5 Falling Profitability, Higher Borrowing Costs, and Chaebol Finances During the Korean Crisis
- David Coe, and Se-Jik Kim
- Published Date:
- September 2002
Anne O. Krueger and Jungho Yoo*
Despite the similarities among the countries hit by the Asian financial crisis of 1997, there were some differences. Among them, there were allegations that Korean chaebol had misallocated resources and become overindebted; by contrast, the role of “crony capitalism,” in which friends and relatives of the politically powerful received loans that they were not expected to repay, was a focal point.
A number of factors contributed to the crises, and there is no broad consensus as to their relative importance. “Crony,” or “chaebol,” capitalism, the weakness of banking systems pre-crisis, financial liberalization and opening of capital accounts, and fixity of exchange rate regimes have all been singled out.
While all these factors obviously contributed, their relative importance quantitatively, and the interactions between them, are little understood. It is the purpose of this paper to delve, insofar as is feasible, into the role of the chaebol and the weak banking system in bringing about the Korean crisis. For that purpose, we trace the earlier history of credit rationing, and of the chaebol, the build-up of domestic credit and foreign indebtedness prior to the crisis, and the role of the chaebol in bringing about the crisis. Obviously, the crises in other countries may have had different relative contributions of financial and other variables and the Korean case cannot provide generalizations as to causation. But an understanding of the role of the chaebol and their financing can help, when experience of other countries is similarly examined, to enhance understanding of the vulnerabilities of countries in which the various factors listed above all play a role.
Such an understanding cannot be achieved, however, without reference to the role and relative importance of each of the key variables and their interaction. For example, if exchange rate depreciation was forced as a consequence of maintaining an unsustainable nominal exchange rate for a long period of time prior to the crisis and was quantitatively the largest factor in leading to the deterioration of the banks’ portfolios, resort in the future to a genuinely floating exchange rate and/or preventing uncovered liabilities denominated in foreign exchange should greatly reduce the likelihood of future crises, regardless of the indebtedness of the chaebol. Likewise, if bank lending practices would have resulted in a rapidly increasing proportion of nonperforming loans (NPLs) in the banking system even had the exchange rate not been a significant factor and even in the absence of chaebol indebtedness, the relative importance of improving bank lending practices as a preventive measure for future crises looms much larger. And if rigidities in the banking/financial system, resulting from failure to liberalize and/or regulate sufficiently, were a major contributing factor, the policy lessons would focus on the urgent need for liberalizing and strengthening banking and financial systems in emerging markets.
In a first section, we briefly sketch the roles that each of these factors can play in theory in financial crises. We then provide background on the Korean economy and the evolution of the banking and financial systems, the chaebol, and linkages to the international economy, which are essential building blocks for our later analysis. The third section examines the history of financing of the chaebol and their role in the Korean economy. A fourth section then provides estimates of the extent to which there were NPLs in the banking system prior to the crisis, and the share of the chaebol in those loans. A fifth section examines the role of foreign-currency denominated debt in intensifying the crisis. A final section then provides our best judgment as to the relative importance of the variables widely pointed to as contributing to crisis.
Domestic Credit Expansion, Lending to Chaebol and/or Cronies, Exchange Rate Depreciation, Capital Account Opening, and Crises1
The problem for analysis of the Asian crises is not the lack of explanations: it is that there are too many. In those crises, and in the Mexican crisis of 1994, a foreign exchange crisis and a financial crisis occurred almost simultaneously, and have come to be termed “twin crises.” As will be seen, there are theoretical grounds to anticipate that these twin crises are likely to have a far more severe impact on a domestic economy than either a financial or a currency crisis alone, and it is not coincidental that their onset is virtually simultaneous. In this section, we briefly review the role of each of the possible culprits in precipitating and intensifying twin crises. Once that is done, focus turns to interactions between them. Thereafter, we focus on the chaebol, and the likely orders of magnitude of the interactions of those variables with others frequently blamed for the crises.
Exchange Rate Pegging
Although any nominal exchange rate could, in theory, be associated with the appropriate real exchange rate,2 empirical evidence shows that governmental policies with respect to nominal exchange rates over periods of 3-5 years, if not longer, significantly affect real exchange rates. Whether this is because of long lags in adjustment or because of the unwillingness of the domestic authorities to adopt the monetary and fiscal policies consistent with their choice of nominal exchange rate is not relevant for present purposes.
Empirically, if the authorities intervene in the foreign exchange market for purposes other than smoothing short-term fluctuations (such as maintaining a fixed nominal exchange rate), the real exchange rate appreciates relative to major trading partners when domestic inflation exceeds the inflation rate in the rest of the world. Likewise, if for any reason (such as changes in the terms of trade or rapid growth of domestic demand for imports) the real exchange rate would adjust in a well-functioning free market but is prevented from doing so, there can be imbalances between the demand for, and supply of, foreign exchange. As long as the authorities can meet this demand, buying or selling foreign exchange as demanded, they can maintain their exchange rate policy.
All of the countries afflicted with twin crises in the 1990s had intervened heavily in their foreign exchange market in one way or another to achieve target nominal exchange rates. In the cases of Mexico and Thailand, the nominal exchange rate had either been fixed, or adjusted according to a formula which resulted in significant appreciation of the real exchange rate. In Indonesia and Korea, terms of trade shocks probably called for a significant real exchange rate depreciation at a time when there was some degree of real appreciation—as will be seen below for Korea. For Korea, the American government’s concern about its bilateral trade deficit with Korea resulted in significant political pressure to prevent depreciation of the won in nominal terms.
When government officials implicitly or explicitly indicate that they will maintain an exchange rate policy that results in an appreciating currency in real terms, they provide individuals and firms with a strong incentive to access the international capital market—the real interest rate is typically lower than in the domestic market.3 When domestic residents have access to the foreign capital market, or when domestic banks can borrow abroad, the result is an increase in the nation’s liabilities, and the exchange rate policy means that the government is increasing its contingent liabilities. The unsustainability of the nominal exchange rate policy results in a buildup of domestic credit and foreign liabilities until the time when either domestic residents and foreigners anticipate that the exchange rate will alter and attempt to get out of domestic money and into foreign currency and/or the public or private debt-servicing obligations denominated in foreign exchange are not voluntarily met. At that point, either the “run on the currency” results in a “currency crisis” or the prospective inability to continue voluntary debt-servicing forces the same outcome. Resolving the crisis almost always involves an alteration in the exchange rate, and usually in exchange-rate policy. 4
Of course, there can be a “pure” currency crisis without a financial crisis. This is more likely if there is a reasonably sound banking/financial system or a pre-existing highly restrictive set of capital controls that prevented the buildup of significant foreign indebtedness. Brazil’s devaluation in 1999 is one good example of a currency crisis in which there was no serious domestic financial spillover.
Financial Weakness and Crisis
If there is a continuing build-up of NPLs in the banking system, a financial crisis will result unless effective measures are taken to reverse the buildup. NPLs can come about for several reasons: (1) there can be an unforeseen macroeconomic disturbance (originating abroad or domestically) that leads to unfavorable outcomes for borrowers; (2) domestic credit expansion may be so rapid that banks are unwilling or unable to exercise normal prudence in lending and a disproportionate number of borrowers fail to be able to service their debts (often after a macroeconomic downturn); (3) banks may be directed or induced to lend to politically well-connected cronies, who do not service their outstanding loans; and, finally, (4) banks may lend to favored (economically important) enterprises who do not or cannot service their debt obligations. This last case includes the circumstance in which banks provide “evergreen” accounts for large businesses that are indebted to them, rolling over existing debt and extending credits to finance interest payments on it.
For Indonesia, it is thought that the third explanation—obligatory lending to politically well-connected friends and relatives of the President—was a significant factor in the NPLs of the banking system. In Thailand (and to a degree in Korea as will be seen below), rapid expansion of domestic credit contributed. In Japan in the late 1980s, where currency crisis was not a factor, a large negative macroeconomic shock when the rapid inflation of asset prices was reversed, was the trigger for difficulties in the banking system accounts.
Here, the important point is that once NPLs become significant in a bank’s portfolios, serious difficulties are likely to result in the absence of sufficient provisioning or capital. A bank with sizable NPLs must charge higher interest rates on its lending in order to cover its costs over a smaller proportion of its business. As such, if it has more NPLs than its competitors, only those unable to obtain cheaper credit at banks with healthier balance sheets will borrow from it, thus increasing the riskiness of its portfolio. At the same time, as depositors learn of the bank’s difficulties, they are likely to attempt to withdraw their deposits. And if the banks lend into arrears (so that the NPLs may not even show up as such on bank balance sheets), the problem of NPLs can mushroom rapidly. When many domestic banks have these difficulties at the same time, domestic credit can contract sharply. If there are foreign competitors (or if creditworthy borrowers can borrow abroad), the entire domestic banking system can be threatened.
Domestic Credit Expansion
Domestic credit can expand unduly rapidly because of government direction of credit or because of macroeconomic concerns. But it can also expand rapidly because of the incentives provided by the exchange rate regime, or simply because government monetary and fiscal policy is very loose for whatever reason. Rapid expansion of credit is dangerous: on one hand, it is inflationary, which means that for a while a permissive environment will enable borrowers to service their debts until tighter monetary policy is adopted to curb the resulting inflation; on the other hand, accelerated lending is associated with a deteriorating quality of borrower, both because there are simply not enough sound borrowers to finance such a rapid expansion and because banks do not have the capacity to evaluate lending at such an increasing rate. Rapid expansion of domestic credit was a feature of the pre-crisis period in Mexico, Indonesia, Thailand, Malaysia, and Korea. In the Indonesian case, the expansion of domestic credit exceeded 20 percent of GDP in the pre-crisis years.
Capital Account Liberalization
Many observers have blamed the opening of the capital account for the twin crises of the 1990s. The simple argument goes that without an open capital account, indebtedness could not have built up. However, there have been many experiences with foreign exchange crises in countries where the capital account was relatively closed. The degree to which cross-border financial flows must be regulated to prevent speculative flows when exchange rates are greatly misaligned is more restrictive than is compatible with a relatively open trading regime.
Moreover, there are many countries with open capital accounts that have not experienced the difficulties that the Asian countries did. Economies such as those of Taiwan Province of China and Singapore, where there were current account surpluses and high levels of foreign exchange reserves relative to trade volumes, did not experience difficulties.
To the extent the opening of the capital account results in difficulties, there are more complex avenues than those associated with real appreciation of the currency. First, when the capital account is open and the nominal exchange rate is fixed without appropriate supportive monetary and fiscal policies as discussed above, there are strong incentives for banks and/or private entities to incur foreign-exchange denominated liabilities (capital inflow) because of lower borrowing costs. When they view the government as having guaranteed the exchange rate, they may not match their future foreign exchange liabilities with foreign exchange assets. Second, banks may not have sufficient incentives for appropriate prudence in their lending policies, due either to a lack of capital adequacy (and existing NPLs) or to an absence of appropriate supervision.
The Korean Economy, the Chaebol, Credit Rationing, and Growth
Korean Economic Growth after 1960
Korea’s spectacular economic growth from the 1960s to the 1990s is well known. In the 1950s, Korea was one of the poorest countries in the world, and was widely regarded as a country without serious growth prospects. After economic policy reforms began in the early 1960s, Korea began growing at sustained rates previously unheard of in world history. 5 Real GDP grew at an average annual rate of 10 percent per annum in the decade starting in 1963. By the mid-1990s, Korea’s real per capita income was nearly nine times what it had been in the early 1960s (Table 1).
Economic liberalization took place throughout the first 35 years of Korea’s rapid growth. In 1960, the country had had the usual developing-country mix of overvalued (and multiple) exchange rates supported by quantitative restrictions on imports (and a black market in foreign exchange), consequent high walls of protection for domestic manufacturers, price controls on many key commodities, credit rationing, a large fiscal deficit, one of the highest rates of inflation in the world, and a huge (averaging around 9 percent of GDP over the period 1953-58) current account deficit financed largely by foreign aid inflows.6
First steps in reform included moving to a more realistic (and constant real) exchange rate for exports, and the relaxation of restrictions on importing by exporters. Thereafter, general imports began to be liberalized in the late 1960s and the exchange regime was unified. Other major reforms also took place, including a major fiscal and tax reform in 1964, gradual removal of price controls, and a shift from a regime discriminating against agriculture to a protective one. Quantitative restrictions on imports were greatly eased in the late 1960s, tariffs were lowered in several steps, and further trade liberalization took place in the 1980s. In the early years of rapid growth, however, the banking system remained tightly controlled. Even after a reform in 1965 (which resulted in a positive real rate of interest for borrowers), credit was rationed and the curb market rate was well above the controlled interest rate.7 Only in the late 1980s did interest rates begin to be deregulated, although the apparent gap between demand and supply of loanable funds was declining over time (see below).
In 1960, Korea’s exports were only about 3 percent of GDP, while imports were about 13 percent. Policy makers therefore began focusing on measures to increase exports. They did so by encouraging all exports uniformly,8 but nonetheless held something that might be regarded as close to an “export theory of value.” Any firm that could export was rewarded in proportion to the foreign exchange receipts from exporting. And many of the firms that were initially successful were chaebol (although they were very small at the time and some Korean analysts today do not regard the Hyundai, Samsung, etc., of the 1960s as chaebol at all). Because they were successful, they grew rapidly. They received new loans as their exports grew and as they expanded into new exporting activities.9 Given the underdeveloped state of the Korean financial markets at that time (and in the absence of measures to strengthen them), access to credit was vital for expansion.
The chaebol were successful exporters and, for the first decade or more of Korean growth, were regarded almost as the “heroes” of Korean development. They were rewarded for export performance, and were highly profitable. Hong (1979) estimates the real rate of return on capital to have been about 35 percent or more in the first decade following the start of reforms. The chaebol were generally encouraged to enter whatever export markets they could, but when the authorities wanted a venture undertaken, the chaebol were asked to do so. They undertook these ventures with the implicit guarantee of the government that credit, tax exemptions, and other support would be available to make the venture profitable.10 But the chaebol were on the whole remarkably profitable and had little difficulty with servicing their (subsidized) debt.
The extent to which the Korean economy changed structure is remarkable (Table 2). Exports and export earnings (the dollar price index of traded goods being stable in the 1960s) grew at over 41 percent annually for the period 1959-69 and continued growing almost that rapidly thereafter. Exports of goods and services as a percentage of GDP rose from 3 percent in 1960 to 14 percent in 1970 to 33 percent in 1980; imports also rose, from 10 percent of GDP in 1960 to 41 percent in 1980. Hence, the Korean economy was becoming much more open.11
At the start of reforms, rationed credit financed a large fraction of new investment, especially in the manufacturing sector. The subsidies implicit in this credit served as a stimulus to industry, and permitted much more rapid expansion than would have been possible had companies had to rely on reinvesting their own profits.12 Exporters were allocated preferential credit based upon their export performance. The real rate of return was so high that all the chaebol would happily have borrowed more had they been able to; most of them, as reported by Hong (1981), borrowed additional funds at the much-higher curb market rates. Thus, lending at controlled interest rates was, at least in the early years, equivalent to an intra-marginal subsidy to the chaebol. However, as the data presented below show, the chaebol did very well even without subsidies. Indeed, given the huge distortions in the economy that prevailed in the late 1950s, it is likely that in the 1960s, at least, almost any reasonably sensible venture into unskilled labor-intensive exportable production had a high real rate of return.
As already mentioned, by 1964 the borrowing rate from the banks was positive in real terms, although below a market-clearing rate. Over the following three decades, the banking system was further liberalized as the real interest rate charged for loans rose, and the gap between the controlled rate and what might have cleared the market diminished (see below). At the same time, the real rate of return on investments naturally fell as the very high initial returns obviously could not be sustained. We trace the decline in real returns and the increase in the real cost of credit in the next section.
When policy reforms began in the early 1960s, the Korean savings rate was very low—and even negative by some estimates. As growth accelerated and per capita incomes rose, domestic savings began to increase rapidly, rising from around 0 percent of GDP in 1960 to 18 percent by 1970 and to 24 percent by 1980 (see Table 1).13 But at least until the late 1970s, profitable investment opportunities greatly exceeded domestic savings. As a result, domestic savings were supplemented by borrowing from abroad, equaling as much as 13 percent of GDP in the late 1960s.14 But, despite the large capital inflows, the debt-service/exports and debt/GDP ratios did not increase because of the rapid rate of growth of export earnings and real GDP.
The Korean government guaranteed these credits, and determined the maximum that could be borrowed, allocating borrowing rights among exporting firms. Since the foreign interest rate was well below the domestic interest rate (especially in the curb market) and the real exchange rate was fairly stable for exporters, there was intense competition for foreign loans.
As domestic savings rose, the proportionate reliance on foreign resources for supplementing domestic savings to finance investment fell. By the 1980s, the domestic savings rate was in excess of 30 percent, and the current account went into surplus for several years in the mid-1980s.15 Beginning at this time, the American government in bilateral trade negotiations began to pressure the Koreans to let the won appreciate in order to reduce the bilateral trade surplus with the U.S.16 Most Korean economists by the mid-1990s believed that it would be in Korea’s best interest to have some real depreciation of the won, but the pressures not to do so prevented it. While the won exchange rate was not fixed, the range within which it fluctuated was relatively narrow: it appreciated from 890 won per dollar at the end of 1985 to 679 won per dollar in 1989, and thereafter gradually depreciated to 808 won per dollar in 1993, appreciating again to 788 won per dollar in 1995. At the end of 1996 it stood at 844 per dollar, and of course depreciated almost 50 percent in 1997.17 But, for the decade prior to the 1997 crisis, there had been little change in the real exchange rate.
Thus, by the mid-1990s, Korea had sustained three and a half decades of rapid growth. While there had been periods of difficulty—both slowdowns and overheating—Korean policy makers had met their challenges successfully. As noted by the OECD, the country had come from being one of the poorest developing countries in 1960 to having a per capita income equal to some OECD countries, with a higher rate of economic growth. 18
The late 1980s had witnessed the introduction of democracy into Korea. The elected governments chose to liberalize further, including especially the financial sector and international capital flows.19 In 1992-93 there was a “growth recession,” as the growth rate slowed to just over 5 percent (contrasted with rates over 9 percent in the preceding two years and an average rate above 8 percent in the preceding decade). One response was to ease monetary policy: domestic credit expanded by over 18 percent in 1994, 14 percent in 1995, and 21 percent in 1996.20 Real GDP growth responded, exceeding 8 percent in 1994 and 1995. But, as will be argued below, underlying weaknesses were not addressed, and the stimulus to the economy, through expansion of domestic credit and other measures, increased the vulnerability of the financial system later on.
Export earnings failed to maintain their growth rate in 1996, increasing only 3 percent in dollar terms, as falling prices for semiconductors and a number of other factors resulted in the slowdown. Then, a number of events took place early in 1997 that surely eroded confidence. One of the large chaebol, Hanbo, went bankrupt early in the year. Given that it had been widely believed that the large chaebol were “too big to fail,” this in and of itself must have resulted in some loss of confidence and a reexamination of Korea’s creditworthiness. Moreover, 1997 was an election year, with the Presidential elections set to be held early in December. That the market anticipated difficulties is reflected in the fact that the Korean stock exchange index fell from 981 in April 1996 to 677 by the end of March 1997 and to 471 at the end of October, even before the outbreak of the currency crisis.
However, while the net and gross foreign (and especially short-term) liabilities of the banking and financial systems were continuing to increase, there was no visible evidence of crisis until the final quarter of the year. The Thai crisis had exploded in June, and the Indonesian crisis had begun during the summer of 1997, but most foreign observers were confident, given Korea’s past history, that Korea would not be affected.21 Korea’s offshore banks were holding paper from Indonesia, Russia, and other countries with dollar liabilities, which would further deteriorate the net foreign asset position, but that was not widely known at the time.
Capital flight began early in the fourth quarter of the year. In many instances, it was simply a refusal to roll over short-term debt. But other factors contributed: Korea’s sovereign risk status was downgraded by Standard and Poor’s in October; reported NPLs in the banking system doubled from the end of 1996 to the fourth quarter of 1998, reaching 7.5 percent of total loans by that time, owing largely to the bankruptcy of six chaebol and the sharp drop in the Korean stock exchange. But, once it became known that reserves were decreasing, others sought to get out of won, and the capital outflow intensified rapidly.22 Total reserves less overseas branch deposits and other unusable foreign exchange were $22.3 billion at the end of October and fell to $7.3 billion by the end of November.23 It is reported that, by the time the IMF was approached, gross reserves were being depleted at a rate so rapid that they would have approached zero within 48 hours. In the program presented to the IMF Board, it was reported that usable reserves had dropped from $22.5 billion on October 31 to $13 billion on November 21, and to $6 billion on December 2.24
The IMF Program25
All three Presidential candidates had declared repeatedly that under no circumstances would they approach the IMF. When the government did approach the IMF, the IMF’s problem was complicated by several things: (1) it was not known who the new president would be, and hence with whom the IMF would have to deal on the economics team; (2) there was very little time to put together a program, and both because Korea had been viewed as “sound” until recently and because the candidates had all said they would not approach the Fund, there had been less preliminary work done than was usually the case;26 (3) the exchange rate was depreciating sharply after the end of October, and when the band was widened to 10 percent on November 19, the rate of depreciation began accelerating rapidly; and (4) as already mentioned, the government was rapidly running out of foreign exchange reserves, and would soon be forced to default on its obligations.27 The high short-term indebtedness meant that foreigners could get out of won simply by refusing to roll over outstanding debt.28
The first (hastily put-together) program set forth as its objectives: “building the conditions for an early return of confidence so as to limit the deceleration of real GDP growth to about 3 percent of GDP in 1998, followed by a recovery towards potential in 1999; containing inflation at or below 5 percent; and building international reserves to more than two months of imports by end-1998.”29 The staff memorandum stated that there were three pillars to the government’s program: the macroeconomic framework;30 restructuring and recapitalizing the financial sector; and reducing reliance of corporations and financial institutions on short-term debt.
For present purposes, the specifics of the Fund program are not relevant. However, understanding those aspects of the program that were important in affecting the severity of the downturn is necessary if an assessment of the role of the various factors leading in the downturn is to be made. In attempting to stem the speculative pressures, the exchange rate was allowed to float, and the won depreciated from the mid-800s level per dollar to almost 1800 per U.S. dollar.31 The liquidity which had been introduced into the financial system in prior weeks (in an effort to support the chaebol) was removed, and money market rates were raised sharply. In the words of the staff these rates would “be maintained at as high a level as needed to stabilize markets” (p.5). Day-to-day monetary policy was to be geared to exchange rate and short-term interest rate movements, while exchange rate policy was to be flexible with intervention “limited to smoothing operations.”
The 1998 budget as passed by the government had projected a surplus of about 0.25 percent of GDP. But Fund staff estimated that lower growth and the altered exchange rate would reduce the balance by 0.8 percent of GDP, and that it would require 5.5 percent of GDP to recapitalize the banks to meet the Basle minimum capital standards. It was assumed that these funds would have to be borrowed, and interest costs (0.8 percent of GDP) were therefore also included in the altered budget estimates. These factors would, on Fund estimates, have shifted the fiscal account into deficit to about 1.5 percent of GDP in 1998. As stated by staff, “In order to prevent such a deficit and alleviate the burden on monetary policy in the overall macroeconomic adjustment, fiscal policy will be tightened to achieve at least balance and, preferably, a small surplus.” The program therefore called for fiscal changes approximately offsetting the negative anticipated changes, and thus for maintenance of the fiscal stance as anticipated prior to the crisis, with the 1.5 percent of GDP cuts equally distributed between government expenditures and revenues. The government initially raised some taxes to yield about 0.5 percent of GDP.
The second leg of the program was financial restructuring. As already indicated, NPLs were large and increasing prior to the crisis. The depreciation of the exchange rate increased debt-servicing obligations for chaebol and financial institutions, as did the increase in interest rates that came about with monetary tightening. An exit policy was to be adopted to close down weak financial institutions, and the remaining banks were to be recapitalized (through merger or other means). A blanket deposit guarantee was to be phased out at the end of December 2000 and replaced with deposit insurance for small depositors only.32
Bank restructuring required a prior, or at least concurrent, restructuring of the chaebols’ finances. Given their very high debt-equity ratios (for one chaebol at the height of the crisis, the debt-equity ratio reached 12:1),33 financial viability where feasible at all would surely require swaps of debt by the chaebol to the banks, giving the banks equity in return. For this reason, it was predictable that it would require time. Data on the finances of the chaebol are given in the next section. The stand-by also addressed corporate governance and corporate-financial-structure issues, focusing on improving incentives and supervision for banking operations and reforming bankruptcy laws. The government also agreed to refrain from providing financial support, providing tax privileges, or forcing mergers for individual companies.
A final issue of concern here is the projected magnitude of the financial support for the Korean program. The current account deficit was expected to decline markedly in 1997 to about 3 percent of GDP, and then—with export growth and won depreciation—to about 0.5 percent of GDP in 1998. However, the very high level of short-term debt was seen to be worrisome. As stated in the standby (p. 12), “It is difficult to estimate with any certainty the likely developments in capital flows…given the uncertainty surrounding the rolling over of private sector short-term debt and the recent collapse in market confidence…The working assumption is that, on the basis of the beneficial effects on market confidence of the announced program and the large financing package, the bulk of the short-term debt will be rolled over. Under this scenario, the purpose of the exceptional financing would be largely to reconstitute reserves. For this outcome to materialize, it is critical that the financing package provided is adequately large and the program is perceived to be strong. It is anticipated that a comprehensive financing package of about $55 billion will be provided on a multilateral and bilateral basis …”
The Severity of the Crisis
For at least two weeks after the announcement of the Fund program, questions remained as to whether the downward slide had been halted.34 By late December, however, the exchange rate had stabilized, and by mid-January, foreign banks announced a $24 billion package of rollovers and new money.35
Domestic economic activity slowed markedly in 1998. For the year as a whole, real GDP fell by 6.7 percent, contrasted with the Fund’s projected 3 percent. The unemployment rate, which had been 2.2 percent at the end of the third quarter of 1997 rose throughout 1998 and peaked in the first quarter of 1999 at 8.4 percent. The seasonally adjusted industrial production index fell by 15 percent from the end of 1997 to the second quarter of 1998. Thereafter, it rose, reaching its pre-crisis level by the end of 1998 and 144.9 at the end of 1999.
The external accounts improved markedly. There was a sharp drop in imports in immediate response to the crisis, and a much-increased current account balance: while exports were slightly lower in dollar terms in 1998 than in 1997, imports fell 22.4 percent and the current account balance was equal to an astonishing 12.5 percent of GDP for the year. Foreign exchange reserves rose in response, reaching $74 billion by the end of 1999 and $83.6 billion by the end of the first quarter of 2000. The decline in real GDP ended in mid-1998, and by the end of the year, real GDP had exceeded its pre-crisis level. For 1999, real GDP growth exceeded 9 percent, and is projected to attain that same rate for 2000.
After early 1998, the nominal exchange rate appreciated in dollar terms, entering the year 2000 at around 1100 to the dollar, contrasted with 1800 to the dollar at the peak of the crisis. Moreover, prices at the end of 1998 were about 7 percent higher than at the end of 1997; in 1999 the rate of inflation was just 0.8 percent, as measured by the consumer price index.
Progress in restructuring the financial sector was necessarily considerably slower. Although interest rates had fallen below their pre-crisis levels by the end of 1998, restructuring of chaebol and financial institutions met considerable resistance.36 Government policy pronouncements and actions have continued to push reforms, but the pace of reform has been much slower than with regard to the balance of payments and external finances.
But by any measure, the negative impact of the crisis and measures to address it were felt most heavily in 1998. By early 2000, the Korean recovery was more rapid and more pronounced than had been anticipated by any.37
Estimating the Role of Financial and Other Variables in Leading to Crisis
Financial restructuring was absolutely essential—first to make the reforms credible (or capital outflows would have continued) and second as a prerequisite for economic recovery. And because the devaluation and higher interest rates would both weaken the financial sector in the short run (and this was understood by the markets), failure to address the issue of financial restructuring would clearly have increased the severity of the recession and delayed, if not aborted, the recovery. And financial restructuring could not be satisfactorily undertaken without addressing the very high debt/equity ratios of the chaebol. How much this intensified the downturn however, cannot be addressed until consideration of the finances of the chaebol and the financial system are considered.
Either a financial crisis or a currency crisis must be addressed with measures that will cause economic pain in the short run. But when the two interact, the resulting costs are much higher. To see how this played out in Korea, we start with an examination of the finances of the chaebol prior to late 1997. An overview of their evolution, and the problems that developed, will be useful before turning to detail. As mentioned earlier, the chaebol had earlier contributed enormously to Korea’s rapid economic growth. By the early 1990s, the largest 30 chaebol accounted for 49 percent of assets and 42 percent of sales in the manufacturing sector. While they had received subsidized credit, this implicit subsidy was probably mostly intramarginal in the 1960s and 1970s, and probably simply increased overall profitability and reinvestment rates. However, over time, the chaebols’ profitability necessarily diminished, while the real interest rate at which they borrowed was increasing.
Table 3 gives data on lending rates of deposit money banks (DMBs) from 1961 to 1987, the period during which interest rates were controlled. In 1987, the proportion of preferential loans in total reached a peak and began to be reduced, and the Bank of Korea stopped reporting the interest rates by those loan categories separately. To estimate how much of a subsidy was involved in DMB lending, it is necessary to contrast the rates with an estimate of what a market-clearing real interest rate might have been.38 To that end, Table 4 gives the curb market interest rates, inflation rates, and growth rates over the years from 1961 to 1987. We then construct an estimate of what a realistic real borrowing rate might have been by adding the inflation rate to the growth rate and calculating a three-year moving average.
We then derive estimates of the subsidy through DMB loans in the first column of Table 5. The estimates are made by multiplying the volume of DMB loans with the difference between the reference interest rate and the actual borrowing rate.39 Also shown in Table 5 are similarly derived estimates of the subsidy through loans to the manufacturing sector from the Korea Development Bank, a nonbank financial institution which lent for investment in public utilities, infrastructure, equipment for manufacturing, and other purchases deemed desirable for developmental purposes. The sum of these estimates should be compared with the final column of Table 5, which gives all manufacturing firms’ ordinary incomes (reported on their balance sheets). As can be seen, the estimated subsidy component of loans exceeded ordinary income in some years, and represented a substantial portion of it in others.
There was almost certainly an element of subsidy in bank lending after 1988 and even in lending at nonpreferential rates prior to that date. Estimating its magnitude is considerably more difficult, as there are no records of the interest rates at which loans were extended. An estimate was made, using the “lending rate” reported by the IMF in International Financial Statistics (Table 3, last column) and taking the difference between the reference rate and that rate times the volume of loans outstanding. The results of those estimates are available upon request to the authors. Unlike the estimates used here, those estimates probably represent upper bounds as to the magnitude of the subsidy implicit in bank loans both because some loans may have been extended at higher interest rates and because the reference rate may overstate the “true” interest rate, especially during periods of falling inflation. Nonetheless, even by our most conservative measure, the subsidy component of lending was large, and constituted an important element of reported profits for those who had access to bank loans, including chaebol.
Figure 1 shows the rates of return on assets and on equity in manufacturing from 1962 to 1997.40 For the 1962-82 period for which we have estimates of the subsidy component of loans, estimates are given as to the rates of return that would have prevailed, all else equal, had there been no subsidy implicit in borrowing. Three things should be noted. First, there were declining rates of return over time. Second, there were earlier periods during which the returns to firms would have been negative had it not been for the subsidized credit. Third, it is small wonder that chaebol and Korean firms in general were highly leveraged: given the incentive to use debt financing entailed in the loans, they were more profitable for doing so, and their founders could retain a stronger controlling interest.
Figure 1.Rates of Return, Manufacturing Sector Total
Figure 2 compares the return on assets of Korean manufacturing sector with those in Germany, Japan, Taiwan Province of China, and the United States. The figure shows that the average return for all Korean manufacturing firms fell from above 8 percent in the 1960s, to about 4 percent in the 1970s under 3 percent in the 1980s, and to 2 percent in the 1990s before the crisis. By contrast, the rates of return in the United States, Germany, and Taiwan Province of China were both higher and more sustained with the exception of the recession years, 1991 and 1992. In Japan, the returns fell but were still about 2.3 percent in 1998—after the impact of the Asian financial crisis. The high rates of return in the 1960s may not have been sustainable for long. They may have been high then, perhaps because many unexploited or under-exploited business opportunities became profitable thanks to policy reforms in the early years of the export drive. And, it may have been inevitable that the rates declined, as better opportunities got exploited first and grew fewer in number. If this had been the only reason, then a decline from 8 percent to 2 percent was too great and the three and half decades have been too long a period for the decline to persist in view of other countries’ experience. Not only in the advanced industrial countries but also in Taiwan Province of China, which may be considered more comparable to Korea, the rates of return fluctuate but do not show any long-term declining trend. The only possible exception is Japan in the 1990s, when it had a significantly lower return than in previous decades.
Figure 2.International Comparison of ROA’s, Manufacturing
Figure 3 shows highly leveraged financial structure of Korean manufacturing firms compared with their counterparts in the same four countries. The average debt-equity ratio of all manufacturing firms sharply rose from around 1.0 in the early 1960s to above 3.0 in the early 1970s. Since then the ratio usually remained between 3.0 and 4.0, once reaching as high as 5.0. In contrast, the same ratio in other countries typically remained under two, with the exception of Japan. Thus, the average debt-equity ratio in Korea has usually been two to three times as large as in other countries. Again, this does not seem to have much to do with Korea being behind in economic development compared with industrial countries. A much lower debt-equity ratio has been the rule in Taiwan Province of China as well as in the advanced industrial countries. The credit policy the government followed in pursuit of development objectives provided incentives to rely on debt financing and had a consequence on the financial structure of Korean firms.
Figure 3.International Comparison of Debt-Equity Ratios, Manufacturing
Chaebol Finances and Falling Profitability41
The next step in the analysis is to consider the chaebol and their profitability in the years leading up to the crisis. For this purpose three groups of firms are identified. Two are groups of chaebol firms: the “Big 5” and the “Big 30,” which respectively refer to those firms belonging to the largest 5 and 30 chaebols. The third group is “non-chaebol,” refering to those firms that do not belong to the Big 30.42
Table 6 shows the average debt-equity ratios of the firms belonging to these three groups in all sectors of the economy and in the manufacturing sector alone, and also the debt-equity ratio of the manufacturing sector as a whole for reference. Clearly, chaebol firms have been much more highly leveraged than non-chaebol firms, whether firms in all sectors of the economy are considered or only those in the manufacturing sector. In the manufacturing sector, the chaebol’s debt-equity ratios were greater than 3.5 until the early 1990s, tended to decline somewhat in the middle of the decade, and to climb back above 4.0 just before the crisis. During the period covered in the table, with few exceptions, the chaebol’s debt-equity ratio was usually 30 to 60 percent higher than non-chaebol’s in the manufacturing sector, and even higher in all sectors. The Big 30 tended to be more highly leveraged than the Big 5, especially when all firms in the economy are considered.
|All Sectors||Manufacturing Sector|
|Big 30||Big 5||Non-chaebol||All firms||Big 30||Big 5||Non-chaebol|
Obviously, highly leveraged firms are vulnerable to shocks, such as increases in the cost of capital, sharp changes in macroeconomic conditions, and sudden drops in foreign demand. The vulnerability of the chaebol was especially dangerous, given their importance to the Korean economy. The situation was even worse as the chaebol firms were closely linked to each other financially. Firms belonging to the same chaebol tended to invest in each other and guarantee the repayment of bank loans for each other. While this may make sense for the individual chaebol, from the economywide viewpoint, there were risks. Chaebol activities that should have been closed down could continue operating, given financial support from their chaebol affiliates. When difficulties were short-run, this support was evidently warranted. But problems arose because there was little way to determine when difficulties were temporary or more long-term, and components of the chaebol remained in business regardless of their own situation, reducing the profitability of the chaebol as a whole. Because of this, the high leverage combined with close inter-firm links must have resulted in declining rates of return for chaebol over time.43
We turn, then, to Table 7 that compares estimates of the returns on assets and equity between chaebol and non-chaebol firms since the mid-1980s. The returns on assets (ROA) were in general falling during the period except for the years of 1994 and 1995, as we saw earlier in Figure 1. (The sudden surge in external demand for semiconductors and cyclical boom in these two years will be discussed below.) Returns on equity (ROE) show the same pattern, with more pronounced fluctuations.
|Return on Assets|
|All Sectors||Manufacturing Sector|
|Big 30||Big 5||Non-chaebol||All firms||Big 30||Big 30*||Big 5||Big 5*||Non-chaebol|
|Returns on Equity|
|All Sectors||Manufacturing Sector|
|Big 30||Big 5||Non-chaebol||All firms||Big 30||Big 30*||Big 5||Big 5*||Non-chaebol|
The rates of return differed much across the three groups of firms. In the mid-1980s the ROA averaged around 4.5 percent for all manufacturing firms, but it averaged 3 percent or less for Big 30, 3-4 percent for Big 5, and above 6 percent or higher for non-chaebol firms. Thus, the non-chaebol firms’ ROAs were nearly three percentage points higher than chaebol firms’, or twice as high. The rates of return declined from the mid-1980s to the early 1990s. The average return for the manufacturing sector fell below 2 percent. While this decline was taking place, the non-chaebol’s ROAs in comparison with chaebol’s remained higher, but in absolute terms the gap in profitability between the two narrowed from around 3 percentage points or more to 2 percentage points or less.
In terms of return on equity, too, the gap in profitability was large in favor of non-chaebol. As shown in Table 7, non-chaebol’s ROE was typically 50 percent or more higher than Big 30, and 30 percent or more higher than Big 5 until the 1990s. Thus, until the early 1990s, the Big 30 on average had higher debt-equity and lower rates of return on assets than the Big 5, and the Big 5 in turn had higher debt-equity and lower rates of return than non-chaebol.
In 1994 and 1995, however, chaebol firms experienced a sudden rise in the returns on assets and on equity. Consequently, the gap in profitability turned in favor of chaebol. One might wonder, taking an agnostic point of view, if the rise in chaebol firms’ returns was a sign of reversal of the long-term decline and beginning of sustained rise in their profitability, which failed to materialize only because of the subsequent crisis. However, it seems quite certain that this was not the case. As Table 7 shows, the sudden rise in returns was most pronounced for the Big 5 chaebol firms in manufacturing. Among them were four major producers and exporters: Samsung Electronics, Hyundai Electronics Industries, LG Semiconductor, and Daewoo Electronics, which benefited greatly from a surge in demand in the world market for semiconductors in 1994 and 1995. When the rates of return were recalculated with these four electronics firms excluded, the returns on assets and equity for chaebol manufacturing firms were more than halved. As shown in Table 7 the return on assets for the Big 5 dropped from 4.8 percent to 2.3 percent for 1994, and from 7.4 percent to 2.4 percent for 1995. The drops in ROA for the Big 30 were similarly drastic. The effect of the change in external conditions on chaebol profitability can also be seen in the returns on equity (Table 7).
The recalculated rates of return for chaebol, although halved, were nevertheless greater than those in the preceding years, still suggesting a possibility of a subsequent upturn in their profitability. However, the economy was in a cyclical upswing in 1994 and 1995.44 Thus, the rise in chaebol profitability in those two years was a temporary phenomenon reflecting a temporary change in external conditions and the cyclical economic boom. It seems quite certain that the trend of declining profitability for chaebol and non-chaebol was continuing into 1997 when the economy was engulfed by the crisis.
At the same time, the reestimated rates of return show that the profitability gap between non-chaebol and chaebol persisted throughout the 1990s except for the four semiconductor producers in 1994 and 1995. This is a fact, but it is not yet clear if the low profitability is intrinsic to chaebol. It might have something to do with industry-specific factors. For example, it is possible that their profitability was low because chaebol firms were concentrated in those industries that had low profitability. To address this, a simple ordinary least squares regression was run with a dummy for chaebol firms and dummies for industries. The equation without a disturbance term and with the subscripts for firms omitted was:
where ROA is return on assets for a firm, expressed in percentage terms; C is a chaebol dummy equal to one, if a given firm belonged to one of the 30 largest chaebol and zero otherwise; and Dj is an industry dummy equal to one if a given firm is in the j-th industry and zero otherwise. There were 22 two-digit industries in the manufacturing sector, with the classification numbers running from 15 to 36. Of these, the regression included 21 industries, omitting the 16th, tobacco, where no private firm operated.
Table 8 shows the estimated coefficients for the chaebol dummy and the probability that the coefficient takes the estimated value, if it were zero, based on t-distribution. (A more detailed report of the regression results is available upon request to the authors.) When the industry difference in rates of return is taken into account, the chaebol’s profitability was still much lower than non-chaebol anywhere between one to five percentage points, usually with very high statistical significance. The statistical significance was lower in the mid-1990s. But it turned highly significant again in 1994 and 1995, when the aforementioned four electronics firms are excluded. Therefore, the low profitability of chaebol firms does not seem due to their concentration in low profitability industries.
Growth of Chaebol Assets, Monetary Policy, and the Financial Conditions of the Banks
Table 9 gives estimates of the assets for all manufacturing firms and the Big 5 chaebol firms. What is striking, given the chaebols’ high debt-equity ratios and low rates of returns, is the fact that the growth of their assets has been incomparably more rapid than that of the manufacturing sector as a whole. As can be seen in column 4, the assets of the Big 5 have been growing in value terms at 20 to 30 percent annually since the mid 1980s, while the combined assets of all firms was growing usually around 20 percent per annum or less (column 3). Thus, the Big 5’s manufacturing assets in 1997 at the time of the financial crisis was 19 times as large as in 1985, whereas the total assets of the sector was 8.5 times as large. As a result, the chaebol’s assets accounted for an increasing proportion of the sector’s total. In 1985, the Big 5 chaebol firms held 16 percent of the assets in the manufacturing sector; the proportion rose to 40 percent in the late 1990s.
|Asset Growth Rates||Big 5’s|
The disproportionate increase in lending to chaebol by the banks and other financial institutions, despite their lower profitability, seems to reflect the banks’ preference for lending to the chaebol in the later period. From the banks’ viewpoint, the chaebol were relatively safer borrowers, as they were likely to have better collateral, and repayments were often guaranteed by other member firms of the same chaebol. Indeed, the government intervened and set a minimum quota in bank lending that should go to small and medium-sized firms so that their access to bank credits might not be unduly restricted.
However, government policy was not repressive toward the chaebol. They had come into being supported by policy favors, especially during the so-called Heavy and Chemical Industry Drive of the 1970s. For, as they grew in assets, sales, employment, exports, etc., and increased their relative importance in the economy, they became indispensable and appeared to be “too big to fail.”
In this regard, an episode of interest rate cuts in the early 1990s provides an interesting case. In January 1993 and again in March 1993, interest rates were cut. The cuts were the policy response to sharply deteriorating economic conditions, especially falling investments (in part in response to the U.S. recession of 1990-91). But it is noteworthy that these cuts coincided with a period of financial difficulty for the chaebols. The return on assets for both the Big 5 and the Big 30 was barely one percent in 1991 (see Table 7), and the annual growth rate of assets that used to be around 30 percent or higher dropped to a mere 6 percent in 1992 (Table 9).
In two steps, the Bank of Korea lowered the rediscount rates under its control by two percentage points “to counter the slowdown of economic growth and contraction of firms’ equipment investment.” In addition to lowering its own discount rate, the Bank of Korea “encouraged” the deposit money banks to lower their loan rates twice, one percentage point at a time. Each time, the their loan and deposit rates were reduced.45
This is significant because the 1993 action was similar to those of earlier years when the ROA had fallen (in 1971 and in 1980-82, see Figure 1). If all manufacturing firms, including the chaebol, had had to pay interest two-percentage points more on all their debts, their income would have dropped almost 3.8 trillion won, wiping out their incomes for that year. The interest rate cuts preceded the cyclical boom of 1994 and 1995, when credit expansion in their aftermath resulted in rapid economic growth.
We conclude that, by 1997, the chaebol were highly vulnerable to negative shocks. Their profitability had been falling and was low. In 1996, the Big 30’s return on assets was already less than 1 percent, and the Big 5’s not much greater than one. Since firms’ ordinary income, the numerator of this rate of return, is an income before tax, there was little margin for a reduction in cash flow or an increase in debt-servicing costs. Yet debt-servicing obligations were mounting, and cash flow does not appear to have been increasing commensurately. The large increase in lending by the commercial banks would appear to have had a significant element of “evergreening” to it. Had the interest rate risen in 1994 or 1995 because of macroeconomic conditions, it seems reasonable to conjecture that NPLs would have increased substantially (or evergreening increased significantly) at that time. The chaebol were over-leveraged and vulnerable to interest rate increases.46
We turn now to the banking side of the picture. Table 10 shows the rates of return for the commercial banks during the 1990s. As can be seen, total assets of the banks rose dramatically during the 1992-97 period, more than tripling. Net income, however, peaked in 1994 and turned negative by 1997. The rate of return on assets was falling continuously during the period, as was the rate of return on equity.
|Interest income, net||3,088.1||3,127.0||3,426.7||4,920.2||6,059.5||7,871.2||6,777.2||9,046.8|
|Interest paid (less)||7,383.2||6,983.0||8,882.0||13,401.6||15,696.3||24,072.8||31,165.9||25,970.7|
|Fees paid (less)||184.1||175.9||237.9||372.8||650.1||8,039.4||11,849.0||5,292.1|
|Other non-interest income||1,139.7||1,453.1||2,407.9||2,353.9||2,569.1||2,696.9||614.7||444.1|
|Operating expenses (less)||3,176.5||3,649.8||4,362.6||6,033.0||6,982.0||8,093.9||7,587.3||6,445.6|
|Of which: personnel expenses||2,221.3||2,595.4||3,187.4||4,228.8||4,964.4||5,609.0||5,596.0||2.885.9|
|Increase in loss provision (less)||942.5||1,023.4||2,371.8||2,319.7||2,342.0||6,192.7||7,780.4||7,487.3|
|Income before income tax||1,217.0||1,322.6||1,598.3||987.0||1,094.1||-3,780.7||-12,458.2||-5,565.8|
|Income Tax (less)||285.5||433.6||550.1||119.2||247.2||139.2||52.4||430.2|
By 1998 the combined net loss of the banks was 46 percent of their equity. The changes up to and including the crisis year reflect three things: (1) the loss provision for NPLs peaked in 1994 and was declining until it rose sharply in 1997 and 1998; (2) provision for valuation loss on securities was steadily increasing; and (3) the part of non-operating income in non-interest income dropped by more than 2.4 trillion won in 1997.47
There was little prior indication of the deterioration in the banks’ assets. Interest had been paid, although it is difficult to estimate how much of this may have reflected “evergreening” lending to enable chaebol to service their debts. The sudden jump in NPLs in 1997 would seem to suggest that evergreening had been taking place in earlier years.48
Not all banks collapsed in 1997, and some had, for all practical purposes, been in difficulty earlier. Table 11 shows the changes in net income in 1993-98 for two of the six largest nationwide commercial banks. As can be seen, Seoul Bank reported virtually zero net income in 1995, and Korea First in 1996 before other banks experienced income losses in 1997. Their plight seems unrelated to the currency crisis in the region or to the sudden and sharp depreciation of won that occurred in the last month in 1997.
There is thus considerable evidence of a weakening of the quality of the banks’ portfolios prior to the crisis, in the sense that the financial health of the borrowers was deteriorating. Nonetheless, the proportion of NPLs in their portfolios was generally stationary or falling until the crisis, although this may in part have reflected the evergreening of accounts. After the crisis, the proportion of NPLs rose sharply, and they were then assumed by the asset management company and the banks booked their losses. The key question is whether those losses were already there and being “evergreened,” or whether the events associated with the exchange rate crisis itself precipitated the financial crisis. Certainly, the chaebol were highly leveraged, and a small change in either their profitability or in interest charges would have been enough to tip them into nonperforming status.
The Foreign Currency Vulnerability of the Banks
Table 12 gives data on foreign-currency denominated assets and liabilities of the commercial banks. As can be seen, foreign-currency denominated assets were slightly below liabilities throughout the 1990s. At their peak in February 1998—post-crisis—commercial banks’ liabilities denominated in foreign currency were 25.1 percent of total liabilities, while assets were 21.8 percent. Interestingly, both the assets and liabilities had risen by about the same percentage during the crisis months, although the gap between them was about two percent wider in early 1998 than it had been in mid-1997.
A question that these data do not answer is the extent to which the quality of the assets and the liabilities were similar. At the time of the crisis, there were reports that many of the loans denominated in foreign currency were to Indonesia, Thailand, and Russia, and that one of the factors precipitating the Korean crisis was the nonperformance of those loans. The data may therefore understate the differential between foreign currency assets and liabilities when risk-adjusted. Even so, it is not evident that the differential was so large that exchange rate changes should have triggered a major decline in the banks’ balance sheets. To the extent there was deterioration caused by the exchange rate change, it would have had to be either in the chaebol’s ability to service their outstanding debts or in the failure of foreign debtors to continue servicing their loans to Korean banks.
The chaebol were in a weak financial condition long before the crisis. While the data do not indicate an increase in NPLs, the rapid increase in assets combined with their deteriorating profitability certainly seems to indicate that the banks were “evergreening” the chaebol’s outstanding debt. If even a quarter of the net increase in chaebol borrowing from the banks was evergreened, the banks were in very bad shape prior to the Korean crisis in 1997.
In an important sense, the vulnerability of the system was extreme. While very favorable conditions—increased semiconductor prices on world markets, falling world interest rates, a pickup in economic activity in the rest of the world—might have prevented the crisis and enabled the chaebol to regain profitability and reduce the degree to which they were leveraged, their behavior during the boom of 1994 and 1995 does not suggest that they were inclined to do so. Instead, in the boom years, they continued borrowing and increasing their assets, while the rate of return remained low with only a slight cyclical upturn.
The conclusion must be that the Korean crisis was a disaster waiting to happen: when very favorable circumstances did not materialize, the needed increase in evergreening was more rapid than the system could tolerate. The foreign exchange crisis itself probably did not trigger the financial crisis: rather, the increase in interest rates did.
The chaebol’s debts to the banks are the chief culprit. And since the chaebol were major exporters, the change in the exchange rate per se probably did not harm their ability to service their debts. However, the increased interest rate clearly did.
In the short run, therefore, more exchange rate depreciation and less interest rate increase—as was in fact the chosen stabilization path—was probably appropriate. Failure to raise the interest rate at all would surely have resulted in larger capital outflows and perpetuated the foreign exchange crisis. Indeed, as was seen, there were doubts over the several weeks after the first IMF program that the package as undertaken was enough. However, further increases in the interest rate (which probably would have reduced the magnitude of exchange rate depreciation) would surely have intensified the financial crisis.
At an analytical level, the impact of the exchange rate depreciation on the banks’ balance sheets either directly or indirectly through the ability of the chaebol to service their debts must be deemed to have been relatively small in the Korean case. The fundamental problem was the magnitude of the leveraging the chaebol had had pre-crisis. That, in turn, made the post-crisis workout of the banking system extremely difficult because of the necessity of restructuring the finances of the chaebol first.
Our conclusion raises more questions than it answers. The corporate sector’s profitability was declining from the early 1960s to the end of 1990s, while no such long-term decline is discernable in advanced industrial countries or in Taiwan Province of China. Clearly, at the heart of the problem were the highly leveraged financial structure of chaebol and their low profitability. This could have been a small problem had the chaebol’s weight in the economy been not so large. As we saw, 40 percent of all assets in the manufacturing sector was in the hands of the five biggest chaebol. This observation raises three related questions: Why did the chaebol keep accumulating assets and expanding business, despite falling profitability? Why did the financial sector, mainly banks, keep financing the chaebol expansion, despite their lower profitability compared with non-chaebol? Why was financial supervision so lax as to allow the system to become so fragile and vulnerable?
As a suggestion for further research, we elaborate the first question. To what extent was the chaebol’s falling profitability the result of inevitable forces as Korea developed? Lucrative business opportunities are likely to be exploited earlier than others and grow fewer in number as time passes. In addition, the capital-labor ratio rises and diminishing returns set in. To what extent is the falling profitability the result of constraints or distortions in the incentive system the chaebol faced? Was there an agency problem, as a head (or a family) of chaebol managed all the firms in the group as if he had a complete ownership while he did not?
The profitability of the non-chaebol was also falling, although it was consistently higher than the chaebol’s before the crisis. How likely is it that this was due to the same, aforementioned inevitable forces? Given that the chaebol had readier access to bank loans and other credits than the non-chaebol, the difficulty in the access must have been comparatively more important for small and medium-sized firms’ falling profitability than were the constraints and other problems that chaebol faced. To what extent is the fall in their profitability related to the rise in chaebol’s weight in the economy in the sense that they had to compete against chaebol for resources, especially for financial resources?
We suggested that the government’s credit policy provided incentives for the chaebol to rely on debt financing as long and as much as possible. Their consistently high debt-equity ratios compared with non-chaebol suggest that they could borrow more per unit of equity than others. Furthermore, their assets grew faster despite their falling profitability. This suggests that the chaebol were better than others at the game of adapting to and taking advantage of the business environment, including government policy. It is not surprising that one that is better in such a game survives and thrives in business as in nature. So, was the provision of cheap credit the sufficient reason for the rise of chaebol and their eventual domination of the economy? Or, were there other non-economic reasons that explain the rise of chaebol? These are institutional questions, the answer to which may have important policy implications.
Similarly, questions need to be asked and further research conducted regarding, on the one hand, the incentive system in the financial sector, governance and agency problems of banks and other financial institutions, their apparent inability to evaluate risks, and the credit culture in general; and, on the other hand, the reasons why prudential regulations over the financial system had been lax. What happened and why needs to be better understood, if lessons are to be learned and the crises are not to be repeated.
Comments on Papers 4 and 5
The first paper by Drs. Chung and Kim analyzes the effects of interest rate policy on the exchange rate. The authors conclude that the high interest rate policy in Korea during the crisis contributed to stabilizing the exchange rate.
There are a number of Korean and foreign studies that have tried to investigate the effects of the high interest rate policy during the Korean crisis, but there have been mixed opinions on whether the policy was effective in stabilizing the exchange rate or not. The most distinctive feature of Chung and Kim’s paper is its methodology, which allows for a non-linear relationship between the interest rate and exchange rate, in contrast to previous studies that have assumed a linear relationship. In this regard, I think the study makes a significant contribution. In particular, exchange rate movements directly affect inflation in Korea, so the study has implications for the central bank, which is responsible for attaining price stability. Moreover, the study will be a great reference for related studies in the future.
I have a number of specific comments on Chung and Kim’s paper:
First, as pointed out by the authors, the study is based on a bivariate model of the interest rate and the exchange rate, and does not consider other variables that can effect the exchange rate, such as capital flows, the corporate bankruptcy rate, the current account balance, the risk premium, and foreign exchange reserves. The results might have been different if the model included other variables. In this regard, the credibility of the results could have been enhanced if the robustness of its estimation results had been demonstrated, for example by alternatively adding one variable at a time, considering the restriction of the number of observations in a non-parametric estimation.
Second, the study does not test for the stationarity of the interest rate differentials. The first difference of the exchange rate is used, which can be assumed to be stationary. However, we cannot assume that the interest rate differential between Korea and the United States is stationary.
Third, the paper analyses only the relationship between interest rate policy and the exchange rate. The paper would be more persuasive if it included an interpretation of the mechanism through which changes in interest rate policy stabilized the exchange rate.
I would like to thank Drs. Chung and Kim for their excellent work. This paper will allow future studies of the effects of the high interest rate policy on the exchange rate during a crisis to proceed with more depth.
The second paper by Drs. Krueger and Yoo analyses various time series data, including the rates of return on assets and on equity, the scale of assets and their growth in the corporate and banking sectors, and corporate debt ratios. The authors conclude that the “chief culprit” in the Korean crisis was the chaebol’s continued heavy build-up of debt despite their worsening profitability. The paper notes that the background to the chaebol’s heavy debt accumulation was the incentives provided by the government’s credit policy, weak governance, poor credit screening skills of financial institutions, and lax prudential regulation of the financial system.
It may well be true that the heavy debts of the chaebol were one of the important causes of the Korean crisis, and that the aforementioned factors contributed to the rapid increase in chaebol indebtedness. However, the rigidity of the exchange rate system, the accumulated current account deficit, the increased ratio of short-term foreign debt to total foreign debt, and contagion from crises in southeast Asian countries are other factors that affected the outbreak of the Korean crisis.
In this regard, the Korean government has pushed ahead with economic reforms since the crisis. Among them are the adoption of a free-floating exchange rate, lowering of large companies’ debt/equity ratios, prohibition of large companies’ new payment guarantees that can be used as collateral, lowering of ceilings on financial institutions’ credits to a single company, introduction of prompt corrective actions to financial institutions by the supervisory authorities, and strengthening of the standards for accounting and public disclosure. Although we cannot expect that long-established behaviors in the corporate and financial sectors will change in one day with the introduction of these reforms, policy makers in Korea have been trying to put these systems on the right track and will continue to do so.
Drs. Krueger and Yoo argue that the Bank of Korea’s easy monetary policy in response to the slowdown of economic growth in the early 1990s also contributed to the large increase in the chaebol’s debts. The easing of the monetary policy stance at that time may well have acted as one of the factors resulting in the expansion of financial support to the chaebol. However, the more accommodating policy stance was adopted not only in response to the economic slowdown but also to defuse financial market instability arising from the launch of the mandatory use of real names in all financial transactions in August 1993 and the step-by-step deregulation of interest rates between November 1993 and December 1994.
Finally, I appreciate the authors’ attempt to present a comprehensive analysis of chaebol’s problems and expect this paper to stimulate further research on the role of the chaebol in the Korean economy.
Bijan B. Aghevli
Before commenting on the two excellent papers by Krueger and Yoo and by Chung and Kim, let me first tell you a fable:
Once upon a time and far away, a litter of little tiger cubs was born. And foreigners came from far and wide to admire the cubs, and comment upon the growth potential in their sturdy limbs and the irresistible allure of their shiny coats. The keepers were proud of their small charges, and set up open zoos with no restrictions and invited everyone to come. Foreign visitors rushed in bearing a formula of milk fortified with vitamins and growth hormones for the cubs, and clamored for the privilege of filling their bowls which, fashioned for a trickle, cracked and leaked under the sudden surge. The cubs grew at an astounding rate, and newspapers and scholarly journals marveled at the miracle of their growth. Before long, the cubs were tigers, and still they grew until they burst out of their flimsy cages and roamed unfettered through the streets, howling for the surfeit of milk and hormones on which they had come to depend. Seeing the tigers at such close range, the visitors began to criticize their less-than-perfect coats, their tendency to purr, their unorthodoxvalues. The food which hadflowed like a river began to dry up. The keepers, in their confidence that the flow was endless, had not repaired the bowls to safeguard the milk they still had, and watched helplessly as the tigers sickened and gnashed their teeth. Then the IMF doctors were called in to wean the tigers off their rich diet and offer alternative lifestyles. As the tigers limped back into their shattered cages, a rowdy crowd of professors gathered to denounce the cruelty of international civil servants in white coats.
More seriously, I would like to discuss the causes of the crisis and, in particular, the role of monetary policy during the adjustment, which is the focus of the two papers. Let me start out by agreeing fully with Krueger and Yoo that “the problem for analysis of the Asian crisis is not the lack of explanations: it is that there are too many.” By now, there is a consensus on the main culprits, but there is still a heated debate on the appropriateness of the remedies that were adopted. The immediate cause of the crisis was the large current account deficit in the Asian countries that was financed largely by short-term debt. Once the crisis hit and financing dried up, exchange rates had no where to go but to collapse. But the large current account deficit was not so much the cause as the effect. Krueger and Yoo have outlined the role of weak financial institutions that led to ever greening of credit to corporations. They underscore that, in the case of Korea, the corporate sector was already in trouble by 1996 as rate of return to assets had declined to a low level. Intermediation of large external funds through an inefficient financial system to an already over-leveraged corporate sector was at the heart of the crisis. Krueger and Yoo, therefore, support the focus of IMF programs on structural reforms and, in particular, on financial sector. It is hardly surprising that I fully endorse their conclusion that: “failure to address the issue of financial restructuring would clearly have increased the severity of the recession and delayed, if not aborted, the recovery.”
Krueger and Yoo note that, as the corporate sector was over-leveraged, higher interest rates had a more damaging impact than exchange rate depreciation. This point is, of course, true under normal circumstances. But I would like to emphasize, as do Chung and Kim, the nonlinear nature of this relationship. While a moderate exchange rate depreciation may not be that damaging to the economy, the situation changes drastically when, as happened in Korea, the won moved from 900 to the U.S. dollar to nearly 2000 in a very short time. In this situation, failure to deal with the exchange rate crisis could quickly degenerate into a vicious cycle of depreciation and inflation. In raising interest rates, an appropriate balance must therefore be struck between preventing the exchange rate from plunging, while not forcing corporations into bankruptcy.
Chung and Kim apply a semi-nonparametric method to derive a nonlinear impulse response function. Applying highly sophisticated econometrics is, of course, a young man’s sport, and I am not in a position to provide critical comments on their paper. By contrast to their highly technical approach, when the IMF team arrived in Seoul in late 1997, we relied mainly on back of the envelope calculations to assist in the formulation of monetary policy. But it is rewarding to be confirmed by Chung and Kim that our basic intuition was correct. That is, a sharp increase in interest rates is effective in stabilizing the exchange rate. I would only make one point on their observation that other critics of the IMF got the opposite results because they did not use a nonlinear method. I suspect that those critics would get the results they wanted using any methodology.
Frankly, it’s very difficult to see how interest rates can be reduced in the middle of a currency crisis. A number of academics have made the point that, in a recession, the orthodox policy would be to lower interest rates and allow the exchange rate to slide to boost economic activity. For example, one commentator has observed that “we cheerfully let the dollar slide from 240 yen to 140, from three Deutsche marks to 1.8; the Fed even helped the process along by cutting interest rates.” But this observation overlooks an important fact: the dollar decline was spread over a period of one and half years, while the drop in the won from less than 1,000 to the dollar to nearly 2,000 took place in only one month. In such an extreme situation, the first priority has to be to stabilize the exchange rate before a vicious inflationary cycle sets in. Once domestic prices begin to skyrocket, the monetary tightening required to re-establish price stability would be extremely costly.
The strategy pursued in these countries was to raise short-term interest rates to arrest the deterioration in the exchange rate, and then gradually reduce it as the exchange rate stabilized. Here, I would like to point out an important fact that has been lost in the debate. Contrary to general perceptions, the initial rise in interest rates in the Asian crisis countries was quite moderate and short lived: in Thailand, short-term rates rose to a peak of 25 percent, in Korea, to 35 percent, and they stayed at these peaks for only a few days before declining rapidly to their pre-crisis level. Furthermore, taking into account the impact of the sharp exchange rate depreciation on inflation expectations, the increase in interest rates was significantly lower in real terms than in nominal terms. Real interest rates (based on the consensus forecast of inflation as a measure of inflation expectations), which were in the range of 7-8 percent before the crisis, rose to short-lived peaks of 20-25 percent before dropping sharply. In both countries, real rates were above 15 percent for only two months, and they are presently about zero. At the same time, both the won and the baht appreciated substantially after the initial crisis, vindicating the approach adopted in these countries.
By contrast, Indonesia’s earlier efforts to stabilize the rupiah turned out to be futile. Indonesia’s experience, however, is the exception that proves the rule. During the first week of the program, the authorities engaged in unsterilized intervention and allowed short-term interest rates to double to 30 percent. As a result, the rupiah appreciated sharply. But within two days, contrary to understandings with the Fund, Bank Indonesia was instructed to cut interest rates back to their initial level. The subsequent liquidity expansion, together with strong signals from the highest levels of the government that commitments under the IMF program would not be fulfilled, led to the subsequent plunge of the rupiah. The resulting high inflation has necessitated much higher interest rates to reestablish financial stability. The adjustment cost would have been drastically smaller had the government persevered with the original program in November 1997.
To be sure, the weakness of the banking and corporate sectors in the Asian countries constrained the scope for raising interest rates. However, while many critics of Fund programs have pointed to the adverse impact of higher interest rates on domestic borrowers, they have neglected to take into account the impact of exchange rate depreciation on holders of external debt. A precipitous drop in the exchange rate raises the corresponding burden of external debt on the banking and corporate sectors to an intolerable level and undermines financial stability. Thus, the trade-off between exchange rate depreciation and interest rate increase shifts drastically in the presence of exchange rate overshooting. The negative impact of the exchange rate depreciation was particularly pronounced for Indonesia, Korea, and the Philippines, which had a relatively higher ratio of external debt to domestic credit.
I should also underscore that the liquidity squeeze in these countries was not just a consequence of high interest rates. Banks’ reluctance to roll over credits also reflected their large nonperforming loans and the weak position of the corporate sector. It is instructive to note that the credit squeeze has not been alleviated even as interest rates in Korea and Thailand have fallen to well below their pre-crisis levels. An even clearer example of this phenomenon is Japan, where even zero short-term interest rates have not restarted the flow of credit.
Finally, I would be remiss if I did not say a few words about Stan Fischer, as he has announced his intention to leave the IMF. The Asian crisis during 1997-98 was an incredibly difficult period for the IMF. As Thailand, the Philippines, Indonesia, and Korea fell in quick succession, IMF staff essentially had to work around the clock. All mission leaders negotiating IMF programs with the Asian countries can tell you many instances of Stan calling them long after midnight to discuss negotiating positions, and then calling them again at the early hours of the next day to be briefed on the results. There is an old adage that no one is indispensable, but Stan comes awfully close. He will be missed by all staff, the current ones as well those who have left.
Bank of TaiwanZhong hua min guo tai wan di gu gong ye cai wu zhuang kuang bao gao (Report on Financial Conditions of Industries, Taiwan Province of China), various issues.
BoughtonJamesM.2000 “From Suez to Tequila: The IMF as Crisis Manager,” Economic Journal pp. 273-291.
Bureau of Census U.S. Department of CommerceQuarterly Financial Report for Manufacturing Mining and Trade Corporationsvarious issues.
ChoDongchul and Kenneth D.West1999“The Effect of Monetary Policy in Exchange Rate Stabilization in Post-Crisis Korea,”KDI Conference onThe Korean Crisis: Before and After (October15).
ChoiSung-No1995 and 1996The Analysis of the 30 Korean Big Business GroupsKorea Economic Research InstituteSeoul.
ChoiSung-No1997 1998 and 1999Big Business Groups in KoreaCenter for Free Enterprise, Seoul.
ClineWilliam R. and SidneyWeintraub editors1981Economic Stabilization in Developing Countries (Washington, D.C.: Brookings Institution).
CooperRichard N.1971“An Assessment of Currency Devaluation in Developing Countries,”in GustavRains ed.Government and Economic Development (New Haven: Yale University Press).
FeldsteinMartin1998 “Refocusing the IMF,” Foreign Affairs (March/April) pp. 20-33.
FrankCharlesR. Jr. Kwang SukKim and LarryWestphal1975Foreign Trade Regimes and Economic Development: South Korea (New York: Columbia University Press).
FurmanJ. and StiglitzJ.E.1998 “Economic Crises: Evidence and Insights from East Asia,” Brookings Papers on Economic Activity(2) pp. 1-114.
HahmJoon-Ho and Frederic S.Mishkin1999“Causes of the Korean Financial Crisis: Lessons for Policy,”The Korean Crisis: Before and After (October; Seoul: Korea Development Institute).
HarbergerArnoldC.1999“The Indonesian Crisis Revisited,”mimeo (December).
HongWontack1981 “Export Promotion and Employment Growth in South Korea,” in Anne O.KruegerHal B.LaryTerryMonson and NarongchaiAkrasanee eds.Trade and Employment in Developing Countries Vol. 1Individual Studies (Chicago: University of Chicago Press) pp. 341-391.
International Monetary Fund1995International Capital Markets: Developments Prospects and Policy IssuesWashington, D.C.
JamesHarold1996International Monetary Cooperation Since Bretton Woods (New York: Oxford University Press).
KruegerAnne O.1979“The Developmental Role of the Foreign Sector and Aid,”Studies in the Modernization of the Republic of Korea: 1948-1975 (Cambridge, MA.: Harvard University Press).
KruegerAnneO.1997 “Lessons for Policy Reform in Light of the Mexican Experience,” in JitendralalBorkakoti and ChrisMilnereds.International Trade and Labour Markets (Houndmills, Basingstoke and London: MacMillan Press) pp. 44-61.
LittleI.M.D.Richard N.CooperW. MaxCorden and SarathRajapatirana1993Boom Crisis and Adjustment: The Macroecononic Experience of Developing Countries (New York: Oxford University Press).
MussaMichael and MiguelSavastano2001 “The IMF Approach to Economic Stabilization,” National Bureau of Economic ResearchMacroeconomics Annual 1999 (Cambridge, Massachusetts: MIT Press) pp. 79-122.
SachsJeffrey1997“The Wrong Medicine for Asia,”New York Times (November3).
We are indebted to Jeong-sam Yang, II-Chong Nam, Inseok Shin, and Duk-Hoon Lee for the generous help they provided in our search for statistics and other data, and to participants in the International Workshop at Stanford University and in the NBER Conference on Currency Crises for helpful comments on the subject matter of this paper.
This section draws heavily on our companion paper, “Chaebol Capitalism and the Currency-Financial Crisis in Korea,” presented at the NBER Conference on Financial Crises held in January 2001 at Cheeca Lodge, Florida.
This would require that the domestic authorities refrain from using monetary and fiscal policies in pursuit of domestic economic objectives and instead allowed inflation or deflation to occur as the “equilibrium” real exchange rate changed. Thus, if from an initial position of balance, the terms of trade deteriorated and warranted a real depreciation of the currency, the domestic price level would have to be allowed to decline to achieve that real depreciation.
Lowering the domestic nominal interest rate would result in more domestic inflation and is thus eschewed by the authorities. See Krueger (1997) for calculation of Mexican real interest rates during the pre-crisis period when a nominal anchor exchange rate policy was followed.
It should be noted that not all exchange rate changes will immediately quell the crisis. In the Mexican case, there was already a significant capital outflow when the authorities announced a nominal devaluation. In the view of most market participants, the magnitude of the announced devaluation was too small and the run on the currency intensified. It was not until the exchange rate was permitted to float that the immediate crisis subsided.
Taiwan Province of China’s rate of economic growth was equally rapid. Prior to the crisis of the late 1990s, most observers would have claimed that the major difference between the Taiwanese and Korean economies was the relatively small scale of Taiwanese enterprises contrasted with the large share of the Korean chaebol in the Korean economy. But there were other differences: perhaps because of greater strategic insecurity, the Taiwanese held very large foreign exchange reserves in relation to the size of their trade or their economy; the Taiwan dollar showed no tendency for real appreciation; and Taiwan Province of China’s current account had been consistently in surplus. The Taiwanese financial system appears to have been considerably sounder than that of Korea in the late 1990s, and the rate of expansion of domestic credit at that time was much lower than that in Korea.
See Hong (1981).
A11 exporters were given an “export subsidy” of a specified number of won per dollar of exports (the number being altered from time to time as conditions were deemed to warrant), an “interest subsidy,” and a tax subsidy, each of a given amount per dollar of export. In addition, exporters were permitted to import goods for their use in generous quantities, which undoubtedly permitted some profits by using the excess for domestic sales. To a significant degree, these “incentives” offset the duties and other charges on imports, and resulted in reasonably uniform incentives for import-competing and exportable production.
Some of these activities were chosen by the chaebol. On occasion, however, the authorities suggested to chaebol owners that they should move into certain lines of production. This attempt to “pick winners” was not always successful; when it reached its height in the heavy and chemical industry (HCI) drive of the mid-1970s, the rate of economic growth and of export expansion slowed substantially and policies were reversed by the late 1970s. When chaebol incurred losses while undertaking these mandated activities, the banks were directed to extend additional credit to the chaebol, thus setting a precedent for later difficulties.
It is important to underscore that these government “rewards” were there in the context of the export drive. When chaebol could not produce competitive exports, there was little support. Even in the HCI drive—the most industry-specific interventionist phase of Korean policy—the output from HCI industries was to be exported within a specified period. When it became clear that that performance test was not being passed, the entire thrust of policy was reevaluated.
Some of the increase in imports was of course intermediate goods used in the production of exportables. But the percentage import content of exports remained fairly stable at around 35 percent of the value of exports over the period of rapid growth. From 1960 onwards, exporters were entitled to import virtually anything that they might use in producing exportables with little paperwork; in addition, they were permitted to import a “wastage” allowance, which they were free to sell on the domestic market. Thus, the de facto liberalization exceeded that which took place because of removal of quantitative restrictions and lowering of tariffs. With an average tariff rate in the tariff schedule of around 15 percent in 1970, average tariff collections as a percent of imports were about 6 percent.
In much of the public discussion of the reliance of firms in crisis countries on borrowing, what seems to be forgotten is that, starting from very low levels of income and development, there is very little equity and a large fraction of investment must therefore be financed through other channels.
In 1960, it is estimated that private saving was a positive 3. 2 percent of GDP while government saving was a negative 2 percent of GDP. Foreign sources financed 78 percent of in vestment, which was 10 percent of GDP (see Krueger, 1979, pp. 206-7). In 1960, most foreign resources were foreign aid.
Most of the capital inflow was from the private sector—largely commercial bank lending—by the late 1960s. Foreign aid had peaked in 1958 and was less than 2 percent of GDP by the mid-1960s. The current account deficit was sustainable because of the profitability of investment and the declining debt-service ratio that resulted from such rapid growth of exports and of real GDP.
Korean policy makers viewed the emergence of the current account surplus as a transitory phenomenon explicable by “three lows”: the fall in oil prices in the mid-1980s, the drop in world interest rates (so that debt-servicing costs declined), and the low dollar (or high yen). The current account turned positive in 1986, rose to a peak of 8.5 percent of GDP in 1988, fell to 2.4 percent of GDP in 1989, turned negative (−0.5 percent) in 1990, and remained negative in the −2 percent range until 1997 when the deficit increased to 4.7 percent of GDP.
Korea was running a bilateral surplus with the U.S. and a bilateral deficit with Japan, and policy makers resisted as far as they could these pressures. One response was to ask the American authorities whether they should devalue with respect to the yen while they appreciated with respect to the U.S. dollar!
Exchange rates, savings rates, and current account deficit data are all taken from various issues of the IMF’s International Financial Statistics unless otherwise noted.
For an account of the Korean economy in the mid-1990s reflecting this consensus view, see OECD (1994).
See OECD (1994) for a description of the five-year financial liberalization plan.
This rate was not markedly faster, however, than it had been over the entire preceding decade. Hahm and Mishkin (1999, p. 21) reject the notion that liberalization of the capital account was responsible for the increase in domestic credit, but note that it did play a role in permitting the banks to take on greater exposures to foreign exchange risk.
However, many Korean economists and policy analysts were very concerned. One of us (Krueger) was at a conference of Korean economic policy makers in August 1997 and the mood was one of deep gloom. Many of the participants were extremely pessimistic about the chaebol, the state of the financial system, and the potential for reforms of economic policy.
However, even in November, the Finance Ministry was issuing reassuring statements, and private forecasters were minimizing the likelihood that Korea would approach the IMF. For a representative account, see John Burton, “Korean Currency Slide Shakes Economy,” Financial Times, November 12, 1997, p. 5.
Other factors also contributed. A financial reform bill, proposed by a blue ribbon committee, had been turned down by Parliament, and it was not clear whether the government had legally guaranteed the foreign exchange liabilities of the financial institutions. While interest rates had risen by about 200 basis points, the Bank of Korea was nonetheless injecting liquidity into the system, which reversed the increase.
The IMF documents cited in this section may be found at http://www.imf.org/external/country/KOR/index.htm.
The fact that the Thai and Indonesian crises had already occurred no doubt diverted some of the attention that Korea otherwise might have received. At that time, too, it must have been anticipated that there would be Malaysian and Philippine programs.
See Boughton (1998).
Hahm and Mishkin (1999, p. 25) point out that “the speculative attack was not in the usual form of direct currency attack to exploit expected depreciation. Due to the tight regulation on currency forwards which should be backed by corresponding current account transactions and the absence of currency futures markets inside Korea at the time, opportunities for direct speculative attack had been much limited. Rather, the drastic depreciation of Korean won was driven by foreign creditors’ run on Korean financial institutions and chaebol to collect their loans, and by foreign investors to exit from the Korean stock market.”
IMF, Korea, “Request for Standby,” December 3, 1997, p. 5.
Much of the controversy surrounding the Korean program centers on whether the program tightened fiscal policy too much. This is discussed below. It should be noted that the Fund staff’s introduction of the macroeconomic program indicated that the program would involve “a tighter monetary stance and significant fiscal adjustment” (Request for Standby, p. 5).
As stated in the Request for Standby (pp. 5-6), “The inflation target reflects a very limited pass-through of the recent depreciation of the won to the aggregate price level. In order to achieve the inflation objective, the government will aim to reduce broad money growth (M3) from an estimated 16.4 percent at end-September to 15.4 percent at end-December 1997, and to a rate consistent with the inflation objective in 1998.”
There were a number of other significant measures, which are less important for present purposes. For example, transparency was to be increased in a variety of ways. Large firms were to be audited by international accounting houses. Supervisory functions were to be reorganized and the Bank of Korea was given much greater independence. Importantly, the government undertook to refrain from attempting to influence lending decisions, leaving those to the financial institutions. But these actions had little impact on the short-run downturn.
hese high debt-equity ratios were public knowledge. The Financial Times published data on debt-equity ratios for 20 chaebol on August 8, 1997. The highest was Sammi with 33.3 times as much debt as equity; Jinro had 85 times as much debt as equity and Halla 20 times; Hyundai’s debt was 4.4 times its equity, and so on. Profits were relatively small as a percentage of assets or sales. In Samsung’s case, for example, net profits were 179.5 billion won on sales of 60 trillion won and total assets of 51 trillion won. Nine of the 20 chaebol listed in the Financial Times, on that day had taken losses.
Because of this, it is very difficult to accept the argument that the Fund program was “too stringent.” Indeed, given those uncertainties it is more plausible to argue that the program might have been even more restrictive initially.
Financial Times, January 30, 1998, p. 11.
See, for example, John Burton, “Boxed into a Corner,” Financial Times, November 23, 1998, p. 17, where the header read “South Korea’s chaebol are fighting a stiff rearguard action against government reforms but the conglomerates are being forced to change their ways.”
This is not to say that corporate and financial restructuring had been completed. At the time of writing in late 2000, unprofitable chaebol activities, including some large entities are still being closed down, with attendant concerns about a slowing down of the rate of growth in 2001.
The curb market rate, given in column 1 of Table 4, provides an alternate “reference interest rate.” As can be seen, the estimated subsidy to borrowers would be considerably higher if the difference between the borrowing rates and the curb market rate were used. The two move together, however, and it seems reasonable that some part of the curb market rate would have been to adjust for additional risk. Our estimates of the implicit subsidy must, however, probably be taken as a lower bound on the value of loans to their recipients.
The subsidy estimation starts from 1963, as the preferential rate began to be applied to “loans for trade” in the year, and it ends in 1982, as the preferential rates applied to various loans ceased to exist (Table 3). In the early 1980s the government lowered interest rates on ordinary loans more than it did the rates on policy-designated preferential loans, narrowing the gap between the two. Data on the volume of DMB loans is available upon request to the authors.
Return on assets (ROA) is the ratio of “ordinary income” and average of assets at the beginning and end of a given year, income and assets reported in various issues of Financial Statement Analysis. Since “ordinary income” was not available for 1963-75, “net income” was used for the period as the numerator, instead. Return on equity (ROE) is similarly estimated except that the denominator is equity.
The debt-equity ratios, rates of return, and asset growth rates reported in this section were estimated on the basis of financial statements of nonfinancial firms subject to the requirement of external audit, compiled by the National Information and Credit Evaluation agency (NICE). This source is used throughout this paper, unless otherwise noted.
The Fair Trade Commission (FTC) of the Korean government each year designates the 30 largest chaebol in terms of assets and lists the firms belonging to them. The list changes over time. The list used in this paper is the same for each year as that which the FTC designates, and therefore changes over time. The Big 5 includes Hyundai, Samsung, Daewoo, LG, and SK.
It should be noted that the practice not only increased vulnerability and lowered the rates of return for the chaebol, but it also doubtless resulted in the banks turning down loan applications from small firms that might have had very high rates of return.
As discussed below, the rediscount rate of the Bank of Korea and loan rates of commercial banks were reduced by two percentage points in 1993. Given chaebol’s high debt-equity ratios, much higher than non-chaebol’s, their profitability would have been still lower, had there been no interest rate cuts.
Bank of Korea, Quarterly Economic Review, March 1993, p. 12 and June 1993, p. 14.
Most of the chaebol sold large proportions of their products overseas. For that reason, they were almost surely less vulnerable to exchange rate changes, as their won sales would have increased significantly in response to a currency depreciation.
This loss reflects the losses banks suffered when they had to sell their NPLs to Korea Asset Management Company (KAMCO), a public enterprise charged with clearing the financial institutions’ balance sheets of their bad loans.
The NPLs of the commercial banks, as reported by Financial Supervisory Commission, were:
|percent of loans||7.00||7.10||7.40||5.80||5.20||4.10||6.00||7.40|