Republic of Korea: Selected Issues

This Selected Issues paper highlights that the authorities in the Republic of Korea recognize the pension policy challenges ahead, and a first wave of reforms has already been launched. Despite the reforms to date, much remains to be done. Without further reforms, the public pension systems in Korea are not financially sustainable. This paper considers options for moving to a funded first-pillar pension system. The main results show that a sustainable, funded pension system can be achieved in Korea with reasonably modest changes to key parameters and extra financing.


This Selected Issues paper highlights that the authorities in the Republic of Korea recognize the pension policy challenges ahead, and a first wave of reforms has already been launched. Despite the reforms to date, much remains to be done. Without further reforms, the public pension systems in Korea are not financially sustainable. This paper considers options for moving to a funded first-pillar pension system. The main results show that a sustainable, funded pension system can be achieved in Korea with reasonably modest changes to key parameters and extra financing.

IV. Household Credit in Korea—A Macroprudential Analysis1

A. Introduction

1. The growth in household indebtedness in Korea has been high in recent years, accompanied by a rapid run-up in some housing prices. The rise in household lending followed a restructuring of the financial sector after the 1997–98 crisis, as banks sought to raise profits and strengthen their balance sheets. These developments have supported domestic demand, contributing to the relative strength of Korean GDP growth in 2001–02. However, Korea, like a number of other countries, has previously experienced boom-bust cycles in property markets, which had often been associated with financial system distress and adverse effects on activity.2 In view of these risks, the Korean authorities are monitoring developments in household credit closely, and a range of policy measures were adopted in 2002.

2. This chapter seeks to assess the potential for weakness in household balance sheets that could undermine the financial sector and macroeconomic prospects. The next section reviews recent developments in household credit and related macroeconomic and housing market trends. Structural changes in the financial sector underpinning these developments and data on credit quality are discussed in Sections C and D. Macroprudential indicators for Korean households are compared with other OECD countries in Section E, and Section F presents a sensitivity analysis to possible interest rate shocks. Section G concludes.

B. Household Credit Developments

3. Household credit rose at an annual rate of 34 percent in the first three quarters of 2002, an acceleration from the strong growth in 2000–01 (Table IV.1).3 This rapid growth lifted the ratio of credit to disposable income for the individual sector from 82 percent at end-1998 to an estimated 120 percent in September 2002.4 The debt ratio is somewhat lower at 110 percent if non-interest bearing liabilities are excluded. Part of this rapid increase was a reversal of the exceptional reduction in credit outstanding to households in 1998 at the time of the crisis, as credit had already reached 97 percent of disposable income in 1997.5

Table IV.1.

Household Credit Developments

article image
Source; Bank of Korea.

Loans from financial institutions and the government. Excludes trade credit and miscellaneous.

Excludes self employed persons and non-incorporated businesses.

Staff estimates for 2001-02—data on disposable income of the individual sector have not been released.

4. High credit growth has been associated with a fall in household saving rates and a recovery in residential construction. After a sharp 14 percentage point increase in 1998, the saving rate returned to its average 1995–97 level of 22½ percent in 1999. Continued strong growth in private consumption expenditure in 2000–02, at a time of modest growth in disposable income, pushed the saving rate down to an estimated 9 percent in 2002.6 Residential construction declined in 1998–2000, but there was a strong rebound in 2001–02, with construction returning to pre-crisis levels. The strength of consumption growth, together with the residential construction rebound, helped cushion Korean growth from the impact of the global slowdown in 2001.


Expenditure and Income Growth

(y/y, in percent)

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A004

5. Housing prices began to rise rapidly from mid-2001, but have slowed in recent months following the adoption of measures in the fall of 2002. Overall house price inflation has been running at 16–18 percent (y/y) since March 2002, led by surging apartment prices in the southern Seoul area of Gangnam, which rose 43.4 percent (y/y) in September 2002. The factors behind the rise in housing prices, and measures taken to cool the market, are discussed in the annex. In the months following the adoption of housing market measures in September and prudential measures in October (Box IV.1), rises in the national house price index slowed, and small price falls were recorded in Seoul, especially in areas like Gangnam. From a longer-term perspective, the level of real housing prices is not historically high (Figure IV. 1).

Figure IV-1.
Figure IV-1.

Housing Price Developments, 1986–2002

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A004

Sources: CEIC database and IMF staff calculations.

6. Most bank loans are to finance house or apartment purchases, as is apparent in recent macroeconomic and housing market developments. In April 2002 the BOK surveyed those banks with the largest amounts of new lending to households, covering about 350 thousand loans from January 2001 to March 2002. Home purchases were the predominant use of bank loans to households, at 56 percent of loans in 2002Q1, up from 30 percent in 2001Q1. Repayment of other credit is the second largest use of bank loans, likely due to the lower interest rate on bank loans, but this has declined from 30 percent of loans in 2001 Q1 to 9 percent in 2002Q1. Some 7½ percent of household loans are used for small business financing, while reported use for consumption seems implausibly low at 2-3 percent—a substantial share of borrowing (17–20 percent) is unallocated—although households may rely on credit card loans to finance consumption. Borrowing for investment purposes, including in equities, is relatively low at 7–8 percent.

C. Structural Change in the Market for Household Credit

7. Household credit expansion has primarily reflected entry by banks into this market after the financial sector restructuring stemming from the 1997–98 crisis.7 While the banking system was largely privately owned, lending activities in most banks were focused on the corporate sector until directed lending was abolished in 1998. A strategy of expanding household lending was adopted by a number of banks as a means to strengthen their balance sheets and profitability. With increased competition, interest margins on household loans relative to the overnight call rate fell from 4.8 percent in 2000 to 3 percent in 2003. Household and housing loans rose from 35 percent of won denominated bank loans in 1997 to 50 percent in 2001, which is similar to the share typical in advanced economies.


Commercial Bank Loans in Won by Sector (in percent)1

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A004

1The Housing and Construction bank was added to commercial banks in 1997, accounting tor the sharp rise in loans fur housing.

Prudential Measures in 2002 Regarding Household Loans

The authorities took a series of actions during 2002:

  • February: the FSC/FSS indicated their intention to develop measures related to provisioning and reporting of loan delinquencies, and to establish a credit information bureau, early warning systems, and a consumer credit rehabilitation system.

  • March: the BOK reduced the availability of credit from the subsidized aggregate credit ceiling facility (2.5 percent interest rate) to banks with increased exposure to household loans. This measure was effective in May, although in practice its impact was mainly as a signaling effect.

  • Effective in May: the FSC/FSS increased minimum provisioning for household loans, and strengthened asset classification criteria by requiring that loans overdue by three months or longer be classified as substandard if they exceeded 60 percent of collateral value. The newly required provisioning was estimated at W 0.7 trillion, relative to existing provisions of W 2 trillion,

  • June: the FSC/FSS required two banks to make comprehensive improvements in household credit risk management, while 14 other banks were required to make specific improvements.

  • September: the FSC/FSS lowered the guideline for the ceiling on loan-to value (LTV) ratios for apartments in Seoul and parts of neighboring Kyonggi Province to 60 percent from 70–80 percent, as part of a package of measures to curb real estate speculation.

Minimum Mandatory Loan Loss Provisioning Ratios

(in percent)

article image

The pace of household lending slowed after May, but picked up strongly in August–September.

A major package of measures was announced in October:

(1) Required provisioning was further increased effective January 2003, with the increase also applied to insurance and finance companies.

(2) Credit evaluation for real estate secured loans was required to be conducted in a similar manner to unsecured loans, taking into account borrower income and ability to pay. Collateral appraisal was strengthened by requiring the use of the lower estimate from two to three appraisers.

(3) BIS risk weighting increased for some real estate secured loans. From a base of 50 percent, the risk weight on new loans is raised to 60 percent if the loan is overdue more that 30 days in one year, or if the borrowers1 debt exceeds 250 percent of income, and to 70 percent if both conditions are met. The higher risk weight will be common on new loans as 60–70 percent of household borrowers are estimated to have debt exceeding 250 percent of income.

(4) LTV guideline lowered to 60 percent throughout the country. Seven banks were found to have average LTV above 70 percent, compared with the overall bank average of 67 percent. These banks were required to adjust their average LTV to less than 70 percent by June 2003.

(5) Greater borrower education and protection. The FSC/FSS will encourage financial institutions to participate more actively in the individual credit rehabilitation program.

This package of measures appears to have been effective, as both house prices and household lending slowed in November—December. The authorities continue to monitor developments closely.

8. Consequently, there have been major shifts in the market shares of banks and other sources of household credit. Growth in bank loans to households exceeded 40 percent in 1999–2002, accounting for 60–80 percent of household credit expansion. As a result, the share of household credit provided by banks rose from 32 percent in 1998 to 52 percent by 2002Q3. A secondary contributor to the expansion in household credit was the boom in credit card use, which benefited from tax and other incentives (Box IV.2). Credit card debt outstanding doubled in 2000 and it rose to 16 percent of total household credit by 2002Q3, from 6 percent at end 1998. Traditionally, savings institutions were major sources of household credit, but credit from savings institutions fell during 1999–2001, and their market share collapsed from 34 percent to 13 percent.


Source, of Household Credit

(in percent)

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A004

9. This rapid change in the structure of financing, along with the aggregate shift in the supply of funds to households, may have increased the risk of loan losses. In a discussion of the causes of the Nordic banking crises in the early 1990s, Pesola (2001) notes that a surge in loan losses is more likely when rapid growth in lending reflects a supply shift, as commonly occurs after the financial sector is liberalized. In this case, banks tend to reduce interest rates and lower minimum credit standards, as happened in Korea in recent years. The post-crisis restructuring also has some aspects of a liberalization given the abolition of directed lending. In addition, shifts in the structure of financing sources imply the formation of many new lending relationships, which could impair credit risk analysis.

D. Developments in Household Credit Quality

10. Delinquency rates on bank loans are relatively low in Korea, although this may partly reflect the atypical structure of household loans. Delinquency rates have remained stable at around 1½ percent on bank loans to households. This is well below the U.S. rate of 2.3 percent on residential mortgages on bank books, and just under 5 percent for securitized mortgages (Federal Reserve, 2002). However, the lack of regular amortization on most bank loans to households may lower delinquency ratios and tend to delay their response to underlying financial distress (Box IV.3).

The Credit Card Boom and Regulatory Action

The credit card industry has grown rapidly in recent years. Credit card usage of W 307 trillion in the first half of 2002 was up 54 percent from the same period in 2001, and up six-fold from 1999. Credit card issuance has risen from 39 million in 1999 to 105 million in mid-2002, implying the average Korea adult has more than three credit cards. About two-thirds of credit card transactions are related to a loan in the form of a cash advance, with the remainder corresponding to purchases of merchandize; this the reverse of the pattern typical of other advanced countries.

Incentives and regulatory changes contributed to the boom in credit card usage:

  • the government removed the ceiling on cash advances in May 1999, which had previously accounted for about half of total turnover.

  • to reduce the number of cash transactions and bring merchants into the VAT net, the government initiated a lottery for credit card users in 2000. One lottery “ticket” is granted for each purchase.

  • also with the aim of increasing the tax net, the authorities have introduced a deduction from taxable income, equal to 20 percent of credit card purchases in excess of 10 percent of gross income. The maximum deduction is W 5 million.

The credit card business has been highly profitable. The credit card industry consists of seven specialized credit card firms plus 19 bank affiliates, but three companies have over half of the market. The return on equity of the specialized card companies was 55 percent in the first half of 2001, and their net profits in 2001 were up 175 percent from 2000.

Delinquency rates on credit cards have risen by 5–6 percentage points during 2002 to 12 percent (Table IV.2). Along with a tightening of access to new credit, the lack of revolving credit contributed this rapid increase, as delinquencies on credit card debt cannot be postponed by paying the monthly minimum. Moreover, any late payments are included, partly accounting for the high level of delinquency—on the U.S. basis where delinquent payments must be at least 30 days overdue, the delinquency rate would be about 8 percent compared with 5 percent in the U.S. on average.

Table IV.2.

Indicators of Household Credit Quality

article image
Sources: FSS and Korean Federation of Banks.

The classification threshold for delinquency was increased from W 50,000 to W 300,000 in July 2002.

Table IV.3.

Estimated Impact of Interest Rate Shocks on Household Indicators in 2004

article image

Household credit is assumed to grow in line with disposable income in 2003-04, so the ratio of credit to income is stable at the estimated level at end 2002 or 113 percent.

Interest bearing financial liabilities, excludes trade credit and miscellaneous.

Based on parameters estimated by Benito et al. (2001)

Based on parameters estimated by Pesola (2001).

Household credit is assumed to grow 20 percent faster than disposable income in 2003, as in 2002, raising the credit to income ratio to 136 percent, but credit grows in line with income thereafter.

A wide range of measures were adopted by the FSC/FSS to address rising delinquency rates:

  • a range of administrative steps were taken in May 2002, including: requirements to verify the identity and incomes of all new customers; a ban on aggressive marketing; a ban on unwarranted debt collection methods; and promotion of the use of debit cards.;

  • credit card issuers were required to cut the portion of cash advances to 50 percent or less in July 2002, to be fully phased in by 2004, and enforcement was tightened in November 2002;

  • capital adequacy requirements were tightened effective April 2003, with minimum capital raised to 6 percent of risk-weighted assets from 4 percent;

  • the criteria for prompt corrective actions (PCA) for credit card companies were tightened in November 2002, such that that they could be banned from issuing new cards if their delinquency rates exceeded 15 percent for over one month, and they were also subjected to the same thresholds on capital adequacy as banks for PCA;

  • minimum provisioning against unused credit limit was set at 1 percent for bank affiliated companies in June 2002, and for specialized companies in January 2003.

There are some preliminary signs of changing behavior in the credit card industry. Domestic banks had 28.1 million outstanding cards at end 2002, down 0.2 million from the previous month, the first decline in three years, as banks addressed delinquent accounts. Simultaneously, their delinquency ratio was reduced by 0.4 percentage points from the peak level of 12.2 percent in November.

11. In contrast, the credit card delinquency rate has risen sharply in 2002, to some 12 percent. Measured by U.S. criteria, the credit card delinquency rate is about 8 percent, compared with the U.S. average of 5 percent (see Box IV.2). There is a possibility that this rise in credit card delinquency could be an early indicator of a broader decline in household credit quality. However, credit cards are more widely used, including by younger and lower income persons with limited savings who do not have access to bank loans, so it is unclear whether the deterioration in the quality of credit card receivables has significant implications for the quality of bank loans.

E. International Comparison of Household Financial Ratios

12. Korean household credit indicators are comparable to other advanced economies (Figure IV.2).8 The most commonly used indicator of household debt burdens is the stock of debt relative to household disposable income.9 The sharp recent rise in the debt burden of Korean households contrasts with the broad stability of debt ratios in Japan and the U.K., and the steady upward trends in the U.S. and Germany. These ratios cluster in the 110–120 percent range in recent years—with Japan being the notable exception—and Korea entered this range in 2002.

Figure IV.2.
Figure IV.2.

Household Financial Ratios in Selected Advanced Economies

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A004

Sources: BOK, Flows of Funds, and OECD.

Evolving Household Lending Practices

Typical household loans have short maturities and do not require amortization, and are based mainly on collateral rather than credit risk evaluation:

  • Household credit is relatively short-term and usually does not require amortization. Bank loans to households for housing and other purposes usually have only a one- to three-year maturity. “Bullet” loans, with the full principal is due at maturity are most common, and the usual practice is to rollover or extend the loan at maturity. Interest rates are mostly floating, with more than one-half of loans linked to the three month CD rate. Long-term mortgages are rare in Korea, whereas they account for about two-thirds of household debt in the U.S., Canada, and Germany.

  • Bank lending practices have relied on collateral. In the first three quarters of 2002, 71 percent of household loans were secured by real estate, and another 8 percent had guarantees.

  • Credit risk analysis has not been standard practice for loans with collateral. For loans against collateral, borrowers often did not need to provide evidence of income. About 60–70 percent of household borrowers are estimated to have debt exceeding 250 percent of income.

  • Loan-to-value ratios are modest. Average loan-to-value ratios were 67 percent across commercial banks in the third quarter of 2002, although the FSS found problems in collateral valuation. A low LTV may be necessary to compensate for the lack of amortization payments, and high leverage of borrowers relative to income, which imply a smaller cushion against a decline in collateral values.

  • Credit information sharing has been weak. For example, the Korea Federation of Savings Banks found only 23 percent of borrowers have debt to only one savings bank, most having liabilities to a number of different savings banks.

Practices are reportedly evolving, partly due to regulatory pressure:

  • The use of loan amortization is beginning. In early 2003, some major banks have begun to require that individuals redeem a percentage of their original debts to apply for debt extensions, e.g., Kookmin Bank requires a 10 percent repayment by borrowers with any recent overdue payments on credit cards or loans, and Woori Bank requires borrowers with poor credit records to redeem amounts exceeding 70 percent of collateral value when extending the maturity of debts—as usually happens at maturity.

  • Interest rate schedules have been adjusted to reflect the size of loans. A number of banks have introduced a higher rate on mortgage loans to borrowers with debt more than two and a half times their annual income. This reflects the introduction of higher risk weights on such loans.

  • Credit information sharing has been strengthened. Starting in September 2002, all financial firms, banks, insurers, and credit cards firms can share information on debtors.

  • Banks have tightened their credit screening systems. A number of banks arc refusing or limiting noncollateralized loans to persons with three or more credit cards, substantial credit card debt, or records of delinquency. This may account for the timing of the rise in credit delinquency, as borrowers are less able to take out new loans to clear their debts.

  • Settlement fees are being reintroduced by some banks. Fees of ½–1 percent of the loan were dropped by banks in 2001 due to competitive pressures.

  • Banks’ business plans for 2003 include slower growth rates in household lending. The plans for 2003 are for expansion of 10–15 percent, compared with the business plans for 2002 that called for growth of about 30 percent.

13. The rise in Korean household debt burdens over the medium term has been comparable to Australia. The household debt burden in Korea is estimated to have risen 34 percentage points in the six years ended 2002. This compares with six year rises of 37 percentage points in Australia (1994–2000) and 41 percentage points in New Zealand (1992–99). Neither of these countries has experienced notable financial sector or macroeconomic problems as a consequence. There is clearly a need to examine other indicators before reaching the judgment that the sharp rise in household debt burdens in Korea is a potential source of macroeconomic or financial sector risk.

14. Other macroprudential indicators for Korean households are consistent with those in advanced economies, and they have been relatively stable over time, aside from the effects of the 1997–98 crisis:

  • The rise in Korean household debt has only modestly reduced household financial net worth. The difference between households’ financial assets and liabilities rose from 102 percent of disposable income in 1997 to 133 percent in 2000, before easing to an estimated 121 percent in 2002.10 This fall may largely be a correction of the rapid rise induced by the crisis. In addition, there has been a significant rise in housing values, so overall net wealth would have been more stable.11 While Korean household net financial wealth is well below the G-7 average of 270 percent in 2001, it is not far below Germany and Australia, which also hold a smaller share of financial assets in total assets.

  • Household liquidity has eased more significantly, but remains comparable to Australia and Germany. The coverage of financial liabilities by financial assets—sometimes interpreted as in indicator of “redemption ability”—is low compared with the U.S., U.K., and Japan, but is similar to Australia and Germany. Again, part of the decline in this ratio is a reversal of the post-crisis increase, and some reduction in liquidity is to be expected when access to credit improves.

  • The interest payment burden on Korean households has been stable even as debt has risen (Figure IV.3). Interest payments as a share of disposable income settled at 10 percent in 1995–97, but rose to 11.6 percent in 1998 due to the sharp rise in interest rates. In 2001–02 the interest ratio is estimated to have returned to 10 percent. Effective interest rates for households have fallen as policy interest rates fell after the crisis, with lending rates falling even further as competition for household lending reduced margins, and as households used bank loans to repay credit from more expensive sources. The interest payment burden is at the top of the range for the OECD countries for which data are available, but it is well below the levels in Sweden and Norway when they experienced banking crises.

Figure IV.3.
Figure IV.3.

Household Interest Burdens

(Interest Payments in Percent of Household Disposable Income)

Citation: IMF Staff Country Reports 2003, 080; 10.5089/9781451822137.002.A004

Source: OECD and BOK.

15. Overall, the international comparison does not give any clear warning signs. Nonetheless, it should be recognized that macroprudential analysis of households has only developed in recent years, and that international experience of deregulated financial markets may not be long enough to gauge what constitutes prudent debt levels (Woolford, 2001).

F. Sensitivity Analysis to Interest Rate Shocks

16. Macroeconomic shocks may undermine the financial position of households, adversely affecting the health of financial institutions. The magnitude of financial sector losses due to shocks to the household sector will depend on the indebtedness of households, not only because this implies greater financial sector exposure to the household sector but also because of likely increases in the sensitivity of household financial soundness to shocks. This section presents some analysis of the potential magnitude of such impacts.

17. The impact of interest rate shocks on the household interest payment burden is the focus of the analysis.12 Stress testing in the context of Korea’s participation in the Financial Sector Assessment Program indicated that the response of interest rates to macroeconomic shocks was the key determinant of their impact on household financial soundness, so the following sensitivity analysis focuses on interest rate shocks. While no single ratio can summarize the household financial position, this analysis uses the interest payment burden of households as the key indicator:

  • As the debt service burden rises: (i) the likelihood of default rises as households have a smaller buffer against unforeseen expenses or income losses; and (ii) the incentives to declare bankruptcy also increase because the immediate consumption gains from bankruptcy are larger, as are the eventual benefits of personal debt restructuring (Davis, 1992).

  • Benito et al. (2001) report that the interest burden (income gearing) is one of the main determinants of mortgage arrears in the U.K., with a 1 percentage point rise in income gearing estimated to increase the share of loans in arrears by 0.46 percentage points in the long run. Similarly, based on a panel estimation on four Nordic countries, Pesola (2001) finds that loan losses increase by 0.6 percentage points in the long run for a 1 percentage point rise in the ratio of household interest payments to GDP.

  • The stability of household delinquency rates in Korea in the face of the rise in indebtedness is consistent with the stability of the interest payment burden.

18. Interest rate shocks are simulated under two scenarios.13 A lasting increase in interest rates of 100–200 basis points over 2003–04 is possible if growth remains strong in Korea, e.g., 5 percent or higher. A 300 basis point hike is relatively high, and could be necessary if there was an exceptionally strong recovery. The sensitivity analysis is over a two-year period ending in 2004, to allow three-quarters of the shock to be passed through into the effective interest rate paid by households. The scenarios for the shocks are:

  • (i) household credit growth slows in 2003–04, such that the credit to income ratio stabilizes at the level estimated for end-2002, of 113 percent;14

  • (ii) household credit growth in 2003 continues at the same rate as in 2002, lifting the credit to income ratio to 136 percent, but credit grows in line with income thereafter.

19. Continued high credit growth would significantly increase the impact of any interest rate hikes. Under scenario (i), a 100–200 basis point rate hike would leave the interest burden at levels similar to the 11½ percent recorded in 1998. A 300 basis point hike would lift the ratio to 13 percent, somewhat outside historical experience. Based on the estimates reported above, this would increase aggregate household loan delinquency by about 1.2 percentage points, almost doubling current levels. Loan losses are estimated to rise by 1 percent of loans, which is substantial but within current provisioning levels. When the interest rate shock is combined with continued rapid credit growth, as in scenario (ii), the impacts are significantly larger, with large interest rate hikes then estimated to raise loan delinquency and loan losses by over 2 percentage points compared with the level in 2002.

20. This sensitivity analysis is preliminary in a number of respects. The monitoring of macroprudential indicators has only begun recently as countries like the U.K. and Norway publish regular financial stability reviews, so research in this area is relatively limited. This analysis relied on parameters estimated for countries with different lending practices than in Korea, including different approaches to personal bankruptcy, which will likely alter the sensitivity of credit quality to shocks. In addition, the second scenario, with continued high credit growth, does not allow for the possibility of an associated boom-bust cycle in house prices, which may significantly increase the likely loan losses. Such an analysis would require more disaggregated data on loan-to-value ratios and loan-to-income ratios.

G. Conclusion

21. While some decline in household credit quality may occur in coming years, this deterioration seems likely to be within manageable limits:

  • the level of household indebtedness is comparable to other advanced economies, and the rate of increase since before the crisis is not outside the experience of other advanced economies, being similar to the increases in Australia and New Zealand;

  • other macroprudential indicators of household liquidity, net worth, and debt service burdens do not indicate a lack of financial soundness;

  • indicators of credit quality have remained stable except for credit card receivables, which account for only one-sixth of household credit;

  • loan-to-value ratios on housing loans are conservative, and as of end-2002 there does not appear to be a significant risk of substantial generalized declines in house prices;

  • from the estimated financial position of households at end-2002, the sensitivity analysis suggests that a relatively large rise in interest rates would be required to substantially raise delinquency rates and loan losses relative to current provisioning levels.

22. The risks of significant loan losses would rise if high household credit growth were to continue, but at this stage, the recently adopted prudential measures appear to have averted this risk. The sensitivity analysis also found that further rapid growth in household indebtedness could significantly increase the impact of interest rate hikes on loan delinquency and losses. Prudential and other measures adopted during 2002, culminating in the package adopted in October, appear to have been effective in slowing credit growth in recent months. If this more sustainable pace of credit growth continues, the macroprudential indicators for households would tend to stabilize.

23. Further progress in household credit risk assessment and longer-term household lending are desirable to increase the robustness of the household credit market. With the aid of regulatory pressure, banks are beginning to strengthen their credit risk management systems. Supervisors will need to monitor loan performance to assess the results of these efforts and to reassess the adequacy of provisioning over time. The current reliance on “bullet” loans (see Box IV.3), without regular amortization, tends to obscure possible debt service problems. Moreover, the short maturities mean that a significant fraction of loans needs to be rolled over in any period. This creates the potential for rollover risk, e.g., if housing values fell, exacerbating housing price declines when this collateral is realized. Amortized loans, with longer durations, would not suffer from these difficulties, and the authorities should continue to encourage the use of such loans, especially for housing finance purposes.


  • Benito, Whitely, and Young, 2001, “Analyzing Corporate and Household Balance Sheets,Bank of England, Financial Stability Review (December).

    • Search Google Scholar
    • Export Citation
  • Chopra, Ajai, Kenneth Kang, Meral Karasulu, Hong Liang, Henry Ma, and Anthony Richards, 2002, “From Crisis to Recovery in Korea: Strategy, Achievements, and Lessons,” in Coe and Kim, eds., Korean Crisis and Recovery, International Monetary Fund and Korean Institute for International Economic Policy.

    • Search Google Scholar
    • Export Citation
  • Davis, E. P., 1992, Debt, Financial Fragility, and Systemic Risk, Clarendon Press, Oxford.

  • Federal Reserve, 2002, “Profits and Balance Sheet Developments at U.S. Commercial Banks in 2001,Federal Reserve Bulletin (June).

  • Hilbers, Lei, and Zacho, 2001, “Real Estate Market Developments and Financial Sector Soundness”, IMF Working Paper WP/01/129 (Washington, D.C.: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Pesola, Jarmo, 2001, “The Role of Macroeconomic Shocks in Banking Crises,Bank of Finland Discussion Papers, No. 6.

  • Sundararajan, Enoch, San José, Hilbers, Krueger, Moretti, and Slack, 2002, “Financial Soundness Indicators: Analytical Aspects and Country Practices,IMF Occasional Paper 212 (Washington, D.C.: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Woolford, Ian, 2001, “Macrofinancial Stability and Macroprudential Analysis,Reserve Bank of New Zealand, Bulletin Vol. 64, no. 3, p. 40.

    • Search Google Scholar
    • Export Citation

ANNEX Developments in the Housing Market

Housing prices have risen strongly since mid-2001, especially in parts of Seoul (see Figure IV.1). After being flat for some years after a crisis-related decline, housing prices began to rise in mid-2001. Overall house price inflation has been running at 16–18 percent (y/y) since March 2002. The most rapid increases were in apartment prices in the southern Seoul area of Gangnam, which rose 43.4 percent (y/y) in September 2002.

Recent developments in house prices reflected both supply and demand factors:

  • The supply of new housing collapsed after the crisis, partly reflecting the bankruptcy of many construction companies. From almost 600 thousand units in 1997, construction fell by almost half in 1998, and remained depressed in 1999–2000. A strong recovery began in 2001 with 530 thousand units built, and 667 thousand units were constructed in 2002.

  • On the demand side, interest rate cuts in 2001 and stock market weakness made housing an attractive investment.15 This investment was facilitated by easy credit conditions as banks competed for household loans and as the BOK cut interest rates, such that the household loan rate fell from percent at end 2000 to 7¼ percent a year later. The rise in housing prices began in mid-2001, after an almost 50 percent collapse in the Korean stock market index during 2000.

  • Trading in contracts (bunyankwan) for the right to purchase apartments prior to their construction was liberalized in 2000 to stimulate the real estate market.

  • From a longer-term perspective, the level of real housing prices was relatively low in 2001. The Korea Research Institute for Human Settlements found that house affordability had increased, with average housing costs equal to five years of earnings, compared with nine to ten years earnings in 1991.

The housing market in Korea includes some unique features that make the market more prone to speculation:

  • Rather than making regular rent payments, renters under the jonse system pay a deposit equal to about 60 percent of the property’s value up front, which is returned in full at the end of the lease. This is an interest-free loan to the property owner. The jonse system possibly developed because of the limited availability of mortgage financing in the past. With the fall in interest rates in recent years, a more standard system of regular rental payments is becoming more prevalent.

  • When purchasing a new dwelling, the future owners buy a right—a bunyankwan—to occupy the property when completed. In effect the future owners of the property help finance construction costs. The holder of the bunyankwan makes four or five payments to the developer over the construction period which typically lasts two years.

  • Korea experienced an asset price bubble in late 1980s and early 1990s, but this affected all types of real estate nationwide. Market observers suggest speculators hoarded bunyankwan to drive up prices. A speculator can also obtain higher leverage by using jonse funds to increase the number of apartments they own.

The authorities took various steps to cool off the housing market in 2002:

  • In early 2002, capital gains on bunyankwan were made taxable, and two installments were required to be paid on bunyankwan before it could be traded. Tax audits on speculators were imposed.

A wide range of steps were taken in April, including:

  • the National Tax Service increased the standard value—used for tax purposes—of 5 million apartment units by an average of 10 percent, with increases of 50 percent in parts of Seoul. This advanced the usual increase by three months, aiming to dampen real estate speculation by making buyers pay higher acquisition and registration taxes;

  • the major construction companies agreed to refrain from price increases and the tax office launched a probe of construction companies; and

  • the Housing Finance Credit Guarantee Fund replaced 100 percent guarantees on housing loans with 90 percent guarantees. This step was designed to encourage banks to be more cautious in extending housing loans and reduce losses on guarantees. It affected only a fraction of the market, as outstanding guarantees were W 16 trillion in 2001.

A second round of targeted measures was introduced in September:

  • limits on resale of bunyankwan contracts were introduced in designated speculative regions, primarily Seoul and Kyonggi province. Resale of contracts was prohibited until one year after signing the contract, or after paying a fee, unless the owner needed to move, and apartment subscribers were prohibited from signing for additional apartments for five years;

  • capital gains tax exemptions were made more restrictive in Seoul and Kyonggi province, as owners must reside in the house at least one year and possess it for at least three years to qualify for an exemption; and

  • market prices are to be used for capital gains taxation if the value of an apartment exceeds W 600 million.

These measures appear to have been effective in recent months. In October 2002–January 2003, the national house price index slowed more than would have been expected due to seasonal factors, and small price falls were recorded in Seoul, especially in areas like Gangnam.


This paper was prepared by Craig Beaumont (APD), incorporating contributions from Paul F. Gruenwald and Sung Eup Yoon (Resident Representative Office in Korea).


Hilbers, Lei, and Zacho (2001) analyze a number of country experiences.


Household credit is broadly measured, including credit from banks, savings institutions, insurance companies, credit card companies, finance companies, merchandise companies, and the National Housing Fund.


The individual sector covers the self employed and non-incorporated businesses in addition to households. Income data are available for the individual sector but not for households separately, so all ratios to disposable income are based on the financial liabilities of the individual sector from the BOK Flow of Funds survey.


Credit has also partly been used to replace informal private loans not captured in the official data, although this market was reportedly more important for SMEs than for households.


Much of this fall in household savings may be lasting, as increased access to credit has reduced the need for households to accumulate wealth before purchasing or renting housing.


The restructuring of the banking system is discussed in Chopra et al. (2002).


Data for Korean household financial indicators are staff estimates based on the individual sector of the Flow of Funds statistics produced by the Bank of Korea. The individual sector is used as a proxy for the household sector for these financial ratios. This Flow of Funds data also covers the self employed and non-incorporated businesses so differences in the level or trends of the financial ratios for these groups relative to households may distort the cross-country comparison. For the purposes of the international comparison, loans from the financial sector and government are included, but trade credit and miscellaneous financial liabilities are excluded. Nonfinancial assets are estimated on the assumption that financial assets are 41 percent of total assets, as was found in the 1997 National Wealth Survey—which is made only every 10 years—similar to the share in Australia (45 percent) and Germany (40 percent).


Macroprudential analysis of the household sector is discussed in Sundararajan et al, (2002).


Unlike the U.S. and U.K., equity valuation gains have not been an important contributor to rising household net worth in Korea.


The aggregate market value of Seoul apartments was W 258 trillion in August 2002, a 55 percent rise from end-2000, according real estate analysts. This increase is equivalent to an increase in household net worth to disposable income of roughly 19 percentage points.


The simulations do not allow for a fall in house prices. The recent acceleration of house prices has been relatively brief, and while prices in some areas are historically high, house prices at the national level appear to be quite low in real terms (see Figure IV.1). Moreover, banks have generally modest loan-to-value ratios (see Box IV.3), so that even a substantial fall in house prices would leave collateral in excess of debt.


It was not feasible to calibrate these shocks based on historical volatility given the extreme interest rate adjustments in the Asian crisis and interest rate controls earlier in the 1990s.


The effective interest rate is estimated to have fallen 150 basis points from 2001, with the fall in bank lending rates of some 100 basis points, and further reductions due to shifting to cheaper sources of credit. There may be continued savings from loan restructuring in future, but their magnitude is unclear, and no further reduction is projected.


Research by the KDI finds that changes in apartment prices generally are reflected in developments in household loans, and that apartment prices are significantly influenced by overall interest rates rather than the amount of household loans.

Republic of Korea: Selected Issues
Author: International Monetary Fund