that the weakening of the yen directly impacted upon capital

adequacy ratios of Japanese banks as 80% of overseas lending is denominated in dollars.

Indeed, a decline of the yen reduces the capital adequacy ratio by raising the value of risky

assets in yen terms (Kwan op.cit.).

The third factor which is not so specific to Japanese banks is the increasing amounts of bad

loans. As a matter of fact, the devaluation of local currencies in the Asian 4, increased the

debt burden of local borrowers and thus their ability to repay. Repayments were all the more

threatened because of the loss of access to refinancing and because of monetary policy

tightening. This

13 Interview material

25

These three factors lead some Japanese banks and other international banks to pull out. This

was accompanied by a loss of confidence by international banks mainly Europeans and

Americans.

In sum, we have shown that the reversal in bank loans from international banks during the

Asian crisis has several causes. They are both bank specific and country specific. As far as

banks are concerned, we have found clear differences between European and Japanese banks.

The behaviour of the former seems to confirm the common lender channel and the panic

hypotheses. Indeed, banks loans fall similarly between banks and across borrowers with the

exception of HSBC and BNP.

As for Japanese banks, the story appears more complex than just a common lender channel as

we do not find any clear similarities in their behaviour. We have proposed a number of

explanations. First, Japanese banks have long term relationships with their customers -

affiliated or non affiliated to the group. These long term relations provide, in normal times, a

guaranty for long term finance. However, in times of crises, one needs to take into

consideration the likelihood of repayment on the side of the borrower and the capital asset

ratio on the side of the lender. Japanese banks usually treat bad loans differently whether they

originate from affiliates or non affiliates, as in the first case they are backed by their parent

company. However, as the crisis unfolded, the impact of the devaluation on the level of bad

loans might have interrupted the rolling over of loans even to affiliates. On the supply side,

the Japanese banking system crisis worsened the capital asset ratios of many banks. This was

aggravated by the declining yen. As a result, banks were forced to reduce their exposure.

However, and as pointed out earlier, the fact that some banks are healthier than others is

another reason for finding differences between banks.

The following figure sums up the different scenarios and their potential impact on lending

decisions in the case of Japanese banks.

26

Figure 1 : Impact of the level of the capital asset ratio and bad loans on lending decisionsLEVEL OF BAD LOANS

Affiliates Non Affiliates

DemandSide

SupplySide LOW HIGH LOW HIGH

Capital

Asset ratio

< 8%

+ ? - - ? -

Capital

Asset ratio

> 8%

+ + ? + - ?

Note: + means that banks roll over their loans- means that banks do not roll over-? means the impact on decisions is unclear but the tendency is more for the bank to pull out+? means the impact on decisions is unclear but the tendency is more for the bank to roll over the loans? changes in decision as the crisis unfolds

The figure shows that a combined deterioration in both supply side (capital asset ratio) and

demand side (increasing bad loans) factors eventually leads to the non renewal of existing

loans. Although there are differences of treatment between affiliates and non-affiliates, these

differences seem to fade away as the crisis deepens.

5. Policy implications : a discussion of the new Basle proposal and suggestions for

more counter-cyclical policies

In the last two sections we have shown that the Basle capital accord implemented by the Bank

for International settlements in 1988 might have played a role in both the build up of short

term lending and its reversal. In this section, we review the 1988 accord and the possible

avenues for reforms that are currently under discussion. We focus on a new proposal by the

BIS published in June 1999 and assess its potential impact on lending to emerging markets in

terms of volume, maturity as well as volatility. This section draws from a presentation made

by Colin Miles at a meeting held in London in November 199914.

14 Colin Miles is Head of Economic Data Risk Analysis at the UK Financial Services Authority

27

In June 1999, the Basel Committee on banking supervision published a consultative paper

proposing amendments to the original capital accord which was created in 198815. Although

it is not clear, at the time of writing, that the whole proposal will be implemented as such, it

will certainly set a new framework from which future reforms will be derived. It deserves all

the more attention because the European Commission also published in late 1999 a very

similar proposal reviewing the regulatory capital requirements of European financial

institutions (EU, 1999).

The most significant change is the attempt to depart from the simple original classifications of

risks to make them more in line with the changes observed in the international financial

environment.

The new proposal has four objectives : (1) to improve the way regulatory capital requirements

reflect underlying risks, (2) to maintain the overall capital adequacy of the banking system,

(3) to strengthen competitiveness and (4) while focusing on "internationally active" banks, the

underlying principles should have universal application.

The proposal is built on three pillars : minimum capital requirements, supervisory review and

market discipline.

As far as the supervisory review is concerned, the proposal underlines that a bank's capital

position should be consistent with its overall risk profile. Supervisors should have the power

to vary capital requirements depending on particular circumstances. This would require banks

to have internal capital assessment and targeting powers that are externally verifiable by

supervisors.

Concerning the pillar on market discipline, banks should be more transparent in the way they

present themselves to the market16. More specifically, banks should provide better

information on (i) capital structure (including components of capital, bad debt provisions,

accounting policies etc.) (ii) risk expenses (credit, market, liquidity, operational) (iii) capital

adequacy (including minimum regulatory ratios, details of internal assessment process).

15 The consultative paper A New Capital Adequacy Framework issued by the Basel Committee in June 1999 isavailable at www.bis.org.16 The BIS published in January 2000 a new document A New Capital Adequacy Framework: Pillar 3 - MarketDiscipline which has been designed to support the third pillar of the 1999 consultative paper.

28

It is clear that in order to comply with these new requirements, a number of developing

countries would need a transition period to build capacities both in the private and official

sector.

The pillar which has received most of the attention is dedicated to the preparation of a revised

grid of risk weightings. The grid under the 1988 system (pre reform) is as follows:

Table 8 : Risk weights under the 1988 accordZERO RISKWEIGHTINGS

CashClaims on cenbanks and national govts denominated in national currencyand funded in that currencyClaims on OECD central govt and cenbanksClaims collateralised by cash of OECD central govt or guaranteed byOECD central govts.

20% RISKWEIGHTING

Claims on multilateral development banks and claims guaranteed by orcollateralised by securities issued by these banksClaims on banks incorporated in the OECD and loans guaranteed by OECDincorporated banksClaims on banks incorporated in countries outside the OECD with residualmaturity of up to 1 yrClaims on non domestic OECD public sector entities (excluding centralgovt) and loans guaranteed by these entitiesCash items in the process of collection

50% RISKWEIGHTING

Loans fully secured by mortgages on residential property that is eitherowner occupied or rented.

100% RISKWEIGHTING

Claims on the private sector ·Claims on banks incorporated outside the OECD with residual maturity ofover 1 yrClaims on central govts outside the OECD (unless denominated/funded innational currency)Claims on commercial companies owned by the public sectorPremises, plant and equipment and other fixed assetsCapital instruments issued by other banks

Source : Ideaglobal, London

The purpose of the new approach is to overcome the shortcomings of the current one by

moving away from the OECD membership criteria and the short-term bias. The answer

proposed is to shift from what is seen as an arbitrary assessment of risk to the use of external

ratings both through rating agencies (and internal ratings for "some sophisticated" banks) and

export insurance companies (EIC) which provide a refined typology of risk. Table 9 compares

the new risk weights to the existing system.

29

Table 9 : Risk weights under the new proposal

Current Proposed

OECD Non-OECD AAA

to AA-

A+ to A- BBB+ to

BBB-

BB+ to B- Below B- Un-rated

Sovereigns 0 100 0 20 50 100 150 100

Banks 20 20 if up to 1 year, 100

otherwise

Option 1(1) 20 50 100 100 150 100

Option 2(2)

Under option 2, noclaims can receive a riskweight less than thatapplied to claims on itssovereign.

> 6 months :

< 6 months :

20

20

50

20

50

20

100

50

150

100

50

20

Corporates 100 20 100 100 100 150 100

(1) The ratings apply here to the sovereign, not the bank. (2) The ratings apply here to the banksSource : Adapted from Miles

30

In the new proposal, external rating agencies are to be relied upon when lending to sovereigns

or central banks. As can be seen from table 9, the weights would range from 0% (investment

grade) to 150% (for ratings below B-) : "risk weights applied to such claims [would] be

benchmarked to the assessment results of eligible external credit assessment institutions"

(BIS, 1999, p27).

Claims on corporates would also be treated differently under the revised system. "The

Committee now proposes that the standard weighting of claims on corporate remains at

100%, but that a weighting of 20% be given to claims on corporate of a very high quality and

that a weighting of 150% be given to claims on corporates which are of very low quality".

(See Table 9) (BIS, 1999, p30).

Claims on banks are dealt with in the following section.

It is worth noting that there are two important qualifications to be eligible for a risk weighting

lower than 100% : First, a sovereign borrower must subscribe to IMF's Special Data

Dissemination Standard. A number of developing countries have expressed some reservation

on that point. Indeed, as discussed in chapter XX (which devotes a section to standards) it is

quite likely that the implementation of some standards will impose a heavy burden on these

countries. Second, a bank's home supervisor should have implemented or endorsed the Basel

Core Principles for Banking Supervision.

5.1 Risk weightings on banks and impact on maturity

In the previous sections, we have shown the crucial role of short-term inter-bank lending both

in the build-up and unfolding of the East Asian crisis. Other authors have also underlined the

need to reform the short-term bias in the accord (see for example Cornford, 2000). It is thus

very important that any reform of the 1988 capital accord be looking very carefully at

weights applied to this type of lending. Changes in that area should also be implemented very

rapidly to eliminate any distortion.

As can be seen from table 9, the BIS has proposed two options. Under the first one, "claims

on banks would be given risk weights based on the weighting applied to claims on the

31

sovereign in which the bank is incorporated. The weight applied to the bank would be one

category less favourable than that applied to the country" (BIS, 1999, p28).

"The second option would be to use ratings assigned directly to banks by an external credit

assessment institution" (BIS, 1999, p29). Under the second option, the short term versus long

term distinction would still apply. Loans of a short term maturity (less than 6 months) would

receive one category more favourable than the usual risk weight on the bank's claim (with

floor of 20%). For example, if a claim on a bank would be weighted at 50%, a short-dated

claim on that bank would be weighted at 20%.

It is worth underlining that the second option accentuates even more the gap between short-

term and long-term. Indeed, the proposed definition of short-term is of six months compared

to one year under the 1988 system. The proposal of the European Commission is also still

favouring a preferential treatment on short-term lending (EU, 1999). This reflects, to some

extent, the emphasis being put on the stability of one single lending bank rather than the

overall stability of the borrower. Indeed, it is generally agreed that from the lender’s

perspective short-term loans have usually a higher probability of repayment. However, if

several lenders lend to the same borrower on a short-term basis, a systemic risk arises due to

potentially unsustainable levels of short-term liabilities, as clearly exemplified by the East

Asian crisis. This in turn can threaten the stability of the banking system on the supply side. It

is thus important that the issue of short-term lending takes into consideration the broader

issue of systemic risk.

5.2 Impact of the proposed reform on the volume of lending to emerging economies

Table 10 provides an indication of the likely impact of the new system on selected

developing countries. As mentioned above, the impact on a sovereign depends on its rating

and whether it is an OECD member.

32

The "winners" are non-OECD countries with above BB+ ratings (using Standard and Poors

rating grid)17. As shown in table 10, and given end 1999-beginning 2000 ratings, the big

winners will be Hong Kong, Taiwan and Singapore as they benefit from very high ratings.

Other countries which will gain include Chile and Thailand which are rated respectively A-

and BBB-.

Countries which will lose are OECD countries which have a rating below AA-. This is the

case of all emerging economies which are OECD members. Other losers are countries which

are unrated. Among the countries which will suffer most are Mexico and Turkey because

they benefit today from the OECD membership (0% weight) while having ratings which

would require a 100% weighting.

Table 10 : Sovereign Winners and LosersS&P Rating Current Proposed

Big losers (+100%) Mexico BB 0 100Turkey B 0 100

Losers (+50%) Pakistan SD 100 150Korea BBB- 0 50Poland BBB 0 50Czech Republic A- 0 50

Losers (0% to 50%) Indonesia CCC+ 100 150Russia SD 100 150

Not affected Brazil B+ 100 100Saudi Arabia NR 100 100

Winners (0% to -50%) Egypt BBB- 100 50S. Africa BB+ 100 50Thailand BBB- 100 50

Winners (-50% to -80%) Chile A- 100 20China BBB 100 50

Big winners (-80% to -100%) HK SAR A 100 20Taiwan AA 100 0Singapore AAA 100 0

Source : Miles

17 "Winners" in the sense that banks lending to those countries will not have to set aside as much capital asunder the 1988 accord and will thus be encouraged to lend more

33

Table 10 also shows that it is difficult to have a clear picture of the overall impact of the new

accord on total lending to emerging markets. The increase in funding costs is a factor which

could contribute to the lowering of the aggregate lending volume to developing countries.

This would be beneficial in the long-term as it would mean lower level of borrowing and

better pricing of risk.

5.3 Impact on volatility and cyclicality

It is worth noting that apart from the issue of a potential reduction in aggregate lending to

developing countries, other concerns about the new proposals have been expressed during the

consultation phase. We focus our analysis here only on the two which are of relevance to our

analysis. There is a growing agreement that the reliance on rating agencies to determine

weightings will (1) introduce excessive volatility in lending flows and (2) aggravate the pro-

cyclical nature of flows.

Table 11 reports the degree of stability of sovereign ratings over a period of one year. It

appears clearly that there is a high stability in ratings at the high end of creditworthy

countries (especially for countries of a rating higher than A). However, the correlation is

much smaller for countries which do not enjoy such high ratings. Countries like Mexico,

Brazil or Turkey which have ratings between B and BB are likely to face changes in their

rating within a period of a year. It is worth mentioning that in the case of Mexico and Turkey,

the new system could both reduce the amounts of loans to those countries but also increase

their volatility. The impact on volatility is nonetheless expected to be small given the

structure of ratings presented in table 9. Indeed, the variations in ratings, if they occur in

normal times, are likely to be of a small magnitude and could thus leave the weightings

unchanged.

34

Table 11 : Average One Year Transition Rates for Sovereign Foreign Currency RatingsRating at year-end

AAA AA A BBB BB B CCC SD or DAAA 97.3 2.7 0 0 0 0 0 0AA 0.8 96.9 0.8 0 0.8 0.8 0 0A 0 4.6 92.3 3.1 0 0 0 0BBB 0 0 5.1 88.1 5.1 1.7 0 0BB 0 0 0 6 85.1 6 1.5 1.5B 0 0 0 0 20 75 0 0

Rating atbeginning

of year

CCC 0 0 0 0 0 0 0 0S&P Ratings from 1975-1998

The story might be different in crisis time as demonstrated by the very rapid and large

changes in the ratings of the most affected Asian countries (see table 12). For example,

between June 1997 and December 1999, Korea's rating went from 0% to 50%, reaching

100% in June 1998.

Table 12 : Risk weightings for sovereignsCredit rating Risk weight Unchanged*

1997-99June 97 June 98 Dec 99

AAA to AA- 0% SingaporeTaiwan

Korea

A+ to A- 20% Hong KongThailandChileCzech Rep.

MalaysiaThailand

Malaysia

BBB+ to BBB- 50% ChinaColombiaHungaryPoland

IndonesiaColombia

ThailandColombia

Korea MalaysiaThailand

BB+ to B- 100% IndiaPhilippinesArgentinaBrazilMexicoPeruVenezuelaSouth AfricaTurkey

Russia KoreaRussia

Colombia

below B- 150% Indonesia IndonesiaRussia

* In some cases, the rating has changed by a notch or two but not enough to imply a change in risk weighting.Note: This table does not allow for other considerations (e.g. a country would, under the draft proposal, only beeligible for a risk weighting below 100% when it subscribes to the IMF's Special Data Dissemination Standard(SDDS)).Sources: Turner (2000), Basel Committee (1999), Standard & Poor's.

35

We now turn to the second important issue. There is growing evidence that rating agencies

act in a pro cyclical fashion. If that were the case, the reliance on ratings in the new system

would exacerbate boom-bust cycles and undermine the stability of the financial system.

The most recent evidence of this pro-cyclical pattern is the Asian crisis. Indeed, as pointed by

various authors (see for example Turner 2000, Cornford 2000 and Reisen, 1999), rating

agencies failed to downgrade the East Asian countries before the crisis but then worsened it

because they brought down the ratings as the crisis unfolded. Perhaps the most serious

criticism is that the rating agencies were backward-looking rather than forward-looking in

their evaluations; for example, they view depreciations as a factor for downgrading rather

than seeing a more competitive exchange rate as a sign of improved competitiveness.

Reisen and von Maltzan (1999) provide further evidence of the pro-cyclical nature of ratings.

Indeed, the authors assess the impact on the market of the publications of ratings by the main

rating agencies (Standard and Poors, Moody's and Fitch Ibca) and find that sovereign ratings

lag rather than lead the market.

These findings should not, however, question the need for reforming the 1988 accord. Indeed,

we have seen in table 9, the current system has fixed weightings which do not adjust with the

cycle. In the event of a recession the increased amounts of bad loans (which are usually not

fully covered by provisions) will impact upon the lending bank's capital and will therefore

lead to decreased lending if the bank is already facing a relatively low capital asset ratio, and

- as is likely in a recession - the bank is unable to raise new capital. There is indeed a strong

evidence that banks cut lending in recessions for this reason (see Jackson, 1999).

Additionally, the current system is biased towards short-term lending. It is thus important that

the new system becomes more flexible and more in line with the new international financial

environment. However, it is not clear that the proposed system will actually succeed in

bringing more stability in international and domestic banking systems. As a matter of fact, we

have seen that the reliance on rating agencies could possibly play the opposite role as it could

lead to higher volatility in bank lending.

The answer thus may lie in the implementation of an explicit counter-cyclical mechanism

which would, in boom periods, and in contrast to ratings, dampen excess lending. On the

contrary, in periods of slowdown and of scarcity of finance the new mechanism should not

36

further accentuate the decline in lending as exemplified by the 1997-1998 Asian crisis but

rather encourage it.

There would be two linked objectives for introducing elements of counter-cyclical regulation.

One would be to help smooth capital flows and the other would be to smooth the domestic

impact of volatile capital flows on the domestic financial system and therefore on the real

economy.

Introducing counter-cyclical elements into regulation would help build a link between the

more micro-economic risks on which regulators have tended to focus till recently and the

macro-economic risks which are becoming increasingly important, both nationally and

internationally18.

Counter-cyclical elements in regulation related to bank lending could be applied, either

internationally, nationally or at both levels.

Several mechanisms could be used to introduce a counter-cyclical element into regulation of

bank lending. One mechanism would be to get the required capital ratio higher in times of

boom, and to allow banks to use the additional cushion provided by the higher capital ratio,

so they could sustain lending in times of recession at a lower capital asset ratio (when

increased bad loans are likely to be reducing their capital). Higher capital requirements in

boom times seem an appropriate regulatory response, as capital requirements are intended

precisely to protect a bank from unexpected losses. Some practical difficulties may arise in

implementing such a mechanism, of which the most serious one may be getting international

agreement on a general formula for cyclically adjusted capital asset ratios. However such

agreement could be facilitated by the logic of the case. A second problem is that it is rather

difficult to distinguish between a sustainable increase in the rate of economic growth and a

temporary and unsustainable boom. Even if a boom is identified (and there are a number of

useful indicators, such as very rapid growth of domestic bank lending, sharp rises in stock

markets and property prices, rapidly growing current account deficits, etc) more narrow

technical issues remain about the moment in which capital adequacy ratios should be required

18 We thank Andrew Crockett for his suggestive remarks on this point.

37

to increase, and the amounts of the increase. Answers to the latter more narrow technical

issues, however, could be helped by empirical research on past booms.

A second mechanism for introducing counter-cyclical elements in bank lending regulation is

for regulators to encourage that higher general provisions be made for possible loan losses

(i.e. subtracted from equity capital in the books of the bank) to cover normal cyclical risks

(Turner, 2000). This would allow for provisions built up in good times to be used in bad

times, without affecting reported capital. The way to ensure this would be to maintain higher

general provisioning that applies to all loans. The main problem for this mechanism,

according to Turner, op.cit, may be that tax laws often limit the tax deductibility of

precautionary provisioning, as such deductibility is often only on losses that have occurred.

However, it is possible to change such tax laws, as indeed was done in the late eighties in the

UK, with the specific purpose of encouraging external debt reduction.

A third mechanism, relevant particularly for domestic bank lending, is for regulators to place

caps on the value of assets (such as real estate or stocks and shares) to be acceptable as

collateral, when the value of such assets has risen sharply in a boom and is at risk of

declining sharply in a recession. Rules could be used such as averaging values for the last

five years, or accepting only 50% of current prices in the peak period of a boom. The latter

mechanism seems to have the least problems of implementation (indeed, reportedly it is

already applied in some jurisdictions, e.g. Hong Kong). However, as pointed out above, it is

mainly relevant for domestic bank lending.

A fourth possible counter-cyclical mechanism is that, as suggested by McKinnon and Pill,

op.cit, monetary authorities could monitor and try to limit or discourage lending for property,

construction and personal consumption, as these items tend to increase substantially - and

often even be a major factor - in booms. Reportedly, such measures were discussed - but not

implemented - by the Mexican monetary authorities during 1993 and early 1994. A possible

implementation problem about this mechanism would be that it may be difficult to verify

final use of credit, and that therefore such measures could possibly be evaded to some extent.

Another problem may be that selective measures of this type seem to be less fashionable in

current practice.

38

Each of these mechanisms or combinations of them would be helpful to curb excessive

lending during booms. Particularly the first three mechanisms would also provide valuable

cushions to help avoid harmful contraction of lending during recessions.

Furthermore, regulators should be flexible in the downturn, particularly to allow banks to

easily use cushions (e.g. of capital or of provisioning) in times of recession; it may even be

advisable, if a recession is very serious, to allow ratios to fall below normally required levels,

(to help sustain lending), in the understanding that they will be rebuilt as soon as the

economy starts recovering. A tension may arise here between the regulatory concerns about

individual bank liquidity and solvency and the macro-economic externalities of their actions,

particularly in recessions. Indeed, a more relaxed regulatory attitude to individual banks'

problems in a recession may ensure higher lending; this will ensure more rapid recovery of

the economy, which in its turn may be the best way to ease the individual banks' financial

situation, as the level of bad loans will be significantly reduced.

To summarise. Introducing counter-cyclical elements in bank lending regulation could be

very important both in helping smooth bank lending to developing countries and in

smoothing bank lending domestically within these countries. This would make both

excessive surges of capital inflows and costly sharp reversals (as well as domestic booms and

busts) less likely, which would in its turn make banking crises less likely.

Specific issues seem to require further research. How best can the distinction between a

temporary boom and a permanent increase in growth be made? After what period of "boom",

should regulatory changes be introduced? How large should such changes be? What are the

best mechanisms through which counter-cyclical measures should be introduced (flexible

capital adequacy ratios, higher provisioning against losses, more "realistic" pricing of

collateral)? Should such measures be introduced for both international and domestic lending,

or preferably for one of them? Our paper provides only initial thoughts on these important

issues. Our main aim is to make the case for counter-cyclical elements in bank regulation.

39

REFERENCES

Aoki, M., and Patrick, H., (1994), The Japanese main bank system, its relevance for

developing and transforming economies, Oxford University Press.

Bank of Tokyo-Mitsubishi, (1999), Annual report, Tokyo.

Basle Committee on Banking Supervision, (1999) "Supervisory lessons to be drawn from the

Asian Crisis", Working Paper No. 2, June (available at www.bis.org).

BIS, (1999), A New Capital Adequacy Framework Consultative paper, Basel Committee,

June.

Chang, R. and Velasco, A. (1998), “Financial crises in emerging markets : a canonical

model”, Federal Reserve Bank of Atlanta Working Paper No 98-10, July.

Cornford, A., (2000), “The Basle committee’s proposals for revised capital standards:

rationale, design and possible incidence”, Paper prepared for the Meeting of the

Technical Committee of the Intergovernmental Group of 24, Lima, Peru, 1-3 March.

Corsetti, G., Pensenti, P. and Roubini, N., (1998), "What caused the Asian currency and

financial crisis?", Mimeo, Stern University.

DKB, (1998), Annual report, Tokyo.

EU, (1999), “A review of regulatory capital requirements for EU credit institutions and

investment firms”, Consultation document, European Commission internal market

directorate general, November

Griffith-Jones, S., with Cailloux, J. and Pfaffenzeller, S., (1998). "The East Asian Financial

Crisis : a reflection on its causes, consequences and policy implications". IDS

Discussion Paper No. 367.

IIF (1999), Capital flows to emerging market economies, IIF, Washington, April.

IIF, (1999), "Testing for moral hazard in emerging market lending", IIF Research Paper No.

99-1.

Ito, T., (1999), "Capital flows in Asia", NBER Working Paper Series, No. 7134, May.

40

Jackson, P. et al (1999) "Capital requirements and bank behaviour: the impact of the Basle

Accord" Basle Committee on Banking Supervision Working Paper 1. April.

Krugman, P. (1989), “Private capital flows to problem debtors”, in Sachs, J. (ed), Developing

country debt and economic performance, NBER, Vol. 1.

Krugman, P. (1999), “Balance sheets, the transfer problem and financial crises”, Mimeo,

MIT.

Krugman, P., (1998), "What happened to Asia", Mimeo, Harvard University.

Kwan, C.H., (1999), "Sayonara dollar peg". Paper presented at the T5 meeting, Royal

Institute of International Affairs, October.

McCauley, R., and Yeaple, S., (1994), "How lower Japanese asset prices affect Pacific

financial markets", Federal Reserve Bank of New York quarterly review, Spring,

p.19-33

Peek, J. and Rosengren, E., (1998), "Japanese banking problems: implications for Southeast

Asia". Paper prepared for the second annual conference of the Central Bank of Chile,

Banking, financial integration and macroeconomic stability, held in Santiago, Chile,

September 3-4, 1998.

Peek, J., and Rosengren, E., (1997), "The international transmission of financial shocks: the

case of Japan", American Economic Review, Vol. 87, September, pp 495-505.

Radelet, S. and Sachs, J. (1998), “The onset of the East Asian crisis”, Mimeo, Harvard

University.

Reisen, H., (1999), "Revisions of the Basel accord and sovereign ratings", Mimeo, OECD

Development Centre, Paris.

Reisen, H; and von Maltzan, J., (1999), "Boom and bust and sovereign ratings", OECD

Development Centre Technical Paper No. 148, Paris

Turner, P., (2000), "Procyclicality of regulatory ratios?", Note prepared for a seminar "A

world financial authority: a new structure for international financial regulation", 21-23

January 2000, Queens' College, Cambridge.

41

Annex 1:

Table 1 : International claims by nationality of reporting banks on Korea (in millions of US dollars); Grand

Total% from

previoussemester

Austria % fromprevioussemester

Belgium % fromprevioussemester

Canada % fromprevioussemester

Finland % fromprevioussemester

France % fromprevioussemester

Germany % fromprevioussemester

end June 99 63482 -2.8 525 -6.9 583 -19.6 .. 13 -40.9 8752 17.9 7605 -7.8end Dec 98 65293 -9.9 564 -9.2 725 -52.2 1512 -5.7 22 -21.4 7425 -6.2 8250 -1.8end June 98 72444 -23.1 621 -16.2 1517 0.0 1603 -8.2 28 -54.1 7913 -28.9 8400 -12.6end Dec 97 94206 -9.5 741 -38.9 1517 -61.1 1747 31.8 61 -42.5 11135 3.2 9616 -10.9end June 97 104148 4.2 1212 -4.5 3899 4.5 1325 -2.2 106 -37.6 10789 21.4 10794 8.2end Dec 96 99953 13.5 1269 1.0 3731 61.4 1355 55.9 170 15.6 8887 27.1 9977 17.0end June 96 88027 13.5 1257 137.2 2312 18.3 869 2.2 147 63.3 6994 0.1 8529 16.5end Dec 95 77528 8.5 530 -3.5 1954 -15.0 850 15.2 90 1.1 6990 22.5 7318 25.3end June 95 71430 549 2298 738 89 5705 5840

GrandTotal

% fromprevioussemester

Italy % fromprevioussemester

Japan % fromprevioussemester

Lux. % fromprevioussemester

Nethe. % fromprevioussemester

Spain % fromprevioussemester

UK % fromprevioussemester

US % fromprevioussemester

end June 99 63482 -2.8 422 -29.8 15018 -11.3 2127 -16.4 213 46.9 4685 -15.6 6420 2.1end Dec 98 65293 -9.9 601 -29.3 16925 -10.6 367 0.0 2544 -0.8 145 -35.3 5551 -1.5 6291 -15.1end June 98 72444 -23.1 850 5.5 18934 -6.6 367 -10.7 2564 32.2 224 -57.4 5634 -18.6 7409 -22.3end Dec 97 94206 -9.5 806 -41.1 20278 -14.6 411 -22.2 1939 11.7 526 -3.7 6924 14.2 9533 -4.3end June 97 104148 4.2 1369 13.3 23732 -2.4 528 -2.0 1736 -9.9 546 16.4 6064 7.5 9961 6.5end Dec 96 99953 13.5 1208 18.0 24324 8.0 539 20.3 1926 16.7 469 31.4 5643 36.3 9355 -2.4end June 96 88027 13.5 1024 1.4 22512 4.9 448 4.9 1651 45.8 357 19.4 4140 7.2 9582 26.2end Dec 95 77528 8.5 1010 0.9 21454 2.8 427 64.9 1132 19.2 299 1.0 3861 -48.6 7590 6.6end June 95 71430 1001 20874 259 950 296 7518 7122The grand total in the first column of the table also includes the international claims of domestically owned banks in Switzerland, which are not shown separately in thistable, as well as those of local subsidiaries and branches of banks which have their head- office outside the BIS reporting area.

42

Table 2 : Consolidated cross-border claims in all currencies and local claims in non-local currencies by maturity and sector. South Korea Total Maturities Sectors

Up to andinc. one

year

over oneyear upto twoyears

over twoyears

Unallocated Banks PublicSector

Non-bank;privatesector

Unallocated

end June 99 63482 34067 7192 11043 11180 36465 5204end Dec 98 65293 29607 8276 16591 10819 37149 5447 22665 32end June 98 72444 33161 9264 18919 11100 41300 4797 26261 86end Dec 97 94206 59465 5222 16440 13079 55957 3923 34249 77end June 97 104148 70897 4139 16366 12746 68007 4389 31680 72end Dec 96 99953 67506 4107 15884 12456 65896 5677 28310 70end June 96 88027 62332 3438 13434 8823 57852 5940 24141 94end Dec 95 77528 54275 2553 12027 8673 49949 6169 21375 35end June 95 71430 51439 2719 10801 6471 47583 5518 18308 21

43

Table 3 : International claims by nationality of reporting banks on Indonesia (in millions of US dollars); Grand

Total% from

previoussemester







end June 99 43764 -2.4 1426 -6.4 511 -36.9 .. 208 -1.9 3967 2.4 7542 33.8end Dec 98 44827 -10.8 1524 9.4 810 -71.1 471 -27.2 212 271.9 3874 -3.4 5638 -4.1end June 98 50268 -13.9 1393 -5.8 2805 0.0 647 -29.1 57 -42.4 4009 -16.0 5876 -4.8end Dec 97 58384 -0.6 1478 1.7 2805 -2.3 913 8.9 99 -20.2 4773 -0.3 6174 10.1end June 97 58733 5.8 1453 7.1 2870 -3.8 838 42.3 124 14.8 4787 7.3 5610 1.9end Dec 96 55523 12.6 1357 17.4 2982 27.1 589 43.0 108 44.0 4463 22.2 5508 13.7end June 96 49306 10.7 1156 6.3 2346 62.8 412 19.1 75 44.2 3652 11.1 4843 24.4end Dec 95 44528 10.2 1087 13.1 1441 72.6 346 174.6 52 10.6 3288 6.8 3893 19.3end June 95 40411 961 835 126 47 3079 3262

GrandTotal









end June 99 43764 -2.4 167 17.6 14043 -14.4 0 3366 1.4 160 -4.8 3428 -10.1 3724 5.3end Dec 98 44827 -10.8 142 -6.6 16402 -13.8 47 0.0 3319 -3.7 168 -22.9 3814 -3.9 3537 9.6end June 98 50268 -13.9 152 -17.4 19030 -13.6 47 -7.8 3445 15.8 218 -9.9 3967 -11.7 3226 -34.1end Dec 97 58384 -0.6 184 -1.6 22018 -4.9 51 -27.1 2975 5.4 242 7.6 4492 3.7 4898 6.7end June 97 58733 5.8 187 79.8 23153 5.1 70 29.6 2823 14.8 225 57.3 4332 13.0 4591 -13.0end Dec 96 55523 12.6 104 7.2 22035 1.9 54 -15.6 2458 -0.3 143 22.2 3834 17.6 5279 48.7end June 96 49306 10.7 97 51.6 21622 3.1 64 28.0 2466 4.1 117 67.1 3260 19.5 3551 27.8end Dec 95 44528 10.2 64 -25.6 20974 2.3 50 78.6 2368 19.1 70 -7.9 2727 -13.9 2778 21.0end June 95 40411 86 20512 28 1988 76 3166 2296The grand total in the first column of the table also includes the international claims of domestically owned banks in Switzerland, which are not shown separately in thistable, as well as those of local subsidiaries and branches of banks which have their head- office outside the BIS reporting area.

44

Table 4 : Consolidated cross-border claims in all currencies and local claims in non-local currencies by maturity and sector. Indonesia Total Maturities Sectors

Up to andinc. one

year


over twoyears



Unallocated


45

Table 5 : International claims by nationality of reporting banks on Malaysia (in millions of US dollars); Grand

Total% from

previoussemester







end June 99 18623 -10.6 195 20.4 202 -3.3 .. 22 175.0 2225 -4.7 2837 -38.6end Dec 98 20826 -9.5 162 -10.0 209 -54.4 450 82.2 8 -75.8 2335 -2.3 4618 -10.5end June 98 23024 -16.4 180 -5.3 458 0.0 247 -7.8 33 -5.7 2391 -17.1 5160 -28.3end Dec 97 27533 -4.4 190 -7.3 458 45.9 268 0.8 35 -10.3 2883 -1.7 7197 25.9end June 97 28800 29.5 205 201.5 314 26.1 266 111.1 39 56.0 2934 11.1 5716 48.2end Dec 96 22234 10.6 68 106.1 249 114.7 126 63.6 25 -30.6 2641 9.7 3857 20.7end June 96 20100 19.8 33 -10.8 116 56.8 77 -33.6 36 33.3 2408 17.5 3195 42.1end Dec 95 16781 14.0 37 12.1 74 15.6 116 176.2 27 -20.6 2049 7.2 2249 -4.8end June 95 14722 33 64 42 34 1912 2363

GrandTotal









end June 99 18623 -10.6 74 -14.0 6056 -8.6 934 0.4 110 6.8 2059 0.9 1074 25.2end Dec 98 20826 -9.5 86 -36.3 6623 -16.2 87 0.0 930 10.6 103 368.2 2040 26.5 858 -25.3end June 98 23024 -16.4 135 -39.2 7905 -7.6 87 13.0 841 -6.9 22 -12.0 1613 -19.9 1149 -35.7end Dec 97 27533 -4.4 222 -29.3 8551 -18.5 77 5.5 903 -14.7 25 257.1 2014 0.1 1786 -25.0end June 97 28800 29.5 314 88.0 10489 27.8 73 21.7 1059 45.1 7 600.0 2011 41.9 2380 1.8end Dec 96 22234 10.6 167 -10.7 8210 1.0 60 5.3 730 18.9 1 -87.5 1417 16.3 2337 23.3end June 96 20100 19.8 187 92.8 8131 11.1 57 26.7 614 32.9 8 300.0 1218 5.2 1896 24.5end Dec 95 16781 14.0 97 1.0 7321 20.2 45 40.6 462 -1.7 2 -83.3 1158 -21.2 1523 41.9end June 95 14722 96 6091 32 470 12 1470 1073The grand total in the first column of the table also includes the international claims of domestically owned banks in Switzerland, which are not shown separately in thistable, as well as those of local subsidiaries and branches of banks which have their head- office outside the BIS reporting area.

46

Table 6 : Consolidated cross-border claims in all currencies and local claims in non-local currencies by maturity and sector. Malaysia Total Maturities Sectors

Up to andinc. one

year


over twoyears



Unallocated


47

Table 7 : International claims by nationality of reporting banks on Thailand (in millions of US dollars); Grand

Total% from

previoussemester







end June 99 34694 -14.9 378 -16.4 495 -5.5 .. 725 2.0 2922 -17.7 4632 -1.2end Dec 98 40749 -12.9 452 -20.3 524 -48.4 539 -8.0 711 767.1 3552 -9.9 4687 -11.3end June 98 46801 -20.5 567 -12.1 1016 0.0 586 -23.7 82 -35.4 3943 -16.4 5286 -12.3end Dec 97 58854 -15.2 645 4.5 1016 -12.5 768 -24.2 127 -39.8 4718 -7.3 6028 -20.2end June 97 69375 -1.1 617 -7.6 1161 -10.6 1013 -9.1 211 -16.6 5089 11.0 7557 9.3end Dec 96 70147 1.1 668 8.4 1298 -0.6 1114 8.9 253 3.7 4583 4.9 6914 8.4end June 96 69409 10.5 616 19.8 1306 12.5 1023 25.8 244 29.8 4367 17.7 6381 28.2end Dec 95 62818 17.2 514 -4.1 1161 30.3 813 87.8 188 3.3 3711 35.0 4977 56.5end June 95 53604 536 891 433 182 2749 3180

GrandTotal









end June 99 34694 -14.9 216 -59.2 18278 -18.5 1261 -20.9 79 -46.6 1476 -16.8 1232 -9.3end Dec 98 40749 -12.9 529 67.9 22437 -14.1 101 0.0 1594 -11.0 148 -10.3 1775 -15.0 1358 -22.7end June 98 46801 -20.5 315 -6.8 26120 -21.3 101 -19.2 1792 7.0 165 16.2 2088 -11.6 1757 -30.6end Dec 97 58854 -15.2 338 -21.6 33180 -12.1 125 -17.8 1675 2.5 142 6.0 2361 -16.2 2533 -36.6end June 97 69375 -1.1 431 -6.5 37749 0.6 152 -7.3 1634 4.3 134 30.1 2818 -9.9 3997 -20.8end Dec 96 70147 1.1 461 6.5 37525 -0.1 164 2.5 1566 10.0 103 13.2 3128 1.9 5049 13.9end June 96 69409 10.5 433 20.9 37552 1.9 160 6.7 1424 45.8 91 85.7 3070 8.8 4433 8.2end Dec 95 62818 17.2 358 10.5 36845 12.9 150 94.8 977 19.3 49 6.5 2822 -37.5 4097 32.0end June 95 53604 324 32628 77 819 46 4518 3104The grand total in the first column of the table also includes the international claims of domestically owned banks in Switzerland, which are not shown separately in thistable, as well as those of local subsidiaries and branches of banks which have their head- office outside the BIS reporting area.

48

Table 8 : Consolidated cross-border claims in all currencies and local claims in non-local currencies by maturity and sector. Thailand Total Maturities Sectors

Up to andinc. one

year


over twoyears



Unallocated


INTERNATIONAL BANK LENDING INTERNATIONAL BANK LENDING AND THE EAST ASIAN CRISIS 1 Jacques Cailloux...

Documents

Transcript of INTERNATIONAL BANK LENDING INTERNATIONAL BANK LENDING AND THE EAST ASIAN CRISIS 1 Jacques Cailloux...