This Selected Issues paper discusses the fiscal policy in the Korean business cycle, and examines the usefulness of the equations for inflation forecasting at horizons consistent with the Bank of Korea 's inflation-targeting framework. It analyzes the causes and macroeconomic consequences of Korea's dual labor market; discusses the government's reform proposals; and the nonbank financial sector restructuring to date.


This Selected Issues paper discusses the fiscal policy in the Korean business cycle, and examines the usefulness of the equations for inflation forecasting at horizons consistent with the Bank of Korea 's inflation-targeting framework. It analyzes the causes and macroeconomic consequences of Korea's dual labor market; discusses the government's reform proposals; and the nonbank financial sector restructuring to date.

IV. Taking Stock of Nonbank Financial Sector Restructuring in Korea1

A. Introduction

1. The crisis in Korea revealed weaknesses across the financial sector. These weaknesses had accumulated over a number of years, and reflected, inter alia, directed lending, poor supervision, and a limited ability to assess and price risk (Chopra, et al., 2002). The financial sector was, in turn, heavily exposed to a corporate sector that was dangerously over-leveraged and more attentive to market share than profitability. Moral hazard problems were prominent owing to a legacy of government intervention and perceptions that many chaebol were “too big to fail.” The result was an unprofitable and over-extended financial system that was highly vulnerable to external shocks and corporate sector distress. In the event, the crisis revealed double-digit nonperforming loans and a large number of nonviable entities across the financial sector.

2. The authorities concentrated most of their initial reform efforts on the banking sector. The initial phase focused mainly on restructuring commercial banks and those nonbank financial entities deemed to be of systemic importance. The results of banking sector restructuring were impressive: nonperforming loans fell sharply, capitalization increased to well above Basel minima, both the number of institutions and employment were reduced significantly, and profitability was restored.

3. The next phase was to tackle problems at the specialized banks and then move to the remainder of the nonblank financial sector. Progress in restructuring nonblank financial institutions is, however, generally less advanced than in the banks. In particular, the Financial Sector Stability Assessment (FSSA) for Korea (IMF, 2003) notes that nonbank deposit taking institutions face soundness problems, the insurance sector remains financially weak, and supervisory oversight in a number of subsectors should continue to be strengthened.

4. The “mini-crisis” sparked by accounting fraud at SK Global in early 2003 underscored that incomplete reform of the nonbank sector can have significant costs. The crisis first engulfed the investment trust companies (ITCs), and then spread to credit card companies, adversely affecting profitability at banks. The initial shock was exacerbated by illiquidity in the bond market and deficiencies in financial market infrastructure, and the negative impact on consumption and investment contributed to the economic slowdown in the first half of 2003. This pattern was not new: corporate difficulties spreading to the ITCs and the wider economy were at the epicenter of the “Daewoo crisis” in mid-1999.

5. This paper takes stock of nonbank financial sector restructuring to date. While the merchant bank sector has been restructured, and progress has been made with merchant banks, mutual savings banks, and credit unions, many parts of the nonbank financial sector remain largely unrestractured. The SK Global crisis showed that key nonbank sectors remain weak, that channels of contagion between nonbank sectors are a source of concern, and that the banking sector itself is not immune to adverse developments in the nonbank sector. While the credit card sector has faced significant difficulties recently, it is not the focus of this paper, as efforts to restructure this sector began only in 2003.

6. The rest of the paper is organized as follows. Section B sets the stage by describing the nonbank financial sector post-crisis. Section C provides details on nonbank restructuring to date using a number of standard metrics. Section D explores the reasons for delaying nonbank reform and the consequences. Section E provides a summary assessment of nonbank reforms and a “to do” list of further actions.

B. The Nonbank Financial Sector Post-Crisis

7. This section provides a condensed picture of the nonbank financial sector once the plan for banking sector restructuring was largely in place—that is, around late 1997.2 A listing of the types of nonbank entities and a functional description of their main activities appears in Box IV.1.

8. Among the nonbank subsectors, the merchant banks were in the worst shape post-crisis. The condition of the merchant banks had deteriorated sharply in late 1997 reflecting the rapid withdrawal of foreign currency deposits by foreigners—only commercial banks and merchant banks were allowed to accept such deposits. Weak confidence in the merchant banks was affected by their exposure to chaebol as well as their losses in the currency, fixed income, and equity markets. The vast majority of the 30 merchant banks were deemed to be nonviable, and were closed or merged. At that point, the merchant bank sector ceased to be of systemic concern.

9. A number of small nonbank financial institutions also proved nonviable. These included many deposit-taking institutions, credit cooperatives, and mutual savings institutions. Many of these entities were closed following the outbreak of the crisis, and the adoption of a unified financial supervisory apparatus in the form of the Financial Supervisory Service and Financial Supervisory Commission helped to set and enforce similar standards across the remaining entities.

10. The life insurance and leasing sectors were both characterized by considerable stress. The life insurance sector comprised 33 institutions and was reportedly the sixth largest in the world measured by premia. It was heavily involved in quasi-banking activities with relatively short term liabilities and a heavy exposure to commercial lending. The authorities sought to rehabilitate the sector by imposing, inter alia, new provisioning and loan classification rules similar to those of commercial banks. The leasing sector was also quite large, reported to be the fourth largest in the world. Many leasing companies were associated with commercial banks, and the bulk were closed.

Nonbank Financial Institutions in Korea: Functional Descriptions

Nonbank depository institutions

Merchant banks offer a variety of financial services excluding payment settlement, stock brokerage, insurance, and household lending. Funding comes mainly through the issuance of their own paper, while the main uses are discounting commercial paper and investment in securities.

Mutual savings and finance companies engage in a narrow range of financial activities, mostly deposit taking and lending. Unlike banks, they have no role in payment settlements and are unable to establish branches. The customer base is the general public and small enterprises.

Credit cooperative institutions comprise: (i) credit unions, which, like mutual saving and finance companies, deal largely in deposit taking and lending—however, lending is restricted to members; (ii) mutual credit facilities, which are similar to credit unions, and include cooperatives for agriculture, fisheries, livestock, forestry, and ginseng; and (iii) community credit cooperatives, which were developed from credit unions and undertake similar financial activities except for the discount of commercial bills.

Postal savings is a state-run system serving agricultural and fishery regions as well as low-income urban households. The main source of funds is deposits and the uses are investments in the Public Fund Management Fund and securities.

Insurance companies

Life insurance companies underwrite insurance policies against death or disease, and to provide for retirement. Funding comes almost solely from insurance premiums, while uses are led by investments in securities and loans.

Nonlife insurance companies underwrite insurance policies against accidents involving fire, automobile collisions, and maritime disasters. Sources and uses of funding are similar to life insurance companies.

Postal insurance is essentially state-run life insurance. The insured are drawn from the low income brackets. Funding comes primarily from premia, while uses are largely deposits in financial institutions.

Securities companies deal, broker, underwrite, and arrange securities subscriptions and sales. Funding is diversified between paid-in capital, deposits, and short-term borrowing. Uses include securities, fixed assets, and cash and deposits.

Collective investment schemes

Investment trust companies (ITCs) purchase securities using funds raised from investors, to whom they distribute the accrued profits. There exist stock-, bond-, and mixed-type funds as well as money market funds. Bond-type funds account for almost one-half of the total, with mixed-type and money market funds at around one-quarter.

Securities investment companies are corporate vehicles (including for restructuring) which cannot establish branches or employ staff. They are obliged to entrust their assets to administrative trustee companies. These companies have the same types of funds as the ITCs.

Trust accounts of banks handle the operation of cash and property entrusted through securities and loans and distribute the accrued profits to customers. Most banks in Korea undertake trust business, 85 percent of which is money in trust.

Other financial institutions

Credit-specialized financial companies perform lending activities including consumer finance, e.g., specialized credit card companies, leasing, and venture financing that are not handled by other financial institutions. These companies do not take deposits.

Remaining institutions: futures companies, money brokerage companies, investment advisory companies.

11. The investment trust industry was seen as the weakest subsector and the one posing the most systemic risk. The three largest ITCs were insolvent and illiquid.3 They retained very large losses on their propriety trading accounts stemming from government-directed equity investments in the late 1980s. Moreover, at the time of the crisis they suffered from maturity mismatches as well as heavy exposure to the chaebol. ITCs generally did not mark to market and the opacity of their accounting served, at least initially, to disguise their losses. The balance sheets of ITCs worsened after the crisis. Assets tripled between January 1998 and June 1999 as funds fled from banks during the restructuring process, rising from 40 percent of M2 pre-crisis to 80 percent. The ITCs used their increased funding to acquire substantial amounts of chaebol debt, which set the stage for the Daewoo crisis in mid-1999.

12. The credit card sector was insignificant at the time of the crisis. The sector took off only in 2000 following the introduction of promotions (lotteries) and tax incentives for credit card use, and a relaxation of restrictions on cash advances. The objectives in promoting credit card usage were to bolster the consumer sector, which had been relatively repressed under Korea’s development model, and to try to bring the “curb” lending market into the formal sector and the tax net.4

C. Progress to Date in Nonbank Restructuring

The structure of the financial system

13. The composition of the Korean financial sector is relatively diverse for an Asian economy, and has remained broadly stable over the past decade. In contrast to the “Asian model,” banks do not dominate. Measured by assets, banking institutions—including specialized banks5—currently comprise just over one-half of the financial system, largely unchanged from 1990 (Table IV.1). The composition of shares across nonbank subsectors has also remained broadly stable. The share of ITCs in total financial system assets has increased gradually to over 10 percent over the past decade, reflecting the rising popularity of money market funds. The share of life insurance assets has remained in the 8–9 percent range, while that of bank trust accounts fell sharply after the crisis due to a reduction in products with guaranteed returns as well as the exclusion of these assets from deposit insurance. There are no relatively large players among nonbank depository institutions; mutual credit facilities have the largest share of assets at 6 percent of the financial sector.

Table IV.1.

Assets of Financial Institutions

(In billions of won and percent, at end-period)

article image
Sources: Financial System in Korea, Bank of Korea (2002); and authorities’ estimates.

14. State ownership remains an important feature of the financial system, although foreign ownership is on the rise in the banking sector. In order to ensure financial stability at the peak of the crisis, significant public funds were used to recapitalize—and in some instances nationalize—systemically important institutions. As a result, the share of government ownership in the commercial banks rose from 17 percent before the crisis to almost 60 percent at end-1998, before falling back to around one-fifth by end-2003 as the reprivatization program advanced. Public funds were used to recapitalize specialized banks as well.6 Mainly as a result of the government’s reprivatization efforts, foreign ownership in the banking sector has increased from below 10 percent at end-1998 to 30 percent at present, with a few banks now majority foreign owned. In contrast, the foreign-owned share of nonbank sector assets remains minimal.

Use of public funds

15. To date, Korea has used public funds equivalent to one-quarter of current year GDP to help restructure the financial system. The W 161 trillion spent by end-September 2003 was split in roughly equal amounts between banks and nonbanks (Table IV.2).7 On a functional basis, recapitalization undertaken by the Korea Deposit Insurance Corporation (KDIC) has accounted for one-third of the total while the purchase of distressed assets undertaken largely by the Korea Asset Management Corporation (KAMCO) accounted for one-quarter. The pace of public fund injections has tapered off in recent years, reflecting the need for large recapitalizations soon after the crisis broke out, as well as the expiration in November 2002 of KAMCO’s mandate to purchase distressed assets.

Table IV.2.

Use of Public Funds for Financial Sector Restructuring

(In trillions of Korean won)

article image
Source: Ministry of Finance and Economy.

16. In contrast to banks, public fund usage in the nonbanks has focused primarily on deposit repayment. Reflecting the closure soon after the onset of the crisis of most merchant banks and, to a lesser extent, mutual savings banks, deposit repayment constitutes the largest share of public fund usage to date in the nonbank sector, at around 40 percent; the comparable figure for banks is zero. As with banks, recapitalization explains a large share of public funds for nonbanks, at over one-third, with insurance companies and securities companies receiving the bulk. Finally, the purchase of distressed assets accounts for one-fifth of the total, with ITCs receiving the largest share of these funds after the Daewoo crisis. Over the past few years, deposit repayments to credit union customers has been the largest use of public funds reflecting the authorities stepped-up efforts to rationalize this subsector.

Reductions in financial sector institutions and employment

17. The number of financial institutions has been reduced by more than one-third since the crisis (Table IV.3). The number of institutions has declined by over 700 since end-1997 to 1,344 at end-2003. More than 600 entities have exited the market and a further 150 have been merged; there have been around 50 new entries. Consolidation has been slightly more pronounced in the banking sector (although banks comprise only 1½ percent of financial institutions in Korea).

Table IV.3.

Number of Financial Institutions in Korea

article image
Sources: Bank of Korea Quarterly Bulletin (March 2003), and authorities’ estimates.

18. Progress in consolidation across the nonbank financial sector has been significant, but uneven. Overall, the number of nonbank financial institutions has fallen by 35 percent since end-1997 compared to a 42 percent reduction in the number of banks.8 As noted above, the number of merchant banks and many small depository institutions fell sharply in the period following the crisis, while, more recently, progress has been made in reducing the number of credit unions. The number of ITCs has remained unchanged since the crisis. Reflecting, in part, the dynamism of financial services, including the growth of derivatives related activities, the number of securities companies and nonlife insurance companies has increased by 25 percent and 14 percent, respectively, since end-1997.

19. Financial sector restructuring has entailed a significant reduction in employment (Table IV.4). Between end-1997 and end-June 2003, total employment in the financial sector fell by one-third to 210 thousand. Although employment at the banks accounted for less than one-half of total employment at the onset of the crisis, reductions have been larger at the banks. Specifically, employment at banks fell by almost 40 percent or 57 thousand over the period, while nonbank employment fell by 28 percent or 48 thousand. Aggregate employment in the nonbank financial sector dropped by 30 percent (more or less across the board) in the year following the outbreak of the crisis, and has fluctuated since. Employment in most subsectors has fallen by 30–45 percent since the crisis, with the exception of the securities subsector, which rose by 22 percent.

Table IV.4

Employment in the Financial Sector

(At end period)

article image
Source: Financial Supervisory Commission, Financial Supervision in Korea 2003.

Includes credit card companies.

Progress in reducing impaired assets

20. Bad loans have been reduced significantly across the financial sector since the crisis, although the ratio for nonbanks remains relatively high. Loans classified as substandard and below (SBLs) in the financial sector as a whole have fallen from 10.4 percent at end-2000 to 4.3 percent at end-September 2003 (Table IV.5).9 The trend decline in SBLs has been attributed to purchases of nonperforming loans by KAMCO (mostly for banks), a general reorientation toward market-based lending, and stronger supervision, which curtailed the flow of new bad loans. A strong recovery from the crisis also served to reduce bad loans. In the banking sector, SBLs fell to a post-crisis low of 2.3 percent at end-2002 before rising one percentage point over the next three quarters. The SBL ratio at the five specialized banks has been slightly below the ratio at the commercial banks. SBLs at the nonbanks were 9.2 percent at end-September 2003, also down substantially from almost 24 percent at end-2000. Unlike for banks, aggregate SBLs at nonbanks continued to fall during the first three quarters of 2003.

Table IV.5.

Impaired Assets of Major Financial Subsectors

(At end period)

article image
article image
Source: Financial Supervisory Service.

The recently introduced category “Agriculture, Fisheries, and Forest Federations” is excluded owing to the lack of a published time series.

21. SBL ratios across the nonbank financial sector vary widely. At the end of the sample period, nonbank SBLs range from under 4 percent for insurance companies to 25 percent for securities firms and ITCs. This high variation also holds within the group of nonbank lending institutions, the largest nonbank subsector with 60 percent of total lending. SBL ratios are in double digits and rising for mutual savings banks and credit unions, while SBLs in credit card companies have risen sharply, but remain below 10 percent (Table IV.6).10 The contribution of credit card companies to total lending of this subsector has been volatile, peaking at almost one-half at end-2002 before falling back to around one third by end-September 2003. Significant progress has been made in recent years in reducing the SBL ratio in leasing companies and installment finance companies, although, again, the SBL ratio in this groups remains several times higher than for banks.

Table IV.6.

Impaired Assets of Nonbank Lending Institutions 1/

(At end period)

article image
article image
Source: Financial Supervisory Service.

Data for credit card servicing companies, installment companies, and technology venture capital companies available only from 2001 onward.

Profitability and net income

22. Profitability returned to the financial sector beginning in 2001, but performance weakened markedly during 2003. Reflecting the benefits of the restructuring process described in previous paragraphs, net income in the financial sector recorded a dramatic turnaround from minus W 9.4 trillion in 2000 to W 10.8 trillion in 2001. At the aggregate level, the turnaround was broadly similar for banks and nonbanks. The uniform improvement reflected in particular stronger capital positions, bad asset disposition, and robust consumer demand. However, there was a sharp drop in profitability in the first three-quarters of 2003 stemming from the rapid run-up in exposure to unsecured household credit along with losses on exposure to SK Global. At end-September, aggregate profits at banks were down two-thirds from the same period in 2002, while aggregate profits at nonbanks fell by almost 90 percent.

Korea: Net Income in the Financial Sector 1/

(In billions of won at end period)

article image
Source: Financial Supervisory Service.

Nonbank data are for the fiscal year (April 30–March 31).

23. Stripping out the substantial losses at credit card companies, net income at nonbank financial institutions has held up reasonably well. After falling by 90 percent in 2002, net income of credit card companies turned sharply negative over the first three quarters of 2003, and was almost 1 percentage point of GDP worse than for the same period in 2002. The rest of the nonbank sector in aggregate actually performed fractionally better through end-September 2003 than in the same period in 2002. While there was some deterioration in net income at the ITCs, there was a sharp turnaround in the performance of securities companies and relatively stable performance of the nonbank depository institutions other than credit card companies.

D. Why Was Nonbank Reform Delayed and What Have Been the Consequences?

24. Comprehensive, up-front restructuring of the financial sector was likely to have been infeasible due to political economy considerations. While such an approach may have been preferable on grounds of economic efficiency, the size of the impaired asset problem was arguably too large for the political process to digest (Cho, 2002). An important element of this constraint was the National Assembly, which held the purse strings to public funds. If the necessary funds for restructuring were not forthcoming, the reform effort would have been compromised.

25. The alternative strategy was to adopt a systemic risk approach. Under this “asymmetric” restructuring approach, the sequencing of financial sector restructuring was dictated by the immediacy of a sector’s threat to financial system integrity (Lee, 2003). Thus, the first wave of restructuring included commercial banks and merchant banks, since weaknesses at these institutions threatened the payments system. These initial efforts were followed quickly by restructuring the specialized banks and mutual savings institutions. The intent was to then turn to the remainder of the nonbanks.

26. As a result of the asymmetric approach, loanable funds shifted from banks the nonbanks, diluting the intended strengthening of the financial system, and slowing corporate restructuring. Many institutions, particularly ITCs, took advantage of the delay in restructuring the nonbank sector (and regulatory forbearance) by aggressively expanding their balance sheets (Cho, op. cit.). A positive effect, however, was to alleviate the credit crunch that followed the crisis, at least to the chaebol.

27. In the end, the asymmetric approach not only delayed, but increased, the cost of nonbank financial sector restructuring. The trade-off for cushioning the post-crisis downtown was additional bad asset creation in the nonbanks. This came about as the loanable funds shifted from the banks were largely channeled to chaebol with weak balance sheets, in particular the Daewoo Group, which collapsed in mid-1999. The asymmetric approach also served to lengthen the period of financial sector restructuring, with negative effects on the level of political support for the process. As a result, some economists have concluded that the costs of delaying restructuring the nonbanks exceeded the short-term benefit of dampening the post-crisis recession (Oh and Rhee, 2002).

E. Assessment and the Remaining Agenda

28. Despite good progress in some areas, restructuring the nonbank financial sector remains unfinished. Accomplishments to date include the closure of most merchant banks and many mutual savings companies soon after the crisis, the pruning of the number of credit unions in 2002–03, and some improvement in reducing high SBL ratios. As with the banks, profitability has returned, at least at the aggregate level. However, progress has suffered in recent years as reform fatigue and the political calendar have slowed the pace of reform. Key subsectors have experienced recurring financial turbulence, double-digit SBL ratios are not uncommon, and key subsectors have yet to shed either institutions or employees.

29. The policy challenge is to complete the remaining nonbank restructuring agenda. The main elements would include the following:

  • restoring the financial soundness of credit card companies, by injecting new capital and strengthening risk management;

  • tackling the longstanding problem of the ITCs through restructuring and reprivatization, which should help promote a broader restructuring of the asset management industry;

  • addressing the longstanding soundness problems in insurance companies and securities houses;

  • improving supervision of nonbanks; and

  • strengthening financial market infrastructure, such as upgrading the quality of credit rating agencies and improving bond market liquidity.


  • Bank of Korea, 2003, Quarterly Bulletin (March).

  • Beaumont, Craig, 2003, “Household Credit in Korea—A Macroprudential Analysis,” Chapter IV in Selected Issues, IMF Country Report No. 03/80.

    • Search Google Scholar
    • Export Citation
  • Cho, Yoon-Je, 2002, “What Have We Learned from the Korean Economic Adjustment Program?,” in Korean Crisis and Recovery, Coe and Kim, eds., International Monetary Fund and Korea Institute for International Economic Policy.

    • Search Google Scholar
    • Export Citation
  • Cho, Yoon-Je and Millard Long, 2000, “Restructuring the Financial System in Korea—Key Issues,” Sogang University Institute of International and Area Studies, Working Paper No. 00-01.

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

    • Search Google Scholar
    • Export Citation
  • Financial Supervisory Commission/Financial Supervisory Service, 2003, Financial Supervision in Korea 2003.

  • Financial Supervisory Service, Weekly Newsletter, various issues, available on

  • International Monetary Fund, 2003, Financial Sector Stability Assessment, IMF Country Report No. 03/81, available on

  • Lee, Jang-Yung, 2002, “Comments on Papers 2 and 3,” in Korean Crisis and Recovery, Coe and Kim, eds., International Monetary Fund and Korea Institute for International Economic Policy.

    • Search Google Scholar
    • Export Citation
  • Lee, Jang-Yung, 2003, “Financial Sector Reform in Korea: A Systemic Risk Approach,” Financial Economics Series No. 2003–02, Korea Institute of Finance.

    • Search Google Scholar
    • Export Citation
  • Lee, Jong Kyu, 2002 “Financial Reform in Korea: A Response to the Economic Crisis in 1997,” Bank of Korea mimeo.

  • Oh, Gyutaeg, and Rhee, Changyong, 2002, “The Role of Corporate Bond Markets in the Korean Financial Restructuring Process,” in Korean Crisis and Recovery, Coe and Kim, eds., International Monetary Fund and Korea Institute for International Economic Policy.

    • Search Google Scholar
    • Export Citation

This paper was prepared by Paul F. Gruenwald, former Resident Representative in Korea.


Much of the material in this section follows Chopra, et al. (op. cit.).


After the crisis, the ITCs were split into investment trust management companies, and investment and securities companies; however, they are still referred to as ITCs.


Beaumont, 2003, discusses the credit card boom and the rise in loan delinquencies, especially Box IV.2.


These specialized banks comprise: the Korea Development Bank, Export-Import Bank of Korea, the Industrial Bank of Korea, the National Agricultural Cooperative Federation, and the National Federation of Fisheries Cooperatives. The specialized banks’ share of financial system assets has remained broadly stable at 15–17 percent.


The government also nationalized the two largest ITCs in 1999, and has recently announced that it will reprivatize these entities in 2004. Moreover, in November 2003, the government reached agreement to sell an 80 percent equity stake Hyundai Investment and Securities and Hyundai Investment Trust Management Company, the third largest ITC. While privately owned, Hyundai has been under the control of the Financial Supervisory Service following the failure of its self-rescue plan.


The amount of public funds allocated by the National Assembly has been virtually exhausted, meaning that any significant future injections will either require an additional allocation by the legislature or be funded by reflows from sale of bad loans, or privatization proceeds. As of end-2003, approximately one-third of injected public funds had been recovered, and the ultimate recovery ratio is officially projected at 56 percent.


The average rate of consolidation for nonbanks across subsectors is not meaningful since credit unions comprise over 80 percent of the number of nonbank institutions.


The Financial Supervisory Service began to apply forward-looking criteria to loan classification in 2000, making earlier figures for nonperforming loans incomparable with subsequent ones.


This figure understates the true extent of impaired assets since it excludes rescheduled credit card loans. The share of overdue receivables and rescheduled loans was 42 percent in November 2003, up from 13 percent at end-2002.